University of Southern California Dissertations and Theses

Reevaluation of the pregnenolone biosynthesis pathway in the human brain

(USC Thesis Other)

Reevaluation of the pregnenolone biosynthesis pathway in the human brain

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2022 Yiqi Christina Lin
REEV ALUATION OF THE PREGNENOLONE BIOSYNTHESIS
PA THW A Y IN THE HUMAN BRAIN
by
Yiqi Christina Lin

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MOLECULAR PHARMACOLOGY AND TOXICOLOGY)
August 2022

ii
Acknowledgements
First and foremost, I would like to express my deep and sincere gratitude to my
supervisor, Dr. Vassilios Papadopoulos, for his continuous support of my PhD research. Not only
did he provide extremely insightful guidance and advice, but he also gave me immense freedom
to explore and build my project to the way it is today. His patience, knowledge, and enthusiasm
has helped me through every step of the process, from hypothesis construction to manuscript
preparation. I could not have asked for a better advisor and mentor for my PhD studies.
I would like to thank the rest of my thesis committee Dr. Martine Culty and Dr. Daryl
Davies for their encouragement and thoughtful comments. I would also like to thank Dr. Enrique
Cadenas, Dr. Roger Duncan, and Dr. Stan Louie for their challenging questions and insightful
suggestions during my qualifying exam.
My sincere thanks to Christian Rabôt, Ethan Canfield, Dr. Alireza Abdolvahabi, Dr. Isaac
Asante, Edward Daly, and Lorne Taylor, for their help in developing the protocol for and running
the mass spectrometry experiments. Special thanks to current and former members of the
Papadopoulos Lab—Dr. Yuchang Li, Dr. Lu Li, Chantal Sottas, Dr. Edith Porter and Dr. Simone
Romano—who have all taught me many techniques and given me a lot of troubleshooting advice
for my experiments. I also would like to thank my friends Dr. Amy Tran-Guzman and Dr.
Melanie Galano for the stimulating science discussions, emotional support and all the fun we had
inside and outside of lab.
Lastly, I would like to thank my parents and my fiancé Garett Cheung for their
unconditional support throughout this journey. They have comforted and encouraged me through
all the stressful times and helped me make it through this PhD feeling happy and accomplished.

iii
Table of Contents
Acknowledgements ....................................................................................................................... ii
List of Tables .................................................................................................................................. v
List of Figures ............................................................................................................................... vi
Abstract ...................................................................................................................................... viii
I. Introduction and Literature Review ...................................................................................... 1
1.0 Abstract ............................................................................................................................... 1
1.1 Neurosteroids: the significance of pregnenolone .............................................................. 2
1.1.1 Function of pregnenolone in the CNS ............................................................. 3
1.1.2 Diseases and conditions with altered pregnenolone levels in the CNS ........... 5
1.1.3 Therapeutic potential of pregnenolone in CNS disorders ............................... 8
1.1.4 Drugs that can alter pregnenolone levels in the CNS .................................... 10
1.1.5 Summary and Discussion .............................................................................. 12
1.2 Steroidogenesis: overview of the CYP11A1 enzyme ........................................................ 14
1.2.1 Biochemistry .................................................................................................. 16
1.2.2 Molecular Biology ......................................................................................... 17
1.2.3 Phylogeny ...................................................................................................... 21
1.3 The discrepancies surrounding CYP11A1 in the brain .................................................. 23
1.3.1 Cyp11a1 in the rat brain ................................................................................ 23
1.3.2 Cyp11a1 in the mouse brain .......................................................................... 30
1.3.3 CYP11A1 in the human brain ........................................................................ 31
1.3.4 CYP11A1 in different brain cell types .......................................................... 34
1.3.5 Transcription regulation ................................................................................. 40
1.3.6 Discussion ...................................................................................................... 43
1.4 Summary ........................................................................................................................... 46
II. Evidence for a CYP11A1-Independent Pathway for Pregnenolone Synthesis in the
Human Brain ................................................................................................................................ 47
2.0 Abstract .............................................................................................................................. 47
2.1 Introduction ....................................................................................................................... 48
2.2 Materials & Methods ......................................................................................................... 52

iv
2.3 Results ................................................................................................................................ 61
2.3.1 Evaluating the steroidogenic potential of glial cells ........................................ 61
2.3.2 Evaluating CYP11A1 mRNA expression in the human brain .......................... 65
2.3.3 Human glial cells synthesize pregnenolone that can be stimulated by
TSPO ligand XBD-173 and hydroxycholesterols .................................................... 75
2.3.4 Pregnenolone synthesis by glial cells is not inhibited by CYP11A1
inhibitors ................................................................................................................... 80
2.4 Discussion .......................................................................................................................... 88
III. Investigation of the alternative pathway for pregnenolone synthesis ............................... 93
3.0 Abstract .............................................................................................................................. 93
3.1 Introduction ....................................................................................................................... 94
3.2 Materials and Methods .................................................................................................... 100
3.3 Results .............................................................................................................................. 104
3.3.1 Pregnenolone synthesis in glial cells is not dependent on reactive
oxygen species ........................................................................................................ 104
3.3.2 Overexpression of CYP11A1b did not alter pregnenolone secretion
unlike overexpression of CYP11A1a ..................................................................... 111
3.3.3 The alternative pathway for pregnenolone synthesis likely involves
another cytochrome P450 ....................................................................................... 117
3.4 Discussion ........................................................................................................................ 123
IV . Conclusion ............................................................................................................................ 126
4.1 Summary of overall findings ............................................................................................ 126
4.2 Limitations ....................................................................................................................... 129
4.3 Future directions .............................................................................................................. 131
References ................................................................................................................................... 134

v
List of Tables
Table 1: Rat brain regions with Cyp11a1 expression .................................................................... 29
Table 2: Evidence for CYP11A1 in different brain cell types ....................................................... 39
Table 3: Sequences of primers used for qRT-PCR ........................................................................ 55
Table 4: List of cell lines used ....................................................................................................... 61
Table 5: Effects of AMG and KC on steroid synthesis .................................................................. 88
Table 6: Sequences of primers used to measure CYP11A1 variants ........................................... 102

vi
List of Figures
Figure 1: Functions and pharmacological regulations of pregnenolone in the CNS. .................... 13
Figure 2: Schematic diagram of the neurosteroid synthesis pathway. ........................................... 15
Figure 3: Graphic representation of the CYP11A1 gene. .............................................................. 20
Figure 4: Graphic summary for CYP11A1 expression in the CNS. .............................................. 33
Figure 5: mRNA expression of steroidogenesis machinery in H295R-S1 and human glial
cells. ............................................................................................................................................... 64
Figure 6: mRNA expression of CYP11A1 in human brain tissues. .............................................. 68
Figure 7: RNAscope in situ hybridization analyses of human brain tissue. .................................. 71
Figure 8: RNA and protein expression of CYP11A1 and its co-factors in glial cells. .................. 72
Figure 9: Western blot analyses of CYP11A1 in human glial cells. .............................................. 73
Figure 10: Confocal images of human cells for CYP11A1 expression. ........................................ 74
Figure 11: Mass spectrometry detection and quantification of pregnenolone production by
cells. ............................................................................................................................................... 77
Figure 12: Pregnenolone secretion by MGM-1 cells over time. ................................................... 78
Figure 13: Human glial cells produce pregnenolone and the production can be altered by
TSPO ligands. ................................................................................................................................ 79
Figure 14: Effect of aminoglutethimide (AMG) on basal and hydroxycholesterol-stimulated
steroid production in peripheral cells and glial cells. .................................................................... 82
Figure 15: Steroid measurements of H295R-S1 and MGM-1 supernatants by mass
spectrometry. .................................................................................................................................. 83
Figure 16: Effect of ketoconazole (KC) on basal and hydroxycholesterol-stimulated steroid
production in peripheral and glial cells. ........................................................................................ 86

vii
Figure 17: Expression of CYP11A1, FDXR, and POR in MGM-1 cells after treatments that
increase pregnenolone synthesis. ................................................................................................... 87
Figure 18: Effect of antioxidant Trolox on pregnenolone secretion and intracellular ROS in
MGM-1 cells. ............................................................................................................................... 105
Figure 19: Effect of antioxidant Trolox on pregnenolone secretion and intracellular ROS in
MGM-3 cells. ............................................................................................................................... 107
Figure 20: Effect of oxidants on pregnenolone secretion and intracellular ROS in MGM-1
cells. ............................................................................................................................................. 109
Figure 21: Effect of oxidants on pregnenolone secretion and intracellular ROS in MGM-3
cells. ............................................................................................................................................. 110
Figure 22: RNA Expression of CYP11A1 variants in the human CNS and peripheral tissues
and cells. ...................................................................................................................................... 112
Figure 23: Localization of CYP11A1 isoforms. .......................................................................... 114
Figure 24: CYP11A1a but not CYP11A1b overexpression leads to increased pregnenolone
production that can be inhibited by AMG. .................................................................................. 116
Figure 25: Effect of iron chelation on pregnenolone secretion in MGM-1 cells. ........................ 118
Figure 26: POR expression in brain cells and effect of POR inhibition on pregnenolone
synthesis. ...................................................................................................................................... 120
Figure 27: Effects of POR and FDXR knockdown on pregnenolone synthesis in MGM-1
cells. ............................................................................................................................................. 122

viii
Abstract
The central and peripheral nervous systems synthesize steroid hormones, called
neurosteroids, which have physiological and modulatory functions in the nervous system. One
such neurosteroid is pregnenolone, the precursor to all other neurosteroids. In classical
steroidogenesis, pregnenolone is synthesized from cholesterol by the CYP11A1 enzyme.
However, there has not been sufficient evidence to conclude that the steroidogenesis pathway is
identical between the central nervous system (CNS) and the periphery, given that pregnenolone
is one of the most abundant neurosteroids in the brain but CYP11A1 is difficult to detect in the
CNS. Therefore, this study aims to provide experimental evidence for whether CYP11A1 is used
by the human brain to produce pregnenolone.
We found that the expression of CYP11A1 in the human brain tissues and cells is more
than 1000-times lower than that in the adrenals. We did not observe detectable CYP11A1 protein
by immunoblotting in four human glial cell lines. However, the glial cells produced detectable
levels of pregnenolone that was not inhibited by the CYP11A1 inhibitor DL-aminoglutethimide
or the non-specific CYP450 inhibitor ketoconazole. Pregnenolone synthesis was increased with
addition of hydroxycholesterols, suggesting the involvement of a desmolase activity. These
results suggest that pregnenolone synthesis in the human brain does not require CYP11A1
activity.
Thus, we explored three possibilities for the CYP11A1-independent alternative pathway:
1) pregnenolone synthesis is dependent on reactive oxygen species; 2) pregnenolone is
synthesized by the second CYP11A1 isoform; 3) pregnenolone is produced by another CYP450
enzyme. Our data did not support the first and second proposed mechanisms, as treatment with
oxidants and antioxidants as well as overexpression of CYP11A1 isoform b had no significant

ix
effects on pregnenolone synthesis. In support of the third proposed mechanism, we found that
nitric oxide, iron chelators, and knockdown of the mitochondrial CYP450 co-factor ferredoxin
reductase significantly inhibited pregnenolone production. Altogether, our data suggest that
pregnenolone is synthesized by a mitochondrial cytochrome P450 enzyme other than CYP11A1
in the human brain.

1
I. Introduction and Literature Review
This section is adapted from:
Lin, Y. C., & Papadopoulos, V. (2021). Neurosteroidogenic enzymes: CYP11A1 in the central nervous
system. Front Neuroendocrinol, 62, 100925. doi:10.1016/j.yfrne.2021.100925

Lin, Y. C., Cheung, G., Espinoza, N., & Papadopoulos, V. (2022). Function, regulation, and
pharmacological effects of pregnenolone in the central nervous system. Current Opinion in
Endocrine and Metabolic Research, 22. doi:10.1016/j.coemr.2021.100310

1.0 Abstract
Neurosteroids, steroid hormones synthesized locally in the nervous system, have
important neuromodulatory and neuroprotective effects in the central nervous system.
Pregnenolone is one of the most abundant neurosteroids and the precursor to all neurosteroids
found in the brain, and thus is arguably one of the most important neurosteroids. Pregnenolone
can modulate myelination, neuroinflammation, neurotransmission and neuroplasticity in the
brain, implying its importance in cognition, aging, and addiction. Evidence for its therapeutic
potential has emerged following studies reporting altered pregnenolone levels in psychiatric and
pain disorders as well as therapeutic benefits of pregnenolone adjunct therapy in these
conditions. However, there is insufficient evidence to support the assumption that
steroidogenesis is exactly the same between the nervous system and the periphery, particularly
regarding the synthesis pathway for pregnenolone. CYP11A1 is the only enzyme currently
known to catalyze the first reaction in steroidogenesis to produce pregnenolone. Although
CYP11A1 mRNA has been found in the brain of many mammals, the presence of CYP11A1
protein has been difficult to detect, particularly in humans. Here, we highlight the discrepancies
in the current evidence for CYP11A1 in the central nervous system and propose new directions
for understanding neurosteroidogenesis, which will be crucial for developing neurosteroid-based
therapies for the future.

2
1.1 Neurosteroids: the significance of pregnenolone
The central and peripheral nervous systems can produce steroid hormones that can act
locally to alter physiology and behavior. Although hormonally active steroids were shown to
have neurological effects in 1941 (Selye, 1941), the term “neurosteroids” was not introduced
until 1981. Neurosteroids refers to steroids synthesized in the nervous system by either by de
novo synthesis from cholesterol or by metabolism in situ from serum precursors (Baulieu, 1981).
The term was coined when dehydroepiandrosterone (DHEA) sulfate was found to accumulate in
rat brains after adrenalectomy with castration had eliminated all peripheral steroid synthesis,
highlighting the ability of brain to produce steroids independently of peripheral gland secretion.
Introduction of the term “neurosteroids” triggered a profusion of studies investigating their
function and synthetic pathways.
Neurosteroids, like steroids made in the periphery, have biological effects in brain from
embryogenesis to adulthood via classical genomic actions as well as non-genomic ones (Mellon
& Griffin, 2002). Genomic actions involve the interaction of neurosteroids with nuclear receptors
to modulate transcription. For example, estrogen, testosterone, and their metabolites can dampen
the expression of pro-inflammatory factors and enhance expression of anti-inflammatory proteins
in brain to modulate neuroinflammation, which is implicated in diseases such as multiple
sclerosis, Alzheimer’s disease, and Parkinson’s disease (Giatti, Boraso, Melcangi, & Viviani,
2012). Non-genomic actions involve mediation of neurotransmitter receptors. The interactions of
neurosteroids with GABAA receptors, in particular, have been extensively studied.
Allopregnanolone is one of the most potent positive modulators of GABAA receptors, with
beneficial effects in stress, anxiety, depression, and epilepsy (Bali & Jaggi, 2014; Reddy, 2011).
Brexalonone, an allopregnanolone-based drugh, was approved by the U.S. Food and Drug

3
Administration for treatment of postpartum depression (Walton & Maguire, 2019), the first
neurosteroid clinically translated into an approved therapy. On the other hand, neurosteroids such
as pregnenolone sulfate and DHEA sulfate can have opposite effects to allopregnanolone by
inhibiting GABAA receptors (M. Wang, 2011). Neurosteroids can also modulate N-methyl-D-
aspartate (NMDA) receptors to affect learning and memory, with pregnenolone sulfate as a
positive allosteric modulator (Ratner, Kumaresan, & Farb, 2019).
The biosynthetic pathway for neurosteroids start with cholesterol, which is converted by
the cytochrome P450 11A1 (CYP11A1) enzyme to the major steroid precursor, pregnenolone.
Not only is pregnenolone the precursor to all other neurosteroids, it is also one of the most
abundant neurosteroids found in the brain (Weill-Engerer et al., 2002) and has important
modulatory activities by itself. However, compared to downstream steroids such as pregnenolone
sulfate, progesterone, dehydroepiandrostenedione (DHEA), and allopregnanolone, pregnenolone
is understudied. In the following sections, a summary and discussion of what is currently known
about the function, regulation, and pharmacology of pregnenolone in the CNS will be presented.

1.1.1 Function of pregnenolone in the CNS
Apart from being a neurosteroid precursor, pregnenolone itself has valuable functions in
maintaining neuronal health. First, pregnenolone may be important for myelination by
facilitating Schwann cell differentiation into the myelinating phenotype (Zhu & Glaser, 2008). It
is likely that pregnenolone can assist myelination in the CNS as well, since oligodendrocytes
have the highest expression of CYP11A1 and produce the most pregnenolone compared to other
brain cells (Zwain & Yen, 1999). Pregnenolone also has anti-neuroinflammatory effects by

4
promoting degradation of toll-like-receptor adaptor proteins in microglia and suppressing pro-
inflammatory cytokine secretion, implying a role in restoring immune homeostasis in
inflammatory conditions within the CNS (Murugan, Jakka, Namani, Mujumdar, &
Radhakrishnan, 2019). In addition, pregnenolone can protect hippocampal neurons against
glutamate- and amyloid β protein-induced cell death by preventing disruption of calcium
homeostasis (Gursoy, Cardounel, & Kalimi, 2001; Kato-Negishi & Kawahara, 2008).
Pregnenolone can also have beneficial effects on cognition. Pregnenolone showed
memory-enhancing effects and dose-dependently attenuated learning deficits induced by amyloid
β in rodents, suggesting anti-amnesic effects (Flood, Morley, & Roberts, 1992; Maurice, Su, &
Privat, 1998). Its ability to dose-dependently increase the rate and extent of microtubule-
associated protein (MAP-2)-induced tubulin assembly in rat neuron cultures indicates its
importance for neuroplasticity, brain development, and aging (Murakami, Fellous, Baulieu, &
Robel, 2000).
It is important to note the distinction between pregnenolone and pregnenolone sulfate, the
latter being reported as a potentiator of some NMDA receptors as well as an inhibitor at other
neurotransmitter receptors (Ratner et al., 2019). Although both neurosteroids play roles in
neuroplasticity (Murakami et al., 2000; Ratner et al., 2019), there is little evidence of
pregnenolone’s interactions with neurotransmitter receptors, suggesting that these molecules act
on the CNS by different mechanisms. Many studies have focused on pregnenolone sulfate’s
actions in the brain, but pregnenolone’s function in the brain has not been well established.
Therefore, more studies are needed to better understand pregnenolone’s function in the CNS,
which also calls for accurate detection techniques to distinguish between pregnenolone and
pregnenolone sulfate in biological samples.

5
1.1.2 Diseases and conditions with altered pregnenolone levels in the CNS
1.1.2.1 Neurodegenerative Diseases
Despite the interest in using neurosteroids as biomarkers, changes in pregnenolone levels
in neurodegenerative diseases remain unclear. A few studies reported significantly higher levels
of pregnenolone in brains from Alzheimer’s disease (AD) patients, specifically in the temporal
cortex and hypothalamus (R. C. Brown, Han, Cascio, & Papadopoulos, 2003; Naylor et al.,
2010), with non-statistically significant positive associations between pregnenolone levels and
AD neuropathological disease stage (Naylor et al., 2008). Meanwhile, others found no
differences in brain pregnenolone levels between AD patients and nondemented controls (S. B.
Kim et al., 2003; Naylor et al., 2008; Weill-Engerer et al., 2002). To complicate the matter,
significantly lower brain pregnenolone levels were seen in rats injected intracerebrally with Aβ
peptides to induce AD (S. Liu et al., 2013). Overall, there is no consensus to the changes in brain
pregnenolone levels in AD. Similarly, significant increases in hindbrain pregnenolone levels and
significant decreases in spinal cord pregnenolone synthesis were seen in rat models of multiple
sclerosis, but no significant changes were observed in human patients (Giatti et al., 2020;
Noorbakhsh et al., 2011). In a rat model of Parkinson’s Disease, pregnenolone levels were
significantly decreased in the striatum but not the cortex (Melcangi et al., 2012); however, there
is little information about changes in pregnenolone in human patients. Thus, individual variations
in brain pregnenolone levels are high and it is unlikely that pregnenolone will be a reliable
biomarker for neurodegenerative diseases.

6
1.1.2.2 Psychiatric Disorders
Although significant changes in pregnenolone levels in the blood have been observed in
patients with schizophrenia, the direction of change has not been consistent across studies, even
though antipsychotic treatments tend to reverse the changes in blood pregnenolone levels (Cai et
al., 2018; M. Ritsner, Maayan, Gibel, & Weizman, 2007). The discrepancy in pregnenolone level
changes in blood could be due to differences in peripheral metabolism, indicating the importance
of studying local neurosteroid levels both in the periphery and CNS. Studies with access to CNS
samples thus provide valuable insights: significantly increased levels of pregnenolone were
found in post-mortem brain tissue of patients with schizophrenia and bipolar disorder (Marx,
Stevens, et al., 2006). In patients with depression, pregnenolone levels in the CSF were
significantly lower in individuals that were in depressed states at the time of their lumbar
puncture (George et al., 1994). Therefore, although allopregnanolone has typically been the
neurosteroid of interest in psychiatric disorders, these studies highlight the potential significance
of pregnenolone as well.

1.1.2.3 Nerve Injury and Pain
The role of neurosteroids in neuropathic pain has mainly been studied in the peripheral
nervous system, but evidence for effects in the CNS has emerged. Pregnenolone content was
increased in all regions of the rat spinal cord after spinal cord transection, which was not affected
by adrenalectomy with castration, suggesting an upregulation of local pregnenolone synthesis as
a response to injury (Labombarda et al., 2006). Similarly, in rats with chronic neuropathic pain
induced by sciatic nerve ligation, pregnenolone levels are increased in the dorsal horn of the
spinal cord (Patte-Mensah et al., 2006). Recently, it was found that spared nerve injury, used to

7
induce neuropathic pain, increased levels of pregnenolone, progesterone, and allopregnanolone
in the rat hippocampus and thalamus, which can produce anxiolytic and analgesic effects (M.
Zhang et al., 2017; M. Zhang et al., 2016). Supporting these findings, upregulated gene
expression in the pregnenolone biosynthesis pathway was found in the medial prefrontal cortex
and nucleus accumbens of rats with chronic neuropathic pain (Descalzi et al., 2017). Altogether,
these studies suggest that increased pregnenolone levels during nerve injury and neuropathic pain
may be a mechanism to reduce sensitivity to pain.

1.1.2.4 Others
Although diabetic neuropathy is typically associated with changes in peripheral nerves,
induction of diabetes in rat models have revealed changes in neurosteroids within the CNS. In
particular, pregnenolone levels are decreased in the cerebral cortex, cerebellum, and spinal cord
of diabetic rats, which show an opposite trend to the changes in the sciatic nerve (Pesaresi et al.,
2010; Romano et al., 2018). In addition to diabetes, changes in pregnenolone levels in the brain
may be seen in diseases such as HIV-AIDS and acute liver failure (Ahboucha et al., 2008;
Belanger, Desjardins, Chatauret, Rose, & Butterworth, 2005; Maingat et al., 2013). However,
these changes are likely consequences of altered peripheral secretion, and more studies would be
needed to confirm whether brain pregnenolone levels play a role in progression of these diseases.

8
1.1.3 Therapeutic potential of pregnenolone in CNS disorders
1.1.3.1 Psychiatric Disorders
In the past decade, numerous clinical trials have tested the effects of pregnenolone
adjunct therapy in psychiatric disorders. In two separate trials, pregnenolone adjunct treatment
improved depressive symptoms in patients with bipolar disorder but showed no significant
improvement with cognition (E. S. Brown et al., 2014; Osuji, Vera-Bolanos, Carmody, & Brown,
2010). In schizophrenia, even though changes in blood pregnenolone levels in patients varied
across studies, pregnenolone adjunct therapy consistently showed benefit in treating
schizophrenia. Pregnenolone reduced negative symptoms in schizophrenia patients (Marx et al.,
2009; M. S. Ritsner, Bawakny, & Kreinin, 2014) and allowed significant cognitive improvement
compared to antipsychotic monotherapy (Cho et al., 2019). Patients with lower baseline levels of
allopregnanolone, pregnenolone, and pregnenolone sulfate showed more improvements in their
symptoms (Marx et al., 2009), while patients with minimal baseline symptoms did not benefit
from pregnenolone therapy (Cho et al., 2019). Higher levels of serum pregnenolone and
allopregnanolone after treatment also predicted better symptom assessment scores. Overall,
pregnenolone has therapeutic potential as an adjunct therapy for psychiatric disorders, but its
efficacy may be limited to a subset of patients.

1.1.3.2 Nerve Injury and Neuropathy
Pregnenolone may be beneficial in treating nerve injury by promoting recovery.
Following spinal cord injury in rats, a combination treatment of pregnenolone, liposaccharide,
and indomethacin allowed significant reductions in lesion size and cavitation, as well as
improved motor function, suggesting that pregnenolone may help coordinate neural, metabolic,

9
and immune systems to promote recovery (Guth, Zhang, & Roberts, 1994). In another nerve
injury model, pregnenolone increased guided regeneration of rabbit nerves (Chavez-Delgado et
al., 2005), which is supported by evidence for pregnenolone’s role in myelination and tubulin
assembly (Murakami et al., 2000; Zhu & Glaser, 2008). Of course, the effects of pregnenolone
on regeneration of CNS neurons, particularly human neurons in the brain, need to be studied
further. Pregnenolone may also benefit pain disorders through reducing the sensation of pain;
recently, pregnenolone treatment for chronic low back pain in veterans showed promising effects
in reducing pain intensity (Naylor et al., 2020).

1.1.3.3 Challenges and Important Considerations
Despite pregnenolone’s apparent therapeutic potential, the results from these studies
should be considered with important caveats. First, the observed effects of neurosteroids are
likely not due to the action of pregnenolone alone. Only a small proportion of studies have
explicitly demonstrated the direct effects of pregnenolone (Flood et al., 1992; Maurice et al.,
1998; Murugan et al., 2019; Zhu & Glaser, 2008). On the other hand, downstream steroids such
as allopregnanolone and progesterone may have more potent pharmacological effects. Thus, it
can be argued that pregnenolone may not be the most optimal neurosteroid to be used as a drug
and more studies would be needed to confirm its exact effect by use of metabolic inhibitors or
non-metabolized analogues. Second, the mechanisms of action for pregnenolone effects are not
always well understood. Although neurosteroid interactions with glutamate and GABA receptors
are well known, dopamine and serotonin signaling pathways may be involved as well (Muneoka
& Takigawa, 2002; Wong et al., 2012). Given that schizophrenia is commonly treated using
dopaminergic drugs, it is possible that pregnenolone’s therapeutic benefit in schizophrenia is

10
through modulation of dopaminergic pathways. In the context of pain, it is uncertain whether
pregnenolone’s analgesic effect is due to direct action on peripheral sensory neurons or on pain
signaling pathways in the brain through cannabinoid receptors. Finally, when administering
pregnenolone as a drug, it is crucial to consider physiological levels present in the body and the
concentration that will reach the CNS to produce a therapeutic effect. With systemic
pregnenolone administration, the presence of pregnenolone binding proteins (Roy et al., 2021;
Strott, 1977) may also affect actual therapeutic doses, which can create more challenges.

1.1.4 Drugs that can alter pregnenolone levels in the CNS
1.1.4.1 TSPO Ligands
The role of translocator protein (TSPO) in the brain is often associated with interactions
with benzodiazepines and anxiolytics, but also with stimulation of steroidogenesis through
enhanced cholesterol transport into the mitochondria, which can be mimicked using TSPO
ligands (Rupprecht et al., 2010). Etifoxine, an FDA approved anxiolytic and anticonvulsant drug,
binds TSPO and dose-dependently increases pregnenolone levels in the rat brain, even upon
removal of peripheral secretion, indicating a CNS-specific mechanism (Liere, Pianos, Oudinet,
Schumacher, & Akwa, 2017; Verleye et al., 2005). Similar effects can be seen with established
TSPO ligands: XBD-173 increased pregnenolone synthesis in glial cells in a TSPO-dependent
manner, while FGIN-1-27 and 4’-chlorodiazepam increased brain pregnenolone levels in
adrenalectomized-castrated rats without altering pregnenolone levels in the periphery and plasma
(Bader et al., 2019; Costa, Auta, Guidotti, Korneyev, & Romeo, 1994; Korneyev et al., 1993;
Lejri et al., 2019). YL-IPA08, a high affinity TSPO ligand with antidepressant and anxiolytic

11
effects, also increased pregnenolone in rat astrocytes (L. M. Zhang et al., 2014). Interestingly, PK
11195 has been reported to inhibit the stimulatory effects of other TSPO ligands, such as that of
FGIN-1-27 (Costa et al., 1994). Other ligands that can increase pregnenolone synthesis in brain
cells include imidazoquinazolinone-based TSPO ligands 2a and 2b and N,N-dialkyl-2-
phenylindol-3-ylglyoxylamide (PIGA) compounds (Da Pozzo et al., 2016; Grimm et al., 2020;
Lejri et al., 2019). In in vitro and in vivo models, the reported effects of TSPO ligands on
steroidogenesis have mainly been either stimulatory or none with little evidence of inhibition.
However, it is important to consider that TSPO ligands may be agonistic or antagonistic
depending on dose, thus introducing challenges to using TSPO ligands as drugs to alter
neurosteroid levels.

1.1.4.2 Drugs of Abuse
Interestingly, drugs of abuse have been found to change pregnenolone levels in the brain.
Tetrahydrocannabinol (THC), the principal psychoactive component in cannabis, dose-
dependently increased the levels of pregnenolone in multiple brain areas of treated mice (Vallee
et al., 2014). Acute nicotine and morphine treatments also led to increases in cerebrocortical and
plasma levels of pregnenolone in rats, which were attenuated by inducing withdrawal (Concas et
al., 2006). These changes in brain pregnenolone levels appear to be due to modulation of the
hypothalamic-pituitary-adrenal axis to alter peripheral synthesis rather than changes in local
synthesis in the brain, but the possibility exists that morphine and nicotine may increase
neurosteroid synthesis in a specific subpopulation of neurons. In humans, increased
pregnenolone levels were seen in post-mortem brain tissue of alcoholics (Karkkainen et al.,
2016). Longer use of cocaine also correlated with decreases in pregnenolone levels in the serum

12
(Milivojevic, Covault, Angarita, Siedlarz, & Sinha, 2019). In addition, pregnenolone has been
suggested to amplify the rewarding effects of cocaine (Romieu, Martin-Fardon, Bowen, &
Maurice, 2003). Taken together, these studies imply that pregnenolone could play a role in
addiction through modulation of the sigma-1 receptor and cannabinoid receptors (Romieu et al.,
2003; Vallee et al., 2014).

1.1.4.3 Others
Other drugs that can increase pregnenolone levels in the brain include L-ascorbic acid
and caffeine, but by different mechanisms of action. L-ascorbic acid requires endogenous 5-HT
and adenylate cyclase activity to increase neurosteroid synthesis (Roscetti et al., 1998).
Meanwhile, caffeine does not directly affect neurosteroid production, but rather modulates the
hypothalamic-pituitary-adrenal axis to increase peripheral pregnenolone secretion (Concas et al.,
2000). Finally, drugs that can impair mitochondria function, such as rotenone, can inhibit
pregnenolone production by brain cells (Avallone et al., 2020).

1.1.5 Summary and Discussion
Pregnenolone is not just a precursor to other neurosteroids; it has important modulatory
activities by itself in the CNS. The function, regulation, and pharmacology of pregnenolone is
summarized in Figure 1. However, since pregnenolone is readily metabolized to downstream
steroids, many of the reported functions of pregnenolone cannot be simply attributed to the direct
effects of pregnenolone if comparisons to downstream steroids have not been assessed. Thus,

13
investigating pregnenolone’s role in the CNS may require additional considerations such as using
metabolic inhibitors or non-metabolizable forms of pregnenolone. Nevertheless, in a
physiological setting, pregnenolone can still be considered neuroprotective and anti-
inflammatory and can modulate neuronal functions, since the CNS expresses the enzymes to
convert pregnenolone into potentially more potent metabolites. Currently, the drugs that can most
consistently alter pregnenolone synthesis in the CNS are TSPO ligands, although some changes
can be seen with drugs of abuse, indicating pregnenolone’s role in reward and addiction. Even
though pregnenolone is unlikely to become a biomarker for neurodegenerative diseases, its levels
in the CNS are altered in mood disorders and nerve injury/neuropathic pain, for which
pregnenolone has shown some degree of therapeutic potential.

Figure 1: Functions and pharmacological regulations of pregnenolone in the CNS.
Etifoxine: 6-Chloro-N-ethyl-4-methyl-4-phenyl-4H-3,1-benzoxazin-2-amine; XBD-173
(emapunil): N-Ethyl-7,8-dihydro-7-methyl-8-oxo-2-phenyl-N-(phenylmethyl)-9H-purine-9-
acetamide; FGIN-1-27: N,N-Dihexyl-2-(4-fluorophenyl)indole-3-acetamide; PIGA: N,N-dialkyl-
2-phenylindol-3-ylglyoxylamide.

14
1.2 Steroidogenesis: overview of the CYP11A1 enzyme
Since different neurosteroids can have opposing effects on various receptors, the ultimate
biological effect would depend on how much each neurosteroid is metabolized and the summed
effects from each parent neurosteroid and its metabolites. Therefore, it is important to understand
steroidogenesis in the nervous system. The process of steroidogenesis is complex, with the
involvement of multiple cytochrome P450s and hydroxysteroid dehydrogenases (Payne & Hales,
2004)(Figure 2). The rate of steroid synthesis is also limited by the rate of cholesterol transfer
into the mitochondria, the first step of steroidogenesis, which is mediated by translocator protein
(TSPO) and steroidogenic acute regulatory protein (STAR) (J. Liu, Rone, & Papadopoulos,
2006). The brain had not been considered a steroidogenic organ until evidence showed that brain
cells could catalyze conversion of radioactive precursors to respective steroid products (Baulieu
& Robel, 1990). Although studies have provided evidence for steroidogenic enzymes in brain
(Le Goascogne et al., 1987; Mellon & Deschepper, 1993), the steroidogenic pathway in the CNS
was assumed to be the same as the periphery and, therefore, not well-studied separately.
Due to their low expression, steroidogenic enzymes can be difficult to detect in brain,
particularly CYP11A1, the enzyme responsible for the first catalytic step in steroidogenesis. As
discussed earlier, the product of CYP11A1, pregnenolone, is the precursor to all other
neurosteroids. This enzymatic reaction by CYP11A1 is also a rate-determining step in
steroidogenesis (Miller & Auchus, 2011), meaning that alterations to CYP11A1 activity can have
significant downstream effects. The next sections will summarize the information currently
known about CYP11A1, followed by a discussion of the current evidence for the presence of
CYP11A1 mRNA, protein, and activity in the CNS of various species, but focusing on rat,
mouse, and human. The uncertainties from current experimental data will be highlighted, as well

15
a discussion of new avenues for investigation that would allow better understanding of
steroidogenesis in the CNS.

Figure 2: Schematic diagram of the neurosteroid synthesis pathway.
Cholesterol is transported into the mitochondria, mediated by translocator protein (TSPO) and
steroid acute regulatory protein (STAR), to be metabolized to neurosteroids by various
cytochrome P450s and non-CYP450 enzymes. The three steps for converting cholesterol to the
major steroid precursor pregnenolone by CYP11A1 is detailed in the blue box. Names of
neurosteroids are written in bold. Enzymes in the steroidogenic pathway include: cytochrome
P450 side-chain cleavage (CYP11A1), cytochrome P450 17α-hydroxylase/C17,20-lyase
(CYP17A1), aromatase (CYP19A1), 21-hydroxylase (CYP21A2), 3β-hydroxysteroid
dehydrogenase (HSD3B1, HSD3B2), 17β-hydroxysteroid dehydrogenase (HSD17B1,
HSD17B2, HSD17B3, HSD17B4), 5α-reductase (SRD5A1, SRD5A2), 3α-hydroxysteroid
dehydrogenase (AKR1C1, AKR1C2, AKR1C3, AKR1C4), steroid sulfatase (STS), and
sulfotransferase (SULT2A1, SULT2B1).

16
1.2.1 Biochemistry
The CYP11A1 enzyme, also known as cytochrome P450 side-chain cleavage (P450scc),
is located in the inner mitochondrial membrane (Vickery, 1993). It catalyzes the side-chain
cleavage reaction that converts cholesterol into the primary steroid precursor pregnenolone.
Catalysis of cholesterol to pregnenolone by CYP11A1 involves three steps: two hydroxylation
reactions followed by a desmolase activity that cleaves the cholesterol side-chain (Burstein,
Middleditch, & Gut, 1975; Chaudhuri, Harada, Shimizu, Gut, & Dorfman, 1962) (Figure 2). Two
intermediates, 22(R)-hydroxycholesterol and 20α, 22(R)-dihydroxycholesterol, are produced in
the process, both of which do not accumulate and remain bound to the enzyme until release of
the final product pregnenolone (Simpson & Boyd, 1967). Steroid hydroxylation activity by
CYP11A1 requires electron transport from two NADPH-specific flavoproteins: ferredoxin and
ferredoxin reductase (also known as adrenodoxin and adrenodoxin reductase, respectively)
(Lambeth, Seybert, Lancaster, Salerno, & Kamin, 1982). The electron transport begins with
electron transfer from NADPH to FAD on ferredoxin reductase, which complexes with
ferredoxin to transfer the electron onto the iron-sulfur center of ferredoxin. Reduction of the
iron-sulfur center promotes dissociation of ferredoxin from ferredoxin reductase, allowing
ferredoxin to interact with the heme group of the membrane-bound CYP11A1. Even though
CYP11A1 can accept electrons from various electron-transfer proteins, including ferredoxin
reductase and P450 oxidoreductase, the mitochondrial environment appears to be required for its
activity (Black, Harikrishna, Szklarz, & Miller, 1994; Huang & Miller, 2001). It has been shown
CYP11A1-ferredoxin-ferredoxin reductase fusion proteins exhibited detectable enzymatic
activity when targeted to the mitochondria but not when they are targeted to the endoplasmic
reticulum (Black et al., 1994).

17
In addition to cholesterol, CYP11A1 can also metabolize other substrates, including
hydroxycholesterols, 7-dehydrocholesterol, vitamin D, ergosterol, and lumisterol (Slominski et
al., 2015). CYP11A1 can metabolize 20α-, 24-, 25-, and 26-hydroxycholesterols to pregnenolone
at higher rates than cholesterol, as polar hydroxycholesterols are not limited by the rate of
transfer through the mitochondrial membrane like cholesterol (Craig, Mason, Suckling, & Boyd,
1982; Mason, Arthur, & Boyd, 1978). A five- to ten-carbon cholesterol side chain is also required
for CYP11A1 side-chain cleavage reactions, with hydroxyl groups on C22, C23, C24, and C25
shown to improve reaction rates (Morisaki, Duque, Ikekawa, & Shikita, 1980).
Dihydroxycholesterols are generally poor substrates for CYP11A1 with the exception of the
classical substrate 20α,22(R)-dihydroxycholsterol.

1.2.2 Molecular Biology
The first cloning study for CYP11A1 was performed using RNA isolated from bovine
adrenals and corpus luteum, showing that this enzyme is encoded by a single gene (John et al.,
1984). The CYP11A1 mRNA is approximately 2000 base pairs long, the majority of which is
polyadenylated. A later study showed that the bovine CYP11A1 precursor contains 520 amino
acids, while the mature protein contains 481 amino acids (Morohashi et al., 1984). The first 39
amino acids at the N-terminus functions as a mitochondrial-signaling peptide that is cleaved
upon mitochondrial entry. Side-directed mutagenesis studies have identified the lysine and
arginine residues important for complex formation with ferredoxin, such as K403, K405, and
R426 (Usanov et al., 2002). Other residues such as K109 and R425 have been found to play
important roles in protein folding and heme insertion in bovine CYP11A1 (Azeva, Gilep,

18
Lepesheva, Strushkevich, & Usanov, 2001; Lepesheva, Azeva, Strushkevich, Gilep, & Usanov,
2000).
Crystallization studies have revealed that the structure of bovine CYP11A1 shows a
classical CYP450 fold with two additional helices (Mast et al., 2011). The active site is a banana-
shaped tunnel where cholesterol within the active site is bent at the C20 and C22 positions (Mast
et al., 2011; Strushkevich et al., 2011). The hydrogen on C22 points towards the oxyferryl
oxygen on the enzyme to allow preferential hydroxylation at this position (Strushkevich et al.,
2011). The 3β-hydroxyl groups of cholesterol form hydrogen bonds with water molecules and
enzyme residues to prevent the substrate from sliding out of the active site. The intermediate
22(R)-hydroxycholesterol occupies approximately 2/3 of the volume of the active site and is
pushed towards the entrance of the site upon conformational changes triggered by interactions
between CYP11A1 and reduced ferredoxin (Mast et al., 2011). Once a reactive oxyferryl
intermediate is formed, 22(R)-hydroxycholesterol moves back towards the heme group of the
enzyme to allow hydroxylation on C20 to form 20α,22(R)-dihydroxycholesterol. The two
hydroxyl groups on 20α,22(R)-dihydroxycholesterol were shown to be in close proximity to the
heme group on CYP11A1 but do not directly interact (Strushkevich et al., 2011). The oxidant
that is ultimately responsible for mediating the carbon-carbon bond cleavage during the last step
of the CYP11A1 enzymatic reaction remains unknown. Superimposition of bovine and human
CYP22A1 bound with 22(R)-hydroxycholesterol shows similarities in shape and volume of the
active site, as well as the positions of water molecules at the entrance to the active site (Mast et
al., 2011).
The human CYP11A1 gene is located on chromosome 15, and the mRNA sequence was
found to contain 1821 bases, excluding the poly-A tail (Chung, Matteson, V outilainen,

19
Mohandas, & Miller, 1986). Similar to bovine, the human genome has a single CYP11A1 gene
(Matteson, Chung, Urdea, & Miller, 1986). Two isoforms of CYP11A1 have been reported,
although no studies have determined differences in expression levels or functions between the
two isoforms (Gene ID: 1583, NCBI Gene, 2021). The mRNA transcripts encoding these
isoforms differ in the sequence of exon 1 but have identical sequences from exon 2 to exon 9
(Figure 3A). CYP11A1 isoform a represents the classical 521-amino-acid CYP11A1 enzyme,
with a 39-amino-acid mitochondrial targeting sequence at the N-terminus. Isoform b is a 363-
amino-acid protein that lacks the first 158 N-terminal amino acids found in isoform a but is still
believed to be localized to the mitochondria despite the lack of a mitochondrial targeting
sequence. However, more studies are needed to characterize the function and subcellular
localization of isoform b.
Similar to humans, two CYP11A1 isoforms have been reported in mice (Gene ID: 13070,
NCBI Gene, 2021). The transcript variants for these isoforms share identical sequences in the
last 7 exons (Figure 3B). In both mouse and humans, the second transcript variant includes one
or more exons that are not translated, but the factors involved in this translation-control process
are unknown. The translated mouse CYP11A1 isoform 1 is a 526 amino acids protein with a
mitochondrial transit peptide at residues 1 – 36, while isoform 2 is a 371 amino acid protein that
has a shorter N-terminus. Nevertheless, both isoforms are believed to be mitochondrial. On the
other hand, only one isoform for CYP11A1 has been reported in rats (Gene ID: 29680, NCBI
Gene, 2021), which has the same protein length as mouse isoform 1 (Figure 3C).

20
Figure 3: Graphic representation of the CYP11A1 gene.
Corresponding mRNA transcript variants and protein isoforms for Homo sapiens (A), Mus
musculus (B), and Rattus norvegicus (C). Exons are depicted as boxes in the mRNA transcript,
with the nucleotide number at the beginning and end of each exon indicated. Exons with
identical sequences between transcript variants are depicted in the same color. Introns are drawn
as a line between exons. The mitochondria targeting peptide (MTP) is depicted as a purple box.
Alignment of the protein with the translation start site is shown as a dotted line.

21
1.2.3 Phylogeny
CYP11A1 is only found in vertebrates (Slominski et al., 2015). The structure of bovine
CYP11A1 is more similar to that of other animal microsomal CYP450s than to that of bacterial
mitochondrial CYP450s, suggesting the early divergence of bacterial and eukaryotic CYP450
systems, followed by divergence of microsomal and mitochondrial CYP450s within eukaryotes
(Morohashi et al., 1984). Mitochondrial CYP450s are found only in the animal kingdom and
possess a C-GRR—E enzyme motif that interacts with ferredoxin for enzymatic reactions (Fan &
Papadopoulos, 2013). Phylogenetic analysis shows that mitochondrial cytochrome P450
enzymes CYP11A1 and CYP11B1 form a branch within all vertebrates, and likely resulted from
a gene duplication event during evolution. Although CYP11A1 and CYP11B1 both bind to the
inner mitochondrial membrane, they were found to be dissimilar by functional divergence
analysis. While CYP11A1 activity is important in all steroidogenic organs by producing the
major steroid precursor pregnenolone, CYP11B1 activity is associated with more downstream
adrenal steroidogenesis by producing cortisol and corticosterone (Parajes et al., 2010). Animal
CYP11A1 was also found to be similar to the CYP44A1 mitochondrial enzyme from C. elegans
and insects, but CYP11A1 favors cholesterol binding while CYP44A1 favors 7-
dehydrocholesterol binding (Fan & Papadopoulos, 2013).
Phylogenetic analyses of CYP11A1 shows three distinct branches within animals:
mammals, fish, and birds, with the mammalian group branching off before fish and birds diverge
(Xu et al., 2015). One study comparing the amino acid sequence for chicken CYP11A1 to that of
other species found 52% homology to humans, 51% homology to bovines, 54% homology to
pigs, 52% homology to rats, and 59% homology to rainbow trout (Nomura, Nakabayashi,

22
Nishimori, & Mizuno, 1997). A total of 179 common residues were found in the six species, but
no conserved residues were observed in the N-terminal mitochondrial-targeting sequence.
The coding sequence of the human CYP11A1 gene is 82% homologous to the bovine
sequence, with 72% homology in the amino acid sequences (Chung et al., 1986). The amino acid
sequence for sheep CYP11A1 shows approximately 72% homology to human, 95% homology to
bovines, and 70% homology to rats (Okuyama, Okazaki, Furukawa, Wu, & Ichikawa, 1996).
Amino acid residue alignment analyses show that sequence homology is highest (> 75%) in the
heme and steroid-binding domains among species (Kazeto, Ijiri, Adachi, & Yamauchi, 2006). For
example, CYP11A1 in Syrian hamsters share approximately 89% identity with that of mice and
rats, while the steroid-binding and heme-binding domains show 100% homology between the
three species (Vilchis, Chavez, Larrea, Timossi, & Montiel, 2002). Studies for the promoter
region for CYP11A1 also revealed highly conserved sequences among rats, mice, humans, and
bovines, with remarkable similarity in the 5’ flanking sequences between rats and mice (Oonk,
Parker, Gibson, & Richards, 1990).
Another commonly studied species for CYP11A1 is the zebrafish, which shows 78%
amino acid sequence similarity to trout CYP11A1 and 58% sequence similarity to human
CYP11A1 (Lai, Hsiao, Guiguen, & Chung, 1998). Interestingly, zebrafish have a second
CYP11A1 gene, named CYP11A2, that shares 80% protein sequence identity to the zebrafish
CYP11A1 protein (Parajes et al., 2013). Zebrafish CYP11A1 isoenzymes share a common
ancestor with other teleosts, and the closest relative is the stickleback. CYP11A1 is required for
gastrulation in zebrafish, and the expression in adults is restricted to the ovary. On the other
hand, CYP11A2 is important for de novo steroidogenesis and is expressed in the head kidney,

23
gonads, and brain of adult zebrafish. CYP11A2 in zebrafish is suggested to be more functionally
equivalent to human CYP11A1 in steroidogenic tissues (Parajes et al., 2013).

1.3 The discrepancies surrounding CYP11A1 in the brain
1.3.1 Cyp11a1 in the rat brain
1.3.1.1 Cyp11a1 Protein Presence
The majority of the evidence for CYP11A1 in brain was found in rats, with the earliest
studies demonstrating positive immunohistochemical staining and CYP11A1 enzymatic activity.
Brain sections from male Sprague Dawley rats were shown to stain positively for CYP11A1 in
the caudate nucleus, olfactory bulb, and white matter of the cerebellum (Le Goascogne et al.,
1987). In the olfactory bulb, CYP11A1 staining was restricted to myelinated fibers, with very
few positively stained cell bodies. In the entorhinal cortex and cingulate cortex, clusters of cells
were labelled by CYP11A1 antibody, where the signal was restricted to the cytoplasm. Positive
cells in the entorhinal cortex were large with big, spherical nuclei, while those in the cingulate
cortex were small with irregular shapes. Le Goascogne et al. also found that staining for
ferredoxin was found in the same brain regions as positive CYP11A1 staining (Le Goascogne et
al., 1987). A later immunohistochemical study observed that CYP11A1 could be found in the
myelinated regions of the corpus callosum, fornix, anterior commissure, thalamus,
hypothalamus, midbrain, olfactory bulb, and cerebellum white matter (Iwahashi et al., 1990).
Immunoblot analysis showed that the cortical white matter is immunopositive for CYP11A1,
while the cerebral cortex is negative. The concentration of CYP11A1 in rat brains was estimated
to be 3 – 4 pmol per mg tissue protein, which is 10% of the concentration found in the adrenal

24
glands. No differences in immunochemical staining patterns were found between male and
female rats, as well as between the left and right side of the corpus callosum.
Later immunohistochemistry studies located CYP11A protein expression to more specific
cell types within various brain regions. In the cerebellum, intense immunopositive signals were
found in Purkinje cells located in the narrow zone between the molecular and granular layers in
the cerebellar cortex (Ukena, Usui, Kohchi, & Tsutsui, 1998). In the hippocampus,
immunohistochemical staining with CYP11A1 antibody was found in pyramidal neurons in the
CA1-CA3 regions and granule cells in the dentate gyrus (Kimoto et al., 2001). However, only a
small proportion of neurons express CYP11A1. In the brainstem, cell bodies in the raphe magnus
nucleus, dorsal raphe nucleus, and parabrachial nucleus were intensely stained with CYP11A1
antiserum (Patte-Mensah, Kappes, Freund-Mercier, Tsutsui, & Mensah-Nyagan, 2003). Moderate
staining was found in the positive reticular nucleus, while no signal could be detected in the
locus coeruleus. CYP11A1 staining could also be found in the primary and secondary
somatosensory cortex in both cell bodies and fibers. However, the motor cortex contained very
sparse CYP11A1-positive fibers.
Not only was CYP11A1 found in adult rat brains, immunohistochemical staining for
CYP11A1 was observed in rat embryos (E15.5 – E19.5) in the hippocampus, pituitary,
diencephalon, and cortex, although the signals were weak (Compagnone, Bulfone, Rubenstein, &
Mellon, 1995). In contrast, faint staining was found in the pituitary, olfactory tract, hypothalamic
perifornical areas, and hippocampus in adult brains. The staining pattern for CYP11A1-positive
cells appear to change over the course of development (Ukena et al., 1998). At postnatal day 0,
positive cells were scattered throughout the cerebellum, while at days 3 to 7 the signals became
restricted to the Purkinje cell layers of the cerebellum. The changes in immunohistochemical

25
signals also matched the pregnenolone levels found in the cerebellum, which were highest at day
0 and decreased until a plateau on day 7.
The expression of CYP11A1 in rat brain appears to increase with pregnancy; immunoblot
analysis revealed that CYP11A1 antibodies could only detect bands in the mitochondrial
fractions of olfactory lobes and hypothalamic preoptic area (HPOA) in pregnant animals but not
control animals (Warner, Tollet, Stromstedt, Carlstrom, & Gustafsson, 1989). Using hydrophobic
chromatography, this study showed that mitochondrial and microsomal fractions of control rat
brains yielded very low levels of cytochrome P450s, but the yield increased 8-fold in fractions
from the HPOA and olfactory lobes of pregnant rats.

1.3.1.2 Cyp11a1 Enzyme Activity
The first indication of CYP11A1 enzyme activity in brain was finding that pregnenolone
accumulated in brain after removal of steroidogenic gland secretion, suggesting that a local
mechanism in the rat brain is responsible for synthesizing pregnenolone (Corpéchot et al., 1983;
Robel et al., 1987). Pregnenolone concentrations were especially high in the olfactory bulb
compared to the forebrain and cerebellum in adrenalectomized and castrated (A-C) rats
(Korneyev et al., 1993). This implies more CYP11A1 activity in the olfactory bulb, which
supports previous findings of CYP11A1 staining in this region. Furthermore, treating A-C rats
with the TSPO ligands FGIN-1-27 and 4’chlorodiazepam resulted in a 70 – 100% increase in
pregnenolone concentration in the rat forebrain. These results indicate that modulation of
cholesterol transport into mitochondria promotes increased conversion of cholesterol into
pregnenolone, which is presumably catalyzed by CYP11A1.

26
More convincing evidence for the presence of CYP11A1 activity in rat brain emerged
when studies demonstrated metabolism of radioactive pregnenolone precursors in brain extracts.
Rat oligodendrocyte mitochondria were found to convert [
3
H]cholesterol to [
3
H]pregnenolone at
a rate of 2.5 pmol per mg protein per hour (Z. Y . Hu, Bourreau, Jung-Testas, & Robel, 1987).
This was approximately six times slower than adrenal mitochondria, which have a conversion
rate of 14 – 15 pmol per mg protein per hour. The formation of radio-labelled pregnenolone and
20-hydroxy-pregnenolone was completely inhibited by aminoglutethimide (AMG), a CYP11A1
inhibitor. Interestingly, using mitochondria prepared from whole rat brain instead of isolated
oligodendrocytes did not result in any detectable radioactive metabolite, suggesting that
oligodendrocytes may be the primary cells responsible for pregnenolone synthesis in brain.
Cultured oligodendrocytes incubated with [
3
H]mevalonate had detectable levels of radioactive
cholesterol, pregnenolone, and 20-hydroxy-pregnenolone (Jung-Testas, Hu, Baulieu, & Robel,
1989). Moreover, addition of AMG to primary cultures of glial cells incubated with
[
3
H]mevalonate resulted in accumulation of [
3
H]cholesterol (Z. Y . Hu, Jung-Testas, Robel, &
Baulieu, 1989), suggesting blockage of CYP11A1 catalytic activity.
The C6 rat glioma cell line is a commonly used in vitro model for studying neurosteroid
synthesis, and CYP11A1 protein can be detected in the mitochondrial fraction of C6 cells via
immunoblot analysis (Papadopoulos, Guarneri, Kreuger, Guidotti, & Costa, 1992). Pregnenolone
synthesis in C6 cells is dose-dependently stimulated by 22(R)-hydroxycholesterol, suggesting the
presence of CYP11A1 desmolase activity, which can be inhibited by AMG. Other
hydroxycholesterols such as 25-hydroxycholesterol and 20α-hydroxycholesterol also induced
increases in pregnenolone formation in C6 cells.

27
1.3.1.3 Cyp11a1 mRNA Presence
Interestingly, evidence for the presence of Cyp11a1 mRNA in rat brain emerged later than
detection of Cyp11a1 protein. Mellon and Deschepper reported that RNAse protection assays
could not detect Cyp11a1 mRNA in whole rat brains, primary rat glial cultures, or C6 rat glioma
cells even though mRNA for ferredoxin can be detected using this assay (Mellon & Deschepper,
1993). Other studies have also supported the inability of RNAse protection assays to detect
Cyp11a1 mRNA in rat brains (Compagnone et al., 1995; Furukawa, Miyatake, Ohnishi, &
Ichikawa, 1998). On the other hand, Mellon and Deschepper were able to detect Cyp11a1 mRNA
in rat brains and primary glial cultures using the more sensitive RT-PCR followed by Southern
blotting, although 100 times more cDNA was needed for brain compared to adrenals (Mellon &
Deschepper, 1993). The highest levels of Cyp11a1 mRNA were found in the cortex, followed by
the hippocampus, amygdala, and midbrain. No signal could be found in the cerebellum and
hypothalamus in this study.
Most studies reporting Cyp11a1 mRNA expression in brain agree that the concentration
of mRNA is extremely low, especially compared to levels found in the testes and adrenals
(Furukawa et al., 1998; MacKenzie et al., 2000; Sanne & Krueger, 1995; Stromstedt &
Waterman, 1995). However, perhaps due to the low levels of Cyp11a1 mRNA in the brain, there
is little consensus as to which brain region has the highest expression. For example, Stromstedt
and Waterman found that Cyp11a1 mRNA can be found in the cerebral cortex, cerebrum,
cerebellum, and brainstem, all at similar levels (Stromstedt & Waterman, 1995). These RT-PCR
results were confirmed using in situ hybridization, which showed the most intense signal in
myelinated regions in the above-mentioned structures. No specific signal could be observed in
gray matter. These findings contrast with those of Mellon and Deschepper, where higher levels of

28
Cyp11a1 mRNA was found in the cortex and none found in the cerebellum. In general, most
studies have found Cyp11a1 mRNA in the rat cerebellum and cerebral cortex, while mRNA
presence in other regions, such as the brain stem, hippocampus, and diencephalon, was not
commonly reported (Furukawa et al., 1998; Kohchi, Ukena, & Tsutsui, 1998; Sanne & Krueger,
1995; Ukena et al., 1998).
The levels of Cyp11a1 mRNA in rat brain can also change during development, although
the trend is not identical to protein level changes reported by Ukena et al (1998). In both male
and female rats, the expression of Cyp11a1 mRNA in the cerebellum increased from postnatal
day 1 to postnatal day 10, with decreasing expression thereafter (Lavaque, Mayen, Azcoitia,
Tena-Sempere, & Garcia-Segura, 2006). The peak in Cyp11a1 mRNA levels at postnatal day 10
was more than 3-fold higher in males than in female rats. The levels of Cyp11a1 mRNA changed
in parallel with expression of other steroidogenic proteins such as STAR and aromatase
(CYP19A1). Despite the differences between these results and the Cyp11a1 expression peak at
day 0 (Ukena et al., 1998), both studies agree that Cyp11a1 expression decreases as rats age. This
suggests that CYP11A1 could play an important role in cerebellar development, as the highest
expression is seen during the timeframe for granule cell proliferation and migration, Purkinje cell
maturation, and formation of synapses in the molecular layer. Table 1 summarizes the different
brain areas shown to express Cyp11a1 mRNA and protein.

29
Table 1: Rat brain regions with Cyp11a1 expression
Brain Region Cyp11a1 Presence References
Cerebellum Protein
mRNA
Le Goascogne et al. 1987
Iwahashi et al. 1990
Stromstedt and Waterman 1995
Sanne and Krueger 1995
Furukawa et al. 1998
Kohchi et al. 1998
Ukena et al. 1998
Lavaque et al. 2006
Karri et al. 2007 (isolated cells)
Striatum
(caudate nucleus)
Protein Le Goascogne et al. 1987
Olfactory bulb Protein Le Goascogne et al. 1987
Iwahashi et al. 1990
Compagnone et al. 1995
Corpus callosum Protein Iwahashi et al. 1990
Diencephalon
(thalamus,
hypothalamus)
Protein
mRNA
Iwahashi et al. 1990
Campagnone et al. 1995
Kohchi et al. 1998
Ukena et al. 1998
Brain Stem Protein
mRNA
Iwahashi et al. 1990
Mellon and Deschepper 1993
Stromstedt and Waterman 1995
Kohchi et al. 1998
Ukena et al. 1998
Patte-Mensah et al. 2003
Karri et al. 2007 (isolated cells)
Cerebral cortex Protein
mRNA
Iwahashi et al. 1990 (white matter only)
Mellon and Deschepper 1993
Compagnone et al. 1995
Stromstedt and Waterman 1995
Sanne and Krueger 1995
Kohchi et al. 1998
Ukena et al. 1998
Zwain and Yen 1999 (isolated cells)
Amygdala mRNA Mellon and Deschepper 1993
Hippocampus mRNA
Protein
Mellon and Deschepper 1993
Compagnone et al. 1995
Furukawa et al. 1998
Kimoto et al. 2001

30
1.3.2 Cyp11a1 in the mouse brain
In contrast to rats, very few studies have shown the presence of CYP11A1 in the mouse
brain. Cyp11a1 mRNA can be detected in the cerebrum, cerebellum, and brain stem in ICR mice,
with bands of intermediate intensity in Southern blotting (Stromstedt & Waterman, 1995). The
CYP11A1 protein appears to co-localize with cells that stain positively for the steroidogenic
protein STAR in mouse brain and is found in the cerebellum, pons, hypothalamus and cerebral
cortex (King et al., 2002). However, the signal in the cerebral cortex appears to be weaker than
that in the cerebellum and pons. Moreover, only a small percentage of cells in brain express
STAR and CYP11A1, suggesting that few cells in the brain are “steroidogenic.” Specific neurons
in the trigeminal ganglion and colliculus inferior were also found to be stained by CYP11A1
antibodies (Apaja, Harju, Aatsinki, Petaja-Repo, & Rajaniemi, 2004). Single-cell RNAseq of
mouse whole cortex and hippocampus reveals an apparently higher expression in the cortex than
hippocampal areas (Yao Z. et al., Whole Cortex & Hippocampus – 10X Genomics with 2010X-
SMART-SEQ Taxonomy, Allen Brain Institute of Science, 2020). Within the hippocampus, the
most CYP11A1-positive cells are found in the dentate gyrus compared to other hippocampal
regions. Similar to rats, CYP11A1 expression can be observed by immunohistochemistry during
embryogenesis; CYP11A1 staining is present in the neural crest of mouse embryos starting at
E9.5, with faint staining in the forebrain, thalamic region, pituitary and hippocampus at E167.5 –
E18 (Compagnone et al., 1995).

31
1.3.3 CYP11A1 in the human brain
The expression of CYP11A1 in human brain has been more difficult to detect than in
rodents. Most of the evidence points to the presence of CYP11A1 mRNA, but very few studies
have demonstrated protein expression. There appears to be large variation in the reported
concentrations of CYP11A1 mRNA in the human brain; levels range from 200-fold to 100,000-
fold lower than adrenal mRNA (Beyenburg et al., 1999; MacKenzie et al., 2008; Yu, Romero,
Gomez-Sanchez, & Gomez-Sanchez, 2002). Competitive RT-PCR has detected CYP11A1 mRNA
in hippocampal tissue, with higher levels seen in women but no difference between children and
men (Beyenburg et al., 1999). In the frontal lobe cortex, levels of CYP11A1 mRNA were higher
in women than men, while levels in the temporal lobe subcortical white matter showed no sex
difference (Watzka, Bidlingmaier, Schramm, Klingmuller, & Stoffel-Wagner, 1999). No
differences were observed in CYP11A1 mRNA levels between the temporal lobe cortex, temporal
lobe subcortical white matter, frontal lobe cortex, or hippocampus tissue. Another study reported
that CYP11A1 mRNA levels are highest in the corpus collosum, followed by the caudate nucleus,
thalamus, hippocampus, amygdala and the cerebellum in decreasing order of expression level
(Yu et al., 2002). Meanwhile, transcriptomic studies have found significantly higher levels of
CYP11A1 mRNA in the basal ganglia compared to all other brain regions, including the cerebral
cortex, olfactory region, cerebellum, thalamus, amygdala and hypothalamus [dataset](Uhlen et
al., CYP11A1, The Human Protein Atlas, 2015). Although alterations in pregnenolone levels
have been reported in brains of Alzheimer’s Disease (AD) patients, there appears to be no
significant differences in CYP11A1 mRNA expression between control and AD brains
(MacKenzie et al., 2008; Naylor et al., 2010; Weill-Engerer et al., 2002).

32
To date, only two studies have demonstrated CYP11A1 protein in human brain. The first
demonstrated positive immunohistochemical staining in the white matter of the cerebellum (Le
Goascogne et al., 1989). Signals for ferredoxin and ferredoxin reductase were also found in the
same region. In the second study, neurons in the frontal cortex were found to be
immunohistochemically stained by CYP11A1 and STAR antibodies (King et al., 2002).
However, there are limitations to the results of both studies; the images showing positive staining
do not reveal any morphological information about the brain area being examined. In the study
by King et al, only four stained cells were shown, which is not representative of the vast number
of cells found in the frontal cortex. Furthermore, in Le Goascogne’s study, the tissue used
appears to have non-specific reactivity to IgG; incubating the cerebellum tissue with non-
immune IgG resulted in a weaker but similar staining pattern to that of CYP11A1 antibody. It is
worth noting that other studies have explicitly noted the lack of detectable CYP11A1 protein in
human brain tissue (Yu et al., 2002)(Uhlen et al., CYP11A1, The Human Protein Atlas, 2015).
Figure 4 illustrates the brain regions where CYP11A1 mRNA and protein have been
found in rat, mouse, and human.

33

Figure 4: Graphic summary for CYP11A1 expression in the CNS.
mRNA and protein expression for CYP11A1 in rat (A), mouse (B), and human (C) CNS.
Regions found to express CYP11A1 mRNA are shaded red (left) and those found to have
CYP11A1 protein are shaded blue (right). Areas shaded gray indicate areas where CYP11A1
mRNA or CYP11A1 protein have not been reported. CNS regions depicted: olfactory bulb
(OB), cerebral cortex (CTX), striatum (ST), corpus callosum (CC), hippocampus (HC), thalamus
(TH), hypothalamus (HT), amygdala (AM), midbrain (MB), pituitary gland (PG), pons (PO),
medulla obloganta (MO), cerebellum (CB), and spinal cord (SC). The frontal, parietal, temporal,
and occipital lobes are also indicated on the human brain. Figure not drawn to scale.

34
1.3.4 CYP11A1 in different brain cell types
Another aspect under discussion is which brain cell type expresses CYP11A1. Initially,
CYP11A1 was thought to be mainly expressed in oligodendrocytes, as rat oligodendrocyte
mitochondria were found to form radioactive pregnenolone from [
3
H]cholesterol, while the same
experiment on brain slices did not yield any radioactive metabolites (Z. Y . Hu et al., 1987). The
results were explained to be due to the relatively low number of oligodendrocytes in the brain. A
later study supported these results by demonstrating that cultured oligodendrocytes isolated from
the forebrain of rats showed intense immunostaining for CYP11A1, while astrocytes (GFAP-
positive cells) in mixed glial cultures did not stain for CYP11A1 (Jung-Testas et al., 1989).
Cyp11a1 mRNA was also found in rat oligodendrocytes, with RT-PCR analysis showing a
stronger band in oligodendrocytes than astrocytes or neurons (Sanne & Krueger, 1995; Zwain &
Yen, 1999). Furthermore, incubating oligodendrocytes, astrocytes, and neurons isolated from rat
cortex with cholesterol led to dose-dependent increases in pregnenolone secretion by all three
cell types, although the highest production was seen in oligodendrocytes (Zwain & Yen, 1999).
With treatment of 10
-6
M cholesterol, oligodendrocytes produced three- to eight-fold higher
pregnenolone than astrocytes or neurons. Similarly, human oligodendrocytes in vitro were able to
synthesize radiolabeled pregnenolone from mevalonactone, while astrocytes and neurons could
not, despite detectable CYP11A1 mRNA in both oligodendrocytes and astrocytes (R. C. Brown,
Cascio, & Papadopoulos, 2000).
However, other studies have presented contradictory results. Kimoto et al. found no co-
localization between CYP11A1 and myelin basic protein or glial fibrillary acidic protein (GFAP)
in the hippocampus of adult male rats, suggesting the lack of CYP11A1 expression in astrocytes
and oligodendrocytes in the hippocampus (Kimoto et al., 2001). In the rat brainstem, only 10%

35
of CYP11A1-positive cells co-stained with galactocerebroside (GalC), an oligodendrocyte
marker (Patte-Mensah et al., 2003). Furthermore, oligodendrocytes appear not to express
CYP11A1 during embryogenesis, as CYP11A1-positive cells do not stain for GalC in brains of
rat embryos (Compagnone et al., 1995). This poses the question of whether other glial cells
express CYP11A1, since glial cells were the first brain cells shown to produce pregnenolone.
Although some studies have found an absence of CYP11A1 staining in GFAP-positive cells
(Jung-Testas et al., 1989; Kimoto et al., 2001; Ukena et al., 1998), others have suggested that
astrocytes do express CYP11A1 and can synthesize pregnenolone (Karri, Dertien, Stocco, &
Syapin, 2007; Mellon & Deschepper, 1993; Patte-Mensah et al., 2003; Zwain & Yen, 1999). RT-
PCR and Southern blot analysis showed that RNA isolated from rat mixed-glial cultures and
purified type I astrocyte cultures had the same abundance of Cyp11a1 mRNA (Mellon &
Deschepper, 1993). Immunohistochemical staining also revealed clustering of the CYP11A1
protein in a region around the nucleus in type I astrocytes. Although astrocytes did not produce
as much pregnenolone as oligodendrocytes, these cells still showed a clear band for Cyp11a1
RNA after RT-PCR and secreted detectable amounts of pregnenolone without incubation with
exogenous cholesterol (Zwain & Yen, 1999). In rat brainstem, it appears that more astrocytes
express CYP11A1 than oligodendrocytes, as 20% of CYP11A1-positive cells co-stained with
GFAP, while only 10% co-stained with GalC (Patte-Mensah et al., 2003). In human cortex, the
highest CYP11A1 expression can be found in astrocytes compared to other brain cell types,
although the majority of astrocytes showed low or no expression (Multiple Cortical Areas –
SMART-SEQ (2019), Allen Institute for Brain Science).
Neurons were not initially thought to express CYP11A1, as no signal could be found in
neuronal cell bodies or their dendrites in any of the brain regions that stained strongly for

36
CYP11A1 (Iwahashi et al., 1990). However, later studies have consistently shown otherwise. In
rat embryos, a proportion of CYP11A1-positive cells also stained positive for neuron-specific
enolase (NSE), although the signals were too weak to identify the exact cell subtypes
(Compagnone et al., 1995). Neurons isolated from rat cortex were shown to express Cyp11a1
mRNA, and dose-dependent increases in their pregnenolone production was shown with
incubation of cholesterol (Zwain & Yen, 1999). In the rat hippocampus, CYP11A1 staining co-
localized with NeuN signals in a small percentage of cells, which were identified as pyramidal
neurons in the CA1-CA3 regions and granule cells in the dentate gyrus (Kimoto et al., 2001). In
rat brainstem, approximately 70% of CYP11A1-positive cells co-stained with microtubule-
associated protein 2 (MAP-2), with the signals mainly localized to cell bodies (Patte-Mensah et
al., 2003). Both cell bodies and fibers in rat primary and secondary somatosensory cortex also
showed CYP11A1 staining, with homogenates of these areas displaying the ability to convert
radioactive cholesterol to pregnenolone. A small number of neurons in the motor cortex were
also found to be CYP11A1-positive, although the staining was restricted to the fibers.
The brain region that has the most evidence for CYP11A1 expression in neurons is the
cerebellum. Granule cells in the cerebellum express Cyp11a1 mRNA, but at lower levels than
primary glia cultures and oligodendrocytes (Furukawa et al., 1998; Sanne & Krueger, 1995). The
somata and dendrites of Purkinje cells in the cerebellum also stain positive for CYP11A1, with
concentrated signals in granules, indicating localization to intracellular organelles (Furukawa et
al., 1998; Ukena et al., 1998). Similar to rats, neurons from mice also express CYP11A1,
although the regional localization is not as well described as in rats (Apaja et al., 2004; King et
al., 2002). CYP11A1 expression in human neurons is less clear. While neurons differentiated
from the Ntera2/D1 cell line do not express detectable CYP11A1 mRNA and do not synthesize

37
pregnenolone (R. C. Brown et al., 2000), single-cell RNAseq of human cortical areas revealed
that both excitatory and inhibitory neuronal populations have cells with positive CYP11A1
expression (Multiple Cortical Areas – SMART-SEQ (2019), Allen Institute for Brain Science).
In the past decade, an increasing number of studies have proposed a link between
neuroinflammation and neurosteroid levels, indicating the possibility that microglia may also be
steroidogenic (Yilmaz et al., 2019). In a pilocarpine-induced epilepsy model, CYP11A1
expression was induced in glial cells in the CA3 region of the rat hippocampal formation
(Biagini et al., 2009). The glial cell type that stained most highly for CYP11A1 was astrocytes,
but HO-1 positive cells (i.e. activated microglia) also displayed increased CYP11A1 staining. In
comparison, a lower number of oligodendrocytes were positive for CYP11A1 in this study.
Moreover, resting BV-2 mouse microglia cells produced approximately 16.5 pg pregnenolone per
10
5
cells, which can be increased with rotenone treatment (Avallone et al., 2020). Secretion of
pregnenolone by resting BV-2 cells was confirmed in a more recent study, which further showed
that the TSPO ligand XBD-173 significantly stimulates pregnenolone secretion by BV-2 cells
(Bader et al., 2019). Although C20 and HMC3 human microglia cells also produce pregnenolone
under basal conditions, stimulation of their pregnenolone production by TSPO ligands was
inconsistent unlike rodent microglia (Germelli et al., 2021; Milenkovic et al., 2019). Immunoblot
analysis of C20 and HMC3 cells revealed a very faint band for CYP11A1, which is supported by
immunofluorescence results using the same antibody that show mitochondrial localization of the
signal (Germelli et al., 2021). Nevertheless, the ability for microglia to produce detectable
amounts of pregnenolone suggests the presence of CYP11A1, even though expression was not
directly measured in most of these studies.

38
Unfortunately, there is an equal amount of evidence that argues against the presence of
CYP11A1 in microglia. In FACS-sorted microglia from both adult rat brains and primary
microglia cultures from P2 rats, RT-PCR analyses could not detect expression of CYP11A1 or
other steroidogenic enzymes, including CYP17A1, aromatase, or 3β-hydroxysteroid
dehydrogenases (Gottfried-Blackmore, Sierra, Jellinck, McEwen, & Bulloch, 2009).
Furthermore, treatment with lipopolysaccharide and/or interferon γ to induce inflammatory
conditions did not lead to any changes in CYP11A1 expression nor the steroid secretion profile.
Single-cell RNA-seq of the whole cortex and hippocampus of mice showed that CYP11A1
expression appears to be higher in neuronal populations than in glia. While oligodendrocyte,
oligodendrocyte precursor cell, and astrocyte clusters consist of a very small number of
CYP11A1-positive cells, no positive cells can be found in microglia clusters [dataset](Yao Z. et
al., Whole Cortex & Hippocampus – 10X Genomics with 2010X-SMART-SEQ Taxonomy, Allen
Brain Institute of Science, 2020). Primary human microglia isolated from temporal lobe brain
tissue also did not express CYP11A1 (Owen et al., 2017).
Table 2 summarizes the findings for the presence or absence of CYP11A1 in neurons,
astrocytes, oligodendrocytes, and microglia. CYP11A1 expression has been found in all four cell
types, but not all studies agree on the presence or absence in a particular cell type. The
contradictory results may be due to differences in the brain regions studied, as well as the low
percentage of each cell type that express CYP11A1 within the brain.

39
Table 2: Evidence for CYP11A1 in different brain cell types

“+” represents a positive finding and “-“ represents a negative finding for the presence of
CYP11A1 protein (P), mRNA (R), or activity (A) in a particular cell type

Reference Species Neuron Astrocyte Oligodendrocyte Microglia
Hu et al. 1987 Rat + A
Jung-Testas et al. 1989 Rat - P + P, A
Iwahashi et al. 1990 Rat - P
Mellon and Deschepper
1993
Rat + R, P
Sanne and Krueger 1995 Rat + R + R + R
Compagnone et al. 1995 Rat + P - P
Furukawa et al. 1998 Rat + R
Ukena et al. 1998 Rat + P - P
Zwain and Yen 1999 Rat + R, A + R, A + R, A
Kimoto et al. 2001 Rat + P - P - P
Patte-Mensah et al. 2003 Rat + P + P + P
Karri et al. 2007 Rat + P, A
Biagini et al. 2009 Rat + P + P + P
King et al. 2002 Mouse + P + P
Apaja et al. 2004 Mouse + P
Gottfried-Blackmore et al.
2009
Mouse - R
Avallone et al. 2020 + A
Whole Cortex &
Hippocampus – 10X
Genomics (2020), Allen
Institute for Brain Science
Mouse + R + R + R - R
Brown et al. 2000 Human - R, A + R, P
- A
+ R, P, A
King et al. 2002 Human + P
Owen et al. 2017 Human - R
Milenkovic et al. 2019 Human + A
Multiple Cortical Areas –
SMART-SEQ (2019),
Allen Institute for Brain
Science
Human + R + R + R - R
Germelli et al. 2021 + P + P, A

40
1.3.5 Transcription regulation
Due to the low levels of CYP11A1 mRNA in brain, it has been suggested that this gene is
transcriptionally regulated differently than in the periphery (P. Zhang, Rodriguez, & Mellon,
1995). In peripheral steroidogenic organs, CYP11A1 expression is regulated by interactions
between sequence-specific nuclear binding proteins and DNA sequences located a few hundred
bases from the transcription initiation site. The 5’ flanking sequence for CYP11A1 contains
bindings sites for steroidogenic factor 1 (SF-1), cAMP response element-binding protein
(CREB), activator protein 1 (AP-1), and transcription factor Sp1, as well as a TATA box (Chung,
Guo, & Chou, 1997). In particular, SF-1 is considered an important regulator of steroidogenic
gene expression and is essential for hormonal stimulation of CYP11A1 expression in the
adrenals and gonads (M. C. Hu, Hsu, Pai, Wang, & Chung, 2001). However, SF-1 expression is
less clear for the nervous system. In rats, early studies reported an absence of SF-1 expression in
C6 glioma cells and in both CNS and PNS of embryos (Compagnone et al., 1995; P. Zhang et al.,
1995). Very weak signals were observed for SF-1 mRNA from parts of the rat cerebrum in
Southern blots (Stromstedt & Waterman, 1995). Currently, the consensus appears to be that SF-1
is exclusively expressed in the hypothalamic region within in the brain, although its expression
has been associated with functions such as neuronal development and energy homeostasis rather
than steroidogenesis (Choi, Fujikawa, Lee, Reuter, & Kim, 2013; K. W. Kim et al., 2011; Roselli,
Jorgensen, Doyle, & Rønnekleiv, 1997). The fact that SF-1 expression is not found in brain
regions believed to have higher CYP11A1 levels (such as the cerebellum and cortex) suggests
that expression of CYP11A1 is not dependent on SF-1 as in the periphery.
Using various reporter genes and cloned mutants, researchers have begun to tease out the
differences in transcriptional regulation of CYP11A1 between the brain and periphery.

41
Constructs with rat Cyp11a1 5’ flanking DNA were able to activate transcription of reporter
genes in Y-1 mouse adrenal cells, MA-10 mouse Leydig cells, and C6 rat glioma cells but not in
GC rat pituitary cells or GT1 mouse neurosecretory cells (P. Zhang et al., 1995). In MA-10 and
Y-1 cells, the minimum length of 5’ flanking DNA needed for reporter gene activity was 94 base
pairs. Although this construct could activate both basal and cAMP-induced transcription in both
cell lines, it did not yield any reporter gene activity in C6 cells. The nuclear protein that binds to
this sequence was proposed to be SF-1. The DNA sequence between -94 and -130 is important in
activating both basal and cAMP-induced transcription in C6 cells but could only mediate basal
transcription in MA-10 cells. These results suggest that different nuclear factors are involved in
the transcription of Cyp11a1 in glial cells than peripheral steroidogenic cells.
The proteins that bind to the -130/-94 region of the rat Cyp11a1 gene include the Ku
complex proteins p70 and p86, as well as the transcription factors Sp1 and Sp4 (Hammer,
Compagnone, Vigne, Bair, & Mellon, 2004). Immunohistochemical analysis of mouse brains
showed co-localization between the two Ku complex proteins and CYP11A1 in the olfactory
bulb, Purkinje cell layer of the cerebellum, medulla, and cortex. Using various mutant constructs
of the Cyp11a1 -130/-94 region, nucleotides at -120/-118 were important for protein-DNA
interactions that could drive transcription. However, co-transfecting C6 cells with the Cyp11a1 5’
flanking DNA constructs in addition to Ku p70 and p86 did not lead to any changes in transgene
activity compared to transfection with the constructs alone, making the importance of p70 and
p86 for CYP11A1 transcription inconclusive. The authors also found co-localization between
CYP11A1 and the transcription factors Sp1 and Sp4 during embryogenesis in rats. Sp1 was able
to activate transcription in cells transfected with constructs containing 130 bp of the 5’ flanking
DNA from rat Cyp11a1. The transcriptional activation was further enhanced by co-transfection

42
of both Sp1 and Ku, suggesting that Sp1 may have a direct effect on Cyp11a1 transcription, while
Ku synergistically enhances the effects of Sp1. However, further studies are needed to elucidate
the details of the interactions.
The human CYP11A1 promotor has also been investigated for its ability to drive
transgene expression. Although the human CYP11A1 promotor could yield high transgene
expression in the adrenal cortex, testes, and ovaries of transgenic mice, it did not consistently
activate transgene expression in brain (Wu et al., 2007). Transgene activity was also not detected
in areas where CYP11A1 expression had been consistently reported such as the cerebral cortex
and cerebellum. On the contrary, transgene activity was found in areas with little reported
endogenous CYP11A1 expression such as the arcuate nucleus and the dorsomedial part of the
ventromedial hypothalamic nucleus. A follow-up study by the same group used an improved
transgene construct that yielded transgene activity in most brain areas, including the
diencephalon and midbrain (Chiang et al., 2011). However, again no activity was found in the
cerebellum, suggesting that different brain regions could have unique regulatory mechanisms for
CYP11A1 expression.
Overall, it is clear that the transcriptional regulation of CYP11A1 differs between the
brain and periphery. However, the proteins and DNA sequences involved in transcription
activation in brain, as well as the factors that contribute to the low expression there, are still
unclear.

43
1.3.6 Discussion
1.3.6.1 Limitations in Current Experimental Data
The low concentration of CYP11A1 in brain has made it difficult for researchers to
provide conclusive evidence for the regional localization and expression level of this enzyme.
For example, in Sanne and Krueger’s study, although in situ hybridization showed wide-spread
distribution of Cyp11a1 mRNA in the rat brain, the signals were very faint and required an 8-day
exposure time to detect (Sanne & Krueger, 1995). Similarly, the in situ hybridization assay by
Furukawa et al. required more than a 1-month exposure time to reveal any signals for CYP11A1
in the rat cerebellum (Furukawa et al., 1998). Older studies utilizing RNAse protection assays
could not detect any Cyp11a1 mRNA in brain tissue or cells (Compagnone et al., 1995;
Furukawa et al., 1998; Mellon & Deschepper, 1993), which highlights the need for very sensitive
techniques. With the more sensitive RT-PCR technique, an increasing number of studies have
found signals for Cyp11a1 mRNA in different brain regions. However, there appears to be no
consensus on where this enzyme is most highly expressed. As seen from Table 1, the cerebellum
and cerebral cortex appears to be the brain areas most consistently found to express CYP11A1,
but studies have also reported no expression or the lowest expression in these two areas
(Iwahashi et al., 1990; Mellon & Deschepper, 1993; Yu et al., 2002). Quantitative PCR could not
give accurate data on the differences in amounts of mRNA between major brain areas, as the
Cyp11a1 mRNA levels were approximately 5000-fold lower than the testes and adrenals (Sanne
& Krueger, 1995). Therefore, more sensitive in situ hybridization and quantitative PCR
techniques are needed to accurately determine CYP11A1 mRNA expression differences between
brain areas.

44
Another major limitation of the current evidence for the presence of CYP11A1 protein in
brain is the questionable specificity of CYP11A1 antibodies used for immunohistochemistry and
immunoblot studies. Almost all data indicating CYP11A1 protein expression in brain came from
immunohistochemistry, but for a low expression protein such as CYP11A1,
immunohistochemical data is not sufficient proof for its presence in the CNS (Warner &
Gustafsson, 1995). A common problem with immunohistochemical staining is the amount of
cross reaction with unrelated proteins (Warner & Gustafsson, 1999). Although most studies use
antigen pre-absorption to enhance specific binding, pre-absorption does not necessarily assure
that the correct protein was recognized by the antibody in the tissue section.
Immunohistochemistry staining should be correlated with quantification using immunoblot
analysis or measurements of catalytic activity, but as seen from Table 2, almost all studies have
investigated activity or staining separately. In fact, some studies have detected CYP11A1 protein
but failed to observe catalytic activity in the same brain sample (Le Goascogne et al., 1987;
Warner & Gustafsson, 1995). Another problem with CYP11A1 antibodies is that they can result
in bands of incorrect sizes in immunoblot analysis of brain homogenates (Warner & Gustafsson,
1995). Based on our experience, many commercial antibodies are able to reliably detect
CYP11A1 in adrenal cells and tissue; however, the same antibodies used on brain samples often
lead to non-specific bands. Since CYP11A1 appears to be expressed in a very small percentage
of cells in the brain, it is highly likely that the strong bands observed for CYP11A1 on
immunoblots for brain homogenates are due to non-specific binding.
Given the low concentration of CYP11A1 in brain and the problems with antibodies,
more sensitive and specific techniques are needed to confirm the results from
immunohistochemical studies. Methods such as mass spectrometry that are less dependent on

45
antibody specificity may yield more reliable results that would establish the presence and
localization of CYP11A1 protein in brain.

1.3.6.2 Limitations of Animal Studies
As illustrated in Figure 4, there appears to be many differences in CYP11A1 expression
patterns between rats, mice, and humans. In rodents, the cerebellum has consistently been shown
to express CYP11A1 and tends to be one of the regions with highest expression. On the other
hand, in humans the cerebellum appears to have lower mRNA expression than other areas such
as the hippocampus (MacKenzie et al., 2008; Yu et al., 2002). Although the rat is the most
studied species in terms of steroidogenesis in the CNS, it may not be the most ideal animal
model. Only one CYP11A1 isoform has been observed for rats, while both mice and humans
have two isoforms. To our knowledge, there have been no studies comparing the differences in
function, subcellular localization, tissue expression levels, and/or transcriptional regulation
between the isoforms. It is unclear whether primers or antibodies used to detect CYP11A1
expression have been designed to differentiate between the two isoforms. Therefore, it is highly
likely that observations from CYP11A1 catalytic activity and expression studies in brain are the
combined results from both isoforms, even though researchers intended to study only the
classical isoform a. Future studies would need to consider the role of both CYP11A1 isoforms,
which could make the mouse arguably a better model than the rat for representing neurosteroid
synthesis in humans.
The difficulties in accessing good-quality human brain tissue limit the amount of human
data that can be generated regarding CYP11A1 in the CNS. As most human samples come from
excised surgical tissue, many studies are not able to compare expression levels directly between

46
brain regions using the same conditions. Experimental protocols and interindividual variations
make it difficult to accurately determine the amount of CYP11A1 mRNA and protein in human
brain, especially because the concentration is so low. Given these circumstances, animal studies
are very valuable, but the species should be carefully selected in order to more accurately
represent human CYP11A1 expression, activity, and transcriptional regulation.

1.4 Summary
It is clear that CYP11A1 mRNA is expressed at very low levels in brain of multiple
mammalian species; however, the problem lies in the detection of CYP11A1 protein. The
reliability of immunohistochemical data is limited by questionable antibody specificity and lack
of corresponding catalytic activity. Better understanding of neurosteroidogenic pathways has
become increasingly important with the approval of the first neurosteroid-based treatment, i.e.
brexanolone. Given the apparent discrepancy between the levels of CYP11A1 and pregnenolone
concentrations found in brain, we hypothesize that other pathways could be used to synthesize
pregnenolone in the CNS. With the emerging therapeutic potential of neurosteroids as
neuroprotective and neuromodulating molecules in conditions like neurodegenerative diseases
and mood disorders, understanding neurosteroidogenesis can lead to therapies that alter
endogenous neurosteroid production. Since systemic administration of steroid-modulating drugs
can also alter peripheral steroidogenesis to cause adverse effects and local administration would
be highly invasive, modulation of CNS-specific steroid synthesis pathways might be a promising
avenue for investigation.

47
II. Evidence for a CYP11A1-Independent Pathway for Pregnenolone
Synthesis in the Human Brain

This section is adapted from:
Lin, Y . C., Cheung, G., Porter, E., & Papadopoulos, V . The neurosteroid pregnenolone is
synthesized by a mitochondrial P450 enzyme other than CYP11A1 in human glial cells.
Journal of Biological Chemistry, 102110. doi:10.1016/j.jbc.2022.102110 (In Press)

2.0 Abstract
Neurosteroids, modulators of neuronal and glial cell functions, are synthesized by the
nervous system from cholesterol. In peripheral steroidogenic tissues, cholesterol is converted to
the major steroid precursor pregnenolone through activity of the CYP11A1 enzyme, which
catalyzes two hydroxylations followed by cleavage of the cholesterol side chain. Although
pregnenolone is one of the most abundant neurosteroids found in the brain, expression of
CYP11A1 has been difficult to detect in the human brain. We found that the expression of
CYP11A1 in the human brain is more than 1000-times lower than that in the adrenals, with less
than 1% of cells in the cerebral cortex and cerebellum expressing CYP11A1 mRNA. Therefore,
we investigated the ability of four human glial cell lines to synthesize pregnenolone in a
CYP11A1-independent manner. These cells produced pregnenolone, detectable by mass
spectrometry and ELISA, despite no observable CYP11A1 protein by immunoblot analysis.
Unlike in testicular and adrenal cortical cells, pregnenolone production in glial cells were not
inhibited by treatment with the CYP11A1 inhibitor DL-aminoglutethimide or the non-specific
CYP450 inhibitor ketoconazole. Pregnenolone synthesis was increased with addition of the
substrates 22R-, 22S-, and 20α-hydroxycholesterols, suggesting the involvement of a desmolase
activity, which was neither blocked by DL-aminoglutethimide nor ketoconazole.

48
2.1 Introduction
Neurosteroids are steroid hormones synthesized in the central or peripheral nervous
system, either de novo from cholesterol or metabolically in situ from precursors in the blood
(Baulieu & Robel, 1990). These neurosteroids—such as pregnenolone, dehydroepiandrosterone
(DHEA), and their sulfates—can accumulate in the nervous system independently of peripheral
gland secretion and have important roles in modulating neuronal functions and behavior.
Changes in neurosteroid levels have been implicated in neurological and psychiatric disorders,
such as Alzheimer’s Disease and mood disorders respectively (Ratner et al., 2019). Levels of
pregnenolone and DHEA can be elevated in brains of Alzheimer’s Disease patients compared to
cognitively intact control subjects (Marx, Trost, et al., 2006; Naylor et al., 2010), while levels of
the downstream neurosteroids allopregnanolone, pregnenolone sulfate, and DHEA sulfate were
found to be lower in brains of Alzheimer’s Disease patients (Marx, Trost, et al., 2006; Naylor et
al., 2010; Weill-Engerer et al., 2002). In mood disorders such as depression and bipolar disorder,
neurosteroid levels in the brain have generally been found to be reduced, and this effect can be
reversed with antidepressant or lithium treatment (Carta, Bhat, & Preti, 2012). With the approval
of brexanolone, an allopregnanolone-based drug, for treatment of post-partum depression
(Walton & Maguire, 2019), there is increasing interest in pharmacological development of
neurosteroids to treat neurological disorders.
In the brain, neurosteroids exert their effects not only through classical genomic
mechanisms but also through modulation of neurotransmitter receptors, such as GABAA and
glutamate receptors. For example, allopregnanolone is a potent positive allosteric modulator of
GABAergic neurotransmission and was found to have anxiolytic effects (Akwa, Purdy, Koob, &
Britton, 1999; Bitran, Shiekh, & McLeod, 1995). Neurosteroids such as DHEA and

49
pregnenolone sulfate can modulate activity of N-methyl-D-aspartate (NMDA) receptors and
have been proposed to play important roles in neuronal plasticity and neuronal cell
differentiation (Bowlby, 1993; Compagnone & Mellon, 1998; Smith, Gibbs, & Farb, 2014).
Pregnenolone not only serves as the precursor to all other neurosteroids, but also has
physiological and pharmacological effects on its own (for review, see (Lin, Cheung, Espinoza, &
Papadopoulos, 2021)). For example, administration of pregnenolone can enhance memory in
mice (Flood et al., 1992; Melchior & Ritzmann, 1996). In humans, administration of
pregnenolone was shown to improve symptoms in patients with bipolar disorder and
schizophrenia (E. S. Brown et al., 2014; Marx et al., 2009), suggesting that this steroid has a
neuromodulatory role in the brain. Pregnenolone may also be involved in modulating neuronal
shape and plasticity by stimulating microtubule assembly (Fontaine-Lenoir et al., 2006;
Murakami et al., 2000). Furthermore, pregnenolone can protect neurons from glutamate- and
amyloid beta-induced toxicity (Gursoy et al., 2001). In general, neurosteroids are believed to
reduce neuroinflammation and be neuroprotective (Borowicz, Piskorska, Banach, & Czuczwar,
2011; Yilmaz et al., 2019). The overall effects of neurosteroids on the brain depends on the
extent that they are metabolized, and how the parent steroid and its metabolite(s) affect
intracellular or extracellular receptors (Ratner et al., 2019). Therefore, understanding the
metabolic pathway of neurosteroids and how the major neurosteroid precursor pregnenolone is
synthesized should allow better understanding of how neurosteroid-modulating drugs may affect
brain function.
Steroid synthesis involves activity of multiple enzymes, most of which are highly
expressed in the adrenals and gonads but have also been shown to be present in the brain (Mellon
& Griffin, 2002; Stoffel-Wagner, 2001). In the classical steroidogenic pathway, de novo synthesis

50
of steroids requires conversion of cholesterol to pregnenolone by the cytochrome P450 11A1
enzyme (CYP11A1). Cholesterol is first transported into the mitochondria, mediated by
translocator protein 18 kDa (TSPO) and steroid acute regulatory protein (STAR) (Papadopoulos,
Lecanu, Brown, Han, & Yao, 2006). In the inner mitochondrial membrane, CYP11A1 performs
two successive hydroxylations on C22 and C20 of cholesterol, forming the intermediates 22(R)-
hydroxycholesterol and 20α,22(R)-dihydroxycholesterol, followed by cleavage of the cholesterol
side chain between C20 and C22 to form pregnenolone (Payne & Hales, 2004). This enzyme
activity requires one molecule of oxygen and one molecule of NADPH for each reaction. Unlike
CYP450 enzymes in the endoplasmic reticulum that use NADPH-cytochrome P450 reductase
(POR) as a redox partner, mitochondrial CYP11A1 activity involves two cofactors ferredoxin
(FDX1) and ferredoxin reductase (FDXR), also known as adrenodoxin and adrenodoxin
reductase, respectively, similar to other mitochondrial CYP450s (Nebert, Wikvall, & Miller,
2013). CYP11A1 uses electrons from the electron transport chain to carry out reactions,
delivered through NADPH (Chien, Rosal, & Chung, 2017). The electron on NADPH is
transferred to ferredoxin reductase, which subsequently donates it to ferredoxin. Ferredoxin then
diffuses into the mitochondrial matrix to shuttle the electron to the heme group on CYP11A1,
allowing the enzyme to carry out its side chain cleavage activity.
Even though pregnenolone is the most abundant steroid in the brain (Weill-Engerer et al.,
2002), CYP11A1 expression has been difficult to detect, particularly in the human brain. Studies
have reported the presence of Cyp11a1 mRNA in rat brains at approximately 0.01% of adrenal
Cyp11a1 mRNA levels (Mellon & Deschepper, 1993). CYP11A1 appears to be more highly
expressed in oligodendrocytes in the rat brain compared to astrocytes and neurons, which is
supported by higher pregnenolone production in oligodendrocytes (Zwain & Yen, 1999). This led

51
to the hypothesis that oligodendrocytes produce the majority of steroid precursors, which are
metabolized to other neurosteroids by astrocytes and neurons. In human brains, CYP11A1 mRNA
was reported to be present in the particular brain regions such as the hippocampus, but levels
range from 200 to 10000 times lower than that of the adrenals (Beyenburg et al., 1999; Yu et al.,
2002). Despite multiple reports of CYP11A1 mRNA presence in the human brain and of Cyp11a1
protein in rat brains, protein expression of this enzyme in the human brain has been inconclusive
(see (Lin & Papadopoulos, 2021) for review). To our knowledge, there has only been two reports
of CYP11A1 protein in the human brain, but the presence was only demonstrated in small areas
of the cerebellum and frontal cortex in a few number of cells (King et al., 2002; Le Goascogne et
al., 1989). The levels of pregnenolone found in the brain and the ability for the brain to produce
increased amounts of pregnenolone in response to treatment with TSPO ligands (Lin et al., 2021;
Rupprecht et al., 2010) creates a discrepancy with the local CYP11A1 levels, suggesting that
alternative pathways may be used to produce pregnenolone in the brain (Lin & Papadopoulos,
2021). To examine this discrepancy, we determined the expression of CYP11A1 in four human
glial cell lines. We found that CYP11A1 protein can only be detected by immunocytochemistry
with very weak signals and, thus, investigated whether a CYP11A1-independent pathway is used
by brain cells to produce pregnenolone.

52
2.2 Materials & Methods
Cell Culture
The human glioma cell lines MGM-1 and MGM-3 were a gift from Dr. Hiroaki Kataoka
(University of Miyazaki, Japan). MGM-1 and MGM-3 cells were grown in Dulbecco’s modified
Eagle medium (DMEM; Gibco, Thermo Fischer Scientific #11965092) with 10% heat-
inactivated fetal bovine serum (FBS, Sigma #12306C) plus 100 IU/mL penicillin and 100 μg/mL
streptomycin (Gibco, Thermo Fischer Scientific #15140122) at 37◦C and 5% CO2. The microglia
cell line HMC3 (#CRL-3304) and the adrenal cortical carcinoma cell line NCI-H295R (referred
to as H295R-S1; #CRL-2128) were purchased from ATCC. HMC3 cells were grown in
DMEM/F-12, GlutaMAX (Gibco, Thermo Fischer Scientific #10565042) with 10% FBS, 100
IU/mL penicillin and 100 μg/mL streptomycin at 37◦C and 5% CO2. H295R-S1 cells were grown
in DMEM/F-12, GlutaMAX with 2.5% NuSerum (Corning, #355100), 100 IU/mL penicillin, 100
μg/mL streptomycin, plus ITS+ Premix Universal Culture Supplement at a concentration
recommended by the manufacturer (Corning, #354352). Normal human astrocytes (NHA) were
purchased from Lonza (#CC-2565). NHA cells were grown at 37°C and 5% CO2 in Astrocyte
Growth Medium BulletKit (Lonza, #CC-3186): Astrocyte Basal Medium plus rhEGF, insulin,
ascorbic acid, GA-1000, L-glutamine, and 3% FBS. MA-10 cells were a gift from Dr. Mario
Ascoli (University of Iowa, USA). MA-10 cells were grown in DMEM/F12, GlutaMAX with 5%
FBS, 2.5% horse serum (Gibco, Thermo Fischer Scientific #26050088), 100 IU/mL penicillin
and 100 μg/mL streptomycin at 37◦C and 3.5% CO2. NHA cells were passaged using
ReagentPack Subculture Reagents (Lonza, #CC-5034) and all other cell lines were passaged
using trypsin/EDTA (Gibco, Thermo Fischer Scientific #25200056). A maximum of 10 passages
were used for all cell lines.

53
Cell Treatments for Steroid Measurements
FGIN-1-27 (#18461) and 22(S)-hydroxycholesterol (#21399) were purchased from
Cayman Chemicals (Ann Harbor, MI, USA). Other compounds used to treat cells were
purchased from Sigma: XBD173 (#SML1223), DL-aminoglutethimide (#A9657), ketoconazole
(#K1003), 22(R)-hydroxycholesterol (#H9384), 20α-hydroxycholesterol (#H6378).
For steroid synthesis experiments, cells were seeded in 24 or 96 well plates and grown to
70 - 80% confluency. Before treatment, cells were washed three times with PBS to remove
steroids from the serum in complete media. Drugs used for treating cells were dissolved in
DMSO, PBS, or ethanol depending on solubility to make stock solutions with at least 200X
concentration. Stock solutions were diluted in serum-free base media to treatment concentrations
immediately before treatment. Appropriate solvent controls were made based on concentration of
solvent in the highest concentration treatment condition. Cells were incubated with treatment
media for 2 hours, after which the media was collected and stored at -80°C until used for steroid
measurement. Cells were then lysed with 0.1M NaOH solution and protein quantity was
measured using the Bradford assay (VWR, #97065-020) to normalize steroid measurements.
The levels of major steroids produced by cells were measured by performing ELISA on
collected media. Pregnenolone was measured in media from MGM-1, MGM-3, NHA, HMC3,
and H295R-S1 cells using pregnenolone ELISA kits (Abnova, Taipei, Taiwan; #KA1912).
Progesterone was measured in media from MA-10 cells using progesterone ELISA kits (Cayman
Chemical; #582601). ELISA assays were performed according to protocols provided by
manufacturers.

54
RNA Extraction and qRT-PCR
Total RNA was extracted from cell pellets consisting of 10
6
cells and DNAse treatment
was performed to remove genomic DNA using the RNAqueous-Micro Kit (Thermo Fischer
Scientific, #AM1931). Human CNS tissue RNA for total brain (#636530), cerebellum
(#636535), cerebral cortex (#636561), parietal lobe (#636571), occipital pole (#636570),
temporal lobe (#636564), and spinal cord (#636554) were purchased from Takara Bio. RNA
(1000 ng) were reverse transcribed into cDNA using PrimeScript RT Master Mix (Takara,
#RR036B). qRT-PCR was then performed in 384-well plates using SYBR Select Master Mix
(Thermo Fischer Scientific, #4472908) with 100 nM forward and reverse primers (Integrated
DNA Technologies). Primers were designed using the NCBI Primer Blast tool, where primers
pairs that spanned exon-exon junctions were preferentially selected. Plates were assayed on a
CFX384 Touch Real-Time PCR Detection System (Bio-Rad). Sequences of all forward and
reverse primers are listed in Table 3. For each gene, 20 ng cDNA was used for detection. Data
were analyzed using Bio-Rad CFX Maestro software.

55
Table 3: Sequences of primers used for qRT-PCR
Gene Forward Primer Reverse Primer
CYP11A1 GCTTTGCCTTTGAGTCCATCA CTCGGGGTTCACTACTTCCTC
FDX1 TTCAACCTGTCACCTCATCTTTG TGCCAGATCGAGCATGTCATT
FDXR CTGAGGCAGAGTCGAGTGAAG CCCGAAGCTCCTTAATGGTGA
TSPO GCCATACGCAGTAGTTGAGTG CCTGCTCTACCCCTACCTGG
STAR ACAGACTTCGGGAACATGC TGAGTAGCCACGTAAGTTTGG
CYP17A1 GTTGTTGGACGCGATGTCTA TTCGTATGGGCACCAAGACT
CYP19A1 TGGAAATGCTGAACCCGATAC AATTCCCATGCAGTAGCCAGG
CYP21B AAGGACAGGTCCGGGTAGTT CCAAGAGGACCATTGAGGAA
CYP11B1 GAGGCCTGAGCGCTATAACC TGGAGGTGTTTCAGCACATGG
CYP11B2 TTCAACCGCCCTCAACACTAC GGAAACGCTGTCGTGTCCA
HSD3B1 CACATGGCCCGCTCCATAC GTGCCGCCGTTTTTCAGATTC
HSD3B2 AGAACGGCCACGAAGAAGAG TGGGTCTTAACGCACAAGTGT
HSD17B1 ACGTGAATGTAGTAGGGACTGT GCGCAATAAACGTCATTGAAAGG
HSD17B2 AGCTGAGGAATTGCGAAGAAC GCAACCTTGCTGTAAGCATCT
HSD17B3 CTGGCGAAGTGCGTGAGATT GAGTACGCTTTCCCAATTCCAT
HSD17B4 TGAGGGATCGTTCCTTTGCTA CGTGTCACTTGGAATGAACCC
AKR1C1 TAGGCAACTGTGTCATGGTGG AAAGGCAGCGAAGGATTCAGA
AKR1C2 GTTGCCAGCTCATTGCTCTT CCAGGACAGGCATGAAGTGA
AKR1C3 GTCATCCGTATTTCAACCGGAG CCACCCATCGTTTGTCTCGTT
AKR1C4 AGGTGAGACGCCACTACCAA GTTTGCTCTGGTTGAGGTAAGG
SRD5A1 TCAGACGAACTCAGTGTACGG CGTAGTGGACGAGGAACATGG
SRD5A2 ACTGCTCAATCGAGGGAGG CACCCAAGCTAAACCGTATGTC
STS CCTACTGTTCTTTCTGTGGGAAG CGAGGTCGTCAGCCATCAC
SULT2A1 CGTGATGAGTTCGTGATAAGGG GGCAGAGAATCTCAGCCAACC
SULT2B1 GTTGCCAGGTGAATACTTCCG CCCGCACATCTTGGGTGTT
HSD11B1 GCCTGCTTAGGAGGTTGTAG CCTTGGAGCATCTCTGGTCT
HSD11B2 GGACCTGACCAAACCAGGAG TTCACCTCCATGCAGCTACG
TUBA1 TCGATATTGAGCGTCCAACCT CAAAGGCACGTTTGGCATACA
ACTB CCTTGCACATGCCGGAG GCACAGAGCCTCGCCTT
POR TTTCGCTCATCGTGGGTCTC CGATGATGTTCCTCCCCGTT

Cell Lysate Preparation and Western Blotting
Cell pellets consisting of 2 x 10
6
cells were lysed in RIPA buffer with 2% protease
inhibitor (Thermo Fischer Scientific, #A32955). 15 μg of total proteins were resolved on 4 to
20% precast polyacrylamide gels (Bio-Rad, #4561096) and transferred to PVDF membranes
(Sigma, #ISEQ00010). After transfer, membranes were blocked for 30 minutes with blocking

56
solution consisting of 5% BSA (Equitech-Bio, #BAH65) dissolved in PBST. Membranes were
then incubated with specific primary antibodies diluted in blocking solution overnight at 4°C,
washed 3 times with PBST, incubated with corresponding secondary antibodies for 1 hour at RT,
and washed 3 times with PBST. Antibodies were detected using Clarity Western ECL Substrate
system (BioRad, #1705061), which allows detection of signals for > 0.3 ng protein recognized
by antibodies, and visualized using a western blot imaging system (Azure Biosystems, c600).
The following primary antibodies were used: anti-CYP11A1 rabbit mAb (Cell Signaling
Technology, #14217, 1:1000 dilution, antigen: human CYP11A1 N-terminus), anti-CYP11A1
rabbit pAb (Proteintech, # 13363-1-AP, 1:1000 dilution, antigen: CYP11A1 aa 1-300), anti-
CYP11A1 rabbit pAb (Abcam, #ab75497, 1:500 dilution, antigen: human CYP11A1 aa 288 –
337), anti-CYP11A1 rabbit pAb (Abcam, #ab232763, 1:1000 dilution, antigen: CYP11A1 aa 350
– 521), anti-GAPDH rabbit mAb (Cell Signaling Technology, #2118, 1:2000 dilution), anti-
FDXR rabbit pAb (Proteintech, #15584-1-AP, 1:1000 dilution), anti-ADX rabbit mAb (Abcam,
#ab108257, 1:1000 dilution), anti-beta actin mouse mAb (Abcam, #ab8226, 1:3000 dilution).

Immunocytochemistry
Cells were seeded onto glass coverslips in 12 well plates and grown to 50% confluency.
For immunocytochemistry, cells were incubated with 250 nM MitoTracker™ Red CMXRos
(Thermo Fischer Scientific, #M7512) in complete media for 40 minutes. Then, cells were fixed
in 4% paraformaldehyde for 10 minutes and permeabilized with 0.1% Triton X for 10 minutes,
with three PBS washes between each step. Blocking was then performed for 30 minutes with 5%
donkey serum (Sigma, #D9663) + 0.5% BSA (Equitech-Bio, #BAH65). Cells were incubated
overnight at 4°C with anti-CYP11A1 rabbit pAb (Proteintech, #13363-1-AP, 1:400 dilution in

57
PBS). Following primary antibody incubation, cells were washed three times with PBS and
incubated with Alexa Fluor™ 488 goat anti-rabbit IgG (H+L) secondary antibody (Thermo
Fischer Scientific, #A11008, 1:1000 dilution in PBS) for 1 hour at room temperature. After three
PBS washes, the coverslips were mounted onto microscope slides with Vectashield® Vibrance
mounting medium with DAPI (Vector Laboratories, #H-1800). The slides were then imaged at
63x magnification using Zeiss LSM 880 with Airyscan Confocal Microscope.

RNAscope in situ Hybridization
Formalin-fixed paraffin-embedded cerebellum and cerebral cortex tissue slices from a 63-
year-old male, as well as cerebral cortex tissue slices from a 92-year-old female, were obtained
from ACD Bio. Both donors were free from any neurological disorders. Two probes were used
for the duplex ISH assay: one custom probe targeting exon 2 – 8 of CYP11A1 and a second
probe targeting MBP from the ACD Bio catalogue (#411058). The ISH assay was performed and
images were taken by ACD Bio according to previously published procedures (F. Wang et al.,
2012).

Shotgun Proteomics Mass Spectrometry
H295R-S1, MGM-1 and NHA cells were pelleted by centrifuging at 200 x g after three
PBS washes. Pellets consisted of 5 x 10
6
cells and were lysed using RIPA buffer with 2%
protease inhibitor (Thermo Fischer Scientific, #A32955). Eighty μg of total proteins were
resolved on 12% acrylamide gels (Bio-Rad, #1610185). Protein bands were visualized by
staining with Coomassie blue (Bio-Rad, #1610786) for 1 hour. Bands at 40 kDa, 50 kDa, and 60
kDa were excised and shipped to Creative Proteomics (NY , USA) for shotgun proteomics. A

58
summary of the protocol was provided by Creative Proteomics. Briefly, the protein bands were
digested with trypsin and resuspended in 0.1% formic acid before LC-MS/MS analysis
performed on a Nanoflow UPLC system. A full scan was performed between 300 – 1650 m/z at
resolution of 60,000 at 200 m/z. Raw MS files were analyzed and searched against human
protein database using Maxquant (1.6.2.6) to obtain a list of matching proteins in each band.

Mass Spectrometry for Pregnenolone Detection
Steroids in the media were extracted twice with 1-chlorobutane (Sigma, #34958) using
media:solvent ratio of 1:2. The top organic layer was collected and dried using a rotary
evaporator. Pregnenolone was derivatized with 1-amino-4-methylpiperazine (AMP; Sigma,
255688) to increase detection sensitivity based on a published method (Ke, Gonthier, & Labrie,
2017). Briefly, 400 µL of 2 mg/mL AMP in methanol and 400 µL of 1% acetic acid in methanol
were added to dried samples and vortexed for 10 minutes before evaporation under a stream of
nitrogen for 35 minutes. Dried, derivatized samples were reconstituted in 200 µL of 50%
methanol + 0.1% formic acid solution, where 100 µL of sample solution was transferred to LC-
MS vials. Cell pellets were lysed in RIPA buffer with 2% protease inhibitor (Thermo Fischer
Scientific, #A32955), and protein concentration was quantified using the BCA assay to use for
normalization (Thermo Fischer Scientific, #23225).
To generate a calibration curve, 400 µL of 1 µg/mL pregnenolone standard (Sigma,
#P9129) were dried, derivatized as described above and reconstituted in 200 µL 50% methanol
for a final concentration of 2000 ng/mL pregnenolone-AMP. An internal standard (IS) stock
solution was prepared in the same way using d4-pregnenolone (Sigma, #809845). Final
concentrations for the calibration curve were prepared at 100, 50, 20, 10, 5, 2, 1, 0.5, and 0.25

59
ng/mL pregnenolone with 50 ng/mL IS in 50% methanol + 0.1% formic acid using serial
dilutions.
The LC–MS/MS measurements were performed with an Agilent Poroshell 120 EC-C18
column (2.7 µm, 2.1 x 150 mm) using Agilent Infinity II HPLC connected to a QTRAP 6500+
mass spectrometer (AB Sciex). Mobile phase A was 0.1% formic acid in H2O, while mobile
phase B was 0.1% formic acid in acetonitrile. MS detection was performed in positive mode.
MRM transitions used were: precursor ion (m/z = 414) à product ion (m/z = 99) for
pregnenolone-AMP and precursor ion (m/z = 418) à product ion (m/z = 99) for IS. MS
parameters were optimized to achieve optimal sensitivity: ion spray voltage was 4500 V , source
temperature was 300°C, curtain gas/ion gas 1/ion gas 2 were 20 psi, declustering potential was
79.5 V , collision energy was 26.3, and dwell time was 50 ms. The data were acquired and
analyzed using Analyst 1.7 software.

Mass Spectrometry for Detection of Multiple Steroids
H295R-S1 and MGM-1 cells were treated for 2 hours with 76 mM AMG or 100 µM
ketoconazole with or without 50 µM 22(R)-hydroxycholesterol stimulation in serum-free media.
One mL of supernatant was collected for each sample and lyophilized using FreeZone 1 Liter -
50C Freeze Dryers (Labconco, #7740021). Lyophilized samples were extracted with 200 µL of
methyl tert-butyl ether (MTBE) three times and the combined organic layers were dried using a
Speed Vac. Samples were reconstituted in 200 µL of 10+90 (v/v) mixture of dH
2
O:ethanol with
0.1% formic acid (aqueous) and transferred to HPLC vials. LC-MS/MS was performed by the
Proteomics Service, Research Institute of McGill University Health Centre (RIMUHC) on an AB
Sciex 5600 Time of Flight Mass Spectrometer connected to a Nexera XR LC-20 ADXR™

60
UHPLC system using a Zorbax™ (Agilent) Eclipse C18 (2.1 x 50 mm, 1.8 µm) column and
autosampler. Mobile phase A was 0.1% formic acid in water, and mobile phase B was 0.1%
formic acid in acetonitrile. MS parameters were as followed: ion spray voltage was 5500 V ,
source temperature was 500°C, curtain gas 30 L/min, ion source gas 1 and 2 were 50 L/min,
collision energy was 10, and declustering potential was 100 V . Product ions scanned were:
androstenedione (m/z = 287.2), testosterone (m/z = 289.2), DHEA (m/z = 289.2), progesterone
(m/z = 315.2), and pregnenolone ([M-18]+; m/z = 229.2). Analytical standards in solution were
used to establish linearity and dynamic range. Data acquisition and analysis were done using
MultiQuant 3.0.2 and Analyst TF 1.7 (AB Sciex).

Statistics
Statistical analyses of steroid measurements were performed using GraphPad Prism 7.
Statistical significance was determined using one-way ANOV A followed by Dunnett’s multiple
comparison test. For multiple comparison tests, each value was compared to the no-drug solvent
control group unless otherwise indicated.

61
2.3 Results
2.3.1 Evaluating the steroidogenic potential of glial cells
Previous studies suggested that oligodendrocytes produce more pregnenolone than
astrocytes and neurons (R. C. Brown et al., 2000; Zwain & Yen, 1999); thus, we used the human
glioma cell lines MGM-1 and MGM-3 as the main models for this study. We also used the
human astrocyte cell line NHA and human microglia cell line HMC3 as non-cancerous glial cell
models to confirm our findings. For comparison to peripheral steroidogenic cells, we used the
human adrenal cortical carcinoma cell line H295R-S1 and mouse Leydig tumor cell line MA-10.
The characteristics of each cell line used are summarized in Table 4.

Table 4: List of cell lines used
Cell Line Species Phenotype Source
MA-10 Mouse Leydig cell Leydig cell tumor
H295R-S1 Human Adrenocortical cell Adrenocortical carcinoma
MGM-1 Human Oligodendrocyte (MBP+) Glioblastoma
MGM-3 Human Oligodendrocyte (MBP+) Glioblastoma
NHA Human Astrocyte (GFAP+) Primary cells
HMC3 Human Microglia (Iba1+) Immortalized primary cells

MBP: myelin basic protein; GFAP: glial fibrillary acidic protein; Iba1: allograft inflammatory
factor 1

62
To determine whether human glial cells express the components of the machinery used
for peripheral steroid biosynthesis, we performed qRT-PCR to evaluate expression of genes
important for steroidogenesis. In general, enzymes in the classical steroidogenesis pathway
(Figure 5A) are expressed at very low levels in glial cells (Figure 5B, Figure 6B). Enzymes
further upstream in the steroidogenesis pathway are more highly expressed. Glial cells express
low but detectable amounts of mRNA for CYP11A1, CYP17A1, CYP19A1, 3β-hydroxysteroid
dehydrogenases (HSD3B1, HSD3B2), 17β-hydroxysteroid dehydrogenases (HSD17B1,
HSD17B2, HSD17B3, HSD17B4, AKR1C3), 3α-hydroxysteroid dehydrogenase (AKR1C1,
AKR1C2), 5α-reductase (SRD5A1, SRD5A2), steroid sulfatase (STS), and sulfotransferase
(SULT2B1). No expression for CYP11B1, CYP11B2, CYP21A2, AKR1C4, and SULT2A1 could
be detected in glial cells.

63

64
Figure 5: mRNA expression of steroidogenesis machinery in H295R-S1 and human glial cells.
(A) Schematic diagram of the steroidogenesis pathway with summary of steroidogenic enzyme
expression in human glial cells. Steroids are written in bold while the gene names of
steroidogenic enzymes responsible for producing the steroid are written in italics. Enzymes with
detectable mRNA expression in MGM-1, MGM-3, NHA, or HMC3 cells are written in black,
while enzymes without detectable expression in glial cells are written in red. The mRNA for all
the enzymes could be detected in H295R-S1 cells. (B) qRT-PCR analyses of genes important for
steroidogenesis in H295R-S1, MGM-1, MGM-3, NHA, and HMC3 cells. Gene expression is
shown as relative expression to α-tubulin. Data are presented as mean ± SD, N=3. Each data
point represents total RNA extracted from cells of a different passage for each cell line. Enzymes
involved in the steroidogenesis pathway include cytochrome P450 side-chain cleavage
(CYP11A1), cytochrome P450 17α-hydroxylase/C17,20-lyase (CYP17A1), aromatase
(CYP19A1), 21-hydroxylase (CYP21A2), 11β-hydroxylase (CYP11B1), aldosterone synthase
(CYP11B2), 3β-hydroxysteroid dehydrogenase (HSD3B1, HSD3B2), 17β-hydroxysteroid
dehydrogenase (HSD17B1, HSD17B2, HSD17B3, HSD17B4), 11β-hydroxysteroid
dehydrogenase (HSD11B1, HSD11B2), 5α-reductase (SRD5A1, SRD5A2), 3α-hydroxysteroid
dehydrogenase (AKR1C1, AKR1C2, AKR1C3, AKR1C4), steroid sulfatase (STS), and
sulfotransferase (SULT2A1, SULT2B1). Translocator protein (TSPO) and steroid acute
regulatory protein (STAR) mediate cholesterol transport into the mitochondria, which is the rate-
limiting step for classical steroid synthesis.

65
2.3.2 Evaluating CYP11A1 mRNA expression in the human brain
We determined the overall levels of CYP11A1 mRNA in various regions of the human
CNS using qRT-PCR. The highest expression of CYP11A1 was found in the spinal cord, followed
by the occipital pole and the parietal lobe (Figure 6B). The expression in the temporal lobe,
cerebral cortex, cerebellum, and total brain were approximately the same. CYP11A1 expression
in CNS tissue is more than 1000 times lower than in the adrenals and more than 100-times lower
than in the testes (Figure 6C).
To further investigate the tissue and cell type localization of CYP11A1 in the brain, we
performed in situ hybridization (ISH) analyses on human cerebellum and cerebral cortex tissue
slices using the RNAscope technology described by Wang et al. (F. Wang et al., 2012). A duplex
chromogenic ISH assay using probes targeting CYP11A1 or MBP was conducted. In the
cerebellum, expression of CYP11A1 was limited to a very small number of cells in the granule
layer and molecular layer (Figure 6D, E). Interestingly, we found that a larger proportion of
Purkinje cells had positive expression for CYP11A1 (Figure 6E). We could not observe the
presence of CYP11A1 mRNA in white matter of the cerebellum (Figure 6G). Similar to the
cerebellum, extremely low levels of CYP11A1 mRNA were seen in the cerebral cortex gray
matter with even lower expression in white matter (Figure 6F, Figure 7). While the majority of
the cortex white matter did not have detectable CYP11A1 mRNA (Figure 7A, C), a very scarce
number cells showed positive CYP11A1 signals in the cortex white matter of one donor (Figure
7D). We estimate that the proportion of CYP11A1-positive cells is less than 1% of cells in both
brain regions. Each CYP11A1-positive cell showed no more than 2 molecules of CYP11A1
mRNA in our ISH assay. There was also little co-localization between MBP and CYP11A1,

66
suggesting that oligodendrocytes and myelinated fibers are not likely to express the most
CYP11A1 in the cerebellum and cortex.

67

68
Figure 6: mRNA expression of CYP11A1 in human brain tissues.
(A) Schematic diagram describing the cholesterol side-chain cleavage reaction by CYP11A1.
qRT-PCR analyses of CYP11A1 in (B) total brain, cerebellum, cortex, parietal lobe, occipital
pole, temporal lobe, and spinal cord and (C) human testes, adrenal glands, brain, and spinal cord.
Data for brain and spinal cord in (C) are the same as the total brain and spinal cord data in (B)
respectively. Gene expression is shown as relative expression to α-tubulin. Data are presented as
mean ± SD, N=3. Each data point represents a replicate of the same RNA sample. (D – G)
RNAscope in situ hybridization analyses of human cerebellum and cerebral cortex tissue (male,
63-yrs old). A chromogenic assay was performed with two targets: CYP11A1 (red) and myelin
basic protein (MBP; teal). Each punctate dot represents one molecule of mRNA. CYP11A1-
positive cells can be found in the granule layer (D), molecular layer and Purkinje layer (E) of the
cerebellum, as well as gray matter of the cortex (F), indicated by black arrows. No CYP11A1-
positive cells were observed in white matter (G). Little co-localization between MBP and
CYP11A1 can be observed.

69

70

71

Figure 7: RNAscope in situ hybridization analyses of human brain tissue.
A chromogenic assay was performed with two targets: CYP11A1 (red) and myelin basic protein
(MBP; teal). Each punctate dot represents one molecule of mRNA. Tissues from two donors
were used: cerebral cortex tissue from a 63y male and from a 92y female. CYP11A1-positive
cells can be found in the gray matter of the cortex (B) but little to none in white matter (A, C, D),
as indicated by black arrows. CYP11A1-positive cells were only observed in the cortex white
matter of one donor (D). (E) Representative positive control for assay (red = RNA polymerase II
subunit a (Polr2a), teal = peptidylprolyl isomerase B (PPIB))

Next, we measured CYP11A1 expression in the glial cell lines (Figure 8). All four glial
cell lines expressed low but detectable amounts of CYP11A1 mRNA, with relative expression
between 10
3
to 10
4
times lower than H295R-S1 cells (Figure 8A). We also confirmed that all four
glial cell lines express mRNA for the CYP11A1 co-factors, FDX1 and FDXR (Figure 8B, C).
Despite the presence of low CYP11A1 mRNA levels in the glial cells, we could not detect any
CYP11A1 protein expression by immunoblotting using four different commercially available

72
antibodies (Figure 8D, Figure 9A - D). No specific signal for CYP11A1 could be observed on
immunoblots even when up to 80 μg protein was loaded for glial cells (Figure 9E). When protein
bands around 40, 50 and 60 kDa were subjected to shotgun proteomics mass spectrometry
analysis, no CYP11A1 protein/peptide sequences were found, confirming the lack of specific
CYP11A1 bands in the immunoblots. Low levels of FDX1 and FDXR protein can be detected in
the four glial cell lines (Figure 8E).

Figure 8: RNA and protein expression of CYP11A1 and its co-factors in glial cells.
(A – C) qRT-PCR analyses of CYP11A1, FDX1, and FDXR in human glial cells, with H295R-S1
as a positive control. Gene expression is shown as relative expression to α-tubulin. Data are
presented as mean ± SD, N=3. Each data point represents total RNA extracted from cells of a
different passage for each cell line. (D, E) Representative immunoblots of CYP11A1, FDX1, and
FDXR in human glial cells, with MA-10 and H295R-S1 as positive controls. GAPDH was used
as a loading control. 15 μg of total cell lysate was loaded in each lane. No specific bands for
CYP11A1 can be observed in glial cells.

73

Figure 9: Western blot analyses of CYP11A1 in human glial cells.
MA-10 and H295R-S1 were used as positive control. (A) Full membrane image of immunoblot
shown in Figure 1B. Anti-CYP11A1 antibody was raised against the N-terminus sequence of
human CYP11A1. (B-D) Immunoblots using different anti-CYP11A1 antibodies. 15 μg of total
cell lysate was loaded into each lane, with GAPDH as a loading control. Antigens used to raise
each antibody are: (B) CYP11A1 aa 1– 300, (C) human CYP11A1 aa 288 – 337, (D) human
CYP11A1 aa 350 – 521. To confirm antibody specificity for blots (C) and (D), protein bands at
40, 50, and 60 kDa were excised and analyzed by shotgun proteomics; however, no peptide
sequence corresponding to CYP11A1 could be found (see Dataset S1). (E) Western blot analyses
of CYP11A1 in human glial cells using the same antibody as (A), with 20, 40, 60, or 80 μg of
total cell lysate loaded into each lane. No specific bands for CYP11A1 can be observed for any
of the glial cell lines. β-actin was used as a loading control.

We were, however, able to detect low levels of CYP11A1 staining using
immunofluorescence (Figure 10). In glial cells, although the strongest punctate stains for
CYP11A1 co-localized with mitochondria, the fluorescent signals were not clearly
distinguishable from background fluorescence.

74

Figure 10: Confocal images of human cells for CYP11A1 expression.
Cells stained with anti-CYP11A1 antibody (green), with mitochondria labelled with Mitotracker
Red and nucleus stained with DAPI (blue). Images were taken at 63x magnification. H295R-S1
cells stain very strongly for CYP11A1, which overlaps entirely with mitochondrial staining. Very
faint CYP11A1 staining co-localizing with mitochondria can be seen in MGM-1, MGM-3, NHA,
and HMC3 cells, with the strongest signal in MGM-3 cells.

75
2.3.3 Human glial cells synthesize pregnenolone that can be stimulated by TSPO ligand XBD-
173 and hydroxycholesterols
Although CYP11A1 protein expression was barely detectable, we found that glial cells
produced pregnenolone. Secreted pregnenolone in culture media was measured using ELISA and
confirmed using mass spectrometry (Figure 11). In addition, pregnenolone levels accumulated in
the culture media over time, confirming novel synthesis and secretion of pregnenolone by glial
cells (Figure 12A, B). We found that pregnenolone production can be stimulated by the TSPO
ligand XBD-173, but not the TSPO ligand FGIN-1-27 at 50 µM (Figure 13). In MGM-3 cells,
FGIN-1-27 inhibited pregnenolone production, but this trend was not observed in the other three
glial cell lines. Pregnenolone production was also increased when cells were given CYP11A1
substrates, such as 22(R)-hydroxycholesterol (22(R)-HC) and 20α-hydroxycholesterol (20α-HC)
(Figure 14) (Lambeth, Kitchen, & Farooqui, 1982). However, pregnenolone production was
increased only in glial cells and not peripheral cells when cells were treated with 22(S)-
hydroxycholesterol (22(S)-HC) (Figure 14). This result suggests the presence of desmolase
activity that can cleave the bond between C20 and C22 in hydroxycholesterols to produce
pregnenolone. Taken together, these data indicate that human glial cells have the ability to
synthesize pregnenolone and pregnenolone production can be stimulated by some of the
compounds that increase steroid production in peripheral steroidogenic cells.

76

77
Figure 11: Mass spectrometry detection and quantification of pregnenolone production by cells.
(A, C) Product ion scan for pregnenolone-AMP in derivatized samples of 200 ng/mL
pregnenolone standard (A) and MGM-1 supernatant extract (C). (B, D) Extracted ion
chromatograms (XIC) for pregnenolone-AMP (414.90 Da) in 1 ng/mL pregnenolone standard
(B) and MGM-1 supernatant extract (D). (E) Calibration curve of pregnenolone-AMP. Under
basal conditions, normalized pregnenolone-AMP concentration is 14.87 ng/mg protein in
H295R-S1 supernatant and 0.32 ng/mg protein in MGM-1 supernatant, calculated based on the
calibration curve.

78

Figure 12: Pregnenolone secretion by MGM-1 cells over time.
Measurement of pregnenolone in MGM-1 cell supernatants after 2, 6, 12, and 24 hours at basal
levels (A) or with treatment of 50 µM XBD-173 (B), 0.76 mM aminoglutethimide (C), or 50 µM
ketoconazole (D). An accumulation of pregnenolone over time was observed in all conditions,
indicating new synthesis and secretion of pregnenolone.

79

Figure 13: Human glial cells produce pregnenolone and the production can be altered by TSPO
ligands.
ELISA measurements of pregnenolone in culture media of MGM-1 (A), MGM-3 (B), NHA (C),
and HMC3 (D) cells treated with DMSO control, 50 μM XBD-173 or 50 μM FGIN-1-27 for 2
hours. Data are presented as mean ± SD, N=3. Statistics performed compared to control group.
Each data point represents the average of one experiment, where each treatment was performed
in triplicate within each experiment. XBD-173 significantly stimulated pregnenolone production
in MGM-1, NHA, and HMC3 cells (** p < 0.01, *** p < 0.001) but not MGM-3 cells. FGIN-1-
27 inhibited pregnenolone production in MGM-3 cells (* p < 0.05) but had no effect in other
glial cells.

80
2.3.4 Pregnenolone synthesis by glial cells is not inhibited by CYP11A1 inhibitors
To confirm whether an undetectable but small amount of CYP11A1 enzyme in glial cells
is responsible for producing pregnenolone, we treated the cells with CYP11A1 inhibitors. The
specific CYP11A1 inhibitor DL-aminoglutethimide (AMG) (Vanden Bossche, 1992)
significantly inhibited steroid production in a dose-dependent manner in H295R-S1 and MA-10
cells (Figure 14A-H, Figure 15A). In the peripheral cells, AMG also significantly inhibited
steroid formation in response to 22(R)-HC, 22(S)-HC and 20α-HC treatment. However, in all
four glial cell lines, AMG failed to inhibit steroid production at both the basal level and when
induced with hydroxycholesterols (Figure 14I-X, Figure 15B).

81

82
Figure 14: Effect of aminoglutethimide (AMG) on basal and hydroxycholesterol-stimulated
steroid production in peripheral cells and glial cells.
ELISA measurements of secreted progesterone by MA-10 cells (A-D) and secreted pregnenolone
by H295R-S1 (E-H), MGM-1 (I-L), MGM-3 (M-P), NHA (Q-T) and HMC3 (U-X) cells when
treated with 4 different doses of AMG for 2 hours. Effects of AMG on secreted steroids under
basal (A,E,I,M,Q,U), 50 μM 22(R)-hydroxycholesterol stimulation (B,F,J,N,R,V), 50 μM 22(S)-
hydroxycholesterol stimulation (C,G,K,O,S,W), and 50 μM 20α-hydroxycholesterol stimulation
(D,H,L,P,T,X) conditions are shown. Each data point represents the average of one experiment,
where each treatment was performed in triplicate within each experiment. Data are presented as
mean ± SD, N=3. Statistics performed compared to 0 mM AMG group within each panel. AMG
significantly inhibited steroid production in MA-10 and H295R-S1 cells under basal and
hydroxycholesterol-stimulated conditions; however, AMG did not inhibit pregnenolone
production in the four glial cell lines at low doses under basal and hydroxycholesterol-stimulated
conditions. High doses of AMG increased pregnenolone production in glial cells. (* p < 0.5, ** p
< 0.01, *** p < 0.001)

83

Figure 15: Steroid measurements of H295R-S1 and MGM-1 supernatants by mass spectrometry.
Measurement of pregnenolone, DHEA, and progesterone in H295R-S1 (A) and MGM-1 (B)
supernatants with or without 50 µM 22(R)-hydroxycholesterol stimulation and/or treatment with
0.76 mM AMG or 100 µM KC. Trends were similar to those observed using ELISA, where
AMG and KC significantly inhibited steroid production in H295R-S1 cells but not MGM-1 cells.
Pregnenolone levels produced by MGM-1 cells at basal levels (i.e., without hydroxycholesterol
stimulation) were too low to detect using this mass spectrometry method. (* p < 0.05, ** p <
0.01, *** p < 0.001)

84
To confirm these results, we used a non-specific inhibitor of cytochrome P450 (CYPs)
enzymes, ketoconazole (KC). KC inhibits multiple CYPs along the steroidogenic pathway—
including CYP11A1, CYP17A1, CYP19A1, CYP11B1, and CYP11B2, as well as other CYPs
such as CYP3A4 (Fleseriu & Castinetti, 2016; Gal & Orly, 2014). The effect of KC may indicate
whether another CYP is responsible for pregnenolone synthesis. As expected, KC significantly
inhibited steroid production in H295R-S1 and MA-10 cells (Figure 16A-H). However, similar to
AMG, KC failed to inhibit steroid production in the four glial cell lines at basal and
hydroxycholesterol-stimulated conditions (Figure 16I-X). Interestingly, high doses of AMG and
KC appeared to increase pregnenolone synthesis in glial cells. A 24-hour time course study
revealed that pregnenolone levels increased with time in cell supernatants after treatment of
AMG or KC (Figure 12C, D), suggesting that these drugs stimulate pregnenolone production in
glial cells rather than inhibiting it as in peripheral cells. This induced increase in pregnenolone
secretion does not appear to be due to upregulation of CYP11A1 expression or upregulation of
CYP450 co-factors (Figure 17) but rather by another mechanism. A summary of these results is
presented in Table 5.

85

86
Figure 16: Effect of ketoconazole (KC) on basal and hydroxycholesterol-stimulated steroid
production in peripheral and glial cells.
ELISA measurements of secreted progesterone by MA-10 cells (A-D) and secreted pregnenolone
by H295R-S1 (E-H), MGM-1 (I-L), MGM-3 (M-P), NHA (Q-T) and HMC3 (U-X) cells when
treated with 4 different doses of KC for 2 hours. Effects of KC on secreted steroids under basal
(A,E,I,M,Q,U), 50 μM 22(R)-hydroxycholesterol stimulation (B,F,J,N,R,V), 50 μM 22(S)-
hydroxycholesterol stimulation (C,G,K,O,S,W), and 50 μM 20α-hydroxycholesterol stimulation
(D,H,L,P,T,X) conditions are shown. Each data point represents the average of one experiment,
where each treatment was performed in triplicate within each experiment. Data are presented as
mean ± SD, N=3. Statistics performed compared to 0 μM KC group within each panel. KC
significantly inhibited steroid production in H295R-S1 cells starting at 1 μM and in MA-10 cells
starting at 0.1 μM. However, in glial cells, KC did not inhibit pregnenolone production at any
concentration. In NHA and HMC3 cells, 10 μM KC increased pregnenolone secretion. In all glial
cells, 50 μM KC increased pregnenolone secretion under almost all conditions. (* p < 0.5, ** p <
0.01, *** p < 0.001)

87

Figure 17: Expression of CYP11A1, FDXR, and POR in MGM-1 cells after treatments that
increase pregnenolone synthesis.
(A – C) qRT-PCR analysis of CYP11A1 expression (A) in addition to expression of CYP450 co-
factors, FDXR (B) and POR (C), following treatment of 0.76 mM AMG, 50 µM ketoconazole, or
50 µM 22(R)-hydroxycholesterol for 2 hours in MGM-1 cells. Gene expression is shown as
relative expression to β-actin. Data are presented as mean ± SD, N=3. (D) Representative
immunoblot showing protein expression of CYP11A1, FDXR, and POR following the
aforementioned treatments. β-actin was used as a loading control. No significant changes were
observed for either RNA or protein expression of CYP11A1, FDXR, and POR following AMG,
KC, and 22(R)-HC treatments.

88
Table 5: Effects of AMG and KC on steroid synthesis

Aminoglutethimide Ketoconazole
Basal +22(R)-
HC
+22(S)-
HC
+20α-
HC
Basal +22(R)-
HC
+22(S)-
HC
+20α-
HC
MA-10
↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
H295R-S1
↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
MGM-1
↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑
MGM-3
↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑
NHA
↑ ↑ ↑ ↑ ↑ ↑ - ↑
HMC3
↑ - ↑ - ↑ ↑ ↑ ↑
↑ indicates a significant increasing and ↓ indicates a significant decreasing trend in steroid
production. – indicates no significant change.

2.4 Discussion
We found CYP11A1 mRNA in various parts of the human brain and four human glial
cell lines, with levels between 1,000 – 10,000 times lower than that in human adrenal cells
and tissue. The difference in expression levels agrees with previous studies comparing
CYP11A1 mRNA levels between human brain and adrenal tissues (MacKenzie et al., 2008;
Yu et al., 2002). The expression for ferredoxin and ferredoxin reductase is significantly
higher than CYP11A1 in the glial cells, suggesting that any present CYP11A1 activity would
not be limited by availability of co-factors. Although the microglia cell line HMC3 appears
to have the highest expression of ferredoxin reductase compared to the other glial cell lines,
it produces the least amount of pregnenolone. HMC3 cells also have the lowest mRNA
expression for most steroidogenic enzymes compared to the other glial cell lines, which

89
indicates that microglia are not likely the main steroidogenic cell type in the human brain.
Interestingly, although it is believed that steroidogenesis is involved in the modulation of
inflammatory responses by microglia due to upregulation of TSPO during
neuroinflammation (Pozzo et al., 2019), very few studies have measured steroid production
in human microglia. In fact, some studies reported the lack of CYP11A1 in mouse and
human microglia (Gottfried-Blackmore et al., 2009; Owen et al., 2017). TSPO ligand XBD-
173 was shown to increase pregnenolone synthesis in mouse microglia cell line BV-2 (Bader
et al., 2019), but it failed to induce pregnenolone production in the human microglia cell
line C20 (Milenkovic et al., 2019). In contrast, our results here show that the human
microglia cell line HMC3 can produce pregnenolone and production can be enhanced by
both XBD-173 and hydroxycholesterols. The discrepancy between our results and those of
Milenkovic et al. may be due to the higher concentration of XBD-173 and shorter
stimulation period to prevent desensitization used in our study, as well as differences in the
cell lines. Nevertheless, our results provide additional evidence that human microglia are
indeed steroidogenic.
Despite many studies, including our current one, showing the presence of mRNA for
CYP11A1 in the human brain, there is a lack of studies demonstrating CYP11A1 protein in
this tissue. The only two reports showing CYP11A1 protein in the human brain found very
limited positive immunohistochemical staining in the prefrontal cortex (King et al., 2002)
and cerebellum (Le Goascogne et al., 1989). Our ISH data revealed that CYP11A1 mRNA is
present in less than 1% of cells in the cortex and cerebellum, with each positive cell having
no more than 2 mRNA molecules. CYP11A1-positive cells were mostly in the gray matter of
the cerebellum and cerebral cortex, as well as the Purkinje cell layer in the cerebellum. In

90
contrast to previous rat studies that found CYP11A1 protein localization in the white matter
of various brain regions (Iwahashi et al., 1990; Le Goascogne et al., 1987), we observed
very little CYP11A1-positive cells in the white matter. With such a low mRNA concentration
for CYP11A1 in the human brain, it is unlikely that CYP11A1 protein would be detectable in
human brain tissue without using ultrasensitive methods.
Previous studies have found weak immunocytochemical staining of CYP11A1 in
MGM-1, MGM-3, and NHA cells (R. C. Brown et al., 2000). We obtained similar results
using immunofluorescence, with MGM-3 showing the highest CYP11A1 protein expression
among the glial cell lines. However, we could not detect any specific bands for CYP11A1
using up to 80 μg of total cell lysate in immunoblotting for all four glial cell lines.
Considering the detection limit of our immunoblotting system, which allows identification
of signals from > 0.3 ng protein, we estimate that the concentration of CYP11A1 in glial
cells is less than 4 picograms per microgram total protein. This indicates that CYP11A1 can
only be detected in human glial cells with more sensitive methods. To elucidate whether
such low amounts of CYP11A1 could be responsible for the pregnenolone production in
human glial cells, we investigated whether an alternative pathway may be involved in
pregnenolone synthesis in glial cells.
Using the two CYP11A1 inhibitors AMG and KC, we found that pregnenolone
production in the glial cells were not dose-dependently inhibited by CYP11A1 inhibitors,
unlike adrenal H295R-S1 and MA-10 Leydig cells. Only when we overexpressed CYP11A1
in glial cells did we see a significant inhibitory effect by AMG on pregnenolone synthesis,
which indicates that endogenous pregnenolone production by glial cells is independent of
CYP11A1. In MA-10 and H295R-S1, progesterone and pregnenolone production were

91
highest when cells were stimulated with 22(R)-HC, followed by 20α-HC, whereas the
stereoisomer 22(S)-HC had no effect as expected (Benelli, Michel, & Michel, 1982;
Venugopal et al., 2016). This is in line with the side-chain cleavage reaction by CYP11A1
where hydroxycholesterols with closest resemblance to the intermediates of the reaction are
better metabolized to pregnenolone. However, in glial cells, 22(S)-HC is metabolized to
pregnenolone at equal or higher rates than 22(R)-HC, suggesting that unlike adrenal and
gonadal cells in the testis, there is no stereospecificity in the glial desmolase activity.
Furthermore, 20α-HC appears to stimulate pregnenolone production more than the other two
hydroxycholesterols in glial cells. This suggests that an alternative pathway to the CYP11A1
side chain cleavage reaction may be used to metabolize hydroxycholesterols in glial cells as
compared to peripheral steroidogenic cells.
Our study highlights the discrepancies between the amount of pregnenolone
produced by brain cells and the amount of CYP11A1 that is found in the brain. It has been
reported that mitochondria from rat brains can convert radiolabeled cholesterol to
pregnenolone at a rate of approximately 2.5 pmol/mg protein/hour while adrenal
mitochondria have a conversion rate of approximately 15 pmol/mg/protein/hour (Z. Y. Hu et
al., 1987). Even though the rate of CYP11A1 activity in the brain is only about 6 times
lower than that of the adrenals, the mRNA and protein levels of CYP11A1 in the brain are
more than 100 times lower (MacKenzie et al., 2008; Mellon & Deschepper, 1993; Yu et al.,
2002). We question whether this small amount of CYP11A1 can account for the levels of
pregnenolone found in human brains range, which range from 5.7 ng/g to 127.44 ng/g tissue
with variations seen between different brain regions and between non-demented and
Alzheimer’s Disease patients (Lanthier & Patwardhan, 1986; Marx, Trost, et al., 2006;

92
Naylor et al., 2010; Weill-Engerer et al., 2002). Together with our data, we propose that an
alternative pathway to CYP11A1 contributes to a major portion of the pregnenolone
production in brain.
This is not the first time an alternative pathway for neurosteroid synthesis has been
proposed. Early experiments have shown that more pregnenolone and DHEA can be isolated
from whole brain extracts when treated with FeSO 4, FeCl3, lead tetraacetate, imidazole, or
triethylamine (Prasad, Vegesna, Welch, & Lieberman, 1994). This idea was discussed by
Lieberman et al., who suggested that the intermediates for pregnenolone synthesis found in
vivo may be unstable oxygen radical products rather than stable intermediates that were
isolated in in vitro experiments (Seymour Lieberman, 2008; Seymour Lieberman &
Kaushik, 2006; S. Lieberman & Lin, 2001). The broad substrate specificity of CYP450
enzymes and the possibility that steroid synthesis can occur in multienzyme complexes pose
uncertainties to the generally accepted pathway. Although there is not enough experimental
evidence to support the potential reaction pathways proposed by Lieberman et al., their
analyses indicate that the generally accepted-steroidogenesis pathway is not without
assumptions either. In conclusion, our data has provided evidence to support the notion of a
pregnenolone synthesis pathway in the human brain that is independent on CYP11A1
activity.

93
III. Investigation of the alternative pathway for pregnenolone synthesis
This section is adapted from:
Lin, Y. C., & Papadopoulos, V. (2021). Neurosteroidogenic enzymes: CYP11A1 in the central
nervous system. Front Neuroendocrinol, 62, 100925. doi:10.1016/j.yfrne.2021.100925

Lin, Y . C., Cheung, G., Porter, E., & Papadopoulos, V . The neurosteroid pregnenolone is
synthesized by a mitochondrial P450 enzyme other than CYP11A1 in human glial cells.
Journal of Biological Chemistry, 102110. doi:10.1016/j.jbc.2022.102110 (In Press)

3.0 Abstract
Our data has shown that human brain cells can synthesize the neurosteroid pregnenolone
independently of the activity of the classical CYP11A1 enzyme. Therefore, in this study, we
explored three different possibilities for an alternative pathway for glial-cell pregnenolone
synthesis: 1) one regulated by reactive oxygen species, 2) another using a different CYP11A1
isoform, and 3) one using another CYP450 enzyme. First, oxidants and antioxidants had no
significant effects on pregnenolone synthesis, suggesting that pregnenolone synthesis is not
regulated by reactive oxygen species. Second, overexpression of CYP11A1 isoform b did not
alter pregnenolone synthesis, indicating that use of another CYP11A1 isoform is unlikely. Lastly,
nitric oxide and iron chelators deferoxamine and deferiprone significantly inhibited
pregnenolone production, indicating that a CYP450 enzyme was involved. Knockdown of
endoplasmic reticulum co-factor NADPH-cytochrome P450 reductase had no effect, while
knockdown of mitochondrial CYP450 co-factor ferredoxin reductase inhibited pregnenolone
production. These data suggest that pregnenolone is synthesized by a mitochondrial cytochrome
P450 enzyme other than CYP11A1 in human glial cells.

94
3.1 Introduction
Although it has been well-established that the brain can synthesize neurosteroids locally,
the mechanism through which brain cells produce the major steroid precursor pregnenolone is
unclear, as the brain may not use the same enzyme (CYP11A1) as the periphery. The current
evidence for CYP11A1 presence in the brain shows that only a very small percentage of cells in
the brain express this steroidogenic enzyme. The transcriptional regulation of CYP11A1 appears
to be different from that of peripheral steroidogenic organs; however, the factors that determine
which neuron, astrocyte, oligodendrocyte, or microglia cells can express CYP11A1 are unknown
and require further investigation. The possibility exists that CYP11A1 could perform a different
role in brain other than steroid synthesis. Despite the abundance of studies showing the presence
of various steroidogenic enzymes in brain, the idea of the steroidogenic pathway in the CNS
being identical to that in the periphery is an assumption that can be challenged by results from
multiple studies, including our own (see Chapter II).
The most apparent issue is that the levels of pregnenolone found in the brain do not
always match the amount of CYP11A1 in the brain. For example, Ukena et al. found no changes
in CYP11A1 expression during diurnal cycles despite changes in pregnenolone levels throughout
(Ukena et al., 1998). These findings were replicated recently in human A172 glioma cells, where
pregnenolone secretion in the glioma cells displayed an oscillatory pattern that was dependent on
the circadian clock, but CYP11A1 expression did not change with the circadian rhythm (Witzig
et al., 2020). The daily changes in pregnenolone levels could be due to changes in CYP11A1
activity rather than expression, but there are still discrepancies between the levels of
pregnenolone that are found in brain and the expression levels of CYP11A1. The mRNA levels
for CYP11A1 are reported to be more than 1000 times—and up to 100,000 times—lower than

95
that in the adrenals, and brain mitochondria were observed to convert cholesterol to
pregnenolone at a six times slower rate than adrenal mitochondria (Z. Y . Hu et al., 1987;
MacKenzie et al., 2008; Yu et al., 2002). With this low concentration and slow conversion rate, it
would appear unrealistic that local synthesis by CYP11A1 in brain could account for the
pregnenolone concentrations found in various brain areas, which range from 5.7 ng/g tissue to
127.44 ng/g tissue (Lanthier & Patwardhan, 1986; Marx, Trost, et al., 2006; Naylor et al., 2010;
Weill-Engerer et al., 2002). This notion has been discussed by Warner and Gustafsson (Warner &
Gustafsson, 1995, 1999). Although studies show that oligodendrocytes can metabolize
radioactive cholesterol to pregnenolone at higher rates than other brain cells, it is unclear that the
low CYP11A1 activity in oligodendrocytes can account for pregnenolone levels seen in rat
brains after adrenalectomy with castration given the low number of oligodendrocytes present in
brain (Warner & Gustafsson, 1995). Furthermore, the rate and magnitude of pregnenolone
stimulation that can be induced by TSPO ligands does not match the CYP11A1 levels in brain. In
one study, intravenous injection of the TSPO ligand 4’-chlorodiazepam in rats led to a 70 – 100%
increase in pregnenolone concentrations in the forebrain within 5 minutes of administration
(Korneyev et al., 1993). As analyzed by Warner and Gustafsson, these results would only be
possible if the pregnenolone synthesis rate in the rat brain is 150 pmol/mg mitochondrial
protein/min, which is unreasonably high and would indicate that another mechanism must be
responsible for producing pregnenolone (Warner & Gustafsson, 1995).
There can be many explanations for the discrepancy between the levels of pregnenolone
in the CNS and the expression of CYP11A1. One consideration is the interaction of
steroidogenesis between the periphery and CNS. Under normal conditions, both peripheral gland
secretion and local synthesis contribute to the pregnenolone pool in the CNS, which would

96
explain why there will be higher amounts of pregnenolone in the CNS in relation to CYP11A1
expression. However, in adrenalectomized and castrated rats, the pregnenolone concentration in
the brain was 60 – 70% of that in control rats even though plasma pregnenolone levels were
more than 10 times lower (Korneyev et al., 1993). This suggests that a significant proportion of
pregnenolone in the brain is locally derived. Another source for pregnenolone in the brain could
be metabolism of peripheral precursors that bypass the need for CYP11A1 activity. Given that
pregnenolone is the first step in steroidogenesis, the only probable precursor that may be derived
from the periphery is pregnenolone sulfate, which can be converted to pregnenolone by
sulfatases (Qaiser et al., 2017). The caveat is that this pathway is still dependent on peripheral
steroidogenesis and thus cannot be a significant source of pregnenolone in adrenalectomized and
gonadectomized animals. Therefore, even though peripheral steroidogenesis may contribute to
the discrepancy between pregnenolone levels and CYP11A1 levels in the brain, it is not likely
the major factor.
Another explanation could be that CYP11A1 is rapidly activated under conditions when
steroidogenesis is needed, similar to how aromatase activity can be rapidly modulated by
phosphorylation in the brain (Charlier et al., 2015). Post-translational modifications can allow
alterations to enzyme activity without significant changes to protein expression, which could
allow higher levels of pregnenolone to be measured in the brain even with such low CYP11A1
expression. Stimulation of CYP11A1 activity by phosphorylation has been reported in
mitochondria from peripheral steroidogenic organs (Aguiar, Masse, & Gibbs, 2005); however,
post-translational modifications of CYP11A1 in the CNS has not been studied. The conditions
for CYP11A1 activation are also unknown, although it may be similar to aromatase, where
signaling by various neurotransmitters modulate enzyme activity (Charlier et al., 2015).

97
Furthermore, the activity of CYP11A1 could also be regulated by the metabolic rate of
downstream steroidogenesis enzymes. Conversion of pregnenolone to pregnenolone sulfate,
DHEA, or progesterone by the respective enzymes could reduce the pregnenolone pool and
trigger feedback mechanisms to upregulate CYP11A1 activity. This idea is supported by studies
in reconstituted enzyme systems showing that the presence of pregnenolone inhibits CYP11A1-
ferredoxin complex formation, while the presence of DHEA sulfate increases the rate of
pregnenolone production (Neunzig & Bernhardt, 2014; Yablokov et al., 2021). However, the
exact mechanism of how downstream steroids can modulate CYP11A1 activity is still unclear.
Overall, the higher pregnenolone levels in the brain may be in part accounted for by activation of
CYP11A1 activity that is independent of changes in expression.
However, in our opinion, the most plausible explanation for the discrepancy between
CYP11A1 expression and pregnenolone levels in the brain is that CYP11A1 activity is not the
sole mechanism of pregnenolone synthesis in the CNS. Some studies have shown that steroids
can be synthesized by brain cells and tissue independently of the activity of the corresponding
enzyme, raising the possibility of an alternative steroidogenesis pathway. For example, treatment
of rat brain extracts with organic bases (triethylamine and imidazole) or iron compounds yielded
more pregnenolone than simple steroid extraction from brains (Prasad et al., 1994). The same
results were observed for DHEA with treatment of triethylamine and imidazole but not with
treatment of iron compounds. Prasad et al. suggested that the brain contains non-peroxide labile
precursors that can be used to produce pregnenolone and DHEA. The reaction proposed was a
one-electron reduction of the O-O bond of peroxide followed by fragmentation of the resulting
alkoxy radical. Supporting this notion, human brain cells were found to synthesize DHEA
independently of CYP17A1 activity via a mechanism dependent on reactive oxygen species

98
(ROS) (R. C. Brown et al., 2000). ROS-dependent DHEA synthesis was also observed in C6 rat
glioma cells, primary rat glial cells, and human brain tissue (R. C. Brown et al., 2003; Cascio et
al., 2000; Cascio, Prasad, Lin, Lieberman, & Papadopoulos, 1998). Moreover, the currently
recognized intermediates in the side chain cleavage process by CYP11A1 have been questioned
and suggested to actually be unstable, non-isolatable compounds instead of the canonical 22(R)-
hydroxycholesterol and 20α,22(R)-dihydroxycholesterol (Seymour Lieberman & Kaushik, 2006;
S. Lieberman & Lin, 2001).
Although a complete alternative pathway has yet to be elucidated, the potential existence
of a CYP11A1-independent mechanism for pregnenolone synthesis is worthy of consideration.
Based on the information presented, we propose three possibilities for the alternative pathway(s)
for pregnenolone synthesis. First, the mechanism for pregnenolone formation in the brain could
depend on reactive oxygen species, as explained above from the studies by Prasad et al., Brown
et al., and Lieberman et al. (R. C. Brown et al., 2000; R. C. Brown et al., 2003; Seymour
Lieberman & Kaushik, 2006; S. Lieberman & Lin, 2001; Prasad et al., 1994). It is possible that
pregnenolone can be made from unstable precursors generated by reactive oxygen species in the
brain, which would imply that conversion to pregnenolone can potentially serve as a protective
mechanism. The second possibility is that the second CYP11A1 isoform may be the main
enzyme metabolizing cholesterol to pregnenolone in the CNS rather than the classical isoform.
Since studies trying to detect CYP11A1 mRNA and/or protein in the CNS have not specifically
designed primers or antibodies to differentiate between the two isoforms, it is possible that
expression of the second isoform could be significantly higher but has gone undetected with the
experimental methods used. A drawback for this second possibility, however, is that it would be
only applicable to species that have two CYP11A1 isoforms, such as mouse and human, and

99
would not be able to account for potential alternative pathways in species such as rats. The third
possibility is that another cytochrome P450 other than CYP11A1 could be responsible for
synthesizing pregnenolone in the CNS, as cytochrome P450 enzymes are known to metabolize a
wide range of substrates. The following study aims to provide experimental data to determine
whether one or more of these proposed pathways are used to produce pregnenolone in the CNS.

100
3.2 Materials and Methods
Cell Culture
Cell line sources and culture methods are identical to those presented in Section 2.2.

Cell Treatments for Steroid Measurements
Compounds used to treat cells were purchased from Sigma: DL-aminoglutethimide
(#A9657), ketoconazole (#K1003), 22(R)-hydroxycholesterol (#H9384), (±)-6-Hydroxy-2,5,7,8-
tetramethylchromane-2-carboxylic acid (Trolox; #238813), hydrogen peroxide solution
(#216763), iron (II) sulfate (#F8633), S-Nitroso-N-acetyl-DL-penicillamine (SNAP; #N3398),
deferoxamine mesylate (#1166003), 3-Hydroxy-1,2-dimethyl-4(1H)-pyridone (deferiprone;
#379409), and diphenyleneiodonium chloride (#D2926). Cell treatment and steroid measurement
procedures are the same as those presented in Section 2.2.

CYP11A1 Isoforms Overexpression
MGM-1 cells were seeded into 24 well plates at a density of 50,000 cells/well. The cells
were then transfected with 1 μg of pCMV-entry mammalian expression vector (Origene,
#PS100001), CYP11A1 transcript variant 1 (Myc-DDK-tagged; CYP11A1a) expression vector
(Origene, #RC207121), or CYP11A1 transcript variant 2 (Myc-DDK-tagged, CYP11A1b)
expression vector (Origene, #RC211728) using Lipofectamine 3000 (Thermo Fischer Scientific,
#L3000001) and Opti-MEM (Gibco, Thermo Fischer Scientific, #31985062), according to
manufacturer’s instructions. After 24 hours, culture media was replaced. Transfected cells were
then selected with 1 mg/mL geneticin (Gibco, Thermo Fischer Scientific, #10131027) for 7 days.
qRT-PCR, immunofluorescence, and western blot were used to confirm CYP11A1a or

101
CYP11A1b expression. Steroid measurements for transfected cells were performed as described
in Section 2.2.

siRNA Knockdown of FDXR and POR
MGM-1 cells were seeded into 12 well plates at a density of 60,000 cells/well or into 6
well plates at a density of 120,000 cells/well. The cells were transfected with 5 nM scrambled
negative control siRNA duplexes or gene-specific siRNA duplexes (Origene, #SR303637 for
POR and #SR301560 for FDXR) using siTran 2.0 transfection reagents (Origene, #TT320001),
according to manufacturer’s instructions. Culture media was replaced 18 hours after transfection.
Transfected cells in 12 well plates were collected 48 hours after transfection for RNA analysis by
qRT-PCR. After 72 hours post-transfection, transfected cells in 6 well plates were washed and
treated with serum-free media for 2 hours, after which media was collected for steroid
measurement by ELISA and the cell pellet collected for protein quantification and western
blotting. Knockdown efficiency was determined by both qRT-PCR and immunoblot analyses.

RNA Extraction and qRT-PCR
RNA extraction and qRT-PCR procedures are identical to those described in Section 2.2.
Relative quantification analysis was performed using the 2
-ΔCT
method except when comparing
transfected cells to negative controls, where the 2
-ΔΔCT
method was used instead. CYP11A1
expression was detected by the CYP11A1 exon 3 – 4 primers listed in Table 6, unless a particular
variant was specified.

102
Table 6: Sequences of primers used to measure CYP11A1 variants
Gene Forward Primer Reverse Primer
CYP11A1 exon
3 – 4
GCTTTGCCTTTGAGTCCATCA CTCGGGGTTCACTACTTCCTC
CYP11A1
variant 1 exon 1
CACGCTCAGTCCTGGTCAAA GGGGATCTCATTGAAGGGGC
CYP11A1
variant 2 exon 1
ATTCCGAAATTGGGTCCGCC GAGTCCGTGTTGGGGAACTT

Cell Lysate Preparation and Western Blotting
Protein extraction and western blotting methods were the same as those described in
Section 2.2. The following primary antibodies were used: anti-FDXR rabbit pAb (Proteintech,
#15584-1-AP, 1:1000 dilution), anti-beta actin mouse mAb (Abcam, #ab8226, 1:3000 dilution),
anti-cytochrome P450 reductase rabbit pAb (Abcam, #ab13513, 1:1000 dilution), anti-Myc-tag
rabbit mAb (Cell Signaling Technology, #2278S, 1:1000 dilution).

ROS Measurements
Changes in ROS levels were measured using a Cellular ROS Assay Kit (Abcam,
#ab186027), performed according to manufacturer’s instructions. Briefly, cells were seeded in 96
well plates at 70-80% confluency and incubated with ROS Red Stain working solution provided
in the kit for 1 hour prior to treatment. Treatment solutions made at 10x concentration were then
added to the wells to achieve the appropriate treatment concentrations. After a 2-hour incubation,
fluorescence was measured at 520 nm/605 nm using a Biotek Synergy H1 Hybrid Multi-Mode
Microplate Reader. ROS measurements were normalized to that of the control condition.

103
Immunocytochemistry
Procedures for cell staining and imaging were the same as those described in Section 2.2,
with the exception that the primary antibody anti-Myc-tag rabbit mAb (Cell Signaling, #2278S,
1:200 dilution) was used instead.

Statistics
Statistical analyses of steroid and ROS measurements were performed using GraphPad
Prism 7. Statistical significance was determined using one-way ANOV A followed by Dunnett’s
multiple comparison test. For multiple comparison tests, each value was compared to the no-drug
solvent control group unless otherwise indicated.

104
3.3 Results
3.3.1 Pregnenolone synthesis in glial cells is not dependent on reactive oxygen species
We first examined the involvement of reactive oxygen species in pregnenolone synthesis
by glial cells. As presented in Chapter II, our results showed that high concentrations of the two
CYP11A1 inhibitors aminoglutethimide (AMG) and ketoconazole (KC) seemed to increase
rather than inhibit pregnenolone production in glial cells both at the basal and
hydroxycholesterol-stimulated conditions. AMG and KC have also been reported to increase
cellular ROS (Siraki, Bonini, Jiang, Ehrenshaft, & Mason, 2007; Won et al., 2015). Furthermore,
previous studies of steroidogenesis in MGM-1, MGM-3, and NHA cells have revealed that
DHEA synthesis by these cells is stimulated by oxidative stress and is independent of CYP17A1
activity in the classical pathway (R. C. Brown et al., 2000). The dose-dependent stimulation in
DHEA production by various oxidants was most apparent in MGM-1 cells. Thus, we postulated
that pregnenolone may be synthesized in a ROS-dependent manner and tested this possibility in
MGM-1 cells.
First, we examined the effects of the antioxidant Trolox, an analog of Vitamin E, on
pregnenolone production and cellular ROS in MGM-1 cells (Figure 18). We also co-treated the
cells with high doses of AMG or KC to see whether an antioxidant could reduce their stimulatory
effects on pregnenolone production. In MGM-1 cells, treatment with Trolox did not result in any
significant changes in pregnenolone levels (Figure 18A) or cellular ROS (Figure 18D) overall.
The increase in pregnenolone production with treatment of 2 mM Trolox was marginally
significant (p = 0.0477). High dose AMG (0.76 mM) and high dose KC (50 μM) treatments led
to higher pregnenolone production but no significant changes in cellular ROS (Figure 18B, C, E,
F). Furthermore, the addition of Trolox had no significant effect on the increased pregnenolone

105
production induced by AMG and KC (Figure 18B, C). Low doses of Trolox decreased the levels
of cellular ROS induced by KC, but this effect was blunted at the highest dose of 2 mM (Figure
18F). Trolox also did not induce significant changes in ROS when cells were co-treated with
AMG (Figure 18E).

Figure 18: Effect of antioxidant Trolox on pregnenolone secretion and intracellular ROS in
MGM-1 cells.
ELISA measurements (A-C) of secreted pregnenolone and intracellular ROS measurements (D-
F) when MGM-1 cells were treated with different doses of Trolox for 2 hours, either alone (A,
D) or combined with 0.76 mM AMG (B,E) or 50 μM KC (C, F). Each data point represents the
average of one experiment, where each treatment was performed in triplicate within each
experiment. Data are presented as mean ± SD, N=3. Statistics performed compared to 0 mM
Trolox (A, D), 0 mM Trolox + 0.76mM AMG (B, E), or 0 mM Trolox + 50 μM KC (C, F).
Trolox did not significantly change secreted pregnenolone or ROS except for a marginally
significant increase in pregnenolone production with the 2 mM Trolox treatment. Although 0.76
mM AMG and 50 μM KC significantly increased pregnenolone production in MGM-1 cells,
neither drug significantly altered intracellular ROS. Low doses of Trolox decreased ROS when
MGM-1 cells were treated with 50 μM KC, but this effect was blunted at the highest dose of 2
mM Trolox. The trend seen with low dose Trolox combined with 50 μM KC did not correlate
with changes in pregnenolone production. (* p < 0.05, ** p < 0.01, *** p < 0.001)

106
To confirm these results, we tested the same conditions in MGM-3 cells. Similar to
MGM-1 cells, Trolox did not significantly affect pregnenolone synthesis and cellular ROS by
itself or in combination with 0.76 mM AMG in MGM-3 cells (Figure 19). On the other hand, 50
μM KC significantly increased cellular ROS in MGM-3 cells, and ROS induction was decreased
by higher doses of Trolox (Figure 19F). However, only lower doses of Trolox reduced the
increase in pregnenolone when cells were treated with 50 μM KC (Figure 19C). Therefore, our
results indicate that antioxidants do not affect pregnenolone production in the glial cells.

107

Figure 19: Effect of antioxidant Trolox on pregnenolone secretion and intracellular ROS in
MGM-3 cells.
ELISA measurements (A-C) of secreted pregnenolone and intracellular ROS measurements (D-
F) when MGM-3 cells were treated with different doses of Trolox for 2 hours, either alone (A,
D) or combined with 0.76 mM AMG (B,E) or 50 μM KC (C, F). Each data point represents the
average of one experiment, where each treatment was performed in triplicate within each
experiment. Data are presented as mean ± SD, N=3. Statistics performed compared to 0 mM
Trolox (A, D), 0 mM Trolox + 0.76mM AMG (B, E), or 0 mM Trolox + 50 μM KC (C, F).
Trolox treatment alone did not significantly change secreted pregnenolone or ROS at any
concentration. Although 0.76 mM AMG and 50 μM KC significantly increased pregnenolone
production in MGM-3 cells, only 50 μM KC significantly increased cellular ROS. Low doses of
Trolox decreased pregnenolone production when MGM-3 cells were treated with 50 μM KC, but
this effect was not seen at the highest dose 2 mM Trolox. Higher doses of Trolox decreased the
elevation of cellular ROS induced by 50 μM KC. However, the trends in pregnenolone
production and ROS when MGM-3 cells were treated with 50 μM KC combined with Trolox do
not correlate. (* p < 0.05, ** p < 0.01, *** p < 0.001)

108
Since antioxidant treatment failed to induce significant changes in pregnenolone
synthesis, we aimed to confirm the role of ROS in pregnenolone synthesis by treating MGM-1
cells with two oxidants: hydrogen peroxide (H2O2) and ferrous sulfate (FeSO4). These two
oxidants were chosen because they were previously shown to increase pregnenolone production
in C6 rat glial cells, with 10 mM H2O2 inducing a 2-fold increase and 10 mM FeSO4 inducing a
3.6-fold increase (Cascio et al., 1998). However, in MGM-1 cells, H2O2 had no significant effect
on pregnenolone production (Figure 20A) despite significantly increasing cellular ROS starting
at 100 μM (Figure 20B). We also found that FeSO4 did not have any significant effects on
pregnenolone synthesis (Figure 20C). However, after two hours of FeSO4 treatment at 500 μM,
the highest dose, there was a significant increase in cellular ROS, with a non-significant
increasing trend at lower concentrations (Figure 20D). In our preliminary studies where we pre-
treated MGM-1 cells with FeSO4 for 24 hours, we had observed a significant increase in ROS
starting at 50 μM but there were still no significant changes in pregnenolone production (data not
shown). The same results were seen in MGM-3 cells where both H2O2 and FeSO4 failed to
induce significant changes in pregnenolone production (Figure 21).
To confirm whether other forms of oxidative stress may be involved in pregnenolone
synthesis, we treated MGM-1 cells with the nitric oxide donor S-nitroso-N-acetylpenicillamine
(SNAP). Interestingly, SNAP significantly inhibited pregnenolone production by the cells
(Figure 20E), despite dose-dependently increasing cellular ROS levels the same as other oxidants
we tested (Figure 20F). Given that there is no correlation between changes in intracellular ROS
levels and pregnenolone production, we believe that the inhibitory effects of nitric oxide on
pregnenolone synthesis is due to another mechanism, which will be discussed further in a later
section. Taken together, our antioxidant and oxidant experimental results provide evidence

109
against our first possible pathway; thus, the alternative pathway for pregnenolone synthesis does
not appear to be dependent on ROS.

Figure 20: Effect of oxidants on pregnenolone secretion and intracellular ROS in MGM-1 cells.
ELISA measurements (A, C) of secreted pregnenolone and intracellular ROS measurements (B,
D) when MGM-1 cells were treated with different doses of hydrogen peroxide (H2O2; A, B) or
ferrous sulfate (FeSO4; C, D). Each data point represents the average of one experiment, where
each treatment was performed in triplicate within each experiment. Data are presented as mean ±
SD, N=3. Statistics performed compared to the no treatment group within each panel. Hydrogen
peroxide treatment significantly increased intracellular ROS starting at 100 μM but did not alter
pregnenolone secretion by MGM-1 cells. Ferrous sulfate treatment did not significantly increase
intracellular ROS except at 500 μM and did not change pregnenolone production by MGM-1
cells. (* p < 0.05, ** p < 0.01, *** p < 0.001)

110

Figure 21: Effect of oxidants on pregnenolone secretion and intracellular ROS in MGM-3 cells.
ELISA measurements (A, C) of secreted pregnenolone and intracellular ROS measurements (B,
D) when MGM-1 cells were treated with different doses of hydrogen peroxide (H2O2) (A, B) or
ferrous sulfate (FeSO4) (C, D) for 2 hours. Each data point represents the average of one
experiment, where each treatment was performed in triplicate within each experiment. Data are
presented as mean ± SD, N=3. Statistics performed compared to the no treatment group within
each panel. Hydrogen peroxide treatment significantly increased intracellular ROS starting at
100 μM but did not alter pregnenolone secretion by MGM-1 cells. Ferrous sulfate treatment did
not significantly increase intracellular ROS except at 500 μM and did not change pregnenolone
production by MGM-1 cells. (* p < 0.05, *** p < 0.001)

111
3.3.2 Overexpression of CYP11A1b did not alter pregnenolone secretion unlike overexpression
of CYP11A1a
Next, we examined the roles of different CYP11A1 isoforms in pregnenolone synthesis.
A second CYP11A1 isoform has been reported in mouse and humans (Lin & Papadopoulos,
2021), but the function, expression, and localization of this isoform has not yet been studied. For
isoform a (i.e. the classical CYP11A1 enzyme), CYP11A1a will be used to denote the protein
while CYP11A1 variant 1 will be used to indicate the mRNA transcript (NCBI Gene, 1988 - ,
Gene ID: 1583). Similarly, CYP11A1b will be used to denote the protein for isoform b and
CYP11A1 variant 2 will be used to indicate the mRNA transcript. In comparison to isoform a,
isoform b is 158 amino acids shorter from the N-terminus. This second isoform could
potentially be resistant to the inhibitory effects of AMG and KC due to possible structural
differences but may still be able to produce pregnenolone. Therefore, human glial cells may
use the second CYP11A1 isoform to synthesize pregnenolone.
We first examined the levels of the CYP11A1 variants in human brain and in brain
cells (Figure 22). Since the mRNA transcripts differ only by the sequence of exon 1, we
conducted qRT-PCR to compare the expression of the variants using primers designed for
exon 1 from each respective variant. The total CYP11A1 expression was determined using
primers complementary to sequences from the common exons. In peripheral steroidogenic
organs and H295R-S1 adrenal cells, it was very clear that CYP11A1 variant 1 is the
predominant variant and makes up almost the entirety of the total CYP11A1 expression
(Figure 22A, C). Although the relative expression of CYP11A1 variant 2 appears to be lower
than that of CYP11A1 variant 1 in all CNS tissues and brain cells, the total CYP11A1
expression was higher than the summed expression of the variants except for in MGM-3 and

112
HMC3 cells (Figure 22B, C). This may suggest an early truncation of the RNA transcript for
CYP11A1b, since exon 1 is not translated into the final protein for this isoform. We could
not accurately determine the protein levels for CYP11A1b in the human brain cells because
CYP11A1b does not have a unique amino acid sequence that can differentiate it from
CYP11A1a, and commercial antibodies that only bind CYP11A1a did not produce clear
bands on immunoblots for glial cells to allow for subtraction (Figure 9A, B).

Figure 22: RNA Expression of CYP11A1 variants in the human CNS and peripheral tissues and
cells.
qRT-PCR analyses of total CYP11A1 (gray), CYP11A1 variant 1 (blue), and CYP11A1 variant 2
(red) expression in human peripheral steroidogenic tissues (A), human CNS tissues (B), and cell
lines (C). Gene expression is shown as relative expression to β-actin. Data are presented as mean
± SD.

113
To characterize the CYP11A1 isoforms better, we transfected MGM-1 cells with a
CYP11A1a or CYP11A1b expression vector with a myc-DDK tag. After drug selection and
colony expansion, we verified that the transfected MGM-1 cells highly expressed the
corresponding CYP11A1 isoforms (Figure 23, Figure 24A-C). As expected, CYP11A1a is
strictly localized to the mitochondria while CYP11A1b has a more diffused localization
throughout the cell and limited co-localization with mitochondria (Figure 23). The CYP11A1
variant 2 expression vector used did not contain the unique exon 1 sequence as it is not translated
into the final protein; thus, successful transfection was observed through increased total
CYP11A1 expression without an increase in CYP11A1 variant 1 exon 1 (Figure 24A, B).
CYP11A1b protein has a molecular weight of approximately 42 kDa on immunoblots while
CYP11A1a appears as an approximately 50 kDa protein (Figure 24C).
Cells overexpressing CYP11A1a showed approximately a 27-fold increase in basal
pregnenolone production and a 3-fold increase in 22(R)-HC-stimulated pregnenolone production
compared to control cells (Figure 24D, E). Unlike WT MGM-1 cells where AMG had no
inhibitory effect (Figure 14I, J), pregnenolone production in CYP11A1a+ cells was significantly
inhibited by AMG under both basal and hydroxycholesterol-stimulated conditions (Figure 24F,
G). These results suggest that overexpressing the classical CYP11A1a in MGM-1 cells
significantly increases pregnenolone synthesis. This enhanced production can be inhibited by
AMG similar to adrenal H295R-S1 and MA-10 Leydig cells that use the classical
steroidogenesis pathway. These data imply that MGM-1 cells are not inherently resistant to the
effects of AMG, but rather endogenous pregnenolone production by MGM-1 cells is independent
of classical CYP11A1a. On the other hand, overexpression of CYP11A1b did not significantly
alter pregnenolone production compared to control cells (Figure 24D, E), and the effects of AMG

114
on CYP11A1b+ cells were similar to that of WT cells, where AMG failed to inhibit
pregnenolone synthesis and instead increased pregnenolone levels at higher doses (Figure 24H,
I). These results indicate that CYP11A1b is not involved in pregnenolone production unlike
CYP11A1a, and that the second possibility listed above of another isoform involved in
pregnenolone production is unlikely.

Figure 23: Localization of CYP11A1 isoforms.
Representative confocal images of MGM-1 cells transfected with CYP11A1a-myc-DDK or
CYP11A1b-myc-DDK expression vector. Cells transfected with empty plasmid were used as a
negative control. Cells were stained with anti-myc-tag antibody (green), with mitochondria
labelled with Mitotracker Red and nuclei stained with DAPI (blue). Images were taken at 63x
magnification. CYP11A1a is strictly localized to the mitochondria. CYP11A1b appears more
dispersed throughout the cell and has limited co-localization with mitochondria.

115

116
Figure 24: CYP11A1a but not CYP11A1b overexpression leads to increased pregnenolone
production that can be inhibited by AMG.
(A, B) qRT-PCR analysis of CYP11A1 expression in MGM-1 cells transfected with empty
plasmid (negative control) or expression vectors for CYP11A1a or CYP11A1b with Myc-DDK
tag. Gene expression is shown as fold change versus negative control, normalized to the
expression of β-actin. Data are presented as mean ± SD, N=3. Each data point represents total
RNA extracted from pooled cells from different passages. (A) Transfected MGM-1 cells showed
more than 200-fold increase in total CYP11A1 mRNA expression. (B) Expression changes of
CYP11A1 variant 1 exon 1 indicate that the correct CYP11A1 isoform had been expressed. (C)
Representative immunoblots for myc-tag in transfected MGM-1 cells. β-actin was used as a
loading control. Thirty μg of protein were loaded into each lane. A band corresponding to
CYP11A1a was observed at around 50 kDa and a band corresponding to CYP11A1b was
observed at around 42 kDa. No specific bands could be observed in WT or transfection control
MGM-1 cells. (D, E) ELISA measurements of pregnenolone secreted by transfected MGM-1
cells, with or without 22(R)-hydroxycholesterol stimulation. (F - I) ELISA measurements of
pregnenolone secreted by transfected MGM-1 cells treated with different doses of AMG for 2
hours under basal (F, H) and 22(R)-hydroxycholesterol stimulated (G, I) conditions. Each data
point represents the average of one experiment, where each treatment was performed in triplicate
within each experiment. Data are presented as mean ± SD, N=3. Statistics performed compared
to control (D, E) or to 0 mM AMG group (F - I). CYP11A1a+ cells synthesized significantly
more pregnenolone than controls and CYP11A1b+ cells, which were significantly inhibited by
AMG. CYP11A1b+ cells behaved similarly to WT MGM-1 cells for pregnenolone synthesis. (**
p < 0.01, *** p < 0.001)

117
3.3.3 The alternative pathway for pregnenolone synthesis likely involves another cytochrome
P450
We demonstrated above that the nitric oxide donor SNAP significantly inhibited
pregnenolone synthesis. Given that there was no correlation between changes in intracellular
ROS and pregnenolone production, we propose that the effect of SNAP on pregnenolone
synthesis may be due to inhibition of cytochrome P450 activity (Minamiyama et al., 1997). Since
cytochrome P450s rely on iron redox reactions for their activities, we tested the effects of iron
chelation on pregnenolone synthesis. In WT MGM-1 cells, deferoxamine inhibited pregnenolone
production, but this effect was only statistically significant when cells were stimulated with
22(R)-HC (Figure 25A, B). In CYP11A1a+ MGM-1 cells, deferoxamine inhibited pregnenolone
production in a dose-dependent manner under both basal and 22(R)-HC stimulated conditions,
demonstrating effective inhibition of CYP450 activity (Figure 25C, D). To confirm the effects of
iron chelation on pregnenolone synthesis, we treated the cells with another structurally different
iron chelator, deferiprone (Simonart et al., 2000). Indeed, pregnenolone production was
significantly inhibited by deferiprone at high doses in MGM-1 cells (Figure 25E, F). The
inhibitory effects of SNAP and iron chelation on pregnenolone secretion by MGM-1 cells
suggest that the alternative pathway for pregnenolone synthesis may involve activity of another
CYP450.

118

Figure 25: Effect of iron chelation on pregnenolone secretion in MGM-1 cells.
ELISA measurements of secreted pregnenolone when MGM-1 WT (A, B, E, F) and CYP11A1a+
(C, D) cells were pre-treated with different doses of deferoxamine (A - D) or deferiprone (E – F)
for 24 hours. Each data point represents the average of one experiment, where each treatment
was performed in triplicate within each experiment. Data are presented as mean ± SD, N=3.
Statistics performed compared to the no treatment group within each panel. Deferoxamine
significantly inhibited pregnenolone production in CYP11A1+ cells and 22(R)-HC stimulated
WT cells but had no statistically significant effect on WT cells under basal conditions.
Deferiprone significantly inhibited pregnenolone production at high doses in basal and 22(R)-HC
stimulated WT cells.(* p < 0.05, ** p < 0.01, *** p < 0.001)

119
To further test this possibility, we explored the effects of inhibiting CYP450 co-factors,
specifically POR and FDXR. CYP450 enzymes in the endoplasmic reticulum accept electrons
from NADPH via POR while enzymes in the mitochondria accept electrons via FDXR and
FDX1 (Nebert et al., 2013). Therefore, the effects of inhibiting POR and FDXR may confirm
whether the alternative pathway for pregnenolone synthesis is indeed dependent on CYP450
activity and determine the localization of the particular CYP450 involved. We first ascertained
the expression of POR in the cell lines. Although H295R-S1 adrenal cells have higher POR
expression than glial cells (Figure 26A, B), RNA expression of POR in glial cells is higher than
that of FDXR (Figure 26A, Figure 8C), and POR protein is clearly detectable by immunoblot
(Figure 26B, Figure 17A). However, treatment with diphenyleneiodonium, an inhibitor of POR,
failed to alter pregnenolone production in glial cells under both basal and 22(R)-HC stimulated
conditions (Figure 26C, D).

120

Figure 26: POR expression in brain cells and effect of POR inhibition on pregnenolone
synthesis.
(A) qRT-PCR analysis of POR expression in human cell lines. Gene expression is shown relative
to β-actin. (B) Representative immunoblot of POR in human cell lines, with β-actin as a loading
control. Forty µg total protein were loaded into each lane. (C, D) ELISA measurements of
pregnenolone levels in culture media after MGM-1 cells were pre-treated with different doses of
diphenyleneiodonium for 24 hours under both basal (C) and 22(R)-HC stimulated (D)
conditions. Data are presented as mean ± SD, N=3. Statistics performed compared to the no
treatment group. No significant changes in pregnenolone production were found after
diphenyleneiodonium treatment.

121
These data appear to suggest that pregnenolone synthesis in glial cells is not dependent
on POR and thus not dependent on activity of CYP450s in the endoplasmic reticulum. However,
in addition to POR, diphenyleneiodonium also inhibits other enzymes such as NADPH oxidase,
nitric oxide synthase, and xanthine oxidase, which have more varied subcellular localizations
including the cytoplasm, cytoskeleton, vesicular organelles, nuclei, and other organelles
depending on cell type (Battelli, Polito, Bortolotti, & Bolognesi, 2016; Glass et al., 2006; Park et
al., 2007; Villanueva & Giulivi, 2010). Therefore, more specific inhibition of POR would be
needed to confirm its lack of involvement in pregnenolone synthesis. Thus, we performed siRNA
knockdown of POR in MGM-1 cells and compared its effect to FDXR knockdown (Figure 27).
We were able to achieve > 60% knockdown of both POR and FDXR (Figure 27A-C, E-G).
Knockdown of POR failed to alter pregnenolone synthesis (Figure 27D), supporting the results
observed with diphenyleneiodonium. On the other hand, knockdown of FDXR significantly
decreased pregnenolone synthesis by MGM-1 cells (Figure 27H). Altogether, our results suggest
that our third possibility is most probable: the alternative pathway for pregnenolone production
in the brain involves activity of CYP450 enzyme(s) other than CYP11A1 and the responsible
enzyme(s) is likely localized to the mitochondria as they are dependent on FDXR activity.

122

Figure 27: Effects of POR and FDXR knockdown on pregnenolone synthesis in MGM-1 cells.
siRNA knockdown of POR (A – D) or FDXR (E – H) in MGM-1 cells. Knockdown efficiency
was determined by qRT-PCR (A, E) and immunoblot (B, C, F, G), with expression normalized to
β-actin. (B, F) Representative immunoblots indicating successful knockdown. Twenty µg total
protein were loaded into each lane (C, G) Quantification of protein levels in immunoblots in all
experiments. (D, H) ELISA measurements of secreted pregnenolone in transfected MGM-1 cells.
Data are presented as mean ± SD, N=3. Statistics performed using unpaired t-test. (* p < 0.05, **
p < 0.01, *** p < 0.001)

123
3.4 Discussion
Based on our literature review (Chapter I) and our experimental results (Chapter II),
we proposed three possibilities for an hypothesized alternative pathway for CYP11A1-based
steroid synthesis in brain (Lin & Papadopoulos, 2021): 1) pregnenolone synthesis may be
regulated by reactive oxygen species in glial cells, 2) a second CYP11A1 isoform may be
responsible for producing pregnenolone in the brain, and 3) another CYP450 enzyme other than
CYP11A1 may be involved in synthesizing pregnenolone in the brain. Since there have not been
any previous studies that could prove or disprove these three possibilities, we sought to provide
experimental evidence that would help identify the alternative pathway for pregnenolone
synthesis.
Previous studies in MGM-1 on steroidogenesis showed that DHEA synthesis by these
cells can be dose-dependently induced by FeSO 4 above 0.3 mM (R. C. Brown et al., 2000).
The same trend was also observed in MGM-3 cells, but the dose-dependent increase was not
as clear. Although we hypothesized that pregnenolone synthesis could be increased by
FeSO4 as well, our results from the present study show that FeSO4 had no effect on
pregnenolone production in both MGM-1 and MGM-3 cells. The previous study also
showed that FeSO4 dose dependently increased cellular ROS in MGM-1 cells up to 3 mM.
Here, we observed that FeSO4 only significantly increased cellular ROS at 500 μM in
MGM-1 cells, with no significant increase at higher concentrations (data not shown).
Nevertheless, in combination with our H2O2 results, our data suggests that pregnenolone
synthesis by human glial cells is not dependent on ROS unlike DHEA synthesis. These results
are in contrast to studies in C6 rat glial cells, where 10 mM H2O2 and FeSO 4 both increased
pregnenolone production by at least 2-fold (Cascio et al., 1998). This difference could be

124
due to species variations, as treatments of these two oxidants above 2 mM caused significant
toxicity in MGM-1 and MGM-3 cells. Consequently, we have excluded these high
concentrations from our analyses. Our data here provided evidence against our first
possibility that the alternative pathway is dependent on ROS.
To our knowledge, our study is the first to characterize CYP11A1b. Our qRT-PCR
results suggest that exon 1 of the mRNA transcript for CYP11A1b may be spliced early,
resulting in low CYP11A1b exon 1 expression but a total CYP11A1 expression higher than
the sum of both isoforms. This may also indicate that CYP11A1b contributes to a non-
negligible amount of total CYP11A1 expression in the brain, unlike in the periphery where
CYP11A1a is clearly the dominant isoform. Given that the brain is rich in cholesterol and
that all of CYP11A1b’s amino acid residues are found in CYP11A1a, we speculate that
CYP11A1b could have roles in cholesterol homeostasis that is not steroidogenesis, but more
studies will be needed determine its exact function. We also observed that CYP11A1b is not
strictly localized to the mitochondria unlike CYP11A1a, which is in line with current
knowledge that CYP11A1b lacks the N-terminal mitochondrial-signaling peptide. However,
we were unable to quantify CYP11A1b protein levels in the human brain and in glial cells
due to the lack of unique amino acid peptides in CYP11A1b, lack of specificity of
antibodies that could detect total CYP11A1 protein, and lack of detectable signal in
immunoblots using antibodies specific to CYP11A1a. Nevertheless, transfection of glial
cells with CYP11A1b expression vectors did not alter pregnenolone synthesis while
overexpression of the classical CYP11A1a significantly increased pregnenolone production.
Therefore, our data did not support our second proposed mechanism as CYP11A1b does not
appear to be involved in pregnenolone synthesis.

125
However, our results did support the third possibility that another CYP450 enzyme
with desmolase activity in the brain can be used to synthesize pregnenolone. Although
CYP46A1 could be a candidate given its high expression in brain, involvement in
cholesterol metabolism, and structural similarity to CYP11A1, ketoconazole (KC) inhibits
CYP46A1 activity potentially more effectively than CYP11A1 activity (Mast, Linger, &
Pikuleva, 2013). Our data showed that the CYP450(s) responsible for producing
pregnenolone is(are) not inhibited by AMG and KC. Furthermore, our preliminary studies
indicated that CYP46A1 expression in glial cells is lower than that of CYP11A1 (data not
shown), thus removing CYP46A1 as a likely candidate. Given that knockdown of FDXR
resulted in decreased pregnenolone production while knockdown of POR did not, our data
point toward a mitochondrial CYP450, of which there are seven known enzymes in humans:
CYP11A1, CYP11B1, CYP11B2, CYP24A1, CYP27A1, CYP27B1 and CYP27C1 (Nebert
et al., 2013). Our qRT-PCR data showed that enzymes in the CYP11 family have low
expression in brain; therefore, CYP24A1 and CYP27 family enzymes are more probable
candidates. Although we are currently investigating these enzymes to determine which of
them are responsible for producing pregnenolone in brain, we cannot eliminate the
possibility that an undiscovered isoform of mitochondrial CYP450 may be involved in
synthesizing pregnenolone. Furthermore, CYP1A1, CYP2E1, CYP2D7, CY1B1, and
CYP2U1 have also been reported to have some mitochondrial localization despite being
found in the cytoplasm and/or endoplasmic reticulum as well (UniProt, 2021), which adds to
the list of potential CYP450 candidates. Thus, more extensive investigations into each of the
mitochondrial CYP450s would be needed to elucidate the full alternative pathway for
pregnenolone synthesis in the human brain.

126
IV. Conclusion
Portions of this section are adapted from:
Lin, Y . C., Cheung, G., Porter, E., & Papadopoulos, V . The neurosteroid pregnenolone is
synthesized by a mitochondrial P450 enzyme other than CYP11A1 in human glial cells.
Journal of Biological Chemistry, 102110. doi:10.1016/j.jbc.2022.102110 (In Press)

4.1 Summary of overall findings
Our literature review highlighted the discrepancies regarding the presence of
CYP11A1 in the CNS. Most studies measured CYP11A1 RNA, protein, or activity in the
brain independently, without showing the presence of all three. There is also no consensus
as to which cell type in the brain expresses the most CYP11A1. Given that many anti-
CYP11A1 antibodies cannot produce specific signals in brain cells and tissues, we question
whether the reported presence of CYP11A1 protein is accurate. Thus, we hypothesized that
the brain may be able to synthesize pregnenolone independently of CYP11A1.
We measured extremely low levels of CYP11A1 mRNA in human brain tissues and
cells but could not detect any CYP11A1 protein in human glial cells by immunoblotting.
Our study also confirmed that although commercial antibodies can reliably measure
CYP11A1 protein in the classical steroidogenic tissues that have high CYP11A1 expression,
they often lead to non-specific bands in brain tissues or brain cell lysates. Despite having
very low CYP11A1 expression, glial cells produce detectable levels of pregnenolone, which
are not inhibited by CYP11A1 inhibitors. The ability for hydroxycholesterols to induce
pregnenolone synthesis suggests the presence of a desmolase activity, which was also not
inhibited by CYP11A1 inhibitors.
As the data supports our hypothesis, we attempted to identify the alternative pathway
for pregnenolone synthesis in the brain. We proposed three possibilities for the alternative

127
pathway based on the literature: 1) pregnenolone synthesis is dependent on ROS; 2)
CYP11A1 isoform b is responsible for producing pregnenolone; 3) pregnenolone is
produced by a CYP450 enzyme other than CYP11A1. Our consequent studies revealed that
the third possibility is the most likely mechanism.
Although previous studies have found that ferrous compounds and hydrogen
peroxide can induce steroid production in brain homogenates and brain cells, we did not
observe any alterations in pregnenolone synthesis in glial cells following treatment with
oxidants and antioxidants. The only exception was nitric oxide, which inhibited
pregnenolone production; however, given the lack of correlation between ROS levels and
pregnenolone changes, we concluded that pregnenolone synthesis does not appear to be
dependent on ROS.
Next, we found that CYP11A1 isoform b also does not appear to be involved in
pregnenolone synthesis. We overexpressed both CYP11A1a and CYP11A1b in glial cells
and observed an increase in pregnenolone production only in CYP11A1a+ cells.
Nevertheless, to our knowledge, our study is currently the only one to measure and
characterize CYP11A1b. Our results suggest that total CYP11A1 in the brain is not
comprised purely of CYP11A1a unlike the adrenals and testes, indicating that further
exploration of the roles of different CYP11A1 isoforms in the brain may be valuable. Our
study also highlights the importance of species selection for animal models. Most studies on
CYP11A1 in the brain has been done on rats but rats only have one isoform while mice have
two isoforms similar to humans. This could denote that the findings in rats regarding
Cyp11a1 may not be fully translatable to humans and that mice may be the better animal
model.

128
Finally, our data supported the third proposed possibility that pregnenolone synthesis
involves activity of another CYP450 enzyme, as indicated by the inhibitory effects of nitric
oxide and iron chelators on pregnenolone production in glial cells. To narrow down
potential CYP450 enzymes, we investigated the effects of inhibiting the CYP450 redox
partner POR in the endoplasmic reticulum versus inhibiting the mitochondrial redox partner
FDXR. Both drug- and siRNA-induced inhibition of POR had no effects on pregnenolone
production while FDXR knockdown significantly reduced pregnenolone synthesis by glial
cells. This suggests that the CYP450 enzyme involved is located in the mitochondria.
In sum, our study demonstrates that the human brain and human glial cells express
very low levels of CYP11A1. Our data provides evidence for a CYP11A1-independent
pathway for pregnenolone synthesis that is also independent of ROS, and we propose that
another mitochondrial CYP450 other than CYP11A1 is responsible for synthesizing
pregnenolone in the human brain. This suggests that it may be possible to specifically
modulate CNS steroid synthesis pathways. Given that adverse effects can arise from
systemic administration of steroid-modulating drugs and local administration of steroids is
invasive, the alternative pathway for pregnenolone synthesis may be a promising
investigative direction for CNS-specific neurosteroid-modulating therapies.

129
4.2 Limitations
One of the biggest limitations of our study is that our experiments have been mostly
performed using glial cell lines. Although our findings have been consistent between both
cancerous and non-cancerous cell lines used, we cannot assume that these the cell lines behave
exactly like the cells in the normal human brain. Furthermore, there may be more complicated
regulatory mechanisms for neurosteroid synthesis in the physiological setting that we cannot
mimic in our in vitro cell culture. However, given the exploratory nature of our studies and the
low steroidogenic ability of brain cells, it would be difficult to culture and maintain primary
human brain cells at the scale needed for this study. Obtaining enough fresh human brain tissue
needed to study steroidogenesis would also be unfeasible. Therefore, although the results from
our cell line studies may not be completely representative of the biochemistry and physiology of
the human brain, we can still extrapolate our finding of CYP11A1-independent pregnenolone
synthesis to the human brain to some extent.
Second, it is unclear whether the precursors for pregnenolone are the same as that in
peripheral steroidogenic organs (i.e. cholesterol). The rate of cholesterol synthesis is believed to
be decreased in the mature brain after myelination has completed, with astrocytes being the
major supplier of cholesterol for the brain (Sodero, 2021; J. Zhang & Liu, 2015). Apart from
cholesterol, other sterols such as desmosterol and oxysterols are also found throughout different
regions of the brain (Yutuc et al., 2020). In our preliminary studies, inhibition of cholesterol
synthesis by pravastatin did not lead to decreased pregnenolone synthesis in MGM-1 and NHA
cells and delivering cholesterol through low-density lipoprotein (LDL) also did not alter
pregnenolone production in these cells (data not shown). This contrasts with H295R-S1 cells,
where inhibiting cholesterol synthesis reduced pregnenolone secretion while delivering

130
cholesterol led to increased pregnenolone synthesis. This implies the possibility that
pregnenolone produced by MGM-1 and NHA cells is not dependent on de novo synthesis of
cholesterol and that alternative precursor may be utilized. We attempted to trace the precursors of
pregnenolone using
13
C-labelled acetate and mass spectrometry screening, but due to the low
steroidogenic ability of glial cells and complexity of the sterol biosynthesis process, we were
unable to accurately detect
13
C-labelled pregnenolone and distinguish relevant mass peaks that
could be potential precursors. Therefore, exploration of pregnenolone precursors in the brain
may be a research avenue to pursue in the future when better labeling methods and more
sensitive detection techniques become available.
As shown in our study, CYP11A1 is not responsible for pregnenolone synthesis in brain
cells even though CYP11A1a overexpression through transfection increases pregnenolone
production. This raises the question: what is the function of CYP11A1 in the brain if it is not
pregnenolone synthesis? Our qRT-PCR results showed that both CYP11A1 variant 1 and variant
2 are expressed in the human brain. Although we speculated that the mRNA transcript for
CYP11A1 variant 2 may be truncated early at exon 1, we cannot assume that no additional
splicing variations or truncations occur for CYP11A1 mRNA in the brain. Such potential
truncations may lead to altered CYP11A1 protein structure and folding, hence leading to
inability of anti-CYP11A1 antibodies to produce detectable signals that allow accurate
measurement of CYP11A1 protein levels. Therefore, it is possible that more isoforms of
CYP11A1 have yet to be discovered. We theorize that these CYP11A1 isoforms may be involved
in cholesterol homeostasis in the brain, but more studies would be needed to determine their
exact function. Furthermore, the distribution of CYP11A1 isoforms in the brain is unknown as

131
there are currently no reliable methods to detect the protein for the different CYP11A1 isoforms,
which highlights the need for more accurate and sensitive techniques.

4.3 Future directions
Although we were able to show that a mitochondrial CYP450 enzyme may be
responsible for synthesizing pregnenolone in the brain, our current data is not sufficient to
identify the exact enzyme involved. The seven known mitochondrial CYP450 enzymes in
humans are CYP11A1, CYP11B1, CYP11B2, CYP24A1, CYP27A1, CYP27B1 and
CYP27C1 (Nebert et al., 2013). Out of these seven, the CYP11 family enzymes are unlikely to
be the responsible CYP450 (as shown by their near-undetectable expression in the human brain),
thus leaving CYP24 and CYP27 family enzymes to be potential candidates. CYP24A1,
otherwise known as vitamin D 24-hydroxylase, catalyzes the conversion of 25-hydroxyvitamin D
to 24,25-dihydroxyvitamin D as a mechanism to clear active vitamin D metabolites (Carpenter,
2017). Interestingly, CYP11A1 is also able to metabolize vitamin D (Guryev, Carvalho, Usanov,
Gilep, & Estabrook, 2003), suggesting that there could be active site and substrate similarities
between CYP11A1 and CYP24A1. Moreover, enzymes in the CYP27 family have similar
substrates as well: CYP27A1 converts cholesterol to 27-hydroxycholesterol while CYP27B1 is
involved in vitamin D metabolism alongside CYP24A1 (Carpenter, 2017; Dubrac et al., 2005).
Therefore, we theorize that CYP24 and/or CYP27 family enzymes may be the enzyme(s)
responsible for synthesizing pregnenolone in the brain and we are currently investigating this
possibility. In addition, the enzyme candidate list can be extended to include CYP1A1,
CYP2E1, CYP2D7, CY1B1, and CYP2U1, all of which have some mitochondrial

132
localization despite also being present in the endoplasmic reticulum and/or cytoplasm
(UniProt, 2021). The possibility that these CYP450s may be involved in pregnenolone
synthesis is worth further investigation as well.
As discussed in the previous section, the role of CYP11A1 isoforms in the brain is not
clear. The function of CYP11A1 has been well-established by studies in peripheral steroidogenic
tissues, but the lack of studies on CYP11A1 in the CNS could undermine the assumption that this
enzyme plays the same role in the brain. Indications of potential differences between brain and
peripheral CYP11A1 was already revealed through studies of its promoter region, where
different transcription factors and DNA sequences may be involved in CYP11A1 transcription as
discussed in Section 1.3.5. The disparity becomes increasingly apparent when CYP11A1b is
considered. In the adrenals and testes, CYP11A1b expression is negligible compared to
CYP11A1a. However, in the brain, CYP11A1a does not make up the majority of total CYP11A1
expression, which could imply that whatever role CYP11A1b plays is not trivial. Therefore,
future studies could investigate the distribution of CYP11A1 isoforms throughout the brain,
possibly using non-antibody-based methods such as mass spectrometry. Our MGM-1
CYP11A1a+ and CYP11A1b+ cell lines may also be useful tools for future endeavors to explore
the functions of the CYP11A1 isoforms in the brain.
Lastly, elucidating the full pathway for pregnenolone synthesis in the brain would be
another direction for further research. As discussed in earlier, the precursors to pregnenolone
may not be identical to those in the periphery. Various types of sterols are found in the brain and
they may be used to synthesize pregnenolone instead of cholesterol. Understanding the source of
pregnenolone precursors will further the knowledge of neurosteroidogenesis and lipid
homeostasis in the brain. Furthermore, discovery of a brain-specific pathway for neurosteroid

133
synthesis could lead to new therapeutic targets that will allow neurosteroid modulation
specifically in the brain. With the approval of the drug brexanolone as well as more and more
neurosteroid-based therapies in development, the understanding of how neurosteroids are made
will become increasingly more important.

134
References
Aguiar, M., Masse, R., & Gibbs, B. (2005). Regulation of Cytochrome P450 by Posttranslational
Modification. Drug Metabolism Reviews, 37(2), 379-404. doi:10.1081/dmr-200046136
Ahboucha, S., Jiang, W., Chatauret, N., Mamer, O., Baker, G. B., & Butterworth, R. F. (2008).
Indomethacin improves locomotor deficit and reduces brain concentrations of
neuroinhibitory steroids in rats following portacaval anastomosis. Neurogastroenterol
Motil, 20(8), 949-957. doi:10.1111/j.1365-2982.2008.01132.x
Akwa, Y ., Purdy, R. H., Koob, G. F., & Britton, K. T. (1999). The amygdala mediates the
anxiolytic-like effect of the neurosteroid allopregnanolone in rat. Behavioural Brain
Research, 106(1-2), 119-125.
Apaja, P. M., Harju, K. T., Aatsinki, J. T., Petaja-Repo, U. E., & Rajaniemi, H. J. (2004).
Identification and structural characterization of the neuronal luteinizing hormone receptor
associated with sensory systems. J Biol Chem, 279(3), 1899-1906.
doi:10.1074/jbc.M311395200
Avallone, R., Lucchi, C., Puja, G., Codeluppi, A., Filaferro, M., Vitale, G., . . . Biagini, G.
(2020). BV-2 Microglial Cells Respond to Rotenone Toxic Insult by Modifying
Pregnenolone, 5alpha-Dihydroprogesterone and Pregnanolone Levels. Cells, 9(9).
doi:10.3390/cells9092091
Azeva, T. N., Gilep, A. A., Lepesheva, G. I., Strushkevich, N. V ., & Usanov, S. A. (2001). Site-
directed Mutagenesis of Cytochrome P450scc. II. Effect of Replacement of the Arg425
and Arg426 Residues on the Structural and Functional Properties of the Cytochrome
P450scc. Biochemistry (Mosc), 66(5), 564-575.
Bader, S., Wolf, L., Milenkovic, V . M., Gruber, M., Nothdurfter, C., Rupprecht, R., & Wetzel, C.
H. (2019). Differential effects of TSPO ligands on mitochondrial function in mouse
microglia cells. Psychoneuroendocrinology, 106, 65-76.
doi:10.1016/j.psyneuen.2019.03.029
Bali, A., & Jaggi, A. S. (2014). Multifunctional aspects of allopregnanolone in stress and related
disorders. Prog Neuropsychopharmacol Biol Psychiatry, 48, 64-78.
doi:10.1016/j.pnpbp.2013.09.005
Battelli, M. G., Polito, L., Bortolotti, M., & Bolognesi, A. (2016). Xanthine Oxidoreductase in
Drug Metabolism: Beyond a Role as a Detoxifying Enzyme. Curr Med Chem, 23(35),
4027–4036. doi:10.2174/09298673236661607250
Baulieu, E. E. (1981). STEROID HORMONES IN THE BRAIN: SEVERAL MECHANISMS?
In K. Fuxe, J.-Å. Gustafsson, & L. Wetterberg (Eds.), Steroid Hormone Regulation of the
Brain (pp. 3-14): Pergamon.
Baulieu, E. E., & Robel, P. (1990). Neurosteroids: A new brain function? J Steroid Biochem Mol
Biol, 37(3), 395-403.
Belanger, M., Desjardins, P., Chatauret, N., Rose, C., & Butterworth, R. F. (2005). Mild
hypothermia prevents brain edema and attenuates up-regulation of the astrocytic
benzodiazepine receptor in experimental acute liver failure. J Hepatol, 42(5), 694-699.
doi:10.1016/j.jhep.2004.12.029
Benelli, C., Michel, O., & Michel, R. (1982). Effect of thyroidectomy on pregnenolone and
progesterone biosynthesis in rat adrenal cortex. J Steroid Biochem, 16(6), 749-754.
doi:10.1016/0022-4731(82)90031-0

135
Beyenburg, S., Stoffel-Wagner, B., Watzka, M., Blumcke, I., Bauer, J., Schramm, J., . . . Elger, C.
E. (1999). Expression of cytochrome P450scc mRNA in the hippocampus of patients with
temporal lobe epilepsy. NeuroReport, 10(14), 3067-3070.
Biagini, G., Longo, D., Baldelli, E., Zoli, M., Rogawski, M. A., Bertazzoni, G., & Avoli, M.
(2009). Neurosteroids and epileptogenesis in the pilocarpine model: evidence for a
relationship between P450scc induction and length of the latent period. Epilepsia, 50
Suppl 1, 53-58. doi:10.1111/j.1528-1167.2008.01971.x
Bitran, D., Shiekh, M., & McLeod, M. (1995). Anxiolytic Effect of Progesterone is Mediated by
the Neurosteroid Allopregnanolone at Brain GABAA Receptors. Journal of
Neuroendocrinology, 7(3), 171-177. doi:10.1111/j.1365-2826.1995.tb00744.x
Black, S. M., Harikrishna, J. A., Szklarz, G. D., & Miller, W. L. (1994). The mitochondrial
environment is required for activity of the cholesterol side-chain cleavage enzyme,
cytochrome P450scc. Proceedings of the National Academy of Sciences of the United
States of America, 91(15), 7247-7251. doi:10.1073/pnas.91.15.7247
Borowicz, K. K., Piskorska, B., Banach, M., & Czuczwar, S. J. (2011). Neuroprotective actions
of neurosteroids. Front Endocrinol (Lausanne), 2, 50. doi:10.3389/fendo.2011.00050
Bowlby, M. R. (1993). Pregnenolone sulfate potentiation of N-methyl-D-aspartate receptor
channels in hippocampal neurons. Molecular Pharmacology, 43(5), 813-819.
Brown, E. S., Park, J., Marx, C. E., Hynan, L. S., Gardner, C., Davila, D., . . . Holmes, T. (2014).
A randomized, double-blind, placebo-controlled trial of pregnenolone for bipolar
depression. Neuropsychopharmacology, 39(12), 2867-2873. doi:10.1038/npp.2014.138
Brown, R. C., Cascio, C., & Papadopoulos, V . (2000). Pathways of neurosteroid biosynthesis in
cell lines from human brain: regulation of dehydroepiandrosterone formation by
oxidative stress and beta-amyloid peptide. J Neurochem, 74(2), 847-859.
doi:10.1046/j.1471-4159.2000.740847.x
Brown, R. C., Han, Z., Cascio, C., & Papadopoulos, V . (2003). Oxidative stress-mediated DHEA
formation in Alzheimer's disease pathology. Neurobiol Aging, 24(1), 57-65.
doi:10.1016/s0197-4580(02)00048-9
Burstein, S., Middleditch, B. S., & Gut, M. (1975). Mass Spectrometric Study of the Enzymatic
Conversion of Cholesterol to (22R)-22-Hydroxycholesterol, (20R,22R)-20,22-
Dihydroxycholesterol, and Pregnenolone, and of (22R)-22-Hydroxycholesterol to the
Glycol and Pregnenolone in Bovine Adrenocortical Preparations. J Biol Chem, 250(23),
9028-9037.
Cai, H., Zhou, X., Dougherty, G. G., Reddy, R. D., Haas, G. L., Montrose, D. M., . . . Yao, J. K.
(2018). Pregnenolone-progesterone-allopregnanolone pathway as a potential therapeutic
target in first-episode antipsychotic-naive patients with schizophrenia.
Psychoneuroendocrinology, 90, 43-51. doi:10.1016/j.psyneuen.2018.02.004
Carpenter, T. O. (2017). CYP24A1 loss of function: Clinical phenotype of monoallelic and
biallelic mutations. J Steroid Biochem Mol Biol, 173, 337-340.
doi:10.1016/j.jsbmb.2017.01.006
Carta, M. G., Bhat, K. M., & Preti, A. (2012). GABAergic neuroactive steroids: a new frontier in
bipolar disorders? Behavioural and Brain Functions, 8(61). doi:doi.org/10.1186/1744-
9081-8-61
Cascio, C., Brown, R. C., Liu, Y ., Han, Z., Hales, D. B., & Papadopoulos, V . (2000). Pathways of
dehydroepiandrosterone formation in rat brain glia. J Steroid Biochem Mol Biol, 75(2-3),
177-186. doi:10.1016/s0960-0760(00)00163-1

136
Cascio, C., Prasad, V . V ., Lin, Y . Y ., Lieberman, S., & Papadopoulos, V . (1998). Detection of
P450c17-independent pathways for dehydroepiandrosterone (DHEA) biosynthesis in
brain glial tumor cells. Proc Natl Acad Sci U S A, 95(6), 2862-2867.
doi:10.1073/pnas.95.6.2862
Charlier, T. D., Cornil, C. A., Patte-Mensah, C., Meyer, L., Mensah-Nyagan, A. G., & Balthazart,
J. (2015). Local modulation of steroid action: rapid control of enzymatic activity. Front
Neurosci, 9, 83. doi:10.3389/fnins.2015.00083
Chaudhuri, A. C., Harada, Y ., Shimizu, K., Gut, M., & Dorfman, R. I. (1962). Biosynthesis of
Pregnenolone from 22-Hydroxycholesterol.pdf. The Journal of Biological Chemistry,
237(3), 703-704.
Chavez-Delgado, M. E., Gomez-Pinedo, U., Feria-Velasco, A., Huerta-Viera, M., Castaneda, S.
C., Toral, F. A., . . . Garcia-Estrada, J. (2005). Ultrastructural analysis of guided nerve
regeneration using progesterone- and pregnenolone-loaded chitosan prostheses. J Biomed
Mater Res B Appl Biomater, 74(1), 589-600. doi:10.1002/jbm.b.30243
Chiang, Y . F., Lin, H. T., Hu, J. W., Tai, Y . C., Lin, Y . C., & Hu, M. C. (2011). Differential
regulation of the human CYP11A1 promoter in mouse brain and adrenals. J Cell Physiol,
226(8), 1998-2005. doi:10.1002/jcp.22534
Chien, Y ., Rosal, K., & Chung, B. C. (2017). Function of CYP11A1 in the mitochondria. Mol
Cell Endocrinol, 441, 55-61. doi:10.1016/j.mce.2016.10.030
Cho, M., Lee, T. Y ., Kwak, Y . B., Yoon, Y . B., Kim, M., & Kwon, J. S. (2019). Adjunctive use of
anti-inflammatory drugs for schizophrenia: A meta-analytic investigation of randomized
controlled trials. Aust N Z J Psychiatry, 53(8), 742-759. doi:10.1177/0004867419835028
Choi, Y . H., Fujikawa, T., Lee, J., Reuter, A., & Kim, K. W. (2013). Revisiting the Ventral Medial
Nucleus of the Hypothalamus: The Roles of SF-1 Neurons in Energy Homeostasis. Front
Neurosci, 7, 71. doi:10.3389/fnins.2013.00071
Chung, B. C., Guo, I. C., & Chou, S. J. (1997). Transcriptional regulation of the CYP11A1 and
ferredoxin genes. Steroids, 62(1), 37-42. doi:10.1016/s0039-128x(96)00156-0
Chung, B. C., Matteson, K. J., V outilainen, R., Mohandas, T. K., & Miller, W. L. (1986). Human
cholesterol side-chain cleavage enzyme, P450scc: cDNA cloning, assignment of the gene
to chromosome 15, and expression in the placenta. Proc Natl Acad Sci U S A, 83(23),
8962-8966.
Compagnone, N. A., Bulfone, A., Rubenstein, J. L., & Mellon, S. H. (1995). Expression of the
steroidogenic enzyme P450scc in the central and peripheral nervous systems during
rodent embryogenesis. Endocrinology, 136(6), 2689-2696.
Compagnone, N. A., & Mellon, S. H. (1998). Dehydroepiandrosterone: A potential signalling
molecule for neocortical organization during development. Proc Natl Acad Sci U S A,
95(8), 4678-4683. doi:10.1073/pnas.95.8.4678
Concas, A., Porcu, P., Sogliano, C., Serra, M., Purdy, R. H., & Biggio, G. (2000). Caffeine-
induced increases in the brain and plasma concentrations of neuroactive steroids in the
rat. Pharmacol Biochem Behav, 66(1), 39-45. doi:10.1016/s0091-3057(00)00237-9
Concas, A., Sogliano, C., Porcu, P., Marra, C., Brundu, A., & Biggio, G. (2006). Neurosteroids in
nicotine and morphine dependence. Psychopharmacology (Berl), 186(3), 281-292.
doi:10.1007/s00213-005-0111-7
Corpéchot, C., Synguelakis, M., Talha, S., Axelson, M., Sjövall, J., Vihko, R., . . . Robel, P.
(1983). Pregnenolone and its sulfate ester in the rat brain. Brain Res, 270(1), 119-125.
doi:10.1016/0006-8993(83)90797-7.

137
Costa, E., Auta, J., Guidotti, A., Korneyev, A., & Romeo, E. (1994). The pharmacology of
neurosteroidogenesis. The Journal of Steroid Biochemistry and Molecular Biology, 49(4),
385-389. doi:10.1016/0960-0760(94)90284-4
Craig, I. F., Mason, J. I., Suckling, K. E., & Boyd, G. S. (1982). The Influence of the Side Chain
on Sterol Side-Chain Cleavage in Rat Adrenal Glands. Biochimica et Biophysica Acta,
711(1), 123-127.
Da Pozzo, E., Giacomelli, C., Costa, B., Cavallini, C., Taliani, S., Barresi, E., . . . Martini, C.
(2016). TSPO PIGA Ligands Promote Neurosteroidogenesis and Human Astrocyte Well-
Being. Int J Mol Sci, 17(7). doi:10.3390/ijms17071028
Descalzi, G., Mitsi, V ., Purushothaman, I., Gaspari, S., Avrampou, K., Loh, Y . E., . . . Zachariou,
V . (2017). Neuropathic pain promotes adaptive changes in gene expression in brain
networks involved in stress and depression. Sci Signal., 10(471), eaaj1549.
doi:10.1126/scisignal.aaj1549
Dubrac, S., Lear, S. R., Ananthanarayanan, M., Balasubramaniyan, N., Bollineni, J., Shefer,
S., . . . Erickson, S. K. (2005). Role of CYP27A in cholesterol and bile acid metabolism.
J Lipid Res, 46(1), 76-85. doi:10.1194/jlr.M400219-JLR200
Fan, J., & Papadopoulos, V . (2013). Evolutionary origin of the mitochondrial cholesterol
transport machinery reveals a universal mechanism of steroid hormone biosynthesis in
animals. PLoS One, 8(10), e76701. doi:10.1371/journal.pone.0076701
Fleseriu, M., & Castinetti, F. (2016). Updates on the role of adrenal steroidogenesis inhibitors in
Cushing's syndrome: a focus on novel therapies. Pituitary, 19(6), 643-653.
doi:10.1007/s11102-016-0742-1
Flood, J. F., Morley, J. E., & Roberts, E. (1992). Memory-enhancing effects in male mice of
pregnenolone and steroids metabolically derived from it. Proc Natl Acad Sci U S A,
89(5), 1567-1571.
Fontaine-Lenoir, V ., Chambraud, B., Fellous, A., David, S., Duchossoy, Y ., Baulieu, E. E., &
Robel, P. (2006). Microtubule-associated protein 2 (MAP2) is a neurosteroid receptor.
Proc Natl Acad Sci U S A, 103(13), 4711-4716. doi:10.1073/pnas.0600113103
Furukawa, A., Miyatake, A., Ohnishi, T., & Ichikawa, Y . (1998). Steroidogenic acute regulatory
protein (StAR) transcripts constitutively expressed in the adult rat central nervous
system: colocalization of StAR, cytochrome P-450SCC (CYP XIA1), and 3beta-
hydroxysteroid dehydrogenase in the rat brain. J Neurochem, 71(6), 2231-2238.
Gal, M., & Orly, J. (2014). Selective inhibition of steroidogenic enzymes by ketoconazole in rat
ovary cells. Clin Med Insights Reprod Health, 8, 15-22. doi:10.4137/CMRH.S14036
George, M. S., Guidotti, A., Rubinow, D., Pan, B. S., Mikalauskas, K., & Post, R. M. (1994).
CSF neuroactive steroids in affective disorders: Pregnenolone, progesterone, and DBI.
Biological Psychiatry, 35(10), 775-780. doi:10.1016/0006-3223(94)91139-8
Germelli, L., Da Pozzo, E., Giacomelli, C., Tremolanti, C., Marchetti, L., Wetzel, C. H., . . .
Costa, B. (2021). De novo Neurosteroidogenesis in Human Microglia: Involvement of the
18 kDa Translocator Protein. International Journal of Molecular Sciences, 22(6).
doi:10.3390/ijms22063115
Giatti, S., Boraso, M., Melcangi, R. C., & Viviani, B. (2012). Neuroactive steroids, their
metabolites, and neuroinflammation. J Mol Endocrinol, 49(3), R125-134.
doi:10.1530/JME-12-0127
Giatti, S., Rigolio, R., Diviccaro, S., Falvo, E., Caruso, D., Garcia-Segura, L. M., . . . Melcangi,
R. C. (2020). Sex dimorphism in an animal model of multiple sclerosis: Focus on

138
pregnenolone synthesis. J Steroid Biochem Mol Biol, 199, 105596.
doi:10.1016/j.jsbmb.2020.105596
Glass, M. J., Huang, J., Oselkin, M., Tarsitano, M. J., Wang, G., Iadecola, C., & Pickel, V . M.
(2006). Subcellular Localization of NADPH Oxidase Subunits in Neurons and Astroglia
of the Rat medial Nucleus Tractus Solitarius: Relationship with Tyrosine Hydroxylase
Immunoreactive Neurons. Neuroscience, 143(2), 547–564.
doi:10.1016/j.neuroscience.2006.08.051
Gottfried-Blackmore, A., Sierra, A., Jellinck, P. H., McEwen, B. S., & Bulloch, K. (2009). Brain
Microglia Express Steroid-Converting Enzymes in the Mouse. J Steroid Biochem Mol
Biol, 109(1-2), 96-107. doi:10.1016/j.jsbmb.2007.12.013
Grimm, A., Lejri, I., Halle, F., Schmitt, M., Gotz, J., Bihel, F., & Eckert, A. (2020). Mitochondria
modulatory effects of new TSPO ligands in a cellular model of tauopathies. J
Neuroendocrinol, 32(1), e12796. doi:10.1111/jne.12796
Gursoy, E., Cardounel, A., & Kalimi, M. (2001). Pregnenolone protects mouse hippocampal
(HT-22) cells against glutamate and amyloid beta protein toxicity. Neurochem Res, 26(1),
15-21. doi:10.1023/a:1007668213330
Guryev, O., Carvalho, R. A., Usanov, S., Gilep, A., & Estabrook, R. W. (2003). A pathway for the
metabolism of vitamin D3: Unique hydroxylated metabolites formed during catalysis
with cytochrome P450scc (CYP11A1). Proc Natl Acad Sci U S A, 100(25), 14754-14759.
Guth, L., Zhang, Z., & Roberts, E. (1994). Key Role for Pregnenolone in Combination Therapy
That Promotes Recovery After Spinal Cord Injury. Proc Natl Acad Sci U S A, 91(25),
12308-12312.
Hammer, F., Compagnone, N. A., Vigne, J. L., Bair, S. R., & Mellon, S. H. (2004).
Transcriptional regulation of P450scc gene expression in the embryonic rodent nervous
system. Endocrinology, 145(2), 901-912. doi:10.1210/en.2003-0125
Hu, M. C., Hsu, N. C., Pai, C. I., Wang, C. K., & Chung, B. C. (2001). Functions of the upstream
and proximal steroidogenic factor 1 (SF-1)-binding sites in the CYP11A1 promoter in
basal transcription and hormonal response. Mol Endocrinol, 15(5), 812-818. doi:doi:
10.1210/mend.15.5.0636
Hu, Z. Y ., Bourreau, E., Jung-Testas, I., & Robel, P. (1987). Neurosteroids: oligodendrocyte
mitochondria convert cholesterol to pregnenolone. Proc Natl Acad Sci U S A, 84(23),
8215-8219.
Hu, Z. Y ., Jung-Testas, I., Robel, P., & Baulieu, E. E. (1989). Neurosteroids: Steroidogenesis in
primary cultures of rat glial cells after release of aminoglutethimide blockade.
Biochemical and BIophysical Research Communications, 161(2), 917-922.
Huang, M. C., & Miller, W. L. (2001). Creation and Activity of COS-1 Cells Stably Expressing
the F2 Fusion of the Human Cholesterol Side-Chain Cleavage Enzyme System.
Endocrinology, 142(6), 2569-2576.
Iwahashi, K., Hiroki, S. O., Tsubaki, M., Ohnishi, J., Takeuchi, Y ., & Ichikawa, Y . (1990).
Studies of the immunohistochemical and biochemical localization of the cytochrome
P450-scc-linked monooxygenase system in the adult rat brain. Biochimica et Biophysica
Acta (BBA) - General Subjects, 1035(2), 182-189.
John, M. E., John, M. C., Ashley, P., MacDonald, R., Simpson, E. R., & Waterman, M. R. (1984).
Identification and characterization of cDNA clones specific for cholesterol side-chain
cleavage cytochrome P-450. Proc Natl Acad Sci U S A, 81(18), 5628-5632.

139
Jung-Testas, I., Hu, Z. Y ., Baulieu, E. E., & Robel, P. (1989). Neurosteroids : Biosynthesis of
Pregnenolone and Progesterone in Primary Cultures of Rat Glial Cells. Endocrinology,
125(4), 2083-2091.
Karkkainen, O., Hakkinen, M. R., Auriola, S., Kautiainen, H., Tiihonen, J., & Storvik, M. (2016).
Increased steroid hormone dehydroepiandrosterone and pregnenolone levels in post-
mortem brain samples of alcoholics. Alcohol, 52, 63-70.
doi:10.1016/j.alcohol.2016.03.002
Karri, S., Dertien, J. S., Stocco, D. M., & Syapin, P. J. (2007). Steroidogenic acute regulatory
protein expression and pregnenolone synthesis in rat astrocyte cultures. J
Neuroendocrinol, 19(11), 860-869. doi:10.1111/j.1365-2826.2007.01600.x
Kato-Negishi, M., & Kawahara, M. (2008). Neurosteroids block the increase in intracellular
calcium level induced by Alzheimer’s β-amyloid protein in long-term cultured rat
hippocampal neurons. Neuropsychiatr Dis Treat., 4(1), 209-218. doi:10.2147/ndt.s2059
Kazeto, Y ., Ijiri, S., Adachi, S., & Yamauchi, K. (2006). Cloning and characterization of a cDNA
encoding cholesterol side-chain cleavage cytochrome P450 (CYP11A1): tissue-
distribution and changes in the transcript abundance in ovarian tissue of Japanese eel,
Anguilla japonica, during artificially induced sexual development. J Steroid Biochem Mol
Biol, 99(2-3), 121-128. doi:10.1016/j.jsbmb.2005.12.004
Ke, Y ., Gonthier, R., & Labrie, F. (2017). A sensitive and accurate LC-MS/MS assay with the
derivatization of 1-Amino-4-methylpiperazine applied to serum allopregnanolone,
pregnenolone and androsterone in pre- and postmenopausal women. Steroids, 118, 25-31.
doi:10.1016/j.steroids.2016.11.007
Kim, K. W., Zhao, L., Donato, J., Jr., Kohno, D., Xu, Y ., Elias, C. F., . . . Elmquist, J. K. (2011).
Steroidogenic factor 1 directs programs regulating diet-induced thermogenesis and leptin
action in the ventral medial hypothalamic nucleus. Proc Natl Acad Sci U S A, 108(26),
10673-10678. doi:10.1073/pnas.1102364108
Kim, S. B., Hill, M., Kwak, Y . T., Hampl, R., Jo, D. H., & Morfin, R. (2003). Neurosteroids:
Cerebrospinal fluid levels for Alzheimer's disease and vascular dementia diagnostics. J
Clin Endocrinol Metab, 88(11), 5199-5206. doi:10.1210/jc.2003-030646
Kimoto, T., Tsurugizawa, T., Ohta, Y ., Makino, J., Tamura, H. O., Hojo, Y ., . . . Kawato, S.
(2001). Neurosteroid synthesis by cytochrome p450-containing systems localized in the
rat brain hippocampal neurons: N-methyl-D-aspartate and calcium-dependent synthesis.
Endocrinology, 142(8), 3578–3589.
King, S. R., Manna, P. R., Ishii, T., Syapin, P. J., Ginsberg, S. D., Wilson, K., . . . Lamb, D. J.
(2002). An Essential Component in Steroid Synthesis, the Steroidogenic Acute
Regulatory Protein, Is Expressed in Discrete Regions of the Brain. J Neurosci, 22(24),
10613-10620.
Kohchi, C., Ukena, K., & Tsutsui, K. (1998). Age- and region-specific expressions of the
messenger RNAs encoding for steroidogenic enzymes P450scc, P450c17 and 3β-HSD in
the postnatal rat brain. Brain Research, 801(1-2), 233-238.
Korneyev, A., Pan, B. S., Polo, A., Romeo, E., Guidotti, A., & Costa, E. (1993). Stimulation of
Brain Pregnenolone Synthesis by Mitochondrial Diazepam Binding Inhibitor Receptor
Ligands In Vivo. J Neurochem, 61(4), 1515-1524.
Labombarda, F., Pianos, A., Liere, P., Eychenne, B., Gonzalez, S., Cambourg, A., . . . Guennoun,
R. (2006). Injury elicited increase in spinal cord neurosteroid content analyzed by gas

140
chromatography mass spectrometry. Endocrinology, 147(4), 1847-1859.
doi:10.1210/en.2005-0955
Lai, W. W., Hsiao, P. H., Guiguen, Y ., & Chung, B. C. (1998). Cloning of Zebrafish cDNA for
3beta-hydroxysteroid Dehydrogenase and P450scc. Endocr Res, 24(3-4), 927-931.
Lambeth, J. D., Kitchen, S. E., & Farooqui, A. A. (1982). Cytochrome P-450scc-Substrate
Interactions. The Journal of Biological Chemistry, 257(4), 1876-1884.
Lambeth, J. D., Seybert, D. W., Lancaster, J. R., Salerno, J. C., & Kamin, H. (1982).
Steroidogenic electron transport in adrenal cortex mitochondria. Molecular and Cellular
Biochemistry, 45(1), 13-31.
Lanthier, A., & Patwardhan, V . V . (1986). Sex steroids and 5-en-3 beta-hydroxysteroids in
specific regions of the human brain and cranial nerves. J Steroid Biochem, 25(3), 445-
459. doi:10.1016/0022-4731(86)90259-1.
Lavaque, E., Mayen, A., Azcoitia, I., Tena-Sempere, M., & Garcia-Segura, L. M. (2006). Sex
differences, developmental changes, response to injury and cAMP regulation of the
mRNA levels of steroidogenic acute regulatory protein, cytochrome p450scc, and
aromatase in the olivocerebellar system. J Neurobiol, 66(3), 308-318.
doi:10.1002/neu.20221
Le Goascogne, C., Gouezou, M., Robel, P., Defaye, G., Chambaz, E., Waterman, M. R., &
Baulieu, E. E. (1989). The cholesterol side-chain cleavage complex in human brain white
matter. Journal of Neuroendocrinology, 1(3), 153-156.
Le Goascogne, C., Robel, P., Gouezou, M., Sananes, N., Baulieu, E. E., & Waterman, M. (1987).
Neurosteroids: Cytochrome P-450scc in Rat Brain. Science, 237(4819), 1212-1215.
doi:10.1126/science.3306919
Lejri, I., Grimm, A., Halle, F., Abarghaz, M., Klein, C., Maitre, M., . . . Eckert, A. (2019). TSPO
Ligands Boost Mitochondrial Function and Pregnenolone Synthesis. J Alzheimers Dis,
72(4), 1045-1058. doi:10.3233/JAD-190127
Lepesheva, G. I., Azeva, T. N., Strushkevich, N. V ., Gilep, A. A., & Usanov, S. A. (2000). Site-
directed Mutagenesis of Cytochrome P450scc (CYP11A1). Effect of Lysine Residue
Substitution on Its Structural and Functional Properties. Biochemistry (Mosc), 65(12),
1409-1418.
Lieberman, S. (2008). The generally accepted version of steroidogenesis is not free of
uncertainties: Other tenable and possibly superior renditions may be invented. The
Journal of Steroid Biochemistry and Molecular Biology, 109(1), 197-199.
doi:10.1016/j.jsbmb.2008.02.001
Lieberman, S., & Kaushik, G. (2006). Other conceivable renditions of some of the oxidative
processes used in the biosynthesis of steroid hormones. The Journal of Steroid
Biochemistry and Molecular Biology, 101(1), 31-41. doi:10.1016/j.jsbmb.2006.06.005
Lieberman, S., & Lin, Y . Y . (2001). Reflections on sterol sidechain cleavage process catalyzed by
cytochrome P450(scc). J Steroid Biochem Mol Biol, 78, 1-14.
Liere, P., Pianos, A., Oudinet, J. P., Schumacher, M., & Akwa, Y . (2017). Differential effects of
the 18-kDa translocator protein (TSPO) ligand etifoxine on steroidogenesis in rat brain,
plasma and steroidogenic glands: Pharmacodynamic studies. Psychoneuroendocrinology,
83, 122-134. doi:10.1016/j.psyneuen.2017.05.022
Lin, Y . C., Cheung, G., Espinoza, N., & Papadopoulos, V . (2021). (In press). Function,
Regulation, and Pharmacological Effects of Pregnenolone in the Central Nervous System
Current Opinion in Endocrine and Metabolic Research.

141
Lin, Y . C., & Papadopoulos, V . (2021). Neurosteroidogenic enzymes: CYP11A1 in the central
nervous system. Front Neuroendocrinol, 62, 100925. doi:10.1016/j.yfrne.2021.100925
Liu, J., Rone, M. B., & Papadopoulos, V . (2006). Protein-protein interactions mediate
mitochondrial cholesterol transport and steroid biosynthesis. J Biol Chem, 281(50),
38879-38893. doi:10.1074/jbc.M608820200
Liu, S., Wu, H., Xue, G., Ma, X., Wu, J., Qin, Y ., & Hou, Y . (2013). Metabolic alteration of
neuroactive steroids and protective effect of progesterone in Alzheimer's disease-like rats.
Neural Regen Res, 8(30), 2800-2810. doi:10.3969/j.issn.1673-5374.2013.30.002
MacKenzie, S. M., Clark, C. J., Fraser, R., Gomez-Sanchez, C. E., Connell, J. M., & Davies, E.
(2000). Expression of 11b-hydroxylase and aldosterone synthase genes in the rat brain. J
Molecular Endocrinology, 24.
MacKenzie, S. M., Dewar, D., Stewart, W., Fraser, R., Connell, J. M., & Davies, E. (2008). The
transcription of steroidogenic genes in the human cerebellum and hippocampus: a
comparative survey of normal and Alzheimer's tissue. J Endocrinol, 196(1), 123-130.
doi:10.1677/JOE-07-0427
Maingat, F. G., Polyak, M. J., Paul, A. M., Vivithanaporn, P., Noorbakhsh, F., Ahboucha, S., . . .
Power, C. (2013). Neurosteroid-mediated regulation of brain innate immunity in
HIV/AIDS: DHEA-S suppresses neurovirulence. F ASEB J, 27(2), 725-737.
doi:10.1096/fj.12-215079
Marx, C. E., Keefe, R. S., Buchanan, R. W., Hamer, R. M., Kilts, J. D., Bradford, D. W., . . .
Shampine, L. J. (2009). Proof-of-concept trial with the neurosteroid pregnenolone
targeting cognitive and negative symptoms in schizophrenia. Neuropsychopharmacology,
34(8), 1885-1903. doi:10.1038/npp.2009.26
Marx, C. E., Stevens, R. D., Shampine, L. J., Uzunova, V ., Trost, W. T., Butterfield, M. I., . . .
Lieberman, J. A. (2006). Neuroactive steroids are altered in schizophrenia and bipolar
disorder: relevance to pathophysiology and therapeutics. Neuropsychopharmacology,
31(6), 1249-1263. doi:10.1038/sj.npp.1300952
Marx, C. E., Trost, W. T., Shampine, L. J., Stevens, R. D., Hulette, C. M., Steffens, D. C., . . .
Lieberman, J. A. (2006). The neurosteroid allopregnanolone is reduced in prefrontal
cortex in Alzheimer's disease. Biol Psychiatry, 60(12), 1287-1294.
doi:10.1016/j.biopsych.2006.06.017
Mason, J. I., Arthur, J. R., & Boyd, G. S. (1978). Control of Sterol Metabolism in Rat Adrenal
Mitochondria. Biochem J, 174(3), 1045-1051. doi:10.1042/bj1741045
Mast, N., Annalora, A. J., Lodowski, D. T., Palczewski, K., Stout, C. D., & Pikuleva, I. A.
(2011). Structural basis for three-step sequential catalysis by the cholesterol side chain
cleavage enzyme CYP11A1. J Biol Chem, 286(7), 5607-5613.
doi:10.1074/jbc.M110.188433
Mast, N., Linger, M., & Pikuleva, I. A. (2013). Inhibition and stimulation of activity of purified
recombinant CYP11A1 by therapeutic agents. Mol Cell Endocrinol, 371(1-2), 100-106.
doi:10.1016/j.mce.2012.10.013
Matteson, K. J., Chung, B. C., Urdea, M. S., & Miller, W. L. (1986). Study of Cholesterol Side-
Chain Cleavage (20,22 Desmolase) Deficiency Causing Congenital Lipoid Adrenal
Hyperplasia Using Bovine-Sequence P450scc Oligodeoxyribonucleotide Probes.
Endocrinology, 118(4), 1296-1305.

142
Maurice, T., Su, T. P., & Privat, A. (1998). Sigma1 (sigma 1) receptor agonists and neurosteroids
attenuate B25-35-amyloid peptide-induced amnesia in mice through a common
mechanism. Neuroscience, 83(2), 413-428. doi:10.1016/s0306-4522(97)00405-3
Melcangi, R. C., Caruso, D., Levandis, G., Abbiati, F., Armentero, M. T., & Blandini, F. (2012).
Modifications of neuroactive steroid levels in an experimental model of nigrostriatal
degeneration: potential relevance to the pathophysiology of Parkinson's disease. J Mol
Neurosci, 46(1), 177-183. doi:10.1007/s12031-011-9570-y
Melchior, C. L., & Ritzmann, R. F. (1996). Neurosteroids Block the Memory-Impairing Effects
of Ethanol in Mice. Pharmacol Biochem Behav, 53(1), 51-56. doi:10.1016/0091-
3057(95)00197-2.
Mellon, S. H., & Deschepper, C. F. (1993). Neurosteroid biosynthesis: genes for adrenal
steroidogenic enzymes are expressed in the brain. Brain Research, 629, 283-292.
Mellon, S. H., & Griffin, L. D. (2002). Neurosteroids: biochemistry and clinical significance.
TRENDS in Endocrinology & Metabolism, 13(1), 35-43.
Milenkovic, V . M., Slim, D., Bader, S., Koch, V ., Heinl, E. S., Alvarez-Carbonell, D., . . . Wetzel,
C. H. (2019). CRISPR-Cas9 Mediated TSPO Gene Knockout alters Respiration and
Cellular Metabolism in Human Primary Microglia Cells. Int J Mol Sci, 20(13).
doi:10.3390/ijms20133359
Milivojevic, V ., Covault, J., Angarita, G. A., Siedlarz, K., & Sinha, R. (2019). Neuroactive
steroid levels and cocaine use chronicity in men and women with cocaine use disorder
receiving progesterone or placebo. Am J Addict, 28(1), 16-21. doi:10.1111/ajad.12828
Miller, W. L., & Auchus, R. J. (2011). The molecular biology, biochemistry, and physiology of
human steroidogenesis and its disorders. Endocr Rev, 32(1), 81-151. doi:10.1210/er.2010-
0013
Minamiyama, Y ., Takemura, S., Imaoka, S., Funae, Y ., Tanimoto, Y ., & Inoue, M. (1997).
Irreversible Inhibition of Cytochrome P450 by Nitric Oxide. J Pharmacol Exp Ther.,
283(3), 1479-1485.
Morisaki, M., Duque, C., Ikekawa, N., & Shikita, M. (1980). Substrate Specificity of
Adrenocortical Cytochrome P-450scc--I. Effect of Structural Modification of Cholesterol
Side-Chain on Pregnenolone Production. J Steroid Biochem, 13(5), 545-550.
Morohashi, K., Fujii-Kuriyama, Y ., Okada, Y ., Sogawa, K., Hirose, T., Inayama, S., & Omura, T.
(1984). Molecular cloning and nucleotide sequence of cDNA for mRNA of mitochondrial
cytochrome P-450(SCC) of bovine adrenal cortex. Proc Natl Acad Sci U S A, 81(15),
4647-4651. doi:10.1073/pnas.81.15.4647
Muneoka, K. T., & Takigawa, M. (2002). A neuroactive steroid, pregnenolone, alters the striatal
dopaminergic tone before and after puberty. Neuroendocrinology, 75(5), 288-295.
doi:10.1159/000057338
Murakami, K., Fellous, A., Baulieu, E. E., & Robel, P. (2000). Pregnenolone Binds to
Microtubule-Associated Protein 2 and Stimulates Microtubule Assembly. Proc Natl Acad
Sci U S A, 97(7), 3579-3584.
Murugan, S., Jakka, P., Namani, S., Mujumdar, V ., & Radhakrishnan, G. (2019). The
neurosteroid pregnenolone promotes degradation of key proteins in the innate immune
signaling to suppress inflammation. J Biol Chem, 294(12), 4596-4607.
doi:10.1074/jbc.RA118.005543
Naylor, J. C., Hulette, C. M., Steffens, D. C., Shampine, L. J., Ervin, J. F., Payne, V . M., . . .
Marx, C. E. (2008). Cerebrospinal fluid dehydroepiandrosterone levels are correlated

143
with brain dehydroepiandrosterone levels, elevated in Alzheimer's disease, and related to
neuropathological disease stage. J Clin Endocrinol Metab, 93(8), 3173-3178.
doi:10.1210/jc.2007-1229
Naylor, J. C., Kilts, J. D., Hulette, C. M., Steffens, D. C., Blazer, D. G., Ervin, J. F., . . . Marx, C.
E. (2010). Allopregnanolone levels are reduced in temporal cortex in patients with
Alzheimer's disease compared to cognitively intact control subjects. Biochim Biophys
Acta, 1801(8), 951-959. doi:10.1016/j.bbalip.2010.05.006
Naylor, J. C., Kilts, J. D., Shampine, L. J., Parke, G. J., Wagner, H. R., Szabo, S. T., . . . Marx, C.
E. (2020). Effect of Pregnenolone vs Placebo on Self-reported Chronic Low Back Pain
Among US Military Veterans: A Randomized Clinical Trial. JAMA Netw Open, 3(3),
e200287. doi:10.1001/jamanetworkopen.2020.0287
Nebert, D. W., Wikvall, K., & Miller, W. L. (2013). Human cytochromes P450 in health and
disease. Philos Trans R Soc Lond B Biol Sci, 368(1612), 20120431.
doi:10.1098/rstb.2012.0431
Neunzig, J., & Bernhardt, R. (2014). Dehydroepiandrosterone sulfate (DHEAS) stimulates the
first step in the biosynthesis of steroid hormones. PLoS One, 9(2), e89727.
doi:10.1371/journal.pone.0089727
Nomura, O., Nakabayashi, O., Nishimori, K., & Mizuno, S. (1997). The cDNA Cloning and
Transient Expression of a Chicken Gene Encoding Cytochrome P-450scc. Gene, 185(2),
217-222.
Noorbakhsh, F., Ellestad, K. K., Maingat, F., Warren, K. G., Han, M. H., Steinman, L., . . .
Power, C. (2011). Impaired neurosteroid synthesis in multiple sclerosis. Brain, 134(Pt 9),
2703-2721. doi:10.1093/brain/awr200
Okuyama, E., Okazaki, T., Furukawa, A., Wu, R. F., & Ichikawa, Y . (1996). Molecular Cloning
and Nucleotide Sequences of cDNA Clones of Sheep and Goat Adrenocortical
Cytochromes P450scc (CYP11A1). J Steroid Biochem Mol Biol, 57(3/4), 179-185.
Oonk, R. B., Parker, K. L., Gibson, J. L., & Richards, J. S. (1990). Rat Cholesterol Side-Chain
Cleavage Cytochrome P-450 (P-450scc) Gene. Structure and Regulation by cAMP in
Vitro. J Biol Chem, 265(36), 22392-22401.
Osuji, I. J., Vera-Bolanos, E., Carmody, T. J., & Brown, E. S. (2010). Pregnenolone for cognition
and mood in dual diagnosis patients. Psychiatry Res, 178(2), 309-312.
doi:10.1016/j.psychres.2009.09.006
Owen, D. R., Narayan, N., Wells, L., Healy, L., Smyth, E., Rabiner, E. A., . . . Moore, C. S.
(2017). Pro-inflammatory activation of primary microglia and macrophages increases 18
kDa translocator protein expression in rodents but not humans. J Cereb Blood Flow
Metab, 37(8), 2679-2690. doi:10.1177/0271678X17710182
Papadopoulos, V ., Guarneri, P., Kreuger, K. E., Guidotti, A., & Costa, E. (1992). Pregnenolone
biosynthesis in C6-2B glioma cell mitochondria: regulation by a mitochondrial diazepam
binding inhibitor receptor. Proc Natl Acad Sci U S A, 89(11), 5113-5117.
doi:10.1073/pnas.89.11.5113
Papadopoulos, V ., Lecanu, L., Brown, R. C., Han, Z., & Yao, Z. X. (2006). Peripheral-type
benzodiazepine receptor in neurosteroid biosynthesis, neuropathology and neurological
disorders. Neuroscience, 138(3), 749-756. doi:10.1016/j.neuroscience.2005.05.063
Parajes, S., Griffin, A., Taylor, A. E., Rose, I. T., Miguel-Escalada, I., Hadzhiev, Y ., . . . Krone, N.
(2013). Redefining the initiation and maintenance of zebrafish interrenal steroidogenesis

144
by characterizing the key enzyme cyp11a2. Endocrinology, 154(8), 2702-2711.
doi:10.1210/en.2013-1145
Parajes, S., Loidi, L., Reisch, N., Dhir, V ., Rose, I. T., Hampel, R., . . . Krone, N. (2010).
Functional consequences of seven novel mutations in the CYP11B1 gene: four mutations
associated with nonclassic and three mutations causing classic 11{beta}-hydroxylase
deficiency. J Clin Endocrinol Metab, 95(2), 779-788. doi:10.1210/jc.2009-0651
Park, S. E., Song, J. D., Kim, K. M., Park, Y . M., Kim, N. D., Yoo, Y . H., & Park, Y . C. (2007).
Diphenyleneiodonium induces ROS-independent p53 expression and apoptosis in human
RPE cells. FEBS Lett, 581(2), 180-186. doi:10.1016/j.febslet.2006.12.006
Patte-Mensah, C., Kappes, V ., Freund-Mercier, M. J., Tsutsui, K., & Mensah-Nyagan, A. G.
(2003). Cellular distribution and bioactivity of the key steroidogenic enzyme, cytochrome
P450side chain cleavage, in sensory neural pathways. J Neurochem, 86(5), 1233-1246.
doi:10.1046/j.1471-4159.2003.01935.x
Patte-Mensah, C., Kibaly, C., Boudard, D., Schaeffer, V ., Béglé, A., Saredi, S., . . . Mensah-
Nyagan, A. G. (2006). Neurogenic pain and steroid synthesis in the spinal cord. J Mol
Neurosci, 28(1), 17-31. doi:10.1385/JMN/28:01:17
Payne, A. H., & Hales, D. B. (2004). Overview of steroidogenic enzymes in the pathway from
cholesterol to active steroid hormones. Endocr Rev, 25(6), 947-970. doi:10.1210/er.2003-
0030
Pesaresi, M., Maschi, O., Giatti, S., Garcia-Segura, L. M., Caruso, D., & Melcangi, R. C. (2010).
Sex differences in neuroactive steroid levels in the nervous system of diabetic and non-
diabetic rats. Horm Behav, 57(1), 46-55. doi:10.1016/j.yhbeh.2009.04.008
Pozzo, E. D., Tremolanti, C., Costa, B., Giacomelli, C., Milenkovic, V . M., Bader, S., . . .
Martini, C. (2019). Microglial Pro-Inflammatory and Anti-Inflammatory Phenotypes Are
Modulated by Translocator Protein Activation. Int J Mol Sci, 20(18).
doi:10.3390/ijms20184467
Prasad, V . V ., Vegesna, S. R., Welch, M., & Lieberman, S. (1994). Precursors of the
neurosteroids. Proc Natl Acad Sci U S A, 91(8), 3220-3223. doi:10.1073/pnas.91.8.3220
Qaiser, M. Z., Dolman, D. E. M., Begley, D. J., Abbott, N. J., Cazacu-Davidescu, M., Corol, D.
I., & Fry, J. P. (2017). Uptake and metabolism of sulphated steroids by the blood-brain
barrier in the adult male rat. J Neurochem, 142(5), 672-685. doi:10.1111/jnc.14117
Ratner, M. H., Kumaresan, V ., & Farb, D. H. (2019). Neurosteroid Actions in Memory and
Neurologic/Neuropsychiatric Disorders. Front Endocrinol (Lausanne), 10, 169.
doi:10.3389/fendo.2019.00169
Reddy, D. S. (2011). Role of anticonvulsant and antiepileptogenic neurosteroids in the
pathophysiology and treatment of epilepsy. Front Endocrinol (Lausanne), 2, 38.
doi:10.3389/fendo.2011.00038
Ritsner, M., Maayan, R., Gibel, A., & Weizman, A. (2007). Differences in blood pregnenolone
and dehydroepiandrosterone levels between schizophrenia patients and healthy subjects.
Eur Neuropsychopharmacol, 17(5), 358-365. doi:10.1016/j.euroneuro.2006.10.001
Ritsner, M. S., Bawakny, H., & Kreinin, A. (2014). Pregnenolone treatment reduces severity of
negative symptoms in recent-onset schizophrenia: an 8-week, double-blind, randomized
add-on two-center trial. Psychiatry Clin Neurosci, 68(6), 432-440. doi:10.1111/pcn.12150
Robel, P., Bourreau, E., Corpechot, C., Dang, D. C., Halberg, F., Clarke, C., . . . Baulieu, E. E.
(1987). Neuro-steroids 3 beta-hydroxy-delta 5-derivatives in rat and monkey brain.
Journal of Steroid Biochemistry, 27(2-6), 649-655.

145
Romano, S., Mitro, N., Giatti, S., Diviccaro, S., Pesaresi, M., Spezzano, R., . . . Melcangi, R. C.
(2018). Diabetes induces mitochondrial dysfunction and alters cholesterol homeostasis
and neurosteroidogenesis in the rat cerebral cortex. J Steroid Biochem Mol Biol, 178,
108-116. doi:10.1016/j.jsbmb.2017.11.009
Romieu, P., Martin-Fardon, R., Bowen, W. D., & Maurice, T. (2003). Sigma 1 receptor-related
neuroactive steroids modulate cocaine-induced reward. J Neurosci, 23(9), 3572-3576.
doi:10.1523/JNEUROSCI.23-09-03572.2003
Roscetti, G., Del Carmine, R., Trabucchi, M., Massotti, M., Purdy, R. H., & Barbaccia, M. L.
(1998). Modulation of neurosteroid synthesis/accumulation by L-ascorbic acid in rat
brain tissue: inhibition by selected serotonin antagonists. J Neurochem, 71(3), 1108-1117.
doi:10.1046/j.1471-4159.1998.71031108.x
Roselli, C. E., Jorgensen, E. Z., Doyle, M. W., & Rønnekleiv, O. K. (1997). Expression of the
orphan receptor steroidogenic factor-1 mRNA in the rat medial basal hypothalamus.
Brain Res Mol Brain Res, 44(1), 66-72. doi:10.1016/s0169-328x(96)00187-8
Roy, S., Sipthorp, J., Mahata, B., Pramanik, J., Hennrich, M. L., Gavin, A. C., . . . Teichmann, S.
A. (2021). CLICK-enabled analogues reveal pregnenolone interactomes in cancer and
immune cells. iScience, 24(5), 102485. doi:10.1016/j.isci.2021.102485
Rupprecht, R., Papadopoulos, V ., Rammes, G., Baghai, T. C., Fan, J., Akula, N., . . . Schumacher,
M. (2010). Translocator protein (18 kDa) (TSPO) as a therapeutic target for neurological
and psychiatric disorders. Nat Rev Drug Discov, 9(12), 971-988. doi:10.1038/nrd3295
Sanne, J. L., & Krueger, K. E. (1995). Expression of cytochrome P450 side-chain cleavage
enzyme and 3 beta-hydroxysteroid dehydrogenase in the rat central nervous system: a
study by polymerase chain reaction and in situ hybridization. Journal of neurochemistry,
65(2), 528-536. doi:10.1046/j.1471-4159.1995.65020528.x
Selye, H. (1941). Anesthetic Effect of Steroid Hormones. Experimental Biology and Medicine,
46(1), 116-121. doi:10.3181/00379727-46-11907
Simonart, T., Degraef, C., Andrei, G., Mosselmans, R., Hermans, P., Van V ooren, J. P., . . .
Heenen, M. (2000). Iron chelators inhibit the growth and induce the apoptosis of Kaposi's
sarcoma cells and of their putative endothelial precursors. J Invest Dermatol, 115(5), 893-
900. doi:10.1046/j.1523-1747.2000.00119.x
Simpson, E. R., & Boyd, G. S. (1967). The Cholesterol Side-Chain Cleavage System of Bovine
Adrenal Cortex. European Journal of Biochemistry, 2(3), 237-285.
Siraki, A. G., Bonini, M. G., Jiang, J., Ehrenshaft, M., & Mason, R. P. (2007).
Aminoglutethimide-induced protein free radical formation on myeloperoxidase: a
potential mechanism of agranulocytosis. Chemical research in toxicology, 20(7), 1038-
1045. doi:10.1021/tx6003562
Slominski, A. T., Li, W., Kim, T. K., Semak, I., Wang, J., Zjawiony, J. K., & Tuckey, R. C.
(2015). Novel activities of CYP11A1 and their potential physiological significance. J
Steroid Biochem Mol Biol, 151, 25-37. doi:10.1016/j.jsbmb.2014.11.010
Smith, C. C., Gibbs, T. T., & Farb, D. H. (2014). Pregnenolone sulfate as a modulator of synaptic
plasticity. Psychopharmacology (Berl), 231(17), 3537-3556. doi:10.1007/s00213-014-
3643-x
Sodero, A. O. (2021). 24S-hydroxycholesterol: Cellular effects and variations in brain diseases. J
Neurochem, 157(4), 899-918. doi:10.1111/jnc.15228
Stoffel-Wagner, B. (2001). Neurosteroid metabolism in the human brain. Eur J Endrocrinol,
145(6), 669-679.

146
Stromstedt, M., & Waterman, M. R. (1995). Messenger RNAs encoding steroidogenic enzymes
are expressed in rodent brain. Molecular Brain Research, 34(1), 75-88.
Strott, C. A. (1977). A pregnenolone-binding protein in soluble fraction of guinea pig adrenal
cortex. Journal of Biological Chemistry, 252(2), 464-470. doi:10.1016/s0021-
9258(17)32739-4
Strushkevich, N., MacKenzie, F., Cherkesova, T., Grabovec, I., Usanov, S., & Park, H. W.
(2011). Structural basis for pregnenolone biosynthesis by the mitochondrial
monooxygenase system. Proc Natl Acad Sci U S A, 108(25), 10139-10143.
doi:10.1073/pnas.1019441108
Ukena, K., Usui, M., Kohchi, C., & Tsutsui, K. (1998). Cytochrome P450 side-chain cleavage
enzyme in the cerebellar Purkinje neuron and its neonatal change in rats. Endocrinology,
139(1), 137-147.
UniProt, C. (2021). UniProt: the universal protein knowledgebase in 2021. Nucleic Acids Res,
49(D1), D480-D489. doi:10.1093/nar/gkaa1100
Usanov, S. A., Graham, S. E., Lepesheva, G. I., Azeva, T. N., Strushkevich, N. V ., Gilep, A.
A., . . . Peterson, J. A. (2002). Probing the Interaction of Bovine Cytochrome P450scc
(CYP11A1) With Adrenodoxin: Evaluating Site-Directed Mutations by Molecular
Modeling. Biochemistry, 41(26), 8310-8320.
Vallee, M., Vitiello, S., Bellocchio, L., Hebert-Chatelain, E., Monlezun, S., Martin-Garcia,
E., . . . Piazza, P. V . (2014). Pregnenolone can protect the brain from cannabis
intoxication. Science, 343(6166), 94-98. doi:10.1126/science.1243985
Vanden Bossche, H. (1992). Inhibitors of P450-dependent steroid biosynthesis: From research to
medical treatment. J Steroid Biochem Mol Biol, 43(8), 1003-1021.
Venugopal, S., Martinez-Arguelles, D. B., Chebbi, S., Hullin-Matsuda, F., Kobayashi, T., &
Papadopoulos, V . (2016). Plasma Membrane Origin of the Steroidogenic Pool of
Cholesterol Used in Hormone-induced Acute Steroid Formation in Leydig Cells. J Biol
Chem, 291(50), 26109-26125. doi:10.1074/jbc.M116.740928
Verleye, M., Akwa, Y ., Liere, P., Ladurelle, N., Pianos, A., Eychenne, B., . . . Gillardin, J. M.
(2005). The anxiolytic etifoxine activates the peripheral benzodiazepine receptor and
increases the neurosteroid levels in rat brain. Pharmacol Biochem Behav, 82(4), 712-720.
doi:10.1016/j.pbb.2005.11.013
Vickery, L. E. (1993). Cholesterol Side Chain Cleavage Cytochrome P450 (P450scc). In J. B.
Schenkman & H. Greim (Eds.), Cytochrome P450 (pp. 651-665). Berlin, Heidelberg:
Springer Berlin Heidelberg.
Vilchis, F., Chavez, B., Larrea, F., Timossi, C., & Montiel, F. (2002). The cDNA Cloning and
Tissue Expression of the Cytochrome P450scc From Syrian Hamster (Mesocricetus
Auratus). Gen Comp Endocrinol, 126(3), 279-286.
Villanueva, C., & Giulivi, C. (2010). Subcellular and cellular locations of nitric oxide synthase
isoforms as determinants of health and disease. Free Radic Biol Med, 49(3), 307-316.
doi:10.1016/j.freeradbiomed.2010.04.004
Walton, N., & Maguire, J. (2019). Allopregnanolone-based treatments for postpartum depression:
Why/how do they work? Neurobiol Stress, 11, 100198. doi:10.1016/j.ynstr.2019.100198
Wang, F., Flanagan, J., Su, N., Wang, L. C., Bui, S., Nielson, A., . . . Luo, Y . (2012). RNAscope:
a novel in situ RNA analysis platform for formalin-fixed, paraffin-embedded tissues. J
Mol Diagn, 14(1), 22-29. doi:10.1016/j.jmoldx.2011.08.002

147
Wang, M. (2011). Neurosteroids and GABA-A Receptor Function. Front Endocrinol (Lausanne),
2, 44. doi:10.3389/fendo.2011.00044
Warner, M., & Gustafsson, J. A. (1995). Cytochrome P450 in the Brain: Neuroendocrine
Functions. Front Endocrinol, 16(3), 224-236.
Warner, M., & Gustafsson, J. A. (1999). Cytochrome P450 in the Central Nervous System. In E.
E. Baulieu, P. Robel, & M. Schumacher (Eds.), Neurosteriods: A New Regulatory
Function in the Nervous System (Contemporary Endocrinology) (V ol. 16). Totowa, NJ:
Humana Press.
Warner, M., Tollet, P., Stromstedt, M., Carlstrom, K., & Gustafsson, J. A. (1989). Endocrine
regulation of cytochrome P-450 in the rat brain and pituitary gland. J Endocrinol, 122(1),
341-349.
Watzka, M., Bidlingmaier, F., Schramm, J., Klingmuller, D., & Stoffel-Wagner, B. (1999). Sex-
and Age-Specific Differences in Human Brain CYP11A1 mRNA Expression. Journal of
Neuroendocrinology, 11, 901-905.
Weill-Engerer, S., David, J. P., Sazdovitch, V ., Liere, P., Eychenne, B., Pianos, A., . . . Akwa, Y .
(2002). Neurosteroid quantification in human brain regions: comparison between
Alzheimer's and nondemented patients. J Clin Endocrinol Metab, 87(11), 5138-5143.
doi:10.1210/jc.2002-020878
Witzig, M., Grimm, A., Schmitt, K., Lejri, I., Frank, S., Brown, S. A., & Eckert, A. (2020).
Clock-Controlled Mitochondrial Dynamics Correlates with Cyclic Pregnenolone
Synthesis. Cells, 9(10). doi:10.3390/cells9102323
Won, K. J., Lee, K. P., Yu, S., Lee, D., Lee, D. Y ., Lee, H. M., . . . Kim, B. (2015). Ketoconazole
induces apoptosis in rat cardiomyocytes through reactive oxygen species-mediated parkin
overexpression. Arch Toxicol, 89(10), 1871-1880.
Wong, P., Chang, C. C., Marx, C. E., Caron, M. G., Wetsel, W. C., & Zhang, X. (2012).
Pregnenolone rescues schizophrenia-like behavior in dopamine transporter knockout
mice. PLoS One, 7(12), e51455. doi:10.1371/journal.pone.0051455
Wu, H. S., Lin, H. T., Wang, C. K., Chiang, Y . F., Chu, H. P., & Hu, M. C. (2007). Human
CYP11A1 promoter drives Cre recombinase expression in the brain in addition to
adrenals and gonads. Genesis, 45(2), 59-65. doi:10.1002/dvg.20266
Xu, Q., Song, Y ., Chen, Y ., Liu, R., Zhang, Y ., Li, Y ., . . . Chen, G. (2015). Molecular cloning and
expression patterns of the cholesterol side chain cleavage enzyme (CYP11A1) gene
during the reproductive cycle in goose (Anas cygnoides). J Anim Sci Biotechnol, 6, 54.
doi:10.1186/s40104-015-0053-9
Yablokov, E. O., Sushko, T. A., Kaluzhskiy, L. A., Kavaleuski, A. A., Mezentsev, Y . V ., Ershov, P.
V ., . . . Strushkevich, N. V . (2021). Substrate-induced modulation of protein-protein
interactions within human mitochondrial cytochrome P450-dependent system. J Steroid
Biochem Mol Biol, 208, 105793. doi:10.1016/j.jsbmb.2020.105793
Yilmaz, C., Karali, K., Fodelianaki, G., Gravanis, A., Chavakis, T., Charalampopoulos, I., &
Alexaki, V . I. (2019). Neurosteroids as regulators of neuroinflammation. Front
Neuroendocrinol, 55, 100788. doi:10.1016/j.yfrne.2019.100788
Yu, L., Romero, D. G., Gomez-Sanchez, C. E., & Gomez-Sanchez, E. P. (2002). Steroidogenic
enzyme gene expression in the human brain. Molecular and Cellular Endocrinology, 190,
9-17.
Yutuc, E., Angelini, R., Baumert, M., Mast, N., Pikuleva, I., Newton, J., . . . Griffiths, W. J.
(2020). Localization of sterols and oxysterols in mouse brain reveals distinct spatial

148
cholesterol metabolism. Proc Natl Acad Sci U S A, 117(11), 5749-5760.
doi:10.1073/pnas.1917421117
Zhang, J., & Liu, Q. (2015). Cholesterol metabolism and homeostasis in the brain. Protein Cell,
6(4), 254-264. doi:10.1007/s13238-014-0131-3
Zhang, L. M., Zhao, N., Guo, W. Z., Jin, Z. L., Qiu, Z. K., Chen, H. X., . . . Li, Y . F. (2014).
Antidepressant-like and anxiolytic-like effects of YL-IPA08, a potent ligand for the
translocator protein (18 kDa). Neuropharmacology, 81, 116-125.
doi:10.1016/j.neuropharm.2013.09.016
Zhang, M., Liu, J., Zhou, M. M., Wu, H., Hou, Y ., Li, Y . F., . . . Wan, Y . (2017). Anxiolytic
effects of hippocampal neurosteroids in normal and neuropathic rats with spared nerve
injury. J Neurochem, 141(1), 137-150. doi:10.1111/jnc.13965
Zhang, M., Liu, J., Zhou, M. M., Wu, H., Hou, Y ., Li, Y . F., . . . Wan, Y . (2016). Elevated
Neurosteroids in the Lateral Thalamus Relieve Neuropathic Pain in Rats with Spared
Nerve Injury. Neurosci Bull, 32(4), 311-322. doi:10.1007/s12264-016-0044-7
Zhang, P., Rodriguez, H., & Mellon, S. H. (1995). Transcriptional regulation of P450scc gene
expression in neural and steroidogenic cells: implications for regulation of
neurosteroidogenesis. Molecular Endocrinology, 9(11), 1571-1582.
Zhu, T. S., & Glaser, M. (2008). Regulatory role of cytochrome P450scc and pregnenolone in
myelination by rat Schwann cells. Mol Cell Biochem, 313(1-2), 79-89.
doi:10.1007/s11010-008-9745-1
Zwain, I. H., & Yen, S. S. (1999). Neurosteroidogenesis in astrocytes, oligodendrocytes, and
neurons of cerebral cortex of rat brain. Endocrinology, 140(8), 3843-3852.
doi:10.1210/endo.140.8.6907

Online Database References
Gene [Internet]. Bethesda (MD): National Library of Medicine (US), National Center for
Biotechnology Information; [1988] - . Gene ID: 1583. CYP11A1 cytochrome P450
family 11 subfamily A member 1 [ Homo sapiens (human) ]. [Cited 2021 Feb 12].
Available from: https://www.ncbi.nlm.nih.gov/gene/1583
Gene [Internet]. Bethesda (MD): National Library of Medicine (US), National Center for
Biotechnology Information; [1988] - . Gene ID: 13070. Cyp11a1 cytochrome P450,
family 11, subfamily a, polypeptide 1 [ Mus musculus (house mouse) ]. [Cited 2021 Feb
14]. Available from: https://www.ncbi.nlm.nih.gov/gene/13070
Gene [Internet]. Bethesda (MD): National Library of Medicine (US), National Center for
Biotechnology Information; [1988] - . Gene ID: 29680. Cyp11a1 cytochrome P450,
family 11, subfamily a, polypeptide 1 [ Rattus norvegicus (Norway rat) ]. [Cited 2021
Feb 14]. Available from: https://www.ncbi.nlm.nih.gov/gene/29680
Allen Institute for Brain Science. Multiple Cortical Areas – SMART-SEQ (2019). [Cited 2021
Jan 29]. Available from: https://celltypes.brain-map.org/rnaseq/human_ctx_smart-seq
Yao, Z., et al. (2020). "A taxonomy of transcriptomic cell types across the isocortex and
hippocampal formation." bioRxiv: Allen Institute for Brain Science. Whole Cortex &
Hippocampus - 10X Genomics (2020) with 2010X-SMART-SEQ Taxonomy (2020).
Available from: https://celltypes.brain-map.org/rnaseq/mouse_ctx-hip_2010x

149
Uhlen, M., et al. (2015). "Proteomics. Tissue-based map of the human proteome." Science
347(6220): 1260419. CYP11A1, Human Protein Atlas. Available from
http://www.proteinatlas.org.

Software Used
Images in Figures 1, 2, and 4 were illustrated using Krita 4.4.0. Chemical structures in Figures 2
and 6 were made using PubChem Sketcher V2.4 (available from:
https://pubchem.ncbi.nlm.nih.gov//edit3/index.html). All graphs were generated using GraphPad
Prism 9.2.0. Final figures were generated using Microsoft PowerPoint.

Abstract (if available)

Abstract The central and peripheral nervous systems synthesize steroid hormones, called neurosteroids, which have physiological and modulatory functions in the nervous system. One such neurosteroid is pregnenolone, the precursor to all other neurosteroids. In classical steroidogenesis, pregnenolone is synthesized from cholesterol by the CYP11A1 enzyme. However, there has not been sufficient evidence to conclude that the steroidogenesis pathway is identical between the central nervous system (CNS) and the periphery, given that pregnenolone is one of the most abundant neurosteroids in the brain but CYP11A1 is difficult to detect in the CNS. Therefore, this study aims to provide experimental evidence for whether CYP11A1 is used by the human brain to produce pregnenolone.
We found that the expression of CYP11A1 in the human brain tissues and cells is more than 1000-times lower than that in the adrenals. We did not observe detectable CYP11A1 protein by immunoblotting in four human glial cell lines. However, the glial cells produced detectable levels of pregnenolone that was not inhibited by the CYP11A1 inhibitor DL-aminoglutethimide or the non-specific CYP450 inhibitor ketoconazole. Pregnenolone synthesis was increased with addition of hydroxycholesterols, suggesting the involvement of a desmolase activity. These results suggest that pregnenolone synthesis in the human brain does not require CYP11A1 activity.
Thus, we explored three possibilities for the CYP11A1-independent alternative pathway: 1) pregnenolone synthesis is dependent on reactive oxygen species; 2) pregnenolone is synthesized by the second CYP11A1 isoform; 3) pregnenolone is produced by another CYP450 enzyme. Our data did not support the first and second proposed mechanisms, as treatment with oxidants and antioxidants as well as overexpression of CYP11A1 isoform b had no significant effects on pregnenolone synthesis. In support of the third proposed mechanism, we found that nitric oxide, iron chelators, and knockdown of the mitochondrial CYP450 co-factor ferredoxin reductase significantly inhibited pregnenolone production. Altogether, our data suggest that pregnenolone is synthesized by a mitochondrial cytochrome P450 enzyme other than CYP11A1 in the human brain.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Role of steroidogenic acute regulatory protein (STAR) and sterol carrier protein-x (SCPx) in the transport of cholesterol and other lipids

PDF

Eicosanoid regulation of spermatogonial stem cell development

PDF

Explore the influence of CYP21A2 silencing on H295R human adrenal cortical carcinoma cell gene expression function

PDF

Identification of functional pathways altered in testis by in utero exposure to low doses of genistein and DEHP

PDF

The role of translocator protein (TSPO) in mediating the effects of mitotane in human adrenocortical carcinoma cells

PDF

Role of translocator protein (18 kDa) in cell function and metabolism

PDF

Effects of endocrine disrupting chemicals and pharmaceuticals on rodent immature Sertoli cell functions

PDF

Disruption of immature Sertoli cell functions by endocrine disruptors and analgesic drugs

PDF

Fetal exposure to genistein and DEHP mixtures alters the expression of genes involved in critical testicular functions in adult rats

PDF

Exploring the impact of elevated bile acids in metabolic dysfunction- associated steatotic liver disease in vitro

PDF

MAO a deficient mice exhibit an altered immune system in the brain and prostate

PDF

Energy metabolism and inflammation in brain aging: significance of age-dependent astrocyte metabolic-redox profile

PDF

Dihydromyricetin improves social isolation anxiety induced synaptic loss and astrogliosis in mice hippocampus

PDF

Mitochondrial dynamics regulate Leydig cell health and integrity

PDF

Effect of acetaminophen and ibuprofen on spermatogenesis and cell signaling mechanisms

PDF

Regulation of mitochondrial bioenergetics via PTEN (phosphatase and tensin homolog deleted on chromosome 10)/estrogen-related receptor alpha (ERRα) signaling

PDF

Metabolic shift in lung alveolar cell mitochondria after exposure to environmental toxicants

PDF

C-jun N-terminal Kinase (JNK) mediated inhibition of Pyruvate Dehydrogenase (PDH) activity and its effect on mitochondrial metabolism during brain aging

PDF

Development of glycine and GABAA ultra-sensitive ethanol receptors (USERs) as novel tools for alcohol and brain research

PDF

Role of purinergic P2X7 receptors in inflammatory responses in the brain and liver: a study using a mouse model of chronic ethanol and high-fat diet exposure

Asset Metadata

Creator Lin, Yiqi Christina (author)

Core Title Reevaluation of the pregnenolone biosynthesis pathway in the human brain

School School of Pharmacy

Degree Doctor of Philosophy

Degree Program Molecular Pharmacology and Toxicology

Degree Conferral Date 2022-08

Publication Date 07/19/2022

Defense Date 06/08/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag brain,CYP11A1,CYP450,glia,human brain,neurosteroid,OAI-PMH Harvest,pregnenolone,steroidogenesis

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Papadopoulos, Vassilios (committee chair), Culty, Martine (committee member), Davies, Daryl (committee member)

Creator Email yiqichrl@usc.edu,yqchristinalin@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111373211

Unique identifier UC111373211

Legacy Identifier etd-LinYiqiChr-10858

Document Type Dissertation

Format application/pdf (imt)

Rights Lin, Yiqi Christina

Type texts

Source 20220719-usctheses-batch-955 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu