Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Extracellular heat shock protein-90alpha (eHsp90α): mechanisms of secretion, quantitation and function
(USC Thesis Other)
Extracellular heat shock protein-90alpha (eHsp90α): mechanisms of secretion, quantitation and function
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Extracellular Heat Shock Protein-90alpha (eHsp90α):
Mechanisms of Secretion, Quantitation and Function
by
Cheng Chang
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
(CRANIOFACIAL BIOLOGY)
May 2024
Copyright 2024 Cheng Chang
ii
Acknowledgements
I want to acknowledge Dr. Wei Li for his support. I want to thank Dr. Wei Li for his
patient mentoring and guidance over the past seven years. I want to thank the former
members of Dr. Li’s lab including Dr. Jiacong Guo, Dr. Xin Tang for their advice, help and
company. I want to thank Dr. Mei Chen and members of her lab including Jon Cogan and
Yinping Hou for their selfless help. I also want to thank Dr. Yanzhuang Wang of the UMICH
and Dr. Chengyu Liang of the Wistar Institute for their generous provision of experimental
materials. Also, I would like to thank my committee members Dr. Jianfu Chen, Dr. Jian Xu,
Dr. Mark Frey, Dr. Dechen Lin and my former committee member Dr. Baruch Frenkel. And
finally, I want to thank my parents for their endless support, encouragement, and love
iii
TABLE OF CONTENTS
Acknowledgements……………………………………………………………………………...ii
List of figures……………………………………………………………………………………..v
Abstract………………………………………………………………………………………....viii
Chapter 1: INTRODUCTION…………………………………………………………………...1
1.1 Cytosolic heat shock protein 90 known as a chaperon……….…………….1
1.1.1 The chaperone cycle……………………………………………………...1
1.1.2 Differences between Hsp90α and Hsp90β……………………………..2
1.1.3 Distinct roles for Hsp90α and Hsp90β during mouse development…4
1.2 Extracellular Heat shock protein 90 alpha (eHsp90α)……………………..5
1.2.1 Discovery of eHsp90α……………………………………………………5
1.2.2 The two main in vitro functions of eHsp90α……………………….……6
1.2.3 Mechanism of action by eHsp90α……………………………………....7
1.2.4 Two evolutionary conserved lysine residues distinguish eHsp90α
and eHsp90β for extracellular functions………………………………..9
1.2.5 Roles of eHsp90α during wound healing and tumor progression…...10
1.2.6 The mechanism of eHsp90α secretion………………………….…….12
1.3 Summary of two main protein secretion pathways…………………...……14
1.3.1 Conventional protein secretion pathway………………………..…….14
1.3.2 Unconventional protein secretion (UPS) pathway ……………..…….15
1.4 Secretory autophagy (Type III UPS)……..…………………………………19
1.4.1 Introduction of secretory autophagy…………………………………...19
1.4.2 The role of Golgi -reassembly stacking protein (GRASP)……………22
1.4.3 The role of autophagosome biogenesis genes……………………….24
1.5 Specific aims……………...…………………………………………………..27
Chapter 2: MATERIALS AND METHODS……………………………………………………29
2.1 Cell lines and cell culture…………………………………………………….29
2.2 Antibodies and reagents……………………………………………………..29
2.3 Western immunoblotting analysis and protein quantitation………………30
2.4 Lentiviral systems for up- or down-regulation of target genes……………31
2.5 Fractionation of cell-conditioned medium into three fractions……………34
2.6 Trypsin digestion assay…………………………………………..................35
2.7 Animal model and tissues………..…………………………………………..36
2.8 Wound healing in mice……………………………………………………….36
2.9 Histology and immunohistochemistry………………………………………36
2.10 Statistical analysis………………………………………………................37
Chapter 3: EXPERIMENTAL DESIGN AND RESULTS…………………………………….38
iv
3.1 Hsp90α secretion is triggered by stress, not due to dead cell leakage………………..38
3.1.1 Introduction…………………………………………………………………38
3.1.2 Experimental design……………………………………………………….38
3.1.3 Results………………………………………………………………………40
3.1.4 Discussion…………………………………………………………………..44
3.2 Distribution and quantitation of Hsp90α secretion………………………………………45
3.2.1 Introduction…………………………………………………………………45
3.2.2 Experimental design……………………………………………………….47
3.2.3 Results………………………………………………………………………48
3.2.4 Discussion…………………………………………………………………..52
3.3 Discovery of the “Cell number to Interstitial Fluid (CIF)” ratio for human tissue………55
3.3.1 Introduction…………………………………………………………………55
3.3.2 Experimental design……………………………………………………….57
3.3.3 Results………………………………………………………………………58
3.3.4 Discussion…………………………………………………………………..62
3.4 Golgi reassembly stacking protein 55 regulates partial secretion of Hsp90α………...63
3.4.1 Introduction…………………………………………………………………63
3.4.2 Experimental design……………………………………………………….65
3.4.3 Results………………………………………………………………………67
3.4.4 Discussion…………………………………………………………………..72
3.5 Autophagosome biogenesis genes regulate the remaining portion of Hsp90α
secretion………………………………………………………………………….............74
3.5.1 Introduction…………………………………………………………………74
3.5.2 Experimental design……………………………………………………….75
3.5.3 Results………………………………………………………………………77
3.5.4 Discussion…………………………………………………………………..80
3.6 eHsp90α plays an essential role during wound healing………………………………..82
3.6.1 Introduction…………………………………………………………………82
3.6.2 Experimental design……………………………………………………….83
3.6.3 Results………………………………………………………………………84
3.6.4 Discussion…………………………………………………………………..87
Chapter 4: DISCUSSION……………………………………………………………………...89
4.1 Summarizing discussion……………………………………………………..89
4.2 Future directions……………………………………………………………...98
Reference……………………………………………………………………………………..103
v
List of figures
Figure 1.1 Hsp90 chaperone cycle…………………………………………………………….2
Figure 1.2 Variation in location of Hsp90α and Hsp90β………………………………………3
Figure 1.3 Hsp90α CRISPR/Cas9 knockout mice are phenotypically indistinguishable
from wild-type mice………………………………………………………………………..........5
Figure 1.4 Two proposed mechanisms of action by eHsp90α: ATPase-dependent and
ATPase-independent hypotheses……………………………………………………………..9
Figure 1.5 Evolutionary conserved dual lysin motif K-270 and K-277 determines the
non-chaperone function of eHsp90α…………………………………………………………10
Figure 1.6 The plasma level of Hsp90α between cancer patients and healthy group……12
Figure 1.7 Schematic illustration of conventional protein secretion pathways……………15
Figure 1.8 Schematic illustration of four types of unconventional protein secretion
pathways………………………………………………………………………………………..16
Figure 1.9 Schematic illustration of the difference between degradative autophagy and
secretory autophagy…………………………………………………………………………...21
Figure 1.10 Schematic illustration of GRASP55 function in UPS and autophagy………..24
Figure 1.11 Schematic model of autophagosome biogenesis……………………………..27
Figure 2.1. Schematic illustration of isolating three fractions from MDA-MB-231
cell-conditioned medium………………………………………………………………………35
Figure 3.1 TGFα stimulates the secretion of Hsp90α in human keratinocytes……………41
Figure 3.2 Starvation and hypoxia (1%O2) stimulates the secretion of Hsp90α in 293T
cells and human dermal fibroblast………………………………………………………...…42
Figure 3.3 Tumor cell constitutively secretes Hsp90α even under normoxia……………..43
Figure 3.4 eHsp90α is stimulated under serum starvation in a time dependent manner
and its secretion is not caused by dead cell leakage…………...…………………………..44
Figure 3.5 Biogenesis of extracellular vesicles……………………………………………...46
Figure 3.6 Distribution of eHsp90α in three fractions that isolated and collected from
vi
cell-conditioned medium………………………………………………………………………49
Figure 3.7 Quantitation of eHsp90α in cell conditioned medium…………………………..50
Figure 3.8 The liner regression curve of positive control determines the concentration
of Sup-associated Hsp90α…………………….……………………………………………...51
Figure 3.9 Supernatant-associated eHsp90 is a naked protein and it is not contained
any small vesicles…………….………………………………………………………………..52
Figure 3.10 Distribution and quantitation of eHsp90α in cell-conditioned medium………54
Figure 3.11 Composition of body fluids………………………………………………………56
Figure 3.12 Cell number per cubic centimeter in three types of cells……………………...59
Figure 3.13 Establishment of “Cell number to Interstitial Fluid (CIF)” ratio………………..61
Figure 3.14 The tissue concentration of eHsp90α secreted by different pathways………61
Figure 3.15 GRASP55 but not GRASP65 knockout, partially blocked eHsp90α in
HeLa cell line……………………….…………………………………………………………..68
Figure 3.16 The inhibition of Hsp90α secretion is not due to global inhibition of protein
secretion……………………..………………...………………………………………………68
Figure 3.17 Re-introducing of GFP-GRASP55 into HeLaGRASP55KO cells rescued
Hsp90α secretion………………...……………………………………………………………70
Figure 3.18 Downregulation of GRASP55 but not GRASP65 partially reduced Hsp90α
secretion in MDA-MB-231 cell line…...………………………………………………………71
Figure 3.19 Down-regulation of GRASP55 partially reduced Hsp90α in
immortalized human keratinocytes (IKC)…………………………………………………….72
Figure 3.20 Autophagosome biogenesis gene Atg5 positively, but Atg7 negatively
regulates eHsp90α secretion………………………………………………………..……….78
Figure 3.21 Screening of five autophagosome biogenesis genes…………………………79
Figure 3.22 Downregulation of both GRASP55 and Atg5 blocked majority of Hsp90α
secretion……………………..………………………………………………………………...80
Figure 3.23 Deposition of eHsp90α in both wounded and tumor growing tissue……….85
Figure 3.24 eHsp90α is essential for wound healing………………………………………..86
vii
Figure 3.25 H&E staining of wedged wound biopsies in both wild type and
Hsp90α-KO mice……..………………………………………………………………………..87
Figure 4.1 Schematic illustration of the secretory autophagy regulates the functional
eHsp90α that drives wound healing………………………………………………………....98
Figure 4.2 Hypothetical model of conventional macroautophagy and alternative
macroautophagy……………………..……………………………………………………...100
viii
Abstract
The Heat shock protein 90 (Hsp90) family proteins (α and β) are historically
recognized as ATP-binding-dependent molecular chaperone that assists the proper
folding and support the functionality of client proteins during homeostasis and stress
responses. This understanding has served as the foundation for both laboratory research
and cancer clinical trials targeting the intracellular function of Hsp90 family proteins.
Although several publications reported cell surface bound Hsp90 related molecules as
early as the 70s, few in Hsp90 community credited the finding of non-chaperone form of
Hsp90 and regraded the observations as a result of dead cell leakage. In the 2000s two
independent laboratories discovered secreted Hsp90α in cell conditioned medium of
normal and tumor cells. Extracellular Hsp90α (eHsp90α) acts by binding to the cell
surface receptor, LRP-1 (low-density lipoprotein receptor-related protein-1), leading to
activation of the Akt as well as other intracellular pathways. The two main functions of
eHsp90α are 1) to prevent cells from undergoing apoptosis and 2) to stimulate cell motility.
Pre-clinical studies demonstrated that topical application of recombinant human Hsp90α
(rhHsp90α) strongly promotes closure of trauma, burn, and diabetic wounds in mice and
pigs. Similarly, injection with either rhHsp90α or anti-Hsp90α antibody promoted or
inhibited tumorigenesis in mice. Clinical studies demonstrated that higher levels of plasma
eHsp90α correlated with the later stages of various cancers. Despite above findings, the
mechanism of secretion and more importantly quantitation of eHsp90α remained unclear
and became my thesis project. During my thesis study, I demonstrated that eHsp90α is
ix
actively secreted outside and it is not because of dead cell leakage during cell culture. In
disagreement with previous reports that exosomes are the major secretion pathway
account for Hsp90α secretion, my results showed that less than 5% of secreted Hsp90α
was associated with exosomes, while the majority remained in EV-depleted supernatant
of conditioned medium. To quantitate eHsp90α in different factions of cell conditioned
medium, I established the so-called “cell number to interstitial fluid” (CIF) ratio for human
tissue environment. Calculation by the CIF ratio revealed a non-exosome pathway
supplies 178 µg/mL of Hsp90, which perfectly matches the reported 100-300 µg/mL of
Hsp90 for promoting wound healing in vivo. Fourth, I demonstrated that the secretory
autophagy pathway with autophagosome biogenesis gene, Beclin-1, Atg5, Atg16 and LC3,
together with the Golgi reassembly-stacking protein gene-55 (GRASP55) regulates
exosome-independent secretion of Hsp90α. Finally, in collaboration with Dr. Xin Tang, we
provided direct evidence that eHsp90α is essential during wound healing. This thesis
analyzed the extracellular distribution of secreted Hsp90α, quantitated the tissue
concentration of eHsp90α derived from different extracellular fractions, and reveled that
it is secretory autophagy rather than exosome secretion regulates the functional
concentration of eHsp90α required for wound healing.
1
Chapter 1. Introduction
1.1 Cytosolic heat shock protein 90 known as a chaperone
1.1.1 The Chaperone cycle
The family members of heat shock proteins (Hsp) are classified according to
molecular weight including Hsp60, Hsp70, Hsp90 and so on. Heat shock protein 90
(Hsp90) is one of the most conserved and abundant Hsp presents from the prokaryotes
to eukaryotes, and it accounts for 2-3% of the total cellular proteins in normal cells and it
can be induced up to 7% of the total cellular proteins in cancer cells [1]. There are two
isoforms of cytosolic Hsp90 in vertebrates, constitutively expressed Hsp90β and stresses
inducible Hsp90α, and they share 86% amino acid sequence identity. The other two
Hsp90 isoforms are organelle-residing with Grp94 residing in the endoplasmic reticulum
and HSP75/tumor necrosis factor receptor associated protein 1 (TRAP1) in the
mitochondrial matrix [2].
The structure and cytosolic chaperone function of Hsp90 has been well-studied.
Hsp90 consist of three domains: 1) An N-terminal ATPase domain responsible for ATP
binding and ATP hydrolysis, 2) a middle domain connected via a linker region, and 3) the
C-terminal domain, which is responsible for Hsp90 dimerization [3]. As Figure 1.1 shown,
Hsp90 begins the chaperone cycle in an open conformation that is only dimerized at the
C-terminal domain. Co-chaperones such as Hop, Hsp40, and Hsp70 recruit client proteins
to the middle domain of Hsp90. Subsequent binding of ATP at the N-terminal results in a
closed conformation and promotes client protein activation. Upon the hydrolysis of ATP,
2
Hsp90 reverts to the open conformation and the activated client protein is released.
Utilizing its chaperone activity, cytosolic Hsp90 can assist other proteins to fold properly,
stabilizing proteins against stress, and aiding in protein degradation.
Figure 1.1 Hsp90 chaperone cycle. This schematic illustration shows the Hsp90 chaperone cycle. By using the energy
of ATP hydrolysis, the chaperone cycle allows Hsp90 to maintain the proper structure of client proteins against various
stresses [4].
1.1.2 Differences between Hsp90α and Hsp90β
The human heat shock protein 90 alpha (Hsp90α) and human heat shock protein 90
beta (Hsp90β) share 86% amino acid sequence identity and differ at a total of 99 amino
acid residues along their full length of 732 (Hsp90α) and 724 (Hsp90β) amino acid
3
sequences (Figure 1.2). Sequence differences between Hsp90α and Hsp90β include 58
conservative substitutions, 41 non-conservative substitutions, and 12 deletions in Hsp90β.
Notability, a 40% reduced homology between the two isoforms comes from the linker
region (LR), in which 21 amino acid substitutions and 3 amino acid deletions exist within
a 32 amino acids segment [2].
Figure 1.2 Variation in location of Hsp90α and Hsp90β. This figure shows the location (black colored) of 99 amino acid
substitutions and deletions between human Hsp90α and Hsp90β [2].
In lower organisms, such as budding yeast (Saccharomyces cerevisiae), a study
showed that a high level of Hsp90 was correlated with high tolerance of temperature for
cell growth. Mutation of both HSP82 (Hsp90α) and HSC82 (Hsp90β) leads to cell death
at any temperature [5]. Moreover, purified HSP82 and HSC82 exhibit variable ATPase
activity and conformational cycle, distinct client interactomes, and sensitivity of single
isoform-expressing cells in growth under environmental stress [6]. In mammalian cells,
many studies compared the various function of cytosolic Hsp90α and Hsp90β. Kuo et al.
reported that Hsp90β regulates CpG-ODN signaling in preventing apoptosis [7].
Bouchier-Hayes et al. reported that Hsp90α negatively regulates heat shock-induced
caspase-2 activation, indicating that Hsp90α plays a role in anti-apoptosis [8]. Chatterjee
et al. reported that Hsp90β plays a more critical role than Hsp90α in mediating myeloma
4
cell survival [9]. Taherian et al. reported that although Hsp90α and Hsp90β show similar
interactions with co-chaperones, they exhibit significantly different interactions with client
proteins under stress conditions [10].
1.1.3 Distinct Roles for Hsp90α and Hsp90β during mouse development
Since 2000, several mouse genetic studies have showed that although Hsp90α and
Hsp90β share 86% amino acids homology, they have distinct and non-compensating
roles during mouse development. Voss et al. generated Hsp90β mutant mice by gene
trap insertion of Hsp90β in exon 9, and they showed that heterozygous mice were
phenotypically normal as wild-type mice. However, homozygous embryos developed
normally until embryonic day 9.0/9.5, where the mutant embryo failed to form a fetal
placental labyrinth and eventually died. The studies of Voss et al. indicated that Hsp90β
has a key role in placental development that cannot be compensated by endogenous
Hsp90α [11]. Grad et al. showed that mice with a C-terminal 36 amino acid deletion in
Hsp90α, which is unable to dimerize, displayed little phenotypic difference from wild type
mice. However, C-terminal truncated Hsp90α specifically abolished the ability of mice to
produce sperm even in the presence of endogenous Hsp90β [12]. Tang et al. showed that
although Hsp90α CRISPR/Cas9 knockout mice are phenotypically indistinguishable from
wild-type mice (Figure 1.3), it induced a destabilization of HIF-1α in mouse testis,
resulting in the shrinkage of mouse testis and infertility in male mice [13]. Therefore, these
genetic mouse studies indicated that as an intracellular chaperone, Hsp90β is clearly
more essential than Hsp90α in mediating cell homeostasis during mouse development.
5
Figure 1.3 Hsp90α CRISPR/Cas9 knockout mice are phenotypically indistinguishable from wild-type mice [13].
1.2 Extracellular Heat shock protein 90alpha (eHsp90α)
1.2.1 Discovery of eHsp90α
For decades, Hsp90 family proteins have been recognized as intracellular ATPbinding-dependent molecular chaperone that stabilizes other proteins. In the late 1970s,
several studies reported that a glucose regulated 90 kDa protein was presence both on
the cell surface and in the cell-conditioned medium of tumor-virus infected mouse and
human fibroblasts [14-18]. In 1983, Hughes et al. showed that Hsp90 proteins are located
on the cell surface of macrophages and mouse embryo 3T3 cells [19]. Ullrich et al. and
the follow-up study by Thangue et al. reported that the presence of Hsp90 and Hsp90-
related proteins on the external surface of Meth A tumor, NIH3T3 cells, and HSV-infected
cells [20,21]. However, due to a lack of evidence for the secreted form of Hsp90 and this
6
undefined surface bound of Hsp90, the observations of “cell surface Hsp90” were thought
to be released by dead cells.
In the early 2000s, two laboratories in different research fields, Jay’s group and Li’s
group, first discovered and defined the function of extracellular Hsp90α in promoting
cancer cell invasion and skin wound healing. In 2004, Jay’s laboratory identified
extracellular Hsp90α by using functional proteomic screens, but not Hsp90β, interacts
and activates the matrix metalloproteinase 2 (MMP2) to promote cancer invasion [22]. In
2007, Li’s laboratory identified extracellular form of Hsp90α by using chromatography,
from the cell-conditioned medium of hypoxia stressed human dermal fibroblasts and
human keratinocytes. They observed that extracellular Hsp90α stimulated skin cell
migration in vitro and wound healing in vivo [23,24]. These findings garnered more
research interest to “cell-surface bound”, “cell-released” or “cell-secreted” Hsp90. To
provide a common terminology for secreted Hsp90, Issac’s group recommended using
“eHsp90” to represent extracellular Hsp90 [25], and it has been widely accepted by the
Hsp90 community. In this thesis, “eHsp90α” will be also used to represent “extracellular
Hsp90α”.
1.2.2 The two main in vitro functions of eHsp90α
Several studies have identified the presence of either eHsp90α alone or eHsp90α
and eHsp90β together in the cell-conditioned medium of various cell types [34,35].
However, Cheng et al. provided strong evidence that recombinant human Hsp90α
(rhHsp90α), but not recombinant human Hsp90β (rhHsp90β), promoted human
7
keratinocyte migration [24]. Jayaprakash et al. showed that rhHsp90α, but not rhHsp90β,
promoted wound healing in pig models [36]. Zou et al. also demonstrated that the
intravenous injection of rhHsp90α, but not rhHsp90β promoted tumor formation and lung
metastasis in mouse models [37]. In addition to the promotility activity of eHsp90α, Dong
et al. identified another function of eHsp90α, which is promoting cancer cell survival under
a hostile hypoxic environment. eHsp90α protected tumor cells from hypoxia-triggered
apoptosis, whereas neutralizing eHsp90α function with a monoclonal antibody enhanced
hypoxia-induced cell apoptosis. Application of rhHsp90α prevented the death of Hsp90α
knockout MDA-MB-231 cells under hypoxia [38]. Taken together, unlike intracellular
Hsp90β which functions as a molecular chaperone and maintains proteostasis, there are
two main biological functions of eHsp90α: 1) promoting cancer cell survival and 2)
promoting cell motility.
1.2.3 Mechanism of action by eHsp90α
Two major parallel mechanisms of action were proposed for eHsp90α including 1)
chaperone dependent mechanism and 2) chaperone independent mechanism. Eustace
et al. showed that the DMAG-N-oxide, a cell membrane-impermeable geldanamycin/17-
AAG-deviered inhibitor that targets the ATPase activity of Hsp90, inhibits tumor cell
invasion [22]. Tsutsumi et al. showed that DMAG-N-oxide inhibitor reduced the invasion
of several cancer cell lines in vitro and lung colonization by B16 melanoma cells in mice
[26]. Sim et al. showed that blocking ATPase using ATP-gamma S increased the ability of
rhHsp90α to activate MMP2 in vitro [27]. Baker-Williams et al. showed that TIMP2 and
8
AHA1 act as a molecular switch for eHsp90α that controls the inhibition and activation of
the eHsp90α client protein MMP2 [28]. Taken together, these studies indicate that the Nterminal ATPase domain of Hsp90α remain essential for eHsp90α outside of the cells.
Unlike chaperone dependent mechanism, the chaperone independent mechanism
has mainly focused on eHps90α/LRP-1 signaling pathway [29]. First, Cheng et al.
reported that the ATPase defective mutants of Hsp90α showed an indistinguishable
degree of promotility activity from the Hsp90α-wt protein on primary human skin cells [24].
Second, they further narrowed down the promotility activity of eHsp90α to a 115-amino
acids fragment between LR and middle domain of Hsp90α, so called F-5. They
demonstrated that F-5 fragment alone promote cell migration in vitro and wound healing
in vivo as effectively as the full length Hsp90α [30]. Third, Tsen et al. illustrated the
eHsp90α/LRP-1 signaling pathway as followed: 1) the extracellular part of LRP-1
(subdomain II) received eHsp90α signal; 2) the NPVY motif in the cytoplasmic tail of LRP1 that connects the eHsp90α signaling to the serine-473, but not threonine-308
phosphorylation in Akt kinase and 3) activates Akt1 and Akt2 to promote cell migration
[31]. In addition to the activation of Akt, Gopal et al. demonstrated a crosstalk mechanism
involving eHsp90α-LRP-1 dependent regulation of EphA2 function, in which the eHsp90αLRP-1 signaling axis regulates Akt signaling and EphA2 activation during glioblastoma
cell invasion [25]. Furthermore, Tian et al. showed that clusterin potentiated the eHsp90αLRP-1 signaling pathways on activation of Akt, Erk, and NF-κB to promote epithelial-tomesenchymal transition (EMT) and metastasis in breast cancer cells [32]. These studies
9
revealed the chaperone independent mechanism of eHsp90α, which is acted by the
binding of LRP-1, leading to the activation of Akt as well as other intracellular pathways.
Taken together, the chaperone dependent mechanism such as the activation of MMP2,
and the chaperone independent mechanism such as eHsp90α/LRP-1 signaling pathways
are schematically depicted in Figure 1.4, which represent two parallel mechanisms of the
action by eHsp90α.
Figure1.4 Two proposed mechanisms of action by eHsp90α: ATPase-dependent and ATPase-independent
hypotheses [33].
1.2.4 Two evolutionarily conserved lysine residues distinguish eHsp90α and
eHsp90β for extracellular functions
Zou et al. used CRISPR/Cas9 to knockout Hsp90α and Hsp90β in MDA-MB-231
breast cancer cells. They found that Hsp90β knockout cancer cells died during drug
selection, whereas Hsp90α knockout had little effect on cancer cell survival. Additionally,
Hsp90α knockout specifically nullified the migration and invasive ability of cancer cells in
10
vitro and tumorgenicity in vivo. Most importantly, the lost ability of migration and
invasiveness in vitro and tumorigenicity in vivo could be fully rescued by extracellular
supplementation or injection to blood circulation with recombinant human Hsp90α but not
Hsp90β. Moreover, Zou and his colleagues identified that the dual lysine motif K-270 and
K-277 determine the non-chaperone function of extracellular Hsp90α [37]. As Figure 1.5
shows, the dual lysine motifs only present in Hsp90α and substituted with G-262 and T269 in Hsp90β. Substituting Hsp90α with G-262 and T-269 resulted in the loss of
extracellular function. Replacing G-262 and T-269 with dual lysine grants Hsp90β the
extracellular function of Hsp90α.
Figure 1.5 Evolutionarily conserved dual lysin motif K-270 and K-277 determines the non-chaperone function of Hsp90α.
Dual lysine motif K270G and K277T determine the extracellular function of Hsp90α. Replacing G262K and T269K with
dual lysine motif grants Hsp90β the extracellular functions of Hsp90α [37].
1.2.5 Role for eHsp90α during wound healing and tumor progression
When skin is injured and nearby broken blood vessel is clotted, cells surrounding the
wound bed suffer from starvation and hypoxia, so-called ischemia. Bhatia et al. made fullthickness skin wounds in pigs, biopsied the wound, and immuno-stained the tissue
sample with anti-Hsp90α antibody. They found a time-dependent increase of eHsp90α in
11
both the epidermis and dermis [39], indicating the two main functions of eHsp90α during
wound healing: supporting cell survival and promoting skin cell migration to close the
wound. A series of studies in Li’s group also demonstrated that topical application of
rhHsp90α, but not rhHsp90β, promoted wound healing in excision, burn, and diabetic
wounds in mice and pigs [23,36,39,40]. Bhatia et al. used transgenic mice that express
N-terminal truncated Hsp90α (chaperone function nullified), and they found that
transgenic mice heal skin wounds as efficiently as wild type, indicating that the chaperone
function of Hsp90α is unnecessary in wound healing [41]. Moreover, Cheng et al. also
narrowed down the functional region of Hsp90α to a 115 amino acid fragment between
LR and middle domain of Hsp90α, so-called F-5 fragment. The F-5 fragment contains the
dual lysine motif that grants Hsp90α extracellular functions as previously discussed. They
showed that the F-5 fragment alone replicated the wound healing-promoting effect of fulllength Hsp90α [24] and F-5 fragment currently is undergoing clinical trials for the
treatment of diabetic foot ulcers.
Since 2008, nearly twenty clinical studies have compared the eHsp90α level in blood
circulation between healthy humans and patients with different types of cancers and
inflammatory diseases. Figure 1.6 shows the clinical studies of eHsp90α in human
plasma [42-58]. In these studies, the plasma level of eHsp90α is dramatically elevated in
patients with various cancers. The healthy group has an average of 0-50 ng/mL eHsp90α
in their blood and the cancer patients have 100-500 ng/mL eHsp90α in their blood. In
addition, the level of eHsp90α is positively correlated with the progression of cancer stage.
12
Therefore, plasma eHsp90α could be a potential therapeutic target for monoclonal
antibodies to block cancer metastasis or an ideal serum prognostic marker for cancer
detection.
Figure 1.6 The plasma level of Hsp90α between cancer patients and healthy group. Compared with healthy group, the
plasma level of eHsp90α is dramatically elevated in patients with various cancers. [42-58]
1.2.6 The mechanism of eHsp90α secretion
After the discovery of eHsp90α and definition of its biological functions, the
mechanism of eHsp90α secretion has been debated. The Hsp90 community initially
believe that eHsp90α comes from the leakage of intracellular Hsp90 from a small number
of dead cells during culture. While it is technically difficult to prove that eHsp90α does not
13
result from dead cell leakage, evidence supports that eHsp90α is actively secreted by
living cells. Li et al. showed that eHsp90α collected from the normal cell conditioned
medium was nearly undetectable under physiological conditions [23]. In comparison,
significantly levels of eHsp90α were detected from the cell-conditioned medium of the
same normal cells under various types of stress including reactive oxygen species (ROS),
heat shock, irradiation, hypoxia, and tissue injury [38,59,60]. In addition to this, Cheng et
al. provided strong evidence that TGFα, a growth factor that promotes cell survival and
cell proliferation, stimulates Hsp90α translocation to the plasma membrane and secretion
to extracellular space in human keratinocytes [24]. Moreover, Tang et al. provided direct
evidence that Hsp90α is present on the outer surface of tumor-secreted exosomes,
indicating that exosomes are one of the mechanisms that account for Hsp90α secretion
[61]. Taken together, these studies indicated that Hsp90α is an actively secreted molecule,
and secretion is not due to dead cell leakage.
Several studies have provided different explanations for the secretion mechanisms
of Hsp90α. Wang et al. reported secretion of Hsp90α is regulated by its C-terminal amino
acid EEVD motif [62]. Tsutsumi et al. reported the secretion of Hsp90α is determined by
a conserved hydrophobic motif in a β-strand at the boundary between the N domain and
charger linker region of Hsp90 [63]. Zhang et al. identified that Rab coupling protein (RCP)
mediates mutant p53 induced Hsp90α secretion [64]. Moreover, studies provide further
evidence for exosomes as a secretion mechanism of Hsp90α based on the observation
that exosome inhibitor Dimethyl amiloride (DMA) blocked Hsp90α secretion in both HIF-
14
1α overexpression tumor cells and TGF-α stimulated human keratinocytes [65,66]. Guo
et al. further identified the proline-rich Akt substrate of 40 kDa (PRAS40) as a unique
downstream regulator of TGF-α but not EGF that binds to EGFR and mediates Hsp90α
secretion via exosomes [67]. The study by Tang et al. confirmed that Hsp90α is associated
with exosomes by providing direct evidence that Hsp90α is present on the external
surface of exosomes [61]. Collectively, the exosome secretion pathway is a widely
accepted mechanism that accounts for Hsp90α secretion.
1.3 Summary of two main protein secretion pathways
1.3.1 Conventional protein secretion pathway
Protein secretion is the cellular process present in both prokaryotes and eukaryotes
that delivers intracellular soluble proteins, cargo, and cellular wastes to the extracellular
space. Prokaryotic cells secrete cellular wastes and other virulence factors through
various protein secretion apparatuses [68]. In contrast, eukaryotes secrete signal peptidecontained protein that is freshly synthesized through the endoplasmic reticulum (ER) to
Golgi apparatus and eventually reaches the plasma membrane and is secreted outside.
As Figure 1.7 shows, the signal peptide contained proteins are recognized by signal
recognition particle (SRP) and transported into the ER lumen. In ER, the signal peptides
are cleaved off and proteins are stabilized by molecular chaperones and then packed into
COPII-coated vesicles and delivered into cis-Golgi. Next, proteins are transported through
secretory vesicles (SV) to the plasma membrane and finally secreted to extracellular
space [69].
15
Figure 1.7 Schematic illustration of conventional protein secretion pathways. Signal peptide tagged protein are
recognized by the signal recognition particle (SRP). Then SRP interacts with receptor SR and signal peptide tagged
protein is translocated into the ER lumen. In the ER, the signal peptide is cleaved off and proteins are packed into
COPII vesicles. Then, COPII vesicles are delivered to the Golgi apparatus for the protein post-modification. At transGolgi network (TGN), proteins that destined to be secreted outside are sorted in various secretory vesicles. Finally,
secretory vesicles fuse with plasma membrane and delivered proteins to the plasma membrane or secreted proteins
to the extracellular environment [69].
1.3.2 Unconventional protein secretion (UPS) pathway
16
In conventional protein secretion pathway, proteins with signal peptide are
translocated to the plasma membrane through ER and Golgi, but numerous leadless
proteins that lack signal peptide (including Hsp90α) are secreted without entering the ERGolgi pathway [70]. For the past two decades, studies have uncovered the existence of
four types of unconventional protein secretion pathways (Figure 1.8) including: 1) type I,
plasma membrane pore formation mediated protein secretion; 2) type II, ABC-transported
based protein secretion; 3) type III, organelle-based protein translocation and 4) type IV,
Golgi-bypass.
Figure 1.8 Schematic illustration of four types of UPS. Type I UPS requires secreted proteins to be able to selfoligomerization and form a secretion channel to transport through plasma membrane. Type II UPS only pump out small
molecules by using the energy of ATP hydrolysis. Type III UPS is proposed to be endosome and autophagosome
mediated protein secretion. Type IV UPS transports signal peptide contained proteins directly from ER to plasma
membrane and bypass Golgi [71].
Type I UPS describes the direct protein translocation across the plasma membrane
17
mediated by the pore formation. The best studied example is fibroblast growth factor 2
(FGF2). Steringer et al. reported that the membrane translocation of FGF2 is involved in
PtdIns(4,5)P2-induced self-oligomerization at the inner leaflet, followed by insertion of
oligomerized FGF2 into the plasma membrane and secreted outside [72]. Zeitler et al.
found HIV TAT protein undergoes PtdIns(4,5)P2-induced self-oligomerization and plasma
membrane insertion that allows itself to be translocated across the plasma membrane
[73]. In addition to FGF2 and HIV TAT, direct translocation across the plasma membrane
is also triggered by inflammation, which leads to the release of cytokines from
macrophages. The best-studied example is IL-1β reported by Martín-Sánchez et al.
Unlike FGF2, IL-1β does not bind to PtdIns(4,5)P2 and form pores. Rather, IL-1β release
is regulated by the hyper permeabilization of the macrophage plasma membrane, which
eventually leads to cell lysis [74].
Type II UPS is mediated by the ABC transporter. By using the energy of ATP
hydrolysis, ABC transporter can pump out small molecules across the plasma membrane
including amino acids, peptides, ions, sugars, vitamins, hormones, and lipids [75]. In 1989,
McGrath et al. reported that the yeast pheromone, a hydrophobic a-factor lipopeptide, is
exported by an ABC transporter called Ste6p [76]. In 1997, Christensen et al. reported
that the farnesylated m-factor of Schzoaccharomyces pombe is exported by ABC
transporter Mam1. In 2009, Ricardo et al. reported that the secretion of Drosophila germ
cell attractant is mediated by ABC transporter named Mdr49. Moreover, acylated proteins
such as hydrophilic acylated surface protein B (HASPB) and plasmodium falciparum
18
Ca2+
-dependent protein kinase 1 (PfCDPK1) are exported by ABC transporter in parasites
[77,78]. These findings indicate that in a wide range of eukaryotes, different kinds of
lipidated peptides are exported by specific ABC transporters.
Type III UPS is defined by organelle-based translocation of leaderless proteins.
These organelles are proposed to be endosomes and autophagosomes [79]. Early
endosomes are formed by plasma membrane invagination. Then, the maturing endosome
continues to bud inward and generate intraluminal vesicles (ILVs). Some of the late
endosomes fuse with lysosomes, and the others fuse with plasma membrane and release
ILVs, which are called exosomes [80,81]. Exosomes have a unique protein composition
depending on their cellular origin. Over hundreds of proteins, mRNAs, and microRNAs
have been identified from a variety of cells and body fluids, including membrane adhesion
proteins, membrane trafficking proteins, cytoskeletal proteins, lysosomal markers,
antigen-presenting proteins, tumor antigens, death receptors, cytokines, iron transporter,
metabolism enzymes, heat shock proteins and so on [82]. Exosomes have been wellstudied for two decades, conversely, secretory autophagosome is a relatively new protein
secretion mechanism that is classified as type III UPS. The first leadless protein involved
in secretory autophagosome was identified during the study of yeast Acb-1. Duran et al.
reported that starvation induced secretion of Acb-1 requires Golgi associated protein
GRASP and autophagosome biogenesis gene Atg5, indicating its secretion is mediated
by secretory autophagosomes [83]. Furthermore, studies have shown that the secretion
of IL-1β is inhibited by autophagy inhibitor Wortmannin, as well as the downregulation of
19
Atg5 in Hek293T cells [84].
Different from type I/II/III UPS, type IV UPS describes the signal peptide contained
proteins that are synthesized in ER but bypass Golgi when they are delivered to the
plasma membrane. For example, treatment with brefeldin A, an inhibitor that blocks ER
to Golgi transportation did not prevent delivery of the mutated transmembrane proteins
such as Phe508-deleted cystic fibrosis transmembrane conductance regulator (CFRT) or
S723T pendrin to the plasma membrane under ER stress. Studies have shown the
mechanism behind Golgi bypass is related to Hsp70 and DNAJC14. Firstly, Hsp70
interacts directly with H723R pendrin. Then, DNAJC14, a co-chaperone of hsp70, is
upregulated under ER stress and mediates delivery of H723R to the plasma membrane
[85,86].
1.4 Secretory autophagy (Type III UPS)
1.4.1 Introduction of secretory autophagy
Autophagy or macroautophagy is usually defined as a cytoplasmic degradative
process that carries out cytoplasmic quality control and nutritional functions by degrading
and digesting disused organelles, misfolded proteins, and invading microbes in
autolysosomes [87]. However, studies in the past ten years have revealed a new function
of autophagy in regulating the unconventional protein secretory pathway. Besides the
intracellular degradative role, studies have shown autophagy plays a critical role in the
extracellular delivery of many leaderless proteins that do not enter the conventional ER
to Golgi protein secretory pathway. As Figure 1.9 shows, the degradative autophagy
20
starts from the nucleation of omegasomes on the ER exit point by the aiding of GRASP.
Omegasome then facilitates the formation of degradative autophagosome by membrane
elongation and closure, which requires the presence of LC3 and ATG family proteins. The
degradative cargo can be sorted into degradative autophagosomes through autophagy
receptors such as sequestosome 1/p62-like receptor (SLRs) and tripartite motif (TRIMs).
After the closure and maturation of degradative autophagosome, it fuses with lysosome
and degrades its cargo [88,89]. Studies have shown that degradative and secretory
autophagy share the common process of autophagosome biogenesis [90-92]. Similarly,
GRASP mediates and facilities the formation of omegasomes at the ER exit point.
Regulating by Beclin-1, ATG family proteins, and LC3, omegasome elongates and mature
to the secretory autophagosome. After the maturation, in contrast to degradative
autophagosome that fuse with lysosome and form autolysosome to degrade its cargo,
secretory autophagosomes migrate along the microtubes and eventually fuse with plasma
membrane and release their cargo to the extracellular space [93].
21
Figure 1.9 Schematic illustration of the difference between degradative autophagy and secretory autophagy. At early
stage of autophagy, degradative and secretory autophagy share the common process of autophagosome biogenesis.
At ER exit point, GRASP55 and Beclin-1 mediates and facilities the formation of omegasomes. Then, mediated by Atg
family proteins and LC3, omegasome elongates and encloses to form mature autophagosomes. In degradative
autophagy, degradative autophagosomes fuse with lysosome to form autolysosome and resulting in the degradation of
engulfed cargos. In contrast, rather than fuse with lysosome, secretory autophagosomes has been shown to migrate
along the microtubules and eventually fuse with plasma membrane and release their cargos into the extracellular
environment [93].
22
1.4.2 The role of Golgi-reassembly stacking protein (GRASP)
In mammalian cells, the Golgi apparatus has developed a unique stacked, ribbonlike structure for processing and sorting protein cargo. The Golgi apparatus is composed
of three stacks including 1) cis-Golgi network (CGN) that faces the nucleus and receives
all the output transported from the ER; 2) the stacked cis-, medial-, and trans- Golgi
cisternae that contain various enzymes for the protein post-translational modification; and
3) trans-Golgi network (TGN) that faces the plasma membrane and transports cargo
proteins to their destination inside or outside the cell [94]. Studies have shown that Golgi
reassembly stacking proteins (GRASPs) play an essential role in regulating Golgi
structural organization, especially the Golgi stack and ribbon formation. In mammalian
cells, there are two GRASPs including GRASP65, which is present in cis-Golgi, and
GRASP55, which is present in medial- and trans- cisternae [95,96]. Single depletion of
either GRASP65 or GRASP55 only reduces the number of cisternae per Golgi stack,
while the double depletion of both GRASP65 and GRASP55 leads to the disassembly of
the entire stack, indicating that GRASP65 and GRASP55 plays a complementary role in
mediating Golgi stack formation [97,98].
In addition to the regulation of Golgi stack formation, GRASP65 and GRASP55 have
specific functions in regulating various cellular events. Ahat et al. showed that GRASP65
is more important for cell attachment and migration than GRASP55 by comparing the
single depletion of each GRASP in HeLa and MDA-MB-231 cells [99]. Cheng et al.
showed GRASP65 but not GRASP55 is a key target for caspase-3 cleavage, which
23
induces Golgi fragmentation and eventually leads to apoptosis. Expression of a caspase
resistant GRASP65 mutant protein inhibits Golgi fragmentation and protects cells from
apoptosis [100], indicating GRASP65 plays a vital role in mediating apoptosis.
In comparison, GRASP55 plays a diverse role in mediating the unconventional
protein secretion pathway and autophagy (Figure 1.10). Duran et al. reported that the
starvation-induced secretion of Acb-1 requires GRASP [83]. Ahat et al. reported that
GRASP55 facilitates huntingtin (Htt) secretion by tethering autophagosome to lysosome
and promoting autophagosome maturation and subsequent lysosome secretion under
stress, indicating GRASP55 is involved in type III UPS [101]. Marioara et al. reported that
depletion of GRASP55 blocks the secretion of leaderless protein interleukin-1β (IL-1β) via
autophagosomes in macrophages [102]. As Figure 1.10 (A) shows, under starvation, the
mature form of IL-1β is translocated into the ER-Golgi intermediate compartment (ERGIC),
and then delivered into secretory autophagosome mediated by GRASP55. Eventually,
secretory autophagosomes fuse with the plasma membrane and release IL-1β to the
extracellular space.
In addition to type III UPS, studies also showed that GRASP55 regulates type IV
UPS. Cystic fibrosis transmembrane conductance regulator (CFTR) is a signal peptide
containing proteins that translocated to plasma membrane through classical ER to Golgi
protein secretion pathway. Mutation of transmembrane CFTR caused protein misfolding
and aggregation in the ER and subsequent ER stress-induced phosphorylation and
monomerization of GRASP55. The GRASP55 monomer binds to mutant CFTR at the ER
24
exit site, bypassing Golgi and facilitating it translocation into the plasma membrane
[103,104]. Collectively, GRASP65 plays a more important role in Golgi rearrangement
and apoptosis, whereas GRASP55 act as a stress sensor and an effector to mediate the
secretion of leaderless proteins and autophagy in response to stress.
Figure 1.10 Schematic illustration of GRASP55 function in UPS and autophagy. (A) Schematic model of GRASP55
function in UPS. Transmembrane proteins, such as CFTR are conventionally transported to plasma membrane through
conventional protein secretion pathway. However, mutant CFTR has been shown to be delivered to the plasma
membrane through GRASP55 dependent type IV UPS (Golgi bypass). In addition to type IV UPS, GRASP55 also has
been shown to mediate the secretion of IL-1β through secretory autophagy dependent type III USP. (B) Schematic
model of GRASP55 function in autophagy. In physiological condition, GRASP55 together with GRASP65 mediates
Golgi stacking. Under starvation, GRASP55 is translocated to the autophagosome to facilitate autophagosomelysosome tethering [105].
1.4.3 The role of autophagosome biogenesis genes
25
As Chapter 1.4.1 introduced, the omegasome is formed from the exit point of the ER.
The cup shaped omegasome elongates to engulf cytoplasmic components and finally
closes to form a double membrane autophagosome. The matured autophagosome then
either fuses with lysosome to degrade its cargo or fuses with the plasma membrane to
release its cargo to extracellular space. Regardless of the final destiny of autophagosome,
both degradative autophagy and secretory autophagy share the same autophagosome
biogenesis process at the early stage of autophagy. There are three steps of
autophagosome biogenesis during early stage, including 1) initiation (nucleation), 2)
elongation, and 3) maturation. Regulation is provided by five core protein complexes and
systems (Figure 1.12), including 1) Atg1/ULK1 protein-kinase complex, 2) Atg9/Atg12-
Atg18 complex, 3) Vps34-Atg6/beclin-1 class III phosphoinositide 3-kinase (PI3-kinase)
complex, 4) Atg12 conjugation system, and 5) Atg8/LC3 conjugation (lipidation) system
[106].
The initiation, or the nucleation, of autophagosome biogenesis requires cooperation
of the first three protein complexes. In mammalian cells, Atg1/ULK1 complex is composed
of ULK1 kinase, Atg13, FIP200 and Atg101 [107,108]. Hosokawa et al. showed that the
C-terminal EAT domain of ULK1 kinase is required for the translocation of Atg1/ULK1
complex to the nucleation site of omegasome [108]. Chan et al reported that Atg13 is
required for the formation of omegasome by promoting lipid recruitment [109]. Mercer et
al. showed that the function of FIP200 and Atg101 is to maintain the stability of
ULK1/Atg13 phosphorylation [110]. Besides ULK1 C-terminal EAT domain, Judith D et al.
26
showed that Atg9/Atg12-Atg18 complex has the same function to mediate the
translocation of the ULK complex to omegasome nucleation site [111]. Compared to the
other two complexes, the Vps34-Atg6/beclin-1 class III PI3-kinase complex is more well
studied in mammalian cells. Xie et al. showed Vps34 uses phosphatidylinositol (PI) as a
substrate to generate phosphatidyl inositol triphosphate (PI3P), which is important for
omegasome elongation and recruitment of subsequent a ubiquitin-like conjugation
system to the omegasome. Moreover, the interaction between beclin-1 and Vps34
promotes its catalytic activity and increases the level of PI3P to further promote
omegasome elongation [112].
Two ubiquitin-like conjugation systems, Atg12 conjugation system and Atg8/LC3
conjugation system, play a direct role in omegasome elongation and maturation. In the
Atg12 conjugation system, Atg12 is activated by E1 ubiquitin activating enzyme Atg7.
Atg12 then conjugates with Atg5 and forms Atg12-Atg5 conjugates. Atg12-Atg5
conjugates pair with Atg16L to form a multimeric Atg5-Atg12-Atg16L complex which
associates and localizes to the elongating omegasome [113,114]. The Atg12 conjugation
system is closely interrelated to the second ubiquitin-like conjugation system Atg8/LC3
[115]. Mizushima et al. showed that Atg5 deficiency results in the defect of Atg8/LC3
lipidation [116]. Fujita et al. showed that Atg16L determines the site of autophagosome
formation [117]. In the later stages of omegasome elongation, the Atg5-Atg12-Atg16L
complex gradually disassociates from omegasome. Atg8/LC3 is subsequently cleaved
and activated by two enzymes, Atg4 and Atg7, to form LC3-II and finally localizes and
27
inserts to the extending omegasome until it forms a mature autophagosome [114].
Figure 1.12 Schematic model of autophagosome biogenesis. Atg13-ULK1 and Beclin-Vps34 complexes regulate the
initiation of autophagosome biogenesis. Two ubiquitin-like conjugation systems (Atg5–Atg12 conjugation and LC3–
phosphatidyl ethanolamine (PE) conjugation) are involved in the elongation and maturation of autophagosome
biogenesis [118].
1.5 Specific Aims
In comparison to intracellular Hsp90 is chaperone function that maintains client
proteins stability and functionality, eHsp90α has been revealed to perform different
functions in both normal and tumor cells. Normal cells only secrete Hsp90α under stress,
however, tumor cells constitutively secrete Hsp90α driven by oncogenes. In both cases,
the two core functions of eHsp90α are: 1) preventing cell death under stress and 2)
promoting cell motility. These two functions play essential roles during wound healing and
cancer progression. In pre-clinical studies, recombinant human Hsp90α was showed to
promote wound healing in pig and mouse models and tumorigenesis in mouse models.
28
In clinical studies, higher plasma level of Hsp90α was reported to correlated with later
cancer stage of human patients. However, given the importance of eHsp90α, few studies
focused on the secretion mechanisms of eHsp90α. Exosomes that belongs to type III
UPS have been reported to be one mechanism that accounts for Hsp90α secretion.
However, my preliminary experiments (shown in Chapter 3) showed that a majority (~95%)
of eHsp90α is not associated with exosomes, and rather present in extracellular vesicle
(EV)-depleted cell-conditioned medium. Therefore, understanding the secretion
mechanism for this large pool (~95%) of eHsp90α can give us new insights into the
therapeutic potential of eHsp90α in development of wound healing drug and anti-cancer
drugs.
In conclusion, the aim of this study was to investigate the secretion mechanism(s) of
eHsp90α to promote wound healing. To achieve this, following specific aims and
experiments were designed and performed:
1. To prove eHsp90α is triggered by stress but not dead cell leakage by using western
immunoblotting analysis.
2. To investigate the distribution of eHsp90α collected from cancer cell-conditioned
medium by using sequential centrifugations.
3. To quantitate the amounts of eHsp90α in different fractions isolated by sequential
centrifugations.
4. To analyze protein secretion pathways that most likely contribute to the functional
concentration of Hsp90α secretion.
29
Chapter 2. Materials and Methods
This chapter illustrates the methodology used to complete the various projects within
this thesis. All reagents were purchased from Thermo Fisher Scientific (Thermo Scientific,
MA), Amerham, Inc (Marlborough, MA), VWR international (Radnor, PA) and Millipore
Sigma (Burlington, MA) unless otherwise specified.
2.1 Cell lines and cell culture
The MDA-MB-231, HeLa, HeLaGRASP55 knockout, HeLaGRASP65 Knockout,
293T, and human dermal fibroblasts were cultured in high glucose Dulbecco’s modified
Eagle medium (DMEM) with penicillin-streptomycin (100U/mL-0.1mg/mL) and 10% fetal
bovine serum (FBS) (Thermo Scientific, MA, USA). Immortalized human keratinocyte
were cultured in Epilife medium with added growth factor supplements (Thermo Scientific,
MA, USA). MDA-MB-231 were obtained from the laboratories of Michael Press (University
of Southern California, Los Angeles, CA). HelaGRASP55 knockout, HeLaGRASP65
knockout cell line was obtained from the laboratory of Yanzhuang Wang (University of
Michigan, Ann Arbor, MI). All cells were tested to ensure that they were mycoplasma free
every 2 months at USC Tissue Culture Core. The third and fourth passages of the primary
human cells, such as human dermal fibroblasts, were used in this study. To investigate
intracellular proteins, 90% confluence of cells were prepared in either 10 cm or 15 cm
culture dishes. To collect extracellular Hsp90α or extracellular vesicles, 80% confluence
of cells were prepared in 15 cm culture dishes.
2.2 Antibodies and reagents
30
TGF-α was purchased from Fitzgerald Industries International (Action, MA, USA).
Trypsin (2.5%), phenol red was purchased from Thermo Fisher Scientific (Thermo
Scientific, MA, USA). ECL Western blotting detection reagent (no. RPN2106) was
purchased from Amersham, Inc. (Marlborough, MA). The antibodies used within this
thesis including anti-Hsp90α antibody (NB120-2928) was purchased from Novus
Biologicals (Littleton CO); anti-AKT2 (D6G4; no.3063), anti-S6 Ribosomal protein (5G10;
no. 2217), anti-STAT-3(no.9132), anti-Raptor (23C12; no.2280), anti-CD9
(D3H4P;13403), anti-phospho-ERK (D11A8; no.5683), anti-Atg5 (D5F5U; no.12994T),
anti-Atg7 (D12B11; no.8558T), anti-Atg16L1 (D6D5; no.3495T), anti-Beclin-1 (D40C5;
no.3495T) and anti-LC3A (D50G8; no.4599T) antibody was purchased from Cell
Signaling Technology (Beverly, MA); anti-β-actin antibody (AC038 ) was purchased from
Transduction Laboratories (San Jose, CA); and anti-GRASP55 (10598-1-AP) and antiGRASP65 (66651-1-Ig) was purchased from Proteintech Group (Rosemont, IL).
2.3 Western immunoblotting analysis and protein quantitation
Presence of Hsp90α, β-actin, CD9, GRASP55, GRASP65, Atg5, Atg7, Atg16, LC3A,
Beclin-1, PERK, S6 Ribosomal protein, Stat3, Akt2, and Raptor under various
experimental conditions was verified using Western immunoblotting analysis. Total cell
lysates were prepared by using either 500 µL lysis buffer for 10 cm dish or 1 mL for 15 cm
dish, homogenized by vertexing and centrifugated at 14,000 × g for 10 minutes at 4 °C.
Subsequent supernatant was transferred to a clean tube and store at -20°C. The lysis
buffer contained the cocktail of 20 mM Tris Base, 50 mM NaCl, 50 mM Napyrophosphate,
31
30 mM NaF, 100 µM NaOrthovanadate, 5 µM Zinc Chloride, 2 mM Iodoacetic Acid, and
1.0% Triton-X-100. The pH was adjusted to 7.4 with 1N HCl and add ddH2O to final
volume of 1 L and filtered with 0.45 µm MEC membrane. The protease inhibitor,
phenylmethane sulfonyl fluoride (PMSF 1mM), was added to the lysis buffer prior to use.
The concentration of cell total lysates was determined and equalized by using BCA
protein assay kit (Thermo Scientific). Sample buffer (with 10% β-mercaptoethanol) was
added to samples in a 1:3 ratio and boiled for 5 minutes. The samples such as cell lysate,
exosomes and extracellular vesicle-depleted supernatant were separated by SDS-PAGE
and transferred to a nitrocellulose membrane. Ponceau S solution (0.2%) staining was
used to confirm the transfer efficiency. The primary antibodies against the indicated
proteins were used as instructed by the manufactures. Secondary anti-rabbit IgG
(AP307P, Millipore Sigma, MA) and anti-mouse IgG (sc-516102, Santa Cruz, TX) were
used as instructed by the manufactures. The intensity of protein bands was quantitated
using NIH imageJ software.
2.4 Lentiviral system for up-or down regulation of target genes
The concentrations of packaging, overexpression and downregulation vectors used
in this study were measured by Eppendorf Biospectrometer (Hamburg, Germany). The
lentiviral infection system pHAGEII-pEF2a was used to overexpress exogenous
GRASP55 cDNA (provided by Dr. Yanzhuang Wang from University of Michigan, Ann
Arbor, MI.) The pHR-CMV-puro RNAi delivery system was used to deliver shRNA against
human GRASP55 and human GRASP65. The sequence of shRNA and cDNA used in this
32
thesis are listed in followed table:
Target Gene Plasmid Name Vector
Backbone
Sequence
GRASP55 pHR-CMVshGRAPS55
pHR-CMV-puro CCACCTGAAGACTTGTGTTAA
GRASP65 pHR-CMVshGRASP65
pHR-CMV-puro GTGAAGCTGGAGGTGTTCAAT
GFP-GRASP55 pHAGEII-pEF2aGRAPS55-GFP
pHAGEII-pEF2a See Appendices in page ( )
ATG5 pGIPZ-shATG5 pGIPZ CTTGGAACATCACAGTACA
ATG7 pGIPZ-shATG7 pGIPZ TGCTGTTGACAGTGAGCGACCAGCTATTGGAA
CACTGTATTAGTGAAGCCACAGATGTAATACAG
TGTTCCAATAGCTGGGTGCCTACTGCCTCGGA
ATG16L1 pGIPZ-shATG16 pGIPZ TGCTGTTGACAGTGAGCGACCAACAGAACTTG
ATTGTAAATAGTGAAGCCACAGATGTATTTACAA
TCAAGTTCTGTTGGGTGCCTACTGCCTCGGA
BECLIN-1 pGIPZ-shBeclin-1 pGIPZ TGCTGTTGACAGTGAGCGAGCCAATAAGATGG
GTCTGAAATAGTGAAGCCACAGATGTATTTCAG
ACCCATCTTATTGGCCTGCCTACTGCCTCGGA
LC3A pGIPZ-shLC3A pGIPZ GTCATTGTCCCTCTGCAGA
Table 2.1 Targeting sequence of up-or downregulation of genes.
The protocol of using lentiviral systems for up- or down- regulation of gene of interest,
including virus packaging, isolation and infection is described below.
33
The transfection kit “ProFection® Mammalian Transfection System- Calcium
Phosphate” is purchased from Promega Corporation (no. E1200, Madison, WI). Following
the protocol instructed by the manufacture, in day one, subculture the 80% confluence
293T cells (in 10 cm culture dish) with 1:3 dilution to a new 10 cm culture dish and cultured
overnight. In day two, change 10 mL fresh medium for 10 cm culture dish or 20 mL fresh
medium for 15 cm culture dish in the early morning and preform transfection in the
afternoon between 3:00 pm to 5:00 pm. To preform transfection, DNA solution need to be
prepared and mixed well, including 15 µg packaging vector pCMV-dR8.2 drpr, 5 µg VSVG
envelope expressing vector pMDG, 15 µg DNA vector and 63 µL 2M CaCl2. Then, using
nuclease-free water add up volume to 500 µL in a 5 mL polystyrene round-bottom tube.
To preform lentivirus packaging, DNA solution need to be added to 500 µL 2 × Hepesbuffered saline dropwise with shake. A fine precipitate should be formed. Keep the tube
in room temperature for 15 minutes (min). After 15 min, add the precipitate solution to
293T cells dropwise slowly and gently and incubate at room temperature for another 15
min. Then move the 293T cells back to the 37°C incubator and incubate overnight. In day
three, change 6 mL fresh medium with 10mM Sodium Butyrate in early morning, incubate
293T cells at 37°C for 8 hours. After 8 hours, change 6 mL fresh medium without Sodium
Butyrate and incubate at 37°C overnight. In day four, to isolate lentivirus, collect the
medium with the syringe then filter the medium with 0.45 µm MCE membrane to remove
dead cells and cell debris. Immediately use the fresh lentivirus or freeze the lentivirus at
-80°C for storage. To preform infection, 70-80% confluence of host cells needs to be
34
prepared. Add the virus medium with 10 µg/µL polybrene to the host cells and incubate
at 37°C for 8 hours. After 8 hours, change fresh medium and post-infection for 48 hours.
Then GFP-fluorescence should be able to be observed after 48 hours post-infection.
Subculture the host cells to 6-wells culture and filter the cell with 2 µg/mL puromycin for
2 weeks. The up-or down-regulation of genes should be further confirmed by western
immunoblotting analysis.
2.5 Fractionation of cell conditioned medium into three fractions
Cells are seeded in approximately 2 million per 15 cm cell culture dish at 37 °C in a
humidified incubator with 5% CO2 in DMEM culture medium with high glucose
supplemented with 10 % FBS and 1% P/S. When cell growth reaches 80% confluence,
serum-containing medium was aspirated, the cells were washed gently three times with
10 mL per dish of 37°C pre-warmed DPBS and incubated in 12 mL of serum-free DMEM
for additional 48 hours. Cell conditioned medium was collected and centrifuged at
2,000×g at 4 °C for 10 min to remove floating cells. The supernatant was subjected to
10,000×g centrifugation at 4 °C for 30 min to collect the pellet of apoptotic bodies and
micro-vesicles. Then, the supernatant was transferred into a 10 mL Backman
ultracentrifuge tube and centrifuged at 100,000×g at 4 °C for 90 min to pellet exosomes.
The exosomes fraction was washed in 10 mL PBS and centrifuged again at 100,000×g
at 4 °C for 90 min to remove any particles. The extracellular vesicle-depleted supernatant
was concentrated into desired volume by using Amicon® Ultra-4 Centrifugal filters,
Ultracel-50K (UFC805024, Millipore Sigma, MA) for subsequent analyses.
35
Figure 2.1. Schematic illustration of isolating three fractions from MDA-MB-231 cell-conditioned medium.
2.6 Trypsin digestion
5 µg of isolated exosomes collected from conditioned medium of MDA-MB-231 cells
were re-suspend in final volume of 0.1mL of PBS with indicated concentrations of trypsin
and incubated 30 min at 37°C with moderate agitation (Thermomixer 5436, Eppendorf).
In comparison, 50mL of extracellular vesicle-depleted supernatant was concentrated to
0.1ml by using Amicon® Ultra-4 Centrifugal filters, Ultracel-50K (UFC805024, Millipore
Sigma, MA), and incubated in PBS with indicated concentrations of trypsin for 30 minutes
at 37°C with moderate agitation. Purified recombinant human Hsp90α protein (5µg) was
also incubated in trypsin for 30 min at 37°C with moderate agitation as positive control.
36
2.7 Animal models and tissues
Hsp90-/- C57BL/6 mice were generated by CRISPR-cas9 technology. All
experimental animal protocols were approved by USC Institutional Animal Care and Use
Committee. Experimental procedures have followed the federal guidelines for the care
and use of laboratory animals (US Department of Health and Human Services, US
Department of Agriculture). Sections of pig skin wounds, control and bleomycin-treated
mouse lung and control and tumor-bearing mouse liver were obtained from our own
stocks, Dr. Beiyun Zhou and Dr. Bangyan Stiles (USC), respectively, and immune stained
with a monoclonal antibody against Hsp90 (without cross reactions for Hsp90).
2.8 Wound healing in mice
Full-thickness, 8-mm wounds were created using punch biopsy and surgical scissors.
For topical treatments, rHsp90α protein (30 µg per wound) was mixed in 1:1 ratio with 15%
sterile CMC in final volume of 100 µL volume (300 µg/mL). Digital photographs were
taken individually of the wounds with a metric ruler next to them on the indicated days
from a fixed distance by a preset tripod. Biopsies were collected on indicated days and
stored in 10% paraffin for sectioning. Photographs of 15 randomly selected images per
condition were examined using planimetric measurements for objective evaluation for
wound closure rates. The area of an open wound was calculated as height (cm) × width
(cm) to give rise to cm2 of the wound. Means of more than five wounds were used for the
presentation. The area of an open wound on a given day was measured and compared
to the area of the wound on day 0 from the same animal, using the software AlphaEase
FC, version 4.1.0 (Alpha Innotech Corporation, Miami, FL).
2.9 Histology and immunohistochemistry
37
The histological and immunohistochemical analyses were carried out for wedged
biopsies measuring 1 cm × 1 cm were taken on the indicated days for skin wounds, lung,
liver, and skin biopsies without or with injury or tumor growth stress. All tissue samples
were fixed in 10% formalin (VWR, Randor, PA), and placed in paraffin blocks for
sectioning. Immunohistochemical analyses were conducted with anti-Hsp90α antibody.
Fifteen randomly selected images under each condition were visualized under
microscope. To carry out semi-quantitation of the staining, we used Gabriel Landini’s
“colour deconvolution” plugin for ImageJ analysis. Using the Image > Color > Colour
Deconvolution, H DAB as vector and Color 2 as DAB image, measurements were carried
out to convert intensity to optical density (optical density = log (max intensity / mean
intensity, where max intensity = 255 for 80-bit images).
2.10 Statistical analyses
All numerical results in triplicates were reported as means and standard deviations
(SD). Statistical significance was determined by two-tailed Student’s t test and two-way
analysis of variance (ANOVA) or one-way repeated measures analysis of variance (RMA).
Final presentation as means and SD was based on at least three independent and
corroborating experiments. Conformation of a difference as statistically significant
requires rejection of the null hypothesis of no difference between means obtained from
replicates sets. A p value equals to or less than 0.05 was considered statistically
significant.
38
Chapter 3. Experimental Design and Results
3.1 Hsp90α secretion is triggered by stress, not due to dead cell leakage.
3.1.1 Introduction
As introduced in Chapter 1.2.6, the mechanism of eHsp90α secretion outside of cells
has been debated. The Hsp90 community initially believed eHsp90α comes from the
leakage of intracellular Hsp90 from a small number of dead cells during culture. Although
several publications have showed that Hsp90α is actively secreted into extracellular
space [23,38,59-63], there is still a lack of direct evidence to rule out the possibility of
eHsp90α is due to dead cell leakage. Therefore, the first step of my thesis is to design
experiments to 1) replicate the conclusion that eHsp90α is result of external stress on
normal cells and internal stress in tumor cells, and 2) provide direct evidence that
eHsp90α is not a result of dead cell leakage.
3.1.2 Experimental design
Three different types of environmental stress, including: 1) TGFα, 2) hypoxia, and 3)
serum starvation, and three normal cell lines including: 1) human keratinocyte (HKC), 2)
human embryonic kidney (293T) cells, and 3) human dermal fibroblast (HDF) were
selected to replicate the result that eHsp90α is a result of external stress on normal cells.
TGFα is a growth factor that remains absent in intact skin and dramatically increased only
when skin is wounded [24]. Therefore, TGFα could be treated as an “environmental stress”
and test whether TGFα stimulates Hsp90α in HKC. In this experiment, three identical
dishes of HKC with 80% confluence were prepared. All three dishes were serum starved
39
overnight and treated with either control (FBS-), 20 ng/mL EGF, or 20 ng/mL TGFα for 16
hours. The cell-conditioned medium (CM) of each group were concentrated and subjected
to western immunoblotting analysis with anti-Hsp90α antibody.
In addition to an increase of TGFα in the wounded site due to the disruption and
clotting of the surrounding blood vessel, both nutriment starvation and hypoxia also occur
at the site of tissue injury. Therefore, I used serum starvation and hypoxia to treat 293T
cells and HDF. In this experiment, two identical dishes of 293T and HDF cells with 80%
confluence were prepared, respectively. Then, 293T cells were either serum starved for
0 or 24 hours. The cell-conditioned medium was collected, processed, and analyzed as
before. Similarly, two identical dishes of HDF with 80% confluence were serum starved
and either untreated or treated with hypoxia (hypoxia chamber with 1%O2) for 16 hours.
The cell-conditioned medium was collected, processed, and analyzed as before.
To prove eHsp90α is a result of internal stress on tumor cells driven by oncogenes,
the secretion of Hsp90α under normoxia and hypoxia was compared in two cancer lines
including 1) MDA-MB-231 triple negative breast cancer and 2) HeLa cervical cancer cells.
If cancer cells secrete Hsp90α under normoxia, it means that secretion of Hsp90α is
constitutively stimulated by cancer oncogenes. In this experiment, two identical dishes of
MDA-MB-231 cells and HeLa cells with 80% confluence were prepared, respectively.
Then cells were cultured under either normoxia or hypoxia with serum starvation for 16
hours. The cell-conditioned medium was collected, processed, and analyzed as before.
To provide direct evidence that eHsp90α is not a result of dead cells. I planned to
40
serum starve six identical dishes of MDA-MB-231 cells for 0, 24, 48, 72, 96, and 120
hours and check whether there was a leakage of intracellular signaling molecule. Similarly,
after each designed serum starvation, the cell-conditioned medium was concentrated to
and subjected to western immunoblotting analysis with anti-Hsp90α, anti-Raptor, antiStat3, anti-Akt, anti-Erk1/2, and anti-S6R antibody. Cell lysates of MDA-MB-231 cells at
each time point were also prepared and subjected to western immunoblotting analysis
blotted with same antibodies as CM.
3.1.3 Results
3.1.3.1 eHsp90α is a result of external stress on normal cells
As shown in Figure 3.1, three identical dishes of human keratinocytes (HKCs) were
serum starved overnight and treated with either control -FBS (lane 1), 20 ng/mL EGF
(lane 2), or 20 ng/mL TGFα (lane 3) for 16 hours. The cell-conditioned media and total
lysates were collected and subjected to western immunoblotting analysis with antibodies
against Hsp90α (panel a and b) and β-actin (panel c). In panel a/ lane 1, I confirmed that
human keratinocytes do not secrete Hsp90α under physiological conditions. In panel
a/lane 2&3, the results show that TGFα but not EGF stimulated a secretion of Hsp90α in
human keratinocytes, indicating that the secretion of Hsp90α in normal cell requires
environmental stress. Panel b shows that neither EGF nor TGFα induced an increase or
decrease of intracellular Hsp90α. Panel c shows the sample loading of the three samples
was equal.
41
Figure 3.1 TGFα stimulates the secretion of Hsp90α in human keratinocytes. Three identical dishes of human
keratinocytes were serum starved overnight and treated with either control -FBS (lane 1), 20 ng/mL EGF (lane 2), or
20 ng/mL TGFα (lane 3) for 16 hours. The cell-conditioned media and total lysates were collected and subjected to
western immunoblotting analysis with antibody against Hsp90α (panel a and b) and β-actin (panel c). Similar results
were obtained from three independent experiments.
As shown in Figure 3.2, two identical dishes of 293T and HDF cells with 80%
confluence were prepared. For 293T cells, two dishes were either serum starved 0 or 24
hours and the CMs were collected and subjected to western immunoblotting analysis with
antibody against Hsp90α. Panel a/lane 1 shows that 293T cells did not secrete Hsp90α
under physiological conditions. Panel a/lane 2 shows that serum starvation stimulated
Hsp90α secretion in 293T cells. Two identical dishes of HDF cells with 80% confluence
were prepared and serum starved either under normoxia or hypoxia (1% O2) for 16 hours,
and the CMs were collected and subjected to western immunoblotting analysis with
antibody against Hsp90α. In panel b/lane 1, HDF cells did not secrete Hsp90α under
normoxia. However, in panel b/lane 2, Hsp90α secretion was significantly induced by
hypoxia, indicating the secretion of Hsp90α requires environmental stress.
42
Figure 3.2 Starvation and hypoxia (1%O2) stimulates the secretion of Hsp90α in 293T cells and human dermal fibroblast.
Two identical dishes of either 293T or HDF cells with 80% confluence were prepared. Two dishes of 293T cells were
either serum starved 0 (panel a/lane1) or 24 hours (panel a/ lane2) and then CMs were collected and subjected to
western immunoblotting analysis with antibody against Hsp90α. Two dishes of HDF were either treated under normoxia
(21%O2, panel b/lane 1) or hypoxia (1%O2, panel b /lane 2) for 16 hours and CMs were collected and subjected to
western immunoblotting analysis with antibody against Hsp90α.
3.1.3.2 eHsp90α is a result of internal stress in tumor cells
As shown in Figure 3.3, two identical dishes of MDA-MB-231 cells and HeLa cells
with 80% confluence were prepared. For both cell lines, the two dishes were serum
starved and either treated under normoxia or hypoxia for 16 hours. The CMs and total
lysates were collected and subjected to western immunoblotting analysis. In comparison
to HDF cells which only secreted Hsp90α under hypoxia, MDA-MB-231 cells and HeLa
cells constitutively secreted Hsp90α even under normoxia as seen in panels a and d/
line1. Panel b shows hypoxia did not influence the expression of intracellular Hsp90α.
Panel c shows the sample loading of the two samples were equal.
43
Figure 3.3 Tumor cell constitutively secretes Hsp90α even under normoxia. Two identical dishes of MDA-MB-231 and
Hela cells with 80% confluence were prepared. For both cell lines, two dishes of cells were serum starved and then
either treated under normoxia (21%O2, lane 1) or hypoxia (1%O2, lane 2) for 16 hours. The CMs and total lysates were
collected and subjected to western immunoblotting analysis with antibody against Hsp90α (panel a, b, and d) and βactin (panel d).
3.1.3.3 The secretion of eHsp90α is not due to dead cell leakage
As shown in Figure 3.4, six identical dishes of MDA-MB-231 cells with 80% cell
confluence were prepared. The six dishes of cells were serum starved for 0, 24, 48, 72,
96, and 120 hours, respectively. The CMs for each time point was collected and subjected
to western immunoblotting analysis with antibody against Hsp90α and five intracellular
markers including Raptor, Stat3, Akt, Erk1/2, and S6 ribosomal protein. Total lysates were
also collected and subjected to western immunoblotting analysis and blotted with the
same antibodies. As shown in panel a, Hsp90α secretion induced by serum starvation
was detected in CMs in a time-dependent manner. The five intracellular markers, however,
were not detected in same CMs. These results indicating that the secretion of Hsp90α
was not due to leakage of intracellular Hsp90α from cell death.
44
Figure 3.4 eHsp90α is stimulated under serum starvation in a time dependent manner and its secretion is not caused
by dead cell leakage. Six identical dishes of MDA-MB-231 cells with 80% confluence were prepared. The six dishes of
cells were serum starved for 0, 24, 48, 72, 96, and 120 hours, respectively (from lane 2 to lane 7). Then the CMs (panel
a) and total lysates (lane 1) were collected and subjected to western immunoblotting analysis with antibodies against
Hsp90α (panel a), Raptor (panel b), Stat3 (panel c), Akt (panel d), Erk1/2(panel e), and S6 Ribosomal protein (panel f).
3.1.4 Discussion
Even though multiple studies have shown that eHsp90α is actively secreted into
extracellular space, there is still a lack of evidence to rule out the possibility that eHsp90α
is due to dead cell leakage. Therefore, to investigate the mechanism of Hsp90α secretion,
the priority experiment that I designed was to provide direct evidence to show eHsp90α
is not caused by dead cell leakage. A comparison of the secreted proteins and intracellular
proteins would provide one evidence to support this. If there is dead cell leakage,
intracellular proteins that are not secreted outside the cell should also be detected along
with eHsp90α in the cell-conditioned medium. If eHsp90α is not caused by dead cells,
intracellular proteins should not be detected in the same cell-conditioned medium.
Therefore, I designed a time-dependent experiment to serum starve MDA-MB-231 cells
45
for one day to five days. The results showed that while increased eHsp90α in the
conditioned medium was detected in a time-dependent manner, none of the intracellular
proteins could be detected in the same conditioned medium. Therefore, the eHsp90α
detected was not a result of intracellular Hsp90α leakage from dead cells.
Moreover, I also provided evidence that eHsp90α is the result of external
environmental stress on normal cells and internal stress in tumor cells. The environmental
stresses included serum starvation, hypoxia, and injured tissue released cytokine-TGFα,
stimulated secretion of Hsp90α in 293T, HDF, and HKC cells, respectively. In contrast,
under physiological conditions, none the normal cells secreted Hsp90α. In addition to
normal cells, I also treated two cancer cell lines with either normoxia or hypoxia. The
results showed that even under normoxia, cancer cells secreted Hsp90α constitutively,
indicating that oncogenes provide an internal stress that stimulates Hsp90α secretion in
tumor cells. Taken together, this section illustrated that the secretion of eHsp90α is an
active process of living cells and not caused by dead cell leakage.
3.2 Distribution and quantitation of Hsp90α secretion
3.2.1 Introduction
Extracellular vesicles (EVs) are cell-derived lipid bilayers ranging from 30 nm to 2000
nm in diameter according to their origin [119]. As Figure 3.5 shows, based on the size,
morphology, markers, contents, and origins, EVs can be classified into three categories,
including 1) apoptotic bodies, 2) microvesicles (MV), and 3) exosomes (Exo) [120-123].
46
Figure 3.5 Biogenesis of extracellular vesicles. Based on the size, morphology, markers, contents, and origins,
extracellular vesicles can be classified into three categories, including 1) apoptotic bodies, 2) microvesicles, and 3)
exosomes. Apoptotic bodies are largest vesicles with diameters ranging from 100 nm to 1000 nm. During apoptosis,
the dying cell begins to condensate its nuclear chromatin and disintegrated the cellular content into distinct membraneenclosed vesicles, so-called apoptotic bodies. Microvesicles are medium size vesicles with diameters ranging from 50
nm to 200 nm. Enclosed microvesicles containing cytoplasmic cargos bud outward directly from the plasma membrane.
Exosomes are the smallest vesicles with diameters ranging from 30 nm to 150 nm [124].
Previous publications identified Hsp90α as one of the cargo proteins associated with
exosomes [63-67,126-128]. However, these studies neither quantitated the amount of
exosome-associated eHsp90α nor investigated other possible transportation routes for
eHsp90α. Therefore, my first aim was to analyze the distribution and quantitation of
eHsp90α in cell-conditioned medium. This required the isolation of different EVs from cellconditioned medium. There were three commonly used methods for EV isolation,
including 1) sequential centrifugation, 2) ultrafiltration, and 3) immunoaffinity isolation.
Sequential centrifugation is the most-commonly used method for EV isolation. Therefore,
47
to analyze the distribution and quantitation of eHsp90α in cell-conditioned medium, I
modified the sequential centrifugation protocol (Chapter 2.5) [129]. Taking advantage of
these modified three steps of sequential centrifugations allows the collection of three
different fractions from CMs, including 1) microvesicle (MV), 2) exosome (Exo), and 3)
EVs-depleted supernatant (Sup).
3.2.2 Experimental design
To screen the distribution of eHsp90α in cell-conditioned medium, I carried out
sequential centrifugation to isolate three fractions from cell-conditioned medium including
microvesicles, exosomes, and EV-depleted supernatant. Three 150 mm culture dishes of
MDA-MB-231 cells with 80% confluence were prepared and serum starved for 16 hours.
CMs were collected and subjected to sequential centrifugation. The three fractions were
collected and subjected to western immunoblotting analysis with antibody against Hsp90α
and CD9 (exosome marker), respectively.
To quantitate eHsp90α in EV-depleted supernatant, I used different amounts of
rhHsp90α to generate a standard curve. The concentrations of rhHsp90α used were 30,
60, 120, 240, and 480 ng. Using the standard curve, I aimed to quantify the eHsp90α in
cell conditioned medium of MDA-MB-231 cells. In three independent experiments, one
150 mm dish of MDA-MB-231 cell with 80% confluence was prepared and starved for 16
hours and CM was collected and concentrated. Samples collected from the three
independent experiments along with the rhHsp90α standards were subjected to western
immunoblotting analysis with antibody against Hsp90α for a direct comparison.
48
To rule out the possibility that EV-depleted supernatant (Sup)-associated Hsp90α is
contained inside any unidentified small vesicles, I also designed a trypsin digestion assay
to treat Sup-associated Hsp90α. If the eHsp90α is a “naked” protein, the eHsp90α should
be fully digested by trypsin. If the eHsp90α is inside small and unidentified vesicles, the
eHsp90α should not be fully digested by trypsin. One 150 mm culture dish of MDA-MB231 cells with 80% confluence was prepared and serum starved for 16 hours. Then, the
exosome and EV-depleted supernatant was isolated and collected by using sequential
centrifugations. An equal amount of rhHsp90α (500 ng) was treated with varying
concentrations of trypsin (0, 50, 150, 450 µg/mL) as a positive control. One dish of
exosomes was treated with same varying amounts of trypsin as a negative control, since
the exosome marker CD9 is a transmembrane protein, and it cannot be fully digested by
trypsin. Similarly, one dish of Sup was also treated with same varying amounts of trypsin
to see whether eHsp90α in Sup is a “naked” protein.
3.2.3 Results
3.2.3.1 The majority of eHsp90α is not associated with extracellular vesicles
As shown in Figure 3.6, three 150 mm culture dishes of MDA-MB-231 cells were
prepared and serum starved for 16 hours. CMs were then collected and subjected to
sequential centrifugations. Three fractions including microvesicle (MV), exosome (Exo),
and EV-depleted supernatant (Sup) were isolated and subjected to western
immunoblotting analysis. To visualize the signal of EV-associated Hsp90α by western blot,
the sample loading of EVs (MV and Exo) and Sup were not equal. Sample preparation of
49
Sup was reduced to a quarter of a 150 mm culture dish (~5 millions of cells). In contrast,
the sample preparation of EVs came from three 150 mm culture dishes (~60 millions of
cells) to ensure a detectable signal. The distribution of eHsp90α in each fraction are
shown in panel a. According to the sample loading discrepancy and band intensity
analyzed by image J, the data in panel a shows that approximately 95% of eHsp90α are
not associated with exosomes. Surprisingly, only 3% of eHsp90α is associated with
exosomes and 2% associated with microvesicles. The CD9 exosome maker was also
included to indicate the success of the fractionation (panel b/ lane 2).
Figure 3.6 Distribution of eHsp90α in three fractions that isolated from cell-conditioned medium. After starvation, CM
was collected and subjected to sequential centrifugation to isolate and collect microvesicles (MV, lane 1), exosome
(Exo, lane 2), and EVs-depleted supernatant (Sup, lane 3). Three fractions were then subjected to western
immunoblotting analysis with antibody against Hsp90α (panel a) and CD9 (panel b). Band intensities were measured
and analyzed by image J.
3.2.3.2 Quantitation of eHsp90α in cell-conditioned medium
As shown in Figure 3.7, different amounts of rhHsp90α were used as reference to
quantitate Sup-associated Hsp90α (lane 1 to lane 5). Three independent experiments
were carried out to isolate and collect eHsp90α in CMs from five million MDA-MB-231
50
cells after 16 hours starvation. eHsp90α from the three samples and rhHsp90α protein
standard was immunoblotted with an anti-Hsp90α antibody. The band intensities were
analyzed by image J for further quantitation.
Figure 3.7 Quantitation of eHsp90α in cell conditioned medium after 16 hours starvation. Different amounts of
recombinant human Hsp90α (30 ng, 60 ng, 120 ng, 240 ng, and 480 ng) were used as reference to quantitate eHsp90α
(from lane 1 to lane 5). Three independent experiments were carried out to isolate and collect eHsp90α (from lane 6 to
lane 8). Samples collected from five million cells together with rhHsp90α were subjected to western immunoblotting
analysis with antibody against Hsp90α.
As shown in Figure 3.8, to quantitate the specific concentration of eHsp90α, I
performed densitometry on the eHsp90α bands in each independent replicate (showed in
lane 6 to lane 8) and rhHsp90α (showed in lane 1 to lane 5) using image J. A linear
regression was done using the rhHsp90α intensities and concentration to generate a
standard curve. The X-axis shows the concentrations of rhHsp90α (30 ng, 60 ng, 120 ng,
240 ng, and 480 ng), and the Y-axis shows the corresponding band intensities (measured
from Figure 3.7). The resulting linear equation based on these six coordinates was:
y=0.0019x + 0.2353 (R2=0.9905)
Based on the standard curve above and the band intensities of eHsp90α, the
amounts of eHsp90α quantified from the three replicates were:
51
Expt.#1: 913 ng
X= Expt.#2: 1071 ng
Expt.#3: 576 ng
In average: (913+1071+576) ÷ 3 ≈ 0.85 µg/ 5 × 106 cells
In conclusion, the average amount of eHsp90α protein collected from the CM of three
independent replicates of five million MDA-MB-231 cells was 0.85 µg.
Figure 3.8 The liner regression curve of positive control determines the concentration of Sup-associated Hsp90α. Based
on the band intensities of increasing amounts of rhHsp90α measured by image J, a linear equation was calculated as
a reference to predict and calculate the Sup-associated Hsp90α. X-axis shows the different amounts of rhHsp90α
including, 30 ng, 60 ng, 120 ng, 240 ng, and 480 ng. Y-axis shows the band intensities of each amount of rhHsp90α.
Six coordinates can be formed according to each rhHsp90α and corresponding band intensity. The liner regression
curve was y=0.0019x + 0.2353 (R2=0.9905).
3.2.3.3 eHsp90α from EV-depleted supernatant is a “naked” protein.
To rule out the possibility that EV-depleted supernatant (Sup)-associated Hsp90α is
contained inside any unidentified small vesicles, one 150 mm culture dish of MDA-MB231 cells with 80% confluence was prepared and starved for 16 hours. Exosome and EV-
52
depleted supernatant was collected. An equal amount of rhHsp90α (500 ng) was treated
with varying concentrations of trypsin (0, 50, 150, 450 µg/mL) as a positive control. One
dish of exosomes was treated with same amounts of trypsin as a negative control and
one dish of Sup was treated with same amounts of trypsin to see whether eHsp90α in
Sup is a “naked” protein.
As shown in Figure 3.9, rhHsp90α protein was sensitive to as low as 50 µg/ml trypsin
digestion (panel a). Sup-associated Hsp90α was also sensitive to as low as 50µg/ml
trypsin digestion (panel b). As expected, however, CD9 was much less sensitive to trypsin
digestion since it is a transmembrane protein associated with exosomes (panel c). This
data indicated that Sup-associated Hsp90α is a “naked” protein and it is not contained in
any small unidentified vesicles.
Figure 3.9 Supernatant-associated eHsp90 is a naked protein and it is not contained any small vesicles. One dish of
MDA-MB-231 cell with 80% confluence was prepared and serum starved for 16 hours. Exosomes and EV-depleted
supernatant were isolated and collected by sequential centrifugations. rhHsp90α (500 ng) and exosomes (one culture
dish) were treated with same increasing amounts of Trypsin as positive and negative control, respectively (panel a and
panel c). Sup-associated Hsp90α was also treated with same increasing amount of trypsin (panel b).
3.2.4 Discussion
53
Studies have identified exosome secretion as one mechanism supporting the
secretion of Hsp90α [63-67,126-128]. However, these studies neither quantitated the
amount of exosome-associated eHsp90α nor investigated other possible transportation
routes for eHsp90α. A modified sequential centrifugation protocol allowed three different
fractions including microvesicle, exosome, and EV-depleted supernatant to be isolated so
that the distribution and quantitation of eHsp90α could be analyzed. The exosome
pathway was once thought to be the major route of Hsp90α secretion. Further analysis of
the extracellular fractions collected from CM revealed that majority of eHsp90α (~95%) is
not associated with either exosome (~3%) or microvesicles (2%). Rather, 95% of
eHsp90α was observed in the EV-depleted cell-conditioned medium (Figure 3.10). Supassociated eHsp90α was sensitive to trypsin digestion as opposed to exosome marker
CD9. Therefore, the majority of eHsp90α appears to be present in the cell-conditioned
media as a “naked” protein and neither within nor on the surface of vesicles. In other
words, the majority of eHsp90α is regulated by unknown protein secretion pathway(s)
apart from the minor (3%) portion accounted by exosomes. At this point, the study
direction of this thesis focused on the investigation of the unknown pathways that regulate
the majority secretion of eHsp90α.
54
Figure 3.10 Distribution and quantitation of eHsp90α in cell-conditioned medium. The 95% of eHsp90α is regulated by
unknown pathway(s). Only 5% of eHsp90α are associated with extracellular vesicles.
Given the distribution of eHsp90α, I also sought to quantitate the concentration of
eHsp90α in each fraction. Using standard curve from known concentration of rhHsp90α
as reference, I calculated that for 5 million MDA-MB-231 breast cancer cells under
starvation for 16 hours secrete approximately 0.85 µg of eHsp90α. The purpose was to
determine which fraction I collected from the cell-conditioned medium could provide the
functional concentration of eHsp90α necessary to drive wound healing in vivo. According
to our previous studies [36,62], 100-300 µg/mL of Hsp90α is required to promote wound
healing in mice and pigs. Each 150 mm culture dish could grow approximately 20 million
MDA-MB-231 cells with 10 mL serum-free medium to cover and treat the cells. Therefore,
the concentration of eHsp90α collected from one 150 mm culture dish of cells can be
extrapolated as follows:
20 (million cells) ÷ 5 (million cells) ×0.85 (µg eHsp90α) ÷10 (mL medium) = 0.34
µg/mL eHsp90α
Therefore, the concentration of eHsp90α collected from one 150 mm culture dish
is only 0.34 µg/mL and it is far less than the functional concentration of Hsp90α required
55
to promote wound healing in vivo. The explanation of this result is that, when the total
number of cells with 100% confluence in the culture dish is converted to the volume in
cubic centimeters (cm3
) and correspond to the same number, and same volume of cells
in tissues in vivo, the actual volume of tissue fluid surrounding these cells are far more
less than the cell culture medium. In other words, the 20 million cells in our actual body
are not cultured in 10 mL medium, but a very little volume of interstitial fluid. Therefore, in
terms of physiological relevance, the actual concentration of a given molecule in
conditioned medium of the cells must be re-calculated.
3.3 Discovery of the “Cell number to Interstitial Fluid (CIF)” ratio for human tissues
3.3.1 Introduction
As shown in Figure 3.11, Body fluids can be classified into two categories:
intracellular fluid and extracellular fluid. Intracellular fluid, which is defined as the
cytoplasm of cells, accounts for approximately 40% of total body weight. To maintain an
appropriate osmolality, the extracellular fluid accounts for approximately 20% of total body
weight and can be further subcategorized as plasma and interstitial fluid. The plasma
accounts for 5% of total body weight and interstitial fluid accounts for 12% of total body
weight [130]. Taking a human adult male with a 70 kg body weight as an example, the
total body volume is 7×104 cm3 with total of 3.7×1013 cells. The number of blood cells is
2.5×1013
, and the number of non-blood cells is 1.2×1013 [131,132]. The total body fluid
can be calculated to be 42 L. Therefore, the volume of intracellular fluid is 28 L. The
volume of extracellular fluid is 14 L with 3 L plasma and 11 L interstitial fluid [133].
56
Figure 3.11 Composition of body fluids. Body fluid can be categorized as intracellular fluid and extracellular fluid.
Extracellular fluid can be subcategorized as plasma and interstitial fluid.
In Chapter 3.2, the study of the distribution and quantitation of eHsp90α showed that
eHsp90α was detected in both extracellular vesicles and EV-depleted supernatant.
Surprisingly, as opposed to previous studies which reported that exosome is the protein
secretion pathway accounts of eHsp90α, my data (Figure 3.6) indicated that
approximately 95% of eHsp90α is not secreted with EVs but secreted to extracellularly
as a “naked” protein. This finding urged me to investigate the concentration of eHsp90α
associated with each extracellular fraction, since at least 100-300 µg/ml of Hsp90α is
required to promote wound healing in vivo [36, 62]. After the quantitation of eHsp90α by
using increasing amounts of rhHsp90α as reference (Figure 3.7), I calculated that five
million MDA-MB-231 cells under serum starvation for 16 hours, secrete 0.85 µg of Hsp90α.
However, based on this in vitro data, the concentration of eHsp90α collected from one
150 mm culture dish (with 10 mL medium) is only 0.34 µg/mL and it is far less than the
functional concentration of Hsp90α required to promote wound healing in vivo. Therefore,
57
what does this in vitro concentration mean for promoting wound healing in vivo?
While investigation of the concentration of secreted proteins that collected from CM,
it has been overlooked that the ratio of the cell number to the volume of medium in a
cultured dish is irrelevant to the same ratio in vivo. In other word, the volume of interstitial
fluid in vivo is actually much less than the culture medium in vitro for an equal number of
cells, resulting the drastically decreasing concentration of cell-secreted protein in vitro.
Therefore, to determine which extracellular fraction from cell-conditioned medium
provides the functional concentration of eHsp90α to promote wound healing in vivo, I
established so-called a “Cell number to Interstitial Fluid (CIF)” ratio to accurately predict
tissue concentration of eHsp90α based on in vitro data.
3.3.2 Experimental design
To prove the hypothesis that the concentration of secreted protein is highly diluted in
vitro because of the discrepancy of the volume of cell culture fluids between in vitro and
in vivo, I planned to count the cell number in each 1 cm3
for three types of cells including
MDA-MB-231 breast cancer cells human dermal fibroblasts (HDF), and human
keratinocytes (HKC) and then average the three. Ten 150 mm culture dishes of three
types of cell lines were prepared, respectively. The cells were trypsinized and pelleted in
1.5 mL Eppendorf tubes until the cell volume reached 0.1 cm3
. 100µl (0.1 cm3
) PBS was
used as control. When the volume of cell reached 0.1 cm3
, the cell numbers of each cell
line were counted and three number were averaged. The averaged cell number in each
cubic centimeter (1mL) would be used to mimic the tissue cell number in the same volume
58
so that I could calculate how many folds of secreted proteins were diluted in vitro.
Now that the observed data 0.85 µg eHsp90α / 5 million cells concentration in vitro
has been proposed to not reflect the actual concentration of eHsp90α in vivo, the actual
concentration of eHsp90α in conditioned medium of the cells must be recalculated. The
volume of interstitial fluid and the number of cells within the same volume (cm3
) in vivo
was considered to ensure physiological relevance. The discrepancy of the volume of cell
culture fluids between in vitro and in vivo could be rectified with two statistics: 1) the
number of cells per cm3
in human body, 2) the volume of interstitial fluid per cm3
in the
human body. These two parameters allow the calculation of the cell number per each 1
mL of interstitial fluid in the human body, which would thereby allow me to recalculate the
tissue concentration of eHsp90α based on in vitro data.
3.3.3 Results
3.3.3.1 The concentration of secreted protein in vitro does not reflect the actual
concentration in vivo.
As shown in Figure 3.12, three different types of cells including MDA-MB-231,
human dermal fibroblast, and human keratinocytes were resuspended and centrifuged in
1.5 mL Eppendorf tube until the volume of the cell pellet reached 0.1 cm3
(100µl). 0.1 cm3
(100 µl) PBS was used as a control. The cell numbers were counted by using
hemocytometer. The number of MDA-MB-231 cells, HDF, and HKC in 1 cm3
is 4.5 × 108
,
1.8 × 108
, and 2.7 × 108
, respectively. The average of number of cells per cubic centimeter
in vitro is calculated as follows:
59
(4.5 × 108 + 1.8 × 108 + 2.7 × 108
) ÷ 3 ≈ 3 × 108
/ cm3
Therefore, the projected cell number of 1 cm3 volume would be approximately 3×108
with no more than 1 mL of surrounding fluid. To culture 3×108 cells, however, at least thirty
15 cm dishes with totally 600 mL medium would be required (if assuming 1×107 cells in
each 15 cm dish with 20 mL of culture medium). Therefore, the ratio of medium volume
is 600 : 1 between same number of cells in a culture dish and 3-D cubic volume (mimicked
tissue cells in EP tube). In other words, the concentration of secreted protein in CM of cell
culture is at least 600 times diluted in comparison to its concentration in the 3-D volume.
In conclusion, according to the ratio of medium between same number of cells in a
culture dish and 3D- cubic volume, the concentration of secreted protein in CM of cell
culture is highly diluted. Therefore, the in vitro concentration of eHsp90α cannot reflect
the actual concentration of eHsp90α in vivo.
Figure 3.12 Cell number per cubic centimeter in three types of cells. Ten 150 mm culture dishes of MDA-MB-231,
human dermal fibroblast, and human keratinocyte were resuspended and centrifuged in 1.5 ml Eppendorf tube until the
pellet reached 0.1 cm3
(100µl). 0.1 cm3
(100 µl) PBS was used as a control. The cell numbers of three cells were
counted using hemocytometer. Experiment was repeated three times to ensure the accuracy of cell counting.
3.3.3.2 Establishment of “Cell-Interstitial Fluid (CIF)” ratio
60
As shown in Figure 3.13, taking a human adult male of 70 kg body weight as an
example, the total body volume (BV) is 7×104 cm3
, the total cell number (TC) is 3.7×1013
,
the number of blood cells (BC) is 2.5×1013 and the number of non-blood cells (NBC) is
1.2×1013 [131, 132]. The human adult male contains a total body fluid (TBF) of 42 L, which
can be classified into 28 L intracellular fluids (INF) and 14 L extracellular fluid (EXF). EXF
can be further subcategorized as 3 L plasma (PL) and 11 L interstitial fluid (IF) [133].
Therefore, based on these known human parameters, the number of cells per 1 cm3 can
be calculated as follows:
First, since I only consider the cells surrounded by interstitial fluid, the number of
blood cells needs to be excluded:
1) Tissue cells (NBC) = TC – BC = 3.7×1013 – 1.2× 1013 = 1.5 ×1013
The average number of tissue cells, or NBC, per cm3 can be calculated as follows:
2) NBC/vol. = NBC ÷ BV = 1.5 ×1013 ÷ 7×104 = 1.7 ×108 (cells/cm3
)
Second, the average volume of tissue fluid per cm3 can be calculated as follows:
3) IF/vol. = IF ÷ BV = 11 ÷ 7×104 = 0.15 (mL/cm3
)
Finally, the human CIF ratio can be calculated as follows:
4) CIF ratio = NBC/vol. ÷ IF/vol. = 1.7 ×108 ÷ 0.15 = 1.1 ×109 cells/mL IF
Taken together, the CIF ratio represents the cell number per 1 mL interstitial fluid in
the human body. After calculation, I concluded that there are approximately 1.1 billion
cells in each 1 mL interstitial fluid.
61
Figure 3.13 Establishment of human “Cell number to Interstitial Fluid (CIF)” ratio.
3.3.3.3 The unknown pathway regulates the functional concentration of eHsp90α
Now that a human CIF ratio of 1.1 ×109 cells/mL IF has been established, and 0.85
µg of eHsp90α can be secreted by 5 ×106 cancer cells under serum starvation, the actual
tissue concentration of eHsp90α can be calculated as follows:
1.1 ×109 cells/mL ÷ 5 ×106 cells × 0.85 µg ≈ 187 µg eHsp90α/mL IF
Therefore, the average tissue concentration of eHsp90α that secreted by MDAMB231 cells under stress is 187 µg/mL.
According to the conclusion I made in Figure 3.10, 95% of eHsp90α is associated
with Sup, and only 5% eHsp90α is associated with extracellular vesicles. The
concentration of eHsp90α in each part can be approximated as shown in Figure 3.14:
Figure 3.14 The tissue concentration of eHsp90α associated with different fractions.
62
Therefore, the concentration of eHsp90α regulated by unknown pathway is 178
µg/mL. The concentration of eHsp90α associated with extracellular vesicles is 9.35 µg/mL.
If we only consider the lowest concentration that required for wound healing and do not
take the potency of EV-associated eHsp90α into account, then this data indicated that it
is the unknown pathway, but not EVs, provide the functional concentration of eHsp90α to
promote wound healing in vivo.
3.3.4 Discussion
Exosomes are one of the secretion mechanisms that account for eHsp90α, whereas
my preliminary data showed that only 5% of eHsp90α is associated with extracellular
vesicles. The other 95% of eHsp90α is regulated by an unknown pathway. After
quantitation, five million MDA-MB-231 cells secrete 0.85 µg of eHsp90α after 16 hours
starvation. However, the concentration of eHsp90α calculated from the in vitro study does
not reflect the actual tissue concentration of eHsp90α. To culture the same number of
cells, the volume of cell culture medium required for covering cells in culture dish is much
larger than the volume of interstitial fluid that surrounds tissue cells in vivo. Functionally,
this drastically decreases the concentration of cell-secreted protein in vitro. Therefore, to
calculate the tissue concentration of a given molecule based on in vitro data, I must know
how many cells are cultured in 1 mL interstitial fluid in human body, so called human CIF
ratio. Based on several known human parameters, I calculated that for each 1 mL of
interstitial fluid in the human body, there are approximately 1.1 ×109 cells. Furthermore,
the average tissue concentration of eHsp90α in the human body was calculated based
63
on the CIF ratio to be 187 µg/mL.
The tissue concentration of eHsp90α that associated with EVs and Sup was also
calculated to be 9.35 µg/mL and 178 µg/mL, respectively. Therefore, exosome secretion
only provides 9.35 µg/mL of eHsp90α, which is far below the 100-300 µg/mL of eHsp90α
to promote wound healing in vivo. In contrast, the unknown secretory pathway provides
178 µg/mL of eHsp90α, which is enough to promote wound healing in vivo. The potency
of Exo-associated Hsp90α could be a variable that affect the actual functionality of
eHsp90α to the receiving cells. Considering exosome is a well-known carrier for
extracellular protein transportation, there is still a possibility that Exo-associated Hsp90α
is easier for cells to take in or activate LRP-1 receptor in cell surface. However, technically,
it is hard to compare the potency of Exo-associated Hsp90α to the Sup-associated
Hsp90α since Hsp90α is not the only one cargo inside exosome. Hundreds of other
proteins and small RNAs also could be variables contributing to the potency of exosome
in driving wound healing and tumorigenesis [82]. Therefore, if I only focus on the
concentration of given eHsp90α in different fractions rather than the delivery potency, I
concluded that it is the unknown pathway but not extracellular vesicles which provide the
functional concentration of eHsp90α. Further investigation of the mechanism of the
unknown pathway regulating the functional concentration of eHsp90α was required.
3.4 Golgi reassembly stacking protein 55 regulates partial secretion of eHsp90α.
3.4.1 Introduction
Protein secretion pathways can be classified into two categories. The first is the
64
conventional protein secretion pathway (CPS) regulated by the ER and Golgi, which
requires the presence of a signal peptide at the N-terminus of proteins to be secreted [69].
The second is unconventional protein secretion pathways (UPS), which control the
secretion of leaderless proteins lacking a signal peptide. UPS can be further
subcategorized into four different types, including: 1) type I, plasma membrane pore
formation mediated protein secretion; 2) type II, ABC-transported based protein secretion;
3) type III, organelle-based protein translocation, and 4) type IV, Golgi-bypass [71]. Type
I UPS requires the secreted protein to be able to self-oligomerize and then insert into the
plasma membrane and be secreted outside [72]. In regard to Hsp90α, it can only form
dimers by C-terminal dimerization, excluding it from Type I UPS. Type II UPS is mediated
by ABC-transporter utilizing the energy of ATP hydrolysis to pump out small molecules
such as sugars, ions, amino acids, and small peptides [75]. However, Hsp90α is a 90 kDa
protein and is therefore too large to transport through ABC-transporter. Type IV UPS,
otherwise known as Gogi bypass, occurs by the signal peptide containing protein being
directly transported from ER to plasma membrane and bypassing Golgi [85]. However,
Hsp90α does not have a signal peptide on its N-terminus. Therefore, the most likely
explanation for the non-exosome secretion of eHsp90α is Type III UPS, which is
organelle-based protein translocation.
The organelles involved in type III UPS are proposed to be endosomes and
autophagosomes [79]. Late endosomes have been shown to fuse with the plasma
membrane and then release its intralumenal vesicles, so called exosomes to the
65
extracellular space [124]. However, my data showed that less than 5% of eHsp90α is
associated with exosomes. Therefore, the most likely explanation for the 95% of eHsp90α
is autophagosomes.
Autophagy, once viewed as a cytoplasmic auto-digestive process, has a nondegradative function, called secretory autophagy. Studies showed that degradative
autophagy and secretory autophagy share the same biogenesis by the aiding of GRASP
at the exit point of ER. After maturation, rather than fuse with lysosome to degrade its
cargo, secretory autophagosomes fuse with plasma membrane and release its cargo to
extracellular space [93].
Golgi reassembly stacking protein (GRASP) plays a central role in controlling Golgi
stacking and organizing Golgi structure. GRASP65 is the first identified GRASP protein
that regulates Golgi stack formation, cell migration, and apoptosis [97,98]. Unlike
GRASP65, GRASP55 has been shown to play a distinct role in regulating unconventional
protein secretion and secretory autophagy [105]. Therefore, to investigate whether
secretory autophagy regulates the 95% of eHsp90α, I planned to target GRASP55 first.
3.4.2 Experimental design
First, to study whether GRASP55 regulates the 95% of eHsp90α, our laboratory
collaborated with Dr. Yanzhuang Wang from the University of Michigan. They provided us
HeLa GRASP55 and HeLa GRASP65 gene knockout cell lines. The HeLa GRASP55
knockout cell line was used to study the secretion of eHsp90α. The HeLa GRASP65
knockout cell line was also included as a control since GRASP65 does not play critical
66
role in unconventional protein secretion pathway and autophagy [105]. In this experiment,
three identical dishes of HeLa wild-type (WT) cells, HeLa GRASP55 knockout (KO) cells,
and HeLa GRASP65 knockout (KO) cells with 80% cell confluence were prepared and
serum starved for 16 hours. Then, cell-conditioned medium from the three cell lines was
collected and subjected to sequential centrifugation to remove floating cells and
extracellular vesicles. The EV-depleted supernatants were then concentrated and
subjected to western immunoblotting analysis with anti-Hsp90α antibody. The total lysates
of three cell lines were collected and subjected to western immunoblotting analysis with
anti-Hsp90α, anti-GRASP55, anti-GRASP65, and anti-β-actin to measure the expression
level of intracellular Hsp90α and the knockout efficiency of GRASP55 and GRASP65.
Second, to verify that the inhibition of eHsp90α was specifically due to the absence
of GRASP55, I planned to re-introduce exogenous GFP-GRASP55 cDNA into the HeLa
GRASP55KO cell line. In this experiment, three identical dishes of HeLa WT cells, HeLa
GRASP55KO cells, and HeLa GRASP55KO+GFP-GRASP55 cells with 80% cell
confluence were prepared and serum starved for 16 hours. Cell-conditioned medium from
the three cell lines was collected and subjected to sequential centrifugation to remove
floating cells and extracellular vesicles. The EV-depleted supernatants were then
concentrated and subjected to western immunoblotting analysis with anti-Hsp90α
antibody. The total lysate of the three cell lines was collected and subjected to western
immunoblotting analysis with anti-Hsp90α, anti-GRASP55, and anti-β-actin to verify that
GFP-GRASP55 was re-introduced into the HeLa GRASP55 knockout cell line.
67
Third, to test whether the inhibition of eHsp90α in HeLa GRASP55KO cells was cell
line specific, I carried out shRNA-mediated knockdown (KD) of GRASP55 in both MDAMB-231 cells and immortalized human keratinocytes (IKC). In these experiments, both
WT and GRASP55KD cells for MDA-MB-231 cells and IKC with 80% confluence were
prepared and serum starved for 16 hours. Cell-conditioned medium from three cell lines
were collected and subjected to sequential centrifugation to remove floating cells and
extracellular vesicles. The EVs-depleted supernatants were then concentrated and
subjected to western immunoblotting analysis with anti-Hsp90α antibody. The total lysates
of above cell lines were collected and subjected to western immunoblotting analysis with
anti-Hsp90α, anti-GRASP55, and anti-β-actin to verify the knockdown efficiency of
GRASP55.
3.4.3 Results
3.4.3.1 GRASP55 knockout partially reduced Hsp90α secretion in HeLa cell line
As shown in Figure 3.15, The evidence of GRASP55KO and GRASP65KO are
shown in panel a and panel b. Panel c shows that the intracellular expression of Hsp90α
was not affected by GRASP55KO or GRASP65KO. Panel f shows intracellular β-actin as
sample loading control. Emphasized by red arrow, Panel d provides evidence that
GRASP55KO, but not GRASP65KO, partially blocked Hsp90α secretion in HeLa cells.
More interestingly, as shown in Figure 3.16, the Coomassie blue staining of EV-depleted
supernatant indicated that the inhibition of Hsp90α secretion in the HeLa GRASP55KO
cell was not due to the global inhibition of protein secretion since the overall protein
68
secretion by the GRASP55KO cells was indistinguishable from the wildtype or
GRAPS65KO cells.
Figure 3.15 GRASP55 but not GRASP65 knockout, partially blocked eHsp90α in HeLa cell line. In GRASP-55-knockout
HeLa cells and with the related but functionally distinct GRASP65KO cells as a specificity control. The evidence of
GRASP55 knockout (panel a, lane 2) and GRASP65 knockout (panel b, lane 3), as well as unchanged intracellular
Hsp90 (panel c) and -actin (panel f), were included. Under these conditions, the total amount of eHsp90 (secreted
Hsp90) was dramatically reduced (panel d, lane 2 vs. lanes 1 and 3, as indicated by a red arrow).
Figure 3.16 The inhibition of Hsp90α secretion is not due to global inhibition of protein secretion. EV-depleted Sup
collected from each three cell lines were subjected to SDA-PAGE gel followed by Coomassie blue staining. No
significant difference was detected between each lane (lane1 vs. lane2 vs. lane3), indicating GRASP55KO and
GRASP65KO do not inhibit the global protein secretion.
69
3.4.3.2 The re-introduction of GFP-GRASP55 rescued Hsp90α secretion in HeLa
GRASP55KO cells
To verify the inhibition of Hsp90α secretion was specifically due to the absence of
GRASP55, exogenous GFP-GRASP55 cDNA was introduced into HeLa GRASP55KO
cells through lentiviral infection. After the stable cell line was established, three identical
dishes of HeLa WT (lane 1), HeLa GRASP55 (lane 2), and HeLa GRASP55KO-GFPGRASP55 (lane 3) were prepared and serum starved for 16 hours when cell confluence
reached 80%. EV-depleted Sup was collected and subjected to western immunoblotting
analysis with anti-Hsp90α antibody. As shown in Figure 3.17 panel a and panel b, the
evidence of the GRASP55 KO (panel a, lane 2), the re-introduction of GFP-GRASP55
(panel a, lane 3), and beta-actin (panel b) were included. The intracellular expression of
Hsp90α was not affected by GRASP55KO or re-introduction of GFP-GRASP55 (panel c).
β-actin was also included as a loading control (panel e). As the red arrow emphasizes in
panel d, the expression of GFP-GRASP-55 completely rescued the secretion of eHsp90
in the GRASP55KO cells, indicating that the inhibition of Hsp90α secretion was
specifically due to the absence of GRASP55.
70
Figure 3.17 Re-introducing of GFP-GRASP55 into HeLaGRASP55KO cells rescued Hsp90α secretion. To verify the
inhibition of Hsp90α secretion was specifically due to the absence of GRASP55, exogenous GFP-GRASP55 cDNA was
introduced into HeLa GRASP55KO cell line. Wildtype HeLa cells express abundant GRASP55) and the GRASP55-KO
cells lack any detectable GRASP55 (panel a, lane1 vs. lane 2). Lentiviral infection-mediated re-expression of a GFPGRASP-55 shows a similar level to the endogenous GRASP55 (panel a, lane 3 vs. lane 1). Under these conditions,
the expression of GFP-GRASP55 completely rescued the secretion of eHsp90 in the GRASP55-KO cells (panel d,
lane 3 vs. lane 2). The endogenous Hsp90 (panel c) and -actin (panel b and f) were included as protein loading
controls.
3.4.3.3 GRASP55 knockdown partially reduced Hsp90α secretion in both MDA-MB-231
cells and immortalized human keratinocytes.
To confirm that inhibition of Hsp90α secretion in the HeLa GRASP55KO cell line is
not cell type specific, I downregulated GRASP55 in both MDA-MB-231 cells and
immortalized human keratinocytes. As shown in Figure 3.18 A, a partial knock down of
GRASP55 was achieved and β-actin was included as loading control. As shown in Figure
3.18 B, the intracellular expression of Hsp90α was not affected by either GRASP55 or
GRASP65 knockdown (panel c). β-actin was also included as a loading control (panel e).
71
Under these conditions, as shown in panel d, downregulation of GRASP55 but not
GRASP65 partially reduced the amount of eHsp90 in MDA-MB-231 cells (panel d, lane
2 vs lanes 1 and 3, as indicated by a red arrow).
Figure 3.18 Downregulation of GRASP55 but not GRASP65 partially reduced Hsp90α secretion in MDA-MB-231 cell
line. The partial downregulation of GRASP55 (Figure A, panel a) and β-actin as a loading control are shown.
Downregulation of GRASP55 but not GRASP65 significantly reduced the amount of eHsp90 (panel d, lane 2 vs lanes
1 and 3, as indicated by a red arrow).
In addition to MDA-MB-231 cells, I also downregulated GRASP55 in immortalized
keratinocytes (IKC). As shown in Figure 3.19, three identical dishes of IKC WT and IKC
GRASP55KD with 80% confluence were prepared and serum starved for 16 hours. EVdepleted Sup was collected and subjected to western immunoblotting analysis with antiHsp90α antibody. The partial downregulation of GRASP55 in IKC (panel a) and β-actin
(panel d) as a loading control are shown. The intracellular expression of Hsp90α was not
affected by GRASP55KD in IKC (panel c). Under these conditions, downregulation of
GRASP55, partially reduced the amount of eHsp90 (panel b, lane 2 vs lane 1 as
indicated by a red arrow).
72
Figure 3.19 Down-regulation of GRASP55 partially reduced Hsp90α in immortalized human keratinocytes (IKC). The
partial downregulation of GRASP55 in IKC (panel a) and β-actin (panel d) as a loading control are shown. The
intracellular expression of Hsp90α was not affected by GRASP55KD in IKC (panel c). Under these conditions,
downregulation of GRASP55 partially reduced the amount of eHsp90 (panel b, lane 2 vs lane 1 as indicated by a red
arrow).
3.4.4 Discussion
Given Hsp90α lacks a signal peptide, only unconventional proteins secretion
pathways (UPS) could explain the secretion mechanism of eHsp90α. Type III UPS, which
is an organelle-mediated protein secretion, is proposed to be exosome or
autophagosome related [79]. However, as shown in Figure 3.10, less than 5% of
eHsp90α associates with exosomes. In contrast, 95% of eHsp90α was regulated by an
unknown protein secretion pathway. Therefore, I hypothesized that it is the
autophagosome but not exosome that regulates this 95% of eHsp90α.
Autophagy, once viewed as a cytoplasmic auto-digestive process, has a nondegradative function so called secretory autophagy [93]. To investigate whether secretory
73
autophagy mediates the 95% of eHsp90α, I first examined the supernatant associated
Hsp90α in HeLa GRASP55KO cells. The HeLa WT and HeLa GRASP65 cell line was
also included as control since GRASP65 does not play a critical role in protein secretion
[105]. The data in Figure 3.15 indicated that GRASP55KO but not GRASP65KO partially
blocked Hsp90α secretion in HeLa cells. Moreover, a Coomassie blue stain indicated that
the inhibition of Hsp90α was not due to global inhibition of protein secretion (Figure 3.16).
To further verify that the inhibition of Hsp90α secretion was specifically due to the absence
of GRASP55, exogenous GFP-GRASP55 cDNA was introduced into HeLa GRASP55KO
cell line. As shown in Figure 3.17, GFP-GRASP55 fully rescued Hsp90α secretion in
HeLa GRASP55KO cells, indicating the specificity of GRASP55 to control Hsp90α
secretion. Validation that the inhibition of eHsp90α in HeLa GRASP55KO cells is not celltype specific, was achieved by lentiviral infection-mediated knockdown of GRASP55 in
both MDA-MB-231 breast cancer cell line and immortalized human keratinocytes. As
shown in Figure 3.18 and Figure 3.19, down-regulation of GRASP55 in both MDA-MB231 cell and IKC partially reduced Hsp90α secretion, indicating that the inhibition of
Hsp90α in GRASP55KO or GRASP55KD is not cell type specific.
Taken together, this section illustrated that as a secretory autophagy-mediated factor,
GRASP55 is partially required for Hsp90α secretion in both tumor and normal cell lines.
Therefore, secretory autophagy is likely the mechanism which accounts for the majority
of Hsp90α secretion. However, only around 50% of eHsp90α was reduced in response to
GRASP55 knock out or knock down. Further mechanistic investigation is required to
74
provide stronger evidence that secretory autophagy is in fact the primary pathway
accounts for Hsp90α secretion. To further investigate this, I decided to target five genes
that directly control the biogenesis of the autophagosome, including Atg5, Atg7, Atg16,
Beclin-1, and LC3.
3.5 Autophagosome biogenesis genes regulate the remaining portion of Hsp90α
secretion.
3.5.1 Introduction
As previous introduced, autophagy can be classified into two categories including 1)
degradative autophagy, and 2) secretory autophagy [93]. Degradative autophagosome
fuses with lysosome to degrade its contents. Secretory autophagosome fuses with
plasma membrane to releases its content into extracellular space. Regardless the final
distinct destiny of degradative autophagosome and secretory autophagosome, both
vesicles share the same autophagosome biogenesis process. The biogenesis of
autophagosome is consist of three steps including 1) initiation, 2) elongation, and 3)
maturation [106].
The initiation of autophagosome biogenesis requires the aiding of the Vps34-
Atg6/beclin-1 class III PI3-kinase complex. The elongation and maturation step of
autophagosome biogenesis requires the aiding of the two ubiquitin-like conjugation
systems, including Atg5-Atg12-Atg16L1 conjugation system, and Atg8/LC3 conjugation
system [112-117]. These three systems control the biogenesis of autophagosome.
In Chapter 3.4, the results showed that approximately 50% of eHsp90α is regulated
75
by GRASP55-mediated secretory autophagy. The partially reduce of eHsp90α due to
GRASP55 absence urged me to re-investigate the role of secretory autophagy plays in
eHsp90α secretion by directly targeting other autophagosome biogenesis genes. Vps34-
Atg6/beclin-1 class III PI3-kinase complex, Atg5-Atg12-Atg16L1 conjugation system, and
Atg8/LC3 conjugation system play a critical role in the initiation, elongation, and
maturation step of autophagosome biogenesis. Therefore, I planned to target these three
essential complexes to disrupt the biogenesis of autophagosome using shRNA and
investigate the level of eHsp90α.
3.5.2 Experimental design
First, I planned to target Atg5 and Atg7 to confirm the conclusion that secretory
autophagy regulates eHsp90α secretion. Atg5 is one of the essential components of Atg5-
Atg12-Atg16L1 conjugation system that regulates the elongation step of autophagosome
biogenesis [113]. Atg7 is an E1 ubiquitin activating enzyme to activate the conjugation of
Atg12 and Atg5 [114]. By using lentiviral shRNA delivery system, I established MDA-MB231 Atg5 and Atg7 knockdown cell lines. In this experiment, three identical 15 cm dishes
of MDA-MB-231 wild type (WT), MDA-MB-231 Atg5 knockdown (KD), and MDA-MB-231
Atg7 knockdown (KD) with 80% cell confluence were prepared and serum starved for 16
hours. Cell-conditioned medium from three cells were collected and subjected to western
immunoblotting analysis with anti-Hsp90α antibody. Total lysates of three cells were
collected and subjected to western immunoblotting analysis with anti-Hsp90α, anti-Atg5,
anti-Atg7, and anti-β-actin to check the intracellular level of Hsp90α and the knockout
76
efficiency of Atg5 and Atg7.
Second, to further confirm secretory autophagy regulates eHsp90 secretion, I
planned to screen three other critical genes that involved in the different steps of
autophagosome biogenesis, including Atg16, Beclin-1, and LC3. Like Atg5, Atg16 is
another critical component of Atg5-Atg12-Atg16L1 conjugation system regulating the
elongation step of autophagosome biogenesis [117]. Beclin-1 is a central component of
Vps34-Atg6/beclin-1 class III PI3-kinase complex regulating the initiation step of
autophagosome biogenesis [112] and LC3 is an important component of Atg8/LC3
conjugation system regulating the elongation and maturation steps of autophagosome
biogenesis [114]. In this experiment, MDA-MB-231 Atg16 knockdown, Beclin-1
knockdown, and LC3 knockdown cell line were established and three identical 150 mm
dishes of MDA-MB-231 WT, Atg5KD, Atg7KD, Atg16KD, Beclin-1KD, and LC3KD cells
were prepared with 80% confluence and serum starved for 16 hours. EV-depleted
supernatants were collected and subjected to western immunoblotting analysis with antiHsp90α antibody. Total lysates of six cells were also collected and subjected to western
immunoblotting analysis with anti-Hsp90α, anti-Atg5, anti-Atg7, anti-Atg16, anti-Beclin-1,
anti-LC3, and anti-β-actin to check the gene knockout efficiency.
Finally, to investigate whether GRASP55 and autophagosome biogenesis genes
regulate two parallel secretory autophagy pathways, I planned to establish GRASP55 and
Atg5 double knockdown cell line. In this experiment, I established MDA-MB-231
GRASP55/Atg5 double knockdown cell line. Three identical 150 mm dishes of MDA-MB-
77
231 WT, MDA-MB-231 GRASP55KD, MDA-MB-231 Atg5KD, and MDA-MB-231
GRASP55/Atg5 double KD cells with 80% confluence were prepared and serum starved
for 16 hours. EV-depleted supernatants were collected and subjected to western
immunoblotting analysis with anti-Hsp90α antibody. Total lysates of MDA-MB-231 Atg5KD
and GRASP55/Atg5 double KD cells were also collected and subjected by western
immunoblotting analysis with anti-Atg5, anti-GRASP55, and anti-β-actin.
3.5.3 Results
3.5.3.1 Autophagosome biogenesis gene Atg5 positively, but Atg7 negatively regulates
Hsp90α secretion
To validate the conclusion that secretory autophagy regulates Hsp90α secretion, I
planned to target autophagosome biogenesis genes, including Atg5 and Atg7. Three
identical dishes of MDA-MB-231 WT, MDA-MB-231 Atg5KD, and MDA-MB-231 Atg7KD
with 80% confluence were prepared and serum starved for 16 hours. EV-depleted Sup
was collected and subjected to western immunoblotting analysis with anti-Hsp90α
antibody. Total lysates of above three cells were also collected and subjected to
immunoblotting analysis. Figure 3.20 A shows the knockdown efficiency of Atg5 or Atg7
in MDA-MB-231 cells (panel a, lane 2 vs lane 1; lane 4 vs lane3). Figure 3.20 B shows
that the knockdown of Atg5 or Atg7 did not affect the intracellular expression of Hsp90α
in MDA-MB-231 cells (panel d). Surprisingly, Atg5KD (panel c, lane2) partially decreased
Hsp90α secretion, whereas downregulation of Atg7 (lane3) slightly enhanced secretion
of Hsp90α.
78
Figure 3.20 Autophagosome biogenesis gene Atg5 positively, but Atg7 negatively regulates eHsp90α secretion. Figure
A shows the knockdown efficiency of Atg5 or Atg7 in MDA-MB-231 cells (panel a). Figure B shows that Atg5 or Atg7
knockdown did not affect the intracellular expression of Hsp90α (panel d) and Atg5 positively, but Atg7 negatively
regulates Hsp90α secretion.
3.5.3.2 Screening of five autophagosome-regulating genes
The contradictory results of Hsp90α secretion between Atg5KD and Atg7KD urged
me to continue investigating the role of autophagosome biogenesis genes played in
mediating Hsp90α secretion in MDA-MB-231 cells. Besides Atg5 and Atg7, Beclin-1,
Atg16, and LC3 also plays indispensable role in regulating the initiation, elongation, and
maturation of autophagosome biogenesis [106]. In this experiment, I established MDAMB-231 Beclin-1KD, Atg16KD, and LC3A KD cell lines. Figure3.21 A shows the
knockdown efficiency of Atg5 (lane 2 vs lane 1), Beclin-1 (lane 4 vs lane 3), Atg16 (lane
6 vs lane 5), LC3A (lane 8 vs lane 7), and Atg7 (lane 10 vs lane 9). Then, I collected EVdepleted supernatants from each five KD cell lines after 16 hours starvation and subjected
it to the western immunoblotting analysis with anti-Hsp90α antibody. As shown in Figure
3.21 B, downregulation of Atg5, Beclin-1, Atg16, and LC3A partially blocked secretion of
79
Hsp90α secretion, whereas downregulation of Atg7 (same as Figure 3.20 B shows)
slightly increased Hsp90α secretion. This result indicated that 1) secretory autophagy
does regulates Hsp90α secretion and 2) lack of Atg7 does not affect the formation of
autophagosome.
Figure 3.21 Screening of five autophagosome biogenesis genes. lentiviral infection-delivered shRNA effectively
downregulated the five genes (Figure A panel a, lanes 1 to 10) in MDA-MB-231 cells. In Figure 7.2 B, downregulation
of Atg5, Atg16, LC3, and Beclin-1 (lane 2, 4, 5, and 6) partially blocked Hsp90α secretion, whereas Atg7 (lane3) slightly
increased Hsp90α secretion in MDA-MB-231 cells.
3.5.3.3 Downregulation of both GRASP55 and Atg5 blocks majority of Hsp90α secretion
Single knockdown of either GRASP55 or autophagosome biogenesis gene in MDAMB-231 cells only partially blocked Hsp90α secretion. Considering these results, I
speculated that double knockdown of both GRASP55 and Atg5 should have addictive
80
effect on inhibition of Hsp90α secretion. In this experiment, I established MDA-MB-231
GRASP55/Atg5 double knockdown cell line. EV-depleted supernatants from MDA-MB231 WT, GRASP55KD, Atg5KD and GRASP55/Atg5 double KD were collected and
subjected to western immunoblotting analysis with anti-Hsp90α antibody. As shown in
Figure 3.22, individual knockdown of either GRASP55 or Atg5 partially reduced secretion
of Hsp90 (lanes 2 and 3 vs lane 1). However, double knockdown of both genes almost
completely blocked Hsp90 secretion (lane 4). I did not expect 100% of eHsp90α
blockade because the level of GRASP55 downregulation is less than 80%. Therefore,
this result may indicate that GRASP55 and autophagosome-biogenesis genes are two
parallel pathways mediating secretory autophagy.
Figure 3.22 Downregulation of both GRASP55 and Atg5 blocked majority of Hsp90α secretion. Single downregulation
of either GRASP55 or Atg5 only partially reduced Hsp90α secretion, while double downregulation of both GRASP55
and Atg5 blocked majority of Hsp90α secretion.
3.5.4 Discussion
The biogenesis of autophagosome is consist of three steps including 1) initiation, 2)
elongation, and 3) maturation [106]. Each step of autophagosome biogenesis requires
the regulation of different complex and conjugation systems. Vps34-Atg6/beclin-1 class
III PI3-kinase complex regulates the initiation of autophagosome formation. Atg5-Atg12-
81
Atg16L1 conjugation system regulates the elongation of autophagosome. Atg8/LC3
conjugation system regulates the elongation and maturation of autophagosome [112-117].
Each complex or conjugation system is indispensable during the biogenesis of
autophagosome otherwise the formation of autophagosome will be disrupted [106]. In
Chapter 3.4, my data showed that eHsp90α is partially regulated by GRASP55-mediated
secretory autophagy. Therefore, to further validate the role of secretory autophagy in
mediating Hsp90α secretion, I decided to directly target genes that involved in
autophagosome formation including Atg5, Atg7, Atg16, Beclin-1, and LC3.
To start with, I established MDA-MB-231 Atg5 and Atg7 single KD cell lines.
Surprisingly, although downregulation of Atg5 partially reduced Hsp90α secretion,
downregulation of Atg7 slightly increased Hsp90α secretion. This contradicted result
urged me to further investigate other autophagosome biogenesis genes including Atg16,
Beclin-1, and LC3. The result showed that except downregulation of Atg7 slightly
enhanced Hsp90α secretion, all the other gene downregulations including Atg5, Atg16,
Beclin-1 and LC3 partially reduced Hsp90α secretion. Three conclusions could be made
from this result. First, secretory autophagy regulates Hsp90α secretion. Second, both
single knockdown of GRASP55 and autophagosome biogenesis genes partially regulates
Hsp90α secretion, indicating that these two types of genes mediated secretory autophagy
could be two parallel pathways. Third, Atg7 is dispensable for autophagosome biogenesis.
To verify the hypothesis that GRASP55-mediated secretory autophagy and
autophagosome biogenesis gene-mediated secretory autophagy are two parallel
82
pathways, I double knocked down both GRASP55 and Atg5 in MDA-MB-231 cells. The
addictive effect on inhibition of Hsp90α secretion was observed in GRASP55 and Atg5
double-downregulation cells, indicating that GRASP55 and Atg5 individually mediates two
parallelled secretory autophagy pathways. To verify this hypothesis, I plan to establish
MDA-MB-231 Beclin-1/GRASP55 double KD cell line and HeLa GRASP55KO/Atg5KD
cell line in the future. Similarly, the total lysates of each double KD cells will be collected
and the KD efficiency will be measured. EV-depleted supernatant also will be collected
and subjected to western immunoblotting analysis to measure the level of secreted
Hsp90α.
Taken together, both single downregulation of GRASP55 and autophagosome
biogenesis genes partially reduced Hsp90α secretion in EV-depleted Sup, indicating
secretory autophagy is the secretion mechanism that account for the functional
concentration of eHsp90α. Double downregulation of GRASP55 and Atg5 blocked
majority of Hsp90α secretion, indicating that there are two parallelled secretory autophagy
pathways collectively mediate the functional concentration (~95%) of extracellular
Hsp90α.
3.6 eHsp90α plays an essential role during wound healing.
3.6.1 Introduction
Tissue injury and repair occur all the time in human body throughout life. The
microenvironment of a fresh skin wound quickly becomes hypoxic and shortage of
nutriment supply due to surrounding blood vessel disruption and high oxygen
83
consumption by cells at the wound edge and in the granulation tissue [134-136]. Like
wound healing, when tumor grows too fast to reach nearby blood vessels, cells inside the
tumor suffered from ischemia. Under both pathological conditions above, cells either turn
on stress-responding mechanism to repair the wound tissue or metastasis from hazard
environments to reconnect with new blood vessels. If both attempts fail, cells turn on
apoptotic programs and die [137].
A series of studies of Li’s group demonstrated that topical application of rhHsp90α
(100-300 µg/mL), but not rhHsp90β, promote wound healing in excision, burn, and
diabetic wounds in mice and pigs [23,36,39,40]. However, there has not been direct
evidence that the action of cell secreted Hsp90α is essential for wound healing. To provide
direct evidence, in collaboration with Dr. Xin Tang, we established Hsp90α Crispr/cas9
gene knockout (Hsp90α-KO) mice and preformed wound healing experiments.
3.6.2 Experimental design
To provide direct evidence that the action of secreted Hsp90α is essential for wound
healing, I first aimed to show the result of massive deposition of eHsp90α in wounded
and tumor-growing tissue. The stocked biopsy sections stained with anti-Hsp90α antibody
in our laboratory were chose and showed. Second, after the establishment of Hsp90αKO mice, we created 8 mm full thickness dorsal wounds and compared the speed of
wound closure between wild type and Hsp90α-KO mice. If the speed of wound healing
was delayed in Hsp90α-KO mice, it could provide direct evidence to show that secreted
Hsp90α is essential for wound healing. Moreover, we planned to topically apply rhHsp90α
84
in Hsp90α-KO mice and see whether the delayed wound closure could be rescued.
3.6.3 Results
3.6.3.1 eHsp90α is massively deposited in injured and tumor-growing tissue
As shown in Figure 3.23, the biopsy sections of wounded skin, injured lung, and
tumor growing liver that stained with anti-Hsp90α antibody were utilized to show the
massive deposition of eHsp90α around wounded or tumor-growing tissue. A biopsy of a
full-thickness excision wound on day 0 showed undetectable staining of monoclonal antiHsp90α antibody (panel a), whereas a biopsy of the same wound on day 4 showed
massive staining of the antibody (panel b brown area, as pointed by red arrows). To
confirm the finding in skin excision wounds, we carried out similar studies using biopsies
of bleomycin-injured lung, a widely used model for studying lung fibrosis in mice, and
biopsies of HFD (high fat diet)- induced liver tumor model in mice. We found that uninjured
lung showed little anti-Hsp90 antibody staining (panel c). In contrast, the injured lung
showed massive anti-Hsp90 antibody staining in the entire section of the biopsy (panel
d). Similarly, increased anti-Hsp90 antibody staining was detected around the HDFinduced liver tumors in mice (panel f vs. panel e).
85
Figure 3.23 Deposition of eHsp90α in both wounded and tumor growing tissue. Wedge biopsies of 1.5 cm full-thickness
pig skin wound on day 0 and day 4 were subjected to anti-Hsp90 antibody staining. The red arrows pointed out the
areas of the specific antibody staining (panel b vs. panel a, brown area). Massive anti-Hsp90 antibody staining was
detected in the entire section of the biopsy from bleomycin-treated lung on day 14, in reference to uninjured lung (panel
d vs panel c). Similarly, anti-Hsp90 staining was detected around tumors (panel f vs. panel e), which took several
months to grow in a mouse fatty liver model.
3.6.3.2 eHsp90α is essential during wound healing.
To provide direct evidence that secreted eHsp90α is essential for wound healing, As
shown in Figure 3.24 A, 8 mm full thickness dorsal wounds were created and
representative images of wound closure was compared between wild type and Hsp90-
KO mice. Wounds in wild type mice underwent the fastest closure, by more than 80% on
day 10 and 100% on day 14 (panels a, b, c). In comparison, wounds in Hsp90-KO mice
showed a significant delay, less than 70% on day 10 and 80% on day 14 (panels d, e, f).
Remarkably, the delay of wound closure in Hsp90-KO mice was completely corrected
by topically applied rhHsp90α protein (panels h and i vs. panels e and f). Quantitation of
the wound closure data is shown Figure 3.24 B.
86
Figure 3.24 eHsp90α is essential for wound healing. 8 mm full thickness wounds were created in wild type and Hsp90-
knockout mice with or without topical treatment of 300 µg/ml rhHsp90 protein (Figure 3.24 A). Wound closure was
measured as % of the open wound area over time in reference to day 0 wounds. Quantitation of the wound closure as
shown in Figure 3.24 B, where wounds in Hsp90-knockout mice remained open on day 14 (red bar, p < 0.05)
To further support the in vivo evidence that eHsp90α is essential to wound healing,
we carried out H&E staining of wedged wound biopsies in both wild type and Hsp90α-KO
mice. As shown in Figure 3.25, wounds in wild type mice showed the longest reepithelialization togue (Re-epi T, as indicated by yellow dotted line) (upper panel), wounds
in Hsp90-KO mice showed the shortest Re-epi T (middle panel), and wounds in Hsp90αKO with topically added rhHsp90α protein reduced the delay of re-epithelialization (bottom
panel).
87
Figure 3.25 H&E staining of wedged wound biopsies in both wild type and Hsp90α-KO mice. Section of partial wounds
on day 10 were subjected to H&E staining. Red vertical dashed lines divide the unwounded (left) and wounded (right)
areas. Yellow horizontal dashed lines show the re-epithelialization tongue (Re-epi T), i. e. epidermal cell migration.
3.6.4 Discussion
Although a series of studies has shown that topical application of rhHsp90α (100-300
µg/mL) promotes wound healing in pigs and mice [23,36,39,40], there has not been direct
evidence that the action of secreted Hsp90α is essential for wound healing. Therefore, to
provide the direct evidence, in collaboration with Dr. Xin Tang, we established Hsp90α
Crispr/cas9 gene knockout mice.
First, to provide the evidence of massive deposition of eHsp90α in wounded and
tumor growing tissue, we showed biopsy sections of wounded skin, injured lung, and
tumor growing liver that stained with anti-Hsp90α antibody. Massive staining of eHsp90α
was detected in wounded and tumor growing sections, indicating the secretion of Hsp90α
under pathological conditions. It must be pointed out that the staining of Hsp90α is not
88
due to intracellular Hsp90α. Because as Figure 3.23 panel a and panel b shows, the
staining area of biopsy section was dermis, in which mostly composed of collagens with
scattered individual fibroblast cells. Therefore, this massive deposition of Hsp90α was not
due to intracellular Hsp90α in fibroblasts.
Second, taking advantage of Hsp90α-KO mice, we created 8 mm full thickness dorsal
wounds on both wild type and Hsp90α-KO mice. The wild type mice underwent the fastest
wound closure. In comparison, wound closure in Hsp90α-KO mice was significantly
delayed. Moreover, topically application of rhHsp90α rescued the delayed wound closure
in Hsp90α-KO mice. Although the main mechanism of mouse (loose) skin wound healing
is by wound contraction, completing the final wound closing would require the epidermal
cell (keratinocyte) migration. Therefore, the data that the wounds remained open on day
14 indicated that keratinocyte migration was affected in Hsp90-KO mice. To further
support this conclusion, we also carried out H&E staining of wedged wound biopsies, the
shortest length of re-epithelialization tough in Hsp90α-KO section confirmed the result of
delayed wound closure. Taken together, this section provides direct in vivo evidence that
eHsp90α is essential in promoting wound healing.
89
Chapter 4. Discussion
4.1 Summarizing discussion
For decades, Hsp90 family proteins have been recognized as intracellular ATPbinding-dependent molecular chaperone that stabilizes other proteins. In the early 2000s,
Jay’s group and Li’s group first discovered and defined the function of extracellular
Hsp90α in promoting cancer cell invasion and skin wound healing [36-38]. In addition to
the promotility activity of eHsp90α, Li’s group also identified another function of eHsp90α,
which is promoting cancer cell survival under hypoxic environment [42]. Therefore, unlike
intracellular Hsp90β which functions as a molecular chaperone and maintains
proteostasis, there are two main biological functions of eHsp90α: 1) promoting cancer cell
survival and 2) promoting cell motility.
These two main biological functions grant eHsp90α an essential role during wound
healing and tumorigenesis. In animal studies, Li’s group demonstrated that topical
application of rhHsp90α, but not rhHsp90β, promoted wound healing in excision, burn,
and diabetic wounds in mice and pigs [23,36,39,40]. In addition, nearly twenty clinical
studies (Figure 1.6) have revealed that the plasma level of eHsp90α dramatically elevated
in patients with various cancer and is positively correlated with the progression of cancer
stage [42-58]. Therefore, plasma eHsp90α could be a potential therapeutic target for
monoclonal antibodies to block cancer metastasis or an ideal serum marker for cancer
prognostic.
Despite the well-establishment of eHsp90α biological functions, the secretion
90
mechanism of Hsp90α has been debated and remained unclear. The Hsp90 community
initially believed that eHsp90α comes from the dead cell leakage until several studies
provided different explanations for the secretion mechanism of eHsp90α [62-67]. Since
Hsp90α lacks signal peptide that required for conventional protein secretion, the most
recognized mechanism account for eHsp90α is via exosome secretion. Using modified
sequential centrifugation techniques [129], I was able to collect three fractions from cellconditioned medium including microvesicles, exosomes, and EV-depleted supernatant.
Surprisingly, after screening the protein level of eHsp90α in each fraction, my preliminary
data showed that a majority (~95%) of eHsp90α is not associated with exosomes, and
rather present in EV-depleted cell-conditioned medium, indicating the majority of
eHsp90α is regulated by unknown pathway rather than exosome secretion. Three
questions were raised from this preliminary data: 1) What the actual concentration of
eHsp90α in this large pool (~95%)? 2) Does this large pool of eHsp90α reaches the
functional concentration of 100-300µg/mL that drives wound healing in vivo? 3) If it
reaches functional concentration, what is the mechanism of this unknown pathway that
regulates the majority of Hsp90α secretion? Four specific aims were designed as followed
to answers these questions:
1. To prove eHsp90α is triggered by stress but not dead cell leakage by using western
immunoblotting analysis.
2. To investigate the distribution of eHsp90α collected from cancer cell-conditioned
medium by using sequential centrifugations.
91
3. To quantitate the amounts of eHsp90α in different fractions isolated by sequential
centrifugations.
4. To analyze protein secretion pathways that most likely contribute to the functional
concentration of Hsp90α secretion.
In Chapter 3.1, considering lack of direct evidence to rule out the possibility that
eHsp90α is due to dead cell leakage during cell culture, I designed experiments to 1)
prove Hsp90α secretion is result of external stress on normal cells and internal stress in
tumor cells, and 2) compare the secreted proteins and intracellular proteins in cell
conditioned medium. Three different environmental stresses including serum starvation,
hypoxia, and injured tissue released cytokine-TGFα simulated secretion of Hsp90α in
293T, HDF, and HKC cells (untreated group did not). In addition to normal cells, I treated
two cancer cell lines including HeLa cells and MDA-MB-231 cells with either normoxia or
hypoxia. Results showed that oncogenes are an internal stress that stimulates cancer
cells to constitutively secreted Hsp90α even under physiological conditions. Moreover, I
designed a time-dependent experiment to serum starve MDA-MB-231 cells from one day
to five days. The cell-conditioned medium in each time point were collected and the
comparison of intracellular proteins and extracellular protein were done. The results
showed that while increased eHsp90α in the conditioned medium was detected in a timedependent manner, none of the intracellular proteins could be detected in the same
conditioned medium. Therefore, the eHsp90α detected was not a result of intracellular
Hsp90α leakage from dead cells. Taken together, the Chapter 3.1 illustrated that 1) the
92
Hsp90α is actively secreted outside the cell by either external environmental stresses or
internal oncogenes stress and 2) eHsp90α is not because of dead cell leakage during cell
culture.
In Chapter 3.2, although exosome secretion has been identified as a mechanism
supporting the secretion of Hsp90α, however, studies neither quantitated the amount of
exosome-associated eHsp90α nor investigated other possible transportation routes for
eHsp90α. Therefore, I collected three fractions including microvesicles, exosome, and
EV-depleted supernatant by using modified sequential centrifugations to analyze the
distribution and quantitation of eHsp90α in cell-conditioned medium. Surprisingly,
analysis of the extracellular fractions collected from CM revealed that majority of eHsp90α
(~95%) is not associated with either exosome (~3%) or microvesicles (2%). Rather, 95%
of eHsp90α was observed in the EV-depleted cell-conditioned medium. Given the
distribution of eHsp90α, I sought to further quantitate the concentration of eHsp90α in
each fraction. Using standard curve from known concentration of rhHsp90α as reference,
I calculated that for 5 million MDA-MB-231 breast cancer cells under starvation for 16
hours secrete approximately 0.85 µg of eHsp90α. This number could be an in vitro
reference to further calculate the actual tissue concentration of eHsp90α in vivo. Moreover,
trypsin digestion analysis also indicated that the Sup-associated eHsp90α (~95%)
appears to be a “naked” protein and neither within nor on the surface of vesicles. Taken
together, Chapter 3.2 illustrated that 1) exosome is a minor secretion mechanism account
for eHsp90α. The majority of Hsp90α secretion (~95%) is regulated by unknown pathway;
93
2) For In vitro study, 5 million MDA-MB-231 breast cancer cells under starvation for 16
hours secrete approximately 0.85 µg of eHsp90α.
In Chapter 3.3, since no study has taken account of the dramatic difference between
cell number and liquid volume surrounding the cells in vitro and the same physiological
situation in vivo, I established “Cell number to Interstitial-Fluid (CIF)” ratio to predict the
tissue concentration of eHsp90α in vivo by using in vitro concentration of eHsp90α.
According to several known human parameters, the CIF ratio was calculated to be 1.1 ×
109 cells per 1 mL of interstitial fluid for humans. Therefore, based on the in vitro data of
every 5 million MDA-MB-231 cells secrete 0.85 µg of eHsp90α, the actual tissue
concentration of eHsp90α was calculated to be 187 µg/mL. Furthermore, the exosomeassociated Hsp90α and Sup-associated Hsp90α can also be calculated based on the
distribution of eHsp90α in cell-conditioned medium, which to be 9.35 µg/mL and 178
µg/mL, respectively. Therefore, exosome secretion only provides 9.35 µg/mL of eHsp90α,
which is far below the 100-300 µg/mL of eHsp90α required for promoting wound healing
in vivo [36, 62]. In contrast, the unknown secretory pathway provides 178 µg/mL of
eHsp90α, which is enough to promote wound healing in vivo. Considering exosome is a
well-known carrier for extracellular protein transportation, there is still a possibility that
Exo-associated Hsp90α is much potent than Sup-associated Hsp90α, or exosomes are
easier for cell to take in or to activate the cell surface receptor of Hsp90α such as LRP-1.
However, it is hard to compare the potency of Exo-associated Hsp90α with the Supassociated Hsp90α since Hsp90α is not the only one cargo inside exosome. Hundreds of
94
other proteins and small RNAs also could be variables contributing to the potency of
exosome in driving wound healing and tumorigenesis [82]. Therefore, if I only focus on
molecular concentration rather than delivery potency, I concluded that it is the unknown
pathway but not extracellular vesicles which provide the functional concentration of
eHsp90α to drive wound healing. Taken together, the Chapter 3.3 illustrated the
establishment of CIF ratio to predict the tissue concentration of a given molecule bases
on its concentration in vitro. Furthermore, based on the tissue concentration of Exosomeassociated Hsp90α (9.35 µg/mL) and Sup-associated Hsp90α (178 µg/mL), I concluded
that it is the unknown pathway rather than exosomes who contributes to the functional
concentration of Hsp90α secretion to promote wound healing.
In Chapter 3.4, to investigate which protein secretion pathway contributes to the
functional concentration of eHsp90α for wound healing, I started from analyzing
conventional protein secretion pathway (CPS) and unconventional protein secretion
pathways (UPS). Conventional protein secretion pathway could not be the mechanism
accounts for Hsp90α secretion because it requires the presence of N-terminal signal
peptide in the to be secreted protein, and Hsp90α does not have. The UPS can be
categorized into four types including: 1) type I, plasma membrane pore formation
mediated protein secretion; 2) type II, ABC-transported based protein secretion; 3) type
III, organelle-based protein translocation and 4) type IV, Golgi-bypass [71]. The type I
UPS requires secreted proteins to be able to self-oligomerized [72]. The type II UPS can
only pump out small molecules such as ions and small peptides [75]. The type IV UPS
95
requires signal peptide as well [85]. Therefore, only type III UPS is likely to account for
Hsp90α secretion. The organelle-based type III UPS is proposed to be exosomes and
autophagosomes based protein secretion [79]. My pervious results showed that exosome
only provide minor part (less than 5%) of Hsp90α secretion. So secretory autophagosome
likely to be the mechanism account for the majority of Hsp90α secretion. Considering
Golgi reassembly stacking protein 55 (GRASP55) has been proved to regulate secretory
autophagy and protein secretion [105], I examined the supernatant associated Hsp90α in
HeLa GRASP55KO, MDA-MB-231 GRASP55KD, and IKC GRASP55KD cells. Results
showed that knockout or knockdown of GRASP55 partially reduced Hsp90α secretion in
both normal and cancer cell lines. Moreover, I carried out rescue experiment in HeLa
GRASP55KO cells to re-introduce GFP-GRASP55 and further verify that GRASP55
mediates Hsp90α secretion. Taken together, the Chapter 3.4 illustrated that the secretory
autophagy-mediated factor GRASP55 is partially required for Hsp90α secretion in both
tumor and normal cell lines. Therefore, secretory autophagy is likely the mechanism which
accounts for the majority of Hsp90α secretion. However, only around 50% of eHsp90α
was reduced in response to GRASP55 knockout or knockdown. Further mechanistic
investigation is required to provide stronger evidence that secretory autophagy is in fact
the primary pathway accounts for Hsp90α secretion.
In Chapter 3.5, to further validate the role of secretory autophagy in mediating
Hsp90α secretion, I decided to directly target genes that involved in autophagosome
biogenesis including Atg5, Atg7, Atg16, Beclin-1, and LC3. The biogenesis of
96
autophagosome is consist of three steps including 1) initiation, 2) elongation, and 3)
maturation [106]. Each step of autophagosome biogenesis requires the regulation of
different complex and conjugation systems. Vps34-Atg6/beclin-1 class III PI3-kinase
complex regulates the initiation of autophagosome formation. Atg5-Atg12-Atg16L1
conjugation system regulates the elongation of autophagosome. Atg8/LC3 conjugation
system regulates the elongation and maturation of autophagosome [112-117]. Each
complex or conjugation system is indispensable during the biogenesis of autophagosome.
Therefore, I established MDA-MB-231 Atg5KD, Atg7KD, Atg16KD, Beclin-1KD, and
LC3KD cell line and examined the level of eHsp90α in EV-depleted supernatant. Results
showed that except downregulation of Atg7 slightly enhanced Hsp90α secretion, all the
other gene downregulations including Atg5, Atg16, Beclin-1 and LC3 partially reduced
Hsp90α secretion. In Chapter 3.4, I showed that GRASP55KD partially reduced Supassociated Hsp90α secretion. The similar results of partially deceased of Sup-associated
Hsp90α in both GRASP55 knockdown and autophagosome biogenesis genes
knockdown inspired me an idea that maybe GRASP55 and autophagosome biogenesis
genes control two parallel secretory autophagy pathway rather than one. To verify this
idea, I established MDA-MB-231 GRASP55/Atg5 double KD cell line. Results showed
that although single knockdown of GRASP55 or Atg5 partially reduced Hsp90α secretion,
double knockdown of GRASP55 and Atg5 blocked majority of Sup-associated Hsp90α
secretion, indicating that GRASP55 and Atg5 may individually mediates two parallelled
secretory autophagy pathways.
97
In Chapter 3.6, although a series of studies has demonstrated topical application of
rhHsp90α promotes wound healing in mice and pigs [23,36,39,40], there has not been
direct evidence that the action of cell secreted Hsp90α is essential for wound healing.
Therefore, to provide the answer, in collaboration with Dr. Xin Tang, we established
Hsp90α Crispr/cas9 knockout (Hsp90α-KO) mice and preformed wound healing
experiments. First, the biopsy sections of wound skin, injured lung, and tumor growing
liver that stained with anti-Hsp90α antibody showed massive deposition of eHsp90α in
injured and tumor growing tissue. Second, the results of wound healing experiment
showed a significantly delayed wound closure in Hsp90α-KO mice in comparison with
wild types. Moreover, the delayed wound closure in Hsp90α-KO mice was completely
corrected by topical application of rhHsp90α. These results provide direct evidence that
eHsp90α is essential during wound healing.
All in all, as shown in Figure 4.1, this thesis aims to investigate the protein secretion
mechanism that regulates the functional concentration of eHsp90α required for wound
healing. First, I demonstrated that eHsp90α is actively secreted outside and it is not
because of dead cell leakage during cell culture. Second, in disagreement with previous
reports that exosomes are the major secretion pathway account for Hsp90α secretion,
my results showed that less than 5% of secreted Hsp90α was associated with exosomes,
while the majority remained in EV-depleted supernatant of conditioned medium. Third,
establishment of “cell number to interstitial fluid (CIF)” ratio for human interstitial fluid
revealed a non-exosome pathway supplies 178 µg/mL of Hsp90, which perfectly
98
matches the reported 100-300 µg/mL of Hsp90 for promoting wound healing in vivo.
Fourth, I demonstrated that the secretory autophagy pathway with autophagosome
biogenesis genes including Beclin-1, Atg5, Atg16 and LC3, together with the Golgi
reassembly-stacking protein gene-55 (GRASP55) regulates exosome-independent
secretion of Hsp90α. Finally, in collaboration with Dr. Xin Tang, we provided direct
evidence that eHsp90α is essential during wound healing.
Figure 4.1 Schematic illustration of the secretory autophagy regulates the functional eHsp90α that drives wound healing.
Secretory autophagy pathway with autophagosome biogenesis regulating genes (AR genes) including Beclin-1, Atg5,
Atg16 and LC3, together with the Golgi reassembly-stacking protein gene-55 (GRASP55) regulates exosomeindependent secretion of Hsp90α to drive wound healing.
4.2 Future directions
The results of single/double knockdown of GRAPS55 and Atg5 in MDA-MB-231 cells
99
has implicated that there are likely two paralleled secretory autophagy pathways
regulating Hsp90α secretion. Nishida et al. has shown that mouse cells lacking Atg5 or
Atg7 can still form autophagosomes and preform autophagy-mediated protein
degradation when subjected to stressors [138]. Unlike conventional macroautophagy that
depends on autophagosome biogenesis genes such as Atg5, Nishda et al. showed that
autophagosomes can be generated in a Rab9-dependent manner by the fusion of
isolation membranes with vesicles derived from the trans-Golgi and late endosomes.
They called Rab9-dependent autophagy as alternative autophagy (Figure 4.2).
Considering GRASP55 is in trans-Golgi as well [107,108], it is not hard to connect
GRASP55 with alternative autophagy. If GRASP55 regulates alternative autophagy, it
perfectly explains the results that single downregulation of either GRASP55 or Atg5 only
partially reduced Hsp90α secretion, while double downregulation of both GRASP55 and
Atg5 nearly blocked Hsp90α secretion. Therefore, as shown in Figure 4.2, my new
hypothesis for the future investigation is that GRASP55 regulates alternative secretory
autophagy, together with conventional secretory autophagy controlling the functional
concentration of eHsp90α that drives wound healing.
100
Figure 4.2 Hypothetical model of conventional macroautophagy and alternative macroautophagy (Nishida et al. 2009,
adapted). Two parallel secretory autophagy pathways may collectively regulate protein secretions. Conventional
macroautophagy depends on Atg5 and Atg7, while alternative macroautophagy depends on Rab9. My hypothesis is
that GRASP55 may also play critical role in regulating alternative macroautophagy, contributing to the formation of CUP
from trans-Golgi [138].
To further investigate this hypothesis, I plan to first establish MDA-MB-231 Beclin1/GRASP55 double KD cell line. As this thesis discussed in Chapter 1.4.3, Beclin-1
regulates the initiation of autophagosome biogenesis and Atg5 regulates the elongation
101
of autophagosome formation. Therefore, same as Atg5, Beclin-1 also plays critical role in
mediating conventional autophagy. Double downregulation of both Beclin-1 and
GRASP55 in MDA-MB-231 cells will help me further validate the hypothesis of GRASP55
mediating alternative autophagy. To prove this result is not cell-type specific, I also plan
to establish HeLa GRASP55KO/Atg5 KD or Beclin-1 KD cell line. EV-depleted Sup will
be collected and subjected to western immunoblotting analysis blot with anti-Hsp90α
antibody to investigate the level of eHsp90α.
In this thesis, another discovery and innovation are the establishment of CIF ratio,
which is 1.1 ×109 cells per 1 mL of interstitial fluid for human tissue. The CIF ratio can be
an in vivo guidance to evaluate possible physiological relevance of secreted proteins
identified in CM of cells in vitro in reference to know the actual tissue concentration of the
proteins. However, the application of the CIF ratio is limited to individual secreted proteins
by a single cell type. The CIF ratio calculation may not apply if the tissue concentration of
a secreted protein is contributed by more than one cell types in the tissue
microenvironment. It is also noticed that the CIF ratio could vary, since additional fluid in
an injured tissue area could occur under certain pathological conditions, such as edema.
In contrast, losing of fluid under other pathological conditions which referred to as
dehydration, could also affect the CIF ratio calculation. Any of these factors may have a
significant effect on the authenticity of the proteomic profiles. Therefore, although this
thesis has proven the accuracy of CIF ratio by pointing out the secreted Hsp90α via the
secretory autophagy rather than exosome pathway that is sufficient for promoting wound
102
healing, more investigations on different secreted proteins and in different cells are
needed to further validation of the accuracy of CIF ratio.
A note: This thesis only covers my latest research in Dr. Li’s laboratory - “the mechanism
of Hsp90 secretion”. Prior to this latest project, I have had eight publications from Dr.
Li’s group, including four either first or co-first author research articles (Tang X*, Chang
C* et al. Scientific report 2019; Tang X*, Chang C* et al. Cancer gene therapy 2021; Tang
X*, Chang C* et al. Molecular and cell biology 2021 and Chang C*, Tang X* et al. Scientific
report 2022); three review articles (Guo J*, Chang C* et al. Expert review of proteomics
2017; Chang C*, Tang X* et al. Cells 2023 and Chen M*, Chang C* et al. International
Journal of Molecular Sciences 2023) and one book chapter (Chang C*, Tang X* et al. In
Chaperones: Methods and Protocols 2023). However, due to the reasons of focus,
coherence, and page limitations, I will not describe the work of these publications in any
details in my thesis.
103
Reference
1. Sahu D, Zhao Z, Tsen F, Cheng CF, Park R, Situ AJ, Dai J, Eginli A, Shams S, Chen
M, Ulmer TS. A potentially common peptide target in secreted heat shock protein-90α
for hypoxia-inducible factor-1α–positive tumors. Molecular biology of the cell. 2012
Feb 15;23(4):602-13.
2. Chang C, Tang X, Woodley DT, Chen M, Li W. The Distinct Assignments for Hsp90α
and Hsp90β: More Than Skin Deep. Cells. 2023 Jan 11;12(2):277.
3. Jackson SE. Hsp90: structure and function. Molecular chaperones. 2013:155-240.
4. Chua CS, Low H, Sim TS. Co-chaperones of Hsp90 in Plasmodium falciparum and
their concerted roles in cellular regulation. Parasitology. 2014 Aug;141(9):1177-91.
5. Borkovich KA, Farrelly FW, Finkelstein DB, Taulien J, Lindquist S. hsp82 is an
essential protein that is required in higher concentrations for growth of cells at higher
temperatures. Molecular and cellular biology. 1989.
6. Girstmair H, Tippel F, Lopez A, Tych K, Stein F, Haberkant P, Schmid PW, Helm D,
Rief M, Sattler M, Buchner J. The Hsp90 isoforms from S. cerevisiae differ in structure,
function and client range. Nature communications. 2019 Aug 9;10(1):3626.
7. Kuo CC, Liang CM, Lai CY, Liang SM. Involvement of heat shock protein (Hsp) 90β
but not Hsp90α in antiapoptotic effect of CpG-B oligodeoxynucleotide. The Journal of
Immunology. 2007 May 15;178(10):6100-8.
8. Bouchier-Hayes L, Oberst A, McStay GP, Connell S, Tait SW, Dillon CP, Flanagan JM,
Beere HM, Green DR. Characterization of cytoplasmic caspase-2 activation by
induced proximity. Molecular cell. 2009 Sep 24;35(6):830-40.
9. Chatterjee M, Andrulis M, Stühmer T, Müller E, Hofmann C, Steinbrunn T, Heimberger
T, Schraud H, Kressmann S, Einsele H, Bargou RC. The PI3K/Akt signaling pathway
regulates the expression of Hsp70, which critically contributes to Hsp90-chaperone
function and tumor cell survival in multiple myeloma. Haematologica. 2013
Jul;98(7):1132.
10.Taherian A, Krone PH, Ovsenek N. A comparison of Hsp90 α and Hsp90 β interactions
with cochaperones and substrates. Biochemistry and cell biology. 2008 Feb;86(1):37-
45.
11.Voss AK, Thomas T, Gruss P. Mice lacking HSP90β fail to develop a placental labyrinth.
Development. 2000 Jan 1;127(1):1-1.
12.Grad I, Cederroth CR, Walicki J, Grey C, Barluenga S, Winssinger N, De Massy B,
Nef S, Picard D. The molecular chaperone Hsp90α is required for meiotic progression
104
of spermatocytes beyond pachytene in the mouse. PloS one. 2010 Dec
31;5(12):e15770.
13.Tang X, Chang C, Hao M, Chen M, Woodley DT, Schönthal AH, Li W. Heat shock
protein-90alpha (Hsp90α) stabilizes hypoxia-inducible factor-1α (HIF-1α) in support of
spermatogenesis and tumorigenesis. Cancer gene therapy. 2021 Sep;28(9):1058-70.
14.Stone KR, Smith RE, Joklik WK. Changes in membrane polypeptides that occur when
chick embryo fibroblasts and NRK cells are transformed with avian sarcoma viruses.
Virology. 1974 Mar 1;58(1):86-100.
15.Shiu RP, Pouyssegur J, Pastan IR. Glucose depletion accounts for the induction of
two transformation-sensitive membrane proteinsin Rous sarcoma virus-transformed
chick embryo fibroblasts. Proceedings of the National Academy of Sciences. 1977
Sep;74(9):3840-4.
16.Pouysségur J, Shiu RP, Pastan I. Induction of two transformation-sensitive membrane
polypeptides in normal fibroblasts by a block in glycoprotein synthesis or glucose
deprivation. Cell. 1977 Aug 1;11(4):941-7.
17.Pouysségur J, Yamada KM. Isolation and immunological characterization of a
glucose-regulated fibroblast cell surface glycoprotein and its nonglycosylated
precursor. Cell. 1978 Jan 1;13(1):139-50.
18.McCormic PJ, Keys BJ, Pucci C, Millis AJ. Human fibroblast-conditioned medium
contains a 100K dalton glucose-regulated cell surface protein. Cell. 1979 Sep
1;18(1):173-82.
19.Hughes EN, Colombatti A, August JT. Murine cell surface glycoproteins. Purification
of the polymorphic Pgp-1 antigen and analysis of its expression on macrophages and
other myeloid cells. Journal of Biological Chemistry. 1983 Jan 25;258(2):1014-21.
20.Ullrich SJ, Robinson EA, Law LW, Willingham M, Appella E. A mouse tumor-specific
transplantation antigen is a heat shock-related protein. Proceedings of the National
Academy of Sciences. 1986 May;83(10):3121-5.
21.La Thangue NB, Latchman DS. A cellular protein related to heat-shock protein 90
accumulates during herpes simplex virus infection and is overexpressed in
transformed cells. Experimental cell research. 1988 Sep 1;178(1):169-79.
22.Eustace BK, Sakurai T, Stewart JK, Yimlamai D, Unger C, Zehetmeier C, Lain B,
Torella C, Henning SW, Beste G, Scroggins BT. Functional proteomic screens reveal
an essential extracellular role for hsp90α in cancer cell invasiveness. Nature cell
biology. 2004 Jun 1;6(6):507-14.
23.Li W, Li Y, Guan S, Fan J, Cheng CF, Bright AM, Chinn C, Chen M, Woodley DT.
105
Extracellular heat shock protein‐90α: linking hypoxia to skin cell motility and wound
healing. The EMBO journal. 2007 Mar 7;26(5):1221-33.
24.Cheng CF, Fan J, Fedesco M, Guan S, Li Y, Bandyopadhyay B, Bright AM, Yerushalmi
D, Liang M, Chen M, Han YP. Transforming growth factor α (TGFα)-stimulated
secretion of HSP90α: using the receptor LRP-1/CD91 to promote human skin cell
migration against a TGFβ-rich environment during wound healing. Molecular and
cellular biology. 2008 May 1;28(10):3344-58.
25.Gopal U, Bohonowych JE, Lema-Tome C, Liu A, Garrett-Mayer E, Wang B, Isaacs JS.
A novel extracellular Hsp90 mediated co-receptor function for LRP1 regulates EphA2
dependent glioblastoma cell invasion. PloS one. 2011 Mar 8;6(3):e17649.
26.Tsutsumi S, Scroggins B, Koga F, Lee MJ, Trepel J, Felts S, Carreras C, Neckers L. A
small molecule cell-impermeant Hsp90 antagonist inhibits tumor cell motility and
invasion. Oncogene. 2008 Apr;27(17):2478-87.
27.Sims JD, McCready J, Jay DG. Extracellular heat shock protein (Hsp) 70 and Hsp90α
assist in matrix metalloproteinase-2 activation and breast cancer cell migration and
invasion. PloS one. 2011 Apr 14;6(4):e18848.
28.Baker-Williams AJ, Hashmi F, Budzyński MA, Woodford MR, Gleicher S, Himanen SV,
Makedon AM, Friedman D, Cortes S, Namek S, Stetler-Stevenson WG. Cochaperones TIMP2 and AHA1 competitively regulate extracellular HSP90: client
MMP2 activity and matrix proteolysis. Cell reports. 2019 Aug 13;28(7):1894-906.
29.Li W, Sahu D, Tsen F. Secreted heat shock protein-90 (Hsp90) in wound healing and
cancer. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research. 2012 Mar
1;1823(3):730-41.
30.Cheng CF, Sahu D, Tsen F, Zhao Z, Fan J, Kim R, Wang X, O’Brien K, Li Y, Kuang Y,
Chen M. A fragment of secreted Hsp90α carries properties that enable it to accelerate
effectively both acute and diabetic wound healing in mice. The Journal of clinical
investigation. 2011 Oct 24;121(11).
31.Tsen F, Bhatia A, O'Brien K, Cheng CF, Chen M, Hay N, Stiles B, Woodley DT, Li W.
Extracellular heat shock protein 90 signals through subdomain II and the NPVY motif
of LRP-1 receptor to Akt1 and Akt2: a circuit essential for promoting skin cell migration
in vitro and wound healing in vivo. Molecular and cellular biology. 2013 Dec
1;33(24):4947-59.
32.Tian Y, Wang C, Chen S, Liu J, Fu Y, Luo Y. Extracellular Hsp90α and clusterin
synergistically promote breast cancer epithelial-to-mesenchymal transition and
metastasis via LRP1. Journal of cell science. 2019 Aug 1;132(15):jcs228213
106
33.Jay D, Luo Y, Li W. Extracellular heat shock protein-90 (eHsp90): everything you need
to know. Biomolecules. 2022 Jun 29;12(7):911.
34.Yu X, Harris SL, Levine AJ. The regulation of exosome secretion: a novel function of
the p53 protein. Cancer research. 2006 May 1;66(9):4795-801.
35.Suzuki S, Kulkarni AB. Extracellular heat shock protein HSP90β secreted by MG63
osteosarcoma cells inhibits activation of latent TGF-β1. Biochemical and biophysical
research communications. 2010 Jul 30;398(3):525-31.
36.Jayaprakash P, Dong H, Zou M, Bhatia A, O'Brien K, Chen M, Woodley DT, Li W.
Hsp90α and Hsp90β together operate a hypoxia and nutrient paucity stress‐response
mechanism during wound healing. Journal of cell science. 2015 Apr 15;128(8):1475-
80.
37.Zou M, Bhatia A, Dong H, Jayaprakash P, Guo J, Sahu D, Hou Y, Tsen F, Tong C,
O'Brien K, Situ AJ. Evolutionarily conserved dual lysine motif determines the nonchaperone function of secreted Hsp90alpha in tumour progression. Oncogene. 2017
Apr;36(15):2160-71.
38.Dong H, Zou M, Bhatia A, Jayaprakash P, Hofman F, Ying Q, Chen M, Woodley DT, Li
W. Breast cancer MDA-MB-231 cells use secreted heat shock protein-90alpha
(Hsp90α) to survive a hostile hypoxic environment. Scientific reports. 2016 Feb
5;6(1):20605.
39.Bhatia A, O’Brien K, Guo J, Lincoln V, Kajiwara C, Chen M, Woodley DT, Udono H, Li
W. Extracellular and non-chaperone function of heat shock protein− 90α is required
for skin wound healing. Journal of Investigative Dermatology. 2018 Feb 1;138(2):423-
33.
40.O'Brien K, Bhatia A, Tsen F, Chen M, Wong AK, Woodley DT, Li W. Identification of the
critical therapeutic entity in secreted Hsp90α that promotes wound healing in newly
re-standardized healthy and diabetic pig models. PLoS One. 2014 Dec
2;9(12):e113956.
41.Bhatia A, O'brien K, Chen M, Wong A, Garner W, Woodley DT, Li W. Dual therapeutic
functions of F-5 fragment in burn wounds: preventing wound progression and
promoting wound healing in pigs. Molecular Therapy-Methods & Clinical Development.
2016 Jan 1;3.
42.Wang X, Song X, Zhuo W, Fu Y, Shi H, Liang Y, Tong M, Chang G, Luo Y. The
regulatory mechanism of Hsp90α secretion and its function in tumor malignancy.
Proceedings of the National Academy of Sciences. 2009 Dec 15;106(50):21288-93.
43.Fu Y, Xu X, Huang D, Cui D, Liu L, Liu J, He Z, Liu J, Zheng S, Luo Y. Plasma heat
107
shock protein 90alpha as a biomarker for the diagnosis of liver cancer: an official,
large-scale, and multicenter clinical trial. EBioMedicine. 2017 Oct 1;24:56-63.
44.Shi Y, Liu X, Lou J, Han X, Zhang L, Wang Q, Li B, Dong M, Zhang Y. Plasma levels
of heat shock protein 90 alpha associated with lung cancer development and
treatment responses. Clinical cancer research. 2014 Dec 1;20(23):6016-22.
45.Kasanga M, Liu L, Xue L, Song X. Plasma heat shock protein 90-alpha have an
advantage in diagnosis of colorectal cancer at early stage. Biomarkers in medicine.
2018 Aug;12(8):881-90.
46.Tas F, Bilgin E, Erturk K, Duranyildiz D. Clinical significance of circulating serum
cellular heat shock protein 90 (HSP90) level in patients with cutaneous malignant
melanoma. Asian Pacific Journal of Cancer Prevention: APJCP. 2017;18(3):599.
47.Fredly H, Reikvam H, Gjertsen BT, Bruserud Ø. Disease‐stabilizing treatment with all‐
trans retinoic acid and valproic acid in acute myeloid leukemia: serum hsp70 and
hsp90 levels and serum cytokine profiles are determined by the disease, patient age,
and anti‐leukemic treatment. American journal of hematology. 2012 Apr;87(4):368-76.
48.Liu W, Li J, Zhang P, Hou Q, Feng S, Liu L, Cui D, Shi H, Fu Y, Luo Y. A novel pan‐
cancer biomarker plasma heat shock protein 90alpha and its diagnosis determinants
in clinic. Cancer science. 2019 Sep;110(9):2941-59.
49.Wei W, Liu M, Ning S, Wei J, Zhong J, Li J, Cai Z, Zhang L. Diagnostic value of plasma
HSP90α levels for detection of hepatocellular carcinoma. Bmc Cancer. 2020
Dec;20:1-9.
50.Chen CC, Chen LL, Li CP, Hsu YT, Jiang SS, Fan CS, Chua KV, Huang SX, Shyr YM,
Chen LT, Huang TS. Myeloid-derived macrophages and secreted HSP90α induce
pancreatic ductal adenocarcinoma development. Oncoimmunology. 2018 May
4;7(5):e1424612.
51.Zhou Y, Deng X, Zang N, Li H, Li G, Li C, He M. Transcriptomic and proteomic
investigation of HSP90A as a potential biomarker for HCC. Medical science monitor:
international medical journal of experimental and clinical research. 2015;21:4039.
52.Wang X, An D, Wang X, Liu X, Li B. Extracellular Hsp90α clinically correlates with
tumor malignancy and promotes migration and invasion in esophageal squamous cell
carcinoma. OncoTargets and therapy. 2019;12:1119.
53.Damasiewicz-Bodzek A, Szumska M, Tyrpień-Golder K. Antibodies to heat shock
proteins 90α and 90β in psoriasis. Archivum Immunologiae et Therapiae
108
Experimentalis. 2020 Apr;68(2):9.
54.Han S, Cheng Z, Zhao X, Huang Y. Diagnostic value of heat shock protein 90 α and
squamous cell carcinoma antigen in detection of cervical cancer. Journal of
International Medical Research. 2019 Nov;47(11):5518-25.
55.Burgess EF, Ham AJ, Tabb DL, Billheimer D, Roth BJ, Chang SS, Cookson MS, Hinton
TJ, Cheek KL, Hill S, Pietenpol JA. Prostate cancer serum biomarker discovery
through proteomic analysis of alpha‐2 macroglobulin protein complexes.
PROTEOMICS–Clinical Applications. 2008 Sep;2(9):1223-33.
56.Liang XQ, Li KZ, Li Z, Xie MZ, Tang YP, Du JB, Huang Y, Li JL, Hu BL. Diagnostic and
prognostic value of plasma heat shock protein 90alpha in gastric cancer. International
Immunopharmacology. 2021 Jan 1;90:107145.
57.Yuan Z, Wang L, Hong S, Shi C, Yuan B. Diagnostic value of HSP90α and related
markers in lung cancer. Journal of Clinical Laboratory Analysis. 2022
Jun;36(6):e24462.
58.Burgess EF, Ham AJ, Tabb DL, Billheimer D, Roth BJ, Chang SS, Cookson MS, Hinton
TJ, Cheek KL, Hill S, Pietenpol JA. Prostate cancer serum biomarker discovery
through proteomic analysis of alpha‐2 macroglobulin protein complexes.
PROTEOMICS–Clinical Applications. 2008 Sep;2(9):1223-33.
59.Liao DF, Jin ZG, Baas AS, Daum G, Gygi SP, Aebersold R, Berk BC. Purification and
identification of secreted oxidative stress-induced factors from vascular smooth
muscle cells. Journal of Biological Chemistry. 2000 Jan 7;275(1):189-96.
60.Kuroita T, Tachibana H, Ohashi H, Shirahata S, Murakami H. Growth stimulating
activity of heat shock protein 90α to lymphoid cell lines in serum-free medium.
Cytotechnology. 1992 Jun;8:109-17.
61.Tang X, Chang C, Guo J, Lincoln V, Liang C, Chen M, Woodley DT, Li W. Tumoursecreted Hsp90α on external surface of exosomes mediates tumour-stromal cell
communication via autocrine and paracrine mechanisms. Scientific Reports. 2019 Oct
22;9(1):15108.
62.Wang X, Song X, Zhuo W, Fu Y, Shi H, Liang Y, Tong M, Chang G, Luo Y. The
regulatory mechanism of Hsp90α secretion and its function in tumor malignancy.
Proceedings of the National Academy of Sciences. 2009 Dec 15;106(50):21288-93.
63.Tsutsumi S, Mollapour M, Graf C, Lee CT, Scroggins BT, Xu W, Haslerova L, Hessling
M, Konstantinova AA, Trepel JB, Panaretou B. Hsp90 charged-linker truncation
reverses the functional consequences of weakened hydrophobic contacts in the N
109
domain. Nature structural & molecular biology. 2009 Nov;16(11):1141-7.
64.Zhang S, Wang C, Ma B, Xu M, Xu S, Liu J, Tian Y, Fu Y, Luo Y. Mutant p53 drives
cancer metastasis via RCP-mediated Hsp90α secretion. Cell Reports. 2020 Jul
7;32(1).
65.Savina A, Furlán M, Vidal M, Colombo MI. Exosome release is regulated by a calciumdependent mechanism in K562 cells. Journal of Biological Chemistry. 2003 May
30;278(22):20083-90.
66.Hegmans JP, Bard MP, Hemmes A, Luider TM, Kleijmeer MJ, Prins JB, Zitvogel L,
Burgers SA, Hoogsteden HC, Lambrecht BN. Proteomic analysis of exosomes
secreted by human mesothelioma cells. The American journal of pathology. 2004 May
1;164(5):1807-15.
67.Guo J, Jayaprakash P, Dan J, Wise P, Jang GB, Liang C, Chen M, Woodley DT, Fabbri
M, Li W. PRAS40 connects microenvironmental stress signaling to exosome-mediated
secretion. Molecular and cellular biology. 2017 Oct 1;37(19):e00171-17.
68.Green ER, Mecsas J. Bacterial secretion systems: an overview. Virulence
mechanisms of bacterial pathogens. 2016 Jun 22:213-39.
69.Viotti C. ER to Golgi-dependent protein secretion: the conventional pathway.
Unconventional protein secretion: methods and protocols. 2016:3-29.
70.Rabouille C, Malhotra V, Nickel W. Diversity in unconventional protein secretion.
Journal of cell science. 2012 Nov 15;125(22):5251-5.
71.Maricchiolo E, Panfili E, Pompa A, De Marchis F, Bellucci M, Pallotta MT.
Unconventional pathways of protein secretion: mammals vs. plants. Frontiers in Cell
and Developmental Biology. 2022 Apr 28;10:895853.
72.Steringer JP, Bleicken S, Andreas H, Zacherl S, Laussmann M, Temmerman K,
Contreras FX, Bharat TA, Lechner J, Müller HM, Briggs JA. Phosphatidylinositol 4, 5-
bisphosphate (PI (4, 5) P2)-dependent oligomerization of fibroblast growth factor 2
(FGF2) triggers the formation of a lipidic membrane pore implicated in unconventional
secretion. Journal of Biological Chemistry. 2012 Aug 10;287(33):27659-69.
73.Zeitler M, Steringer JP, Müller HM, Mayer MP, Nickel W. HIV-Tat protein forms
phosphoinositide-dependent membrane pores implicated in unconventional protein
secretion. Journal of Biological Chemistry. 2015 Sep 4;290(36):21976-84.
74.Martín-Sánchez F, Diamond C, Zeitler M, Gomez AI, Baroja-Mazo A, Bagnall J, Spiller
D, White M, Daniels MJ, Mortellaro A, Peñalver M. Inflammasome-dependent IL-1β
release depends upon membrane permeabilisation. Cell Death & Differentiation. 2016
Jul;23(7):1219-31.
110
75.Jones PM, George AM. The ABC transporter structure and mechanism: perspectives
on recent research. Cellular and Molecular Life Sciences CMLS. 2004 Mar;61:682-99.
76.McGrath JP, Varshavsky A. The yeast STE6 gene encodes a homologue of the
mammalian multidrug resistance P-glycoprotein. Nature. 1989 Aug 3;340(6232):400-
4.
77.Stegmayer C, Kehlenbach A, Tournaviti S, Wegehingel S, Zehe C, Denny P, Smith DF,
Schwappach B, Nickel W. Direct transport across the plasma membrane of
mammalian cells of Leishmania HASPB as revealed by a CHO export mutant. Journal
of cell science. 2005 Feb 1;118(3):517-27.
78.Möskes C, Burghaus PA, Wernli B, Sauder U, Dürrenberger M, Kappes B. Export of
Plasmodium falciparum calcium‐dependent protein kinase 1 to the parasitophorous
vacuole is dependent on three N‐terminal membrane anchor motifs. Molecular
microbiology. 2004 Nov;54(3):676-91.
79.Rabouille C. Pathways of unconventional protein secretion. Trends in cell biology.
2017 Mar 1;27(3):230-40.
80.Théry C, Zitvogel L, Amigorena S. Exosomes: composition, biogenesis and function.
Nature reviews immunology. 2002 Aug 1;2(8):569-79.
81.Mashouri L, Yousefi H, Aref AR, Ahadi AM, Molaei F, Alahari SK. Exosomes:
composition, biogenesis, and mechanisms in cancer metastasis and drug resistance.
Molecular cancer. 2019 Dec;18:1-4.
82.Simpson RJ, Jensen SS, Lim JW. Proteomic profiling of exosomes: current
perspectives. Proteomics. 2008 Oct;8(19):4083-99.
83.Duran JM, Anjard C, Stefan C, Loomis WF, Malhotra V. Unconventional secretion of
Acb1 is mediated by autophagosomes. Journal of cell biology. 2010 Feb
22;188(4):527-36.
84.Zhang M, Kenny SJ, Ge L, Xu K, Schekman R. Translocation of interleukin-1β into a
vesicle intermediate in autophagy-mediated secretion. elife. 2015 Nov 2;4:e11205.
85.Jung J, Kim J, Roh SH, Jun I, Sampson RD, Gee HY, Choi JY, Lee MG. The HSP70
co-chaperone DNAJC14 targets misfolded pendrin for unconventional protein
secretion. Nature communications. 2016 Apr 25;7(1):11386.
86.Yoo JS, Moyer BD, Bannykh S, Yoo HM, Riordan JR, Balch WE. Non-conventional
trafficking of the cystic fibrosis transmembrane conductance regulator through the
early secretory pathway. Journal of Biological Chemistry. 2002 Mar 29;277(13):11401-
9.
111
87.Mizushima N. Autophagy: process and function. Genes & development. 2007 Nov
15;21(22):2861-73.
88.Wang CW, Klionsky DJ. The molecular mechanism of autophagy. Molecular medicine.
2003 Mar;9:65-76.
89.Noda NN, Inagaki F. Mechanisms of autophagy. Annual review of biophysics. 2015
Jun 22;44:101-22.
90.Dupont N, Jiang S, Pilli M, Ornatowski W, Bhattacharya D, Deretic V. Autophagy‐
based unconventional secretory pathway for extracellular delivery of IL‐1β. The
EMBO journal. 2011 Nov 30;30(23):4701-11.
91.Liu JY, Lin YH, Leidal AM, Huang HH, Ye J, Wiita AP, Debnath J. GRASP55 restricts
early-stage autophagy and regulates spatial organization of the early secretory
network. Biology Open. 2021 Oct 15;10(10):bio058736.
92.Zhang X, Wang Y. The Golgi stacking protein GORASP2/GRASP55 serves as an
energy sensor to promote autophagosome maturation under glucose starvation.
Autophagy. 2018 Sep 2;14(9):1649-51.
93.Ponpuak M, Mandell MA, Kimura T, Chauhan S, Cleyrat C, Deretic V. Secretory
autophagy. Current opinion in cell biology. 2015 Aug 1;35:106-16.
94.Huang S, Wang Y. Golgi structure formation, function, and post-translational
modifications in mammalian cells. F1000Research. 2017;6.
95.Barr FA, Puype M, Vandekerckhove J, Warren G. GRASP65, a protein involved in the
stacking of Golgi cisternae. Cell. 1997 Oct 17;91(2):253-62.
96.Shorter J, Watson R, Giannakou ME, Clarke M, Warren G, Barr FA. GRASP55, a
second mammalian GRASP protein involved in the stacking of Golgi cisternae in a
cell-free system. The EMBO journal. 1999 Sep 15;18(18):4949-60.
97.Bekier ME, Wang L, Li J, Huang H, Tang D, Zhang X, Wang Y. Knockout of the Golgi
stacking proteins GRASP55 and GRASP65 impairs Golgi structure and function.
Molecular biology of the cell. 2017 Oct 15;28(21):2833-42.
98.Xiang Y, Wang Y. GRASP55 and GRASP65 play complementary and essential roles
in Golgi cisternal stacking. Journal of Cell Biology. 2010 Jan 25;188(2):237-51.
99.Ahat E, Xiang Y, Zhang X, Bekier ME, Wang Y. GRASP depletion–mediated Golgi
destruction decreases cell adhesion and migration via the reduction of α5β1 integrin.
Molecular biology of the cell. 2019 Mar 15;30(6):766-77.
100. Cheng JP, Betin VM, Weir H, Shelmani GM, Moss DK, Lane JD. Caspase
112
cleavage of the Golgi stacking factor GRASP65 is required for Fas/CD95-mediated
apoptosis. Cell Death & Disease. 2010 Oct;1(10):e82-.
101. Ahat E, Bui S, Zhang J, da Veiga Leprevost F, Sharkey L, Reid W, Nesvizhskii AI,
Paulson HL, Wang Y. GRASP55 regulates the unconventional secretion and
aggregation of mutant huntingtin. Journal of Biological Chemistry. 2022 Aug 1;298(8).
102. Chiritoiu M, Brouwers N, Turacchio G, Pirozzi M, Malhotra V. GRASP55 and UPR
control interleukin-1β aggregation and secretion. Developmental cell. 2019 Apr
8;49(1):145-55.
103. Gee HY, Noh SH, Tang BL, Kim KH, Lee MG. Rescue of ΔF508-CFTR trafficking
via a GRASP-dependent unconventional secretion pathway. Cell. 2011 Sep
2;146(5):746-60.
104. Kim J, Noh SH, Piao H, Kim DH, Kim K, Cha JS, Chung WY, Cho HS, Kim JY,
Lee MG. Monomerization and ER relocalization of GRASP is a requisite for
unconventional secretion of CFTR. Traffic. 2016 Jul;17(7):733-53.
105. Zhang X, Wang Y. Nonredundant Roles of GRASP55 and GRASP65 in the Golgi
Apparatus and Beyond. Trends in biochemical sciences. 2020 Dec 1;45(12):1065-79.
106. Tanida I. Autophagosome formation and molecular mechanism of autophagy.
Antioxidants & redox signaling. 2011 Jun 1;14(11):2201-14.
107. Hosokawa N, Hara T, Kaizuka T, Kishi C, Takamura A, Miura Y, Iemura SI,
Natsume T, Takehana K, Yamada N, Guan JL. Nutrient-dependent mTORC1
association with the ULK1–Atg13–FIP200 complex required for autophagy. Molecular
biology of the cell. 2009 Apr 1;20(7):1981-91.
108. Hosokawa N, Sasaki T, Iemura SI, Natsume T, Hara T, Mizushima N. Atg101, a
novel mammalian autophagy protein interacting with Atg13. Autophagy. 2009 Oct
1;5(7):973-9.
109. Chan EY, Longatti A, McKnight NC, Tooze SA. Kinase-inactivated ULK proteins
inhibit autophagy via their conserved C-terminal domains using an Atg13-independent
mechanism. Molecular and cellular biology. 2009 Jan 1;29(1):157-71.
110. Mercer CA, Kaliappan A, Dennis PB. A novel, human Atg13 binding protein,
Atg101, interacts with ULK1 and is essential for macroautophagy. Autophagy. 2009
Jul 1;5(5):649-62.
111. Judith D, Jefferies HB, Boeing S, Frith D, Snijders AP, Tooze SA. ATG9A shapes
the forming autophagosome through Arfaptin 2 and phosphatidylinositol 4-kinase IIIβ.
Journal of Cell Biology. 2019 May 6;218(5):1634-52.
112. Xie Z, Klionsky DJ. Autophagosome formation: core machinery and adaptations.
113
Nature cell biology. 2007 Oct;9(10):1102-9.
113. Mizushima N, Noda T, Ohsumi Y. Apg16p is required for the function of the
Apg12p–Apg5p conjugate in the yeast autophagy pathway. The EMBO journal. 1999
Jul 15;18(14):3888-96.
114. Mizushima N, Noda T, Yoshimori T, Tanaka Y, Ishii T, George MD, Klionsky DJ,
Ohsumi M, Ohsumi Y. A protein conjugation system essential for autophagy. Nature.
1998 Sep 24;395(6700):395-8.
115. Ichimura Y, Kirisako T, Takao T, Satomi Y, Shimonishi Y, Ishihara N, Mizushima N,
Tanida I, Kominami E, Ohsumi M, Noda T. A ubiquitin-like system mediates protein
lipidation. Nature. 2000 Nov 23;408(6811):488-92.
116. Mizushima N, Yamamoto A, Hatano M, Kobayashi Y, Kabeya Y, Suzuki K,
Tokuhisa T, Ohsumi Y, Yoshimori T. Dissection of autophagosome formation using
Apg5-deficient mouse embryonic stem cells. The Journal of cell biology. 2001 Feb
19;152(4):657-68.
117. Fujita N, Itoh T, Omori H, Fukuda M, Noda T, Yoshimori T. The Atg16L complex
specifies the site of LC3 lipidation for membrane biogenesis in autophagy. Molecular
biology of the cell. 2008 May;19(5):2092-100.
118. Ojha R, Bhattacharyya S, Singh SK. Autophagy in cancer stem cells: a potential
link between chemoresistance, recurrence, and metastasis. BioResearch open
access. 2015 Jan 1;4(1):97-108.
119. El Andaloussi S, Mäger I, Breakefield XO, Wood MJ. Extracellular vesicles:
biology and emerging therapeutic opportunities. Nature reviews Drug discovery. 2013
May;12(5):347-57.
120. Zaborowski MP, Balaj L, Breakefield XO, Lai CP. Extracellular vesicles:
composition, biological relevance, and methods of study. Bioscience. 2015 Aug
1;65(8):783-97.
121. Van der Pol E, Böing AN, Harrison P, Sturk A, Nieuwland R. Classification,
functions, and clinical relevance of extracellular vesicles. Pharmacological reviews.
2012 Jul 1;64(3):676-705.
122. Tkach M, Théry C. Communication by extracellular vesicles: where we are and
where we need to go. Cell. 2016 Mar 10;164(6):1226-32.
123. Akers JC, Gonda D, Kim R, Carter BS, Chen CC. Biogenesis of extracellular
vesicles (EV): exosomes, microvesicles, retrovirus-like vesicles, and apoptotic bodies.
Journal of neuro-oncology. 2013 May;113:1-1.
124. Dang XT, Kavishka JM, Zhang DX, Pirisinu M, Le MT. Extracellular vesicles as an
114
efficient and versatile system for drug delivery. Cells. 2020 Sep 29;9(10):2191.
125. Zhang S, Wang C, Ma B, Xu M, Xu S, Liu J, Tian Y, Fu Y, Luo Y. Mutant p53 drives
cancer metastasis via RCP-mediated Hsp90α secretion. Cell Reports. 2020 Jul
7;32(1).
126. McCready J, Sims JD, Chan D, Jay DG. Secretion of extracellular hsp90α via
exosomes increases cancer cell motility: a role for plasminogen activation. BMC
cancer. 2010 Dec;10:1-0.
127. Eguchi T, Ono K, Kawata K, Okamoto K, Calderwood SK. Regulatory roles of
HSP90-rich extracellular vesicles. Heat shock protein 90 in human diseases and
disorders. 2019:3-17.
128. Lancaster GI, Febbraio MA. Exosome-dependent trafficking of HSP70: a novel
secretory pathway for cellular stress proteins. Journal of Biological Chemistry. 2005
Jun 17;280(24):23349-55
129. Chang C, Tang X, Li W. A Modified Differential Centrifugation Protocol for Isolation
and Quantitation of Extracellular Heat Shock Protein 90 (eHsp90). In Chaperones:
Methods and Protocols 2023 Aug 5 (pp. 251-261). New York, NY: Springer US.
130. Brinkman JE, Dorius B, Sharma S. Physiology, body fluids.
131. Bianconi E, Piovesan A, Facchin F, Beraudi A, Casadei R, Frabetti F, Vitale L,
Pelleri MC, Tassani S, Piva F, Perez-Amodio S. An estimation of the number of cells
in the human body. Annals of human biology. 2013 Nov 1;40(6):463-71.
132. Sender R, Fuchs S, Milo R. Revised estimates for the number of human and
bacteria cells in the body. PLoS biology. 2016 Aug 19;14(8):e1002533.
133. Pain RW. Body fluid compartments. Anaesthesia and intensive care. 1977
Nov;5(4):284-94.
134. Singer AJ, Clark RA. Cutaneous wound healing. New England journal of medicine.
1999 Sep 2;341(10):738-46.
135. Gurtner GC, Werner S, Barrandon Y, Longaker MT. Wound repair and
regeneration. Nature. 2008 May 15;453(7193):314-21.
136. Eming SA, Martin P, Tomic-Canic M. Wound repair and regeneration:
mechanisms, signaling, and translation. Science translational medicine. 2014 Dec
3;6(265):265sr6-.
137. Simon MC, Keith B. The role of oxygen availability in embryonic development and
stem cell function. Nature reviews Molecular cell biology. 2008 Apr;9(4):285-96.
138. Nishida Y, Arakawa S, Fujitani K, Yamaguchi H, Mizuta T, Kanaseki T, Komatsu
115
M, Otsu K, Tsujimoto Y, Shimizu S. Discovery of Atg5/Atg7-independent alternative
macroautophagy. Nature. 2009 Oct 1;461(7264):654-8.
Abstract (if available)
Abstract
The Heat shock protein 90 (Hsp90) family proteins (α and β) are historically recognized as ATP-binding-dependent molecular chaperone that assists the proper folding and support the functionality of client proteins during homeostasis and stress responses. This understanding has served as the foundation for both laboratory research and cancer clinical trials targeting the intracellular function of Hsp90 family proteins. Although several publications reported cell surface bound Hsp90 related molecules as early as the 70s, few in Hsp90 community credited the finding of non-chaperone form of Hsp90 and regraded the observations as a result of dead cell leakage. In the 2000s two independent laboratories discovered secreted Hsp90α in cell conditioned medium of normal and tumor cells. Extracellular Hsp90α (eHsp90α) acts by binding to the cell surface receptor, LRP-1 (low-density lipoprotein receptor-related protein-1), leading to activation of the Akt as well as other intracellular pathways. The two main functions of eHsp90α are 1) to prevent cells from undergoing apoptosis and 2) to stimulate cell motility. Pre-clinical studies demonstrated that topical application of recombinant human Hsp90α (rhHsp90α) strongly promotes closure of trauma, burn, and diabetic wounds in mice and pigs. Similarly, injection with either rhHsp90α or anti-Hsp90α antibody promoted or inhibited tumorigenesis in mice. Clinical studies demonstrated that higher levels of plasma eHsp90α correlated with the later stages of various cancers. Despite above findings, the mechanism of secretion and more importantly quantitation of eHsp90α remained unclear and became my thesis project. During my thesis study, I demonstrated that eHsp90α is actively secreted outside and it is not because of dead cell leakage during cell culture. In disagreement with previous reports that exosomes are the major secretion pathway account for Hsp90α secretion, my results showed that less than 5% of secreted Hsp90α was associated with exosomes, while the majority remained in EV-depleted supernatant of conditioned medium. To quantitate eHsp90α in different factions of cell conditioned medium, I established the so-called “cell number to interstitial fluid” (CIF) ratio for human tissue environment. Calculation by the CIF ratio revealed a non-exosome pathway supplies 178 µg/mL of Hsp90α, which perfectly matches the reported 100-300 µg/mL of Hsp90α for promoting wound healing in vivo. Fourth, I demonstrated that the secretory autophagy pathway with autophagosome biogenesis gene, Beclin-1, Atg5, Atg16 and LC3, together with the Golgi reassembly-stacking protein gene-55 (GRASP55) regulates exosome-independent secretion of Hsp90α. Finally, in collaboration with Dr. Xin Tang, we provided direct evidence that eHsp90α is essential during wound healing. This thesis analyzed the extracellular distribution of secreted Hsp90α, quantitated the tissue concentration of eHsp90α derived from different extracellular fractions, and reveled that it is secretory autophagy rather than exosome secretion regulates the functional concentration of eHsp90α required for wound healing.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Mechanism of secretion and function of heat shock protein-90 (Hsp90) family genes
PDF
The mechanism by which extracellular Hsp90α promotes cell migration: implications in wound healing and cancer
PDF
Extracellular Hsp90α in diabetic wound healing: structure, function, and mechanism of action
PDF
Characterization of a fragment in secreted Hsp90α with potential therapeutic benefits in wound healing and cancer
PDF
To investigate whether Hsp90α is present on the surface of exosome
PDF
Study of the mechanism by which hypoxia-inducible-factor-1 alpha (HIF-1α) regulates heat shock protein-90 alpha (Hsp90α) secretion in breast cancer
PDF
The role of secreted Hsp90α in tissue repair and cancer
PDF
TGF-β signaling regulates gingival epithelial wound healing and barrier function
PDF
Amelogenin-ameloblastin protein interaction and function in dental enamel formation
PDF
Identifying and characterizing a unique enhancer for Y537S mutant estrogen receptor (ER) in breast cancer cells
PDF
From mesenchymal stem cell therapy to discovery of drug therapy for systemic sclerosis
PDF
Molecular basis of CD33 receptor function, and effects of a destabilizing transmembrane motif on αXβ2 integrin
Asset Metadata
Creator
Chang, Cheng
(author)
Core Title
Extracellular heat shock protein-90alpha (eHsp90α): mechanisms of secretion, quantitation and function
School
School of Dentistry
Degree
Doctor of Philosophy
Degree Program
Craniofacial Biology
Degree Conferral Date
2024-05
Publication Date
03/27/2024
Defense Date
03/22/2024
Publisher
Los Angeles, California
(original),
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
CIF ratio,extracellular Hsp90,Hsp90,OAI-PMH Harvest,unconventional protein secretion,wound healing
Format
theses
(aat)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Li, Wei (
committee chair
), Chen, Jianfu (
committee member
), Frey, Mark (
committee member
), Lin, Dechen (
committee member
), Xu, Jian (
committee member
)
Creator Email
113272453@qq.com,chan296@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC113858370
Unique identifier
UC113858370
Identifier
etd-ChangCheng-12720.pdf (filename)
Legacy Identifier
etd-ChangCheng-12720
Document Type
Dissertation
Format
theses (aat)
Rights
Chang, Cheng
Internet Media Type
application/pdf
Type
texts
Source
20240327-usctheses-batch-1131
(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.
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
Tags
CIF ratio
extracellular Hsp90
Hsp90
unconventional protein secretion
wound healing