University of Southern California Dissertations and Theses

A novel chemoenzymatic conjugation method for bispecific antibody production

(USC Thesis Other)

A novel chemoenzymatic conjugation method for bispecific antibody production

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2022 Hyo Sun Kim

A Novel Chemoenzymatic Conjugation Method for Bispecific Antibody Production

By

Hyo Sun Kim

A Thesis Presented to the
FACULTY OF THE USC SCHOOL OF PHARMACY
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(PHARMACEUTICAL SCIENCES)

May 2022
ii
Acknowledgements
I would like to thank everyone who was supportive for the completion of my master’s
studies for the past two years. I am very happy that I joined Dr. Yong (Tiger) Zhang’s lab for my
master studies. I believe my experience in this lab has taught me problem solving skills along
with critical thinking skills and most importantly skills needed to work with others in a very
cooperative environment. I thank Dr. Zhang for introducing me into the field of protein
engineering and immunotherapy, providing detailed feedbacks on my research, and filling in
knowledge gap in my background. Communication with him was the driving force of the
research progression.
I also want to thank not only the current members of our lab but also the previous
members who mentored me when I first joined the lab: Qinqin Cheng, Zhefu Dai, Tian Xiao
Zhang, Jiawei Li, Xiao-Nan Zhang, Elisa Stephens, Guoyun Kao, Po-Wen Yu, Kimia Hariri,
Zekun Li, Liang-Chieh Chen, Yuanteng Zhao, and Gabriel Mpilla. Without their help, I would
not be able to adjust myself successfully into the lab.
I want to thank my committee members as well: Dr. Ian Haworth and Dr. Rebecca
Miranda Romero for their feedbacks and supports for my thesis project. Also, the courses taught
by them helped me a lot to have background knowledge for my research.
Lastly, I want to thank my family and friends for their endless support that kept me
sturdy even in the hard times. I hope I could do the same for them in the future.
iii
Table of Contents
Acknowledgements ......................................................................................................................... ii
List of Figures ................................................................................................................................. v
List of Tables ................................................................................................................................. vi
Abbreviations ................................................................................................................................ vii
Abstract .......................................................................................................................................... ix
Chapter 1: Introduction ................................................................................................................... 1
1.1 Immunotherapy ..................................................................................................................... 1
1.2 Bispecific Antibodies as Cancer Immunotherapeutics ......................................................... 4
1.3 Current Production Methods of Bispecific Antibodies ......................................................... 8
1.4 Enzymatic Activity of CD38 .............................................................................................. 14
1.5 Click Reaction between ...................................................................................................... 15
Chapter 2: Materials and methods ................................................................................................ 17
2.1 Molecular Cloning .............................................................................................................. 17
2.2 Protein Expression and Purification .................................................................................... 24
2.3 Biological Activity Assay ................................................................................................... 27
2.4 Protein Conjugation ............................................................................................................ 31
Chapter 3: Results ......................................................................................................................... 33
3.1 Molecular Cloning .............................................................................................................. 33
3.2 Antibody Expression and Purification ................................................................................ 37
3.3 Biological Activity Assay ................................................................................................... 41
3.4 Protein Conjugation ............................................................................................................ 43
Chapter 4: Discussion ................................................................................................................... 48
iv
References ..................................................................................................................................... 53

v
List of Figures
Figure 1. Different Types of Bispecific Antibodies Based Immunotherapy .................................. 7
Figure 2. Clinical Trials Relevant to Bispecific Antibody Research .............................................. 9
Figure 3. Models of Linkers Used for Conjugation and Conjugation Scheme ............................. 16
Figure 4a. Herceptin scFv C terminal Fusion Construct ............................................................... 34
Figure 4b. Herceptin scFv N-terminal Fusion Construct .............................................................. 34
Figure 4c. UCHT1 scFv C-terminal Fusion Construct ................................................................. 34
Figure 4d. UCHT1 scFv N-terminal Fusion Construct ................................................................. 34
Figure 4e. UCHT1 Fab Light Chain C-terminal Fusion Construct .............................................. 34
Figure 4f. UCHT1 Fab Heavy Chain C-terminal Fusion Construct ............................................. 34
Figure 4g. Herceptin Fab Light Chain C-terminal fusion Construct ............................................ 34
Figure 4h. Herceptin Fab Heavy Chain C-terminal fusion Construct .......................................... 34
Figure 5. Overlap Extension PCR Products of anti-CD3-CD38 Fusions ..................................... 35
Figure 6. Colony PCR of DH10B Cells Transformed with pFuse_CD38-Herceptin scFv .......... 36
Figure 7. SDS-PAGE Analysis of Each Antibody Fusion ............................................................ 38
Figure 8a, b. NGD Assay Analysis of CD38 activity ................................................................... 40
Figure 9a-d. Flow Cytometry Histograms for Antibody Fusions ................................................. 42
Figure 10a, b. NGD Assay Analysis for Inhibition of CD38 fused with scFv and Fab ............... 44
Figure 11. SDS-PAGE Analysis of scFv Conjugation ................................................................. 45
Figure 12. Protein Stability Analysis via SDS-PAGE Gel Electrophoresis ................................. 46
Figure 13. Dot Blot Assay for Linker Conjugated Antibodies ..................................................... 47
vi
List of Tables
Table 1. Primers Used for Overlap Extension PCR ...................................................................... 19
Table 2. Parameters of PCR Reactions ......................................................................................... 21
Table 3. Sample Preparation for Flow Cytometry ........................................................................ 29
Table 4. Expected Molecular Weight of Each Antibody Fusion .................................................. 39

vii
Abbreviations
ADC: antibody-drug conjugate
ADCC: antibody-dependent cellular cytotoxicity
APC: antigen presenting cells
araF-NAD
+
: nicotinamide 2’-deoxy-2’-fluoroarabinoside adenine dinucleotide
araNAD
+
: nicotinamide arabinoside adenine dinucleotide
CAR: chimeric antigen receptor
CDC: complement-dependent cytotoxicity
DTT: dithiothreitol
EGFR: epidermal growth factor
ELISA: enzyme-linked immunosorbent assay
EtBr: ethidium bromide
Fab: antigen binding fragment
Fc: fragment crystallizable
Fw: forward primer
HC: heavy chain
IgG: immunoglobulin G
IL-2: interleukin-2
KIR: killer cell Ig-like receptors
LB: Luria broth
LC: light chain
LDS: lithium dodecyl sulfate
MHC-I: major histocompatibility complex class I
viii
NGD: nicotinamide guanine dinucleotide
NK: natural killer cells
NMN: nicotinamide mononucleotide
PBS: phosphate-buffered saline
PCR: polymerase chain reaction
Rv: reverse primer
scFv: single-chain variable fragment
SDS-PAGE: sodium dodecyl sulphate-polyacrylamide gel electrophoresis
SIRPα: signal-regulatory protein α
TAE: tris-acetate-EDTA
THPTA: tris-hydroxypropyltriazolylmethylamine
Tm: melting temperature
WT: wild type

ix
Abstract
One of the challenges that hinders bispecific antibodies from getting approval for its use
in clinical trials is the production in terms of quality, quantity, and stability. There have been
extensive studies trying to develop production methods that could solve the problems, but some
methods result in heterogeneous pool of products with batch-to-batch variations, low yields, and
fluctuations of efficiency. In this study, we tried to develop a chemo-enzymatic conjugation
method for bispecific antibody production that can result in homogeneous batch of products with
constant efficiency in every batch as well as potentially improved pharmacological activities.
This chemo-enzymatic conjugation involves an expression of an enzyme called CD38 on
designated sites of either Fab or scFv antibodies. CD38 is a membrane enzyme expressed on
immune cells that has both ADP-ribosyl (ADPR) cyclase and cADPR hydrolase activities.
However, it can also be covalently inhibited by NAD
+
analogs. Using this characteristic, two
types of functionalized NAD
+
analogs were synthesized: one with an azide functional group and
another with a DBCO functional group. These linkers were then conjugated with both scFv and
Fab format antibodies via enzyme inhibition. The two linkers are expected to go through a click
reaction and facilitate conjugation of two antibodies. In this thesis so far, functional antibody
CD38 fusion expressions have been done and confirmed via different functional assays.
However, one of the NAD
+
analogs had a stability issue, which disabled the conjugation of the
antibodies. Once this stability issue is resolved, the developed conjugation approach is expected
to allow for successful production of functional bispecific antibodies that can recruit effector T
cells in proximity to cancer cells and redirect the immune system toward fighting the tumors.

1
Chapter 1: Introduction
1.1 Immunotherapy
Immunotherapeutic drugs are cancer treatments that utilize the patient’s own immune
system to fight the tumor cells. The idea of using one’s own immune system to eradicate
abnormal cells in the body started with an immunologist in the early 20
th
century, Paul Ehrlich,
who first came up with an idea that tumors consistently arise in the human body, but it gets
eradicated by the immune system that always scans for abnormality in the body (Ehrlich, 1909)
This theory was partially revoked with experimental evidence that there was no increased
incidence of tumor in athymic nude mice compared to wild-type mice during an observation
period of up to 420 days (Rygaard, Acta Pathol Microbiol Scand, 1974). However, this theory
was partially supported with an experimental evidence that tumors could be repressed by the
immune system in the mid 20
th
century by Ludwik Gross’s research in 1943. This research
strongly suggested the existence of tumor-associated antigens (TAAs) that has become the basis
of immunosurveillance theory postulated by Sir Macfarlane Burnet and Lewis Thomas, which
states that immune system recognizes specific antigens displayed on tumor cells and kills
spontaneously developed carcinomas (Burnet, 1957 & 1970) (Thomas, 1959)
With this theory, there have been an extensive effort in understanding the immune
system, accelerating the development of immunotherapeutics that boost the immune system in
the human body (Malmberg, 2004) . There are three big categories of immunotherapy: toll-like
receptor agonist therapies, cell-based immunotherapies, and antibody-based therapies (Borghaei
et al., 2009). Toll like receptors (TLR) are receptors expressed on sentinel cells such as
macrophages and dendritic cells. Because they recognize structurally conserved molecules
derived from microbes and signal for innate and adaptive immune attack, they are considered to
2
be a key first line defense system against microbes in the human body. Therefore, agonists for
TLR2, TLR3, TLR4, and TLR9 can boost inflammatory innate immune response against tumor
cells (Vijay, 2018).
Cell-based therapies involve modification of lymphocytes and injection of them to
improve immune response against cancer cells. Different types of lymphocytes can be used for
this treatment: T cells, Natural Killer (NK) cells, and dendritic cells. The first attempt to use the
T cells was to collect the patient’s own T cells, culture them with cytokines such as IL-2 that
could activate the T cells, and lastly, inject them back into the patient. These attempts resulted in
increasing the number of activated T cells and provide long term immune protection with
minimal autoimmune responses. Some complementary therapies were also used, such as using
chemotherapy and/or total body irradiation for lymphodepletion prior to cell infusion because it
can bring some positive effects for T cell therapy such as reduction of regulatory T-cells and
damage to vascular endothelial cells, making the environment more hospitable for modified T
cells (Budde, 2020). In the recent years, along with the traditional T cell therapy, a targeted
therapy called Chimeric Antigen Receptor (CAR) T cell therapy had been developed for
improved specificity and reduced off-target effect. For this therapy, the T cells collected from the
patients are transduced to express CARs that are specific to a certain type of cancer; thus, they
can direct the CAR T cells to the tumor cells. Along with this therapy, a NK cell therapy was
also developed. NK cells are a type of cytotoxic lymphocytes that are also critical to the innate
immune system. They get activated when two criteria are met: lack of ligand for its inhibitory
receptor called Killer Cell Ig-like receptors (KIR), and the presence of stress-related molecules
(Liu et al, 2021). KIR is specific to major histocompatibility complex class I (MHC-I)
molecules. MHC-I molecules, however, can also be identified by cytotoxic T cells and activate
3
the immune attack from the T cells. Therefore, tumor cells tend to downregulate the expression
of MHC-I molecules in order to avoid recognition by T cells, which in turn activates the immune
response of NK cells. Therefore, injection of NK cells can bring up the inflammatory immune
response. Unlike T cell therapies, NK cells are most of the times collected from donors who are
different from the recipients because individuals inherent different KIR family members, and
mismatch of the donor’s KIR on NK cells and the recipient’s KIR ligand on tumor cells can
improve the efficacy of this therapy (Ruggeri et al.,2002). Lastly, dendritic cell therapy is
another type of cellular therapy that uses dendritic cells, crucial initiators of T cell-dependent
immune responses. Because they are considered to be the most effective antigen presenting cells
(APC), they have been extensively studied as potential vectors that can be loaded with peptide
antigens (Cintolo et al., 2012). For example, dendritic cells can be grown alongside cancer cells
in the lab to obtain the ability to recognize the cancer cells. Then, these cells can be used as
cancer vaccine.
Along with toll-like receptor agonists and cell-based immunotherapeutics, antibody-
based immunotherapeutics have gained attention for its potential to be very effective treatment
due to its potential to be modified extensively. Antibody based therapies for cancer treatment
could be classified into three major categories: antibody-dependent cellular cytotoxicity (ADCC)
therapy, complement-dependent cytotoxicity (CDC) therapy, and intracellular signal induction or
blockade therapy. First of all, for ADCC therapy, antibodies coat a target cell by binding to the
antigens on the surface of the target cell. Then, the fragment crystallizable (Fc) region of the
antibody interacts with effector cells, helping effector cells to attack the cancer cells (Zahavi, et
al., 2018). Even though ADCC can be mediated with all NK cells, monocyte-macrophages, and
neutrophils, NK cells are the principal ADCC effects where CD16 receptors on NK cells bind to
4
Fc regions on IgG antibodies and signal for cytotoxic attack. CDC therapy, on the other hand,
have the Fc region interact with C1q protein, which can form a membrane attack complex
(MAC). Therefore, once the antibody is bound to the target cell surface antigens, protein C1q
binds to the Fc domain and triggers the activation of the complement cascade and forms a MAC.
This complex then forms pores in the cell membrane, leading to apoptosis. After the apoptosis,
complement activation can deposit C3, C3b, and C3b breakdown products on the cell surface
which can be recognized by C3-receptors on effector cells, enhancing cell killing effect
(Kennedy et al., 2003). Antibodies can also alter intracellular signaling pathways. For example,
epidermal growth factor (EGFR) and human epidermal growth factor receptor 2 (HER2)/neu
inhibitors bind to the extracellular component of the EGFR and inhibits growth factor binding to
the receptor, preventing signal transduction pathway for cell division. Also, checkpoint inhibitors
such as inhibitors of programmed cell death protein 1 (PD-1) and programmed death-ligand 1
(PD-L1) can interrupt the binding between PD-1 on effector cells and PD-L1/L2 overexpressed
on the surface of some tumor cells, inhibiting immunosuppression, (Dahlen et al., 2018).

1.2 Bispecific Antibodies as Cancer Immunotherapeutics
Even though antibody-based therapy has worked for some patients, the overall response
rate or the efficacy stayed low for some patients. Therefore, there have been attempts to improve
their efficacy by engineering or modifying the conventional immunoglobulin molecular format.
Among them, what has gained the most attention was bispecific antibody therapies, where the
two different antibodies were conjugated to carry out dual functions. Because bispecific
antibodies can interact with more than one surface antigen, they showed higher efficacy and
safety as cancer treatments. Depending on which two antibodies are designed to be conjugated
5
together, the bispecific antibodies could have different functions such as redirection of cytotoxic
effector cells and enhancement of antitumor activity (Figure 1).
Pan-T cell engagers have two antibodies where one targets CD3 on Pan-T cells and the
other targets antigens that are specific for tumor cells. Therefore, once it is injected, it can bind to
Pan-T cells expressing CD3 antigen and redirect them in proximity to the target cancer cells,
bringing up the immune response against the cancer cells. However, because anti-CD3 bispecific
antibody targets a large proportion of the T cell pool, T cell engagers could be made using
antibody that targets antigens that are specifically expressed on the surface of subsets of T cells
such as CD8 for cytotoxic T lymphocytes (Blanco, 2021). Other than the T cells, NK cells could
also be redirected using bispecific antibodies that bind to antigens on NK cells and antigens on
the tumor cells.
Other than redirecting cytotoxic effector cells, bispecific antibodies can be designed to
restore and enhance antitumor activity. For example, in order to block innate immune
checkpoints, a bispecific antibody can target the tumor cell and signal-regulatory protein α
(SIRPα) on myeloid cell that otherwise binds to CD47 overexpressed on the surface of tumor
cell. Because binding of SIRPα and CD47 lets the tumor cell escape from macrophage-mediated
phagocytosis, interrupting the binding can result in immune attack against the tumor cell.
Another example of T cell activity enhancement could be done with bispecific antibodies that
interrupt the binding of check-point inhibitor to its receptor (PD-1 or CTLA-4), interrupting
immunosuppression. In this case, bispecific antibodies can be designed to have trans co-
engagement (binding to PD-1 on T cells and PD-L1 on tumor cells) or cis co-engagement (PD-1
and CTLA-4 on T cells). Costimulatory receptors of the TNF receptor superfamily (TNFRSF)
such as CD27, OX40, 4-1BB, and GITR can also be targeted by bispecific antibodies because
6
they are expressed on T cells after their activation. Therefore, if agonist antibodies are bound to
them, stimulation or activation of T cells is expected (Duhen et al., 2021). Therefore, the T cells
are stimulated to attack the tumor cells bound to the other part of bispecific antibodies. A
combination therapy could be used to have both costimulatory and co-inhibitory immune
checkpoint receptors bound with a bispecific antibody. It has been shown that T cell co-
stimulation along with PD-L1 checkpoint inhibition resulted in potent antitumor immune
responses in vivo tumor models.
7

SIRPα
Figure 1. Different Types of Bispecific Antibodies Based Immunotherapy (Blanco et al., 2021)
8
1.3 Current Production Methods of Bispecific Antibodies
Even though the idea of bispecific antibody was first suggested in 1964 by Alfred
Nisonoff and his collaborators, its translation rate into clinical trials was quite slow due to
difficulties in basic research and production, having the first bispecific antibody approved by
FDA in 2014 (Przepiorka, 2015). The first bispecific antibody was called Blincyto, an anti-CD19
and anti-CD3 bispecific antibody for treatment of relapsed or refractory B-cell acute
lymphoblastic leukemia. Ever since this approval, there has been a number of clinical trials
registered. Figure 2 depicts the increasing number of clinical trials for bispecific antibody
research from 2011 to 2020, and the cancer types these clinical trials study for.

9

Figure 2. Clinical Trials Relevant to Bispecific Antibody Research (Wang et al., 2021)

10
With its high potential for extensive modification, bispecific antibodies have gained a lot
of attention. Especially the production method of bispecific antibodies has been studied in lots of
research recently because unlike the monoclonal antibodies, it is challenging to get approval for
bispecific antibodies because of struggles with the quantity, quality, and stability. Blincyto, the
first approved bispecific antibody was produced as a single chain recombinant protein. Bispecific
antibodies can be produced using three different methods: recombinant production, Quadroma
technology, and chemical conjugation. For the recombinant production of bispecific IgG, cells
can be genetically modified to produce heavy and light chains of two different antibodies in the
same cell. However, random assembly of these four different fragments can result in
heterogeneous batch of 16 different combinations. Therefore, some new approaches have been
discovered such as knobs-into-holes BsAb IgG, where mutations on CH3 domains force heavy
chain heterodimerization and result in asymmetric antibodies, Ig-scFv fusion, where scFv format
of the second binding moiety is added to either C- or N-terminus of the heavy or light chain of
the first binding moiety (IgG format), diabody-Fc fusion, where the Fab fragment of an IgG is
replaced with noncovalent dimer of scFv (diabody), and Dual-Variable-Domain-IgG , where
heavy and light variable fragments of the first binding moiety are fused into those of the second
binding moiety (Liu et al., 2017). These methods are for production of bispecific antibodies with
Fc region. Besides, some bispecific antibodies without Fc region can be produced recombinantly.
The types of such bispecific antibodies are: tandem scFv, diabodies, DARTs, scFv-CH1/CL-
scFv, Fab-scFv, Fab-Fv, and many other formats. The main characteristic of these antibodies is
that it lacks Fc mediated effector functions such as ADCC, antibody dependent cellular
phagocytosis (ADCP), CDC, and Fc-mediated recycling. However, these kinds of antibodies are
easy to be designed because each binding moiety can be just linked with short peptide linkers.
11
Quadroma technology is based on the somatic fusion of two distinct hybridoma cells.
Each hybridoma produces a unique monoclonal antibody. Once two cells are fused, the fused cell
express immunoglobulin heavy and light chains from both parents. Then random assembly of
each piece occurs resulting in one antibody of desire in 10 different combinations (Kontermann,
2005). In order to reduce the chance of producing non-desired antibody, fusion of two
hybridomas of different species can be done. The resulted antibody is in the conventional
antibody format with Fc region that can bring Fc mediated effector functions.
The last method of antibody production is via chemical conjugation of two purified
antibodies (Dimasi et al., 2021). Chemical conjugation can be done on lysine on full-length
antibodies. The advantage of this method is that it is very simple. For example, reagents like N-
hydroxysuccinimide can be used as a bridging reagent. This reagent can react with amine group
of lysine and add succinimidyl-3 (2-pyridylthiol)propionate (SPDP) on the antibody. Then the
PDP group can be reduced using dithiothreitol (DTT) and react with lysine on the other antibody,
producing heteroconjugate with a disulfide bond. However, because of the presence of a lot of
lysine residues in antibody scaffolds, the conjugation site for this method is non-selective, thus,
resulting in batch-to-batch variation. Also, the overall yield could be very low (~10-29%) (Lee et
al., 2007). Cysteine can also be used for conjugation. The basic is very similar to the lysine
conjugation. For example, a bridging reagent, o-phenylenedimaleimide has two maleimide
groups that can react with thiols in the cysteine. By reacting each maleimide group with different
antibody, two different antibodies can be bridged with this reagent. While lysine conjugation is
commonly used for full-length antibodies, cysteine conjugation is commonly used for Fab
antibodies. One advantage of using cysteine over lysine is that cysteine is less commonly found
12
in antibodies (less propensity of heteroconjugation) and is more tightly tied up in disulfide
bonds.
Another chemical conjugation method is to have site-specific genetic incorporation that
can express unnatural amino acids with some modifications that can be used for conjugation after
purification. For example, as reported in 2012, an evolved tRNA/aminoacyl-tRNA synthetase
pair was used to incorporate p-acetylphenylalanine on some sites of Fab antibodies. In this work,
each Fab antibody mutant was coupled with an alkoxy-amine containing drug linker with an
oxime bond. In this case, one Fab antibody had a drug linker with an azide group on the opposite
side from the alkoxy-amine group, and the other Fab antibody had a drug linker with a
cyclootyne group. Because azide group and cyclootyne group are orthogonal to each other, they
can go through a click chemistry, yielding [3+2] cycloadduct between the linkers coupled with
the antibodies. For ligation of azides and alkynes, Cu(I) catalysts are needed to increase the
reaction rate. However, because cyclootyne group has an alkyne group in a ring, this strain in the
ring can promote the cycloaddition, making it possible to run the reaction without a copper
catalyst (copper-free [3+2] Huisgen cycloaddition) (Agard et al., 2004). This research showed
that in the presence of effector cells the bispecific antibody produced via this method could
successfully induce targeted cell lysis. Also, because this method could incorporate unnatural
amino acids at the designated sites of the antibody, homogeneous and chemically defined
bispecific antibodies could be produced with a high yield (70%). However, even though this
method has such advantages, efficient incorporation of unnatural amino acid is prohibited by the
fact that the efficiency is highly dependent on the sequence context surrounding the in-frame
UAG amber stop codons (Kato et al., 2017). Specifically, if release faction-1 (RF-1) recognizes
UAG for termination of protein synthesis faster than the tRNA that translates the UAG codons
13
into the unnatural amino acid, then that antibody gets produced without the extra unnatural
amino acid. Therefore, it becomes difficult to find the useful conjugation site for each antibody.
This project aims to develop a novel bispecific antibody conjugation method that
produces homogeneous batch of antibodies. This method involves recombinant conjugation of an
enzyme called CD38, which can form a covalent bond with its inhibitor. Because it is
recombinantly fused to the antibody, a conjugation site can be designed to be on specific location
of the antibody, either on C- or N-terminus of either the heavy or light chain of the antibody,
resulting in homogeneous conjugation. Due to this, high yield of desired conjugate is expected,
overcoming the disadvantage of chemical conjugation using lysine and cysteines. This method
also takes the advantage of using the chemical conjugation such as precise linker design because
the covalent inhibitors for CD38 enzyme can be chemically synthesized to be a linker for the
bispecific antibody product. For this project, CD38 covalent inhibitors were modified to have
two different functional groups that were orthogonal to each other, making it possible for two
antibody-CD38 fusions inhibited with those to be conjugated. Because the proteins are
recombinantly produced and the linkers are chemically synthesized, it gets rid of the need to
design many different gene sequences for bispecific antibody and reduces the production time,
overcoming the disadvantage of recombinant conjugation method.

14
1.4 Enzymatic Activity of CD38
CD38 is a transmembrane glycoprotein with an ectoenzymatic function expressed in
many immune cell types including bone marrow progenitors, natural killer cells, monocytes, and
activated T- and B- lymphocytes (Malayasi et al., 2008). As a cyclase, its extracellular domain
metabolizes nicotinamide adenine dinucleotide (NAD
+
) and cytoplasmic nicotinamide adenine
dinucleotide phosphate (NADP) and produces Ca
2+
-mobilizing compounds such as cyclic
adenosine diphosphate (ADP) ribose and nicotinic acid adenine dinucleotide phosphate. ADPR
produced by CD38 is then metabolized by CD203a/PC-1 and CD73 and become
immunosuppressive molecule, adenosine. Therefore, some immunotherapeutics targeting CD38
were developed to downregulate the immunosuppression (Morandi et al., 2018).
However, in 1992, Helene M. Muller-Steffner et al. found out that nicotinamide
arabinoside adenine dinucleotide (araNAD
+
) and nicotinamide 2’-deoxy-2’-fluoroarabinoside
adenine dinucleotide (araF-NAD
+
) are not hydrolyzed by CD38 at measurable rates, suggesting
these kinds of compounds to be inhibitors of CD38. Also, this research concluded that NAD
+

analogues possessing a pyridinium moiety with a high pKa value are poor substrates to CD38,
being good competitive inhibitors. With this result, previous research of our lab found out that
CD38 could be incorporated into antibodies in order to produce site-specific antibody-drug
conjugates (ADCs). In those studies, antibody-CD38 fusions were inhibited using dinucleotide-
based covalent inhibitors conjugated with payload. That project was successful in producing
homogeneous ADCs. With the previous success in incorporating CD38 on specific sites of
antibodies, this project tried to expand the usage of antibody-CD38 fusion and develop a novel
way to produce bispecific antibody using CD38.

15
1.5 Click Reaction between
In this project, dinucleotide-based covalent inhibitors were synthesized as linkers
between two different antibodies. The models of the linker conjugated antibodies (scFv format)
and the scheme of the click reaction is included in Figure 3. From previous research, 2’-Cl-
substituted NAD
+
analogs acted as a covalent inhibitor for CD38 enzymes. Therefore, as
depicted in the structure, the azide linker is composed of 2’-Cl-araNMN. This structure is
designed so that it can form stable arabinosyl-ester bonds with CD38 E226 residue. The DBCO
linker has a 2’-Cl-araNAD
+
for the same reason. Then, the azide linker has the azide group on
the other end of the linker, while the DBCO linker has the dibenzycyclooctyne (DBCO) group.
As described above, the azide group and the DBCO group can go through copper free [3+2]
cycloaddition (click-reaction). Therefore, after linker conjugation with the antibody, two
different antibody-CD38 fusions are planned to be conjugated via click reaction in this project to
form a bispecific antibody. In this project two versions of T cell redirector bispecific antibodies
were produced using this chemo-enzymatic conjugation method: anti-CD3 anti-HER2 scFv
bispecific antibody and anti-CD3 anti-HER2 Fab bispecific antibody.

16

Figure 3. Models of Linkers Used for Conjugation and Conjugation Scheme. The models of
the linker conjugated antibody in scFv format (Left: azide linker; Right: DBCO linker). For the
Fab formats, the scFv format antibodies in dark and light blue color will be substituted with Fab
format antibodies.

CD38
linker
CD38
linker
+
CD38
linker
CD38
linker
17
Chapter 2: Materials and methods
2.1 Molecular Cloning
Polymerase Chain Reaction (PCR)
For this project, both scFv and Fab versions of anti-HER2 antibodies (Herceptin) and
anti-CD3 antibodies (UCHT1) with CD38 enzyme conjugated at the optimal sites were
produced. In order to place enzyme CD38 on different locations of the antibodies and express
them, the genes that encode for CD38 were inserted in the different parts of the DNA sequence
of antibodies. For scFv antibodies, CD38 genes were inserted either on C-termini or N-termini of
the antibody sequences. His6 tag was also added to the C-terminus of the whole sequence for
purification in the future. Because a construct of Herceptin with CD38 conjugated on the C-
terminus was provided from the previous member of the lab, overlap extension PCR was done to
switch the order of Herceptin encoding gene fragment and the CD38 encoding gene fragment
and produce N-terminal fusion. Constructs of both N- and C-terminal fusions of anti-CD3
antibodies were provided by the previous member of the lab.
For Fab antibodies, CD38 genes were inserted in the C-terminus of either the heavy
chain (HC) or light chain (LC) sequence because it was previously proved by the previous
projects in the lab that fusing CD38 on the N-terminus of Fab antibody affects the binding of the
antibodies. Using the CD38 gene fragment from CD38-Herceptin scFv N-terminal fusion,
constructs for UCHT1 Fab HC-CD38 and UCHT1 LC-CD38 were produced via overlap
extension PCR. Using the same CD38 gene fragment, constructs for Herceptin Fab HC-CD38
and Herceptin Fab LC-CD38 fusions were produced by another member in the lab.
In order to develop the constructs needed, two rounds of PCR were used. The first round
was for amplification of both the gene fragments of the antibody and those of CD38 enzyme.
18
Primers were designed to encode each fragment and add an overlapping region where two
fragments have the same sequence of short length. Also, the primers were designed so that the
digestion sites that are included in the pFuse plasmid backbone could also present before and the
after of the gene sequence of interest. Designed primers were ordered from Integrated DNA
Technologies IDT, which were then diluted to be 100 uM in ddH2O as a stock solution.

19
Table 1. Primers Used for Overlap Extension PCR
Construct 1
st
round primer sequence
Tm
(°C)
2
nd
round primer sequence
Tm
(°C)
CD38
–
Herce
ptin
scFv
CD38
Fw
GCCGGCCAGATCTA
GATGGAGGCAACAA
TGGTCAGGC
69.7
Fw
GCCGGCCAGATC
TAGATGGAGGCA
ACAATGGTCAGG
C
69.7
Rv
TGGATGTCACTACCC
CCTCCGCCGATCTCA
CTAGTACATGAACTA
TCCTCTGGGTTTTTG
69.9
Herce
ptin
scFv
Fw
TGAGATCGGCGGAGG
GGGTAGTGACATCCA
GATGACCCAGTCTCC
ATCC
72
Rv
TGGCCAGCTAGC
ACTTATCAATGGT
GGTGGTGATGGT
GAGAGGACACTG
TGACCAGGGTG
72.4
Rv
TGGCCAGCTAGCAC
TTATCAATGGTGGT
GGTGATGGTGAGAG
GACACTGTGACCAG
GGTG
72.4
UCH
T1
Fab –
HC
CD38
UCHT
1 Fab
HC
Fw
TCACGAATTCGGAG
GTCCAGTTACAG
60
Fw

TCACGAATTCGG
AGGTCCAGTTAC
AG

60
Rv
CTTCCGCCACCCCC
TGTGTGAGTTTTGT
CGCAAGATTTGG
58.3
CD38
Fw
ACAAAACTCACACA
GGGGGTGGCG
63.9
Rv

CCAGCTAGCACTT
ATCAGATCTCACT
AGTACATGA

60.9
Rv
CCAGCTAGCACTTA
TCAGATCTCACTAG
TACATGA
60.9
UCH
T1 –
LC
CD38
UCHT
1 LC
Fw
TCACGAATTCGGAT
ATCCAGATGACACA
GAC
60.8
Fw
TCACGAATTCGG
ATATCCAGATGA
CACAGAC
60.8
Rv
CTTCCGCCACCCCC
ACACTCTCCCCTGT
TGAAGCTC
58.8
CD38
Fw
GGGAGAGTGTGGG
GGTGGCGGAAGCA
GA
63.4
Rv
CCAGCTAGCACTT
ATCAGATCTCACT
AGTACATGA
60.9
Rv
CCAGCTAGCACTTA
TCAGATCTCACTAG
TACATGA
60.9

20
For the first round of PCR, 50 uL of the reaction was prepared in PCR tubes with 1.5 uL
of 10 uM forward and reverse primers, 5 uL of 10X AccuPrime™ Pfx Reaction Mix (Invitrogen,
Waltham, MA), 0.4 uL of the AccuPrime™ Pfx DNA Polymerase (2.5 U/µL) (Invitrogen,
Waltham, MA), 100 ng of the DNA template, and enough volume of ddH2O to make up 50 uL
of total volume. Then, the 35 cycles of PCR were done as described in Table 2, where the
annealing step temperature was decided to be 2-3 degrees lower than the melting temperature of
the region on the template that each primer anneals to. PCR product of each fragment was then
analyzed by agarose gel electrophoresis for its size. 1 kb Plus DNA Ladder (New England
BioLabs, Ipswich, MA) was used as reference. 1% agarose gel in TAE with EtBr added at 1 in
20000 ratio was used. Then, the gel piece of the DNA fragment was cut out and melted with
ZymoClean DNA Recovery Kit (Genesee Scientific, San Diego, CA). Two recovered DNA
fragments, one for antibody and the other for CD38, are expected to have a common overlapping
sequence. Therefore, in order to anneal them together, each fragment of 1 to 1 molar ratio was
added to a total of 50 uL reaction that included 5 uL of 10X AccuPrime™ Pfx Reaction Mix
(Invitrogen, Waltham, MA), 0.4 uL of the AccuPrime™ Pfx DNA Polymerase (2.5 U/µL)
(Invitrogen, Waltham, MA), and enough volume of water (100 ng of antibody fragment and
enough weight of CD38 fragment were used at 1:1 molar ratio). The reaction went through 20
cycles of PCR as described in Table 2 without any primers. For this step, the temperature of the
annealing step was adjusted to be 2-3 degrees lower than the melting point of the overlapping
region. Lastly, primers for the second PCR were added to the reaction, and the reaction went
through another 35 cycles of PCR for amplification. The temperature of the annealing step was
then adjusted with respect to the melting temperature of the primers. PCR product once again
went through analysis via 1% agarose gel electrophoresis and purification via gel recovery.
21
Table 2. Parameters of PCR Reactions
Step Time (mm:ss) Temperature (°C)
1. Initialization 5:00 95
2. Denaturation 0:15 95
3. Annealing 0:30 2-3°C lower than Tm
4. Extension 1:00 per kb 68
*Steps 2-4 are repeated for each cycle.

22
Construct Insertion into Vectors
Recovered constructs were then inserted into pFuse expression vector. Digestion of both
the backbone vector and the insert and ligation of the two pieces were performed. Digestion of
each vector and the insert was done in the total of 50 uL reaction with 5 uL of 10X digestion
buffer, 1-5 ug of DNA, 2 uL of each digestion enzymes (20,000 units/mL), and enough volume
of water. For both UCHT1 Fab HC-CD38 and LC-CD38 constructs, double digestion was done
at 37 °C overnight since both EcoRI-HF (New England BioLabs, Ipswich, MA) and NheI-HF
(New England BioLabs, Ipswich, MA) that were used for the digestion of these two constructs
have the optimal activity in the same buffer, rCutSmart (New England Biolabs Inc., MA).
However, for the Herceptin-CD38 scFv construct, sequential digestion was done because
different buffers were needed for optimal activity of BglII (New England BioLabs, Ipswich,
MA) and NheI-HF (New England BioLabs, Ipswich, MA). Digestion with NheI-HF in
rCutSmart was done first at 37 °C for 3 hours, and then the DNA was recovered using DNA
clean and concentration kit (Genesee Scientific, San Diego, CA). The second digestion with
BglII in NEBuffer r3.1 (New England Biolabs Inc., MA) was done at 37 °C overnight. The
digested products were then analyzed and purified via 1% agarose gel electrophoresis and
ZymoClean DNA Recovery Kit (Genesee Scientific, San Diego, CA).
Digested inserts and the backbones were then ligated in the 20 uL reaction where 100 ng
of the backbone and 10 times more moles of insert were added. The reaction also included 1 uL
of the T4 ligase (New England Biolabs Inc., MA), 2 uL of 10X T4 ligase buffer (New England
Biolabs Inc., MA), and enough volume of water to make up 20 uL. For the negative control,
samples with only backbones without the inserts were included and incubated in the same
23
condition. For complete ligation, the reaction ran at 16 °C overnight. Ligation product was then
purified using ZymoClean DNA Clean & Concentrator Kit (Genesee Scientific, San Diego, CA).

Transformation of DH10B Bacterial Cells and Plasmid Isolation
Of 6 uL of each eluted product from the concentration kit, 2 uL was pipetted into 50 uL
of DH10B electrocompetent cells for electroporation. Electroporation was done at 1.3 kV in the
electroporation cuvettes. The cells were incubated in 400 uL of low salt Luria broth (LB)
medium for 40 minutes for recovery. Then, the samples were plated on two low salt LB medium
agar plates (50 uL on one plate and 250 uL on the other) and incubated overnight (~16 hrs).
Colonies that grew on the plate were collected, diluted in 10 uL of ddH2O, and used for colony
PCR for 30 cycles. The PCR mix included 10 uL of 2X GoTaq Green Master Mix (Promega
Corporation, Madison, WI), 1 uL of 10 M forward primer, 1 uL of 10 M reverse primer, 1 uL of
the diluted bacterial sample, and enough volume of water to make up 20 uL. In this case, the
primers used for the second round of PCR earlier were used and the melting temperature of the
primers were used to adjust the temperature of the annealing step. Also, the temperature of the
extension step was also adjusted to 72 °C. The PCR products were then analyzed with 1 %
agarose gel. For the bacterial samples that showed the correct size bands on the gel, the leftover
diluted bacterial samples were incubated in 5 mL of low salt LB medium with zeocin antibiotics
added at the ratio of 1 in 4000 overnight. The next day, 500 uL of the 5 mL bacterial sample was
mixed with 500 uL of the 50 % glycerol to make glycerol stocks. Plasmids from the 4.5 mL of
the bacterial sample was isolated using ZymoPURE Plasmid Miniprep Kit (Genesee Scientific,
San Diego, CA), and sent out to GENEWIZ for sequencing.
24
Once verified for the sequence of the plasmid, glycerol stocks were inoculated in 5 mL of
low salt LB medium with zeocin once again for overnight incubation. Glycerol stocks for wild-
type scFv and Fab from the previous member of the lab were also inoculated. Then, the culture
was transferred to 150 mL of low salt LB medium with zeocin the next day for another overnight
incubation. Larger amounts of plasmids in 150 mL bacterial sample were isolated using
ZymoPURE II Plasmid Maxiprep Kit (Genesee Scientific, San Diego, CA).

2.2 Protein Expression and Purification
Expi293 Cell Transfection
In order to express each scFv and Fab antibodies, the plasmids isolated from maxiprep
were used to transfect Expi293 cells. Expi293 cells maintained in BalanCD HEK293 Medium
(Irvine Scientific, Irvine, CA) at 37 °C in passage 2 to 5 were used for the transfection. For
transfection, the cell density was first adjusted to be 2.5x10
6
cells/mL in 120 mL medium.
However, for this step, only half volume of the medium was added. Then, ExpiFectamine
TM

293/plasmid DNA complex solution, was added. In order to prepare this solution, 120 ug of the
plasmid was added to 6 mL of Opti-mem
TM
I Reduced Serum Medium (Thermo Fisher
Scientific, Houston, TX) and vortexed. Then 480 uL of polyethylenimine transfection reagent
(PEI) (Polysciences, Warrington, PA) was added in and vortexed again. Lastly, this solution was
incubated for 20 minutes at room temperature before added. Note that for scFv antibodies, 120
ug of the scFv plasmids were used while for the Fab antibodies, 60 ug of the heavy chain
plasmid and 60 ug of the light chain plasmid were used. For the Fab antibodies in this project, 60
ug of the UCHT1 Fab HC CD38 fusion plasmid and 60 ug of the wild-type UCHT1 LC plasmid
were mixed for the HC CD38 fusion, and 60 ug of the wild-type UCHT1 Fab HC plasmid and 60
25
ug of the UCHT1 Fab LC CD38 fusion plasmid were mixed for the LC CD38 fusion. Wild-type
scFv and Fab antibody samples were also made. After the addition of the transfection reagent,
the mammalian cell samples were incubated on the shaker at 37 °C for 2 hours. Then, the other
half portion of the cell medium was added to the flasks.

Protein Purification
Transfected cells were then incubated on the shaker at 37 °C for 5 days. Then, the cell
samples were centrifuged at 700 rpm for 10 minutes to collect the supernatant only. Then, the
supernatant was centrifuged at 4000 g for 30 minutes for further purification. Antibodies were
then isolated from the purified sample. scFv antibodies were isolated via Ni-NTA affinity
chromatography. Prior to sample loading to the column, the samples went through two rounds of
1000X dialysis against PBS at 4°C using 30 kDa dialysis bag with stirring (one 6hr dialysis and
one overnight dialysis) in order to get rid of EDTA. For the column, 1 mL of Ni-NTA resin was
added, washed with 10 mL of H2O, and equilibrated with 15 mL of equilibrium buffer (20 mM
Tris, 250 mM NaCl, 20 mM imidazole, pH 7.5). All the samples were loaded onto the column,
and the column was washed with 30 mL of wash buffer (20 mM Tris, 250 mM NaCl, 40 mM
imidazole, pH 7.5). Lastly, 10 mL of elution buffer (20 mM Tris, 250 mM NaCl, 400 mM
imidazole, pH 7.5) was used to elute the proteins. The eluate was then dialyzed again against
PBS twice at 4°C in order to remove high concentration of imidazole. After two rounds of
dialysis, the protein was concentrated to 1-2 mg/mL using 30 kDa cutoff filter concentrator.
Concentration of each protein was determined using Nanodrop (Thermo Fisher Scientific,
Houston, TX). The absorbance data at 280 nm measured by Nanodrop was divided by extinction
26
coefficient of each protein computed by ProtPram (See Table 4). The final products were
aliquoted, frozen using liquid nitrogen, and stored at -80 °C.
Fab antibodies were purified using protein G column. 1 mL of protein G resin was loaded
onto the column, and the column was washed with 10 mL of H2O. The resin was then
equilibrated with 5 mL of PBS. Then, without having to do dialysis, the samples were loaded
onto the column. 15 mL of PBS was used again to wash the column. Lastly, the protein was
eluted with 10 mL of 100 mM glycine (pH 2.7). The eluates were collected 1 mL by 1 mL in
separate tubes. Using nanodrop blanked with 100 mM glycine (pH 2.7), the protein concentration
was measured for each tube, and only the samples that had significant amount of protein were
neutralized with 100 uL of 1 M Tris buffer (pH 8). The neutralized samples were then
concentrated to 3-5 mg/mL using 30 kDa cutoff filter concentrator. The final products were
aliquoted, frozen using liquid nitrogen, and stored at -80 °C.

Expression Analysis via SDS-PAGE
Purified protein samples were confirmed for the expression via SDS-PAGE and
Coomassie blue staining. The samples were prepared with 5-20 ug protein, 4X LDS (1 in 4 total
volume), and 1 ug of 1 M DTT. The samples were boiled for 5 minutes at 98 °C for denaturation.
The samples were then loaded onto 4-12 % Bis-Tris, 15-well gels in MES buffer (GenScript,
Piscataway, NJ), and the electrophoresis was run at 140 V for 45 minutes. All Blue Protein
Marker II (BioLand Scientific LLC, Paramount, CA) was loaded as a ladder. The gel was then
placed in Coomassie blue (enough volume to submerge the gel) and microwaved for 1 minute.
The used Coomassie blue was then poured out into the proper waste bottle, and enough volume
of tap water was added with a piece of Kim Wipe. This was then microwaved for 3 minutes. This
27
step was repeated 2 more times with fresh water and a new piece of Kim Wipe. Lastly, the gel
was left on the shaker overnight for complete destaining. The corresponding size to each band
was analyzed for each protein sample.

2.3 Biological Activity Assay
Enzymatic Activity Assay of CD38
To confirm successful expression of the CD38 enzyme on the antibody, the enzymatic
activity of CD38 was assessed via nicotinamide guanine dinucleotide (NGD) fluorescence-based
assay. In this assay, CD38 catalyzes the formation of fluorescent cyclic GDP-ribose via ADP-
ribosyl cyclase reaction of NGD, and the fluorescent can be measured at wavelength of 300nm
for excitation and 410nm for emission. Therefore, 100 uL solution was prepared in 96-well
plates with each protein with final concentration of 100 nM and NGD with final concentration of
100 uM in 50 mM Tris (pH = 8.5). The fluorescence intensity was measure at 410 nm over 10
minutes by SynergyH1 (Agilent, Santa Clara, CA). 50 mM Tris (pH = 8.5) was used for negative
control, and recombinantly produced CD38 from Expi 293 cells by a previous member of the lab
was used for positive control.

Flow Cytometry
To confirm successful expression of the antibody, the binding affinity of each antibody
towards its target was assessed via flow cytometry. Jurkat cell line was used for UCHT1
antibodies that bind to CD3 on the surface of the Jurkat cells, while HCC1954 cell line was used
for Herceptin antibodies that bind to HER2 receptors on the surface of HCC1954 cells. For both
antibodies, MDA-MB-468 cell line was used as the negative control because this cell line lacks
28
both CD3 and HER2 receptors on its surface. Each cell line was seeded at appropriate seeding
density and maintained in RPMI 1640 media with 10 % FBS (Corning, Manassas, VA) until the
density reaches at the appropriate level (~3 million cells). Each cell line was washed with PBS
and suspended in PBS with 2 % FBS. It was then aliquoted into different tubes, making sure
each tube has 0.1-0.5 million cells suspended in ~500 uL of PBS with 2 % FBS. Then each
antibody with the same number of moles was added to the tube and incubated on ice for 30
minutes. Then the unconjugated antibodies were washed by adding 4 mL PBS with 2% FBS,
centrifuging at 500 xg for 5 minutes, and discarding supernatant. Then the cells were
resuspended in 500 uL PBS with 2 % FBS and the secondary antibodies were added. Then the
sample was incubated for 30 minutes, washed via centrifugation, and lastly incubated with
fluorescent conjugated antibodies. Antibodies used for each sample are noted in Table 3. The
fluorescent intensity of each sample was detected and analyzed with LSRFortessa X20 Cell
Analyzer (BD, San Jose, CA).

29
Table 3. Sample Preparation for Flow Cytometry
Sample
Cell Line
(density)
Primary antibody
(experimental)
Secondary
antibody
Fluorescent
conjugated
antibody
Untreated
HCC1954
N/A
Fluorescent
antibody only
N/A
6xHis Tag
Monoclonal
Antibody
(HIS.H8)
(Invitrogen,
Waltham, MA)

Goat anti-mouse
IgG [H+L] Alexa
Fluor
TM
Plus 488
(Invitrogen,
Waltham, MA)
Herceptin scFv
C-terminal fusion
Herceptin scFv-CD38
Herceptin scFv
N-terminal fusion
CD38-Herceptin scFv
Control
Herceptin-UCHT1
bispecific*
Untreated
MDA-MB-468
N/A
Fluorescent
antibody only
N/A
6xHis Tag
Monoclonal
Antibody
(HIS.H8)
(Invitrogen,
Waltham, MA)

Goat anti-mouse
IgG [H+L] Alexa
Fluor
TM
Plus 488
(Invitrogen,
Waltham, MA)
Herceptin scFv
C-terminal fusion
Herceptin scFv-CD38
Herceptin scFv
N-terminal fusion
CD38-Herceptin scFv
Control
Herceptin-UCHT1
bispecific*
Untreated
Jurkat cell
N/A
Fluorescent
antibody only
N/A
Anti-Ig light chain
k Mouse
Monoclonal
Antibody
(BioLegend, San
Diego, CA)
Goat anti-mouse
IgG [H+L] Alexa
Fluor
TM
Plus 488
(Invitrogen,
Waltham, MA)
UCHT1 Fab HC
fusion
UCHT1 Fab HC-
CD38
UCHT1 Fab LC
fusion
UCHT1 Fab LC-
CD38
UCHT1 WT UCHT1 Fab
Control N/A
Alexa Fluor® 488
anti-human CD3
(BioLegend, San
Diego, CA)
30
Untreated
MDA-MB-468
N/A
Fluorescent
antibody only
N/A
Anti-Ig light chain
k Mouse
Monoclonal
Antibody
(BioLegend, San
Diego, CA)
Goat anti-mouse
IgG [H+L] Alexa
Fluor
TM
Plus 488
(Invitrogen,
Waltham, MA)
UCHT1 Fab HC
fusion
UCHT1 Fab HC-
CD38
UCHT1 Fab LC
fusion
UCHT1 Fab LC-
CD38
UCHT1 WT N/A
Alexa Fluor® 488
anti-human CD3
(BioLegend, San
Diego, CA)
*Recombinantly produced Herceptin-UCHT1 bispecific antibody by a previous member of the
lab was used for positive control

31
2.4 Protein Conjugation
CD38 Inhibition with NAD
+
-derived Linkers
Two types of dinucleotide-based covalent inhibitors were synthesized by a senior
member in our lab: one with an azide functional group and the other with a DBCO functional
group (Figure 3). For CD38 linker conjugation, antibody and the linker were incubated in 50
mM Tris (pH = 8.5) at 1:100 = antibody:linker ratio at room temperature for 30 minutes. CD38
inhibition was measured via NGD assay (100 nM protein and 100 uM NGD in 100 uL). The
fluorescence intensity was measure at 410 nm over 10 minutes by SynergyH1 (Agilent, Santa
Clara, CA). Inhibition reaction was done with ~100 ug of antibody. Once the inhibition was
confirmed with the small scale, a larger scale inhibition was done with > 0.5 mg of antibody at
the same antibody to linker ratio and until complete inhibition was observed via NGD assay.
Then, buffer exchange was done against PBS via 30 kDa cut-off filter concentrator, and the
antibodies were concentrated to >1-2 mg/mL and stored in PBS at -80 °C. Unreacted linkers
filtered out from the reaction were repurified and concentrated by the senior member of the lab.

Click Reaction between Two Antibodies
Antibodies with CD38 conjugated with the linker were then reacted to form bispecific
antibody. In order to do so, each antibody stored in PBS were reacted keeping high
concentrations ( >1-2 mg/mL each antibody in the reaction mix). The molar ratio between two
antibodies were kept at 1:1. Then, the reaction mix was incubated at 37 °C for 2 hours. In order
to keep track of the reaction progress, ~20 ug of the protein was collected at different time and
analyzed for its size via SDS-PAGE (see Expression Analysis via SDS-PAGE for protocol).
Small scale conjugation was done with ~100 ug of each antibody.
32
Dot Blot Assay
To confirm that the linker is still functional when bound to CD38, biotinylating reaction
was done on CD38 fusions conjugated with the linkers. For antibodies conjugated with azide
linker, copper catalyzed reaction was required with an alkyne biotin. Therefore, a reaction mix
was prepared with 0.75 mM CuSO4, 1.5 mM THPTA, 7.75 mM sodium ascorbate, 300 uM of
alkyne biotin in PBS. Specifically, 1.66 uL of 90 mM CuSO4 was mixed with 3 uL of 100 mM
THPTA. Then 6.88 uL of 225 mM sodium ascorbate was added. Lastly, 182 uL PBS and then
the 6 uL of 10 mM alkyne biotin were added. 5 uL of this reaction mix was added to 10 uL of
protein stored in PBS at concentration of 3 uM, and this reaction was incubated at room
temperature for 2 hours. For DBCO linker sample, however, because this reaction does not
require copper catalyst, 10 uL of protein stored in PBS at concentration of 3 uM was mixed with
5 uL of 300 uM azide biotin and incubated at room temperature for 2 hours. Before loading the
biotinylated samples on to the membrane, PVDF membrane was prepared with 15-second
methanol activation, 2-minute H2O wash, 5-minute PBS-T (0.05% Tween 20) equilibration.
Then, the samples were loaded in 3-4 mm diameter avoiding touching each other. The membrane
was dried until no liquid on the membrane was visible (~30 min). Then the membrane was
blocked with 5 % BSA in PBS-T for an hour at room temperature. The 5 % BSA was washed
with PBS-T 3 times. Then, the membrane was incubated for 30 minutes with Streptavidin-HRP
antibodies (R&D systems, Minneapolis, MN) diluted in 5 % BSA at volume ratio of 1:400. The
excess of antibodies was washed with PBS-T for 3 times. HRP chemiluminescent substrate
(ThermoFisher Scientific, MA) was then applied evenly on the membrane, and the signals were
developed and captured via ChemiDoc
TM
touch imaging system (Bio-Rad, Irvine, CA).
33
Chapter 3: Results
3.1 Molecular Cloning
The constructs developed for this project including the ones given by the previous
member of the lab are organized in Figures 4a-h. For scFv antibodies, gene sequence that
encodes CD38 enzyme was inserted in either C- or N-terminus of the sequence that encodes each
antibody as shown in Scheme a-d. His6 tags were also inserted on the C-terminus of the whole
sequence for future purification convenience. On the other hand, Fab antibody constructs had
CD38 encoding gene sequence on C-terminus of either heavy chain or light chain of the antibody
because it was discovered from the previous research that CD38 fusion on the N-terminus of the
Fab antibodies can interrupt the binding of the antibody to its target. Each construct was prepared
via overlap extension PCR, and the product was analyzed via agarose gel electrophoresis for its
size and purity. The DNA gel electrophoresis analysis of overlap extension PCR product of
Herceptin Fab HC CD38 fusion and LC CD38 fusion is depicted in Figure 5. The bands of the
expected sizes were cut out and the DNA was recovered. The DNA was then digested and
ligated into pFuse vectors. The ligation product was then used to transfect DH10B bacterial cells
which were plated on agar plates and analyzed using colony PCR. Colony PCR results of scFv
fusions are depicted in Figure 6. Some of the colonies with the right size band were inoculated
again in the media and miniprepped for sequencing. The sequencing result was then compared
with the vector map for successful construct development. Once confirmed with its sequence, the
colonies with the right plasmid were inoculated and maxi-prepped for larger number of plasmids.

34

Figure 4a. Herceptin scFv C terminal Fusion Construct (given by a previous lab member)

Figure 4b. Herceptin scFv N-terminal Fusion Construct

Figure 4c. UCHT1 scFv C-terminal Fusion Construct (given by a previous lab member)

Figure 4d. UCHT1 scFv N-terminal Fusion Construct (given by a previous lab member)

Figure 4e. UCHT1 Fab Light Chain C-terminal Fusion Construct

Figure 4f. UCHT1 Fab Heavy Chain C-terminal Fusion Construct

Figure 4g. Herceptin Fab Light Chain C-terminal fusion Construct (prepared by another lab
member)

Figure 4h. Herceptin Fab Heavy Chain C-terminal fusion Construct (prepared by another
lab member)

Herceptin scFv CD38
GGGGS
His
Herceptin scFv CD38
His
GGGGS
UCHT1 scFv CD38
His
GGGGS
UCHT1 scFv CD38
His
GGGGS
UCHT1 LC CD38
GGGGS
UCHT1 Fab HC hIg1-Fc2 CD38
GGGGS
Herceptin Fab HC hIg1-Fc2 CD38
GGGGS
Herceptin LC CD38
GGGGS
35

Figure 5. Overlap Extension PCR Products of anti-CD3-CD38 Fusions. After the second
round of PCR, the PCR products for both UCHT1 Fab HC CD38 fusion and LC CD38 fusion
were analyzed by agarose gel electrophoresis. The expected sizes of UCHT1 HC and LC fusions
are 1473bp and 1428bp respectively. Lanes 1 & 2 correspond to heavy chain fusion, and lanes 3
& 4 correspond to the light chain fusion. The bands of the correct size were cut out, and the
DNA inside were recovered for digestion and ligation.

3.0
1.0
2.0
1.2
1.5
Lane 1 Lane 2 Lane 3 Lane 4
444
36

Figure 6. Colony PCR of DH10B Cells Transformed with pFuse_CD38-Herceptin scFv. The
expected size of Herceptin scFv N-terminal CD38 fusion was 1533 bp. All colonies except for
the one on the bottom 3
rd
lane showed bands of correct size. Colonies tested in upper lane 3 & 4
and bottom lane 5 were sent out with its primers for sequencing.
1 2 3 4 5 6
1 2 3 4 5 6 7
37
3.2 Antibody Expression and Purification
Mammalian cells were then transfected with plasmids isolated from maxiprep in order to
express each antibody CD38 fusion. The culture medium was collected 5 days after the
transfection for antibody purification. ScFv antibodies were purified using Ni-NTA column,
while Fab antibodies were purified using Protein G column. Average yield of each scFv antibody
from 240 mL of cell was ~0.5mg, while the average yield of each Fab antibody from 240 mL of
cell was ~4mg. SDS-PAGE with Coomassie blue staining confirmed the expression and purity of
each engineered antibody fusions. Figure 7 shows that each fusion was expressed successfully
with high purity.
The activity of the CD38 enzymes was confirmed via NGD assay. Figures 8a,b show the
increase of fluorescence intensity overtime due to cyclase activity of CD38 with its substrate,
NGD. Extensive activity of CD38 was confirmed in these figures since they show that all NGD
was used up in ~3 minutes after the exposure for scFv antibody fusions and ~1 minute after the
exposure for Fab antibody fusions.
38

Figure 7. SDS-PAGE Analysis of Each Antibody Fusion. SDS-PAGE analysis confirmed that
the right sized proteins were purified for all antibody fusions. a: CD38-Herceptin scFv antibody
(N-terminal fusion); b: Herceptin-CD38 scFv antibody (C-terminal fusion); c: CD3-CD38 scFv
antibody (C-terminal fusion); d: CD38-CD3 scFv antibody (N-terminal fusion); e: wild-type
Herceptin Fab antibody; f: UCHT1 Fab heavy-chain CD38 fusion; g: UCHT1 Fab light-chain
CD38 fusion; h: UCHT1 Fab heavy-chain CD38 fusion without DTT added; i: Herceptin Fab-
light chain CD38 without DTT added.

180
140
100
72
60
45
35
25
15
10
a b c d e f g h i
140
100
72
60
41
35
25
15
10
MW
(kDa)
MW
(kDa)
39

Table 4. Expected Molecular Weight of Each Antibody Fusion
Antibody Molecular Weight (kDa) Extinction Coefficient
(M
-1
cm
-1
)
Herceptin scFv – CD38 57.09 1.79
UCHT1 scFv – CD38 56.78 1.90
Herceptin Fab Wild type 46.64 1.50
Herceptin Fab HC-CD38 77.70 1.70
Herceptin Fab LC-CD38 76.57 1.75
UCHT1 Fab Wild type 47.23 1.67
UCHT1 Fab HC-CD38 78.29 1.67
UCHT1 Fab LC-CD38 77.16 1.70

40

Figure 8a, b. NGD Assay Analysis of CD38 activity. For NGD assays, the flouresence
intensity was measured at 410nm over 10 minutes. Figure 8a compares the CD38 activity of
Herceptin scFv C- and N-terminal fusions, while Figure 8b compares the CD38 activity of
UCHT1 Fab heavy and light chain fusions.
-500
0
500
1000
1500
2000
2500
3000
0:00:00 0:01:26 0:02:53 0:04:19 0:05:46 0:07:12 0:08:38 0:10:05
Fluorescence Intensity
Time (h:mm:ss)
NGD Assay (HERscFv fusions)
Tris Buffer Recombinant CD38
HERscFv C-terminal CD38 fusion HERscFv N-terminal CD38 fusion
-500
0
500
1000
1500
2000
2500
3000
3500
0:00:00 0:01:26 0:02:53 0:04:19 0:05:46 0:07:12 0:08:38 0:10:05
Fluorescence Intensity
Time (h:mm:ss)
NGD Assay (UCHT1 fab fusions)
Tris Buffer Recombinant CD38 UCHT1 Fab HC CD38 UCHT1 Fab LC CD38
a
b
41
3.3 Biological Activity Assay
NGD assay showed that CD38 activity is not a good parameter in terms of determining
which fusion would be better for bispecific antibody conjugation since both C- and N-terminal
fusions of scFv antibodies and both HC and LC fusions of Fab antibodies showed similar CD38
activity. Therefore, affinity of each antibody to its target was compared using flow cytometry.
Figures 9a-d show fluorescence at 488nm shifted for samples with antibodies compared to the
negative controls (untreated and fluorescent antibody only samples). Figures 9a,b compare
binding affinity of Herceptin scFv C- and N-terminal fusions to HER2 receptor on HCC1954
cells. Since the fluorescence peak shifted further for the C-terminal fusion, the binding affinity
was determined to be higher for C-terminal fusion compared to the N-terminal fusion. Therefore,
the C-terminal fusion of Herceptin scFv was chosen to be used for conjugation in the next step.
Also, UCHT1 scFv C-terminal fusion was shown to have higher binding affinity to CD3 on
Jurkat cells according to a previous member of the lab and determined to be conjugated with
Herceptin scFv fusion. On the other hand, Figures 9c,d compare binding affinity of UCHT1 Fab
HC and LC fusions to CD3 on Jurkat cells. In the figures, HC fusion showed binding affinity to
CD3 that is comparable to that of the wild-type, comfirming expressing CD38 on the C-terminal
of heavy chain does not interrupt the binding between the UCHT1 Fab antibody and CD3.
Therefore, The HC fusion was chosen to be used for further conjugation. Another member of the
lab who works on antibody drug conjugate project with Herceptin Fab-CD38 fusions compared
the binding affinity of HC and LC fusions and determined that the LC fusion of Herceptin Fab
had higher binding affinity towards HER2 receptor via ELISA. Therefore, with her data,
Herceptin Fab LC fusion was determined to be conjugated with the UCHT1 Fab HC-CD38
fusion.
42

Figure 9a-d. Flow Cytometry Histograms for Antibody Fusions. Figures 9a,b compare the
affinity of Herceptin scFv fusions to HER2 receptors, while Figures 9c,d compare the affinity of
UCHT1 Fab fusions to CD3. Figures 8a, b show C-terminal fusion of Herceptin scFv had
slightly more shift in the peak, while Figures 8c,d show heavy chain fusion of UCHT1 Fab had
slightly more shift in the peak.

a b
c
d
43
3.4 Protein Conjugation
The first step of antibody conjugation was to conjugate the linkers with CD38 enzymes
expressed on antibodies. In order to confirm the conjugation of all CD38 enzymes with excessive
amounts of linker, NGD assay was done after 30-minute incubation with the linker. Before
moving on to the large scale inhibition, small amount of each antibody (~100 ug) was incubated
with its linker. In this project, both UCHT1 scFv and Fab fusions were conjugated with azide
linkers, and both Herceptin scFv and Fab fusions were conjugated with DBCO linkers.
According to Figures 10a and b, 30-minute incubation yielded full inhibtion of CD38,
confirming the conjugation between CD38 with its ligands. After the confirmation that the batch
of the linker produced can be conjugated with CD38, a larger scale inhibition was done with
>0.5mg of antibodies.
Now that each fusion has its CD38 enzyme conjugated with linkers, the antibodies with
their linkers were conjugated to each other. The progress of conjugation reaction was analyzed
using SDS-PAGE over 54 hours. However, the conjugation did not seem to be working as
depicted in Figure 11. Since denaturation of protein was detected in Figure 11, in order to see if
the temperature is the issue, the stability of each fusion at 37°C overnight was tested using SDS-
PAGE in Figure 12. However, no denaturation was detected in Figure 11. Then, in order to
check the functionality of each linker conjugated to CD38 on antibodies, each linker conjugated
antibody was biotinylated and assessed for the signal using dot bot assay (Figure 13). Since no
signal was detected for Herceptin scFv conjugated with DBCO linkers, mass spectrometry was
done by a senior member of the lab to figure out what might have gone wrong. From the mass
spectrometry result, the peak corresponding for DBCO linker was not detected, indicating that it
was degraded.
44

Figure 10a, b. NGD Assay Analysis for Inhibition of CD38 fused with scFv and Fab.
Inhibition of CD38 was analyzed using NGD assay. Figure 10a demonstrates the inhibition of
CD38 activity of two scFv fusions conjugated with its linker while Figure 10b demonstrates the
inhibition of CD38 activity of two Fab fusions conjugated with its linker after 30-minute
reaction.

-1000
0
1000
2000
3000
4000
5000
0:00:00 0:01:26 0:02:53 0:04:19 0:05:46 0:07:12 0:08:38 0:10:05 0:11:31
Fluorescent Intensity
Time (h:mm:ss)
NGD assay (100 nM Protein)
UCHT1 scFv + DBCO linker HER scFv + azide linker HER scFv UCHT1 scFv
-500
0
500
1000
1500
2000
2500
0:00:00 0:01:26 0:02:53 0:04:19 0:05:46 0:07:12 0:08:38 0:10:05 0:11:31
Fluorescent Intensity
Time (h:mm:ss)
NGD Assay (100 nM Protein)
Tris UCHT1 Fab HER Fab UCHT1 Fab + azide linker HER fab +DBCO linker
45

Figure 11. SDS-PAGE Analysis of scFv Conjugation. Each lane represents the following: a)
Herceptin scFv-CD38 fusion; b) UCHT1scFv-CD38 fusion; c) 0hr conjugation; d) 16hr
conjugation; e) 26hr conjugation; f) 46hr conjugation; g) 54hr conjugation. No conjugation was
observed.

140
100
72
60
41
35
25
a b c d e f g
MW
(kDa)
46

Figure 12. Protein Stability Analysis via SDS-PAGE Gel Electrophoresis. Antibody fusions
were incubated at 37°C for 24 hours in order to check the stability of the proteins. Each lane
represents the following: a) UCHT1 Fab HC-CD38 fusion at t=0h; b) UCHT1 scFv-CD38 fusion
at t=0h; c) Herceptin scFv-CD38 fusion at t=0h; d) UCHT1 Fab HC-CD38 fusion at t=24h; e)
UCHT1 scFv-CD38 fusion at t=24h; f) Herceptin scFv-CD38 fusion at t=24h.

140
100
72
60
41
35
25
MW
(kDa)
a b c d e f
47

Figure 13. Dot Blot Assay for Linker Conjugated Antibodies. Each dot represents the
following: A) UCHT1scFv-CD38 fusion without linker conjugation biotinylated with alkyne
biotin; B) UCHT1 scFv-CD38 fusion conjugated with azide linkers and biotinylated with alkyne
biotin; C) Herceptin scFv-CD38 fusion without linker conjugation biotinylated with azide biotin;
D) Herceptin scFv-CD38 fusion conjugated with DBCO linker and biotinylated with azide
bition. No signals were detected for Herceptin scFv-CD38 conjugated with DBCO linkers.

A B C D
48
Chapter 4: Discussion
In contrast to the conventional monoclonal antibodies, it is harder to produce bispecific
antibodies because it is challenging to have high yields of homogeneous batches of desired
antibody fusions and to design antibodies that are still functional after being fused with another
antibody. This fact has hindered the development and the translation of bispecific antibodies into
clinical trials until very recent years. Even though the development of bispecific antibodies has
been accelerated recently, there is still a demand for more efficient bispecific antibody
production methods. The aim of this project was to develop an antibody conjugation method by
expressing an enzyme on antibodies and inhibiting the enzyme with its ligand. If successful, this
method is expected to result in high yields of homogeneous batch of antibodies since the
enzymes are expressed using genetic means at designated locations on the antibody. Also,
extensive application of this method for producing many different formats of the bispecific
antibodies is expected with the potential in modifying the chemically synthesized linkers.
The project started with designing the constructs where CD38 enzymes are expressed on
different sites of antibodies. Since this project aimed to use the chemo-enzymatic conjugation for
both Fab and scFv versions of antibodies, for scFv antibodies, CD38 was expressed on either C
or N-terminus of them, while it was expressed on the C terminus of either heavy chain or light
chain of Fab antibodies. The final bispecific antibody will then be an immunotherapeutic
bispecific antibody that recruits T cells in proximity to the breast cancer cell. Therefore, an anti-
CD3 antibody, UCHT1 antibody in both scFv and Fab formats and an anti-HER2 antibody,
Herceptin antibody in both scFv and Fab formats were expressed with CD38. Each construct was
inserted into the vector and used to transfect mammalian cells for protein expression. The
expressed protein was then purified using corresponding method: His6-tag affinity
49
chromatography for scFv antibodies and protein-G affinity chromatography for Fab antibodies.
In comparison to the wild-type antibody, CD38 fusions were collected at a comparable yield,
indicating that fusing antibodies with CD38 does not affect the yield of antibodies. Coomassie
blue stained SDS-PAGE gel in Figure 7 confirmed the expression since each band was shown at
expected size of each protein fragment. For examples, for UCHT1 Fab antibodies that were
denatured using DTT before the gel electrophoresis, it was clearly shown that for heavy chain
CD38 fusion, the band corresponding the heavy chain shifted up (became heavier), and for light
chain CD38 fusion, the band corresponding to the light chain shifted up. And the fact that bands
with random sizes were not stained confirmed the purity of the antibody after the purification.
In order to select the optimal CD38 fusion for conjugation, the activity of CD38 on the
antibody fusions and binding affinity of the antibodies were compared. In Figures 8a and 8b,
the NGD assay was done to compare the activity of CD38 on different fusions. It turned out that
the CD38 fused with antibodies had higher activity compared to the recombinantly produced
CD38 proteins. However, in terms of the comparison, since both fusions seemed to react with all
the NGD and had the plateau of fluorescence intensity roughly at the same time (~3 minutes post
reaction), it was determined that expression of CD38 at different locations does not affect the
activity of CD38. Lastly, to check if CD38 expression location affects the binding affinity of the
antibodies, the binding affinity to the cells expressing each antibody’s target was measured via
flow cytometry. For UCHT1 antibodies, the affinity was measured with Jurkat cell line, which
express CD3 on its surface, while for anti-HER2 antibodies, HCC1954 cell line was used since it
expresses a high level of HER2 receptors on its surface. However, for the negative cell line,
MDA-MB-468 cell line was used since it lacked the expression of both HER2 and CD3 on its
surface. As a result, Figure 9a showed that Herceptin scFv-CD38 C-terminal fusion had higher
50
binding affinity to HER2 receptors on HCC 1954 cell line even though it was not as good as
recombinantly produced Herceptin-UCHT1 bispecific antibody. Heavy chain CD38 fusion of
UCHT1 Fab antibody, however, had binding affinity as strong as the wild-type antibody, which
was slightly better than the light chain fusion according to Figure 9c. Since the yield of the
antibody production from the mammalian cells and the CD38 activity of the antibodies did not
show significant difference depending on the structure, the binding affinity became the main
criteria in terms of which fusion to be used for conjugation. Therefore, Herceptin scFv-CD38 C-
terminal fusion was selected over the N-terminal fusion, and UCHT1 Fab-CD38 heavy chain
fusion was selected over the light chain fusion. For UCHT1 scFv, the C-terminal fusion was
selected over the N-terminal fusion since it showed much higher binding affinity towards CD3
on Jurkat cells according to the previous lab member who worked on a relevant project. Lastly,
for Herceptin Fab antibodies, light chain fusion was selected over the heavy chain fusion for its
higher binding affinity towards recombinant HER2 from ELISA assay done by another member
of the lab who also works on Herceptin Fab CD38 fusions.
Now the CD38 enzymes expressed on the C-terminus of Herceptin scFv and UCHT1
scFv, the C-terminus of heavy chain of UCHT1 Fab, and that of the light chain of Herceptin
were then conjugated with the linkers. For simplicity, azide linkers were used for both scFv and
Fab formats of UCHT1 antibodies, and DBCO linkers were used for both scFv and Fab formats
of Herceptin antibodies. After 30-minute conjugation reaction at room temperature, the
inhibition of CD38 was confirmed using NGD assay. Figures 10a and 10b shows that the 30-
minute conjugation successfully inhibited CD38 enzymatic activity. The unreacted linkers were
then removed then the antibody conjugation was done at 1:1 molar ratio for several days. Some
portions of scFv conjugation reaction were taken out time to time and analyzed by SDS-PAGE
51
Coomassie blue staining to keep track of the progression of the conjugation reaction. However,
according to the gel image in Figure 11, the conjugation was not successful since all the proteins
seemed to be denatured after 24 hours of reaction. Therefore, in order to confirm that it’s not the
temperature that makes the protein denatured, each antibody was incubated at 37°C for 24 hours
and checked for the stability in Figure 12. Since the SDS-PAGE gel analysis in Figure 12
showed that the antibodies are stable at 37°C for 24 hours, the functionality of the linkers when
bound to CD38 was measured using dot blot assay in Figure 13. The conjugated antibodies are
supposed to show signal after biotinylating reaction if still functional. However, no signals were
detected for Herceptin scFv conjugated with DBCO linker. Therefore, a senior member of the lab
did a mass-spectrometry analysis for DBCO linker sample. It turned out that the peak of DBCO
functional group was not shown on the mass-spectrometry result, indicating that this linker might
have been denatured either during the concentration process or during storage since the mass
spectrometry analysis confirmed the presence of DBCO group when measured before it was
concentrated down. The denatured pieces of DBCO linker might have affected the stability of the
Herceptin antibodies and the unstable form of anti-HER2 antibody might have accelerated the
denaturation of its partner antibody during the conjugation. This also explains the extra band
detected for Herceptin scFv in Figure 11 in Lanes a and c. Therefore, in the future studies,
another batch of DBCO linker will be synthesized with extra care, and the stability of the linker
will be analyzed using the mass spectrometry both before and after the concentration step. Once
it was confirmed that the new batch of DBCO linkers is stable and ready to be used, the
conjugation will be done once more in a small scale (~100 ug) and in a larger scale (>0.5 mg). If
the conjugation is successful in the future trials, then it will finally be used for cellular and
animal studies, where breast cancer cell lines and mice with implanted breast cancer get treated
52
with the bispecific antibodies prepared via this novel chemo-enzymatic conjugation method. The
efficacy and specificity of the generated bispecific antibodies will be compared with the
recombinantly produced UCHT1-Herceptin scFv bispecific antibodies. Once the viability of the
conjugation method and the efficacy and specificity of the final products are confirmed, this
method is expected to be a new and efficient conjugation method with high yield of
homogeneous batch of bispecific antibodies to be used for many different applications. Also,
there is a potential for linker modifications which will allow for extensive application of this
method for many formats of bispecific antibody production.

53
References
Agard, N. J., Prescher, J. A., & Bertozzi, C. R. (2004). A strain-promoted [3 + 2] azide-alkyne
cycloaddition for covalent modification of biomolecules in living systems. Journal of the
American Chemical Society, 126(46), 15046–15047.

Blanco, B., Domínguez-Alonso, C., & Alvarez-Vallina, L. (2021). Bispecific
Immunomodulatory Antibodies for Cancer Immunotherapy. Clinical cancer research :
an official journal of the American Association for Cancer Research, 27(20), 5457–5464.

Borghaei, H., Smith, M. R., & Campbell, K. S. (2009). Immunotherapy of cancer. European
journal of pharmacology, 625(1-3), 41–54.

Budde E. Optimizing lymphodepletion prior to CAR T cell therapy. Eight annual meeting of
SOHO; Sept 11, 2020

Burnet, M. (1957) Cancer—a biological approach. Br Med J. 1(5023): 841-847.

Burnet, M. (1970) The concept of immunological surveillance. Prog Exp Tumor Res 13:1-27 doi:
10.1159/000386035.

Cintolo, J. A., Datta, J., Mathew, S. J., & Czerniecki, B. J. (2012). Dendritic cell-based vaccines:
barriers and opportunities. Future oncology (London, England), 8(10), 1273–1299.

Dahlén, E., Veitonmäki, N., & Norlén, P. (2018). Bispecific antibodies in cancer
immunotherapy. Therapeutic advances in vaccines and immunotherapy, 6(1), 3–17.

Dai, Z., Zhang, X. N., Nasertorabi, F., Cheng, Q., Li, J., Katz, B. B., Smbatyan, G., Pei, H.,
Louie, S. G., Lenz, H. J., Stevens, R. C., & Zhang, Y. (2020). Synthesis of site-specific
antibody-drug conjugates by ADP-ribosyl cyclases. Science advances, 6(23), eaba6752.

Dimasi, N., Kumar, A., Gao, C. (2021). Generation of bispecific antibodies using chemical
conjugation methods, Drug Discovery Today: Technologies, 40, 13-24,

Duhen, R., Ballesteros-Merino, C., Frye, A. K., Tran, E., Rajamanickam, V., Chang, S. C.,
Koguchi, Y., Bifulco, C. B., Bernard, B., Leidner, R. S., Curti, B. D., Fox, B. A., Urba,
W. J., Bell, R. B., & Weinberg, A. D. (2021). Neoadjuvant anti-OX40 (MEDI6469)
therapy in patients with head and neck squamous cell carcinoma activates and expands
antigen-specific tumor-infiltrating T cells. Nature communications, 12(1), 1047.

Ehrlich, P. (1909) Ueber den jetzigen stand der karzinomforschung. Ned Tijdschr Geneeskd 5

Gross, L. (1943) Intradermal immunisation of C3H mice against a sarcoma that origintated in an
animal of the same line. Cancer Res 3:326
54
Kennedy, A, Solga, M., Schuman, T., Chi, A., Lindorfer, M., Sutherland, W., Foley, P., Taylor,
R. (2003). An anti-C3b(i) mAb enhances complement activation, C3b(i) deposition, and
killing of CD20+ cells by rituximab. Blood, 101(3) 1071-1079

Kim, C. H., Axup, J. Y., Dubrovska, A., et al. (2012). Synthesis of bispecific antibodies using
genetically encoded unnatural amino acids. Journal of the American Chemical
Society, 134(24), 9918–9921.

Kontermann R. E. (2005). Recombinant bispecific antibodies for cancer therapy. Acta
pharmacologica Sinica, 26(1), 1–9.

Lee, R. J., Fang, Q., Davol, P. A., et al. (2007). Antibody targeting of stem cells to infarcted
myocardium. Stem cells (Dayton, Ohio), 25(3), 712–717.

Liu, H., Saxena, A., Sidhu, S., Wu, D., (2017) Fc Engineering for Developing Therapeutic
Bispecific Antibodies and Novel Scaffolds. Frontiers in Immunology.

Liu, S., Galat, V., Galat4, Y. et al. (2021). NK cell-based cancer immunotherapy: from basic
biology to clinical development. J Hematol Oncol, 14(7).

Malavasi, F., Deaglio, S., Funaro, A., Ferrero, E., Horenstein, A. L., Ortolan, E., Vaisitti, T., &
Aydin, S. (2008). Evolution and function of the ADP ribosyl cyclase/CD38 gene family
in physiology and pathology. Physiological reviews, 88(3), 841–886.

Malmberg K. J. (2004). Effective immunotherapy against cancer: a question of overcoming
immune suppression and immune escape?. Cancer immunology, immunotherapy :
CII, 53(10), 879–892.

Morandi, F., Horenstein, A. L., Costa, F., Giuliani, N., Pistoia, V., & Malavasi, F. (2018). CD38:
A Target for Immunotherapeutic Approaches in Multiple Myeloma. Frontiers in
immunology, 9, 2722.

Muller-Steffner, H. M., Malver, O., Hosie, L., Oppenheimer, N. J., & Schuber, F. (1992). Slow-
binding inhibition of NAD
+
glycohydrolase by arabino analogues of beta-NAD. The
Journal of biological chemistry, 267(14), 9606–9611.

Nisonoff, A., Wissler, F., & Lipman, L. (1960). Properties of the major component of a peptic
digest of rabbit antibody. Science (New York, N.Y.), 132(3441), 1770–1771.

Przepiorka, D., Ko, C. W., Deisseroth, A., et al. (2015). FDA Approval: Blinatumomab. Clinical
cancer research : an official journal of the American Association for Cancer
Research, 21(18), 4035–4039.

55
Ruggeri, L., Capanni, M., Urbani, E., Perruccio, K., Shlomchik, W. D., Tosti, A., Posati, S.,
Rogaia, D., Frassoni, F., Aversa, F., Martelli, M. F., & Velardi, A. (2002). Effectiveness
of donor natural killer cell alloreactivity in mismatched hematopoietic
transplants. Science (New York, N.Y.), 295(5562), 2097–2100.

Rygaard, J., Povlsen, C., (1974) The mouse mutant nude does not develop spontaneous tumours:
an argument against immunological surveillance. Acta Pathol Microbiol Scand [B]
Microbiol Immunol 82:99

Thomas, L. (1959) Cellular and Humoral Aspects of the Hypersensitive States: A Symposium at
the New York Academy of Medicine. Jama Network. 170(7):883.

Vijay, K. (2018). Toll-like receptors in immunity and inflammatory diseases: Past, present, and
future. International immunopharmacology, 59, 391–412.

Wang, S., Chen, K., Lei, Q., et al. (2021). The state of the art of bispecific antibodies for treating
human malignancies. EMBO molecular medicine, 13(9), e14291.

Zahavi, D., AlDeghaither, D., O’Connell, A., Weiner, L. (2018). Enhancing antibody-dependent
cell-mediated cytotoxicity: a strategy for improving antibody-based
immunotherapy, Antibody Therapeutics, 1(1), 7–12.

Abstract (if available)

Abstract One of the challenges that hinders bispecific antibodies from getting approval for its use in clinical trials is the production in terms of quality, quantity, and stability. There have been extensive studies trying to develop production methods that could solve the problems, but some methods result in heterogeneous pool of products with batch-to-batch variations, low yields, and fluctuations of efficiency. In this study, we tried to develop a chemo-enzymatic conjugation method for bispecific antibody production that can result in homogeneous batch of products with constant efficiency in every batch as well as potentially improved pharmacological activities. This chemo-enzymatic conjugation involves an expression of an enzyme called CD38 on designated sites of either Fab or scFv antibodies. CD38 is a membrane enzyme expressed on immune cells that has both ADP-ribosyl (ADPR) cyclase and cADPR hydrolase activities. However, it can also be covalently inhibited by NAD+ analogs. Using this characteristic, two types of functionalized NAD+ analogs were synthesized: one with an azide functional group and another with a DBCO functional group. These linkers were then conjugated with both scFv and Fab format antibodies via enzyme inhibition. The two linkers are expected to go through a click reaction and facilitate conjugation of two antibodies. In this thesis so far, functional antibody CD38 fusion expressions have been done and confirmed via different functional assays. However, one of the NAD+ analogs had a stability issue, which disabled the conjugation of the antibodies. Once this stability issue is resolved, the developed conjugation approach is expected to allow for successful production of functional bispecific antibodies that can recruit effector T cells in proximity to cancer cells and redirect the immune system toward fighting the tumors.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Development of engineered antibodies as novel anti-cancer agents

PDF

Development of cyclized bispecific antibodies for enhanced serum stability

PDF

Development of new approaches for antibody modification

PDF

Employing engineered exosomes for combating colon cancer and engineering CD38 as an optimized drug carrier

PDF

Development of site-specific anti-HER2 Fab antibody-drug conjugates

PDF

Biochemical development and analysis of NAD⁺-related biomolecules

PDF

Investigating CD99 as a therapeutic target in acute myeloid leukemia

PDF

Chimeric Antigen Receptor targeting Prostate Specific Membrane Antigen (PSMA)

PDF

Protein ADP-ribosylation: from biochemical characterization to therapeutic applications

PDF

FLT3/CD99 bispecific antibody-based nanoparticles (BiAbs) for acute myeloid leukemia

PDF

Engineered exosomes for immunotherapy

PDF

Site-specific dinitrophenylation of single-chain antibody fragments for redirecting a universal CAR-T cell against cancer antigens

PDF

The role of CD99 in T cells

PDF

Investigating the effects of targeting CD99 on T cells to enhance their antileukemia activity

PDF

Immunotherapy of cancer

PDF

Antibodies and elastin-like polypeptides: cellular and biophysical characterization of an anti-ELP monoclonal and an anti-CD3 single-chain-ELP fusion

PDF

Reprogramming exosomes for immunotherapy of acute myeloid leukemia

PDF

Genome engineering of filamentous fungi for efficient novel molecule production

PDF

NMI: a near infrared conjugated MAO-A inhibitor as a novel targeted therapy for colorectal and other cancers

PDF

Rational design of monoclonal antibody inhibitors targeting human cathepsin B

Asset Metadata

Creator Kim, Hyo Sun (author)

Core Title A novel chemoenzymatic conjugation method for bispecific antibody production

School School of Pharmacy

Degree Master of Science

Degree Program Pharmaceutical Sciences

Degree Conferral Date 2022-05

Publication Date 04/17/2024

Defense Date 05/13/2022

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

Tag antibody conjugation,antibody engineering,antibody production,bispecific antibody,cancer therapy,click chemistry,immunotherapy,OAI-PMH Harvest,protein engineering

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Zhang, Yong (committee chair), Haworth, Ian (committee member), Romero, Rebecca (committee member)

Creator Email hyosunki@usc.edu,hyosunnykim2@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC110964851

Unique identifier UC110964851

Document Type Thesis

Format application/pdf (imt)

Rights Kim, Hyo Sun

Type texts

Source 20220417-usctheses-batch-927 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

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

Repository Name University of Southern California Digital Library

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

Repository Email cisadmin@lib.usc.edu