Yale University
EliScholar – A Digital Platform for Scholarly Publishing at Yale
Yale Medicine Thesis Digital Library
School of Medicine
January 2019
The Application Of Extracorporeal
Photochemotherapy To Head And Neck
Squamous Cell Carcinoma
Alp Yurter
Follow this and additional works at: https://elischolar.library.yale.edu/ymtdl
This Open Access Thesis is brought to you for free and open access by the School of Medicine at EliScholar – A Digital Platform for Scholarly
Publishing at Yale. It has been accepted for inclusion in Yale Medicine Thesis Digital Library by an authorized administrator of EliScholar – A Digital
Platform for Scholarly Publishing at Yale. For more information, please contact elischolar@yale.edu.
Recommended Citation
Yurter, Alp, “The Application Of Extracorporeal Photochemotherapy To Head And Neck Squamous Cell Carcinoma” (2019). Yale
Medicine Thesis Digital Library. 3544.
https://elischolar.library.yale.edu/ymtdl/3544
The Application of Extracorporeal Photochemotherapy to
Head and Neck Squamous Cell Carcinoma
A Thesis Submitted to the
Yale University School of Medicine
in Partial Fulfillment of the Requirements for the
Degree of Doctor of Medicine
by
Alp Yurter
Graduating Class of 2019
TABLE OF CONTENTS
INTRODUCTION ………………………………………………………………………………………………………………………
1
ECP Discovery …………………………………………………………………………………………………………………………
1
ECP’s Mechanism of Action
………………………………………………………………………………………………………
1
ECP’s Evolution ……………………………………………………………………………………………………………………….
4
Potential Application to Head and Neck Squamous Cell Carcinoma ……………………………………………….
5
STATEMENT OF PURPOSE ……………………………………………………………………………………………………..
8
Specific Aims …………………………………………………………………………………………………………………………..
8
Hypothesis ……………………………………………………………………………………………………………………………….
8
MATERIALS AND METHODS ………………………………………………………………………………………………….
9
HPV16 E7 Antigen Sources ……………………………………………………………………………………………………….
9
Peripheral Blood Mononuclear Cells (PBMCs)
…………………………………………………………………………..
10
CD8 T Cells
……………………………………………………………………………………………………………………………
10
Transimmunization (TI) procedure ……………………………………………………………………………………………
13
Cell Stimulation Readouts
………………………………………………………………………………………………………..
15
Statistics…………………………………………………………………………………………………………………………………
16
RESULTS …………………………………………………………………………………………………………………………………
18
REP generates a large population of CD8 T cells with Desired TCR specificity. …………………………….
18
CD8 T cells release IFNg upon direct stimulation with SP. ………………………………………………………….
18
Co-culture of PBMC with E7(11-20) CD8 T cells and E7 Ags results in non-specific IFNg production. .
18
Co-culture of PBMC with E7(11-19) CD8 T cells and E7 Ags results in Ag-specific IFNg production. ..
19
TI, platelets, and E7 Antigen sources induce a pro-inflammatory MoDC phenotype. ………………………
21
PD-1 can be used as surrogate for T cell stimulation.
…………………………………………………………………..
22
DISCUSSION
……………………………………………………………………………………………………………………………
23
Limitations and Future Directions
……………………………………………………………………………………………..
26
REFERENCES …………………………………………………………………………………………………………………………
27
FIGURES………………………………………………………………………………………………………………………………….
31
ABSTRACT
Extracorporeal Photochemotherapy (ECP) is an FDA-approved immunotherapy that has been treating
cutaneous T cell lymphoma (CTCL) for over three decades. ECP’s antitumoral effect is a consequence of
its generation of functional, physiologic, inflammatory monocyte-derived dendritic cells (MoDCs) and
apoptotic, patient-derived tumor, which collectively, stimulate the adaptive immune system. Thus, in
CTCL, ECP serves as a therapeutic dendritic cell vaccine against patient-specific neoantigens. This
mechanism of action suggests ECP’s potential application to other solid tumors. We tested ECP’s
applicability to head and neck squamous cell carcinoma (HNSCC) using a trackable antigen system
involving the constitutively expressed HPV16 E7 oncoprotein. We hypothesized that ECP would
successfully stimulate anti-epitope CD8 T cells, quantified by IFN-gamma ELISA, following processing
and cross-presentation of HPV16 E7+ peptides and tumor cells by MoDCs. The trackable antigen system
employed a commonly cited epitope, E7(11-19). E7+ short peptide and long peptide generated significant
IFNg (p<0.0001) relative to the null control group. Tumor cell line SCC61 T+ (E7hi) demonstrated
significantly elevated IFNg production relative to SCC61 T- (non-E7 expressing tumor), but only in the
presence of platelets, plate-passage, and overnight incubation (p<0.0001). These results suggest an
antigen-specific CD8 T cell response and reiterate critical ECP components that have previously been
shown to facilitate immunogenic MoDC generation. Immunogenic MoDC phenotype was confirmed with
flow cytometry of inflammatory surface markers and intracellular cytokines, all of which were generally
upregulated following ECP. Overall, we have demonstrated a proof-of-principle for ECP’s therapeutic
vaccination against HNSCC. This is particularly relevant because ECP offers unique synergistic potential
with recently FDA-approved checkpoint inhibitors.
ACKNOWLEDGEMENTS
Working in Dr. Richard Edelson’s lab was my most memorable academic experience in medical school. I
fortuitously found my mentor’s lab in my search for a basic science cancer immunotherapy project
following my first year of medical school. After completing my summer project, I knew I would return
for a fifth-year research experience.
Dr. Richard Edelson and Dr. Douglas Hanlon were my primary mentors. Dr. Edelson oversaw the “big
picture”. Without him, this project would not be possible and I would not have the privilege of working
with experts in this field. Dr. Douglas Hanlon and Dr. Olga Sobolev provided their expertise in
immunology and subsequently guided the experimental designs. They also helped in interpreting the
results. Eve Robinson, Renata Filler, and Dr. Kazuki Tatsuno provided their specialized knowledge to
fine-tune experiments. Dr. Olga Sobolev, Eve Robinson, Dr. Douglas Hanlon, Renata Filler, Dr. Kazuki
Tatsuno, and Patrick Han all assisted with running various aspects of the experiments.
The Yarbrough Lab (Dr. Wendell Yarbrough, Dr. Natalia Issaeva, and former members/classmates Cassie
Pan and Tejas Sathe) generously collaborated with us on this project, providing materials and head and
neck squamous cell carcinoma expertise.
The Hinrichs Lab at the NCI/NIH (Dr. Christian S. Hinrichs, Dr. Benjamin Jin, etc.) generously
collaborated with us, providing materials for T cell transductions and expertise in T cell manipulation.
Inger Christiansen and the rest of the nurses of the photopheresis unit generously drew donor blood.
Antonella Bacchiocchi and Dr. Ruth Halaban generously provided a working space in which to transduce
T cells.
Funding support for this thesis was provided by the James G. Hirsch, M.D., Endowed Medical Student
Research Fellowship.
1
INTRODUCTION
An improved understanding of the relationship between immune system dysfunction and the
development, establishment, and progression of cancer has evolved the landscape of immunotherapy.1 In
an immunocompetent host, if the immune system is able to recognize the cancer cell as “foreign” and
avoid suppression, it should succeed in eradicating the malignancy. A high tumoral mutation rate
increases the probability of generating non-synonymous mutations, and consequently, targetable
neoantigens on the cell surface.1-3 Interestingly, extracorporeal photochemotherapy or extracorporeal
photopheresis (ECP), an FDA-approved immunotherapy for advanced cutaneous T cell lymphoma
(CTCL), serves as a therapeutic dendritic cell vaccine which generates a clinically significant cytotoxic T-
cell response against patient-specific neo-antigens.4,5 This unique capability theoretically extends ECP’s
application to other malignancies4-8, particularly those with high mutation rates or poorly characterized
molecular targets.2,3,9,10
ECP Discovery
Edelson and colleagues serendipitously discovered the clinical effects of ECP when they were devising a
blood-directed palliative chemotherapy against CTCL, a group of non-Hodgkins lymphoma marked by
skin infiltrating malignant T cells.11,12 The first patient treated with ECP in 1982 entered lifelong
remission after only 3 monthly treatment cycles of 3% of the total CTCL cells.5 Furthermore, in the 1987
phase I/II clinical study treating 37 resistant CTCL patients, although less than 5% of the patient’s
malignant T cells were treated with photochemotherapy, 73% of the patients responded to treatment and
experienced an average 64% decrease in cutaneous involvement. Remarkably, there were no serious side
effects.11 As a result of this favorable trial, ECP was expeditiously approved by the FDA in 1988, and
subsequently, CTCL was no longer deemed a universally fatal disease.5,12
ECP’s Mechanism of Action
The cellular and cytokine interactions involved in ECP’s immunogenic effect have slowly been defined
2
over the course of several decades since its invention. Edelson and colleagues surmised that the clinical
effects of ECP could not be explained simply by the reduction in the population of circulating malignant
cells, since less than 5% of the patients involved T cells were exposed to the photochemotherapy. ECP’s
mechanism of action became even more nebulous after the medical community realized its beneficial
immunosuppressive effects in organ transplant recipients (ie graft versus host disease or transplant
rejection)13, suggesting bidirectionality of ECP’s immunomodulation. It was not until the discovery that
ECP was able to generate two maturationally distinct subsets of dendritic cells, one immunogenic and one
immunotolerizing4,5,14-16 that this seemingly paradoxical phenomenon was reconciled.
Understanding of ECP’s immunogenic mechanism requires summarizing the therapy’s use in CTCL:
(1) A portion of the patient’s whole blood is removed and the leukocyte fraction containing malignant
CD4 T cells, healthy immune cells and platelets is isolated and combined with 8-methoxypsoralen (8-
MOP).
(2) This cellular/chemotherapy mixture is exposed to UVA radiation as it passes through a 1mm channel
within parallel plastic plates permeable to UVA. UVA light activates 8-MOP into the therapeutically
active form capable of cross-linking DNA and thus inducing cell death through DNA damage. The
photochemotherapy combination, UVA and 8-MOP (PUVA), is titratable, so that it can preferentially
induce apoptosis of leukocytes (ie malignant cells in CTCL) while sparing other immune cell subsets such
as monocytes and dendritic cells.
(3) The cell mixture is returned to the patient, inducing an immunogenic response against the malignancy.
The current understanding of ECP’s mechanism, deriving from years of scientific research, and endorsed
by the ECP community at the recent meeting of the American Council of ECP17, is as follows.
A. Platelet-Monocyte Interactions Produce Monocyte-Derived Dendritic Cells
As the cells and plasma proteins pass through the parallel plates during ECP, fibrinogen coats the surface
of the polystyrene flow chamber, and platelets subsequently adhere to the RGD domains of fibrinogen via
3
αIIβ3 and α5β1 receptors. The adherent/activated platelets translocate P-selectin among other proteins to
their surface, which then binds monocytes via PSGL-1, facilitating monocyte tethering to the platelets,
partial monocyte activation, and monocyte integrin receptor conformational changes. Further interactions
between platelet ligands (including those containing RGD domains) and monocytes induce monocytes
into the dendritic cell (DC) maturational pathway. Flow shear stresses imparted on tethered monocytes,
and platelet density within the ECP plate, are significant factors in transforming monocytes into
monocyte-derived dendritic cells (MoDCs).5,7,18
B. PUVA Effect on Antigenicity and Dendritic Cell Maturation
Concurrent to monocyte activation by platelets, 8-MOP becomes transiently photoactivated upon UVA
exposure, covalently crosslinking pyrimidine bases of DNA and irreversibly damaging exposed nucleated
cells. In ECP, the 8-MOP-UVA (PUVA) dosage is such that as few as three DNA-photoadducts per
million DNA base pairs can induce universal lymphocyte apoptosis, while largely sparing the similarly
exposed monocytes.5,19
In CTCL patients, platelet-monocyte interactions produce functional, immunogenic MoDCs, which
uptake PUVA-induced apoptotic malignant lymphocytes as a patient-specific antigen source, and produce
a clinically significant CD8 T cell response against the malignancy.5,7,8,11,12,20 Within 24 hours of ECP,
monocytes increase expression of 498 genes, nearly 20 of which are associated with mature DC function
and identity, and over 60 encoding for transmembrane signaling proteins. In particular, RNA transcripts
coding for MHC class II presentation (ie DC-LAMP), T cell costimulatory molecules (ie CD80, CD86,
CD40), and other DC maturation markers (ie CD83, Decysin, FPRL2, CCR7, Decysin, OLR1) have been
shown to be significantly upregulated.12 The immunogenic ECP-derived MoDCs upregulate surface
expression of CD83, CD36, and MHC class II molecules, while reducing expression of CD14. Most
importantly, these DCs acquire the functional capacity to mediate antigen cross-presentation to CD8 T
cells and induce antigen-specific T cell proliferation.5,7 In addition, tumor cells treated with PUVA
4
upregulate MHC class I molecule (MHCI) expression and therefore, along with uptake by ECP-derived
MoDCs, also enhance neoantigen peptide processing, improving the TCR-mediated CD8 T cell
response.21
In contrast, if monocytes themselves are sufficiently damaged by PUVA, they instead become immature,
tolerogenic DCs.16,22-24 Under these conditions, PUVA-exposed monocytes upregulate the glucocorticoid-
induced leucine zipper GILZ (a hallmark of tolerogenic DCs), downregulate costimulatory molecules
CD80 and CD86, exhibit resistance to Toll-like receptor-induced maturation, increase production of IL-
10, and decrease production of IL-12.22 Additionally, PUVA can result in DC apoptosis, which furthers
the immunosuppressive environment by inducing regulatory T cells. These cellular mechanisms are
thought to underlie ECP’s immunotolerizing effects in GvHD and organ transplant patients.16
C. MoDCs Generate an Antigen-Specific Response
Current understanding of dendritic cell biology and ECP’s favorable clinical results in the absence of
major systemic side effects suggest patient-specific antigen targeting by T cells, with recent murine and in
vitro studies confirming this.4-6,8,20,25 Immunogenic ECP drives an adaptive immune response against
patient-specific neoantigens, primarily through antigen-containing MoDC activation of CD8 T cells,
though the presence of CD4 T cells and natural killer cells have also recently been shown to be significant
factors in the anti-tumor effect.4 In tolerizing ECP, tolerogenic MoDCs are thought to achieve antigen-
specificity though uptake of non-self antigens arising from tissue damage in GvHD or transplant settings,
and specific stimulation of regulatory T cells.5,16,22
ECP’s Evolution
To improve MoDC maturation and neoantigen uptake, processing, and docking onto APC’s MHC
molecules, studied in the early 2000’s onward adopted a modified ECP protocol termed
“Transimmunization” (TI), which includes an overnight incubation period of PUVA-treated cells.4,6,8
During this period, spatial proximity of PUVA-damaged tumor cells with platelet-activated MoDCs
5
allows for more efficient DC uptake and processing of antigens from dying tumor cells, as evidenced by a
60% clinical response rate of CTCL patients who failed standard ECP and/or other therapies during a
clinical trial of Transimmunization for CTCL.26 Furthermore, ECP’s ability to provide a functional,
therapeutic DC vaccine against undefined patient-specific neoantigens suggests therapeutic potential
against other solid tumors. A modified transimmunization protocol where only tumor cells but not
immune cells are exposed to PUVA, and where PUVA-treated tumor cells are subsequently co-incubated
with MoDCs prior to re-injection, has recently demonstrated a significant CD8 T cell response against
melanoma antigens both ex vivo and in vivo,4,20 with melanoma (YUMM1.7)-inoculated and colon
carcinoma (MC38)-inoculated mice exhibiting significantly reduced tumor volume relative to untreated
controls.4
Potential Application to Head and Neck Squamous Cell Carcinoma
To test ECP’s efficacy in other solid tumors, we sought a malignancy with poor prognosis and easily
trackable antigens to demonstrate therapeutic vaccination proof-of-principle. Head and neck squamous
cell carcinoma (HNSCC) accounts for 550,000 cases and 380,000 deaths annually. One-third of patients
present with early stage disease (stage I or II). These patients, managed with primary surgery or
radiotherapy (RT), carry a five-year overall survival rate of 70-90%. The remaining two-thirds of patients
present with locoregionally advanced disease (stage III or IV) and are treated with a combination of
surgery, RT, and/or chemotherapy. Unfortunately, 60% in this latter group experience local recurrence
and 30% demonstrate distant metastases. Those with recurrent or metastatic disease have a median overall
survival of 10 months.1,27-29 HNSCC’s robust immunosuppressive mechanisms explain the particularly
poor prognosis, at least in part due to the following: generation of regulatory T cells, dysfunctional
antigen processing and presentation, dysregulation of cytokine and chemokine pathways, and a hostile
tumor microenvironment.1,27,30,31
6
HNSCC can be further classified into human papilloma virus (HPV) positive and negative entities, with
25% of all HNSCC and up to 90% of oropharyngeal carcinoma cases being HPV positive.29,30,32 It is well
established that carcinogenic strains of HPV are responsible for the constitutive synthesis of oncoproteins
E6 and E7, inhibitors of tumor-suppressor proteins p53 and Retinoblastoma (Rb), respectively.33 Because
these oncoproteins are uniquely expressed on malignant tissues, they have become attractive therapeutic
targets,30,33-35 and at minimum, can serve as trackable tumor-specific antigens for demonstrating ECP’s
applicability to HNSCC.
We hypothesized that ECP could be a potent therapy against HNSCC given the recent success of adoptive
tumor-infiltrating lymphocyte (TIL) transfer strategies in this disease.3,10,30,36 In fact, Stevanovic et al.
found in patients an immunodominant T cell reactivity against neoantigens, rather than the canonical E6
and E7 epitopes,3 which favorably plays into ECP’s unique mechanism of action.5 Moreover, PUVA’s
ability to upregulate MHCI expression21 may partially overcome the immunosuppressive nature of
HNSCC.1 Finally, the recent FDA-approval of checkpoint inhibitors directed against PD-1 and CTLA-4
receptors on T cells further increased the ECP’s appeal, offering potential for immunotherapeutic
synergism between a dendritic cell-based strategy and immunogenic T cell activation strategies.1
Here, we investigated ECP’s potential to treat HNSCC through the in vitro generation of a CD8 T cell
response against a well-established HPV16 E7 epitope using the TI protocol. In selecting a trackable
antigen, we focused on HPV16 because it is the most commonly implicated HPV strain in HNSCC.29,30,32
With respect to oncoprotein, we chose E7 instead of E6, given that it exhibits less sequence variation and
has superior epitope binding affinity to HLA*A2:01, the most prevalent MHCI molecule in the United
States and Europe.30,37,38 With respect to CD8 T cell receptors, we sought epitope specificity to E7(11-20)
and E7(11-19), highly studied sequences with ability to stimulate CD8 T cells in vitro and in vivo.38,39 We
reasoned that if ECP is able to initially demonstrate successful proof-of-principle CD8 T cell response
specifically against HPV antigens in HNSCC, then the therapy may similarly generate an in vivo response
7
against patient-specific HNSCC neoantigens, which are implicated in clinically effective anti-tumor
responses.3,5,9 This is the first time ECP has been tested in the context of HNSCC.
8
STATEMENT OF PURPOSE
This is a proof-of-principle project to determine ECP applicability to HNSCC by using a trackable
antigen source, in particular, two commonly studied immunogenic HPV16 E7 epitopes.
Specific Aims
1. Generate a CD8 T cell line with TCR specificity to HPV16 E7(11-19) and E7(11-20) epitopes
2. Demonstrate that the updated “Transimmune” version of ECP (TI) can generate an HPV16 E7 Ag
specific CD8 T cell response using HPV16 E7 overlapping and long peptides
3. Demonstrate that TI can generate an HPV16 E7 Ag specific CD8 T cell response using a HNSCC cell
line expressing E7 protein
4. Demonstrate that following TI, monocytes acquire a mature, immunogenic DC phenotype, which
underlies successful Ag cross-presentation
Hypothesis
TI will generate functional, immunogenic MoDCs, defined by their ability to phagocytose, process, and
cross-present HPV+ HNSCC antigens, thus stimulating CD8 T cells in an antigen-specific manner.
9
MATERIALS AND METHODS
HPV16 E7 Antigen Sources
Peptides. Short, long, and overlapping peptides of the HPV16 E7 protein were used as Ag sources. Short
peptide (SP) can directly fit into the TCR groove, bypassing APC phagocytosis, processing, and loading
onto MHCI. Because of this capacity to directly stimulate T cells, SP served as a positive control. In
contrast, long and overlapping peptides served as the experimental Ags, as they require functioning APCs
for CD8 T cell stimulation.39 Short peptides: 9mers (amino acid sequence: YMLDLQPET) and 10mers
(amino acid sequence: YMLDLQPETT) were synthesized at Tufts School of Medicine. Long peptide
(LP): HPV16 E7(1-30) was synthesized at Tufts School of Medicine. Overlapping peptide (OP): PepMix™
HPV 16 Protein E7, a commercially available pool of 22 peptides derived from a peptide scan (15mers
with 11 aa overlap) was used (JPT Peptide Technologies). Nanoparticle encapsulation: Using a previously
described protocol, LP and OP was encapsulated with PLGA, which has been shown to enhance APC
phagocytosis and presentation of the MHCI-peptide complex.40 SIINFEKL (ie OVA257-264) and melanoma
antigen gp100(25-33) were used as negative peptide controls.
Cell Lines
Two SCC61 cells, one transfected to express high levels of HPV16 E6/E7 proteins (SCC61 T+) and one
non-transfected and naturally non-expressive (SCC61 T-), were grown in DMEM/F12 media (1:1) + 10%
FBS + 0.4ug/ml hydrocortisone.41 The SCC090 cell line (SCC90), naturally expressing HPV16 E7
protein, was purchased at ATCC and grown in MEM, 10%FBS, 2mM L-glutamine. Normal human
fibroblasts (NHF), which served as a cell line based negative control for E7 Ag, and NHF line modified to
express HPV16 E6/E7 proteins (NHF T+) which served as a non-tumoral cell line source of E7 Ag, were
grown in DMEM, 10%FBS, 2mM L-glutamine. SCC61 T-, SCC61 T+, NHF, NHF T+ were generously
donated by the Yarbrough Lab at Yale. All cell lines were grown with Pen/Strep and grown to 70-80%
confluence, harvested with trypsinization, and centrifuged at 1400RPM for 7min prior to use.
10
Peripheral Blood Mononuclear Cells (PBMCs)
PBMCs were obtained from healthy, HLA*A2:01+ human donors, in accordance with the guidelines of
the Yale Human Investigational Review Board, and with informed consent obtained under protocol
number 0301023636. Whole blood was collected at a ratio of 100ml whole blood to 500ul heparin, so as
to prevent coagulation but preserve future platelet activation. Whole blood was gently layered over
Isolymph (CTL Scientific Supply Corp.) at a ratio of 35ml:15ml, centrifuged for 30minutes at 1500rpm.
Then, the buffy coat was collected, washed with PBS using the same centrifugation settings, and the
PBMC pellet resuspended for use.
CD8 T Cells
Anti-HPV16 E7(11-20) CD8 T cells. CD8 T cells with anti-HPV16 E7(11-20) TCR specificity were obtained
from a commercial line originating from a female HLA*A2:01+ donor (Astarte Biologics), and expanded
using the rapid expansion protocol (below). Following expansion, CD8 T cell purity was determined with
flow cytometry analysis (below).
Anti-HPV16 E7(11-19) CD8 T cells. Transgenic CD8 T cells with anti-HPV16 E7(11-19) TCR specificity were
generated from a healthy, middle-aged, male HLA-A2+ donor’s PBMC population (Supplementary
Figure 1). CD8 T cells were isolated from PBMCs, transduced to co-express anti-E7(11-19) TCR, and
expanded using a rapid expansion protocol (REP). Following expansion, CD8 T cell purity was
determined with flow cytometry analysis.
TCR retroviral constructs. A retroviral supernatant capable of transducing human T cells to co-express
TCRs with high-affinity binding to the E7(11-19) epitope was generously donated by Dr. Christian Hinrichs
and his colleagues (NCI). The supernatant contains MSGV1 retrovirus, modified to include TCR
nucleotide sequences inserted into its retrovirus backbone and is codon-optimized for expression in
human cells. Moreover, its human TCR constant regions are exchanged for mouse TCR constant regions,
11
allowing for easy flow cytometry confirmation of successful transduction. This retrovirus has been used
in prior clinical trials (Gene Oracle) and its molecular structure has been extensively characterized.30
T cell transduction. Retroviral transduction was carried out based on recommendations from the Hinrichs
Lab. On day 1, human PBMCs were isolated from a healthy HLA-A2+ donor and CD8 T cells were
isolated using negative selection (Miltenyi Biotec). CD8 T cells were seeded into three wells of a tissue-
culture 24-well plate at a ratio of 7.5x106 CD8 T cells per well, with each well containing 2ml media
(50/50 media + 300 IU/mL IL2 + 50 ng/mL OKT3). On day 2, a non-tissue culture 24-well plate was
coated with retronectin (Takara #T100A/B) at final concentration of 20 µg/ml in PBS (ie 500ul
retronectin/well), wrapped and stored overnight at 4C. On day 3, the retronectin plate was centrifuged at
2000xg at 32C for 1 hour, followed by retronectin solution aspiration and blockage with 2ml 2% BSA in
PBS for 30min at room temperature (RT). The retroviral supernatant was spun for 10 minutes at
1000RPM. The retronectin plate was washed 2x with 1ml/well PBS, with the PBS remaining until the
viral supernatant was added. After the 0.5ml/well supernatant was added, the retronectin plate was plastic
wrapped, and spun for 2000xg at 32C for 2 hours. At the 1.5hour mark, T cells from the other plate were
harvested and counted. The retronectin plate’s supernatant was aspirated and 0.25x106 T cells were
added/well in 1 ml of media (50/50 AIMV/RPMI + 300 IU/mL IL2). The plate was wrapped and spun for
10min at 1500RPM with brake set at 1, and incubated overnight. On day 4, the transduced cells were
transferred off the retronectin into a tissue culture-coated 24-well plated. On day 6, cells were transferred
to the appropriate sized flask and replenished with T cell media. On day 7, flow cytometry was used to
assess E7(11-19) TCR positivity (below).
Cell Sorting. T cells were harvested, spun down for 10min at 1500RPM, supernatants were discarded, and
the cells were resuspended. Fc block was added and cells were incubated for 15min on ice. Staining
buffer (1xPBS w/3% BSA) was added, and cells were centrifuged and resuspended identically. Samples
12
were stained for anti-CD8-BV421, anti-CD3-APC, anti-mTCRb-FITC (clone: H57-597), anti-7AAD-
percp-cy5.5, with respective single stain controls; stained cells were incubated for 30min on ice. Cells
were centrifuged and resuspended identically, passed through a 70um filter (Falcon), and brought up at a
25x106/ml sorting buffer (1xPBS w/ 0.1% BSA). Live CD3/CD8/mTCR-β triple-positive were selected
(BD FACSAria).
Rapid Expansion Protocol (REP) of CD8 T cells. REP stimulates T cells with the monoclonal antibody
OKT3 (anti-CD3), IL-2, and irradiated autologous or allogeneic feeder cells to expand lymphocytes up to
1000-fold over 14 days. This protocol was used to expand both E7(11-19) and E7(11-20) CD8 T cells.
Approximately 2x108 feeder cells (ie PBMCs) were obtained from 3 healthy HLA-A2+ donors using the
methods above and subsequently irradiated with 4000cGy (Yale Cesium-137 irradiator) at a cell
concentration of 50*106/ml on ice. 2x108 feeder cells were combined with 1-2x106 CD8 Tcells in a T-150
with the following “50/50” media components: 75ml “complete media” (10% human AB serum (Lonza),
pen/strep, 2.5% HEPES in RPMI), 75ml AIM V media, 4.5ul OKT3 (1mg/ml), 75ul IL-2 (1300IU/ml),
ciprofloxacin (10 ug/ml). This flask was incubated at 37C, 5% CO2 until day 5, during which 3/4 of the
medium was aspirated without disturbing the settled cells, and replenished with fresh 50/50 medium (as
above) but without OKT-3. From this point forward, when the viable cell count reached 0.6x106-
1x106/ml, the cells were split 1:2 and replenished with AIMV media with 5% human AB serum and
1300IU/ml IL-2. Cell count was assessed every two days. The desired effector T cell population was
confirmed with flow cytometry (below) and frozen down at 5x106/ml 10% DMSO aliquots in liquid
nitrogen.
Flow cytometry to characterize post-REP CD8 T cells. Flow cytometry was used to confirm the epitope-
specific T cell population following REP. Fixable apoptosis dye “zombie”-apc-cy7, anti-mTCRb-FITC,
anti-E7(11-20) dextramer-PE, anti-CD8-PacBlue, anti-CD4-PE-Cy5.5, and anti-CD3-APC stains with
respective IgG controls were used (Biolegend). Anti-HPV E7(11-19) CD8 T cells were identified based on
13
the live CD3/CD8/mTCRβ positive population. Anti-HPV E7(11-20) CD8 T cells were identified based on
the live CD3/CD8/E7(11-20) dextramer positive population. Data were acquired with a Stratedigm flow
cytometer (BD Biosciences) and analyzed with FlowJo software (FlowJo).
Transimmunization (TI) procedure
The TI employed closely followed a previously established protocol,4 but was adjusted for different cell
lines. An overall schematic is available as Supplementary Figure 2. Briefly, PBMCs were combined
with PUVA-treated tumor cells, plate-passaged together, and incubated overnight. For peptide Ag groups,
PBMCs alone were passed over a plate and peptide was added directly to the overnight incubation dish.
The effects of plate passage, platelets, and TI (ie plate passage and overnight incubation) on T cell
stimulation were tested. Additionally, each PBMC + CD8 Tcell + Ag group had parallel Ag groups
containing only CD8 T cells or only PBMCs. Allogenic PBMCs were used in E7(11-20) experiments while
autologous PBMCs were used in E7(11-19) experiments.
CD8 T cell Thaw (Day 0). CD8 T cells were thawed and spun down at 700 RPM for 7 minutes. They
were counted and incubated in T-25 flasks with 20ml T cell media (5% human AB serum, pen/strep,
2mM L-glutamine in AIMV) and 8ul/20ml media of 1300IU/ml IL-2 until Day 2.
PUVA (Day 1). The PUVA dose was cell line specific and based on the minimum dosage required to
induce 100% apoptosis. Apoptosis was assessed by proliferation cessation during a 10-day monitoring
period under confocal microscopy. SCC61 T+, NHF T-, and NHF T+ were treated with 8 Joules(J)/cm2
and 200ng/ml 8-MOP while SCC61 T-, SCC90 were treated with 4J/cm2 and 200ng/ml 8-MOP. NHF T+
served as non-tumoral cell source of E7 Ag while NHF T- served as a non-antigenic negative control cell
line. Cell lines were trypsinized, centrifuged and brought up at a concentration of 2.5x106 cells/300ul
FBS. They were incubated in the dark with respective amounts of 8-MOP for 20minutes. To prepare for
UVA treatment, 12-well plates were pre-coated with FBS to prevent cell adhesion. When 8-MOP
14
incubation was completed, FBS was aspirated from the wells, and each well received 2.5x106 cells.
Finally, each cell line received their respective UVA dose.
Plate passage (Day 1). For groups involving cell lines as an Ag source, PUVA treated cells were
combined with 107 cells PBMCs for a total volume of 600ul FBS. For groups involving peptides as an Ag
source or PBMC + Ag groups, up to 4x107 cells PBMCs were brought up in 600ul FBS. For all groups,
two-thirds (400ul) of the 600ul were used to coat the micro-plates for an hour at 37C to maximize fibrin
and platelet adhesion; the remaining one-third remained in the Eppendorf tube. Microplate tubing was
coated with FBS to prevent cell loss. After one hour, 600ul of cells were passaged through the microplate
at 0.09 ml/min using a syringe pump. After collecting cells into respective Eppendorf tubes, microplates
were washed with FBS at 0.49 ml/min in order to maximize cell yield. Overall, plate passage closely
followed a previously established protocol.4
Overnight Incubation (Day 1). To maximize MoDC maturation and Ag uptake, cell groups were
incubated overnight at 37C. After plate passage, cells were counted. The cell count across groups were
normalized based on the group with the lowest yield, typically above 5x106 cells, and each cell group was
transferred into a respective 35mm dish with 15% human AB serum in 2ml RPMI. Groups that were not
plate-passaged were thrown into the 35mm at the same time as their plate-passaged counterparts. In
general, peptides were added at 10uM concentration into their respective dishes; however, nanoparticle
encapsulated particles were added at a concentration of 200ug NP/2ml RPMI, based on prior NP
experiments.40
CD8 T Cell Co-Culture (Day 2). The next day, PBMCs in 35mm dishes and effector CD8 T cells were
harvested, spun down, brought up in TCM, and counted. All groups were seeded as triplicates in 96-well
plates for a total volume of 200ul TCM/well. For PBMC + T cell + Ag groups, 100ul of 0.2x106 PBMCs
15
were combined with 100ul of 0.1x106 effector CD8 T cells. PBMC + Ag and T cell + Ag groups seeded
0.2x106 PBMCs in 200ul TCM and 0.1x106 CD8 T cells in 200ul TCM/well, respectively. In the T cell +
Ag groups, peptide Ags were added at 10uM as before, and cell lines were PUVA treated and added in
proportional numbers to their respective day 1 groups. Coincubation in the 96-well plates lasted 72 hours.
Platelet depletion and “No TI”. In experiments involving transgenic CD8 T cells, the immunogenic
contributions of platelets and TI were investigated. On day 1, platelets were depleted using CD61
microbeads (Miltenyi Biotec) following PBMC isolation. Platelet depletion was confirmed using a
Hemavet (Drew Scientific, Inc.), with a threshold of <104/ul considered successful depletion. Moreover,
autologous serum (BD Vacutainer) was used for overnight incubation media. Subsequent experimental
distribution of platelet-containing and platelet-depleted PBMCs was the same as above. The “no TI”
effect was explored by acquiring fresh PBMCs on day 2 and platelet-depleting a portion of them as done
with TI, and adding them to a PBMC + T cell + Ag co-culture without plate passage and overnight
incubation. “No TI” PBMCs were distributed to their respective groups as described in the “CD8 T Cell
Co-Culture” section above.
Monocyte Isolation. To simplify the APC-T Cell co-culture system and background IFNg noise,
monocytes were isolated from donor PBMCs using the Pan Monocyte Isolation Kit (Miltenyi Biotec) per
manufacturer’s recommendations. In the E7(11-20) system, monocytes were purified either on the day of
PBMC isolation or after overnight incubation and prior to 96-well plate transfer. In the E7(11-19) system,
monocytes were purified immediately following PBMC isolation.
Cell Stimulation Readouts
IFNg ELISA to measure CD8 T cell stimulation. After 72 hour co-cultures, the 96-well plates were spun
down at 2000RPM for 10min at 4C. 180ul of supernatant were transferred into new 96-well plates,
16
wrapped, and frozen at -80C. Supernatants were analyzed with ELISA for interferon-γ (IFNg) production
(BioLegend), which signifies CD8 T cell stimulation.30,42 The ELISA was run according to vendor
guidelines and analyzed with the Spectromax reader.
Flow cytometry to characterize phenotypic changes in monocytes after TI. The following surface markers
which characterize MoDC maturation4,12,43 used to stain PBMC following TI treatment, per
manufacturer’s instructions: CD80-FITC, CD83-APC, CD86-PE, HLA ABC-APC, HLA-DR-
APCEeFlour 780, CD11c-PE/Cy7, CD14-PacBlue, ICAM1-FITC, PLAUR-PE (BioLegend). An
intracellular staining kit (BD Biosciences) was used to characterize associated chemokine expression with
the following stains: CXCL5-PE, MCP-1-APC (BD Biosciences). Singlet, live, CD11c/CD14+ were
gated on. Relative surface marker expression was quantified by change in mean fluorescent intensity
(MFI) or by percentage of positive cells, as most appropriate to the stain. MFI was calculated as the
intensity difference between stain and IgG. Percentage of positive monocytes was used to track
populations expressing intracellular chemokines or surface CD80.
Flow cytometry to characterize PD-1 expression following T cell stimulation. PD-1 is a recently
discovered surrogate for neoantigen-specific CD4 and CD8 T cell stimulation3,36,44-46 and is therefore
useful to ECP experiments, which involve mounting an adaptive immune response to undefined
neoantigens. PD-1-PacBlue staining was used in conjunction with zombie-APC-Cy7, CD3-PE, CD4-
APC, CD8-FITC, with corresponding IgGs (BioLegend) following 2 day stimulation with OKT-3 (30ul
of 1mg/ml stock per 2ml media) and IL-2 (1300IU/ml) of PBMCs.
Statistics
Statistical analyses were performed with Prism 8 (GraphPad Software). Depending on the experiment, 1-
way or 2-way ANOVA with Holm-Sidak’s multiple comparisons test were used to determine statistically
significant IFNg production with respect to antigen, nanoparticle, and/or plate-passage effects. Tukey’s
17
multiple comparison tests were used to determine statistically significant differences in TI permutations in
the E7(11-19) system. A P value of less than 0.05 was considered significant.
18
RESULTS
REP generates a large population of CD8 T cells with Desired TCR specificity.
To assess TI’s APC cross-presentation efficiency, we first needed to generate two large populations of
epitope-specific CD8 T cells using REP. Anti-HPV16 E7(11-20) CD8 T cells were expanded 200-fold from
a commercially available 2 million cell aliquot, of which approximately 81% were viable epitope-specific
CD8 T cells. Following REP, cells were 95% viable, with E7(11-20) CD8 T cells comprising two-thirds of
the compartment (Figure 1A). Since the physiological relevance of the E7(11-20) epitope is not universally
accepted by the HPV research community38, we also generated in-house anti-HPV16 E7(11-19) CD8 T cells,
and in this iteration, we sought a higher cell purity. After cell sorting, we obtained 1 million cells, nearly
100% of which were E7(11-19) CD8 T cell. After REP, cells were expanded 150-fold; up to 93% were
viable, with E7(11-19) CD8 T cell comprising 95% of the compartment (Figure 1B).
CD8 T cells release IFNg upon direct stimulation with SP.
Following generation of epitope-specific CD8 T cell populations, we confirmed T cell functionality in the
simplest IFNg assay, cognate SP stimulation. Both E7(11-19) and E7(11-20) T cell lines demonstrated
appropriate IFNg release upon direct TCR stimulation with the appropriate SP. E7(11-20) T Cell cultures
with antigen showed significant IFNg production exclusively with SP (Figure 2A: p=0.012; Figure 2B:
p<0.0001). In contrast, non-specific gp100 or SIINFEKL peptides, or E7+ Ag sources requiring APC
processing were not able to directly stimulate E7(11-20) T cells (p>0.99; Figure 2). E7(11-19) transgenic CD8
T cells also demonstrated significant IFNg release against their cognate SP (Figure 2C: p<0.0001).
Co-culture of PBMC with E7(11-20) CD8 T cells and E7 Ags results in non-specific IFNg production.
After demonstrating that our CD8 T cell lines could be stimulated with SP, we began TI testing of the
E7(11-20) system to demonstrate APC-mediated Ag cross-presentation. PBMC + Ag cultures were run in
parallel to the co-cultures to determine non-specific IFNg production (Figure 3), set as the dotted lines in
19
Figure 4. E7(11-20) CD8 T cells co-cultured with PBMC and peptide Ags (Figure 4A), did not
significantly increase their IFNg production relative to the “null” (ie no Ag addition) group (p>0.98).
When testing cellular E7 sources with the E7(11-20) T cell line (Figure 4B), again, there was no significant
increase in IFNg production relative to the “null” group (p>0.98). Nanoparticle encapsulation and plate-
passage did not significantly increase IFNg production (p>0.98 for both). Given that one-third of the
“E7(11-20) CD8 T cells” added to the co-cultures were not epitope-specific (Figure 1), and the non-
autologous co-culture system, we suspected that high IFNg levels even in the “null” control group were
due to cross-reactivity between the impure compartment of E7(11-20) CD8 T cells and donor PBMCs. To
minimize non-specific IFNg release, the peptide Ag experiment was repeated by isolating monocytes, the
APC progenitor, from PBMC donors. The PBMC + E7(11-20) T cell + Ag experiment was run in parallel
with the monocyte + E7(11-20) T cell + Ag experiment; corresponding monocyte groups had up to ~40x
reduction in IFNg release, confirming our hypothesis of non-specific allogenic cross-reactivity in this
system (data not shown). Initially, we planned to purify and REP E7(11-20) CD8 T cells for repeat
experiments, but after speaking with HPV experts at the NCI (ie Dr. Christian Hinrichs and colleagues),
we decided to instead move forward with the co-culture system involving E7(11-19) CD8 T cells, as
E7(11-19), and not E7(11-20), is a naturally expressed epitope in HPV16 carcinoma. Additionally, this system
used autologous PBMC as the source of APCs to minimize nonspecific reactivity.38
Co-culture of PBMC with E7(11-19) CD8 T cells and E7 Ags results in Ag-specific IFNg production.
Overall, all E7+ groups seemed to selectively produce IFNg in the E7(11-19) system, suggesting successful
TCR stimulation by direct exogenous loading as well as through APC-internalization/cross-presentation.
First, a preliminary peptide stimulation assay was run comparing “null” no peptide, SP, and LP antigens
in PBMC + E7(11-19) T cell + Ag co-cultures (Figure 5A). In this assay, autologous PBMCs were isolated
and incubated overnight with 25uM of the relevant Ag, and the next day were transferred to a 96-well
plate for 72 hour co-culture with T cells. Both SP and LP groups demonstrated over 40x IFNg production
relative to the null group (p<0.0001). Further, the LP group produced significantly more IFNg than the SP
20
group (p<0.05). Next, a preliminary tumor stimulation assay was run comparing SCC61 T+ (Figure 5B)
and SCC61 T- tumor cells as Ag sources (Figure 5C), and investigating the role of platelets in antigen-
specific T cell stimulation. Following PBMC isolation, a portion of PBMC was platelet-depleted, and
additionally, purified monocytes with and without platelets were prepared from PBMCs. 5x106 PBMCs
with or without platelets were pulsed overnight with 2.5x106 PUVA-treated tumor cells prior to co-culture
with E7(11-19) T cells. Significant IFNg release (p<0.0001) exclusively required the presence of platelets
and an E7+ antigen source (Figure 5B). This preliminary experiment suggested that Ag from tumor cells
could be efficiently cross-presented to anti-tumor T cells.
Following the initial confirmation of successful E7(11-19) T cell line stimulation with peptide and cellular
antigens, a larger experiment investigating the role of TI (ie plate passage and overnight incubation) and
platelets (PLT) in peptide and tumor Ag presentation was conducted (Figure 6). First, corresponding TI
permutations of Ag groups were compared to that of the null group; the statistically significant
differences are mentioned below. Negative control groups (null and SIINFEKL) exhibited similarly
(p>0.78) minimal IFNg release across all tested TI groups. In contrast, positive control SP produced over
10x the maximum IFNg observed in the null group (red vertical asterisks; p<0.0001). With respect to LP,
PLT-/TI- and PLT+/TI- groups demonstrated significantly elevated IFNg (red vertical asterisks;
p<0.0001). Finally, the high E7-expressing SCC61 T+ tumor cell group demonstrated significantly
elevated IFNg production in the PLT+/TI- (p=0.043) and PLT+/TI+ (p<0.0001) setups. Thus, groups
containing E7+ peptide and tumor cells demonstrated antigen-specific CD8 T cell stimulation, as shown
in the preliminary experiment.
To account for potential tumor-influenced inflammatory cytokine production47, IFNg profiles of the E7
expressing groups were compared to those of the negative control cell line, SCC61 T-. The SCC90 group
produced significantly less IFNg (p=0.039) in the PLT-/TI+ permutation but otherwise produced
comparable levels to SCC61 T- (p>0.09). Notably, “full TI” SCC61 T+ (ie PLT+/TI+) produced
significantly higher IFNg than its corresponding SCC61 T- group (red vertical asterisks; p<0.0001).