Expansion of Candidate HPV-Specific T Cells in the Tumor Microenvironment during Chemoradiotherapy Is Prognostic in HPV16⁺ Cancers

Radiotherapy, a cornerstone of treatment for cancers related to infection with human papillomavirus (HPV), functions as an in situ vaccine, generating systemic antigen-specific immunity to tumor-specific antigens (1). Chemoradiotherapy can increase antitumor immunity in immunotherapy-resistant patients, such as those with disease progression during checkpoint inhibitor therapy (2, 3). Preclinical studies have demonstrated that radiotherapy plays a role in enhancing the diversity and abundance of the MHC type I (MHC-1) peptide repertoire, increasing the presentation of cancer neoantigens to T cells, promoting the infiltration of lymphocytes into the tumor, and augmenting the activation of cytotoxic T cells (4–6). However, it is still unclear whether these complex processes have clinical relevance in response to radiotherapy.

In this study, we used HPV-related cancers as a model for investigating antigen-specific immune response to chemoradiotherapy. HPV-related cancers are unique in that they express the viral oncoproteins E6 and E7, which can serve as tumor-specific antigens (7, 8). HPV's integration into the host's cellular genome results in the stable expression of E6 and E7, which inhibits the function of p53 and Rb. This subsequently leads to transformation of the mucocutaneous epithelium into dysplastic and ultimately malignant cells (9). HPV-related squamous cell carcinomas are among the most exquisitely radiation-sensitive solid tumors (10) and can often be eliminated with radiotherapy or chemoradiotherapy alone without requiring surgical resection. Thus, HPV-related squamous cell carcinomas are ideal models for studying the effect of chemoradiotherapy on antigen-specific immune responses.

In the current study, we quantified antigen-specific immune activation during chemoradiotherapy by identifying and tracking candidate HPV-responsive T cells in both the tumor microenvironment (TME) and peripheral blood among a large cohort of patients with diverse HPV-related cancers. We used functional HPV antigen stimulation assays to identify candidate HPV-specific and HPV-responsive T-cell clones after which we periodically followed their presence and proportions in serially sampled tumor and blood from patients with HPV-related cervical, anal, vulvar, or vaginal cancer. We analyzed the correlation between these clones and survival. We also used multiparametric flow cytometry and T-cell receptor (TCR) sequencing to quantify immune subsets, which were subsequently analyzed for associations with survival after chemoradiotherapy.

Eighty-six patients with biopsy-confirmed locally advanced cervical (N = 67), vaginal (N = 4), vulvar (N = 1), or anal cancer (N = 14) were enrolled on a multi-institutional prospective tissue banking protocol for patients treated with standard-of-care chemoradiotherapy at The University of Texas MD Anderson Cancer Center (Houston, TX) and the Harris Health System Lyndon B. Johnson Hospital Oncology Clinic from September 22, 2015 to January 11, 2019. Cases of gynecologic and anal cancers were staged according to the FIGO 2009 staging system and the American Joint Committee on Cancer 7th edition staging system respectively, both of which were relied on when the study began. Patients were required to have a visible tumor and planned standard-of-care treatment for intact cancer. All patients received external beam radiotherapy (EBRT) with intensity modulated radiotherapy for approximately 5–7 weeks with concurrent chemotherapy per institutional standard. Total dose and method of boost, brachytherapy or EBRT, differed according to primary tumor site. For cervical cancer, patients received a brachytherapy boost at completion of treatment, per standard of care. Patients with vaginal cancer received a brachytherapy boost when anatomically feasible, otherwise they received EBRT. This boost was delivered after all on-treatment samples were collected. Patients with vulvar or anal cancers received end-of-treatment boosts with EBRT. All patients with gynecologic cancers received weekly cisplatin at a radiosensitizing dose of 40 mg/m². Patients with anal cancer received 5-fluorouracil in addition to cisplatin, per institutional standard. Patients with any previous pelvic radiotherapy were excluded from the study. Patients receiving EBRT were exposed to a minimum radiation dose of 45 Gy in 25 fractions over 5 weeks. Clinical, demographic, and pathology data were collected prospectively. The patient study was approved by the MD Anderson Institutional Review Board (2014-0543), in compliance with the Belmont Report and applicable ethics guidelines. All patients provided written informed consent.

Radiation or gynecologic oncology clinicians collected blood and tumor swabs and brushes from 86 patients at five time points: baseline, week 1 (after 5 radiotherapy fractions), week 3 (after 10–15 radiotherapy fractions), week 5 (within 1 week before, or at the time of first brachytherapy), and first follow-up (∼12 weeks after treatment). Blood samples were collected from 42 patients and tumor samples were collected from 85 patients. For 2 of the patients with cervical cancer, additional tumor brushes were collected at week 5 for functional expansion of antigen-specific T cells before DNA extraction and TCR sequencing.

For blood samples, a total of 10–20 mL of blood were collected into 10-mL EDTA-containing vacutainers (BD Biosciences, 366643) and were transported at room temperature to the lab within 4 hours. In the lab, sterile PBS was added to the blood in the vacutainers at a 1:1 ratio. The diluted blood was then layered in new conical tubes at a 2:1 ratio onto Lymphoprep (Stemcell Technologies, 07851; or Cosmo Bio USA, AXS1114545) and centrifuged at 400 × g, with the brakes off, for 40 minutes at room temperature. The peripheral blood mononuclear cells (PBMC) were isolated with a serological pipette, placed into new conical tubes, and rinsed once in sterile PBS and twice in sterile complete RPMI1640 media. PBMCs were centrifuged at 400 × g for 10 minutes at room temperature after each wash. Finally, PBMCs were counted using a hemocytometer and aliquoted. Aliquots of 5 × 10⁵ cells were pelleted and stored at −80°C until DNA extraction before TCR sequencing. Aliquots of 0.5 × 10⁵ cells per treatment were immediately used for functional expansion of antigen-specific T cells and flow cytometry.

Tumor swabs were collected with an Isohelix Buccal Swab (Isohelix, DSK-50) by swabbing the tumor and immediate region. Swabs were placed into individual collection tubes and transported at room temperature to the lab within 4 hours. In the lab, 400 μL of stabilization buffer (Isohelix, BFX-25) were added to each tube, which were vortexed for 15 seconds and stored at −80°C until DNA extraction before TCR sequencing and HPV genotyping.

Tumor brushes were collected as described previously (11). Briefly, two Cytobrush Plus Endocervical Samplers (Cooper Surgical, C0012) were rotated against the tumor and immediate region. Brushes were placed into individual conical tubes and immediately transported at room temperature to the lab. In the lab, 10 mL of sterile complete RPMI1640 media, containing 1% penicillin-streptomycin and gentamicin antibiotics (Thermo Fisher Scientific, SH30027FS, SV30010, and BW17-518Z, respectively) and 10% FBS (Mediatech, MT35010CV), was added to each of the tubes, which were then vortexed for 1 minute to dislodge and suspend cells. When large amounts of mucus were present, 5 mL of dithiothreitol solution (1X Hank balanced salt solution, 4% BSA, 2 mmol/L dithiothreitol; Invitrogen, P2325) were added to the cell suspensions, which passed through a 70-μmol/L cell strainer into new conical tubes. Cells were pelleted by centrifugation and resuspended in sterile complete RPMI1640 media to be immediately used for functional expansion of antigen-specific T cells and flow cytometry. For the additional tumor brushes collected at week 5 from the 2 patients with cervical cancer, cells were immediately used for functional expansion of antigen-specific T cells before DNA extraction and TCR sequencing.

DNA was extracted from aliquoted PBMCs with the DNeasy Blood and Tissue Kit (Qiagen, 69504) following the manufacturer's instructions for erythrocytes, with all centrifugation performed at room temperature. Briefly, 5–10 μL of thawed PBMCs were placed into 2-mL tubes and 20 μL of Proteinase K and enough sterile PBS were added to bring the total volume to 220 μL. Next, 200 μL of Ethanol-free Buffer AL were added to each of the tubes, which were vortexed and incubated for 10 minutes at 56°C. Following incubation, 200 μL of ethanol were added to each of the tubes, which were again vortexed, and then the solutions were transferred into individual spin columns. Columns were placed into new 2-mL collection tubes and centrifuged at 6,000 × g for 1 minute. Columns were placed into new collection tubes, 500 μL of Buffer AW1 were added, and the tubes were centrifuged at 6,000 × g for 1 minute. This was done once more, with the final centrifugation at 20,000 × g for 3 minutes to completely dry the column membranes. The collected liquid and the two collection tubes were discarded after each centrifugation. Finally, the dried columns were placed into new 2-mL microcentrifuge tubes, 200 μL of Buffer AE that had been incubated at room temperature for 1 minute, were added to the membrane, and the tubes were centrifuged at 6,000 × g for 1 minute. Resultant DNA isolates were used for TCR sequencing.

DNA was extracted from tumor swabs with the Isohelix Xtreme DNA Isolation Kit (Isohelix, XME-50) within 1 year of collection. Swabs were thawed at room temperature and 20 μL of Proteinase K were added to the collection tubes, which were then vortexed and incubated for 1 hour at 60°C. The manufacturer's instructions were followed for the remaining steps. Resultant DNA isolates were used for TCR sequencing and HPV genotyping.

DNA was extracted from the additional tumor brushes after functional expansion of antigen-specific T cells was performed with the QIAamp UCP DNA Micro Kit (Qiagen, 56204) following the manufacturer's instructions. Samples and buffers were brought to room temperature before centrifugation and all centrifugation was performed at room temperature. Briefly, cells were suspended in 100 μL of sterile complete RPMI1640 media with 10 μL of Proteinase K and 100 μL of Buffer AUL, vortexed, and incubated at 56°C for 35 minutes. Next, 50 μL of ethanol were added and tubes were vortexed and incubated for 3 minutes at room temperature. Following incubation, the solutions were transferred to individual QIAamp UCP MinElute columns in 2-mL collection tubes and centrifuged at 17,800 × g for 5 minutes. Columns were placed into new collection tubes and centrifuged again. The solutions were then transferred to new columns and 500 μL of Buffer AUW1 were added to the columns, which were centrifuged and placed into new collection tubes. The same was done with 500 μL of Buffer AUW2. Next, 100 μL of microbial DNA–free water were added to the columns, which were incubated for 10 minutes at room temperature and centrifuged. This was repeated twice more with 100 μL of Buffer AUE and 50 μLs of Buffer AUE. Resultant DNA isolates were stored at 4°C for less than 1 month until sent for TCR sequencing.

Three aliquots each of PBMCs (0.5 × 10⁵ cells) and cells isolated from tumor brushes (maximum of 1 × 10⁶ cells) were suspended in 100 μL sterile complete RPMI1640 media and placed into individual wells of a 96-well plate. Commercially generated synthetic long peptide sequences (HPV-SLP) of nine E6 and four E7 domains (Supplementary Table S1; Biosynthesis Inc; refs. 12–17) were pooled and added to wells of treatment groups at a final concentration of 10 μg/mL before plates were incubated at 37°C overnight (12–16 hours). The following day, a cell activation cocktail containing phorbol 12-myristate 13-acetate (PMA)/ionomycin (BioLegend, 423301/2) was added to wells of positive control groups and Golgiplug (BD Biosciences, 555029) was added to all wells. The wells containing only cells, media, and Golgiplug served as negative control groups. Plates were incubated at 37°C for 4–6 hours to allow IFNγ to accumulate before staining. Finally, cells were immunostained and analyzed via flow cytometry as described in the below section. For the expanded cells from the additional tumor brushes collected at week 5 from the 2 patients with cervical cancer, cells were harvested for DNA extraction and TCR sequencing.

Lymphocyte immunostaining was performed according to standard protocols. Briefly, cells were fixed using the FOXP3/Transcription Factor Staining Buffer Set (eBioscience, 00-5523-00) and stained with an 18-color antibody panel from BioLegend, BD Biosciences, eBioscience, and Life Technologies (Supplementary Table S2) for 30 minutes at 4°C, and then held in flow cytometry staining buffer (2 mmol/L EDTA, 2% FBS; Corning, MT35010CV). Counting beads (Thermo Fisher Scientific) were used for single-color controls. The cells were analyzed using a 5-laser, 18-color LSRFortessa X-20 Flow Cytometer (BD Biosciences) and FlowJo 10.6.1. The flow gating strategy is listed in Supplementary Fig. S1.

PBMC, tumor swab, and expanded tumor brush DNA isolates were sent for TCR sequencing at the Cancer Genomics Laboratory at The University of Texas MD Anderson Cancer Center (Houston, TX). Multiplex PCR-based deep sequencing of the CDR3 region of TCRβ was performed using the Adaptive Biotechnologies immunoSEQ human T-cell receptor beta (hsTCRB) Kit, Version 3 (Adaptive Biotechnologies, ISK10101). The system uses a library of known forward primers, each specific to a TCR Vβ segment, and reverse primers specific to a TCR Jβ segment.

Survey resolution sequencing was performed for the tumor swab DNA and the tumor cytobrush cells that underwent antigen-specific T-cell expansion. Deep resolution sequencing was performed for the DNA extracted from blood PBMCs. For survey resolution two replicates of 200 ng DNA per sample, and for deep resolution six replicates of 200 ng DNA per sample, were prepared for qPCR with the V- and J-gene specific primers provided in the immunoSEQ hsTCRB kit and the QIAGEN Multiplex PCR Kit (Qiagen, catalog no. 206145). First, 31 cycles of qPCR were performed on all replicates, then, sample manifest barcodes generated with immuSEQ Analyzer and Illumina adapters were added to each PCR replicate for eight additional qPCR cycles. The libraries were purified using a bead-based system to remove residual primers, pooled at equal volume, and checked for quality control with Agilent D1000 screen tapes to determine the size-adjusted concentration. The libraries were quantified with the Applied Biosystems QuantStudio 6 and the KAPA Biosystems library quantification kit, using manufacturer's instructions.

On the basis of the qPCR results, approximately 15 pmol/L of the pooled libraries were loaded onto the Miseq Sequencing System for a single end read which includes a 156-cycle Read 1 and a 15-cycle Index 1 read run. Raw sequences output from the Miseq were transferred to Adaptive's immunoSEQ Data Assistant, where the data were processed to report the normalized and annotated TCRB repertoire profile for each sample.

Tumor swab DNA isolates were applied to the Linear Array HPV Genotyping Test and Linear Array Detection Kit (Roche, 04472209 190 and 03378012 190, respectively).

The Working Master Mix (MMX) was prepared by adding 125 μL of HPV Mg²⁺ to one vial of HPV MMX and mixing by inversion 10–15 times. Then, 50 μL of Working MMX were combined with 50 μL of isolated DNA in each reaction tube. Amplification was performed in an Applied Biosystems Gold-plated 96-Well GeneAmp PCR System 9700 with the following program: HOLD 2 minutes at 50°C; HOLD 9 minutes at 95°C; CYCLE (40 cycles, ramp rate 50%) 30 seconds at 95°C, 1 minute at 55°C, 1 minute at 72°C; HOLD 5 minutes at 72°C; HOLD 72°C indefinitely. Less than 4 hours after amplification, 100 μL of Denaturation Solution were added to the amplification products and mixed by pipetting.

For the hybridization reaction, HPV Strips containing probes were placed in wells of a 24-well tray (Roche, 03140725 001). Working Hybridization Buffer (100 mL SSPE, 12.5 mL SDS, and 388 mL deionized water) and 75 μL of denatured amplicon were added to each well. The tray was hybridized in a shaking water bath at 53°C for 30 minutes with a shaking speed of 60 RPM. As of this step, each buffer was removed from the strips by vacuum aspiration. The strips were first washed with 4 mL of Working Ambient Wash Buffer (133 mL SSPE, 13.3 mL SDS, and 2,520 mL deionized water) by rocking 3–4 times, followed by 4 mL of Working Stringent Wash Buffer in a shaking water bath at 53°C for 15 minutes with a shaking speed of 60 RPM.

To begin the detection process, 4 mL of the Working Conjugate were added to each well and incubated for 30 minutes at room temperature on an orbital shaker at 60 RPM. To wash the conjugate off the strips, three rinses were performed by adding 4 mL Working Ambient Wash Buffer. For the first rinse, the tray was rocked gently 3–4 times. However, for the second and third rinses, the tray was shaken at 60 RPM on an orbital shaker for 10 minutes at room temperature. Afterward, 4 mL of Working Citrate Buffer (25 mL CIT and 475 mL deionized water) were added to each well and the tray was shaken at 60 RPM on an orbital shaker for 5 minutes at room temperature. After removing the final buffer by vacuum filtration, 4 mL of Working Substrate (4 mL SUB A and 1 mL SUB B) were added to the wells and the tray was shaken at 60 RPM on an orbital shaker for 5 minutes at room temperature. The substrate was aspirated from the wells and a final rinse of 4 mL deionized water was applied to each well containing a strip.

Each strip was removed from the tray with forceps and dried on a clean surface for 1 hour. Results were interpreted by aligning each strip with the Linear Array HPV Genotyping Reference Guide. HPV genotypes corresponding to positive bands on the strips were recorded for each sample. The results were validated by confirming that the negative control (included in the kit) showed no bands, the positive control (included in the kit) showed bands for HPV16, and the β-globin internal controls were present on all sample strips and the positive control strip.

The TCR metrics we studied were Total Templates, Productive Templates, Total Rearrangements, Productive Rearrangements, Productive Clonality, Sample Clonality, Productive Entropy, Max Productive Frequency, Max Frequency, and Out of Frame Rearrangements. Both productive templates and nonproductive templates (corresponding to CDR3 regions predicting out-of-frame receptor genes or premature stops) were assessed, but only the productive templates were included in the final analysis. To study changes in the flow characteristics over time, we compared the means for blood and tumor samples at baseline and each subsequent time point by using a paired sample t test. We compared median changes from baseline for TCR characteristics by using a Wilcoxon signed-rank test. We calculated the log₂ fold change of any variable in blood or tumor samples that underwent significant changes from baseline. We then performed a Wilcoxon signed-rank test to assess the difference in degree of change between the blood and tumor samples. We also fit univariate Cox proportional hazards models for each of the flow and TCR variables at the static time points as well as for the fold changes in blood and tumor samples. Clustering of immune variables with and without clinical characteristics was performed using both the machine learning algorithm “mclust” (18) and unsupervised hierarchical clustering. HPV remodeling was defined as a >1.5-fold change in proportions of HPV-responsive clones from baseline to week 5 of chemoradiotherapy.

Public and exclusive repertoires were created for HPV peptide, PMA, and controls based on amino acid overlap in CDR3 sequences. Each sequence was annotated using the mcPAS (19), VDJdb (20), and TBAdb (21) databases, downloaded on March 24, 2020. The R package immunarch (22) was then used to monitor the frequency of HPV-specific CDR3 amino acid sequences over time as well as the most abundant clones overall. We also calculated the numbers of unique samples and patients that had each HPV-specific clone. To study potential changes to the HPV-specific repertoire over time, we grouped clones from 2 patients based on their presence in HPV, PMA, and control wells into seven groups: HPV⁺PMA⁺CTRL⁺, HPV⁺PMA⁺CTRL⁻, HPV⁺PMA⁻CTRL⁺, HPV⁺PMA⁻CTRL⁻, HPV⁻PMA⁻CTRL⁺, HPV⁻PMA⁺CTRL⁻, and HPV⁻PMA⁺CTRL⁺. We calculated the number and proportion of clones present in each sample and defined the clones' presence as a binary variable. We then compared the number and proportions of clones in blood and tumor samples for all seven groups at each time point to those at baseline by using the Wilcoxon signed-rank test. We calculated the log₂ fold change from baseline to determine whether the degree of change was associated with survival. We also assessed whether the number and proportion of clones in these groups' blood and tumor samples were associated with overall survival (OS) and recurrence-free survival (RFS) by using a univariate Cox proportional hazards model.

We performed the full analysis using subsets of data stratified by cancer type (anal vs. cervical, vaginal, and vulvar), histological category (squamous cell carcinoma vs. adenocarcinoma and adenosquamous carcinoma), and HPV⁺ and HPV16⁺ subsets.

Relevant motifs in the CDR3β portion of patient T-cell sequences in comparison to expected frequencies in a reference set of unselected naïve TCRs were identified using the GLIPH (Grouping of Lymphocyte Interactions by Paratope Hotspots) algorithm (23). A local convergence minimum probability score cutoff of 0.001 and local convergence minimum observed versus expected fold change of 10 was used. A simulated resampling depth of 1,000 was used. A minimum motif length of 3 was set, and discontinuous motifs were excluded.

Although we used immunoSEQ TCR sequencing to sequence the β chain of the CDR3 complex, the TCR complex contains both α and β chains as well as the signaling molecules CDδ, CD3γ, CD3ϵ, and CD3ζ. GLIPH is a computational algorithm that predicts significant motif lists and convergence lists of motifs based on individual TCR sequences, including known HLA type when available. The software provides enrichment for each motif, V-gene, CDR3 length, and proliferation counts.

We built univariate Cox proportional hazard models for the counts and proportions of each motif at static time points as well as for the dynamic changes in the motifs from baseline. We also tested for baseline differences in the counts and proportions of the motifs by way of an ANOVA test and conducted post hoc comparisons with a Bonferroni adjustment to identify differences between motifs. In addition, we compared the extent of motif presence or their absence at baseline for each patient and assessed whether they experienced an increase or decrease in motif proportion by week 5 using individual Fisher exact tests. We calculated the overall P value for motifs comparison. The significance was adjusted for the 10 comparisons by dividing the type I error of 0.05 by 10. We tested for correlations between clinical characteristics and CD8⁺ (% Live Lymphocytes), CD4⁺ (% Live Lymphocytes), and TCR characteristics in tumor and blood samples. Age and body mass index (BMI) were each fitted in a simple linear regression model. We used a Wilcoxon signed-rank test to test for associations for nodal status, stage, and histology. Statistical significance was set at an α of 5% for a two-sided P value. All available samples were used for analyses. Analyses were conducted using RStudio 1.2.5033 Orange Blossom (24).

The data generated in this study are publicly available in immuneACCESS at https://doi.org/10.21417/LC2022CIR.

TCR sequencing data generates innumerable potential TCR sequences for each sample, any of which could be HPV specific. To the best of our knowledge, public databases contain limited data on HPV-related sequences. Performing T-cell assays with peptide stimulation allows the identification of candidate clones that are HPV-reactive based on proliferation and enrichment of antigen-specific T-cell clones. We hypothesized that T-cell clones that are specific for HPV antigens could be identified on the basis of ex vivo expansion following incubation with HPV peptides, PMA, or media alone (control). Two HPV16⁺ patients with different HLA types were used for ex vivo expansion experiments and their clones were pooled to identify expanded T-cell clones and scan for these in a larger patient population. Unique clones were identified in each treatment condition for both patients; a minority of clones were identified in multiple samples (Fig. 1A–C). This suggests that the incubation conditions altered the subset of intratumoral clones that proliferated ex vivo. Seven potential categories of clones were identified; T-cell clones identified in HPV peptide–treated cultures (HPV⁺PMA⁻CTRL⁻, n = 1,113; HPV⁺PMA⁺, n = 108; and HPV⁺CTRL⁺, n = 64) were considered “HPV-responsive,” and T-cell clones identified in only the HPV⁺ wells (HPV⁺PMA⁻CTRL⁻; n = 1,113) were considered “HPV-specific.” Clones present in both experimental patients defined as clones with “contradicting” results (negative in 1 patient, positive in another), were not included in the pools of HPV-responsive or HPV-specific clones. Annotations were performed for all known TCR databases and did not yield known HPV-related TCR clones. However, identification and annotation of known HPV-related clones in public TCR databases are extremely limited and thus annotations were not utilized for this analysis.

The overall study design is shown in Fig. 1D for the 2 patients from whom HPV-reactive T cells were identified. Clones identified in the 2 patients (Fig. 1C) were then applied to data from the larger cohort (Fig. 1D). The approach of using individual ex vivo HPV peptide stimulation assays to identify HPV-reactive T cells was initially validated in 1 patient by searching the patient's TCR sequencing data in their corresponding serial tumor samples (Fig. 1E and F). Distinct kinetics were observed for candidate HPV-specific clones as compared with the 100 most abundant clones overall. Candidate HPV-specific clones initially decreased before expanding by week 5, whereas the relative frequency of the most abundant clones overall increased in week 1 and then declined through week 5.

Next, we sought to determine whether these candidate HPV-reactive clones were specific to the patients used in the initial ex vivo expansion experiment or present in other patient samples as well. A total of 86 total enrolled patients had locally advanced HPV-related cancers amenable to definitive chemoradiotherapy. Of these patients, 67 (78%) had cervical cancer, 14 (16.3%) had anal cancer, 4 (5%) had vaginal cancer, and 1 (1%) had vulvar cancer. Patient and tumor characteristics are summarized in Table 1, and samples collected at each time point are described in Supplementary Table S3. HPV and HLA typing for all study patients, including the 2 patients from whom clones were identified, are presented in Supplementary Table S4. Most patients had squamous cell carcinoma (84.9%), positive nodal status (69.8%), and stage II or III disease (38.4% and 33.7%, respectively). Staging systems differed among disease types. Viral strains identified in patients in whom HPV viral DNA was detected included HPV16 (58.6%) and HPV18 (12.9%). This is consistent with prior reports of HPV genotyping for HPV-related cancers (25–27).

Individual clones were present across multiple time points in a median of 9 patients (range, 0–35 patients for unique clones) and at any time point in a median of 12 patients (range, 0–65 patients). The most commonly identified clones are shown in Fig. 1G. The most recurring clone in tumors, CASSLGQGYEQYF, was identified in 30 tumor samples and 28 peripheral blood samples. The second most frequently identified clone in tumors, CASSLGGTEAFF, was identified in peripheral blood samples (n = 36) more often than in tumor samples (n = 26). These two clones were present at baseline in multiple patients in peripheral blood (n = 18; Supplementary Fig. S2A) and tumor (n = 62; Supplementary Fig. S2B), but their overall proportion was quite low in both blood (maximum proportion = 0.0005; Supplementary Fig. S2C) and tumor (maximum proportion = 0.006; Supplementary Fig. S2D).

Next, we reasoned that if the candidate HPV-reactive T-cell clones recognize HPV antigens, they may uniquely expand in the TME during chemoradiotherapy as a result of increased HPV antigen presentation induced by chemoradiotherapy. Non–HPV-specific PMA-responsive clones (PMA⁺; either HPV^+/− or CTRL^+/−) showed no change from baseline in either blood or tumor (Fig. 2A). Counts of HPV-responsive (HPV⁺; either PMA^+/− or CTRL^+/−) clones increased in blood by week 5 (P < 0.05; Fig. 2B) and increased by a greater degree in tumor at each time point and by week 5. Overall, T-cell clonality slightly increased in blood, but decreased in tumor, from baseline to week 5 (Fig. 2C). Only the most HPV-specific group (HPV⁺PMA⁻CTRL⁻), had distinct expansion kinetics in tumor versus blood (Fig. 2D and E), which might account for this difference. Table 2 demonstrates changes from baseline to week 5 in tumor for all T-cell clones, including an increase in the median number of HPV-specific T-cell clone counts from 0 at baseline to 2+ at week 5 (P < 0.01). Non-specific T-cell counts decreased in both blood and tumor from baseline to week 5 (P = 0.02 and P < 0.01, respectively). Supplementary Table S5 demonstrates changes for all T-cell repertoires in blood and tumor over time. Although HPV-specific T-cell counts decreased in blood (6 vs. 1; P < 0.01), they expanded in tumor (0 vs. 2; P < 0.01). The median proportion of HPV-specific T-cell clones in tumor was 0.025% at baseline and 0.044% at week 5, whereas the median proportion of HPV-specific T-cell clones in blood was 0.05% at baseline and 0.034% at week 5. Subset analyses of all HPV⁺ patients and only HPV16⁺ patients showed that by week 5, HPV⁺ clones increased significantly in tumor (HPV⁺ subset: 0.0009 vs. 0.0007, P = 0.014; HPV16 subset: 0.001 vs. 0.0005, P = 0.01) but remained unchanged in blood (HPV⁺ subset: P = 0.42; HPV16 subset: P = 0.5).

Common short amino acid motifs are present within TCRs, which recognize matching or partially matching T-cell antigens. We thus hypothesized that motifs may be present in our candidate HPV-reactive TCRs. We identified five amino acid motifs among HPV antigen–stimulated T-cell populations (minimum P = 0.001). Of these five motifs [QSRANV (Fig. 3A), KTYG (Fig. 3B), GTRF, PIW, and RHH], three (QSRANV, KTYG, and GTRF) were not found in control conditions but were enriched in response to HPV peptides and, to a lesser degree, in response to PMA [QSRANV, 13/4,284 template clones (0.30%) for HPV peptides vs. 6/7,294 clones (0.08%) for PMA; KTYG, 6/4,284 clones (0.14%) vs. 4/7,294 clones (0.05%); GTRF, 4/4,284 clones (0.09%) vs. 1/7,294 clones (0.01%)].

The two most differentially enriched motifs, QSRANV and KTYG, were also the most common among our patient population as they were identified in 40 and 33 patients, respectively (P < 0.01), as compared with 13 for PIW, 9 for GTFR, and 7 for RHH (Supplementary Table S6). The QSRANV motif was present in higher proportion per sample than all other motifs with the exception of the KTYG motif (P < 0.01). These motifs were also more likely to proportionally increase throughout chemoradiotherapy (as noted by an increase in each patient; Fig. 3C and D). The proportion of the QSRANV motif increased significantly more than all others (P < 0.01 for all). Conversely, 12 of 17 patients with the PIW motif witnessed a decrease in the motif proportion in the T cells in the TME, whereas only around 50% of patients with the GTFR and RHH motifs at baseline underwent an increase over time. We also compared the presence of these HPV-specific T-cell clones in a cohort of patients with ovarian cancer not related to HPV infection with all of our cohort of patients with HPV-related cervical cancer. Our analysis determined that the two most differentially enriched motifs, QSRANV and KTYG, were more frequent in HPV-related cervical cancers in comparison with non–HPV-related ovarian cancers (Fig. 3E, P = 0.0013 and Fig. 3F, P = 0.0066, respectively). This supports our hypothesis that these motifs are present in T-cell populations that recognize HPV-derived antigen.

When the CDR3β receptor sequences for these enriched motifs were cross-referenced to known pathogenic TCR sequences using the McPAS-TCR database (19), epitopes to both QSRANV and KTYG mapped to a conserved region of an essential viral oncogenic domain that disables the Rb tumor suppressor. This same domain is necessary for the association of the HPV oncoprotein E7 with the Rb tumor suppressor protein and therefore is a key mediator in the oncogenic transformation of HPV-related cancers (28), validating the finding that these motifs are likely E7 related.

To determine whether expansion of HPV-specific T cells within the TME was associated with survival, we performed univariate Cox proportional hazard modeling for RFS and OS with P-value correction for multiple testing. We found no significant correlation between survival and the count, proportion, or presence of HPV-specific or HPV-responsive T-cell clones at any time point (Supplementary Table S7). Fold changes in HPV-specific or HPV-responsive T-cell clone populations in the TME or blood from baseline to week 5 also were not clearly associated with RFS or OS. Because the clones identified were generated from HPV16⁺ patients, and not all patients in the cohort were HPV16⁺, we also performed a subset analysis exclusively among HPV16⁺ patients with baseline and week 5 samples (Supplementary Table S8; n = 28). We also performed a subset analysis including HPV⁺ patients (Supplementary Table S9) In the subset of HPV16⁺ patients only, increased fold change in HPV-responsive clones was significantly associated with improved RFS [HR, 0.39 (95% confidence interval, CI, 0.18–0.85); P = 0.02] (Supplementary Fig. S3), suggesting that any survival benefit is limited by the HPV subtype.

The static counts, proportions, and fold changes over time of E7-related motifs were not significantly associated with RFS or OS (Supplementary Table S10). Higher fold changes in only one motif, PIW, from baseline to week 5 were significantly associated with improved RFS in the TME [HR, 1.04 (95% CI, 1.01–1.07); P = 0.007]. The PIW motif, which was enriched in HPV-stimulated populations as compared with control populations (0.09% vs. 0.03%) but absent in PMA-stimulated populations, may represent an HPV-specific but non-reactive motif. Repeating this analysis for subsets of HPV⁺ and HPV16⁺ patients yielded consistent findings.

We also profiled the intratumoral and peripheral immune milieu during chemoradiotherapy to identify changes associated with HPV-specific T-cell remodeling. Multiparametric flow cytometry was used to assess samples from 86 patients after stimulation with either HPV peptides or media (control). Markers included those of lymphocyte lineages, which comprise functional markers of activation [CD69, IFNγ, granzyme B (Grzb), Ki67] and exhaustion (CTLA4, PD-1). HPV antigen–specific T-cells were analyzed with intracellular staining for IFNγ-producing CD8⁺ T cells following stimulation with overlapping peptides spanning the E6 and E7 genes from HPV16.

Overall, the most substantial changes in the immune-cell composition of the TME were significantly increased percentages of PD-1⁺CD8⁺ and PD-1⁺CD4⁺ T cells at follow-up as compared with baseline (P = 0.01 and P = 0.03, respectively) (Supplementary Figs. S4 and S5; Supplementary Table S5). The expression levels of markers of early clonal priming and activation (PD-1, CD69, IFNγ, and Grzb) increased during chemoradiotherapy and peaked at either week 3 (PD-1) or at the end of treatment (CD69, IFNγ, and Grzb) and were sustained in the treated tumor at week 12 follow-up (Supplementary Fig. S4; Supplementary Table S5).

In peripheral blood, the percentage of CD8⁺ T cells with the activation markers CTLA-4 and Ki67 increased during chemoradiotherapy (week 3 for CTLA-4, P = 0.03; week 3 and week 5 for Ki67, P = 0.04 and P = 0.04, respectively) as compared with baseline (Supplementary Fig. S5; Supplementary Table S5). There was no significant change in HPV peptide–responsive IFNγ⁺CD8⁺ expression in blood or TME between the start of treatment and week 5. Testing for differences in the fold changes of certain markers between blood and tumor showed that productive clonality of T-cell repertoires increased in blood and decreased in tumor samples at week 3 (blood: 0.08, tumor: −0.97; P = 0.008) and week 5 (blood: 0.04, tumor: −0.53; P = 0.009; Supplementary Table S11).

Peptide-specific activation of T cells from either blood or tumor, measured by IFNγ⁺CD8⁺ expression, was not associated with any clinical variables (age, body mass index, nodal status, disease stage, or tumor histology; Supplementary Tables S12 and S13). Higher overall CD8⁺ counts were associated with lower BMI (P < 0.01) and negative nodal status (P < 0.01). Higher non-specific (media responsive) IFNγ⁺CD8⁺ T cells in blood was associated with younger age (P < 0.01).

We used model-based clustering for CD4⁺ and CD8⁺ markers to better describe the immune phenotypes that were associated with HPV-reactive clone remodeling (Fig. 4A). Clustering revealed several distinct clusters distinguished primarily by these same markers of early priming (PD-1) and activation (Grzb, IFNγ, Ki67; Fig. 4B). One cluster, which was characterized by a rich population of PD-1⁺CD4⁺ and PD-1⁺CD8⁺ subsets, was seen in patients with high HPV-specific remodeling at baseline (Fig. 4C) but disappeared at week 1. By week 5, these patients demonstrated CD8⁺ activation (Ki67⁺CD8⁺ cells, CD69⁺CD8⁺ cells), suggesting an “early priming/activation” pattern.

We also performed unsupervised hierarchical clustering with all clinical variables (Supplementary Fig. S6). Non-squamous cancers were enriched over squamous cancers in immune poor cluster membership versus early priming/activation cluster membership (P = 0.02), with a similar profile at all time points rather than a change over time. For patients with HPV remodeling, defined as a fold change >1.5 in HPV-specific clones in the TME from baseline to week 5, the immune phenotype shifted over time from overall activation at baseline [high CD8⁺ T cells, high CD4⁺ T cells, low T regulatory cells (Treg)] to immune exhaustion at week 1 and then to activation and memory markers at week 5 (CD8⁺Grzb⁺, CD8⁺CD69⁺). Patients with HPV remodeling also exhibited an overall shift from CD8⁺ T-cell activation to CD4⁺ T-cell memory markers.

On univariate analysis, higher populations of FOXP3⁺CD4⁺ Tregs, in peripheral blood at baseline were associated with lower RFS [HR, 1.13 (95% CI, 0.03–0.22); P = 0.01] and OS [HR, 1.12 (95% CI, 1.02–0.21); P = 0.02]. Higher populations of exhausted T cells (CTLA4⁺CD4⁺ cells) in blood samples at baseline were also associated with poorer RFS [HR, 1.29 (95% CI, 0.03–0.48); P = 0.027] and OS [HR, 1.31 (95% CI, 1.05–0.5); P = 0.02] (Supplementary Table S7). Higher populations of activated T cells (IFNγ⁺CD8⁺ cells) in peripheral blood at the end of treatment (week 5) were associated with improved RFS [HR, 2.12 (95% CI, 1.08–4.53); P = 0.03] and OS [HR, 2.49 (95% CI, 1.17–5.31); P = 0.02]. Sensitivity analyses by tumor type (anal, cervical, vaginal, and vulvar) and histology (squamous cell carcinoma vs. all others) showed no differences from the full-set analysis. Subset analyses of HPV⁺ patients and HPV16⁺ patients were also performed, but there were no significant associations between flow parameters and survival.

The key findings of our study are that (i) an ex vivo expansion approach using select HPV16⁺ patients allowed us to identify HPV-related T-cell clones that are relevant in a larger patient population; (ii) these clones selectively expand in the TME during chemoradiotherapy; and (iii) this expansion may provide prognostic information for HPV16⁺ patients after chemoradiotherapy. These findings will inform the development and testing of immune-based therapies for HPV-related cancers, especially in the up-front or definitive setting and in combination with standard-of-care chemoradiotherapy. Chemoradiotherapy affects the equilibrium of the immune system by increasing proportions of Tregs, decreasing proportions of functional, activated cytotoxic T cells, and increasing dendritic cells' presentation of tumor antigens. We were able to identify candidate HPV-reactive T cells using functional T-cell assays using tumor specimens obtained during radiotherapy. The candidate HPV-reactive TCR signature that we have identified may be useful as a biomarker to identify effective immune responses in patients receiving chemoradiotherapy as well as immunotherapy.

The HPV-specific T-cell population we identified within the TME could be a surrogate for tumor antigen–specific immune responses across cancer types. Although chemoradiotherapy can directly cause T-cell apoptosis, it can also upregulate MHC-1 expression, thereby increasing antigen presentation and activated T–cell expansion. This is particularly useful if this response is tumor antigen–specific. In addition, the activation or stimulation of IFN genes through cytosolic DNA pathways induced by radiotherapy also results in the infiltration of T cells into the tumor (6). Consequently, the addition of radiotherapy in patients receiving immunotherapy for progressive disease could spur the release tumor antigens to function as a vaccine to achieve an antigen-specific immune response and reignite immune responses (1).

Furthermore, combining chemoradiotherapy with immunotherapies targeting the HPV-specific T-cell population may improve outcomes. The T-cell populations and motifs at static time points we identified were not strongly associated with outcomes, which suggest that T-cell status alone is not sufficient to achieve a complete tumor response. Rather, HPV16⁺ patients in whom fold changes in HPV-specific T cells occurred had improved survival, suggesting that HPV-specific remodeling is more important than absolute counts or proportions of specific cell types. Santegoets and colleagues (29) also observed that the number of HPV16-specific T cells that expanded from HPV16⁺ cervical carcinomas were unable to predict survival in cervical cancers, as they conversely had in oropharyngeal cancers. The authors of that report hypothesized that this was due to low total CD4⁺ T cells, implying that without in vitro expansion, these cells were unable to impart any strong effect. This is consistent with our findings, in that patients in our study in whom these T cells did expand throughout treatment had significantly improved prognosis. Patients in whom HPV remodeling did not occur demonstrated low CD8⁺ activation throughout treatment, suggesting that these patients lack a priming response and may not significantly benefit from immune checkpoint therapy in combination with chemoradiotherapy. Although higher populations of activated cytotoxic T cells in peripheral blood at the end of treatment were associated with improved RFS and OS, higher populations of Tregs and exhausted T cells at baseline were associated with lower RFS and OS. This suggests that anti-Treg therapy or anti-CTLA4 therapy can potentially facilitate radiation's ability to stimulate antigen-specific immune response. Clustering analysis revealed that patients with high HPV-specific remodeling had one cluster characterized by a rich population of PD-1⁺CD4⁺ and PD-1⁺CD8⁺ subsets at baseline. By week 5, these patients demonstrated CD8⁺ T-cell activation (Ki67⁺CD8⁺ cells, CD69⁺CD8⁺ cells), suggesting an “early priming/activation” pattern. Therefore, this cluster may represent patients in whom the early addition of checkpoint inhibitor therapy might improve immune responses. The change of immune profiles among squamous cancers over time, but not among non-squamous cancers suggests that the immune microenvironment of squamous cancers is more susceptible than that of non-squamous cancers to remodeling by chemoradiotherapy. This helps in explaining the known increased radiosensitivity of squamous cancers over non-squamous cancers, which could potentially be exploited to improve radiosensitivity.

Little is known about the intratumoral T-cell repertoire of HPV-related cancers. Cui and colleagues (30), using TCR sequencing to characterize the peripheral T-cell repertoires of 25 patients with cervical cancer or cervical dysplasia and those of healthy patients, found that differences in spatial heterogeneity and diversity were associated with cervical cancer. However, it is important to note that the samples in that study were primarily obtained from peripheral blood, and sequencing was performed at only one time point. We also identified several of the T-cell clones Cui and colleagues (30) had mapped, as well as motifs found to be associated with Rb in other previous studies, in addition to novel sequences and motifs. Nevertheless, the known E7 clustering motifs were the most frequent in our patient population: for nearly every patient in the current study, these E7-related motifs were increased by the end of chemoradiotherapy. Changes in the known E7-related motifs were not associated with survival; however, one of the unknown motifs, PIW, was associated with RFS. This motif was less common among our patients. PIW may be an HPV-related motif which, owing to its infrequency, may be associated with survival in a larger cohort. Alternatively, it may be a non-functional HPV-responsive motif. Further study of these motifs in a larger population of patients with a specific HPV subtype is warranted. Future studies should include additional ex vivo experiments with patients presenting with other common HPV genotypes, such as HPV18 and HPV45.

One limitation of the current study was its lack of evaluation of the functional capacity of the HPV-specific T cells. Although we performed HPV antigen stimulation and flow cytometry, this approach is less sensitive than functional T-cell assays with TCR sequencing for detecting antigen-responsive T cells (31). Despite the presence of these T cells and their expansion at baseline, they could be dysfunctional or exhausted after chemoradiotherapy. Future studies should undertake long-term tracking of T cells in patients undergoing chemoradiotherapy and attempt to provide further phenotyping of HPV-specific T cells to drive engineered immunotherapy approaches.

L.E. Colbert reports grants from American Society for Clinical Oncology and MD Anderson Cancer Center during the conduct of the study. E.J. Lynn reports grants from HHS|NIH|NCI during the conduct of the study. A. Olvera reports grants from NIH during the conduct of the study. A. Jhingran reports personal fees from Genentech during the conduct of the study, as well as personal fees from Genentech outside the submitted work. L. Lin reports other support from AstraZeneca and Pfizer, and grants from NCI outside the submitted work. A.A. Jazaeri reports other support from Lovance, Bristol Myers Squibb, AstraZeneca, Aravive, Merck, and Eli Lilly and personal fees from Nuprobe, AvengeBio, Genentech-Roche, EMD-Serono, Agenus, Macrogenics, TwoXAR, and Instil Bio outside the submitted work. J.A. Wargo reports other support from Micronoma during the conduct of the study, as well as other support from Imedex, Dava Oncology, Illumina, and PeerView outside the submitted work. A. Reuben reports scientific advisory board and honoraria from Adaptive Biotechnologies. A.C. Koong is a stockholder in Aravive, Inc. E.J. Koay reports personal fees from Apollo Cancer Center, RenovoRx, Taylor and Francis, and AstraZeneca and grants from NIH, DOD, SU2C, and CPRIT outside the submitted work, as well as a patent for 3D printing of customized oral stents pending. P. Das reports personal fees from Adlai Nortye, NCI, and American Society for Radiation Oncology outside the submitted work. No disclosures were reported by the other authors.

L.E. Colbert: Study design, data acquisition, analysis and interpretation, and has drafted and approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. M.B. El Alam: Data acquisition, analysis and interpretation, and has drafted and approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. E.J. Lynn: Data acquisition, analysis and interpretation, and has drafted and approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. J. Bronk: Data analysis and interpretation and has drafted and approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. T.V. Karpinets: Study design, data acquisition, analysis and interpretation, and has drafted and approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. X. Wu: Data acquisition, analysis and interpretation, and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. B.V. Chapman: Data acquisition and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. T.T. Sims: Data acquisition, analysis and interpretation, and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. D. Lin: Data acquisition and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. R. Kouzy: Data acquisition and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. J. Sammouri: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. G. Biegert: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A.Y. Delgado Medrano: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A. Olvera: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. K.J. Sastry: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. P.J. Eifel: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A. Jhingran: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. L. Lin: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. L.M. Ramondetta: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A.P. Futreal: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A.A. Jazaeri: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. K.M. Schmeler: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. J. Yue: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A. Mitra: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. K. Yoshida-Court: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. J.A. Wargo: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. T.N. Solley: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. V. Hegde: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. S.S. Nookala: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A.V. Yanamandra: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. S. Dorta-Estremera: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. G. Mathew: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. R. Kavukuntla: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. C. Papso: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. M. Ahmed-Kaddar: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. M. Kim: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. J. Zhang: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A. Reuben: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. E.B. Holliday: Data acquisition and analysis and has approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. B.D. Minsky: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A.C. Koong: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. E.J. Koay: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. P. Das: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. C.M. Taniguchi: Data acquisition and analysis and has approved the submitted version of the article. He has agreed both to be personally accountable for his contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. A. Klopp: Was involved in the study design, data acquisition, analysis and interpretation, and has drafted and approved the submitted version of the article. She has agreed both to be personally accountable for her contributions and to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

This work was supported in part by Cancer Center Support (Core) Grant P30 CA016672 from the NCI, NIH, to The University of Texas MD Anderson Cancer Center, an American Society for Clinical Oncology (ASCO) Young Investigator Award, The University of Texas MD Anderson HPV Moon Shots Program, and a Stand Up To Cancer–Lustgarten Foundation Pancreatic Cancer Interception Translational Cancer Research Grant (Grant Number: SU2C-AACR-DT25-17). Stand Up To Cancer is a division of the Entertainment Industry Foundation. The indicated SU2C grant is administered by the AACR, scientific partner of SU2C. We also acknowledge assistance from the Department of Scientific Publications, including Joseph Munch and Chris Wogan.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.