Skip to main content
NCBI home page
Journal for Immunotherapy of Cancer logoLink to Journal for Immunotherapy of Cancer
. 2026 Feb 4;14(2):e012969. doi: 10.1136/jitc-2025-012969

Redirecting cytomegalovirus immunity against pancreas cancer for immunotherapy

Remi Marrocco 1,0, Jay Patel 2,0, Rithika Medari 2,1, Philip Salu 2,1, Eduardo Lucero-Meza 1, Catarina Maia 1,2, Simon Brunel 1, Alexei Martsinkovskiy 2, Siming Sun 2, Kevin Gulay 2, Malak Jaljuli 2, Evangeline Mose 2, Andrew Lowy 2, Chris Benedict 1,*,0, Tatiana Hurtado de Mendoza 2,✉,0
PMCID: PMC12927352  PMID: 41638871

Abstract

Background

Immunotherapy has had limited success in pancreatic cancer, largely due to a low mutational burden and immunosuppressive microenvironment. Here we hypothesized that systemic delivery of viral antigens can redirect pre-existing antiviral immunity against pancreatic tumors.

Methods

Cytomegalovirus (CMV, a β-herpesvirus) was chosen, as the majority of the population is infected and it induces an extremely large/broad memory T-cell response. Mice latently infected with murine CMV (MCMV) were orthotopically implanted with pancreatic cancer cells and treated with systemic injections of MCMV T-cell epitopes. Tumor growth was monitored by ultrasound two times a week, and immune cell infiltration was analyzed by histology, flow cytometry and single-cell RNA sequencing (scRNA-seq). Statistical analysis was performed by two-way analysis of variance with Sidak correction.

Results

MCMV peptide-epitope therapy (MCMVp) promoted preferential accumulation of MCMV-specific T cells within pancreatic tumors, delaying tumor growth and increasing survival. Immunophenotyping and scRNA-seq analyses showed these T cells were highly activated and cytotoxic, leading to increased tumor necrosis and caspase-3 activation. Depletion of CD4 and CD8 T cells abolished the impact of MCMVp therapy, indicating the antitumor response is T-cell dependent. Together, these results show that CMV-specific T cells can be repurposed to combat pancreatic cancer.

Conclusions

Our studies reveal that CMV-specific viral memory T cells can be re-directed to control a solid tumor normally refractory to immunotherapy via a simple, intravenous injection of T-cell peptide epitopes. This mutation-agnostic approach has significant potential for the development of “off-the-shelf” therapeutics by stimulating pre-existing antiviral memory, and it is widely applicable due to the high prevalence of CMV.

Keywords: Immunotherapy, Tumor microenvironment - TME, Solid tumor, Viral-specific T cells, Cytokine

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Immunotherapy has mostly failed in the treatment of pancreatic cancer, due to the aggressive nature of these tumors coupled to a low mutational burden and a highly immunosuppressive tumor microenvironment.

WHAT THIS STUDY ADDS

  • Here we show that pre-existing anti-cytomegalovirus (CMV) immune memory can be redirected to pancreatic tumors via systemic injection of T-cell peptide epitopes. Murine cytomegalovirus-peptide therapy resulted in tumor growth control, extended survival and marked tumor infiltration of activated and cytotoxic T cells, associated with profound changes of both immune and tumor cell transcriptomes.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • This study represents a tumor-mutation agnostic approach that can be applicable to the majority of the population, due to the high prevalence of CMV. Additionally, our recent identification of ~200 new human CMV T-cell epitopes will allow for a personalized approach based on a patient’s haplotype.

Background

Immunotherapy has been most successful in the treatment of tumors with high mutational burdens such as melanoma and non-small cell lung cancer.1,4 These mutations result in the generation of neoantigens that the immune system recognizes as “non-self”, eliciting an antitumor immune response that can often be enhanced with immune checkpoint therapy. Other tumors such as pancreatic cancer have a much lower mutational burden5,7 coupled with an immunosuppressive tumor microenvironment (TME).8 9 Consequently, numerous strategies to enhance antitumor immunity through checkpoint blockade and/or by modulating the TME have been largely unsuccessful.10,12 Personalized therapies where tumor biopsies are sequenced for the emergence of potential neoantigens followed by vaccination are currently being tested. However, neoantigen prediction and subsequent therapy is not trivial, and this approach is currently very expensive and time-consuming, oftentimes making it prohibitive.

Prior studies using subcutaneous tumor models by Rosato and Çuburu et al13 14 showed that intratumoral injection of antiviral T-cell peptide epitopes in mice previously infected with murine cytomegalovirus (MCMV) led to growth arrest and even complete remission. However, direct intratumoral injection is not always feasible. Here we hypothesized that antiviral T cells could also be redirected to tumors via systemic epitope delivery, which has the potential to be much more applicable to a variety of tumor types.

iRGD is a tumor targeting peptide,15,17 promoting systemic delivery of covalently linked or co-injected drugs to various solid tumors including breast, pancreatic, gastric, ovarian and others. iRGD interacts with tumor vasculature αv integrins and the NRP-1 receptor, inducing enhanced vascular permeability and delivery of co-administered chemotherapy drugs, peptides, small molecules or even antibodies.18,24 We thus tested whether systemic delivery of MCMV T-cell epitopes, together with iRGD (clinical grade form certepetide, “Cert”), could induce an antitumor immune response.

We chose to repurpose cytomegalovirus (CMV) memory T cells because (1) >80% of the world’s population is infected with human CMV (HCMV), (2) immunodominant CMV-specific memory T cells show broad tissue residency and an effector memory phenotype in people and mice and (3) ~10% of all circulating CD4 and CD8 T cells are CMV-specific.25 26 In the present study, mice were infected with MCMV until viral latency was established (>2 months), were orthotopically implanted with KPC pancreatic cancer cells, and treated with major histocompatibility complex (MHC) class I and II MCMV peptide epitopes (MCMVp therapy)±Cert. Mice responded to treatment as evidenced by enhanced survival, delayed tumor growth, increased tumor apoptosis and massive T-cell infiltration. Surprisingly, the beneficial effects of MCMVp therapy were independent of Cert. Together, these data show that mobilizing pre-existing antiviral memory T cells through systemic peptide-epitope treatment shows clinical efficacy against pancreatic tumors.

Materials and methods

Mice

Wild-type C57BL/6J (B6) mice and C57BL/6J×129S1/SVlmJ (129) F1 hybrid (B6129SF1/J, Strain #:101043) were purchased from Jackson Laboratories at 4–5 weeks of age and infected or not with MCMV at 5–6 weeks of age. This study was carried out in strict accordance with the guidelines of Association for Assessment and Accreditation of Laboratory Animal Care and National Institutes of Health (NIH). All animal protocols used in this study were approved by the Institutional Animal Care and Use Committee of the University of California, San Diego and La Jolla Institute for Immunology (LJI).

MCMV infection

MCMV (Smith strain) was originally produced in 3T3 cells from cloned and sequenced BAC DNA (gift of B Adler27) and was then amplified in vivo in BALB/c mice, extracted from the salivary glands, and titrated on MEF (murine embryonic fibroblast) cells.17 5–6 weeks old mice were infected by intraperitoneal injection with 104 Plaque-Forming Units (pfus) in 100 µL of phosphate-buffered saline (PBS). The mCherry-MCMV was kindly provided by Dr Lars Dolken.

Pancreatic tumor models

MCMV-infected mice (2–3 months post-infection) and age-matched uninfected controls were orthotopically injected with KPC1242 (syngeneic with B6 mice) or KPC46 (syngeneic with B6129SF1/J) pancreatic cancer cells. These cells were cultured in vitro in Rosewell Park Memorial Institute (RPMI) medium with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Once confluent, they were trypsinized, washed and resuspended in Matrigel (Corning Matrigel Matrix) for injection (5,000 KPC1242 or KPC46 tumor cells in 20 µL of matrigel, per mouse). A small incision (~1–1.5 cm) was made into the skin of the abdomen through the peritoneal membrane distal to the pancreas. The tail end of the pancreas was pulled out to directly inject the tumor cell suspension. Tumor growth was monitored initially by palpation and after day 8 post-injection, mice were subjected to ultrasound (Convex L20 HD3, Clarius) two times a week (under isoflurane) for tumor growth monitoring: mice were anesthetized, shaved, and gel was applied (Clear Image Gel Singles) before ultrasound monitoring. Tumor volume was calculated using the ellipsoid volume calculation formula. Mice were enrolled in treatment studies and randomly assigned to the different treatment conditions once most tumors were visible by ultrasound, usually around 9–11 days post injection (d.p.i). For survival studies, the euthanasia criteria consisted of mice losing 20% or more of their initial weight or having tumors reaching 2,000 mm3 or more plus any evident signs of distress.

Treatment studies

MCMV peptides (CD4T: m09133-147; m25409-423; m14224-38; CD8T: IE3416-423; m38316-323; m45985-993) were purchased from Mimotopes (Australia) (purity 90–96%), and ovalbumin peptides (OVA257-264; OVA323-339) were from Eurogentec-AnaSpec, California, USA. iRGD (Cert-Ac-Cys-Arg-Gly-Asp-Lys-Gly-Pro-Asp-Cys-NH2 (Cys:Cys disulfide linkage)) was acquired from LISATA Therapeutics. Mice were treated two times a week by retro-orbital (RO) injections of the indicated peptides: Dose A (50 µg CD4p/1 µg CD8p); Dose B (50 µg CD4p/0.1 µg CD8p); Dose C (10 µg CD4p/3 µg CD8p); Dose D (10 µg CD4p/1 µg CD8p). When Cert was used, mice were first injected with Cert (300 µg) in one eye followed by the mix of MCMV peptides 5 min later in the other eye. When indicated, mice were intraperitoneally injected with anti-programmed cell death protein-1 (PD-1) (10 mg/kg two times a week, Bio X Cell, 29F.1A12), anti-Interleukin 10 Receptor (IL10R) (200 µg, once a week, Bio X Cell, 1B1.3A), anti-CD8 (200 µg, two times a week, Bio X Cell, YTS 169.4), anti-CD4 (100 µg two times a week, Bio X Cell, GK1.5) antibodies, or gemcitabine (5 mg/kg, two times a week, Selleck Chem). Control mice were injected RO with vehicle (PBS/Dimethyl Sulfoxide (DMSO) equivalent to the amount in the MCMVp mix) and saline or the paired IgG controls for the antibodies by intraperitoneal injection.

Randomization and exclusion criteria: prior to treatment mice were subjected to ultrasound and assigned to a treatment group, initial tumor volumes were comparable among groups.

Mice were excluded from the study if they died prior to the endpoint (not applicable for survival studies). For flow cytometry experiments, five to six mice per group were chosen, excluding the biggest and the smallest tumors to have a representative data set.

Mice immunization

8-week-old mice were immunized subcutaneously at the base of the tail with 200 µL of an emulsion containing 50 µg of each indicated peptide (MCMVp or Ovalbumin peptides (OVAp) and Complete Freund’s Adjuvant (CFA) (0.5 mg/mL final; Sigma). Booster immunizations containing peptides+IFA were performed 3 weeks later. Mice were tumor challenged 3 weeks later as previously described.

Histology/immunohistology

On harvesting, tumors were fixed in 10% formalin for 24–48 hours and kept in 70% ethanol until paraffin embedded. Immunohistochemistry was performed as previously described.17 In short, Formalin Fixed Paraffin Embedded (FFPE) sections were de-paraffinized and subjected to steam heat-mediated antigen retrieval using low pH buffer (eBioscience). Blocking was performed with 5% donkey and 5% goat serum. The sections were stained for cleaved caspase 3 (1:250) (Cat.no. 9579S, Cell Signaling), and CD3 (1:250) (Cat.no. 11089, Abcam) and incubated with their respective secondary antibodies. Histological and immunofluorescence imaging was performed using a ZEISS AxioScan Z1 automated slide scanning microscope equipped with a 20× objective (Numerical Aperture (NA) 0.8). For brightfield imaging of H&E-stained specimens, an Light Emitting Diode (LED) illumination system was employed in conjunction with a Hitachi HV-FL202SCL camera. Immunofluorescence imaging was conducted using a Colibri7 LED illumination system and ZEISS single-band filter sets. Fluorescence images were captured using a Hamamatsu ORCA-Flash4.0 V.2 camera (Microscopy Core, LJI). Staining was quantified using the QuPath software. H&E staining was performed also as previously described,17 and necrosis scores were assessed by a histopathologist blinded to the different treatment conditions.

FACS analysis

Tumors were minced into small pieces and incubated at 37°C with rotation for 30 min in 5 mL of digestion buffer consisting of Dulbecco’s Modified Eagle Medium (DMEM) high glucose, 10% Gentle Collagenase/Hyaluronidase (STEMCELL), 10% FBS, and 10% DNaseI (1 mg/mL stock, Roche). Samples were then smashed through a filter. Flow through was spun at 300 g for 10 min and Red Blood Cell (RBC)-lysed (Pharm Lyse, BD, Biosciences). Cells were counted and subjected to magnetic separation of tumor-infiltrating CD45+ cells using EasySep TIL kit (STEMCELL). Spleen and liver cells were harvested and processed as previously described28 and analyzed on an LSR-II Fortessa X20 (BD) or a Cytek Aurora. Dead cells were excluded by staining with LIVE-DEAD (LD) Blue (Invitrogen). Surface staining was performed in Brilliant Stain Buffer (Invitrogen). Fixation/permeabilization was done with Foxp3/Transcription Factor Staining Buffer Set (eBioscience). For antibodies used, see Online supplemental table 1.

NIH-provided biotinylated monomers were tetramerized in-house with streptavidin-Allophycocyanin (APC) or streptavidin-Phycoerythrin (PE) (Agilent), and subsequently used to stain the cells as previously described.27 Tetramer staining was done at Room Temperature (RT) for 1.5–2 hours. LD and surface/intracellular staining were done subsequently. Data was analyzed using the FlowJo software. Tumor samples with less than 10% viability were excluded.

Cytokine stimulation

Total liver or purified CD45+cells from the tumors were incubated for 4 hours at 37°C 5% CO2 in 96-well plates in the presence of the indicated peptides (5 µg/mL for each peptide) and GolgiPlug (Brefeldin A; eBioscience). Control cells were incubated in the presence of GolgiPlug alone.

Single-cell RNA sequencing

MCMV-infected mice were implanted with KPC1242 pancreatic tumor cells ~3 months post-infection and treated with vehicle or MCMVp. Tumors were collected at 23 days post-tumor cell injection and total live-tumor cells (n=3 per group) or tumor-infiltrating T cells (n=3 per group) (some samples were a combination of two tumors) were sorted from the indicated groups, counted, and processed for subsequent 5’scRNAseq using Chromium GEM-X Single Cell 5’ V.3 gene expression kit. Sorting was performed on Aria-1, Aria-5, and Fusion sorter. Total live tumor cells were sorted as LD blue−. T Cells were sorted as LD blue−CD3+CD4+ and LD blue−CD3+CD8+, from a different cohort. See supplementary material for analysis pipeline.

Data analysis

Data is presented as means±SEM. The GraphPad Prism software was used for statistical analyses (V.10). Results were compared by one-way analysis of variance (ANOVA) with Tukey correction, and repeated measures by two-way ANOVA with Sidak correction.

Results

MCMV memory T cells can be redirected to fight pancreatic tumors

In order to induce an MCMV memory T-cell pool, we infected C57BL6/J (B6) mice at 4–5 weeks of age and waited a minimum of 2 months to allow the establishment of latency, defined by maintenance of the viral genome without detectable lytic replication. MCMV-infected and non-infected control mice were surgically implanted with KPC1242 tumor cells in their pancreas. Mice were treated biweekly with either vehicle (DMSO/PBS) or Cert plus MCMVp therapy (three 15-mer CD4 T-cell peptide epitopes derived from the viral proteins m09, m25 and m142 and three 8/9-mer CD8 T-cell peptide epitopes derived from the viral proteins m38, m45 and IE3) (figure 1A).

Figure 1. MCMV memory T cells can be redirected to fight pancreatic tumors. (A) Protocol overview. Each treatment consists of a Cert injection (300 µg) followed by injection of 50 µg of each MCMV peptide. (B) KPC1242 tumor growth over time, monitored by ultrasound. (n=4–5/group). (C) H&E and immunofluorescent analysis of the tumors at endpoint (18 days post tumor implantation) looking at apoptosis by cleaved caspase 3 (CC3-Red) and T-cell infiltration by CD3 (green) staining. Necrosis score was attributed blindly by a histopathologist (n=3–5/group). Results were compared by one-way ANOVA with Tukey correction, and repeated measures by two-way ANOVA with Sidak correction. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. ANOVA, analysis of variance; Cert, certepetide; i.v., intravenous; MCMV, murine cytomegalovirus.

Figure 1

Open in a new tab

Infected mice treated with Cert+MCMVp therapy showed a dramatically reduced tumor growth, while infection alone or treatment of uninfected mice showed no benefit (figure 1B and online supplemental S1A bottom panel). Histological analysis of the tumor tissue showed increased necrosis, cleaved caspase-3 and CD3+ T-cell infiltration (figure 1C). Analysis of the liver showed increased T-cell infiltration but no cleaved caspase 3 staining (online supplemental figure S1). Despite the beneficial effects of MCMVp therapy, some toxicity resulting in significant weight loss was observed in the infected and treated group, possibly due to the high initial doses of MCMVp (50 µg of each) (online supplemental figure S1A top panel). This was associated with a massive expansion of highly proliferative (Ki67) and activated (CD69) MCMV-specific CD8 T cells in the tumor and spleen (>80% of total T cells). Expansion of MCMV CD4 T cells in the spleen was minimal (online supplemental file 2figure S1C, S1D). This prompted us to perform a dose titration of the MCMVp to see if a therapeutic window lacking systemic toxicity could be found (figure 2A). We determined that 1 µg of CD8 peptides and 50 µg of CD4 peptides (Dose A) was optimal, allowing better tumor growth control without inducing measurable weight loss. A survival study showed a median of 25 days in the control group versus 42 days for treated mice (Dose A: 68% increased survival, p=0.0027) (figure 2A). We next assessed whether Cert co-injection was required for MCMVp therapy inhibition of tumor growth. Surprisingly, MCMVp had the same effect on T-cell infiltration, tumor growth and survival whether Cert was included or not (online supplemental figure S2A-C). Similarly, the three immunodominant MCMV-specific T-cell populations (m25 CD4 T and m38/m45 CD8 T cells) were still preferentially enriched in the tumor compared with the spleen and liver despite the absence of Cert (Online supplemental figure S2D). Given these results, the remaining studies were conducted without Cert.

Figure 2. MCMVp therapy results in delayed tumor growth with increased survival and can be enhanced with chemotherapy. (A) KPC1242 tumor growth curves using various doses of MCMVp therapy in combination with Cert (Dose A 50 µg CD4p/1 µgCD8p, Dose B 50 µg CD4p/0.1 µgCD8p, Dose C 10 µg CD4p/3 µgCD8p, Dose D 10 µg CD4p/1 µgCD8p). Normalized mouse weight (middle) and Kaplan-Meier survival graph (right). (n=5–8 mice /group). (B) Dose A without Cert was used in combination or not with Gem. Normalized mouse weight (middle) and tumor weight at endpoint (right) (25 days post-tumor implantation). (n=5–10 mice/group) (C) B6×129 F1 hybrid mice were injected with the KPC46 tumor cell line and treated as in B. Normalized mouse weight (middle) and tumor weight at endpoint (right) (22 days post-tumor implantation). (n=5–7 mice/group). Results were compared by one-way ANOVA with Tukey correction, and repeated measures by two-way ANOVA with Sidak correction. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. ANOVA, analysis of variance; Cert, certepetide; Gem, gemcitabine; MCMV, murine cytomegalovirus; MCMVp, MCMV peptide.

Figure 2

Open in a new tab

Since high PD-1 expression was observed in MCMV-specific tumor-infiltrating lymphocytes (TILs) (online supplemental figure S1C), we assessed whether checkpoint blockade could further enhance MCMVp therapy-dependent tumor control. As KPC tumors contain high levels of interleukin-10 producing myeloid-derived suppressor cells (MDSC), we also combined our treatment with a blocking anti-IL10R antibody. However, no additional benefits to MCMVp therapy were observed, even when both PD-1 and IL10R blockade were combined (online supplemental figureS2E).

We then sought to assess whether combining MCMVp therapy with a chemotherapeutic agent could show additional benefit. Gemcitabine (Gem) is one of the standard of care drugs, known for its potential to reduce the levels of MDSCs and enhance T-cell immunity, especially when used at low doses (5 mg/kg).29,31 We found that while Gem lacked any antitumor effect on its own, co-treatment with MCMVp showed a subtle benefit over MCMVp therapy alone (figure 2B).

To show that the benefits of MCMVp therapy were not restricted to a single pancreatic tumor model, we used the C57Bl6/J (B6)×129SvJ (129) F1 hybrid model implanted with the syngeneic KPC46 tumor cell line. Similar to KPC1242 tumors, only infected mice receiving MCMVp therapy showed a significant decrease in tumor growth (figure 2C). In this model, MCMVp therapy alone showed markedly enhanced efficacy compared with KPC1242 in B6 mice, with Gem having no additional benefit. Taken together, we show that harnessing of pre-existing CMV memory T-cell responses has major therapeutic benefits in two different aggressive, orthotopic pancreatic tumor models.

MCMV T cells expand following treatment and preferentially localize to the tumor.

We next sought to characterize the phenotype and effector function of MCMV-specific T cells in spleen, liver and tumor in the various therapeutic regimens described above (figure 3 and online supplemental figure S3). First, we determined the total number of CD8 T cells and CD4 conventional T cell (Tconv) (non-regulatory T cell (Treg) CD4 T cells) and saw (1) no differences in the spleen, (2) increased CD8 and CD4 Tconv in the liver of infected mice receiving MCMVp therapy (±Gem) and (3) increased CD8 T cells in tumors of the same two therapeutic groups (online supplemental figure S3A). MCMV-specific T cells were identified using MHC-I and MHC-II tetramers labeled with APC and PE, and double-positive cells were considered positive (figure 3A and D). MCMV-specific CD4 Tconv cells were detected exclusively in tumors from infected mice receiving MCMVp therapy. The spleen contained a relatively low proportion of MCMV CD4 T cells, while the liver and tumor had higher and more comparable levels. The dominant epitope in the liver was m142, while tumors were more enriched with m25-specific T cells (figure 3A and B, S3B). Quantification of the absolute number of m09, m25 and m142 specific Tconv cells in tumors (figure 3C), livers and spleens (online supplemental figure 3B) paralleled their proportions shown in figure 3B. CD4+FoxP3+ MCMV-specific Tregs composed a much lower proportion than CD4 Tconv (online supplemental figure 3D), suggesting that in vivo expansion of MCMV-specific CD4 T cells results in the differentiation of a bona fide antiviral effector CD4 T-cell rather than an immunoregulatory one. Of note, there were more CMV-specific Tregs in the liver compared with tumors (online supplemental figure 3D), perhaps accounting for the absence of increased apoptosis in this organ as measured by cleaved caspase 3 (online supplemental figure S1D).

Figure 3. MCMV T cells expand after MCMVp therapy and preferentially localize to the tumor. Mice were implanted with KPC 1,242 cells and treated as previously described in figure 2B. At day 25 post-tumor implantation, spleen, liver (LV), and tumors (TMR) were collected to analyze TILs by flow cytometry. (n=5/group). (A, D) Representative dot-plot of tetramer binding MCMV-specific CD4 Tconv (A) and CD8 T cells (B). (B, E) Mean proportions of MCMV tetramer binding T cells among total CD4 Tconv (B) and CD8 T cells (E). (C, F) Absolute number of tumor-resident MCMV tetramer-binding CD4 Tconv (C) and CD8 T cells (F). Results were compared by one-way ANOVA with Tukey correction. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. ANOVA, analysis of variance; APC-Tetra, Allophycocyanin labeled tetramer; LV, liver; MCMV, murine cytomegalovirus; MCMVp, MCMV peptide; PE-Tetra, phycoerythrin labeled tertamer; Tconv, conventional T cell; TILs, tumor-infiltrating lymphocytes; TMR, tumor.

Figure 3

Open in a new tab

Notably, while low levels of MCMV-specific CD8 T cells were always detectable in tumors from infected mice in the absence of MCMVp therapy (figure 3E), their absolute number was quite low and did not reach statistical significance (figure 3F). After MCMVp therapy we observed an expansion of m38 and m45 CD8 T-cell proportions in the liver as well as the tumor (figure 3D and E), commensurate with an increase in their absolute numbers in tumors (figure 3F) and livers (online supplemental figure S3), while IE3 CD8 T cells did not expand significantly. Taken together, these data show that MCMVp therapy in infected mice selectively enhances the presence of m25, m38 and m45 T cells in the tumor.

Finally, to determine if the beneficial effects of MCMVp therapy were T cell-dependent, we depleted either CD8 or CD4 T cells in vivo. We showed that in the absence of CD4 and especially CD8 T cells, the effects of MCMVp therapy were reduced (online supplemental figure S3E). CD8 T cells were not completely depleted in the spleen, as were CD4 T cells, but a >80% reduction in their number was observed.

MCMV-specific CD8 T cells show a marked effector phenotype

The expression levels of activation markers (CD69), costimulatory receptors (CD226), inhibitory receptors (PD-1, Lymphocyte Activation Gene 3 (LAG3), and T cell Immunoglobulin and Mucin domain 3 (TIM3)) and cytotoxic molecules (Perforin (Prf), Granzymes A and B (GzmA and GzmB) were assessed in tumor-resident MCMV-specific CD8 T cells. All of these markers except GzmA were significantly increased in m45-specific CD8 TILs from infected mice receiving MCMVp therapy (figure 4A). Of note, GzmB levels were significantly higher in mice treated with MCMVp+Gem, providing a possible mechanistic explanation for the enhanced tumor control in this group compared with MCMVp therapy alone. M45-specific CD8 T cells also showed a very similar phenotype in the liver, suggesting a conserved phenotype in multiple tissues (online supplemental figure S4A). While PD-1, LAG3 and TIM3 are often called “exhaustion markers”, they do not define functional exhaustion as they are also upregulated early during T-cell activation to dampen excessive inflammation,32 or in terminally differentiated T cells.33 Consistently, we observed strong expression of activation markers and cytotoxic molecules in T cells expressing these immune checkpoint molecules, arguing against an exhausted state. In vitro re-stimulation of the tumor and liver cells with a mix of the six MCMV peptides used for treatment resulted in robust interferon (IFN)-γ and tumor necrosis factor (TNF) production by CD8 T cells, further indicating these cells were not exhausted (figure 4B, online supplemental S4D). Of note, MCMV-specific CD8 T cells also established tumor residence in infected mice that did not receive MCMVp therapy, as shown by tetramer binding (see figure 3E) and cytokine production (figure 4B). MCMVp therapy further increased both IFN-γ and TNF production by these TILs. This enhanced IFN-γ was most evident in the MCMVp+Gem group, correlating with greater tumor growth control, and the same was true for IFN-γ production by CD4 Tconv (figure 4B).

Figure 4. MCMV T cells are highly activated and cytotoxic within tumors. Immunophenotyping and cytokine production analysis of the TILs described in figure 3 (25 days post-tumor implantation). (n=5/group). (A) Representative histograms of m45-specific CD8 T cells isolated from tumors (top) and MFI quantification (bottom). (B) Tumors were subjected to CD45+TIL magnetic separation and the purified cells were incubated with the six MCMV peptides for 4 hours, in the presence of GolgiPlug. The graphs show IFN-γ and TNF production in CD8 or CD4 Tconv cells. Results were compared by one-way ANOVA with Tukey correction. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. ANOVA, analysis of variance; Gzm A and B, granzymes A and B; IFN, interferon; LAG3, lymphocyte activation gene 3 ; MCMV, murine cytomegalovirus; MFI, mean fluorescence intensity; Prf, perforin; PD-1, programmed cell death protein-1; Tconv, conventional T cell; TILs, tumor-infiltrating lymphocytes; TIM3, T cell immunoglobulin and mucin domain 3; TNF, tumor necrosis factor.

Figure 4

Open in a new tab

Despite the intrinsic differences between m45-specific and m38-specific CD8 T-cell populations (m45TET+: conventional central memory T cell; m38TET+: inflationary effector memory T cells), the phenotype of these populations was similar in both the liver and tumor (figure 4A, online supplemental figure S4A-C). GzmA and GzmB expression was significantly higher in liver-resident m38 compared with m45 T cells in infected mice not receiving MCMVp therapy (online supplemental figure 5A, B). However, once treated with MCMVp±Gem, both CD8 T-cell populations showed similarly high Gzm expression in the liver. In the tumor, both these CD8 T-cell populations expressed lower GzmA than in the liver. Surprisingly, tumor resident m38 T cells downregulated GzmA expression after MCMVp therapy, but upregulated GzmB (online supplemental figure S4C). This downregulation of GzmA may result from the TME, but, importantly, did not prevent the beneficial effect of MCMVp therapy.

MCMVp treatment induces profound changes in the tumor microenvironment

To assess the effect of MCMVp therapy on the TME, we performed single-cell RNA sequencing (scRNAseq) of whole tumors from MCMV-infected mice treated with vehicle or MCMVp therapy (figure 5A). The major tumor cell populations were epithelial cells, macrophages, T cells, fibroblasts, and dendritic cells. As shown (figure 5A and B, online supplemental figure S5A), the most significant change in the MCMVp therapy group was a marked increase in T cells (0.3%Veh vs 5.7%MCMVp). Additionally, a trend was seen for decreased pancreatic tumor cells (80%Veh vs 66%MCMVp), and an increased proportion of macrophages (17%Veh vs 25%MCMVp) and dendritic cells (0.7%Veh vs 1.87%MCMVp). When lymphocytes were subclustered into B cells, CD4/CD8 T cells and natural killer (NK) cells, we observed that the proportion of CD8 T cells was markedly increased following MCMVp therapy, while B and NK cells remained largely stable (online supplemental figure S5B). Gene Set Enrichment Analysis (GSEA) of epithelial cells revealed an increase in type I and type II IFN response genes and a decreased expression of genes involved in angiogenesis, KRAS signaling and epithelial-mesenchymal transition (EMT) (figure 5C). MCMVp therapy induced an increase in several antigen presentation genes such as H2-D1, H2-K1 and β2m (online supplemental figure S5C), suggesting a potential role for direct MHC-I presentation of MCMVp by tumor cells. In contrast, EMT-related genes like Acta2 and genes associated with tumor progression such as Megf10 or pleiotrophin were significantly downregulated (online supplemental figure S5C).

Figure 5. MCMVp treatment induces profound changes in the tumor microenvironment. Infected mice bearing tumors were treated with vehicle or MCMVp as previously described, tumors were collected (23 days post-tumor implantation) and used for 5’scRNAseq (n=3/group, each sample being the pool of 1–2 mice). (A) UMAP distribution. (B) Boxplot of cellular proportions in the tumor. (C) GSEA analysis of epithelial cell DEGs. (D) CellChat analysis of potential cell–cell interactions in the tumor. Arrowheads indicate the directionality of the interactions. (E,F) Relative quantification of the different pathways involved in the potential cell–cell interactions in the tumor shown in D. Double arrows indicate bidirectional interactions, simple arrows indicate unidirectional interactions. DEG, differentially expressed genes; GSEA, gene set enrichment analysis; MCMV, murine cytomegalovirus; MCMVp, MCMV peptide; scRNA-seq, single-cell RNA sequencing; UMAP, uniform manifold approximation and projection.

Figure 5

Open in a new tab

Fibroblasts and other stromal cell populations are known to support tumor growth and shape the TME.34 MCMVp therapy triggered a significant increase in MHC class II genes, such as H2-Aa, H2-Ab1 and H2-Eb1 associated with CD4 T-cell mediated antitumor immune responses (online supplemental figure S5C). In addition, Differentially Expressed Genes (DEGs) in macrophages showed an upregulation of Stat1 and Irf1 (two IFN response-related genes) and Tap1 (involved in peptide processing) consistent with an enhanced immune response in the tumor (online supplemental figure S5C).

We then used CellChat to investigate how cell–cell interactions may be impacted by MCMVp therapy. This analysis suggested reduced interactions between fibroblasts and epithelial cells or macrophages, while T-cell interactions with macrophages, dendritic cells and epithelial cells were increased (figure 5D). This conclusion was based on the apparent increased expression of chemokines, complement, APRIL/Tnfsf13, and Tnf-mediated crosstalk between T cells and macrophages. Decreased Transforming Growth Factor β (TGF-β) and Secreted Phosphoprotein 1 (SPP1)-dependent pathways were also observed (figure 5E), which are involved in maintaining the immunosuppressive TME phenotype.35 36 TGF- β dependent signaling networks are also decreased between T cells, epithelial cells and fibroblasts (figure 5E and F). In addition, MCMVp therapy increased potential FasLg interactions between T cells and fibroblasts, but not between T cells and epithelial cells. Importantly, fibroblast-epithelial cell interactions involving growth factors (Fibroblast Growth Factor (FGF), Vascular Endothelial Growth Factor (VEGF), Epidermal Growth Factor (EGF)) were reduced following MCMVp therapy (figure 5F). Taken together, these results show that our peptide-based therapy induces significant changes in the TME that lead to improved survival and better prognosis.

MCMV therapy leads to tumor infiltration of activated and cytotoxic MCMV TCRαβ cells

To further explore the TIL phenotype following MCMVp therapy, and whether “bystander activation” of T cells that are not virus specific may occur, a second scRNAseq on Fluorescence Activated Cell Sorting (FACS) sorted CD3+CD4+ and CD3+CD8+ T cells was performed (figure 6 and S6). ~50,000 T cells were analyzed for both gene expression and V(D)J CDR3 sequences to assess the overall T-cell receptor (TCR)-αβ repertoire. Bioinformatic analysis yielded 20 distinct T-cell clusters (figure 6A), with 13 composed by CD8 T cells, 5 representing CD4 Tconv and 2 FoxP3+CD4 Treg (figure 6A and B).

Figure 6. MCMV therapy leads to tumor infiltration of activated and cytotoxic MCMV TCRαβ cells. A different cohort of mice treated as in figure 5 was used to purify T cells by FACS sorting for CD3, CD4 and CD8 (n=3/group, each sample being the pool of 1–2 mice) (22 days post-tumor implantation). The purified T cells were used for 5’scRNAseq and TCR sequencing. (A) T-cell clustering and UMAP distribution. (B) Feature plot mapping of CD4, CD8 and regulatory T cells within these clusters and overlay of the two conditions (Vehicle-Blue, MCMVp-Red). (C) Graph showing the frequency of each sample in each of the predefined clusters after normalizing the amount of cells per sample. (D) Dimplot representing the detection of specific MCMV-associated matched CDR3 sequence for TCRα and TCRβ. (E) Proportion of MCMV-specific TCRαβ detected per cluster. (F) Proportion of MCMV-specific TCRβ detected per cluster. (G) Dotplots featuring selected gene sets and expression levels among clusters. Arrows point to clusters containing high proportion of MCMV-specific T cells (red=m45, orange=m38, green=m142, blue=m25). FACS, fluorescence activated cell sorting; MCMV, murine cytomegalovirus; MCMVp, MCMV peptide; scRNA-seq, single-cell RNA sequencing; TCR, T-cell receptor; UMAP, uniform manifold approximation projection.

Figure 6

Open in a new tab

Analysis of T cells in treated mice showed the emergence of five new clusters (1, 7, 12, 14 and 16) (figure 6B–C, online supplemental figure S6A), with cluster 7 being CD4 Tconv and the other four composed of CD8 T cells. CDR3 sequence analysis from TCRα and β chains showed strong enrichment of MCMV-specific T cells in these clusters (figure 6D–F), based on published TCR sequences37 38 and in-house obtained sequences revealed during infection (manuscript in preparation). Cluster 7 was enriched in m25 and m142-associated CDR3s and had intermediate signal for m09. Similarly, we detected a large proportion of m45 and m38-associated CDR3s in clusters 1, 16 and 14, and a weaker signal in cluster 12 (figure 6D and E). Next, we repeated the analysis looking exclusively at TCRβ sequences, since more MCMV-specific TCRβ sequences are known, compared with TCRα sequences (figure 6F, online supplemental figure S6B). Up to 50% of cluster 1—the second most abundant cluster (online supplemental figure S6C)—is composed of MCMV-specific CD8 T cells, half that express m45-specific CDR3, and the other half expressing CDR3 ascribed to both m45/m38 in the literature37 38 (figure 6F).

We then analyzed all 20 clusters for the expression of key transcription factors, cytotoxic molecules, cytokines and co-signaling receptors that regulate T-cell effector function (figure 6G). All clusters induced by MCMVp therapy expressed high levels of Tbx21 (T-bet) and Ifnγ. Cluster 1 expressed very high levels of Prf1, Gzmb, Gzmk and Fasl, indicative of high cytotoxic potential. As seen by flow cytometry (see figure 4 and online supplemental figure S4), elevated expression of co-inhibitory receptors was observed in cluster 1 (Pdcd1, Havcr2, Lag3), as well as Ctla4 and Tigit. However, enhanced levels of several co-stimulatory receptors were also seen (Cd226, Klrk1, Cd27, Cd28, Tnfsf9), consistent with a highly activated/differentiated phenotype. Notably, cluster 7, which contains the highest numbers of m25 and m142 CD4 Tconv, had the most Ifnγ expression of any cluster, and showed high expression of co-stimulatory receptors Cd28, Tnfrsf4 (OX40), Tnfrsf18 (GITR), and Cd40lg. The top five differentially expressed genes in cluster 1 were Havcr2 (TIM3), Pdcd1 (PD1), Rgs16, Lag3, and Cxcr6, and in cluster 7 Maf, Tnfrsf4, Glrx and Izumo1r (online supplemental figure S6D). Glutaredoxin (Glrx) regulates redox homeostasis and cellular oxidative stress, which may help sustain T-cell function in a stressful environment like the pancreatic TME. Of note, cluster 6 did not include any CDR3 sequences associated with MCMV-specific TCRs and was strongly enriched following MCMVp therapy. This cluster expressed high levels of Tbx21, Prf1, Gzma, Gzmb, Gzmk, Fasl and Ifnγ but no exhaustion markers. CD39 (Entpd1), which has been shown to be specifically expressed by tumor-reactive CD8 T cells in humans,39 40 was highly expressed by the MCMVp-specific T-cell clusters,1 7 12 16 but also by the CD8 T cells in cluster 6 (online supplemental figure S6E). This suggests that systemic MCMVp therapy may induce bystander T-cell recruitment and/or epitope spreading.

In addition, we found there is a partial overlap between the TCRαβ sequences found within the tumors of infected mice±MCMV therapy. This suggests that therapy not only amplifies pre-existing intratumoral MCMV-specific T cells but also triggers the expansion of peripheral MCMV-specific T cells and their subsequent migration into the tumor (online supplemental figure S6F).

Immunization-induced T cells localize to tumors, but to a lesser extent than infection-induced T cells

Since MCMV-specific T cells preferentially localized to tumors in the absence of iRGD, we sought to further investigate the underlying mechanism for this. Our first hypothesis was based on publications showing that CMV DNA is present in glioblastoma tumors.41 In our experiments, mice are infected with MCMV and latency is established before tumor cell implantation. Therefore, if viral replication occurs directly in KPC tumor cells, reactivation from latency and subsequent infection of these cells would have to occur. Not unexpectedly, MCMV could productively infect KPC cells in culture, as shown using an MCMV-mCherry reporter virus (figure 7A). Next, we tested whether we could detect MCMV DNA by quantitative PCR (qPCR) in tumors implanted in infected mice. qPCR detected MCMV DNA in livers and spleens from day 4 acutely infected mice, as expected, but was undetectable in KPC1242 (figure 7B) or KPC46 (online supplemental figure S7A) tumors.

Figure 7. Immunization-induced T cells infiltrate tumors, but to a lesser extent than infection-induced T cells. (A) KPC1242 infected in vitro with MCMV-mCherry for 48 hours. (B) IE1 DNA qPCR normalized to actin for KPC1242 tumor cells harvested from uninfected or MCMV-infected animals after tumor implantation. Positive controls are spleen and liver from mice four d.p.i. (C) Protocol for peptide immunization and tumor challenge. (D) Tumor volume measured by ultrasound over time (left) and end point tumor weight (right) in the different treatment conditions after immunization (CT-non immunized+MCMVp, PBS-immunized with CFA/IFA no peptides, + MCMVp, OVAp-immunized+OVA I and OVA-IIp and MCMVp-immunized+MCMVp). (E–I) Mice from (D) were analyzed for T-cell phenotype and function from the draining lymph node, spleen and tumor (n=6 mice per group) (26 days post-tumor implantation). (E) Representative m45-tetramer binding among MCMVp-immunized mice, and proportions quantified over the three tissues. (F) Representative OT-I-tetramer binding among OVAp-immunized mice, and proportions in the three tissues. (G) Phenotype of tumor-infiltrating m45-specific CD8 T cells compared with m45-tetramer-negative. (H) T-cell quantification per gram of tumor, and fold differences seen between control and MCMVp-treated groups after immunization and infection (top, from experiment shown in figure 4). (I) GzmB expression and fold increase comparing control and treated groups after immunization (bottom) and infection (top, from experiment shown in figure 4). Results were compared by one-way ANOVA with Tukey correction. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. ANOVA, analysis of variance; CFA, complete Freund’s Adjuvant; dLN, draining lymph node; d.p.i, days post infection; GzmB, granzyme B; IFA, incomplete Freund’s Adjuvant; LV, liver; MCMVp, murine cytomegalovirus peptide; OVA, ovalbumin; OVAp, ovalbumin peptides; PBS, phosphate-buffered saline; qPCR, quantitative PCR, SPL, spleen; TMR, tumor.

Figure 7

Open in a new tab

As these analyses showed that MCMV was not directly infecting the tumor and promoting virus-specific T-cell recruitment, we next tested whether infection was even required for their tumor-residency. This was done by immunizing mice with MCMV or OVA-derived CD8 and CD4 T-cell epitopes, implanting tumors and recalling these immunization-induced memory T cells with peptide injections once tumors were palpable (figure 7C). Tumor growth was monitored by ultrasound for 2 weeks and then mice were sacrificed for analysis of MCMV or OVA specific T cells from tumors, draining lymph nodes and spleens. Contrary to MCMV-infected mice, MCMVp therapy did not curtail tumor growth to nearly the same extent when memory T cells were induced by immunization. Recalling immunization-induced MCMV or OVA T cells had no impact on tumor volume and only a modest reduction of tumor weight was seen in both groups at the experimental end point (figure 7D). Somewhat unexpectedly, m45 and m38 CD8 T cells were still found to be highly enriched in the tumors of the MCMVp immunized group (figure 7E), with fewer numbers of m25 and m142 CD4 Tconv (online supplemental figure S7B).

These results proved that (1) viral infection is not required for preferential accumulation of MCMV-specific T cells within the tumor in our experimental setup, and (2) T cells accumulate in the tumor in an antigen-agnostic manner (figure 7F). Nevertheless, while antigen-specific T cells infiltrated tumors after immunization, fewer established tumor residence compared with those induced by infection: total CD8 infiltration/gram of tumor increased 11–14-fold following treatment of infected mice, compared with 1.3-fold after treatment of immunized mice. Similarly, m45 CD8 T-cell numbers increased 408–603-fold in infected mice compared with ~35-fold in immunized mice (figure 7H). This may explain the lack of tumor control observed in the context of immunization, despite T-cell infiltration in the tumor.

Phenotyping of immunization-induced tumor resident m45 CD8 T cells showed expression of CD69, PD1, LAG3, Prf and GzmB (figure 7G and I), and ability to produce IFN-γ (online supplemental figure S7C). However, the expression of activation and cytotoxic markers seen in these T cells was lower in immunized mice, as compared with infected mice after peptide recall of memory cells. Specifically, CD226, GzmA and GzmB were not increased to nearly the extent in immunization-induced m45 CD8 T cells after MCMVp therapy as seen in infected mice (eg, GzmB expression by m45 CD8 T cells increased 1.6-fold compared with control group in the immunization protocol, compared with 4.1–5.8-fold in the infection protocol; figure 7I, online supplemental figure S7D). Taken together, these results suggest that while viral infection itself is not required for preferential infiltration of T cells in the tumor after MCMVp therapy, infection-induced memory T cells can be more efficiently recalled by peptide therapy, leading to far greater T-cell infiltration in the tumor and higher cytotoxic potential, providing a rationale for harnessing pre-existing antiviral immunity.

Discussion

In this work, we show that harnessing pre-existing antiviral memory T cells through systemic peptide therapy results in their massive tumor recruitment, substantial changes in the TME and ultimately reduced tumor growth and enhanced survival. We believe CMVp therapy represents a promising new approach to pancreatic cancer treatment in HCMV-seropositive patients. This systemically delivered therapy is a relatively simple “off the shelf” T-cell-based therapy, that is, inexpensive compared with personalized medical approaches and can be used to treat tumors that are not easily accessible for intratumoral injections. Furthermore, systemic delivery would allow the therapy to reach metastatic sites. Clinical translation is facilitated by the pre-existing knowledge of the HCMV peptide sequences recognized by human CD4 and CD8 T cells, some of which have been published by our laboratory. We recently identified ~200 new HCMV T-cell epitopes,25 and are now testing the feasibility of using them for cancer treatment in humanized mouse models. These experiments should help identify HCMV epitopes that cover multiple haplotypes and are broadly applicable to CMVp therapy in people.

Finding that a technically simple peptide therapy can induce such dramatic changes by modifying the TME and increasing T cell-tumor cell interactions is encouraging, as it suggests multiple antitumor pathways are engaged. scRNAseq analysis revealed substantial changes in the transcriptome of tumor epithelial cells after MCMVp therapy, leading to increased IFN-signaling pathways and decreased EMT, angiogenesis and Kras signaling. The highest induced genes in cancer cells were MHC-I-associated messenger RNAs such as β2m, H2-k1 and H2-d1 (online supplemental figure S5C), suggesting tumors may present viral epitopes directly to infiltrating T cells, activating tumor killing. Another possibility is that peptide epitopes are presented by intratumoral or peripheral macrophages or dendritic cells. Analysis of the T-cell clones present in the tumors of MCMV-infected mice±treatment (online supplemental figure S6F) suggested some of the TILs in the treated tumors may have expanded from tumor resident T-cell clones and others may have expanded in the periphery and migrated to the tumor, something we are currently exploring.

Of note, T-cell cluster 6 (figure 6G and online supplemental figure S6E) was present only in treated tumors and did not contain any known MCMV sequences suggesting they may react with tumor neoantigens. Future analyses of the clonality of their TCRαβ sequences should help to elucidate this.

Our immunization-based protocol demonstrated that both MCMV-specific T cells and OVA-specific OT-I CD8 T cells could infiltrate the tumor but had no effect on tumor growth control. Perhaps this is due in part to the fact that peptide immunization does not induce the same “inflationary” MCMV-specific effector-memory CD8 T-cell populations as seen during infection.42 43 In addition, immunization-induced T cells infiltrated tumors at far lower numbers and expressed less GzmB when compared with those produced during infection (figure 7H and I and Online supplemental figure S7D).

While establishing a more efficient immunization protocol may induce more TILs with higher effector function that lead to better tumor control, our strategy to harness infection-induced CMV memory T cells is applicable to the vast majority of the population, given the high prevalence of the virus.

In theory, CMV is not the sole virus model that could be used with our strategy; other prevalent viruses such as Epstein-Barr virus (EBV) that infects >90% of the world population could be explored. However, EBV does not induce nearly as many effector-memory T cells as HCMV and has been implicated in multiple sclerosis.44

One question that remains open is why does MCMVp therapy result in preferential tumor homing of MCMV-specific T cells even in the absence of Cert? We showed that MCMV DNA is not present in tumor tissue, so TILs are not responding to infected cells. Given the fact that multiple MCMV and OVA T cells induced by peptide immunization colonized the tumor, it is unlikely that molecular mimicry is the underlying mechanism for their trafficking to this site. One possibility we have considered is that the enhanced macropinocytosis previously reported in pancreatic tumors45,47 may supersede any requirements for Cert in the uptake of systemically injected peptides. Of note, while Cert was not needed in this context, its usage as co-treatment with MCMVp therapy might still be necessary for the treatment of other tumor types.

In conclusion, this study demonstrates that systemic treatment with short T-cell peptide epitopes can efficiently recall CMV memory T cells and promote their tumor localization leading to tumor control. This strategy is a mutation-agnostic approach that has the potential to benefit the majority of the population that harbor latent CMV. Importantly, our study reveals the benefit this approach provides in aggressive orthotopic pancreatic tumor models, where immunotherapy has failed, suggesting it may be reasonable for clinical translation.

Supplementary material

online supplemental file 1
jitc-14-2-s001.tiff (2.1MB, tiff)
DOI: 10.1136/jitc-2025-012969
online supplemental file 2
jitc-14-2-s002.pdf (8.9MB, pdf)
DOI: 10.1136/jitc-2025-012969

Acknowledgements

NGS, flow cytometry, microscopy, and histology core facilities at the La Jolla Institute for Immunology. Biorepository, microscopy and flow cytometry cores at Moores Cancer Center (supported by NIH S10OD032316 and UCSD Specialized Cancer Center Support P30 Grant 2P30CA023100), NIH Emory tetramer facility for providing MHC-I/II biotinylated monomers.

Footnotes

Funding: This work was supported by Foundation for a Better World grant 001 to T.H.M, 1R21CA286198 to T.H.M. and C.A.B and AI139749 and AI101423 grants to C.A.B. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Provenance and peer review: Not commissioned; externally peer reviewed.

Patient consent for publication: Not applicable.

Ethics approval: This study was carried out in strict accordance with the guidelines of Association for assessment and Accreditation of laboratory Animal Care (AAALAC) and National Institutes of Health (NIH). All animal protocols used in this study, were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of California, San Diego and La Jolla Institute for immunology (LJI).

Data availability statement

Data are available upon reasonable request.

References

  • 1.Schadendorf D, Hodi FS, Robert C, et al. Pooled Analysis of Long-Term Survival Data From Phase II and Phase III Trials of Ipilimumab in Unresectable or Metastatic Melanoma. JCO . 2015;33:1889–94. doi: 10.1200/JCO.2014.56.2736. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Reck M, Rodríguez-Abreu D, Robinson AG, et al. Pembrolizumab versus Chemotherapy for PD-L1–Positive Non–Small-Cell Lung Cancer. N Engl J Med. 2016;375:1823–33. doi: 10.1056/NEJMoa1606774. [DOI] [PubMed] [Google Scholar]
  • 3.Garon EB, Rizvi NA, Hui R, et al. Pembrolizumab for the treatment of non-small-cell lung cancer. N Engl J Med. 2015;372:2018–28. doi: 10.1056/NEJMoa1501824. [DOI] [PubMed] [Google Scholar]
  • 4.Wei SC, Duffy CR, Allison JP. Fundamental Mechanisms of Immune Checkpoint Blockade Therapy. Cancer Discov. 2018;8:1069–86. doi: 10.1158/2159-8290.CD-18-0367. [DOI] [PubMed] [Google Scholar]
  • 5.Alexandrov LB, Nik-Zainal S, Wedge DC, et al. Australian Pancreatic Cancer Genome Initiative, ICGC Breast Cancer Consortium. Nature New Biol. 2013;500:415–21. [Google Scholar]
  • 6.Schumacher TN, Schreiber RD. Neoantigens in cancer immunotherapy. Science. 2015;348:69–74. doi: 10.1126/science.aaa4971. [DOI] [PubMed] [Google Scholar]
  • 7.Evans RA, Diamond MS, Rech AJ, et al. Lack of immunoediting in murine pancreatic cancer reversed with neoantigen. JCI Insight. 2016;1:e88328. doi: 10.1172/jci.insight.88328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Hiraoka N, Onozato K, Kosuge T, et al. Prevalence of FOXP3+ regulatory T cells increases during the progression of pancreatic ductal adenocarcinoma and its premalignant lesions. Clin Cancer Res. 2006;12:5423–34. doi: 10.1158/1078-0432.CCR-06-0369. [DOI] [PubMed] [Google Scholar]
  • 9.Ino Y, Yamazaki-Itoh R, Shimada K, et al. Immune cell infiltration as an indicator of the immune microenvironment of pancreatic cancer. Br J Cancer. 2013;108:914–23. doi: 10.1038/bjc.2013.32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Brahmer JR, Tykodi SS, Chow LQM, et al. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. N Engl J Med. 2012;366:2455–65. doi: 10.1056/NEJMoa1200694. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Foley K, Kim V, Jaffee E, et al. Current progress in immunotherapy for pancreatic cancer. Cancer Lett. 2016;381:244–51. doi: 10.1016/j.canlet.2015.12.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Royal RE, Levy C, Turner K, et al. Phase 2 trial of single agent Ipilimumab (anti-CTLA-4) for locally advanced or metastatic pancreatic adenocarcinoma. J Immunother. 2010;33:828–33. doi: 10.1097/CJI.0b013e3181eec14c. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Rosato PC, Wijeyesinghe S, Stolley JM, et al. Virus-specific memory T cells populate tumors and can be repurposed for tumor immunotherapy. Nat Commun. 2019;10:567. doi: 10.1038/s41467-019-08534-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Çuburu N, Bialkowski L, Pontejo SM, et al. Harnessing anti-cytomegalovirus immunity for local immunotherapy against solid tumors. Proc Natl Acad Sci U S A. 2022;119:e2116738119. doi: 10.1073/pnas.2116738119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Teesalu T, Sugahara KN, Kotamraju VR, et al. C-end rule peptides mediate neuropilin-1-dependent cell, vascular, and tissue penetration. Proc Natl Acad Sci U S A. 2009;106:16157–62. doi: 10.1073/pnas.0908201106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Sugahara KN, Teesalu T, Karmali PP, et al. Coadministration of a tumor-penetrating peptide enhances the efficacy of cancer drugs. Science. 2010;328:1031–5. doi: 10.1126/science.1183057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Hurtado de Mendoza T, Mose ES, Botta GP, et al. Tumor-penetrating therapy for β5 integrin-rich pancreas cancer. Nat Commun. 2021;12:1541. doi: 10.1038/s41467-021-21858-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Sugahara KN, Braun GB, de Mendoza TH, et al. Tumor-penetrating iRGD peptide inhibits metastasis. Mol Cancer Ther. 2015;14:120–8. doi: 10.1158/1535-7163.MCT-14-0366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Liu X, Lin P, Perrett I, et al. Tumor-penetrating peptide enhances transcytosis of silicasome-based chemotherapy for pancreatic cancer. J Clin Invest. 2017;127:2007–18.:92284. doi: 10.1172/JCI92284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Song W, Li M, Tang Z, et al. Methoxypoly(ethylene glycol)-block-poly(L-glutamic acid)-loaded cisplatin and a combination with iRGD for the treatment of non-small-cell lung cancers. Macromol Biosci. 2012;12:1514–23. doi: 10.1002/mabi.201200145. [DOI] [PubMed] [Google Scholar]
  • 21.Deng C, Jia M, Wei G, et al. Inducing Optimal Antitumor Immune Response through Coadministering iRGD with Pirarubicin Loaded Nanostructured Lipid Carriers for Breast Cancer Therapy. Mol Pharmaceutics. 2017;14:296–309. doi: 10.1021/acs.molpharmaceut.6b00932. [DOI] [PubMed] [Google Scholar]
  • 22.Sha H, Li R, Bian X, et al. A tumor-penetrating recombinant protein anti-EGFR-iRGD enhance efficacy of paclitaxel in 3D multicellular spheroids and gastric cancer in vivo. Eur J Pharm Sci. 2015;77:60–72. doi: 10.1016/j.ejps.2015.05.020. [DOI] [PubMed] [Google Scholar]
  • 23.Ding N, Zou Z, Sha H, et al. iRGD synergizes with PD-1 knockout immunotherapy by enhancing lymphocyte infiltration in gastric cancer. Nat Commun. 2019;10:1336. doi: 10.1038/s41467-019-09296-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Agemy L, Friedmann-Morvinski D, Kotamraju VR, et al. Targeted nanoparticle enhanced proapoptotic peptide as potential therapy for glioblastoma. Proc Natl Acad Sci U S A. 2011;108:17450–5. doi: 10.1073/pnas.1114518108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Dhanwani R, Dhanda SK, Pham J, et al. Profiling Human Cytomegalovirus-Specific T Cell Responses Reveals Novel Immunogenic Open Reading Frames. J Virol. 2021;95:e00940-21. doi: 10.1128/JVI.00940-21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Picarda G, Benedict CA. Cytomegalovirus: Shape-Shifting the Immune System. J Immunol. 2018;200:3881–9. doi: 10.4049/jimmunol.1800171. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Brunel S, Picarda G, Gupta A, et al. Late-rising CD4 T cells resolve mouse cytomegalovirus persistent replication in the salivary gland. PLoS Pathog. 2024;20:e1011852. doi: 10.1371/journal.ppat.1011852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Picarda G, Ghosh R, McDonald B, et al. Cytomegalovirus Evades TRAIL-Mediated Innate Lymphoid Cell 1 Defenses. J Virol. 2019;93:e00617-19. doi: 10.1128/JVI.00617-19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Suzuki E, Kapoor V, Jassar AS, et al. Gemcitabine selectively eliminates splenic Gr-1+/CD11b+ myeloid suppressor cells in tumor-bearing animals and enhances antitumor immune activity. Clin Cancer Res. 2005;11:6713–21. doi: 10.1158/1078-0432.CCR-05-0883. [DOI] [PubMed] [Google Scholar]
  • 30.Sasso MS, Lollo G, Pitorre M, et al. Low dose gemcitabine-loaded lipid nanocapsules target monocytic myeloid-derived suppressor cells and potentiate cancer immunotherapy. Biomaterials. 2016;96:47–62. doi: 10.1016/j.biomaterials.2016.04.010. [DOI] [PubMed] [Google Scholar]
  • 31.Eriksson E, Wenthe J, Irenaeus S, et al. Gemcitabine reduces MDSCs, tregs and TGFβ-1 while restoring the teff/treg ratio in patients with pancreatic cancer. J Transl Med. 2016;14:282. doi: 10.1186/s12967-016-1037-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Kissick H, Ahmed R. New epigenetic regulators of T cell exhaustion. Cancer Cell. 2022;40:708–10. doi: 10.1016/j.ccell.2022.06.008. [DOI] [PubMed] [Google Scholar]
  • 33.Wijesinghe SKM, Rausch L, Gabriel SS, et al. Lymph-node-derived stem-like but not tumor-tissue-resident CD8+ T cells fuel anticancer immunity. Nat Immunol. 2025;26:1367–83. doi: 10.1038/s41590-025-02219-2. [DOI] [PubMed] [Google Scholar]
  • 34.Kalluri R. The biology and function of fibroblasts in cancer. Nat Rev Cancer. 2016;16:582–98. doi: 10.1038/nrc.2016.73. [DOI] [PubMed] [Google Scholar]
  • 35.Du Y, Lin Y, Gan L, et al. Potential crosstalk between SPP1 + TAMs and CD8 + exhausted T cells promotes an immunosuppressive environment in gastric metastatic cancer. J Transl Med. 2024;22:158. doi: 10.1186/s12967-023-04688-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Wu J, Shen Y, Zeng G, et al. SPP1+ TAM subpopulations in tumor microenvironment promote intravasation and metastasis of head and neck squamous cell carcinoma. Cancer Gene Ther. 2024;31:311–21. doi: 10.1038/s41417-023-00704-0. [DOI] [PubMed] [Google Scholar]
  • 37.Smith CJ, Venturi V, Quigley MF, et al. Stochastic Expansions Maintain the Clonal Stability of CD8+ T Cell Populations Undergoing Memory Inflation Driven by Murine Cytomegalovirus. J Immunol. 2020;204:112–21. doi: 10.4049/jimmunol.1900455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Dash P, Fiore-Gartland AJ, Hertz T, et al. Quantifiable predictive features define epitope-specific T cell receptor repertoires. Nature New Biol. 2017;547:89–93. doi: 10.1038/nature22383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Simoni Y, Becht E, Fehlings M, et al. Bystander CD8+ T cells are abundant and phenotypically distinct in human tumour infiltrates. Nature New Biol. 2018;557:575–9. doi: 10.1038/s41586-018-0130-2. [DOI] [PubMed] [Google Scholar]
  • 40.Chow A, Uddin FZ, Liu M, et al. The ectonucleotidase CD39 identifies tumor-reactive CD8+ T cells predictive of immune checkpoint blockade efficacy in human lung cancer. Immunity. 2023;56:93–106. doi: 10.1016/j.immuni.2022.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Krenzlin H, Behera P, Lorenz V, et al. Cytomegalovirus promotes murine glioblastoma growth via pericyte recruitment and angiogenesis. J Clin Invest. 2019;129:1671–83.:123375. doi: 10.1172/JCI123375. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Munks MW, Cho KS, Pinto AK, et al. Four distinct patterns of memory CD8 T cell responses to chronic murine cytomegalovirus infection. J Immunol . 2006;177:450–8. doi: 10.4049/jimmunol.177.1.450. [DOI] [PubMed] [Google Scholar]
  • 43.Panagioti E, Redeker A, van Duikeren S, et al. The Breadth of Synthetic Long Peptide Vaccine-Induced CD8+ T Cell Responses Determines the Efficacy against Mouse Cytomegalovirus Infection. PLoS Pathog. 2016;12:e1005895. doi: 10.1371/journal.ppat.1005895. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Lanz TV, Brewer RC, Ho PP, et al. Clonally expanded B cells in multiple sclerosis bind EBV EBNA1 and GlialCAM. Nature New Biol. 2022;603:321–7. doi: 10.1038/s41586-022-04432-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Commisso C, Davidson SM, Soydaner-Azeloglu RG, et al. Macropinocytosis of protein is an amino acid supply route in Ras-transformed cells. Nature New Biol. 2013;497:633–7. doi: 10.1038/nature12138. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Qiu Z, Liu W, Zhu Q, et al. The Role and Therapeutic Potential of Macropinocytosis in Cancer. Front Pharmacol. 2022;13:919819. doi: 10.3389/fphar.2022.919819. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Andreasson C, Ansari D, Ekbom F, et al. Macropinocytosis: the Achilles’ heel of pancreatic cancer? Scand J Gastroenterol. 2021;56:177–9. doi: 10.1080/00365521.2020.1855471. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

online supplemental file 1
jitc-14-2-s001.tiff (2.1MB, tiff)
DOI: 10.1136/jitc-2025-012969
online supplemental file 2
jitc-14-2-s002.pdf (8.9MB, pdf)
DOI: 10.1136/jitc-2025-012969

Data Availability Statement

Data are available upon reasonable request.

Articles from Journal for Immunotherapy of Cancer are provided here courtesy of BMJ Publishing Group

RESOURCES