Simultaneous targeting of PD-1 and IL-2Rβγ with radiation therapy to inhibit pancreatic cancer growth and metastasis

SUMMARY

In pancreatic ductal adenocarcinoma (PDAC) patients, we show that response to radiation therapy (RT) is characterized by increased IL-2Rβ and IL-2Rγ along with decreased ILR2α expression. The bispecific, PD1-IL2v, is a PD-1-targeted IL-2 variant (IL2v) immunocytokine with engineered IL-2 cis-targeted to PD-1 and abolished IL-2Rα binding, which enhances tumor-antigen specific T cell activation while reducing regulatory T cell (Treg) suppression. Using PD1-IL2v in orthotopic PDAC KPC-driven tumor models, we show marked improvement in local and metastatic survival along with profound increase in tumor-infiltrating CD8⁺ T cell subsets with a transcriptionally and metabolically active phenotype, and preferential activation of antigen-specific CD8⁺ T cells. In combination with single dose RT, PD1-IL2v treatment results in a robust, durable expansion of polyfunctional CD8⁺ T cells, T cell stemness, tumor-specific memory immune response, natural killer (NK) cell activation, and decreased Tregs. These data show that PD1-IL2v leads to profound local and distant response in PDAC.

Graphical Abstract

eTOC Blurb

Piper and Hoen et. al show that combination of radiation and PD1-IL2v immunotherapy enhances CD8⁺ T cell polyfunctionality, activation, and immune memory across tumor, lymph node, and blood compartments and results in a durable local and systemic anti-tumor response.

INTRODUCTION

Pancreatic ductal adenocarcinoma (PDAC) is a malignancy known to establish an immunosuppressive tumor microenvironment (TME) making it resistant to conventional targeted and cytotoxic therapies, such as radiation therapy (RT)^1,2. Even in the face of immune-invigorating treatments, responses in this disease type are almost always transient³. Examination of the biological elements contributing to immune escape in the setting of treatment failure, therefore, is imperative to overcoming inherent resistance. Given that PDAC is a systemic disease with a high risk of metastasis, and with only approximately 10% of patients diagnosed at early stages⁴, any treatment aimed at improving response and outcomes must place emphasis on systemic immunity. With the practice of immunotherapy in PDAC still in its infancy, preclinical studies integrating and examining both micro- and macroenvironmental compartments in models that reflect local growth, metastasis, and survival endpoints are critical to the development of novel treatments for the approximately 53% of PDAC patients that are diagnosed with metastatic disease⁵.

The sustained activation and proliferation of cytotoxic T lymphocytes (CTLs) is a well-known requirement for cancer cell kill and immune surveillance in both local and systemic disease control. The role IL-2 plays in promoting the survival of CTLs and facilitating the maintenance of natural killer (NK) cells^6-9 is extensively characterized in both preclinical and clinical settings¹⁰. Most recently, IL-2 signaling has been shown to provide an alternative differentiation path from “stem-like” T cells to what has been termed “better effectors”, which are similar in function and phenotype to effector cells¹¹. However, this cytokine’s strong affinity for the non-signaling IL-2Rα-chain (CD25) and function in Treg maintenance and expansion, as well as clinical concerns for vascular leak syndrome, have made it a more challenging candidate for agonism¹².

To combat this dichotomous effect, the development of therapeutics that simultaneously stimulate IL-2 and inhibit PD-1 signaling on CTLs, all while inhibiting Treg activation and function, have emerged as a priority in immune-oncology. Here, we utilize a murine PD-1-targeted IL-2 variant antibody complex (PD1-IL2v) composed of a high-affinity anti-PD-1 antibody fused to an IL-2v with abolished binding to CD25 (IL-2Rα), which allows for the expansion of tumor-antigen specific T-cells and their differentiation towards functional immune effectors^11,13 (Fig 1A). Using this antibody complex, we report a marked effect on local and distant tumor growth that was further improved with the addition of RT. This response was associated with a profound expansion of proliferative, polyfunctional, and antigen specific CD8⁺ T cell subsets while curbing Treg immunosuppressive activity. RT and PD1-IL2v combination therapy also induced a significant systemic memory immune response resulting in an anti-tumor effect that persists after rechallenge in mice that achieve eradication.

Figure 1: PD1-IL2v treatment enhances survival of pre-clinical murine PDAC models, which is further improved by radiation therapy.

(A) 1.) Conventional IL-2 and 2.) aPD-1 therapy mechanisms. 3.) Model of PD1-IL2v antibody complex mechanism of targeted delivery of IL-2v through cis-binding of PD-1 4.) PD1-IL2v has been optimized for IL-2Rβγ binding without binding in the presence of IL-2Rα. Created with BioRender.com

(B) Kaplan-Meier survival analysis of PK5L1940 orthotopically implanted pancreatic tumor-bearing mice treated with RT, aPD-1, DP47-IL-2v, PD1-IL2v, or a combination therein.

(C) Kaplan-Meier survival analysis of PK5L1940 orthotopically implanted pancreatic tumor-bearing mice untreated and treated with PD1-IL2v with and without RT.

(D) Kaplan-Meier survival analysis of PK5L1940 orthotopically implanted pancreatic tumor-bearing mice untreated and treated with aPD-1 with and without RT.

(E) Kaplan-Meier survival analysis of PK5L1940 orthotopically implanted pancreatic tumor-bearing mice untreated and treated with DP47-IL-2v with and without RT.

(F) Kaplan-Meier survival analysis of PK5L1940 orthotopically implanted pancreatic tumor-bearing mice untreated and treated with aPD-1 + DP47-IL-2v with and without RT.

(G) Kaplan-Meier survival analysis of PK5L1940 orthotopically implanted pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v.

(H) Kaplan-Meier survival analysis of FC1242 orthotopically implanted pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v.

RT administered 7 days post-implantation. PD1-IL2v, DP47-IL-2v, and aCD25 dosed once per week beginning day 7 post-implantation. aPD-1 dosed twice per week beginning day 7 post-implantation. n ≥7 per group. See also Figure S1.

RESULTS

PD1-IL2v results in superior response than its principal components, further improved with radiation therapy

To examine the effect of the murine variant antibody fusion protein, PD1-IL2v, on pancreatic cancer growth, we utilized orthotopic mouse models of KRAS-driven PDAC with PK5L1940 and FC1242 KPC cell lines. We compared the effect of PD1-IL2v to its principal components, with and without radiation therapy (RT). Under this pretense, PD1-IL2v emerged as the clear driver of improved survival in our tumor models (Fig 1B, S1A). Interestingly, response to PD1-IL2v was even further improved with the addition of RT as a cytotoxic agent (Fig 1C, S1B). aPD-1 treatment alone conferred no additional benefit, performing similarly to untreated (Untrx) controls, and combination RT + aPD-1 treatment resulted in marginally worsened responses than each treatment individually (Fig 1D). Furthermore, utilizing a non-targeted IL-2 variant (DP47-IL-2v), we again found that non-specific IL-2 agonism did not improve survival over untreated control, and RT + DP47-IL-2v did not improve survival over each treatment alone (Fig 1E). To ensure the observed response to PD1-IL2v treatment was due to an in cis interaction occurring within the same cell, we also treated tumor bearing mice with the combination of aPD-1 + DP47-IL-2v. In this model system, there was no improvement in survival following independent PD-1 blockade and IL-2 agonism (Fig 1F), suggesting the improved response following PD1-IL2v treatment is due to the specific delivery of IL-2 to PD-1⁺ CTLs.

To investigate cis binding properties of PD1-IL2v, we conducted an in vitro assay examining STAT5 phosphorylation as a measure of IL-2R activation potency¹⁴. The response to PD1-IL2v was compared to an untargeted-IL-2v on IL-2R activation with untargeted-IL-2 wild type (WT) and aPD-1 controls. PD1-IL2v drug treatment resulted in a 40-fold increase in IL-2R activation potency compared to untargeted-IL-2v treatment (Fig S1C). As expected, untargeted delivery of wild type IL-2 resulted in the highest frequency of isolated STAT5⁺ CD4⁺ T cells, due to the upregulated expression of CD25 on activated T cells and the high binding affinity of IL-2 to CD25 (IL-2Rα)¹⁵.

Having established RT + PD1-IL2v as the superior treatment regimen in our tumor model, we next sought to explore its mechanism of action. We hypothesized that due to the PD-1 binding properties of PD1-IL2v, the conjugate antibody could potentially bind to suppressive PD-1⁺ Tregs, initiating IL-2 signaling, and promoting their activation and proliferation. Utilizing a non-IL-2 blocking aCD25 Treg depleting antibody made against IL-2Rα and modified to include ‘activating’ Fc regions, we next treated tumor bearing mice with RT and aCD25, PD1-IL2v, or the combination of antibodies. Using the PK5L1940 cell line, we found that although the addition of aCD25 to RT treatment marginally improved survival and response relative to RT alone, the conjugate PD1-IL2v antibody resulted in a superior response, an effect that was not significantly improved by the addition of aCD25 (Fig 1G). We conducted an in vitro assay to determine if aCD25 treatment prevents the activation of PD1-IL2v on IL-2R in CD4⁺ and CD8⁺ T cells. No reduction in p-STAT5 was observed when PD1-IL2v was simultaneously administered with aCD25 or IL-2R blocking aCD25 control (Fig S1D). Response to treatment was also driven by PD1-IL2v using the KPC-derived, KRAS-driven FC1242 cell line (Fig 1H). Together, these findings suggest not only that PD1-IL2v is being preferentially shuttled to effector T cells rather than Tregs, but that additional Treg inhibition is unnecessary as PD1-IL2v inherently subverts Treg-mediated suppression and expansion.

PD1-IL2v treatment results in significant infiltration, proliferation, and activation of intratumoral CD8⁺ T cells

We next sought to identify the immune subpopulations involved in the responses to these treatment regimens. Starting with monotherapy groups compared to the RT + PD1-IL2v combination, analysis by flow cytometry showed that PD1-IL2v, with or without RT, results in significant CD8⁺ T cell expansion and decreased Tregs across compartments (Fig 2A, S1E-G, S2). Our results indicate these tumors are baseline are type II adaptive immune resistance phenotype, characterized by PD-L1 expressing tumor cells and the presence of immune infiltrates, as defined in a recent review by Kim et al¹⁶ (Fig 2A, S1H). Using a flow cytometric analysis on tumor-infiltrating immune populations across monotherapy and combination therapy experiments, we observed significant expansion of the intratumoral CD8⁺ T cell population following RT + PD1-IL2v treatment that was not seen with RT or RT + aCD25 treatment (Fig 2A-B, S1E, S2A, S3A). Furthermore, although RT + aCD25 treatment was most effective of the combination radio-immunotherapy treatments in reducing the frequency of intratumoral Tregs and IL-10 expression by tumor-infiltrating CD4⁺ T cells (Fig 2C), PD1-IL2v treatment, with and without RT resulted in the largest increases of intratumoral CD8⁺ T cell to Treg ratio (Fig 2D, S2C). Both treatments also led to significantly increased proliferation in the CD8⁺ T cell compartment as shown by Ki67 positivity (Fig 2E, S2D), an effect that was specific to the CD8⁺ T cell population. Accompanied by this enhanced CD8⁺ T cell infiltration and proliferation, RT + PD1-IL2v treatment also resulted in a significant increase in CD8⁺ T cell activation as evidenced by increases in IFNγ IL-2, CD44, and TNFα expression, as well as a decrease in exhausted PD-1⁺TIM3⁺CD8⁺ T cells over RT alone (Fig 2F, S3B).

Figure 2: PD1-IL2v increases infiltration, proliferation, and activation of polyfunctional, intratumoral CD8⁺ T cells.

(A) Quantification of tumor-infiltrating immune subpopulations in untreated, RT, aCD25, PD1-IL2v, or RT + PD1-IL2v treated mice. Quantification determined by gating analysis using FlowJo software. Gating is performed on live CD45⁺ cells, n ≥6 per group. CD4⁺ T cell group does not include Tregs. Significance shown for differences in CD8⁺ T cell population frequency.

(B) Representative histogram of CD8 MFI (left) and representative gating plots (right) of intratumoral leukocytes (CD45⁺ subset) in pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Gating is performed on live CD45⁺ cells.

(C) Flow cytometric analysis of frequency of tumor infiltrating Tregs (top) and IL-10 MFI of intratumoral CD4⁺ T cells (bottom). MFI is calculated on live CD45⁺ CD4⁺ population, n ≥5 per group.

(D) Ratio of CD8⁺ T cells to Tregs in the TME of pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. n ≥5 per group.

(E) Ki67 expression across CD8⁺, CD4⁺, and Treg subpopulations in pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Gating is performed on live CD45⁺ Ki67⁺ cells, n ≥5 per group.

(F) Flow cytometric analysis of the expression of functional and activation markers IFNγ (left), IL-2 (center), PD-1 and TIM3 (right) in intratumoral CD8⁺ T cells. Gating is performed on live CD45⁺ CD8⁺ cells, n ≥5 per group.

(G) Heat map showing the expression profile of FlowSOM-generated intratumoral CD8⁺ T cell clusters collected from tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v (L) Differences in cluster frequencies by treatment (R) Clustering is performed on live CD45⁺CD8⁺ cells.

(H) Flow cytometric analysis of IFNγ expression of intratumoral CD8⁺ T cells in untreated, RT, aCD25, and PD1-IL2v treated pancreatic tumor bearing mice. Gating is performed on live CD45⁺ CD8⁺ cells, n ≥6 per group

Mice were orthotopically implanted with PK5L1940 cells. RT administered 7 days post-implantation. PD1-IL2v and aCD25 dosed once per week beginning day 7 post-implantation. Serum collected at time of sacrifice. Tumors harvested 21 days post-implantation. Data represented as mean ± SEM. P-values calculated with Students T-test, *indicates p<0.05, **<0.01, ***<0.001, ****<0.0001. See also Figures S1, S2, and S3.

To appreciate the impact RT and PD1-IL2v combination treatment has on CTL function, we next performed a viSNE clustering analysis on tumor-infiltrating CD8⁺ T cells. Through this analysis, we identified a proliferative and functional CD8⁺ subset (Cluster 13) defined by high Ki67, TNFα, and Granzyme-B (GnzmB) expression that was increased in abundance with the addition of PD1-IL2v to RT, further emphasizing the CD8-activating properties of the PD1-IL2v immunocytokine. We also found that a polyfunctional CD8⁺ subset (Cluster 8), defined by high Ki67, IL-2, IFNγ, and TNFα expression was increased in all treatment groups vs RT alone (Fig 2G). In a side-by-side comparison with monotherapies, RT + PD1-IL2v resulted in the highest proportion of IFNγ⁺ intratumoral CD8⁺ T cells, further explaining the enhanced survival benefit seen with this radiation combination compared to PD1-IL2v alone (Fig 2H).

As NK cells are another critical CTL population in mediating response to treatment^17-19, and known to express the IL-2Rβ and IL-2Rγ subunits^20,21, we next performed a similar clustering analysis on intratumoral NK cells. Through this analysis, we found that most intratumoral NK cell subsets express GnzmB and DNAM1 (Fig S3C), and RT + PD1-IL2v significantly increases DNAM1 expression compared to RT alone (Fig S3D).

To establish the cellular dependencies of these effects and the improved response to RT, we depleted CD8⁺ T cells and NK cells in tumor-bearing mice treated with RT + PD1-IL2v. Here we found that the beneficial effect of treatment was abrogated following CD8⁺ T and NK cell depletion (Fig S3E), suggesting that not only does PD1-IL2v function to enhance NK and T cell activation, but that this activation is required for the improved response to RT following PD1-IL2v treatment. Given that both CD8⁺ T cells and NK cells are required for PD1-IL2v’s function, we further sought to establish the relationship between these two cell types. In the depletion experiments, we found that NK depletion does not negatively impact CD8⁺ T cell activation and CD8⁺ T cell depletion does not negatively impact NK cell activation (Fig S3F-G). However, depletion of either subtype appeared to increase IFNγ production in the alternative population (Fig S3F-G), suggesting that these two cell types help maintain a pro-inflammatory environment, especially in the absence of one another.

Addition of aCD25 to RT plus PD1-IL2v combination blunts CD8⁺ T cell anti-tumor activity

Given the observation that concurrent administration of aCD25 did not add to the survival benefit of RT + PD1-IL2v (Fig 1G-H) and largely abolished the increase in CD8⁺ T cell/Treg ratio observed for RT + PD1-IL2v (Figure 2D), we compared CD8⁺ T cell subsets across these treatment groups (Fig 2G, Fig S3). Administration of PD1-IL2v increased CD25 (IL-2Rα) expression on intratumoral CD8⁺ T cells (Fig 2G), but the prevalence of these CD25⁺ CD8⁺ T cells was significantly reduced when combined with aCD25 treatment (Fig S4A). To further investigate the functional contribution of CD25⁺ CD8⁺ T cells towards the anti-tumor effect observed in our RT + PD1-IL2v treated mice, we performed proteomics. In comparing CD25⁺ to CD25⁻ CD8⁺ T cells derived from tumor bearing mice treated with RT + PD1-IL2v, we observe features of a more active and mobile phenotype as evidenced by enhanced RNA processing ²² and cytoskeletal reorganization of intermediate filaments, which have been shown to play a key role in rigidity changes that enable T cell tissue migration²³(Fig S4D). We also observe increased interleukin-7 signaling in CD25⁺ CD8⁺ T cells, which has been shown to promote memory T cell survival and self-renewal²⁴ as well as enhanced anti-tumor cytotoxicity^25,26. Altogether, the CD25⁺ CD8⁺ T cells appear to display characteristics suggestive of anti-tumoral response and may partially contribute to the enhanced efficacy observed when RT is added to PD1-IL2v.

Expression of CD25 aside, the triple combination also appeared to increase the frequency of exhausted (PD-1⁺TIM3⁺) CD25⁻ CD8⁺ T cells in the TME of tumor-bearing mice compared to mice treated with RT + PD1-IL2v (Fig S4B). Furthermore, the frequencies of activated (GnzmB⁺) and antigen experienced, stem-like CD8⁺T cells (PD-1⁺TCF1/7⁺)²⁷ are significantly higher in tumors treated with RT + PD1-IL2v compared to RT + aCD25 + PD1-IL2v (Fig S4C).

To isolate the direct effects of aCD25 signaling on CD8 T cell function and activation unconfounded by the indirect immune stimulation following aCD25-mediated Treg depletion, wildtype CD8⁺ T cells isolated from C57BL/6 mice were adoptively transferred into previously irradiated, tumor-bearing Rag1^−/− mice and treated with either PD1-IL2v alone or with concurrent administration of PD1-IL2v and aCD25 (Fig S4E). Flow cytometric analysis showed decreased CD69, CD44, and IL-2 expression on CD8⁺ T cells treated with the triple combination, suggesting that even in the absence of Tregs, the concurrent addition of aCD25 to RT plus PD1-IL2v negatively impacts CD8⁺ T cells’ activation and maturation (Fig S4F). To eliminate competitive binding of aCD25 and PD1-IL2v on the CD8⁺ T cells as a driving rationale for decreased CD8⁺ T cells activation, we performed a pSTAT5 assay and found no impact of aCD25 on PD1-IL2v mediated phosphorylation of STAT5 (Fig S1D).

RT + PD1-IL2v treatment results in an expanded, activated CD8⁺ T cell population in the tumor draining lymph node

Having established a critical role for CD8⁺ T cells in mediating the improved response to RT + PD1-IL2v treatment, we sought to understand the effect of our treatment regimens on CD8⁺ T cell priming and activation in the draining lymph node of tumor-bearing mice which had been orthotopically implanted with PK5L1940 cells. Using flow cytometric analysis, we found a significant increase in nodal CD8⁺ T cell frequency following PD1-IL2v treatment with or without RT (Fig 3A-B, S1F, S2E-F), similar to that in the tumor. Moreover, the RT + PD1-IL2v combination significantly increased polyfunctional (IFNγ⁺ TNFα⁺ GnzmB⁺) CD8⁺ T cells in the lymph node, even when compared to PD1-IL2v alone treatment (Fig 3C), and significantly decreased exhausted (PD-1⁺TIM3⁺) CD8⁺ T cells in the lymph node (Fig 3D). Of note, there were significantly fewer Tregs in the draining lymph node of RT + PD1-IL2v and PD1-IL2v treated mice, as well as reduced expression of FoxP3 and IL-10 in nodal CD4⁺ T cells with both treatments (Fig 3F, S2G), suggesting diminished suppression of effector T cells in the draining lymph node of PD1-IL2v treated mice. Furthermore, the reduced expression of IL-10 on CD4⁺ T cells (Fig 3F, S2G) derived from PD1-IL2v, with and without RT, treatment groups may indicate an increase in Treg fragility^28,29.

Figure 3: RT+ PD1-IL2v simultaneously increases polyfunctional CD8⁺ T cells and reduces Treg immune suppression within the tumor draining lymph node.

(A) viSNE dimensionality reduction analysis of immune subpopulations in the draining lymph node of tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Clustering is performed on live CD45⁺ cells.

(B) Quantification of immune subpopulations the draining lymph node of tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Gating performed on live CD45⁺ cells, n ≥5 per group.

(C) Flow cytometric quantification of polyfunctional (IFNγ⁺ TNFα⁺ GnzmB⁺) CD8⁺ T cells in the draining lymph node of tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v (L) or untreated, RT, aCD25, PD1-IL2v, and RT with PD1-IL2v (R) Gating is performed on live CD45⁺ CD8⁺ cells, n ≥5 per group.

(D) Flow cytometric quantification of exhausted (PD-1⁺TIM3⁺) CD8⁺ T cells in the draining lymph node of tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v

(E) viSNE dimensionality reduction analysis showing expression of IFNγ, CD44, TNFα, GnzmB, and PD-1 on CD8⁺ T cells in the draining lymph node of tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Clustering is performed on live CD45⁺ CD8⁺ cells.

(F) Flow cytometric analysis of the frequency of Tregs as a proportion of CD45⁺ cells (top), and FoxP3 (center) and IL-10 (bottom) MFI of CD4⁺ T cells in the draining lymph node of pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. MFI is calculated on live CD45⁺ CD4⁺ population, n ≥5 per group.

(G) Heat map showing the expression profile of FlowSOM-generated nodal CD8⁺ T cell clusters collected from tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Clustering is performed on live CD45⁺ CD8⁺ cells.

Mice were orthotopically implanted with PK5L1940 cells. RT administered 7 days post-implantation. PD1-IL2v and aCD25 dosed once per week beginning day 7 post-implantation. Lymph nodes harvested 21 days post-implantation. Data represented as mean ± SEM. P-values calculated with Students T-test, *indicates p<0.05, **<0.01, ***<0.001, ****<0.0001. See also Figure S4

We then performed dimension reduction analysis on nodal CD8⁺ T cells to understand changes among specific CD8⁺ subsets. This analysis revealed high expression of the activation and maturation markers GnzmB, CD44, and CD69 across CD8⁺ T cell subsets (Fig 3G). Interestingly, we found CD25 (IL-2Rα) and CD122 (IL-2Rβ) to be highly expressed in two distinct CD8⁺ T cell subsets, suggesting IL-2Rβ signaling can occur independent of IL-2Rα in our model. Together, these results are suggestive of a mechanism of enhanced CTL priming and activation in the lymph node of PD1-IL2v treated mice which is enhanced by radiation treatment.

Enhanced CTL response following PD1-IL2v treatment with and without RT is antigen-specific

We next sought to understand the antigen-specific nature of the response to PD1-IL2v treatment. To accomplish this, we adoptively transferred OVA-specific CD8⁺ T cells isolated from the blood of CD45.1 OT-I mice into wildtype C57BL/6 mice (CD45.2) inoculated with OVA-expressing pancreatic tumors (PK5L1940-OVA). Mice were irradiated within one day of adoptive transfer and PD1-IL2v was subsequently administered according to the previously described treatment regimen (Fig 4A). For purposes of identifying adoptively transferred OT-1 T cells, OVA-specific cells were defined as CD45.1⁺ CD8⁺ cells.

Figure 4: Anti-tumor response of PD1-IL2v with and without RT is antigen-specific.

(A) Experimental timeline of OT-1 CD8⁺ T cell adoptive transfer experiment. Created with BioRender.com

(B) Frequency of CD45.1⁺ CD8⁺ T cells as a proportion of live cells in OVA-expressing tumors of untreated, RT, PD1-IL2v, or RT+PD1-IL2v treated mice following CD45.1⁺ OT-1 adoptive transfer.

(C) Frequency of CD45.1⁺ CD8⁺ T cells as a proportion of live cells in the blood (left) and draining lymph node (right) of OVA-expressing tumor-bearing mice following OT-I CD8⁺ T cell adoptive transfer and no treatment, RT, PD1-IL2v or RT + PD1-IL2v treatment. Gating is performed on live cells.

(D) Flow cytometric analysis of activation and cytotoxicity marker MFI of CD45.1⁺ CD8⁺ T cells in the blood (top), draining lymph node (middle) and tumor (bottom) of OVA-expressing tumor-bearing mice following OT-I CD8⁺ T cell adoptive transfer and no treatment, RT, PD1-IL2v or RT + PD1-IL2v treatment. MFI is calculated on live CD45.1⁺ CD8⁺ population, I Flow cytometric analysis of IFNγ⁺ TNFα⁺ GnzmB⁺ CD8⁺ T cells in the blood (left) and draining lymph node (right) of OVA-expressing tumor bearing mice following OT-I CD8⁺ T cell adoptive transfer and no treatment, RT, PD1-IL2v or RT + PD1-IL2v treatment. Gating is performed on live CD3⁺ CD8⁺ cells.

(F) Flow cytometric analysis of DNAM⁺ NKp46⁺ cells in the blood of OVA-expressing tumor bearing mice following OT-I CD8⁺ T cell adoptive transfer and no treatment, RT, PD1-IL2v or RT + PD1-IL2v treatment. Gating is performed on live cells.

(G) Flow cytometric analysis of activation and cytotoxicity marker MFI of CD45.1⁻ and CD45.1⁺ CD8⁺ T cells in the blood (top) and draining lymph node (bottom) of OVA-expressing tumor-bearing mice following OT-I CD8⁺ T cell adoptive transfer and RT + PD1-IL2v treatment. MFI is calculated on live CD45.1⁻ CD8⁺ and CD45.1⁺ CD8⁺

(H) Pancreatic tumor incidence 5.5 weeks post-implantation in OVA-expressing tumor-bearing mice following OT-I CD8⁺ T cell adoptive transfer and no treatment, RT, PD1-IL2v or RT + PD1-IL2v treatment.

Mice were orthotopically implanted with PK5L1940-OVA expressing cells. PD1-IL2v dosed once per week beginning day 7 post-implantation. All tissue collected 39 days post implantation, n ≥5 per group. Data represented as mean ± SEM. P-values calculated with Students T-test, *indicates p<0.05, **<0.01, ***<0.001, ****<0.0001. See also figure S5.

Prior to tissue harvest, we performed a cheek bleed to capture early systemic immune differences between untreated, RT alone, PD1-IL2v alone, and RT + PD1-IL2v combination treated mice. We observed a trend towards the frequency of circulating OVA specific CD8⁺ T cells when RT was added to PD1-IL2v (Fig S5A). Additionally, PD1-IL2v with and without RT significantly reduced the magnitude of suppressive CD4⁺ T cells in circulation (Fig S5A).

In this system, we observed a significant increase in the number of tumor-infiltrating CD45.1⁺ CD8⁺ T cells in mice that received RT + PD1-IL2v treatment (Figs 4B, S5B). We also observed significant increases in OVA-specific CD8⁺ T cell frequency in both the circulating blood and the draining lymph node (tdLN) of tumor-bearing mice treated with RT + PD1-IL2v relative to untreated and RT alone (Figs 4C, S5C). A pronounced increase in tdLN OVA-specific CD8⁺ T cells was also detected for RT + PD1-IL2v combination compared to PD1-IL2v monotherapy, while blood levels were similar. From a functional perspective, CD45.1⁺ CD8⁺ T cells in the circulating blood of mice treated with PD1-IL2v with or without RT were found to have increased expression of the activation markers IFNγ, IL-2, GnzmB, and CD218¹¹ (Fig 4D). CD44, CXCR3, Ki67 and TNFα expression were also found to be increased in draining lymph node-resident CD45.1⁺ CD8⁺ T cells following RT + PD1-IL2v treatment (Fig 4D). Additionally, both circulating and intranodal polyfunctional CD8⁺ T cells as well as activated circulating NK cells in the blood were increased with RT + PD1-IL2v combination treatment (Fig 4E,F). However, RT + PD1-IL2v treated OVA-specific CD8⁺ cells in the lymph node were found to have decreased CD62L expression, suggesting an increased migratory capacity and characteristic of an effector-memory CD8⁺ T cell phenotype³⁰ (Fig S5D). Furthermore, OVA-specific CD8⁺ T cells derived from tumors treated with PD1-IL2v with and without RT exhibited higher expression of activation markers CD69, IL-2, and IFNγ and lower expression of the exhaustion marker, TIM3 (Fig 4D).

We then narrowed our analysis to mice treated with RT + PD1-IL2v to examine the effect of PD1-IL2v treatment as a function of antigen-specificity. Through this analysis, we found that antigen-specific CD45.1⁺ CD8⁺ T cells had significantly increased expression of polyfunctional, stemness, and activation markers CD44, TCF1/7, TNFα, IFNγ, CD218, GnzmB, and CXCR3 compared to their CD45.1⁻, antigen non-specific counterparts in both the circulating blood and the tdLN of mice treated with RT + PD1-IL2v (Fig 4G). Circulating and nodal antigen-specific CD8⁺ T cells also exhibited less exhaustion by reduced expression of PD-1 and TIM3, as well as reduced CD62L expression compared to antigen non-specific CD8⁺ cells following PD1-IL2v treatment (Fig S5E). This suggests not only a preferential increase in activation and decrease in exhaustion of tumor-educated CD8⁺ T cells, but also a shift toward an effector memory phenotype in antigen-specific CD8⁺ T cells after RT + PD1-IL2v treatment. These observed immunostimulatory effects ultimately correlated with a tumor incidence of 100% , 75%, 62.5%, and 28.6% for untreated, RT, PD1-IL2v, and RT + PD1-IL2v, respectively, at 5.5 weeks post implantation (Fig 4H). These results corroborate with a pilot OVA-experiment conducted for RT vs RT + PD1-IL2v (Fig S5F).

To validate the claim of enhanced function of tumor-specific CD8⁺ T cells, we directly assayed the effect of RT + PD1-IL2v treatment on the cytotoxic potential of antigen-specific CD8⁺ T cells in vitro. To do this, we harvested CD8⁺ T cells treated with RT or RT + PD1-IL2v in vivo from OVA-expressing tumor-bearing mice (PK5L1940-OVA) that contained adoptively transferred OVA specific CD45.1⁺ OT-I T cells. We then incubated these isolated CD8⁺ T cells with OVA-expressing pancreatic cancer cells in vitro and quantified the resulting cancer cell death. We found that CD8⁺ T cells harvested from RT + PD1-IL2v treated OVA-expressing tumor bearing mice induced significantly more cancer cell death compared to their RT treated counterparts (Fig S5G). Taken together, these data are suggestive of a critical antigen-specific nature of the response to RT + PD1-IL2v treatment and illustrate a mechanism by which IL-2 agonism via PD1-IL2v may elicit a durable anti-tumor memory response.

RT + PD1-IL2v combination treatment results in activation of peripheral CTLs, reduced metastatic burden

Given the antigen-specific nature of the CD8⁺ T cell response to PD1-IL2v, as well as the increase in intratumoral and nodal CD8⁺ T cells following PD1-IL2v treatment, we hypothesized that tumor-specific CD8⁺ T cells not only function to control local tumor progression, but also contribute to combatting metastatic spread and disease dissemination. To test this concept, we began by quantifying the frequency of circulating tumor cells (CTCs) in the blood of pancreatic tumor-bearing mice treated with RT and aCD25 or PD1-IL2v. Using a flow cytometry-based analysis, we found that administration of PD1-IL2v, but not aCD25, in combination with RT decreased the presence of tumor cells in circulation (Fig 5A).

Figure 5: PD1-IL2v in combination with radiotherapy increases peripheral activation of anti-tumor immune populations and reduces metastatic burden.

(A) Flow cytometric analysis of the frequency of circulating tumor cells (CTCs) in the blood of pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. CTCs defined as CD45⁻ EpCAM⁺ cells. Gating is performed on all live cells, n ≥5 per group.

(B) viSNE dimensionality reduction analysis of circulating CD4⁺, CD8⁺, and NK cells in the blood of pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Clustering is performed on live CD45⁺ cells.

(C) Quantification of the immune subpopulations in the blood of pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Gating performed on live CD45⁺ cells, n ≥5 per group.

(D) Representative histogram of CD8 MFI (left) and representative gating plots (right) of circulating leukocytes (CD45⁺) in pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Gating is performed on live CD45⁺ cells.

(E) Ki67 expression across peripheral CD8⁺, CD4⁺, and Treg subpopulations in pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. Gating is performed on live CD45⁺Ki67⁺ cells, n ≥5 per group.

(F) viSNE dimensionality reduction analysis performed on the circulating CD8⁺ T cell population showing relative expression of Ki67, TNFα, IFNγ, and CD44 in mice treated with RT, aCD25, and/or PD1-IL2v. Clustering is performed on live CD45⁺ CD8⁺ cells.

(G) Flow cytometric analysis of the frequency of NK cells (left), and NK cell expression of functional markers DNAM1 (center) and GnzmB (right) in the blood of pancreatic tumor-bearing mice treated with RT, aCD25, and/or PD1-IL2v. n ≥5 per group.

(H) Representative images of hepatic metastatic lesions following hemi-splenic implantation at the time of sacrifice in mice untreated or treated with RT, aCD25, and/or PD1-IL2v.

(I) Kaplan-Meier survival analysis of mice implanted with PK5L1940 using a hemi-splenectomy model of metastatic pancreatic cancer. Mice were untreated or treated with RT, aCD25, and/or PD1-IL2v. n ≥7 per group.

Mice were orthotopically implanted with PK5L1940 cells. RT administered 7 days post-implantation. PD1-IL2v and aCD25 dosed once per week beginning day 7 post-implantation. Blood collected 21 days post-implantation. Data represented as mean ± SEM. P-values calculated with Students T-test, *indicates p<0.05, **<0.01, ***<0.001, ****<0.0001. See also Figure S2.

To uncover potential mechanisms of this reduced metastatic burden following RT + PD1-IL2v treatment, we next performed a flow cytometric analysis on circulating immune populations. Like the tumor and tdLNs, CD8⁺ T cells were significantly expanded in the blood of tumor-bearing mice treated with PD1-IL2v with or without RT (Fig 5B-D, S2H). Moreover, we found that circulating CD8⁺ T cells in the RT + PD1-IL2v and PD1-IL2v groups more frequently expressed the proliferative marker Ki67 (Fig 5E, S2I). Dimension reduction analysis revealed peripheral CD8⁺ T cells treated with RT + PD1-IL2v also had increased expression of the functional and activation markers TNFα, IFNγ, and CD44 (Fig 5F).

Due to their role in controlling metastatic spread^18,31, we shifted our analysis to peripheral NK cells. Although the frequency of circulating NK cells was decreased following aCD25 and PD1-IL2v treatments, we found that expression of the functional markers DNAM1 and GnzmB were increased with the addition of PD1-IL2v (Fig 5G,S2J), suggesting PD1-IL2v may also function to reduce metastatic burden by activating NK cells in the periphery.

To directly test the effect of this immune modulation on metastasis and disease dissemination, we next utilized a hemi-splenectomy metastatic model of pancreatic cancer wherein tumor cells are injected through splenic vessels and reproducibly form metastatic lesions in the liver³². Using this model with the PK5L1940 cell line, we found that the addition of PD1-IL2v to the treatment regimen results in a reduction in gross hepatic metastatic lesions at the time of sacrifice and improves survival over untreated, RT, and RT + aCD25 treated mice (Fig 5H-I). Furthermore, a complete response was observed in 1/8 mice in the RT + PD1-IL2v treatment group and 2/8 mice in the RT + aCD25 + PD1-IL2v treatment group, highlighting the robust and systemic nature of the PD1-IL2v induced anti-tumor immune response. Complete response with these therapy combinations was also achieved in this model using the FC1242 cell line (Fig 6G).

Figure 6: RT + PD1-IL2v with and without aCD25 induces an effector memory CD8⁺ T cell phenotype in tumor-eradicated mice.

(A) Tumor growth curves of PK5L1940 flank tumors following tumor rechallenge in tumor-eradicated mice treated with RT + PD1-IL2v or RT + aCD25 + PD1-IL2v.

(B) Flow cytometric analysis of the frequency of circulating CD8⁺ T cells early (21 days post-implant), prior to (4 days), and following (4 days) rechallenge in tumor-eradicated mice in (A)

(C) Flow cytometric analysis of Ki67 MFI in circulating CD8⁺ T cells early (21 days post-implant), prior to (4 days), and following (4 days) rechallenge in tumor-eradicated mice in (A) CD69 (D), Eomes (E), and CD62L (F) MFI in circulating CD8⁺ T cells prior to (4 days) and following (4 days) rechallenge in tumor-eradicated mice in (A)

(G) Kaplan-Meier survival analysis of mice implanted with FC1242 cells using a hemi-splenectomy model of metastatic pancreatic cancer. RT administered 7 days post-implantation. PD1-IL2v and aCD25 dosed once per week beginning day 7 post-implantation Treatment groups: RT alone (n ≥7), RT + aCD25 (n<7), RT + PD1-IL2v (n≥7), and RT + aCD25 + PD1-IL2v (n≥7)

(H) Tumor growth curves of FC1242 flank tumors following tumor rechallenge in tumor-eradicated mice treated with RT + PD1-IL2v or RT + aCD25 + PD1-IL2v. Black circles denote ulcerated tumors.

(I) Flow cytometric analysis of Treg functionality prior to (3 days) and following (4 days) rechallenge in tumor-rechallenge mice in (H)

(J) Flow cytometric analysis of Treg (CD45⁺ CD4⁺ FoxP3⁺ CD25⁺) to CD45⁺ CD8⁺ cell ratio prior to (3 days) and following (4 days) rechallenge in tumor-rechallenge mice in (H)

(K) CD62L MFI in circulating CD8⁺ T cells prior to (3 days) and following (4 days) rechallenge in tumor-rechallenge mice in (H). Gating performed on CD45⁺ CD8⁺ CD62L⁺

(L) Flow cytometric analysis of CD69⁺ circulating CD8⁺ T cells prior to (3 days) and following (4 days) rechallenge in tumor-rechallenge mice in (H)

Data represented as mean ± SEM. P-values calculated with Students T-test, *indicates p<0.05, **<0.01, ***<0.001, ****<0.0001.

RT + PD1-IL2v with and without aCD25 induces an effector memory CD8⁺ T cell phenotype in tumor-eradicated mice

Having observed complete tumor eradication in our metastatic tumor model, we next sought to explore the immunological mechanisms of tumor rejection and determine whether RT + PD1-IL2v combination with and without aCD25 had induced a durable memory response. To do this, we rechallenged cured mice by implanting PK5L1940 tumors in the flank. Upon tumor rechallenge, we observed a brief period of tumor growth followed by tumor rejection (Fig 6A).

To elicit more of an understanding of the immune response leading to tumor rejection of the PK5L1940 tumors, we performed flow cytometry on blood harvested from cured mice prior to and following tumor rechallenge. Through this analysis, we found that circulating CD8⁺ T cell frequency and proliferation (as determined by Ki67 expression) was significantly reduced from the time of initial treatment to the time of rechallenge (Fig 6B-C), presumably due to the discontinued PD1-IL2v administration. However, circulating CD8⁺ T cells were found to have increased CD69 and Eomes expression and decreased CD62L expression (Fig S6D-F) post-rechallenge as compared to pre-rechallenge, consistent with the induction of an effector memory phenotype^30,33,34.

We also performed a flank tumor rechallenge in cured mice of the FC1242 hemispleen model (Fig 6G,H). We again observed a period of tumor growth followed by tumor rejection in this model; however, two of five flank tumors of the triple combination group were euthanized prior to tumor rejection due to flank tumor ulceration (Fig 6H). Interestingly, the anti-tumor immune response towards the rechallenged FC1242 flank tumors exhibited a higher reduction in Treg functionality (Fig 6I). Unlike the PK5L1940 rechallenge, an increase in CD62L MFI of CD8⁺ T cells within the circulating blood, indicative of a lymph node-homing phenotype³⁰, was observed post-rechallenge (Fig 6J), but both rechallenges caused a trend towards higher CD8⁺ T cell activation, evidenced by CD69 expression (Fig 6D,K). Interestingly, the FC1242 tumor rechallenge also resulted in a higher frequency of activated (DNAM1⁺) NK cells in circulation (Fig 6L). Together, these data indicate that RT + PD1-IL2v mediated activation of effector T cells results in not only an initial anti-tumor immune response, but also establishment of a durable T-cell mediated memory response.

RT + PD1-IL2v treatment enhances protein translation and induces metabolic activity in the proteome of CTLs

Given the profound effects of our treatment regimens on CTL expansion and activation, we next performed proteomics analyses on sorted CD8⁺ T cells and NK cells to determine the impact on the cell proteomes of RT alone or in the presence of aCD25, PD1-IL2v or the combination of the two (Fig 7). Interestingly, unsupervised analyses – including principal component analyses (Fig 7A, Fig S6A) and hierarchical clustering analysis of the top 1500 proteins by ANOVA (Fig 7B, Fig S6B) showed a stronger impact of treatment, rather than cell type, on the observed changes in protein expression as 1164 proteins were differentially-expressed based on treatment regimen as compared to 206 differentially-expressed proteins based on cell type (240 proteins were shared among gene sets). More specifically, RT + PD1-IL2v and the triple combination treatment—but not RT + aCD25—had a pronounced effect on the cellular proteomes of both CD8⁺ T cells and NK cells, with similar phenotypes for both. Pathway analysis identified a significant impact of PD1-IL2v on translational capacity (ribosome and spliceosomal components) (Fig 7C) and T cell receptor signaling and metabolism (Fig 7E-F). NF-kβ was identified as the node with the highest degree in the interactome of differentially expressed proteins (Fig 7D). Similarly, several components of the T cell receptor signaling pathway (KEGG ID: hsa04660) were upregulated in response to PD1-IL2v, including CD8α, Akt1 and 2, CD247, CDC42, and NFATC1 (Fig 7E). Specifically, enzymes of the Krebs cycle and enzymes involved in fatty acid metabolism and degradation were found to be down- and upregulated, respectively, in PD1-IL2v and combination treated CD8⁺T and NK cells, consistent with dysregulated fatty acid metabolism in sera (Fig S7G). Together these data uncover a mechanism of enhanced metabolic and translational activity in CTLs following IL-2Rβγ agonism.

Figure 7: Distinct proteome differences observed between mice treated with RT and/or aCD25 versus those treated with PD1-IL2v or combination.

(A) Three-dimensional principal component analysis of proteome in NK and CD8⁺ T cells isolated from mice treated with RT, aCD25, and/or PD1-IL2v (left). Venn diagram showing significantly changed proteins based on treatment regimen and cell type as determined by ANOVA (right).

(B) Hierarchical clustering analysis of the top 1500 proteins in RT, aCD25, and/or PD1-IL2v treated mice as determined by ANOVA.

(C) Table of significantly altered metabolic pathways in CTLs harvested from RT + PD1-IL2v vs. RT treated mice.

(D) Network view of the interactome of differentially expressed proteins in CTLs harvested from RT+ PD1-IL2v vs. RT treated mice.

(E) Quantification of T cell receptor signaling pathway proteins in CD8⁺ T cells harvested from RT, aCD25, and/or PD1-IL2v treated mice.

(F) Quantification of TCA, Glycolysis, and Fatty acid metabolism enzymes in CD8⁺ T cells harvested from RT, aCD25, and/or PD1-IL2v treated mice.

Mice were orthotopically implanted with PK5L1940 cells. RT administered 7 days post-implantation. PD1-IL2v and aCD25 dosed once per week beginning day 7 post-implantation. CD8⁺ and NK cells isolated 17 days post-implantation. n=3 per group. See also Figure S6,7,8.

RT + PD1-IL2v improves response to RT and induces systemic metabolic changes

To assess the systemic effects of the various treatment regimens, we next performed metabolomics analyses of sera from these mice, which showed a significant impact of all treatments as determined by partial least squares discriminant analysis (PLS-DA) (Fig S7A) and hierarchical clustering analysis of the top 50 metabolites by ANOVA (Fig S7B). We found that the branched chain amino acids (BCAAs) leucine/isoleucine and valine were marginally decreased in abundance following aCD25 and PD1-IL2v treatment (Fig S7C). As BCAAs have been shown to play a significant role in the metabolic activity that supports Treg maintenance and have been linked to Treg abundance and production of the immunosuppressive cytokine IL-10³⁵, this finding suggests a less functional Treg phenotype following aCD25 and PD1-IL2v treatment. Furthermore, the abundance of the metabolite kynurenine, a tryptophan catabolite important for Treg induction, expansion, and function³⁵, was unchanged following aCD25 and PD1-IL2v treatment (Fig S7D), again suggesting a reduction in systemic immunosuppression.

Of note, all treatments promoted decreases in the levels of circulating markers of proteolysis/fibrosis (trans-hydroxyproline) (Fig S7E). Most notably, circulating levels of free fatty acids increased following RT + PD1-IL2v treatment (especially saturated 14:0, 16:0, monounsaturated 16:1 and 18:1, poly-unsaturated 18:2, 18:3, 20:3 and highly unsaturated 20:5) (Fig S7G). This finding is of particular relevance as fatty acid metabolism has been recently shown to play a regulatory role in CD8⁺ T cell proliferation³⁶ and function³⁷, especially CD8⁺ memory T cell function³⁸ and activation^39,40. Increases in lactate and decreases in pyruvate were also observed following all treatments, especially PD1-IL2v (Fig S7F).

To follow up on these findings and determine the relative contribution of CD8⁺ T cells to alterations of the circulating metabolome, we performed steady state metabolomics analyses on isolated CD8⁺ T cells from spleen and blood of mice upon RT, aCD25, PD1-IL2v alone or in combination with RT, RT + PD1-IL2v (heat map of top 50 significant metabolites by ANOVA in Fig S8A). Most notable changes were observed in the aCD25 group, followed by PD1-IL2v (alone or in combination with RT) with respect to increases in the levels of 2-hydroxyglutarate compared to RT alone (Fig S8B). This is relevant because 2-hydroxyglutarate is an immunometabolite recognized for its role in bolstering CD8⁺ T cell proliferation, persistence, and anti-tumor activity⁴¹. However, only aCD25 treatment resulted in CD8⁺ T cell decreases in the levels of several amino acids, including serine, threonine, alanine, and – most importantly – glutamate and glycine (Fig S8A), building blocks of the key antioxidant tripeptide glutathione. Indeed, glutathione and its catabolites from the gamma-glutamyl-cycle (e.g., 5-oxoproline) were elevated in CD8⁺ T cells in response to PD1-IL2v treatment, either alone or in combination with RT (Fig S8B).

Given the limitation of steady state measurements, we then performed ex vivo tracing experiments with U-¹³C₆-glucose to determine whether increases in circulating levels of serum lactate were at least partly associated with increases in glycolytic fluxes in CD8⁺ T cells from mice treated with RT alone, aCD25, PD1-IL2v or combination treatment, RT + PD1-IL2v (heat map of top 50 significant metabolites by ANOVA in Fig S8C). Lactate, a byproduct of glycolysis, has been reported to increase CD8⁺ T cell stemness to augment anti-tumor immunity²⁷. We show a significantly increased accumulation of 1,2,3-¹³C₃-lactate in response to PD1-IL2v, especially from animals treated with combination therapy, RT+ PD1-IL2v (Fig S8D), following a 6-hour incubation with labeled glucose. This result corroborates with elevated frequency of TCF1/7⁺ CD8⁺ T cells observed in tumors treated with RT + PD1-IL2v (Figs 4G, S4C). Moreover, labeled glucose-derived carbon flux into the Krebs cycle was not significantly different across groups, except for the metabolite succinate whose ¹³C₂ labeled isotopologue was significantly higher in the combination treatment compared to the other groups (Fig S8D). Notably, succinate accumulation in CD8⁺ T cells has been previously associated with more functional effector and memory cells, providing superior protection against influenza virus in rodent models⁴². The RT + PD1-IL2v results are consistent with an increased flux through glycolysis coupled with incomplete glucose oxidation in the Krebs cycle (pathways depicted in Fig S8E). The observation that LDHA levels decrease and de novo lactate synthesis increases can be interpreted as a function of increased expression of alternative LDH isoforms with differential affinity for pyruvate and lactate (LDHB compared to LDHA), slower glucose oxidation in mitochondria, or increased consumption of alternate substrates to glucose in mitochondria. These data are consistent with the published literature where such dichotomy between lactate and LDHA or LDH2 has been associated with CD8⁺ T cell stemness ^27,43. Also consistent with this is our data showing increased expression of TCF1/7, a stemness marker, in intratumoral CD8⁺ T cells following RT + PD1-IL2v treatment. Similarly, we observe a trend for higher TCF1/7 expression on adoptively transferred, antigen specific CD8⁺ T cells within the OVA experiment (Fig 4). Together, these findings are suggestive of a shift toward a more metabolically active, multi-functional CTL phenotype following PD1-IL2v treatment.

Clinical response to RT is associated with increased IL-2Rβ and IL-2Rγ expression, decreased ILR2α and exhaustion markers

To help predict the translational impact of RT + PD1-IL2v treatment in pancreatic cancer patients, we next performed RNA sequencing on paired patient PDAC tumor tissue samples prior to and following RT. Through this analysis, we observed a significant increase in tumor expression of both PD-L1 and PD-1 (Fig 8A), an increase in the high-affinity IL-2 receptor subunit, IL-2Rα (CD25), and a decrease in the intermediate-affinity subunit, IL-2Rβ (CD 122) post-RT (Fig 8B). Stratifying this analysis by response to treatment, we found that responders had significantly decreased IL-2Rα expression and significantly increased IL-2Rβ and IL-2Rγ (CD132) expression compared to non-responders (Fig 8C), suggesting not only that RT may contribute to PD-1 mediated immune exhaustion, but that additional signaling via IL-2Rβ may overcome treatment resistance. Furthermore, metabolomics analysis on the serum of PDAC trial patients also revealed significant increases in the branched chain amino acids (BCAAs) leucine/isoleucine and valine, as well as the tryptophan metabolite kynurenine (Fig 8D-E) following RT, which are two metabolites that correlate with tumorigenesis and immunosuppression, respectively^44,45, and were not significantly altered following RT + PD1-IL2v treatment in our preclinical models.

Figure 8: Patient clinical response to RT is associated with significant changes in IL-2R signaling.

(A) Normalized expression of IL-2Rα, IL-2Rβ, and IL-2Rγ in patient PDAC tumor tissue samples before (n=26) and after (n=29) RT as determined by RNA sequencing.

(B) Normalized expression of PD-L1 and PD-1 in patient PDAC tumor tissue samples before (n=26) and after (n=29) RT as determined by RNA sequencing.

(C) Normalized transcriptional levels of IL-2Rα, IL-2Rβ, and IL-2Rγ of patient PDAC tumor tissue samples collected before RT stratified by responders (n=11) and non-responders (n=4).

(D) Normalized concentration of branched chain amino acids (BCAAs) leucine/isoleucine and valine in the serum of PDAC patients at baseline (n=12), during (n=12), and post (n=11) RT treatment.

(E) Normalized concentration of the metabolite kynurenine in the serum of PDAC patients at baseline (n=12), during (n=12), and post (n=11) RT treatment.

(F) Normalized enrichment score (NES) of the 20 most upregulated pathways in responders as compared to non-responders to RT treatment. Blue bars represent non-significant changes (p>0.05). Red bars represent significant changes (p<0.05).

(G) Enrichment curve showing enrichment of the IL-2/STAT5 signaling pathway in responders compared to non-responders to RT. NES=1.42; p.adj=0.046.

(H) Bar chart showing the genes most frequently appearing in the top 10 most upregulated pathways in responders relative to non-responders.

(I) viSNE dimensionality reduction analysis of PBMCs collected from PDAC patients performed on live CD45⁺CD3⁺ cells showing expression of the immune cell markers CD8, CD4, and FoxP3, and the exhaustion marker PD-1 before, during, and after RT.

Data represented as mean ± SEM. P-values calculated with Students T-test, *indicates p<0.05, **<0.01, ***<0.001, ****<0.0001. See also Figures S5 and S7.

To provide further evidence for the clinical utility of RT + PD1-IL2v treatment, we next performed gene set enrichment analysis (GSEA) on responders and non-responders to RT treatment. Evaluating our data across the Hallmark gene sets, we found that many of the most significantly upregulated pathways in responders were related to enhanced immune activity including enrichment of the inflammatory response, IFNγ and IFNα signaling, and IL-2/STAT5 signaling (Fig 7F-G). Interestingly, IL-2Rβ was among the genes most frequently appearing in the 10 most upregulated pathways in responders (Fig 8H), further advocating for the integration of IL-2Rβ agonism into standard cytotoxic treatment regimens.

Finally, to understand the systemic effect of RT on the activation state of circulating immune populations, we subjected PBMCs collected and isolated as part of a Phase I radiation dose escalation trial (NCT02873598) before, during, and after RT treatment to a mass cytometry (CyTOF) analysis. Utilizing a viSNE clustering analysis performed on the live CD45⁺CD3⁺ population, we found that PD-1 expression was preferentially increased in the CD8⁺ T cell subset following RT and remained unchanged in the CD4⁺ and FoxP3⁺ populations (Fig 8I), lending evidence to suggest that systemic exhaustion of CD8⁺ T cells may contribute to treatment failure in PDAC. As immune checkpoint inhibitor trials have yet to show success in PDAC⁴⁶, these data suggest that although PD-1 mediated immune exhaustion is a critical component of acquired resistance to RT, sustained IL-2Rβ signaling in cytotoxic immune populations through the use of agents like PD1-IL2v may provide additional benefit in improving the therapeutic efficacy of RT.

DISCUSSION

As meaningful responses remain exceedingly rare in pancreatic cancer, the immunomodulatory mechanisms of responses that do arise require delicate and tedious characterization to fully realize the potential of immunostimulatory treatment modalities. For decades, IL-2 has been shown to be a superkine with the ability to both stimulate and dampen immune responses through the activation of effector and suppressive T cell populations⁴⁷. However, the inability to circumvent IL-2-mediated toxicity and severe side effects has limited its utility in clinical trials^47,48. Here we investigated the use of a novel murine bispecific PD1-IL2v antibody construct designed to selectively bind to IL-2Rβγ, but not IL-2Rα, while simultaneously binding to PD-1 on effector immune cells^11,49, thus overriding effector immune cell exhaustion, enhancing their proliferation and expansion, and stimulating their cytotoxic abilities.

We began this work by assessing the individual contribution of the components of PD1-IL2v treatment to its therapeutic efficacy. We hypothesized that aside from expanding and activating effector T cells, the selective inhibition of PD-1 on non-IL-2Rα expressing cells by PD1-IL2v would have the added benefit of preventing IL-2-mediated maintenance of PD-1⁺Tregs and keep these immunosuppressive cells in an exhausted state. We tested this theory by comparing the effects of PD1-IL2v to aPD-1 and IL-2 agonism, either alone or in combination. In this setting, we found PD1-IL2v to be the primary driver of response, with RT + PD1-IL2v significantly outperforming RT, aPD-1, DP47-IL-2v, or any combination therein. While PD1-IL2v alone treatment improved survival, stimulated anti-tumor CD8⁺ T cell response, and reduced Treg burden, tumor eradication was only achieved using combination RT and PD1-IL2v. This synergistic effect was also demonstrated by the marked increase in polyfunctional (IFNg⁺ GnzmB⁺ TNFα⁺) CD8⁺ T cells within the nodal compartment of the RT + PD1-IL2v combination treatment over PD1-IL2v treatment alone.

Moreover, as PD-1 blockade has recently been shown to contribute to tumor progression by activating tumor-infiltrating PD-1⁺ Tregs⁵⁰, we next postulated that Treg depletion would confer an improved survival advantage in our orthotopic models. To test this hypothesis, we employed a non-IL-2 blocking aCD25 Treg depleting antibody. We found that adding aCD25 to the RT + PD1-IL2v treatment regimen did not result in any measurable improvement in response or CTL activation. As competitive binding of aCD25 and PD1-IL2v on the CD8⁺ T cells was not occurring, we discovered that the addition of aCD25 to PD1-IL2v plus RT reduces an active CD8⁺ CD25⁺ T cell population, leading to an overall decrease in the CD8⁺ T cell to Treg ratio.

Perhaps the most translationally impactful result of this work is the finding PD1-IL2v in combination with RT results in an antigen-specific anti-tumor response. Using genetically engineered mouse models and leveraging the diversity of leukocyte isoforms, we show tumor-specific CD8⁺ T cells more frequently infiltrate the tumor microenvironment, highly express a myriad of activation markers including IFNγ, CD44, and CXCR3, and are more efficient at directly killing cancer cells when treated with RT + PD1-IL2v. As antigen specific T cells frequently have high PD-1 expression⁵¹, these results would suggest shuttling IL-2 to tumor-experienced PD-1⁺ CTLs with the use of agents like PD1-IL2v maintains their proliferative and functional capacity, ultimately resulting in a significantly improved survival. Indeed, recent work has shown the differentiation of antigen-specific, PD-1⁺ TCF1/7⁺ CD8⁺ T cells to be critical to the success of immunotherapies^11,49, providing rationale for the selective agonism of PD-1⁺ leukocytes. Here, we show RT + PD1-IL2v increases the frequency of these antigen-experienced, stem-like CD8⁺ T cells, which have previously been shown to play a critical role in maintaining an anti-tumor immune response ^52,53. Moreover, our metabolic analyses corroborate the increase in stem-like CD8⁺ T cells following RT + PD1-IL2v treatment through observed increased lactate synthesis and decreased LDHA levels.

Furthermore, in this study we show the antigen-specific nature of the response by PD1-IL2v culminates in the differentiation of a CD8⁺ effector memory subpopulation. Using our preclinical models, we found tumor-eradicated mice previously treated with RT + PD1-IL2v successfully reject tumors following rechallenge. Additionally, tumor rechallenge induced an upregulation of CD69 and Eomes and a downregulation of CD62L in circulating CD8⁺ T cells, an expression pattern hallmark of an effector memory phenotype^30,33,34 and suggestive of a tumor-specific memory immune response. Although the biology of memory CD8⁺ T cell maturation is nuanced and highly context-dependent, recent evidence suggests a strong impact of IL-2Rα signaling on the fate of memory cells, with increasing IL-2 driving effector CD8⁺ T cells progressively toward terminal differentiation⁵⁴. Whether the observed effect of PD1-IL2v on memory CD8⁺ populations is due to a lack of affinity for the IL-2Rα subunit or the prevention of terminal exhaustion through the inhibition of PD-1 signaling, the data presented here would suggest sustaining the activation of PD-1⁺ T cells may promote a durable and robust anti-tumor immune response.

Finally, the significant role that RT plays in enhancing response to PD1-IL2v treatment cannot be overlooked. In theory, the therapeutic efficacy of radioimmunotherapies is contingent on radiation-induced release of neoantigen, uptake by dendritic cells, and priming of tumor-specific T cells^55-57. In this work, we show that in the absence of radiation, the response to PD1-IL2v treatment is significantly reduced. We attribute this effect to the likely reduced neoantigen release and fewer tumor-antigen reactive dendritic cells in the tumor-draining lymph node of mice that were not treated with RT. This finding further implicates dendritic cell activity and T cell priming as a critical upstream mediator of response to radioimmunotherapeutic strategies and argues for the use of RT or other cytotoxic therapies in combinatorial approaches to treatment of pancreatic cancer. Further studies are needed to explore this concept.

To summarize, the results presented here illustrate a profound activation of polyfunctional and antigen-specific CTL subsets following RT + PD1-IL2v treatment resulting in significantly improved survival, and advocate for the use of this antibody complex in PDAC translational settings.

STAR METHODS

RESOURCE AVAILABILITY

Lead contact

Additional information and any request for resources or materials should be directed to and will be fulfilled by the lead contact, Dr. Sana Karam (sana.karam@cuanschutz.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE⁶² partner repository with the dataset identifier PXD040266. All metabolomics data have been uploaded to the NIH Common Fund's National Metabolomics Data Repository (NMDR) website, the Metabolomics Workbench⁶³, where it has been assigned Study ID ST002486. RNA sequencing data has been deposited in NCBI's Gene Expression Omnibus⁶⁴ (GEO: GSE225767). This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and subject details

Patients and samples

Written consent was obtained for all tumor sample collection. Studies were performed in accordance with U.S. Common Rule and approved by institutional review board. Patient archival tumor samples were identified and obtained from the University of Colorado PSR biorepository and collected per COMIRB13-0315. Samples were selected from all borderline resectable pancreatic cancer patients seen through University of Colorado pancreatic multidisciplinary clinic between 1/2013-12/2018 and were treated with FOLFIRINOX or gemcitabine based neoadjuvant chemotherapy and stereotactic body radiotherapy (SBRT). Following neoadjuvant therapy, all patients received surgery followed by further adjuvant chemotherapy.

Cell lines and reagents

PK5L1940 mouse pancreatic adenocarcinoma cell line was kindly provided by Dr. Michael Gough (Providence Cancer Institute, Portland, OR). FC1242 mouse pancreatic adenocarcinoma cell line was kindly provided by Dr. David Tuveson (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY).

To generate PK5L1940-OVA cells: Five hundred thousand HEK293-FT cells underwent transfection with 2ug of pLVX-puro-cOVA-IRES-BFP (Addgene plasmid #135074) and 2ug of packaging viral mixed at a 1:2 ratio of psPAX2 (Addgene plasmid #12260) and pMD2.G (Addgene plasmid #12259) to generate lentiviral (LV) particles in a six-well plate. Two ml of LV was collected 3 days post-transfection. One mL LV was used to transduce 500,000 LY2 cells. Media was changed ~24 h post-transduction. Transduced cells were selected with puromycin (1ug/mL) for 5–10 days.

Murine KPC pancreatic cancer cell lines PK5L1940, PK5L1940-OVA and FC1242 were passaged in RPMI1640 supplemented with 10% FBS. Cells were passaged every 2-3 days at a density of 1:4-1:10. Cells were not allowed to grow beyond passage 20.

Mice

Female C57BL6 (6 weeks old) and Rag1−/− (B6.129S7-Rag1tm1Mom/J) (6 weeks old) were purchased from Jackson Laboratories (Indianapolis, In, USA). All mice were cared for in accordance with the ethical guidelines and conditions set and overseen by the University of Colorado, Anschutz Medical Campus Animal Care and Use Committee. Protocols used for animal studies were reviewed and approved by the Institutional Animal Care and Use committee at the University of Colorado, Anschutz Medical Campus.

OT-I transgenic mice (C57BL/6-Tg(TcraTcrb)1100Mjb/J) were mated with B6-SJL (Ptprca Pepcb/BoyJ) expressing CD45.1 allele, both from The Jackson Laboratory. CD45.1⁺ OT-I offspring were screened for maintenance by cytometric analysis of peripheral blood stained with Kb/SIINFEKL-PE tetramers and CD45.1/.2 specific antibodies.

Method details

Local and metastatic cancer cell implantations

Local orthotopic implantations were conducted by first anesthetizing mice using isoflurane and making a 1 cm incision in the left subcostal region. Mouse pancreata were located, externalized, and injected with 200,000 PK5L1940 or FC1242 KPC cells suspended 1:1 in Matrigel (Corning, Corning, NY). Pancreata were then reintroduced into abdomen and mice peritoneum and skin were closed. Protocol described in further detail⁶⁵. Survival and flow cytometric in vivo studies were conducted and analyzed separately.

Metastatic orthotopic implantations were conducted as above with spleen externalization following subcostal incision. Spleens were first ligated with horizon clips and 1 hemispleen was injected with 200,000 KPC cells suspended in 50μl 10% RPMI followed by washout injection of 50 μl PBS. Pancreatic vessels were then ligated with horizon clips and hemispleen was excised prior to closure of peritoneum and skin. Metastatic implantation described in further detail³². For cancer specific mortality, mice determined to have died from other causes were excluded from the analysis.

Tumor rechallenging was conducted by injecting either PK5L1940 cells at a final concentration of 5×10⁵ cells/0.1 ml or FC1242 cells at a final concentration of 1×10⁶ cells/0.1 ml into the right flank of each animal. Flank tumors were measured twice weekly with digital calipers and tumor volumes were estimated using the formula $(V=A\times B^2∕2{\text{mm}}^3)$, where A is the longer and B is the shorter diameter of the tumor.

For adoptive transfer experiments, CD8⁺ T cells were sorted with a CD8-negative selection kit (StemCellTechnologies cat#19853) per manufacturer guidance. Cells were then counted via TC20 Automated Cell Counter(Bio-rad). For pilot PK OVA study (Fig S5B-F), 6E6 viable CD8⁺ T cells were resuspended into 100uL PBS and injected via tail vein into restrained mice. For monotherapy PK-OVA adoptive transfer study (Figure 4) 0.5E6 viable cells CD8⁺ T cells were resuspended into 100uL PBS and injected via tail vein into restrained mice.

RNAseq for tumor samples

RNA sequencing was conducted as previously described⁶⁶. FNA biopsy and post-neoadjuvant surgical patient samples were obtained from the University of Colorado biorepository. All tumors were borderline resectable and treated with neoadjuvant chemotherapy with 30-33.6Gy SBRT, followed by pancreaticoduodenectomy. Tumor samples were reviewed by a pathologist to identify areas of tumor cellularity. Only the pathologist-marked areas of tumor cellularity were scraped and processed per BioSpyder kit instructions. For the human RNA sequencing library preparation, lysates were made from scrapes using BioSpyer FFPE lysis protocol, then diluted in 1:5 in 1X TempO-Seq Lysis Buffer and processed through a TempO-Seq Human Full Transcriptome FFPE Assay 96 sample Kit (BioSpyder, Carlsbad, CA). Reads were aligned and counts generated using Biospyder TempoSeqr platform. Genes with <1 mean raw counts or <1 mean counts per million (CPM) were removed from the data set. Differential expression was calculated using the limma R package⁶⁷. The resulting fold change was used with the fgsea R package to perform gene set enrichment analysis for Hallmark and C2 Curated gene sets⁶⁸.

Mass Cytometry (CyTOF) for tumor samples

Mass cytometry by time-of-flight (CyTOF) analysis was performed on human peripheral blood mononuclear cells (PBMCs). Cells were stained with the following heavy-metal tagged antibodies: CD19-142Nd, CD11c-162Dy, CD127-149Sm, CD16-209Bi, CD25-169Tm, CD27-155Gd, CD45-89Y, CD3e-154Sm, CD4-174Yb, CD8a-168Er, CD11b-153Eu, FoxP3-165Ho, CD15-164Dy, Intercalator Ir-191/Ir-193, and Cisplatin- Pt-195. Samples were run on the Helios Mass Cytometer at the University of Colorado Denver Cancer Center Flow Cytometry Core. Gating was performed on nucleated live cells. Data was analyzed using FlowJo Analysis software.

Murine CD4⁺ T cell isolation and in vitro activation

Spleens of C57BL/6 mice were homogenized to a single cells suspension by mashing the spleen through a 100 uM cells strainer and the erythrocytes were lysed with ACK (ammonium-chloride-potassium) lysis buffer for 5 min at 4°C. CD4⁺ T cells were sorted with a CD4-negative selection Miltenyi beads system following manufacturer instructions. CD4⁺ T cells were seeded into an αCD3/ αCD28 pre-coated plate (5 μg/ml, clone 145-201, BioLegend and 5 μg/ml, clone 37.51 BioLegend) and activated for 3 days.

IL-2R signaling (STAT-5P) in PD-1⁺ and PD-1-blocked CD4⁺ T cells

After 3 days of in vitro activation the cells were harvested and washed multiple times to remove endogenous IL-2. The cells were treated with increasing doses of muPD-1-IL-2v or muFAP-IL-2v for 30min at 37°C. Directly after treatment, the cells were fixed with Phosphoflow Fix Buffer I (BD) and incubated for 30 min at 37°C. The cells were then permeabilized overnight at −80°C with Phosphoflow PermBuffer III (BD) before being stained for 30min at 4°C with anti-STAT-5P-AF647 antibody (clone 47/pY694 (1:20), BD Biosciences).

In vivo drug administration

aCD25 (3mg/kg), PD1-IL2v (muPD1-IL2v)⁴⁹ (0.5mg/kg), and DP47-IL2v (muDP47-IL-2v) (0.5mg/kg) were administered intraperitoneally once per week beginning one week after tumor implantation. aPD-1 was administered intraperitoneally at a dose of 10mg/kg twice per week beginning one week after tumor implantation. aCD8 and aNK1.1 antibodies were dosed at 200ug twice per week for the duration of depletion experiments.

Flow cytometry

For flow cytometric analysis of tumor tissue, tumors were digested into single-cell suspension as previously reported⁶⁹. Briefly, tumors, were finely cut and incubated in HBSS solution with Collagenase III (Worthington) at 37°C. After incubation, tumors were passed through a 70 μm nylon mesh. The resulting cell suspension was centrifuged and re-suspended in red blood cell (RBC) lysis buffer for 5 minutes. RBC lysis buffer was deactivated, cell suspensions were centrifuged, re-suspended, and counted using an automated cell counter. Tumor draining inguinal lymph nodes and spleens were processed into single-cell suspensions as above. For flow cytometric analysis, cells were plated in 24-well plates and cultured for 4 hours in the presence of monensin, PMA, and ionomycin to stimulate cytokine production and block Golgi transport. Cells were then blocked with anti-CD16/32 antibody. A summary of the antibodies used to identify immune cell populations can be seen in Table 1. Where necessary, cells were fixed and permeabilized prior to staining using the FoxP3 Fixation/Permeabilization protocol (eBioscience). Samples were run on the Cytek Aurora Spectral Cytometer at the Barbara Davis Center at the University of Colorado Diabetes Research Center (NIDDK grant #P30-DK116073). Data were analyzed using FlowJo Analysis software. Populations were visualized using FlowSOM within Cytobank software.

Radiotherapy

Image-guided radiotherapy was performed using the X-Rad SmART small animal irradiator (Precision X-Ray, North Bradford CT) at 225kVp, 20mA with 0.3 mm Cu filter. Mice were positioned in the prone orientation and a CT scan was acquired. Radiation was delivered at a dose rate of 5.6Gy/min. A single 8 Gy dose of X-ray radiation was delivered to mouse pancreata using 10mm square beam with field edges at mouse midline and below left ribs. Monte-Carlo simulation was performed using SmART-ATP software (SmART Scientific Solutions, Maastricht, Netherlands) with a CBCT scan of one mouse to determine the appropriate time and current. All mice received identical treatment after repositioning by fluoroscopy. For all in vivo experiments, radiation was given at 7 days post-implantation.

Metabolomics

High-throughput metabolomics analysis was performed at the University of Colorado Cancer Center Mass Spectrometry Shared Resource on murine serum samples and CD8⁺ T cells upon ex vivo incubation with 25 mM U-¹³C₆-glucose (Sigma Aldrich) for 6h in glucose-free RPMI. Samples were thawed on ice and metabolites extracted from serum and CD8⁺ T cells by adding chilled 5:3:2 methanol:acetonitrile:water (v/v/v) to each tube at 1:24 ratio or 1 million cells/ml of solution equivalent, followed by 30 minutes of vortexing and 10 minutes of centrifugation, both at 4°C.⁷⁰ All extracts were analyzed twice (10 μL injections each) by ultra-high-performance liquid chromatography using a Thermo Vanquish UHPLC coupled to a Thermo Q Exactive mass spectrometer in negative and positive polarity modes, as described ^70,71. For each method, the UHPLC utilized a Phenomenex C18 column at a flow rate of 0.45 mL/min with a 5-minute gradient (detailed in ^70,71). Data analysis and peak picking (including ¹³C-labeled isotopologues) was performed via Maven (1.4.20-dev-772). MetaboAnalyst (5.0) was used to perform multivariate analyses including Partial Least Square-Discriminant Analysis (PLS-DA), and hierarchical clustering analyses with heatmap representation of the top 50 metabolites by ANOVA⁷². Graphs were rendered using Graphpad Prism (9.3.1).

Proteomics

NKp46⁺, CD8⁺, CD8⁺ CD25+, and CD8⁺CD25⁻ cell populations were flow sorted via MoFlo XDPFlex70. Populations were digested according to the FASP protocol using a 10 kDa molecular weight cutoff filter. The analyses were performed on the timsTOF SCP instrument (Bruker Daltonics, Bremen) coupled to an Evosep One system (Evosep Biosciences) and used a ~58-minute gradient Whiper 20SPD method that offers a constant flow of 100 nL/min. Raw data files conversion to peak lists in the MGF format, downstream identification, validation, filtering, and quantification were managed using FragPipe version 13.0.

NK and CD8⁺ T cell Analysis

Spectral counts from detected proteins were taken, and associated gene symbols for these proteins were used for downstream analysis. Proteins with low expression were removed, keeping proteins detected in all samples. Hierarchical clustering analysis of the top 50 proteins by ANOVA was performed. Network analyses were generated via OmicsNet 2.0 based on significant changes assessed via proteomics. Pathway enrichment analyses were performed to determine the significantly enriched pathways based upon proteomics changes via the David functional annotation tool⁶¹, against the murine KEGG pathway database, and validated against the better annotated human KEGG pathway database.

CD25⁺ vs. CD25⁻ CD8⁺ T cell Analysis

Spectral counts from detected proteins were taken, and associated gene symbols for these proteins were used for downstream analysis. Proteins with low expression were removed, keeping proteins detected in all samples. Six gene symbols that were mapped to by 2 different protein isoforms each were removed for purposes of pathway analysis (Aak1, Cnbp, Psmd4, Samhd1, Sh3kbp1, Tpm3). The resulting 1266 genes were input into Metaboanalyst 5.0 (https://www.metaboanalyst.ca/) for analysis with the Statistical Analysis [one factor] module with settings: no filtering, sample normalization - normalization by sum, data scaling - auto scaling. In Metaboanalyst, unpaired t-tests and associated adjusted p-values were calculated, and log2-transformed fold changes were calculated.

Upregulated genes with log2fc > 0 and p.adjusted values < 0.05 were assessed by over-representation analysis with R (v4.2.1) using R package clusterProfiler (v4.4.4) using the enrichGO function for GO biological pathways in size range 10–500. Treeplots were generated with R package enrichplot (v1.16.2) with setting nCluster = 5, and clusters were manually labeled according to their GO pathway members.

Cytotoxicity assay

Cytotoxicity assay was performed by pre-staining PK5L1940 pancreatic cancer cell lines using a 2mg/ml calcein solution and isolating CD8⁺ T cells according to the manufacturer’s protocol (Stemcell, #19853A). Cancer cells were incubated in calcein-containing media for 30 minutes at 37C and subsequently plated in a 96 well plate at a 2:1 CD8⁺ T cell to cancer cell ratio. The reaction mixture was then allowed to incubate at 37C for 4 hours. Following the incubation period, supernatant was collected and cancer cell release of calcein was quantified by fluorescence (Ex: 485nm/EM: 530 nm) Protocol described in further detail⁷³.

Statistical significance

Quantitative analyses were performed using a two-sided Student’s t-test, Mann-Whitney test, One-Way ANOVA with multiple comparisons, or the Mantel-Cox test for survival using GraphPad Prism, p-values of <0.05 were considered statistically significant.

Key Resources Table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rat Anti-mouse CD3-BUV805 (17A2)	BD Biosciences	Cat#741982
Rat Anti-mouse CD4-BUV496 (GK1.5)	BD Biosciences	Cat#612952
Rat Anti-mouse CD8-BB515(53-6.7)	BD Biosciences	Cat#564422
Rat Anti-mouse CD11b-BUV661 (M1/70)	BD Biosciences	Cat# 612977
Hamster Anti-mouse CD11c-PE/Cyanine5 (N418)	BioLegend	Cat# 117316
Rat Anti-mouse CD25-BV786 (3C7)	BD Biosciences	Cat# 564368
Rat Anti-mouse/human CD44-BV570 (IM7)	BioLegend	Cat#103037
Anti-mouse CD45.1-BV421(A20)	BioLegend	Cat#110731
Anti-mouse CD45.2-PerCP(104)	BioLegend	Cat#109826
Anti-mouse CD45-PerCP(30-F11)	BioLegend	Cat#103130
Anti-mouse CD69-SuperBright436(H1.2F3)	eBioscience	Cat# 62-0691-82
Hamster Anti-mouse CD103-AF700(2E7)	BioLegend	Cat#121442
Anti-mouse CD122-PerCP-eFluor 710 (TU27)	eBioscience	Cat# 46-1228-42
Anti-mouse CD218a-eFluor 450 (P3TUNYA)	eBioscience	Cat# 48-5183-82
Hamster Anti-mouse CD183(CXCR3)-BUV615 (CXCR3-173)	BD Biosciences	Cat#751457
Rat Anti-mouse CD226(DNAM-1)-BV605 (TX42.1)	BioLegend	Cat#133613
Anti-mouse CD326(EpCAM)-APC (G8.8)	BioLegend	Cat#118214
Anti-mouse FoxP3-Alexa Fluor 532(FJK-16s)	eBioscience	Cat# 58-5773-82
Anti-human/mouse Granzyme B-FITC (QA16A02)	BioLegend	Cat#372206
Anti-mouse Ly-6G/Ly-6C(GR-1)-APCFire810 (RB6-8C5)	Biolegend	Cat#108470
Rat anti-mouse IFNg-BUV737 (XMG1.2)	BD Biosciences	Cat# 612769
Rat Anti-mouse IL10-BV711 (JES5-16E3)	BD Biosciences	Cat# 564081
Anti-mouse IL2-APC(JES6-5H4)	eBioscience	Cat# 17-7021-82
Anti-mouse Ki-67-APC-eFluor 780(SolA15)	eBioscience	Cat# 47-5698-82
Rat Anti-mouse I-A/I-E(MHCII)-PEDazz594 (M5/114.15.2)	BioLegend	Cat# 107648
Rat Anti-mouse CD335 (NKp46)-PECy7 (29A1.4)	BioLegend	Cat#137618
Rat Anti-mouse CD274-BV650 (MIH5)	BD Biosciences	Cat#740614
Hamster Anti-mouse CD279(PD-1)-BUV395(J43)	BD Biosciences	Cat#744549
Anti-mouse anti-TCF-7/TCF-1-PE(S33-966)	BD Biosciences	Cat#564217
Anti-mouse CD366(Tim-3)-Alexa Fluor 647(B8.2C12)	Biolegend	Cat# 134006
Rat Anti-mouse TNF-BV750(MP6-XT22)	BD Biosciences	Cat#566365
Anti-mouse Tbet-BV421 (4B10)	BioLegend	Cat#644816
Anti-mouse NK1.1(PK136)	BioXCell	Cat#BE0036
Anti-mouse CD8a(53-6.7)	BioXCell	Cat#BE0004-1
Biological samples
Human PDAC-- tumor material	University of Colorado Pancreas and Biliary Cancer Multidisciplinary Clinic	Collected per COMIRB13–0315.
Human PDAC – PBMC material	University of Colorado Pancreas and Biliary Cancer Multidisciplinary Clinic	Clinicaltrials.gov: NCT02873598
Chemicals, peptides, and recombinant proteins
Fetal Bovine Serum	Gibco	Cat# 10437028
RPMI 160 Medium	Gibco	Cat#11875-093
RPMI 1640 Medium, no Glucose	Gibco	Cat#11879-020
D-Glucose-13C6	Sigma Aldrich	SKU#389374
eBioscience Cell Stimulation Cocktail Plus Protein Transport Inhibitors	Invitrogen	Ref#00-4975-93
Anti-Mouse CD16/CD31 (Fc Shield)	TONBO biosciences	Cat#70-0161-M001
pLVX-puro-cOVA-IRES-BFP plasmid	Addgene	Plasmid #135074
psPAX2 plasmid	Addgene	Plasmid #12260
pMD2.G plasmid	Addgene	Plasmid #12259
Brilliant Stain Buffer	BD Biosciences	Cat#563794
Trypsin Inhibitor, Soybean Purified, AOF	Worthington Biochemical Corporation	Cat#LS003571
Collagenase Type III	Worthington Biochemical Corporation	Cat#LS004182
eBioscience 1X RBC Lysis Buffer	Invitrogen	Raf#00-4333-57
eBioscience FOXP3/Transcription Factor Fixation/Permeabilization	Invitrogen	Ref#00-5521-00
TempO-Seq Lysis Buffer	BioSpyder	58
Critical commercial assays
Live/Dead Fixable Aqua Dead Cell Stain Kit	Invitrogen	Ref#L34966
EasySep Mouse CD8+ T Cell Isolation Kit	StemCell	Cat#19853A
TempO-Seq FFPE Assay 96 sample kit	BioSpyder	58
Calcein, AM	Invitrogen	Ref#C3099
Deposited Data
Mouse serum metabolomics	This paper	ST002485
Mouse serum proteomics	This paper	PXD040266
Metabolomics of CD8 T cells isolated from blood of treated mice	This paper	ST002486
Metabolomics of CD8 T cells isolated from spleen and lymph node of treated mice	This paper	ST002487
Metabolomics of Carbon-13 tracing, cell culture supernatants	This paper	ST002488
Metabolomics of Carbon-13 tracing, CD8 T cells	This paper	ST002489
Proteomics of CD8 T cells isolated from treated mice	This paper	PXD040266
Metabolomics of PDAC patient plasma	This paper	ST002490
RNA sequencing of surgically resected PDACs	This paper	GEO: GSE225767
CyTOF of surgically resected PDACs	This paper	Data available upon request
Experimental models: Cell lines
PK5L1940		Kindly gifted by Dr. Michael Gough
PK5L1940-OVA	This paper	N/A
FC1242		Kindly gifted by Dr. David Tuveson
Experimental models: Organisms/strains
C57BL/6J(JAX)	The Jackson Laboratory	Strain #:000664
B6.129S7-Rag1tm1Mom/J (JAX)	The Jackson Laboratory	Strain#002216
OT-1 CD45.1		In this study
Software and algorithms
FlowJo 10.8.1	BD Biosciences	N/A
GraphPad Prism 9.5.0	GraphPad Software Inc.	N/A
Cytobank	N/A	59
MetaboAnalyst5.0	N/A	N/A
FragPipe 13.0	Nesvilab	N/A
OmicsNet 2.0	N/A	60
DAVID Gene Functional Classification Tool	N/A	61
R version 4.2.1	R Foundation for Statistical Computing	N/A
Biorender	Biorender.com	N/A

Highlights.

• RT + PD1-IL2v reduces tumor growth, metastasis by enhancing CD8 T cell function.

• RT + PD1-IL2v reduces tumoral Tregs while increasing systemic NK cell frequency.

• RT + PD1-IL2v induces a durable memory response, evidenced by tumor eradication.

• Human PDAC patients upregulate the molecular targets of PD1-IL2v post-RT.

INCLUSION AND DIVERSITY

We support inclusive, diverse, and equitable conduct of research.