CD4+ T cells facilitate the RT-induced abscopal effect by promoting antigen cross-presentation to CD8+ T cells at unirradiated tumor sites

Abstract

Background

The local effect of radiotherapy (RT) is enhanced by CD8⁺ T-cell responses elicited through dendritic cell (DC)-mediated cross-presentation of tumor antigens, facilitated by RT-induced damage-associated molecular patterns. The abscopal effect—regression of non-irradiated tumors—has been observed clinically, particularly in combination with immune checkpoint blockade, although it remains uncommon. To better understand how to enhance this effect, we investigated two RT/α-programmed death 1 (PD-1)-based triple combinations incorporating either α-cytotoxic T lymphocyte-associated protein 4 (CTLA-4) or CD122-targeted interleukin (IL)-2 complexes (IL-2c).

Methods

We tested these regimens in B16 melanoma and C51 colon carcinoma models in mice with one irradiated and one non-irradiated tumor on opposite flanks.

Results

In both models, RT/αPD-1/αCTLA-4 elicited a stronger abscopal response than RT/αPD-1/IL-2c. In the C51 model, RT/αPD-1/αCTLA-4 achieved a 61.5% abscopal cure rate, dependent on both CD8⁺ and CD4⁺ T cells. In contrast, the less effective RT/αPD-1/IL-2c response required only CD8⁺ T cells. The enhanced abscopal effect with RT/αPD-1/αCTLA-4 was associated with increased numbers, effector function, and reduced exhaustion of tumor-specific CD8⁺ tumor-infiltrating lymphocytes (TILs) and of CD4⁺ TILs, along with elevated CD80⁺CD86⁺ DCs in abscopal tumors, as shown by flow cytometry; immunofluorescence confirmed increased T-cell infiltration. CD4⁺ T-cell depletion during RT/αPD-1/αCTLA-4 treatment impaired abscopal but not irradiated tumor control, reducing infiltration of tumor-specific CD8⁺ T cells and conventional (c) DC1s, and diminishing cDC1-mediated cross-presentation in abscopal tumors. Activated CD4⁺ T cells upregulated CD80/CD86 on cDC1s and enhanced cross-presentation, partly via interferon-γ and tumor necrosis factor. Adoptively transferred tumor-specific CD8⁺ T cells from tumor-irradiated donors localized to unirradiated tumors and draining lymph nodes in αPD-1/αCTLA-4-treated recipients, but not in untreated or CD4⁺ T cell-depleted mice.

Conclusions

These results demonstrate that an RT-based combination therapy that robustly induces CD4⁺ T cells alongside CD8⁺ T cells can elicit a strong abscopal response and suggest that CD4⁺ effector T cells act at abscopal sites by promoting DC-mediated cross-presentation of tumor antigens to CD8⁺ T cells originating from the irradiated tumor.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Radiotherapy (RT) can induce systemic, T cell-mediated antitumor responses known as the abscopal effect, but in most clinical settings, this effect is currently insufficient to effectively control distant, non-irradiated tumors.

WHAT THIS STUDY ADDS

• In murine tumor models, an RT-based combination therapy induced a robust, CD4⁺ T cell-dependent abscopal effect that was associated with strong induction of effector-like CD4⁺ T cells. Our data suggest that, at non-irradiated tumor sites, effector-like CD4⁺ T cells enhance dendritic cell (DC)-mediated cross-presentation of tumor antigens to CD8⁺ T cells, including RT-primed CD8⁺ T cells, in part via interferon-γ and tumor necrosis factor. The results support a model in which CD4⁺ T cell-activated cross-presenting DCs restimulate RT-induced CD8⁺ T cells at abscopal sites, with CD4⁺ effector T cells in non-irradiated tumors acting similarly to RT-induced damage-associated molecular patterns by promoting DC-mediated activation of CD8⁺ T cells.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The study highlights that future preclinical and clinical efforts to enhance the abscopal effect should focus on RT combinations that induce effector CD4⁺ T cells alongside CD8⁺ tumor-specific T cells and/or promote DC-mediated tumor antigen cross-presentation at unirradiated (abscopal) sites.

Background

Radiotherapy (RT) not only kills tumor cells directly but can also induce major histocompatibility complex (MHC) class I-restricted, CD8⁺ tumor-specific cytotoxic T cells.^{1 2} The latter requires immunogenic cell death (ICD) of tumor cells, which activates cross-presenting dendritic cells (DCs) capable of directing endocytosed or phagocytosed tumor antigens into the MHC class I antigen processing pathway.³ Tumor antigen cross-presentation after RT depends on damage-associated molecular patterns (DAMPs) generated during ICD.⁴ RT-induced DAMPs—including nuclear and mitochondrial double-stranded DNA, interferon (IFN)-β, ATP, high mobility group box 1 protein (HMGB1), calreticulin, and others—serve as danger signals that recruit and activate cross-presenting DCs. During DAMP-induced activation of DCs, co-stimulatory molecules such as CD80 and CD86 are upregulated.³ These bind CD28, expressed on naïve and non-terminally differentiated T cells, and cytotoxic T lymphocyte-associated protein 4 (CTLA-4) on activated T cells and immunosuppressive regulatory T cells (Tregs).⁵ IFN-β-induced CXCL10 binds CXCR3 to recruit tumor-specific CD8⁺ T cells to irradiated tumors. The immunogenicity of RT is shaped by dose and fractionation.⁶ ATP, HMGB1, and calreticulin are dose-dependently induced across single doses from 2 to 20 Gy.⁷ IFN-β is triggered via the cGAS–STING pathway following release of nuclear or mitochondrial DNA. Hypofractionated RT (hRT), using 6–12 Gy per fraction, is particularly effective at inducing IFN-β.^{6 8} RT-induced CD8⁺ tumor-infiltrating lymphocytes (TILs) express programmed death 1 (PD-1) and secrete IFN-γ, which upregulates programmed death-ligand 1 (PD-L1) on tumor cells and DCs, providing a rationale for combining RT with anti-PD-1/PD-L1 immune checkpoint blockade (ICB).

Preclinical studies show that hRT-induced tumor-specific CD8⁺ T cells improve local tumor control and can mediate regression of non-irradiated tumors—the so-called abscopal effect—particularly in immunogenic models and when hRT is combined with ICB. Although observed in case reports, most clinical trials have yielded negative results.^{1 2} This may reflect the inclusion of ICB-resistant, polymetastatic patients in whom only one or a few lesions were irradiated. Another likely reason is the limited accumulation of RT-induced or RT/ICB-induced CD8⁺ T cells in non-irradiated tumors, especially in cold or immunosuppressed tumors lacking activated DCs.⁹ Properly activated DCs are now recognized as essential not only for T-cell priming but also for sustaining T-cell effector function.^{10 11} Thus, the absence of adequately activated, cross-presenting DCs in non-irradiated tumors may be a key reason for an ineffective abscopal response.

We investigated the abscopal effect using two triple combinations: hRT with anti-PD-1 (αPD-1) plus either anti-CTLA-4 (αCTLA-4) or CD122-directed interleukin (IL)-2 complexes (IL-2c). While both αCTLA-4 and CD122-targeted IL-2 derivatives have gained attention,^{12 13} clinical development of the latter has faced setbacks following negative phase 3 results in combination with αPD-1.13,16 We recently reported that adding CD122-directed IL-2c¹⁷—known to massively expand CD8⁺ T cells¹³—to hRT/αPD-1 yielded an abscopal effect, but with predominantly peripheral (extratumoral) CD8⁺ T-cell expansion. To identify a more effective triple combination on the hRT/αPD-1 backbone, we here compared hRT/αPD-1/IL-2c with a regimen incorporating αCTLA-4. In two murine models, hRT/αPD-1/αCTLA-4 consistently induced a significantly stronger abscopal effect than hRT/αPD-1/IL-2c. Our data show that CD4⁺ T cells are critical for this enhanced abscopal effect and suggest that, at abscopal sites, hRT/αPD-1/αCTLA-4-induced effector-like CD4⁺ T cells promote DC-mediated cross-presentation of tumor antigens to CD8⁺ T cells, including CD8⁺ T cells originating from the irradiated tumor. The results suggest that restimulation of RT-induced CD8⁺ T cells by cross-presenting DCs at non-irradiated tumor sites enhances the abscopal effect and that CD4⁺ effector T cells at these sites exert functions similar to those of RT-induced DAMPs in irradiated tumors by promoting DC-mediated CD8⁺ T-cell activation.

Materials and methods

Mice and cell lines

BALB/c and C57BL/6 mice were purchased from Janvier Labs. CD45.1 C57BL/6 mice were kindly provided by Dr Robert Zeiser, Freiburg. Mice were used at the age of 8–12 weeks. Animals were housed under specific pathogen-free conditions with 12 hours light/dark cycle, 21–25°C, 45–65% humidity. A maximum of five mice per cage were housed with unlimited access to food and water. The protocols and procedures were approved by the animal care committee of the Regierungspräsidium (Federal Ministry for Nature, Environment and Consumers’ Protection) Freiburg, Germany (registration numbers: G-21/106, G-20/016, G-25/030). The C51 colon carcinoma cell line was kindly provided by Dr Mario Paolo Colombo (Milan) and cultured in complete Dulbecco’s Modified Eagle Medium (DMEM) medium (Gibco). B16-CD133 melanoma cells were generated and cultured in Roswell Park Memorial Institute (RPMI) 1640 medium according to previously published protocols.¹⁸ B16-OVA cells, obtained from Dr Vincenzo Cerullo (Finland), were cultured in RPMI-1640 supplemented with 500 µg/mL geneticin for selection. All in vitro T-cell functional assays were conducted in complete RPMI-1640 medium.

Mouse tumor models and treatments

5×10⁵ C51 cells were resuspended in a 1:1 mixture of phosphate-buffered saline (PBS) and Matrigel (Corning) and injected subcutaneously into the right flank (primary tumor) of mice. The same number of cells were injected into the left flank 5 days later (secondary tumor). For the B16-CD133 and B16-OVA melanoma tumor models, 2×10⁵ cells were injected following the same approach, with a 3–4 days interval between primary and secondary tumor. Mice with tumors in a predefined size range were randomly assigned to experimental groups. Tumor irradiation was conducted using the RS2000 X-ray system (Radsource). Anesthetized mice were placed in a custom-designed plastic jig featuring an adjustable opening that allowed the tumor to be positioned within the irradiation field, while the remainder of the body was shielded with lead. To ensure uniform dose distribution, mice were rotated 180° for a second exposure. Primary tumors were irradiated locally with two fractions of 8 Gy (C51 model) or 12 Gy (B16-CD133 and B16-OVA model) on consecutive days when the primary and secondary tumors reached a volume of 175–285 mm³ and 40–75 mm³, respectively. Weekly intraperitoneal (i.p) injections of αPD-1 (200 µg; RMP1-14, Bio X Cell) and αCTLA-4 antibodies (200 µg; 9D9, Bio X Cell) were performed starting on the first day of RT. Mice received i.p injections of 9 µg IL-2c three times per week, starting on day 3 after the initiation of RT. IL-2c were prepared by combining recombinant mouse IL-2 (ImmunoTools) with the S4B6 monoclonal antibody purified from culture supernatants of the S4B6-1 hybridoma (ATCC clone HB-10968, RRID:CVCL_9236) at a molar ratio of 2:1.¹⁷ In depletion experiments, CD8⁺ T cells or CD4⁺ T cells were depleted by injecting 200 µg/mouse of anti-CD8 (clone 2.43, Bio X Cell) or anti-CD4 (clone GK1.5, hybridoma) antibodies intraperitoneally 1–2 days before hRT, on the day of the first RT fraction, and once weekly thereafter. To block costimulatory signals on DCs, anti-CD70 (FR70), anti-CD80 (16–10 A1), and anti-CD86 (GL-1) antibodies (Bio X Cell) were administered intraperitoneally at 200 µg/mouse every 3 days, starting 1 day before hRT. Tumor growth was calculated as length×width×height×π/6 measured by caliper. Survival was defined as the time point after treatment start when either the primary or the secondary tumor had reached a size of 1,000 mm³.

Determination of absolute cell numbers

After preparing single-cell suspensions, viable leukocytes were counted using trypan blue staining, based on their size and morphology. The frequency of each subset within CD45⁺ cells was analyzed by flow cytometry. Absolute numbers were obtained by multiplying the percentage of each subset among CD45⁺ cells by the total number of viable leukocytes. For tumor samples, absolute counts were normalized to tumor weight (cells per gram of tissue); for blood, as cells per milliliter; and for spleen, as total number of cells.

Flow cytometry analysis

Flow-cytometric analyses of TILs and T cells from peripheral blood, spleen, and lymph nodes were performed using the following anti-mouse antibodies/MHC tetramers: AH1 tetramer-PE (H-2L^d, gp70) (from Baylor College of Medicine), M8 tetramer-PE (H-2K^b, MuLV p15E, KSPWFTTL) was used to detect tumor-specific CD8⁺ T cells in C51 and B16-CD133 tumor models, respectively; MHC-II-AF700 (clone M5/114.15.2), CD45-BV510 or CD45-BV650 (clone 30-F11), CD4-BV510 or CD4-PerCP (clone RM4-4), CD3-PE-Cy5.5, CD3-FITC or CD3-APC-eFluor 780 (clone 145–2 C11), CD49b-FITC (clone DX5), CD19-FITC (clone 1D3), CD8-AF700 (clone 53–6.7), Ki-67-BV605 (clone 16A8), CD11c-BV650 (clone N418), CD11b-PE-Cy7 (clone M1/70), CD69-FITC (clone H1.2F3), CD86-BV421 (clone GL1), CD80-APC (clone 16–10 A1), CD103-PE-CF594 (clone M290), Ly6C-PerCP-Cy5.5 (clone HK1.4), TNF-BV421 (clone MP6-XT22), IL-2-BV510 (clone JES6-5H4), PD-1-FITC (clone 29F1.A12), CCL5-PE-Cy7 (clone RANTES) F4/80-FITC (clone BM8), T-bet-PE-Cy7 or T-bet-Pacific Blue (clone 4B10), GATA3-PerCP-Cy5.5 (clone 11B11), CXCR5-APC-Cy7 (clone L138D7), BCL6-PE (clone 7D1) and CD278 (ICOS)-PE-Cy5 (clone 15F9) were from BioLegend. Other antibodies used were: CD101-Pe-Cy7 (clone Moushi101), granzyme B PE-Cy7 (clone NGZB) and IFN-γ-FITC (clone XMG1.2) from Invitrogen; Tim-3-APC (clone REA602), CCL3-PE or APC (clone REA355) and H-2k^b/SIINFEKL-PE (25-D1.16) from Miltenyi; TCF-1 (clone C63D9) from Cell Signaling; F4/80-APC (clone BM8), Foxp3-PE (clone FJK-16s) and CD25-PerCP-Cy5.5 (clone PC61) from eBioscience; CXCL10-AF546 (clone E-2) from Santa Cruz; RORγt-BV421 (clone Q31-378) from BD Biosciences. Isotype control antibodies used were: mouse IgG2b-PE (clone MPC-11), Rat IgG1κ-BV421 (clone RTK2071) and Rat IgG1κ FITC (clone RTK2071) from BioLegend, mouse IgG2bκ-PE-Cy7 (clone eBMG2b) from eBioscience. 7-AAD (eBioscience), Zombie Red or Zombie NIR (BioLegend) were used to exclude dead cells. When needed, cells were fixed with the Foxp3/Transcription Factor Staining Buffer Set (eBioscience) or Intracellular Staining Kit (eBioscience) according to the manufacturer’s instructions. Data were acquired using a CytoFLEX S Flow Cytometer (Beckman Coulter) with CytExpert V.2.4.0.28 software followed by data analysis with FlowJo V.10.4 software.

Ex vivo T-cell re-stimulation

To evaluate CD4⁺ and CD8⁺ T-cell effector functions, tumor single-cell suspensions (0.3×10⁶) were stimulated with phorbol 12-myristate 13-acetate (PMA) (50 ng/mL) plus ionomycin (1 µg/mL) plus Brefeldin A (1:1000, BioLegend) for 4 hours at 37°C, 5% CO₂ in RPMI medium. Thereafter, intracellular cytokine staining was performed followed by flow cytometry analysis.

DC and CD4⁺ T-cell co-culture assay

CD4⁺ T cells were isolated from the spleen of naïve mice with the CD4⁺ T Cell Isolation Kit (Miltenyi Biotec). Fms-like tyrosine kinase 3 ligand–transduced B16 melanoma (B16-Flt3L) cells were provided by Dr Max Schnurr (LMU Munich, Germany). A total of 10⁷ B16-Flt3L cells were injected subcutaneously into C57BL/6N mice. After 9–11 days, mice were sacrificed to harvest the spleens. Splenic DCs were then isolated via CD11c microbeads (Miltenyi Biotec). 96-well flat-bottom plates were coated with anti-CD3 (1 µg/mL) and anti-CD28 (2 µg/mL) antibodies in PBS (200 µL) and isolated CD4⁺ T cells in RPMI-1640 complete medium with 10 ng/mL mIL-12 (from RD system) /mIL-2 (from ImmunoTools) were seeded at 1×10⁵ cells per well in 100 µL. Cells were incubated at 37°C with 5% CO_₂ for 48 hours and analyzed using flow cytometry to confirm activation by CD25 and CD69 markers. Isolated DCs were co-cultured with either naïve or activated CD4⁺ T cells at a 1:1 ratio for 2 hours in RPMI complete medium, followed by the addition of OVA protein (100 µg/mL) or SIINFEKL peptide (100 µg/mL), and lipopolysaccharide (LPS) (50 ng/mL). For antibody blockade experiments, tumor necrosis factor (TNF) blocking and IFN-γ neutralizing antibodies (Selleckchem) (10 µg/mL) were added 30 min later. After 12–16 hours, cells were collected and analyzed by flow cytometry.

CD4⁺ T-cell recognition of MHC-II-presented antigen

CD4⁺ T cells were isolated from secondary tumors of hRT/αPD-1/αCTLA-4-treated B16-OVA mice on day 7 post-treatment. DCs were enriched from spleens of B16-FLT3L tumor-bearing mice using CD11c MicroBeads (Miltenyi Biotec) and stimulated with LPS (100 ng/mL) plus OVA₃₂₃–₃₃₉ peptide (5 µg/mL) in complete RPMI medium for 4 hours at 37°C. After washing, CD4⁺ T cells were added at a 3:1 ratio and co-cultured for 12 hours. CD69 expression was analyzed by flow cytometry. An MHC-II blocking antibody (M5/114, 100 µg/mL) was included to inhibit CD4⁺ T-cell activation.

Ex vivo co-culture of DCs with OT-I CD8⁺ T cells

CD11c⁺MHC-II⁺ DCs were sorted from B16-OVA tumors of hRT/αPD-1/αCTLA-4-treated mice on day 7 post-treatment and pulsed with LPS (50 ng/mL) plus OVA protein (100 µg/mL) in complete RPMI medium for 24 hours. OT-I CD8⁺ T cells were isolated from spleen and lymph nodes of OT-I mice. DCs and OT-I cells were co-cultured at a 1:10 ratio for 72 hours in the presence of low-dose IL-2 (10 U/mL).

Adoptive cell transfer (ACT) experiments

Donor mice were treated with hRT/αPD-1/αCTLA-4 once tumor volume reached 175–285 mm³. Recipient mice received αPD-1 plus αCTLA-4 or αPD-1 plus αCTLA-4 in combination with CD4⁺ cell depletion when tumors reached 40–75 mm³. CD8⁺ T cells were isolated from single-cell suspensions of blood and tumor-draining lymph nodes (TDLNs) of donor mice using a CD8⁺ T Cell Selection Kit (Miltenyi Biotec) on day 7 after treatment initiation. In the C51 tumor model, isolated CD8⁺ T cells from donor mice were labeled with CellTrace Far Red (Invitrogen). A total of 1.5× 10⁶ CD8⁺ T cells in 100 µL PBS were adoptively transferred into recipient mice via intravenous injection. Transferred T cells were analyzed in tumors and lymph nodes of recipient mice by flow cytometry 2–3 days after the transfer.

Immunofluorescence

Formalin-fixed, paraffin-embedded tumor sections were stained using Opal 6-Plex Manual Detection Kit (Akoya Biosciences). The following primary antibodies from Cell Signaling Technology were used: rabbit anti-CD3 (Cell Signaling, 1:200), anti-CD4 (Cell Signaling, 1:200) and anti-CD11c (Cell Signaling, 1:100), anti-CD8 (Abcam, 1:1000). Immunofluorescence images were acquired and analyzed using QuPath software (V.0.5.1).

Statistical analysis

Previous experience served as the basis for calculations of expected averages and deviations used to calculate sample size to detect an effect with a power of 0.8. This typically resulted in a minimum sample size of n=5 per group, depending on the experiment. Data distribution was assessed using the Shapiro-Wilk test. For comparisons between two groups, the unpaired two-tailed Student’s t-test was used for parametric data, and the Mann-Whitney U test for non-parametric data. For parametric data involving more than two groups, one-way analysis of variance (ANOVA) was performed, followed by Dunnett’s multiple comparison test for comparisons to a control group, or Tukey’s test for all pairwise comparisons. In the experiment involving six groups, the two-stage linear step-up procedure of Benjamini-Krieger-Yekutieli was applied to control the false discovery rate following one-way ANOVA. For non-parametric data, the Kruskal-Wallis test was used to compare medians across groups, followed by Dunn’s test for multiple comparison correction. Survival data were analyzed using the log-rank test. P value<0.05 (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001) was considered statistically significant. Statistical analyses were performed using Prism V.9.0 (GraphPad).

Results

Superior antitumor efficacy of triple treatment with hRT/αPD-1/αCTLA-4 over hRT/αPD-1/IL-2c

We evaluated tumor control and survival in two dual-tumor mouse models in which hRT was confined to the primary tumor: the C51 colon carcinoma model and the B16-CD133 melanoma model. C51 represents an immunologically “cold” tumor with minimal baseline T-cell infiltration (online supplemental figure 1A,B). The B16-CD133 melanoma model is highly aggressive; the B16 line transduced to express the human cancer stem cell antigen CD133 shows increased immunogenicity compared with parental B16 wild-type tumors.¹⁸ We first examined the C51 model (figure 1A). Triple therapy with hRT/αPD-1/αCTLA-4 outperformed not only hRT/αPD-1/IL-2c but also dual combinations including hRT/αPD-1, hRT/αCTLA-4, hRT/IL-2c, and αPD-1/αCTLA-4, as well as the monotherapy controls (figure 1B–D; online supplemental figure 2A,B). Long-term tumor control and survival were achieved in 61.5% (8/13) of hRT/αPD-1/αCTLA-4-treated mice (figure 1C,D). Notably, the high complete remission rate included the unirradiated tumor, which is remarkable given the “cold” immune microenvironment of C51 (online supplemental figure 1A,B). Moreover, all mice achieving complete remission with hRT/αPD-1/αCTLA-4 (figure 1C) exhibited long-term tumor control (online supplemental figure 2C), indicating a durable cure. In contrast, the only mouse whose non-irradiated tumor initially responded strongly to hRT/αPD-1/IL-2c (figure 1C) relapsed within several weeks (online supplemental figure 2C). In mice treated with dual ICB (αPD-1/αCTLA-4), rapid primary tumor growth limited the observation period for the secondary tumor. To address this, we treated mice bearing a single tumor with αPD-1/αCTLA-4; at treatment start, these tumors were similar in size (40–75 mm³) to the secondary tumor in the abscopal model (see Materials and methods). Tumor control was weaker than that of the unirradiated tumor in hRT/αPD-1/αCTLA-4-treated mice in the two-tumor setting (figure 1E). These findings confirm that the strong response of the unirradiated tumor to hRT/αPD-1/αCTLA-4 in the two-tumor model was, to a considerable extent, due to an abscopal effect dependent on primary tumor irradiation (see figure 1C and E). A stronger abscopal effect and improved survival with hRT/αPD-1/αCTLA-4 versus hRT/αPD-1/αL-2c were also observed in the aggressive B16-CD133 melanoma model (online supplemental figure 3).

Figure 1. Enhanced abscopal tumor control and survival with hRT/αPD-1/αCTLA-4 versus hRT/αPD-1/IL-2c in the C51 colon cancer model. (A) Scheme for treatments. (B–C) Mean (B) and individual (C) tumor growth curves of primary (irradiated) and secondary (non-irradiated) C51 tumors. (D) Survival of mice. (E) Metastasis-mimicking experiment: mice bearing a single tumor of similar size (40–75 mm³) to the non-irradiated tumor in the two-tumor model received αPD-1/αCTLA-4 treatment (one-tumor model). Responses of non-irradiated tumors from hRT/αPD-1/αCTLA-4-treated mice (two-tumor model) are shown for comparison. Data are presented as mean±SEM and were collected from three independent experiments (B, D, E); p values (ns, not significant; *p<0.05, **p<0.01, ***p<0.001) were determined using one-way ANOVA followed by Dunnett’s multiple-comparison test, unpaired two-tailed Student’s t-test (B, E) or log-rank test (D). In B and E, the comparison time points for tumor volume measurement, that is, the final time point at which one mouse of the compared groups had reached the experimental end point, are indicated. ANOVA, analysis of variance; CTLA-4, cytotoxic T lymphocyte-associated protein 4; hRT, hypofractionated radiotherapy; IL-2c, interleukin-2 complexes; PD-1, programmed death 1.

hRT/αPD-1/αCTLA-4 enhances CD8⁺ and CD4⁺ T-cell infiltration of abscopal tumors

We quantified bulk CD8⁺ and CD4⁺ T cells, as well as tumor-specific CD8⁺ T cells, across key tissues involved in the antitumor T-cell response, including primary and secondary tumors, TDLNs, peripheral blood, and spleen (figure 2A). Bulk CD4⁺ T cells and tumor-specific CD8⁺ T cells increased exclusively in the abscopal tumor of hRT/αPD-1/αCTLA-4-treated mice, but not in the irradiated tumor or in either tumor of hRT/αPD-1/IL-2c-treated mice (figure 2B–D; online supplemental figure 4). In line with our recent findings,¹⁷ hRT/αPD-1/IL-2c treatment instead led to increased numbers of bulk and tumor-specific CD8⁺ T cells in extratumoral compartments, namely the spleen, blood, and lymph nodes draining the unirradiated tumor. Most CD4⁺ TILs in abscopal tumors of hRT/αPD-1/αCTLA-4-treated mice were Th1-like cells (online supplemental figure 5), and this regimen induced not only tumor-specific CD8⁺ but also tumor-specific CD4⁺ T cells in abscopal tumors (online supplemental figure 6). FoxP3⁺ Treg CD4⁺ were reduced in the non-irradiated (abscopal) tumor by day 7 in both hRT/αPD-1/αCTLA-4-treated and hRT/αPD-1/IL-2c-treated mice, relative to untreated controls, and this reduction became more pronounced by day 13 in the hRT/αPD-1/αCTLA-4 group (online supplemental figure 7). Consistent with the flow cytometry data (figure 2B–D), immunofluorescence analysis showed extensive T-cell infiltration in the unirradiated secondary tumor of hRT/αPD-1/αCTLA-4-treated mice, while only limited infiltration was observed in hRT/αPD-1/IL-2c-treated mice and virtually none in untreated controls (figure 2E).

Figure 2. Enhanced tumor-specific CD8⁺ and CD4⁺ T-cell responses in non-irradiated tumors with hRT/αPD-1/αCTLA-4 versus hRT/αPD-1/IL-2c. (A) Scheme for treatments. (B–D) Absolute numbers and frequencies of CD8⁺ T cells (B), AH1-tet⁺ (tumor-specific) CD8⁺ T cells (C), and CD4⁺ T cells (D) in different compartments of mice treated with hRT/αPD-1 (n=5), hRT/αPD-1/IL-2c (n=5–6) and hRT/αPD-1/αCTLA-4 (n=5–6). (E) Immunofluorescence images of DAPI and CD3 co-staining in non-irradiated tumors from untreated, hRT/αPD-1/IL-2c-treated and hRT/αPD-1/αCTLA-4 treated mice. Scale bar, 500 µm. Quantification of the CD3/DAPI ratio and CD3⁺ T cells/mm² is shown (n=4 per group). All samples were collected and analyzed on day 7 after treatment initiation. Data are presented as mean±SEM and were collected from two independent experiments (B–D). P values (ns, not significant; *p<0.05, **p<0.01, ***p<0.001) were determined using Kruskal-Wallis followed by Dunn’s multiple-comparison or one-way ANOVA followed by Tukey’s multiple-comparison test. ANOVA, analysis of variance; CTLA-4, cytotoxic T lymphocyte-associated protein 4; DAPI, 4′,6-diamidino-2-phenylindole; hRT, hypofractionated radiotherapy; IL-2c, interleukin-2 complexes; PD-1, programmed death 1; TDLN, tumor-draining lymph node.

hRT/αPD-1/αCTLA-4 induces more polyfunctional and proliferating TILs than hRT/αPD-1/IL-2c, especially in abscopal tumors

To assess effector function, we analyzed bulk CD4⁺ TILs and tumor antigen-specific CD8⁺ TILs. Striking differences in cytokine-expressing TILs were observed in the unirradiated tumors between the two triple therapies. By day 7, hRT/αPD-1/αCTLA-4 induced more than 10-fold higher numbers of tumor-specific CD8⁺ TILs producing IFN-γ, IFN-γ+TNF, or IFN-γ+IL-2 (ie, polyfunctional TILs) compared with hRT/αPD-1/IL-2c (figure 3A; online supplemental figure 8A,C,E). Notably, a similar pattern was observed for bulk CD4⁺ TILs (figure 3B; online supplemental figure 8B,D,F). Surprisingly, proliferation of both bulk CD4⁺ TILs and tumor-specific CD8⁺ TILs was also higher in hRT/αPD-1/αCTLA-4-treated mice, again predominantly in the unirradiated tumor (figure 3C–E).

Figure 3. hRT/αPD-1/αCTLA-4 induces more polyfunctional and fewer exhausted tumor-specific CD8⁺ and CD4⁺ TILs than hRT/αPD-1/IL-2c in non-irradiated tumors. (A–B) Absolute numbers of cytokine-secreting AH1-tet⁺ CD8⁺ (A) and CD4⁺ (B) TILs from C51 tumor-bearing mice treated with hRT/αPD-1, hRT/αPD-1/IL-2c, or hRT/αPD-1/αCTLA-4 (n=5–9). Cytokine production was assessed 4 hours after ex vivo stimulation with PMA/ionomycin. (C–D) Percentage of Ki67⁺ among AH1-tet⁺ CD8⁺ TILs (C) and CD4⁺ TILs (D) in C51 tumor-bearing mice treated with hRT/αPD-1, hRT/αPD-1/IL-2c, or hRT/αPD-1/αCTLA-4 (n=5 per group). (E) Representative flow cytometry plots showing Ki67 expression on AH1-tet⁺ CD8⁺ and CD4⁺ TILs from tumors of mice treated with hRT/αPD-1, hRT/αPD-1/IL-2c, or hRT/αPD-1/αCTLA-4. (F) Absolute numbers of PD-1⁺ AH1-tet⁺ CD8⁺ and PD-1⁺ CD4⁺ exhausted TIL subset cells per gram of tumor tissue following treatment with hRT/αPD-1, hRT/αPD-1/IL-2c, or hRT/αPD-1/αCTLA-4. Flow cytometry staining examples for identification of TCF1⁻ Tim3⁺ and TCF1⁺ Tim3⁻ subsets among AH1-tet⁺ PD-1⁺ CD8⁺ TILs and PD-1⁺ CD4⁺ TILs from treated mice (bottom panel). All samples were collected and analyzed on day 7 after treatment initiation. Data are presented as mean±SEM and were collected from two to three independent experiments. P values (ns, not significant; *p<0.05, **p<0.01, ***p<0.001) were determined using unpaired two-tailed Student’s t-test or Mann-Whitney test (A–D), and one-way ANOVA followed by Tukey’s multiple-comparison test (F). ANOVA, analysis of variance; CTLA-4, cytotoxic T lymphocyte-associated protein 4; hRT, hypofractionated radiotherapy; IFN, interferon; IL-2c, interleukin-2 complexes; PD-1, programmed death 1; PMA, phorbol 12-myristate 13-acetate; TCF1, T cell factor 1; Tim3, T cell immunoglobulin and mucin-domain containing 3; TIL, tumor-infiltrating lymphocyte; TNF, tumor necrosis factor.

hRT/αPD-1/αCTLA-4 induces more non-terminally differentiated exhausted TILs than hRT/αPD-1/IL-2c, especially in abscopal tumors

Within tumors, chronic exposure of tumor-specific T cells to epitope–MHC complexes on tumor cells leads to T-cell dysfunction or exhaustion, a phenomenon primarily studied in CD8⁺ T cells. We analyzed key markers representing different exhaustion states—stem-like, non-terminally exhausted effector-like, and terminally exhausted T cells. Notably, differences between treatment groups were observed only in the non-irradiated tumor, not in the irradiated one (figure 3F; online supplemental figure 1B). In the CD8⁺ T-cell compartment, we detected a significant difference in the number of non-terminally exhausted, effector-like tumor-specific T cells (T cell factor-1 [TCF-1]⁻ T cell immunoglobulin and mucin-domain containing-3 [TIM-3]⁺ CD101⁻ PD-1⁺). These cells were significantly more abundant in the unirradiated tumors of hRT/αPD-1/αCTLA-4-treated mice than in those of hRT/αPD-1/IL-2c-treated mice (figure 3F, left; bulk CD8⁺ PD-1⁺ TILs: online supplemental figure 8G). Among CD4⁺ TILs, hRT/αPD-1/αCTLA-4 increased the numbers of TCF-1⁻ TIM-3⁺ CD101⁻ PD-1⁺ effector-like cells and additionally elevated the numbers of TCF-1⁺ TIM-3⁻ PD-1⁺ stem-like cells (figure 3F, right). In CD8⁺ T cells, CD101 expression distinguishes non-terminally exhausted (CD101⁻ TCF-1⁻ TIM-3⁺PD-1⁺) from terminally exhausted (CD101⁺TCF-1⁻ TIM-3⁺PD-1⁺) subsets.19,21 We extended this analysis to CD4⁺ TILs by evaluating IFN-γ and TNF expression in CD101⁻ and CD101⁺subsets of TCF-1⁻ TIM-3⁺PD-1⁺ CD4⁺ T cells. Notably, IFN-γ/TNF co-expression was higher in the CD101⁻ subset, suggesting that CD101 may also mark functional, non-terminally exhausted effector-like cells in the CD4⁺ T-cell compartment (online supplemental figure 8H,I).

The strong abscopal effect of hRT/αPD-1/αCTLA-4 treatment depends on both CD8⁺ and CD4⁺ T cells

To determine the contribution of CD8⁺ and CD4⁺ T cells, we performed antibody-mediated depletion in triple-treated mice (figure 4A,B). The response of the irradiated tumor was CD8⁺ T-cell dependent in both hRT/αPD-1/αCTLA-4 and hRT/αPD-1/IL-2c treatment groups (figure 4C, left). The modest abscopal effect observed with hRT/αPD-1/IL-2c also relied on CD8⁺ T cells. In contrast, the robust abscopal effect induced by hRT/αPD-1/αCTLA-4 required both CD8⁺ and CD4⁺ T cells (figure 4C, right).

Figure 4. CD4⁺ T-cell depletion markedly impairs the abscopal effect and reduces tumor-specific CD8⁺ TILs in non-irradiated tumors following hRT/αPD-1/αCTLA-4 treatment. (A) Scheme of treatments. (B) Representative flow cytometry plots of peripheral blood 1 day after administration of CD8⁺ or CD4⁺ T cell-depleting antibodies in mice treated with hRT/αPD-1/αCTLA-4. (C) C51 tumor growth curves for primary (left) and secondary (abscopal) tumors (right) in mice treated with hRT/αPD-1/IL-2c or hRT/αPD-1/αCTLA-4, with or without CD4⁺ or CD8⁺ T cell-depleting antibodies. (D–E) Numbers of AH1-tet⁺ CD8⁺ TILs (D) and exhausted PD-1⁺ AH1-tet⁺ CD8⁺ TIL subsets cells (E) from mice treated with hRT/αPD-1/αCTLA-4, with or without CD4⁺ T-cell depletion. Samples were collected from two independent experiments (C–E) and analyzed on day 7 after treatment initiation (D–E). Data are presented as mean±SEM. P values (ns, not significant; *p<0.05, **p<0.01, ***p<0.001, **** p<0.0001) were determined using one-way ANOVA followed by Dunnett’s multiple-comparison test (C), and unpaired t-test or Mann-Whitney U test (D, E). ANOVA, analysis of variance; CTLA-4, cytotoxic T lymphocyte-associated protein 4; hRT, hypofractionated radiotherapy; IL-2c, interleukin-2 complexes; PD-1, programmed death 1; TCF1, T cell factor 1; Tim3, T cell immunoglobulin and mucin-domain containing 3; TIL, tumor-infiltrating lymphocyte.

CD4⁺ T-cell depletion reduces CD8⁺ tumor-specific TILs, cross-presenting DCs, and impairs cross-presentation in abscopal tumors of hRT/αPD-1/αCTLA-4-treated mice

In hRT/αPD-1/αCTLA-4-treated mice, CD4⁺ T-cell depletion significantly reduced the number of tumor-specific CD8⁺ TILs in the non-irradiated (abscopal) tumor, but not in the irradiated tumor (figure 4D). This reduction primarily affected more differentiated, exhausted CD8⁺ TILs (figure 4E). To examine cross-presenting DCs, we used the B16-OVA model because the monoclonal antibody (“25-D1.16”) enables flow-cytometric detection of cell-surface complexes of the immunodominant OVA epitope SIINFEKL bound to H-2K^b on DCs.²² CD4⁺ T-cell depletion also significantly decreased total DCs, activated CD86⁺ DCs, and cDC1s cross-presenting the OVA antigen, specifically within the abscopal tumor (figure 5A–E). Similarly, in hRT/αPD-1/IL-2c-treated mice, fewer activated CD80⁺/CD86⁺ cross-presenting cDC1s were detected in the unirradiated tumor compared with hRT/αPD-1/αCTLA-4-treated mice (figure 5F). Notably, CD4-expressing DCs were virtually undetectable (online supplemental figure 9), arguing against a role for CD4-expressing DCs in tumor-antigen cross-presentation in unirradiated tumors.

Figure 5. CD4⁺ T-cell depletion diminishes activated cross-presenting DCs and cross-presentation in non-irradiated tumors. (A–B) Flow cytometry gating strategy to identify DC subsets, including SIINFEKL/K^b+ DCs, in secondary tumors. (C–E) Absolute numbers of total DCs (C), SIINFEKL/K^b+ CD86⁺ DCs (D) and SIINFEKL/K^b+ CD86⁺ CD103⁺ cDC1s (E) in B16-OVA melanoma tumors following the indicated treatments (n=6–7). (F) Frequency of CD80⁺CD86⁺ cells among CD103⁺ DCs in primary and secondary tumors of C51 tumor-bearing mice treated with hRT/αPD-1, hRT/αPD-1 plus IL-2c or αCTLA-4 (n=5 per group). All samples were collected and analyzed on day 7 after treatment initiation, pooled from two independent experiments (C–F). Data are presented as mean±SEM. P values (ns, not significant; *p<0.05, **p<0.01, ***p<0.001) were determined using unpaired two-tailed Student’s t-test (C, F). CTLA-4, cytotoxic T lymphocyte-associated protein 4; DC, dendritic cell; hRT, hypofractionated radiotherapy; IL-2c, interleukin-2 complexes; PD-1, programmed death 1.

Activated CD4⁺ T cells enhance cDC1-mediated cross-presentation in a manner partially dependent on IFN-γ and TNF

To better understand how effector-like CD4⁺ T cells may promote the accumulation of tumor-specific CD8⁺ TILs via cross-presenting DCs, we investigated how activated CD4⁺ T cells influence DC activation and antigen cross-presentation in vitro. CD4⁺ splenocytes were stimulated with αCD3 and αCD28 antibodies in the presence of IL-12 and IL-2. Similar to CD4⁺ TILs from non-irradiated tumors of hRT/αPD-1/αCTLA-4-treated mice (figure 3B and online supplementalfigure 10), these αCD3-activated CD4⁺ T cells secreted the TH1 effector cytokines IFN-γ and TNF, along with the chemokines CCL3, CCL5, and CXCL10 (figure 6A), which attract T cells and DCs.²³ Co-culture of DCs with these activated CD4⁺ T cells led to increased expression of the co-stimulatory molecules CD80, CD86, and CD70 on cDC1s (figure 6B and C). Since IFN-γ and TNF are key markers and mediators of TH1 effector-like CD4⁺ T cells, we next tested whether these cytokines directly influence DC activation and cross-presentation. As shown in figure 6D, addition of recombinant IFN-γ and TNF upregulated CD80, CD86, CD70, and MHC class I—but not MHC class II—expression on cDC1s. To assess antigen cross-presentation, we incubated cDC1s with OVA protein and analyzed presentation of the SIINFEKL–K^b complex by flow cytometry.²² Cross-presentation was significantly enhanced in the presence of activated CD4⁺ T cells (figure 6E, red histogram/bar). To evaluate the contribution of IFN-γ and TNF, we repeated the assay with neutralizing antibodies against one or both cytokines. As shown in figure 6E, neutralization of IFN-γ or TNF—especially in combination—substantially reduced the cross-presentation capacity of cDC1s. DCs isolated from unirradiated tumors of hRT/αPD-1/αCTLA-4-treated mice activated CD8⁺ T cells via cross-presentation (online supplemental figure 11A,B), and combined CD70/CD80/CD86 blockade reduced tumor-specific effector CD8⁺ TILs in unirradiated tumors (online supplemental figure 11C–E).

Figure 6. CD4⁺ T cells enhance cDC1-mediated cross-presentation in vitro. (A) Isolated CD4⁺ T cells were activated with αCD3 and αCD28 antibodies for 48 hours, followed by flow cytometric analysis. MFI quantifications are shown for indicated cytokines and chemokines produced by naïve or activated CD4⁺ T cells. (B) Splenic DCs isolated from B16 tumor-bearing mice expressing Flt3L were co-cultured with naïve or activated CD4⁺ T cells for 16 hours and analyzed by flow cytometry. Shown are the percentages of CD70⁺ or CD80⁺CD86⁺ cells among CD8α⁺ DCs (cDC1) for DCs alone (white), or co-cultured with naïve (gray) or activated (red) CD4⁺ T cells. (C) Representative gating and marker expression profiles for cDC1. (D) DCs were incubated with recombinant IFN-γ (10 µg/mL) plus TNF (10 µg/mL) for 48 hours, followed by flow cytometric analysis; shown are the frequencies of the indicated markers expressed on cDC1. (E) Naïve or activated CD4⁺ T cells were co-cultured with DCs in the presence of OVA protein and LPS (50 ng/mL) for 16 hours, with or without IFN-γ neutralizing and/or TNF-blocking antibodies, followed by flow cytometric analysis. Representative MFI histograms (left) and quantification of SIINFEKL/K^b+ cDC1s (right) are shown. Data pooled from three to four independent experiments with three technical replicates (A, B, D, E) each are presented as mean±SEM. P values (ns, not significant; *p<0.05, **p<0.01, ***p<0.001) were determined using unpaired two-tailed Student’s t-test (A, D) or one-way ANOVA followed by Dunnett’s multiple-comparison test (B) or the Benjamini-Krieger-Yekutieli procedure (E). ANOVA, analysis of variance; DC, dendritic cell; IFN, interferon; LPS, lipopolysaccharide; MFI, median fluorescence intensity; MHC, major histocompatibility complex; rIFN, recombinant interferon; rTNF, recombinant tumor necrosis factor; TNF, tumor necrosis factor.

CD4⁺ T cells facilitate the accumulation of adoptively transferred, RT-induced tumor-specific CD8⁺ T cells at non-irradiated tumor sites

To further investigate the role of CD4⁺ T cells in facilitating the accumulation of RT/ICB-induced tumor-specific CD8⁺ T cells in unirradiated tumors, we performed adoptive transfer experiments (figure 7). CD8⁺ T cells were isolated from blood and TDLNs of hRT/αPD-1/αCTLA-4-treated donor mice (bearing one irradiated tumor) and transferred into recipient mice carrying only a non-irradiated tumor. Recipients received either no treatment, αPD-1/αCTLA-4, or αPD-1/αCTLA-4 combined with CD4⁺ T-cell depletion. Studies were conducted in both the C51 (BALB/c) and B16-CD133 (C57BL/6) tumor models, using CellTrace or the congenic marker CD45.1 to track donor CD8⁺ T cells. No lymphodepleting conditioning was applied prior to transfer. 2–3 days post-transfer, the presence of donor CD8⁺ T cells in the unirradiated tumor and its TDLNs was assessed by flow cytometry. In untreated recipients, virtually no donor CD8⁺ T cells were detected in either the tumor or TDLNs (figure 7A–D and E–H). In contrast, αPD-1/αCTLA-4-treated recipients showed clear accumulation of donor-derived CD8⁺ T cells, including tumor-specific (tetramer⁺) CD8⁺ T cells, at these sites. This accumulation was significantly reduced on CD4⁺ T-cell depletion. Moreover, nearly all donor-derived tumor-specific CD8⁺ T cells recovered from αPD-1/αCTLA-4-treated recipients expressed the activation/exhaustion marker PD-1 (figure 7F and I).

Figure 7. CD4⁺ T cells facilitate the accumulation of adoptively transferred, RT-induced tumor-specific CD8⁺ T cells at non-irradiated tumor sites. (A, E) Schematic representation of ACT in the C51 and B16-CD133 tumor models, respectively. (B) Representative flow cytometry plots of CTFR-labeled transferred CD8⁺ T cells or AH1-tet⁺ CTFR⁺ CD8⁺ T cells in TDLNs of recipient mice at 48 hours post-ACT (n=3). (C–D) Numbers of CTFR⁺ CD8⁺ T cells (C) and CTFR⁺ AH1-tet⁺ CD8⁺ (D) T cells in TDLNs of recipient mice. (F) Representative flow cytometry plots of transferred CD45.1⁺ CD8⁺ T cells, M8-tet⁺ CD45.1⁺ CD8⁺ T cells, and PD-1⁺ M8-tet⁺ CD45.1⁺ CD8⁺ T cells in tumors of recipient mice at 72 hours post-ACT (n=3). (G–I) Percentage of CD45.1⁺ among CD8⁺ TILs (G), M8-tet⁺ among CD45.1⁺ CD8⁺ TILs (H), and PD-1⁺ among CD45.1⁺ M8-tet⁺ CD8⁺ TILs (I). Data are presented as mean±SEM. P values (ns, not significant; *p<0.05, **p<0.01, ***p<0.001) were determined using unpaired two-tailed Student’s t-test. ACT, adoptive cell transfer; CTFR, CellTrace Far Red; CTLA-4, cytotoxic T lymphocyte-associated protein 4; hRT, hypofractionated radiotherapy; PD-1, programmed death 1; TDLN, tumor-draining lymph node; TIL, tumor-infiltrating lymphocyte.

Discussion

RT can induce DAMP-mediated ICD, leading to activation of tumor-specific CD8⁺ T cells via DC-mediated cross-presentation.1,4 Clinically, enhancing the systemic, T cell-driven abscopal effect remains a highly desirable goal.¹ However, limited infiltration of non-irradiated (abscopal) tumors by RT/ICB-induced CD8⁺ T cells—potentially due to insufficient restimulation at these sites—may represent a key limiting factor. We show that an RT/immunotherapy combination inducing effector-like CD4⁺ T cells in addition to CD8⁺ T cells (hRT+αPD-1+αCTLA-4) markedly increases CD8⁺ T-cell accumulation in non-irradiated tumors, including circulating tumor-specific CD8⁺ T cells primed by hRT. The induced CD4⁺ T cells play a key role at abscopal sites, enhancing both the number and function of tumor-specific CD8⁺ T cells, as well as the activation, abundance, and cross-presenting capacity of cDC1. We further show that activated CD4⁺ T helper cells promote cDC1 maturation, activation, and cross-presentation, partly through IFN-γ and TNF. Together, these findings suggest that robust induction of effector-like CD4⁺ T cells enables effective restimulation of RT-induced, tumor-specific CD8⁺ T cells by cDC1 at non-irradiated sites, thereby increasing their abundance and effector functions and enhancing the abscopal effect (online supplemental figure 12).

Growing evidence suggests that T-cell (re)stimulation by DCs within tumors is important, in addition to the well-established role of DCs in priming tumor-specific T cells in lymph nodes.^{10 11 24 25} These DCs must be activated, expressing costimulatory molecules such as CD80/86.^{24 26} However, immunologically cold or progressive tumors often lack activated DCs, particularly activated cross-presenting subsets.^{3 9} 

The activation and maturation of DCs (particularly cross-presenting DCs) has been linked to pathogen-associated molecular pattern, DAMPs, and CD4⁺ T cell help.27,32 In our model using hRT+αPD-1+αCTLA-4, the abscopal response not only depended on CD8⁺ T cells but also partially on CD4⁺ T cells, whereas the local response of the irradiated tumor was CD8⁺ T cell-dependent only. The observed correlation between CD4⁺ T-cell presence and increased infiltration of non-irradiated tumors by activated cross-presenting DCs, enhanced cross-presentation, and effector-like CD8⁺ T cells suggests functional collaboration between CD4⁺ T cells, cross-presenting DCs, and CD8⁺ T cells at abscopal sites.

Early-differentiated, stem-like CD8⁺ T cells from lymph nodes of hRT-treated tumors appear crucial for the abscopal effect.³³ However, to exert the abscopal effect, these stem-like CD8⁺ T cells must infiltrate non-irradiated tumors, differentiate, and expand as effector cells. The intratumoral stem-to-effector CD8⁺ T-cell differentiation is strongly promoted by activated CD80⁺/86⁺ DCs.²⁴ Notably, CD28—the receptor of CD80/86—is expressed on stem-like T cells.^{34 35} Growing evidence highlights the importance of triadic interactions among cross-presenting DCs, CD8⁺, and CD4⁺ tumor-specific T cells within the tumor microenvironment.^{11 28 29 32} CD4⁺ T cells are thought to license or activate cross-presenting DCs,^{28 36 37} thereby enhancing their ability to support CD8⁺ T-cell responses. In viral infection models, early T-cell activation occurs in distinct DC-mediated steps, with the later step involving cDC1 that co-present antigen via MHC-I and MHC-II, allowing CD4⁺ T helper cells to augment CD8⁺ T-cell responses.³⁸ Recent in vitro studies show that activated CD4⁺ T cells stimulate cDC1s, but not other DC subsets, to cross-present antigen to CD8⁺ T cells,³⁷ in part via IFN-β production.³⁹ Although preferential recruitment of antitumoral CD4⁺ T cells into unirradiated tumors cannot be excluded, current evidence instead supports preferential induction and expansion of tumor-specific CD4⁺ T cells on αCTLA-4 or aPD-1+αCTLA-4 therapy, with αCTLA-4 or combined αCTLA-4/αPD-1—but not αPD-1 alone—inducing TH1-like CD4⁺ effector-like T cells.^{40 41} Our findings are consistent with these observations. In the context of hRT/αPD-1/αCTLA-4 treatment, we propose that, following initial DC-mediated priming or activation in lymph nodes draining the irradiated tumor, a second round of DC stimulation of early differentiated, RT-induced CD8⁺ tumor-specific T cells occurs in the abscopal tumor, facilitated by CD4⁺ effector-like T cells (online supplemental figure 12). Additionally, Th1 effector-like CD4⁺ TILs (see online supplemental figure 5) may enhance recruitment of tumor-specific CD8⁺ T cells into non-irradiated tumors by producing T cell-recruiting chemokines (see online supplemental figure 10)2342,44 and by secreting IFN-γ (see figure 3B), which is known to induce such chemokines.⁴⁵

With hRT/αPD-1/CD122-directed IL-2c, we observed weaker and less polyfunctional T-cell infiltration in abscopal tumors compared with hRT/αPD-1/αCTLA-4. This aligns with our earlier findings that hRT/αPD-1/IL-2c induces strong expansion of CD8⁺ T cells—including stem-like tumor-specific CD8⁺ T cells—mainly in extratumoral compartments.¹⁷ The comparatively low tumor T-cell infiltration may help explain the failure of recent phase 3 trials combining CD122-targeted IL-2 derivatives with αPD-1.14,16 Notably, IL-2 and CD122-targeted IL-2 promote proliferation and activation across all cDC subsets.⁴⁶ This raises the question of whether IL-2-based and αCTLA-4-based therapies differentially induce or activate cross-presenting DCs. We found that the stronger abscopal effect from hRT/αPD-1/αCTLA-4 depended on CD4⁺ T cells, while the weaker abscopal response with hRT/αPD-1/IL-2c was CD4⁺-independent. High-dose IL-2 can paralyze primary CD4⁺ T-cell responses.⁴⁷ Thus, although CD122-directed IL-2 may induce and activate DCs, the formation of DC/CD8⁺ T cell/CD4⁺ T cell triads—structures critical for effective DC-mediated CD8⁺ T-cell activation—could be impaired.

Several preclinical studies have explored strategies to enhance T-cell infiltration of abscopal tumors. Two independent studies showed that adding low-dose external RT to abscopal tumors—alongside high-dose hRT to the primary tumor and ICB—enhances T-cell numbers and activity within abscopal lesions.^{48 49} Similar effects were observed on combining low-dose systemic radioligand therapy with hRT and ICB.⁵⁰ Adding low-dose cytotoxic chemotherapy (cisplatin or doxorubicin) to hRT+ICB also enhanced T-cell infiltration and improved response of abscopal tumors.^{42 51 52}

The abscopal effect on hRT/αPD-1/αCTLA-4 treatment has been studied in preclinical models, though with a distinct mechanistic focus. One study highlighted complementary effects: hRT to the primary tumor expanded the T-cell repertoire, αPD-1 limited T-cell exhaustion, and αCTLA-4 reduced CD4⁺ immunosuppressive Tregs.⁵³ The combination of hRT to the primary tumor, T cell-attracting LDRT to the abscopal tumor, and dual ICB with αPD-1 and αCTLA-4 was shown to depend on CD4⁺ T cells. However, the specific therapeutic component driving the CD4⁺ T-cell dependence as well as the mechanistic implications was not examined.⁴⁸ Since we compared two triple combinations sharing the hRT/αPD-1 backbone (hRT/αPD-1/αCTLA-4 vs hRT/αPD-1/IL-2c), differing only by the addition of αCTLA-4 or IL-2c, and CD4⁺ T-cell dependence of the abscopal effect was observed only with the αCTLA-4 triple combination, this dependence can be attributed to the αCTLA-4 component.

Dual ICB with αPD-1 and αCTLA-4 has shown improved clinical outcomes in cancers such as melanoma and renal cell carcinoma. However, many patients are initially resistant, and responders often develop adaptive resistance.⁵⁴ The combination of potentially immunogenic RT with dual ICB (αPD-1 and αCTLA-4) has been investigated in several clinical trials. A recent meta-analysis of prospective trials suggests that combining RT with dual αPD-1/αCTLA-4 ICB improves 6-month overall survival compared with RT with single ICB.⁵⁵ However, to date, only a few randomized trials with or without RT have been completed. These trials were conducted in patients with αPD-1-resistant non-small cell lung cancer,⁵⁶ relapsed small cell lung cancer,⁵⁷ and castration-resistant (typically ICB-non-responsive) prostate cancer,⁵⁸ all with multiple metastases, where only one or a few tumor lesions were irradiated. Early-phase studies in hepatocellular carcinoma and microsatellite-stable colorectal cancer have shown encouraging systemic responses to combined RT and dual ICB.^{59 60} Notably, the hepatocellular carcinoma study provided preliminary evidence of abscopal tumor regression.⁵⁹ It is conceivable that, in tumors with sufficient antigenicity and immunogenicity and limited ICB resistance, clinical abscopal responses could be enhanced by optimal lesion selection, use of predictive biomarkers, refinement of other RT/ICB parameters (including RT-ICB sequencing,^{18 61 62} and—consistent with our findings—promotion of CD4⁺ T-cell responses in unirradiated lesions. Analyses of serial biopsies from non-irradiated tumors in clinical trials specifically designed to elicit an abscopal effect (eg, with RT/αPD-1/αCTLA-4 or other combinations that elicit both CD8⁺ and CD4⁺ antitumor responses) could help validate our preclinical observations in patients.

In summary, our data demonstrate that robust RT-dependent abscopal tumor regression can be achieved with a combination therapy that induces CD4⁺ T cells alongside CD8⁺ tumor-specific T cells. The results suggest that CD4⁺ effector T cells act at abscopal sites by promoting DC-mediated cross-presentation of tumor antigens to CD8⁺ T cells, including those originating from the irradiated tumor. Our findings support the concept that restimulation of RT-induced CD8⁺ T cells by cross-presenting DCs in non-irradiated tumors enhances the abscopal effect, with CD4⁺ effector T cells likely playing a role analogous to RT-induced DAMPs by facilitating DC-mediated CD8⁺ T-cell activation.

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.