Releasing the brakes of tumor immunity with anti-PD-L1 and pushing its accelerator with L19-IL2 cures poorly immunogenic tumors when combined with radiotherapy

Abstract

Background: Poorly immunogenic tumors are hardly responsive to immunotherapies such as immune checkpoint blockade (ICB) and are, therefore, a therapeutic challenge. Combination with other immunotherapies and/or immunogenic therapies, such as radiotherapy (RT), could make these tumors more immune responsive. We have previously shown that the immunocytokine L19-IL2 combined with single-dose RT resulted in 75% tumor remission and a 20% curative abscopal effect in the T cell-inflamed C51 colon carcinoma model. This treatment schedule was associated with the upregulation of inhibitory immune checkpoint (IC) molecules on tumor-infiltrating T cells, leading to only tumor growth delay in the poorly immunogenic Lewis lung carcinoma (LLC) model.

Methods: We aimed to trigger curative therapeutic responses in three tumor models (LLC, C51 and CT26) by "pushing the accelerator" of tumor immunity with L19-IL2 and/or "releasing the brakes" with ICB, such as antibodies directed against cytotoxic T lymphocyte associated protein 4 (CTLA-4), programmed death 1 (PD-1) or its ligand (PD-L1), combined with single-dose RT (10 Gy or 5 Gy). Primary tumor endpoint was defined as time to reach four times the size of tumor volume at start of treatment (4T×SV). Multivariate analysis of 4T×SV was performed using the Cox proportional hazards model comparing each treatment group with controls. Causal involvement of T and natural killer (NK) cells in the anti-tumor effect was assessed by in vivo depletion of T, NK or both cell populations. Immune profiling was performed using flow cytometry on single cell suspensions from spleens, bone marrow, tumors and blood.

Results: Combining RT, anti-PD-L1 and L19-IL2 cured 38% of LLC tumors, which was both CD8⁺ T and NK cell dependent. LLC tumors were resistant to RT +anti-PD-L1 likely explained by the upregulation of other IC molecules and increased T regulatory cell tumor infiltration. RT+L19-IL2 outperformed RT+ICB in C51 tumors; effects were comparable in CT26 tumors. Triple combinations were not superior to RT+L19-IL2 in both these models.

Conclusions: This study demonstrated that combinatorial strategies rationally designed on biological effects can turn immunotherapy-resistant tumors into immunologically responsive tumors. This hypothesis is currently being tested in the international multicentric randomized phase 2 trial: ImmunoSABR (NCT03705403).

Keywords: immunotherapy; radiotherapy; tumor microenvironment.

Figure 1

Therapeutic effect of RT combined with L19–IL2 and/or ICB treatment. (A) Treatment schedule. Each mouse was injected with tumor cells on the right flank on day 8 (C51 and CT26 models) or day 5 (LLC model). Blood was withdrawn 3 days before and 6 days after treatment start for flow cytometric analysis. Fraction of tumors not reaching four times start tumor volume in LLC (n=7–8 mice per treatment arm) (B), CT26 (n=7–9) (C) and C51 (n=7–9) (D) tumor models. C51 and CT26 tumors were irradiated with a single dose of 5 Gy and LLC tumors with an isoeffective single dose of 10 Gy. Treatments in the LLC and the CT26 models were performed in one single experiment, while in the C51 model two independent experiments were performed (see online supplemental figure 4). Combinations with anti-PD-1 in the C51 (n=6–8) and CT26 (n=5–6) were tested in an independent experiment. Differences between treatment groups for each tumor model are summarized in table 1. Treatments: blue: RT, red: RT + L19-IL2, green: RT + ICB, black: RT + L19-IL2 + ICB. ICB, immune checkpoint blockade; IgG, immunoglobulin G; LLC, Lewis lung carcinoma; PBS, phosphate-buffered saline; RT, radiotherapy.

Figure 2

RT+L19–IL2+anti-PD-L1 (trimodal) therapy effect is associated with high infiltration of NK and antigen-experienced CD8⁺ CD44⁺ T cells in the LLC model. (A) Treatment schedule for the evaluation of immunological parameters in the LLC model (n=5 mice per treatment arm), experiment was performed once. Stainings, flow cytometry data acquisition, and analysis of the samples were done in independent duplicates. (B) Representative flow cytometry dot plots of tumor-infiltrating CD8⁺ CD44⁺ (gated on CD3⁺ T cells) T cells 6 days after the start of treatment and quantification. (C) Representative flow cytometry dot plots of tumor-infiltrating T (CD3⁺ NK1.1^-), NKT (CD3⁺ NK1.1⁺) and NK (CD3^- NK1.1⁺) cells and quantifications of NKT and NK cell percentages. IgG, immunoglobulin G; LLC, Lewis lung carcinoma; PBS, phosphate-buffered saline; RT, radiotherapy.

Figure 3

Changes in pretreatment immunological blood parameters (expression of PD-1 and PD-L1 on CD4⁺ and CD8⁺ T cells) during triple therapy (RT+L19–IL2+anti-PD-L1). Blood samples were collected 3 days before treatment (day 3) and on day 6 after treatment start (n=8–9 mice per tumor model, see treatment scheme in figure 1A). Data are reported as the day 6/day -3 ratio of the MFI of PD-L1 on CD4⁺ (A) and CD8⁺ (B) T cells and of PD-1 on CD4⁺ (C) and CD8⁺ T cells (D). Ratio >1 indicates an increase and ratio <1 indicates a decrease in the parameter during therapy. (E) and (F) Representative flow cytometry histograms of (C) and (B), respectively. MFI, median fluorescence intensity.

Figure 4

CD8⁺ T and NK cells play a key role in the anti-tumor effect of triple therapy. (A) Treatment schedule of the depletion study. (B) Flow cytometric analysis of CD8⁺ T cell and NK cell presence in the blood to confirm their depletion after 3 doses of depleting antibodies. (C) Fraction of tumors not reaching four times start tumor volume in LLC tumor-bearing mice (n=9 mice per treatment arm) treated with 10 Gy+L19–IL2+anti-PD-L1 depleted of CD8⁺ T (red), NK (green) or both cell types (blue) or IgG-treated (black) according to schedule in (A). This depletion study was performed once. IgG, immunoglobulin G; LLC, Lewis lung carcinoma.

Figure 5

PD-L1 blockade upregulates the expression of other co-inhibitory receptors on tumor-infiltrating T cells. Representative flow cytometry histograms and quantifications of PD-1⁺ (A), Tim-3⁺ (B) and CD39⁺ (C) tumor-infiltrating CD4⁺ and CD8⁺ T cells from LLC tumors at day 6 after the start of treatment depicted in figure 2A. (D) Representative flow cytometry histograms of CD45− PD-L1⁺ LLC tumor cells and CD3− NK1.1− CD11b⁺ LLC tumor-infiltrating cells and quantifications. Experiment was performed once. Stainings, flow cytometry data acquisition and analysis of the samples were done in independent duplicates. LLC, Lewis lung carcinoma; RT, radiotherapy.

Figure 6

RT+anti-PD-L1 increases tumor-infiltrating Tregs. (A) Representative flow cytometry dot plots of tumor-infiltrating Tregs and quantification of CD4⁺ CD25⁺ FoxP3⁺ Tregs, CD4⁺ CD25^- FoxP3⁺ Tregs and CD8⁺ CD44⁺ T cell to CD25⁺ Treg ratio in LLC tumors at day six after the start of treatment depicted in figure 2A. (B) Representative flow cytometry histograms of PD-1⁺ and Tim-3⁺ tumor-infiltrating CD25⁺ Tregs and quantifications. Experiment was performed once. Stainings, flow cytometry data acquisition and analysis of the samples were done in independent duplicates. LLC, Lewis lung carcinoma; MFI, median fluorescence intensity; RT, radiotherapy; Tregs, regulatory T cells.

Figure 7

Anti-PD-L1-based trimodal therapy causes an accumulation of memory T cell subsets in lymphoid organs of cured LLC tumor-bearing mice. Representative flow cytometry dot plots and quantification of CD8⁺ CD44⁺ CD127⁺ memory T cells (A) and CD44⁺ CD62L⁺ central memory CD8⁺ T cells (B) from spleen and bone marrow of LLC tumor-bearing mice treated with 10 Gy+L19–IL2+anti-PD-L1 harvested 43±2.1 days after tumor cure (12 days after tumor rechallenge) or endpoint (T4×SV) had been reached. Cells were obtained from one experiment performed in figure 1B. Stainings, flow cytometry data acquisition, and analysis of the samples were done in independent duplicates. (C) Representative flow cytometry dot plots and quantifications of CD8⁺ CD44⁺ T cells expressing IFNγ and granzyme B from splenocytes of cured or non-cured LLC tumor-bearing mice treated with 10 Gy+L19–IL2+anti-PD-L1, or from naïve mice after co-culture with irradiated LLC target cells or GL261 as non-specific target cells. Cells were obtained from the same experiment as in (A) and (B). Co-culture assay, stainings, flow cytometry data acquisition, and analysis of the samples were done in independent duplicates. LLC, Lewis lung carcinoma.