Triple Therapy with MerTK and PD1 Inhibition Plus Radiotherapy Promotes Abscopal Antitumor Immune Responses

Lung cancer is the leading cause of cancer mortality among women and men throughout the world (1). At present, the standard of care for advanced nonmetastatic lung cancer is radiotherapy combined with chemotherapy, but this approach can be highly toxic and is not effective for controlling metastases, the cause of death for most patients with lung cancer. Traditionally, the benefit of radiotherapy was seen as resulting from its ability to damage tumor-cell DNA (2), but more recently interest has been expressed in its ability to achieve immunogenic cell death (3). To date, studies of this effect have focused mainly on using radiation to “prime” the immune system so as to enhance the ability of radio-immunotherapy combinations to elicit systemic, rather than strictly local, control. Radiation also has immunomodulatory effects via evoking production of type I IFNs and upregulating MHC-I molecules (4), which in turn activate CD8⁺ T-effector cells and release danger signals such as high-mobility-group box 1 (HMGB1) and ATP, which can activate macrophages (5), and upregulate Fas molecules on tumors, which leads to their apoptotic death. Radiation has also been shown to promote changes in the inflammatory microenvironment (6, 7) such as increasing antigen presentation by myeloid cells within the tumor stroma, which in turn enhances T-cell–mediated killing of tumor cells (8). We recently found that resistance to immune checkpoint inhibitors such as anti-PD1 was due in part to downregulation of MHC-I and that radiation was capable of upregulating MHC-I molecule expression and re-sensitizing tumors to anti-PD1 therapy (4).

On the other hand, radiation also has immunosuppressive effects via its upregulation of Tregs (7), myeloid-derived suppressor cells (MDSC), and M2 macrophages in the tumor microenvironment (9, 10). Tregs act to suppress a variety of immune cells, including B cells, NK cells, NKT cells, CD4⁺ T helper cells, CD8⁺ T-effector cells, monocytes, and dendritic cells, which dampen antitumor immune responses (11). In addition to Treg-depleting Abs, therapies involving depletion of M2 macrophages or polarization of macrophages towards the M1 phenotype also represent a rational therapeutic approach that can be combined with radiotherapy. Another approach is to use an Ab to the MER proto-oncogene tyrosine kinase (MerTK), a macrophage-specific phagocytic receptor, blockade of which suppresses the phagocytosis of apoptotic bodies after radiotherapy, thereby tilting the balance towards secondary necrosis and induction of inflammatory and immunogenic cell death (12, 13). MerTK is a cell surface receptor of the TAM-RTK family (Tyro3/AXL/MerTK receptor tyrosine kinases). Its overexpression on tumor cells has been linked with poor prognosis (14). MerTK initiates efferocytosis by macrophages, which is crucial for the efficient clearance of apoptotic material and intracellular antigens (15). MerTK signaling also alters macrophage gene expression so as to suppress inflammatory cytokine production and polarize macrophages towards the wound-healing, anti-inflammatory M2 phenotype (16, 17). During engulfment of apoptotic cells, MerTK suppresses the release of proinflammatory cytokines such as IL12, IL6, and TNF by M1 macrophages, partly by diminishing NF-κB signaling (18–20); MerTK also enhances the M2 macrophage-induced production of anti-inflammatory cytokines such as IL10, TGFβ, and IL4 (17, 21, 22). Therefore, MerTK inhibition promotes re-polarization of macrophages toward the M1 antitumor phenotype and enhances antigen presentation, a crucial step in T-cell priming and activation. Although previous studies have shown that MerTK antibodies can promote antitumor effects when combined with radiation (23), the optimal timing and sequence of radiation with α-MerTK mAb and α-PD1 mAbs have not been evaluated. Here, we tested the efficacy of triple therapy (RT + α-PD1 + α-MerTK) in a murine model of non–small cell lung cancer (NSCLC) that includes a KrasG12D mutation. We hypothesized that radiation and anti-PD1 can be made more immunogenic by the addition of α-MerTK antagonist, and indeed we observed improved abscopal tumor control from this triple therapy. We further found this effect to be dependent on CD8⁺ T cells and NK cells and to be mediated by increased numbers of CD8⁺ CD103⁺ tissue-resident memory cells (T_RM) at the tumor site.

We used 344SQ parental tumor cells (344SQ-P), a murine metastatic adenocarcinoma NSCLC cell line derived from a spontaneous subcutaneous metastatic lesion in p53R172HΔg/+K-rasLA1/+ mice. This cell line was a generous gift from Dr. Jonathan Kurie (MD Anderson). Cells were cultured in RPMI1640 supplemented with 100 U/mL penicillin, 100 mg/mL streptomycin, and 10% heat-activated FBS and incubated at 37°C in a 5% CO₂ atmosphere. The mAbs were all from Bristol Myers Squibb (α-MerTK, 4E9.E6; α-PD1, clone 4H2; and mouse IgG1 isotype control) and were prepared for injection by dilution in PBS at pH 7.4. For the depletion studies, CD8⁺ cells and NK cells were depleted by anti-CD8 (clone 53–6.7) from BioXcell and anti-Asialo GM1 from Wako Chemicals, respectively.

Eight- to 12-week-old 129Sv/Ev syngeneic female mice were purchased from Taconic Biosciences and bred at a mouse colony maintained by the Department of Experimental Radiation Oncology at The University of Texas MD Anderson Cancer Center. All animal procedures followed the guidelines of the Institutional Animal Care and Use Committee.

Mice were inoculated by subcutaneous injection of 344SQ-P tumor cells into the right hind leg (to establish primary tumors) and the left hind leg (to establish abscopal tumors). Tumor growth was monitored twice a week with digital calipers and recorded as the change in average tumor diameter. Primary tumors were irradiated with a Cesium source on days 7, 8, and 9, when the tumors had reached about 7 mm in diameter. Local tumor irradiation involved positioning the mice on a jig that shields the mouse's body except for the leg that is to receive the designated dose. Tumor-bearing mice were injected intraperitoneally with α-PD1 mAb (200 μg/injection) on days 5, 8, 11, and 14; and with α-MerTK mAb (200 μg/injection) on days 7, 10, 12, 15, 19, and 22. For the depletion studies, anti-CD8 (500 μg/injection) and anti-asialo NK depletory were injected interperitoneally on days 5 and thereafter once/week to maintain the pressure. Mice were euthanized when the average tumor diameter reached 14 mm in any dimension or became ulcerated.

Tumors were harvested and tissues digested with 250 μg/mL of Liberase (Roche) and incubated for 30 minutes at 37°C while shaking at 105 rpm. FBS was added to stop the digestion reaction, samples were filtered, and TILs were enriched by using Histopaque 1077 (Sigma; Catalog No. H8889). Cells were then blocked with anti-CD16/CD32 before being stained for flow cytometry. Stains included fluorochrome-conjugated anti-CD4 APC (Catalog No. 100412), anti-CD8 PercpCy5.5 (Catalog No. 100734), anti-CD45 Pacific blue (Catalog No. 103126); anti-CD103 BV510 (Catalog No. 121423), anti-PD1 FITC (Catalog No. 135214), anti-CD49b PE (Catalog No. 108907), anti-CD69 PE-Cy7 (Catalog No. 104511), anti-Gr1 BV510 (Catalog No. 108437), anti- CD11b APC Fire750 (Catalog No. 101262), and anti-F4/80 FITC (Catalog No. 123108) for flow cytometry (LSRII). All antibodies were purchased from BioLegend. Flow cytometry data were analyzed with FlowJo software.

TILs were isolated as described above, and lysed for RNA extraction. For quality assurance, a fragment analysis system was used to assess nucleic acid fragmentation, with optimal performance defined as 50% of samples of more than 300 nucleotides in length. Samples were analyzed with the NanoString nCounter Immune Panel by the Genomic & RNA Profiling Core at Baylor College of Medicine.

Tumors were harvested 7 days after the last fraction of radiation. Tumor samples were fixed in 10% neutral buffered formalin for 24 hours then washed in PBS at room temperature. Caspase 3 activity, for assessment of apoptosis, was evaluated using Polink-2 plus HRP anti-Rabbit Detection Kit for IHC (Catalog No. D39–110). The whole process was performed at MD Anderson IHC cores. Digital quantification was performed using ImageJ software.

All statistical analyses were done with Graph Pad Prism 7 software. Tumor growth curves were compared with multiple t tests. Mouse survival was analyzed by using the Kaplan–Meier method and compared with log-rank tests. Statistical significance was defined as P ≤ 0.05.

Protocols for animal use, treatment, and euthanasia were approved by the Institutional Animal Care and Use Committee of The University of Texas MD Anderson Cancer Center.

The schedule for creating the mouse model of 344SQ lung adenocarcinoma is shown in Fig. 1A. Briefly, mice were first injected subcutaneously with 0.5 × 10⁶ 344SQ-P cells in their right hind legs (to establish the primary tumor), and received a second subcutaneously injection of 0.1 × 10⁶ 344SQ-P cells in the left hind legs (to establish the abscopal tumor). Four α-PD1 doses were given intraperitoneally at 200 μg/dose on days 5, 8, 11, and 14; high-dose radiation (36 Gy in 3 12-Gy fractions) was given to the primary tumors on days 7, 8, and 9; and 6 doses of α-MERTK were given (also intraperitoneally at 200 μg/dose) on days 7, 10, 12, 15, 19, and 22. The radiation dose was picked after optimization for apoptosis rates (Supplementary Fig. S1). At day 21, half of each treatment group was evaluated for tumor growth delay and survival and the other for tumor-infiltrating leukocytes (TIL).

The triple therapy extended mouse survival beyond day 55 (P < 0.0001; Fig. 1B). The control, α-MerTK-only, and α-MerTK + α-PD1 conditions did not delay growth of the primary (irradiated) tumor (Fig. 1C). The triple therapy significantly controlled the primary tumors in all mice comparing to control group. The triple therapy led to substantial growth delay at the abscopal site (Fig. 1D) as compared with RT only (P < 0.001) and suppressed the number of lung metastases (P < 0.001; Fig. 1E) in this model.

To explore the mechanism underlying the delay in abscopal tumor growth from the triple therapy, we collected primary and abscopal tumors on day 21 and evaluated them for myeloid cell populations including tumor-associated macrophages (TAM) and granulocyte (G)-MDSCs by flow cytometry. The triple therapy did not affect the number of TAMs in the primary tumors (Fig. 2A), but greatly increased the number of TAMs at the abscopal tumors relative to the radiation-only condition (P = 0.0402; Fig. 2B). Similarly, triple therapy did not affect numbers of G-MDSCs in the primary tumors (Fig. 2C) but greatly decreased the G-MDSCs in the abscopal tumor relative to radiation-only (P = 0.0351; Fig. 2D). Finally, the triple therapy significantly increased the percentage of M1 macrophages at both the primary tumors (P = 0.0279) and the abscopal tumors (P = 0.0256; Fig. 2E and F).

No differences were found among the various treatment groups in percentages of CD4⁺ lymphocytes in the tumor microenvironment at either the primary or abscopal tumor sites (Fig. 3A and B). Similarly, no differences were found in CD8⁺ T cells at the irradiated (primary) tumor sites between the triple therapy and radiation-only conditions (Fig. 3C), although an increase may have been present at the unirradiated (abscopal) tumor sites (P = 0.0639; Fig. 3D).

Next, we quantified activated cytotoxic CD8⁺ T cells and pan-NK cells among TILs isolated from the primary and abscopal tumors by gating those TILs on CD8⁺ Granzyme B⁺ and CD49b⁺ Granzyme B⁺ populations. No changes in CD8⁺ cells were found in the primary (irradiated) tumors in any treatment condition, but the triple therapy led to a significant increase in activated CD8⁺ cells versus RT only group in the abscopal (unirradiated) tumors (P = 0.0072; Fig. 3E and F). Similarly, activated NK cells were only observed in the abscopal tumors with the triple therapy (P = 0.0028; Fig. 3G and H).

Further investigation at the level of gene expression using Immuno-NanoString showed that triple therapy led to significant increases (relative to radiation-only) in Pecam (CD31) and STAT4 expression at the abscopal tumor sites (Fig. 3I). Abscopal tumors also showed increased expression of the IFN regulator factor-1 (IRF-1), which activates the expression of IFNβ (24) and p53 (25), after triple therapy (Fig. 3I). Additionally, there was significant increase in IL12 and TNFα expression with the triple therapy compared with no treatment, although IL6 expression was not significantly modulated (Supplementary Fig. S2).

Next, to evaluate the role of activated CD8⁺ effector T cells and activated NK cells in the response to triple therapy, we depleted CD8⁺ cells, NK cells, or both and analyzed the effects on survival and tumor control at the primary (irradiated) and abscopal (unirradiated) tumor sites. Survival was enhanced in the CD8⁺/NK⁺-proficient condition relative to the CD8⁺-depletion condition (P = 0.0196), the NK-depletion condition (*, P = 0.0074), and the combined CD8⁺- and NK-depleted condition (P = 0.0064; Fig. 4A). Triple therapy without CD8⁺ and NK⁺ depletion also suppressed tumor growth at the primary and abscopal sites (Fig. 4B and C). Finally, depletion of CD8⁺ and NK⁺ cells abrogated the suppressive effect of triple therapy on the number of lung metastases (P = 0.0022; Fig. 4D).

We harvested both primary and abscopal tumors from mice on day 21, isolated TILs, and gated those cells on CD8⁺ CD103⁺ to identify T_RM cells. The triple therapy increased the percentage of CD8⁺ CD103⁺ cells at the primary tumor site relative to the radiation + α-PD1 condition (P = 0.0142; Fig. 5A). Triple therapy also increased the percentage of CD8⁺ CD103⁺ cells at the abscopal tumor site relative to the radiation-only condition (P = 0.0456) and relative to the radiation + α-PD1 condition (P = 0.0663; Fig. 5B).

Our discovery of CD8⁺ CD103⁺ T_RM cells in the tumor microenvironment and their expression of PD1 (Supplementary Fig. S3) led us to hypothesize that the effects of the triple therapy could be enhanced (in terms of extending survival) if we continued the α-PD1 treatment throughout the experiment, rather than stopping at 4 doses, which presumably would protect the T_RM cells from exhaustion by blocking the PD1–PDL1 interaction. Indeed, we found that continuing the α-PD1 after the triple therapy extended survival to day 63 (P < 0.0001) (Fig. 5C).

To confirm our 344SQ-P results, we performed an additional experiment using another cell line (pancreatic cell line PANC-02) in C57BL/6 mice. The mice were inoculated by subcutaneous injection of PANC-02 tumor cells into the right hind leg (to establish primary tumors) and the left hind leg (to establish abscopal tumors). Tumor growth was monitored twice a week with digital calipers. Four α-PD1 doses were given intraperitoneally at 200 μg/dose on days 5, 8, 11, and 14; high-dose radiation (36 Gy in 3 12-Gy fractions) was given to the primary tumors on days 7, 8, and 9; and 6 doses of α-MERTK were given (also intraperitoneally at 200 μg/dose) on days 7, 10, 12, 15, 19, and 22 following the same experimental timeline. The triple therapy extended mouse survival beyond day 60 (P < 0.0001; Fig. 6A). The control, α-MerTK-only, and α-MerTK + α-PD1 conditions did not delay growth on both the primary (irradiated) and abscopal tumor sites (Fig. 6B) whereas the triple therapy significantly inhibited the tumor growth mainly on the abscopal tumor sites (Fig. 6C). We harvested both primary and abscopal tumors from the triple therapy group on day 60, isolated TILs, and subjected the samples to flow cytometry analysis. The triple therapy increased the percentage of CD8⁺ CD103⁺ cells at both tumor sites relative to no treatment condition (Fig. 6D).

Standard therapy for early-stage lung cancer usually involves some form of local therapy, either radiation or surgical resection (1). However, the use of radiation is also being explored for its ability to elicit systemic responses to control tumors outside the radiation fields (i.e., “abscopal effects”; ref. 26). Abscopal responses are rare when radiation is used alone or in combination with chemotherapy, and thus data on survival outcomes for patients whose tumors show abscopal effects are rare as well (27). Our prior preclinical (4) and clinical studies (NCT02402920) indicate that combining high (stereotactic) radiation doses with targeted immune checkpoint blockade such as α-PD1 induces synergetic effects that both control tumor growth at the local level and stimulate a systemic antitumor response by upregulating MHC-I molecules on the tumor cells through the induction of type I IFNs (4).

Our findings from the current study suggest that adding α-MerTK and α-PD1 further augments the adaptive immune-mediated effect of radiation by increasing the percentage of CD8⁺ CD103⁺ T_RM cells at the abscopal tumor sites. These cells are a tumor-specific cytotoxic T lymphocyte subpopulation that accumulates in the tumor microenvironment (28, 29) and express CD103 [αE (CD103) β7] and CD49a [α1 (CD49a) β1] integrins along with the C-type lectin CD69, which collectively help these cells to migrate into tumor islets by binding to E-cadherin, leading ultimately to T-cell receptor–dependent cell killing. Our NanoString molecular analysis results showed that the Pecam (CD31) gene, known to participate in integrin activation (30), was upregulated at the abscopal tumor sites.

On the other hand, radiation is also known to upregulate PDL1 on tumor cells (31). We found here that CD8⁺ CD103⁺ T_RM cells have high expression of PD1, which provides a rationale for adding α-PD1 at the induction phase (before RT), concurrent with RT, and in the maintenance phase (after RT and MerTK blockade). This is consistent with other studies showing that long-term treatment with α-PD1 protects against CD4 and CD8 T-cell exhaustion (32). In our study, we found that continuing α-PD1 as maintenance therapy extended survival time, presumably by preserving the adaptive immune effects and function of the observed T_RM cells (33).

Macrophages expressing MerTK have been shown to engulf apoptotic bodies of dying tumor cells and convert into M2 phenotype (17). MerTK blockade was essential to shift the apoptotic cell death caused by radiation towards secondary necrosis (34) and re-polarize M2 macrophages towards M1 phenotype at both primary and abscopal tumor sites to strengthen antitumor immunogenic cell death. We also noted that α-PD1, in combination with α-MerTK, acts to augment long-lasting adaptive immunity of radiation by increasing the percentage of tumor-specific memory cells at abscopal tumor sites, which were not subject to the long-term suppressive effect of high-dose radiation (35, 36).

We further found that the significant tumor growth delays, improved survival rates, and limitation of metastatic spread resulting from triple therapy is mediated by activated NK and CD8⁺ T lymphocytes, as verified in our depletion studies. The stability of NK-cell and CD8⁺ T-cell activation at the abscopal metastatic tumor sites is possibly related to the presence of CD8⁺ CD103⁺ T_RM cells. Our phenotyping data of abscopal tumors after the triple therapy showed significant increase in percentages of T_RM cells that was associated with tumor response and shrinkage. However, we did not specifically deplete the T_RM population that may reflect direct mechanistic link. Our NanoString data showed significant upregulation in IL12 accompanied with upregulation in STAT4 expression. In agreement with our findings, others have shown that IL12 may act on CD8⁺ CD103⁺ T_RM cells to produce IFNγ via STAT4 transcription, which enhances antigen-specific immunity (37, 38).

The general lack of effect of the triple therapy at the primary tumor site (relative to the abscopal tumor results) seems to validate our previous studies that high-dose radiation upregulates Tregs (CD4⁺ CD25⁺; ref. 39), which are known to suppress CD8⁺ T cells and immune responses at the primary tumor. Our data also showed significant decrease in the percentage of the immunosuppressive G-MDSCs only at the abscopal site, with no effect on the primary tumor site, by adding α-PD1 in combination with α-MerTK to the radiation treatment.

Increasing the percentage of CD8⁺ CD103⁺ T_RM cells would have robust clinical application because the presence of these cells would influence survival among patients with NSCLC. CD8⁺ CD103⁺ T_RM cells reside in tumor islets, where they mediate long-term tumor-specific memory (29); we believe that increasing the numbers of these cells would be of particular benefit for patients who develop recurrence after radiation plus α-PD1 therapy.

We propose the following conceptual framework (Fig. 7) as one possible explanation for the results of this study. First, radiation initiates the tumor-death process while blockade of MerTK shifts the manner of cell death in the tumor microenvironment from tolerogenic (apoptotic) to nontolerogenic (immunogenic; Fig. 7), and α-PD1 sustains the resulting immune effects. Radiotherapy basically primes the CD8⁺ T-effector lymphocytes by binding to the upregulated MHC-I molecules on the tumor surface. The triple of radiation, anti-PD1, and anti-MerTK maximizes this priming and leads to generation of CD8⁺ CD103⁺ T_RM cells in the tumor islets at the abscopal tumor sites. The triple therapy also enhances the activation of NK and CD8⁺ T cells, mainly at the abscopal tumor sites. Finally, continuing the anti-PD1 treatment would prolong outcomes elicited by the radiation and α-MerTK and overcome T_RM-cell exhaustion.

M. Quigley, T.E. Spires, and T.P. Reilly are employees/paid consultants for and hold ownership interest (including patents) in Bristol-Myers Squibb. J.W. Welsh is an employee/paid consultant for Reflexion Medical, Molecular Match, Onco Response, Checkmate Pharmaceuticals, Mavu Pharmaceuticals, Alpine Immune Sciences, AstraZeneca, Merck, Incyte, and Aileron, reports receiving commercial research grants from GlaxoSmithKline, Bristol-Myers Squibb, Merck, Nanobiotix, Mavo Pharma, and Checkmate Pharmaceuticals, and holds ownership interest (including patents) in Helios Oncology, Molecular Match, OncoResponse, Radscopal, and is listed as a co-inventor on patent applications on MP470 and MRX34 regulation of PDL1. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M.S. Caetano, A.I. Younes, T. Spires, T.P. Reilly, T.R. Cushman, J.E. Schoenhals, A. Li, J.W. Welsh

Development of methodology: M.S. Caetano, A.I. Younes, T.P. Reilly, T.R. Cushman, J.W. Welsh

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.S. Caetano, A.I. Younes, T.P. Reilly, T.R. Cushman, T.R. Cushman, J.E. Schoenhals, J.W. Welsh

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.S. Caetano, A.I. Younes, H.B. Barsoumian, M. Quigley, H. Menon, T.P. Reilly, T.R. Cushman, J.W. Welsh

Writing, review, and/or revision of the manuscript: M.S. Caetano, A.I. Younes, H.B. Barsoumian, M. Quigley, H. Menon, C. Gao, T.P. Reilly, A.P. Cadena, T.P. Reilly, T.R. Cushman, A. Li, Q.-N. Nguyen, M. Angelica Cortez, J.W. Welsh

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.S. Caetano, A.I. Younes, M. Quigley, J.W. Welsh

Study supervision: M. Angelica Cortez, J.W. Welsh

Other (illustrations/figures): T.P. Reilly

This work was supported and funded by Bristol-Myers Squibb, and further supported by the family of M. Adnan Hamed, the Susan and Peter Goodwin Foundation, the Orr Family Foundation (to MD Anderson Cancer Center's Thoracic Radiation Oncology program). We also have support from the MD Anderson Knowledge Gap award, Doctors Cancer Foundation Grant, The Lung Cancer Research Foundation, Cancer Center Support (Core) Grant CA016672 from the NCI (to The University of Texas MD Anderson Cancer Center). We would like to thank Christine F. Wogan for reviewing and editing the manuscript.

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