Multiple lines of evidence indicate a critical role of antigen cross-presentation by conventional BATF3-dependent type 1 classical dendritic cells (cDC1) in CD8-mediated antitumor immunity. Flt3L and XCL1, respectively, constitute a key growth/differentiation factor and a potent and specific chemoattractant for cDC1. To exploit their antitumor functions in local immunotherapy, we prepared Semliki Forest Virus (SFV)–based vectors encoding XCL1 and soluble Flt3L (sFlt3L). These vectors readily conferred transgene expression to the tumor cells in culture and when engrafted as subcutaneous mouse tumor models. In syngeneic mice, intratumoral injection of SFV-XCL1-sFlt3L (SFV-XF) delayed progression of MC38- and B16-derived tumors. Therapeutic activity was observed and exerted additive effects in combination with anti–PD-1, anti-CD137, or CTLA-4 immunostimulatory mAbs. Therapeutic effects were abolished by CD8β T-cell depletion and were enhanced by CD4 T-cell depletion, but not by T regulatory cell predepletion with anti-CD25 mAb. Antitumor effects were also abolished in BATF3- and IFNAR-deficient mice. In B16-OVA tumors, SFV-XF increased the number of infiltrating CD8 T cells, including those recognizing OVA. Consistently, following the intratumoral SFV-XF treatment courses, we observed increased BATF3-dependent cDC1 among B16-OVA tumor-infiltrating leukocytes. Such an intratumoral increase was not seen in MC38-derived tumors, but both resident and migratory cDC1 were boosted in SFV-XF–treated MC38 tumor-draining lymph nodes. In conclusion, viral gene transfer of sFlt3L and XCL1 is feasible, safe, and biologically active in mice, exerting antitumor effects that can be potentiated by CD4 T-cell depletion.
These findings demonstrate that transgenic expression of sFLT3L and XCL1 in tumor cells mediates cross-priming of, and elicits potent antitumor activity from, CD8 T lymphocytes, particularly in combination with CD4 T-cell depletion.
Cancer immunotherapy is in the limelight of oncology therapeutics due to the efficacy of systemic administration of checkpoint inhibitors and chimeric antigen receptor–transduced T cells (1). Intratumoral approaches with immunotherapy agents are feasible (2), and include local administration of Toll-like receptor or STING agonists (3, 4) and recombinant oncolytic viruses (5) or viral vectors (6). Most immunotherapy approaches necessarily rely on the activation of CD8 T lymphocytes by mature dendritic cells (DC) presenting cognate tumor antigens (7). A subset of DCs dependent on the transcription factors BATF3 and IRF8 for their ontogeny is critical for the activation of CD8 T lymphocytes (8, 9) and crucial for the antitumor efficacy of treatment with anti-PD1 and anti-CD137 mAbs in mouse models (10). BATF3-dependent DCs are also termed type 1 conventional DCs (cDC1) and excel in uptaking antigens from dead cells and presenting their peptides on MHC-I molecules (cross-presentation), leading to the activation/expansion of specific CTLs (cross-priming). Two subsets of mouse cDC1 have been identified. One of these resides in T-cell zones of lymphoid organs (CD11c+CD8α+CD103−Clec9a+; ref. 11) and the other (CD11c+CD8α−CD103+Clec9a+) is deployed in peripheral tissues and migrates toward lymphoid tissue once activated (7, 12). Migratory CD103+ cDC1s have been observed to carry tumor antigen to tumor-draining lymph nodes (TDLN) for cross-presentation (10, 13, 14). Flt3L is a critical growth/differentiation factor for this DC subpopulation (15) and XCL1 a chemokine that chemoattracts cDC1s, which exclusively express the XCL1 receptor (XCR1; ref. 16) to allow for cDC1 rendezvous with natural killer (NK) and CD8 T cells (17, 18). cDC1s are endowed with abundant TLR3 expression that drives their activation/maturation once challenged with dsRNA denoting viral infection (19).
Local gene transfer into experimental tumors with Semliki Forest Virus (SFV)-derived vectors is feasible and has an attractive immunotherapeutic potential. Although SFV vectors are not replication-competent viruses, they induce catastrophic death of infected cells (20), release abundant viral dsRNA (21), induce local IFNα/β production (21), and are safe. Indeed, a vector encoding IL12 (SFV-IL12) is highly efficacious in murine (22) and woodchuck (23) models of cancer and synergizes with other immunotherapies such as treatment with anti–PD-1 (24) and anti-CD137 (25) immunomodulatory mAbs.
Transfection of sFlt3L (26) or XCL1 (27) into tumor cells has been previously tested in culture and in vivo with immunotherapy purposes, achieving excellent vaccination effects in the case of sFlt3L (26).
In this study, repeated injections of an SFV vector simultaneously expressing sFlt3L and XCL1 were tested in an attempt to attract and expand cDC1 cells, while killing a fraction of tumor cells and providing viral RNA–mediated activation of innate immunity (28). Partial antitumor activity was substantiated against transplantable established tumors. This antitumor effect was dependent on CD8 T cells and on the integrity of the BATF3 and IFNAR genes in tumor-bearing mice.
Materials and Methods
Cell lines and culture conditions
MC38 cells were a kind gift from Dr. Karl E. Hellström (University of Washington, Seattle, WA) in September 1998. B16-OVA cells were provided by Dr. Lieping Chen (Yale University, New Haven, CT) in November 2001. B16F10 cells were purchased from the ATCC in June 2006. CT26 cells were purchased from ATCC in 2011. These cell lines were authenticated by Idexx Radil (Case 6592-2012) in February 2012. Panc02 tumor tissue was obtained from the NCI DCTDC Tumor Repository (Frederick, MD). A cell line was isolated from trypsinization of Panc02-grafted tumor tissue (29). MC38, PANC02, CT26, B16F10, and B16-OVA cells were cultured in RPMI medium (Gibco) supplemented with 10% decomplemented and filtered FBS (Sigma-Aldrich), containing 50 μmol/L β-mercaptoethanol, 100 U/mL penicillin, and 100 μg/mL streptomycin (all from Gibco). Baby hamster kidney (BHK) cells were cultured in GMEM–BHK21 medium (Gibco) supplemented with 5% decomplemented and filtered FBS (Sigma-Aldrich), containing 20 mmol/L Hepes (Invitrogen), 10% tryptose phosphate broth, 2 mmol/L glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin (all from Gibco). When indicated, BHK cells were cultured in Chinese hamster ovary (CHO) medium (Sigma) supplemented with the same components as indicated for BHK, except the FBS. For infection, cells were incubated in Minimum Essential Medium (Gibco) containing 0.2% BSA (Sigma).
Functional assays for transgene products
BHK cells were infected with SFV vectors at a multiplicity of infection of 10 as described above and incubated overnight in serum-free CHO medium (Sigma) for XCL1 bioactivity testing or GMEM BHK-21 (Gibco) for Flt3L bioactivity testing. Supernatants were collected and kept frozen until use. For Flt3L testing, bone marrow (BM) cell suspensions were flushed out of hind limb bones and cultured in RPMI medium conditioned with 20% infected BHK-derived supernatants. After 9 days, classical BM-DC (CD11c+CD11b+) and plasmacytoid BM-DC (CD11c+CD11b− B220+) cells were assessed by flow cytometry to demonstrate sFlt3L-dependent differentiation. For XCL1 testing, standard transwell chemotaxis assays were performed on iCD103 BM-DCs (30). A total of 1 × 105 iCD103 cells were suspended in serum-free CHO medium and plated onto 5-μm Transwell inserts (Costar). Cells were allowed to migrate for 4 hours toward infected BHK-derived supernatants and the total number of cells in the lower well was quantitated by flow cytometry.
Mice and in vivo tumor experiments
Experiments involving mice were carried out in the animal facility of CIMA (Pamplona, Spain) under study approvals 150/12 and 082/16 from the University of Navarra Ethics Committee. C57Bl/6 Batf3tm1Kmm/J [Batf3 knockout (KO); ref. 8], Tmem173gt/J (STING KO; ref. 31), and IFNα/bRo/o (IFNAR KO; ref. 32) mice were bred at CIMA in specific pathogen-free conditions. C57Bl/6 mice were obtained from Envigo. Batf3 KO, STING KO, and IFNAR KO mice were kindly provided, respectively, by Dr. Kenneth M. Murphy (Washington University, St. Louis, MO), by Dr. Gloria González Aseguinolaza (CIMA, Pamplona, Spain), and by Dr. Matthew Albert (Institut Pasteur, Paris, France). Cultured tumor cells were cultured and trypsinized for injection before reaching confluence. A total of 5 × 105 MC38 or B16-OVA cells were injected subcutaneously in 50 μL PBS into the right flank of 6- to 12-week-old mice. SFV viral particles (VP) were diluted in PBS and kept ice-cold until administration. Intratumoral injection of 50 μL suspension containing 1 × 108 VPs or vehicle control was performed using 29G syringes and under inhalatory anesthesia. When indicated, 100 μg anti-CD137 (1D8) or anti–PD-1 (RMP1-14) were administered intraperitoneally in PBS. Depletion of lymphocyte subsets was performed by intraperitoneal injection of anti-CD4 (GK1.5, Bioxcell), anti-CD8 (H35-17.2, in-house), or anti-NK1.1 (PK136, in-house) mAbs. Two-hundred μg of each mAb was injected 2 days before SFV administration; 100 μg on SFV treatments days and 3 days after the last SFV administration. A single dose of 300-μg anti–CD25-Rat IgG (PC61, in-house) or 200-μg anti-CD25-mIgG2a (33) was administered 2 days before SFV administration. Depletions were verified by peripheral blood flow cytometry staining. One-hundred μg of p60 peptide (34) was administered intraperitoneally daily for 10 days, starting 2 days before SFV administration. Tumor area was measured twice weekly and calculated as the product of orthogonal diameters.
Additional materials and methods
Construction of SFV-derived vectors, mRNA quantitative analysis, Western blotting, tissue processing and flow cytometry, and software and statistical analyses are detailed in Supplementary Materials and Methods.
Characterization of SFV-derived vectors encoding sFLt3L and XCL1
Nonreplicative SFV vectors were constructed by replacing the viral structural proteins with the mouse sequences of XCL1 or sFlt3L, generating vectors SFV-XCL1 and SFV-sFlt3L, respectively (Fig. 1A). An SFV vector expressing β-galactosidase encoded by LacZ gene (SFV-LacZ) was used for control purposes. An SFV vector encoding both XCL1 and sFlt3L as a single open reading frame (ORF) was made by placing a 2A cis-protease sequence to permit posttranslational efficient proteolytic separation of both transgene products. A furin cleavage site was also inserted to eliminate the remaining 2A target sequence from XCL1. Three cell lines were infected in culture with the different SFV vectors and qRT-PCR detected strong transcription of the transgenes (Fig. 1B). Moreover, gene expression was readily detected in subcutaneous MC38-derived tumors excised 24 hours postintratumoral injection of the corresponding SFV vectors (Fig. 1C). Of note, both in vitro and in vivo, the vector expressing the two transgenes showed comparatively lower quantities of each transgene mRNA as compared with the single-gene SFV vectors, indicating less efficient expression in the double-transgene vector. Translation was confirmed by analyzing tissue culture cell lysates of 24-hour–infected BHK cells by Western blot analysis (Fig. 1D). The differences in the sizes of the detected proteins encoded by the single-transgene and double-transgene vectors are due to the presence of a C-terminal myc tag from the XCL1 parental expression plasmid. Because of the cloning strategy used, the tag is present in the C-terminus of the XCL1 protein from SFV-XCL1 and of the sFlt3L protein from SFV-XF, thus slightly modifying their detected molecular weights in the Western blot analysis.
Next, we examined the functionality of the expressed transgenes (Fig. 1E). For this purpose, we analyzed the chemotactic activity of XCL1 from the tissue culture supernatants of SFV-infected BHK cells on iCD103 DCs derived in culture from bone marrow precursors as described previously (Fig. 1F; ref. 30). sFlt3L bioactivity was assessed by studying the ability of infected BHK culture supernatants to promote the differentiation of bone marrow cell suspensions into the conventional and plasmacytoid DCs (cDCs and pDCs; Fig. 1G). In both instances, transgene products appeared to be fully functional.
Antitumor activity of SFV vectors encoding sFlt3L and/or XCL1
To study the antitumor effects of the constructed SFV vectors, a single injection of 1 × 108 VPs was given into day 8–established MC38 subcutaneous tumors (Fig. 2A). A certain degree of tumor growth retardation was observed with all sFlt3L-containing SFV vectors, but it was more prominent with the vector encoding both XCL1 and sFLt3L (SFV-XF). To enhance the antitumor effects, three doses of vectors were given every 2 days starting at day 8 after tumor cell inoculation. Again, MC38 tumors were more efficiently delayed in their growth by the SFV-XF vector (Supplementary fig. S1A). In a series of experiments represented in Fig. 2B and C, evident tumor growth delays were achieved by repeated intratumoral administration of SFV-XF into the established MC38 (Fig. 2B) and B16-OVA (Fig. 2C) tumors. This treatment resulted in survival prolongation in both models but seldom in tumor eradication. Treatment of subcutaneous B16F10-derived melanomas, Panc02-derived pancreatic carcinomas, and CT26-derived colon carcinomas with three viral doses also showed the therapeutic effects of SFV-XF (Supplementary fig. S1B-D) over saline control and SFV-LacZ, although SFV-LacZ showed some degree of activity on some of the transplanted tumor models.
Given the clinical success of immunomodulatory mAbs, we explored whether local SFV therapeutic activity could be potentiated by its combination with systemic antagonist anti–PD-1 or anti–CTLA-4 mAb, as well as with agonist anti-CD137 mAbs. As shown in Fig. 3, although the anti-CD137 mAb was able to delay tumor growth in both models, anti–PD-1 was almost ineffective against either model (Fig. 3A and B). Anti–CTLA-4 monotherapy was effective against MC38 tumors, but not against B16-OVA tumors. In combination with a single dose of SFV-XF, systemic anti-CD137 mAb enhanced efficacy against MC38 and B16-OVA tumors, whereas systemic anti–PD-1 enhanced activity only against B16-OVA. Anti–CTLA-4 mAb effects were potentiated both by SFV-XF and SFV-LacZ against B16-OVA (Fig. 3).
Antitumor activity of SFV-XF was dependent on CD8 T cells but markedly enhanced by CD4 T-cell depletion
To study the cellular requirements for the activity of SFV-XF, selective depletion of T-cell subsets and NK1.1+ NK and NKT cells were performed prior to treatment in MC38 tumor-bearing mice. As shown in Fig. 4A, depletion of CD8β cells abolished therapeutic activity, whereas CD4 and NK1.1 depletion enhanced the therapeutic effects, leading to extended survival. This result indicates that the antitumor effect mediated by SFV-XF is mainly mediated by CD8+ T cells.
One interpretation of the enhanced antitumor activity following CD4 depletion is the ensuing elimination of CD4+ Tregs; however, predepletion of Tregs with an anti-CD25 mAb (35) or inhibition of Foxp3 with an antagonist peptide (34) did not enhance the therapeutic effects (Supplementary Fig. S2A). In contrast, CD4 T-cell depletion gave rise to 4 of 5 mice eradicating their tumors upon intratumoral treatment with SFV-XF in this experiment. Moreover, when an anti-CD25 mouse IgG2a that optimally depletes intratumoral Tregs was used, it did not show therapeutic synergy with SFV-XF against MC38 tumors, even though it exerted strong activity by itself (Supplementary Fig. S2B). Completeness of depletions in peripheral blood was checked by flow cytometry (Supplementary Fig. S2C).
Because Treg depletion did not appear to be the mechanistic explanation, we focused on the effector CD4 T-cell cytokine production profile. In experiments shown in Supplementary Fig. S3, B16-OVA–bearing mice were adoptively transferred with TCR-transgenic OVA-reactive OT-II cells on day 7 and one day later intratumorally treated with SFV-XF, SFV-LacZ, or saline. Intracellular cytokine staining for IFNγ, IL10, and IL17 in endogenous (Supplementary Fig. S3A) and adoptively transferred (Supplementary Fig. S3B) CD4 T cells from TDLNs was interrogated following an ex vivo 4-hour stimulation with PMA + ionomycin. Both in endogenous CD4 and OT-II T cells, a decrease of IFNγ- and IL17-producing fractions was observed, whereas there were increases of IL10 in the OT-II cells. These data suggest hampered Th1 differentiation upon SFV vector treatment potentially leading to deleterious effects of non-Treg CD4 subsets.
In mice bilaterally engrafted with MC38 tumors, SFV-XF treatment in the context of CD4 T-cell, but not NK1.1 depletion, undoubtedly delayed the growth of the concomitant distant noninjected tumors (Fig. 4B and C). Similar effects on the contralateral tumors were observed in the B16-OVA model (Supplementary Fig. S4A and S4B). SFV-XF as a single agent did not have therapeutic effects on distant tumors, even though a trend for delay of tumor growth was observed in some of the experiments (Fig. 4C; Supplementary Fig. S4B).
SFV-XF increases effector T-cell infiltration of the B16-OVA tumor microenvironment
Upon treatment with SFV-XF of B16-OVA–derived tumors, there was an increase of CD4 and CD8 T-cell content in the tumor microenvironment detectable in directly treated tumors, but also in concomitant nondirectly injected contralateral lesions (Fig 5A). In these B16-OVA tumors, we observed a rapid increase in the number of H-2Kb-tetramer–positive CD8 T cells recognizing the OVA-specific SIINFEKL epitope contingent on SFV-XF treatment (Fig. 5B). However, these effects in the tumor microenvironment were not found under similar conditions of treatment in MC38-derived tumors (Supplementary Fig. S5A and S5B), indicating model-inherent differences in spite of the fact that both models respond to treatment.
Interestingly, in bilaterally engrafted B16-OVA-bearing mice in which CD4 depletion was induced 48 hours prior to intratumoral SFV vector injection to one of the tumors, there were more prominent bilateral increases in the content of antigen-specific CD8 cells among tumor-infiltrating T cells (Fig. 5B). CD4 depletion also markedly increased bilaterally the percentage of proliferating Ki67+ CD8 T cells (Fig. 5C). In the MC38 bilateral model, intratumoral percentage of proliferating Ki67+ CD8 T cells also increased in response to the CD4 depletion and SFV-XF (Supplementary Fig. S5C). These results indicate increases in tumor-reactive CTLs consistent with the results of CD8 depletion experiments and support that CD4 depletion further potentiates tumor-specific CD8 T cells in the context of intratumoral SFV-XF treatment. In light of the effects of SFV-XF in combination with immunostimulatory mAbs in B16-OVA and MC38 bilaterally engrafted mice, we analyzed the percentage of expression of CD137, CTLA-4, PD-1, and LAG-3 on CD8 and non-Treg CD4 tumor-infiltrating T cells following intratumoral treatment with SFV vectors. In the case of CD8 T cells, expression was also monitored upon concomitant CD4 depletion. As shown in Supplementary Fig. S6A, CD8 T cells expressing CTLA-4, PD-1, and LAG-3 increased in the treated tumor with moderate increases of CD137 expression. Percentages of cells expressing these checkpoints were further increased by CD4 codepletion. Percentages of PD-1, CTLA-4, and LAG-3 expressing CD4 T cells were also increased by SFV-XF treatment (Supplementary Fig. S6B). Again, such changes were not observed in MC38-bearing mice (Supplementary Fig. S6C and S6D). FACS-gating strategies for tumor-infiltrating T-cell studies are shown in Supplementary Fig. S7A. At least in the B16-OVA model, additive effects of systemic immunostimulatory mAbs with intratumoral SFV-XF could be explained in part by enhanced expression of the coinhibitory or costimulatory T-cell targets.
SFV-XF therapeutic activity is contingent on BATF3-dependent DC integrity and causes cDC1 accumulation in TDLNs
Experiments were performed in mice deficient in BATF3, which are virtually devoid of cDC1s (8). In these animals, the antitumor effects of SFV-XF seen in wild-type (WT) control mice were completely lost (Fig. 6A and B). The integrity of the type-I IFN(IFN-I) system is required for the function of BATF3-dependent DCs (36) and for CD8 immunity (37). As seen in Fig. 6A, efficacy was also lost, when treatment was given to Ifnar−/− mice. However, tumor growth delay was preserved to some degree in STING KO mice, indicating at least partial independence of our therapy of the cGAS–STING pathway.
Given the activity of the SFV-encoded transgenes, we expected tumors to become infiltrated by cDC1s, a feature reported to correlate with better prognosis in human cancer (38, 39). In B16-OVA, tumor increases not only in cDC1, but also in cDC2, were substantiated in terms of cells per mg of the tissue (Fig. 7A). However, as seen in Fig. 7B, in MC38-derived tumors, DC infiltrates did not significantly change following three intratumoral doses of SFV-XF over control or SFV-LacZ. In contrast, harvested TDLNs from SFV-XF–treated MC38 tumors showed marked increases in absolute numbers of both migratory (CD11c+IAbhiCD103+CD11b−) and resident (CD11chiIAb+CD8α+CD11b−) cDC1 cells (Fig. 7C). In addition, there was a detectable increase in CD11b+ cDC2 cells (Fig. 7B). FACS-gating strategies for analysis are shown in Supplementary Fig. S7B and S7C.
In conclusion, dependency on BATF3 and the increase of cross-presenting DCs in TDLNs are consistent with the immunotherapeutic activity of XCL1 and sFlt3L as SFV-encoded transgenes.
In this study, SFV vectors engineered to increase cross-priming of tumor antigens were tested following intratumoral injection. Although all SFV constructions encoding sFlt3L delayed the tumor growth, the combination of the chemokine XCL1 and sFlt3L showed more marked antitumor effects on five different transplantable mouse models. Bioactivity of both transgenes was confirmed using the supernatant of infected cells in culture.
Intratumoral injection of viral vectors including HSV (40), measles virus (41), vaccinia virus (42), vesicular stomatitis virus (43), and reovirus (44) is gaining momentum in tumor immunotherapy (6). Their intratumoral administration frequently leads to meaningful therapeutic effects, particularly when combined with anti–CTLA-4 or anti–PD-1 checkpoint inhibitors (5, 45). In the case of alphavirus vectors, an SFV virus encoding IL12 exerts potent antitumor effects dependent on CD8 T-cell antitumor immunity (22). SFV-XF was therapeutically less potent than an SFV vector encoding IL12, although it has the advantage that uncontrolled production of IL12 might have safety problems, as reported in human patients systemically given the recombinant protein (46). In this regard, Flt3L recombinant protein is reportedly safe in humans following repeated subcutaneous administration (47).
The original objective of the SFV-XF vector was to enhance tumor antigen cross-presentation by means of attracting and differentiating cDC1s and thereby enhancing CD8 T-cell cross-priming. Indeed, the SFV-XF–encoded transgenes exert these effects on cells in culture. We had previously shown two important features of SFV-based local immunotherapy: (i) it provides abundant viral RNA that enhances TLR3 and helicase-dependent innate signals (28), and (ii) it enhances local IFNα/β through these mechanisms (28). These two effects, in conjunction with a more prominent cDC1 infiltrate and function, should prime and sustain the cellular antitumor immunity. In this context, it was surprising that SFV-XF showed a rather modest antitumor activity, although most tumors were delayed in their growth after treatment. Additive effects were observed upon combination with anti-CD137, anti-PD1, and anti–CTLA-4 immunostimulatory mAbs that probably involve some degree of induced expression of their lymphocyte targets on tumor-reactive T cells. Of note, intratumoral SFV-IL12 is reportedly highly synergistic with these immunomodulatory antibodies (24, 25).
Experiments upon depletion of CD8 T cells were consistent with a necessary involvement of CTLs in the antitumor effects. Surprisingly, CD4 T-cell depletion and NK/NKT depletion gave rise to enhanced therapeutic activity. Furthermore, we have used a mIgG2a version of the anti-CD25 mAb to ensure intratumoral Treg depletion and this treatment did not potentiate SFV-XF antitumor activity, although this antibody had a high antitumor effect by itself (33). Having ruled out a simple explanation based on the elimination of Tregs by CD25 depletion and Foxp3 inhibition, our next hypothesis was that lymphopenia, secondary to CD4 depletion, augmented the availability of homeostatic cytokines such as IL7 or IL15 for CD8 T cells. However, we were unable to detect increased circulating levels of these cytokines following depletion. It is of note that the CD4+Foxp3− and CD4+Foxp3+ infiltrate increases 48 hours following SFV-XF intratumoral administration (Fig. 5A), potentially affecting locally the therapeutic actions of the recombinant virus.
Next, we examined whether the SFV-XF intratumoral treatment altered the cytokine production profile of effector CD4 T cells in TDLNs. These experiments showed an SFV-induced reduction in the number of CD4 T cells able to produce IFNγ and an increase of those producing IL10 (Supplementary Fig. S3). In this context, CD4 depletion is likely to be counteracting this deleterious effect for antitumor activity. Observations on tumor-infiltrating CD8 T cells in the B16-OVA tumor model are consistent with a relief of an inhibitory mechanism by CD4 depletion.
The mechanistic interplay of NK and NKT cells to dampen the efficacy of SFV-XF remains to be elucidated, although some reports suggest an inhibitory activity of NK cells on recently activated CD8 T-cell blasts (48, 49).
SFV-XF intratumoral treatment in the bilateral B16-OVA tumor model gave rise to marked increases in activated and tumor-specific CD8 T cells taking place both in the directly treated tumor and in the noninjected contralateral side. Indeed, such increases were markedly enhanced by CD4 concomitant depletion. In spite of therapeutic activity, such tumor microenvironment observations were not reproduced in the MC38 model, a fact that could reflect the intrinsic differences between the models or differences in the time course of T-cell infiltration changes.
In keeping with the function of the XCL1 and sFlt3L transgenes, antitumor effects were contingent on BATF3-dependent DCs. Notably, we observe an increase in such DCs in the tumor microenvironment following the SFV-XF intratumoral administration to B16-OVA-derived tumors. Increased numbers of DC included not only cDC1 but also cDC2. It remains to be seen whether such increased intratumoral DC subsets are properly performing tumor–antigen presentation. Intratumoral DCs were not increased in the MC38 model, but there were marked increases found in TDLNs that were also seen, to a much lesser extent, in TDLNs from nontreated distant subcutaneous tumors obtained in the same way, probably due to an accumulation of the circulating Flt3L after repeated SFV-XF administrations. Such increased cDC1 cells belonged to both resident and migratory phenotypes, suggesting that perhaps part of these cDC1 cells seen in TDLNs might have been in the tumor tissue at some earlier time points. In fact, innate effects of SFV RNA may induce maturation and rapid migration of cDC1s out of tumor tissue to TDLNs, thus explaining that cDC1 numbers are not increased in the tumor microenvironment of MC38 tumors. Further enhancement of cDC1 content in the tumor microenvironment warrants future research, for instance using other chemokines known to attract this DC subset, such as CCL4 (50). However, in two recent articles, XCL1 and FLT3L as produced by NK cells were found to dictate cDC1 presence in the tumor microenvironment (18, 51).
The striking therapeutic effect of SFV-XF combination with CD4 depletion, which led to a certain degree of efficacy against distant tumors (Fig. 4B; Supplementary Fig. S4B), is difficult to translate into the clinic, because CD4 depletion is highly immunosuppressive and in practice could only be induced transiently. CD4 T-cell immunity is complex and encompasses both antitumor and protumor activities. Transplanted tumors in mice, as opposed to human malignancies, grow fast in the two weeks following tumor cell inoculation and the mechanism of action of SFV-XF, relying on cross-priming, might take longer to properly begin. In fact, DC numbers kept increasing in TDLNs from the treated mice over time. Little is known about the functional interplay of CD4 T cells and cDC1s, and our results call for an in-depth study.
All in all, our results indicate interesting immunobiological effects of SFV-mediated XCL1 and sFlt3L local gene transfer into tumors that might find suitable combination partners for effective cancer immunotherapy. The strategy is of much interest due to its effects on antigen-presenting cells specialized in CD8 T-cell cross-priming.
Disclosure of Potential Conflicts of Interest
I. Melero reports receiving a commercial research grant from BMS, Bioncotech, Alligator, and Roche; has received speakers bureau honoraria from MSD; and is a consultant/advisory board member for BMS, Roche, Genmab, F-Star, Bioncotech, Bayer, Alligator, and Merck Serono. No potential conflicts of interest were disclosed by the other authors.
Conception and design: C. Smerdou, I. Melero
Development of methodology: J.I. Quetglas, M.E. Rodriguez-Ruiz, A. Sánchez-Arráez, I. Melero
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.R. Sánchez-Paulete, A. Sánchez-Arráez, E. Bolaños, M.C. Ballesteros-Briones
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.R. Sánchez-Paulete, A. Teijeira, A. Sánchez-Arráez, S. Labiano, P. Berraondo, I. Melero
Writing, review, and/or revision of the manuscript: A.R. Sánchez-Paulete, A. Teijeira, M.E. Rodriguez-Ruiz, I. Etxeberria, P. Berraondo, D. Sancho, C. Smerdou, I. Melero
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.I. Quetglas, A. Azpilikueta, E. Bolaños, N. Casares, S.A. Quezada
Study supervision: C. Smerdou, I. Melero
We are in debt to Eneko Elizalde for excellent animal facility care. Critical reading and suggestions by Drs. Sandra Hervás, Carmen Ochoa, Jose L. Perez-Gracia, Ana Rouzaut, and Juan José Lasarte are also much appreciated. This work was supported by grants MINECO SAF 2014-52361-R and SAF 2017-83267-C2-1R and Cancer Research Institute (CRI) CLIP Grant 2017 and the Horizon 2020 Programme from the European Comission (PROCROP project, ref. #635122) to I. Melero; FIS (PI14/01442 and PI17/01859 (to C. Smerdou); and Fundación Echébano fellowship (to M.C. Ballesteros-Briones).
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.