Treatment of tumors with cytokines to boost antitumor immune responses is limited in the clinic due to systemic adverse events. Intratumoral delivery of cytokines may overcome these issues and improve therapeutic efficacy. Hotz and Wagenaar et al. thoroughly investigated the efficacy of local delivery of mRNA encoding a mixture of cytokines alone or combined with checkpoint blockade in murine models to take advantage of these powerful immune stimulants. Their work was published in Science Translational Medicine.

Based on in vivo screens, the researchers systematically established a cocktail of saline-formulated mRNAs encoding four murine cytokines: IL-12 single chain, granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-15 sushi, and IFN-α4. Local administration of these mRNAs resulted in cytokine expression in the tumor microenvironment (TME), with limited expression in other organs, suggesting localized translation. In the TME, both CD45⁺ and CD45^- cells were able to take up and translate the mRNA.

Repeated intratumoral injection resulted in tumor growth control, with complete regression (CR) in the majority of B16F10 and CT26 tumors, resulting in long-term survival. The cytokine expression also resulted in the induction of IFNγ and IP-10. To investigate the mechanistic contribution of each cytokine in the mixture, the researchers performed experiments with bone marrow cells and splenocytes that were exposed to the four cytokines and mixtures in which one of the cytokines was left out. When IL-15 sushi or IFN-α4 were missing, CD8⁺ T cells, NK cells, and DCs expressed lower levels of activation markers, and when GM-CSF or IL-12sc were missing there was less induction of IFNγ. Therapeutic responses were reduced most when IL-12sc was omitted, but all four were necessary for maximal effects. This was also shown in vivo, where all three-component mixtures of mRNAs resulted in varying levels of CT26 tumor control, with removal of IL-12sc mRNA having the most pronounced effect. In the B16F10 model, some of the CRs developed localized vitiligo, suggesting induction of T cell responses toward shared tumoral and melanocyte antigens. In addition, mice with CR were resistant to tumor rechallenge, suggesting the development of immunologic memory. This was further confirmed in immune cell depletion experiments, which showed that the antitumor activity was dependent on the presence of NK cells and CD8⁺ and CD4⁺ T cells.

Post-treatment, there was a higher infiltration of CD4⁺ and CD8⁺ T cells in the TME, and these TIL expressed higher levels of granzyme B, TNFα, and IFNγ. RNAseq of treated tumors showed upregulation of proinflammatory signatures, including the IFNγ and TNFα signaling pathways. In addition, cytokine mRNA treatment resulted in PD-L1 upregulation on immune and non-immune cells in the TME, while levels of PD-1 on CD8⁺ T cells remained unaffected. Further confirming the essential role of the IFNγ pathway in the therapeutic effects, IFNγ knockout mice with B16F10 tumors did not respond to therapy.

Hotz and Wagenaar et al. then investigated how the cytokines may affect T cell priming by using a B16F10-ovalbumin (OVA) model together with OVA-specific OT-I CD8⁺ T cells. Treatment with the mRNA mixture resulted in higher OT-I proliferation in the tumor-draining lymph nodes. Similar results were found in a CT26 model expressing the antigen gp70, where treatment resulted in the expansion of peripheral and tumoral gp70-specific T cells. Rechallenge experiments with CT26 tumors lacking gp70 revealed that immune memory directed against dominant and subdominant antigens was formed.

Treatment of subcutaneous B16F10 tumors also suppressed the growth of lung metastases, resulting in reduced lung tumor burden. In these metastatic lesions, NK cells, but not CD8⁺ T cells, upregulated CD69 and KLRG1. In a B16F10 model with subcutaneous tumors on both flanks, treatment in one flank was effective for both tumors, resulting in longer overall survival, suggesting local treatment results in systemic antitumor responses.

The researchers then investigated whether the combination of the mRNA treatment protocol with checkpoint blockade could further enhance efficacy. Although anti-CTLA-4 alone showed efficacy in this model, adding an anti-CTLA-4 antibody to a single intratumoral injection of the cytokine mRNA mixture in the CT26 and B16F10 models resulted in improved antitumor activity and survival. When mice with larger tumors were treated, there were more complete responses after combination treatment than after single treatment with mRNA. When PD-1 blockade was combined with the mRNA therapy in the B16F10 and MC38 models with tumors at both flanks or the B16F10 metastasis model, survival increased; in the metastasis model, 40% of mice experienced CR.

Given that many tumors are resistant to checkpoint inhibition, therapy efficacy was also assessed in a model in which β2M was knocked out of MC38 cells by CRISPR to model antigen presentation loss. These mice did not respond to anti-PD-1 single treatment. Injection of the cytokine mRNA mixture prolonged survival, but no additional effect in efficacy was observed when it was combined with checkpoint blockade. To model heterogeneous tumors, mice were inoculated with the MC38 β2M knockout cells on one flank, and MC38 WT tumor cells on the other flank. Anti-PD-1 alone was ineffective in this model. The knockout tumors were injected with the mRNA mixture, which temporarily improved outcomes, but the combination with anti-PD-1 further improved survival, suggesting abscopal effects with the combination.

Finally, the researchers prepared human mRNAs for the four cytokines to inject into a human A375 melanoma xenograft in SCID mice. All four cytokines were expressed in the 24 hours after injection, after which levels lowered to baseline at 72 hours post-injection. In vitro experiments using PBMCs from patients with advanced-stage cancer revealed that treatment with supernatants from individual cytokine mRNA-transfected cells induced the release of high levels of IFNγ. When combined with anti-CD3 stimulation, treatment resulted in increased CD8⁺ T cell proliferation in 3 out of 4 samples and CD4⁺ T cell proliferation in 2 out of 4 samples.

Overall, local intratumoral injection of a saline-formulated mRNA mixture encoding 4 distinct cytokines could improve the immune environment in tumors – an effect that was further enhanced by combining this strategy with checkpoint blockade. Clinical studies with this mRNA mixture are currently underway.

Write-up by Maartje Wouters, image by Lauren Hitchings