Capitalizing on the messenger: Intra-tumoral delivery of RNA with a systemic effect

Abstract

As reported recently in Science Translational Medicine, intra-tumoral administration of mRNAs encoding 4 cytokines formulated in saline mediates regression of injected and distant murine tumors. The anti-tumor effect, also seen in the anti-PD-1 resistant setting and enhanced by checkpoint blockade, is mediated by tumor-specific T cells and Natural Killer cells.

Intralesional cancer therapy aims to localize immune stimulating agents within tumors, thus minimizing systemic toxicity. Given the systemic nature of cancer, a main goal of this local immunotherapy should arguably also be the induction of a systemic anti-tumor response thereby treating distant tumor metastases. This in situ vaccination, by harnessing the proximity of the immune stimulus and the tumor - and hence, its antigens - has the potential of inducing and expanding anti-tumor effector T cell responses, thereby leading to durable anti-tumor immunity. Several intralesional therapies have been tested in the clinic including Toll-Like Receptor (TLR) Agonists, oncolytic viral agents, and inflammatory cytokines. These locally administered approaches can induce systemic anti-tumor immunity as evident by the observation of increased CD8⁺ T cell frequencies associated with clinical responses and objective responses in non-injected metastases (Ribas et al., 2017, Andtbacka et al., 2015, Ribas et al., 2021). However, early evidence from clinical trials and the experience with the oncolytic virus talimogen laherparepvec in patients with advanced melanoma and other cancers indicate that the systemic effect of intralesional therapies has been modest, highlighting a need for improvement. Cytokines such as interleukin-2 and Interferon-α have been in use for the treatment of melanoma and renal cell cancer for more than 2 decades, but are hampered by modest clinical efficacy and high systemic toxicity (Fyfe et al., 1995, Kirkwood et al., 1996).

Messenger RNA-based therapies, long under investigation as a potential treatment modality, have recently leapfrogged into the clinic with the approval and wide adoption of SARS-CoV-2 vaccines for the prevention of COVID-19, triggering considerable excitement about this technology for many other applications in medicine, including cancer therapy. Hotz and Wagenaar et al recently reported the results from extensive pre-clinical studies conducted by investigators at BioNTech, Sanofi, and Translational Oncology at the University Medical Center of Johannes Gutenberg University (TRON) testing local delivery of cytokines encoding mRNA(Hotz et al., 2021).

Iterative in vivo screening of lead cytokine mRNAs in the B16F10 melanoma and CT26 colon carcinoma models resulted in selection of mRNAs encoding 4 cytokines for intra-tumoral administration: IL-12 single chain (IL-12sc), granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-15 fused with the α-chain of its receptor (IL-15 sushi), and IFN-α4 (Figure 1). This 4-cytokine mixture induced substantial anti-tumor immunity in both the B16F10 and the CT26 models as evident by complete tumor regression in a large proportion of animals. Although all 4 cytokines were necessary for optimal activity, IL-12sc appeared to have a particularly important role. Bone marrow and splenocytes cultured with conditioned supernatants containing the cytokine mRNA without IL12sc secreted substantially lower amounts of IFN-γ. Furthermore, omission of IL-12sc from the 4-cytokine mRNA regimen compromised anti-tumor activity in the CT26 model to a larger extent compared to exclusion of any of the other 3 cytokines. A role for intratumoral IL-12 has also previously been shown with mRNA delivery or by delivery of tumor-localizing versions of IL-12 (Mansurov et al., 2020, Momin et al., 2019), further validating the importance of this systemically toxic cytokine. The 4-cytokine regimen also demonstrated higher activity against B16F10 tumors compared to each of the 4 cytokines administered individually. Importantly, animals with complete response after cytokine treatment were protected from tumor re-challenge, suggesting the induction of a memory immune response. Antibody depletion experiments showed the key role of CD8+ T cells, but also CD4+ and Natural Killer (NK) cells in mediating the anti-tumor responses. Cytokine mRNA-treated B16F10 mice exhibited increased frequencies of CD4⁺ and CD8⁺ T cells (but not regulatory CD4⁺ T cells) and anti-tumor activity was critically dependent on the IFN-γ pathway, an observation corroborated in IFN-γ knockout mice. Ovalbumin-specific CD8⁺ T cells (OT-I) adoptively transferred into mice bearing B16F10-ovalbumin tumors proliferated vigorously in the draining lymph node upon treatment with a single dose of the 4-cytokine mRNA regimen, indicating successful priming of tumor-antigen specific CD8⁺ T cells. Furthermore, CD8⁺ T cells specific for the immunodominant endogenous retroviral envelope protein (gp70) antigen in the CT26 model also expanded in the blood and tumor. Notably, re-challenge of mRNA cytokine-treated mice with CT26 tumors lacking gp70 were protected from tumor growth, suggesting induction of a broadened immune response now including subdominant antigens. To address the critical question of systemic anti-tumor immunity, the efficacy of intratumorally administered cytokine mRNA was assessed in B16F10 mice bearing subcutaneous and pseudo-metastatic lung lesions as well as a B16F10 mice with tumors on both flanks. Injection of the cytokine mRNA mixtures into one of the two flank tumors generated systemic anti-tumor immunity as evident by regression of the lung pseudo-metastases or the contralateral flank tumors. Combination with anti-CTLA-4 or PD-1 therapy augmented both the local and systemic anti-tumor effect of intratumorally injected cytokine mRNA in the CT26 and B16F10 models. Direct anti-tumor activity of the mRNA cytokine mixture was also seen in tumors lacking β−2 microglobulin (β2M), a key mechanism of resistance to PD-1 inhibition in humans, suggesting an impact of stimulation of innate responses alone. Impressively, anti-tumor activity was observed with cytokine mRNA either alone or in combination with PD-1 inhibition in 9 additional syngeneic murine tumor models with varying degrees of resistance to PD-1 inhibition.

Figure 1:

Injection of murine tumors with RNAs encoding 4 cytokines leads to in vivo translation and induction of a local and systemic tumor-specific immune response, mediating regression of both injected and distant tumors. Immune responses are comprised of several immune cell populations including T cells, dendritic cells (DCs), and Natural Killer (NK) cells. Anti-tumor activity is dependent on the IFN-γ pathway, consistent with a key role for IL-12sc in the cytokine mixture, and tumor antigen-specific T cells are expanded intratumorally.

Whereas previous approaches employing mRNA for intra-tumoral delivery of co-stimulatory molecules and cytokines have employed lipid-based formulations, in the current study the mRNAs were administered in a saline solution, which was demonstrated to still result in sufficient translation but may reduce systemic exposure and toxicity(Haabeth et al., 2019, Hewitt et al., 2019). This focus on transient and local delivery of cytokines is in line with its intended mechanism of action i.e. the triggering of an effective systemic anti-tumor response at the tumor site - an in situ vaccination. Given the compelling anti-tumor activity of the 4-cytokine mRNA regimen alone, with relatively modest additional efficacy when administered in combination with PD-1 inhibition, exploration of additional mRNAs encoding mechanistically distinct immune-stimuli is warranted, given that in this study only IL-2 was additionally explored. Notably, in the study by Haabeth et al(Haabeth et al., 2019) anti-tumor activity of both injected and non-injected metastatic sites and induction of a tumor-specific memory T cell response was also observed after intralesional delivery of lipid nanoparticle encapsulated mRNA encoding a different set of cytokines (IL-23 and IL-36γ) in addition to the T cell co-stimulation agonist OX-40L. Similar to the work by Hotz and Wagenaar et al, this regimen induced strong systemic tumor-specific immunity as well as anti-tumor activity – both as monotherapy and in combination with PD-1 inhibition in various murine models, including anti-PD-1 resistant tumors. A direct comparison of these distinct RNA based intra-tumoral therapies (nanoparticle-based versus saline solution formulation, different sets of cytokines with and without co-stimulatory agonism) would be desirable, potentially even prior to moving the approach into the clinic. The versatility of the mRNA approach – as is impressively demonstrated by the systematic analysis of different cytokine combinations identifying IL-12 as its key ingredient in the study by Hotz and Wagenaar - presumably lends itself to an “iterative learning process” that could be extended to additional complementary therapeutic targets. The inclusion of carefully curated sets of key tumor antigens into the mRNA constructs may in principle even obviate the need for intra-tumoral administration, which may substantially expand the clinical applicability of this promising in situ RNA vaccination strategy for cancer patients.