Intratumoral immunotherapy with mRNAs encoding chimeric protein constructs encompassing IL-12, CD137 agonists, and TGF-β antagonists

Intratumoral immunotherapy strategies for cancer based on interleukin-12 (IL-12)-encoding cDNA and mRNA are under clinical development in combination with anti-PD-(L)1 monoclonal antibodies. To make the most of these approaches, we have constructed chimeric mRNAs encoding single-chain IL-12 fused to single-chain fragment variable (scFv) antibodies that bind to transforming growth factor β (TGF-β) and CD137 (4-1BB). Several neutralizing TGF-β agents and CD137 agonists are also undergoing early-phase clinical trials. To attain TGF-β and CD137 binding by the constructions, we used bispecific tandem scFv antibodies (taFvs) derived from the specific 1D11 and 1D8 monoclonal antibodies (mAbs), respectively. Transfection of mRNAs encoding the chimeric constructs achieved functional expression of the proteins able to act on their targets. Upon mRNA intratumoral injections in the transplantable mouse cancer models CT26, MC38, and B16OVA, potent therapeutic effects were observed following repeated injections into the tumors. Efficacy was dependent on the number of CD8+ T cells able to recognize tumor antigens that infiltrated the malignant tissue. Although the abscopal effects on concomitant uninjected lesions were modest, such distant effects on untreated lesions were markedly increased when combined with systemic PD-1 blockade.


INTRODUCTION
Combinations of immunotherapy agents are one of the most avidly pursued routes to improve efficacy against cancer. 1 In this regard, agents endowed with powerful intrinsic activity often pose serious hurdles in terms of safety.A strategy to mitigate such problems is to locally deliver the immunotherapy agents to malignant lesions in an attempt to maximize therapeutic effects and avoid systemic side effects. 2 The expectations are that local interventions raise immune responses that would control distant untreated tumors (abscopal or nonenestic effects). 3The most widely used immunotherapy agents via intratumoral routes are Toll-like receptor (TLR) agonists, recombinant viruses, and different types of nucleic acids encoding cytokines. 4mRNA transfer has several advantages, including rapid dose-dependent expression that, although transient, can be sustained by means of repetition with subsequent doses. 5mRNAs encoding immune transgenes have shown efficacy in preclinical models and are being tested in clinical trials. 6,7Naked mRNA dissolved in buffers containing Ca 2+ could deliver mRNAs to be transiently expressed in the tumor. 8,9However, formulation optimizations are desirable to increase delivery and expression, including nanolipoformulation with ionizable lipid mixtures. 6,10Furthermore, circularization and optimizations of sequences can be implemented for more efficient clinical development.Self-replicating RNA constructs also offer some advantages. 11tratumoral delivery of a number of cytokines is reportedly efficacious against mouse cancer models. 12Moreover, a number of these agents are being tested in the clinic.4][15] Therefore, recombinant versions of IL-12 constructed to be monomeric by a linker (singlechain IL-12 [scIL12]) 16 have been used in multiple gene therapy approaches to locally treat tumor-bearing mice. 11,17These efforts include lipoformulated and naked IL-12-encoding mRNAs. 5,8,180][21] The efficacy of scIL-12 mRNA results from its activity on T lymphocytes and natural killer (NK) cells, whereupon it induces activation and IFNg production, especially if IL-18 is also present. 5,22IL-12 has a key role at inducing Th1 differentiation from Th0 CD4 + T cells.In mouse models, the activation of STAT4 by IL-12 was found to be essential for the differentiation of Th1 while inhibiting the differentiation toward a Th2 phenotype. 23L-12 local gene transfer can be combined with systemic immunotherapy agents such as anti-PD-(L)1 monoclonal antibodies (mAbs). 5,24Moreover, various cytokines can be locally delivered together in the form of multiple mixed mRNAs.This approach has been followed with scIL-12, IFNa, granulocyte-macrophage colonystimulating factor (GM-CSF), IL-15, OX40L, IL-23, IL-36g, and IL-18. 5,8,25,26 addition to cytokines and checkpoint inhibitors, other immunotherapy targets are being preclinically and clinically pursued.Among them, agonist T cell costimulatory mAbs directed to CD137 (4-1BB). 27In the case of the current study, bispecific constructs targeting 4-1BB crosslinking to the tumor microenvironment have been tested to avoid dose-limiting liver inflammation as was observed with the agonist antibody urelumab given systemically to patients with cancer. 28In preclinical modeling, the anti-CD137 mAb 1D8 is active against various mouse transplantable tumors but also induces a certain degree of liver inflammation in the mice. 29Importantly, new CD137 bispecific antibodies targeted to the tumor show evidence for clinical activity. 30,31e tumor microenvironment is enriched in soluble immunosuppressive factors that are known to curtail the functions of the immune cells involved in antitumor responses, 32 including T and NK lymphocytes, as well as dendritic cells.In that regard, dimeric active TGF-b is reportedly a dominant soluble mediator at downregulating antitumor immunity, 33,34 as reviewed by E. Batlle and J. Massagué. 35The group of Richard Flavell showed that gene transfer of a dominant-negative variant of TGFBRII to CD8 + T cells markedly enhances antitumor immune responses. 36,37Hence, multiple approaches are being followed to neutralize TGF-b activity. 38A mAb termed 1D11 that neutralizes all TGF-b isoforms shows antitumor activity in mice. 39,40A humanized version of this mAb is in the clinic (fresolimumab) but has dose-limiting side effects. 41Other forms of TGF-b antagonists such as TGF-bR inhibitors and TGF-b traps are also under clinical development via systemic delivery.[44][45] In this study, we sought to exploit chimeric constructs encoded by mRNA for intratumoral delivery, seeking synergistic effects of the various functional components in the constructions.Single-chain antibody fragments and scIL-12 were used.A tri-specific construct has been demonstrated to lead to neutralization of TGF-b and immunostimulation via IL-12R and CD137 ligation.Local delivery intends to maximize immunogenicity in a given tumor territory, acting as an in situ vaccine. 46Therapeutic efficacy results were observed that can be synergized simultaneously using PD-1 blockade.

RESULTS
mRNA-encoded trifunctional immunotherapy agents encompassing scIL-12, anti-CD137, and anti-TGF-b scIL-12 is reportedly active upon intratumoral injection of an mRNA encoding this cytokine. 18In order to improve activity, we cloned single-chain fragment variable (scFv) variants of 1D8 and 1D11 mAbs that, respectively, act agonistically on CD137 or antagonize TGFb. 39,47To make the constructs flexible, spacing linkers were included between the sequence of the synthetic cytokine and the scFv sequences as schematized in Figure 1A, 48 which represents the mRNA constructs and the predicted proteins.It has been shown that scIL-12 remains functional if present at the N terminus of fusion proteins, 16 and two versions of the triple constructs were generated with the 1D8 scFv and the 1D11 scFv linked in alternative order (IL-12-tandem Fv1 [taFv1] and IL-12-taFv2) as shown in Figure 1A.
To study the proteins encoded by these in-vitro-synthesized mRNAs, we transfected several cell lines with the mRNA lipocomplexed using TransIT.Figure 1B shows that in the supernatant of the transiently transfected cells, we could quantify IL-12, denoting the presence of the chimeric proteins.Western blot analyses of the supernatants developed with anti-IL-12 mAb revealed bands of the predicted molecular sizes (Figure S1A).Moreover, the secreted IL-12 was active Sandwich ELISA assay as schematized in which recombinant TGF-b1 was coated to the plate surface and, following incubation of the indicated supernatants, was developed by biotinylated anti-IL-12 mAb+streptavidin-HRP. (E) The same supernatants were assayed to neutralize luciferase expression in mink lung epithelial cells carrying a TGF-b-sensitive luciferase reporter system.1D11 anti-TGF-b mAb was used as a positive control for inhibition.Recombinant TGF-b was added to every condition at 2 ng/ mL.(F) Similar sandwich ELISA as in (D) but coating recombinant mouse CD137 to detect binding of the constructs in the culture supernatants of the indicated cell lines transfected with mRNAs encoding the corresponding constructs.(G) Scheme of indirect staining and flow cytometry analysis of CD137 binding to activated mouse T cells expressing CD137.(H) FACS histograms showing the IL-12 coating of activated T cells from WT and CD137 À/À mice incubated in the presence of the indicated supernatants containing the trispecific constructs.Results are representative of three experimental replicates.In (C) and (E), data are expressed as mean ± SEM.See also Figures S1-S3.
since the conditioned culture supernatants could induce IFNg release from mouse splenocytes preactivated for 24 h with anti-CD3 mAb when a 48-h conditioned culture in the presence of the supernatants was carried out (Figure 1C).Next, we studied if the supernatants contained constructs binding TGF-b1 using ELISAs on TGF-b1-coated plates.As shown in Figure 1D, binding of the constructs to recombinant TGF-b1 was substantiated in the analysis of the supernatants of all the transfected mouse cell lines.Moreover, the supernatants successfully inhibited luciferase activity controlled by a TGF-b reporter system based on luciferase as a reporter gene (Figure 1E).TGF-b bioactivity was measured using TGF-b-sensitive mink lung epithelial cells (MLEC) transfected with the reporter system. 49In the conditions tested, the supernatants were able to abolish the effects of 2 ng/mL recombinant TGF-b1 to induce luciferase in the reporter system.Similar ELISA binding experiments were carried out on plasticbound recombinant mCD137, indicating binding to the target (Figure 1F).Furthermore, binding to CD137 on the surface of activated mouse splenocytes by CD3+CD28 stimulation was observed (Figures 1G and 1H).Flow cytometry staining for extracellular scIL-12 was used to indirectly develop the flow cytometry assay, and activated splenocytes from CD137 À/À mice were used as a negative control (Figures 1G, 1H, and S1B).
To ascertain if these constructs could be expressed in vivo, we used hydrodynamic gene transfer to the liver of mice with the corresponding cDNA constructs in a cytomegalovirus (CMV) promotercontrolled expression plasmid (Figure S2A).16-18 h following hydrodynamic gene transfer, the serum samples of the mice contained measurable concentrations of IL-12 by ELISA (Figure S2B).Using these sera, it was also possible to demonstrate by ELISA techniques the binding to TGF-b1 that was coated to plastic plates (Figure S2C).The sera with the 1D11 scFv-containing constructs were also able to functionally neutralize TGF-b in the MLEC bioactivity assay (Figure S2D).Binding to recombinant mouse CD137 by the constructs in the serum was also revealed in ELISA assays on mCD137-coated plates (Figure S2E).
The liver can also be gene transferred in a transient manner using lipoplexed mRNA following intravenous administration of the synthetic mRNAs complexed with TransIT reagents.Again, following intravenous injection, readily detectable levels of the constructs could be detected over 6 to 48 h (Figure S3A, left panel), which gave rise to increases in circulating IFNg (Figure S3A, right panel).Of note, this treatment, given three times to B16OVA-tumor-bearing mice, resulted in some degree of tumor growth control (Figure S3B).However, such mice often succumbed due to systemic toxicity reflected by weight loss and elevated circulating transaminases (Figures S3C-S3E).
Overall, our results show that functional trimeric constructs could be encoded by mRNA and transferred in vivo to achieve antitumor immunotherapeutic effects.Given the worrisome safety profile of systemic IL-12, 13 we sought to investigate the potential of these constructs for intratumoral immunotherapy. 2 Therapeutic intratumoral delivery of chimeric mRNAs encoding scIL-12, anti-TGF-b, and anti-CD137 mRNAs encoding the constructs were formulated in Ringer's lactate and were used to directly inject established subcutaneous B16OVA-and MC38-derived tumors. 5Our group has previously demonstrated that the intratumoral administration of IL-12 mRNA in mice bearing subcutaneous tumors is feasible and does not result in observable toxicity. 5In B16OVA, substantial concentrations of IL-12 could be measured in the interstitial fluid of tumors injected with the mRNA encoding IL-12 and IL-12-taFv1 (Figure S4A, left panel), and as a result, increased concentrations of IFNg were also measurable in such recovered tumor interstitial fluid (Figure S4A, right panel).Much lower concentrations of IL-12 were found in peripheral blood (Figure S4B, left panel), and no IFNg was detected in the circulation (Figure S4B, right panel).This shows the greater safety profile of intratumoral injection.Similar observations were made using mice bearing MC38-derived tumors (Figure S4C).
Next, we studied the therapeutic effect of repeated intratumoral injections of the mRNA constructs into B16OVA-derived tumors on days 6, 9, and 12 post-tumor cell subcutaneous engraftment (Figure 2A).As shown in Figures 2B and 2C, both IL-12-taFv1 and IL-12-taFv2 exerted antitumor activity.No such activity was seen with control mRNA encoding luciferase, and less efficient activity was observed when scIL-12 was injected at equimolar doses.The advantage in terms of complete regression and survival of IL-12-taFv1 led us to decide to further develop such a construct (Figure 2D).Of interest, successful treatment was associated with vitiligo in the area of the rejected experimental melanomas (Figure 2E).The therapeutic effects of intratumoral IL-12-taFv1 were also observed against CT26-derived tumors implanted in syngeneic Balb/c mice (Figures 2F-2I), in which we also observed complete regression of the injected tumors (4 out of 6), again highlighting the advantage over equimolar doses of mRNA encoding scIL-12 (Figure 2G).
An important next step was to determine if the constructs exerted effects on distant concomitant tumors that did not receive intratumoral treatment.Experiments in Figure 3A show clear effects on the tumors injected with the mRNA constructs in the bilateral B16OVA model that were more favorable in the case of IL-12-taFv1.More importantly, 3 out of 11 complete regressions were observed in the concomitant distant tumors (Figures 3A and 3B).Similar experiments in mice bearing bilateral MC38 tumors also demonstrated the more efficacious bilateral effect of IL-12-taFv1 (Figures 3C and 3D), thus providing another reason to select this construct for subsequent experimental development.

Efficacy of local immunotherapy with the chimeric mRNA IL-12-taFv1 construct is dependent on the function of CD8 + T cells
To study the immune requirements underlying the efficacy of our mRNA construct, we performed selective lymphocyte depletion experiments in mice receiving treatments for B16OVA tumors such as those in Figure 3A.Following CD8 b + T cell depletion or double CD8 b + T cell and CD4 + T cell depletion, the therapeutic effects on the directly treated and distant tumors were almost absent (Figures 4A and 4B).Single depletion of CD4 + T cells had no effect and even seemed to enhance efficacy against the contralateral tumor, perhaps as a result of regulatory T cell (Treg) depletion. 50xt, we investigated whether CD137 was involved in our therapeutic effects by using CD137 À/À C57BL/6 recipient mice in comparison to cohoused wild-type control mice.As can be seen in Figure 4C, the therapeutic effect of intratumoral mRNA encoding IL-12-taFv1 was reduced (Figures 4C and 4D).
In this experimental setting of bilateral B16OVA-bearing mice, we also studied the T cell infiltrates upon treatment as indicated in Figure 5A.A two-dose regimen was implemented in this case in order to prevent the complete eradication of the treated tumors, to permit exams of the tumor microenvironment.As a result of therapy, numbers of CD8 + T cells increased in the treated tumors, while Tregs were markedly reduced.Moreover, in the contralateral tumors, a similar trend was observed regarding an increased abundance of infiltrating T cell numbers (Figure 5B).
To establish if CD8 + T lymphocytes were systemically recognizing their tumor-associated cognate antigen, IFNg-enzyme-linked immunosorbent spot (ELISpot) assays were performed using splenocytes from treated mice, as indicated in Figure 5A.Images of IFNg-ELISpot wells (Figure 5C) and quantitative data (Figure 5D) show that the numbers of CD8 + T cells recognizing the canonical ovalbumin (OVA) epitope (SIINFEKL) presented by H-2K b were clearly increased.In this vein, antigen-stimulated OT-I and/or OT-II T lymphocytes exposed in culture to IL-12-taFv1 or IL-12 enhance their production of IFNg (Figure S5).
As a whole, our data indicate that local immunotherapy injection of the mRNA-encoded trispecific chimeric construct exerts both local and systemic T cell-mediated effects, which are therapeutically beneficial to intratumorally treated mice.the combination with anti-PD-1 mAb given systemically (Figure 6A).Results in the B16OVA bilateral model confirmed the efficacy on injected tumors but again showed modest, albeit observable, effects on the distantly implanted tumors (Figure 6B).In this setting, anti-PD-1 mAb given systemically did not achieve any measurable efficacy (Figures 6B and 6C).However, the combination of the intratumoral mRNA IL-12-taFv1 construct with systemic anti-PD-1 mAb resulted in excellent bilateral efficacy, giving rise to complete regression and long-term survival of half of the treated mice (Figures 6A-6D).
Mice surviving long term in these experiments that had remained tumor free for at least 90 days were bilaterally rechallenged with B16OVA in one flank and with B16F10 in the contralateral flank.
Naive age-matched mice were used for comparison.As can be seen in Figure 6E, cured mice showed excellent memory to reject B16OVA rechallenge and some degree of cross-reactivity with B16F10, which delayed progression of such tumors.Collectively, our results demonstrate that chimeric immunotherapeutic mRNAs encompassing IL-12, anti-TGF-b, and anti-CD137 can be used for intratumoral approaches, potentiating the effect of checkpoint inhibitors.

DISCUSSION
In a search for more efficacious alternatives for intratumoral cancer immunotherapies, we tested chimeric mRNAs encoding immunotherapeutic products to be delivered as mRNAs.Based on previous evidence, we sought to combine the effects of IL-12 with TGF-b neutralization and CD137 agonism.3][54][55][56][57] IL-12 is efficacious upon intratumoral administration of recombinant viruses and IL-12-encoding nucleic acids, 13 thus providing the rationale to enhance their function with alternative complementary mechanisms.Using mRNA in this setting is advantageous since a more sustained local presence of the protein is achieved, and the resulting translated product has a better chance of locally mediating its immunomodulatory activity.Conceivably, bioavailability of the encoded proteins in tumor-draining lymph nodes is favored.Furthermore, mRNAs are more easily and economically produced than complex recombinant chimeric proteins.mRNA can be used for intratumoral purposes, but liver gene transfer can also be attained using this organ as an endogenous factory of therapeutic proteins. 58,59 can be envisioned that the IL-12-taFv1 construct will simultaneously bind dimeric TGF-b, act on IL-12 receptors, and, as a consequence, be crosslinked to stimulate CD137.Indeed, crosslinking by dimeric TGF-b could be a major advantage of IL-12-taFv1 as a therapeutic agent.Moreover, the coexpression of the targets may result in selective biodistribution of the locally released construct that would be retained in malignant tissues.
The alternative to these chimeric products would have been the use of a mixture of mRNAs.Such an approach has been followed with other mRNAs encoding cytokines and has some advantages.However, it runs up against the unpredictable dominance of some mRNAs over others in terms of expression, less straightforward kinetics, and less simple clinical development due to the need to handle multiple RNA moieties.The chimeric construct also has the advantage of physically linking various counter receptors and potentially giving rise to synthetic and synergistic biology.Indeed, we observe potent effects on injected tumors that are attributable to the mutual potentiation of the effects of each component in the chimeric construct.mRNA as a toolbox allows the delivery of complex chimeric proteins that are otherwise difficult to manufacture and purify. 6For instance, in our case, the construction of a human reactive version would be feasible based on known sequences for the scFv of urelumab, 60 on the interspecies TGF-b crossreactivity of the 1D11 scFv, and on the available scIL-12 sequences. 18chimeric molecule encompassing a TGF-b trap (TGFBRII) and type I IFN had been previously tested as encoded by mRNA for in vivo treatment with a rationale similar to our studies.61 Local and regional interference with TGF-b is likely to locally improve dendritic cell functions 62,63 and interfere with Treg differentiation 64 and Treg suppressive functions.65,66 These effects are desirable in addition to the mentioned functional release of CD8 T lymphocytes and NK cells upon TGF-b blockade, 36,37,67 either directly or through modulation of myeloid-derived suppressor cells.68 Of important note, TGF-b blockade reportedly synergizes with agonist mAbs to costimulatory members of the TNFR family including OX40 69 and 4-1BB.40,55 In our experiments, we observe clear antitumor effects that are weak against concomitant tumors that are not directly injected with the mRNA constructs.This is in spite of marked increases in CD8 + T cell infiltrates in the mRNA-treated tumors, with a similar trend in the contralateral tumors growing in the group treated with the chimeric molecule.Intriguingly, the percentage of CD4 + T cells in such distant noninjected tumors slightly declined, while no changes were observable in the treated tumors. Thse observations require further investigations focused on lymphocyte migration cues and on the fate of CD4 + T cells in the mice receiving treatment.Moreover, the in situ vaccination approach is conducive to the production of functional antigen-specific cytotoxic lymphocytes (CTLs) abundantly present in the spleen and detectable by IFNg-ELISpot.As expected, the antitumor effects as exerted by the mRNA-encoded chimeric molecule were weaker in mice lacking CD137, thus providing evidence for the relevance of CD137 ligation by the chimeric construct.
Given that IL-12 would elicit IFNg and thereby promote PD-L1 expression, we tested the combination of the intratumoral mRNA constructs and systemic PD-1 blockade. 51The synergistic bilateral effect on difficult-to-treat mouse tumor models that are completely resistant to PD-1 monotherapy is truly remarkable.In addition, PD-1 blockade is standard of care for multiple malignancies and easy to combine in early clinical development. 70IFNg is also well known for the induction of surface major histocompatibility complex (MHC) class I expression, [71][72][73] thereby facilitating tumor cell recognition by cognate CD8 + T lymphocytes.
Intratumoral immunotherapy faces several obstacles, including logistics and the biological and antigenic heterogenicity of metastatic lesions, but offers the opportunity for tolerable synergistic combinations. 74Locally releasing more than one bioactivity in the chimeric constructs aims to produce synergistic effects that can be further potentiated using systemically delivered agents.In our case, we intend to keep TGF-b at bay precisely in the tumor tissue area to which we are providing immunostimulation with IL-12 and anti-CD137 agonists.8][79] However, other options are possible including engineering CD28, ICOS (inducible T cell costimulator), or CD27 costimulation.
We used naked mRNA for the proof of concept, but to treat larger tumors in humans, we are investigating lipoformulations in order to maximize delivery and expression.This sort of chimeric construct could also be considered to be launched from oncolytic RNA viral vectors or self-replicating RNAs (srRNA). 80,81srRNAs have advantages since small doses may result in larger quantities of transgene expression, as recently demonstrated for IL-12 with a Semliki Forest virus-based srRNA. 82rther improvements in mRNA are currently being investigated to optimize the expression of these relatively long messengers. 10Transient systemic leakage because of liver expression could be advantageous but also dangerous.To mitigate this potential issue, microRNA (miR) targets can be incorporated into the mRNAs to prevent hepatic expression. 18,83Once the proof of concept for these immunocytokines launched from mRNA is made, other costimulatory mAbs (i.e., anti-OX40, ICOS, CD27) can be considered as well as other cytokines in similar mRNA-encoded formats.
All things considered, our work provides evidence that chimeric constructs encompassing cytokines and two mAb fragments can be locally and systemically delivered by the synthetic mRNA moieties that encode for them.The strategy holds promise for clinical translation improving the effects of existing intratumoral mRNA-based immunotherapies and showing synergistic effects with systemic delivery of checkpoint inhibitors.
Cell lines and tumor mouse models HEK293T (293T) and CT26 cells were procured from the ATCC.MC38 cells were kindly provided by Dr. Karl E. Hellström (University of Washington, Seattle, WA, USA) in September 1998. 84B16OVA cells were a kind gift from Dr. Lieping Chen (Yale University, New Haven, CT, USA) in November 2001. 84The MLEC cell line was provided by Dr. Fernando Pastor (CIMA, Universidad de Navarra, Pamplona, Spain). 49Cell lines were cultured in RPMI 1640 medium (Gibco) supplemented with 10% FBS (Sigma-Aldrich), 100 U/mL penicillin, 100 mg/mL streptomycin (Gibco), and 5 Â 10 À5 mol/L 2-mercaptoethanol (Gibco).HEK293T and MLEC cells were maintained with DMEM high glucose (Gibco) supplemented with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin.B16OVA and MLECs cells were supplemented with 400 and 250 mg/mL geneticin (Gibco), respectively.All cell lines were grown in a humidified incubator with 5% CO 2 at 37 C for at least 7 days before inoculation into mice.All cell lines were routinely tested for mycoplasma contamination using the MycoAlert Mycoplasma Detection Kit (Lonza).
For the B16OVA, MC38, and CT26 unilateral tumor model, C57BL/6 mice were subcutaneously injected with 5 Â 10 5 tumor cells in the right flank.For the B16OVA and MC38 bilateral tumor model, C57BL/6 mice were subcutaneously injected with 5 Â 10 5 tumor cells in the right flank and 1.5 Â 10 5 cells in the left flank on day 0. For tumor rechallenge, mice that were B16OVA tumor free at least 90 days after treatment received a subcutaneous injection of 5 Â 10 5 B16OVA and B16F10 cells in the right and left flanks, respectively, in a total volume of 50 mL PBS.

Plasmids, mRNA synthesis, and ex vivo transfection
Variable heavy (V H ) and variable light (V L ) sequences for the anti-TGF-b and anti-CD137 were obtained by sequencing of each hybridoma.The variable domains of the 1D11 or 1D8 fragments were fused via a (G 4 S) 5 linker, and the two scFvs of 1D11 and 1D8 were fused via a SG 4 S linker. 48scIL-12 was fused to the taFvs antibodies via a (G 4 S) 3 linker. 16,85The protein sequences of IL-12-taFvs were designed as following starting from N terminus IL-12-taFv1: Protein sequences of taFv1 and taFv2 used as controls were the same as the IL-12-taFvs but were devoid of scIL-12.
The cDNA sequences encoding mRNAs were cloned by GeneScript in the pUC57-Kan vector containing a T7 promoter upstream of the cDNAs and followed by 2 tandem repetitions of the 3 0 UTR sequence of the human b2-globin cDNA and a 60 poly A tail.The cDNA sequence encoding construct DNAs were cloned by GeneScript in the pcDNA3.1(+)vector.
The mRNA production was performed as previously described. 5For some experiments, the mRNAs were formulated with TransIT-mRNA Transfection Kit (Mirus Bio) for cell transfections, according to the manufacturer's instructions.

MLEC luciferase assay
For the luciferase assay, 4 Â 10 4 MLEC cells were plated in a 96-well plate and allowed to attach for 5 h in a humidified incubator with 5% CO 2 at 37 C. Supernatants from mRNA-transfected cells or serum from hydrodynamically injected mice were incubated with rhTGF-b1 (Peprotech cat.100-21) at 2 ng/mL for 30 min at 37 C.Such samples with rhTGF-b1 were added to MLEC cells and then incubated in a humidified incubator with 5% CO 2 at 37 C for 1 h.Cells were then washed with PBS and incubated overnight in 0.1% BSA DMEM.After 16-18 h, the luciferase assay was performed using Bio-Glo Reagent (Promega, cat.G7940) according to the manufacturer's instructions.Firefly luciferase signal was detected in an Orion L Microplate Luminometer (Berthold Detection System).

Western blot assay
For direct protein visualization, supernatants from mRNA-transfected cells were separated on 8% acrylamide SDS-PAGE gel and transferred with Trans-Blot Turbo Transfer System to a PVDF blot membrane.IL-12 protein was detected using an anti-mIL-12 antibody (AP-MAB0853).
In vivo transient mRNA and DNA gene transfer mRNAs were administered by intravenous or intratumoral injections.Given the different molecular weights of the mRNAs and to ensure equimolar administration, the same amount of mRNA moles (8.7 mmol) was injected for each treatment.
For the intravenous administration in B16OVA tumor-bearing mice, mRNA was formulated in TransIT-mRNA transfection kit (Mirus Bio) reagent and intravenously injected into mice, as previously reported. 48For intratumoral administration, mRNAs were formulated in Ringer's lactate (Grifols) and then injected into the tumor in a total volume of 50 mL.
Hydrodynamic tail-vein injections were performed administrating 10 mg of the indicated cDNA plasmids diluted in 2 mL sodium chloride 0.9% at room temperature in less than 8 s.

Design of mRNA in vivo treatment experiments
To determine IL-12 or IFNg concentrations in tumor extracts and blood, mRNAs were intravenously or intratumorally injected.To quantify IL-12 and IFNg tumoral protein concentrations, tumors were excised 6 h following the intratumoral mRNA injections and homogenized using VWR Disposable Pellet Mixers in PBS containing cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail (Roche).For quantification in blood, 100-150 mL of peripheral blood were collected in 50 mL of Heparin (Hospira).
To evaluate the toxicity of systemically injected mRNAs, mice bearing B16OVA subcutaneous tumors were intravenously injected on days +6, +9, and +12 after tumor engraftment.Biochemical analysis was performed on day +13 on mice plasma, and body weights were recorded every day.
To evaluate the therapeutic efficacy of mRNA-encoded proteins, tumor-bearing mice were intratumorally injected with mRNAs formulated in Ringer's lactate on days +6, +9, and +12 after tumor cell subcutaneous implantation.Tumor sizes were measured twice a week.
To investigate the effects of mRNA treatments on tumor T-cell infiltrate, mice bearing B16OVA tumors were intratumorally injected with the indicated mRNAs diluted in Ringer's lactate on days +9 and +12 after tumor cell subcutaneous implantation.On day +14, mice were euthanized, and primary and contralateral untreated tumors and spleens were excised, and the cell suspensions obtained were analyzed by flow cytometry and the ELISpot assay.
To evaluate PD-L1 expression in mRNA-treated tumors, mice bearing B16OVA tumors were intratumorally injected on days +8 and +9 with 26.1 mmol saline-formulated mRNAs.24 h following the last injection, tumors were excised and cell suspensions analyzed by flow cytometry.
For immune cell depletion studies, 100 mg control (IgG), anti-CD8 b (53-5.8),anti-CD4 (GK1.5), or a combination of anti-CD8 b and anti-CD4 antibodies were intraperitoneally administered 1 day before the first mRNA intratumoral treatment, twice a week for the next 2 weeks, and once a week thereafter until the end of the experiment.Efficient depletion of CD4 + , CD8 + , or both populations was checked on day +14 in peripheral blood samples by flow cytometry.
All the antibodies used for in vivo studies were purchased from BioXcell.

ELISA determinations of cytokine concentrations and ELISpot assays
IL-12p70 protein levels in cell supernatants, supernatants of minced tumor tissue, and serum samples were quantified using commercially available ELISA kits (BD OptEIA Mouse IL-12 (p70) ELISA Set, cat.555256), according to the manufacturer's instructions.
For IFNg determination following splenocyte stimulation with supernatants, splenocytes from wild-type (WT) mice were preactivated with plate-bound anti-CD3 (1 mg/mL) for 24 h and exposed to B16OVA transfection medium.Supernatants were collected after 48 h of incubation and analyzed by ELISA (BD OptEIA Mouse IFNg ELISA Set, cat.555138) according to the manufacturer's instructions.The same ELISA kit was used to detect IFNg in supernatants of minced tumor tissue and mouse serum samples.
The OVA-specific CD8 + T cell response was assessed ex vivo using a mouse IFNg-ELISpot Assay kit (BD 551083).Splenocytes depleted of erythrocytes were added to the wells (8 Â 10 5 cells) and then stimulated with synthetic OVA 257-264 peptide (1 mg/mL) for 16-24 h.IFNg-producing cells were assessed by counting the spots with reference to input cells according to the manufacturer's instructions.
All the samples were treated with FcR-Block (anti-CD16/32 clone 93; BD Biosciences) before the cytometry staining.

Figure 1 .
Figure 1.Construction, expression, and function of mRNA-encoded chimeric trifunctional constructs encompassing scIL-12 and scFv anti-TGF-b and anti-CD137 (A) Scheme of the mRNA constructions used that include the incorporated sequences from scIL-12, 1D11 anti-TGF-b mAb, and 1D8 anti-CD137 mAb with the adequate linkers.Representation of the encoded proteins is provided for scIL-12, IL-12-taFv1, and IL-12-taFv2.(B) ELISA-determined concentration of IL-12 in the culture supernatants of the cell lines 293T, B16OVA, MC38, and CT26 transfected with the mRNAs encoding the indicated constructs formulated in TransIT.(C) The indicated supernatants from B16OVA were used to stimulate preactivated spleen T lymphocytes, and IFNg production in the culture was quantified by ELISA upon 48-h culture.(D) Sandwich ELISA assay as schematized in which recombinant TGF-b1 was coated to the plate surface and, following incubation of the indicated supernatants, was developed by biotinylated anti-IL-12 mAb+streptavidin-HRP. (E) The same supernatants were assayed to neutralize luciferase expression in mink lung epithelial cells carrying a TGF-b-sensitive luciferase reporter system.1D11 anti-TGF-b mAb was used as a positive control for inhibition.Recombinant TGF-b was added to every condition at 2 ng/ mL.(F) Similar sandwich ELISA as in (D) but coating recombinant mouse CD137 to detect binding of the constructs in the culture supernatants of the indicated cell lines transfected with mRNAs encoding the corresponding constructs.(G) Scheme of indirect staining and flow cytometry analysis of CD137 binding to activated mouse T cells expressing CD137.(H) FACS histograms showing the IL-12 coating of activated T cells from WT and CD137 À/À mice incubated in the presence of the indicated supernatants containing the trispecific constructs.Results are representative of three experimental replicates.In (C) and (E), data are expressed as mean ± SEM.See also Figures S1-S3.

Figure 2 .
Figure 2. Intratumoral injections of the chimeric mRNA-encoded constructs exert antitumor effects on transplantable mouse tumor models (A) B16OVA tumor cells were engrafted in syngeneic mice to produce established tumors and mRNAs in Ringer's lactate were injected intratumorally at equimolar doses.(B) Individual follow up of tumor sizes with the fraction of mice achieving complete regressions.(C) Compiled data with statistical comparisons.(D) Overall survival of the mice in the experimental groups.(E) Representative image showing vitiligo in the area where the rejected tumors used to be.(F) Scheme of similar experiments in CT26 tumors engrafted in syngeneic Balb/c mice.(G) Individual tumor follow up in the indicated groups of treatment.(H) Compiled data and statistical comparisons.(I) Overall survival of mice in the indicated treatment groups.Results are representative of two replicates with comparable outcome.In (C) and (H), dotted lines represent the dates of mRNA treatments, and data are expressed as mean ± SD.Longitudinal data were fitted to a third-order polynomial equation and compared with an extra sum-of-squares F test (C and H).Statistical comparisons in (D) and (I) were made using the log-rank test.Statistical significance: *p % 0.05, **p % 0.01, ****p % 0.0001.See also Figure S4.

Figure 3 .
Figure 3. Modest but observable abscopal effects of the mRNA encoding chimeric immunotherapeutic constructs on concomitant untreated tumors In mice bearing bilateral B16OVA-derived tumors for 6 days in opposite flanks, treatment was only given to the right implanted tumor.(A) Individual follow up of tumor size in groups of mice intratumorally treated with the indicated mRNAs.The top panels represent injected tumors and the bottom panels the concomitant noninjected counterparts.(B) Compiled data and statistical comparisons.(C) Similar experiments in mice bearing bilateral MC38-derived tumors for 6 days treated with the indicated mRNA constructs.(D) Compiled data and statistical comparisons.Results are representative of two repetitions with comparable outcomes.In (B) and (D), dotted lines represent the dates of mRNA treatments, and longitudinal data were fitted to a third-order polynomial equation and compared with an extra sum-of-squares F test.Data are expressed as mean ± SD.Statistical significance: *p % 0.05, ****p % 0.0001.

Figure 4 .
Figure 4. CD8 + T cells and CD137 requirements for antitumor activity (A) Selective depletions with anti-CD4 and anti-CD8 b monoclonal antibodies of B16OVA bilaterally engrafted mice intratumorally injected with the indicated IL-12-taFv1 mRNA construct on days 6, 9, and 12. Individual sizes of directly treated tumors and concomitant distant tumors are shown.(B) Compilation of data and statistical comparisons.(C) Experiments performed in either WT or CD137 À/À mice bearing B16OVA tumors as indicated.Tumors were intratumorally treated with Ringer's lactate (vehicle) or the IL-12-taFv1 mRNA construct and measured over time.(D) Compiled data and statistical comparisons.In (B) and (D), dotted lines represent the dates of mRNA treatments.Results are representative of two repetitions with a comparable outcome.Longitudinal data were fitted to a third-order polynomial equation and compared with an extra sum-of-squares F test (B and D).Data are expressed as mean ± SD.Statistical significance: **p % 0.01, ***p % 0.001, ****p % 0.0001.

Figure 5 .
Figure 5. Intratumoral treatment with the mRNA chimeric constructs results in an increase of CD8 + T cells both in injected and noninjected lesions and results in specific CD8 systemic immunity (A) Scheme of experiments in mice bearing bilateral B16OVA-derived tumors.(B) Flow cytometry quantification of the percentages of CD8 + , CD4 + , and CD4 + CD25 + FoxP3 + over CD45 + leukocytes in the corresponding treated and contralateral tumors following intratumoral injection of the indicated mRNA constructs or vehicle control.(C) Images of IFNg-ELISpot assays using splenocytes from treated mice as in (A) that were restimulated in a 24-h culture without antigen or with synthetic SIINFEKL peptide.(D) Quantitative data from the ELISpot assays and statistical comparisons.Results are representative of two repetitions with comparable outcomes.In (B), one way ANOVA tests followed by Sidak's post-test were used for statistical comparisons across groups.In (D), statistical comparisons were made by two-way ANOVA followed by Tukey posttest.Data are expressed as mean ± SD.Statistical significance: *p % 0.05, **p % 0.01, ***p % 0.001, ****p % 0.0001.

Figure 6 .
Figure 6.Intratumoral injections of mRNA encoding chimeric constructs encompassing IL-12, anti-TGF-b, and anti-CD137 synergize with systemic PD-1 blockade (A) Schematic representation of the experiments in mice bearing bilateral B16OVA tumors.Mice were intratumorally treated with the IL-12-taFv1-encoding mRNA and systematically with anti-PD-1 mAb as indicated.(B) Individual tumor sizes followed over time of the indicated groups of treatment.The fraction of mice in which complete regression of the tumors was attained is provided for the directly treated and contralateral noninjected tumors.In (C), compiled data and statistical comparisons are provided, and dotted lines represent the dates of mRNA treatments (D) Overall survival of the mice in the indicated groups of treatment.(E) Cured mice from experiments in (B) were rechallenged with tumor cells at least 90 days after being declared tumor free.Mice received B16OVA cells in the right flank and B16F10 cells in the left flank.Tumor sizes were monitored, and naive mice were used as a control.Experiments in (B) and (C) were repeated twice.Data were fitted to a third-order polynomial and compared using an extra sum-of-squares F test (C and E).In (D), statistical comparisons were made using the log rank test.Statistical significance: **p % 0.01, ***p % 0.001, ****p % 0.0001.
Statistical methodsFlow cytometry analyses were performed with CytExpert software.Means and standard deviations of the mean are presented as averages and error bars unless otherwise indicated in the figure legends.GraphPad Prism v.8 (La Jolla, CA, USA) was used for statistical analysis as indicated in figure legends.When differences are statistically significant, the significance is represented with asterisks according to the following values: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.the manuscript; C.A.D.T. performed experiments and analyzed results; I.O.performed experiments and analyzed results; G.G. performed experiments; J.G.-V.performed experiments and analyzed data; B. Pinci performed experiments; S.G. performed experiments and analyzed data; S.S.-G.performed experiments and analyzed data; A.A. performed experiments and analyzed data; I.E.-S.provided key reagents and methods; K.V. performed experiments and analyzed data; B. Palencia managed project and data; M.A. designed and analyzed experiments; M.C.O.designed and analyzed experiments; A.T. supervised experiments and wrote the article; P.B. designed experiments, cosupervised the project, and wrote the article; I.M. supervised the project, designed experiments, analyzed data, and wrote the article.