Oncolytic Alphaherpesvirus and Myeloid-Tropic Cytomegalovirus Cooperatively Enhance Systemic Antitumor Responses

doi:10.21203/rs.3.rs-2817662/v1

Download PDF

Research Article

Oncolytic Alphaherpesvirus and Myeloid-Tropic Cytomegalovirus Cooperatively Enhance Systemic Antitumor Responses

https://doi.org/10.21203/rs.3.rs-2817662/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 03 Nov, 2023

Read the published version in Molecular Therapy →

Version 1

posted

You are reading this latest preprint version

Oncolytic virotherapy aims to activate host antitumor immunity, but the effectiveness of current intratumoral herpes-simplex virus (HSV) based monotherapy is limited. Here, to boost the potency of cancer cell targeting α-herpesviruses derived from HSV or Varicella-Zoster Virus (VZV), we coinjected them with myeloid cell-tropic β-herpesviruses derived from cytomegalovirus (CMV) which can powerfully modulate immune responses. Co-administered CMV was shown to promote intratumoral replication of HSV, increase local release of pro-inflammatory cytokines, inhibit adaptive antiviral immune responses, enhance systemic antitumor responses, and lead to superior control of both injected and contralateral tumors. CMV-encoded myeloid-tropic genes, NK evasins (i.e. the vMHC-I homolog), and the m02-m16 region (involved in downregulating classical MHC-I) were implicated in enhancing antitumor activities. CMV is a promising adjuvant to reshape and enhance the antitumor immune responses following α-herpesvirus therapy.

Oncolytic viruses (OVs) proliferate selectively within tumors, kill cancer cells, and activate systemic anti-tumor immune responses ¹. T-VEC and DELYTACT, both herpes-simplex virus type 1 (HSV-1) based virotherapies, have been approved for the treatment of melanoma and malignant glioma, respectively. Following intratumoral administration of an oncolytic virus, tumor antigens (tAgs) released by dead or dying tumor cells can be phagocytosed, processed, and presented by locally resident antigen-presenting cells (APCs) to stimulate tAg-reactive cytotoxic lymphocytes (CTL) ². The strength and durability of the virus-induced tumor-specific immune response are the major determinants of tumor control, both in pre-clinical and clinical studies ³. However, activation of antitumor immunity is often insufficient to impact treatment outcomes following intratumoral virotherapy, being limited by factors such as restricted intratumoral viral propagation, a highly immunosuppressive tumor microenvironment (TME), and/or insufficient recognition of tumor antigens ⁴. To enhance virus-induced antitumor responses after oncolytic α-herpesvirus therapy we co-injected oncolytic HSV or VZV with a myeloid-tropic cytomegalovirus (CMV) which targets tumor-associated myeloid cells including tumor-associated macrophages (TAMs) which may account for up to 50% of tumor mass and are highly pro-tumorigenic, pro-angiogenic and immunosuppressive ⁵. Aside from their distinct cell tropisms, it was hypothesized that the complex mechanisms by which a- and β-herpesviruses modulate the host immune response might prove complementary and/or synergistic in the oncolytic virotherapy setting.

The human α-herpesviruses have a characteristic short reproductive cycle, rapidly kill infected cells, and are able to infect diverse tissues ⁶. For the current study, Herpes-simplex virus type 1 (HSV-1) and varicella-zoster virus (VZV) reverse genetics systems were established in the laboratory and were used to generate oncolytic α-herpesviruses. Both platforms are highly lytic towards epithelial and neural cells ^{7, 8} and both encode immune evasion proteins (i.e. HSV ICP47 and VZV ORF66) that downregulate the transporter associated with antigen processing (TAP) thereby inhibiting MHC-I antigen presentation and the activation of CD8 + T cells. In addition, by suppressing the expression of NK activating ligands or upregulating the expression of NK inhibitory ligands, both of these α-herpesviruses decrease the susceptibility of infected cells to NK cell attack ⁹. However, despite these suppressive mechanisms, responses of plasmacytoid dendritic cells (pDCs), NK cells, and TAP-MHC-I dependent CD8 + T cells are still maintained and rapidly control virus infection ¹⁰. In contrast to HSV and VZV, CMV directly infects myeloid cells, modulates the behavior of APCs, and powerfully reshapes host immune responses ^{11, 12}. Moreover, immunological imprinting by CMV affects host immunity against both viral and non-viral stimuli ¹³ making this virus an interesting candidate with which to impact the induction of tAg-reactive immune responses.

CMV is a slowly replicating β-herpesvirus with a strong tropism for fibroblasts and myeloid cells ¹⁴. The virus has a 235kb double-stranded linear DNA genome encoding over 200 polypeptide or RNA products, more than 40 of which are known to modulate host immune responses, most notably by interfering with NK functions and/or the antigen presentation activities of professional APCs ¹⁵. The US1-US11 region of CMV encodes a number of proteins that inhibit the classical TAP-dependent MHC-I and MHC-II antigen presentation pathways, increasing the vulnerability of the infected cell to NK cell attack. To compensate for this, at least 12 HCMV genes are known to inhibit NK cell activation ¹⁵. The UL18 gene product is a CMV-encoded peptide-binding vMHC-I homolog that can send inhibitory signals to NK cells by engaging with the LIR-1 receptor ¹⁶. Interestingly, while CMV efficiently blocks peptide antigen presentation via classical MHC-I, it also redirects and reshapes the antigen-specific immune response by expanding CD94/NKG2C⁺ cells and HLA-E restricted NK- cytotoxic T lymphocytes (NK-CTLs) as well as a subset of non-exhausted CD8 + T cells recognizing non-canonical peptides ^17–20. Similar to human strains of CMV, mouse cytomegalovirus (mCMV) is myeloid-tropic and encodes multiple immune evasins that interfere with the classical MHC-I/II antigen presentation pathway and NK cell functions ²¹. In this regard the m04 gene product blocks K^b-restricted antigen presentation, the m06 product redirects MHC-I for degradation, and m144 encodes a vMHC-I homolog ²². In the absence of conventional T cell responses primed via the classical MHC-I antigen presentation pathway during mCMV infection, non-classical CD8 + T cells are induced and appear sufficient to provide meaningful antiviral protection ²³. Interestingly, an intratumorally delivered mCMV was recently shown to infect tumor associated macrophages (Mφ) and to work in concert with PD-L1 blockade to amplify the antitumor CD8 + T cell response in a melanoma model, even though the virus alone failed to control tumor growth ²⁴.

In light of the above findings, we hypothesized that CMV could be used as a viral adjuvant to reshape and boost the systemic antitumor immune responses and hence the antitumor potency of an intratumorally administered oncolytic a-herpesvirus. Our studies show that, whereas the a-herpesviruses preferentially targeted tumor cells, the coinoculated CMV preferentially targeted tumor associated Mφ/DCs and upregulated a variety of pro-inflammatory cytokines. CMV also promoted the replication of attenuated HSVs but suppressed the adaptive anti-HSV immune response, at the same time enhancing the systemic antitumor immune responses, a finding that was validated both in vivo and in vitro. Our mechanistic studies suggested that the myeloid tropism and multiple immune evasion genes of CMV were implicated in triggering these enhanced antitumor responses.

mCMV enhances systemic antitumor efficacy when combined with an oncolytic α-herpesvirus.

Using recombinant KOS-BAC viruses (Supplementary Fig. 1 and Supplementary Fig. 3a-c) carrying NIS and bhCG reporter transgenes, we initially assessed the efficacy of HSV-1 monotherapy. KOS-BAC R2.3 (wild-type backbone), KOS-BAC R4.1 (ΔICP47) and KOS-BAC R15.1 (ΔICP47ΔUL41) viruses were tested in parallel in immune competent mice with bilateral subcutaneous EMT6 tumors. Following IT virus delivery (Supplementary Fig. 4a), the WT KOS-BAC R2.3 demonstrated the best efficacy, whereas viruses with ΔICP47 or ΔICP47ΔUL41 showed decreased efficacy. Interestingly, when administered IT + IV, antitumor efficacy was not seen in the WT virus group, and the ΔICP47 virus was most effective with the DICP47ΔUL41 virus showing diminished anti-tumor activity (Supplementary Fig. 4b).

Based on the above findings, the combination of KOS-BAC R4.1 (ΔICP47) with mouse CMV (Smith strain) was evaluated as intratumoral therapy in two mouse bilateral syngeneic tumor models (EMT6 breast cancer and A20 B cell lymphoma). Increased antitumor efficacy was observed for both injected and uninjected tumors when mCMV was co-administered with HSV at a 1:10 ratio (Fig. 1a,b). To determine whether the HSV vector could be further optimized for combination therapy with mCMV, additional HSV variants were constructed on the KOS-BAC R2.3 backbone by deleting ICP47, ICP34.5, UL41, ICP0, or ICP6, or mutating gB to obtain a hyperfusogenic variant ²⁵ (Supplementary Fig. 1 and Supplementary Fig. 3c). These HSV variants were co-injected with mCMV into EMT6 tumors (Fig. 1c and Supplementary Fig. 5). Interestingly, the two WT viruses (KOS R2.3 and mCMV) were antagonistic in suppressing tumor growth, and the replication of KOS R2.3 declined quickly after 3 dpt. In contrast, all of the KOS variants with attenuating mutations displayed improved antitumor efficacy. The KOS R29.1 virus carrying hyperfusogenic gB^R858H mutation showed improved antitumor efficacy and elevated virus replication especially on 6 dpt compared to the WT virus (Fig. 1c). The KOS R24.11 (ΔICP0) was highly effective against injected tumors but did not significantly impact uninjected tumors. Notably, KOS R4.1 (ΔICP47) and KOS R20.11 (ΔICP34.5) exhibited superior activities against both injected and uninjected tumors, with KOS R20.11 demonstrating the highest efficacy (Fig. 1c and Supplementary Fig. 5). Based on these results, a new vector KOS-BAC R29.31 carrying both gB^R858H mutation and ΔICP34.5 was constructed for use in subsequent experiments (Supplementary Fig. 1).

To determine whether CMV could boost the antitumor activity of other α-herpesviruses, we turned our attention to VZV, which is a highly cell-associated pathogen. At least 70 ORFs are encoded by VZV, and all but six of them have homologs in HSV. In addition, VZV lacks homologs for nine HSV genes, including ICP34.5, US11, and ICP47 ²⁶. Initially, two reporter genes (NIS and bHCG) were introduced into the vaccine strain vOka backbone; however, only the vOka-BAC R2.2 with bHCG can grow properly. Based on R2.2, the vOka-BAC-bHCG-gBsyn (gB^Y881F) was constructed (Supplementary Fig. 2a and Supplementary Fig. 3d). The vOka R2.2 virus showed potent antitumor efficacy in human melanoma MeWo model, whereas the hyperfusogenic gB^Y881F mutation significantly reduced both virus replication and antitumor effects (Supplementary Fig. 6) ²⁷. Furthermore, the insertion of scIL12 into vOka-BAC prevented virus propagation. Therefore, a drug-controllable Tet-off scIL12 cistron was added to VZV-Ellen-BAC vector to obtain Ellen R5 (Supplementary Fig. 3e). In the bilateral B16-F10 model, the combination of Ellen R5 with mCMV administered intratumorally was proven superior in inhibiting tumor growth of both injected and uninjected tumors, and prolonging animal survival compared to monotherapy groups (Fig. 1d).

mCMV targets tumor-associated myeloid cells, promotes replication of oncolytic HSV, and stimulates expression of multiple pro-inflammatory cytokines.

To identify the in vivo cellular targets of each virus, KOS-BAC R4.1 (GFP) and mCMV-BAC (dTomato) (Supplementary Fig. 2b) were co-injected into EMT6 tumors. Tumors were collected at 2 dpt and virus infected cells were co-stained with antibodies recognizing immune markers. Co-staining with myeloid markers reveals that the two viruses targeted different cells, and that most of the CMV infected cells were CD11b+, many of them also expressing CD11c and/or F4/80 (Supplementary Fig. 7b). Considering that both Mφ and DCs are positive for these three markers, and that these cell types are known to be the primary targets of HCMV (31, 32), we conclude that mCMV targeted the tumor associated Mφ/ DCs which may in turn transport the virus to tumor draining lymph nodes to facilitate spreading ^28–30. Precise definition of the specific cell subsets targeted by CMV is challenging because the virus can strategically impair the expression of immune markers ³¹. Co-staining of CMV infected cells with fibroblast and endothelial markers (CD90.2 and CD31) was not observed, and HSV-infected cells did not co-stain with CD45 (Supplementary Fig. 7a,b).

To evaluate how HSV and CMV interact with each other when co-injected into the same tumor, HSV reporter gene expression was monitored. Blood and tumor bHCG levels show that the co-administration of mCMV improved the replication of KOS-BAC R4.1 (DICP47) at early time points (Fig. 2a). However, beyond 3 dpt, the replication of KOS-BAC R4.1 rapidly declined in both the monotherapy and combination groups. Comparing combination group versus monotherapy groups q-PCR analyses revealed an increase of intratumoral KOS-BAC R4.1 gene expression and a downregulation of mCMV expression on 3 dpt (Fig. 2b). Cytokine array screening assays were used to determine the relative levels of intratumoral cytokines in the combination group versus single agent groups. On 3 dpt, several pro-inflammatory cytokines were slightly upregulated in the combination group comparing to other three groups (Fig. 2c). In contrast, a variety of pro-inflammatory cytokines were upregulated in the combination group on 9 dpt (Fig. 2d and Supplementary Fig. 9). ELISA analyses verified the upregulation of CCL4 and CXCL13 on 3 dpt in the combination group, also the upregulation of CCL4, IL12-p70, and pigment epithelium-derived factor (PEDF) on 9 dpt by the combination treatment (Fig. 2e). Functions of these regulated cytokines include recruiting and activating immune cells (CCL4, CXCL13, and IL12) as well as inhibiting angiogenesis (PEDF) ^32–34. We further observed that HCMV Merlin or AD169r-BAC (Supplementary Fig. 2c) infection in monocyte-derived macrophages upregulated the expression of a range of pro-inflammatory, pro-proliferation, and anti-proliferation cytokines while downregulating many anti-inflammatory and pro-angiogenesis cytokines (Supplementary Fig. 8 and Supplementary Fig. 9). The altered tumor cytokine environment triggered by CMV may be related to the increased replication of KOS-BAC R4.1 at early time points ³⁵.

Oncolytic HSV + mCMV combination induces robust systemic antitumor immune responses and suppresses antiviral immunity.

Before the evaluation of combination therapy induced antitumor immune responses, PEDF, single chain IL12-p70 (scIL12), and IL15Ralpha sushi/IL15 (rIL15RA/IL15) were inserted to KOS-BAC R29.31 vector as therapeutic transgenes (Supplementary Fig. 1). Coadministration of the KOS R30.4 (scIL12) with mCMV in the unilateral A20 model resulted in a 100% CR rate and robust systemic anti-A20 cell immune responses in the spleens cells from CR responders (Fig. 3a,b and Supplementary Fig. 10a). Since the virus-released tAgs have not been characterized in the A20 model, tumor cells were used directly as tAg-presenting cells in the Elispot analyses ³⁶. We went on to determine whether simultaneously expressing multiple therapeutic transgenes could further improve antitumor efficacy. The KOS viruses R30.4 (scIL12), R30.5 (rIL15RA/IL15), and R30.44 (scIL12-PEDF) were injected alone or mixed. Neither injection of the KOS R30.44 virus nor co-delivery with the KOS R30.5 virus resulted in enhanced effects compared to the R30.4 (scIL12) group. Co-injection of mCMV significantly improved the efficacy of KOS R30.44, but this enhancement was negated when the rIL15RA/IL15 expressing virus was used (Fig. 3c and Supplementary Fig. 10b). The rIL15RA/IL15 might be antagonistic to the enhancing effect of mCMV co-infection since it may activate CMV-reactive NK cells ³⁷. No direct correlation was seen between blood IL12-p70 levels and CR rate (Supplementary Fig. 10c). IFN-γ elispot analyses using target tumor cells mixed with autologous spleen cells from CR mice revealed no statistically significant variations in the number of spots across treatment groups (Fig. 3d and Supplementary Fig. 11a). Interestingly, anti-HSV and anti-mCMV immune responses were both significantly diminished in the KOS R30.44 + mCMV group in comparison to other treatment groups (Fig. 3d and Supplementary Fig. 11b,c) and antiviral responses were restored when using a HSV encoding rIL15RA/IL15.

Tumor-reactive immune cells induced by R30.4 plus mCMV combination therapy demonstrate potent tumor-killing ability.

Using the bilateral B16-F10-nectin1 model, systemic antitumor responses induced by the intratumoral administration of KOS R30.4, mCMV or both viruses in combination were examined. Improved antitumor activity was demonstrated, and higher bHCG levels were detected on 3dpt in the combination group (Fig. 4a-c). Elispot analysis indicates that spleen cells from the combination group (collected on 34 dpt) exhibited increased antitumor responses against the B16-F10-nectin1 tumor cells compared to spleen cells from the monotherapy groups (Fig. 4d,e). When spleen cells from all of the treatment groups were expanded first and then stimulated using B16F10 tumor cells (effector: target (E:T) ratio at 10:1), the stimulated spleen cells from the combination group showed considerably higher cell viability compared to spleen cells from KOS R30.4 monotherapy group (Fig. 4f). Further in vitro cytotoxicity assays against B16-F10 and B16-F10-nectin1 tumor cells (E:T ratio at 1:1) revealed a considerably enhanced tumor cell killing activity for the stimulated spleen cells obtained from the combination group (Fig. 4g,h). Single-cell RNA sequencing was performed to phenotype the stimulated spleen cells from combination therapy group (cells from the R30.4 single agent group failed to meet QC requirements for sequencing due to lower cell viability). Twelve cell clusters exhibiting different cell features were plotted based on fold changes of marker genes. Differential levels of immune markers were observed for the cell clusters (Supplementary Fig. 12 and Supplementary Table 2). All the cell clusters were positive for CD45, CD3, and αβ-TCR. Low levels of γδ-TCR were also detected for cell clusters 1–4 and 11. CD8 and NK markers were observed in most clusters except clusters 8 and 12. These CD8 and NK positive clusters also showed the expression of activation markers (CD69, CD25, and IFN-γ) and cytotoxic molecules (granzyme B and perforin 1). Among all the cell clusters, cluster 4 showed the highest levels of the proliferation markers Ki67 and proliferating cell nuclear antigen (PCNA). There was no or very limited expression of B cell markers, CD4, myeloid markers, or the Treg marker Foxp3 in all the cell clusters (Supplementary Fig. 12b). The results suggest that the effector cells showing potent tumor cell-killing ability might include specific subsets of cytotoxic T cells potentially with NK or γδ-T cell features. Anti-tumor activity was also tested in immune deficient mice bearing SKOV-3 xenografts but there were no significant differences in tumor growth control or animal survival between KOS R30.4 monotherapy versus combination therapy with mCMV plus KOS R30.4 (Supplementary Fig. 13). The lack of additive or synergistic efficacy in this xenograft model further confirm that the enhanced antitumor of the mCMV plus HSV combination therapy is immune system dependent.

CMV-encoded NK evasins and MHC-I targeting genes are implicated in the enhancement of alphaherpesvirus antitumor activity.

To probe the identity of CMV determinants implicated in the enhancement of HSV-induced antitumor activity, various CMV deletion mutants were constructed (Supplementary Fig. 2b). Starting from the mCMV-BAC-R0 with Δm145-m155 (m145-m155 stabilizes the inhibitory NKR-P1B ligand Clr-b and downregulates CD86 and NKG2D ligands), recombinant mCMVs were constructed with additional deletion of m129-m131 (controls macrophage tropism as part of a gH/gL/MCK-2 complex), m132-m135 (function unknown), m136-m138 (inhibits NK activation by downregulating MULT-1, RAE-1varepsilon, and H60), m144 (vMHC-I homolog, sends inhibitory signals to NK cells), or m02-m16 (involved in targeting classical MHC-I molecules) ^{21, 38–44}. Each of these mCMV variants was co-injected with KOS R30.4 intratumorally in a bilateral subcutaneous CT26 tumor model (Fig. 5a,b). Neither m145-m155 nor m132-m135 deletion reduced the antitumor effectiveness of the combination therapy or impacted blood bHCG levels. In contrast, when m129-m131, m144, or m02-m16 were removed, the anticancer effects were reduced. While these differences in tumor response rate were not statistically significant, the trending efficacy attenuation of specific mCMV recombinants suggest that the CMV encoded myeloid-tropic genes and the vMHC-I homolog which confers ability to escape NK activation were likely contributing to the antitumor activity of the virus combination. Additionally, the m02-m16 region which contains MHC-I targeting genes (i.e. m04 and m06) may have contributed to the enhanced antitumor responses. As an assessment of antiviral immunity, the titers of antiviral neutralizing antibodies were measured in all treatment groups on day 26 (Fig. 5c). Anti-HSV antibody titers were considerably lowered in all combination therapy groups compared to the single agent KOS R30.4 group, whereas only the mCMV R3 (Δm129-m131) group showed a significant decrease in anti-mCMV antibody titers.

Pre-existing antiviral immunity does not inhibit the antitumor activity of herpesvirus combination therapy.

Given the high prevalence of herpesvirus seropositivity in human populations ⁴⁵, we sought to determine whether pre-existing anti-herpesviral immunity could negatively impact virus-induced antitumor responses (Fig. 6a). Immune competent mice were pre-immunized with HSV or CMV and subcutaneous EMT6 tumors were implanted bilaterally in both flanks. Prior to intratumoral virus administration, ganciclovir (GCV) was administered to stop the replication of pre-inoculated WT viruses. Antiviral neutralizing antibody titers were determined on 1 day before virus treatment to confirm successful immunization (Fig. 6b). After treatment with KOS R30.4 + mCMV combination, monitoring of blood bHCG levels and animal survival times revealed no significant differences between the virus-immune and nonimmune groups (Fig. 6c). Antitumor and antiviral immune responses were detected in spleen cells from all treatment groups (Fig. 6d). These results indicate that pre-immunization with WT herpesviruses neither prevented re-infection nor significantly reduced virus-induced antitumor immune responses.

T-VEC is an HSV based virotherapy that has been approved to treat relapsing stage 3 or 4 melanoma and has a reported overall response rate (ORR) of ~ 25% with complete response rate (CRR) of ~ 10% ⁴⁶. However, expanding the T-VEC label to other cancers has proven challenging even though responses have been observed when the virus is combined with checkpoint inhibitor therapy ⁴⁷.

Here we tested the hypothesis that CMV might enhance the antitumor activity of a cancer cell-targeting oncolytic α-herpesvirus, such as HSV, by infecting and modulating the activities of tumor-associated myeloid cells. In line with this hypothesis, enhanced antitumor activity was observed both against injected and uninjected tumors when CMV was combined with intratumoral HSV or VZV α-herpesviral therapy in multiple bilateral mouse tumor models. We further observed that co-injection of CMV enhanced the intratumoral replication of attenuated HSVs (but not of wild-type HSV), upregulated expression of pro-inflammatory cytokines within treated tumors, suppressed adaptive anti-HSV immune responses and at the same time substantially enhanced anti-tumor CTL responses compared to single agent HSV therapy. The efficacy boosting activity of co-injected CMV was confirmed using several different HSV constructs encoding various immune-enhancing payloads (of which HSV-IL12 was the most potent).

To probe deeper into the molecular mechanism(s) whereby co-administered CMV is able to enhance the efficacy of intratumoral HSV oncolytic virotherapy, we generated and tested multiple CMV deletion mutants and tested them in the (bilateral) CT26 mouse model. All of these CMV deletion mutants had comparable in vitro replicative fitness and, although the differences in survival between the various treatment groups did not reach statistical significance after in vivo testing of HSV adjuvant activity, there were clear trends in the direction of diminished efficacy when using CMV recombinants lacking m129-m131, m144, or m02-m16. These findings suggest that the CMV encoded myeloid-tropic genes, the vMHC-I homolog which confers ability to escape NK cell attack and the genes that target MHC-I may be important contributors to the antitumor activity of the virus combination (Supplementary Fig. 14).

Previous studies aiming to precisely define the antitumor mechanisms of action of oncolytic viruses have proven challenging. Indeed, the mechanism of action of intratumoral T-VEC as melanoma therapy has still not been fully elucidated, but available evidence points to the local recruitment and activation of cytotoxic T lymphocytes capable of recognizing tumor-associated antigens ⁴⁸. While our findings with the Dm129-131 and Dm144 CMV mutants suggest that CMVs ability to boost the potency of an oncolytic HSV depends on its ability to generate chronically infected, NK-resistant myeloid lineage APCs, the question arises as to which functions of these CMV infected APCs may be contributing to the enhancement of the antitumor effect. In this regard, a potentially highly relevant “immune combat” function of CMV is to interfere with the presentation of certain types of MHC peptide complexes on the surface of the infected APCs. The capability of CMV to block both classical MHC-I and MHC-II antigen presentation pathways has been extensively studied and provides a well-established mechanism for the suppression of adaptive antiviral immunity ^{21, 49}. Our data show that adding CMV to intratumoral HSV therapy diminishes the adaptive immune response to the virus and virus-associated antigens but at the same time promotes the recognition and killing of tumor cells and tumor-associated antigens. Interestingly, while ELISPOT analyses of splenocytes from virus treated animals showed similar abundance of tumor cell reactive spleen cells for both monotherapy (HSV or CMV) and combination groups, potent tumor cell killing by antigen- stimulated spleen cells was observed only in the combination group. This finding suggests that the addition of CMV may have led to the “reshaping” of the antitumor immune response, perhaps via the engagement of nonconventional antitumor T cells or the unmasking of nonconventional tumor antigens. To further probe this possibility, studies are ongoing to further expand the virus-induced tumor-reactive immune cells and characterize their phenotypes.

Returning to our studies using CMV deletion mutants, the m02-m16 region of mCMV, which comprises MHC-I targeting viral genes, was also implicated in the HSV-boosting antitumor effects. This finding further supports a potential role for non-classical MHC-I pathways in presenting tAgs when the classical MHC pathway is blocked (in APCs) by CMV. Previously published studies have shown that the cross-presentation of non-canonical tAgs can be enhanced when the presentation of canonical tAgs is suppressed in tumors with a defective TAP-MHC-I pathway ^{50, 51}. Indeed, in 40–90% of human malignancies, the TAP-MHC-I pathway is downregulated, which is typically associated with cancer metastasis and poor prognosis ⁵², although in the absence of CMV infection APCs continue to present antigens in a conventional manner since they are not part of the malignant clone. Based on this hypothesis, future research will focus on characterizing the tAg-presenting MHC pathways (classical or non-classical MHC-I pathway) under the combination therapy.

A further notable aspect of the current study is that the efficacy enhancing activity of CMV was observed at a very low virus dose. This is fortuitous since in vitro amplification and preparation of high titer stocks of unmodified and/or recombinant CMVs is technically challenging and limits the maximum dose that can be applied during in vivo testing. For the studies reported in this manuscript we employed a CMV to HSV ratio of 10:1 infectious particles, typically administering just 10⁵ CMV infectious units intratumorally in combination with 10⁶ HSV infectious units. From our immunohistochemical analysis of infected tumors, it is apparent that, at this low CMV dose, only a small percentage of the myeloid cells in a given injected tumor were infected but this nevertheless resulted in a profound impact on the antitumor immune response. We therefore propose that subversion of the antigen presenting pathways, even in just a small subset of the tumor-infiltrating APCs, can have a profound and therapeutically beneficial impact on the nature of the host antitumor immune response.

Pre-existing antiviral immunity might be expected to attenuate the antitumor activity of oncolytic virus therapy and this has been demonstrated in several previously studied tumor models, particularly when viruses are delivered via the intravenous route ⁵³. Interestingly, in the current study, pre-existing immunity to HSV and/or CMV had no measurable impact on the antitumor activity when these two agents were administered in combination via the intratumoral route. This is of significance because human infections with HSV and CMV are both very common resulting in correspondingly high seroprevalence and T cell memory against both of these viruses among human cancer patients. Our results therefore support the further study and development of CMV as a viral adjuvant to enhance antitumor immune responses for current α-herpesvirus monotherapy.

Recombinant viruses

Genetic manipulations of herpesviruses were conducted using the bacterial chromosome vector (BAC) based lambda red recombineering system as described previously ⁵⁴. Viral gene deletions, mutations, and insertion of transgenes to the herpesvirus genomes have been verified by PCR and Sanger sequencing. The expression of transgenes has been confirmed by ELISA assays. Supplementary Fig. 1 and Supplementary Fig. 3a-c demonstrate the structures and in vitro growth of recombinant HSV-1 KOS-BAC viruses. Supplementary Fig. 2a and Supplementary Fig. 3d-f illustrate the structures and growth of recombinant VZV-BACs, while Supplementary Fig. 2b,c list the structures of recombinant mCMV-BACs and HCMV-BACs. Human sodium iodide symporter (NIS) and beta subunit of human chorionic gonadotropin (bHCG) were inserted into the HSV-1 and VZV genomes as reporter genes ⁵⁵.

Construction of recombinant HSV-1s: The eGFP expression cassette was amplified from the pcSV40KeGFP plasmid using primers No. 1-2 followed by the digestion with PciI/ KasI, and was then ligated into the PciI/ KasI treated pBeloBAC11 vector to obtain the BAC vector pSG. The BamHI site in pSG was removed by PCR amplification primers No. 3-4 followed by a blunt-end ligation, the obtained vector was designated as pSG2. To insert pSG2 between UL43 and UL44 of the HSV-1 KOS genome, the left and right homology arms were PCR amplified from KOS genome using two sets of primers, primers No. 5-6 and primers No. 7-8. The two homology arms were then digested with BamHI and HindIII, followed by being ligated into the BamHI/HindIII digested pSG2 vector, the obtained vector was designated as pSG3. The pSG3 vector was linearized with BamHI and then co-transfected with linear KOS genome into Vero cells (1 of linearized pSG3 and 1 of KOS genome were added per 5 10⁵ cells). At 3 days post transfection, GFP+ plaques were picked and transferred to Vero cells. KOS-BAC episome DNA was isolated from the GFP+ cells using the Hirt method as described previously ⁵⁴ and then transformed into the MegaX DH10B competent bacterial cells. Colonies containing full length KOS genome sequence were selected. Sequence integrity of the extracted KOS-BAC plasmid was confirmed by next generation sequencing (Genewiz Illumina MiSeq, 2 250bp configuration). The KOS-BAC plasmid was further transferred into the SW102 recombineering bacterial strain. To construct the NIS-P2A-bHCG cassette, the NIS and bHCG fragments were PCR amplified from pAAV-NIS plasmid and pAAV-bHCG plasmid using primers No. 9-12, and then were digested with HindIII/AflII and AflII/XbaI respectively, before being ligated into the HindIII/ XbaI digested pcSV40KeGFP vector, obtained plasmid was designated as pcSV40-pCMV-NIS-P2A-bHCG. The pCMV-NIS-P2A-bHCG cassette was amplified using primers No. 13-14 and was digested with SalI/ EcoRI before being ligated into the SalI/ EcoRI digested PL451 plasmid, obtained plasmid was designated as PL4512. The pCMV-NIS-bHCG-FRT-Kan-FRT cassette was amplified from the PL4512 plasmid using primers No. 15-16 and was then inserted between UL10 and UL11 of KOS-BAC within SW102 cells. The KanR sequencing in the KOS-BAC was then removed within the SW105 cells containing the ara-inducible Flpe gene ⁵⁴, obtained vector was designated as KOS-BAC-NIS-bHCG (R2.3). For deletion of viral genes with only one copy of ORF in R2.3, the pEM7-galK cassette was PCR amplified from the pGalK plasmid with two long primers, each contains 50nt sequence at the 5´ terminal as homology arm. Then the whole ORF of viral gene was replaced by the galK cassette mediated by the homology arms. The galK cassette inserted into viral genome was eventually removed using a 100nt single stranded DNA with the same sequence of homology arms. Positive and negative selections for galK+and galK- colonies were performed as described previously ⁵⁴. Recombinant viruses with deletion of viral genes (ICP47, UL41, ICP6, or ICP27) were constructed using the primers No. 17-28. RL1(ICP34.5) and RL2(ICP0) were replaced with the galK cassette and the KanR cassette (amplified from plasmid pRSV-rev-Kan) respectively, using primers No. 29-33 and primers No. 34-38. To construct the gB R858H point mutation, the nucleotide G at position 2573 of gB ORF was replaced with the galK cassette, which was then replaced with nucleotide A, primers used were No. 39-41. Transgenes (PEDF, scIL12 (single chain IL12-p70), and rIL15RA/IL15 (IL15Ralpha sushi/IL15)) were cloned from vectors (reagents No. 13, 16, 17, and 18) into the pSelect-zeo-fcy::fur plasmid using primers No. 46-47, and 52-59. PEDF was cloned into the pcSV40KeGFP plasmid using primers No. 64-67. Transgene cassettes were then inserted into the KOS-BAC backbone using primers No. 42-45. Recombinant HSV-1s were reconstituted in the Vero cells, or in Vero-cre cells (for removing the BAC sequence from virus constructs). The KOS construct carrying ICP27 deletion was failed to be reconstituted.

Construction of recombinant VZVs: To generate a VZV capturing BAC vector, two homology arms were amplified from the vOka episomal DNA using primers No. 70-73, and then were cloned into the pSG2 vector using restriction sites BamHI and HindIII, obtained vector was pSG6. The chloramphenicol resistance gene (CmR) cassette in the pSG6 vector was then replaced with the neomycin resistance cassette pPGK-pEM7-NeoR using primers No. 74-77, obtained vector was pSG7. pSG7 was linearized with BamHI and transfected into APRE-19 cells (2 of linearized pSG7 was added per 5 10⁵ cells), at 24h post transfection live VZV virus was added to the transfected cells with multiplicity of infection (MOI)=0.05. At 48h post infection, G418 (Geneticin) was added to cells with final concentration of 800 /ml to enrich the GFP+ infected cells. The circular VZV BAC DNA was extracted from the GFP+ cells and transferred to DH10B competent cells. BAC colonies containing full length VZV genome were selected. VZV-BAC DNA was then transformed into the SW102 cells. Sequence integrity of the VZV-BACs was verified by next generation sequencing (Genewiz Illumina MiSeq, 2×250bp configuration). The pCMV-NIS-P2A-bHCG cassette, pCMV-NIS cassette and pCMV-bHCG cassette were amplified from plasmids pcSV40-pCMV-NIS-P2A-bHCG, pcSV40-pCMV-NIS and pcSV40-pCMV-bHCG using primers No. 78 and 79, and then inserted between ORF60 and ORF61 of vOka genome within SW102 cells. The scIL12 expression cassette was amplified and inserted between ORF60 and ORF61 of vOka genome using primers No. 80 and 82 or used to replace the AmpR sequence of the pcSV40-pCMV-bHCG cassette inserted using primers No. 80 and 81. The Tet-off drug-controllable pRSV-tTA- pTight-scIL12 cassette was amplified from vector pTet-off-mIL12BA-ZeoR and inserted between ORF60 and ORF61 of VZV genome using primers No. 82 and 113. The vOka carrying gB^Y881F hyperfusogenic mutation was constructed using primers No. 117-119. Recombinant VZVs were reconstituted in the ARPE-19 cells, or in ARPE-19-cre cells (for removing the BAC sequence from virus constructs).

Construction of recombinant mCMVs: To generate a mCMV capturing BAC vector, dTomato expression cassette was amplified using primers No. 94 and 96, and then ligated to the PCR amplified linear SG vector (with primers No. 93 and 95), obtained vector was pET. Left and right homology arms were then amplified from the mCMV Smith genome using primers No. 97-100 and were inserted into the pET vector to construct the pET2 capture vector. pET2 was then linearized with AvrII and co-transfected with mCMV linear genome into NIH-3T3 cells (1 of linearized pET2 and 1 of Smith genome per 5 10⁵ cells). dTomato+ plaques were picked and enriched. Circular mCMV-BAC DNA was extracted and transformed into SW102 cells for genetic manipulations. Sequence integrity of the selected mCMV-BAC plasmid was confirmed by next generation sequencing (Genewiz Illumina MiSeq, 2×250bp configuration). For deletion of viral genes from mCMV-BAC, primers No.101-110 were used to amplify the galK cassette for replacing the viral genes m129-m131, m132-m135, m136-m138, m144, and m02-m16, respectively. The galK cassette within viral genome was then removed using a synthesized single stranded DNA consist of the same sequence of both homology arms. Recombinant mCMVs were reconstituted in NIH3T3-cre cells to remove the BAC sequence.

Construction of recombinant HCMVs: To generate a HCMV capturing BAC vector, left and right homology arms were amplified from the HCMV AD169 genome using primers No. 83-86, and then were cloned into the pSG2 vector, obtained vector was pSG4. pSG4 was linearized with BamHI and co-transfected with the HCMV linear genome (1 of linearized pSG4 and 1 of AD169 genome per 5 10⁵ cells) into MRC-5 cells. GFP+ plaques were picked and enriched. Circular HCMV-BAC DNA was extracted and transformed into SW102 cells for genetic manipulations. Sequence integrity of the HCMV-BAC was confirmed by next generation sequencing (Genewiz Illumina MiSeq, 2 250bp configuration). The UL1-UL20 region and sequences (UL148, UL147A, UL147, UL146, UL145, UL139, UL138, UL136, UL135, UL133, UL148A-D, UL150 and UL150A) within the UL/b' region were found lost in the AD169-BAC, the reference AD169 sequence used for alignments was GenBank: FJ527563.1. To replace the UL131A like sequence in the AD169-BAC with the UL131A ORF from Merlin strain, the AD169 UL131A like sequence was first replaced by the galK cassette which was further replaced with the Merlin UL131 ORF, primers used were No. 87-90.

HSV-1s and CMVs were purified as described previously ⁵⁶ with modifications. The supernatant of infected cells in forty 100mm dishes was collected when more than 90% of cells showed cytopathic effect (CPE) and was centrifuged at 3000 x g, 4 for 15min to remove large cell debris. Virus in the supernatant was then pelleted down using a Beckman Coulter Optima L-100K Ultracentrifuge with a SW32Ti rotor at 100,000 x g, 4 for 1h. Virus pellets were then incubated in PBS overnight at 4 , before being resuspended and loaded onto a continuous 15%-50% (w/v) sucrose density gradient in SW41 centrifuge tubes. The tubes were then centrifuged at 80,000 x g,4 for 1 h with a SW41 rotor. Virus bands were collected and diluted with PBS in SW41 centrifuge tubes, and were then centrifuged at 100,000 × g, 4 for 1h to pellet viral particles. Finally, the obtained virus pellets were incubated and resuspended in ice-cold PBS, before being aliquoted and stored at -80 . Virus titers were determined in Vero cells (for HSV-1), APRE-19 cells (for VZV), NIH3T3 cells (for mCMV) or MRC-5 cells (for HCMV) with plaque formation assay.

Cell lines

The Vero cells, Vero-cre cells, APRE-19 cells, ARPE-19-cre cells, NIH3T3 cells, NIH3T3-cre cells, MRC-5 cells, CT-2A-luc, B16-F10, and B16-F10-nectin1 were cultured in Dulbecco's Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS) and 100 units/ml of penicillin/streptomycin. The EMT6 cells and EMT6-nectin1 cells were cultured in Waymouth's MB 752/1 Medium with 2mM L-glutamine, 100 units/ml of penicillin/streptomycin, and 15% FBS. The CT26 cells and CT26-nectin1 cells were cultured in RPMI-1640 Medium with 10% FBS and 100 units/mL of penicillin/streptomycin. The A20 cells were cultured in RPMI-1640 Medium with 10% FBS, 0.05 mM 2-mercaptoethanol, and 100 units/ml of penicillin/streptomycin. The human ovarian adenocarcinoma SK-OV-3 cells were cultured in McCoy's 5a Medium with 10% fetal bovine serum (FBS) and 100 units/ml of penicillin/streptomycin. The nectin1 stable mouse tumor cell lines were generated with lentivirus transduction followed by puromycin selection and fluorescence-activated cell sorting (FACS). Lentivirus expressing human nectin1 was produced with the lentivirus vector pLV-EF1-nectin1-PGK-puro.

Flow cytometry

The puromycin enriched tumor cells were resuspended in ice cold FACS Buffer (1 mM EDTA, 25 mM HEPES pH 7.0, and 1% BSA in Ca/Mg2+ free Hanks Buffer) (1×10⁶ cells in 1ml buffer per tube), and then were stained with PE anti-human CD111 (Nectin-1) antibody or PE Mouse IgG1, κ Isotype Ctrl antibody together with the Live-or-Dye™ 350/448 for 30 min at room temperature. The nectin1-positive live cell populations were sorted with a BD FACSCanto Flow Cytometer and then were expanded in growth media at 37°C, 5%CO₂. To analyze the expression of nectin1 in the expanded stable cells, 1×10⁶ stable cells in 1ml ice cold FACS Buffer was stained with PE anti-human CD111 antibody or PE Mouse IgG1, κ Isotype Ctrl antibody together with the Live-or-Dye™ 350/448 for 30 min at room temperature. Wild type cells were stained parallelly as control. Stained cells were analyzed with a Bio-Rad ZE5 Cell Analyzer.

I^-125 uptake assay

Vero cells were plated in six-well plates (5×10⁶ cells per well) and cultured overnight before being infected with HSVs at MOI=0.01. Three replicates of wells were used for each group. Immediately prior to the uptake assay, cells were incubated at 37°C for 1h in 1 ml uptake buffer (Hank’s balanced salt solution (HBSS) containing 10 mM HEPES, pH 7.4). The radioactive substrate Na¹²⁵I in 0.1 M NaOH was diluted in uptake buffer, and then 5 μl contained ~1,000,000 counts per minute was added to each well. Cells were incubated at 37°C for 50 minutes and then were washed twice with 2 ml ice-cold uptake buffer. Cells were lysed with pre-warmed 1 ml 1M NaOH. Cell lysate was transferred to tubes for gamma counting using a Wallac 1480 gamma counter.

Animal studies

All mice experiments were approved by Mayo Clinic Institutional Animal Care and Use Committee (IACUC) and were performed in compliance with Mayo Clinic IACUC guidelines. For studies using syngeneic immune-competent tumor models, mouse tumor cells suspended in PBS (5×10⁷ cells/ml for A20 model, 1×10⁷ cells/ml for all other models) were subcutaneously implanted to both flanks (100µl per side of flank) of 5~6 weeks old female mice (Balb/c or C57BL/6J). When average tumor diameter reached 5 mm, viruses suspended in PBS were intratumorally (IT) delivered into the tumors on the right flank (50µl per IT dose) or were intravenously (IV) delivered (100µl per IV dose). Totally three doses were given for each mouse at indicated time points. For studies using immune-comprised tumor models, human tumor cells suspended in PBS (5×10⁷ cells/ml) were subcutaneously implanted to the right flanks (100µl per mouse) of 6 weeks old female nude mice. When average tumor diameter reached 5 mm, viruses suspended in PBS were IT delivered into the tumors (50µl per IT dose). The body weight and tumor volumes were monitored three times per week post virus injection. Blood and tumor samples were collected at indicated time points. Animals were euthanatized according to predetermined criteria (weight loss equal to or exceeding 20% of baseline, tumor burden that equals or exceeds 10% of body weight, development of hind limb paralysis or other signs of neurotoxicity, or focal motor deficits). Mice with ulcerated tumors will be monitored daily for body weight, tumor size and status of ulceration. Spleens were collected when animals were euthanatized. With no specific indications, the doses for IT delivery were 1×10⁶ plaque formation units (PFUs) of HSV-1, 1×10⁶ PFUs of HSV-1+1×10⁵ PFUs of CMV, or 1×10⁵ PFUs of CMV in 50ul PBS per injection. When a ten-fold higher dose was injected for each virus, the doses used were: 1×10⁷ PFUs of KOS R30.4, 1×10⁷ PFUs of KOS R30.44, 1×10⁷ PFUs of KOS R30.44+1×10⁷ PFUs of KOS R30.5, 1×10⁷ PFUs of KOS R30.44+1×10⁶ PFUs of mCMV, or 1×10⁷ PFUs of KOS R30.44+1×10⁷ PFUs of KOS R30.5+1×10⁶ PFUs of mCMV in 50ul PBS per IT injection.

RT-qPCR

Tumors were cut into small pieces and crushed on ice. Total RNA was extracted from around 100 mg tissue per sample using the RNeasy Midi Kit according to the manufacturer’s instructions. Power SYBR™ Green RNA-to-CT™ 1-Step Kit was used for semi-quantifying the viral gene expression levels, and 100 ng of total RNA was added for each reaction. Primers used were listed in Supplementary Table 1 primer No. 120-125. Relative viral gene expression levels (fold change) were analyzed using the ∆∆Ct method and presented as mean ± SEM from replicate samples.

Immunohistochemistry analysis

Tumors collected on day 2 post virus injection were fixed in 4% paraformaldehyde (PFA) for 3 days before being dehydrated with 30% sucrose in PBS at 4°C for additional 2 days. The tumor samples were then subjected to cryosectioning with a Leica CM1860 Cryostat, sickness of tumor slices was set at 40µm. Tumor slices were penetrated with 1% Triton X-100 in TBS buffer and stained with 4′,6-diamidino-2-phenylindole (DAPI) and conjugated antibodies (Supplementary Table 1 Reagent #57-66). Images were collected with a Zeiss LSM 780 confocal microscope and analyzed with the software ZEN 3.3. Images with maximum intensity projection were presented.

Cytokine array assays

On 3 dpt, tumors with similar size (100-120 mm³) were collected from each group. On 9 dpt, tumors with the smallest size in each group were collected. Tumors were crushed and homogenized in PBS with protease inhibitors (10 μg/ml Aprotinin, 10 μg/ml Leupeptin, and 10 μg/ml Pepstatin). Triton X-100 was added to the lysate to a final concentration of 1%. Tumor samples were then freezed at -80°C, thawed, and centrifuged at 10,000×g for 5 min to remove cell debris. Cytokine levels in tumor lysate were detected using the Proteome Profiler Mouse XL Cytokine Arrays and Mouse Cytokine Panel A Arrays according to manufacturer’s instructions. Protein concentrations of samples were determined with BCA assay, and 3 mg and 1 mg of total protein were loaded onto the membranes for 3 dpt and 9 dpt tumor lysate samples, respectively. Cytokine levels in supernatant of cultured human macrophages were detected using the Proteome Profiler Human XL Cytokine Array Kit. Cell culture supernatant samples were loaded onto the membranes with 100 µg total protein per sample. Membranes for each experiment were developed with autoradiography on a single film. Pixel densities of dots on developed X-ray films were scanned using the ImageJ software. Signals between membranes were normalized using the pixel density of reference dots on each membrane. Cytokines that showed more than 1.25× or less than 0.75× fold regulation were displayed.

ELISA analyses

ELISA analyses were performed according to the manufacturer's instructions. ELISA kits were listed in Supplementary Table 1.

Culture of human monocyte-derived macrophages

To generate monocyte-derived macrophages, human peripheral blood CD14+ monocytes in monocyte attachment medium were plated to six-well plates (1×10⁶ cells in 1ml medium per well) and cultured at 5% CO₂ and 37°C for 1h to let the monocytes attach. Adherent cells were washed with pre-warm monocyte attachment medium for three times and then cultured with the macrophage generation medium (RPMI 1640 with 10% human AB serum, 1μg/ml hIL-4, 1μg/ml hIL-10, and 1μg/ml hM-CSF) for 7 days, medium and cytokines were renewed on day 3. Differentiated macrophages were detached with macrophage detachment solution, counted, and seeded to plates in macrophage generation medium for in vitro analyses. Immunofluorescence staining using the conjugate antibodies against human macrophage markers CD206 and CD163 was performed.

Spleen cell expansion and cytotoxicity assays

Spleens were collected when condition of mice meet predetermined euthanasia criteria or at experimental terminal. Freshly collected spleens were crushed in RPMI-1640 Medium, and tissue debris was removed with 70µm cell strainers. Cell suspension passed through the strainers was centrifuged at 1,800 rpm for 5min. Cells were resuspended and incubated in 5ml RBC Lysis Buffer for 5min before being pelleted down at 1,800 rpm. Spleen cell pellets were washed twice with PBS and then stored at -80°C.

To analyze the cytotoxicity of spleen cells against tumor cells, the isolated spleen cells were first expanded with spleen cell expansion medium (TexMACS Medium with 10%FBS, 60ng/ml mIL2, 1µM 2-mercaptoethanol, and 100 units/ml of penicillin/streptomycin) at 5% CO₂ and 37°C until more than 1×10⁸ cells can be obtained for each spleen sample. The expanded spleen cells were then stimulated with target tumor cells in six-well plates. Briefly, tumors cells were seeded to six-well plates (5×10⁵ cells per well) and cultured overnight, then spleen cells were added to the wells (5×10⁶ cells per well) and co-incubated for 4 days, media was renewed every day. On day 4 post co-culture, the suspension cells in each well were collected, counted, and designated as stimulated spleen cells, and the viability of the stimulated spleen cells were determined with trypan blue exclusion. Tumor cells and the stimulated spleen cells were sequentially seeded to 96-well plates (2×10⁴ tumor cells and 2×10⁴ live stimulated spleen cells per well), and were co-cultured in spleen cell expansion medium at 5% CO₂ and 37°C. The control wells were loaded only with 2×10⁴ tumor cells per well. At day 4 post co-culture, cell medium was renewed to remove the suspension cells, and cytotoxicity was analyzed with TACS MTT Cell Proliferation Assay Kit.

Elispot assays

IFN-γ elispot assays were conducted using the Mouse IFN-gamma ELISpot Kit as described previously with modifications ³⁶. Spleens were collected when condition of mice meet predetermined euthanasia criteria or at experimental terminal. Spleen cells suspended in RPMI-1640 Medium (supplemented with 10% FBS, 10 ng/ml mIL2, 1µM 2-mercaptoethanol, and 100 units/ml of penicillin/streptomycin) were loaded to the 96-well plates (1×10⁵ or 5×10⁵ live spleen cells per well). Two technical replicates were set for each sample. Antitumor immunity was analyzed by loading live tumor cells (1×10⁵ tumor cells per well) to the spleen cell pre-seeded wells. Antiviral immunity was analyzed by adding UV-inactivated viral particles (1×10⁶ inactivated particles per well) to the spleen cell pre-seeded wells. After tumor cells or viral particles were loaded, the plates were cultured at 5% CO₂ and 37°C for 3 days before spots were developed and counted. Phytohaemagglutinin P (PHA-P) (100 µg/ml) was added in the positive control groups. To prepare UV-inactivated HSV-1 and CMV particles, purified and titrated infectious virus stocks were UV-inactivated (120 mJ/CM², 30min) with a CL-1000 Ultraviolet Crosslinker. Inactivation of virus was confirmed by titration.

Antiviral neutralizing antibody titration for serum samples

Plaque reduction neutralization test (PRNT) was performed to quantify the anti-HSV-1/mCMV neutralizing antibody titers in the mouse serum samples. Briefly, Vero or NIH-3T3 cells were plated to 24-well plates (2×10⁵ cells per well) and cultured overnight to form cell monolayers. Mouse serum samples were heat-inactivated (56°C for 30 min) and 2-fold diluted beginning with a 1:25 dilution (1:25, 1:50, 1:100, 1:200, 1:400, and 1:800), and were added to 96-well V-bottom plates. A total of 60 PFUs of HSV-1 KOS-BAC or mCMV Smith-BAC virus were added per well. The virus-serum mixtures (150 µl per well) were inoculated at 37°C for 1h before being transferred to the Vero or NIH-3T3 monolayers. Virus without serum was added to the virus-only control wells. After 1h’s absorption, the inoculums were removed and the DMEM/agarose overlay medium (containing 10% FBS, 1% low-melting point agarose, 0.375% sodium bicarbonate, and 100 units/ml of penicillin/streptomycin) was added to the cells. The plates were incubated at 5% CO₂ and 37°C before GFP or dTomato positive plaques were counted for all groups. PRNT₅₀ was calculated with “R 4.2.2” package using the algorithm described previously ⁵⁷.

Single cell RNA sequencing of spleen cells

Spleen cells after expansion and stimulation with tumor cells as described above were counted and subjected to 10X Genomics® Chromium™ 3' gene expression RNA-seq provided by Azenta. Samples must pass the minimal QC requirement (more than 70% cell viability after dead cell removal) for library preparation. Target number of cells per sample was 3,000, and target number of reads per cell was 50,000. Data was analyzed by ROSALIND® (https://rosalind.bio/). Quality scores were assessed using FastQC. Cell Ranger was used to align reads to the Mus musculus genome build GRCm38, count UMIs, call cell barcodes, and perform default clustering. Individual sample reads were normalized via Relative Log Expression (RLE) using DESeq2 R library ⁵⁸, DEseq2 was also used to calculate fold changes and p-values and perform optional covariate correction. The t-distributed stochastic neighbor embedding (t-SNE) Projection of Cells was colored by the cluster identification for each cell-barcode. The axes correspond to the 2-dimensional embedding produced by the t-SNE algorithm. Cells with high variance were removed from this plot.

Statistics

Statistical analyses of data were performed using GraphPad Prism 8 software. Values were presented as mean ± Standard Error of Mean (SEM). All data sets passed Normality test (Shapiro-Wilk test or Kolmogorov-Smirnov test) before being subjected to one-way ANOVA (Dunnett's multiple comparisons test), two-way ANOVA (Tukey's multiple comparisons test or Dunnett's multiple comparisons test), Unpaired t-test (two tailed), or Multiple t-tests to calculate p-values for group comparisons. p < 0.05 was considered significant. ns, not significant, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Author contributions

H.J. and S.J.R. designed the study, interpreted the data, and prepared the manuscript. H.J. constructed the recombinant viruses and stable cell lines, and performed the qPCR, ELISA, PRNT, and IHC analyses. Y.M., E.M., and C.F. prepared and analyzed the recombinant viruses. H.J., R.N., and L.Z. conducted the animal studies. H.J. and E.A. performed the cytokine array analyses. H.J., T.F.C., and N.P. conducted the Elisopt and in vitro spleen cell cytotoxicity analyses. K.W.P. interpreted the data.

Acknowledgements

This work was supported by New Frontiers Program of Vyriad, Inc. (Funding ID: FP00108613).

Competing interests

A patent applicant (Inventors: Haifei Jiang and Stephen J. Russell) was filed in the US Patent and Trademark Office by Mayo Foundation for Medical Education and Research, as application PCT/US2022/048950. The application covers the combination of oncolytic HSV/VZV with HCMV.

Lawler, S.E., Speranza, M.C., Cho, C.F. & Chiocca, E.A. Oncolytic Viruses in Cancer Treatment: A Review. JAMA Oncol 3, 841-849 (2017).
Russell, S.J. & Barber, G.N. Oncolytic Viruses as Antigen-Agnostic Cancer Vaccines. Cancer Cell 33, 599-605 (2018).
Macedo, N., Miller, D.M., Haq, R. & Kaufman, H.L. Clinical landscape of oncolytic virus research in 2020. J Immunother Cancer 8 (2020).
Russell, S.J., Bell, J.C., Engeland, C.E. & McFadden, G. Advances in oncolytic virotherapy. Commun Med (Lond) 2, 33 (2022).
Zhou, J. et al. Tumor-Associated Macrophages: Recent Insights and Therapies. Front Oncol 10, 188 (2020).
Kuny, C.V. & Szpara, M.L. Alphaherpesvirus Genomics: Past, Present and Future. Curr Issues Mol Biol 42, 41-80 (2021).
Peters, C. & Rabkin, S.D. Designing Herpes Viruses as Oncolytics. Mol Ther Oncolytics 2 (2015).
Zhang, Z. et al. Genome-wide mutagenesis reveals that ORF7 is a novel VZV skin-tropic factor. PLoS Pathog 6, e1000971 (2010).
Koyanagi, N. & Kawaguchi, Y. Evasion of the Cell-Mediated Immune Response by Alphaherpesviruses. Viruses 12 (2020).
Chew, T., Taylor, K.E. & Mossman, K.L. Innate and adaptive immune responses to herpes simplex virus. Viruses 1, 979-1002 (2009).
Brinkmann, M.M. et al. Cytomegalovirus immune evasion of myeloid lineage cells. Med Microbiol Immunol 204, 367-382 (2015).
Bayer, C. et al. Human cytomegalovirus infection of M1 and M2 macrophages triggers inflammation and autologous T-cell proliferation. J Virol 87, 67-79 (2013).
Reddehase, M.J. Adverse immunological imprinting by cytomegalovirus sensitizing for allergic airway disease. Med Microbiol Immunol 208, 469-473 (2019).
Adler, B., Sattler, C. & Adler, H. Herpesviruses and Their Host Cells: A Successful Liaison. Trends Microbiol 25, 229-241 (2017).
Patro, A.R.K. Subversion of Immune Response by Human Cytomegalovirus. Front Immunol 10, 1155 (2019).
Yang, Z. & Bjorkman, P.J. Structure of UL18, a peptide-binding viral MHC mimic, bound to a host inhibitory receptor. Proc Natl Acad Sci U S A 105, 10095-10100 (2008).
Verweij, M.C. et al. Modulation of MHC-E transport by viral decoy ligands is required for RhCMV/SIV vaccine efficacy. Science 372 (2021).
Pietra, G. et al. HLA-E-restricted recognition of cytomegalovirus-derived peptides by human CD8+ cytolytic T lymphocytes. Proc Natl Acad Sci U S A 100, 10896-10901 (2003).
Fruh, K. & Picker, L. CD8+ T cell programming by cytomegalovirus vectors: applications in prophylactic and therapeutic vaccination. Curr Opin Immunol 47, 52-56 (2017).
Hanley, P.J. et al. CMV-specific T cells generated from naive T cells recognize atypical epitopes and may be protective in vivo. Sci Transl Med 7, 285ra263 (2015).
Kavanagh, D.G., Gold, M.C., Wagner, M., Koszinowski, U.H. & Hill, A.B. The multiple immune-evasion genes of murine cytomegalovirus are not redundant: m4 and m152 inhibit antigen presentation in a complementary and cooperative fashion. J Exp Med 194, 967-978 (2001).
Cretney, E. et al. m144, a murine cytomegalovirus (MCMV)-encoded major histocompatibility complex class I homologue, confers tumor resistance to natural killer cell-mediated rejection. J Exp Med 190, 435-444 (1999).
Anderson, C.K., Reilly, E.C., Lee, A.Y. & Brossay, L. Qa-1-Restricted CD8(+) T Cells Can Compensate for the Absence of Conventional T Cells during Viral Infection. Cell Rep 27, 537-548 e535 (2019).
Erkes, D.A. et al. Intratumoral Infection with Murine Cytomegalovirus Synergizes with PD-L1 Blockade to Clear Melanoma Lesions and Induce Long-term Immunity. Mol Ther 24, 1444-1455 (2016).
Pataki, Z., Sanders, E.K. & Heldwein, E.E. A surface pocket in the cytoplasmic domain of the herpes simplex virus fusogen gB controls membrane fusion. PLoS Pathog 18, e1010435 (2022).
Cohen, J.I. The varicella-zoster virus genome. Curr Top Microbiol Immunol 342, 1-14 (2010).
Oliver, S.L., Yang, E. & Arvin, A.M. Dysregulated Glycoprotein B-Mediated Cell-Cell Fusion Disrupts Varicella-Zoster Virus and Host Gene Transcription during Infection. J Virol 91 (2017).
Farrell, H.E. et al. Murine Cytomegalovirus Spreads by Dendritic Cell Recirculation. mBio 8 (2017).
Farrell, H.E. et al. Lymph Node Macrophages Restrict Murine Cytomegalovirus Dissemination. J Virol 89, 7147-7158 (2015).
Khairallah, C., Dechanet-Merville, J. & Capone, M. gammadelta T Cell-Mediated Immunity to Cytomegalovirus Infection. Front Immunol 8, 105 (2017).
Thiel, N. et al. The Mouse Cytomegalovirus Gene m42 Targets Surface Expression of the Protein Tyrosine Phosphatase CD45 in Infected Macrophages. PLoS Pathog 12, e1006057 (2016).
Patterson, S.J. et al. T regulatory cell chemokine production mediates pathogenic T cell attraction and suppression. J Clin Invest 126, 1039-1051 (2016).
Kowarik, M.C. et al. CXCL13 is the major determinant for B cell recruitment to the CSF during neuroinflammation. J Neuroinflammation 9, 93 (2012).
Belkacemi, L. & Zhang, S.X. Anti-tumor effects of pigment epithelium-derived factor (PEDF): implication for cancer therapy. A mini-review. J Exp Clin Cancer Res 35, 4 (2016).
Ellermann-Eriksen, S. Macrophages and cytokines in the early defence against herpes simplex virus. Virol J 2, 59 (2005).
Malyguine, A. et al. A modified human ELISPOT assay to detect specific responses to primary tumor cell targets. J Transl Med 2, 9 (2004).
Mendez-Lagares, G. et al. Cytomegalovirus mediates expansion of IL-15-responsive innate-memory cells with SIV killing function. J Clin Invest 131 (2021).
Stahl, F.R. et al. Mck2-dependent infection of alveolar macrophages promotes replication of MCMV in nodular inflammatory foci of the neonatal lung. Mucosal Immunol 8, 57-67 (2015).
Wagner, F.M. et al. The viral chemokine MCK-2 of murine cytomegalovirus promotes infection as part of a gH/gL/MCK-2 complex. PLoS Pathog 9, e1003493 (2013).
Arapovic, J., Lenac Rovis, T., Reddy, A.B., Krmpotic, A. & Jonjic, S. Promiscuity of MCMV immunoevasin of NKG2D: m138/fcr-1 down-modulates RAE-1epsilon in addition to MULT-1 and H60. Mol Immunol 47, 114-122 (2009).
Loewendorf, A. et al. Identification of a mouse cytomegalovirus gene selectively targeting CD86 expression on antigen-presenting cells. J Virol 78, 13062-13071 (2004).
Aguilar, O.A. et al. Mouse Cytomegalovirus m153 Protein Stabilizes Expression of the Inhibitory NKR-P1B Ligand Clr-b. J Virol 94 (2019).
Natarajan, K. et al. Crystal structure of the murine cytomegalovirus MHC-I homolog m144. J Mol Biol 358, 157-171 (2006).
Sgourakis, N.G. et al. A Novel MHC-I Surface Targeted for Binding by the MCMV m06 Immunoevasin Revealed by Solution NMR. J Biol Chem 290, 28857-28868 (2015).
Olsson, J., Kok, E., Adolfsson, R., Lovheim, H. & Elgh, F. Herpes virus seroepidemiology in the adult Swedish population. Immun Ageing 14, 10 (2017).
Masoud, S.J., Hu, J.B., Beasley, G.M., Stewart, J.H.t. & Mosca, P.J. Efficacy of Talimogene Laherparepvec (T-VEC) Therapy in Patients with In-Transit Melanoma Metastasis Decreases with Increasing Lesion Size. Ann Surg Oncol 26, 4633-4641 (2019).
Kaufman, H.L., Shalhout, S.Z. & Iodice, G. Talimogene Laherparepvec: Moving From First-In-Class to Best-In-Class. Front Mol Biosci 9, 834841 (2022).
Saha, D., Wakimoto, H. & Rabkin, S.D. Oncolytic herpes simplex virus interactions with the host immune system. Curr Opin Virol 21, 26-34 (2016).
Yunis, J. et al. Murine cytomegalovirus degrades MHC class II to colonize the salivary glands. PLoS Pathog 14, e1006905 (2018).
Dhatchinamoorthy, K., Colbert, J.D. & Rock, K.L. Cancer Immune Evasion Through Loss of MHC Class I Antigen Presentation. Front Immunol 12, 636568 (2021).
Oliveira, C.C. & van Hall, T. Alternative Antigen Processing for MHC Class I: Multiple Roads Lead to Rome. Front Immunol 6, 298 (2015).
Cornel, A.M., Mimpen, I.L. & Nierkens, S. MHC Class I Downregulation in Cancer: Underlying Mechanisms and Potential Targets for Cancer Immunotherapy. Cancers (Basel) 12 (2020).
Lauer, U.M. & Beil, J. Oncolytic viruses: challenges and considerations in an evolving clinical landscape. Future Oncol (2022).
Paredes, A.M. & Yu, D. Human cytomegalovirus: bacterial artificial chromosome (BAC) cloning and genetic manipulation. Curr Protoc Microbiol Chapter 14, Unit14E 14 (2012).
Li, H., Nakashima, H., Decklever, T.D., Nace, R.A. & Russell, S.J. HSV-NIS, an oncolytic herpes simplex virus type 1 encoding human sodium iodide symporter for preclinical prostate cancer radiovirotherapy. Cancer Gene Ther 20, 478-485 (2013).
Dai, X. & Zhou, Z.H. Purification of Herpesvirus Virions and Capsids. Bio Protoc 4 (2014).
Bewley, K.R. et al. Quantification of SARS-CoV-2 neutralizing antibody by wild-type plaque reduction neutralization, microneutralization and pseudotyped virus neutralization assays. Nat Protoc 16, 3114-3140 (2021).
Love, M.I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550 (2014).

SupplementaryFiguresandLegends.pdf
Supplementary Figures and Legends
SupplementaryTables.pdf
Supplementary Tables
GA.png

Download PDF

Journal Publication

published 03 Nov, 2023

Read the published version in Molecular Therapy →

Version 1

posted

You are reading this latest preprint version

Oncolytic Alphaherpesvirus and Myeloid-Tropic Cytomegalovirus Cooperatively Enhance Systemic Antitumor Responses

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Online Methods

Declarations

References

Supplementary Files

Status:

Journal Publication

Version 1