Enhancing the tumor selectivity of a picornavirus virotherapy promotes tumor regression and the accumulation of infiltrating CD8+ T-cells

Abstract

Picornaviruses have emerged as promising cancer therapies due to their ability to drive cytotoxic cellular immune responses and for promoting oncolysis. These properties include preferential replication in tumor cells, the induction of strong innate and adaptive immune responses and the ease with which their genomes can be manipulated. We have developed Theiler's murine encephalomyelitis virus (TMEV) as an immunotherapy vector that promotes strong adaptive immune responses to tumor antigens embedded within its genome. To further explore its usefulness as cancer therapy we investigated whether direct intratumoral delivery of TMEV could promote tumor regression. We generated several picornavirus hybrids using substrains of TMEV that have unique immunopathologic characteristics, despite their extensive sequence homology. These hybrids exhibit a unique propensity to infect and replicate in melanoma. We have identified GD7-KS1, a virus that is particularly effective at replicating and infecting B16 melanoma in vitro and provides benefit as an oncolytic therapy in vivo after intratumoral injection. In addition, this virus promotes the mobilization and accumulation of CD8+ T-cells within treated tumors. Altogether, these findings demonstrate that picornavirus substrains can be used to rationally design virus hybrids that promote antitumor responses and add to the known strategies identified by us and others to further enhance the therapeutic potential of vectors used to treat cancer.

Introduction

The use of viruses to directly target and kill tumor cells has emerged as a promising treatment for a variety of cancers and has been associated with some remarkable clinical successes (1, 2). This success has emerged from an increased understanding of virus attachment, entry and replication in non-transformed and transformed cells (3, 4). Consequently, much effort has been placed on defining and modifying the molecular attributes associated with virus attachment and replication and understanding how viruses can be manipulated to increase tumor specific killing. A drawback to their use is that receptor expression levels can limit their efficacy and broader use (5-8). Therefore it is important to not only identify new vectors but to identify strategies that best enhance their effectiveness as cancer therapy. Evidence further suggests that the immune system is critical for driving these therapies (9), suggesting that engineering viruses to promote strong immune responses may enhance their effectiveness as well. Questions remain in regards to this role and whether the immune response to oncolytic therapy focuses primarily on virus clearance or whether this therapy can support secondary responses to tumor antigens as well.

Theiler's murine encephalomyelitis virus (TMEV) is a naturally occurring pathogen that displays a similar neurotropism and pathology to human poliovirus in wild-type mice (10). Several TMEV isolates have been identified and are divided into two subgroups, GDVII and TO, based on their virulence after intracranial infection (11). The GDVII subgroup viruses cause a severe acute infection that is often lethal and the TO subgroup causes a less severe encephalitis that often persists in the central nervous system (CNS). Although TMEV is recognized as a model of CNS disease after direct intracranial infection, natural CNS infections are rare and little is known about how this virus migrates to the brain and which organ systems are primarily involved in propagating the virus after natural infection. Although these two substrains share extensive homology, the major differences between these two viruses are within the virus capsid regions, suggesting that cell entry and attachment may differ, a characteristic consistent with the differences in pathology induced using these related virus strains(12-14).

We have developed TMEV as an immunotherapy vector that drives strong T-cell responses to tumor antigens embedded within its genome (15, 16). Our vector was designed using the genetic backbone of the TO subgroup member Daniel's strain (DA), a strain that is most often readily cleared from intracranially infected mice after the development of an immunodominant CD8+ T-cell response (17). These studies identified a strategy for driving cytotoxic T-cell responses that subsequently target tumors and inhibit their outgrowth. This strategy has been shown to effectively inhibit melanoma, breast cancer and glioblastoma outgrowth when delivered systemically (15, 16, 18). Although our studies focused on the immune potential of this vector, it is unclear whether direct infection of tumors with TMEV could provide further therapeutic benefit for the treatment of cancers.

In the current study we examine the oncolytic potential of TMEV for use as a virotherapy vector for breast cancer and melanoma. We have generated a hybrid TMEV vector that integrates virulence determinants from the GDVII subgroup and find that this modification enhances its ability to target and inhibit melanoma outgrowth when used as an intratumoral virotherapy. In addition, we find that direct tumoral injection of GD7-KS1 promotes the accumulation of an increased number of T-cells including a substantial number of CD8+ T-cells that target an immunodominant virus antigen. These results reveal GDVII virus as a dual threat able to both mobilize activated CD8+ T cells and directly infect tumor cells.

Materials and Methods

Tumor lines, cell lines, and animals

The tumor lines EL4, TUBO, 4T1, B16, B16-F10 and B16-OVA were maintained in DMEM supplemented with 10% fetal calf serum (GIBCO Invitrogen, Grand Island, NY). B16, B16-OVA and B16-F10 cells lines were originally obtained from Dr. Richard Vile (Mayo Clinic, Rochester, MN) in 2005 and frozen stocks have been maintained in our laboratory. EL4 and TUBO were obtained from Dr. Keith Knutson (Mayo Clinic, Jacksonville, FL) in 2007 and the 4T1 cell line was obtained from Dr. Haidong Dong (Mayo Clinic, Rochester, MN) in 2014. The fibroblast lines BHK and L929 (ATCC, Manassas, VA) were obtained in 2012 and were used to propogate virus and to titer virus by plaque forming assay were maintained in DMEM-10% FBS. B16-OVA cells were grown in media additionally supplemented with G418 (Life Technologies, Carlsbad, CA). No authentication of the described cell lines was performed by the authors.

C57BL/6 and Balb/c mice were purchased from Jackson Laboratories (Bar Harbor, ME). All animals were housed in the Mayo Clinic Department of Comparative Medicine and cared for according to institutional and NIH guidelines for animals use and care.

Recombinant viruses, in vitro killing, virus quantitation and growth kinetics

Recombinant vectors and viruses were generated using techniques previously described (15). The wild-type Daniel's strain TMEV (DA) used here was generated using the pciDAFL3 vector (15). To generate GDVII chimeric viruses we cloned a KpnI-StuI fragment corresponding to nucleotides 932 to 3941 of GDVII (NC_001366) from cDNA generated from virus supernatant (Dr. Charles Howe, Mayo Clinic, Rochester, MN). This fragment was subsequently cloned into the vector pciDAFL3 to generate the vector encoding GD7-KS1. Subsequently, 3’ and 5’ flanking regions of the capsid coding region (nucleotides 1296 to 3833) were replaced with the DA genome to produce the vectors GD7-P1S1, GD7-KP1 and GD7-P1. All viruses generated from these vectors were subsequently plaqued on L929 cells. Plaque sizes were determined using ImageJ software (19) on scanned images of viral plaques. Virus growth kinetics for the wild-type DA and GD7-KS1 virus were performed on BHK and B16 cells. Viruses were added to wells at an MOI of 0.01 and allowed to adhere for 1 hour at RT. Supernatants were removed and cells were washed twice with media before incubation at 37° Celsius for designated times. Time zero samples were those not placed at 37°. At each time point plates were freeze/thawed twice before sonication and pelleting of cell debris. Supernatants were tested for plaque forming units.

In vitro killing of cell lines and virus quantitation by semi-quantitative RT-PCR were performed as previously described (15).

Tumor experiments and virotherapy

The implantation of B16, B16-OVA and TUBO tumors into the flank of B6 and Balb/c mice was performed as described previously (15, 16). Mice were monitored daily and tumor size was measured every second or third day after virus treatment. Balb/c mice bearing TUBO tumors were injected intratumorally with 5×10⁵ PFU of wild-type TMEV-DA on day 9 after tumor implantation. C57BL/6 mice bearing B16 or B16-OVA tumors were injected with either a single injection of TMEV-DA (6×10⁶ PFU) or with six consecutive injections of 2×10⁵ PFU of GD7-KS1 or GD7-P1 starting on day 7 after implantation. Control animals were injected with PBS or media absent virus. Tumor index was calculated at √w·h. Mice bearing tumors with a diameter in excess of 17 mm were killed in accordance with Mayo IACUC requirements.

Flow cytometric analysis of tumor infiltrating lymphocytes

To analyze the tumor infiltrating lymphocyte populations after administration of GD7-KS1 virus, tumor bearing mice were intratumorally injected with virus on day 7 and day 8. On day 13 tumors were harvested, physically dissociated into a single cell suspension and sieved through a 100 μM mesh filter. After red blood cell lysis with ACK, cells were pelleted and resuspended in FACS buffer for staining of appropriate markers. We stained with mouse CD45-PerCP (BD Pharmingen, San Jose, CA; Clone 30-F11) to discriminate infiltrating lymphocytes from tumor cells. Antibodies to mouse CD8β (Ebiosciences, San Diego, CA; Clone H35-17.2) and mouse CD4 (Ebiosciences; Clone GK1.5) were used to identify T-cell populations. Quantitation of absolute numbers of infiltrating CD8 and CD4 T-cells was performed using CountBright Absolute Counting Beads (Molecular Probes Inc., Eugene, OR). The H-2K^b/OVA8 and H-2D^b/VP2_121-130 tetramers were kindly provided by Dr. Aaron Johnson (Mayo Clinic, Rochester, MN) and have been described previously (17). Samples were run on a BD LSRII flow cytometer (BD Biosciences) and data were analyzed using FloJo Software version 7.6.5 (Tree Star, Ashland, OR).

Statistics

Mean and standard error values were calculated using Excel 2010. All statistical analysis was performed using Sigmaplot for Windows version 11.0. All parametric data were analyzed by t-test or ANOVA with individual comparisons performed using the Student-Newman Keuls test. Survival analysis was analyzed by Kaplan-Meier Log Rank Survival Analysis. Significance was determined by P < 0.05.

Results

Direct intratumoral injection with DA does not significantly inhibit breast cancer or melanoma outgrowth

Since our previous studies demonstrated that a TMEV vector harboring tumor antigens could effectively inhibit both breast cancer and melanoma outgrowth we chose to study these models to discover whether TMEV could provide a therapeutic benefit through direct oncolysis of established tumors. To test this we analyzed tumor growth in two implantable models of cancer, the rat her2/neu expressing breast cancer TUBO and the melanoma tumor B16. After 9 days of growth in vivo breast cancer tumors were treated by intratumoral injection with either the wild-type DA or with vehicle. We found that our strategy failed to inhibit the outgrowth of breast cancer at any of the time points measured when compared to the vehicle control group (Figure 1A). We then tested this strategy on established melanoma tumors using the tumor line B16 and found a modest but insignificant delay in tumor outgrowth and survival in the time points measured (Figure 1B).

Figure 1. The use of Daniel's strain of TMEV as oncolytic therapy fails to inhibit outgrowth of breast cancer and melanoma.

(A) Left: Average tumor index of TUBO breast tumors in Balb/c mice beginning on day of intratumoral injection of TMEV-DA (d. 9). Right: Individual tumor growth over time. No increase in survival due to virus therapy was observed by Kaplan Meier Analysis. Median survival: TMEV-DA – 19 days, Vehicle-Ctrl. – 34 days. (B) Left: Average tumor index of B16 tumors treated with intratumoral TMEV-DA (d.7). NS - Not significant at any timepoint tested (Students t-test). Right: Growth of individual tumors in TMEV-DA and vehicle control groups. No significant survival advantage by Kaplan Meier Analysis was observed. Median survival: TMEV-DA – 22 days, Vehicle-Ctrl. – 17 days.

Engineering DA-GDVII chimeric vectors and the generation of recombinant viruses with unique plaque phenotypes

Having determined that wild-type DA does not provide a significant therapeutic benefit in models of breast cancer or melanoma when used as oncolytic therapy, we chose to design and engineer a series of virus vectors that encoded fragments of the GDVII genome, a substrain previously characterized for its enhanced neurovirulence (10). The initial vector generated consists of the DA backbone vector with a subgenomic fragment of the GDVII virus cloned into KpnI and StuI restriction sites present in both virus genomes (Figure 2A). This fragment consists of the entire capsid coding region of GDVII (P1) along with a portion of the 5’ UTR, the leader protein coding region as well as a 3’ fragment that encodes the non-capsid proteins 2A, 2B and a portion of 2C. Subsequently vectors were generated that contained the P1 region and 3’ flanking region, 5’ flanking region and P1 region and a P1 only region (Figure 2A).

Figure 2. Generation of chimeric viruses composed of genomic elements from the TMEV substrains Daniel's and GDVII.

(A) Cloning strategy and map of DA/GDVII chimeric vectors. A 3 kilobase Kpn I and Stu I restriction fragment was used to generate an initial DA/GDVII vector. Site-directed mutagenesis and directional cloning were used to replace capsid flanking regions with wild-type TMEV-DA sequences. GD7-P1 contains the complete capsid structure of GDVII and all non-capsid encoding elements are derived from TMEV-DA. (B) Virus plaques generated after transfection of virus encoding vectors into BHK cells. (Right) Plaque measurements from each virus. (* Significant by ANOVA p<0.05).

These vectors were transfected into BHK cells to generate the viruses GD7-KS1, GD7-P1S1, GD7-KP1 and GD7-P1 (Figure 2B). Since both capsid and non-capsid regions have been implicated in promoting virulence in the TMEV substrains (20, 21) we hypothesized that changes in these regions would modulate cell tropism, replication efficiency or the lytic potential of this virus. We used plaque morphology on L929 cells as an indicator of infection, viral replication and lytic potential. To determine how each of the GDVII fragments contributes to changes in plaque morphology we measured plaque diameters after 3 days of in vitro infection. When compared to DA all GD7 chimeric viruses demonstrated increased plaque size, demonstrating that components of GD7 capsid and non-capsid sequences contribute to these changes. Further, the GD7-KS1 virus had the largest plaque size, GD7-P1S1 virus had the second largest plaques and the GD7-P1 and GD7-KP1 had smaller plaques than both of the other GD7 chimeras (Figure 2B). These viruses provide a variety of plaque phenotypes for the design and implementation of vectors with modified cellular tropism and potential for enhanced virulence and immunogenicity when used as immunotherapy or as virotherapy.

The GD7-KS1 virus preferentially kills and replicates in melanoma tumor lines

To determine whether the large plaque virus GD7-KS1 or the small plaque virus TMEVDA preferentially replicates in specific tumor lines we tested infectivity in melanoma, breast cancer and lymphoma lines. These viruses killed all melanoma lines tested by 48 hours after infection, however only the DA line killed breast cancer and neither killed the thymic lymphoma EL4 (Figure 3A). Consistent with these results, both viruses replicate their virus genome in melanoma, DA replicated in breast cancer and neither replicated significantly in EL4 (Figure 3B). To determine whether there was preferential replication of either virus in the B16 melanoma line we performed one-step growth kinetics of the two viruses in the cell line BHK which is used to propagate both viruses and in the melanoma line B16. Both viruses propagated to similar levels after infection in BHK cells, in contrast the GD7-KS1 virus propagated to higher titers by 12 hours post infection when compared to DA. In addition, at the 24 hour time point the GD7-KS1 virus had replicated 1.5 logs more virus than DA and the overall yield of virus was 370 PFU/cell compared to 9 PFU/cell with DA (Figure 3C).

Figure 3. GD7-KS1 demonstrates enhanced cytotoxicity towards and increased replication in melanoma.

(A) Melanoma lines (B16, B16-OVA and B16-F10), breast cancer lines (TUBO and 4T1) and the thymic lymphoma line (EL4) were exposed to TMEV-DA or GD7-KS1 for 24 and 48 hours. At the given time points MTT assays were performed to assess percent survival. (B) Relative fold change in VP2 specific virus transcript levels assessed in melanoma, breast cancer and lymphoma at time 0 and 24 hours. Values are relative to uninfected control. (C) Growth kinetics of TMEV-DA and GD7-KS1 in the virus propagating cell line BHK and in B16 melanoma assessed at given time points and expressed as log₁₀ virus plaque forming units. Cells were infected at an MOI of 0.01. PFU/cell were calculated based on plating 10⁶ cells and the total virus yield from 2 mL of media (* Significant by t-test, p<0.05). Data points and standard deviation for A-C represent triplicate independent measurements.

Oncolytic viruses containing the GDVII capsid region inhibit B16 outgrowth after direct intratumoral injection

Since the GD7-KS1 virus preferentially replicates in melanoma we tested whether this virus could delay tumor outgrowth when used as an oncolytic therapy. We find that the GD7-KS1 virus inhibits tumor outgrowth with a significant delay detected 6 days after initial intratumoral delivery of the virus compared to vehicle only controls (Figure 4A). We find that GD7-KS1 treatment provides a survival advantage when compared to DA, with an increase in overall survival of 5 days. To determine the role of the capsid region in promoting this increase we tested oncolytic potential of the GD7-P1 virus to delay tumor outgrowth. Similar to the result obtained with GD7-KS1 we find that GD7-P1 inhibits outgrowth of melanoma and increases survival by 7 days (Figure 4B), thus demonstrating that the increased effectiveness of this therapy over wild-type DA is likely dependent on incorporation of the GDVII capsid region.

Figure 4. GDVII capsid containing viruses delay tumor outgrowth and promote increased survival when used as oncolytic therapy.

(A) Tumor bearing mice were treated with GD7-KS for six days beginning on day 8. Tumor index was calculated until overwhelming tumor burden was reached (n=5/group). (Right) Individual tumor growth curves through completion of the study. The median survival was 18 days for vehicle control and 23 days for GD7-KS1 treated animals (* Significant by Log-Rank, p=0.021). (B) Tumor bearing mice were treated with the GDVII capsid only virus GD7-P1 for 6 days beginning on day 7 and tumors were monitored as above (n=5/group). (Right) Tumor growth curves for individual animals. The median survival was 18 days for vehicle control and 25 days for the GD7-P1 treated group (* Significant by Log-Rank, p=0.003).

GD7-KS1 virus promotes the accumulation of activated CD8+ T-cells specific for virus and tumor antigen

Since the TMEV virus GD7-KS1 promotes tumor regression after direct intratumoral delivery and directly kills B16, we asked whether infection of the tumor promotes the mobilization and accumulation of T-cells within tumors that had responded to GD7-KS1 therapy (Figure 5A). Using dissociated tumor tissue, we stained for CD8 and CD4 to determine the percentage and number of T-cells that had infiltrated into the B16 tumor. We found that this therapy increased the percentage and absolute number of CD8+ T-cells within the tumor, but did not alter the overall number of CD4+ T-cells (Figure 5B). To further characterize this population we tested for the presence of CD8+ T-cells that recognize the immunodominant virus antigen VP2. We found that over 30% of the infiltrating CD8+ T-cells were specific for the immunodominant H-2D^b antigen VP2_121-130 (Figure 5C), demonstrating that therapy drives a strong influx of CD8+ T-cells including a large percentage that are specific for the immunodominant VP2 peptide.

Figure 5. Intratumoral delivery of GD7-KS1 promotes the accumulation of CD8+ T-cells.

(A) Tumor index of B16 bearing hosts treated with intratumoral vehicle or GD7-KS1. Tumors were measured until tumor harvest (* Significant by t-test, p<0.05). (B) Percentage (top) and absolute number of CD8+ and CD4+ T-cells recovered from B16 host tumors treated with vehicle or GD7-KS1 in (A) (Significant by t-test, p<0.05). (C) Tumor infiltrating CD45+ gated cells were assessed for CD8 and tetramer for VP2 antigen (* Significant by t-test, p<0.05).

Although the GD7-KS1 treatment did show an enhancement in the accumulation of CD8+ T-cells within the tumor we wanted to test this therapy using the immunogenic melanoma model B16-OVA to determine if virotherapy modulates tumor-specific immunity in tumors responding to virotherapy (Figure 6A). The average tumor index for the control tumors at harvest was 9.3+1.4 and was 7.7+0.4 for tumors treated with GD7-KS1 (p<0.05 by Student's t-test). We analyzed the percentage and total number of CD8+ and CD4+ T-cells that had infiltrated these tumors. We found that there was an increase in the percentage of CD8+ T-cells but not CD4+ T-cells on day six after treatment and the ratio of CD8+ T-cells to CD4+ T-cells was increased with GD7-KS1 treatment (Figure 6B). Since this tumor expresses the immunogenic protein ovalbumin we tested for the presence of tumor specific CD8+ T-cells by using H-2K^b/SIINFEKL tetramers. We found that the percent and absolute number of H-2K^b/SIINFEKL specific cells was not different between the groups (Figure 4B), however the cell surface expression level of CD8 on the CD45 population was decreased in the GD7-KS1 treated group (Figure 4D), a marker for activation (22), demonstrating that GD7-KS1 modulates the tumor environment in a manner that promotes tumor regression and tumor-specific T-cell activation.

Figure 6. GD7-KS1 promotes the activation of tumor specific CD8+ T-cells.

(A) Tumor index of B16-OVA tumors implanted into B6 hosts treated with control vehicle alone or GD7-KS1 virus. Tumors were measured until harvest. (B) Tumors from (A) were dissociated and analyzed by FACS. Total CD45+ cells were gated and analyzed for the percentage of CD8+ and CD4+ cells within dissociated tumors. Ratio of CD8 cells to CD4 cells comparing control to GD7-KS1 treatment. (C) Percent of CD8+ T-cells specific for the tetramer H-2K^b-SIINFEKL. Cells within plot are from CD45+ gate. (D) Histogram of CD8 staining comparing vehicle to GD7-KS1 and mean fluorescence intensity of groups. * Significant by t-test, p<0.05.

Discussion

The goal of this work was to determine the efficacy of an engineered picornavirus vector as an oncolytic therapy. In the absence of an appreciable affect using TMEV-DA, we sought to identify a strategy for enhancing the potential of the picornavirus as an oncolytic. Using two substrains of TMEV we generated a chimeric virus that contains genomic material from both the DA strain and the GDVII strain. The chimeric virus GD7-KS1 showed an increased ability to kill melanoma in vitro and in vivo. In addition, we find that the introduction of capsid coding regions primarily account for this increase in efficacy, implicating enhanced virus infectivity and replication in B16 melanoma as the driver of enhanced therapy. Although direct virus killing of the tumor may account for this therapeutic affect we find that the virus increases the accumulation of virus specific CD8+ T-cells and promotes activation of tumor-specific CD8+ T-cells. These findings demonstrate that TMEV can be rationally designed to target an established tumor and that oncolytic therapy using this vector modulates the immune cell infiltrate associated with tumor regression.

Although precise receptors for TMEV substrains have not been identified, infectivity of the low pathogenic strains like DA is associated with sialic acid moieties on N-linked oligosaccharides (23) whereas GDVII attachment is thought to use heparan sulfate as a co-receptor (24). Tumor specific restriction of picornaviruses has consequently focused on modulating replication once the virus has entered the target cell. Several strategies including the incorporation of microRNA targets or tumor antigens and modification of non-capsid regions have been used to restrict virus replication (25-27). Although these strategies are effective, they may also decrease replication capacity, promote less efficient infectivity or tumor killing and may disrupt the genomic stability of the vector (15). We find that the introduction of the GDVII capsid region enhances melanoma-specific killing of GD7-KS1, suggesting that it replicates more efficiently in B16 melanoma and is more cytolytic in vivo than DA, consistent with what we find in vitro. Similar to previous findings (21), we find that the virus assembly machinery from the DA strain can package and assemble GDVII capsids and these virus hybrids have increased replication capacity and assembly upon infection.

The association with sialylation or surface polysaccharides and TMEV attachment to cells may provide an opportunity for specific targeting of non-human pathogens to human cancers (28-30). Variations in cell surface glycosylation are observed in various cell types, providing unique functions to cells and potentially making them more or less susceptible to virus infection (31, 32). Cancer metastasis and progression are often associated with altered regulation of post-translational modifications including glycosylation (33, 34). This provides an opportunity for designing therapeutics to target these patterns, including viruses that preferentially target human cancers through their dysregulated glycosylation. Although a specific receptor has not been identified for TMEV, much is known about the virus capsid structure and the specific amino acids within these structures that contribute to virus attachment and cell entry (23, 35). Several virus mutations within the outer capsid regions have been identified and these are often associated with virus attachment (36, 37). Our results here demonstrate that the GDVII capsid proteins can be interchanged with DA and provide a more specific targeting of melanoma.

The use of viruses for the treatment of cancer has led to the development of vectors that not only target tumors through direct lysis but have also been used as tools to directly stimulate immunity against tumors. Although direct tumor killing by viruses is often the goal, the mechanisms for protection from tumor outgrowth are often linked to the immune system (38, 39). We find that intratumoral delivery of GD7-KS1 promotes the accumulation of CD8+ T-cells specific for the virus antigen VP2 at the tumor site. This immunodominant response is critical for virus clearance (40), providing a basis for CD8 mediated killing of virus infected melanoma cells. In addition to the virus specific T-cells that accumulate after infection, we find that tumor specific CD8+ T-cells are present prior to treatment however there activation status is diminished compared to CD8+ T-cells acquired after intratumoral virus. This suggests that the development of immunity to virus and the accumulation of virus specific T-cells provides a proinflammatory environment and provides the potential to overcome the immunosuppressive tumor microenvironment. Studies to explore the role of viruses in overcoming factors that include T-cell exhaustion, myeloid derived suppressor cells and T-regulatory cells will be important for understanding the underlying immunologic mechanisms that drive oncolytic immunotherapy.

Unlike other viruses, the replication, translation and assembly of picornaviruses occurs exclusively in the cytoplasm of infected cells (41). In addition, their genomes are translated as one polyprotein that is cleaved by specific proteases embedded within this long protein. This makes the rational design of picornaviruses challenging since the virus replication and assembly machinery is most often encoded within the carboxyl-terminus and must specifically recognize cleavage, assembly and replication signals that may be unique to a particular virus. Nevertheless, several picornavirus hybrids have been described. One focus of this work has been on the interchange of 5’ UTR sequences between several picornavirus strains that aid in attenuating their replication or restricting cell specific replication (42-45). Although these hybrids have shown demonstrated cell type and target specificity, the ability to attach and enter specific cell types is unaltered using this approach. Alternatively, our findings here demonstrate that virus tropism can be altered through the introduction of related virus substrain sequences similar to results obtained with coxsackie B3 capsid variants (46). In addition, we find that manipulation of TMEV capsid sequences promotes enhanced targeting of melanoma using this approach and that direct delivery of a hybrid virus enhances its potential as an oncolytic therapy.

The use of viral oncolytics has shown promise in clinical settings, however the precise contributions of direct virus infection, infection of host tissues, immune responses to the virus or to the tumor are not fully understood. We sought to determine whether the direct infection of tumors with the picornavirus TMEV could inhibit tumor outgrowth. We found that TMEV substrains provide useful tools for engineering tumor specific vectors that can effectively target and inhibit melanoma. This virus provides a unique set of properties that will allow us to further investigate its use as on oncolytic therapy and studies using TMEV will provide further mechanistic insight to the role of infection and immunity and how they contribute to tumor regression.