|Home | About | Journals | Submit | Contact Us | Français|
Naturally occurring strains of Newcastle disease virus (NDV) are currently being investigated in multiple clinical trials for oncolytic cancer therapy in the US and abroad. We have previously reported for the first time the development of recombinant NDVs designed for enhanced cancer therapeutic efficacy. Specifically, we have shown that NDV engineered to express IL-2 generates a robust therapeutic response associated with increased tumor specific T cell infiltration after intratumoral administration in mice. We have now demonstrated that this therapeutic response is dependent on T cells and we have investigated the potential to focus the NDV-induced immune response towards a tumor-associated antigen (TAA) in order to further enhance the inherent therapeutic efficacy of NDV. We found that intratumoral treatments of tumor-bearing mice with recombinant NDV expressing a model TAA elicited an enhanced tumor specific response, resulting in a significant increase in the number of complete tumor regressions as compared to control NDV. Additionally, coadministration of NDV expression a model TAA with NDV expressing IL-2 enhanced the TAA directed response and led to more complete tumor regressions. Our results show that TAA directed immunotherapy by oncolytic recombinant NDV alone or in combination with IL-2 results in an enhanced therapeutic efficacy and warrant consideration in the development of cancer therapies based on the use of oncolytic NDV.
As a result of genetic alterations, tumors express unique antigens that can be recognized by T cells. Tumor associated antigens (TAAs) can be tumor-specific antigens, shared antigens specific from particular cell lineages, differentiation antigens, mutated proteins, antigens derived from oncogenes or viral antigens (1). These antigens are processed and presented by major histocompatibility complex (MHC) class I and MHC class II molecules for priming and activation of CD4+ and CD8+ T-cells (2–4). Numerous studies indicate that the immune system has the potential of eliciting a T cell mediated tumor-specific immune response capable of large tumor destruction. Despite this fact, cancer cells, even when expressing these antigens, may not elicit a robust immune response. In some cases cancer cells escape immune detection because they originate from normal cells and do not activate APCs (5). Non-activated APCs that present unique cancer antigens without co-stimulation could lead to tolerance (6, 7). Specifically, T cells that encounter TAAs in the absence of co-stimulation may become ignorant, anergic, or apoptotic (8, 9). Therefore, although a robust TAA-specific immune response is capable of tumor eradication, the tolerance generated during oncogenic development may lead to large tumor growth (10).
Although TAA tolerance presents a significant challenge to cancer immunotherapy, this barrier can be overcome. Interestingly, several viruses are strong inducers of an immune response and offer an attractive strategy for cancer therapy by stimulating the patient's immune system within the tumor in order to overcome these immunologic barriers. The ability of oncolytic viruses to induce tumor specific immune responses has been well documented, tested in clinical trials (11), and was also recently reported by our group when intratumoral injections of a recombinant NDV caused complete tumor regression in tumor-bearing mice, and protected these mice against subsequent tumor challenge with the same parental tumor cells (12). In our previous report we found that intratumoral injection of NDV/F3aa resulted in complete regression in 20% of mice bearing subcutaneously implanted CT26 autologous colon carcinoma cells. In addition, recombinant expression of IL-2 by NDV/F3aa enhanced the therapeutic efficacy of this treatment and resulted in 60% of the mice undergoing complete regression, consistent with previous reports using HSV expressing IL-2 (13, 14). Mice that underwent complete regression in either treatment were protected from subsequent tumor rechallenge highlighting the induction of a potent and long lasting anti-tumor adaptive immune response. Mice treated with NDV/F3aa-IL-2 also had increased T cell infiltration and enhanced tumor specific adaptive immune responses, correlating with the enhanced therapeutic response. We have now investigated whether expression of a TAA by a genetically engineered NDV could focus the NDV-induced adaptive T cell response towards the TAA and enhance the NDV therapeutic efficacy.
The quality of the immune response to an antigen is dependent on many interrelated factors including tissue distribution, pre-existing tolerance, T-cell repertoire, and immune activation. In order to study the ability of NDV to induce a TAA specific response we have used a tumor model system in which β-galactosidase (β-gal) is expressed by murine CT26 tumor cells. In this model, expression of a β-gal-derived MHC class I H-2 Ld epitope (amino acid sequence TPHPARIGL) by NDV allowed the evaluation of a single CD8 T cell antigenic determinant in enhancing the regression of established CT26 tumors by NDV. This CD8 T cell epitope has previously been reported to be an effective TAA target for the generation of T cells against CT26 colon carcinoma cells expressing β-gal (15). Because β-gal is a xenogeneic antigen, this model is an example of tumors that express unique non-self antigens such as viral antigens, mutated self antigens (16), or frame-shift mutations (17).
A cDNA encoding the β-gal-specific CD8 T cell epitope TPHPARIGL (minigal) was cloned between the NDV P/V and M genes of a full-length infectious clone of NDV (NDV/F3aa) (Fig. 1). We chose the NDV/F3aa backbone, containing a multibasic cleavage site in the F protein of the virus, based on our previous studies showing increased anti-tumor efficacy of this virus as compared to the wild-type parental NDV (B1 strain). For proper expression by the NDV replication/transcription machinery, the minigal epitope was flanked by NDV-specific transcriptional start and stop signals (18). In addition, the epitope was expressed after an ER-insertion sequence which was known to enhance the immunogenicity of the encoded epitope by directing the epitope to the ER, most likely enhancing antigen presentation (19–21). ER-targeted CD8 T cell epitopes are at least as immunogenic as the full length proteins from which they are derived (22). The recombinant NDV/F3aa-minigal was rescued as previously described (18), plaque purified, and the growth kinetics of this virus in A549 cells were comparable to the previously reported recombinant NDV/F3aa-IL-2 (Fig 1b).
We previously reported that oncolytic therapy with recombinant NDV/F3aa expressing IL-2 associated with an increase in the number of T cell tumor infiltrates and in the adaptive antitumor immune response (12). Based on our previous results, recent publications (23), and because IL-2 is a known T cell growth factor, we investigated whether NDV/F3aa-IL-2 oncolytic therapy is dependent on T cells. For this purpose, we tested the therapeutic efficacy of recombinant NDV/F3aa-IL-2 in a nude mouse model. Nude mice are severely deficient in T cells and produce only small numbers of T cells due to a genetic mutation in the Foxn1 gene that causes a developmental deterioration of the thymus (24). Nude mice maintain normal numbers of macrophages, NK cells, antigen-presenting cell (APC) functions, and normal complement activity. BALB/c nude mice were subcutaneously injected with the 5×105 syngeneic CT26 colon carcinoma cells. The injected cells were allowed to establish solid tumors until tumors reached between 5–8 mm in diameter (approximately 12 days). Mice were then treated with 1×107 pfu of recombinant NDV every other day for a total of four intratumoral injections. As seen in Fig. 2, nude mice treated with NDV/F3aa-IL-2 did not undergo a single complete regression and tumor growth was not impaired as compared to PBS-treated nude mice. Same results were obtained in nude mice treated with NDV/F3aa-minigal virus, which in this case was used as a control virus, since the used CT26 cells did not express β-gal. This is in contrast to our previously reported results where 6/10 immunocompetent BALB/c mice underwent complete regression following NDV/F3aa-IL-2 intratumoral treatments and 2/10 mice following NDV/F3aa treatment (12). These results highlight the dependency for T cell function in both treatment groups and illustrate that T cells are critical for therapeutic efficacy.
Knowing that the oncolytic properties in vivo of NDV/F3aa are T cell dependent, we next investigated whether expression of a model TAA by NDV/F3aa could enhance its anti-tumor efficacy after intratumor injections of immunocompetent wild type BALB/c mice. BALB/c mice were subcutaneously implanted with 5×105 CT26.CL25 or CT26.wt syngeneic cancer cells. CT26.CL25 cells represent a tumor cell line derived from CT26.wt cells but expressing β-gal under the long-terminal repeat of the Moloney murine leukemia virus cDNA (25). After approximately 12 days when tumors were between 5–8 mm in diameter, mice were intratumoraly treated with 1×107 pfu of NDV/F3aa-GFP, NDV/F3aa-minigal, NDV/F3aa-IL-2 or PBS every other day for a total of four injections. As seen in Fig. 3, intratumoral injection of NDV/F3aa-minigal resulted in higher percentage of complete regressions in mice harboring CT26.CL25 tumors (60%) compared to treatment with NDV/F3aa-GFP control virus (20%) (p=0.0253). This enhanced response is absent in CT26.wt tumor-bearing mice treated with NDV/F3aa-minigal. Importantly, all of the mice that had complete tumor regressions were protected when rechallenged on the contralateral flank 60 days post tumor treatment initiation with 5 × 105 of the same tumor cells as they original bore (data not shown). Both NDV/F3aa-minigal and NDV/F3aa-GFP treated CT26.wt tumor bearing mice that had underwent complete tumor regressions were capable of inducing protective immunity against CT26.wt cells suggesting that inherent tumor-specific antigens present on this cell line are capable of mediating adaptive immune recognition and challenge protection. This result highlights the induction of a potent adaptive immune response capable of rejecting tumor challenge.
In order to assess the induction of a tumor specific cellular immune response, three mice per group were treated as in Fig. 3 and the tumor draining lymph node (LN) cells were harvested and co-cultured with irradiated CT26.CL25 cells. Supernatants from these cocultures were assayed for IFN-γ release and measured by ELISA. The results show that after three days of coculture with CT26.CL25 cells, LN cells from NDV/F3aa-minigal treated mice produced significantly higher levels of IFN-γ than LN cells from PBS or NDV/F3aa-GFP treated mice (Fig. 4A). However, LN cells from mice treated with NDV/F3aa-IL-2 produced more IFN-γ when cocultured with CT26.CL25 cells than LN cells from mice treated with NDV/F3aa-minigal. The IFN-γ response correlated with the increased percentage of complete tumor regressions following NDV/F3aa-IL-2 treatment compared to NDV/F3aa-minigal treatment (Fig. 3).
In a separate experiment, LN cells were cultured with the β-gal TAA peptide minigal (TPHPARIGL) or with an irrelevant control peptide (AMQMLKETI ) and IFN-γ production was measured from supernatants after 3 days of incubation. As seen in Fig. 4B, LN cells from NDV/F3aa-minigal treated mice had a minigal specific induction of IFN-γ production. This response was undetectable when cultured with the control peptide (data not shown). Again, LN cells from CT26.CL25 tumor bearing mice that were treated with NDV/F3aa-IL-2 also produced a minigal specific immune response which was more potent than this in NDV/F3aa-minigal treated mice.
Next, we wanted to determine whether co-administration of NDV/F3aa-minigal and NDV/F3aa-IL-2 had a synergistic therapeutic effect against CT26.CL25 tumors expressing β-gal 10 mice per group were implanted with CT26.CL25 tumor cells and treated with NDV/F3aa-minigal alone of in combination with NDV/F3aa-IL-2. As control virus, we used NDV/F3aa-GFP. As shown in Fig. 5, treatment of CT26.CL25 tumor bearing mice with NDV/F3aa-minigal again resulted in a statistically significant number of tumor regressions than treatment with control NDV/F3aa-GFP (p=0.0015). Moreover, cotreatment with NDV/F3aa-minigal and NDV/F3aa-IL-2 resulted in 9/10 mice undergoing complete regressions as compared to 5/10 mice treated with NDV/F3aa-minigal alone or 5/10 in the NDV/F3aa-GFP and NDV/F3aa-IL-2 control cotreatment group (p=0.0114). All of the mice that had complete tumor regressions were protected when rechallenged on the contralateral flank 60 days post tumor treatment initiation with 5 × 105 CT26.CL25 cells (Fig 5b). On day 140 five mice were challenged with a syngeneic renal cell carcinoma cell line (Renca) as a control for specificity to CT26 cells. All of the mice CT26.CL25 protected mice (5/5) and naïve control mice (5/5) developed visible tumors between 15–20 days after Renca cell implantation (data not shown).
In order to assess the induction of a tumor-specific immune response by co-administration of NDV/F3aa-minigal and NDV/F3aa-IL-2, LN cells from treated mice were harvested and co-cultured with either CT26.wt or CT26.CL25 cells and supernatants were measured for IFN-γ release. The results show that after three days of coculture with β-gal expressing tumor cells (CT26.CL25), LN cells from mice that were treated with NDV/F3aa-minigal and NDV/F3aa-IL-2 together produced significantly more IFN-γ than treatment with one of these viruses plus the control NDV/F3aa-GFP alone. This synergistic effect was specific for the minigal antigen, as it was not seen when CT26.wt cells were used as stimulator cells (Fig 6A) and it was also observed when LN cells were stimulated only by minigal peptide (Fig. 6B). In order to negate the contribution of IFN-γ by NK cells, we repeated this assay with LN cells depleted of NK cells by Magnetic Cell Sorting and Separation (MACS) and observed similar results (Fig. 6C). Additionally, by FACS analysis we observed internal IFN-γ cytokine staining from CD3+CD8+ T cells 14 hrs post coculture with CT26.CL25 cells (Fig 6D). Our findings suggest that the T cell response to CT26.CL25 tumor cells induced by NDV treatment can be increased by expression of IL-2 and of a known CD8 T cell antigen present on the tumors, resulting in enhanced tumor regression.
Here we investigate the use of reverse genetics to enhance the therapeutic anti-tumor efficacy of NDV by eliciting a TAA focused immune response. We first demonstrate the dependence of T cells for NDV oncolytic therapy in the CT26 murine colon carcinoma tumor model. We previously reported the induction of an adaptive immune response against tumor cells resulting in 20% of CT26 tumor-bearing mice undergoing complete tumor regression following intratumoral injections of NDV/F3aa. Importantly, the induced anti-tumor adaptive response in surviving treated mice was capable of protecting from subsequent tumor rechallenge (12). This is in contrast to the results in Fig. 2 where nude mice subcutaneously implanted with CT26 cells did not undergo tumor regressions upon NDV oncolytic therapy (0/15 mice). Additionally we found that tumor bearing nude mice treated with NDV/F3aa expressing the T cell growth factor IL-2, were also unresponsive to treatment therapy (0/15). While we cannot exclude that the role of T cells in these experiments might be supportive of other cellular functions different from direct lysis of tumor cells, we conclude that the presence of T cells is necessary for anti-tumor efficacy by NDV, and highlights the importance of T cells in NDV oncolytic therapy. This is in contrast to other virus oncolytic agents, such as VSV, which can mediate destruction of tumors in mice in the absence of T cells (26). This is probably due to the more robust cytolitic activity and higher levels of replication of VSV in tumor cells as compared to NDV.
Next we investigated the potential to enhance the T-cell mediated oncolytic efficacy of NDV by expression of a TAA in NDV, which would result in overexpression of this TAA in NDV-infected cells. This approach resulted in enhanced therapeutic efficacy of NDV against CT26 tumors expressing the same TAA. Specifically, intratumoral inoculations of oncolytic NDV expressing a model TAA (minigal) induced an enhanced TAA specific immune response resulting in an increased percentage of regressions of tumors expressing the same TAA (CT26.CL25) (five of ten mice, compared to two of ten mice in NDV/F3aa-GFP control treated mice). Furthermore, tumor-bearing mice that did not express the model TAA (CT26.wt tumor bearing mice) responded similar to the treatment with NDV/F3aa-minigal or with NDV/F3aa-GFP, demonstrating the specificity of the response for the TAA. We also found that coexpression of the T cell growth factor IL-2 and TAA by NDV significantly improved the T cell directed therapeutic response, resulting in nine of ten mice undergoing complete tumor regression (p=0.0114). These mice had a significant increase in their tumor-specific T cell immune response and an increased T cell specific response to the β-gal TAA peptide. This enhanced therapeutic response might be a result of an increase in T cell infiltration within the tumor of mice treated with NDV/F3aa-IL-2, resulting in increased anti-tumor specific adaptive responses, as we reported previously (12) or of a reverse of anergy in T cells upon IL-2 signaling (27). Nonetheless, the increased TAA specific adaptive response upon IL-2 treatment translates into an increased therapeutic response by NDV, consistent with a previous report observed using poxvirus TAA-directed therapy (28). Interestingly, high levels of IFN-γ against CT26.wt cells were produced by CT26.CL25 tumor bearing mice treated with NDV/F3aa-minigal. We speculate that this high level of IFN-γ production can be explained by a broadening effect that others have observed when immunizing against a specific antigen. For example, Markiewicz et al have shown that cured mice vaccinated with p815 tumor peptide rejected a tumor derived cell line that did not express the vaccine peptide (29). Also, mice vaccinated with DNA encoding her-2/neu rejected her-2/neu negative tumors in 50% of mice (30). This phenomenon, referred to as antigen cascade, has also been observed in several other preclinical reports as well as clinical reports (31–34).
NDV possesses numerous characteristics that make it an ideal vaccine vector and is recognized as an effective vaccine vector capable of eliciting a potent immune response targeted to the encoded vaccine antigens. The ability of recombinant NDV vectors to be used as a prophylactic vaccine vector has been harnessed by several groups and utilized for vaccination against respiratory syncytial virus (35), highly pathogenic avian influenza virus (36–38), simian immunodeficiency virus (39), human parainfluenza virus type 3 (40), infectious bursal disease virus (41), as well as severe acute respiratory syndrome-associated coronavirus (42). These studies have found that NDV vectors expressing foreign antigens are capable of inducing protective immunity against several pathogens in multiple animal models including murine, avian, and non-human primates. Importantly, NDV vaccine vectors induced both humoral and cellular immunity against expressed recombinant antigens. Recombinant virus-based immunotherapy is recognized as one of the most potent inducers of TAA directed CTLs (43). These tumor specific CTL mediated responses are traditionally sought in therapeutic cancer development as well as cancer vaccines, and are important in mediating antitumor responses in vivo (44). Oncolytic NDV has demonstrated therapeutic efficacy in several clinical trials and administration of ex vivo NDV-infected oncolysates to cancer patients results in an increased tumor specific CD8+ T cell response and is associated with prolonged survival (11). While generating an antigen specific immune response may more easily translate into prophylactic therapies for the prevention of a number of cancers, the necessity for a therapeutic agent given subsequent to cancer diagnosis is traditionally the focus of cancer therapy. Here we demonstrate the ability to direct a targeted TAA-immune response by recombinant NDV encoding a model TAA. Virus-based cancer vaccines targeting specific TAAs for clinical cancer therapy have entered clinical trials and have reported statistically significant therapeutic responses (45). However, the presence of pre-existing neutralizing antibodies to some of these viral vectors is a major concern for their future development. This concern is less important in NDV-TAA vaccine therapy due to the low prevalence of prior exposure in humans to this avian paramyxovirus. NDV is antigenically distinct from common human pathogens and vaccines, making it immunogenic to the general human population without the concern of pre-existing neutralizing antibodies. However NDV causes hemagglutination and binds to salic acid thereby possibly limiting the effectiviness of intravenous delivery. For this reason we investigated intratumoral injections of oncolytic NDV as our delivery route. Nonetheless intravenous delivery of NDV has demonstrated therapeutic efficacy in both preclinical and clinical studies, and may be a possible route of delivery for some forms of cancer.
Previously we reported for the first time the use of reverse genetics to enhance the cancer therapeutic efficacy of oncolytic NDV. We have now investigated the potential to enhance the therapeutic efficacy of recombinant NDV by TAA directed therapy. NDV encoding a model CD8 T cell TAA epitope elicited a targeted TAA directed adaptive immune response resulting in a statistically significant increase in therapeutic efficacy. Additionally, this response could be further enhanced by cotreatment with NDV expressing IL-2, resulting in 90% of mice undergoing complete tumor regression in our tumor model system. The results presented here are the first to demonstrate the ability of recombinant NDV expressing a TAA to be used as a therapeutic cancer vaccine vector.
The CT26.wt and CT26.CL25 cell lines were generously provided by Dr Nicholas P. Restifo (National Cancer Institute, NIH, Bethesda, MD) and were maintained in RPMI 1640 13 supplemented with 10% heat-inactivated fetal bovine serum (FBS; Sigma-Aldrich), 100 ug/ml streptomycin, 100 ug/ml penicillin, 0.03% L-glutamine, and 400 ug/ml G418. A549 and MDBK cells are maintained in DMEM and REM respectively. Chicken embryo fibroblasts (CEF) cells are prepared as previously described and maintained in MEM. DMEM, MEM, and REM are supplemented with 10% FPS, 100 ug/ml streptomycin, and 100 ug/ml penicillin.
The NDV cDNA sequence was derived from the Hitchner B1 lentogenic strain, which is commonly used as a live attenuated vaccine in chickens and recently utilized in a phase I/II oncolytic clinical trial (46). The recombinant NDV viruses were generated as previously described (47) and sequenced by reverse transcription-PCR for fidelity. Viruses are plaque purified and grown in 10-day old pathogen-free chicken eggs (Charles River Laboratories, SPAFAS). Virus stock preparations are tested for contamination by streaking them on sheep blood agar plates and incubating them at 37° C overnight (data not shown). We engineered rNDV/F3aa to express the model TAA, GFP, and IL-2 from a gene cassette insert as diagramed in Figure 1. Virus preparations were titrated by plaque purification in MDBK cells and growth curves were assessed by immunofluorescence in A549 cells using a polyclonal antibody against NDV.
All procedures involving animals followed NIH guidelines and were approved by and performed according to specific guidelines of the Institutional Animal Care and Use Committee of Mount Sinai School of Medicine. Six-week-old female BALB/c mice were purchased from Taconic Farms and housed in a pathogen-free environment. BALB/c wild type or BALB/c nude mice were subcutaneously implanted with either 5 × 105 CT26.wt or CT26.CL25 cells. When tumors reached a size of 5–8mm in diameter (approximately 12 days) mice were intratumorally treated with PBS or 1×107 pfu of the indicated virus. Tumor volumes were monitored every other day using a digital caliper in two dimensions. Tumor volumes were calculated using the following formula: tumor volume (V) = 4/3 × π × S2/2 × L/2, where S is the smallest measured diameter and L is the larger diameter. Mice were intratumorally injected with 1×107 pfu of recombinant NDV (in 100 µl) every other day for four injections total. Mice in the cotreatment group received 1×106 pfu of the NDV/F3aa-IL-2 or NDV/F3aa-GFP, supplemented with 9×106 pfu of the cotreatment virus in a total volume of 100 µl. Animals were culled when tumor size reached 18 mm in any dimension or at defined experimental time points.
Tumor draining lymph nodes were dissected from 3 mice per group on day 14 post-treatment initiation and manually dissociated into a single cell suspension. 250,000 LN cells were then seeded into 96 wells in triplicate and cocultured with irradiated CT26.wt or CT26.CL25 cells at a ratio of 5:1. Supernatants were collected after 3 days of incubation and release of IFN-γ was assayed by ELISA kit (R&D Systems) following the supplier’s protocol. Peptide specific IFN-γ response was measured by incubating 250,000 LN cells per group with 5ug/ml of either the minigal specific peptide (TPHPARIGL) or a control peptide (AMQMLKETI from HIV). Supernatants were collected after 3 days of incubation and IFN-γ was assayed by ELISA kit (R&D Systems) following the supplier’s protocol. Where indicated, LN cells from two mice per treatment group were depleted of NK cells by MicroBeads separation using DX5 anti-NK Abs according to manufacture’s protocol (Miltenyi Biotec) and assayed for IFN-γ upon coculture with irradiated CT26.CL25 cells for 3 days.
Purified LN cells from two mice per treatment group were cocultured with irradiated CT26.CL25 cells for 14hrs in the presence of GolgiStop (BD Pharmingen). At the end of the incubation, cells were collected, stained with CD3 and CD8 (17A2 and 53–6.7 respectively,BD Pharmingen) and then permeabilized with Cytofix/Cytoperm (BD Pharmingen) followed by anti-IFN-γ staining (XMG1.2, eBiosciences). Data was analyzed by flow cytometry in a Cytomics FC500 machine (Beckman Coulter)
This work was supported in part by NIH fellowship F31CA110209 and F31AI056678 (A Vigil and M.A. Chua respectively), NIH training grant T32AI07647 (A. Vigil), NIH grants (A. Garcia-Sastre) and National Institute of Allergy and Infectious Diseases grant U19AI62623 (Center to Investigate Viral Immunity and Antagonism).
We thank Dr Nicholas P. Restifo for the CT26.wt and CT26.CL25 cell lines and Richard Cadagan for excellent technical assistance.