C57BL/6 mice were obtained from Charles River Laboratories (L’Arbresle, France) or Jackson Laboratory (Sacramento, CA, USA). At the start of the experiments, mice were 6 to 8 weeks old. Mice were housed in individually ventilated cages (IVC) under specific pathogen-free (SPF) conditions in the animal facility of Leiden University Medical Centre (LUMC, The Netherlands) or Oregon Health and Science University (OHSU, Oregon, USA). All animal experiments were approved by the institutional Animal Experiments Committee and were executed according to the institutional animal experimentation guidelines and were in compliance with the guidelines of the Institutional, European and Federal USA animal care and usage committees.

Viral titers of virus stocks or infected tissues were determined by plaque assays as previously described [25]. MCMV-IE2-E7 (RAHYNIVTF) and MCMV-IE2-OVA (SIINFEKL) were generated as described [26]. MCMV-IE2-E6/E7 full length fusion, MCMV-IE2-E6/E7 full length replace, MCMV-M79-FKBP-E7 (E7 epitope RAHYNIVTF, IE2 fusion), MCMV-M79-FKBP-gB (gB epitope SSIEFARL, IE2 fusion) were constructed as follows. MCMV BACs were maintained in SW105 E. coli and modified by metabolic galactokinase (galK) selection [27], essentially as described [28]. In brief, bacteria containing the MCMV backbone were first transformed with a PCR product of galK/kanamycin (kan) and flanking homology to target a specific region in the MCMV BAC. Transformed bacteria were selected for kan resistance, grown and subjected to a second transformation to replace the galK/kan cassette with the final insert. These included annealed oligos used for IE2 peptide fusion or the PCR product for M79-FKBP [29] and HPV E6/E7 insertions containing the desired modification with the same MCMV flanking homology to insert the galK/kan cassette. Recombinant bacteria were counter-selected on chloramphenicol 2-deoxy-galactose (DOG) minimal media plates with glycerol as the carbon source. MCMV BAC constructs were characterized by restriction digest, PCR screening, and Sanger and NGS sequencing. Virus was reconstituted by either Lipofectamine 3000 (ThermoFisher Scientific) transfection or electroporation (250 V and 950 uF) of NIH 3T3s.

Tissue culture-derived stocks of the MCMV vectors were amplified and titered in NIH 3 T3 cells grown in complete growth media (DMEM, FBS, PSG). FKBP-tagged viruses were grown in complete growth media supplemented with Shield-1 at a final concentration of 1 uM and added every 48 h [29]. Cell free virus was obtained from supernatant of infected cells, clarified at 3.000 rpm for 20 min and virus was pelleted at 24.000 rpm for 1 h through a sorbitol cushion (10% D-sorbitol, 0.05 M Tris pH 7.4, 1 mM MgCl₂). Virus pellet was resuspended in PBS. For virus quantification, plaque assays were performed in 24-well plates by infection with appropriate serial virus dilution in 0.2 mL of media and then incubated at 37 °C for 2 h rocking. Following incubation, the infected cells were overlaid with 1 mL complete media supplemented with carboxymethylcellulose. After 5 to 6 days, the cells were fixed in 3.7% formaldehyde in PBS and stained with 0.001% aqueous methylene blue. The plaques were counted by light microscopy. Multi-step virus replication curves were performed in NIH 3 T3 cells at MOI 0.1 in 6 well plates, 3 replicates per virus per time-point. Virus was incubated at 37 °C for 2 h, washed 3 times with PBS and then 2 mL of media was added. Supernatant was harvested at 1, 3, 5, and 7 days post-infection, stored at − 80 °C and titered by plaque assay. FKBP-tagged viruses were grown in complete growth media supplemented with Shield-1 at a final concentration of 1 uM and added every 48 h.

The tumor cell line TC-1 (a kind gift from T.C. Wu, John Hopkins University, Baltimore, MD) was generated by retroviral transduction of C57BL/6 lung epithelial cells with the HPV16 E6/E7 and c-H-ras oncogenes [30] and cultured as previously described [31]. The tumor cell line C3 was developed by transfection of mouse embryonic cells with the HPV16 genome and an activated-ras oncogene and maintained as previously described [32]. The MC38-OVA tumor cell line is generated by a retroviral infection of the MC38 parental cell-line with PMIG/MSCV-IRES-GFP plasmid encoding cytoplasmic bound OVA [33]. Iscove’s Modified Dulbecco’s Media (IMDM) (Lonza, Basel, Switzerland) supplemented with 8% fetal calf serum (FCS) (Greiner), 2 mM L-glutamine (Life Technologies, Carlsbad, CA, Unites States), 50 IU/ml Penicillin (Life Technologies) and 50 μg/ml Streptomycin (Life Technologies) was used to culture tumor cell lines. Cells were cultured in a humidified incubator at 37 °C and 5% CO₂. Mycoplasma tests that were frequently performed for all cell lines by PCR were negative.

Treatment schedule of experiments are indicated in the respective figures and legends. Mice were vaccinated with MCMV vectors via the intraperitoneal (IP), intranasal (IN) or subcutaneous (SC) route with the indicated inoculum size. In tumor experiments, mice were inoculated subcutaneously in the flank with 0.25–1 × 10⁵ TC-1 tumor cells, 5 × 10⁵ C3 tumor cells or with 2.5 × 10⁵ MC38-OVA in 200 μl PBS containing 0.2% BSA on day 0. Tumor size was measured two times a week using a caliper. Mice were euthanized when tumor size reached > 1000 mm³ in volume or when mice lost over > 20% of their total body weight (relative to initial body mass).

CD8 T cell depleting monoclonal antibodies (clone 2.43) were purchased from Bio-X-Cell (West Lebanon, NH, United States) and administered IP twice weekly (200 μg/mouse) for 2–3 weeks. CD8 T cell depletion was started 4 days before tumor challenge. Depletion was checked by staining for CD3 and CD8 marker expression followed by flow cytometric analysis.

Blood collection and processing was performed as described [34]. Cells were re-suspended in staining buffer (PBS + 2% FCS + 0.05% sodium azide) and incubated with various fluorescently labelled antibodies detecting CD8 (clone 53–6.7), CD62L (clone MEL-14), CD44 (clone IM7), KLRG1 (clone 2F1), CD3 (clone 500A2), CD127 (clone A7R34). Antibodies were obtained from eBioscience (San Diego, CA, United States), BD Biosciences (San Jose, CA, United States) and Biolegend (San Diego, CA, United States). For dead cell exclusion, 7-Aminoactinomycin D (Invitrogen, Carlsbad, CA, United States) was used. To measure the MCMV-specific and tumor antigen-specific T cell responses, PE and APC-labelled class I-restricted multimers (tetramers or dextramers) with the peptide epitopes OVA_257–264 (SIINFEKL), HPV E7_49–57 (RAHYNIVTF), MCMV M45_985–993 (HGIRNASFI), MCMV M38_316–323 (SSPPMFRV) were used. Tetramers were produced as described [35] and dextramers were obtained from Immudex. Samples were analyzed with a BD LSRII or LSRFortessa flow cytometer, and results were analyzed using FlowJo software (Tree Star, Ashland, OR, United States).

Splenocytes of naïve CD45.1 (Ly5.1) mice were isolated and loaded with MHC class I restricted-E7_49–57 or OVA_257–264 peptides for 1 h at 37 °C at the final concentration of 1 μg/ml. After extensive washing, cells which were pulsed with specific peptide or irrelevant peptide were labelled with high and low concentrations of CFSE (Invitrogen), respectively. Next, 5 × 10⁶ peptide pulsed cells were pooled and injected intravenous (IV) via the retro-orbital route into the mice that have been vaccinated 70 days earlier with MCMV-based vaccines. After 24 h, spleens of the recipient mice were isolated, stained for CD45.1, and subjected to flow cytometry. The cytotoxic capacity was calculated relative to naïve mice by using the following formula: [100 – ((percentage of pulsed peptide in infected mice/percentage of unpulsed peptide in infected mice)/(percentage of pulsed peptide in naive mice/percentage of unpulsed peptide in naive mice)] × 100.

Splenocytes of individual vaccinated mice were isolated and cultured on ELISpot plates (R&D systems, Minneapolis MN) pre-coated with anti-IFN-γ antibodies. The indicated peptides (E7_49–57, RAHYNIVTF; or a pool of peptides spanning the E7 sequence, 15-mers overlapping by 11) were added at 1 μg/ml. 36 h later cells were removed and secondary antibodies added. Plates were developed according to the manufactures instructions. Spots were counted using an AID ELISpot reader.

Statistical analyses were performed using GraphPad Prism (La Jolla, CA, Unites States). Survival data were analyzed by Kaplan-Meier and the log-rank (Mantel-Cox) test. Statistical significance was determined by Mann Whitney. P-values of ≤0.05 were considered statistically significant.

To determine effective regions to insert foreign sequences, we generated recombinant MCMV-based vaccine vectors containing the entire E6 and E7 antigens of human papilloma virus type 16 (HPV16). In these vectors the E6/E7 genes replaced exon 3 of IE2 (MCMV-IE2-E6/E7 full length replace) or E6/E7 was placed in the end part of the IE2 region (MCMV-IE2 E6/E7 full length fusion) (Fig. 1a). To analyze the T cell responsiveness to the inserted foreign sequences induced by these vaccine vectors, we IP vaccinated mice and measured the E7-specific T cell responses by IFN-γ ELISPOT, following restimulation with the immunodominant peptide epitope E7_49–57 or with a pool of overlapping peptides covering the whole E7 protein, and by staining with MHC class I multimers. Immunization with MCMV-IE2-E6/E7 full length fusion induced significantly higher E7-specific T cell responses (~ 0.5–2% of the total CD8⁺ T cell population) than MCMV-IE2-E6/E7 full length replace (< 0.3%) at 10 and 21 weeks after vaccination (Fig. 1b and c). To assess the anti-tumor immunity induced by these recombinant viruses we challenged the vaccinated mice with TC-1 tumor cells, which are transformed by expression of the HPV16 E6 and E7 oncoproteins [30]. Vaccination with MCMV-IE2-E6/E7 full length fusion prevented tumor progression in most tumor-challenged mice, while the majority of the mice vaccinated with MCMV-IE2-E6/E7 full length replace developed tumors (Fig. 1d). Thus, consistent with our previous report we found that fusion of foreign sequences to the C-terminus of the IE2 protein works superior in eliciting T cell responses, and this event relates to higher amounts of the peptide epitope presented in peptide-MHC complexes [26].Fig1Fig. 1

Position and epitope context of heterologous antigens in MCMV-vectored vaccines impact vaccine-elicited T cell responses and associated tumor protection. a Schematic of MCMV-IE2-E6/E7 full length replace/fusion and MCV-IE2-E7. The MCMV genome region at ∼kb 186–187 corresponds to the MCMV gene ie2 (enlarged). Constructs encoding the E6/E7 fusion protein or the E7 epitope were generated by BAC recombineering. DNA sequences encoding the full-length E6/E7 fusion protein were inserted either to replace the ie2 gene (full length replace) or in frame at the very end of the ie2 gene (full length fusion). The E7_49–57 epitope RAHYNIVTF was fused to the end of ie2. b, c C57BL/6 mice were vaccinated IP with 1 × 10⁵ PFU MCMV-IE2-E6/7 full length fusion, MCMV-IE2-E6/7 full length replace or MCMV-IE2-E7. Splenocytes were harvested 10 (b) or 21 (c) weeks after vaccination and stimulated with the immunodominant E7_49–57 peptide or with a pool of 15-mer peptides covering the entire E7 protein for 36 h. IFN-γ production was measured by ELISpot assay. Shown is the number of spots per 5 × 10⁵ cells ± SEM. c Frequency of E7_49–57-specific T cells identified by MHC class I multimers within the total CD8⁺ cell population at 21 weeks post vaccination. Data represents mean values ± SEM (n = 4–5 mice per group). *, P < 0.05; **, P < 0.01. d Tumor outgrowth of TC-1 tumor cells in mice challenged with TC-1 tumor cells 21 weeks after vaccination with 1 × 10⁵ PFU MCMV-IE2-E6/7 full length replace, MCMV-IE2-E6/7 full length fusion or MCMV-IE2-E7 (5–10 mice per group). The number of tumor-free/total mice is indicated in each tumor out growth graph

To investigate the impact of the route of inoculation on the application of MCMV-based vaccine vectors for vaccine-elicited immunity against cancer, we compared the magnitude and kinetics of the antigen-specific CD8⁺ T cell response in the blood upon IP, IN and SC inoculation. The E7-specific CD8⁺ T cell responses elicited with MCMV-IE2-E7 increased gradually to ~ 5% after IP immunization, while the responses were steadily ~ 0.6% after SC inoculation. Upon IN inoculation however, the E7-specific CD8⁺ T cell responses were minute throughout infection (< 0.3%) (Fig. 2a and b). The M38_316–323 epitope-specific CD8⁺ T cell response after IP and SC inoculation showed similar magnitude and kinetics as the response to the E7 epitope in the same vector (Fig. 2a and b). Upon IN immunization, the M38-specific CD8⁺ T cell response was barely detectable and did not show signs of inflation. Upon immunizations via the IP and SC routes, the non-inflationary response to the M45_985–993 epitope showed the typical response pattern of rapid expansion followed by swift contraction and stable memory formation (Fig. 2a and b). IN vaccination induced only ~ 0.1% M45-specific CD8⁺ T cells, and during the memory phase this response was close to the detection limit (Fig. 2a and b).Fig2Fig. 2

Vaccination with recombinant MCMV vectors via the IP and SC route induces potent specific T cell responses. Mice were vaccinated with MCMV-IE2-E7 via IP, SC or IN route or kept unvaccinated (control), and the CD8 T cell responses were examined in blood. a Representative flow cytometry plots for MHC class I multimer staining to E7, M38 and M45 epitopes in blood of MCMV-IE2-E7 infected mice via IP during acute phase (day 8 post vaccination) and memory phase (day 64 post vaccination). Numbers represent the percentage of Ag-specific CD8⁺ T cells within the total CD8⁺ T cell population. Control flow cytometry plots show the staining for E7 multimer in unvaccinated mice. Flow cytometry plots show similar numbers of total cells in each plot except for control. b Frequency of Ag-specific CD8⁺ T cells to E7, M38 and M45 epitopes over time. c Frequency of effector-memory (EM) cells (KLRG1⁺, CD62L⁻) within E7, M38 and M45 MHC multimer⁺ CD8⁺ T cells over time. Data represents mean values ± SEM (n = 7–8 mice per group) and are representative of two independent experiments with similar results

Although CMV has the capacity to re-infect the host despite the presence of CMV-specific T and B cell responses elicited upon primary infection, CMV-specific immunity clearly controls viral pathogenesis of primary and secondary infections as shown by the fact that CMV mainly causes disease in immature or immunocompromised situations, and pre-existing immunity strongly limits viremia upon secondary infection [11, 12]. To address the impact of pre-existing MCMV immunity on vaccine-induced CD8⁺ T cell responses and hence the induction of protective immunity, we examined the kinetics of the MCMV vaccine-specific T cell responses in different settings of pre-existing immunity. To mimic the condition of pre-existing immunity in the TC-1 tumor challenge setting, we infected mice first with MCMV-IE2-OVA, which provokes pre-existing immunity but does not elicit tumor-specific T cell immunity. After the establishment of pre-existing immunity, mice were vaccinated with MCMV-IE2-E7. This experimental setup allowed us to follow the antigen-specific T cell response following the first infection and after the subsequent immunization, in which a) the OVA-specific T cell response is primary after the infection and not boosted, b) the M38-specific CD8⁺ T cell response is primary after the infection and then boosted by the immunization, and c) the E7-specific T cell response (which is tumor-specific in case of TC-1 tumor challenge) is not induced during the first infection, hence is primary upon the immunization. Here, we confirmed that IP infection with MCMV-OVA induces higher OVA and M38-specific CD8⁺ T cell responses (~ 5% OVA and ~ 1.5 M38) in blood (Fig. 6a) compared to infection via IN (~ 0.5% OVA and ~ 0.4% M38). Next, we investigated the E7 (tumor)-specific T cell response in these high (IP) and low levels (IN) of pre-existing immunity. Vaccination with MCMV-IE2-E7 via the IP and SC route boosted the M38-specific T cell response (~ 3% M38 compared to 1.5 and 0.5% after primary infection) and this increase was more pronounced in the mice with low pre-existing immunity compared to high pre-existing immunity. Importantly, a higher E7-specific T cell response was also determined in the mice with low pre-existing immunity (0.7–1% versus 0.3%) (Fig. 6a). The CD8⁺ T cell response against the OVA_257–264 epitope did not change upon subsequent immunization with MCMV-IE2-E7. Together, these data suggest that super-infection occurs more readily upon primary inoculation via the IN route versus the IP route. Moreover, the level of pre-existing immunity affects the magnitude of the T cell responses elicited to MCMV vectored antigens.Fig6Fig. 6

Pre-existing immunity against the MCMV vector affects the strength of T cell responses and anti-tumor efficacy. a Percentage of OVA, M38 and E7-specific CD8⁺ T cells, identified using MHC class I multimers. Mice were infected with 1 × 10⁵ PFU via IP or IN MCMV-IE2-OVA at day 0. After 35 days, the mice were vaccinated with 1 × 10⁵ PFU MCMV-IE2-E7 via IP route or with 5 × 10⁵ MCMV-IE2-E7 via the SC route as shown in the schematic. T cell responses were followed for 70 days in blood. Data represent mean values ± SEM. b and c Tumor outgrowth (b) and survival (c) of mice infected with 1 × 10⁵ PFU MCMV-IE2-OVA via IP or IN and vaccinated with 1 × 10⁵ PFU MCMV-IE2-E7 via IP or IN route or with 5 × 10⁵ MCMV-IE2-E7 via SC as shown in the schematic. After 70 days, all the vaccinated and control mice were challenged with 1 × 10⁵ TC-1 tumor cells. The number of tumor-free/total mice is indicated above each tumor outgrowth graph. d The percentage of E7-specific CD8⁺ T cells within the total CD8⁺ T cell population in the blood (y-axis) plotted against the E7-specific CD8⁺ T cell response ranked from low to high (x-axis) of the mice that were either tumor-free (tumor-) or tumor positive (tumor+). Data are representative of two independent experiments with similar results

Non-attenuated CMV may cause disease in immunosuppressed individuals. To facilitate the translation of CMV-based vaccine vectors into the clinic, we generated an attenuated E7-epitope-expressing MCMV vector by fusing the FKBP-degradation domain to the essential M79 protein (MCMV-M79-FKBP-E7), which restricts viral replication to a single-cycle in absence of the stabilizing ligand Shield-1 similar to corresponding HCMV FKBP constructs targeting the homologous UL79 gene (Additional file 2: Figure S2). Vaccination with the spread-deficient MCMV-M79-FKBP-E7 elicited E7-specific CD8⁺ T cell responses that were > 0.5% of the total CD8⁺ T cell population at 4 weeks and even up to 13 months after vaccination (Fig. 7a and b). In addition, the E7-specific CD8⁺ T cells displayed an EM phenotype at late time-points post immunization (Fig. 7c). IP immunization with MCMV-M79-FKBP-E7, but not with a control single-cycle virus expressing the immunodominant HSV gB epitope (MCMV-M79-FKBP-gB), prevented TC-1 tumor progression at 4 weeks, 21 weeks and 13 months post-vaccination indicating long-lasting specific anti-tumor immunity (Fig. 7d). In a similar fashion, MCMV-M79-FKBP-E7 vaccination prevented outgrowth of C3 and TC-1 tumors in most mice (Fig. 7e, Additional file 3: Figure S3A and data not shown). Low level pre-existing immunity elicited via IN infection with MCMV-IE2-OVA did not impact the anti-tumor efficacy of MCMV-M79-FKBP-E7 (Fig. 7e). In mice that succumbed to the tumor challenge, the E7-specific CD8⁺ T cell responses were all < 0.3%, while all (except one) of the protected mice had responses > 0.3% (Fig. 7f).Fig7Fig. 7

Single cycle replication of MCMV based vector vaccine is sufficient to induce anti-tumor effects in prophylactic and therapeutic settings. a. Mice were vaccinated IP with 1 × 10⁶ PFU MCMV-M79-FKBP-E7 or control virus (MCMV-M79-FKBP-gB). Splenocytes were harvested 4, 21 weeks and 13 months after vaccination and stimulated with dominant peptide of E7 or pooled E7 peptides for 36 h (a). IFN-γ production was measured by ELISPOT assay. Shown is the number of spots per 5 × 10⁵ cells ± SEM. b Frequency of E7_49–57-specific CD8⁺ T cells within total CD8⁺ T cell population at 4, 21 weeks and 13 months post vaccination. Data represents mean values ± SEM (n = 5 mice per group). c KLRG1 and CD127 marker expression on E7-specific CD8⁺ T cells at day 7 and 60 post vaccination with MCMV-M79-FKBP-E7. d TC-1 tumor outgrowth of the mice that were vaccinated with 1 × 10⁶ PFU MCMV-M79-FKBP-E7 or control virus MCMV-M79-FKBP-gB at 4, 21 weeks and 13 months before tumor challenge. Tumor out growth was followed for 23 or 40 days. e and f C3 tumor outgrowth of the mice were infected with 1 × 10⁵ PFU MCMV-IE2-OVA via IN route or kept uninfected. After 35 days, the mice were vaccinated with 1 × 10⁶ PFU MCMV-M79-FKBP IP or kept unvaccinated. Mice were challenged with C3 tumor cells after 35 days. Tumor outgrowth was followed for 60 days (e). The number of tumor-free mice from the total mice is indicated above each tumor out growth graph. f The percentage of E7-specific CD8⁺ T cells within the total CD8⁺ T cell population in the blood (y-axis) plotted against the E7-specific CD8⁺ T cell response ranked from low to high (x-axis) of the mice that were either tumor-free (tumor-) or tumor positive (tumor+) (f). g Mice were challenged with TC-1 tumor cells, and after 8 days when tumors were palpable mice were treated with 1 × 10⁶ PFU MCMV-m79-FKBP-E7 or control virus MCMV-m79-FKBP-gB via IP, 1 × 10⁵ PFU MCMV-IE2-E7 via IP or 5 × 10⁵ PFU MCMV-IE2-E7 via SC or kept unvaccinated. Tumor out growth was followed for 40 days. The number of tumor-free mice from the total mice is indicated. Data are representative of two independent experiments with similar results