Characterization of a live-attenuated HCMV-based vaccine platform

Abstract

Vaccines based on cytomegalovirus (CMV) demonstrate protection in animal models of infectious disease and cancer. Vaccine efficacy is associated with the ability of CMV to elicit and indefinitely maintain high frequencies of circulating effector memory T cells (T_EM) providing continuous, life-long anti-pathogen immune activity. To allow for the clinical testing of human CMV (HCMV)-based vaccines we constructed and characterized as a vector backbone the recombinant molecular clone TR3 representing a wildtype genome. We demonstrate that TR3 can be stably propagated in vitro and that, despite species incompatibility, recombinant TR3 vectors elicit high frequencies of T_EM to inserted antigens in rhesus macaques (RM). Live-attenuated versions of TR3 were generated by deleting viral genes required to counteract intrinsic and innate immune responses. In addition, we eliminated subunits of a viral pentameric glycoprotein complex thus limiting cell tropism. We show in a humanized mouse model that such modified vectors were able to establish persistent infection but lost their ability to reactivate from latency. Nevertheless, attenuated TR3 vectors preserved the ability to elicit and maintain T_EM to inserted antigens in RM. We further demonstrate that attenuated TR3 can be grown in approved cell lines upon elimination of an anti-viral host factor using small interfering RNA, thus obviating the need for a complementing cell line. In sum, we have established a versatile platform for the clinical development of live attenuated HCMV-vectored vaccines and immunotherapies.

Figure 1

Construction and in vitro characterization of HCMV TR3. (A) Sequential construction steps resulting in TR3. The original isolate TR was cloned by replacing the US2-6 region with a BAC cassette resulting in TR-BAC. TR-GFP was generated by inserting loxP sites flanking the BAC cassette, replacing US7 with a GFP expression cassette and inserting the US2-7 region of AD169. TR3 was generated from TR-GFP by deleting the GFP cassette, inserting a Cre-expression cassette into the BAC cassette and by replacing the defective UL97 with intact UL97 of AD169. The BAC cassette in TR3 is self-excising resulting in a residual loxP site upon reconstitution of virus. (B) TR3 is sensitive to Ganciclovir. Human MRC-5 fibroblasts were infected with the indicated viruses at MOI = 0.5 in the presence of decreasing concentrations of Ganciclovir. Culture supernatants were harvested when controls showed full cytopathic effect and the amount of viral progeny produced was quantified by standard plaque assay. Results are shown as percent of untreated control. (C) US6 expression by TR3. MRC-5 cells were infected or mock infected with the indicated viruses at MOI = 0.5. At 96 h after infection, cells were harvested and cell lysates were electrophoretically separated and probed for expression of the indicated viral or host proteins by immunoblot. The blots for each protein are shown as cropped from different parts of the same gel. (D) Design of TR3ΔUL78gag and TR3ΔUL82gag. (E) Viral protein expression in the absence of pp71. MRC-5 cells were infected or mock infected with the indicated viruses at MOI = 0.5 in the presence or absence of DAXX-targeting siRNA. Cells were harvested at 96 h post-infection. Cell lysates were electrophoretically separated and probed for the indicated viral and cellular proteins by immunoblot. The blots for each protein are shown as cropped from different parts of the same or different gels separating the same lysate. HIVgag was detected using a polyclonal antiserum to the p24 subunit.

Figure 2

In vitro growth of pp71-deficient HCMV TR3. (A) MOI-dependent growth deficiency of UL82-deleted TR3 in fibroblasts. MRC-5 cells were inoculated with TR3 or TR3ΔUL82gag at a multiplicity of infection (MOI) of 1 or 0.001 and cell culture supernatants were collected at the indicated times. Where indicated cells were transfected with 10 nM of Daxx siRNA and, after 24 h, infected with HCMV. After 2 h, the inoculum was removed, cells were washed 3 times with PBS and fresh media was added. A second round of transfection was performed at day 10 post-infection. Virus was titered on BJ-pp71 cells. The average titer of triplicate experiments (+/− SD) is shown. (B) Growth deficiency of UL82-deleted TR3 in non-fibroblasts. Endothelial cells (HUVEC), epithelial cells (ARPE-19) and astrocytic cells (CCF-STTG1) were infected with the indicated MOI of TR3 or TR3ΔUL82gag. The supernatants were harvested at the indicated days and the virus titer determined on BJ-pp71 cells.

Figure 3

Genetic stability of pp71 deficient HCMV TR3. (A) Stable expression of HIVgag upon multiple passages. MRC-5 cells were mock-infected or infected with TR3ΔUL78gag or TR3ΔUL82gag at an MOI of 0.01 and the supernatant was harvested at full CPE. For sequential passages (p1-p20), TR3ΔUL82gag in the supernatant was titered on BJ-pp71 cells and used for the next round of infection at MOI = 0.01. Lysates from cells harvested at the indicated passages were electrophoretically separated and probed for expression of the shown viral proteins or cellular actin by immunoblot. The blots for each protein are shown as cropped from different parts of the same gel. (B) SNP analysis of TR3ΔUL82gag upon serial passaging. Viral DNA was harvested at the indicated passages from the same samples as in (A) and subjected to Next Generation Sequencing (NGS) on a MiSeq platform. The position and frequency of SNPs are shown along the viral genome. Two SNPs were enriched upon passaging: T-C at position 59972 and C-T at position 82959 resulting in an Asn-Asp change in UL45 and an Asp-Asn change in UL55. (C) SNP analysis of TR3 upon serial passaging. TR3 was serially passaged as in (A) and viral DNA was harvested at the indicated passages and subjected to NGS. The position and frequency of SNPs are shown along the viral genome. One SNP was enriched upon passaging: C-A at position 82489 resulting in a Ser-Ile change in UL55.

Figure 4

pp71-deficient TR3 establishes latency, but does not reactivate in vivo. (A) Latency and reactivation of TR3 and pp71-deleted TR3 in humanized mice. NOD-scidIL2Rγc null (NSG) mice engrafted with human CD34+ stem cells (n = 10 per group) were inoculated IP with MRC-5 fibroblasts infected with TR3 or TR3ΔUL82gag. Eight weeks post‐infection, human hematopoietic stem cells were mobilized by G-CSF treatment and the viral load was measured in liver and spleen one week later. The viral DNA copy number was determined by quantitative PCR and is shown per microgram of total DNA. Values in the absence of granulocyte colony stimulating factor (G-CSF) represent the latent viral load and values after G-CSF stimulation represent the reactivation of virus emerging from latency. (B) Latency and reactivation of PC-deficient and PC-intact AD169 compared to TR-BAC in humanized mice. NSG mice were inoculated with fibroblasts infected with AD169 and AD169rUL131a, or TR-BAC and genome copy numbers were determined as in (A). Statistical significance was determined using two-way analysis of variance, followed by Bonferroni’s posthoc test (P values are shown).

Figure 5

TR3-derived vectors elicit HIVgag-specific T_EM in RM. (A) TR3 elicits T cell responses to HIVgag in RM. Two RM were inoculated with 5 × 10⁶ FFU of TR3ΔUL78gag at day 0. PBMC were collected at the indicated days and HCMV-lysate as well as HIVgag-specific CD4+ and CD8+ T cell responses were measured by intracellular cytokine staining for TNFα and/or IFNγ. The frequency of responding T cells is shown as percent of total memory T cells. (B) pp71-deleted TR3 elicits T cell responses to HIVgag. RM were inoculated with 10² FFU (n = 2), 10⁴ FFU (n = 2) or 10⁶ FFU (n = 4) of TR3ΔUL82gag at day 0 and HCMV lysate and HIVgag-specific T cell responses were determined as in A) at the indicated time points post-infection. (C) Frequency of memory populations within the HIVgag-specific CD4+ and CD8⁺ memory T cells in peripheral blood of the RM inoculated with 10⁴ FFU or 10⁶ FFU of TR3ΔUL82gag. Memory differentiation state was based on CD28 vs. CCR7 expression, delineating central memory (+/+T_CM), transitional effector memory (+/− T_TrEM), and effector memory (−/− T_EM), as designated. The same colors and symbols are used in B) and in C) for the same animals.

Figure 6

T cell responses to pentamer-deficient TR3 vectors in RM. (A) In vitro growth curve of TR3 and TR3ΔUL128-131gag. MRC-5 cells were inoculated with TR3 or TR3ΔUL128-131gag at a MOI = 0.01 and cell culture supernatants were collected at the indicated times. Virus was titered by plaque assay on MRC-5 cells. The average titer of triplicate experiments (+/− SD) is shown. (B) UL128-131-deleted TR3 does not elicit T cell responses in RM. Two RM were inoculated with 5 × 10⁶ FFU TR3ΔUL128-131gag, one RM was inoculated with 5 × 10⁶ FFU TR3ΔUL78gag and the T cell response to HCMV-lysate or HIVgag was measured at the indicated days. (C) In vitro growth curve of PC-deficient TR3. MRC-5 cells (top) or ARPE cells (bottom) were inoculated with the indicated viruses at MOI = 0.01 (MRC-5) or MOI = 5 (ARPE) and cell culture supernatants were collected at the indicated times. Virus was titered as above and the average titer of triplicate experiments (+/− SD) is shown. (D) HIVgag-specific T cell response elicited by PC-deficient TR3. RM (n = 2 per recombinant) were inoculated with 5 × 10⁶ FFU of the indicated PC-deficient vectors on day 0 and the HIVgag-specific or HCMV-lysate CD4+ and CD8+ memory T cell response frequencies were determined in PBMC by intracellular cytokine staining for TNFα and/or IFNγ using overlapping HIVgag peptide pools at the indicated days. (E) Frequency of memory populations within the HIVgag-specific CD8⁺ memory T cells in peripheral blood of the six RM in (D) inoculated with the indicated viruses. Memory differentiation state was based on CD28 vs. CCR7 expression, delineating central memory (+/+T_CM), transitional effector memory (+/− T_TrEM), and effector memory (−/− T_EM), as designated.

Figure 7

In vitro and in vivo characterization of pentamer and pp71-deficient TR3 vectors. (A) HIVgag expression and lack of pp71 and UL130 expression for TR3ΔUL82gagΔUL128-130. Lysates from MRC-5 cells infected with the indicated viruses (MOI = 0.5) for 96 h were electrophoretically separated and probed for the indicated proteins by immunoblot. The blots for each protein are shown as cropped from different parts of the same gel. (B) Viral growth in fibroblasts in the presence or absence of DAXX siRNA. MRC-5 cells were infected with the indicated viruses at MOI = 1 or 0.001 in the presence or absence of siRNA targeting DAXX. Virus titers in the supernatant were determined at the indicated days on BJ-pp71 cells. (C) Latency and reactivation in humanized mice. HuNSG mice were infected with the indicated viruses and treated with G-CSF as shown. Infections and determination of viral genome copy numbers as well as statistical analysis were as described in Fig. 4. (D) Immunogenicity in RM. RM were inoculated with 10⁴ FFU, 5 × 10⁴ FFU, 10⁵ FFU, 5 × 10⁵ FFU, 10⁶ FFU or 5 × 10⁶ FFU (n = 2 each dose) of TR3ΔUL82gagΔUL128-130 at day 0 and HIVgag-specific or HCMV-lysate T cell responses were determined as described in Fig. 5. The frequency of responding T cells is shown as percent of total memory T cells. E) Frequency of memory populations within the HIVgag-specific CD4+ and CD8+ memory T cells in peripheral blood of all RM in (D) demonstrating HIVgag-specific T cell responses (≥1 × 10⁵ FFU inoculum). Memory differentiation state was based on CD28 vs. CCR7 expression, delineating central memory (+/+T_CM), transitional effector memory (+/− T_TrEM), and effector memory (−/− T_EM), as designated. The same colors and symbols are used in (D) and in (E) for the same animals.