Targeting the host transcription factor HSF1 prevents human cytomegalovirus replication in vitro and in vivo

Abstract

FDA-approved antivirals against HCMV have several limitations, including only targeting the later stages of the viral replication cycle, adverse side effects, and the emergence of drug-resistant strains. Antivirals targeting host factors specifically activated within infected cells and necessary for viral replication could address the current drawbacks of anti-HCMV standard-of-care drugs. In this study, we found HCMV infection stimulated the activation of the stress response transcription factor heat shock transcription factor 1 (HSF1). HCMV entry into fibroblasts rapidly increased HSF1 activity and subsequent relocalization from the cytoplasm to the nucleus, which was maintained throughout viral replication and in contrast to the transient burst of activity induced by canonical heat shock. Prophylactic pharmacological inhibition or genetic depletion of HSF1 prior to HCMV infection attenuated the expression of all classes of viral genes, including immediate early (IE) genes, and virus production, suggesting HSF1 promotes the earliest stages of the viral replication cycle. Therapeutic treatment with SISU-102, an HSF1 inhibitor tool compound, after IE expression also reduced the levels of L proteins and progeny production, suggesting HSF1 regulates multiple steps along the HCMV replication cycle. Leveraging a newly developed human skin xenograft transplant murine model, we found prophylactic treatment with SISU-102 significantly attenuated viral replication in transplanted human skin xenografts as well as viral dissemination to distal sites. These data demonstrate HCMV infection rapidly activates and relocalizes HSF1 to the nucleus to promote viral replication, which can be exploited as a host-directed antiviral strategy.

Keywords: Cytomegalovirus; Heat shock factor 1; Host-directed antiviral; Stress response; Viral replication.

Fig. 1.. HCMV induces a persistent activation of HSF1 during lytic infection.

(A to D) HEL 299 fibroblasts were mock infected or infected (MOI 5) with HCMV strain TB40E (A, B), UV-inactivated HCMV strain TB40E (UV-HCMV) (A, B), or HCMV strain Towne (C, D) for indicated time points post infection. As a positive control, fibroblasts were subjected to heat shock (HS) for 15 min at 42°C. Total HSF1 and phosphorylation of HSF1 at Ser³²⁶ were detected by Western blot (A, C) and quantified (B, D). β-actin was used as a loading control. Western blots and densitometry are representative of at least 3 biological replicates per group. ns, non-significant, *P<0.05, **P < 0.005, ****P < 0.0001, by one-way ANOVA with Tukey’s HSD post hoc test.

Fig. 2.. Inhibition of HSF1 reduces HCMV IE protein abundance.

(A to E) HEL 299 fibroblasts were prophylactically treated with SISU-102 (A, B, E) or ganciclovir (C, D) at the indicated concentrations for 24 h. (F, G) Fibroblasts were transfected with a scrambled or HSF1 siRNA for 48 h. Following drug treatment or siRNA knockdown, cells were mock infected or infected (MOI 1) with IE2-eGFP for an additional 24 h. GFP intensity was measured using a fluorescent plate reader (A, C). Cytotoxicity was determined by SRB colorimetric assay (B, D, G). Total HSF1 and IE1 were detected by Western blot at 24 hpi (E, F). IE1 was quantified by densitometry (F). β-actin was used as a loading control. All data are representative of at least 3 biological replicates per group.

Fig. 3.. HSF1 inhibition attenuates the levels of HCMV L proteins during lytic replication.

(A to C, E, F) HEL 299 fibroblasts were prophylactically treated with SISU-102 (A, B, C) or ganciclovir (E, F) at the indicated concentrations for 24 h. (D) Fibroblasts were transfected with a scrambled or HSF1 siRNA for 48 h. Following drug treatment or siRNA knockdown, cells were mock infected or infected with UL99-GFP for 96 h at an MOI of 1 (A to F). At 96 hpi, GFP intensity was measured using a fluorescent plate reader (A, E). Cytotoxicity was determined by SRB colorimetric assay (C, F). Total HSF1 and gL were detected by Western blot (B, D). β-actin was used as a loading control. (G, H) Confluent Nuff-1 cells were infected with UL99-eGFP or TB40Egan^RA594V, both of which express the fluorescent marker mCherry with IE kinetics. Cells were then treated with DMSO control vehicle, SISU-102 (8 μM) (G), or ganciclovir (1 μM) (H). Viral spread was quantified by mCherry area of infection using the BioTek Cytation 5 cell imaging multimode reader. All data are representative of at least 3 biological replicates per group.

Fig. 4.. HSF1 activity is required for efficient viral replication and progeny production.

(A to C) HEL 299 fibroblasts were prophylactically treated with SISU-102 or ganciclovir at the indicated concentrations for 24 h. (D) Fibroblasts were transfected with a scrambled or HSF1 siRNA for 48 h. Following drug treatment or siRNA knockdown, cells were mock infected or infected (MOI 1) with WT HCMV strain TB40E (A, B, D) or HCMV strain Towne (C) for 96 h. At 96 hpi, viral genome copy number was assessed by qPCR analysis for UL123 (as a marker of the viral genome) and GAPDH (A). TCID₅₀ assays were performed on the supernatants to measure progeny virus production (B to D). All data are representative of 3 to 4 biological replicates per group. ****P < 0.0001, by one-way ANOVA with Tukey’s HSD post hoc test (A). *P<0.05, by Mann Whitney t-test (One-tailed) (B to D).

Fig. 5.. Therapeutic inhibition of HSF1 reduces HCMV L protein abundance and progeny production.

(A to F) HEL 299 fibroblasts were infected (MOI 1) with UL99-GFP (A to D) or WT HCMV (E, F) for 24 h. Infected cells were then prophylactically treated with SISU-102 or ganciclovir at the indicated concentrations for an addition 72 h. At 96 hpi, GFP intensity was measured using a fluorescent plate reader (A, C). Cytotoxicity was determined by SRB colorimetric assay (B, D). Viral genome copy number was assessed by qPCR analysis for UL123 and GAPDH (E). TCID₅₀ assays were performed on the supernatants to measure progeny virus production (F). All data are representative of 3 to 4 biological replicates per group. ****P < 0.0001, by one-way ANOVA with Tukey’s HSD post hoc test (E). *P<0.05, by Mann Whitney t-test (One-tailed) (F).

Fig. 6:. Prophylactic inhibition of HSF1 attenuates HCMV infection in vivo.

(A) Timeline for skin implantation, drug treatment phase, and infection period of athymic mice. (B to G) Mice were treated daily beginning 2 days prior to infection with vehicle controls, SISU-102 (5 mg/kg/day), or valganciclovir (50 mg/kg/day). At day 2 following first drug administration, the bottom xenograft in each mouse was mock infected or infected with 10⁶ plaque forming units (pfu) of ΔUL18-fLuc. Bioluminescence imaging and quantification (total flux) of the inoculated (B) and uninoculated (G) skin xenografts were measured at the indicated time points using an IVIS Spectrum bioluminescence imaging system. Body weights of the mice were measured at the indicated time points throughout the drug treatment and infection phase (C). Bioluminescence and body weight data are representative of 7 to 22 mice per treatment group. Skin xenografts at day 16 post infection were collected, fixed, and processed for immunohistochemistry (IHC) to detect GFP (red) and nuclei (purple) (D). IHC data are representative of 5 biological replicates per group. At 16 days post infection, skin xenografts were harvested and homogenized into a single cell suspension (E, F). Viral genome copy number from each skin xenograft was determined by qPCR analysis for UL123 and GAPDH (E). qPCR data are representative of 3 to 4 biological replicates per group. GFP, total HSF1, UL44, and gL were detected by Western blot (F). β-actin was used as a loading control. Western blots are representative of 5 biological replicates per group. *P<0.05, ****P < 0.0001, by one-way ANOVA with Tukey’s HSD post hoc test.