Viral gene drive in herpesviruses

Herpesviruses are common human pathogens with significant clinical impact, especially in immunocompromised individuals and the elderly. Human cytomegalovirus (HCMV) contributes to morbidity via chronic infection and reactivation, and current treatments face issues of resistance and side effects, motivating novel antiviral strategies. Gene drives—typically engineered using CRISPR-Cas9—bias inheritance to spread specific genetic sequences through populations, but have been considered limited to sexually reproducing organisms. The authors hypothesize that, because herpesviruses replicate their genomes in the nucleus and undergo frequent homologous recombination during coinfection, a CRISPR-based gene drive could be engineered to spread an engineered cassette among viral genomes without sexual reproduction. They propose and test a viral gene drive in HCMV capable of replacing wildtype (WT) viruses and, when targeting a gene important for replication under immune pressure, suppressing infection.

The paper situates its work within prior CRISPR-based gene drive developments in sexually reproducing organisms (e.g., mosquitoes) and yeast, and discusses concerns about resistance allele formation in such systems. It highlights that herpesviruses are amenable to CRISPR editing and naturally undergo high-frequency homologous recombination during replication. Evidence from the literature indicates coinfection and recombination occur in vivo among herpesviruses (including HSV-1, VZV, and CMV). A prior report of a bacterial gene drive is noted but did not aim to spread a trait through a population. This background supports the feasibility of a directed, CRISPR-enabled recombination strategy to drive engineered sequences through viral populations.

Design and construction: A CRISPR-Cas9 gene drive cassette was inserted into the HCMV UL23 locus, chosen as dispensable for replication in fibroblasts. The donor plasmid contained UL23 homology arms, Streptococcus pyogenes Cas9, an mCherry reporter, and a gRNA targeting the UL23 5′ UTR, with Cas9 driven by the UL23 promoter. Human foreskin fibroblasts were transfected with the donor and infected with HCMV TB40/E-bac4 to generate mCherry-expressing gene drive viruses (GD-mCherry) via homologous recombination and plaque purification. A control TB40/E strain with an mCherry reporter in a neutral locus was also generated. A Towne strain expressing eGFP (Towne-eGFP) served as a WT comparator. Growth and spread assays: Replication kinetics of GD-mCherry versus TB40/E and Towne-eGFP were measured by plaque assays over time. To assess recombination and drive transfer, fibroblasts were coinfected with GD-mCherry and WT strains; supernatants were then used to infect fresh cells at very low MOI (<0.001) to isolate progeny viruses. Fluorescent plaque phenotypes (eGFP-only, mCherry-only, dual eGFP+mCherry) were quantified. Molecular analyses: Episomal viral DNA was extracted from infected cells using a modified HIRT method. PCR assays targeted reporter genes and homology arms; Sanger sequencing verified integration and recombination junctions. Long-read Oxford Nanopore sequencing of viral DNA enabled reconstruction of recombination events and genome segment origins using SNP maps distinguishing Towne and TB40/E strains; reads were mapped (e.g., GraphMap) and visualized (IGV) to infer recombination histories. Selective pressure/defective drive experiments: Because UL23 blocks IFN-γ responses, GD-mCherry viruses lacking a UL23 start codon were tested for sensitivity to human IFN-γ. Cells were treated with increasing IFN-γ concentrations (e.g., 10, 100 ng/mL) and viral titers assessed by plaque assays. Coinfection experiments under IFN-γ treatment evaluated whether a defective drive could spread when complemented in trans by WT virus, and whether subsequent generations would be suppressed by IFN-γ. Statistics and simulations: Viral titers (PFU/mL) were compared using two-way ANOVA with Sidak’s multiple comparisons on log-transformed data and non-parametric tests where applicable. Numerical simulations modeled gene drive spread under varying fitness costs and coinfection rates to contextualize in vivo feasibility. Cell culture conditions followed standard DMEM + 10% FBS with antibiotics; infections were typically performed at MOI ~0.1 unless stated.

- Successful construction of a CRISPR-Cas9 gene drive cassette at UL23 (GD-mCherry) that replicated with slightly slower dynamics than TB40/E but reached similar titers. - Coinfection with WT HCMV (Towne-eGFP) yielded progeny plaques expressing eGFP only, mCherry only, or both reporters, indicating recombination and integration of the drive cassette into WT genomes. - Long-read sequencing of 38 individual viral genomes after coinfection showed: 6% lacked the drive cassette; 18% were pure donor (TB40/E) genomes; 74% were recombinants between Towne-eGFP and GD-mCherry, and all recombinants contained the gene drive cassette. U/S regions frequently originated from the TB40/E strain, revealing asymmetric recombination patterns. - In population spread assays, the gene drive efficiently invaded WT viral populations across initial frequencies of GD-mCherry at 50%, 10%, and as low as 0.1%, with incorporation into the majority of eGFP-expressing viruses; dual-fluorescent viruses propagated as new gene drive viruses. Final proportions of drive-containing viruses reached up to approximately 95% in culture. - Targeting UL23 created a defective drive under immune pressure: GD-mCherry replication was strongly inhibited by IFN-γ, showing ~250-fold and ~8000-fold titer reductions at 10 and 100 ng/mL IFN-γ, respectively, while WT viruses were minimally affected. - Despite this defect, coinfection allowed complementation and initial spread of the defective drive; however, upon passaging supernatants to fresh cells under IFN-γ, titers dropped ~180-fold (10 ng/mL) and ~420-fold (100 ng/mL), demonstrating that subsequent generations were severely suppressed. - Numerical modeling indicated that with realistic in vivo coinfection rates (~2–10%), defective drive viruses could still spread in a viral population, consistent with observed in vivo coinfection/recombination rates in herpesviruses.

The study demonstrates that a CRISPR-based gene drive can be engineered to spread in a non-sexually reproducing organism—herpesviruses—by leveraging coinfection and homologous recombination. The findings directly address the hypothesis that directed recombination during viral replication can duplicate a gene drive cassette into WT genomes, enabling replacement of WT populations in culture even from very low initial frequencies. Importantly, by inserting the drive into UL23, the authors show that the approach can be tuned to create a therapeutically valuable phenotype: drive viruses that become highly sensitive to IFN-γ, allowing suppression of infection in subsequent generations under immune pressure. The work underscores that recombination in herpesviruses is frequent but can be made precise and directional through Cas9-mediated cleavage and homology-directed repair, distinguishing this approach from random recombination. Considerations for in vivo translation include evidence that coinfection and recombination occur in natural infections, and simulations suggesting that with 2–10% coinfection, even defective drives may spread. Potential emergence of drive-resistant alleles is acknowledged; targeting essential, conserved viral genes may both drive spread and minimize viable resistance. Biosafety is emphasized, with experiments conducted in laboratory strains and alignment to guidelines; future deployment would require careful risk–benefit assessment.

This work provides a proof of concept that gene drives can be designed for DNA viruses, specifically HCMV, to transfer engineered sequences between strains and efficiently replace WT populations in vitro. By targeting UL23, the authors demonstrate a strategy to couple drive spread with functional attenuation under IFN-γ, enabling suppression of infection in subsequent passages. The approach opens avenues to develop antiviral gene drives against other herpesviruses or DNA viruses. Future research should: evaluate efficacy and safety in animal models; optimize targets to essential, conserved genes to reduce resistance; investigate dynamics under varying coinfection rates and immune contexts; and address ecological and biosafety considerations for translational applications.

- Demonstrations are in cell culture using laboratory HCMV strains; in vivo dynamics, immune interactions, and tissue-specific coinfection rates may differ. - The drive relies on coinfection; spread would be limited in predominantly latent infections with low replication and coinfection. - Targeting UL23 yields a strong phenotype only under exogenous IFN-γ; more universally essential targets may be required for robust therapeutic effect without external modulation. - Evidence of potential drive-resistant viruses (non-converted sequences) suggests resistance could emerge and be selected; long-term evolutionary dynamics were not fully characterized. - Recombination patterns were complex and asymmetric between strains, which may influence predictability of genome outcomes. - Pre-existing immunity and superinfection exclusion could limit superinfection in some hosts, although literature indicates superinfection occurs in immunocompromised settings. - Statistical power was limited in some assays, and multiple analyses relied on log-transformed data due to small sample sizes.