A forward phenotypically driven unbiased genetic analysis of host genes that moderate herpes simplex virus virulence and stromal keratitis in mice

Abstract

Both viral and host genetics affect the outcome of herpes simplex virus type 1 (HSV-1) infection in humans and experimental models. Little is known about specific host gene variants and molecular networks that influence herpetic disease progression, severity, and episodic reactivation. To identify such host gene variants we have initiated a forward genetic analysis using the expanded family of BXD strains, all derived from crosses between C57BL/6J and DBA/2J strains of mice. One parent is highly resistant and one highly susceptible to HSV-1. Both strains have also been fully sequenced, greatly facilitating the search for genetic modifiers that contribute to differences in HSV-1 infection. We monitored diverse disease phenotypes following infection with HSV-1 strain 17syn+ including percent mortality (herpes simplex encephalitis, HSE), body weight loss, severity of herpetic stromal keratitis (HSK), spleen weight, serum neutralizing antibody titers, and viral titers in tear films in BXD strains. A significant quantitative trait locus (QTL) on chromosome (Chr) 16 was found to associate with both percent mortality and HSK severity. Importantly, this QTL maps close to a human QTL and the gene proposed to be associated with the frequency of recurrent herpetic labialis (cold sores). This suggests that a single host locus may influence these seemingly diverse HSV-1 pathogenic phenotypes by as yet unknown mechanisms. Additional suggestive QTLs for percent mortality were identified--one on Chr X that is epistatically associated with that on Chr 16. As would be anticipated the Chr 16 QTL also modulated weight loss, reaching significance in females. A second significant QTL for maximum weight loss in male and female mice was mapped to Chr 12. To our knowledge this is the first report of a host genetic locus that modulates the severity of both herpetic disease in the nervous system and herpetic stromal keratitis.

Figure 1. Representitive results with several BXD strains.

Shown are representitive data from four different BXD lines. The top Y axis in each panel is the percent survival and the bottom Y axis is the percent weight loss. The X axis is the time in Days post infection (p.i.). The scattergrams on the right show the tear film titers of individual mouse eyes at 4 days p.i. Red circles and dashed lines  =  female; Black squares and solid lines =  male. In the experiment shown, which is a small subset of the data generated, four male and four female mice of each BXD strain were employed.

Figure 2. Quantitative trait loci for percent maximum weight loss.

Mice were infected as described in methods and weighed every other day. Analysis tools in genenetwork (GN) available at www.genenetwork.org were employed to search for possible quantitative trait loci (QTL) in male and female mice. The maximum percent weight loss was employed to search for genome wide association. Shown is a screen capture from GN for female mice. The X axis shows the mouse chromosomes from chromosome 1 to X from the centromere to the end (mouse chromosomes have only one arm). The long ticks are at base pair number 1 of each chromosome and the short ticks are every 25 megabases. The Y axis is the likelihood ratio statistic (LRS) score. 2000 permutation tests were employed to estimate the genome wide adjusted suggestive LRS (p = 0.37, grey horizontal line) and the genome wide adjusted significant LRS (p = 0.05, red horizontal line). The wavy blue line indicates the local LRS at various SNPs etc. across the entire mouse genome. The colored lines show the additive effect of the influence of the locus (red lines indicate association of DBA/2J with trait values, green lines indicate that C57BL/6J alleles increase trait values). Significant QTL for percent maximum weight loss were mapped to Chr12 for both male and female mice and on Chr16 for female mice (shown). The QTL on 16 was suggestive in male mice (not shown). A fully functional interactive analysis and graphic interface available to the public can be generated on GN for trait ID:16194 as detailed in methods. On GN the user can restrict the output to the specific chromosomal region under the peak of the QTL and explore the genes, single nucleotide polymorphisms (SNPs) etc. present in the location (for example see fig. 7).

Figure 3. Quantitative trait loci for percent survival.

Percent survival of male and female mice combined was employed to search for QTL associated with mortality. See the legend for fig. 2 for an explanation of the lines in the figure. A significant QTL was mapped to the distal end of Chr 16 and suggestive QTL were localized to Chr 4, 11, 12 and X (panel A). A pair scan analysis revealed a possible interaction between the locus on 16 and the suggestive locus on X (not shown). The marker regression function was used to identify SNPs with high LRS scores on Chr 16 and X, and the composite scan analysis was then employed to control for the Chr 16 or the Chr X locus. Controlling for the QTL on 16 (green arrow and line) revealed that the suggestive QTL on X was significant (panel B).

Figure 4. Quantitative trait loci associated with herpetic stromal keratitis.

BXD strains of mice were infected via the cornea and the severity of stromal keratitis in male and female mice combined determined by two independent observers using a 5 point scale (0 =  no disease, 5 =  penetrating stromal keratitis). Analysis tools in genenetwork available at www.genenetwork.org were employed to search for possible quantitative trait loci (QTL). See the legend for fig. 2 for an explanation of the lines in the figure. A significant quantitative trait locus (QTL) for severity of herpetic stromal keratitis was identified on Ch 16. A fully functional interactive analysis and graphic interface is available on GN for trait ID:16186.

Figure 5. Representitive data in BXD lines that differ at the HSK QTL on Chr 16.

Groups of male and female mice were infected as described in methods with 1×10⁶ pfu of HSV-1 strain 17syn+. The indicated parameters were assayed as described in the text and eyes were scored at 30 days p.i. by two independent investigators. The top micrographs are of cresyl violet stained corneal sections processed at 30 days p.i. Deranged and thickened corneal epithelium, neovascularization, derangned stromal keratocytes and greatly thickened stromal layer with immune cellular infiltrate is seen in the BXD102 strain. The BXD70 cornea appears normal. The bottom photos are of representitive live mouse eyes 30 days p.i. The BXD70 mouse eye appears normal. Neovascularization and stromal opacity covering more than 75% of the cornea is seen in the BXD102 eye.

Figure 6. Correlations between phenotypes that map to the distal end of Chr 16.

Spearman Rank correlations were performed bewteen percent mortality and herpetic stromal keratitis datasets (left panel, p = 1.52×10⁻⁰⁵) or percent maximum weight loss of both male and female mice combined and percent mortality of male and female mice combined (right panel, p =  1.75×10⁻⁰⁶).

Figure 7. Schematic representation of the region of mouse Chr 16 from 75 to 90 mb.

Suggestive (grey horizontal line) and significant (red horizontal line) LRS for HSV virulence (HSE) in male and female mice combined are indicated. The wavy blue line is the local LRS. The gold seismic graph along the X axis indicates local SNP density. The mouse susceptability QTL region is deliniated by a solid blue rectangle just above the SNP seismic graph. The region equivalent to the human cold sore frequency QTL is designated by a solid red rectangle just above the seismic graph at 75–82 mb. The location of the mouse equivalent (D16Ertd472e ) to the human “cold sore” gene (C21orf91) is indicated by a red triangle. A blue triangle indicates the 3’ end of the Grik1 gene that is an expression QTL for Grik1 in the nervous system. C21orf91 is a short gene shown as a gold box, which is not clearly discernable at this magnification (Top panel). Bottom panel. A three megabase region comprising the peak of the QTL is shown. Grik1, which is 395 kb in length is also shown as a gold rectangle and the other variously colored shapes indicate known or predicted genes and transcripts in the region. In GN, hovering the mouse cursor over the colored shapes (indicated by the open arrow in fig. 7, bottom) gives brief information about each gene. Clicking on the shapes calls up all related entries in the NCBI Gene database. The bands along the top are clickable and either increase the magnification, go to the respective region of the UCSC genome browser, or go to the respective region of the Ensemble genome browser. A fully functional and interactive figure can be generated and further explored on genenetwork.org for trait 16185 by performing the interval mapping function for Chr16.