Introduction
Infection with herpes simplex virus type 1 and 2 (HSV-1/2) causes orofacial, genital, cerebral, and ocular disorders. In particular, HSV-1 has been implicated in diverse acute and chronic neurological diseases including meningitis, encephalitis, and epilepsy (Gilden et al.
2007; Kleinschmidt-DeMasters and Gilden
2001).
Herpesviruses are thought to enter the brain via the olfactory pathway, as argued for HSV-1 (Mori et al.
2005) and other neurotropic viruses (Harberts et al.
2011), and then migrate to linked susceptible regions. HSV meningitis typically affects the temporal cortex, and infection of the limbic system including the hippocampus (Hpc) is implicated in HSV-1 encephalitis (Damasio and Van Hoesen
1985). Epilepsy associated with HSV-1 may also involve the Hpc because origins of focal epilepsy are typically found in the temporal lobe (reviewed in Blair
2012). An immunological study reported the presence of HSV antigen within temporal lobe and Hpc of human herpes encephalitis, with further evidence for antigen in amygdala and olfactory cortex (Esiri
1982). In mouse, inoculation of HSV into Hpc gave significantly more severe disease than inoculation into cerebellum (McFarland and Hotchin
1987). However, the tropism of HSV for the temporal brain remains unexplained.
A major factor determining species and tissue specificity of viruses is the cellular receptor(s) mediating viral entry. In contrast to most other viruses, HSV-1 entry into epithelial or neuronal cells is a complex process requiring multiple viral glycoproteins and cellular receptor molecules, including the four viral glycoproteins, glycoprotein D (gD), glycoprotein B (gB), glycoprotein H (gH), and glycoprotein L (gL), as well as the cellular receptors for at least gD and gB. Entry involves a series of concerted events based on the interaction between ligands and receptors that eventually lead to fusion of the viral and cellular membranes (Campadelli-Fiume et al.
2012a). The initial step of HSV-1 entry is low-affinity attachment of viral particles to cell-surface heparan sulfate proteoglycans (HSPGs) via gH and gB. HSPGs are very widely expressed by a broad range of different cells including neurons. This initial attachment is followed by high-affinity binding of the viral gD glycoprotein to one of the gD receptors, either poliovirus receptor-like protein 1 (PVRL1, also known as nectin 1) or tumor necrosis factor (TNF) receptor superfamily member 14 (TNFRSF14), also known as herpesvirus entry mediator (HVEM). In mouse, PVRL1 knockout was reported to attenuate but not abolish HSV infection (Taylor et al.
2007), and PVRL1 was essential for lethal brain infection when inoculated directly (Kopp et al.
2009), but not peripherally. By contrast, TNFRSF14 (HVEM) was not essential (Kopp et al.
2009). The current view is that receptor-bound gD leads to activation of gH/gL, which in turn transforms gB into a fusion-competent state (Campadelli-Fiume et al.
2012a). Virus-induced membrane fusion, however, is only possible if gB is bound by an additional specific gB receptor. Three gB receptors are currently known, including (i) the paired immunoglobulin-like type 2 receptor α (PILRA), (ii) another paired-type receptor with homology to PILRA, myelin-associated glycoprotein (MAG), and (iii) myosin heavy chain 9 (MYH9, non-muscle, also known as NMMHC-IIA). There is also some evidence that HSV-1 entry may require the binding of the gH/gL heterodimer to a cellular receptor (Campadelli-Fiume et al.
2012b; Karasneh and Shukla
2011).
In vitro, HSV-1 infects a diverse spectrum of different human and non-human cells. In vivo, however, infection is much more restricted to epithelial and neuronal cells. Despite extensive characterization of HSV-1 receptors, little is known about their tissue expression patterns in the human brain. In mouse, PVRL1/nectin 1 expression has been reported in limbic regions, frontal association cortex, and olfactory system (e.g., Horvath et al.
2006), but few studies are available in human. Guzman et al. (
2006) and Geraghty et al. (
1998) reported expression of PVRL1/nectin 1 in a variety of normal and transformed human cells, but the brain distribution was not analyzed. In the most detailed study of human brain, Prandovszky et al. (
2008) used immunohistochemistry to inspect fetal human tissues and reported PVRL1 expression in endometrium, cornea, and cortex, with strong staining in pyramidal cells of the Hpc. However, no studies have addressed the expression of HSV-1 receptors other than PVRL1 in human brain. In mice, there are marked changes in receptor expression between the fetal and postnatal periods (Horvath et al.
2006; Prandovszky et al.
2008), but the patterns of HSV-1 receptor expression in adult human brain have not been examined. We therefore used different online transcriptome databases to map the expression patterns of all known HSV-1 gD and gB receptors in adult and developing human brain.
Discussion
Type 1 herpes simplex virus (HSV-1) is a neurotropic virus that affects select brain regions including particular brainstem nuclei and, most notably, the temporal brain including the Hpc, both in mouse and in human. HSV-1 is unusual among viruses in that it requires the expression of two different receptors on the same target cell, which respectively interact with viral glycoproteins gB and gD (“
Introduction” section). We present for the first time a systematic analysis of the distribution of both gB and gD HSV-1 receptors in human brain. Our results confirm and extend previous results, in mouse, that the gene encoding the HSV-1 receptor PVRL1 is selectively expressed in Hpc (Horvath et al.
2006), and we now report that, in adult human, at least three HSV-1 receptors are differentially expressed in Hpc,
MYH9 (gB receptor),
PVRL1 (gD), and
TNFRSF14 (gD), whereas
PILRA (gB) is most abundantly expressed in human cerebellum. For
MYH9,
PVRL1, and
TNFRSF144, the level of upregulation in Hpc was comparable to that observed for the mineralocorticoid receptor, the archetypical Hpc-specific transcript in mouse (discussed below).
MAG (gB) expression may also be enriched in Hpc, but there was major discrepancy between the two databases we consulted (ABA and HBT), with high levels of
MAG expression being seen in postnatal Hpc in the HBT database, whereas this region was not apparently enriched over mean brain levels in ABA. We have no explanation for this result, but possible explanations include the use of different hybridization probes that might detect alternative transcript splicing and/or differential background hybridization, noting that ABA is in general based on two or more probes per gene, whereas HBT relies on single probes.
Although the different human brains sampled in ABA gave slightly different expression patterns for the same hybridization target, in this study the overall profile was largely conserved (Fig.
1); technical issues with tissue preparation combined with genetic and environmental effects may underlie this variability.
In addition, patterns were substantially conserved between male and female brain, with selective HSV receptor expression in Hpc in both male and female. There were potentially some minor differences (“
Results” section); however, the ABA microarray database is based on a single female individual, and studies on a greater diversity of individuals will be necessary to address possible male/female differences in HSV receptor expression.
Both databases confirm that
PILRA (gB receptor) is prominently expressed in cerebellum; however, there was no parallel enrichment of gD receptors: lack of gD receptors could explain the observation, in mouse, that inoculation of HSV into cerebellum did not give rise to severe disease (McFarland and Hotchin
1987).
Significant gB/gD receptor expression was also observed (ABA) in adult globus pallidus (Gp), trigeminal nuclei, brainstem, and red nucleus. This is of interest because HSV-1 brain infection can sometimes selectively involve the Gp (Hu et al.
2003). The trigeminal nuclei are also of note because they are connected to the trigeminal ganglia where HSV-1 latently persists (Baringer and Swoveland
1973; Croen et al.
1987; Furuta et al.
1992; Theil et al.
2001), and selective involvement of the brainstem in some cases of HSV-1 encephalitis has been reported (Livorsi et al.
2010). Finally, the presence of receptors in the red nucleus is notable because this region is functionally connected with the Hpc (Dypvik and Bland
2004; Nioche et al.
2009). The relevance of other brain regions highlighted in Fig.
1 to the pathobiology of HSV-1 infection is not known and warrants further investigation.
The finding that HSV-1 receptors are expressed at only low levels in gestation, but rise markedly during the postnatal period, is notable because the fetal human brain is not generally subject to infection with HSV-1, in contrast to other viruses (e.g., cytomegalovirus, rubella, Zika virus) that can establish damaging brain infection during gestation (Driggers et al.
2016; Guillemette-Artur et al.
2016; McCarthy et al.
2011; Naeye and Blanc
1965). Although exposed to HSV-1 in utero of HSV-1-positive mothers, overt pathology appears to be generally limited to postnatal ages. Our results suggest that the fetal human brain may be partially refractory to HSV-1 infection owing to low levels of receptor expression.
This study has some limitations. First, the data presented address transcript levels rather than protein expression. Brain proteomic analysis demonstrates expression of PVRL1 in adult human Hpc (Hondius et al.
2016), but other brain regions were not studied; comprehensive proteomics of human brain (an emerging project at
www.hupo.org/human-brain-proteome-project) will be necessary to confirm the regional distribution of HSV-1 receptors. Second, cell-surface receptors are not the only determinants of virus tropism, and multiple levels of control on virus entry, genome transcription, and translation govern the outcome of infection. Additional restriction factors include routing factors such as αVβ3 integrin (Campadelli-Fiume et al.
2012a), which affect virus internalization, and also proteins such as B5 that influence HSV-1 translation (Cheshenko et al.
2010; Perez et al.
2005). However, without expression of receptors for both gD and gB, effective virus entry cannot take place. Third, much of our analysis addresses the normalized
z score or fold-change values instead of absolute expression levels. However, this is not necessarily misleading. To illustrate, mineralocorticosteroids are known to selectively target the Hpc, and, for the mineralocorticoid receptor
NR3C2, the top-ranked gene for selective expression in mouse Hpc (Lein et al.
2007), both the
z and fold change scores were comparable to those generated by both gD and gB receptors (Fig.
1 and Table
2).
Further, because HSV-1 entry requires the simultaneous expression of at least two different receptors, conjoint overrepresentation of both receptor types in a single tissue even by a small amount may be anticipated to synergistically and disproportionately influence the statistical likelihood of encountering both types of receptor on the same cell, further amplified during serial transfer between cells, and thus favor selective tissue tropism. For example, if there is only a 2-fold excess of receptors for gB (MYH9) and gD (PVRL1 plus TNFRSF14), the differential efficiency of primary infection could rise 8×, and to 64× for a further round of infection.
In conclusion, the selective virus receptor gene expression profiles reported here may contribute to the tropism of HSV-1 for particular postnatal brain regions in human, centrally including the temporal brain and Hpc.