(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Herpes simplex virus spreads rapidly in human foreskin, partly driven by chemokine-induced redistribution of Nectin-1 on keratinocytes [1]

['Hafsa Rana', 'The Westmead Institute For Medical Research', 'Westmead', 'New South Wales', 'Faculty Of Medicine', 'Health', 'The University Of Sydney', 'Sydney', 'Naomi R. Truong', 'Blake Johnson']

Date: 2024-06 

HSV infects keratinocytes in the epidermis of skin via nectin-1. We established a human foreskin explant infection model to investigate HSV entry and spread. HSV1 entry could only be achieved by the topical application of virus via high density microarray projections (HD-MAPs) to the epidermis, which penetrated beyond one third of its thickness, simulating in vivo microtrauma. Rapid lateral spread of HSV1 to a mean of 13 keratinocytes wide occurred after 24 hours and free virus particles were observed between keratinocytes, consistent with an intercellular route of spread. Nectin-1 staining was markedly decreased in foci of infection in the epidermis and in the human keratinocyte HaCaT cell line. Nectin-1 was redistributed, at the protein level, in adjacent uninfected cells surrounding infection, inducible by CCL3, IL-8 (or CXCL8), and possibly CXCL10 and IL-6, thus facilitating spread. These findings provide the first insights into HSV1 entry and spread in human inner foreskin in situ.

Herpes Simplex Virus (HSV) infects 3.7 billion people globally, leading, in some, to lifelong recurrent disease and co-infection with viruses such as HIV. There is no cure or vaccine. Animal models and in vitro cell lines do not accurately represent the initial events that take place during HSV infection in human genital mucosa. We have successfully established a model of acute HSV-1 infection within explants of human inner foreskin, a common site for sexual transmission. Topical microtrauma penetrating a third of the way into the epidermis was essential for HSV infection, indicating that the two most superficial strata were refractory. There was rapid spread of HSV1 particles through and around epidermal keratinocytes over 24 hours. HSV1 particles also interacted with and were taken up by epidermal Langerhans Cells (LCs) and Dendritic Cells (Epi DCs). Focal HSV1 infection of keratinocytes induced a redistribution of nectin-1, the HSV entry receptor, in a collar surrounding the foci via specific chemokines which further facilitated viral spread. These results provide an insight into the initial events that lead to HSV1 infection and spread within human genital mucosa and defines a time window of opportunity to target the virus before it enters epidermal nerves and becomes latent.

Funding: A.L.C. received project grant 1163748 and investigator grant 1177942 from the National Health and Medical Research Council of Australia, URL: https://www.nhmrc.gov.au . H.R. received a Research Training Stipend from The University of Sydney. The Westmead Scientific Platforms was supported by the Westmead Research Hub, the Westmead Institute for Medical Research, the Cancer Institute New South Wales, the National Health and Medical Research Council and the Ian Potter Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here we introduced three substantial innovations to more closely simulate the events of genital herpes and to more accurately track viral spread. Firstly, we developed a HSV infection model using human inner foreskin explants, a clinically relevant site of infection. Inner foreskin is a type II mucosal tissue with a thin stratum corneum and contains an abundance of cytokine secreting immune cells making it more vulnerable to pathogens such as HIV [ 28 – 30 ]. Secondly, we studied the role of microtrauma in infection by introducing HSV into the tissue using High-Density Microarray Patches (HD-MAPs). Thirdly, in addition to using GFP-labelled virus, we utilized a more sensitive detection system–RNAscope in situ hybridization (ISH), to detect HSV1 DNA. LCs and Epi DCs interacted with HSV within foci of infection but were reduced in density. HSV only infected keratinocytes when microneedles punctured at least a third of the depth of the epidermis, below the stratum granulosum, and the virus spread rapidly within 24 hours suggesting an intercellular route in addition to cell to cell spread. HSV infection caused redistribution of nectin-1 in situ, with reduced staining in infected cells. In HSV1 + HaCaT cell cultures, there was a redistribution of nectin-1 in the uninfected cells surrounding HSV foci, via keratinocyte-secreted chemokines, especially CCL3, CXCL10, IL-6 and IL-8, thus facilitating viral spread.

Keratinocytes have three major cell junction protein complexes that play different roles in adhesion and epidermal barrier formation; tight junctions, adherens junctions, and desmosomes. In skin, all cell-cell junction proteins are located in close proximity to each other within the keratinocyte membrane to regulate and maintain this barrier [ 17 , 18 ], with tight junctions being most superficial in nongenital skin. Nectin-1, the major entry receptor for HSV, also functions as an adherens junction protein to maintain the epidermal barrier and regulate the exchange of fluids and other molecules, as well as in the formation of desmosomes [ 19 , 20 ]. Desmosomes are complex multi-protein structures, including the adaptor protein, Plakoglobin, which is also found in adherens junctions [ 20 ]. Inflammation results from the depletion of Ca 2+ and disruption of Ca 2+ -dependent tight junctions in skin [ 21 , 22 ]. This causes a breakdown in epidermal integrity and redistribution of nectin-1 from the adherens junction to the entire cell surface which leads to increase binding of HSV1 gD [ 23 ]. In human and murine skin explant studies, mechanical trauma also led to viral infection around the wounded area, due to tight junction disruption, but did not spread to cells with intact tight junctions [ 4 , 5 , 24 , 25 ]. However the epithelial distribution of these three intercellular junctions in cervico-vaginal mucosa and in inner foreskin differs markedly from abdominal or murine trunk skin, likely to reduce the barrier function of tight junctions in the upper epidermis and permit luminal exchange [ 26 , 27 ].

Uptake of HSV1/2 by epidermal LCs and DCs is essential for initiation of the immune response to HSV infection. We recently showed that the two MNPs, LCs and DCs are infected by HSV1 via different routes [ 14 ]. LCs internalised HSV1 via langerin-mediated, pH-dependent endocytosis and became productively infected which triggered apoptosis. LCs migrated into the dermis while apoptosing and clustered with dermal type 1 conventional DCs (cDC1s) [ 3 ], which have a propensity to cross-present viral antigens to CD8 + T cells. Epidermal DCs (Epi DCs) internalised HSV1 via pH-independent pathways, yet still apoptosed due to productive viral infection similar to LCs [ 14 ]. Epidermal MNPs can directly sample HSV as it penetrates the skin, or via infected keratinocytes.

Studying the mechanisms and kinetics of HSV entry and spread within the epidermis of inner foreskin will elucidate 1) The time window prior to viral entry into the sensory nerve endings within the epidermis and subsequent transport to the dorsal root ganglia (DRG) where it establishes lifelong latency and 2) The kinetics of initiation of the immune response via MNPs. Rapid infection and spread in keratinocytes, to the ‘sanctuary’ of terminal axons of sensory neurons provides HSV with an efficient strategy to evade the immune response [ 12 , 13 ]. Thus, understanding these events could guide strategies for vaccines or immunotherapy to prevent viral spread to the nerves and establishment of a lifelong infection.

Herpes simplex viruses (HSV) types 1 and 2 cause genital herpes. HSV2 is the most common cause of recurrent herpes, but HSV1 is an increasingly common cause of initial genital herpes and predominant in some populations [ 1 ]. As shown in biopsies of primary and recurrent lesions [ 2 , 3 ], both viruses target the keratinocytes in the stratified squamous epithelium that makes up the epidermis of skin, in addition to two types of mononuclear phagocytes (MNPs): Langerhans cells (LCs) and dendritic cells (DCs). Viral spread is restricted to the epidermis. This epidermis consists of the deep stratum basale and successively superficial layers of the stratum spinosum, stratum granulosum and stratum corneum, with the upper two layers providing barrier protection from pathogens. Thus, a breach or inflammation of the epithelium in skin appears to be required to facilitate viral penetration [ 4 , 5 ]. Sexual intercourse, with either anal or vaginal penetration has been shown to induce microabrasions and trauma to anogenital mucosa [ 6 , 7 ]. In separate studies, 55% of women who engage in consensual sex and up to 73% of both women and men have been shown to undergo some type of anogenital injury [ 8 , 9 ]. Studies in Kenya also showed 64–66% of men self-reported penile injuries during consensual sex, with a higher risk of injury in uncircumcised men [ 10 , 11 ].

Treatment of HaCaT cells for 6 hours with CCL3, CCL5, (both targeting CCR1 expressed by keratinocytes), CXCL1, IL-8 (targeting CXCR2), CXCL9 and CXCL10 (targeting CXCR3) and IL-6 (targeting IL-6R) followed by IF labelling of nectin-1 showed a significant increase in the extent of nectin-1 redistribution in the HaCaT cell monolayer at cytokine concentrations of 10 and/or 50–100 pg/mL ( Fig 5B and 5C ). CCL2 and CXCL5 did not significantly increase nectin-1 staining, thus this data was excluded. These studies took into account the cycling of nectin-1 expression in HaCaT cells (minimal at 18 hours post seeding, S7 Fig ). RT-qPCR was performed to measure changes in nectin-1 expression in HaCaT cells at the transcriptional level in response to IL-8 and CCL3, however no significant change was seen in comparison to mock-treated cells ( Fig 5D ). This strongly suggests that nectin-1 redistribution, not up or downregulation, is occurring at the protein level. Thus, it appears that HaCaT cells respond to the early stages of HSV infection by producing these pro-inflammatory cytokines and chemokines, but this is diminished by 24 hours, probably an effect of HSV-induced inhibition of transcription of cellular RNAs and subsequent cell death. This increase in pro-inflammatory cytokine and chemokine release coincided with the increase in nectin-1 at 12 h.p.i.

(A) HaCaT cells cultured for 12 or 24 hours with either serum-free medium (mock) or HSV1-GFP (MOI 1). Supernatants were analysed for an array of cytokines and chemokines via the LEGENDplex Assay (BioLegend). Concentrations (pg/mL) of detected cytokines and chemokines are displayed as mean ± S.D. (n = 4). * = p<0.05 determined by an ordinary one-way ANOVA with Dunnett’s multiple comparisons test. (B) HaCaT cells were treated with serum-free media (mock) or CCL3, CCL5, CXCL1, CXCL5, CXCL9, CXCL10, IL-6 and IL-8 at 10, 50 or 100 pg/mL for 6 hours and labelled with anti-Nectin-1 (magenta) antibody and DAPI nuclear stain (grey). Fold change in the percentage of the area that is nectin-1 + , and relative to the mock, is displayed as mean ± S.D. (CCL3: n = 4, all others: n = 3). * = p<0.05, ** = p<0.01, *** = p<0.001, determined by repeated measures one-way ANOVA with Dunnett’s multiple comparisons or single unpaired parametric t tests. (C) Representative images of HaCaT cells treated with the concentrations required for optimal response are displayed. Scale bars = 100 μm. (D) HaCaT cells were treated with serum-free media (mock) or CCL3 or IL-8 at 10 pg/mL for 2 and 6 hours. RNA was extracted and RT-qPCR was performed using primers to nectin-1 and GAPDH. The fold change in nectin-1 expression by IL-8 and CCL3 treatment normalised to GAPDH is shown, displayed as mean ± S.D. (n = 2). Statistical analysis to assess data significance was determined by repeated measures two-way ANOVA (Treatment x Time) with Dunnett’s multiple comparisons test.

We reasoned that the collar of increased nectin-1 staining in uninfected HaCaT cells, around HSV1 plaques was likely the result of diffusing soluble mediators, probably chemokines or cytokines. To assess the mechanism behind this paracrine effect, the supernatants of HaCaT cells infected with HSV1-GFP for 12 hours or 24 hours were examined for various inflammatory cytokines and chemokines. HSV1-infected HaCaT cells at 12 h.p.i. produced significantly increased levels of CCL3 and IL-8, with trends in CCL2, CCL5, CXCL1, CXCL5 and CXCL10, all of which decreased by 24 h.p.i., especially for CCL2, CCL3, CXCL5 and IL-8 ( Fig 5A ). Only IL-6 was produced at both 12 and 24 h.p.i. in most samples. Interestingly, CXCL9 and CXCL11 were not detected at either timepoint in HSV1-infected HaCaT cells, although both chemokines, along with CXCL10, are ligands of CXCR3. We observed similar trends in the supernatants of HSV1-GFP infected foreskin explants at 24 h.p.i. ( S6 Fig ).

To determine whether these changes in nectin-1 staining were due to productive HSV1 infection or a direct interaction with gD on the virus (as shown previously by Barghava et al. [ 33 ], HaCaT cells were treated with HSV1-GFP for 24 hours and compared to cells treated with Foscarnet for 1 hour prior to HSV1 infection ( Fig 4D ). Only a few single cells were HSV1 + within the Foscarnet treatment, as expected (indicated by yellow arrows and lower gD expression as shown in Fig 4E ). However there was no early increase in nectin-1 staining in infected cells, nor the signature collar of nectin-1 + cells surrounding the infected cells at 24 h.p.i.. This was further proven by quantification of the nectin-1 + area (100±50mm 2 ) and maximum pixel intensity (20 000±15 000) when treated with HSV1 alone, compared to a decrease of about 50% in both when pre-treated with Foscarnet ( Fig 4E ). UV-inactivated HSV1-GFP was also used to determine whether nectin-1 redistribution could be induced by direct contact and entry of the inactivated virus (via surface gD) with keratinocytes. The pattern of staining of nectin-1 did not follow the serial increase and decrease in expression in HSV1-GFP-infected cells, nor was there any nectin-1 staining in uninfected cells ( Fig 4D and 4E ), indicating that direct interaction of the cell and virus surface glycoproteins (especially gD) is not sufficient to induce changes in nectin-1. Thus, HSV1 replication, late protein production and/or spread is essential for the redistribution of nectin-1 in foci and in surrounding cells.

We next used HaCaT cells to further investigate the dynamics of nectin-1 localisation which are similar phenotypically to the keratinocytes of the stratum spinosum, as determined by their expression of keratins 10 and 14 and function [ 34 , 35 ]. HaCaT cells grown on coverslips were mock or HSV1-GFP-infected for 24 hours ( Fig 4C and 4D ). Within mock-infected HaCaT cells, nectin-1 staining was absent in most cells, and where dim, was usually localised to the cell membranes ( Fig 4D ), presumably due to their location within adherens junctions, which may prevent efficient antibody binding [ 23 ]. Foci were first detected by ICP27 and GFP staining between 6 and 12 hours ( S5 Fig ). Nectin-1 staining was bright in some of the infected cells which co-expressed ICP27 + and/or GFP + indicative of earlier stage infection ( Fig 4C ). Nectin-1 staining was reduced in foci within 24 hours post infection (with increased expression of HSV1-GFP). Of particular note, the cells immediately surrounding the infected foci (ICP27 - GFP - ), showed increased nectin-1 staining. This was confined to the edges of the foci forming a defined collar or rim around them ( Fig 4D ).

(A) Cryosectioned 24-hour HSV1-GFP infected inner foreskin from two samples labelled by RNAscope to detect HSV1 DNA (red), anti-GFP (green), mouse anti-human Nectin-1 primary conjugated in house to Cyanine5 (Lumiprobe) and mouse anti-human Plakoglobin (PG, cyan) primary followed by donkey anti-mouse AF755 secondary antibodies and DAPI nuclear stain (grey). Scale bars = 100 μm. (B) Pixel intensity of nectin-1 within HSV1-GFP + or negative epidermal regions within the same tissue presented as mean ± standard deviation (n = 3). * = p<0.05, determined by an unpaired t test. (C) 24h HSV1-GFP infected HaCaT cells with arrows indicating ICP27 + GFP + Nectin-1 + (orange), ICP27 - GFP + Nectin-1 + (white) and ICP27 - GFP - Nectin-1 + (magenta) cells. Scale bars = 100 μm unless stated otherwise. (D) HaCaT cells cultured for 24 hours with either serum-free medium only (mock), HSV1 or HSV1-GFP (MOI 0.1), pre-treated with Foscarnet (1 mg/mL) for 1 hour (and remaining in the culture thereafter), or treated with UV-inactivated HSV1-GFP (UV-HSV1-GFP). Cells were labelled with anti-gD (green; when treated with untagged HSV1), anti-ICP27 (red) and anti-Nectin-1 (magenta) antibodies and DAPI nuclear stain (grey). Yellow arrows indicate gD -/low ICP27 + Nectin-1 - cells. Scale bars = 100 μm. (E) The average area (mm 2 ) and maximum pixel intensity for (i) nectin-1 and (ii) gD expression in mock, and HSV1 vs HSV1 + Foscarnet treated HaCaT cells, n = 2 and iii) nectin-1 and iv) GFP expression in mock, HSV1-GFP and UV-HSV1-GFP treated HaCaT cells. * = p<0.05 determined by an ordinary one-way ANOVA with Tukey’s multiple comparisons test.

To determine the effects of microtrauma and HSV1 infection on the distribution of nectin-1, we aimed to visualise nectin-1 and its localisation to the intercellular junctions between keratinocytes in inner foreskin epidermis. There are three types of intercellular junctions; tight junctions, adherens junctions, and desmosomes. We used Plakoglobin, which resides within both desmosomes and adherens junctions [ 20 ], as a control marker for intercellular junctions, to distinguish specific effects on nectin-1 (contained in adherens junctions) rather than the integrity of all intercellular junctions. RNAscope for HSV1 DNA was performed along with HSV1-GFP detection and anti-nectin-1 and anti-plakoglobin IF labelling in inner foreskin at 24 h.p.i.. Within the HSV1 + regions of all infected explants, nectin-1 staining was markedly decreased at 24 hours, where the distinct reticulated pattern marking the interface of the cell membranes observed in uninfected control epidermis, was completely absent on infected cells ( Fig 4A ). However, plakoglobin staining remained unchanged within the keratinocyte junctions, indicating that this phenomenon was specific to nectin-1, although plakoglobin staining did decrease in areas where epithelial integrity was lost. Quantitatively, there was a significant decrease in the staining intensity of nectin-1 in HSV1 + regions compared to HSV1 - regions (quantified as the average pixel intensity) within the tissue ( Fig 4B ). Nectin-1 has previously been shown to be reduced in HSV1 + cell lines [ 32 , 33 ]. Here, for the first time we observed a similar effect on HSV1-infected keratinocytes in situ in human foreskin tissue.

(A and B) Cryosectioned 24-hour HSV1-GFP infected inner foreskin from two samples labelled by RNAscope to detect HSV1 DNA (red), anti-GFP (green), anti-CD11c (blue) and anti-Langerin (magenta) antibodies and DAPI nuclear stain (grey). (A) Yellow arrows indicate HSV1 DNA + LCs, white arrows indicate HSV1 DNA + Epi DCs. Scale bars = 20 μm unless indicated otherwise. (B) Yellow arrows indicate HSV1 DNA surrounding a langerin + CD11c - LC and localised to the nucleus. Scale bars = 10 μm. (C) High resolution Z-stack with orthogonal view inset showing the colocalisation of DAPI (blue) and HSV1 DNA (red) overlapping to appear magenta, in a langerin + (yellow) LC. Image acquired on Olympus VS200 Slide Scanner at 100x magnification. Scale bars = 10 μm. (D) Pair-wise comparisons of the mean densities of LCs in HSV - (black circles) and HSV + (magenta circles) regions, and Epi DCs in HSV - (black circles) and HSV + (blue circles) regions. * = p<0.05, determined by paired t tests with Wilcoxon’s matched pairs signed rank test. (E) Percentage of HSV DAPI + DNA + and DAPI + DNA - Epi DCs of total DCs within foci presented as mean ± S.D. (n = 3). * = p<0.05, determined by paired parametric t tests.

LCs (CD11c - langerin + ) and Epi DCs (CD11c + langerin +/- ) interacted with HSV1 and HSV1-infected keratinocytes in multiple samples, indicated by yellow and white arrows respectively ( Fig 3A ). Epidermal LCs and DCs were observed in the infected regions as well as the surrounding areas, and as shown by RNAscope, and LCs were observed engaging in uptake of and infection by HSV1 ( Fig 3B and 3C ). High resolution Z-stacking confirmed the colocalization of HSV1 DNA within the nucleus of LCs, demonstrating these cells were infected ( Fig 3C ). Previous studies have shown LCs emigrate out of the tissue in response to external stimuli including viral infection [ 31 ]. We observed that there was no significant decrease in the density of Epidermal LCs and DCs over 24h in untreated tissue, suggesting that only a minority of cells appeared to migrate out of the tissue after patch treatment ( S4A and S4B Fig ). Thus, there was no significant further overall cell loss after treatment with patches, although a decreased density around punctures was noted ( S4C Fig ). Overall, there was very little change in the distribution of the epidermal LCs and DCs over time with or without patch-treatment, indicating that these cells remain dispersed in the epidermal tissue and are still prime targets for viral infection, even in the presence of microtrauma. However, both subsets showed a decrease in cell density within HSV1-infected regions with a significant decrease in LCs ( Fig 3D ). Quantification of the number of HSV1 DNA + DAPI + cells showed 90±5% of LCs within infected foci contained HSV1 DNA, as opposed to 45±40% of Epi DCs ( Fig 3E ).

Of the 13 samples processed, approximately half became infected, as assessed by HSV1 DNA and/or HSV1-GFP expression ( Table 1 ). Further analyses of the puncture wounds were performed to determine if the punctures had affected the level of virus penetration into the epidermis in infected and non-infected samples ( Fig 2F ). The HSV + samples all showed a deeper puncture depth (60±5 μm) than the HSV - samples (38±5 μm) which corresponded to 40±5% and 25±1% of total epidermal thickness in these samples respectively. These results indicate that the microneedles needed to penetrate at least 30–35% of the epidermal thickness, and pass the most superficial layers of keratinocytes (the strata corneum and granulosum), to establish detectable intra-epidermal infection.

(A-D) Cryosectioned 24-hour HSV1-GFP infected inner foreskin was labelled by RNAscope to detect HSV1 DNA (red), anti-GFP antibody (green) and DAPI nuclear stain (grey). Scale bars = 100 μm unless indicated otherwise. (A) HSV1 infected keratinocytes of the mid-epidermis, but not upper layers, as indicated by HSV1 DNA. (B) HSV1 infection cascade is shown with HSV1 DNA only detected in lower epidermis, and both DNA and GFP detected in upper layers. (C) Single virions or small aggregates of particles (right inset: zoomed in image of smaller particles) of HSV1 DNA observed within the puncture region (dotted line; left inset: zoomed in image in different Z plane) indicated by the yellow arrow at 100x magnification, imaged using the Olympus VS200 Slidescanner. (D) Nuclei of infected cells show various infectious states within keratinocytes; nuclear globular DNA alone (green arrow), nuclear globular DNA with cytoplasmic GFP (white arrows), single punctate virions and cytoplasmic GFP (yellow arrow). (E) A table of averages for the measurements of all HSV1 + explants infected with HSV1 including; the average number of HSV1 + regions per mm of epidermis, proportion of HSV1 + area, and average size of the HSV1 + foci in width (μm) and in number of cells (n = 5). (F) 24-hour HSV1-GFP infected inner foreskin samples were classified as HSV - (no expression of GFP or DNA) or HSV + (expressing GFP and/or DNA) (i) the puncture depth and (ii) the puncture depth proportional to the total epidermal thickness were measured. Data presented as mean ± S.D. (HSV - : n = 3, HSV + : n = 4). * = p<0.05, determined by unpaired parametric t tests with Welch’s correction assuming unequal variances.

HSV1 DNA was also targeted using RNAscope and showed a large quantity of viral DNA being produced by infected keratinocytes ( Fig 1C ). Using both RNAscope and GFP-labelling, which labelled a late stage protein, facilitated the detection of different stages of infection within the explants, as HSV1 DNA in foci was detected in some samples that did not express any GFP signal and also in others at the periphery of GFP + foci, indicating that those keratinocytes were at an earlier stage of infection ( Table 1 and Fig 2A and 2B ). Occasionally, individual virions or small viral aggregates were seen within the epidermis, most clearly within the punctures (indicated by yellow arrow and dotted line), but also between keratinocytes ( Fig 2C ). DNase pretreatment of HSV1-infected tissue showed that the presence of these virions or aggregates was not due to free DNA from the viral inoculum ( S3 Fig ). HSV1 DNA was detected within infected cells and often in cells co-expressing GFP, indicating late viral protein production ( Fig 2D ). HSV1 DNA was mostly detected within the nuclei of infected keratinocytes where it appeared focal in some cells and diffuse in others, and in some cells also within the cytoplasm, in the presence or absence of GFP. Analysis of six HSV1 + foreskin samples showed that at 24 h.p.i. there was a mean of 0.42 (with a range of 0.05–2.8) HSV1 + regions per mm of epidermis ( Fig 2E ). The average HSV1 + area made up 6.25% (with a range of 0.8–29%) of the total area of epidermis and the average width of an infected region was 326 (with a range of 14–860) μm. There was great variability in lateral spread of the infection away from the puncture within 24 hours, with an average spread of 13 cells from the puncture in two dimensions, and a range of 1–47 cells within one section. As the HD-MAPs were 1cm x 1cm with a density of 10 000 microneedles, one row contains 100 microneedles and with an average of 5.46 microneedle punctures per mm of epidermis. This lead to approximately 8% of microneedles inducing HSV1 foci. Where foci were small these were usually located within the stratum spinosum of the mid-epidermis, indicating infection could be initiated in this region. Larger foci were spread across almost all layers of the epidermis and spread laterally as well. Thus, microtrauma facilitated the establishment of large regions of intra-epidermal HSV1 infection in human inner foreskin tissue with substantial lateral viral spread within 24 hours.

As microtrauma is known to facilitate viral entry into human skin, HD-MAPs were used to induce microtrauma in inner foreskin epidermis ( Fig 1A ). These HD-MAPs were applied using an automatic mechanical applicator and induced punctures in adult epidermis to a depth of 40±20 μm or approximately 40% or less of the epidermal thickness ( S2 Fig ), thus inducing consistent punctures that did not penetrate beyond the basement membrane (BM) into the dermis, allowing efficient epidermal viral entry. Analysis of multiple child foreskin samples (ages 5–14 y.o.) and adults (ages 17 y.o. and above) showed that the thickness of the epidermis increases with age ( S2A and S2B Fig ). The child inner foreskins had a thinner epidermis of 100±3 μm in comparison to adult inner foreskins which had a thickness of 150±25 μm. Infant foreskins (under age 5) with even thinner epidermis were omitted from this study as most (80±5%) punctures penetrated the BM ( S2B Fig ) and could facilitate atypical infection within the dermis. After patch-treatment, topical application of the virus did not result in viral entry into the epidermis. Thus, the HD-MAPs were coated with a solution containing HSV1-GFP (tagged to US9, a late infection protein) before application to the tissue ( Fig 1A ). Intra-epidermal infection of human inner foreskin was detected by GFP signal at 24 hours post infection (h.p.i.) in association with microneedle punctures ( Fig 1B ). The virus infected the keratinocytes surrounding the punctured regions as indicated by the yellow arrows, forming viral plaques within the epidermis, similar to herpetic lesions.

In preliminary experiments, RNAscope was used to detect intracellular HSV1 DNA, specifically the long unique region (UL) -30, and individual HSV1 particles after topical application of HSV1 to explants of child and adult inner foreskin. Explants were infected with HSV1 via a cloning cylinder attached to the surface of the epidermis of the inner foreskin and cultured for 24 hours. HSV1 was detected but only on the mucosal surface, without any keratinocyte infection ( S1A Fig ). Therefore, without a breach in the epidermis, especially of the stratum corneum, HSV was not able to enter the skin of children or adults.

Discussion

This study aimed to develop a model of acute HSV infection at the site of sexual transmission and determine the cellular and cytokinetic factors that enhance viral spread within the epidermis. After topical administration of HSV to explants of human inner foreskin, HSV did not enter the intact epidermis, and microtrauma, via HD-MAPs, was required for entry into this tissue by bypassing the two upper strata of keratinocytes. Entry was probably also aided by initial redistribution of the surface expressed nectin-1 out of sequestration in the adherens junctions on keratinocytes abutting the punctures. Pre-coating of the projections of HD-MAPs with HSV1 prior to application to the tissue was strictly necessary to establish infection. No infection was established if the punctures were introduced prior to adding the virus. This suggests that for transmission, virus containing fluids must be present at the time of microtrauma during intercourse in vivo. Infection of inner foreskin explants from multiple samples showed that HSV1 spread laterally away from the puncture at an average of 13 cells in each infected focus within 24 hours. As a single viral replication cycle of HSV1 takes approximately 12–18 hours [36,37], the duration of infection in situ represented 1–2 rounds of cell infection at most and yet managed to spread across 13 cells, on average. This suggests that virus is not only spreading cell-to-cell via de novo replication, but also through the interstitial spaces or across the surfaces or cell membranes of adjacent keratinocytes without infection. This rapid spread could involve viral ‘surfing’ [38] or partly by other mechanisms.

This study expanded our previous studies that used GFP labelled virus by incorporating RNAscope, enabling the detection of HSV1 DNA in individual particles and small aggregates in punctures, within and between cells for the first time. RNAscope was used in our previous study to detect HIV interactions with epidermal LCs and DCs within inner foreskin epidermis [30]. Using RNAscope for HSV1 DNA we were able to show earlier stages of infection in several different distribution patterns; 1) Focal nuclear DNA detection in infected keratinocytes, presumably representing the earliest accumulation within ND10 domains [39,40]; 2) Diffuse nuclear staining, probably mostly as nuclear capsids; 3) Nuclear and cytoplasmic staining, the latter presumably in capsids and assembling virions; 4) Nuclear DNA and cytoplasmic US9-GFP in infected keratinocytes. Some DNA-expressing groups of keratinocytes were also seen that did not express any HSV1-GFP which may be due to early stages of infection, or that only early infection was established (i.e. abortive infection).

At first glance our results appear to be in contrast to reports from the Knebel-Morsdorf laboratory [4,5] which found that in explants of human abdominal and breast skin and oral mucosa and murine skin, mechanical trauma was insufficient for topical HSV-1 entry and infection. They established successful HSV infection of skin by stripping the dermis and submerging the epidermis in a viral solution, resulting in infection of the basal epithelial layers and spread superficially, but only to the stratum spinosum as the stratum granulosum did not become infected even when treated with mechanical abrasions. However, this does not appear to explain sexual transmission in vivo. Although we could initiate topical infection of genital mucosal explants by automated patch delivery of HSV, we also found that only epidermal layers below the stratum granulosum (i.e. below 30–35% of the depth of the epidermis) could be infected. Such infection could be initiated in the stratum spinosum as well as the deeper stratum basale. Indeed small and larger foci of HSV infection were often confined to the stratum spinosum in the mid-epidermis without involving basal layers, The stratum granulosum is probably refractory to HSV1 infection due to marked biological changes including increased accumulation of keratin and lipids [5,25,41,42].

Recent results from the same group, using reconstituted epidermis from dissociated human keratinocytes, also suggested that tight junction formation superficial to the adherens junctions containing nectin-1 was a major barrier to HSV interacting with nectin-1 after topical addition. However, conditions simulating atopic dermatitis allowed such infection. There are several reasons for the differences to our system. Firstly, we could achieve topical infection by viral coating of the HD-MAPs. Secondly, spread of HSV from punctures and through interstitial spaces clearly led to expansion of foci of infected keratinocytes consistent with the marked differences in the distribution of intercellular junctions in the upper strata of inner foreskin epithelium resulting in a reduced barrier function, compared to external skin [26,27]. Thirdly, our RNAscope technique is far more sensitive in detecting viral infection, including individual or aggregated virions, than conventional immunofluorescence staining. Together, the results from the two laboratories are consistent with human in vivo HSV infection and disease which is common in the genital tract but uncommon in external skin, unless the patient has an underlying inflammatory skin disease, such as atopic dermatitis.

Our study also contributes further understanding of the interaction between HSV and resident LCs/DCs in the epidermis of human genital mucosa after topical infection, adding to our previous work. The Epi DCs and LCs within inner foreskin epidermis, were distributed throughout the epidermis, allowing them to encounter incoming viruses and pathogens. In situ quantification showed a similar concentration of both epidermal subsets [30]. We previously showed HSV replicates within LCs to express structural proteins such as GFP-labelled pUL37/US9 in inner foreskin [3], using IF microscopy after topical application to the surface of infant inner foreskin tissue. Unlike RNAscope used here, this technique does not detect particle uptake nor early infection. In the absence of microtrauma, infection of LCs, but only of occasional keratinocytes, were detected. Probably, the thinner upper epidermal strata in infants allowed such LC infection but the sequestration of nectin-1 in adherens junctions in the absence of trauma explained the lack of keratinocyte infection. In this study, within HSV-treated epidermis, the cell density for both LCs and Epi DCs was lower adjacent to the HSV1+ regions than to HSV1- regions, suggesting that the infected cells migrate out of the epidermis to interact with dermal cells, as previously shown [3]. In a more recent study, we showed that both infected LC and Epi DCs undergo apoptosis in vitro [14] and thus we presume both cell types migrate to the dermis in response to HSV1 infection, while undergoing apoptosis, to be taken up by dermal DCs. This will be addressed in future studies.

A limitation of this foreskin explant model is its ‘closed-system’ nature, being isolated from the vascular system carrying soluble factors, maintaining tissue oxygenation and immigration of circulating inflammatory cells, such as plasmacytoid DCs, monocytes, and T cells which control HSV infection. However, this does not diminish the biological relevance of such a model as it very closely resembles the tissue microenvironment that HSV would initially encounter at early times in vivo. We found that the epidermis remains intact during ex vivo culture for duration of our experiments (up to 24 hours) as determined by histology. Cellular composition and relationships within the epidermis were also not significantly altered within this time. Another limitation of this model is sample variability in tissue samples, requiring increased number of replicates for significant results.

The role of the HSV1/2 receptor nectin-1 is essential to our understanding of the initial events of HSV infection within genital tissue. It is highly expressed on keratinocytes and LCs/DCs [4,16,30,33,43]. Nectin-1 staining was completely and specifically diminished in HSV1-infected areas within the epidermis, however the normal reticular staining pattern remained intact within the surrounding uninfected cells. Plakoglobin expression was maintained in the foci, suggesting disruption of adherens junctions but not desmosomes between HSV1-infected keratinocytes [5,17,24]. We chose to model the effects of HSV1 infection on nectin-1 redistribution in the human keratinocyte HaCaT cell line as they can express abundant nectin-1 and HSV spread in these cells has been studied by a number of laboratories [44,45]. Although HaCaT cells are an in vitro model, they are representative of the suprabasal layers of the epidermis, characterised by their expression of keratins 1 and 10 [46,47] where HSV1 infection was established. There was natural cycling of nectin-1 in HaCaT cell cultures (although not at the transcriptional level), suggesting this cycling was due to nectin-1 protein redistribution between adherens junctions in the cell membrane and the cytoplasm (S7 Fig). We quantified HSV1 and chemokine effects at the time of minimal nectin-1 staining (i.e. 18 hours post-seeding), clearly showing a marked reduction in nectin-1 staining within infected foci and its inhibition by Foscarnet. The collar of redistributed nectin-1 in the adjacent cells surrounding the area of infection were ICP27-GFP-, indicating that they were not yet infected (or not yet expressing viral proteins), but exposed to nectin-1-inducing soluble factors, probably cytokines and chemokines secreted from infected foci. For comparison, changes in nectin-1 also occurred in the majority of HSV-transfected mouse melanoma cells within a monolayer but were induced by relatively few virions, perhaps suggesting that the interaction induces a paracrine effect on surrounding cells [33].

Chemokines from HSV1-infected HaCaT keratinocytes showed a tendency to increase at 12 h.p.i. then decrease again by 24 h.p.i., correlating with the kinetics of nectin-1 redistribution. Our studies in infected HaCaTs and to some degree infected explants expanded our previous data on cytokine and chemokine production by HSV infected keratinocytes and correlates with those in recurrent herpes vesicle fluid [48]. Both studies showed production of IL-6, -10 and -12 and CCL2, -3, and -5 and changes over time in culture (or for Mikloska et al. [48], times of sampling of vesicle fluid). However, our data also shows enhanced production of IL-6, IL-8 and a much wider range of chemokines. Addition of CCL3 and CCL5, which bind to keratinocyte-expressed CCR1; CXCL1, CXCL5 and IL-8 which bind to CXCR2; and CXCL9 and 10, which bind to CXCR3, showed increased staining of nectin-1 by the HaCaT cell monolayer. This was significant for CCL3 and IL-8 with marked trends for the others. CXCL9, 10 and 11 were also shown to play a critical role in HSV infection within mice [49,50]. The individual variability in chemokine effects, may be due to chemokine receptor-targeting by alternate chemokines in different people. The signalling pathways downstream from these chemokine receptors intersect, leading to NFκB activation and suggesting a potential mechanism for nectin-1 redistribution, which will be tested in future experiments [51–55]. IL-6, the ligand for IL-6 receptor (IL-6R) also induced nectin-1 redistribution. However, whether this receptor-ligand interaction activates NFκB has been variable in previous reports [56–60]. Nevertheless, crosstalk between STAT3, PI3K and NFκB pathways has been reported previously and may explain the redistribution of nectin-1 within keratinocytes [61,62]. Therefore, our data shows that keratinocytes are producing cytokines and chemokines that enhance nectin-1 redistribution on neighbouring cells and lead to the redistribution of the entry receptor from the cellular membrane to facilitate viral spread.

In conclusion using a novel combination of HSV DNA detection, an authentic genital mucosal explant and a system for reproducible microtrauma, this study provides insight into events that resemble those occurring in vivo. The results strongly suggest that sexual transmission of HSV requires induction of microtrauma penetrating the stratum spinosum during sexual intercourse. There was rapid spread of virions probably via intercellular spaces within 24 hours to produce infected foci of keratinocytes on average 13 cells wide and of both types of epidermal MNPs, suggesting prevention or therapy should be aimed at free virus as well as infected cells. In HaCaT culture models, spread was facilitated by chemokine induction of a collar of redistributed nectin-1 in uninfected keratinocytes surrounding the foci of infected cells which secreted the relevant chemokines CCL3 and IL-8, as well as IL-6, CXCL1, CXCL5, CXCL9 and CXCL10. LCs and Epi DCs also take up virus and appear to migrate out of the epidermis, consistent with the pathway published previously on the early steps of the immune response [3]. Further work is underway to confirm that Epi DCs also undergo apoptosis and migrate to the dermis as do LCs.

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1012267

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/