Eighty-eight patients diagnosed with HSV-2 infection were recruited at the STD clinic at Sahlgrenska University Hospital and Infectious Disease clinics at Sahlgrenska University Hospital, Skaraborg Hospital and Södra Älvsborg Hospital. A diagnosis of HSV-2 meningitis required symptoms of viral meningitis and CSF pleocytosis levels above 5 X 10⁶ leukocytes/L with an HSV-2 infection etiology. The etiology was verified by (1) detection of HSV-2 DNA in the CSF by PCR, (2) preceding or concurrent virologically verified HSV-2 lesions in the genital or lumbosacral region, or (3) a history of previous aseptic meningitis of herpetic or unknown origin and positive HSV-2 serology in the acute phase serum sample. For criteria 2 and 3, negative results in bacterial culture and PCR for enterovirus in CSF and negative serology results for tick-borne encephalitis virus were compulsory. A diagnosis of HSV-2 genital herpes required symptoms of genital herpes and an HSV-2 infection etiology confirmed by HSV-2 type-specific PCR. All patients were enrolled either during remission or while on prophylactic anti-viral treatment. All patients donated 20 ml of blood collected in heparin tubes. Permission for this study was granted by the Ethics committee of University of Gothenburg, and all patients gave informed consent. The patients were divided into two groups based on their clinical status, HSV-2 meningitis or HSV-2 genital herpes only ( Table 1 ).

Levels of HSV-2 glycoprotein mgG2-specific antibodies were determined in plasma with a semi-quantitative routine analysis (LIAISON HSV-2 IgG; Diasorin) at the Clinical Microbiology unit, Sahlgrenska University hospital. Co-infection with HSV-1 was assessed using an HSV-1 ELISA kit (HerpesSelect1 ELISA IgG; Focus Technologies). DNA was purified from plasma using Qiagen DNeasy Blood and Tissue Kit (Qiagen) and HSV-2 DNA was detected using qPCR with the HSV-2 TaqMan probe Vi04646232_s1 and TaqMan™ Universal PCR Master Mix (ThermoFisher) for 40 cycles.

The lab-adapted strain HSV-2–333 was obtained as described (15). For antigen preparation, virus was inactivated with UV-light for 30 minutes.

Freshly isolated PBMC from HSV-2-infected patients (1x10⁶ cells/ml) were cultured in 96-well plates in a total volume of 200 µl x-vivo medium (Lonza, Verviers, Belgium) supplemented with 1% L-glutamine in the presence or absence of HSV-2 or UV-inactivated HSV-2 (corresponding to 4x10⁵ p.f.u/ml). Culture supernatants were collected after 48 hours and stored at -20 C until assayed for content of eighteen immune factors (IFN-α, IFN-β, IFN-γ, IFN-ω, IL-2, IL-29, IL-6, IL-8, LIF, TNF-α, CXCL9, CXCL10, CXCL11, CX3CL1, Perforin, Granzyme B, PD-1, TIM3) using a Human Custom ProcartaPlex 18-plex beads array from Invitrogen (ThermoFisher Scientific, Bender MedSystems GmbH, Vienna, Austria) according to the manufacturer’s instructions. In one experiment, PBMC from two HSV-infected donors were exposed to 2 μg/ml Anifrolumab (Saphnelo^®, Astrazeneca, kind gift from Professor Sören Paludan) or isotype control (Ultra-LEAF™ Purified Human IgG1 Isotype Ctrl Recombinant, Biosite) 60 min before and continuously during incubation with HSV-2 or UV-inactivated HSV-2. Total RNA was isolated from cells after 6 hours using RNase Mini kit (74106, Qiagen). TaqMan RNA-to C_t 1-Step kit, and Taqman probes (Life Technologies) were used to measure the mRNA expression levels of IFNγ (Hs00989291), IFNα2 (Hs00265051), IFNβ1 (Hs01077958) and TATA-box binding protein (TBP) (Hs00427620).

Multivariate analyses of pattern recognition “Orthogonal Projections to Latent Structures by means of Partial Least Squares Discriminant Analysis” (OPLS-DA) were performed using the SIMCA-P (version 15.0.2) statistical package (MKS Data Analytics Solutions, Malmö, Sweden). The quality of the models was evaluated by their explanatory power (R2Y) and robustness (Q2Y). Mann Whitney signed rank test was used to compare continuous data and Fisher´s exact test was used to compare categorical data comparing HSV-2 meningitis and HSV-2 genital herpes patients. Kruskal-Wallis one-way ANOVA non-parametric test, with Dunn’s post-test, was used to compare cytokine secretion patterns in PBMC from HSV-2-infected patients in response to different stimuli, two-way ANOVA with Šidák’s multiple comparison test was used to compare the relative mRNA levels of IFNα2, IFNβ1 and IFNγ in PBMC treated with Anifrolumab, and simple linear regression was used to analyze the relationship between age of the patients and cytokine secretion by patient PBMC. These different statistical analyses were performed using GraphPad Prism software 10.3.0 (GraphPad, San Diego, CA, US). A P-value <0.05 was statistically significant.

Eighty-eight patients with a clinical diagnosis of HSV-2 infection were recruited to the study. The median age at diagnosis was 34 years. 73% were women and 55% had been diagnosed with HSV-2 meningitis (of which 50% also had genital and/or sacral herpes) and the remaining 45% with HSV-2 genital herpes only. 34% of the patients were co-infected with HSV-1 at the time of blood sampling and 19% received prophylactic anti-viral treatment ( Table 1 ). The HSV-2 meningitis patients were significantly older that the HSV-2 genital herpes patients, both at diagnosis and at time of sampling. However, there was no statistically significant difference between the groups regarding time from diagnosis to sampling ( Table 1 ).

Firstly, we examined the immune activation in PBMC exposed to either UV-inactivated HSV-2 (“HSV-2 antigens”) or to replication-competent HSV-2 (“HSV-2”) and measured the secretion of eighteen immune factors in culture supernatants after 48 hours by multiplex analysis. We observed three distinct response patterns for the different immune factors; i) no upregulation in response to HSV-2, ii) stronger response to replication-competent HSV-2 compared to HSV-2 antigens, and iii) equally strong response to HSV-2 antigens and replication-competent HSV-2. Only one of the 18 immune factors, IFN-β, was not altered after exposure to HSV-2 antigen or to HSV-2 ( Figure 1A ). Five of the immune factors (IFN-α, IFN-ω, CXCL10, IL-8 and perforin) were secreted to a higher degree after exposure to replication-competent HSV-2, indicating that these responses required viral replication ( Figures 1B–F ). The remaining twelve immune factors were induced to a similar degree by HSV-2 antigens and replication-competent HSV-2, indicating that these responses were antigen-specific and did not require viral replication ( Figures 2A–L ).

Secondly, we examined if demographic factors (age, sex, prophylactic treatment, HSV-1 co-infection) and immunological parameters (anti-HSV-2 antibody levels, steady-state secretion of cytokines, chemokines and cytotoxic proteins and de-novo production of cytokines, chemokines and cytotoxic proteins in response to replicating HSV-2 and to HSV-2 antigens) could be used to distinguish HSV-2 meningitis from HSV-2 genital herpes using a multivariate OPLS-DA model. This approach rendered a stable model (Q2Y=0.47) with an explanatory power of 66% (R2Y=0.66) which indicates that these two patient groups differed significantly from one another ( Supplementary Figure S1A ). Meningitis patients tended to be older and female whereas the patients with genital herpes were more often HSV-1 positive and received prophylactic treatment ( Supplementary Figure S1B ). Meningitis patients had considerably stronger steady-state production of several anti-viral cytokines, chemokines and cytotoxic proteins and had an overall higher de-novo production of cytokines and chemokines in response to replicating HSV-2 ( Supplementary Figure S1B ).

To rule out that demographic data dictated the separation of the meningitis and genital herpes cohorts in the OPLS-DA analysis, we performed a second analysis excluding age, sex, prophylactic treatment, and co-infection with HSV-1. Exclusion of demographic data somewhat improved both the stability (Q2Y=0.49) and the explanatory capacity (R2Y=0.67) of the model ( Figure 3A ) indicating that the immune parameters account for the major differences between patients with HSV-2 meningitis and HSV-2 genital herpes. The second model confirm that both the steady-state production of anti-viral cytokines proteins and the de-novo production of cytokines and chemokines in response to replicating HSV-2 is higher in meningitis patients ( Figure 3B ). The levels of anti-HSV-2 antibodies had a minor role in distinguishing the two cohorts and did not differ significantly between patients with meningitis and genital herpes in either of the models ( Supplementary Figure S1B ; Figure 3B ) and thus contributed very little to the separation of the cohorts.

Next, we performed univariate statistical analyses of the major immune parameters that were identified in the OPLS-DA model ( Figure 1B ). We thus confirm that PBMC from HSV-2 meningitis patients spontaneously secrete significantly higher levels of several potent antiviral cytokines (IFN-α, IFN-β, IFN-ω, IL-29, CXCL9, and perforin) during steady state (i.e. in the absence of viral proteins and viral replication) ( Figures 4A–F ) indicating ongoing anti-viral activation of the immune system also during remission. Secondly, PBMC from HSV-2 meningitis patients had significantly higher de-novo production of Th1 and inflammatory cytokines (IFN-γ, IL-6, IL-8 and LIF) after exposure to replicating HSV-2 ( Figures 5A–D ) which indicate that PBMC from HSV-2 meningitis patients react to the virus with a stronger inflammatory and recall T-cell mediated immune response than PBMC from patients with HSV-2 genital herpes. The IFN-γ and LIF (but not IL-6 and IL-8) responses to HSV-2 antigen were also higher in PBMC from meningitis patients, indicating that these two cytokines represented recall responses ( Figures 5E–H ).

Given that meningitis patients were older and more often women whereas genital herpes patients were more likely to be HSV-1 seropositive and on anti-viral prophylactic treatment, we also assessed if the levels of cytokines produced differed with age, gender, HSV-1 co-infection and prophylactic treatment. PBMC from females spontaneously secreted significantly higher levels of IL-29 compared to PBMC from men ( Supplementary Figure S2 ), while there were no statistically significant sex-related differences in the spontaneous secretion of IFN-α, IFN-β, IFN-ω, CXCL9, or perforin, nor in the de-novo production of IFN-γ, IL-6, IL-8 or LIF (data not shown). The only statistically significant age-related difference in cytokine production was spontaneous secretion of IFN-β, which degreased with increasing age (p=0.043). There were no significant differences in the production of any of these cytokines between patients that did, or did not, receive prophylactic treatment, nor did the cytokine secretion patterns differ between patients that were, or were not, HSV-1 seropositive (data not shown).

To establish if there is a possible functional connection between the enhanced de-novo production of type I IFNs in HSV-2 meningitis patients and the enhanced HSV-2-induced production of IFN-γ, we blocked signaling through the type I IFN receptor in PBMC using Anifrolumab and then exposed the cells to HSV-2 antigen or replication-competent HSV-2 and measured the transcription levels of IFN-α, IFN-β and IFN-γ. As expected, Anifrolumab-treatment significantly reduced the mRNA levels of both IFN-α and IFN-β ( Supplementary Figure S3 ). Anifrolumab also significantly reduced the transcription levels of IFN-γ by PBMC in response to HSV-2 antigen and replication-competent HSV-2 ( Figure 6 ), indicating that signaling though the type I IFN receptor promote recall IFN-γ production.

To assess if the enhanced spontaneous production of type I IFNs by PBMC from meningitis patients reflected ongoing viral exposure, we measured presence of HSV-2 DNA in plasma from the meningitis patients. However, none of the convalescent meningitis patients had any detectable HSV-2 DNA in their plasma (data not shown).