VZV Infection of Primary Human Adrenal Cortical Cells Produces a Proinflammatory Environment without Cell Death

Virus infection of adrenal glands can disrupt secretion of mineralocorticoids, glucocorticoids, and sex hormones from the cortex and catecholamines from the medulla, leading to a constellation of symptoms such as fatigue, dizziness, weight loss, nausea, and muscle and joint pain. Specifically, varicella zoster virus (VZV) can produce bilateral adrenal hemorrhage and adrenal insufficiency during primary infection or following reactivation. However, the mechanisms by which VZV affects the adrenal glands are not well-characterized. Herein, we determined if primary human adrenal cortical cells (HAdCCs) infected with VZV support viral replication and produce a proinflammatory environment. Quantitative PCR showed VZV DNA increasing over time in HAdCCs, yet no cell death was seen at 3 days post-infection by TUNEL staining or Western Blot analysis with PARP and caspase 9 antibodies. Compared to conditioned supernatant from mock-infected cells, supernatant from VZV-infected cells contained significantly elevated IL-6, IL-8, IL-12p70, IL-13, IL-4, and TNF-α. Overall, VZV can productively infect adrenal cortical cells in the absence of cell death, suggesting that these cells may be a potential reservoir for ongoing viral replication and proinflammatory cytokine production, leading to chronic adrenalitis and dysfunction.


Introduction
Virus infection of adrenal glands is an important, yet under-recognized, cause of adrenal dysfunction. The adrenal glands are part of the hypothalamic-pituitary-adrenal (HPA) axis, which is a major neuroendocrine system that regulates a spectrum of physiological functions. The adrenal gland produces glucocorticoids, mineralocorticoids, and sex hormones in the outer adrenal cortex and catecholamines in the inner adrenal medulla; these hormones regulate metabolism, the immune system, blood pressure, response to stress and other essential functions. Virus infection can disrupt hormone secretion, leading to a constellation of non-specific symptoms such as fatigue, dizziness, abdominal pain, weight loss, nausea, vomiting, and muscle and joint pain.
Recent reports indicate that varicella zoster virus (VZV) may be an important pathogen contributing to adrenal disease. VZV is latent in 6% of human adrenal glands [1] and may also produce adrenal disease during primary infection or reactivation. Support for VZV's contributions to adrenal dysfunction during primary infection is provided by case reports of individuals developing bilateral adrenal hemorrhage (Waterhouse-Friderichsen Syndrome) associated with varicella [2][3][4][5]. Additionally, in non-human primates (NHP) with spontaneous disseminated simian varicella virus (SVV; the NHP homolog of VZV) infection, adrenal glands contain both SVV DNA and antigen, predominantly in the cortex, associated with Cowdry A inclusion bodies, cellular necrosis, multiple areas of hemorrhage, and varying amounts of polymorphonuclear cells [6]. However, the mechanisms by which VZV affects the adrenal glands are not well-characterized. Thus, we determined if primary human adrenal cortical cells (HAdCCs) infected with VZV support viral replication and produce a proinflammatory environment, potentially contributing to adrenal dysfunction.

Cells and Virus
Primary human adrenal cortical cells (HAdCCs; ScienCell, Carlsbad, CA, USA) were used; cell type was confirmed by an immunofluorescence antibody assay showing the presence of glial fibrillary acidic protein (GFAP; present in adrenal cells in the cortex) and absence of tyrosine hydroxylase (TH; present in adrenal chromaffin cells in the medulla). HAdCCs were seeded at 5000 cells/cm 2 in mesenchymal medium containing 5% fetal bovine serum (FBS), 1% mesenchymal stem cell growth supplement, and 1% 100X penicillinstreptomycin (ScienCell). After 24 h, media was changed to media with 0.1% FBS and 1% 100X penicillin-streptomycin and was replenished every 72 h for 7 days, establishing quiescence (qHAdCC). On day 7, uninfected (mock) or VZV-infected qHAdCCs (60 plaqueforming units/cm 2 ; VZV Gilden strain [GenBank accession number MH379685] [7]) were added to qHAdCCs. For analysis, cells were harvested at 1, 2, or 3 days post-infection (DPI), and conditioned supernatants were collected at 3 DPI.

DNA Extraction and Quantitative PCR (qPCR)
At 1, 2, and 3 DPI, DNA was extracted from mock-and VZV-infected qHAdCCs using the DNeasy Blood and Tissue Kit per manufacturer's instructions (Qiagen, Germantown, MA, USA). DNA was analyzed with custom qPCR primers (Integrated DNA Technologies, Coralville, IA, USA) corresponding to VZV open reading frame 68 and glyceraldehyde-3phosphate-dehydrogenase (GAPdH) as described [8]. Known concentrations of VZV DNA were used as qPCR standards to determine DNA copy number. Data were quantified using a standard curve for VZV and normalized to ng of DNA loaded.

Statistical Analysis
Statistical analysis was performed using graphing software (GraphPad Prism; Graph-Pad; San Diego, CA, USA). Cytokine differences between mock-and VZV-infected qHAdCC supernatant were determined by multiple unpaired t-tests with a False Discovery Rate (q-value 0.05). Significant differences in viral titer were determined using a 1-way ANOVA with a Tukey's multiple comparisons test.

VZV Infection of qHAdCCs Does Not Cause Cell Death
Many non-neuronal cell types undergo cell death during VZV infection [11][12][13][14][15][16]. To examine whether VZV-infected qHAdCCs undergo cell death, mock-and VZV-infected qHAdCCs were fluorescently labeled with a TUNEL stain at 3 DPI ( Figure 2A). There was no difference in abundance or intensity of fluorescent label between the two conditions. To further determine if apoptosis was occurring, we examined mock-and VZV-infected cell lysates with Western blot using antibodies against PARP and caspase 9. Cleaved caspase 9 and PARP are indicative of apoptosis (37 kDa and 89 kDa, respectively). No-tominimal cleaved caspase 9 and PARP was seen in VZV-infected cell lysates, similar to that found in mock-infected lysates, indicating that VZV-infected qHAdCCs are not undergoing differing amounts of apoptosis over mock-infected cells. As a positive control for apoptosis, HLFs were treated with 1 µM saurosporine; bands for both full length and cleaved caspase 9 and PARP were detected, indicative of apoptosis. Taken together, these results demonstrate that qHAdCCs are resistant to cell death during VZV infection.

VZV Infection of qHAdCCs Does Not Cause Cell Death
Many non-neuronal cell types undergo cell death during VZV infection [11][12][13][14][15][16]. To examine whether VZV-infected qHAdCCs undergo cell death, mock-and VZV-infected qHAdCCs were fluorescently labeled with a TUNEL stain at 3 DPI (Figure 2A). There was no difference in abundance or intensity of fluorescent label between the two conditions. To further determine if apoptosis was occurring, we examined mock-and VZV-infected cell lysates with Western blot using antibodies against PARP and caspase 9. Cleaved caspase 9 and PARP are indicative of apoptosis (37 kDa and 89 kDa, respectively). No-to-minimal cleaved caspase 9 and PARP was seen in VZV-infected cell lysates, similar to that found in mock-infected lysates, indicating that VZV-infected qHAdCCs are not undergoing differing amounts of apoptosis over mock-infected cells. As a positive control for apoptosis, HLFs were treated with 1 µM saurosporine; bands for both full length and cleaved caspase 9 and PARP were detected, indicative of apoptosis. Taken together, these results demonstrate that qHAdCCs are resistant to cell death during VZV infection.

Discussion
Herein, we found that VZV infects adrenal cortical cells and produces a proinflammatory environment, in the absence of cell death, under our experimental conditions (analysis at 3 DPI in a qHAdCC monoculture). Specifically, VZV DNA increased over time, indicative of productive virus infection. When compared to supernatant from mock-infected cells, elevated IL-6 and IL-8 were seen in supernatant from VZV-infected cells. Further, IL-12p70, IL-13, IL-4, or TNF-α were absent in mock-infected qHAdCC supernatant but present in VZV-infected qHAdCC supernatant. Taken together, these observations shed light on the potential clinical manifestation of VZV infection of human adrenal glands.
Conditioned supernatant from VZV-infected cells contained elevated IL-6 that can potentially affect adrenal function. While cortisol and aldosterone levels could not be measured in mock-and VZV-infected cells due to paucity of available adrenal cortical cells, elevated IL-6 has been reported to increase plasma cortisol and aldosterone levels in several studies for full review see ref. [17]. Steensberg and colleagues [18] showed IL-6 induces an increase in cortisol and consequently an increase in circulating neutrophils. It is important to note that Jarosinski and colleagues [19] have documented elevated levels of IL-6 in human fetal skin explant models, which suggests IL-6 may be a hallmark of VZV infection in multiple tissues and cell types. Chronic increased cortisol levels can lead to symptoms of Cushing's syndrome (e.g., obesity, fatigue, muscle weakness, hypertension, and hyperglycemia), whereas increased aldosterone (hyperaldosteronism) can lead to high blood pressure, extreme thirst, muscle spasms/weakness, and vascular inflammation. However, the persistence of VZV-induced IL-6 in vivo and its effects on these hormones remain to be determined.
Other cytokines (IL-12p70, IL-13, IL-4, or TNF-α) that may impact adrenal function were seen in supernatant of VZV-infected cells but not supernatant of mock-infected cells. Specifically, IL-4 was found only VZV-infected cells and is associated with increased ACTHstimulated cortisol release in bovine adrenal cortical cells, which may play an important role in coordinating adrenal response to inflammatory stressors [20]. The relationship between cortisol and TNF-α is less clear, with reports in rats that cortisol can potentiate TNF-α levels [21]. However, in humans it seems TNF-α may be similar to IL-6, which increases cortisol levels [17,22]. The exact effect IL-12p70 or IL-13 may have on adrenal function is less clear. Other herpesviruses, such as cytomegaloviruses (CMV), have been shown to have correlative changes in cytokine production and adrenal hormone changes during infection. Specifically, infection of mice with murine CMV has been shown to increase levels of IL-12, INF-γ, TNF-α, and IL-6. During the time of cytokine production, they also showed an increase in corticosterone and ACTH [23]. Interestingly, Ruzek and colleagues [23] did not find elevated IL-1β during CMV infection, consistent with what we have found in human adrenal cortical cells ( Figure 3A, Table 1). Significantly increased IL-8 was seen in supernatant from VZV-infected cells compared to mock-infected cells; to our knowledge, there are no reports of IL-8 directly affecting secretion of adrenal hormones. However, the resultant increase in neutrophil infiltration into adrenal tissue could impact function indirectly.
The absence of VZV-induced adrenal cortical cell death in our study is surprising given that many non-neuronal cells undergo apoptosis during VZV infection [11,24]. Only neurons have been consistently reported to be resistant to VZV-induced cell death [25]; rare reports of corneal epithelial cells and spinal astrocytes resistant to VZV-induced cell death have also been observed [9,16]. The absence of cell death in VZV-infected adrenal cortical cells could be attributed to the time of analysis (longer infection times may lead to cell death) or absence of immune cells that may infiltrate adrenal glands and induce cell death in vivo. Indeed, our histopathological examination of SVV-infected adrenal glands show immune infiltrates, hemorrhage, and cellular necrosis [6]. Alternatively, adrenal cortical cells may be resistant to VZV-induced cell death, implying that they may be a potential reservoir for persistent viral infection. This is particularly interesting given that VZV is latent in 6% of human adrenal glands [1] and can reactivate to potentially produce a persistent infection. Indeed, the adrenal gland has been known to be a reservoir for other pathogens, most notably Mycobacterium tuberculosis [26]. Whether VZV or other viruses do the same remains to be explored in future studies.
Chronic inflammation of the adrenal gland (adrenalitis) can alter production of glucocorticoids and mineralocorticoids, thereby disrupting homeostasis and producing disease such as Addison's disease, characterized by fatigue, weight loss, gastrointestinal problems, salt cravings, arthralgias and myalgias. Adrenalitis has been primarily attributed to an autoimmune etiology, yet pathogens have also been reported to affect adrenal glands, including HIV, coronaviruses, enteroviruses, and human herpesviruses (EBV, CMV, and HSV-1) [27][28][29][30]. More recently, SARS-CoV-2 has been proposed to interfere with the HPAaxis, leading to "long haul symptoms" reminiscent of chronic fatigue syndrome (myalgic encephalomyelitis) [31,32]. Critically ill COVID-19 patients have elevated cortisol levels [33,34] and postmortem analysis of COVID-19 patients show adrenal gland infection with virus [35]. In one case report, a 51-year-old man with COVID-19 was diagnosed with adrenal insufficiency following confirmed SARS-CoV-2 infection [36]. These observations suggest that long haul symptoms could be attributed in part to SARS-CoV-2 or other viral infections of adrenal glands, leading to increased cortisol levels with disruption of the HPA axis.
Myalgic encephalomyelitis, similar to COVID-19 long haul symptoms, have been reported following VZV infection. An epidemiological study found that the incidence of chronic fatigue syndrome was higher in zoster versus non-zoster patients [37]. Another study found patients with myalgic encephalomyelitis and Gulf War syndrome had significantly elevated antibodies towards many human herpesviruses [38,39]. Given that VZV infection has been known to manifest with many neurological symptoms, with or without rash [40], VZV should be considered in the differential diagnosis of idiopathic adrenal dysfunction/adrenalitis, including disorders consistent with myalgic encephalomyelitis.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.