Coinfection of Dermal Fibroblasts by Human Cytomegalovirus and Human Herpesvirus 6 Can Boost the Expression of Fibrosis-Associated MicroRNAs

Tissue fibrosis can affect every type of tissue or organ, often leading to organ malfunction; however, the mechanisms involved in this process are not yet clarified. A role has been hypothesized for Human Cytomegalovirus (HCMV) and Human Herpesvirus 6 (HHV-6) infections as triggers of systemic sclerosis (SSc), a severe autoimmune disease causing progressive tissue fibrosis, since both viruses and antiviral immune responses toward them have been detected in patients. Moreover, HCMV or HHV-6A infection was reported to increase the expression of fibrosis-associated transcriptional factors and miRNAs in human dermal fibroblasts. However, it is unlikely that they have separate effects in the infected host, as both viruses are highly prevalent in the human population. Thus, our study aimed to investigate, by quantitative real-time PCR microarray, the impact of HCMV/HHV-6A coinfection on the expression of pro-fibrotic miRNAs in coinfected cells, compared to the effect of single viruses. The results showed a possible synergistic effect of the two viruses on pro-fibrotic miRNA expression, thus suggesting that HCMV and HHV-6 may enhance each other and cooperate at inducing enhanced miRNA-driven fibrosis. These data may also suggest a possible use of virus-induced miRNAs as novel diagnostic or prognostic biomarkers for SSc and its clinical treatment.


Introduction
Tissue fibrosis is a pathological feature characterized by the massive accumulation of extracellular matrix (ECM), which can occur in nearly every tissue of the body. It occurs in most chronic inflammatory diseases, and when progressive, the process can lead to scarring, organ malfunction, and death, as seen in several end-stage diseases affecting the kidneys, lungs, and heart. Fibrosis can also influence tumor invasion and metastasis, as well as many chronic autoimmune diseases, including scleroderma or systemic sclerosis (SSc) [1]. Although they have a remarkable impact on morbidity and mortality, the mechanisms of the fibrogenesis process are not yet clarified, and specific therapeutic options targeted to fibrosis pathogenesis are extremely scarce, if not absent.
Among fibrosis-related autoimmune diseases, SSc has a prevalence of 7.2-33.9 per 100,000 individuals in Western countries and affects 2-3 times more females than males [2]. The disease is characterized by complex manifestations, but the pathology always includes three main features: the presence of both humoral and cellular immune system alterations, a severe vasculopathy, and an over-production and deposition of ECM by altered fibroblasts, which are the main target cells involved in the disease, together with endothelial cells (ECs), leading to pronounced tissue and multiorgan fibrosis [3][4][5][6].
To date, the etiological agents of SSc are still unclear, but it is widely accepted that the disease occurs as a result of a multistep and multifactorial process, involving genetic predisposition of the host and environmental factors, the latter including both physicochemical agents and infectious agents [6,7]. Among environmental factors, viral infections have been consistently suggested as possible etiological factors, mainly those which persist in the host and can reactivate in susceptible subjects, such as Human Cytomegalovirus (HCMV) and Human Herpesvirus 6 (HHV-6) [8][9][10][11][12][13][14]. Both viruses belong to the Herpesviridae family, Beta-herpesvirinae subfamily, and are ubiquitous and highly prevalent in the human population. Primary infection occurs in early childhood, and afterwards, they establish a latent infection in the host, reactivating mostly asymptomatically in the healthy adult, whereas symptomatic reactivations happen in immune-dysregulated hosts, where they have been associated with several autoimmune diseases, including those involving the connective tissue [11,12,[15][16][17][18].
An important body of evidence supports the role of HCMV and HHV-6 in SSc etiopathogenesis: both viruses have been found reactivated in the skin of SSc subjects (the HHV-6A species, in particular, has the more marked tissue tropism) [10,19], who also were characterized by a strong activation of antiviral humoral immunity, showing significantly higher amounts of antibodies directed toward viral antigens, such as UL94 (HCMV) and U94 (HHV-6), compared to healthy controls [10,[20][21][22][23]. In addition, a molecular mimicry has been observed with anti-UL94 autoantibodies recognizing fibroblasts and EC membrane receptors, resulting in fibroblasts' activation and the stimulation of EC apoptosis [20,24]. Regarding the cellular immunity, the presence of specific HCMV antigendriven CD8+ T cells has been observed in SSc patients [9]; an impaired NK response against HHV-6A/B was observed in a subset of SSc patients expressing KIR2DL2 receptor [10].
Both viruses can establish a productive infection in fibroblasts and endothelial cells [8,13,14,[25][26][27][28]. The in vitro infection of human primary dermal fibroblasts with HCMV and HHV6 has been shown to have a strong impact on the expression of profibrotic factors [8]. Interestingly, we recently observed that the simultaneous presence of HCMV and HHV-6 determines a more prominent and sustained expression of transcriptional factors associated with fibrosis and apoptosis pathways, compared to that observed in singly infected cells, highlighting for the first time the potential cooperation of these beta-herpesviruses in sclerodermic disease [14].
Indeed, fibrosis can affect every type of tissue or organ, possible leading to organ malfunction and failure; however, the factors modulating and influencing this process are still unclarified. In addition, the hypothesized role of infectious agents and other possible environmental triggers (such as mechanical injury, radiation, and toxic compounds) is not fully elucidated, and there is a lack of validated molecules and markers potentially useful in the diagnosis, prognosis, and therapeutic approach for this group of diseases.
In order to understand the molecular pathways involved in SSc pathogenesis, recent attention has been given to the role of epigenetic factors, and in particular to mi-croRNAs (miRNAs), short RNA sequences 20-23 nucleotides in length that are recognized to have a fundamental importance in the regulation of gene expression at the posttranscriptional level [29,30]. Interestingly, several miRNAs associated with vasculopathy and fibro-proliferative alteration have also been found to be dysregulated in SSc patients compared to controls [31]. In parallel, β-herpesvirus infection has been reported to alter miRNA expression in different tissues and cellular types [32,33], and we recently found that the individual infection with HCMV or HHV-6A could induce a remarkable modulation of miRNAs expressed in human primary dermal fibroblast cells [13]. Based on these considerations, here we aimed to investigate the impact of coinfection by HCMV and HHV-6 on the modulation of fibrosis-associated miRNAs in order to evidence any enhanced effect potentially involved in SSc onset and/or progression. For this purpose, primary human dermal fibroblasts were infected in vitro with individual HCMV or HHV-6A inocula or coinfected simultaneously with HCMV and HHV-6A, and miRNAs levels were evaluated in control uninfected, singly infected, and dually infected cells.
The TB40E HCMV strain (a kind gift from Prof. Thomas Mertens, Ulm University, Ulm, Germany) was propagated in MRC5 fibroblast cells (ECACC 05072101, Merck Life Science, Milan, Italy). Briefly, for HCMV titration, ten-fold serial dilutions of HCMV suspensions were used to infect MRC5 fibroblast monolayers grown in 60 mm Petri dishes, performing the titration in triplicate for each dilution. After adsorption (37 • C for 2 h), virus inocula were removed and replaced with fresh Earle's modified Minimum Essential Medium with 1% l-glutamine, 1% non-essential amino acids, 10% fetal calf serum, supplemented with 0.6% agarose (Merck KGaA, Darmstadt, Germany). Plates were then incubated at 37 • C for 7 days, then cell monolayers were stained with the vital dye neutral red for 2 h, then the medium was discarded and cells were fixed with 10% formalin for 10 min at room temperature. Plaques were counted and titers expressed as mean PFU values/mL. HCMV virus stock contained 10 9 PFU/mL.
The U1102 strain of HHV-6A was grown in human J-Jhan T cells. The 6A species of HHV-6 was utilized based on our previous findings showing its presence in the skin of SSc patients, supporting its higher tissue tropism compared to the HHV-6B species [10]. Virus titers were assessed as previously reported [13,35]. The HHV-6A U1102 strain was directly titrated by real-time quantitative PCR (qPCR) targeted to the viral U94 gene after propagation in human lymphoid J-Jhan T cells, as previously described. Briefly, total DNA was extracted from J-Jhan cells at complete cytopathic effect and used to quantify the HHV-6A virus genome number as previously described [13]. HHV-6A virus stock contained 10 10 genome equivalents/mL. The same viral stocks of HCMV and HHV-6A were used for all infection experiments.

Virus Infection
Cell infection experiments were carried out in primary human dermal fibroblasts by using HCMV (T40E) and HHV-6A (U1102) strains. Infections and coinfections were performed in 90% confluent primary human dermal fibroblasts, using a multiplicity of infection (M.O.I.) corresponding to 0.1 PFU/cell for HCMV and 1.0 genome equivalent/cell for HHV-6A, as previously described [8,13]. Virus adsorption was performed for 2 h at 37 • C, then virus inocula were removed and replaced with complete fibroblast cell medium. Cells were then incubated at 37 • C and collected at 0, 1, 2, 4, and 7 days post-infection (d.p.i.). Briefly, cells were centrifuged at 1000× g for 5 min at 4 • C, then cell pellets were washed in PBS to remove any presence of extracellular miRNAs and immediately frozen in liquid nitrogen. Pelletized samples were kept at −80 • C until use. Duplicate samples were analyzed for each experiment.

Nucleic Acid Extraction
Total nucleic acids (TNAs) were extracted from aliquots of 10 6 infected cells by using the AllPrep DNA/RNA/miRNA kit (Qiagen, Hilden, Germany), which allows the simultaneous isolation of total DNA and RNA, including the miRNA fraction (<200 nucleotides). Extracted TNAs were quantified by spectrophotometric analysis reading at 260/280 nm by using a Nanodrop instrument and kept at −80 • C until analysis. The total extracted RNA, including the miRNA fraction, underwent DNase I treatment (Thermo Fisher Scientific, Waltham, MA, USA) in order to eliminate any DNA contamination. The absence of DNA contamination was checked by PCR amplification of human β-actin gene using 10 ng of extracted RNA as a template.

Virus Quantitation in Infected Cells
The virus presence in infected cells was assessed and quantified by specific quantitative real-time PCR (qPCR), using 100 ng of total extracted DNA as a template. Specifically, the HCMV DNA amount was quantified by the qPCR CMV ELITe MGB ® Kit (ELITechGroup, Turin, Italy), designed to detect the HCMV DNA exon 4 region of the immediate-early (IE)1 gene, following the manufacturer's instructions, as already described [8,14]. HHV-6A presence was assessed by a specific qPCR targeting the U94 viral gene, as previously described [17,35]. In addition, the detection and quantification of the house-keeping human RNase P gene was carried out as a control of nucleic acids' quality and normalization of viral genome copies per cell number. Both qPCR assays were performed in a 7500 Real-time PCR system (ABI PRISM, Applied BioSystems, Milan, Italy), and results were expressed as number of genome copies per µg of DNA.

miRNA Analysis
The analysis of the expression of miRNA in control uninfected, single-infected, and double-infected fibroblasts was carried out by qPCR microarray. In short, 10 ng aliquots of RNA were retrotranscribed by the miRCURY LNA miRNA RT Kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. Then, the cDNA products were analyzed by the LNA miRNA Focus PCR Panel (Qiagen, Hilden, Germany), which allows simultaneous identification and quantification of 84 human microRNAs associated with cell fibrosis. Each panel included six controls of the process (RNA spikein assays controls, interplate calibrators) and four miRNA reference genes (SNORD44, SNORD38B, SNORD49A, U6 snRNA) to allow data normalization. Each microarray was run on a Quant Studio 5 real-time PCR system (Thermo Fisher Scientific, Milan, Italy), and miRNA data were analyzed with the free Qiagen software (Qiagen Gene Globe, https://geneglobe.qiagen.com/ca/analyze (accessed on 6 January 2023)). Results were expressed as fold-change expression value of infected versus uninfected control cells at each time post-infection, after normalizing for the house-keeping endogenous controls. The analysis threshold was set at 2-fold up-or down-modulation compared to control uninfected cells. Duplicate samples from two independent experiments were analyzed.
Ten miRNAs, among those most up-or down-regulated by coinfection, as judged by microarray analysis, were further analyzed by individual miRCURY LNA miRNA PCR Assays, able to quantify specific miRNAs with high sensitivity using LNA-optimized, SYBR ® Green-based miRNA PCR (Qiagen, Hilden Germany). The constitutively expressed cellular miR-RTC and miR-SNORD11 were also included in the analysis as controls.

Statistical Analyses
Paired t-test was used to analyze the significance of differential miRNA levels between infected and control cells. Bonferroni correction for multiple comparisons was applied. A p value ≤ 0.05 was considered as statistically significant.

HCMV and HHV-6A Coinfection in Primary Human Dermal Fibroblasts
First, the effect of simultaneous coinfection of HCMV and HHV-6A was analyzed with respect to the ability of viruses to replicate together in human primary dermal fibroblasts. For this purpose, cells were seeded at optimal density in 25 cm 2 flasks three days before infection, and then they were infected with HCMV (TB40E strain) and HHV-6A (U1102 strain) at a M.O.I. of 0.1 PFU per cell and 1 genome equivalent per cell, respectively. Cell samples from single-infected, double-inflected, and control uninfected fibroblasts were collected at 0, 1, 2, 4, and 7 d.p.i. At each timepoint, total DNA was extracted from collected cell samples and analyzed by specific qPCR targeting HCMV IE-1 and HHV-6 U94 genes, respectively, in order to quantify the amount of the intracellular virus genomes. The results, summarized in Figure 1, confirmed that the primary human dermal fibroblasts used in the experiments were permissive for both HCMV and HHV-6A productive replication, as also previously reported [14], based on the clear increase in viral genome copies inside infected cells over time.
the simultaneous presence of both viruses could enhance their respective replication was paralleled by an earlier cytopathic effect (CPE), confirming what was prev reported [14]. In fact, in individually infected fibroblasts, HCMV replication produ lytic infection with an evident CPE appreciable starting from 4 d.p.i. until the end experiment. On the other hand, as also previously reported, HHV-6A did not induc evident CPE, and likely established a latent infection at 7 d.p.i, as judged by the l further increase in virus DNA at this timepoint. Despite the lack of CPE inducti HHV-6A, the simultaneous presence of HCMV induced an earlier HCMV-associated onset, with fibroblasts showing clear morphological alterations as early as 2 d.p. shown). Notably, the genome copy number appeared increased by about 1 Log in coinfected cells with respect to the amount of virus detected in single-infected cells, suggesting that the simultaneous presence of both viruses could enhance their respective replication. This was paralleled by an earlier cytopathic effect (CPE), confirming what was previously reported [14]. In fact, in individually infected fibroblasts, HCMV replication produced a lytic infection with an evident CPE appreciable starting from 4 d.p.i. until the end of the experiment. On the other hand, as also previously reported, HHV-6A did not induce any evident CPE, and likely established a latent infection at 7 d.p.i, as judged by the lack of further increase in virus DNA at this timepoint. Despite the lack of CPE induction by HHV-6A, the simultaneous presence of HCMV induced an earlier HCMV-associated CPE onset, with fibroblasts showing clear morphological alterations as early as 2 d.p.i. (not shown).

Effect of HCMV and HHV-6A Coinfection on the Expression of Fibrosis-Associated miRNAs
The modulation of microRNA expression in control uninfected, individually infected, and coinfected fibroblast cells was analyzed by a specific qPCR microarray able to identify and quantify simultaneously a panel of 84 fibrosis-associated miRNAs. The results (  whereas the other miRNAs were expressed between −2.14 and −9.83-fold with respect to controls. Of note, significant differences were observed by comparing the levels of miRNA expression induced by single and double infection in fibroblast cells (Table S1).
Briefly, three different conditions were observable. For miRNAs that were up-regulated or down-regulated by both individual infections and coinfection, the level of alteration was invariably higher in coinfected cells compared to single-infected cells, supporting the enhancing joint effect of coinfection compared to single infection on miRNA expression. An example of this was miR-1 at 7 d.p.i., which was much more up-regulated by coinfection than by individual viruses. Similarly, the following miRNAs were more induced by coinfection compared to single infections at the indicated times p. In a similar way, miRNAs that were down-regulated by both single and double infection were expressed at a lower level in coinfected cells compared to single-infected cells. These included miR-141 at 2 d.p.i., miR-211 at 7 d.p.i., miR-338 at 1 d.p.i., and miR-5011 at 0, 1, and 2 d.p.i.
Several miRNAs exhibited a biphasic behavior, with different alterations (increase or decrease) at early and late times p.i., and the effect of coinfection could sometimes be opposite to that observed with one or both the infecting viruses individually used. This condition was observed at various timepoints p.i. for the following miRNAs: miR-1, miR-129, miR-142, miR-18a, miR-200a, miR-203a, miR-211, miR-223, miR-449a, miR-5011, and miR-661.
Among the most up-and down-regulated miRNAs in coinfected cells, 10 miRNAs were further analyzed by individual assay to validate and verify the alterations observed by microarray analysis. Specifically, miR-1, miR-122, miR-19b, miR-200b, miR-32, miR-328, miR-375, and miR-155 were tested by individual assay to confirm the up-regulation and miR-200a and miR-223 for down-regulation by coinfection, as observed by microarray analysis. Two control constitutively expressed miRNAs were also included in the analysis as controls (miR-RTC and miR-SNORD11). The results, after normalization for the housekeeping miRNAs, confirmed the data obtained by microarray, indeed showing in most cases superimposable or more significant levels of variations compared to that observed by microarray qPCR (Figure 3). 129, miR-142, miR-18a, miR-200a, miR-203a, miR-211, miR-223, miR-449a, miR-5011, and miR-661. Among the most up-and down-regulated miRNAs in coinfected cells, 10 miRNAs were further analyzed by individual assay to validate and verify the alterations observed by microarray analysis. Specifically, miR-1, miR-122, miR-19b, miR-200b, miR-32, miR-328, miR-375, and miR-155 were tested by individual assay to confirm the up-regulation and miR-200a and miR-223 for down-regulation by coinfection, as observed by microarray analysis. Two control constitutively expressed miRNAs were also included in the analysis as controls (miR-RTC and miR-SNORD11). The results, after normalization for the housekeeping miRNAs, confirmed the data obtained by microarray, indeed showing in most cases superimposable or more significant levels of variations compared to that observed by microarray qPCR (Figure 3).

Discussion
The beta-herpesviruses HCMV and HHV-6 have been repeatedly hypothesized to have a role in the onset and/or progression of SSc, based on virological and immunological

Discussion
The beta-herpesviruses HCMV and HHV-6 have been repeatedly hypothesized to have a role in the onset and/or progression of SSc, based on virological and immunological clues. Both viruses are highly prevalent and ubiquitous in the human population, and their reactivation in a susceptible host has been associated with the development of several symptomatic diseases, including the onset of diverse autoimmune diseases, where their concurrent reactivation has been invariably correlated with a worse clinical outcome [36][37][38][39].
Indeed, individual infection by HCMV or HHV-6A in human dermal fibroblasts, which represent the main target cells of SSc disease, has been shown to produce a remarkable impact on the expression of cell factors involved in the progression toward tissue fibrosis [8]. Last, the simultaneous presence of these viruses in coinfected fibroblast cells has been recently reported to induce a higher, pronounced, and sustained expression of cell factors correlated with fibrosis and apoptosis processes compared with single-infected cells [14].
With regard to cell fibrosis, several miRNAs have been reported to be involved in key pathways linked to fibrotic alterations; they have been detected to be significantly altered in tissues and blood of SSc patients, and their expression was similarly found to be modulated in human fibroblasts infected with HCMV or HHV-6 [13,40].
Based on the observation that it is unlikely that these viruses infect or reactivate separately in the host and that currently there are no data on their joint effect on miRNA expression in coinfected cells, our study aimed to investigate the modulation of miRNA expression following coinfection in primary human dermal fibroblasts, comparing it with control uninfected cells and individually infected cells.
The intracellular miRNA profiling of 84 fibrosis-related miRNAs by microarray evidenced a huge enhancing effect on miRNA expression in HCMV/HHV-6A coinfection compared to what was observed in single-infected cells. This was likely correlated to the reciprocal boost of virus replication associated with the simultaneous presence of both viruses, as demonstrated by the 1-2 Log increase in viral genome copy number in coinfected cells compared to what was detected in single-infected cells, despite the equal amount of viral inocula used. This highlights that HCMV and HHV-6 can enhance each other's replication and supports previous reported data on the observed synergism of beta-herpesviruses in the induction of serious clinical manifestation and worse patient outcome [38,41,42].
Consistently, the co-presence of HCMV and HHV-6A resulted in a strengthened effect on miRNA expression. In fact, a higher level of dysregulation of miRNA expression was detected compared to single-infected cells for several fibrosis-associated miRNAs. Specifically, the pro-fibrotic miR-1, miR-19b; miR-122, miR-129, miR-150, miR-155, miR-192, miR-200b, miR-215, miR-216a, miR-32, miR-328, miR-375, miR-449b, miR-491, miR-661, and miR-7 were all more induced than in single-infected cells, showing values suggesting true synergism between the viruses rather than a mere additional effect. The pro-fibrotic action of the up-regulated miRNAs was well documented in several reports; miR-1 was reported as involved in increased fibrosis of the cartilage in acetabular dysplasia [43] and in liver fibrosis [44]; miR-19b-3p was associated with hypertrophic and fibrosis indexes in acute heart failure [45]; miR-122 has been related with fibrosis-induced cardiovascular remodeling [46] and, in this regard, it is intriguing that both HCMV and HHV-6A have been associated with cardiovascular diseases [25,[47][48][49]; miR-150 was correlated with renal fibrosis and miR-150 antagonists can ameliorate the pro-fibrotic pathway in a mouse model [50]; miR-155 is essential in fibrosis and it is consistently up-regulated in fibrotic disorders [51]; miR-192 can promote fibrosis by transforming growth factor beta (TGF-β) activation, which is recognized as a major mediator of fibrosis [52], and its circulating levels are increased in patients with hypertrophic cardiomyopathy and diffuse myocardial fibrosis [53]; miR-200b was found to be up-regulated in fibrotic liver samples compared to non-fibrotic ones [54]; miR-216a accelerates fibrogenesis in cardiac fibroblasts [55]; miR-32 has been reported to mediate the glucose-induced hepatic fibrosis [56]); miR-328 was found to be up-regulated in cardiac fibrosis and shown to directly stimulate TGF-β1 signaling, promoting collagen production in cultured fibroblasts [57], although its pro-fibrotic role it is not so clear, since it has been also reported to prevent renal fibrogenesis [58]; miR-375 was recognized to promote cardiac fibrogenesis by accelerating the ferroptosis of cardiomyocytes through mediating glutathione peroxidase 4 (GPX4) [59]; the miR-449 family was detected to be up-regulated in cystic fibrosis [60] and in bleomycin-induced lung fibrosis [61] and able to activate TGF-β1 in nasopharyngeal carcinoma [62]; miR-491 is induced by TGF-β1 during renal fibrosis [63]; miR-661 has been recently shown to accelerate fibrosis by increasing fibroblast growth factor 2 (FGF2) [64]; miR-7, besides its recognized role as a tumor suppressor in the liver, has been recently reported to promote Fibroblast growth factor receptor 4 (FGFR4) activation and associated liver fibrosis [65].
Consistent with the hypothesized pro-fibrotic action of HCMV/HHV-6A coinfection, it concomitantly decreased the expression of miRNAs with putative or recognized antifibrotic effect, such as miR-145, miR-18a, miR-194, miR-200a, miR-223, miR-338, and miR-449a. Other antifibrotic miRNAs were instead variably regulated or indeed induced by HCMV/HHV-6A coinfection. For example, miR-215 was down-regulated by coinfection at earlier times p.i. but up-regulated at later times, and published data suggest an anti-fibrotic action, showing that its dampening correlates with increased fibroblastic production of matrix in ocular pterygium [55]. Similarly, miR-133a, up-regulated by coinfection, despite its induction by TGF-β1, was identified as an anti-fibrotic factor functioning as a feedback negative regulator of TGF-β1 pro-fibrogenic pathways [66]. In addition, miR-141 was reported to reduce cardiac fibrosis and improve cardiac function [67]. The results were also in agreement with previous published results obtained in individually infected cells by a microRNA two-card array set able to identify and quantify simultaneously a total of 754 miRNA by qPCR TaqMan assays [13]. Most of the differentially expressed miRNAs in the present study were in fact also identified in the previous analysis, among those up-regulated or down-regulated by HCMV or HHV-6A in in vitro infected primary human dermal fibroblasts. These included miR-1, miR-19b, miR-122, miR-192, miR-200b, miR-32, miR-449b, and miR-7, which were also among the most up-regulated by individual infections [13]. By contrast, miR-150, miR-155, miR-215, miR-216a, miR-328, miR-375, miR-491, and miR-661 appeared noticeably more induced only by coinfection, highlighting the synergistic effects related to the simultaneous presence of both HCMV and HHV-6A, rather than a mere summation effect. Quantitative differences observed in single-infected cells in the two studies are likely due to the different methodologies used. However, despite the diverse analytical method employed, the results confirmed what was previously reported, underlining the reliability of the microarray results obtained here, as well as the significant differences that emerged between single-and double-infected fibroblasts in terms of the modulation of their microRNome.
Although it is difficult to draw conclusions on the impact of virus coinfection on a possible miRNA-mediated fibrosis, it is interesting to note that we previously demonstrated that HCMV/HHV-6A simultaneous infection increased TNFα and other fibrosis-associated factors and that miR-19b and miR-155, both induced by coinfection, were reported to trigger endothelial IL-6 and TNFα production [68]. In addition, miR-129 was reported to be involved in an IL-6/TNFα signaling axis in neuroinflammation [69]. miR-150 overexpression promotes cell apoptosis, inhibited cell proliferation, and increased secretion of pro-inflammatory cytokines such as IFNγ and TNFα [70]. miR-155 may play an important role in the pathogenesis of fibrosis by pro-inflammatory factors IL-1β and TNFα [71], by which it is also induced [72]. MiR-32 could enhance cell proliferation and epithelial-tomesenchymal transition via an miR-32/CXCR4 axis [73], and CXCR4 was found to be up-regulated by HCMV/HHV-6A coinfection. On the other hand, miR-200a, found to be down-regulated by virus coinfection, was reported to suppress epithelial-to-mesenchymal transition via BMP7 [74], another factor significantly affected by virus coinfection [14]. Similarly, miR-223 is considered a regulator of inflammation via CCL3 and CXCR4 regulation [75,76], which were instead detected to be overexpressed by us in coinfected cells [14], consistent with miR-223 down-regulation. Other signaling pathways appear less clear and somehow contradictory, such as for example those related to the induction of miR-1 and miR-19b, which have been both reported to decrease the expression of CXCR4 [77,78], rather than enhancing its production. Overall, functional studies focused on the most important factors and miRNAs impacted by HCMV/HHV-6A coinfection would be needed to assess which pathways are responsible for the hypothesized virus-induced pro-fibrotic effect.

Conclusions
Tissue fibrosis is a complex and multistep process, involving several cellular factors and signaling. Among them, miRNA expression has been recognized as a crucial mechanism underlying the development of fibrosis. Our data show that HCMV/HHV-6A coinfection, which is a highly probable condition in vivo, can boost the expression of miRNAs associated with fibrosis compared to individual viruses [13], strengthening the hypothesis of a significant role of the simultaneous presence of these two beta-herpesviruses in fibrosis-associated diseases, such as SSc. Of note, only some of the induced pro-fibrotic miRNAs have been already recognized to be altered in SSc patients' tissues or blood (such as miR-133, miR-145, and miR-223), while several others have been associated with other fibrosis-associated diseases, extending the potential putative involvement of HCMV and HHV-6A virus coinfection in more fibrosis-related pathologies. Indeed, the data highlighted a high induction of miRNAs involved in cardiovascular diseases, supporting the hypothesized role of these viruses in the pathogenesis of cardiac and vascular troubles, and suggesting their role as possible triggers of fibrosis for these specific tissues. Of note, miRNA deregulation also was consistent with previously reported data on the expression of cellular factors associated with the fibrosis process. In conclusion, these data show for the first time that coinfection by HCMV and HHV-6A can result in high and sustained dysregulation of miRNA expression in coinfected cells, thus potentially opening the way to considering the use of these virus-induced miRNAs as novel diagnostic or prognostic biomarkers for SSc, as well as for other fibrosis-related diseases, with potential usefulness in the assessment of disease onset and progression to differentiate and foresee the diverse clinical outcomes.