Gut microbiota modulation induced by Zika virus infection in immunocompetent mice

Gut microbiota composition can modulate neuroendocrine function, inflammation, and cellular and immunological responses against different pathogens, including viruses. Zika virus (ZIKV) can infect adult immunocompetent individuals and trigger brain damage and antiviral responses. However, it is not known whether ZIKV infection could impact the gut microbiome from adult immunocompetent mice. Here, we investigated modifications induced by ZIKV infection in the gut microbiome of immunocompetent C57BL/6J mice. Adult C57BL/6J mice were infected with ZIKV and the gut microbiota composition was analyzed by next-generation sequencing of the V4 hypervariable region present in the bacterial 16S rDNA gene. Our data showed that ZIKV infection triggered a significant decrease in the bacteria belonging to Actinobacteria and Firmicutes phyla, and increased Deferribacteres and Spirochaetes phyla components compared to uninfected mice. Interestingly, ZIKV infection triggered a significant increase in the abundance of bacteria from the Spirochaetaceae family in the gut microbiota. Lastly, we demonstrated that modulation of microbiota induced by ZIKV infection may lead to intestinal epithelium damage and intense leukocyte recruitment to the intestinal mucosa. Taken together, our data demonstrate that ZIKV infection can impact the gut microbiota composition and colon tissue homeostasis in adult immunocompetent mice.

Zika Virus (ZIKV) is an Arbovirus member of the Flaviviridae family that is mainly transmitted by the bite of Aedes genus mosquitoes 1 . During the 2015 outbreak, ZIKV spread quickly in America, mainly in Brazil 2 . During ZIKV dissemination, non-vector born routes of infection were reported, including sexual transmission 3 . A great concern about ZIKV's impact worldwide was the association of post-infection disorders, such as Guillain-Barré syndrome 4 , and the development of congenital malformations 5 .
As a means of better understanding ZIKV pathogenesis, most of the molecular mechanisms associated with the infection were elucidated with the use of genetically-induced immunodeficient mice models 6 . In those reports, infected mice presented weight loss, high viremia, detectable signs of illness, intensive neuronal loss, immune system activation on neuronal surroundings, and severe testicular inflammation [6][7][8] . In contrast, the impact of ZIKV on immunocompetent adult mice models showed detectable viral loads in the serum, and effects on different organs or lethality after infection through different routes 9,10 . Despite this, ZIKV induces innate and adaptive immune responses that are essential for protecting the organism against the establishment of disease 11,12 . The immune system activation demonstrates that specific parameters are modulated during ZIKV infection such that, regardless of the milder symptoms, the virus still impacts on the host 13 . These findings lead to new questions regarding if other alterations could be linked to viral activation of the host's immune system and modulation of physiological functions not noticed before.
The gut microbiota has been described as a strong modulator of inflammatory and immune responses, both locally and systemically 14,15 , playing essential roles in triggering host responses against pathogen infections 16 . In recent years, studies have shown that the gut microbiota influences and is influenced by viral infections 17 . This study reports that enteric viruses can lead to substantial disturbances in gut microbiota composition, impacting ZIKV infection decreases the abundance of Actinobacteria and Firmicutes in the gut microbiota. The taxonomic analysis identified Actinobacteria, Bacteroidetes, Firmicutes, Proteobacteria, and Spirochaetes as the most abundant phyla in both ZIKV-infected and uninfected mice ( Fig. 2A). In the ZIKV-infected group, the abundance of Actinobacteria significantly decreased from 1.8 to 0.3%, (Fig. 2B), and Firmicutes significantly decreased from 41.8 to 18.2% (Fig. 2C) relative to the uninfected controls. In contrast, ZIKV infection increased the relative abundance of Deferribacteres from undetected to 0.5% (Fig. 2F) and Spirochaetes from 4.2 to 42.9% (Fig. 2G) when compared to uninfected mice. The relative abundances of Bacteroidetes (Fig. 2E) and Proteobacteria (Fig. 2D) were not modulated by ZIKV infection. Thus, ZIKV infection triggered a significant change in bacterial community composition in the immunocompetent mice at the phylum level.
ZIKV infection increases the abundance of the Deferribacteraceae and Spirochaetaceae families in the gut microbiota. At the family level, we also observed an important modulation caused by ZIKV infection in the gut microbiota of the immunocompetent mice (Fig. 3A). The heat map demonstrates the polarization of some family groups (Fig. 3B). The levels of the Coriobacteriaceae, a family within the Actinobacteria phylum, significantly decreased to 0.3% after ZIKV infection in comparison to the uninfected group (1.8%) (Fig. 3C). Similarly, the Enterobacteriaceae family, within the Proteobacteria phylum, significantly decreased in abundance approximately one hundred times in ZIKV-infected mice (0.1%) when compared to uninfected (10.6%) (Fig. 3D). The Helicobacteraceae family, also within the Proteobacteria phylum, seemed to be negatively modulated during ZIKV infection, but not significantly (Fig. 3E). The Peptostreptococcaceae family also decreased its abundance in the ZIKV-infected mice (undetectable, 0%) compared to the uninfected group (2.6%) (Fig. 3F). Also, other families within the Firmicutes phylum, such as the Clostridiaceae (Fig. 3G) and Lactobacillaceae (Fig. 3H), seem to have their abundance negatively impacted on the ZIKV-infected group. However, for these other families, no significant difference relative to the uninfected controls was found.
In contrast, the Deferribacteraceae, a family within the Deferribacteres phylum, increased in the ZIKVinfected group (0.5%) in comparison to the uninfected control group (undetectable) (Fig. 3I). Interestingly, the most drastic impact of ZIKV infection on the gut microbiome of infected immunocompetent mice was on the Spirochaetaceae components, a family within the Spirochaetes phylum, which increased about tenfold (42.9%) relative to the uninfected control group (4.2%; Fig. 3J). Our data did not show any significant modulation of the Bacteroidaceae family in the gut microbiome of the ZIKV-infected mice group (Fig. 3K). Other families were analyzed but were not impacted by ZIKV infection ( Table 1).

The modulation of gut microbiota induced by ZIKV infection induces leukocyte infiltration in the colon.
After confirming the changes in the gut microbiota composition induced by ZIKV infection, we asked whether ZIKV could be present in the serum and the gut of mice 14 days post-infection (dpi). To address this question, we performed colon tissue and serum RNA extraction of immunocompetent mice infected or not with ZIKV and performed qPCR. Our results did not demonstrate the presence of ZIKV either in the serum (Fig. 4A) and the colon 14 dpi (Fig. 4B).
Next, we investigated whether ZIKV-induced gut microbiota modulation could induce local inflammation in the intestine. To investigate this, we collected portions of the gut colon of the animals and analyzed both Scientific Reports | (2021) 11:1421 | https://doi.org/10.1038/s41598-020-80893-y www.nature.com/scientificreports/ pro-inflammatory and anti-inflammatory cytokine production in this tissue. However, no significant levels of IL-12, TNF-α, IFN-γ, IL-1β, IL-10, and IL-33 were observed between ZIKV-infected and uninfected mice, indicating that ZIKV-induced gut microbiota modulation did not modulate the secretion of these cytokines 14 dpi in immunocompetent mice (Fig. 4C). Although microbiota modulation induced by ZIKV did not induce any significant alteration in cytokine production in the colon, we detected significant leukocyte infiltration (Fig. 4E) and intestinal epithelium changes (Fig. 4F). HE-stained colon sections displayed moderate inflammatory cell infiltration, which can be seen in mucosa and submucosa of the gut colon, and mild epithelial changes as goblet cell loss, hyperplasia, and erosion on the surface epithelium in the ZIKV-infected mice when compared with uninfected animals (Fig. 4D).

Discussion
Herein, we showed that wild-type (WT) C57BL/6J male mice presented leukocyte colon infiltration, intestinal epithelial damage, and gut microbiome composition alteration 14 days after ZIKV infection. Although Thackray and colleagues displayed the key roles of intestinal microbiota in host immunity against flaviviruses and suggested intestinal bacterial depletion to be related to exacerbated infection severity 20 , no previous studies have explored the effects of flavivirus infection on the relative abundance of specific bacterial taxa in the mouse gut.   24 , possibly explaining the reported GI involvement in some ZIKV patients. In contrast, ZIKV is only capable of effectively replicating in mice that show impaired antiviral immune response 25 . This disrupted immunity favors ZIKV survival, which results in a pathology showing intestinal involvements including intestine inflammation 26 and bowel dilation 27 , alterations that indicate ZIKV as capable of inducing inflammatory processes in the mouse GI tract. Although immunocompetent animals do not show macroscopic intestinal complications, in this study we detected evidence of inflammatory responses in colon tissue, as we discovered the occurrence of epithelial damage and leukocyte infiltration in the colon of mice 14 dpi. The inflammatory activity has already been linked with gut microbial alterations, as Lupp and colleagues indicated that host-mediated inflammation triggered by infection agents can alter the colonic microbial community 28 . As pro-inflammatory bacteria and pathogens increase their abundance in the gut 29 , increased gut permeability, immune dysfunction, and intestinal epithelial cells damaging are observed 30 . As further detailed here, the communities we discovered to increase in ZIKV-infected mice are related to detrimental health phenotypes. Moreover, bacterial taxa commonly associated with GI tract homeostasis seem to be negatively impacted by the inflammatory responses during the acute phase of infection 29 .

ZIKV-infected human colon cells connected this virus with intestinal inflammatory abnormalities and colitis
In the present work, we showed that ZIKV infection drastically diminished the abundance of Firmicutes members in the gut microbiome of WT mice. Among the most represented communities in the gut, Firmicutes are the main producers of butyrate 31,32 . This short-chain fatty acid (SCFA) shows anti-inflammatory properties as it impairs leukocyte migration 33 , diapedeses 34 , and enhances intestinal barrier integrity by facilitating tight junction assembly 35 and influencing mucus production 36 .
Another taxon that dramatically decreased in ZIKV-infected animals is the phylum Actinobacteria. Although less abundant than Firmicutes, Actinobacteria are one of the four major phyla in the intestinal microbiome. They also play key roles in maintaining gut homeostasis through the secretion of SCFAs that protect against enteropathogenic infection 37 , and contribute to intestinal barrier integrity 38 . The fact that members of this phylum are currently used as probiotics highlights their putative beneficial roles for the maintenance of intestinal health 39 .  www.nature.com/scientificreports/ In contrast, the Deferribacteres and Spirochaetes phyla significantly increased in infected animals, are associated with detrimental pathology in animals 40,41 . Experiments using dextran sodium sulfate (DSS)-treated mice, a colitis animal model, showed that these animals display increased abundance of taxa within the Deferribacteres in the gut compared to untreated animals, strongly associating this phylum with intestinal inflammation 42 . The most severe mouse colitis model, DSS is widely used due to its simplicity and resemblance to human ulcerative colitis (UC) 43 . Although the mechanisms through which DSS affects intestinal homeostasis are not yet fully described, this chemical colitogenic is believed to induce colitis after damaging the epithelial monolayer in the large intestine 44 .
Within the Deferribacteres phylum, Mucispirillum sp. (Table 2), enriched in ZIKV-infected mice, is considered an indicator phylotype for DSS treatment 42 . Bacteria from this genus express secretion systems and secrete proteins that modulate intestinal mucosa gene expression, including inflammatory processes 40 . Specifically, the pathobiont M. schaedleri is a potent oxygen scavenger, which may enable this species to survive and proliferate in harsh conditions such as inflammatory environments 45 . Also increased in the infected group, the Spirochaetes phylum displayed an increased abundance in the gut when compared with other phyla. As several taxa within the Spirochaetaceae family and Treponema genus have been identified as disease-causing 46 and that an increased presence of Treponema in the stool is generally classified as unhealthy 41 , we believe this increase may be relevant for deleterious GI effects during ZIKV infection.
We hypothesize that ZIKV infection of WT C57BL/6 mice triggers intestinal inflammation during an early phase of the infection that leads to gut dysbiosis, which is characterized by the outgrowth of pathobionts and disruption of beneficial bacterial communities 47 . As we show data that indicate that ZIKV infection favors the outgrowth of bacterial phyla associated with deleterious pathology and drastically reduces bacterial phyla implicated in intestinal barrier integrity and gut homeostasis, we suggest that these alterations may have influenced the occurrence of leukocyte colon infiltration and tissue damage detected by our group.
In this study we did not detect any modulation of colon secreted cytokines; however, microbial communities may influence the secretion of other mediators, such as IL-8 and MCP-1 31,32 . Also, quantification of a broader set of cytokines during the acute phase of ZIKV infection tends to confirm the relevance of the intestinal microbial dysbiosis in this infection. Once immunocompromised mice develop macroscopic intestinal complications, gut microbial analyses of these animals should corroborate the relevance of our gut microbiome data.
Once microbiome composition and function impacts on a plethora of other pathologies 48 , future research should more thoroughly determine the influence of ZIKV infection on this community, as through using female mice or immunodeficient mice, alternative infection routes as subcutaneous or intraperitoneal, other ZIKV strains, and other times of infection. Besides, further experimental efforts should assess GI function as a means of linking microbiome alterations with intestinal involvement, as White and others implicated flavivirus infection with intestinal dysmotility syndromes 27 . Moreover, as oral antibiotic administration was shown to augment mice susceptibility to severe flavivirus infection through interfering with antiviral T cell responses 20 and possibly exacerbating inflammatory processes 49,50 , employing specific antibiotic treatments must be useful for dissecting the connections between ZIKV infection, dysbiosis, intestinal involvement, and immunity. Uncovering the influence of ZIKV infection on gut microbial communities may prove valuable to determine therapeutic targets for ZIKV infection, a pathology intimately related to the life-threatening conditions of congenital microcephaly 51 and Guillain-Barré syndrome (GBS) 52 .
In conclusion, our study shows for the first time the modulation of the gut microbiota composition by ZIKV in immunocompetent mice (Fig. 5). We show that ZIKV infection decreases the abundance of bacterial families important for the maintenance of gut permeability and gut homeostases, such as Clostridiaceae, Enterobacteriaceae, and Coriobacteriaceae. Besides, ZIKV infection increases the abundance of bacterial families that associate with inflammation, such as Deferribacteraceae, and allows for the emergence of members of pathogenic genera, such as Treponema sp. Moreover, even though microbiota modulation induced by ZIKV infection did not induce any significant alteration in cytokine production, we observed that microbiota modulation triggered notable leukocyte infiltration and intestinal epithelium damage in the colon of immunocompetent mice infected by ZIKV.

Viral load by quantitative RT-PCR.
For RNA extraction, 20 mg of colon tissue from infected and uninfected mice was collected and carefully washed with saline. The viral RNA was extracted using the RNeasy Mini Kit (QIAGEN). For better access to the RNA and avoiding inappropriate tissue lysis, TRIzol and chloroform were used before the RNeasy Mini Kit and the top solution phase was driven to the next viral RNA isolation step, with the above-mentioned kit. After RNA isolation and purification following the manufacture's instruction, viral load was accessed by one-step quantitative reverse transcriptase PCR (RT-q PCR) as previously described 53 . A published primer set was used to detect ZIKV RNA 57 : Fwd, 5′-CCG CTG CCC AAC ACAAG-3′; Rev, 5′-CCA CTA ACG TTC TTT TGC AGA CAT -3′; Probe, 5′-FAM/AGC CTA CCT TGA CAA GCA GTC AGA CAC TCA A/3IABkFQ/-3′ (Integrated DNA Technologies). ZIKV antigens produced in immunocompromised mice brain was provided by the Central Laboratory of Federal District and was used as a positive control.
Histology analysis. The colon from mice was collected, carefully washed with saline buffer, and fixed in 10% Formalin. Thereafter, the tissue was cut into slices, dehydrated, and embedded with paraffin. The slides were stained with hematoxylin and eosin (HE) (Sigma) following standard procedures 58 . Sections were examined by light microscope Zeiss Lab. A1 Axiocam 105 color and photomicrographs were scanned using the ZEN program from Zeiss. Tissue samples were well-oriented with longitudinally cut crypts to precisely assess alterations in the overall intestinal tissue architecture. The slides were blindly scored based on a semiquantitative scoring system that includes the main alterations observed: (I) inflammatory cell infiltrate: severity and extent, (II) Epithelial changes: hyperplasia, goblet cell depletion, and erosion. Each parameter could receive 0-5 in the score index (0: normal; 1: minimal; 2: mild; 3: moderate; and 4 or 5 extensive) 59 .

Statistical analysis.
Results contained in the present work were reported by mean ± standard deviation (SD). Statistical differences among the two compared groups were made using Student's t-test with Bonferroni corrections provided by GraphPad PRISM Software version 6.00 or QIIME software version 1.9.1. p values are represented by asterisks: p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***) and p ≤ 0.0001 (****).

Data availability
Sequence data are deposited in the Sequence Read Database (SRA) with the accession number SUB7941842.