The Entero-Mammary Pathway and Perinatal Transmission of Gut Microbiota and SARS-CoV-2

COVID-19 is a severe respiratory disease threatening pregnant women, which increases the possibility of adverse pregnancy outcomes. Several recent studies have demonstrated the ability of SARS-CoV-2 to infect the mother enterocytes, disturbing the gut microbiota diversity. The aim of this study was to characterize the entero-mammary microbiota of women in the presence of the virus during delivery. Fifty mother–neonate pairs were included in a transversal descriptive work. The presence of SARS-CoV-2 RNA was detected in nasopharyngeal, mother rectal swabs (MRS) and neonate rectal swabs (NRS) collected from the pairs, and human colostrum (HC) samples collected from mothers. The microbiota diversity was characterized by high-throughput DNA sequencing of V3-16S rRNA gene libraries prepared from HC, MRS, and NRS. Data were analyzed with QIIME2 and R. Our results indicate that several bacterial taxa are highly abundant in MRS positive for SARS-CoV-2 RNA. These bacteria mostly belong to the Firmicutes phylum; for instance, the families Bifidobacteriaceae, Oscillospiraceae, and Microbacteriaceae have been previously associated with anti-inflammatory effects, which could explain the capability of women to overcome the infection. All samples, both positive and negative for SARS-CoV-2, featured a high abundance of the Firmicutes phylum. Further data analysis showed that nearly 20% of the bacterial diversity found in HC was also identified in MRS. Spearman correlation analysis highlighted that some genera of the Proteobacteria and Actinobacteria phyla were negatively correlated with MRS and NRS (p < 0.005). This study provides new insights into the gut microbiota of pregnant women and their potential association with a better outcome during SARS-CoV-2 infection.


Introduction
The actual well-known Coronavirus disease 2019  is produced by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), causing a worldwide pandemic that began in 2020. At the population level, pregnant women are a vulnerable group, with a greater risk of COVID-19 morbidity and mortality, increasing the risk of adverse pregnancy outcomes [1]. In this context, statistics from the UK reported a total of 9% of pregnant women and 6-week postpartum COVID-19 admissions to intensive care units (ICU) [2], while a review highlighted that most infections occurred during the third trimester, 11% of pregnant women with COVID-19 required admission to ICU, and 8% required mechanical ventilation [3]. It has also been reported that prevalence among women is variable, rising from 0.5% to 5% in the span of 2 weeks [4]. Regarding the Mexican population, it was reported that during the second peak of the pandemic, approximately 12% of asymptomatic pregnant women were positive for the virus [1].
Recently, several studies have demonstrated the ability of SARS-CoV-2 to infect and replicate in enterocytes of the human small intestine, in addition to RNA virus detection in fecal samples and the altered gut microbiota structure in SARS-CoV-2 positive patients [5,6]. It is known that SARS-CoV-2 infection is related to the angiotensin-converting enzyme receptor 2 (ACE2), whose activity is influenced and, in turn, affects the functionality of the gut microbiota [7]. For this reason, the association between gut microbiota and SARS-CoV-2 infection has been studied, given that the gut microbiota not only supports mucosal immunity but also modulates the systemic immune response in the host [8]. Reduction in the abundance of key immunomodulatory species of the gut microbiota, such as Bifidobacteria, Faecalibacterium prausnitzii, or Eubacterium rectale, has been reported in COVID-19 patients [5,9].
The alteration in the gut microbiota of pregnant women is an important aspect of the study since the maternal gut microbiota plays a crucial role in the first microbiota and immunity of the newborn. It has been reported that the maternal microbiota induces the activation of B cells [10,11] and other immunological components that migrate to the mammary gland to be transferred to the neonate through human milk [12]. Moreover, there is an interesting hypothesis for the translocation of internal bacteria from the mother's gastrointestinal (GI) tract into the mammary gland, driven by immune cells and occurring during the late stages of pregnancy. In this manner, the maternal GI tract is a source of bacteria for the milk microbiota, and the maternal gut microbiota is vertically transferred to the infant via human milk [13,14]. Human milk contains specific microbiota comprising mostly bacteria such as Staphylococcus, Pseudomonas, Streptococcus, Propionibacterium, and Sphingomonas, important anaerobic bacteria such as Bifidobacterium, Faecalibacterium, and Akkermansia, and remarkable lactic acid bacteria such as Lactobacillus [15,16]. This microbiota is the source of the first colonizing bacteria of the neonate's GI tract, playing important roles in the metabolism of milk sugar, influencing the development of the immune system, producing secondary metabolites such as short-chain fatty acids (SCFA), vitamins, contributing to the reduction of infections by potential pathogens, producing antimicrobial components, and improving the function of the neonate's intestinal barrier [17].
Despite the important role of maternal gut microbiota in the human milk microbiota and neonatal health, limited information has been published with respect to the COVID-19 influence on the maternal microbiota or the effect of SARS-CoV-2 maternal infection on the human milk bacterial diversity and microbial populations. The aim of this work was to analyze the microbiota from rectal swab samples, taken during delivery, of women with detectable presence of SARS-CoV-2 genome RNA. The diversity of the bacterial microbiota found in their human colostrum and in rectal swab samples from their newborns was also characterized to compare with the bacterial community found in samples from healthy women. The results obtained in this study provide interesting insight into changes occurring in bacterial community diversity in the infection state and its influence on newborn health.
We hypothesize that an alteration of the gut microbiota profile during SARS-CoV-2 infection in the mother is associated with changes in the bacterial community in the human colostrum, which potentially will influence bacterial gut colonization in the newborn.

Presence of SARS-CoV-2 Genomic RNA in the Mother-Neonate Pairs
The sampled women were on average 25 years old, with an age range from 16 to 38 years, while the gestational age revealed an average of approximately 39 weeks. Their heights and weights were within common ranges for women in the Mexican population, as well as the BMI data, which indicated a high abundance of overweight and obesity conditions in the sample ( Table 1). The 50 neonates included in the study were less than 6 days old, 32% of them were female, and 68% were male, with an average weight of 2.92 (±0.42) kg and size of 49.51 (±1.47) cm. The mother rectal swab (MRS), human colostrum (HC), and neonate rectal swab (NRS) samples of the 50 mother-neonate pairs were initially characterized for the presence of SARS-CoV-2 genome RNA using RT-ddPCR to detect the N gene. The RT-ddPCR test results were positive for two-thirds of mothers at the time of delivery, whereas 44 of them (88%) had no COVID-19 symptoms ( Table 1). The RT-qPCR results for detection of SARS-CoV-2 in nasopharyngeal swabs of the same mothers were positive for nearly one-third of the RT-ddPCR positive women, while 100% of the RT-ddPCR negative women were also negative based on RT-qPCR. No other data collected from the same women, such as blood tests, risk factors, parity, or socioeconomic data, showed any interesting association or tendency with the presence of SARS-CoV-2 RNA in MRS (Table 1). For the HC samples, less than half of the women expressed positive results for the virus using RT-ddPCR (Table 1). On the other hand, SARS-CoV-2 genome RNA detection by digital PCR in NRS showed that half of neonates carried the viral genome. Interestingly among the positive cases, only one-fifth were positive for the presence of the virus by RT-qPCR detection of SARS-CoV-2 in nasopharyngeal swabs (Table 2). No other collected variable in the neonates, such as the qualification status at birth and the somatometric data, showed any interesting association or tendency with the presence of SARS-CoV-2 RNA in MRS (Table 2).

Diversity of Microbiota Taxa and SARS-CoV-2 Positivity
The microbiota diversity in MRS, HC, and NRS samples was characterized by semiconductor sequencing of V3-16S rRNA gene libraries. The total reads had an average of 2 million for each type of sample, with a median quality score of 30 ( Table 3). The rarefaction plots also indicated satisfactory sequencing (Supplementary Materials Figure  S1). At the phylum level, the microbiota diversity showed a predominance of Bacteroidetes and Firmicutes phyla in the MRS samples; a large relative abundance of Firmicutes phylum in HC, and a dominance of Firmicutes and Proteobacteria phyla in the NRS samples (Figure 1A). There was no statistically significant difference in the phyla's relative abundances associated with the detection of SARS-CoV-2 by RT-ddPCR (Supplementary Materials Table  S2). At the family level, for the MRS samples, the bacterial diversity was characterized by a predominance of the Prevotellaceae (Bacteroidetes), Ruminococcaceae (Firmicutes), and Enterobacteriaceae (Proteobacteria) families ( Figure 1B). However, only Bifidobacteriaceae (Actinobacteria) and Oscillospiraceae (Firmicutes) were more abundant in the SARS-CoV-2 positive cases by digital PCR, while Microbacteriaceae (Actinobacteria) was more abundant in the negative cases with statistical significance (Supplementary Materials Table S3). The diversity at the family level in the HC samples showed a clear predominance of Staphylococcaceae (Firmicutes) in both positive and negative samples for SARS-CoV-2 by digital PCR ( Figure 1B), with no statistical significance (Supplementary Materials Table S4). On the other hand, the family abundance in the NRS samples showed a predominance of Streptococcaceae (Firmicutes) and Enterobacteriaceae in the negative samples for SARS-CoV-2 by RT-ddPCR, and a predominance of Enterobacteriaceae in the positive samples ( Figure 1B). In this case, the comparatively higher abundance of Streptococcaceae (Firmicutes) in the negative, and higher abundance of Ruminococcaceae in the positive, samples for SARS-CoV-2 by RT-ddPCR were statistically significant (Supplementary Materials  Table S5).  Tables S3-S5), while "Other" groups families with < 2.5% relative abundance (Supplementary Materials Table S6).   Tables S3-S5), while "Other" groups families with < 2.5% relative abundance (Supplementary Materials Table S6). Table 3. Sequencing summary for all samples of this study.

Alfa and Beta Diversity of the Bacterial Microbiota and SARS-CoV-2 Positivity
The analyses characterizing the alfa diversity of MRS, HC, and NRS samples did not show statistically significant difference for observed number of species, Chao 1, Shannon, and Simpson indexes (Supplementary Materials Table S7); there was only a slight tendency for higher observed number of species in the MRS, HC, and NRS positive samples for SARS-CoV-2 by RT-ddPCR ( Figure 2A). The beta diversity, on the other hand, clustered apart the MRS from the HC samples with statistical significance, regardless of the SARS-CoV-2 detection by RT-ddPCR ( Figure 2B). For the MRS samples, there was statistically significant difference between the beta diversity associated with the detection of SARS-CoV-2 by RT-ddPCR (Supplementary Materials Figure S2A).

Alfa and Beta Diversity of the Bacterial Microbiota and SARS-CoV-2 Positivity
The analyses characterizing the alfa diversity of MRS, HC, and NRS samples did not show statistically significant difference for observed number of species, Chao 1, Shannon, and Simpson indexes (Supplementary Materials Table S7); there was only a slight tendency for higher observed number of species in the MRS, HC, and NRS positive samples for SARS-CoV-2 by RT-ddPCR ( Figure 2A). The beta diversity, on the other hand, clustered apart the MRS from the HC samples with statistical significance, regardless of the SARS-CoV-2 detection by RT-ddPCR ( Figure 2B). For the MRS samples, there was statistically significant difference between the beta diversity associated with the detection of SARS-CoV-2 by RT-ddPCR (Supplementary Materials Figure S2A).  Table S6). (B) Beta diversity of bacteria in MRS and HC. The graphics show beta-diversity analyses calculated by dissimilarity metrics using features tables and Unweighted UniFrac analysis. The scatter plots were generated using principal coordinates analysis (PCoA) in three different axes showing the percentage of total differences. The positive (+) or negative (−) SARS-CoV-2 genome detection by RT-ddPCR is shown beside the type of labels identified for the samples on the right side of the graphics. MRS and HC differed significantly according to ANOSIM (p = 0.001).

SARS-CoV-2 Positivity Is Associated with an Increased Abundance of Several Bacterial Taxa
The LEfSe analysis of the bacterial community present in the MRS, HC, and NRS samples revealed bacterial taxa with large relative abundances associated with the presence of the SARS-CoV-2 genome as detected by RT-ddPCR. The negative MRS samples had an increased abundance of bacterial members of the phyla Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria, while the positive samples exhibited an increased abundance of the same phyla plus members of the phylum Verrucomicrobiota ( Figure  3A). Remarkably, there was an increase in the abundance of bacteria of genera such as Akkermansia, Butyricimonas, and families Lachnospiraceae and Ruminococcaceae associated with the detection of SARS-CoV-2. The negative HC samples showed an increase in the abundance of bacterial members of the Actinobacteria and Firmicutes phyla, while the  Table S6). (B) Beta diversity of bacteria in MRS and HC. The graphics show beta-diversity analyses calculated by dissimilarity metrics using features tables and Unweighted UniFrac analysis. The scatter plots were generated using principal coordinates analysis (PCoA) in three different axes showing the percentage of total differences. The positive (+) or negative (−) SARS-CoV-2 genome detection by RT-ddPCR is shown beside the type of labels identified for the samples on the right side of the graphics. MRS and HC differed significantly according to ANOSIM (p = 0.001).

SARS-CoV-2 Positivity Is Associated with an Increased Abundance of Several Bacterial Taxa
The LEfSe analysis of the bacterial community present in the MRS, HC, and NRS samples revealed bacterial taxa with large relative abundances associated with the presence of the SARS-CoV-2 genome as detected by RT-ddPCR. The negative MRS samples had an increased abundance of bacterial members of the phyla Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria, while the positive samples exhibited an increased abundance of the same phyla plus members of the phylum Verrucomicrobiota ( Figure 3A). Remarkably, there was an increase in the abundance of bacteria of genera such as Akkermansia, Butyricimonas, and families Lachnospiraceae and Ruminococcaceae associated with the detection of SARS-CoV-2. The negative HC samples showed an increase in the abundance of bacterial members of the Actinobacteria and Firmicutes phyla, while the positive HC samples had an increase in bacteria from the same phyla plus Cyanobacteria and Proteobacteria ( Figure 3B).
In this case there was an increase in the abundance of bacteria of genera such as Lawsonella and families such as Ruminoccocaceae associated with SARS-CoV-2. Finally, the NRS negative samples showed an increase in the abundance of bacterial taxa of the Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria phyla, whereas the NRS positive samples included only Actinobacteria, Firmicutes, and Proteobacteria ( Figure 3C). There was an increase in the genera Bifidobacterium, Coprococcus, and Phascolarctobacterium associated with SARS-CoV-2 detection. The change in the abundance of all mentioned taxa was statistically significant (Supplementary Materials Table S8). positive HC samples had an increase in bacteria from the same phyla plus Cyanobacteria and Proteobacteria ( Figure 3B). In this case there was an increase in the abundance of bacteria of genera such as Lawsonella and families such as Ruminoccocaceae associated with SARS-CoV-2. Finally, the NRS negative samples showed an increase in the abundance of bacterial taxa of the Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria phyla, whereas the NRS positive samples included only Actinobacteria, Firmicutes, and Proteobacteria ( Figure 3C). There was an increase in the genera Bifidobacterium, Coprococcus, and Phascolarctobacterium associated with SARS-CoV-2 detection. The change in the abundance of all mentioned taxa was statistically significant (Supplementary Materials Table  S8).  Table S8 for full taxon descriptions and LDA score p-values and q-values.

Shared Taxa and Differential Frequency of Taxa among Samples
An analytical inspection of the feature table obtained for each type of sample showed 102 bacterial taxa shared between MRS and HC; 162 taxa shared between HC and NRS, and 168 bacterial taxa shared between MRS and NRS samples ( Figure 4A). The analysis of high-throughput sequencing data for the microbiota based on the negative binomial distribution showed that in the MRS SARS-CoV-2 positive samples by RT-ddPCR, there was an increase in the abundance of members of Firmicutes such as Clostridia_UCG-014, Megasphaera, Ruminococcaceae_NK4A214_group, Ruminococcus, Ruminococcaceae_UCG-005, Ruminococcaceae_Uncultured, Christensenellaceae_R−7_group, Erysipelotrichaceae_UCG-003,  Table S8 for full taxon descriptions and LDA score p-values and q-values.

Shared Taxa and Differential Frequency of Taxa among Samples
An analytical inspection of the feature table obtained for each type of sample showed 102 bacterial taxa shared between MRS and HC; 162 taxa shared between HC and NRS, and 168 bacterial taxa shared between MRS and NRS samples ( Figure 4A). The analysis of high-throughput sequencing data for the microbiota based on the negative binomial distribution showed that in the MRS SARS-CoV-2 positive samples by RT-ddPCR, there was an increase in the abundance of members of Firmicutes such as Clostridia_UCG-014, Megasphaera, Ruminococcaceae_NK4A214_group, Ruminococcus, Ruminococcaceae_UCG-005, Ruminococcaceae_Uncultured, Christensenellaceae_R-7_group, Erysipelotrichaceae_UCG-003, Roseburia, Ruminococcaceae_UCG-002; phylum Actinobacteriota such as Bifidobacterium; phy-lum Verrucomicrobiota such as Akkermansia, and phylum Bacteroidetes such as Bacteroides, and a decrease in the abundance of Actinobacteriota, genus Pseudonocardia. For the MRS SARS-CoV-2 negative samples by RT-ddPCR, the increase was for Firmicutes such as Ezakiella and Bacteroidetes such as Porphyromonas ( Figure 4B). The NRS SARS-CoV-2 positive samples had an increase in the abundance of the genus Blautia in Firmicutes, while the genera Streptococcus (Firmicutes) and Mesorhizobium (Proteobacteria) were increased in the MRS SARS-CoV-2 negative samples ( Figure 4B). On the other hand, only the HC samples showed a decrease in the Firmicutes Blautia by this analysis (Figure 4B).  Figure 4B). The NRS SARS-CoV-2 positive samples had an increase in the abundance of the genus Blautia in Firmicutes, while the genera Streptococcus (Firmicutes) and Mesorhizobium (Proteobacteria) were increased in the MRS SARS-CoV-2 negative samples ( Figure 4B). On the other hand, only the HC samples showed a decrease in the Firmicutes Blautia by this analysis ( Figure 4B).

Origin of Bacterial Taxa Present in the Human Colostrum Samples
The potential origin of bacteria present in the HC was explored. SourceTracker analysis estimated that approximately 21% of the identified bacteria in the HC had a common source with the bacteria characterized in MRS, while the remaining 79% came from unknown sources (p < 0.006, Wilcoxon signed-rank test) ( Figure 5A). The relative abundance of bacteria found in HC classified by the analysis of MRS origin, showed that members of phyla Firmicutes Staphylococcus (46%), and Streptococcus (6%), Proteobacteria Escherichia (11%), Actinobacteria Microbacterium (7%), and Bacteroidetes Prevotella (5%) were more abundant, while the relative abundance of taxa in the Unknown source was dominated by the Firmicutes Staphylococcus (47%), Lactobacillus (4%) and Actinobacteria Cutibacterium (5%) ( Figure 5B and Supplementary Materials Table S9).

Origin of Bacterial Taxa Present in the Human Colostrum Samples
The potential origin of bacteria present in the HC was explored. SourceTracker analysis estimated that approximately 21% of the identified bacteria in the HC had a common source with the bacteria characterized in MRS, while the remaining 79% came from unknown sources (p < 0.006, Wilcoxon signed-rank test) ( Figure 5A). The relative abundance of bacteria found in HC classified by the analysis of MRS origin, showed that members of phyla Firmicutes Staphylococcus (46%), and Streptococcus (6%), Proteobacteria Escherichia (11%), Actinobacteria Microbacterium (7%), and Bacteroidetes Prevotella (5%) were more abundant, while the relative abundance of taxa in the Unknown source was dominated by the Firmicutes Staphylococcus (47%), Lactobacillus (4%) and Actinobacteria Cutibacterium (5%) ( Figure 5B and Supplementary Materials Table S9).  Table S9).

Bacterial Taxa Correlating with the Presence of SARS-CoV-2 and Metadata Collected in Mothers
The correlation of the bacterial taxa abundance with numerical clinical metadata was explored. The Spearman analyses showed that in the MRS samples with a significance of p < 0.001, members of bacterial taxa such as Firmicutes (Colidextribacter, Lachnospiraceae, and Flavonifractor) were positively correlated with the presence of SARS-CoV-2 as detected by RT-ddPCR, while Lactococcus (Firmicutes) was negatively correlated with the presence of SARS-CoV-2. On the other hand, the SARS-CoV-2 virus genomic RNA detection by RT-qPCR in nasopharyngeal swabs collected from the mothers was positively correlated with members of the phyla Proteobacteria (Undibacterium) and Firmicutes (Oscillospiraceae family and Faecalibacterium UBA1819), and negatively correlated with Proteobacteria (Beijerrinckiaceae and Bradyrhizobium), Actinobacteria (Slackia and Microbacterium), Firmicutes (Fenollaria), and Proteobacteria (Sphingomonas and Mesorhizobium) (Figure 6A). For the case of HC samples, bacteria of phylum Actinobacteria (Lawsonella) correlated positively with the detection of SARS-CoV-2 by RT-ddPCR, while there was also a positive correlation of SARS-CoV-2 detection by RT-qPCR in nasopharyngeal swabs collected from the mothers with members of the phyla Proteobacteria (Enterobacter, Comamonas, Pseudomonas) and Firmicutes (Gemella). There was, in addition, an interesting positive correlation of gestational weeks with members of the phylum Firmicutes (Blautia) and a negative correlation of blood glucose levels with Firmicutes (Oscillospiraceae UCG-002), Bacteroidetes (Bacteroides), and Proteobacteria (Aeromonas) ( Figure 6B). Finally, the NRS samples showed a negative correlation of SARS-CoV-2 detection by RT-ddPCR with Actinobacteria (Brevibacterium), an interesting positive correlation of members of phylum Deinococcota (Meiothermus) with the delivery mode (vaginal), and correlation of members of the phyla Firmicutes (Subdoligranulum, Paraclostridium) and Proteobacteria (Sphingomonas) with the male sex of newborns ( Figure 6C).  Table S9).

Bacterial Taxa Correlating with the Presence of SARS-CoV-2 and Metadata Collected in Mothers
The correlation of the bacterial taxa abundance with numerical clinical metadata was explored. The Spearman analyses showed that in the MRS samples with a significance of p < 0.001, members of bacterial taxa such as Firmicutes (Colidextribacter, Lachnospiraceae, and Flavonifractor) were positively correlated with the presence of SARS-CoV-2 as detected by RT-ddPCR, while Lactococcus (Firmicutes) was negatively correlated with the presence of SARS-CoV-2. On the other hand, the SARS-CoV-2 virus genomic RNA detection by RT-qPCR in nasopharyngeal swabs collected from the mothers was positively correlated with members of the phyla Proteobacteria (Undibacterium) and Firmicutes (Oscillospiraceae family and Faecalibacterium UBA1819), and negatively correlated with Proteobacteria (Beijerrinckiaceae and Bradyrhizobium), Actinobacteria (Slackia and Microbacterium), Firmicutes (Fenollaria), and Proteobacteria (Sphingomonas and Mesorhizobium) ( Figure 6A). For the case of HC samples, bacteria of phylum Actinobacteria (Lawsonella) correlated positively with the detection of SARS-CoV-2 by RT-ddPCR, while there was also a positive correlation of SARS-CoV-2 detection by RT-qPCR in nasopharyngeal swabs collected from the mothers with members of the phyla Proteobacteria (Enterobacter, Comamonas, Pseudomonas) and Firmicutes (Gemella). There was, in addition, an interesting positive correlation of gestational weeks with members of the phylum Firmicutes (Blautia) and a negative correlation of blood glucose levels with Firmicutes (Oscillospiraceae UCG-002), Bacteroidetes (Bacteroides), and Proteobacteria (Aeromonas) ( Figure 6B). Finally, the NRS samples showed a negative correlation of SARS-CoV-2 detection by RT-ddPCR with Actinobacteria (Brevibacterium), an interesting positive correlation of members of phylum Deinococcota (Meiothermus) with the delivery mode (vaginal), and correlation of members of the phyla Firmicutes (Subdoligranulum, Paraclostridium) and Proteobacteria (Sphingomonas) with the male sex of newborns ( Figure 6C). Columns show the bacterial taxa, while rows show the numerical metadata. The color keys from blue (−1, negative) to red (+1, positive) measure the correlation. The plus symbol "+" denotes a significance of p < 0.001. For the "Sex" variable, the blue color indicates a correlation of the bacteria with male children; for the "Delivery mode" variable, the red color indicates a correlation with vaginal delivery. Two MRS, two HC, and four NRS samples were not included in these analyses since they lacked suitable metadata.

Discussion
COVID-19, caused by SARS-CoV-2, affects the health of people worldwide, severely affecting adults, and has caused changes in social habits in the population, with priorities placed on prevention and the care of the elderly. In addition, the pandemic has also raised concerns for pregnant women and newborns, and prevention measures were increased in the hospitals [18]. During pregnancy, intestinal dysbiosis associated with COVID-19 Columns show the bacterial taxa, while rows show the numerical metadata. The color keys from blue (−1, negative) to red (+1, positive) measure the correlation. The plus symbol "+" denotes a significance of p < 0.001. For the "Sex" variable, the blue color indicates a correlation of the bacteria with male children; for the "Delivery mode" variable, the red color indicates a correlation with vaginal delivery. Two MRS, two HC, and four NRS samples were not included in these analyses since they lacked suitable metadata.

Discussion
COVID-19, caused by SARS-CoV-2, affects the health of people worldwide, severely affecting adults, and has caused changes in social habits in the population, with priorities placed on prevention and the care of the elderly. In addition, the pandemic has also raised concerns for pregnant women and newborns, and prevention measures were increased in the hospitals [18]. During pregnancy, intestinal dysbiosis associated with COVID-19 inflammation could affect the composition of mothers' and newborns' pioneering bacterial communities [19], through the entero-mammary pathway and vertical transmission to the neonate. During COVID-19, there is an interaction between the gastrointestinal and respiratory tract that might cause changes in the gut microbiota. In fact, previous studies have observed alterations in the host microbiota after viral lung infections, resulting in increases in the Bacteroidetes and Firmicutes ratio [20]. Specifically, for COVID-19 patients, significantly lower bacterial diversity and higher relative abundance of opportunistic pathogens have been reported [21].
In our work, upon admission to the hospital, 28 of 33 women with MRS positive for SARS-CoV-2 by RT-ddPCR (MRS+) were asymptomatic, while five had symptoms related to COVID-19 disease; for the group of 17 women with negative MRS, only one exhibited SARS-CoV-2 symptoms. The observed symptoms were breathlessness, cough, diaphoresis, diarrhea, fever, headache, and runny nose, which were lower in comparison to other reports. In a study in the UK population, 342 of 424 (80%) women in the third trimester of pregnancy or peripartum had symptoms on hospital admission, with fever, cough, and breathlessness the most commonly reported [22]. However, pregnant women with COVID-19 were less likely to be symptomatic than non-pregnant counterparts because the risk factors for severe disease include being overweight or obese, older than 35 years of age, and having pre-existing comorbidities [2]. We found a positive trend in SARS-CoV-2 RNA as detected by both tests (RT-qPCR and RT-ddPCR) in male compared to female newborns ( Table 2). In other studies, the ratio of Indian male to female infected newborns was 2:1, as was observed in our study [23]. A review of different studies of newborns of different nationalities positive for SARS-CoV-2 reported a male-to-female ratio of 2.8, indicating that male infants were more susceptible to the viral infection than females [24]. The basis of male susceptibility might be explained by the results obtained in murine models, where males are more susceptible to the infection due to the viral recognition of the androgen receptor present in male mice [25]. On the other hand, possible reasons for the differences in SARS-CoV-2 detection among the MRS, HC, and NRS samples might be due to the pathophysiology of the disease, which proceeds from the upper respiratory tract to the lungs, reaching other organs of the lower body including epithelial tissue in the intestinal tract [26]. In addition, differences in viral detection might be associated with the type of sample. A report on SARS-CoV-2 detection using RT-PCR in different types of clinical specimens showed that the rectal swab had a higher positive rate of detection compared to the nasopharyngeal swab, which had a moderate detection rate. The use of rectal swab sampling is recommended for clinical diagnosis of COVID-19 and can be considered a representative gastrointestinal specimen [27]. Additionally, although detection with the RT-qPCR technique is possible with at least 10 genomic copies, more accurate quantification of lower concentrations is obtained via digital PCR, which is based on multiple reactions [28]. All of these might explain the different results for viral detection in the samples of our study, including a higher number of SARS-CoV-2 positive MRS and NRS samples compared to nasopharyngeal swab samples.
The present study reported differences between the microbial composition of SARS-CoV-2 positive versus negative MRS, NRS, and HC samples, analyzed by RT-ddPCR test, obtained from Mexican pregnant women and their neonates. The most abundant phyla found in the different samples were Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria, which represented approximately 97% of the microbial composition in each type of sample regardless of the SARS-CoV-2 positivity; these phyla were similar to those in other reports [5,15,29,30]. In the MRS+ samples, members of the Bifidobacteriaceae and Oscillospiraceae families were enriched, while Microbacteriaceae was decreased with a significant difference, suggesting an alteration in the intestinal microbiota due to the presence of the SARS-CoV-2 virus in infected pregnant women. A previous study of nasopharyngeal swab samples taken from Spanish pregnant women with COVID-19 reported a bacterial diversity at the phylum level similar to that in our work, where the Prevotellacea family had the highest abundance [31]. In another comparative study in the Italian population carried out using rectal swab samples from healthy individuals versus COVID-19 patients at different phases of the disease, the comparison of the microbial composition pointed out significant changes in the bacterial communities during progressive phases of the disease [29].
Further, in our work, three representative taxa of the phylum Firmicutes (Lachnospirraceae, Colidextribacter, Flavonifractor) were positively correlated with the presence of SARS-CoV-2 in MRS, and Faecalibacterium UBA1819 was positively correlated with SARS-CoV-2 detection in nasopharyngeal swabs by Spearman analysis. Interestingly, Colidexibacter massiliensis species have been isolated from the human colon of obese patients [32], and an increase in the abundance of members of the families Ruminococacceae, Lachnospiraceae, and Enterocacaceae has been linked to the diagnosis of gestational diabetes mellitus (GDM) [33,34]. On the other hand, Flavonifractor plautii, a bacterium involved in the degradation of anticancer flavonoids, has been found in patients with colorectal cancer [35]. The abundance of members of the family Oscillospiraceae was significantly increased in the MRS+ samples in our work. Members of the same family as Faecalibacterium prausnitzzi present in healthy intestinal microbiota decreased their abundance in the dysbiosis associated with colitis, and it is reported that these bacteria exert an anti-inflammatory action, acting as immunomodulators [5,29,36,37]. A decrease in the abundance of Faecalibacterium is associated with GDM [34]; specifically, Faecalibacterium UBA1819 has been negatively correlated with metabolic syndrome in model mice after dietary fiber supplementation [38]. The genus Oscillobacter was found to increase after dietary fiber supplementation in a patient with postacute COVID-19 syndrome [38], and bacteria of the NK4A214 group have been linked to low insulin resistance [39]. In other work, authors reported that gut microbiota of a COVID-19 group was dominated by the genera Streptococcus, Rothia, Veillonella, Eryspelatoclostridium, and Actinomyces, whereas the healthy group was dominated by genera Romboustia, Faecalibacterium, Fusicatenibacter, and Eubacterium halli [20]. Clostridium ramosum, Coprobacillus, and Clostridium hathewayi correlated with COVID-19 severity, while Faecalibacterium prausnitzii was negatively correlated with disease severity. Bacteroides thetaiotaomicron, B. massiliensis, B. dorei, and B. ovatus were inversely correlated with SARS-CoV-2 in fecal samples from patients [20,30].
Moreover, Oscillospiraceae UCG-010, Oscillospiraceae UCG-005, and Anaerotruncus have been found as potential biomarkers of inflammation [40][41][42], while Oscillospiraceae UCG-002 might play determinant roles in gut microbial community structure and function leading to the development of IgE-mediated food allergy [43]. Some strains of Akkermansia (phylum Verrucomicrobia) were associated with lower insulin sensitivity [37]; A. muciniphila was detected in COVID-19 patients during hospitalization [5]. The genus Desulfovibrio (Proteobacteria), which was positively correlated with glucose in our work, is associated with GDM in women with high fat intake in their diet, causing an increase in the abundance of Parabacteroides, Prevotella, Escherichia coli, Sutterella, and Desulfovibrio, which weaken the intestinal epithelial permeability [37,44].
On the other hand, Blautia, which we found to be negatively correlated with BMI, was found depleted in SARS-CoV-2 infection [45] and was associated with reduced inflammatory response in patients with COVID-19 [5]. Parvimonas, associated in our study with MRS negative for SARS-CoV-2 by RT-ddPCR, was also found to increase in abundance after SARS-CoV-2 vaccination in oral samples [46]. Sphingomonas, associated in our study with nasopharyngeal swabs negative for the virus in the MRS samples, was correlated with altered serum metabolites in symptomatic COVID-19 patients in other studies [47]. The genera Butyrcimonas (Odoribacteriaceae family) and Porphyromonas (Porphyromonadacea family), associated with negative detection of SARS-CoV-2 in MRS samples, are bacterial producers of butyric, isobutyric, isovaleric, acetic, and propionic acids [48], and are reported to increase along with COVID-19 recovery [49].
Previous studies have reported that the main phyla in human milk are Firmicutes and Proteobacteria, with differences in abundance between them depending on the milk stages (colostrum, transitional milk, or mature milk). The main phyla detected in the colostrum samples of this study were Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria, representing 95% of samples negative for SARS-CoV-2 by RT-ddPCR, and 97% of those samples that tested positive, with no significant differences between HC samples positive or negative SARS-CoV-2. Firmicutes was the most abundant phylum in both groups, with more than 50% represented by members of the Staphyloccaceae family in the negative samples, and 40% in the positive samples. In comparison to other studies, similarities and differences have been found in relation to the bacterial profiles of the colostrum samples. In a study of the Mexican population, the most abundant phyla in human milk samples were Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria, representing more than 97% of the total; however, unlike our study, the most abundant phyla were Proteobacteria (55%) and Firmicutes (25%), represented by the families Pseudomonadaceae (71%) and Staphylococaceae (82%) [15]. Similarly, in a study carried out using human milk of Caucasian women, most phyla mentioned above were also found, with Proteobacteria the most abundant (67%), followed by Firmicutes (26%) [50]. In a different study using human milk manually collected 1 month after delivery, Firmicutes was found to be the most abundant phylum, followed by the Staphylococaceae family [51]. A report comparing the bacterial population present in human milk and the skin areola found that the most abundant phylum in milk was Proteobacteria, while for the areola it was Firmicutes, to which we could relate the dominance of the phylum Firmicutes in colostrum samples to the presence of skin flakes in the milk [52]. Additionally, in human milk samples obtained 30 days postpartum, Firmicutes (79%) and Proteobacteria (14%) were the most frequent phyla, and the predominant families were Streptococcaceae (50%), Gemellaceae (15%), and Staphylococcaceae (11%) [53]. Finally, in other studies, in human milk, the most abundant taxa at the genus level were Streptococcus (16%), Ralstonia (5%), and Staphylococcus (5%) [51]; or Streptococcus (73%) and Staphylococcus (10%) [54].

Study Type and Selection of Subjects
This study was transversal, descriptive, and included 50 mother-neonate pairs recruited at General Hospital "Dr. Gustavo Baz Prada" located in the municipality of Ciudad Netzahualcoyotl (19 • 25 19 N 99 • 00 53 W) at "Instituto Materno Infantil del Estado de Mexico" (IMIEM) located in Toluca-de-Lerdo, Mexico (19 • 16 02.0 N 99 • 39 41.2 W) during the epidemiological wave between 17 July 2020 and 13 October 2020. Inclusion criteria were Mexican women from 16 to 38 years old in the third trimester of pregnancy, with no hormonal and no antibiotic treatments. Entries with incomplete or inadequate data were excluded. Based on a survey, sociodemographic and clinical information for mothers and neonates were collected (maternal age, gestational age at delivery, delivery mode, sex, age, clinical data, SARS-CoV-2 diagnosis, blood test, and socioeconomic data). The study was approved by the hospital's Bioethics Committee in Research, with registry number 208C0101110500T_2020-08. All participants consented to the collection of data and signed informed consent in accordance with the Declaration of Helsinki.

Sample Collection
For this study, 250 samples were obtained by medical staff. For SARS-CoV-2 RNA detection by RT-qPCR, 100 pharyngeal and nasopharyngeal swabs of mothers and neonates were collected according to the Standardized Guidelines for Epidemiological and Laboratory Surveillance of COVID-19 of the Health Secretary of the Mexican Government [55,56]. For SARS-CoV-2 RNA RT-ddPCR detection, 50 mother rectal swabs (MRS) and 50 neonate rectal swabs (NRS), were collected at the time of delivery in accordance with the "Method of collecting a rectal swab" [57]. Fifty colostrum or transition milk samples (HC) were collected from each mother between 0-6 days postpartum. The nipple area was cleaned using a sterile swab and water, and approximately 0.1-1 mL of the sample was collected in a sterile tube. All samples were immediately stored at −20 •   sion High-Fidelity DNA Polymerase (Thermo Scientific Baltics UAB, Vilnius, Lithuania), and deionized H 2 O). Each primer had a previous heat treatment (95 • C, 3 min followed by 4 • C, 3 min). The PCR cycling program consisted of 6 min, 98 • C of denaturing, followed by 25 thermo-cycles (12 s, 98 • C; 15 s, 62 • C; 10 s, 72 • C), and 5 min, 72 • C extension in a 720 Thermal Cycler (Applied Biosystems, Singapore). For the positive control DNA, a mixture of identified probiotics was used (Bioleven, Wilmington, DE, USA). Simulated DNA extractions were used as negative controls; in these cases, the~281 pb amplicon was not observed and consequently not sequenced. Finally, the concentration of every single 1-100 barcoded amplicon was determined, normalized, and pooled to build the final library.

ASV Determination and Taxonomic Annotation
QIIME 2022.2 [60] was used for ASV determination and taxonomic annotation. ASVs were determined with qiime dada2 denoise-single plugin; the option -p-trunc len 170 was used to trim and filter the sequences at this value. Taxonomic annotation was performed using the QIIME 2022.2 [60] feature-classifier classify-consensus-blast plugin, with a percentage of identity of 97%. Silva 138 database (accessed in July 2022) was used with the weighted pre-trained classifier (Weighted Silva 138, 99% OTUs full-length sequences) [61]. Run commands are provided in Supplementary Materials QIIME2 workflow.md.

Bacterial Relative Abundance and Diversity Analyses
For further analyses, R 4.2.0 [62] in RStudio 2022.02.32+492 [63] was used. Phyloseq 1.4.0 [64] package was used for analyses of microbial communities; for instance, alpha diversity Observed species, Chao1, Shannon, and Simpson indexes were calculated, and beta diversity was assessed with Unifrac distance. Qiime artifacts were imported using qiime2R 0.99.6 [65] [74]. Vegan 2.6-2 package [75] was employed for the analysis of similarities (ANOSIM) of beta diversity. Venn diagram and Spearman correlation were performed using microbiome 1.18.0 [76] and eulerr 6.1.1 [77] packages. The full script is provided in the Supplementary Materials phyloseq workflow.R.

SourceTracker 2 and LEfSe Analyses
SourceTracker analysis was conducted with sourcetracker2 giibs plugin [78]; rarefaction depth of 950 samples and p-no-loo parameter were utilized. Linear discriminant analysis effect size (LEfSe) was carried out with the Galaxy module from The Huttenhower Lab [79] using the microbial abundance calculated by QIIME2.

Statistical Analyses
Epidemiological, clinical, and biochemical variables were analyzed using descriptive statistics of the groups. Student's t-test was applied to assess the differences between the metadata, and microbial data of the given groups (SPSS Statistics, version 25). Data were represented with the mean ± SD. Statistics with p-values < 0.05 were considered significant. Spearman's rank correlation coefficient was used to assess the relationship between sample metadata and microbial relative abundance. Differences in beta diversity between groups were assessed with an analysis of similarities (ANOSIM). Differential gene expression analysis based on the negative binomial distribution analysis (DESeq2) was used to highlight relevant taxa in the groups. Data were processed with R, as described above.

Sequence Accession Numbers
The sequence FASTQ files and the corresponding mapping file for all samples used in this study were deposited in the NCBI BioProject ID PRJNA856971 Link https://www. ncbi.nlm.nih.gov/sra/PRJNA856971 (accessed on 8 August 2022).

Conclusions
The results of our work permit the conclusion that defined changes in the bacterial microbiota diversity are associated with the presence of SARS-CoV-2 in rectal swabs collected from mothers and neonates. In addition, the detection of the virus is also associated with changes in the microbiota present in human milk. Our results and conclusions are valid for the sample; however, they should be interpreted with caution due to the limited number of samples, the sampling method by rectal swab, and the fact that neonates were not breastfed by the mothers during the time that the MRS, NRS and HC samples were collected as a consequence of the hospital protocols due to the COVID-19 pandemic.