key: cord-0692542-xfxmv7ym authors: Xu, R.; Lu, R.; Zhang, T.; Wu, Q.; Cai, W.; Han, X.; Jin, X.; Zhang, Z.; Zhang, C.; Wan, Z. title: Temporal dynamics of human respiratory and gut microbiomes during the course of COVID-19 in adults date: 2020-07-22 journal: nan DOI: 10.1101/2020.07.21.20158758 sha: 429efe5630c303492597dd9ff02df864cbac9931 doc_id: 692542 cord_uid: xfxmv7ym SARS-CoV-2 infects multiple organs including the respiratory tract and gut. Whether regional microbiomes are disturbed significantly to affect the disease progression of COVID-19 is largely unknown. To address this question, we performed cross-sectional and longitudinal analyses of throat and anal swabs from 35 COVID-19 adults and 15 controls by 16S rRNA gene sequencing. The results allowed a partitioning of patients into 3-4 categories (I-IV) with distinct microbial community types in both sites. Lower-diversity community types often appeared in the early phase of COVID-19, and synchronous fast restoration of both the respiratory and gut microbiomes from early dysbiosis towards late near-normal was observed in 6/8 mild COVID-19 adult patients despite they had a relatively slow clinical recovery. The synchronous shift of the community types was associated with significantly positive bacterial interactions between the respiratory tract and gut, possibly along the airway-gut axis. These findings reveal previously unknown interactions between respiratory and gut microbiomes, and suggest that modulations of regional microbiota might help to improve the recovery from COVID-19 in adult patients. Study cohort 66 The study subjects included 35 adult COVID-19 patients from 17 to 68 years of age, 15 healthy 67 adults, and 10 non-COVID-19 patients with other diseases. All COVID-19 patients (except patient 68 p09) had mild clinical symptoms. A total of 146 specimens including 37 pairs of both throat and anal 69 swabs were collected from COVID-19 patients ( Supplementary Fig. S1 ). High-throughput Table 76 S1). Six throat microbial community types (or clusters) were identified using the Dirichlet 77 Multinomial Mixtures (DMM) modelling based on lowest Laplace approximation (Fig. 1a) and 78 visualized by Nonmetric Multidimensional Scaling (NMDS) based on Bray-Curtis distance (Fig. 1b) . 79 Healthy adults (H) and non-COVID-19 patients (NP) formed independent clusters. The vast majority 80 of the specimens of COVID-19 patients were divided into four community types, called I-IV, and 6 81 specimens were included in the NP type (Fig. 1a) . All COVID-19-related community types, as well 82 as the NP type, were significantly separated from the H type. Community types III and IV were not 83 only significantly separated from the types I and II, but also from each other (Fig. 1b) . Consistently, 84 significantly lower richness and evenness were observed in community types III and IV, compared 85 with the H type (Fig. 1c) . 86 To more directly demonstrate that the variation of throat microbial composition is an indicator 87 of COVID-19 disease stages, the community type-specific indicator taxa were identified based on the 88 top 30 microbial genera (Fig. 1d) . The type H was characterized by bacterial genus Bacteroides 89 (predominant taxa in the lung of healthy individuals) and unclassified Comamonadaceae, whereas 90 the NP type was marked by pro-inflammatory Enterobacteriaceae members. In contrast, the indicator 91 bacteria of four COVID-19-related community types were Alloprevotella in I, Porphyromonas, 92 Neisseria, Fusobacterium and unclassified Bacteroidales in II, Pseudomonas in III, and that the indicators in the types III and IV belong to putative pathogenic (e.g. Pseudomonas) and 95 opportunistic pathogenic bacteria (e.g. Rothia). Community type I contained Alloprevotella genus, as 96 well as abundant Bacteroides and Prevotella that typically present in the H type (Fig.1a) . Pathogenic 97 bacteria Pseudomonas, Saccharibacteria incertae sedis and Rothia that are associated with 98 pneumonia and various human diseases were significantly enriched in community types III and IV, 99 reflecting an impaired microbiome (dysbiosis) in the respiratory tract [15] [16] [17] . Apart from beneficial 100 commensals (e.g. Bacteroidales), some opportunistic pathogenic bacteria such as Porphyromonas, 101 Fusobacterium, and Neisseria that typically exist in the nasopharynx, and are associated with Table S1 and Fig. S2 ). According to indicator characteristics, community types I to IV reflected a 105 progressive imbalance of the respiratory microbiome. 106 Longitudinal analysis showed that lower-diversity community types often appeared in early 107 specimens (Fig. 1e) , and throat microbiota was dominated by few putative pathogenic and 108 opportunistic bacteria at the early phase of COVID-19 ( Supplementary Fig. S2 ). Prominent 109 microbiome community type shifts from early lower-diversity community types (NP, IV or II) 110 towards later higher-diversity types (II or I) were observed in 9/24 COVID-19 adults who had 111 specimens at two or more time points. For example, an obvious throat microbiome progression from 112 type (IV or II) in early specimens to type I in late specimens was observed in five patients (p17, p25, 113 p13, p11 and p05) with 4 or more consecutive specimens (Fig. 1e) . Accompanied with the restoration 114 of throat microbiota, beneficial commensals appeared to occupy the niches, and the bacterial diversity 115 increased accordingly ( Supplementary Fig. S2 ). However, a reversed pattern was observed in four 116 patients who had microbiome composition shift from early higher-diversity types (I or II) to later 117 lower-diversity type (II-IV), implying a worsening of the throat microbiome. For example, in the only 118 severe case (p09), the community type II on day 10 was shifted to type IV on day 27, and sustained 119 to at least day 33 after symptom onset (Fig. 1e) , and pathogenic bacteria Saccharibacteria incertae 120 sedis and Rothia were significantly enriched at late stage ( Supplementary Fig. S2 ). Furthermore, a 125 Gut microbiome dynamics in COVID-19 126 To expand the scope of this research, a total of 1,940 ASVs were recovered from the 16S-rRNA 127 gene sequences of all anal swabs, representing 13 known phyla including 182 known genera 128 (Supplementary Table S1 ). The gut microbial communities of COVID-19 patients formed three 129 distinct community types I-III ( Fig. 2a-b) . The richness and evenness of the gut microbiome 130 decreased from type I to III (Fig. 2c) . Indicator analyses showed that type I was primarily 131 characterized by healthy gut bacteria including Bacteroides genus and several known butyrate-132 producing bacteria (e.g. Faecalibacterium, Roseburia, Blautia, and Coprococcus) and one 133 opportunistic pathogenic bacterium (Finegoldia) (Fig. 2d) [21] [22] [23] [24] [25] [26] [27] . The indicators of type II mainly 134 contain various pathogenic or opportunistic pathogenic bacteria (e.g. Neisseria and Actinomyces). In 135 community type III, the gut microbiota was dominated by Pseudomonas, implying a severe dysbiosis. 136 We also used the gut community types I-III to reflect the progressive worsening of the microbiome. A shift of the gut microbiome from the lower-diversity community type (II or III) towards a 138 higher-diversity type (I or II) was observed over time in 8/10 patients who had anal swabs at different 139 time points (Fig. 2e) . Accompanied with the shift, a clear trend of increase of bacterial diversity and 140 the relative abundance of beneficial commensals (e.g. Bacteroides and Faecalibacterium) was 141 observed in the gut microbiota from early to late stages of COVID-19 ( Supplementary Fig. S3) , 142 indicating the restoration of gut microbiota. Only one patient had a reverse shift from higher-diversity 143 type II to lower-diversity community type III. Association between the respiratory and gut microbiomes in COVID-19 146 Most paired throat and anal swabs showed the same or similar community type levels (Fig. 3a) . In particular, the direction of shift over time of the microbiome appeared to match between the throat 148 and the gut in 7/8 patients who had two or more paired specimens at different time points (Fig. 3a) . Synchronous improvements of both the respiratory and gut microbiomes from early lower-diversity 150 community type towards late higher-diversity type occurred in six patients (p05, p17, p13, p11, p25 151 and p29). One patient (p33) showed an improved respiratory microbiome but a stable gut community 152 type on day 24. Only one case (p07) had a worsen gut microbiome from day 24 to day 35 but remained 153 a stable respiratory community type. These results suggested that the microbial community dynamics 154 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 22, 2020. . https://doi.org/10.1101/2020.07.21.20158758 doi: medRxiv preprint of either niches may reflect disease development and flora restoration. Because of no available anal 155 specimens, we were unable to assess whether the gut microbiota, like the respiratory microbiota, 156 shifted from higher-diversity type to lower-diversity type over time in this severe case (p09) (Fig. 1e ). Except for the duration of COVID-19, the respiratory and gut microbial community divergence 158 seemed not to be significantly correlated with age, gender, antibiotics use, and detection of SARS-159 CoV-2 RNA (Supplementary Figs. S4 and S5). 160 We further assessed the changes of the relative abundance of several representative bacteria, To further investigate the association between the respiratory and gut microbiomes, we 181 performed co-occurrence network analysis using paired specimens from 13 patients. Significantly CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 22, 2020. . https://doi.org/10.1101/2020.07.21.20158758 doi: medRxiv preprint between different community types (Fig. 4 ). There were four patterns of co-occurrence networks 185 (Throat-IV, Throat-III, II-II and I-I) that had been identified, and only bacteria from the same 186 community types formed significant co-occurrence networks regardless of whether they were in the 187 throat or gut. Under the dysbiosis condition, a few dominant pathogenic bacteria (e.g. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 22, 2020. . https://doi.org/10.1101/2020.07.21.20158758 doi: medRxiv preprint Whether SARS-CoV-2 infection alters microbiota to affect COVID-19 disease progression is an 201 important question that needs answers. In this study, we made three major observations. First, the 202 respiratory and gut microbiota compositions of COVID-19 adults can be characterized by four (I-IV) 203 and three (I-III) community types, respectively, and these types reflect different levels of balance 204 between the near-normal microbiota (type I) and dysbiosis (type III/IV). Second, lower-diversity 205 community types III/IV often appears in the early phase of COVID-19, and a consistent pattern of CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2020. . https://doi.org/10.1101/2020.07.21.20158758 doi: medRxiv preprint was relatively late in that study (about 14 days after symptom onset), therefore unable to determine 230 whether the baseline microbiome status is a consequence of SARS-CoV-2 infection, or a cause of 231 disease severity. We also observed alterations of the gut microbiota during COVID-19 in adults, and 232 found some pathogenic bacteria (e.g. Streptococcus, Rothia, Veillonella, Actinomyces, Bacteroides 233 and Actinomyces) similar to those reported in the previous observations 13, 14 . However，distinct from 234 the previous studies, three community types I-III were identified to characterize the changes of gut 235 microbiome over time, and low-diversity community types II and III often appeared in early 236 specimens, supporting the early effect of SARS-CoV-2 on the microbiome. A fast restoration with 237 the community type shifted from low-diversity type II to high-diversity type I over time was observed 238 in at least 4 patients. Moreover, the Pseudomonas-dominated community type III appeared in earlier 239 specimens of patients, and showed a slow improvement towards community II in three patients. Of Gut microbiota plays an important role in human health by shaping local immunity and 258 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2020. . https://doi.org/10.1101/2020.07.21.20158758 doi: medRxiv preprint remodeling mucosal tissues 37 . It is relatively more stable and plastic than the respiratory microbiota, 259 and it may affect the latter by cross-talk between these two organs along the airway-gut axis 34,36 , as We noted that Pseudomonas-dominated bacterial community type III was difficult to restore CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2020. . https://doi.org/10.1101/2020.07.21.20158758 doi: medRxiv preprint communities in both organs by mass replacement of pathogenic bacteria (e.g. Rothia and Neisseria) 289 in the type I stage ( Fig. 3b and Supplementary Fig. S7 ). However, a progressively worsening in the 290 respiratory and gut microbiome might be associated with severe cases of COVID-19. 291 In summary, we revealed for the first time the associations between the respiratory and gut 292 microbiota and COVID-19 disease progression, and observed early dysbiosis towards later restoration 293 to near-normal microbiota in a proportion of adults with mild COVID-19. In the absence of specific CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2020. subsampled at an even depth of 4700 and 3000 sequences per sample, respectively. The ASV coverage 371 of 82.6% (gut) and 77.2% (throat) were sufficient to capture microbial diversity of both sites. "The ANOSIM statistic "R" compares the mean of ranked dissimilarities between groups to the mean 386 of ranked dissimilarities within groups. An R value close to "1.0" indicates dissimilarity between 387 groups, whereas an R value close to "0" indicates an even distribution of high and low ranks within 388 and between groups". The ANOSIM statistic R always ranges between −1 to 1. The positive R values 389 closer to 1 suggest more similarity within sites than between sites, and that close to 0 represent no 390 difference between sites or within sites 56 . ANOSIM p values that are lower than 0.05 imply a higher 391 similarity within sites. Richness (Observed OTUs/ASVs) and Pielou's or Species evenness for each 392 community type were calculated for estimating the difference of alpha diversity. The analyses of alpha 393 diversity, NMDS and ANOSIM were performed using R package "vegan" v2.5-6. Dynamic change 394 of community types was showed according to collected dates of specimens with R package 'P 395 heatmap'. For association between community types and potential confounding factors such as sex, 396 age, virus existence and antibiotic use, the fisher exact test based on sample count was performed and 397 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2020. . https://doi.org/10.1101/2020.07.21.20158758 doi: medRxiv preprint the association with FDR-corrected p value <0.05 was considered significant. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2020. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2020. . is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2020. Ⅰ Ⅱ Ⅲ Ⅲ Ⅰ Ⅱ Ⅲ Ⅰ Ⅱ Ⅲ Ⅰ Ⅱ Ⅲ Ⅰ Ⅱ Ⅲ Days after symptom onset 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China Gut microbiota and Covid-19-possible link and implications The influence of the microbiome on respiratory 516 health COVID-19: towards understanding of pathogenesis Microbes, metabolites, and the gut-lung axis Gut microbiota, metabolites and host immunity Cross talk between neutrophils and the microbiota COVID-19: treating and managing severe cases Epidemiological and clinical characteristics of COVID-19 patients in Nantong A comprehensive benchmarking study of protocols and sequencing platforms for 16S 529 rRNA community profiling 16S rRNA gene amplicon sequencing library preparation method for the Illumina HiSeq platform UPARSE: highly accurate OTU sequences from microbial amplicon reads Best practices for analysing microbiomes DADA2: High-resolution sample inference from Illumina amplicon data Deblur rapidly resolves single-nucleotide community sequence patterns. mSystems QIIME allows analysis of high-throughput community sequencing data Naïve Bayesian Classifier for Rapid Assignment of rRNA 544 Sequences into the New Bacterial Taxonomy Ribosomal Database Project: data and tools for high throughput rRNA analysis Dirichlet Multinomial Mixtures: Generative Models for Microbial