key: cord-0273960-068qqvcf authors: Bikel, Shirley; López-Leal, Gamaliel; Cornejo-Granados, Fernanda; Gallardo-Becerra, Luigui; Sánchez, Filiberto; Equihua-Medina, Edgar; Ochoa-Romo, Juan Pablo; López-Contreras, Blanca Estela; Canizales-Quinteros, Samuel; Leyva, Adrian Ochoa title: Gut Phageome Analysis Reveals Disease-Specific Hallmarks in Childhood Obesity date: 2020-07-29 journal: bioRxiv DOI: 10.1101/2020.07.29.227637 sha: b9c94ce8f157ed968478876d89a3a800bb0c4512 doc_id: 273960 cord_uid: 068qqvcf Changes in the composition of the human gut microbiome are recognized to have a significant association with obesity and metabolic syndrome. Mexico leads worldwide childhood-obesity rankings representing an epidemic problem for public health. To this date, it is still unclear how the gut phageome, the bacteriophage component of the virome, influences childhood obesity and obesity with metabolic syndrome. We characterized the gut phageome of 28 school-age children with healthy normal-weight (NW), obese (O), and obese with metabolic syndrome (OMS) profiles, using metagenomic sequencing of virus-like particles (VLPs) from fecal samples. Viromes derived from VLPs were mainly dominated by Caudovirales, and only Inoviridae family was significantly increased in obesity. The three groups showed a similar number of VLPs, while a significant increase in phage richness and diversity was found in obesity groups compared NW. Few phage contigs dominated the phageome composition in NW, being increased in obesity groups. Interestingly, the majority of the phageome was shared among all individuals, establishing a core and common phageome, which abundances correlated with anthropometric and biochemical traits and bacteria previously associated with obesity and metabolic syndrome. We also established a healthy core phageome shared in >80% of NW samples, with a decreased prevalence in the O and OMS groups. Our data support that changes in the gut phageome may contribute to obesity and metabolic syndrome development via bacterial dysbiosis. We consider the phageome characterization provides the basis for novel diagnostic and therapeutic strategies for managing obesity and preventing metabolic syndrome development in obese children through potential phage manipulation. To the best of our knowledge, this study represents the most in-depth sequenced dataset of human bacteriophages, demonstrating for the first time that alterations of the gut phageome characterize obesity. We used the feces collected from 28 children, ten healthy normal-weight (NW), ten obese (O), and eight obese with metabolic syndrome (OMS) children, aged 7-10 years old and paired by gender and age based on previously collected data 1 8 (Extended Table 1 there was no significant difference in the number of VLPs among groups, obtaining an average of 1.42x10 9 ± 1.65 x10 9 , 4.88x10 9 ± 5.34x10 9 and 3.51x10 9 ± 4.22x10 9 VLP's for NW, O and OMS, respectively (Fig. 1A) . Unlike previous studies, DNA extracted from VLPs was directly sequenced without prior whole-genome amplification to avoid the typical PCR amplification biases. After quality controls, we obtained an average of 4,871,075 paired-end sequences per sample (Extended Table 2 ). These 11.23 Gb of data, represent the most in-depth sequenced dataset of human bacteriophages to date (Extended Table 3 ). To only select the sequences derived from VLPs, the reads mapped to bacterial (average of 28%), and human (average of 15%) genomes were removed for further analysis (Extended Table 2 ). Although, the removal of potential bacterial contamination risks also removing viral reads from a prophage state, we preferred to avoid potential bacterial DNA contamination. As a result, we obtained an average of 2,673,548 quality-filtered sequences per sample with no significant sequence depth differences among the three groups (Extended Fig. 1 and Extended Table 2 ). The annotation of the viral reads to the KEGG database showed that 96.2±11.89% mapped to genes with unknown function (Fig. 1B) , which is in agreement with a previous virome report 32 , supporting the viral enrichment in our VLPs purification. To assess the potential viral richness before assigning a taxonomical classification, we conducted 1,000 randomly subsampled exercises of 149,000 viral reads (according to the smallest sample) and clustered them at 95% identity to generate unique clusters of sequences for each sample. These rarefactions at the same sequence depth showed a significant increase (p-value <0.0001) of unique clusters in OMS and O, followed by NW ( Fig. 1C ). This increased viral-reads richness in obesity groups compared to healthy normal-weight was not dependent on sequencing depth or taxonomy classification. All the sequences for each group were clustered at 95% identity for each sample to remove redundant reads and to generate "unique" viral sequences. After that, we obtained a reduction of 68% ± 8%, resulting in an average of 856,825 unique sequences per sample (Extended Table 2 ). After mapping, only 2.95±0.95% of the unique reads matched against the viral NR Refseq protein database (Extended Fig. 2A showed no significant differences between these families among the three groups (Extended Fig. 2B ). Considering that the variable length of the phage genomes can affect their real abundance 32 , we assembled the potential viral genomes for further in-depth sequence analysis. We used the viral reads for de Novo assembly of the human gut virome. We identified the statistically differential abundant phages using edgeR with the recruitment matrix mentioned in the previous section. This procedure is analogous to using (Fig. 3) . The shared phage contigs belonged to Caudovirales (90%), followed by non-determined phages (6%), and Inoviridae (2%). The shared phage contigs in both two obesity groups suggest they could be associated with the core of obesity disease. It would be interesting to validate this assumption by analyzing these shared phages at functional level 36 , as well as to analyze the functional profile of the over-abundant unique phages in O and OMS groups. To assess the relationship within-sample and between-sample phageome diversity, we examined α -diversity and β -diversity, respectively. The α -diversity metrics showed that phage richness and diversity significantly increased in the obesity groups compared to NW ( Fig. 4A and B) . The O group exhibited higher richness followed by OMS and NW (Fig. 4B ), while the OMS group exhibited higher diversity followed by O and NW (Fig. 4A) . The between-sample diversity comparison observed in a principal component analysis (PCoA) based on Bray-Curtis distances showed no defined clusters by group (Fig. 4C) . Interestingly, when all the obesity samples were tagged together (O + OMS), obesity samples tend to cluster separately from NW samples (Fig. 4D ). We observed that 488 (10.58%) phage contigs accounted for 70% of the normalized reads in the NW samples, while 679 (14.73%) and 831 (18.02%) phage contigs, accounted the 70% of the normalized reads in O and OMS groups. This is in accordance with the higher diversity observed in the obesity groups and suggested a considerable change in the number of dominant phage contigs in the gut microbiome, being the highest is obesity with metabolic syndrome, followed by obesity and normal weight. The recruitment matrix was analyzed to assess the composition of the phageome in all the samples independently of the health status. From the 4,611 phage contigs, 2,605 (56.50%) were present in 20-50% (termed "common phages") and 1,088 (23.56%) were present in >50% individuals (termed "core phages") and two (0.04%) were present in all individuals. Contrary, 847 (18.37%) phage contigs were detected in 2-19% of individuals (termed "low overlap phages"), and 71 (1.54%) in only one individual (termed "unique phages") ( Fig. 5A) . Interestingly, two of the "core phages" were identified as putative "crass-like" phages. The identification of "crass-like" phages supports our phageome definition of "core phages", due to these are the most abundant phages reported in the human gut from adult individuals 34 . We The Spearman correlation between the 48 phage contigs present in ≥ 80% of all the samples and the anthropometrical and biochemical parameters showed positive correlations between the abundance of several phage contigs (Extended Fig. 3B ) and LDL levels, total cholesterol, glucose, waist circumference, BMI, and weight (Fig. 6B ). In contrast, we also found a negative correlation between the abundance of several phage contigs (Extended Fig. 3B ) and BMI, HDL levels, and triglycerides (Fig. 6B) . These correlations suggest an association between phage contigs and typical host affected anthropometrical and clinical parameters by obesity. To analyze if the NW core phageome was related to a good health status, we compared the prevalence of core phage contigs in NW children (n = 10) to those in O (n= 10) and OMS (n = 8). To this end we selected the phage contigs shared among >80% of the NW samples. Therefore, we established a NW core phageome of 52 phage contigs, and compared its prevalence with respect to O and OMS groups. NW core phageome was found in an average of 91.54% NW individuals, and it was significantly reduced to 76.35% (p-value= <0.0001) and 68.27% (p-value= <0.0001) on average in O and OMS patients, respectively (Fig. 5B) . These results show the gut core phageome of normal-weight children was significantly altered in children with obesity, being most affected in in children with obesity with metabolic syndrome. Microbiome changes are widely accepted to have a significant association with human health, while metagenomic analysis of viruses, one of the most poorly understood components of the human gut microbiome, has recently revolutionized our view of gut microbiome 20,38-40 , highlighting the critical role of the interrelationships between gut phages and bacteria in health and disease. We report a viral metagenome with one of the largest 37 sequencing depths, ~75M of viral reads, analyzing the phageome of school-age children with obesity and obesity with metabolic syndrome and comparing to normal-weight controls. Instead of solely using the sequencing reads, we used a de novo approach to construct the phageome community, We detected a low inter-individual variability of phageome. Only ~20% of phages were low-overlap or unique, while ~80% represented common and core phages, independently of the disease. In this regard, we also identified that the prevalence of the core phageome that was shared in >80% of the normal-weight individuals was significantly reduced in obesity, an even more in OMS. This suggests that the loss of some normalweight phage contigs in obesity could play a role into encouraging the children's progression to obesity. Our observation is in agreement with previous studies where a core phageome was proposed for a set of healthy humans and; also, it was reduced in patients with inflammatory bowel disease 37 . The concept of a core virome has also received support from a recent study with adult monozygotic twins, in which 18 contigs were found to be present in all individuals 32 . In contrast to these findings, the compilation of a large-scale gut virome database called into question of the existence of a human core gut virome 19,50 . This disparity is mainly due to the well-established belief that the human gut virome is highly individual- Also, the increased abundance of phage contigs207 and 540 in NW was associated with a decreased abundance of Erysipelotrichaceae and BMI. This bacterial family was significantly increased in OMS and positively correlated with waist circumference 18 . Moreover, the increased abundance of this family has been associated with host dyslipidemia in the context of obesity, metabolic syndrome, and hypercholesterolemia 56 Compelling evidence supports the concept that shifts in the microbial system during infancy may increase the risk for obesity later in life 58, 59 . Therefore, manipulating gut microbiota using phages at an early stage might offer possibilities for the prevention and treatment of the dysbiosis associated with the obesity. Fecal microbiota transplantation (FMT) revealed that phages could be co-transferred with bacteria 60 . Even, successful treatments against Clostridium difficile using bacteria-free fecal filtrate provided the first evidence that phageome manipulation may be an effective therapeutic strategy to stabilize the eubiosis of the bacteria in the microbiome 61 . The role that phages play in obesity and obesity with metabolic syndrome is critical to understanding their contribution to the microbiota dysbiosis. We believe that our study provides a better knowledge of the phage-bacteria dynamics of the gut microbiome. We propose studying specific overabundant phages in O and OMS as possible biomarkers related to the development of metabolic syndrome in obese children. Also, the development of in vivo models to test the phage-bacteria dynamics in obesity will undoubtedly be an essential area that could help to complement the understanding of microbiome dysbiosis associated with obesity. A better knowledge of the intestinal phageome composition and its interaction with the microbiota and immune host system will allow us to diminish the prevalence of childhood obesity and metabolic syndrome. Fecal samples and the biochemical parameters information used in this study were Whole-Virome Analysis Sheds Light on Viral Dark Matter in Inflammatory Bowel Disease Biology and Taxonomy of crAss-like Bacteriophages, the Most Abundant Virus in the Human Gut Benchmarking viromics: an in silico evaluation of metagenome-enabled estimates of viral community composition and diversity Gnotobiotic mouse model of phage-bacterial host dynamics in the human gut Parabacteroides distasonis Alleviates Obesity and Metabolic Dysfunctions via Production of Succinate and Secondary Bile Acids Association between composition of the human gastrointestinal microbiome and development of fatty liver with choline deficiency Clostridium difficile phages: Still difficult? Differences in gut microbiota composition between obese and lean children: A cross-sectional study The intestinal microbiota composition and weight development in children: The KOALA Birth Cohort Study Long-term colonisation with donor bacteriophages following successful faecal microbial transplantation Efficacy of Sterile Fecal Filtrate Transfer for Treating Patients With Clostridium difficile Infection Fast and accurate long-read alignment with Burrows-Wheeler transform Ultrafast metagenomic sequence classification using exact alignments MEGAN analysis of metagenomic data IDBA-UD: A de novo assembler for single-cell and metagenomic sequencing data with highly uneven depth QIIME allows analysis of high-throughput community sequencing data We thank Juan Manuel Hurtado Ramírez for the informatics technical support and to Abigail Hernández-Reyna for his helpful support on molecular biology protocols. We thank the National Laboratory Advanced Microscopy in the Biotechnology Institute (IBt) at UNAM for help with epifluorescence microscopy technical support and Dr. Guadalupe