Limited intestinal inflammation despite diarrhea, fecal viral RNA and SARS-CoV-2-specific IgA in patients with acute COVID-19

Abstract

Gastrointestinal symptoms are common in COVID-19 patients but the nature of the gut immune response to SARS-CoV-2 remains poorly characterized, partly due to the difficulty of obtaining biopsy specimens from infected individuals. In lieu of tissue samples, we measured cytokines, inflammatory markers, viral RNA, microbiome composition, and antibody responses in stool samples from a cohort of 44 hospitalized COVID-19 patients. SARS-CoV-2 RNA was detected in stool of 41% of patients and more frequently in patients with diarrhea. Patients who survived had lower fecal viral RNA than those who died. Strains isolated from stool and nasopharynx of an individual were the same. Compared to uninfected controls, COVID-19 patients had higher fecal levels of IL-8 and lower levels of fecal IL-10. Stool IL-23 was higher in patients with more severe COVID-19 disease, and we found evidence of intestinal virus-specific IgA responses associated with more severe disease. We provide evidence for an ongoing humeral immune response to SARS-CoV-2 in the gastrointestinal tract, but little evidence of overt inflammation.

Figure 1

Gastrointestinal (GI) symptoms and serologic parameters in hospitalized COVID-19 patients. (A) Proportion of patients with GI manifestations in this cohort of hospitalized COVID-19 patients in whom stool samples obtained. (B) Serum concentrations of IL-1β, IL-6, IL-8, and TNFα and select laboratory values at the time of admission in patients with and without GI symptoms and (C) in patients with severe compared with non-severe COVID-19 disease. Each point represents an individual value for a patient, the box plot represents the median and the interquartile range and the p values calculated using the Mann–Whitney test with significance defined as p < 0.05.

Figure 2

Detection and sequencing of SARS-CoV-2 virus genome in stool. (A) The median fecal SARS-CoV-2 Ct of each sample, plotted according to the time of collection relative to either the onset of symptoms or the first positive nasal PCR. p value—Kruskal–Wallis. (B) The median fecal SARS-CoV-2 Ct of each sample, plotted according to the relative time of collection: day 28 as a cut off. Relative to either the onset of symptoms or the first positive nasal PCR. p value—Mann–Whitney. (C) The lowest fecal SARS-CoV-2 Ct measured for each donor shows higher peak viral loads in donors who did not survive. p value—Mann–Whitney. (D) The lowest fecal SARS-CoV-2 Ct value measured from each donor during admission, plotted according to the peak severity of COVID19 symptoms—p value—Kruskal–Wallis. (E) Fecal SARS-CoV-2 Ct values in samples from donors with and without diarrhea. p value—Mann–Whitney. (F) The relationship between the fecal SARS-CoV-2 Ct values and the Bristol Score of each sample. Spearman rank test. (G) Phylogenetic relationships of the SARS-CoV-2 isolates from stool (Stool, dark red) and other nasopharyngeal virus isolates from the Mount Sinai Health System (MSHS) in NYC (Nasopharyngeal, salmon), in a background of global strains (grey). Main phylogenetic clades are indicated. The scale bar represents divergence from the root (reference genome NC_045512) as number of single nucleotide variants (SNVs). The panel in the bottom right indicates the clade distribution of the MSHS nasopharyngeal samples from 403 SARS-CoV-2 genomes from February 29 to July 15. (H) Sequence alignment depicting clade and cluster defining (teal), and other (grey) SNVs present in the stool SARS-CoV-2 isolates from stool (S) or nasopharynx (N) of the indicated donors. The clade and amino acid substitutions are indicated when applicable.

Figure 3

Fecal cytokine levels in COVID-19 patients. (A) Concentrations of cytokines in fecal samples from control donors and COVID-19 patients. Where multiple samples from the same donor were collected, the median of these samples is plotted. (B) Concentrations of fecal cytokines in COVID-19 patients with severe disease compared to those with non-severe disease. (C) Concentrations of fecal calprotectin in COVID-19 patients with severe disease compared to those with non-severe disease. (D) Concentrations of fecal calprotectin are unrelated to fecal viral load. Boxplots show median + / − IQR and p values are calculated by Mann–Whitney test.

Figure 4

Gut microbiome and fecal SARS-CoV-2 genome sequencing of COVID-19 patients. (A) PCoA (Bray–Curtis dissimilarity) of 16S rRNA gene amplicon and shotgun metagenomic sequencing of fecal samples from donors stratified by COVID-19 severity. (B) Alpha diversity (Shannon entropy) is not significantly altered in mild, moderate or severe COVID-19. (C) PCoA (Bray–Curtis dissimilarity) of 16S rRNA gene amplicon and shotgun metagenomic sequencing of fecal samples from donors with and without diarrhea. (D) Gut microbiome alpha diversity (Shannon entropy) in patients with and without diarrhea and as relates to the Bristol score of each sample. (E) PCoA (Bray–Curtis dissimilarity) of 16S rRNA gene amplicon and shotgun metagenomic sequencing of fecal samples from patients receiving or not receiving antibiotic therapy. (F). Gut microbiome alpha diversity (Shannon entropy) is reduced in hospitalized COVID-19 patients recently or currently receiving antibiotic therapy (Abx). Boxplots show median + / − IQR.

Figure 5

SARS-CoV-2-specific IgA in stool of COVID-19 patients. (A) Fecal and serum anti-SARS-CoV-2 RBD IgA are correlated. (B, C) Relative titers of anti-SARS-CoV-2 S1 RBD IgA in feces of control and COVID infected individuals. (D) Relative titers of anti-SARS-CoV-2 S1 RBD IgA in feces of COVID infected individuals with varying disease severity. (E) Anti-SARS-CoV-2 S1 RBD IgA and viral load are not correlated. Each point shows data from a single sample. (F) Anti-SARS-CoV-2 S1 RBD IgA and anti-S1 IgA in feces are correlated. Each point shows data from a single sample.