Abstract

   The impact of coronavirus disease 2019 (COVID-19) history on Crohn’s
   disease (CD) is unknown. This investigation aimed to examine the effect
   of COVID-19 history on the disease course, oral-gut microbiota, and
   serum metabolomics in patients with CD. In this study, oral-gut
   microbiota and serum metabolomic profiles in 30 patients with CD and a
   history of mild COVID-19 (positive group, PG), 30 patients with CD
   without COVID-19 history (negative group, NG), and 60 healthy controls
   (HC) were assessed using 16S rDNA sequencing and targeted metabolomics.
   During follow-up, the CD activity index showed a stronger decrease in
   the PG than in the NG (p = 0.0496). PG patients demonstrated higher
   α-diversity and distinct β-diversity clustering in both salivary and
   fecal microbiota compared to NG and HC individuals. Notably, the gut
   microbiota composition in the PG patients showed a significantly
   greater similarity to that of HC than NG individuals. The interaction
   between oral and intestinal microbiota in the PG was reduced. Moreover,
   serum metabolome analysis revealed significantly increased
   anti-inflammatory metabolites, including short-chain fatty acids and
   N-Acetylserotonin, among PG patients; meanwhile, inflammation-related
   metabolites such as arachidonic acid were significantly reduced in this
   group. Our data suggest that the gut microbiota mediates a potential
   beneficial effect of a mild COVID-19 history in CD patients.

   Keywords: Crohn’s disease, COVID-19, oral microbiota, gut microbiota,
   metabolome, inflammatory bowel disease

1. Introduction

   The global coronavirus disease (COVID-19) pandemic, caused by severe
   acute respiratory syndrome coronavirus 2 (SARS-CoV-2) outbreaks in
   2019, has substantially impacted global health [[40]1]. While
   vaccination and public health control measures effectively manage its
   spread, eliminating the virus is challenging. It appears that
   SARS-CoV-2 will remain alongside humans for an extended period. Amid
   the expanding COVID-19 pandemic, there is increasing public concern
   regarding its impact on individuals with chronic conditions, such as
   Crohn’s disease (CD). CD, a subtype of inflammatory bowel disease
   (IBD), is a chronic, recurrent, and incurable inflammatory condition
   that affects the digestive tract, often accompanied by extraintestinal
   manifestations [[41]2,[42]3]. The occurrence of COVID-19 in patients
   with IBD is comparable to that in the general population [[43]4].
   SARS-CoV-2, detectable in stools even after respiratory samples test
   negative, raises concerns about the possible long-term gastrointestinal
   impact [[44]5]. Many researchers have hypothesized, based on
   theoretical mechanisms and pathway theories, that COVID-19 could act as
   a risk factor for either the onset or relapse of IBD. These hypotheses
   are rooted in the inflammatory and immune responses triggered by the
   virus, which could theoretically exacerbate existing IBD conditions or
   unmask new cases [[45]6,[46]7]. However, despite these theories,
   evidence remains lacking to definitively support the notion that
   COVID-19 directly leads to the onset or exacerbation of IBD. Therefore,
   further research is required to investigate the effect of COVID-19 on
   individuals with CD.

   Alterations in the oral and gut microbiota composition in patients with
   COVID-19 were observed at diagnosis, during the disease course, and in
   the follow-up period, in contrast to individuals without COVID-19
   [[47]8,[48]9]. Significantly decreased α diversity was noted in the
   oral and gut microbiota of patients with COVID-19 [[49]8]. Compared
   with healthy controls (HC), patients with COVID-19 exhibited a
   reduction in butyric acid-producing bacteria and an increase in
   lipopolysaccharide-producing bacteria [[50]9]. Compared with seasonal
   flu patients, those with COVID-19 displayed a significant increase in
   opportunistic pathogens, such as Streptococcus and Fusobacterium
   [[51]10]. Furthermore, compared to patients with H1N1 influenza
   infection, those with COVID-19 showed a significantly higher level of
   Actinobacteria and Firmicutes phyla [[52]11]. The abnormal alteration
   of oral-gut microbiota persisted for up to 1 year post SARS-CoV-2
   [[53]8] infection. COVID-19 typically lasts only a few days or weeks;
   nevertheless, its impact on the microbiota is more enduring. The
   crucial role of the oral gut microbiota and their metabolites in
   developing, progressing, and treating CD has been convincingly
   demonstrated. In cases of IBD, the gut microbiota is characterized by
   increased instability, reduced microbial diversity, elevated levels of
   Proteobacteria taxa, and decreased Firmicutes [[54]12]. Notably,
   ectopic gut colonization by oral microbes is a distinctive feature of
   IBD, and the abnormal ectopic location of oral pathogens in the
   intestine exacerbates gastrointestinal inflammation [[55]13,[56]14].
   Host metabolism is significantly affected by COVID-19. Lipid
   alterations in recovered patients of COVID-19 have been observed. In
   these patients, 122 lipid molecules exhibited an increase, while 47
   lipid molecules showed a decrease [[57]9]. Sphingosine-1-phosphate
   (S1P), closely tied to the inflammatory storm in COVID-19, had lower
   levels, indicating a worse prognosis [[58]8]. S1P reached its lowest
   level during the infection period and rose gradually during recovery
   [[59]8]. However, the metabolites and oral-gut microbiota composition
   in vulnerable patients with CD post-COVID-19 remain unclear.

   Thus, we aimed to investigate the impact of COVID-19 history on the
   disease course, the oral-gut microbiota, and metabolites in patients
   with CD. Our data indicate an interplay between the metabolome and
   microbiota that may influence the disease course of CD in patients with
   a history of COVID-19. These results suggest that the gut microbiota
   mediates a potential beneficial effect of COVID-19 history in patients
   with CD.

2. Materials and Methods

2.1. Study Design

   This study continuously enrolled 60 CD patients who visited the Sixth
   Affiliated Hospital of Sun Yat-sen University from November 2022 to May
   2023 and consented to participate, and 60 healthy individuals who
   participated from June 2018 to November 2019. Fecal and saliva samples
   were collected from all participants. Based on their history of
   COVID-19, patients with CD were divided into two groups: 30 with a
   history of mild infection (positive group, PG) and 30 without a history
   of infection (negative group, NG). The CD activity index (CDAI) was
   assessed by clinical experts during sample collection. The patients in
   the PG and NG groups were further classified into subgroups: the active
   group (CDAI ≥ 150), which encompassed mild (CDAI = 150–220), moderate
   (CDAI = 221–450), and severe (CDAI > 450) disease activity, and the
   remission group (CDAI < 150) [[60]15]. Demographic and clinical data
   such as sex, age, body mass index (BMI), medication use, and
   location/behavior were obtained from medical records. In addition, the
   CDAI of the patients with CD was assessed again 6 and 9 months after
   sample collection by clinical experts.

2.2. Eligibility Criteria

   The inclusion criteria were the following: (1) CD diagnosed by
   experienced clinicians based on clinical, pathological, endoscopic,
   laboratory, and radiological features [[61]2,[62]3]; (2) COVID-19
   diagnosis determined based on the positive PCR or antigen test
   [[63]16,[64]17]; (3) HC free of both CD and COVID-19; (4) age of all
   participants: 18–75 years; (5) PG participants all with mild COVID-19.
   Mild COVID-19 was defined as showing mild symptoms such as fever,
   cough, and sore throat, without pneumonia symptoms and pneumonia
   imaging evidence according to the WHO interim guidance (Clinical
   management of COVID-19: interim guidance, 27 May 2020); (6) For
   COVID-19 infection, the PG group only received symptomatic treatment
   such as antipyretics and cough suppressants.

   The exclusion criteria were the following: (1) use of antibiotics,
   probiotics, or proton pump inhibitors within 1 month before sample
   collection; (2) pregnancy; (3) gastrointestinal cancer, tuberculosis,
   or Behcet’s disease; (4) gastrointestinal resection.

2.3. Outcomes

   The outcomes of interest were the following: (1) clinical symptom
   improvement, i.e., changes in CDAI; (2) composition of the microbiota
   in the PG group compared to the HC and NG groups; (3) composition of
   serum metabolic groups in the PG group compared to the NG group.

2.4. Sample Collection

   Fresh fecal samples were collected in sterile tubes immediately after
   defecation, divided into aliquots (200 mg), and promptly stored at −80
   °C until needed. Saliva samples were collected in sterile containers,
   mixed with an equal volume of DNA stabilization buffer, and stored at
   −80 °C until use. The DNA stabilization buffer, totaling 250 mL,
   contained 1.461 g NaCl, 2.5 mL Tris HCl, 5 mL 0.5 M EDTA, 12.5 mL 10%
   SDS, 2.5 mL 20 mg/mL Proteinase K Solution, and 227.5 mL sterilized
   water. All participants were instructed to avoid oral hygiene
   procedures overnight and refrain from drinking or eating for at least 2
   h before saliva collection. Simultaneously, peripheral blood was
   harvested and centrifuged at 3000 rpm for 10 min, and the serum was
   divided into aliquots (100 μL) and stored at −80 °C.

2.5. DNA Extraction and High-Throughput Sequencing

   Microbial genomic DNA (gDNA) was extracted from samples using a stool
   DNA Kit (Magen, Guangzhou, China; cat. D3141B) according to the
   manufacturer’s instructions. Saliva DNA extraction was performed using
   a saliva DNA Kit (Magen, Guangzhou, China; cat. D3134). The V5–V6
   hypervariable region of bacterial 16S rDNA was amplified from gDNA
   using the forward primer 784F (5′-RGGATTAGATACCC-3′) and the reverse
   primer 1064R (5′-CGACRRCCATGCANCACCT-3′). Subsequently, the PCR
   products were purified and subjected to paired-end sequencing using a
   high-throughput sequencer (MGISEQ2000, BGI, Shenzhen, China) following
   standard protocols. The raw data were quality-filtered and spliced
   using FLASH (v1.2.11). The DADA2 in QIIME2 was utilized for denoising
   to obtain amplicon sequence variants (ASVs)—which are sequences with
   100% similarity—and to generate the feature table. For the alpha
   diversity analysis, we calculated the Chao index. For the beta
   diversity analysis, principal coordinate analysis (PCoA) was performed
   using the R language. The Wilcoxon test was utilized to assess
   significantly different bacteria, with a p-value and false discovery
   rate (FDR) < 0.05. Network correlation analysis was based on Spearman’s
   correlation (|correlation coefficient| > 0.6, p < 0.05) generated using
   R (version 4.3.1). The resulting networks were subsequently processed
   and visualized using Gephi software (version 0.9.2).

2.6. Metabolomic Data Generation and Analysis

   Serum samples were submitted to Metabo-Profile Biotechnology (Shanghai,
   China) Co., Ltd. for metabolomics detection. The quantification of
   targeted metabolites was performed using an ultra-performance liquid
   chromatography coupled with a tandem mass spectrometry (UPLC-MS/MS)
   system (ACQUITY UPLC-Xevo TQ-S, Waters Corp., Milford, MA, USA). All
   standards for the targeted metabolites were prepared at a concentration
   of 5.0 mg/mL. Dots exceeding ± 3 standard deviations in score values
   are generally considered potential outliers. If improper sample
   preparation procedures or injections are identified, outlying samples
   would be re-prepared and reanalyzed. Raw data were processed using TMBQ
   software (v1.0, Metabo-Profile, Shanghai, China). Statistical analyses,
   including univariate, multivariate, and pathway enrichment analyses,
   were conducted using the iMAP platform (v1.0, Metabo-Profile, Shanghai,
   China).

2.7. Statistical Analysis

   Descriptive data are presented as mean ± SD or median (25–75
   interquartile range). To assess differences in taxonomic abundance
   between groups, students’ t-tests or Wilcoxon tests were conducted.
   Chi-square and Fisher’s exact tests were performed using SPSS, version
   26.0.

3. Results

3.1. Participant Characteristics

   [65]Table 1 displays the demographic and clinical characteristics of
   the participants. No significant differences were observed in sex (HC:
   76.7% male, NG: 66.7% male, PG: 56.7% male) and age (HC: 21.8
   [19.56–24.5], NG: 30.37 ± 8.65, PG: 33.13 ± 7.80) among the three
   groups. Similarly, the PG and NG did not significantly differ in BMI
   (NG: 18.77 ± 3.35, PG: 20.24 ± 2.82), staging (NG: 73.3% active stage,
   PG: 73.3% active stage), lesion location (NG: 66.7% ileocolon, PG:
   56.7% ileocolon), and other factors. All individuals in the PG
   recovered from COVID-19 within 7–14 days, and the samples were
   collected 2–4 months after recovering from COVID-19. CDAI was assessed
   at the time of enrollment and again after 6 and 9 months; the results
   showed a more pronounced CDAI decline (ΔCDAI) in the PG than NG
   ([66]Figure 1A and [67]Figure S1).

Table 1.

   Participant clinical characteristics.
   Parameter HC NG PG p
   Sex (M/F) 34/26 20/10 23/7 ^a 0.567
   ^b 0.069
   ^c 0.494
   Age (years) 21.8
   (19.56–24.56) ^d 30.37 ± 8.65 ^e 33.13 ± 7.80 ^a 0.199
   ^b 0.543
   ^c 0.114
   BMI (kg/m^2) 22.25 ± 3.51 18.77 ± 3.35 20.24 ± 2.82 ^a 0.071
   ^b 0.008
   ^c 0.000
   Duration (years) - 3.32 ± 3.38 5.17 ± 4.65 0.08
   Smoking - 0 2 0.492
   Biologics -
   - 10 ^f
   12.00 ± 9.15 ^g 16
   13.00 ± 9.28 0.192
   0.79
   Immunosuppressants - 7
   15.83 ± 8.23 3
   18.00 ± 2.00 0.299
   0.676
   Glucocorticoids - 3
   6.00 ± 2.00 0
   - 0.237
   -
   Enteral nutritional agent - 3
   5.83 ± 7.97 6
   8.52 ± 15.26 0.472
   0.79
   5-ASA - 7
   6.83 ± 5.42 3
   2.00 ± 1.73 0.299
   1.88
   Traditional Chinese medicine - 0
   - 2
   10.00 ± 5.66 0.492
   -
   Non-use - 7 1 0.052
   CDAI < 150/CDAI ≥ 150 - 8/22 8/22 1.000
   Mild/Moderate/Severe 14/8/0 15/7/0 1.000
   Location (ileum/ileocolon/colon) - 9/20/1 13/17/0 0.422
   Stricturing - 2 6 0.254
   Penetrating - 6 5 1.000
   Stricturing and penetrating - 1 1 1.000
   Non-stricturing and non-penetrating - 21 18 0.589
   [68]Open in a new tab

   ^a PG vs. NG; ^b PG vs. HC; ^c NG vs. HC; ^d median (25–75
   interquartile range); ^e mean ± standard error of the mean; ^f number
   of patients who received the medication; ^g received treatment duration
   (month); BMI, body mass index; HC, healthy control group; NG, negative
   group; PG, positive group.

Figure 1.

   [69]Figure 1
   [70]Open in a new tab

   Recovery of clinical activities and microbial community. (A) CDAI
   changes from the initial enrollment to the 6-month follow-up. (B,C)
   Differences of α-diversities. (D,E) β-diversities calculated using
   UniFrac-based unweighted principal coordinate analysis (PCoA). (F)
   Relative abundance of bacterial phyla. (G) Bray Curtis distance of gut
   microbiota between HC and NG or PG. CDAI, Crohn’s disease activity
   index; HC, healthy control; NG, negative group; PG, positive group. * p
   < 0.05; ** p < 0.01; **** p < 0.0001.

3.2. Altered Oral and Gut Microbiota Composition in COVID-19-Infected
Patients with CD

   A total of 176 samples, consisting of 88 saliva and 88 fecal samples,
   underwent 16S rDNA gene sequencing, revealing 23 phyla, 39 classes, 73
   orders, 136 families, 424 genera, and 897 species. The diversity of
   microbial communities within (α diversity) and between (β diversity)
   samples in the three groups was analyzed at the ASV level. Alpha
   analyses revealed that the PG had significantly higher oral (HC: 127.0
   [98.0–165.8], NG: 102.0 [84.5–199.3], PG: 256.0 [227.8–305.8]; HC vs.
   PG: p < 0.0001, NG vs. PG: p < 0.0001) and fecal (HC: 155.0
   [111.3–209.8], NG: 89.5 [50.75–181.0], PG: 241.5 [18.3–277.3]; HC vs.
   PG: p < 0.0001, NG vs. PG: p < 0.0001) microbial diversity than the
   other groups based on the Chao index, indicating that COVID-19
   substantially altered the diversity and richness of the oral and gut
   microbiota ([71]Figure 1B,C). Regarding β diversity, principal
   coordinate analysis (PCoA) based on unweighted UniFrac distance was
   conducted to compare microbial communities among the three groups,
   revealing significant differences (saliva: p < 0.0001; feces: p <
   0.0001) ([72]Figure 1D,E). The lower β diversity indicated smaller
   heterogenicity in the microbial communities of the PG.

   Discernible shifts have been observed in gut microbiota compositions in
   the PG according to the stacked bar chart ([73]Figure 1F).
   Specifically, the gut microbiota in the PG was characterized by an
   increase in the Bacillota phylum and a decrease in the Fusobacteriota
   and Pseudomonadota phyla, approaching levels similar to those in the HC
   ([74]Figure 1F). At the genus level, the gut microbiota in the PG was
   characterized by an increase in Megamonas, Faecalibacterium,
   Akkermansia, and Phocaeicola, and a decrease in Fusobacterium,
   Bacteroides, and Veillonella, approaching levels similar to those in
   the HC ([75]Figure S2A). At the species level, the gut microbiota in
   the PG was characterized by an increase in Akkermansia muciniphila,
   Ruminococcoides bili, Megamonas funiformis, Phocaeicola dorei, and
   Faecalibacterium duncaniae, and a decrease in Fusobaccterium varium,
   Klebsiella pneumoniae, and Veillonella dispar, approaching levels
   similar to those in the HC ([76]Figure S2B). The Bray–Curtis distance
   between the PG and HC (0.81 [0.69–0.92]) was significantly lower than
   that between the NG and HC (0.76 [0.66–0.83]), indicating that the gut
   microbiota in the PG was more similar to that in the HC group than in
   the NG ([77]Figure 1G). In the gut, three genera that acted as
   probiotics—Bifidobacterium (NG: 0.48 ± 1.10%, PG: 1.08 ± 2.76%; p =
   0.0196 and FDR = 0.0491), Akkermansia (NG: 0.04 ± 0.10%, PG: 1.41 ±
   4.66%; p = 0.0001 and FDR < 0.0001), and Faecalibacterium (NG: 1.59 ±
   3.42%, PG: 4.44 ± 6.94%; p < 0.0001 and FDR < 0.0001)—were
   significantly enriched in the PG compared to the NG ([78]Figure 2A–C).
   Conversely, the opportunistic pathogens Klebsiella (NG: 3.56 ± 9.07%,
   PG: 1.09 ± 1.12%; p = 0.0005 and FDR = 0.0020) and Veillonella (NG:
   2.41 ± 7.58%, PG: 0.19 ± 0.29%; p < 0.0001 and FDR < 0.0001) were
   enriched in the NG ([79]Figure 2D,E).

Figure 2.

   [80]Figure 2
   [81]Open in a new tab

   Relative abundance of genera Bifidobacterium (A), Akkermansia (B),
   Faecalibacterium (C), Klebsiella (D), and Veillonella (E) in three
   groups.

   Additionally, in the remission and active subgroups, the α diversity of
   oral (HC vs. PG: p = 0.0008, NG vs. PG: p = 0.0221) and gut (HC vs. PG:
   p = 0.0034, NG vs. PG: p = 0.0082) microbiota was significantly
   increased in the PG based on the Chao index ([82]Figure S3A–D).
   Regarding β diversity, significant differences were observed among the
   groups (saliva in remission: p = 0.0084; feces in remission: p <
   0.0001; saliva in active: p < 0.0001; feces in active: p < 0.0001)
   ([83]Figure S3E–H). No significantly different genus was observed
   during the remission stage between the PG and NG. In the active stage,
   there were 31 different salivary genera and 39 different fecal genera
   between the PG and the NG. This suggested that COVID-19 had a more
   pronounced impact on the microbiota of patients with active CD.

3.3. Interaction between the Oral and Gut Microbial Communities

   Saliva and feces shared 86 ASVs in HC, 343 in NG, and 183 in PG
   ([84]Figure 3A,D,E). The Spearman’s correlation network depicting the
   association between the salivary and gut microbiota showed 6
   significant relationships in HC, 730 in NG, and 33 in PG ([85]Figure
   3B,C,F). These results indicate that taxa within the NG exhibited
   elaborate and robust interrelationships compared to the other groups,
   while the oral-gut axis showed signs of recovery in the PG towards HC.

Figure 3.

   [86]Figure 3
   [87]Open in a new tab

   Interaction between oral and gut microbiota. (A) Venn diagram
   illustrating ASVs of oral and gut microbiota in HC. (B) Spearman’s
   correlation network between oral and gut microbiota in HC. (C) Venn
   diagram illustrating ASVs of oral and gut microbiota in NG. (D)
   Spearman’s correlation network between oral and gut microbiota in NG.
   (E) Venn diagram illustrating ASVs of oral and gut microbiota in PG.
   (F) Spearman’s correlation network between oral and gut microbiota in
   PG. The red circle represents fecal microbiota, and the blue circle
   represents salivary microbiota. The size of the circles represents the
   quantity of significant correlation relationships. The red line
   represents positive correlation. The blue line represents negative
   correlation. F_, fecal microbiota; S_, salivary microbiota; HC, healthy
   control; NG, negative group; PG, positive group; ASV, amplicon sequence
   variants.

3.4. Metabolomic Changes in Participant Serum

   Based on the differences in oral and gut microbiota composition between
   the PG and NG groups, we further investigated the serum metabolomics in
   the two groups. A total of 521 metabolites were detected, and the
   composition of their categories is shown in [88]Figure S4: organic
   acids had the highest proportion (32.88%), followed by amino acids
   (23.27%), cholesterol esters (CE, 21.17%), triglycerides (TAG, 13.78%),
   and others. The relative abundance of each metabolite class in the PG
   and NG is illustrated in the stacked bar chart ([89]Figure 4A). There
   were significant differences in amino acids (NG: 23.43%, PG: 23.08%),
   short-chain fatty acids (SCFAs) (NG: 0.67%, PG: 0.90%), carbohydrates
   (NG: 0.244%, PG: 0.238%), peptides (NG: 0.007%, PG: 0.006%), and
   secondary bile acids (NG: 0.001%, PG: 0.003%) between the PG and NG.
   Forty-three potential biomarkers were selected by combining the
   differential metabolites from univariate and multi-dimensional
   statistics, with a threshold value of variable importance in projection
   (VIP) >1 in multi-dimensional statistics and p < 0.05 and |log2FC| ≥ 0
   in univariate statistics ([90]Figure 4B). Among the 43 potential
   biomarkers, only 8 were increased in the PG, while the other 35 were
   decreased ([91]Figure 4B). Notably, inflammation-related metabolites,
   including arachidonic acid (NG: 3.643 [3.212–5.347], PG: 3.332
   [2.427–4.211], p = 0.0162), aspartic acid (NG: 31.68 [26.03–50.56], PG:
   23.29 [19.04–27.74], p = 0.0001), serine (NG: 162.2 ± 35.66, PG: 135.5
   ± 23.64, p = 0.0012), and pyroglutamic acid (NG: 72.83 [56.85–91.16],
   PG: 60.69 [48.70–71.75], p = 0.0415), were significantly reduced in PG,
   while N-acetylserotonin (NG: 0.084 [0.0061–0.1150], PG: 0.096
   [0.0061–0.1408], p = 0.0321) and acetic acid (NG: 78.47 [68.48–90.34],
   PG: 100.7 [75.29–119.9], p = 0.0036), as anti-inflammatory metabolites,
   were significantly increased in PG ([92]Figure 5A–F). Additionally, a
   substantial quantity of amino acids, including lysine, arginine,
   ornithine, glutamic acid, alanine, threonine, creatine, asparagine,
   phenylalanine, and glycine, were markedly decreased in the PG
   ([93]Figure 4B). Pathway enrichment analysis was conducted using
   pathway-associated metabolite sets. Metabolic pathways were notably
   enriched, including glycine and serine metabolism, the urea cycle,
   ammonia recycling, and alanine metabolism (p < 0.05 and FDR < 0.05)
   ([94]Figure S5).

Figure 4.

   [95]Figure 4
   [96]Open in a new tab

   Serum metabolite composition. (A) Relative abundance of each metabolite
   class in the negative and positive groups; * p < 0.05; ** p < 0.01. (B)
   Significantly different metabolites (n = 43); log2FC > 0 represents an
   increase in the PG group, while a negative value indicates a decrease.
   PG, positive group.

Figure 5.

   [97]Figure 5
   [98]Open in a new tab

   Boxplot of serum arachidonic acid (A), aspartic acid (B), serine (C),
   pyroglutamic acid (D), N−Acetylserotonin (E), and acetic acid (F) in
   the NG and PG. NG, negative group; PG, positive group. * p < 0.05; ** p
   < 0.01; *** p < 0.001.

3.5. Correlations between Microbiota and Serum Metabolites

   We conducted a correlation analysis network to understand the potential
   association between the oral gut microbiota and metabolites. We
   identified many interconnected relationships between oral-gut
   microbiota and metabolites in the PG ([99]Figure 6). Gut microbiota
   genera, including Veillonella, Adlercreutzia, Segatella,
   Peptostreptococcus, Hoylesella, Desulfovibrio, Faecalimonas,
   Acidaminococcus, Eubacterium, Lachnoanaerobaculum, and Leyella, showed
   a relatively high correlation with serum metabolites ([100]Figure 6A).
   Saliva microbiota genera, including Centipeda, Ottowia, Haemophilus,
   Bacteroides, and Actinobacillus, strongly correlated with serum
   metabolites ([101]Figure 6B).

Figure 6.

   [102]Figure 6
   [103]Open in a new tab

   Interaction between oral-gut microbiota and serum metabolites. (A)
   Spearman’s correlation network between gut microbiota and serum
   metabolites in PG. (B) Spearman’s correlation network between oral
   microbiota and serum metabolites in PG. The red circle represents fecal
   or salivary microbiota, and the blue circle represents serum
   metabolites. The size of the circles represents the number of
   significant correlation relationships. The red line represents positive
   correlation, and the blue line represents negative correlation. F_,
   fecal microbiota; S_, salivary microbiota; PG, positive group.

4. Discussion

   In this study, we found significant alterations in both oral and gut
   microbiota as well as serum metabolites in patients with CD and a
   history of COVID-19. The PG exhibited a higher diversity of microbes
   and distinct β diversity in their oral and gut microbiota compared to
   the NG. Interestingly, the composition of gut microbiota in patients
   with CD showed more similarity to that of HC. This finding highlights
   the dynamic recovery process of gut microbiota diversity among patients
   with CD following COVID-19 infection.

   In the PG, beneficial microbes were increased and harmful microbes
   decreased. Bacteria that acted as probiotics, such as Bifidobacterium,
   Akkermansia, and Faecalibacterium, were increased, while opportunistic
   pathogens, such as Klebsiella and Veillonella, decreased. These changes
   in microbiota may be associated with improved clinical outcomes in
   patients with CD in the PG compared to the NG. After the gut
   microbiota, the oral microbiota is considered the second most diverse
   microbial community within the human body. In healthy individuals,
   microbiota interactions in different ecological niches, such as oral
   and gut microbiota, are relatively infrequent [[104]18,[105]19].
   However, previous studies have observed that patients with CD exhibited
   a greater similarity between their gut and oral microbiota than HC
   [[106]19,[107]20]. Consequently, it is hypothesized that oral bacteria
   may translocate to the gut and contribute to intestinal inflammation in
   pathological conditions such as CD [[108]13,[109]14,[110]21]. In line
   with this notion, our study also revealed an increased interaction
   between oral and gut microbiota in patients with CD compared to HC.
   Notably, this phenomenon was more pronounced in the NG than in the PG,
   indicating that COVID-19 impeded abnormal crosstalk between oral and
   gut microbial communities.

   Compared to NG, the metabolome was also found to be altered in the PG
   group. Most metabolites showed decreased levels in the PG group,
   consistent with a previous study indicating that patients with COVID-19
   exhibit significant metabolic suppression compared to healthy
   individuals [[111]22]. The inflammation-related metabolites were
   decreased in the PG group. Arachidonic acid—an unsaturated fatty acid
   released through the hydrolysis of membrane phospholipids by
   phospholipase—plays a crucial role as an inflammation mediator through
   its metabolites, such as prostaglandins, leukotrienes, and lipotoxins
   [[112]23]. Serine is a central amino acid in the one-carbon metabolism
   network. The serine-dependent one-carbon metabolism plays a crucial
   role in T cell expansion [[113]24] and the function of inflammatory
   macrophages [[114]25]. It promotes the production of inflammatory
   factors such as interleukin-1β (IL-1β) and interferon
   [[115]25,[116]26]. Aspartic acid has been observed to correlate with
   CDAI and inflammatory cytokines levels and shows decreased levels in
   patients with active CD compared to those in remission [[117]27].
   Pyroglutamic acid levels were found to be elevated both in patients
   with IBD and rats with colitis [[118]28,[119]29,[120]30], which has
   been associated with the upregulation of inflammatory factors and DNA
   damage [[121]30]. Interestingly, these inflammation- or IBD-associated
   metabolites were observed at lower levels in PG than in NG; this
   suggests that COVID-19 infection might promote clinical improvement
   among patients with CD by downregulating these specific metabolites. In
   addition, the levels of anti-inflammatory metabolites such as
   N-acetylserotonin and acetic acid were increased in the PG.
   N-acetylserotonin, derived from tryptophan metabolism, possesses potent
   antioxidant and anti-inflammatory properties through receptors on
   innate and adaptive immune cells. It participates in the
   anti-inflammatory nuclear factor erythroid 2-related factor 2
   (Nrf2)-heme oxygenase (HO)-1 pathway to ameliorate colitis
   [[122]31,[123]32]. N-Acetylserotonin was elevated in the PG compared to
   the NG, indicating that COVID-19 infection might promote clinical
   improvement in patients with CD by upregulating it. Consequently,
   COVID-19 might significantly impact both oral-gut microbiota and serum
   metabolome profiles. We further investigated the correlation between
   microbiota and serum metabolome in the PG, revealing significant
   associations between specific microbes and distinct metabolite levels.
   These findings provide evidence that COVID-19 might impact both
   oral-gut microbiota composition and serum metabolome profiles,
   highlighting their reciprocal influence on each other.

   Our discoveries challenge preconceived notions regarding the adverse
   effects of COVID-19 and highlight its potential positive impact on
   microbial recovery. Tursi et al. [[124]33] and Elbadry et al. [[125]34]
   have identified that SARS-CoV-2 infection is a potential trigger factor
   for the de novo occurrence of IBD. However, it is important to note
   that the number of cases reported in these studies was relatively small
   (n = 6 and n = 2), which may limit the generalizability of their
   findings. In contrast, other researchers have observed no increase in
   the risk for de novo IBD following COVID-19 infection, suggesting that
   the relationship between SARS-CoV-2 and IBD may not be straightforward
   and could vary based on multiple factors [[126]35]. Viganò et al.
   [[127]36] observed an association between severe active IBD and
   COVID-19, suggesting that severe disease exacerbations might be linked
   to COVID-19 infection. However, they also noted that this association
   could be influenced by selection bias. Specifically, patients with
   severe IBD exacerbations who were hospitalized may have been more
   likely to be tested for COVID-19, potentially skewing the observed
   correlation. Additionally, they found that patients with IBD and
   COVID-19 reported more cases of diarrhea compared to patients with IBD
   without COVID-19 [[128]36]. However, it was challenging to
   differentiate whether these gastrointestinal symptoms were primarily
   due to the digestive manifestations of COVID-19 or a result of IBD
   relapse or exacerbation. In contrast to these studies, our research
   focused on patients with CD who had experienced mild COVID-19
   infections and had recovered for several months before the study. Our
   study specifically examined the impact of COVID-19 on existing CD
   rather than on new cases of IBD. Notably, our cohort did not include
   patients with severe active CD. By focusing on this specific
   population, our study aims to provide insights into how a mild,
   recovered COVID-19 infection affects pre-existing CD conditions. In
   agreement with our findings, previous studies have demonstrated that
   patients with COVID-19 with gastrointestinal symptoms exhibit
   significantly improved clinical outcomes [[129]37,[130]38]. This
   improvement has been linked to a notable reduction in inflammatory
   dendritic cells in the lamina propria of these patients [[131]39].
   Concurrently, RNA bulk sequencing demonstrated significant
   downregulation of the pathways related to antigen processing, Th17 cell
   differentiation, and IBD, indicating that SARS-CoV-2 infection in the
   intestines may alter the systemic immune cell landscape by shifting
   pro-inflammatory immune profiles towards a more favorable immune state
   [[132]39]. This evidence suggests a plausible biological mechanism
   through which COVID-19 infection could influence the gut microbiome in
   patients with CD. Specifically, the reduction in pro-inflammatory
   immune responses and the downregulation of IBD-related pathways might
   contribute to an improved gut microbiome profile. The decreased
   circulating levels of critical inflammatory proteins, such as IL-6,
   IL-8, IL-17A, and CC chemokine ligand 28, observed by Livanos et al.
   [[133]40], further indicate that COVID-19-induced changes in the immune
   system could enhance the gut microbiome environment by reducing overall
   inflammation.

   This study has several limitations. First, the sample size was not
   sufficiently large, which may limit the generalizability of our
   findings. Second, we did not include patients with COVID-19 without CD
   in our analysis, which prevents drawing broader comparisons. Third, we
   lacked baseline microbiome and metabolome data, which could have
   provided a more comprehensive understanding of the differences between
   the PG and NG. Lastly, while the groups were followed up for several
   months concerning the CDAI, our study did not include longitudinal data
   on the microbiome and metabolome. This was primarily because most
   patients were from out of town, and follow-up was mainly conducted via
   phone, making it challenging to collect consistent longitudinal
   samples.

   For future research, we recommend increasing the sample size to enhance
   the statistical power and generalizability of the findings. Including a
   broader cohort, such as patients with COVID-19 without CD, would
   provide more robust comparative data. Additionally, efforts should be
   made to establish a more comprehensive cohort that allows for the
   observation of dynamic changes across multiple time points, providing a
   deeper understanding of temporal evolution. Finally, exploring the
   underlying mechanisms driving these changes could offer significant
   insights into the interaction between CD, COVID-19, and the microbiome.

5. Conclusions

   In conclusion, a history of mild COVID-19 is beneficial for patients
   with CD, which could be mediated by alterations in the gut microbiota.
   Specifically, patients with CD and a COVID-19 history displayed higher
   levels of bacteria that act as probiotics, lower levels of
   opportunistic pathogens, an increase in anti-inflammatory metabolites,
   and a decrease in inflammatory metabolites.

Acknowledgments