Abstract
Background: Late-onset breast milk jaundice (LBMJ) is a common form of hyperbilirubinemia, which can result in serious complications for newborns with persistently high bilirubin levels. The aim of this study was to investigate the differences in fecal metabolites between breastfed infants with and without LBMJ in order to elucidate potential biological mechanisms. Methods: Biological samples were collected from 12 infants with LBMJ and 12 healthy individuals. Ultra-high performance liquid chromatography quadrupole time-of-flight tandem mass spectrometry (UHPLC-Q-TOF/MS) was utilized for non-targeted determination of fecal metabolites. Principal components analysis (PCA), cluster analysis, and differential metabolite analysis were performed in both positive ion mode and negative ion mode for the two groups. Additionally, the KEGG database was employed to comprehensively analyze the pathways of differential metabolites. Results: There were no significant differences in maternal and neonatal demographic characteristics between the two groups (p > 0.05). The results of PCA and cluster heat map analysis in both modes showed that there were significant differences in metabolites between the two groups. Among 751 differential metabolites (DMs) detected in positive ion mode, 720 were up-regulated in the case group while 31 were down-regulated. In negative ion mode, 1891 DMs were detected, including 817 up-regulated metabolites and 1074 down-regulated metabolites in the case group. Analysis of differential metabolic pathways showed that the DMs of the two groups were mainly annotated and enriched in Biotin metabolism, N-Glycan biosynthesis, Taurine and hypotaurine metabolism, Pyrimidine metabolism, and Pentose and glucuronate interconversions. Conclusion: Significant differences exist in fecal metabolites between LBMJ infants and healthy controls. The study of differential metabolic pathways provides insights into the mechanism of LBMJ.
Introduction
Jaundice, also known as hyperbilirubinemia, is a common clinical issue in newborns [1]. Neonatal hyperbilirubinemia occurs more frequently in breastfed infants compared to formula-fed infants [2]. Late-onset breast milk jaundice (LBMJ) is a type of jaundice that arises in neonates due to breastfeeding. It typically presents within the second week of life and is characterized by the abnormal accumulation of bilirubin, lasting 4–6 weeks or even 2–3 months. This prolonged elevation of bilirubin levels may lead to serious complications in newborn patients [3]. While LBMJ is generally considered to have a good prognosis, its impact on reduced breastfeeding rates and vaccine coverage has been widely observed in the community. This, in turn, may have adverse consequences for disease prevention and infant growth [1, 4]. The global prevalence of LBMJ has not been extensively documented, but research conducted in countries such as the United States Turkey and Taiwan has revealed that 20 to 30% of newborns exhibit symptoms of breast milk jaundice (BMJ) by the time they reach four weeks of age [5].
The etiology of LBMJ has been extensively researched, with a focus on the nutritional components and active factors present in breast milk. However, there is a lack of studies examining the intestinal flora and fecal metabolites in children with LBMJ.
In recent years, studies have demonstrated the involvement of intestinal microbiota in bilirubin metabolism, with the disruption of intestinal microbiota potentially leading to hyperbilirubinemia [6, 7]. Previous research conducted by our team identified a correlation between Klebsiella and LBMJ; however, the precise mechanism by which microbiota contributes to the development of LBMJ remains unclear [8]. The metabolism of both host and intestinal flora results in the production of numerous small molecules that may influence disease occurrence and progression by impacting metabolic pathways [9]. Metabolomics represents one of the most valuable methods for investigating the metabolism of small molecules, including short-chain fatty acids, vitamins, bile acids, hormones, and amino acids [10].
Recently, there has been a significant focus on the relationship between metabolism and jaundice in several studies. Alterations in metabolic profiles of jaundice-related diseases, particularly bile acid profiles, have been observed in both human and animal experiments [11–13]. As the field of metabolomics research in neonatology continues to advance, there is an increasing number of publications that concentrate on the correlation between metabolic changes associated with the extrauterine environment and metabolic diseases in early neonatal life [11].
So far, several metabolomic studies related to neonatal jaundice have identified various metabolites and metabolic pathways, but the conclusions have been inconsistent [14, 15]. In order to investigate the relationship between gut microbiota metabolites and LBMJ, metabolomics can be utilized to efficiently screen biomarkers and further analyze the molecular mechanism of LBMJ. Ultra-high performance liquid chromatography quadrupole time-of-flight tandem mass spectrometry (UHPLC-Q-TOF/MS) was employed for non-targeted determination of fecal metabolites in this study. A more comprehensive understanding of gut microbiome metabolomics may contribute to a better comprehension of the etiology of LBMJ, ultimately leading to improved prevention and treatment strategies.
Methods
Subjects
Patients with LBMJ and jaundice-free newborns from Peking University People’s Hospital were included in the study during October 1, 2020 to January 1, 2022 based on their participation in a follow-up mother-infant cohort.
Inclusion criteria were as follows:
The mother is between 20 and 45 years old and has a medical record at Peking University People’s Hospital.
Full-term infants born at Peking University People’s Hospital.
Exclusive breastfeeding or predominantly breastfeeding (more than 80% breastfeeding).
Presence of jaundice in the skin or sclera persisting for more than 3 weeks.
Transdermal bilirubin value exceeding 7.87mg/dL on the 42nd day after birth.
Infants were excluded for the following reasons:
The infants were excluded if they had hemolytic disease, blood type incompatibilities, reticulocytosis, abnormal blood smear, polycythemia, a positive Coombs’ test, glucose-6-phosphate dehydrogenase deficiency, skull hematoma, hypothermia, intracranial hemorrhage, cholestasis, neonatal hypothyroidism, phenylketonuria screening that was positive and if they had used antibiotics or probiotics.
Finally, twelve infants with LBMJ and twelve healthy full-term newborns without exposure to antibiotics or probiotics after birth were included in the case group and the control group separately, based on the breast milk bank follow-up cohort established by our research group, which belongs to a prospective nested case-control study. The study received approval from the Ethics Committee of the Peking University People’s Hospital (Approval Number: 2020PHB113-01), and written informed consents were obtained from the guardians of all participants.
Sample collection and storage
Newborn fecal samples are collected once at 42 days after birth without a specific time limit from October 1, 2020 to January 1, 2022. The specimens are obtained from diapers using sterile stool collection tubes, then promptly transferred to the laboratory and stored in a refrigerator at -80°C.
Extraction and detection of fecal metabolites
Frozen fecal samples from all 24 neonates were utilized for metabolomics analysis. The extraction procedure was carried out according to Cheng K et al.21 and Ng JS et al. [16, 17], with some modifications. The fecal samples were ground into powder using liquid nitrogen, with a weight of 100 mg, and then mixed with 500 μL of pre-cooled ultra-pure water. The mixture was vortexed for 30 seconds, followed by centrifugation at 10,000 g for 15 minutes to obtain the supernatant, which was transferred to another centrifuge tube. Subsequently, 500 μL of pre-cooled methanol was added to the remaining sediment, swirled for 30 seconds, and centrifuged at 10,000 g for another 15 minutes. The two resulting supernatants were combined and allowed to stand at -20°C for a duration of 12 hours. Afterward, they were subjected to further centrifugation at 10,000 g for an additional period of time lasting up to15 minutes in order to obtain the final supernatant. This final supernatant was then passed through a filter membrane with a pore size of 0.22μm before being placed into sample bottles. The bottle caps were tightly sealed and wrapped with sealing film before storing them at -20°C until testing could be conducted. Ultra-high performance liquid chromatography quadrupole time-of-flight tandem mass spectrometry (UHPLC-Q-TOF/MS, Agilent Technologies, CA, USA) was utilized for the non-targeted determination of fecal metabolites. An Agilent Poroshell 120 SB-Aq column (100 mm × 2.1 mm i.d., 2.7 μm) was employed to separate the extracts. Quality control (QC) samples were prepared by mixing an equal volume of each sample (5 μL). The samples were used to evaluate the data reliability for metabolomics analysis. Both positive and negative ionization modes were utilized. The mobile phases A and B consisted of water and acetonitrile (0.1% formic acid), respectively. The gradient elution program was as follows: 0–2 min, 5% B; 2–20 min, linear increase from 5% to 100% B; 20–25 min, hold at 100% B; post-run time: 5min. The flow rate was set at a constant value of 300 μL/min, with an injection volume of 2 μL and a column temperature maintained at a steady level of 40°C.
For MS acquisition, an Agilent Q-TOF LC/MS system model number 6545 was employed with the following parameters: capillary voltage set at 4.0 kV for positive ionization mode and 3.5kV for negative ionization mode; gas temperature held constant at 300°C; drying gas flow rate maintained at 9 L/min; nebulizer pressure set to 35 psi; sheath gas temperature kept at 350°C; sheath gas flow rate adjusted to 11 L/min; mass range scanned from 50 to 1700 m/z.
Statistical analysis
The Agilent Profinder software was utilized for retention time correction, peak identification, extraction, integration, alignment, and other data processing tasks on the original mass spectrum data, ultimately generating CEF files. Subsequently, the Agilent Mass Profiler Professional software was employed for statistical analysis, PLSDA and OPLSDA modeling under "supervised" conditions to evaluate the overall sample distribution. Material identification was conducted using the Metlin database, while the volcano plot was utilized to screen for differential compounds. Differential metabolite screening criteria were defined as metabolites with a Fold change > 2 and P value < 0.05 by t-test. MetaboAnalyst 6.0 and KEGG were applied for metabolite pathway analysis.
Results
Clinical characteristics of population
Based on previous research, 24 subjects were ultimately included in this study after excluding those with incomplete data. The study comprised of 12 LBMJ infants in the case group and 12 healthy individuals in the control group.
There were no significant differences observed between the two groups in terms of infants’ sex ratio, birth length, birth weight, 1 min Apgar score, 5 min Apgar score, as well as mothers’ BMI, maternal age, gestational week, delivery mode, parity and pregnancy complications (P>0.05) (Table 1).
Table 1. Clinical data of participants.
| LBMJ (n = 12) | Control (n = 12) | P value | ||
|---|---|---|---|---|
| Infants | ||||
| sex ratio (male/female) | 1.00 (6/6) | 0.50(4/8) | 0.408 | |
| birth length (Mean±SD, cm) | 49.83±1.70 | 49.67±1.67 | 0.811 | |
| birth weight (Mean±SD, g) | 3256.67±391.62 | 3358.33±336.31 | 0.502 | |
| 1 min *Apgar score (Mean±SD) | 10±0 | 10±0 | 1.000 | |
| 5 min *Apgar score (Mean±SD) | 10±0 | 10±0 | 1.000 | |
| 10 min *Apgar score (Mean±SD) | 10±0 | 10±0 | 1.000 | |
| Mothers | ||||
| BMI (Mean±SD, kg/m2) | 22.60±3.29 | 20.81±2.14 | 0.130 | |
| maternal age (Mean±SD, y) | 33.33±2.64 | 31.58±3.29 | 0.165 | |
| gestational week (Mean±SD, w) | 39.13±0.87 | 38.85±0.94 | 0.469 | |
| delivery mode (%) | Vaginal | 10(83.30%) | 9(75.00%) | 0.615 |
| Cesarean | 2(16.70%) | 3(25.00%) | ||
| **parity (%) | 1 | 8(66.70%) | 8(66.70%) | 1.000 |
| 2 | 4(33.30%) | 4(33.30%) | ||
| pregnancy complications (%) | Yes | 2(16.70%) | 5(41.70%) | 0.178 |
| No | 10(83.30%) | 7(58.30%) |
*Apgar score represents birth health including suffocation.
**Data of parity represents times of delivery.
Sample total ion flow chromatogram based on UHPLC-Q-TOF/MS
As depicted in Fig 1, metabolites in fecal samples were effectively separated under the positive and negative ion detection modes, with a retention time of less than 25 minutes.
Fig 1.
a. UHPLC-Q-TOF/MS TIC diagram of QC samples (positive ion mode). b. UHPLC-Q-TOF/MS TIC diagram of QC samples (negtive ion mode).
Principal components analysis (PCA) of case group and control group
PCA was conducted on mixed samples from the case and control groups in order to identify differences in metabolic profiles between them. The PCA plot of positive ion mode (Fig 2A) showed that samples from the case group were clustered together, positioned to the left of the control group samples. In the PCA plot of negative ion mode (Fig 2B), samples from the case group were grouped together, situated to the right of the control group samples. These findings suggest that our analysis was stable and reproducible.
Fig 2.
Two-dimensional PCA plot (a) positive ion mode; (b) negative ion mode. The yellow dots represent the case group and the red dots represent the control group.
Cluster analysis of case group and control group
Differences in the accumulation patterns of metabolites in the two groups of samples can be analyzed using heat maps (the top 50 from T-tests/ANOVA results are shown in Fig 3). The findings indicate significant differences in metabolite levels between the two groups. In the positive ion mode, the metabolites were categorized into two clusters. Metabolites in cluster 1 exhibited higher expression in the control group and lower expression in the case group. With only two exceptions in cluster 2, the majority of metabolites showed higher expression in the control group and lower expression in the case group. Conversely, in the negative ion mode, metabolites were divided into two clusters: both cluster 1 and cluster 2 displayed higher expression levels in the case group and lower expression levels in the control group.
Fig 3.
Heat maps of TOP 50 in (a) positive ion mode and (b) negative ion mode. The samples were grouped by horizontal coordinates, and the color scale from blue to orange indicated the content of differentially expressed metabolites from low to high.
Differential metabolite analysis: Volcano maps
Metabolites that satisfy both FC > 2 or < 0.05 and P < 0.05 simultaneously are considered as differential metabolites. In Fig 4A, there were a total of 720 up-regulated metabolites, indicating higher expression levels in the case group compared to the control group, and 31 down-regulated metabolites, indicating lower expression levels in the case group compared to the control group. In Fig 4B, there were a total of 817 up-regulated metabolites, indicating higher expression levels in the case group compared to the control group, and 1074 down-regulated metabolites, indicating lower expression levels in the case group compared to the control group.
Fig 4.
a. volcano plot in positive ion mode. b. volcano plot in negative ion mode.
Analysis of differential metabolic pathways
The KEGG database was utilized for a comprehensive analysis of the pathways of differential metabolites. As indicated in Tables 2 and 3, 24 and 32 pathways were identified in the positive and negative ion modes, respectively. The total number of compounds in each pathway is presented in column 2; Hits represent the actual matched number from the data; Raw p denotes the original p-value calculated from the enrichment analysis; Impact refers to the pathway impact value calculated from pathway topology analysis.
Table 2. Differential metabolic pathways in positive ion mode.
| KEGG pathway | Total number of compounds | Hits | Differential metabolites* | Raw p | -log10(p) | Impact |
|---|---|---|---|---|---|---|
| Nicotinate and nicotinamide metabolism | 15 | 1 | Nicotinate D-ribonucleoside | 0.003813 | 2.4187 | 0 |
| Caffeine metabolism | 10 | 1 | 7-Methyluric acid | 0.005382 | 2.2691 | 0 |
| Purine metabolism | 70 | 2 | Inosine;8-Hydroxyadenine | 0.006698 | 2.174 | 0.00955 |
| Histidine metabolism | 16 | 1 | Imidazole-4-acetate | 0.009499 | 2.0223 | 0 |
| N-Glycan biosynthesis | 41 | 2 | Dolichyl phosphate; Dolichyl diphosphate; | 0.014307 | 1.8445 | 0.10845 |
| Arginine and proline metabolism | 36 | 1 | Spermidine | 0.017808 | 1.7494 | 0.06628 |
| Glutathione metabolism | 28 | 1 | Spermidine | 0.017808 | 1.7494 | 0.00719 |
| β-Alanine metabolism | 21 | 1 | Spermidine | 0.017808 | 1.7494 | 0 |
| Biotin metabolism | 10 | 1 | Biotinyl-5’-AMP | 0.020436 | 1.6896 | 0.15 |
| Arachidonic acid metabolism | 44 | 1 | 11,12-EET | 0.092701 | 1.0329 | 0.01474 |
| Glycine, serine and threonine metabolism | 33 | 1 | Βine | 0.26142 | 0.58266 | 0.05149 |
| Tryptophan metabolism | 41 | 1 | Indole-3-acetaldehyde | 0.28758 | 0.54124 | 0.0139 |
| Pyrimidine metabolism | 39 | 1 | dTMP | 0.29507 | 0.53007 | 0.08661 |
| Linoleic acid metabolism | 5 | 1 | 12(13)-EpOME | 0.30567 | 0.51475 | 0 |
| α-Linolenic acid metabolism | 13 | 1 | Stearidonic acid | 0.30786 | 0.51165 | 0 |
| Porphyrin metabolism | 31 | 2 | Bilirubin β-diglucuronide; Cob(II)yrinate a,c diamide | 0.49906 | 0.30185 | 0.02795 |
| Sphingolipid metabolism | 32 | 4 | Sphinganine; Globoside; GM3; Phytosphingosine | 0.50413 | 0.29746 | 0.1159 |
| One carbon pool by folate | 9 | 1 | Dihydrofolate | 0.66725 | 0.17571 | 0.01587 |
| Folate biosynthesis | 27 | 2 | THF-polyglutamate; Dihydrofolate | 0.68781 | 0.16253 | 0.00434 |
| Ubiquinone and other terpenoid-quinone biosynthesis | 18 | 1 | Ubiquinol | 0.74079 | 0.13031 | 0 |
| Pantothenate and CoA biosynthesis | 20 | 1 | Pantetheine 4’-phosphate | 0.78493 | 0.10517 | 0.22449 |
| Cysteine and methionine metabolism | 33 | 1 | 4-Methylthio-2-oxobutanoic acid | 0.81111 | 0.090922 | 0.06458 |
| Drug metabolism—other enzymes | 39 | 1 | Irinotecan | 0.8792 | 0.055913 | 0 |
| Lysine degradation | 30 | 1 | Carnitine | 0.92253 | 0.035018 | 0 |
* Recurring compounds are shown in bold.
Table 3. Differential metabolic pathways in negative ion mode.
| KEGG pathway | Total number of compounds | Hits | Differential metabolites* | Raw p | -Log(p) | Impact |
|---|---|---|---|---|---|---|
| Lysine degradation | 30 | 1 | L-Lysine | 5.93E-05 | 4.2272 | 0 |
| Biotin metabolism | 10 | 1 | L-Lysine | 5.93E-05 | 4.2272 | 0 |
| Caffeine metabolism | 10 | 2 | 1-Methyluric acid;5-Acetylamino-6-formylamino-3-methyluracil | 0.001603 | 2.7952 | 0 |
| Pyrimidine metabolism | 39 | 4 | Uridine;Deoxyuridine;Thymidine;Orotate | 0.002098 | 2.6781 | 0.18316 |
| Purine metabolism | 70 | 2 | Guanosine;Urate | 0.003795 | 2.4207 | 0 |
| Fatty acid degradation | 39 | 1 | Octanoyl-CoA | 0.0058 | 2.2365 | 0.02355 |
| Fatty acid elongation | 39 | 1 | Octanoyl-CoA | 0.0058 | 2.2365 | 0.01148 |
| Glycosaminoglycan biosynthesis—chondroitin sulfate / dermatan sulfate | 10 | 1 | Chondroitin | 0.006707 | 2.1735 | 0 |
| Pentose and glucuronate interconversions | 19 | 2 | β-D-Glucuronoside; L-Xylulose | 0.007174 | 2.1442 | 0.10843 |
| Metabolism of xenobiotics by cytochrome P450 | 68 | 2 | 7,12-Dimethylbenz[a]anthracene 5,6-oxide;1a,11b-Dihydro-4,9-dimethylbenz[a]anthra[3,4-b]oxirene | 0.007319 | 2.1356 | 0.08163 |
| Selenocompound metabolism | 20 | 1 | Dimethyl selenide | 0.008723 | 2.0593 | 0 |
| Ascorbate and aldarate metabolism | 9 | 1 | β-D-Glucuronoside | 0.010438 | 1.9814 | 0 |
| Primary bile acid biosynthesis | 46 | 3 | Taurine;3α,7α-Dihydroxy-5β-cholest-24-enoyl-CoA; 3α,7α,12α-Trihydroxy-5β-cholestanoate | 0.012068 | 1.9184 | 0.07167 |
| D-Amino acid metabolism | 15 | 1 | D-Aspartate | 0.020155 | 1.6956 | 0 |
| Taurine and hypotaurine metabolism | 8 | 1 | Taurine | 0.022413 | 1.6495 | 0.42857 |
| Alanine, aspartate and glutamate metabolism | 28 | 2 | D-Aspartate; Citrate | 0.026877 | 1.5706 | 0 |
| Citrate cycle (TCA cycle) | 20 | 1 | Citrate | 0.026886 | 1.5705 | 0.09038 |
| Glyoxylate and dicarboxylate metabolism | 32 | 1 | Citrate; | 0.026886 | 1.5705 | 0.03175 |
| Histidine metabolism | 16 | 1 | Imidazole-4-acetate | 0.06201 | 1.2075 | 0 |
| Starch and sucrose metabolism | 18 | 1 | Sucrose | 0.073103 | 1.1361 | 0.05023 |
| Galactose metabolism | 27 | 1 | Sucrose | 0.073103 | 1.1361 | 0.03888 |
| Biosynthesis of unsaturated fatty acids | 36 | 1 | Octadecanoic acid | 0.10193 | 0.99169 | 0 |
| Steroid hormone biosynthesis | 87 | 2 | Urocortisol; Tetrahydrocortisone | 0.133 | 0.87615 | 0 |
| Porphyrin metabolism | 31 | 1 | Bilirubin β-diglucuronide | 0.25124 | 0.59991 | 0.02795 |
| Linoleic acid metabolism | 5 | 1 | 12(13)-EpOME | 0.25994 | 0.58513 | 0 |
| Pantothenate and CoA biosynthesis | 20 | 1 | N-((R)-Pantothenoyl)-L-cysteine | 0.32818 | 0.48388 | 0 |
| N-Glycan biosynthesis | 41 | 1 | Dolichyl phosphate D-mannose | 0.32818 | 0.48388 | 0.00676 |
| Folate biosynthesis | 27 | 1 | Dihydropteroate | 0.35267 | 0.45264 | 0 |
| Arginine biosynthesis | 14 | 1 | N-Acetyl-L-glutamate | 0.55242 | 0.25773 | 0 |
| Tryptophan metabolism | 41 | 1 | 6-Hydroxymelatonin | 0.6107 | 0.21417 | 0 |
| Glycerophospholipid metabolism | 36 | 1 | Phosphatidylserin | 0.61623 | 0.21025 | 0.04701 |
| Terpenoid backbone biosynthesis | 18 | 1 | Farnesylcysteine | 0.68944 | 0.16151 | 0 |
* Recurring compounds are shown in bold.
The pathway analysis in positive and negative ion mode was separately presented in Fig 5. The figure highlights important KEGG pathways, including Biotin metabolism, N-Glycan biosynthesis, Taurine and hypotaurine metabolism, Pyrimidine metabolism, and Pentose and glucuronate interconversions (Pathway Impact > 0.1, -log10(p) > 1.0).
Fig 5.
Summary of pathway analysis in (a) positive ion mode and (b) negative ion mode.
Discussion
Although an increasing number of studies have focused on LBMJ, the exact etiology of it has not been determined. Metabolomics can be utilized to detect and identify various metabolites in order to aid in the exploration of pathological mechanisms. In this study, we selected 24 LBMJ neonates based on strict metabolomic selection criteria to determine the changes in fecal metabolites and understand the pathogenesis of LBMJ.
The results of PCA and cluster analysis revealed that samples from the case group and the control group exhibited distinct separation into two different regions, indicating significant differences in the accumulation patterns of metabolites between the two groups. A recent study conducted by Yaxuan Li et al. utilized gas chromatography-mass spectrometry (GC/MS) to analyze fecal metabolites, and similarly observed a notable distinction in the metabolic profiles of infant fecal samples between the BMJ group and the healthy control group [18]. After comparing the differential metabolites, it was determined that in the positive ion mode, there were a total of 751 differential metabolites identified in both the case group and the control group. Among these, 720 metabolites were up-regulated and 31 metabolites were down-regulated. In the negative ion mode, a total of 1891 different metabolites were found in both groups, with 817 up-regulated metabolites and 1074 down-regulated metabolites. When compared to relevant studies utilizing targeted metabolomics, such as short-chain fatty acids and bile acids in infant feces, this study revealed a relatively large number of differential metabolites with more complex types [15, 18].
In addition, KEGG pathway analysis identified 56 differential metabolic pathways. The KEGG database is a powerful tool for metabolism analysis and metabolic network research, providing comprehensive information on metabolic pathways involving carbohydrates, nucleosides, amino acids, as well as genes, expression data, and metabolite content studied as a whole network [18, 19]. Based on the comprehensive analysis of P value and Pathway Impact, five main differential metabolic pathways were found between the case group and control group in this study (Fig 5). Representative differential metabolites include Biotinyl-5’-AMP, Dolichyl phosphate, Dolichyl diphosphate, Taurine, Uridine, Deoxyuridine, Thymidine,Orotate, β-D-Glucuronoside, and L-Xylulose. Furthermore, Spermidine, Dihydrofolate, L-Lysine, Octanoyl-CoA, β-D-Glucuronoside, Taurine, D-Aspartate, Citrate and Sucrose were identified multiple times in the pathways summary table (Table 2), indicating their potential significance in the metabolic pathways associated with LBMJ. These findings offer valuable insights for analyzing the metabolomics mechanism of LBMJ, and those differential metabolites mentioned above may be potential pharmacological targets and diagnostic analytes which may be clinical beneficial for LBMJ. Recent research on the etiology of BMJ has highlighted breast milk-related factors as predominant [20]. Breast milk is closely linked to the establishment of infant intestinal flora. The intestinal flora and its metabolites may impact bilirubin metabolism through participation in metabolic pathways, thereby influencing the onset and progression of jaundice. To date, limited studies have explored the fecal metabolomics mechanisms of BMJ.
The enhanced taurine and hypotaurine metabolism in the case group was predicted through untargeted metabolome analysis in our study. Taurine, identified as one of the most crucial differential metabolites in this study, suggests that it plays a vital role in bilirubin metabolism in infants with BMJ. The synthesis or metabolism of taurine is commonly observed in hepatocytes and is fundamental to the "enterohepatic circulation" of bile acids [21]. Despite its numerous diverse biological and important metabolic functions, its role in BMJ metabolomics mechanisms has yet to be fully elucidated. Previous studies have indicated that the concentration of taurine-conjugated bile acids in the stool of children with BMJ is significantly higher than that of healthy infants, and the proportion of serum taurine-conjugated bile acids increases proportionally to serum bilirubin levels [22, 23]. However, taurine has been documented to exhibit protective effects on neuronal cells against ischemia both in vivo and in vitro studies [24]. Recent research has also indicated that the enhancement of taurine and hypotaurine metabolism may have beneficial effects on the regulation of lipid metabolism and the restoration of liver function. This is achieved through its antioxidant and anti-inflammatory properties [25, 26]. These findings suggest the potential of taurine as an effective agent for preventing or treating neuronal damage in neonatal jaundice. Furthermore, they indicate that taurine and hypotaurine metabolism could be a targeted metabolic pathway for intervention therapy in bilirubin-induced neurological dysfunction.
Studies have indicated that taurine has the potential to regulate intestinal microecology by potentially inhibiting the growth of harmful bacteria, expediting the production of short-chain fatty acids, and reducing the concentration of LPS [27]. Additionally, research findings suggest that certain bacteria such as Roseburia, Klebsiella, Clostridium IV, and Ruminococcaceae may impact the development of metabolic diseases by influencing host metabolic pathways like taurine, sphingolipid, and ceramide [28]. Furthermore, a previous study revealed significant differences in the richness and diversity of intestinal microbiota between the LBMJ group and the control group. Notably, there was a marked increase in the dominant genus Klebsiella within the LBMJ group [8]. These findings suggest that Klebsiella may play a significant role in influencing intestinal homeostasis through certain metabolic pathways, ultimately leading to the development of hyperbilirubinemia. Recent research on bacterial sulfoacetaldehyde and taurine metabolism has revealed that IsfD, a reductase from Klebsiella, facilitates the reversible reduction of thioacetaldehyde to the corresponding isotaurine alcohol. This process is a crucial step in detoxifying carbonyl intermediates formed during bacterial nitrogen assimilation from the α-aminoalkanesulfonic acid taurine [29]. Therefore, it can be inferred that the pathogenesis of LBMJ may be linked to the interaction between Klebsiella and taurine metabolism.
Besides Klebsiella, which is one member of Enterobacteriacea in phylum Proteobacteria, Firmicutes, Proteobacteria, Haemophilus, Escherichia, Morganella and Rothia et.cl in infant intestine have also been found relevant with BMJ in several researches [30], and some specific gut microbiota, for example, Lactobacillus had been discovered that it had higher abundance of functional genes involved in taurine and hypotaurine metabolism in mice [31]. Even though the exactly relationship between metabolic pathways and intestinal flora of LBMJ infants is not clear, taurine and hypotaurine metabolism is probably closely related to fecal microbiota metabolism especially Firmicutes and Proteobacteria based on these recent studies [14, 32], thus it provides a future detective direction for us.
Another key differential metabolite identified in this study is β-D-glucuronoside, which plays a crucial role in the pentose and glucuronate interconversions pathway. A functional metabolomics study has shown that D-glucuronic acid is essential for binding unconjugated bilirubin via the pentose and glucuronate interconversions pathway [33]. The enzyme β-glucuronidase is pivotal in the "enterohepatic circulation", as it releases residual bilirubin in the intestine, allowing it to return to the liver through portal vein circulation for recombination [34]. Neonatal β-glucuronidase activity is significantly higher, and the intestinal wall is more permeable, resulting in higher overall unbound bilirubin concentrations and increased enterohepatic circulation. The expression of β-D-glucuronoside was significantly reduced in the case group of this study, suggesting that it may be decomposed under the action of more β-glucuronidase, possibly produced by an increase in Klebsiella [8]. Combining these results leads to a consistent conclusion: decreased intestinal β-D-glucuronoside levels in newborns may contribute to LBMJ occurrence by affecting the pentose and glucuronate interconversion pathway.
Other differentiated metabolites and metabolic pathways are primarily associated with amino acids, pyrimidines, vitamins, carbohydrates, and energy metabolism. Some of these have been confirmed to be linked to bilirubin metabolism in feces or serum or both. Previous studies have demonstrated that citric acid plays a role in the tricarboxylic acid cycle, which is beneficial for maintaining intestinal barrier function integrity and reducing the increase in enterohepatic circulation of bilirubin caused by damage [35, 36]. In this study, the concentration of citrate in the feces of the case group was lower than that of the control group, potentially leading to slower intestinal development and energy synthesis in infants, resulting in bilirubin accumulation and BMJ occurrence. Aspartic acid is involved in uridine diphospho-glucuronosyltransferase (UDPGT) synthesis, a rate-limiting enzyme during bilirubin metabolism. Gourley et al. [37] found that supplementation with aspartic acid could reduce transcutaneous bilirubin levels, promote excretion of fecal bilirubin, and alleviate jaundice. In this study, the fecal metabolite D-aspartate content was lower in infants from the case group compared to those from the control group; this may lead to a deficiency in UDPGT synthesis due to aspartic acid shortage and slow down bilirubin metabolism resulting in its accumulation. Spermidine is crucial for regulating cell growth and differentiation while preventing liver damage and promoting liver tissue regeneration [38]. The down-regulated fecal spermidine content observed in the case group may lead to jaundice due to insufficient metabolic capacity for bilirubin caused by immature liver function. In addition, strong relationship between bilirubin metabolism and some metabolites in blood samples, including Taurine, Citrate and Sucrose, which were found to be significant in this study, has been discovered in several studies [39–41], indicating that there may be some correlations of these key metabolites from different sources of specimens.
It is well-known that the gut microbiome plays a central role in the homeostasis of bile acid metabolism [42], and the present study has shown that there are several differential metabolites related to specific pathways in which some type of intestinal microbiota may play essential roles. The composition and activity of the gut microbiota codevelop with the host and is subject to a complex interplay that depends on the host genome, nutrition, and life-style [43]. While conjugated bile acids are deconjugated by bile salt hydrolases produced by gut microbiota, bile acids in turn have direct antimicrobial effects on gut microbiota and indirect effects through FXR-induced antimicrobial peptides, and decrease of bile acid levels in the gut favors the growth of Gram-negative bacteria [44]. Therefore, variations in microflora and fauna may affect the disorder and vice-versa, and deeper exploration about the causal relationship between them should be performed in the near future.
There are two strengths in this study, namely the unique scheme about fecal metabolism of LBMJ utilizing untargeted metabolomics approach while few studies had focused on it, and the novel differential metabolites discovered which may provide new possible molecular mechanism and clinical strategy of LBMJ. However, the study has two drawbacks. First, in this study, there was a lack of exploration of the influence of maternal factors, especially breast milk composition, maternal diet and other lifestyle factors, on infant intestinal metabolites, because LBMJ is inevitably associated with breast milk which is related with various maternal factors. Second, this is a cross-section study with a small number of subjects. Though subjects had been selected through rigorous criteria and significant differences were observed in this study after metabolomics statistical analyses, the data might not accurately reflect the characteristics of a broader, more diverse population. Thus deeper exploration will take place as the size of mother-infant cohort extends and more maternal factors are taken into account.
Conclusion
Given the high prevalence of LBMJ and its potential impact on infant health, this study is dedicated to exploring the molecular mechanisms underlying the metabolomics aspects of the disease and discovered that there are significant differences in fecal metabolites between infants with LBMJ and healthy controls. The discovery of differential metabolic pathways provides novel insights into the mechanism of LBMJ which may be related to the regulation of infant metabolism especially bilirubin metabolism by influencing the levels of taurine, β-D-glucuronoside, citrate, D-aspartate, thymidine, as well as other metabolites mentioned above in infant intestine, although they were first discovered in this study. This study provides theoretical support for the intestinal metabolomics mechanism of LBMJ, and suggests that intestinal flora and its metabolites may be used as biomarkers for diagnosing and monitoring the severity of LBMJ in the future. Moreover, this study also provides theoretical and experimental basis for the regulation of intestinal flora and its metabolites, which is also a new direction for the treatment strategy of LBMJ.
Supporting information
(XLSX)
Acknowledgments
We express our gratitude to all the participants involved in this study. Additionally, we extend our thanks for the research platform provided by Peking University People’s Hospital.
Data Availability
All relevant data are within the manuscript and its Supporting Information files.
Funding Statement
This research was supported by the Construction of Breast Milk Bank Based on Birth Cohort (2020-Z-19) grant funded by the Ausnutria Maternal and Child Nutrition Research Fund (AU-YJY-B-LX-19-018). The sponsors or funders played no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
References
- 1.Ullah S, Rahman K, Hedayati M. Hyperbilirubinemia in Neonates: Types, Causes, Clinical Examinations, Preventive Measures and Treatments: a Narrative Review Article. Iran J Public Health. 2016;45(5): 558–68. [PMC free article] [PubMed] [Google Scholar]
- 2.Preer GL, Philipp BL. Understanding and managing breast milk jaundice. Arch Dis Child Fetal Neonatal Ed. 2011;96(6): F461–6. doi: 10.1136/adc.2010.184416 [DOI] [PubMed] [Google Scholar]
- 3.Muchowski KE. Evaluation and treatment of neonatal hyperbilirubinemia. Am Fam Physician. 2014;89(11): 873–8. [PubMed] [Google Scholar]
- 4.Wilkins T, Sams R, Carpenter M. Hepatitis B: Screening, Prevention, Diagnosis, and Treatment. Am Fam Physician. 2019;99(5): 314–23. [PubMed] [Google Scholar]
- 5.Maruo Y, Morioka Y, Fujito H, Nakahara S, Yanagi T, Matsui K et al. Bilirubin uridine diphosphate-glucuronosyltransferase variation is a genetic basis of breast milk jaundice. J Pediatr. 2014;165(1): 36–41. doi: 10.1016/j.jpeds.2014.01.060 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cani PD, Van Hul M, Lefort C, Depommier C, Rastelli M, Everard A. Microbial regulation of organismal energy homeostasis. Nat Metab. 2019;1(1): 34–46. doi: 10.1038/s42255-018-0017-4 [DOI] [PubMed] [Google Scholar]
- 7.Lee WJ, Hase K. Gut microbiota-generated metabolites in animal health and disease. Nat Chem Biol. 2014;10(6): 416–24. doi: 10.1038/nchembio.1535 [DOI] [PubMed] [Google Scholar]
- 8.Guo Q, Liu X, Cui M, Li X, Yang C, Zhao S et al. Characteristics of intestinal microbiota in infants with late-onset breast milk jaundice. Front Nutr. 2023;10(1119768. doi: 10.3389/fnut.2023.1119768 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Nicholson JK, Holmes E, Kinross J, Burcelin R, Gibson G, Jia W et al. Host-gut microbiota metabolic interactions. Science. 2012;336(6086): 1262–7. doi: 10.1126/science.1223813 [DOI] [PubMed] [Google Scholar]
- 10.Kennedy AD, Wittmann BM, Evans AM, Miller L, Toal DR, Lonergan S et al. Metabolomics in the clinic: a review of the shared and unique features of untargeted metabolomics for clinical research and clinical testing. J Mass Spectrom. 2018;53(11): 1143–54. doi: 10.1002/jms.4292 [DOI] [PubMed] [Google Scholar]
- 11.Georgakopoulou I, Chasapi SA, Bariamis SE, Varvarigou A, Spraul M, Spyroulias GA. Metabolic changes in early neonatal life: NMR analysis of the neonatal metabolic profile to monitor postnatal metabolic adaptations. Metabolomics. 2020;16(5): 58. doi: 10.1007/s11306-020-01680-4 [DOI] [PubMed] [Google Scholar]
- 12.Long Y, Dong X, Yuan Y, Huang J, Song J, Sun Y et al. Metabolomics changes in a rat model of obstructive jaundice: mapping to metabolism of amino acids, carbohydrates and lipids as well as oxidative stress. J Clin Biochem Nutr. 2015;57(1): 50–9. doi: 10.3164/jcbn.14-147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wang X, Zhang A, Han Y, Wang P, Sun H, Song G et al. Urine metabolomics analysis for biomarker discovery and detection of jaundice syndrome in patients with liver disease. Mol Cell Proteomics. 2012;11(8): 370–80. doi: 10.1074/mcp.M111.016006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.14 Duan M, Han ZH, Huang T, Yang Y, Huang B. Characterization of gut microbiota and short-chain fatty acid in breastfed infants with or without breast milk jaundice. Lett Appl Microbiol. 2021;72(1): 60–7. doi: 10.1111/lam.13382 [DOI] [PubMed] [Google Scholar]
- 15.Yang T, Yang S, Zhao J, Wang P, Li S, Jin Y et al. Comprehensive Analysis of Gut Microbiota and Fecal Bile Acid Profiles in Children with Biliary Atresia. Front Cell Infect Microbiol. 2022;12(914247. doi: 10.3389/fcimb.2022.914247 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Cheng K, Brunius C, Fristedt R, Landberg R. An LC-QToF MS based method for untargeted metabolomics of human fecal samples. Metabolomics. 2020;16(4): 46. doi: 10.1007/s11306-020-01669-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ng JS, Ryan U, Trengove RD, Maker GL. Development of an untargeted metabolomics method for the analysis of human faecal samples using Cryptosporidium-infected samples. Mol Biochem Parasitol. 2012;185(2): 145–50. doi: 10.1016/j.molbiopara.2012.08.006 [DOI] [PubMed] [Google Scholar]
- 18.Li Y, Shen N, Li J, Hu R, Mo X, Xu L. Changes in Intestinal Flora and Metabolites in Neonates with Breast Milk Jaundice. Front Pediatr. 2020;8(177. doi: 10.3389/fped.2020.00177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Tanabe M, Kanehisa M. Using the KEGG database resource. Curr Protoc Bioinformatics. 2012;Chapter 1(1–12. doi: 10.1002/0471250953.bi0112s38 [DOI] [PubMed] [Google Scholar]
- 20.Soldi A, Tonetto P, Varalda A, Bertino E. Neonatal jaundice and human milk. J Matern Fetal Neonatal Med. 2011;24 Suppl 1(85–7. doi: 10.3109/14767058.2011.607612 [DOI] [PubMed] [Google Scholar]
- 21.Chiang JY. Bile acids: regulation of synthesis. J Lipid Res. 2009;50(10): 1955–66. doi: 10.1194/jlr.R900010-JLR200 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.22 Demirkol M, Bohles H. Breast milk taurine and its possible influence on the development of breast milk induced jaundice of the neonate—a hypothesis. Adv Exp Med Biol. 1994;359(405–10. doi: 10.1007/978-1-4899-1471-2_42 [DOI] [PubMed] [Google Scholar]
- 23.Yamada M, Tazawa Y, Nakagawa M, Konno T, Tada K, Goto J et al. Alterations of serum bile acid profile in breast-fed infants with prolonged jaundice. J Pediatr Gastroenterol Nutr. 1985;4(5): 741–5. doi: 10.1097/00005176-198510000-00010 [DOI] [PubMed] [Google Scholar]
- 24.Zhang B, Yang X, Gao X. Taurine protects against bilirubin-induced neurotoxicity in vitro. Brain Res. 2010;1320(159–67. doi: 10.1016/j.brainres.2010.01.036 [DOI] [PubMed] [Google Scholar]
- 25.Li Y, Zhao D, Qian M, Liu J, Pan C, Zhang X et al. Amlodipine, an anti-hypertensive drug, alleviates non-alcoholic fatty liver disease by modulating gut microbiota. Br J Pharmacol. 2022;179(9): 2054–77. doi: 10.1111/bph.15768 [DOI] [PubMed] [Google Scholar]
- 26.Ren Y, Xu Z, Qiao Z, Wang X, Yang C. Flaxseed Lignan Alleviates the Paracetamol-Induced Hepatotoxicity Associated with Regulation of Gut Microbiota and Serum Metabolome. Nutrients. 2024;16(2). doi: 10.3390/nu16020295 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yu H, Guo Z, Shen S, Shan W. Effects of taurine on gut microbiota and metabolism in mice. Amino Acids. 2016;48(7): 1601–17. doi: 10.1007/s00726-016-2219-y [DOI] [PubMed] [Google Scholar]
- 28.Liu H, Chen X, Hu X, Niu H, Tian R, Wang H et al. Alterations in the gut microbiome and metabolism with coronary artery disease severity. Microbiome. 2019;7(1): 68. doi: 10.1186/s40168-019-0683-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Rohwerder T. New structural insights into bacterial sulfoacetaldehyde and taurine metabolism. Biochem J. 2020;477(8): 1367–71. doi: 10.1042/BCJ20200079 [DOI] [PubMed] [Google Scholar]
- 30.Rohwerder T. New structural insights into bacterial sulfoacetaldehyde and taurine metabolism. Biochem J. 2020;477(8): 1367–71. doi: 10.1042/BCJ20200079 [DOI] [PubMed] [Google Scholar]
- 31.Li Y, Zhao D, Qian M, Liu J, Pan C, Zhang X et al. Amlodipine, an anti-hypertensive drug, alleviates non-alcoholic fatty liver disease by modulating gut microbiota. Br J Pharmacol. 2022;179(9): 2054–77. doi: 10.1111/bph.15768 [DOI] [PubMed] [Google Scholar]
- 32.Chen T, Qin Y, Chen M, Zhang Y, Wang X, Dong T et al. Gestational diabetes mellitus is associated with the neonatal gut microbiota and metabolome. BMC Med. 2021;19(1): 120. doi: 10.1186/s12916-021-01991-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sun H, Zhang AH, Song Q, Fang H, Liu XY, Su J et al. Functional metabolomics discover pentose and glucuronate interconversion pathways as promising targets for Yang Huang syndrome treatment with Yinchenhao Tang. RSC Adv. 2018;8(64): 36831–9. doi: 10.1039/c8ra06553e [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Roberts MS, Magnusson BM, Burczynski FJ, Weiss M. Enterohepatic circulation: physiological, pharmacokinetic and clinical implications. Clin Pharmacokinet. 2002;41(10): 751–90. doi: 10.2165/00003088-200241100-00005 [DOI] [PubMed] [Google Scholar]
- 35.Ao T, Cantor AH, Pescatore AJ, Ford MJ, Pierce JL, Dawson KA. Effect of enzyme supplementation and acidification of diets on nutrient digestibility and growth performance of broiler chicks. Poult Sci. 2009;88(1): 111–7. doi: 10.3382/ps.2008-00191 [DOI] [PubMed] [Google Scholar]
- 36.Dibner JJ, Buttin P. Use of organic acids as a model to study the impact of gut microflora on nutrition and metabolism. JOURNAL OF APPLIED POULTRY RESEARCH. 2002;11(4): 453–63. doi: 10.1093/japr/11.4.453 [DOI] [Google Scholar]
- 37.Gourley GR, Li Z, Kreamer BL, Kosorok MR. A controlled, randomized, double-blind trial of prophylaxis against jaundice among breastfed newborns. Pediatrics. 2005;116(2): 385–91. doi: 10.1542/peds.2004-1807 [DOI] [PubMed] [Google Scholar]
- 38.Okumura S, Teratani T, Fujimoto Y, Zhao X, Tsuruyama T, Masano Y et al. Oral administration of polyamines ameliorates liver ischemia/reperfusion injury and promotes liver regeneration in rats. Liver Transpl. 2016;22(9): 1231–44. doi: 10.1002/lt.24471 [DOI] [PubMed] [Google Scholar]
- 39.Brandoni A, Torres AM. Expression of renal Oat5 and NaDC1 transporters in rats with acute biliary obstruction. WORLD JOURNAL OF GASTROENTEROLOGY. 2015;21(29): 8817–25. doi: 10.3748/wjg.v21.i29.8817 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hansen T, Tommarello S. Effect of phenobarbital on bilirubin metabolism in rat brain. BIOLOGY OF THE NEONATE. 1998;73(2): 106–11. doi: 10.1159/000013966 [DOI] [PubMed] [Google Scholar]
- 41.Long Y, Dong X, Yuan Y, Huang J, Song J, Sun Y et al. Metabolomics changes in a rat model of obstructive jaundice: mapping to metabolism of amino acids, carbohydrates and lipids as well as oxidative stress. JOURNAL OF CLINICAL BIOCHEMISTRY AND NUTRITION. 2015;57(1): 50–9. doi: 10.3164/jcbn.14-147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Yan S, Yin X. Gut microbiome in liver pathophysiology and cholestatic liver disease. LIVER RESEARCH. 2021;5(3): 151–63. doi: 10.1016/j.livres.2021.08.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Nicholson JK, Holmes E, Kinross J, Burcelin R, Gibson G, Jia W et al. Host-Gut Microbiota Metabolic Interactions. SCIENCE. 2012;336(6086): 1262–7. doi: 10.1126/science.1223813 [DOI] [PubMed] [Google Scholar]
- 44.Chen W, Wei Y, Xiong A, Li Y, Guan H, Wang Q et al. Comprehensive Analysis of Serum and Fecal Bile Acid Profiles and Interaction with Gut Microbiota in Primary Biliary Cholangitis. CLINICAL REVIEWS IN ALLERGY & IMMUNOLOGY. 2020;58(1): 25–38. doi: 10.1007/s12016-019-08731-2 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(XLSX)
Data Availability Statement
All relevant data are within the manuscript and its Supporting Information files.