Low-grade inflammatory diseases revealed metabolic perturbations that have been linked to various phenotypes, including gut microbiota dysbiosis. Metaproteomics has been used to investigate protein composition profiles at specific steps and in specific healthy/pathologic conditions. Metaproteomics allows researchers to build a more complete overview on protein composition at a specific time (fingerprint) and in specific health conditions, especially when used in combination with meta-omics approaches. Metaproteomics approach and pathway modulation in obesity and diabetes are described.
1. Introduction
Gut microbiota modulates the innate and adaptive immune systems both locally in the intestinal mucosa and outside the gut. Variations in microbial pattern abundance have been associated with certain autoimmune or inflammatory diseases known as ‘metabolic diseases’. In this field, type 1 and type 2 diabetes (T1D and T2D, respectively) and obesity are the most common and prevalent diseases featured by metabolic perturbations also involving gut microbiota dysbiosis
[1][2]. Evidence-based data revealed how changes in gut microbiome contribute to an increased susceptibility to the onset and development of several diseases
[3]. The main actors of these mechanisms are the colonic microbiota, its metabolic products, and the host immune system
[4].
Researchers are here referring to pathologies mainly featured by an imbalance in the Bacteroidetes/Firmicutes ratio
[5]. Some species belonging to these phyla are responsible for the production of short chain fatty acids (SCFAs), such as butyrate, propionate, and acetate
[6]. The imbalance in these phyla abundance can impact gut epithelial integrity, leading to an increased permeability and undermining the immune homeostasis and the inflammatory response
[7]. On the other hand, alterations in the abundance of specific microbial patterns may affect saccharolytic, proteolytic, and lipolytic metabolisms and may influence the expression of involved enzymatic pathways. However, the crosstalk interaction among all the mentioned factors has not been completely clarified yet.
Although shotgun 16S rRNA marker gene sequencing delivers interesting insights on human microbiota communities
[8], it does not provide information about microbiome plasticity, especially when the adaptation to specific and mutable niche conditions is required
[9]. Metaproteomics, instead, provides findings on (i) microbial constituents, (ii) the interaction between gastrointestinal (GI) microbiota and the host proteome, (iii) signal transduction, and (iv) metabolic pathways
[9]. Functional shifts in microbial and human protein profiles can be further detected by using specific and curated databases allowing the identification of novel diagnostic targets and specific disease biomarkers
[9][10].
Obesity and diabetes mellitus are both associated with inflammation of different tissues and organs
[11]. Seeking inflammatory factors related to T1D progression, some studies highlighted findings on C-reactive protein (CRP) levels
[12]. An increase in the monocyte release of interleukin (IL)-1β and superoxide radicals were also reported, suggesting an up-regulation of the inflammatory activity
[13]. Besides, inflammatory processes contribute to insulin resistance in T2D. Considering that obesity is also a risk factor for the development of T2D, a large number of proteins synthesized during the inflammatory state as CRP, adipocyte-derived metabolites such as lipids, fatty acids, adipocytokines, and various inflammatory cytokines (TNF-α, IL-1β, and IL-6), have been linked to the development of insulin resistance
[13][14][15].
Peripheral blood mononuclear cells (PBMC) are also involved in the crosstalk between inflammation response and dysbiosis. Pro- and anti-inflammatory activities of PBMC could be mediated by the exposure to microbial-derived SCFAs
[16] or to lipopolysaccharides (LPS) coating Gram-negative bacteria
[17]. Interesting insights about this mutual interaction, also in individuals affected by metabolic disease as T2D, were reported
[18].
Metaproteomics allows researchers to build a more complete overview on protein composition at a specific time (fingerprint) and in specific health conditions, especially when used in combination with the above-mentioned meta-omics approaches
[18][19].
However, metaproteomics is facing several methodological challenges due both to the ever-increasing amount of data constantly produced and the lack of standardized protocols for downstream data analysis. Advisedly, a standardized workflow is necessary to compare metaproteomics outputs belonging from different studies. This will lead to the inspection of specific associations between gut microbiota functional variations and the obesity/diabetes states. To provide a critical overview on the topic, this entry has been conducted considering those studies that, in the last eleven years, applied metaproteomics to investigate the onset and progression of diabetes and obesity.
2. Metaproteome Alteration in Obesity
A total of four studies assessed metaproteome variations in obesity (
Table 1). Both metagenomics and metaproteomics data in Ferrer et al. revealed an increase of the relative abundances of Firmicutes and a decrease of Bacteroidetes in obese patients compared with lean individuals
[20]. However, in line with Kolmeder et al., 2015, the relative amounts of expressed proteins from both phyla were very similar in obese and lean individuals. In addition, differences between the two subject groups were observed for proteins involved in cell motility, butyrate production, vitamin synthesis (B6 and B12), and starch metabolism (
Table 1). The same study showed that most of the detected alterations were associated with an increased energy production by the obese gut microbiota, as indicated by butyrate production and some pili-forming proteins and flagellins that might facilitate the microbial access to carbohydrates
[20]. Moreover, Kolmeder et al., 2015 reported how peptides derived from proteins involved in C5 and C6 carbohydrate metabolism, (e.g., enolase, ribulokinase, xylulokinase, phosphoketolase, and a specific glycoside hydrolase) were more abundant in non-obese individuals. Additionally, obese subjects had much more proteins involved in starch and pectin metabolism (glucosidase and pectate lyase).
On the other hand, fructose, mannose, galactose, and sucrose metabolisms resulted up-regulated in subjects belonging to the obese cohort investigated by Hernandez et al., 2013
[21]. In detail, a higher total sugar metabolism, assessed by a colorimetric assay with a set of 23 structurally diverse sugars, and a major activity of glycosidase were detected in extracted proteins from stool samples of obese individuals compared to those of lean subjects. Moreover, both Hernandez et al., 2013 and Sanchez-Carrillo et al., 2020
[22][21] highlighted a significant alteration in the expression of proteins linked to metabolic derangement, intestinal damage, and chronic inflammation state (alkaline phosphatase, serpins, and α-amylase more expressed in obese patients than healthy individuals or individual post bariatric surgery). Additionally, Sanchez-Carrillo et al., 2021 found ferritin and ferrous transport protein to be expressed in lean adults (0.46–1.0 ng/g) while both proteins resulted below the detection threshold in individuals with severe obesity.
Table 1. Summary of metaproteome variation in terms of significantly up- and down-regulated proteins in gathered/filtered studies (n = 10) and other metabolism pathways.
Authors and Year |
Disease |
Protein Origin |
↑ Proteins |
↓ Proteins |
Metabolic Pathway/Functionality |
Gavin et al., 2018 |
T1DM |
Microbial |
1. Enzymes for mucin degradation 2. Elongation factor 3. Ferredoxin reductase |
4. Transferases (butyrate synthesis) |
3.↑Ferredoxin catabolism 4. ↑ Butyrate anabolism |
|
|
Human |
1. Galectin-3 2. Fibrillin |
3. CELA-3A, 4. CUZD1 5. CLCA1 6. Neutral ceraminidase 7. IGHA1 |
6.↓ Sphingosine (SPH) and sphingosine 1-phosphate (S1P) 3.4.5. ↓ exocrine pancreas functionality 7.↓ IgA |
Pinto et al., 2017 |
T1DM |
Microbial |
1. ilvE (BCAA transaminase) 2. Glutamate dehydrogenase (AA degradation) 3. Bifunctional GMP synthase 4. Glutamine amido transferase 5. Chaperonin GroEL |
6. Phosphoketolas 7. Glyceraldehyde-3-phosphate dehydrogenase, 8. Transketolase |
1.6.8. ↓ Via penthos phosphate → ↑ BCAA synthesis (Shikmic Acid Pathway) ↓ glycolysis 2.↑ NH4+ (Urea) 7. ↓ Glycolysis →↓ Piruvate ↓ SCFAs |
|
|
Human |
MUC2 precursor |
CELA-3A |
↑ Intestinal mucin-2 ↓ Exocrine pancreas functionality |
Heintz et al., 2016 |
T1DM |
Microbial |
|
Thiamine synthesis cofactor |
↓ Thiamine synthesis |
|
|
Human |
|
↓ AMY2A, AMY2B, CPA1, and CUDZ1 |
↓ Complex sugar degradation |
Singh et al., 2017 |
T1DM |
Human urinary proteome |
1. LGR1 2. CD14 3. CPE 4. CTSB 5. CTSD 6. NAGA |
7. Fibronectin-1 8. Pancreatic α-amylase 9. MUC1 10. PTPRN |
1. ↑Inflammatory pathways (TGF-β) 3.↑AA degradation (↑urea production) 8. ↓ Exocrine pancreas functionality and ↓ complex sugar metabolism |
Zhong et al., 2019 |
T2DM |
Microbial |
1. PTS 2. ABC transporter 3. FMO3 (TMAO producing enzyme) |
4. Ferredoxin oxidoreductase 5. Bacterial ribosomal proteins |
1.↑Phosphorylation and transport of sugar in microbial cells 2.↑ HDL cholesterol 3.↑TMAO synthesis |
Zhou et al., 2019 |
T2M |
Human |
1. IL-1RA 2. CRP 3. A1C |
|
1.↓IL-1 2.↑immune defense mechanism 3.↑ glycaemia |
Ferrer et al., 2013 |
Obesity |
Microbial |
1. Glycoprotein containing FN3 2. Cobaltochelatases 3. B12-dependent methylmalonyl-CoA mutase 4. PduB 5. 3-hydroxybutyryl-CoA dehydratase 6. Butyryl-CoA dehydrogenases 7. Acetyl-CoA acetyltransferases |
7. Pectate lyase 8. Aldose 1-epimerase 9. SOD 10. Pyridoxal biosynthesis lyases |
1.↑ Fibrin and proteoglycans 2.3. ↑ Vitamin B12 and propionate production 4.↑ Propanediol catabolism 5. Butyrate 10. ↓ Vitamin B6 |
Kolmeder et al., 2015 |
Obesity |
Microbial |
1. α-glucosidase 2. Pectate lyase 3. Aminoacyl-histidine dipeptidase 4. Bacteroidetes proteins |
|
1.2. ↑Starch and pectin metabolism 3. ↑AA metabolism 4. ↑SCFAs |
|
|
Human |
1. Trehalase (intestinal injury and inflammation marker) 2. Alkaline phosphatase (AP) 3. Serpins (serina protease inhibitors) 4. α-amylase |
|
1.↑ Threalosie 4.↑Starch digestion |
Sanchez-Carrillo et al., 2020 |
Obesity |
Microbial (pre-BS) |
1. Enzymes involved in gluconeogenesis (glyceraldehyde 3-phosphate dehydrogenase, pyruvate orthophosphate dikinase, PEP carboxykinase, fructose-bisphosphate aldolase, glutamate dehydrogenase) 2. Enzymes involved in Acetyl-CoA synthesis (Formate C-acetyltransferase, acetyl-CoA synthase, carbon-monoxide dehydrogenase) 3. Ferredoxin oxidoreductase |
4. Ferritin 5. Ferrous ion transport protein 6. Porphobilinogen synthase |
1.↑Pyruvate 2.↑ Acetyl-CoA (WL pathway) 4.5. ↓Iron synthesis |
|
|
Microbial (post-BS) |
1. AdhE 2. OhyA 3. SOD and perodoxins (involved in maintenance of redox balance) |
|
1. ↑ Acetyl Acteyl-CoA → ethanol 1. Saturated fatty acid |
Hernandez et al., 2013 |
Obesity |
Microbial |
1. α-polyglucose (refined carbohydrate digestion) 2. Proteins involved in pentose phosphate metabolism (PPP) 3. Proteins involved in TCA cycle |
|
1.2. ↑ Fructose, mannose, galactose, sucrose, starch, amino sugar, and nucleotide sugar metabolism 3. ↑ Via pentose phosphate |
3. Metaproteome Alteration in T1D
As above described, four studies investigated the metaproteome of subjects affected by T1D
[23][24][25][26] (
Table 2).
Table 2. Baseline characteristics of the included studies and Newcastle-Ottawa Quality Assessment Scale (NOS).
Authors and Year |
Size Sample and Characteristics |
Subjects Characteristics (Sex, Age, Country) |
Scope of Study |
Study Design |
Metaproteomics Techniques Used |
Other “Omics” Techniques Used |
Limitations |
NOS Score |
Gavin et al., 2018 |
101 subjects: 33 NO, 17 SP, 29 SN, 22 CO |
Denver, Colorado 46 females and 55 males Age: 9–12 |
Investigate functional interactions host-microbiota in subjects with T1D risk |
Cross-sectional |
LC-MS/MS |
|
No information about dietary intake. Wide age range. |
7 |
Pinto et al., 2017 |
6 subjects: 3 healthy and 3 T1D children |
Portugal 2 males and 4 females Age: T1D children 9.3 ± 1.5 and control children 9.3 ± 0.6 years |
Identify differences in the activity of intestinal microbiota between healthy and T1D children |
Case-control |
SDS-PAGE and LC-MS/MS (using LTQ Orbitrap) |
|
Small number of T1D children. |
6 |
Heintz et al., 2016 |
20 subjects from 4 families of at least 2 generations presenting at least 2 cases of T1D |
Luxembourg 7 males 13 females Age: 5–62 |
Resolution of the taxonomic and functional attributes of gut microbiota and evaluation of the effect of family on gut microbiota composition |
Longitudinal study (4 month) |
LC (Nano-2D-UPLC-Orbirtap MS system) and MS (TopN-MS/MS method) |
Metagenomics and metatranscriptomics |
Need for large-scale studies. |
6 |
Singh et al., 2017 |
223 subjects: 110 T1D children/adolescents and 113 healthy siblings |
Washington D.C. 115 males 108 females Age: 13.9–14.5 |
Detection of gut microbial differences and evaluation of lysosomal dysfunctions |
Case-control |
LC-MS/MS |
|
Imperfect glycemic control or subclinical inflammation in T1D patients. No information about eating habits and lifestyle. |
7 |
Zhong et al., 2019 |
254 subjects: 77 TN-T2D, 80 Pre-DM, and 97 NGT |
Suzhou, China 173 females 81 males Age: 41–86 |
Investigate compositional and functional changes of the gut microbiota to characterize different disease stages |
Cross-sectional |
iTRAQ-coupled- LC-MS/MS |
Metagenomics |
Limitations of MS-based proteomics. Confounding variables: age, drugs (CCB, hypertension, and dyslipidemia), diet, and health conditions. |
7 |
Zhou et al., 2019 |
106 subjects: healthy and pre-diabetic adults |
Standford, California 55 females and 51 males Age: 25–75 |
Understand how healthy individuals and those at risk of T2D, change over time, in response to perturbations and in relation to different molecules and microorganism |
Longitudinal study (4 years) |
SWATH-MS |
Metagenomics, metatranscriptomics, and metabolomics |
Limited studies of microbial changes. No information about diet and exercise. Heterogeneous data. |
6 |
Ferrer et al., 2013 |
2 subjects: 1 lean, 1 obese |
Spain 1 female (lean) and 1 male (obese) Age: 15 |
Identify and analyze active bacterial members and proteins expressed in lean and obese microbiota |
Case-control |
1D-gel electrophoresis and UPLC-LTQ Orbitrap-MS/MS |
Metagenetics |
No information about dietary intake. |
7 |
Kolmeder et al., 2015 |
29 subjects: 9 lean, 4 overweight, 16 obese |
Spain 21 females 8 males Age: 23.1 ± 2.2 (non-obese); 38.6 ± 2.4 (obese) |
Characterization of non-obese and obese fecal metaproteome |
Case-control |
1D-gel electrophoresis RP-HPLC online coupled to MS/MS |
|
Regular medication between obese and non-obese group. |
6 |
Sanchez-Carrillo et al., 2020 |
40 severely obese adults subjected to BS |
Spain Age: 47–60 |
Investigation the impact of bariatric surgery |
Longitudinal study (3 months) |
LC-ESI-MS/MS analysis |
Metabolomics |
Results biased for using pooling strategy. |
6 |
Hernandez et al., 2013 |
13 subjects: 2 adults (β-lactam-therapy), 7 obese adolescents, 5 lean adolescents |
Germany Obese: 3 females and 4 males Lean: 2 males and 3 females Age: 13–16 |
Evaluation of microbial shifts in relation to antibiotic treatment and obesity and measurement of carbohydrate activate enzymes |
Cross-sectional |
96-well plates using a BioTek Synergy HT spectrometer in a colorimetric assay |
|
No information about dietary intake. Wide age range. Small number of subjects. |
6 |
Pinto et al., 2017 found that microbial metaproteome variations of children affected by T1D were originated from Eubacterium, Faecalibacterium, and Bacteroides. The presence of these bacterial taxa is mainly linked to amino acids transport, metabolism and transcription, protein turnover, and chaperones. Specifically, the branched-chain amino acid transaminase (ilvE) and the glutamate dehydrogenase enzymes were detected among those proteins found to be more abundant in T1D subjects than healthy individuals. These proteins are involved in amino acids transport and metabolism. Additionally, regarding host proteins, T1D patients exhibited an increased expression of mucin-2 and a reduction in elastase 3A expression with respect to healthy individuals, suggesting both an increased mucin synthesis in charge of gastrointestinal epithelium protection and a reduction in exocrine pancreas functionality (Table 1).
Singh et al., 2017 considered 223 children and adolescents (age range 5–22) and observed a significant depletion of the genus Enterococcus in T1D subjects with high levels of HbA1c compared with healthy individuals. In this study, the metaproteomics was used to investigate the urinary proteome alterations in T1D subjects compared to their healthy siblings. Increased levels of lysosomal enzymes were associated with HbA1c levels. Together with these, some other proteins involved in inflammatory responses were more expressed in T1D patients. Specifically, LRG1 and CD14 resulted in the adipose tissue inflammation.
The study conducted by Gavin et al., 2018 described the alterations of both host and microbial proteins collected from children and adults affected by T1D. Five human proteins exhibited a lower level in new-onset diabetics (NODs) and seropositive individuals (positive for islet autoantibodies) compared to control subjects; the same trend was reported for three proteins secreted by the exocrine pancreas (Table 1). Moreover, two human proteins associated with inflammation, fibrillin-1, and galectin-3, were overexpressed in the T1D group. As far as concerns microbial proteins, several were differentially expressed in diseased and control individuals. The relative KEGG assignment to specific categories showed how the great part of them belongs to the phosphotransferase system, thermo-unstable elongation factors, ferredoxin hydrogenase class, and butyrate synthesis metabolism. These data revealed that proteins altered in NODs and seropositive individuals were involved in the inflammation onset, increasing the mucus secretion and the defective mucosal barrier function.
Finally, Heintz and co-workers
[25], by applying a multi-omics approach to resolve the taxonomic and functional attributes of gut microbiota at the metagenomic level, found lower levels of specific human exocrine pancreatic proteins in T1D subjects (α-amylase proteins AMY2A, AMY2B, and carboxypeptidase CPA1) compared to the healthy ones. At any rate, the study was unable to identify taxa whose abundance levels correlated with those relative to pancreatic enzymes.
4. Metaproteome Alteration in T2D
Zhong et al., 2019
[18] investigated the compositional and functional changes of gut microbiota in pre-diabetic (Pre-DM), treatment naïve T2D (TN-T2D), and healthy individuals cross-sectionally to elucidate different mechanisms linked to the disease stages (
Table 1). A reduction of pancreatic enzymes in Pre-DM and TN-T2D, compared to healthy individuals, was detected and implied a reduced exocrine pancreas functionality. A substantial number of Pre-DM associated microbial and human proteins were identified at the metagenomics and metaproteomics level. In fact, an enrichment in the structural domains of microbial proteins modules involved in the sugar phosphotransferase system (PTS), ATP-binding cassette (ABC) transporter of amino acids, and bacterial secretion system was detected in Pre-DM compared to normal glucose transport (NGT) individuals.
Besides, alterations in human protein production among the three groups of analyzed individuals were highlighted. The trimethylamine-N-oxide producing enzyme (FM03) was exclusively detected in the TN-T2D group. In the same group, a loss of rasGTPase-activating-like protein (IQGAP1) and unconventional myosin-Ic (MYO1C), related to the impairment of insulin signaling, were detected.
The longitudinal study by Zhou et al., 2019
[27] aimed at understanding the early disease stages of diabetes profiles by inspecting transcriptomes, metabolomes, cytokines, proteomes, and changes in the microbiome of 106 healthy subjects and individuals with pre-diabetes for four years. Correlations between microbial taxa and specific cytokines were highlighted and microbial–cytokine correlations resulted in being significant in insulin-sensitive but not in insulin-resistant participants. Barnesiella was positively associated with IL-1β (q = 0.0054), Faecalibacterium was inversely associated with TNFA (q = 0.0244), and Butyricimonas was negatively associated with four lipids only in insulin-resistant participants (q < 0.05). The study revealed that many host biochemical and microbial components are stable over time in healthy individuals even though they can undergo dynamic and marked changes in response to viral infection or other perturbations. These changes differed between insulin-sensitive and insulin-resistant individuals. By integrating information provided from proteins, cytokines, and metabolites, the pathways associated with defense responses, such as interleukin signaling pathways, mTOR signalling23, and B and T cell receptor signaling, were identified. Furthermore, during a viral infection, inflammatory pathways were differently altered in insulin-resistant participants with respect to insulin-sensitive individuals. This would suggest the presence of alterations in defense responses.
This entry is adapted from the peer-reviewed paper 10.3390/nu14010047