Integrated Label-Free and 10-plex DiLeu Isobaric Tag Quantitative Methods for Profiling Changes in the Mouse Hypothalamic Neuropeptidome and Proteome: Assessment of the Impact of the Gut Microbiome

Abstract

Gut microbiota can regulate host physiological and pathological status through gut-brain communications or pathways. However, the impact of the gut microbiome on neuropeptides and proteins involved in regulating brain functions and behaviors is still not clearly understood. To address the problem, integrated label-free and 10-plex DiLeu isobaric tag-based quantitative methods were implemented to compare the profiling of neuropeptides and proteins in the hypothalamus of germ-free (GF) vs. conventionally-raised (ConvR) mice. A total of 2,943 endogenous peptides from 63 neuropeptide precursors and 3,971 proteins in the mouse hypothalamus were identified. Among these 368 significantly changed peptides (fold changes over 1.5, and p-value < 0.05), 73.6% of the peptides showed higher levels in GF-mice than in ConvR-mice, and 26.4% of the peptides had higher levels in ConvR-mice than in GF-mice. These peptides were mainly from Secretogranin-2, Phosphatidylethanolamine-binding protein-1, ProSAAS, and Proenkephalin-A. Quantitative proteomic analysis employing DiLeu isobaric tags revealed that 282 proteins were significantly up- or down-regulated (fold changes over 1.2, and p-value < 0.05) among the 3,277 quantified proteins. These neuropeptides and proteins were mainly involved in regulating behaviors, transmitter release, signaling pathways, and synapses. Interestingly, pathways including long-term potentiation, long-term depression, and circadian entrainment were involved. In the present study, a combined label-free and 10-plex DiLeu-based quantitative method enable a comprehensive profiling of gut microbiome induced dynamic changes of neuropeptides and proteins in the hypothalamus, suggesting that the gut microbiome might mediate a range of behavioral changes, brain development, learning and memory through these neuropeptides and proteins.

Graphical Abstract

Introduction

Neuropeptides are a class of endogenous peptides that act as long-lasting cell-to-cell signaling neurotransmitters in the nervous system and other target organs. Neuropeptides and their receptors play important roles in several key processes¹, such as modulating neurobehavior², social development³, affecting energy harvest from the diet and energy storage⁴, inducing systemic insulin resistance, and modifying glucose homeostasis and the immune response^{5, 6}. Neuropeptides and their precursors are important mediators both within the nervous system and between neurons and other cell types. When a neuron releases neuropeptides, the binding of the neuropeptide to its receptor on a receiving cell causes conformational changes within the receptor that, depending on the type of receptor, either open ion channels or activate coupled G-proteins that can cause a series of downstream effects within the cell⁷.Furthermore, neuropeptides have been implicated in the regulation of normal biological functions such as circadian rhythm⁸, feeding regulation⁹, temperature fluctuation¹⁰, depression, stress, anxiety¹¹, cognition¹², and other behaviors¹³. For instance, pituitary adenylate cyclase-activating polypeptide (PACAP) has neuroprotective and neurotrophic properties that could slow down Alzheimer’s disease-like pathology; increasing or restoring PACAP and its receptor function might provide Alzheimer’s disease treatment benefits¹⁴.

Over the past decade, accumulating evidence has suggested that the gut microbiota can affect the host’s physiological status through “brain-gut axis” pathways^15–17. It has been demonstrated that central nervous system (CNS) neurotransmission can be profoundly affected by the absence of normal gut microbiota¹⁸. Gut microbiota can regulate mouse behavior through the hippocampal glucocorticoid receptor pathway¹⁹. Several studies indicate that gut microbes regulate host physiological and pathological status through bi-directional communication, along the gut-brain axis. Microbiota leading to CNS influences might be associated with second messenger pathways and synaptic long-term potentiation (LTP) in brain regions^{19, 20}. Bercik et al. reported that gut microbiota influence behavior and brain chemistry through the autonomic nervous system, gastrointestinal-specific neurotransmitters, or inflammation²¹.

These diverse neurochemical messengers and regulators, including neurotransmitters, neuropeptides, hormones, proteins, and receptors, can be completely different under different gut microbiota environments^22–24. Among these gut-influenced brain molecular players, neuropeptides and proteins are most often associated with regulation of behavior, neural interconnection, neurodegeneration, neuronal development, and neuroimmunomodulation²⁵. However, to date, there has been little research on determining the effects of gut microbiota on brain proteomics and neuropeptidomics, or on possible pathways of gut-brain communications at the protein or neuropeptide level.

MS-based quantitative proteomics or peptidomics is used to characterize relative peptide abundances across different biological conditions and can be implemented using label-based or label-free methods^{1, 26, 27}. We have developed our own cost-effective amine-reactive N, N-dimethyl leucine (DiLeu) tags for multiplexed quantification of many samples in a single LC-MS/MS. The DiLeu tags consist of an N, N-dimethyl leucine reporter group, a balance group, and a triazine ester amine-reactive group to selectively label peptide N-termini and lysine side chains. Multiplex DiLeu quantitation is powerful and enables enhanced performance in terms of throughput and reproducibility. Compared with commercial tags, such as tandem mass tags (TMT) and isobaric tags for relative and absolute quantification (iTRAQ), DiLeu isobaric tags offer high multiplexing capacity. DiLeu isobaric tags enable up to 21-plex quantification during high-resolution MS/MS acquisition²⁸. Our DiLeu reagents provided good coefficients of variation (CVs) of 7.9% for 1:1 ratio samples and 11.5%for the 16:1 ratio samples, respectively, and showed good reproducibility and variance of the protein quantitative ratios between 16:1 replicates²⁶. 12-plex DiLeu tags have been employed to develop a highly multiplexed quantitative proteomic and phosphoproteomic method to globally assess protein expression and phosphorylation changes in smooth muscle cells. The multiplexed DiLeu isobaric tags enabled analysis of multiple samples in a single assay to reduce run-to-run variability and facilitate high-throughput quantification²⁹. Compared with current commercial multiplexed isobaric labels, DiLeu isobaric labeling reagents offer an attractive due to their cost-effectiveness, comparable performance, high proteome coverage, quantification efficiency, high multiplexing capability^{26, 28, 30}.

In this study, neuropeptides and proteins were extracted from the same mouse brain hypothalamus region, and the relative quantitative changes of neuropeptides and proteins between different groups were measured and compared based on precursor ion intensity from MS spectra or DiLeu isobaric tag reporter ion intensities from MS/MS spectra. On the one hand, this integrated label-free and 10-plex DiLeu isobaric tag quantitative method enabled profiling of dynamic changes of neuropeptides and proteins simultaneously in the hypothalamus of mice with distinct gut microbiota environments; on the other hand, possible connections and interplay between neuropeptides and proteins could be established with this combined approach. This approach provided further evidence on how the gut microbiota regulates brain signaling molecules (peptides and proteins), potentially revealing novel host functions that are under microbial regulation.

Experimental

Animal Experiment and Tissue Extraction

Animal care and study protocols were approved by the UW–Madison Animal Care and Use Committee. Mice were group-housed by colonization status in temperature-controlled rooms with 12-h light: 12-h dark cycle and received ad-libitum access to water and food. Mice were fed autoclaved LabDiet #5021 (Purina Mills, Inc. Richmond, IN). Twenty-one-week-old male C57BL/6 GF-mice and ConvR-mice (n=5/group) were euthanized after 4-h fast via CO₂ inhalation and exsanguination. Their brains were dissected and immediately rapid-heated via Denator™ to minimize postmortem degradation, after which the hypothalamus region was isolated and snap frozen in liquid nitrogen immediately, and stored in the −80°C until analysis.

Detailed descriptions of tissue extraction, protein digestion, and LC-MS instrument operation methods can be found in the Supporting Information. DiLeu isobaric tag synthesis and labeling are described in a previous publication²⁶.

Statistical Analysis

The Student’s t-test (two-tailed) was performed for comparisons between two groups of independent samples. A p-value <0.05 was considered statistically significant.

Results and Discussion

Label-free quantification of endogenous peptides in the hypothalamus region in GF- and ConvR-mice

An integrated label-free quantification (LFQ) and 10-plex DiLeu isobaric tag-based quantification strategy was developed; the workflow of this approach is illustrated in Figure 1. In total, 2,943 endogenous peptides derived from 63 precursors from the hypothalamus region of both GF- and ConvR-mice were identified. A complete summary of all the identified peptides is provided in Table S1. Highly confident peptide identification was achieved by high-resolution, accurate-mass MS/MS analysis using a nanoLC Orbitrap platform (Figure S1).

Figure 1.

Workflow for the integrated label-free quantification (LFQ) and the 10-plex DiLeu-based strategy. Germ-free (GF)-mice and conventionally-raised (ConvR)-mice were euthanized and the hypothalamus region was harvested. Neuropeptides in hypothalamus were extracted and analyzed, pellets after neuropeptide extraction were lysed, digested, and analyzed as well. Then label-free and DiLeu-label quantitative methods were applied to compare the relative changes in neuropeptides and proteins in the hypothalamus region of GF- and ConvR-mice. Bioinformatic analysis was used to reveal the possible pathways that neuropeptides and proteins are involved under different microbiota status.

Among the 2,943 identified endogenous peptides, as shown in Figure 2A, 880 peptides were detected exclusively in the GF-mice samples and were mainly processing products derived from well-known neuropeptide precursors, including SCG2, pro-opiomelanocortin (POMC), ProSAAS (PCSK1N), phosphatidylethanolamine-binding protein 1 (PEBP1), neuroendocrine protein (SCG5), vasopressin-neurophysin 2-copeptin (AVP), pro-thyrotropin-releasing hormone (TRH), secretogranin-1 (SCG1), and somatostatin (SST) (Figure 2D). A total of 387 peptides were detected exclusively in the ConvR-mice samples and included mainly SCG2, PEBP1, SCG1, proenkephalin-A (PENK), SST, PCSK1N, cholecystokinin (CCK), and VIP peptides (VIP) (Figure 2C). Over half the total, or 1,676 peptides, were identified in both GF- and ConvR-mice samples, in which the SCG2 precursor accounted for the highest number of peptides (535), followed by PEBP1, PCSK1N, SCG1, PENK, SCG5, and SST (Figure 2E). The average length of endogenous peptides identified exclusively in ConvR-mice was 28.2 amino acids, whereas it was 24.0 for GF-mice and 23.0 for peptides identified in both sample sets (Figure 2B).

Figure 2.

Summary of the endogenous peptides identified from mice hypothalamus. (A) Venn diagram of endogenous peptides from Germ-free (GF)-mice and conventionally-raised (ConvR)-mice hypothalamus, 880 peptides were detected exclusively in the GF-mice samples, and 387 peptides were detected exclusively in the ConvR-mice samples. (B) Distribution of endogenous peptide in terms of peptide length in three categories, namely ConvR-mice only, GF-mice only, and Overlap between ConvR-mice and GF-mice. (C)-(E) Distributions of the numbers of peptides per precursor were shown after comma. For example, in panel (D), Scg2, 297 represented that in GF-mice only part, a total of 297 processed peptides with varying lengths were derived from neuropeptide precursor Scg2.

LFQ of endogenous peptides was carried out in this study to compare the level of peptides in the GF- and ConvR-mice hypothalamus. Peptides measured with ≥1.5-fold changes and Student’s t-test p-values <0.05 were determined (Figure 3A). In total, 368 peptides that were significantly changed were selected (Table S2); representative neuropeptides and preprohormone-derived peptides identified from the GF- and ConvR-mice hypothalamus are shown in Table S3. Among the 368 peptides, 271 peptides (73.6%) showed higher levels in GF-mice than in ConvR-mice, whereas 97 peptides (26.4%) showed higher levels in ConvR-mice than in GF-mice. These significantly changed peptides were mainly from SCG2 (82, 22.3%), PEBP1 (69, 18.8%), PCSK1N (72, 19.6%), SCG1 (20, 5.4%), and PENK (18, 4.9%). Peptides derived from SCG1, SCG2, and PCSK1N were generally present at higher levels in GF-mice, whereas peptides derived from PEBP1 and PENK were usually higher in ConvR-mice (Figure S2). Known bioactive neuropeptides were detected in the present study and quantified as well, such as orexin-B, joining peptide (J-peptide), corticotropin-like intermediate peptide (CLIP), γ-lipotropin, β-endorphin, α-MSH, neurokinin A, neurokinin B, galanin, substance P, nociceptin, Big LEN, Little LEN, Big SAAS, Little SAAS, and cerebellin. All of these bioactive neuropeptides were detected at relatively low levels in ConvR-mice, as shown in Figure 3B.

Figure 3.

(A) Volcano plot showing −log₁₀ p-values plotted against log₂ ratio (GF-/ConvR-). The horizontal line represents the t-test threshold of significance assigned (p < 0.05). The vertical lines mark the threshold of 1.5-fold up and down-regulated peptides. Some peptides had been reported previously, such as LLYEKMKGGQ (SCG5), IPVGSLKNEDTPN (SCG2), LVNAVGSGRSQSGPNGDRAA (SCG2), and YDGVAELDQLLHY (SCG1). (B) Relative abundance of endogenous peptides from the hypothalamus. For each peptide, the peak area was calculated and log₁₀ transformed, somatostatin-28 [1–12], Little SAAS, Big SAAS, CLIP, Big LEN, orexin-B, and α-MSH were present in the upper range, whereas deamidated Big LEN, galanin, neurokinin A, neurokinin B, β-endorphin, and cerebellin were present in the lower range. (C) Ratio of the relative abundance of endogenous peptides as determined by LFQ approach in GF- and ConvR-mice. For each peptide, the peak area was calculated and log₁₀ transformed. Values represent means ± SEM. * denotes p < 0.05, ** denotes p < 0.01, and *** denotes p < 0.001.

Neuropeptides are highly enriched in the CNS and involved in most physiological and psychological processes³¹. The hypothalamic-pituitary-adrenal (HPA) axis is a complex set of interactions among the hypothalamus, the pituitary gland, and the adrenal glands. The hypothalamus is considered as the starting point of the HPA axis; as previously reported, neuropeptides in the hypothalamus region play important roles in affecting food-intake behavior³², anxiety-like behavior and fear conditioning³³, body weight regulation³⁴, cardiac function, blood pressure regulation³⁵, circadian function⁸, control of the sleep-wakefulness cycle, and energy homeostasis³⁶. Therefore, we chose to focus on hypothalamus in this study to unveil the interaction between hypothalamus and gut microbiome. Zhang et al. reported that the brain peptidome could be altered by gut microbial regulation through the gut-brain axis³¹, which suggested potential future directions to expand our study of neuropeptidome and proteome changes in other brain regions, such as hippocampus, pituitary, striatum, etc. impacted by gut microbiome.

PCSK1N-derived peptides, such as Big SAAS, Little SAAS, PEN, and Big LEN, significantly decreased after food intake, and some other neuropeptides, such as galanin and SCG2-derived peptides, were decreased as well. Chromogranin-A (CHGA)-, protachykinin-1 (TAC1)-, and CART prepropeptide (CARTPT)-derived peptides significantly increased after food intake32. Most of these food-intake involved neuropeptides were also detected in the present study, as shown in Figure 3C. Several mature neuropeptides derived from the PCSK1N precursor, including Little LEN, Big LEN, Little SAAS, and Big SAAS, and also some truncated PCSK1N-derived neuropeptides, such as GAV [1-20], Big LEN [1-15], PEN [1-18], and Little SAAS [1-16], were quantified. These peptides were all down-regulated in colonized mice.

Representative mature neuropeptides were listed in Table S3, PCSK1N-derived peptides have previously been reported and expressed at a relatively high level in the mouse hypothalamus region. PCSK1N can lead to body weight changes through feeding control and fat deposition³⁴. As one of the most abundant neuropeptides in the brain, PCSK1N-derived Big LEN could regulate body weight by binding G-protein-coupled receptor 171 (GPR171), which might be associated with food intake and metabolism³⁷. A Big LEN-GPR171 system, were involved in mood-related behaviors, such as anxiety-like and fear behaviors³³. In addition, it has been reported that PCSK1N-derived Big SAAS, PEN, Big LEN, Little SAAS, and a truncated Little SAAS decreased about 40% in food intake group³².

Galanin (which decreased by 51.4% in ConvR-mice) detected in the present study was reported to be involved in the regulation of food intake and energy homeostasis³⁸. The neuropeptides nociceptin and nociceptin [1-11] detected in the present study decreased by 46.4% and 58.1% in ConvR-mice, respectively. It has been previously reported that nociceptin and its receptor can regulate learning and memory³⁹. Decreasing the activation of the nociceptin receptor by using a specific endogenous agonist could enhance memory³⁹, but nociceptin or nociceptin receptor agonists might impair memory consolidation⁴⁰. Interestingly, our current data might suggest that the relatively low nociceptin level in ConvR-mice could potentially help the mice enhance their learning and memory, although follow-up studies are needed to confirm or disapprove this speculation.

Neuropeptide orexin-B decreased by 54.4% in the hypothalamus of ConvR-mice compared to GF-mice, and orexin-B was suggested to be involved in the stimulation of food intake, sleep, and energy regulation^{36, 41, 42}. Neuropeptides derived from POMC were significantly decreased in expression levels with the existence of gut microbiota, such as J-peptide (decreased by 86.1%), CLIP (decreased by 83.0%), γ-lipotropin (decreased by 80.2%), β-endorphin (decreased by 77.0%), and α-MSH (decreased by 74.8%). The POMC neuron works in association with energy homeostasis and regulating feeding behavior through releasing peptides⁴³: α-MSH mediates the anorexigenic effect on feeding, whereas on the contrary, β-endorphin can promote food intake. These two neuropeptides might maintain a dynamic balance to regulate feeding behavior. It has also been reported that POMC-derived neuropeptides with feeding and obesity-suppressing functions were up-regulated in the hypothalamus of ConvR-mice compared with GF-mice, which was consistent with the present investigation²⁴.

Besides the well-known mature neuropeptides, many peptides identified in the present study were truncated or longer forms. It has been established that neuropeptides are proteolytic cleavage products from a precursor molecule; a single precursor can be differentially cleaved into different sets of peptides to form mature neuropeptides⁴⁴. These truncated peptides with differential cleavage levels might be related to distinct specificity and activity of the preprohormone convertases involved, but they showed similar change trends in the hypothalamus of GF- and ConvR-mice. For example, orexin-B decreased in ConvR-mice compared to GF-mice; similarly, truncated orexin-B forms, such as orexin-B [1-12], orexin-B [3-28], and orexin-B [14-28], were decreased by 56.4%, 51.7%, and 62.0%, respectively. We also observed similar patterns in truncated Somatostatin-28, truncated Big LEN, and Little SAAS. Notably, C-terminal amidation was observed in several mature neuropeptides, such as orexin-B and C-terminal-amidated orexin-B, neurokinin B, and its C-terminal amidated form. Nevertheless, for the J-peptide, α-MSH, neurokinin A, substance P, and galanin neuropeptides, only their C-terminal amidated forms were identified in the present study. The level of peptide E (fragment) series derived from the precursor PENK was significantly elevated by over 150% in ConvR-mice. Additionally, peptides derived from the precursor SCG1 (DPLQWKN, LFNPYFDPLQWKNSD, DGVAELDQLLHY, YDGVAELDQLLHY, etc.) were elevated from 52.4% to 266.3%. Precursor CCK-derived peptides APSGRMSVLKNLQSLDPSHRIS and APSGRMSVLKNLQSLDPS were elevated by 62.6% and 148.9%, whereas a peptide from SST, APSDPRLRQFLQKSLAAATGKQELAKYFLAE, was elevated by 263.1%.

In addition, a large number of active neuropeptides, truncated neuropeptides, or endogenous peptides with unknown activities were detected and quantified. In total, 84 SCG2-derived peptides were quantified. SCG2-derived peptides are known for their role in reproduction, food intake, and dopamine release⁴⁵. Neuropeptides derived from TAC1 and TAC3, including neurokinin A, neurokinin B, substance P, and neuropeptide K [1-23], decreased by 58.6%, 48.8%, 49.0%, and 40.5%, respectively. TAC1-derived peptides, including neurokinin A, substance P, and neurokinin B (TAC3-derived) are involved with the suprachiasmatic nucleus and are associated with circadian rhythm⁸. In addition, in ConvR-mice, nociceptin, galanin, catestatin, and truncated MCH were decreased by 46.4%, 51.4%, 40.4%, and 40.9%, respectively.

GF-mice gain less weight than ConvR-mice even though their food intake and total fecal calories are similar. GF-mice accumulate less fat compared to colonized counterparts when fed a cholesterol-rich, lard-based, high-fat diet or a high-fat, sugar-rich diet⁴⁶. The mechanism protecting GF-mice from diet-induced obesity might involve the influence of microbiota on both sides of the energy homeostasis⁴⁷, energy harvest from the diet and energy storage in the host⁴. Regulation of hypothalamic mature active neuropeptides levels may represent a novel pathway by which gut microbes affect food intake and energy homeostasis.

GF-mice show increased spontaneous motor activity and decreased anxiety-like behavior compared to ConvR-mice²⁰. Based on the changing profiles of neuropeptides under different gut microbiota environments, it could be speculated that these changes in brain development and behavior-related neuropeptides might act as signal molecules or might be associated with second messenger pathways and synaptic LTP in brain regions, which would lead to behavioral changes. Furthermore, proteins, including receptors, kinases, and cytokines, are also important in response to gut microbiota environment change.

10-plex DiLeu isobaric tag quantitative proteomic analysis of GF- and ConvR-mice

An example tandem MS spectrum of a DiLeu-labeled peptide sequence is shown in Figure 4. Using the DiLeu isobaric tag proteomics approach, a total of 3,971 proteins were identified, and 3,277 proteins were quantified (Table S5). Volcano plotting (Figure S3) was used to compare the fold changes (log₂ ratio GF-/ConvR-) and the statistical significance (−log₁₀ of p-value) for these quantified proteins. A p-value of < 0.05 and a 1.2-fold change were set as thresholds for statistically significant identifications. The 282 up- or down-regulated proteins passing these thresholds are listed in Table S6. Biological process, cellular component, molecular function enrichment analysis, and COG/KOG protein category analysis are shown in Figure 5.

Figure 4.

An example of identification and quantification of a DiLeu-labeled peptide. The intensities of the ten DiLeu reporter ions reflect the relative abundances of peptide LTGFHETSNINDFSAGVANR in ten different hypothalamus samples.

Figure 5.

(A) Among these 282 significantly changed proteins, the functions and characteristics of all the detected and quantified proteins were further investigated by gene ontology (GO) analysis and classification of the proteins into three categories: biological process, cellular component, and molecular function. (B) Analysis of eukaryotic orthologous groups (KOG) and clusters of orthologous groups (COG) of proteins was performed. Among the significantly changed proteins, 66 proteins were involved in signal transduction mechanisms, 30 proteins were related to general function prediction only, 18 proteins were involved in intracellular trafficking, secretion, and vesicular transport, and 16 proteins were related to inorganic ion transport and metabolism.

In the present study, proteomics and neuropeptidomics analysis are implemented simultaneously from the same sample. Due to different sample preparation strategies, five neuropeptide precursors identified in peptidomics analysis can also be identified in DiLeu proteomics analysis, including Pro-opiomelanocortin, Cerebellin-1, Galanin, Neuroendocrine convertase 2, and Chromogranin-A. LFQ results showed that neuropeptides derived from these 5 precursors were all up-regulated in GF-mice compared with those in ConvR-mice. In DiLeu quantitative analysis, only Cerebellin-1 and Pro-opiomelanocortin proteins were up-regulated by 1.12- and 1.43-fold in GF-mice compared with those in ConvR-mice, respectively. Chromogranin-A, Neuroendocrine convertase 2, and Galanin did not show significant up-regulation in DiLeu quantitative analysis, but neuropeptides from these precursors in LFQ method showed fold changes >1.5 (Table S4); Cerebellin-1 and Pro-opiomelanocortin showed 1.12- and 1.43- fold changes in DiLeu analysis, and neuropeptides from Cerebellin-1 and Pro-opiomelanocortin showed greater fold changes from 1.94 to 7.21 using LFQ approach. In general, LFQ and DiLeu quantitative analysis showed consistent changing trends but different fold changes in protein and neuropeptide quantification. Interestingly, for some mature neuropeptides, such as Joining peptide, γ-Lipotropin, CLIP, α-MSH, β-endorphin from Pro-opiomelanocortin were significantly up-regulated in GF-mice using LFQ approach, which was consistent with Pro-opiomelanocortin quantitative results using DiLeu labeling approach.

Interestingly, biological process enrichment results showed that the significantly changed proteins were related to nervous system development and signal transduction, including CNS development, intracellular signal transduction, trans-synaptic signaling, G-protein coupled receptor signaling pathways, and modulation of synaptic transmission; some may be related to learning and memory (Figure 6A). The cell periphery, cell projection, neuron projection, and synapses were all involved in cellular component enrichment (Figure 6B). In molecular function enrichment analysis (Figure 6C), proteins were enriched in various functions, including transporter activity, cation or metal ion trans-membrane transporter activity, and ligand-gated ion channel activity.

Figure 6.

Enrichment results of biological process (A), cellular component (B), and molecular function (C).

KEGG pathway enrichment analysis suggested that these proteins were involved in different pathways. The main pathways that attracted attention were those for signaling (cGMP-PKG, oxytocin, cAMP, GnRH) and synapses (glutamatergic and dopaminergic synapses). In addition, LTP, circadian entrainment pathways were shown to be involved. This analysis showed that proteins were involved in these important pathways. Proteins in the brain, such as receptors and kinases, are important and influence CNS development and behaviors. Compared to normal bacterial colonized mice, Neufeld et al. reported that GF-mice showed a decrease in the N-methyl-D-aspartate receptor subunit NR2B and in serotonin receptor 1A, as well as an increase in a brain-derived neurotrophic factor⁴⁸. It has been well established that the gut microbiota can alter the expression of cAMP-responding element-binding protein (CREB) and protein kinase C (PKC), suggesting that it might facilitate neurodevelopment via the PKC-CREB signaling pathway⁴⁹. Considering the circadian rhythm pathway, for instance, as shown in Figure 7A, proteins including glutamate receptor (AMPAR), glutamate receptor ionotropic, NMDA (NMDAR), inositol 1,4,5-trisphosphate receptor (IR₃P), ryanodine receptor (RyR), guanine nucleotide-binding proteins, adenylate cyclase (AC), 1-phosphatidylinositol 4,5-bisphosphate phosphodiesterase (PLC), protein kinase C (PKC), and calcium/calmodulin-dependent protein kinase (CaMKII) were involved. Interestingly, in the circadian rhythm pathway, PACAP was also identified in the neuropeptide investigation. PACAP and glutamate play a co-regulatory role in modulating the circadian clock⁵⁰. In addition, neuropeptides have been reported to change significantly during the natural circadian rhythm. For example, the PEN-LEN region derived from PCSK1N was significantly more abundant at nighttime, whereas the truncated GAV peptide was significantly more abundant in the daytime. When the neuropeptides and proteins detected in the present study are combined, it becomes evident that gut microbiota can affect proteins and neuropeptides in the brain, which may be associated with circadian rhythm-related physiological changes. In another KEGG pathway instance, LTP and long-term depression (LTD) was noted. As shown in Figure 7B, significantly changed proteins in the hypothalamus were also involved in the LTP and LTD pathways. LTP is the process of strengthening the connections between two neurons, which is associated with learning and memory. LTD has long been considered to be an important contributor to motor learning and memory.

Figure 7.

Neuropeptides and proteins involved in pathways: (A) circadian rhythm pathway, guanine nucleotide-binding protein (Gs), adenylate cyclase (AC), inositol 1,4,5-trisphosphate receptor (IP₃R), ryanodine receptor (RyR), glutamate receptor ionotropic (NMDAR), glutamate receptor (AMPAR), calcium/calmodulin-dependent protein kinase (CaMK II), PACAP were identified; (B) glutamatergic synapse, LTP, and LTD pathways, phosphatidylinositol phospholipase C (PLC), classical protein kinase C (PKC), metabotropic glutamate receptor (mGluR), serine/threonine-protein phosphatase 2B catalytic subunit (CaN), protein phosphatase 1 regulatory subunit 1A (PPP1R1A), serine/threonine-protein phosphatase PP1 catalytic subunit (PPP1C), Ras-related protein (Rap1), glial high affinity glutamate transporter (EAAT) were identified. Red and blue arrows represent up- and down-regulated proteins in GF-mice compared with ConvR-mice, respectively.

In summary, although it has been reported that gut microbiota can affect food intake, anxiety-like behavior, memory, and learning in mice, the molecular basis that leads to these physiological changes remains unclear. Gut microbiota are associated with impaired learning and memory. GF-mice display normal anxiety levels but exhibit an absence of non-spatial and working memory, proving that gut microbiota are important for memory development⁵¹. GF-mice showed more motor activity than ConvR-mice²⁰. GF-mice had lower body and adipose tissue weights compared with age-matched specific pathogen-free (SPF) mice; GF-mice exhibited decreased food intake levels at 4 weeks old but exhibited increased food intake levels compared with SPF-mice at the age of 8 months⁵². The present study enables profiling of changes in hypothalamic proteins and neuropeptides in GF-mice and may provide potential molecular clues to help elucidate possible mechanism by which these phenotypes are affected.

Conclusions

In this work, LFQ-based neuropeptidomics analysis and multiplexed DiLeu isobaric tag-based proteomics analysis were performed to compare the relative abundance changes in neuropeptides and proteins in the hypothalamus region of GF- and ConvR-mice. Interestingly, neuropeptides and proteins exhibiting significant changes in relative abundances were identified, and the functions of these neuropeptides and the pathways of the proteins involved were related to regulation of food intake, circadian rhythm, learning, and memory. The integrated label-free and 10-plex DiLeu isobaric tag quantitative strategy, combined with high-resolution MS, enabled simultaneous identification and quantification of a surprisingly rich repertoire of neuropeptides and proteins. The current study confirmed the importance of the gut microbiota in regulating brain component changes. Using the present method, signaling molecules including neuropeptides and proteins in other brain regions can be identified and quantified, and the relationship of important signaling molecules among these different regions can be elucidated as well. Future studies should focus on targeted validation of these neuropeptide and protein changes and further examination of these components for changes in other brain regions to obtain a more comprehensive understanding about the interplay of these diverse signaling pathways in different brain regions influenced by gut microbiome.