Graphical abstract

   graphic file with name fx1.jpg
   [47]Open in a new tab

Highlights

     * •
       L. reuteri lysate suppresses adipocyte differentiation during early
       adipogenesis
     * •
       L. reuteri lysate inhibits KLF5 and reduces PPARγ expression and
       BCAA catabolism
     * •
       L. reuteri lysate induces metabolic changes via HIF1α and regulates
       lipin-1 isoform
     * •
       KLF5 and HIF1α mediate L. reuteri lysate’s inhibitory effects on
       adipogenesis
     __________________________________________________________________

   Obesity medicine; Microbiology; Microbiome; Omics

Introduction

   Global obesity rates have almost doubled since 1980, with over
   one-third of the global population now overweight or obese, rendering
   obesity a major public health concern.[48]^1 Obesity is defined as an
   abnormal or excessive accumulation of fat to the point of impairing
   health.[49]^2 Excess fat stored in the body increases the risk of
   metabolic disorders, such as type 2 diabetes and cardiovascular
   diseases.[50]^3 Lifestyle and environmental factors play a pivotal role
   in the current rapid increase in obesity prevalence despite the
   significant influence of genetic factors on obesity
   development.[51]^4^,[52]^5 The changes in the lifestyle of modern
   people, including diet and physical activity, ultimately lead to an
   imbalance in energy intake and expenditure, resulting in obesity.[53]^6
   Efforts to prevent obesity are being conducted in many countries, but
   for patients with obesity above a certain level, additional treatment,
   including drugs or surgery, may be required.[54]^7 However, surgical
   intervention (e.g., bariatric surgery) alone is insufficient to meet
   the medical needs of the obese population. Thus, there is increasing
   interest in the development of anti-obesity medications (AOMs), which
   are expected to fulfill the unmet medical demands for obesity
   treatment. Despite many efforts to develop effective AOMs, safety
   concerns due to various side effects have prevented many treatments
   from being approved for use.[55]^8 Therefore, appropriate drug
   candidates that are safe and effective for treating obesity are crucial
   for developing AOMs. In this perspective, probiotics, known to exert
   various positive effects on the host’s energy balance and metabolism,
   are emerging as novel candidates for AOMs.

   Probiotics are defined as live microorganisms that provide health
   benefits to the host when consumed in adequate amounts.[56]^9 Recently,
   the concept has been expanded to include postbiotics, which have
   similar benefits to probiotics.[57]^10 Probiotics generally affect host
   health by improving gut dysbiosis and modulating
   metabolism.[58]^9^,[59]^11 Gut dysbiosis can contribute to obesity
   through alterations in energy and lipid metabolisms.[60]^12 Interest in
   developing probiotics as therapeutic adjuvants is increasing, offering
   promising pharmacotherapy to improve this condition.[61]^13^,[62]^14
   Lactobacillus spp. and Bifidobacterium spp. are representative
   probiotics known for their safety profile and effectiveness on various
   diseases, including obesity, metabolic disorders, and immune
   disorders.[63]^15 Several probiotics, including Lactobacillus spp. and
   Bifidobacterium spp., have been reported to exhibit anti-obesity
   effects in vivo or in vitro, and their potential as novel medications
   for obesity is attracting attention.[64]^16^,[65]^17 Studies on the
   mechanism of action (MOA) of new disease treatments are important for
   validating and assessing their efficacy and predicting possible side
   effects.[66]^18 Therefore, target identification and MOA studies at the
   molecular level are crucial for therapeutic probiotic research.[67]^19
   Nevertheless, identifying the exact mechanism is challenging due to the
   complexity of probiotics, which is more intricate than common drugs,
   and the complex regulation of adipocyte
   differentiation.[68]^19^,[69]^20 Despite the increasing number of
   reports on the anti-obesity effect of probiotics, research on the MOA
   for individual probiotics remains unclear.[70]^21 To date, many studies
   have evaluated the anti-obesity effect of probiotics by analyzing
   single genes and proteins related to lipogenesis in adipocytes and
   adipose tissues. However, to fully recognize the anti-obesity
   characteristics of probiotics, it is essential to identify specific
   biomarkers and develop a method to confirm changes in metabolic
   pathways.[71]^19^,[72]^20^,[73]^21 In this context, there is an
   increasing emphasis on using “multiomics” approaches to improve our
   molecular understanding of the complex MOA of probiotics by providing
   robust data.[74]^21^,[75]^22

   Despite extensive research regarding the multistep transcriptional
   complexity of adipocytes, attempts to analyze lipid suppression
   mechanisms focusing specifically on the timing of adipocyte
   differentiation have not always been successful.[76]^23 Herein, we
   established an advanced multiomics workflow to identify the mechanisms
   inhibiting adipogenesis in adipocytes to assess the anti-obesity effect
   of a complex treatment including probiotics ([77]Figure 1). First,
   adipocyte differentiation was categorized into early (days 0–2),
   intermediate (days 2–4), and late (days 4–8) stages based on changes in
   major transcription factors to determine the inhibitory effect on each
   stage. Subsequently, proteomics was performed on adipocytes at the
   stage where the inhibitory effect was observed, as well as on mature
   adipocytes to compare the changes in metabolic pathways. Based on the
   proteomics results, additional metabolomic and lipidomic analyses were
   performed to confirm the metabolic changes in adipocytes and predict
   the MOA, which was finally validated via mRNA expression analysis. As a
   model study, 15 species of probiotics were randomly prescreened from
   Lactobacillus spp. and Bifidobacterium spp., which are generally
   recognized as safe and previously reported for their antiadipogenic
   potential.[78]^15 Among them, we selected L. reuteri, which most
   effectively inhibits adipocyte lipid accumulation, and identified its
   inhibition mechanism. Following the workflow presented in this study,
   L. reuteri was found to have a significant effect, particularly at the
   early stage of adipocyte differentiation. L. reuteri reduced lipid
   accumulation in adipocytes by inhibiting branched-chain amino acid
   (BCAA) catabolism, specifically targeting early differentiation.
   L. reuteri inhibited the expression of KLF5 during early adipocyte
   differentiation, resulting in the downregulation of PPARγ expression
   and reduction in BCAA catabolism. Simultaneously, L. reuteri inhibited
   adipogenesis by promoting the expression of HIF-1α and downregulating
   the expression of lipin-1 in early adipocyte differentiation. This
   discovery shows that our multiomics strategy can identify new
   mechanisms and target molecules for inhibiting adipogenesis through
   probiotics.

Figure 1.

   [79]Figure 1
   [80]Open in a new tab

   Schematic overview of deciphering the anti-obesity mechanisms of
   probiotics

   An effective probiotic lysate was prepared from the candidate
   probiotics and applied to adipocytes. The adipocytes were categorized
   into early, intermediate, and last stages of adipogenesis to identify
   the stages at which a reduction in lipid accumulation was observed upon
   treatment. Proteomics analysis was conducted focusing on the identified
   stages. Analyses of metabolites, lipids, mRNA were performed to
   validate the proteomic results.

Results and discussion

Screening of bacterial strains and investigating the stages of adipocyte
differentiation

   To be recognized as bacterial strains with antiadipogenic potential,
   bacteria should be able to interfere with lipid accumulation in
   differentiated adipocytes without causing significant changes in the
   cell viability of adipocytes.[81]^24 In our results, probiotic lysates
   except L. johnsonii inhibited lipid accumulation in 3T3-L1 adipocytes
   ([82]Figure 2A), while B. animalis subsp. animalis, B. longum,
   B. animalis subsp. Lactis, and B. bifidum substantially damaged cells
   in the treatment process ([83]Figure 2B). Among other strains,
   L. fermentum (48.56% in 5 × 10^7 colony-forming unit [CFU]/mL) and
   L. reuteri (54.11% in 5 × 10^7 CFU/mL) caused superior lipid
   accumulation inhibition at the same CFU concentration. Therefore, we
   considered L. reuteri (5 × 10^7 CFU/mL), which reduced lipid
   accumulation to the maximum extent compared to the control, as a model
   candidate for studying probiotics to validate the objectives of this
   study.

Figure 2.

   [84]Figure 2
   [85]Open in a new tab

   Comparison of the inhibitory effects of probiotic candidates on
   adipogenesis in 3T3-L1 cells

   (A) Comparison of lipid accumulation in 3T3-L1 cells. Lipid
   accumulation in 3T3-L1 cells treated with probiotic lysates was
   measured using Oil red O staining (n = 3).

   (B) Cytotoxicity of probiotic lysates on 3T3-L1 cells. Cell viability
   was measured by MTT assay (n = 3).

   Data are represented as means ± SD. Statistical significance was
   analyzed using one-way ANOVA with Tukey’s HSD test and indicated as ∗
   for p < 0.05 respectively compared with controls.

   Some strains of L. reuteri have been consistently demonstrated in
   previous studies to exert anti-obesity effects by suppressing lipid
   accumulation in the liver or adipose tissue.[86]^25^,[87]^26^,[88]^27
   The anti-obesity effect of L. reuteri was observed in vivo or in vitro
   not only in live cells[89]^25^,[90]^26 but also in dead cells[91]^26 or
   lysate derived from them.[92]^27 Specifically, the reported strain in
   L. reuteri was identical to the L. reuteri strain tested in this
   study.[93]^27 However, despite numerous reports on the anti-obesity
   properties of L. reuteri strains,[94]^25^,[95]^26^,[96]^27 most studies
   have been based on phenotypic characteristics and have not provided a
   clear mechanism of how L. reuteri strains inhibit adipogenesis through
   metabolic processes. A study was conducted using a multiomics approach
   to evaluate the anti-obesity properties of L. reuteri, investigate its
   mechanism and molecular target in adipocytes, and address uncertainties
   found in previous studies.

   Adipogenesis converts preadipocytes with fibroblast-like phenotype into
   adipocytes capable of accumulating lipids. This adipocyte
   differentiation process is highly regulated involving an intricate
   interaction of various metabolic pathways, and different stages of
   adipocyte differentiation result in different phenotypes.[97]^28
   Therefore, we divided adipocyte differentiation into three stages
   (early, intermediate, and late) and treated the aforementioned lysate
   of L. reuteri at each stage to identify the key steps involved in cell
   differentiation ([98]Figure 3A). First, the treatment conditions of
   days 0–2, days 0–4, and days 0–8 showed a significant difference from
   the control and a reduction in lipid accumulation of about 40%,
   suggesting that L. reuteri showed antiadipogenic potential when treated
   at the early differentiation stage of days 0–2 ([99]Figure 3B). In
   addition, when L. reuteri was treated only from days 0–2 by
   concentration, it showed a reduction in lipid accumulation of 54.4%
   (5 × 10^7 CFU/mL), which was similar to the results in [100]Figure 2A,
   where treatment was administered from days 0–8 ([101]Figure 3C). As
   adipogenesis progressed, adipocytes and lipid droplet size were
   significantly reduced compared to the control ([102]Figure S1).
   Furthermore, image analysis using ImageJ confirmed a reduction in the
   lipid accumulation area of adipocytes by L. reuteri treatment, like
   previous results ([103]Figure S2).

Figure 3.

   [104]Figure 3
   [105]Open in a new tab

   Effect of L. reuteri on adipogenesis inhibition

   (A) Diagram outlining the stages of the adipogenesis process.
   L. reuteri was treated during different time periods of adipogenesis.

   (B) Lipid accumulation was quantified by staining adipocytes with Oil
   red O staining on day 8 (n = 6).

   Data are represented as means ± SD. Statistical significance was
   analyzed using one-way ANOVA with Tukey’s HSD test and indicated as
   ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns: no significant.

   (C) Lipid accumulation in adipocytes treated with L. reuteri under the
   D0/2 condition (n = 3).

   Data are represented as means ± SD. Statistical significance was
   analyzed using one-way ANOVA with Tukey’s HSD test and indicated as
   ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

   (D) Relative mRNA levels of PPARγ and C/EBPα measured in early
   differentiating (day 2) and mature (day 8) adipocytes, respectively
   (n = 4).

   Data are represented as means ± SD. Statistical significance was
   analyzed using unpaired two-tailed Student’s t test and indicated as
   ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

   To investigate the impact of L. reuteri on major transcription factors
   involved in the regulation of adipogenesis in adipocytes, we analyzed
   changes in gene expression of the major factors: peroxisome
   proliferator-activated receptor γ (PPARγ) and CCAAT/enhancer-binding
   protein (C/EBPα).[106]^29 At each time point of days 2 and 8 of
   adipogenesis, mRNA expression of both PPARγ and C/EBPα was decreased
   compared to the control group ([107]Figure 3D). This result suggested
   that treatment with L. reuteri inhibited the expression of PPARγ and
   C/EBPα from the early differentiation stage, thus preventing adipocyte
   differentiation and reducing lipid accumulation. Taken together, our
   findings revealed the crucial stage at which L. reuteri exhibited
   antiadipogenic potential. This was further confirmed by suppressing
   PPARγ and C/EBPα expression during early differentiation.

Proteomic changes in adipocytes in early differentiation and maturation
stages

   To investigate the intracellular proteome changes in adipocytes during
   treatment with L. reuteri lysate from days 0–8, cells were harvested
   and analyzed at two timepoints—days 2 and 8—after induction of
   preadipocyte differentiation. Principal component analysis showed a
   substantial difference in separation between different experimental
   groups and clustering between biological replicates on both days 2 and
   8 ([108]Figure 4A). Additionally, hierarchical clustering analysis of
   the total proteins detected was used to visualize differences between
   groups ([109]Figure S3). A total of 3,535 and 2,641 proteins were
   identified on days 2 and 8, respectively, with 2,340 proteins
   identified in common between the two datasets ([110]Figure 4B). To
   analyze differences in protein expression in adipocytes induced by
   L. reuteri, proteins exhibiting a 1.5-fold change and a p value of less
   than 0.05 were classified as differentially expressed proteins (DEPs).
   On day 2, 128 proteins were classified as DEPs, with 62 upregulated and
   66 downregulated proteins. On day 8, 606 proteins were classified as
   DEPs, with 316 upregulated and 290 downregulated proteins
   ([111]Figure 4E; [112]Table S1). In addition, 40 proteins were
   identified as overlapping DEPs at both time points ([113]Figure 4C;
   [114]Table S2).

Figure 4.

   [115]Figure 4
   [116]Open in a new tab

   Analysis of intracellular proteome changes in early differentiating
   (day 2) and mature (day 8) adipocytes in response to L. reuteri

   (A) Principal component analysis (PCA) from the intracellular proteome
   of adipocytes treated with L. reuteri.

   (B and C) Venn diagrams illustrating all identified proteins at each
   time point and their overlap among differentially expressed proteins.

   (D) Heatmap of protein changes associated with adipocyte
   differentiation, based on a fold change cut-off of >1.5 and p < 0.05.
   The color of heatmap represents the log2 ratio of each protein in each
   replicate.

   (E) Volcano plots showing proteins differentially regulated at each
   point by L. reuteri treatment. Proteins with increased levels compared
   to the control are shown in red, while decreased proteins are shown in
   blue. The number of proteins is indicated at each edge of the plot,
   based on fold change cut-off of >1.5 and p < 0.05.

   (F) Enrichment analysis performed using ShinyGO for proteins with
   significant differences, showing the top 10 KEGG pathways based on fold
   enrichment with an FDR cut-off of <0.05. No significant enrichment was
   found in proteins that decreased on day 2.

   Subsequently, to identify the molecular pathways regulated by
   L. reuteri during adipogenesis, Gene Ontology (GO) term and Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis
   was performed using ShinyGO 0.80 on the four DEP groups categorized by
   protein harvest time and differential expression.[117]^30 KEGG pathway
   enrichment analysis results were shown only for the top 10 with
   FDR < 0.05 based on fold enrichment ([118]Figure 4F). The downregulated
   DEPs on day 2 were excluded as no significant enrichment was found.
   Through GO term and KEGG pathway enrichment analysis, upregulated DEPs
   in adipocytes treated with L. reuteri were mainly involved in iron and
   glucose metabolism, including the HIF-1 and glucagon signaling
   pathways, as well as with the composition of cell and organelle
   membranes. The downregulated DEPs were mainly involved in energy and
   lipid metabolism, including the tricarboxylic acid (TCA) cycle, BCAA
   degradation, and fatty acid metabolism, as well as in mitochondrial
   composition ([119]Figures 4F and [120]S4).

   To further focus on the changes in adipogenesis-associated proteins,
   they were visualized using heatmaps ([121]Figure 4D). The downregulated
   proteins were mainly involved in the PPAR signaling pathway, fatty acid
   metabolism, and TCA cycle, while the upregulated proteins mainly
   included proteins involved in the tumor necrosis factor signaling
   pathway. The changes in most adipogenesis-associated proteins
   (including SerpinA3g and Fasn) in [122]Figure 4D show a similar pattern
   to previously reported protein changes in adipocytes in response to
   other probiotic lysates.[123]^31 This suggested that the MOA of
   antiadipogenic probiotics on adipocytes may be partially
   shared.[124]^32

Inhibition of lipid accumulation through treatment with L. reuteri via
modulation of BCAA catabolism

   To characterize the metabolic changes induced by L. reuteri, we
   investigated enzymatic changes related to energy and lipid metabolisms,
   which are major metabolic pathways in adipocytes and are shown in
   [125]Figure 5A. Proteomic analysis was conducted at two time
   points—early differentiating adipocytes (day 2) and mature adipocytes
   (day 8)—to investigate changes in energy and lipid metabolisms. Changes
   in lipid metabolism were confirmed only in mature adipocytes
   ([126]Figure 5B). Notably, alterations in lipid metabolism were
   exclusively observed in mature adipocytes. Major energy metabolism
   processes, including the TCA cycle and oxidative phosphorylation
   (OXPHOS), occurring in mitochondria exhibited an overall decrease in
   adipocytes (days 2 and 8) treated with L. reuteri, suggesting that it
   may inhibit mitochondrial function ([127]Figure 5B). The decrease in
   mitochondrial function was further confirmed by Search Tool for the
   Retrieval of Interacting Genes/Proteins (STRING)[128]^33 network
   analysis of downregulated proteins in adipocytes (day 8)
   ([129]Figure S5). In mature adipocytes, the major source for lipid
   synthesis is glucose, but BCAAs (isoleucine, leucine, and valine) can
   also contribute to lipid accumulation.[130]^34 BCAAs participate in the
   TCA cycle by catabolizing themselves to acetyl-CoA or succinyl-CoA,
   which is then converted to palmitate (C16:0) and subsequently
   synthesized into triglycerides through de novo lipogenesis to form
   lipid droplets.[131]^34^,[132]^35 Our proteomics data showed an overall
   decrease in enzymes involved in BCAA catabolism ([133]Figure 5B), which
   could result in their reduced degradation and intracellular
   accumulation. Generally, BCAA consumption increases gradually during
   adipocyte differentiation.[134]^35 Because the increase in BCAA
   catabolism in early adipocyte differentiation promotes adipogenesis
   prior to PPARγ activation, the level of intracellular BCAAs in
   adipocytes serves as a crucial indicator for evaluating the inhibition
   of adipogenesis.[135]^36 Metabolomic analysis was performed in
   adipocytes to investigate the alteration of BCAA levels according to
   L. reuteri treatment conditions. Interestingly, a significant increase
   in BCAA levels was observed only in groups (D0/2 and D0/8) treated
   during early adipocyte differentiation ([136]Figure 5D). In addition,
   the inhibition of BCAA catabolism in early adipocyte differentiation
   consequently decreased the production of acetyl-CoA,[137]^34^,[138]^35
   a major precursor in de novo lipogenesis, which was observed in
   adipocytes (D0/2 and D0/8) along with a reduction in malonyl-CoA and
   palmitoyl-CoA ([139]Figure 5D). At the transcriptional and protein
   levels, the expression of most genes involved in adipogenesis,
   including those responsible for transferring acetyl-CoA to
   palmitoyl-CoA (Fasn, Acaca), fatty acid transporter (Fabp4), and
   adipocyte marker (Adipoq) expression, was confirmed to be significantly
   downregulated ([140]Figure 5E). These results confirmed that L. reuteri
   selectively inhibits early adipocyte differentiation, reduces BCAA
   catabolism, and inhibits adipogenesis before PPARγ activation.

Figure 5.

   [141]Figure 5
   [142]Open in a new tab

   Reducing lipid accumulation by modulating BCAA catabolism with
   L. reutri

   (A) Visualization of the major metabolic changes in the adipocytes,
   including energy metabolism and lipid metabolism.

   (B and C) Heatmap of protein changes associated with energy metabolism
   and lipid metabolism in adipocytes based on fold change cut-off of >1.5
   and p < 0.05. The color of heatmap represents the log2 ratio of each
   protein in each replicate (n = 4).

   (D) Relative levels of lipogenesis-related metabolites and BCAAs during
   the L. reuteri treatment period (n = 3).

   Data are represented as means ± SD. Statistical significance was
   analyzed using one-way ANOVA with Tukey’s HSD test and indicated as ∗
   for p < 0.05 respectively compared with controls.

   (E) Relative levels of mRNA and protein in mature adipocytes for major
   factors involved in adipogenesis (n = 4).

   Data are represented as means ± SD. Statistical significance was
   analyzed using unpaired two-tailed Student’s t test and indicated as
   ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

   (F) GC-MS chromatogram of the fatty acid methyl esters (FAMEs) in
   mature adipocytes using heneicosanoic acid (C21:0) as an internal
   standard.

   (G) Quantification of fatty acids in mature adipocytes with and without
   L. reuteri treatment. The identified fatty acids were quantified using
   peak area quantification, and the results were subsequently normalized
   to the internal standard and lipid weight (n = 3).

   Data are represented as means ± SD. Statistical significance was
   analyzed using unpaired two-tailed Student’s t test and indicated as
   ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

   Lipid metabolism in adipocytes is primarily categorized into energy
   storage, including lipid and fatty acid synthesis, and energy
   expenditure, including lipolysis and fatty acid oxidation, and is
   regulated by PPARγ.[143]^37 Inhibition of BCAA catabolism through
   L. reuteri can interfere with adipogenesis, which was confirmed by a
   decrease in enzymes involved in lipid metabolism ([144]Figure 5C). The
   decrease in new lipid droplets indicated a decrease in lipid
   accumulation, which was confirmed by the reduction in Cav1, Plin1, and
   Cavin1 levels, which are membrane proteins of lipid droplets
   ([145]Figure 5C). As a component of triglycerides, the end product of
   lipid metabolism, fatty acids contribute to ATP production or promote
   adipogenesis as signaling molecules.[146]^38^,[147]^39 Total fatty acid
   contents were quantified in mature adipocytes by lipidomic analysis to
   identify changes in fatty acid profiles within adipocytes
   ([148]Figure 5F). Identified fatty acids all decreased except for
   stearic acid (C18:0), with fatty acids such as C16:0, C16:1n7 and
   C18:1n9 collectively constituting approximately 80% of the total fatty
   acids analyzed in each group ([149]Table S3). Because most fatty acids
   that constitute lipid droplets are synthesized from acetyl-CoA,[150]^34
   its decrease caused by L. reuteri can decrease major lipids in
   adipocytes. Furthermore, odd-chain fatty acids (OCFA; C15:0, and C17:0)
   are not synthesized from acetyl-CoA but can be synthesized via BCAA
   catabolism.[151]^40^,[152]^41 The observed reduction in OCFA is
   probably due to inhibition of BCAA catabolism.[153]^35^,[154]^41 Thus,
   our comprehensive multiomics results, including proteomics, lipidomics,
   and metabolomics, provide evidence that L. reuteri inhibits
   adipogenesis.

Metabolic alteration induced by L. reuteri in adipocytes

   Considering that enzymes located in the mitochondria, including most
   enzymes involved in BCAA catabolism, have decreased ([155]Figure S5),
   L. reuteri showed the potential to decrease adipogenesis through
   inhibition of mitochondrial function and subsequently reduce BCAA
   catabolism. In previous studies, upregulation of mitochondrial
   metabolism in adipocytes has been reported to promote adipogenesis by
   producing metabolic intermediates important for essential metabolic
   processes such as de novo lipogenesis rather than simply being a
   consequence of adipogenesis.[156]^34^,[157]^42 Furthermore, the
   maintenance of these metabolic processes depends on a continuous ATP
   supply.[158]^43 Therefore, inhibition of mitochondrial function by
   L. reuteri may decrease adipogenesis by disrupting the supply of
   metabolic intermediates and ATP. Interestingly, although mitochondrial
   function was reduced in adipocytes treated with L. reuteri,
   intracellular ATP levels did not show any significant changes
   ([159]Figure 6A). Based on this, it was hypothesized that ATP levels
   were maintained by an alternative pathway that operated in the cytosol
   rather than the mitochondria. Under normal circumstances, adipocytes
   commonly generate ATP through OXPHOS in the mitochondria, but under
   conditions of inhibited mitochondrial function, glycolysis operates as
   an alternative pathway for ATP production.[160]^44 In our analysis of
   protein profiles, we observed an increase in glycolytic enzymes
   following treatment with L. reuteri. Based on these findings, we
   consider glycolysis an alternative pathway for ATP production
   ([161]Figure 5B). Pyruvate, derived from glucose, can undergo
   mitochondrial metabolism or be converted to lactate, the end
   product.[162]^45 Accumulation of lactate, observed in conditions of
   reduced mitochondrial function, indicates activation of
   glycolysis.[163]^45^,[164]^46 Lactate accumulation is regulated by two
   isoforms of lactate dehydrogenase (LDH). Notably, in our proteomic
   analysis results, LDHA, which converts pyruvate to lactate, increased.
   Conversely, LDHB, which converts lactate to pyruvate, decreased
   ([165]Figure 5B). This modulation of glycolytic activity may result in
   decreased glucose levels and increased lactate levels, consistent with
   changes observed in metabolomic analysis performed in early
   differentiating adipocytes ([166]Figure 6A). As a result, L. reuteri
   inhibited mitochondrial function and induced glycolysis, activating an
   alternative pathway for ATP production. The increase in glycolysis
   resulting from reduced mitochondrial function, induced by L. reuteri,
   indicated the inhibition of adipogenesis through reduced BCAA
   catabolism.[167]^47

Figure 6.

   [168]Figure 6
   [169]Open in a new tab

   Effect of L. reuteri on mitochondrial metabolism in adipocytes

   (A) Relative levels of intracellular or extracellular metabolites
   involved in energy metabolism in early differentiating adipocytes.

   Data are represented as means ± SD. Statistical significance was
   analyzed using unpaired two-tailed Student’s t test and indicated as
   ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

   (B) Schematic for the processing of L. reuteri treatment during
   adipogenesis. Previous results on the differentiation of adipocytes,
   depending on the timing of L. reuteri treatment, were also shown.

   (C) Illustration of metabolic pathways activated during exposure to
   L. reuteri, accompanied by a heatmap of the involved metabolites.

   (D–G) Changes in lactate (D), citrulline (E), arginine (F), and DHAP
   (G) between early differentiating (day 2) and mature (day 8) adipocytes
   in response to the time-point analysis of L. reuteri treatment.

   Our multiomics results confirmed that L. reuteri effectively reduced
   mitochondrial function, attenuating adipogenesis. However, this raised
   the question of whether these actions induced short-term modifications
   or irreversible impairment of mitochondrial metabolism in early
   differentiating adipocytes. Therefore, it was necessary to investigate
   the duration and reversibility of changes induced by L. reuteri. To
   determine whether glycolysis was altered in mature adipocytes depending
   on L. reuteri treatment, additional metabolic analyses were conducted,
   considering whether the lysate was administered on day 2 before or
   after differentiation induction ([170]Figures 6B and [171]S6).
   Interestingly, contradictory results were observed in the D0/2 and D0/8
   groups in terms of changes in intracellular lactate levels at day 8,
   despite both groups experiencing adipogenesis inhibition due to
   L. reuteri treatment during the early differentiation period
   ([172]Figure 6D). First, D0/8 exhibited an increasing trend in lactate
   levels that was sustained by L. reuteri treatment until day 8, even
   after initial differentiation. By contrast, after day 2, the lactate
   levels decreased significantly in D0/2, where the L. reuteri treatment
   was not performed. Additionally, lactate levels increased in D2/8
   treated with L. reuteri from day 2 onward, indicating that lactate
   production was induced by L. reuteri treatment in the D0/8 and D2/8
   groups. Given the earlier observation of increased lactate levels in
   adipocytes undergoing early differentiation following treatment with
   L. reuteri, it seems probable that exposure to the lysate at the time
   of cell acquisition is the determining factor of this alteration,
   suggesting that the observed metabolic changes are a transient effect
   that occurs specifically during exposure to L. reuteri. To provide
   further evidence for this transient metabolic alteration, we performed
   an additional search for metabolites associated with glycolysis that
   showed similar changes in D0/8 and D2/8 ([173]Figure 6C).

   It is possible that the metabolic alteration caused by L. reuteri was
   due to decreased mitochondrial function, similar to the observation on
   day 2. When mitochondrial function is inhibited, nitric oxide (NO)
   production can be stimulated to restore it in cells with activated
   glycolysis.[174]^48 NO is generated from arginine through a process
   that involves its conversion to citrulline.[175]^48 To test our
   hypothesis that L. reuteri inhibits mitochondrial function only during
   transient exposure, we assessed changes in arginine and its precursor
   citrulline—essential components of NO production. We observed
   significant differences only in adipocytes at early differentiating and
   in groups D2/8 and D0/8, which were exposed to L. reuteri at the time
   of cell acquisition ([176]Figures 6E and 6F). The notable decrease in
   arginine and the corresponding increase in citrulline in the adipocyte
   groups exposed to L. reuteri indirectly prove that the mitochondrial
   function of adipocytes was transiently inhibited. In addition, the
   decreased production of dihydroxyacetone phosphate (DHAP), a precursor
   of lipid components produced from glucose, at D2/8 and D0/8, suggests
   the enhancement of the glycolytic pathway leading to lactate
   production, supporting the possibility of transient metabolic changes
   induced by L. reuteri ([177]Figure 6G). These metabolic changes induced
   by L. reuteri can be restored to their previous state by controlling
   the treatment process. This implies that L. reuteri did not induce
   mitochondrial damage and did not negatively affect cell survival. In
   contrast, the inhibition of adipogenesis by L. reuteri at the early
   differentiation stage of adipocytes was irreversible, even when the
   experimental conditions were controlled. These observations suggest
   that the effect of L. reuteri on early differentiating adipocytes is
   important in regulating overall adipocyte differentiation. In
   conclusion, our findings underscore the importance of the timing of
   early adipocyte differentiation for L. reuteri to exert specific and
   temporally controlled inhibition of adipogenesis.

Inhibition of adipogenesis by interference with KLF5 transcription during
early adipocyte differentiation

   Major transcription factors and enzymes were investigated to identify
   potential targets of L. reuteri, focusing on C/EBPβ and C/EBPδ, which
   are crucial regulators expressed during early adipocyte
   differentiation.[178]^49 In addition, enzymes involved in mitotic
   clonal expansion (MCE), a process that occurs during early adipocyte
   differentiation, were also considered potential targets.[179]^50 C/EBPβ
   and C/EBPδ are expressed early in adipogenesis initiated by adipogenic
   cocktails and directly induce expression of PPARγ and CEBPα.[180]^49 In
   our results, we identified decreased PPARγ and CEBPα expression in
   early differentiating adipocytes ([181]Figure 3D). This suggests that
   there is a possibility of inhibition at the pre-expression and
   postexpression stages of C/EBPβ and C/EBPδ before PPARγ and CEBPα
   expression. An adipogenic cocktail containing
   3-isobutyl-1-methylxanthine (IBMX), dexamethasone (DEX), and insulin
   induce the expression of C/EBPβ and C/EBPδ through the activation of
   cAMP response element-binding protein (CREB), glucocorticoid receptor,
   and sterol regulatory element-binding protein 1 (Srebp1),
   respectively.[182]^51 In order to specify the location of the target
   before and after C/EBPβ/δ expression, changes were confirmed by
   examining the mRNA and protein levels of C/EBPβ and C/EBPδ
   ([183]Figure 7B). However, the analysis found no significant changes in
   either transcription factor. Consequently, it can be concluded that
   potential targets of L. reuteri remained after C/EBPβ/δ expression. The
   absence of significant changes in the mRNA expression of Creb1 and
   Srebp1, transcription factors that regulate C/EBPβ and C/EBPδ
   expression, indirectly suggested that adipogenesis may be inhibited
   following C/EBPβ/δ expression ([184]Figure 7C). C/EBPβ is activated
   through a phosphorylation process to upregulate PPARγ to promote
   adipogenesis, and the cell cycle is partially involved in this process.
   Confluent preadipocytes re-enter the cell cycle and undergo MCE, which
   increases the activity of Cdk2. Activated Cdk2 participates in the
   phosphorylation of C/EBPβ, promoting its expression as
   PPARγ.[185]^50^,[186]^52 To evaluate the changes in C/EBPβ activity, we
   investigated the impact of L. reuteri on MCE by measuring cell counts
   during adipogenesis. Cell counts showed normal proliferation, with no
   significant difference observed compared to the control group, and
   proliferation was nearly complete by day 2 ([187]Figure 7D).
   Additionally, changes in cell cycle-related enzymes were not
   significant, suggesting that L. reuteri does not affect C/EBPβ activity
   via MCE ([188]Figure 7E). Therefore, the expression and activation of
   C/EBPβ remained largely unchanged, and it became apparent that the
   reduction in PPARγ expression occurred independently of C/EBPβ.
   Considering the preceding result, we hypothesized that L. reuteri might
   regulate PPARγ expression through modulation of transcription factors
   downstream of C/EBPβ, including Krüppel-like factor 5 (KLF5), which is
   considered a strong potential candidate. During early adipocyte
   differentiation, C/EBPβ binds to the promoter of KLF5 and induces its
   expression.[189]^53 Moreover, C/EBPβ cooperates with induced KLF5 to
   bind directly to the promoter of PPARγ, thereby promoting its
   expression.[190]^53 Indeed, mRNA expression analysis was conducted to
   confirm the hypothesis that KLF5 could be a target of L. reuteri. It
   was found that the expression of KLF5 decreased in early
   differentiating adipocytes ([191]Figure 7C). Consequently, L. reuteri
   reduced PPARγ expression by downregulating KLF5 transcription.
   Additionally, the reduction of KLF5 resulted in the downregulation of
   transcription factors induced by it. In adipocytes, KLF5 promotes BCAA
   catabolism by inducing the expression of Krüppel-like factor 15
   (KLF15).[192]^54^,[193]^55 This suggests that a reduction in KLF5 may
   attenuate BCAA catabolism, consistent with previously observed changes
   in BCAA metabolism ([194]Figure 5A). In conclusion, KLF5 is a crucial
   target of L. reuteri in reducing adipogenesis by interfering with early
   adipocyte differentiation, and the observed changes in adipocyte
   metabolism strongly support this conclusion.

Figure 7.

   [195]Figure 7
   [196]Open in a new tab

   Investigation of transcription factors that regulate early adipocyte
   differentiation and associated processes

   (A) Visualization of overall transcription factor expression and
   processes occurring during early adipocyte differentiation.

   (B) Relative mRNA expression and protein levels of C/EBPβ and C/EBPδ,
   key regulators in early adipocyte differentiation, in early
   differentiating adipocyte (n = 4).

   (C) Relative mRNA expression levels of major factors involved in PPARγ
   expression in early differentiating adipocytes (n = 4).

   (D) Effects of L. reuteri on the proliferation of adipocytes during
   early differentiation (n = 4). 3T3-L1 preadipocytes at day 0 served as
   a control.

   (E) The protein-protein interaction (PPI) network and a table of
   changes in proteins involved in the cell cycle are presented.

   Data are represented as means ± SD. Statistical significance was
   analyzed using unpaired two-tailed Student’s t test and indicated as
   ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns: no significant.

Investigating HIF-1α as a target for L. reuteri in inhibiting adipogenesis

   Based on previous research results, it was validated that L. reuteri
   exhibited specific and temporal effects in inhibiting adipogenesis
   ([197]Figure 6). Considering the metabolic changes induced by
   L. reuteri as transient responses, we hypothesized that the factors
   regulating these metabolic changes could be additional targets. To
   identify these additional targets, we explored the enrichment analysis
   results presented in [198]Figure 4F, focusing on pathways that could
   induce metabolic changes. Treatment with L. reuteri increased iron and
   glucose metabolism in adipocytes on days 2 and 8. Furthermore, an
   increase in HIF-1 signaling, which may regulate these metabolic
   changes, was identified specifically in adipocytes on day 2. This
   suggests that the activation of HIF-1 signaling potentially influences
   metabolic changes in adipocytes. In general, HIF-1α, a crucial
   transcription factor of HIF-1 signaling and a metabolic regulator
   induced by hypoxia,[199]^56 increased in early differentiating
   adipocytes through mRNA expression analysis ([200]Figure 8A).
   Activation of HIF-1 signaling stimulates glycolysis by upregulating
   phosphoglycerate kinase 1 (PGK1) and lactate dehydrogenase A (LDHA)
   while concurrently inhibiting pyruvate dehydrogenase (PDH) through
   increased expression of pyruvate dehydrogenase kinase 1 (PDK1), which
   suppresses the entry of pyruvate into the TCA cycle.[201]^57 This
   metabolic modulation aligns with the observed changes in adipocytes
   treated with L. reuteri and provides evidence supporting the activation
   of HIF-1 signaling ([202]Figure 5B). Moreover, the increase in HIF-1α
   levels in adipocytes can result in reduced BCAA catabolism, consistent
   with prior research findings.[203]^58^,[204]^59 HIF-1 signaling might
   have been activated through hypoxia induced by L. reuteri, but not all
   metabolic processes regulated under hypoxia are dependent on
   HIF-1α.[205]^57 This implies that by investigating alternative
   pathways, it is possible to determine whether the increase in HIF-1α is
   independent of or dependent on hypoxia. Recent studies have
   demonstrated that hypoxia can activate the Nuclear Factor
   κ-light-chain-enhancer of activated B cells (NF-κB) pathway in
   adipocytes, particularly the canonical NF-κB pathway regulated by IκB
   kinase (IKK).[206]^60 In our results, most enzymes involved in the
   canonical NF-κB pathway did not show significant changes
   ([207]Figure 8B). According to the results, HIF-1 signaling is
   activated independently of hypoxia, suggesting that this may be a
   response to specific substances within L. reuteri rather than
   environmental changes resulting from L. reuteri treatment. Given that
   most metabolic changes in adipocytes treated with L. reuteri
   involve HIF-1 signaling, HIF-1α can be considered another potential
   target of L. reuteri contributing to the suppression of adipogenesis.

Figure 8.

   [208]Figure 8
   [209]Open in a new tab

   HIF-1α as a potential target for L. reuteri in adipogenesis inhibition

   (A) Relative mRNA expression of HIF-1α and (n = 4).

   (B) Relative mRNA expression levels of major factors involved in the
   canonical NF-κB pathway in early differentiating adipocytes (n = 4).

   (C) Illustration of lipin-1 isoform expression by RNA splicing.

   (D) Relative protein levels in early differentiating adipocytes (n =
   4).

   (E) Relative mRNA level of lipin-1α in early differentiating adipocytes
   (n = 4).

   Data are represented as means ± SD. Statistical significance was
   analyzed using unpaired two-tailed Student’s t test and indicated as
   ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns: no significant.

   For HIF-1α to be recognized as a target to inhibit adipogenesis, it
   must selectively act in early adipocyte differentiation and interfere
   with adipogenesis. Previous studies investigating the impact of HIF-1α
   on adipogenesis have reported its dual effect, involving lipid
   accumulation and inhibition.[210]^61 The reported adipogenic effect of
   HIF-1α is predominantly association with mature adipocytes,[211]^61
   whereas the antiadipogenic effect is predominantly association with
   differentiating adipocytes.[212]^61 These reports suggest that the role
   of HIF-1α differ depending on the adipocyte differentiation stage. The
   effect of HIF-1α is consistent with the effect of L. reuteri in
   selectively targeting early adipocyte differentiation, which
   consequently means that HIF-1α can be considered a target of
   L. reuteri. To explore the mechanism through which elevated levels of
   HIF-1α may influence early adipocyte differentiation, we investigated
   the transcription factors that potentially affect lipid metabolism and
   are activated by HIF-1α. We focused on the fact that the expression of
   lipin-1 in adipocytes can be induced by HIF-1α.[213]^62 Interestingly,
   HIF-1α reduces the expression of lipin-1 in differentiating adipocytes
   and increases its expression in mature adipocytes.[214]^63 Lipin-1 has
   two isoforms, lipin-1α, which is expressed during early differentiation
   of adipocytes and promotes adipogenesis, and lipin-1β, which is
   expressed in adipocytes and involved in lipid accumulation, and its
   expression is regulated by RNA splicing.[215]^64 This suggests that
   HIF-1α upregulates the expression of lipin-1β while downregulating the
   expression of lipin-1α, suggesting a possible differential effect as a
   result of the regulation of lipin-1 depending on the timing of
   adipocyte differentiation. Based on previous studies, we hypothesized
   that L. reuteri could inhibit early adipogenesis by reducing the
   expression of lipin-1α. Thus, we investigated the changes in
   serine/arginine splicing factor 10 (Srsf10), an enzyme involved in the
   RNA splicing of lipin-1. In our results, Srsf10 was found to be
   significantly reduced in early differentiating adipocytes
   ([216]Figure 8D). This decrease in Srsf10 may eventually lead to a
   reduction in lipin-1α expression, which was confirmed through
   additional mRNA expression analysis ([217]Figure 8E). Therefore, it was
   suggested that L. reuteri can inhibit early adipocyte differentiation
   by reducing lipin-1 through upregulation of HIF-1α and subsequent
   reduction of Srsf10. Furthermore, HIF-1α induces metabolic changes in
   adipocytes in response to L. reuteri, inhibiting adipogenesis during
   early differentiation.

Conclusions

   Our advanced multiomics strategy identifies the critical treatment
   period in which L. reuteri effectively exhibits the anti-obesity effect
   and identifies the molecular mechanism inhibiting adipogenesis through
   multiomics analysis. In exploring the intricate mechanisms underlying
   adipogenesis, this study delves into the specific inhibition of BCAA
   catabolism by L. reuteri during early differentiation—a critical
   juncture that may pivot toward lipid production. Our findings
   illuminate how L. reuteri orchestrates a nuanced modulation of
   adipocyte differentiation by simultaneously decreasing KLF5 and
   increasing HIF-1α, initiating a cascade that leads to the attenuation
   of BCAA catabolism and PPARγ expression, resulting in the suppression
   of adipogenesis. This dual modulation not only highlights KLF5 and
   HIF-1α as pivotal targets of the action of L. reuteri but also unravels
   the connection between KLF5 and BCAA catabolism. Such insights extend
   our understanding regarding the regulatory pathways in adipogenesis,
   suggesting new angles of investigation. Moreover, the efficacy of our
   advanced multiomics strategy in uncovering these novel mechanisms and
   target molecules emphasizes potential of this approach in identifying
   the probiotic-regulated pathways of adipogenesis inhibition. Therefore,
   our advanced multiomics strategy has successfully elucidated the
   complex causal dynamics of adipocytes and microorganisms and identified
   critical targets for probiotics to exhibit anti-obesity properties.
   Probiotics demonstrating beneficial effects on obesity are promising
   candidates for the next generation of obesity treatment. Systematically
   identifying the MOA and targets through validated methods is crucial
   for their development as treatments.

Limitations of the study

   In this study, we proposed a multiomics-based strategy to elucidate the
   suppression mechanism of potential probiotics on lipid formation. By
   providing complementary validation through integrated multiomics data,
   we gained insights into the putative mechanisms by which probiotics
   inhibit lipid formation and identified potential target proteins.
   However, to confirm whether these proposed mechanisms function as
   hypothesized, additional validation through experiments using knock-out
   or overexpression models targeting these proteins is necessary.
   Nevertheless, given the importance of a molecular understanding of the
   complex mechanisms of adipocytes in characterizing anti-obesity
   effects, this multiomics-based in-depth/systematic approach could serve
   as a foundational framework for future research. Additionally, to
   evaluate the anti-obesity effects of live probiotics more robustly, it
   is crucial to consider extending this strategy to host-microbe
   interaction coculture platforms and in vivo models.

Resource availability

Lead contact

   Further information and requests for resources and reagents should be
   directed to and will be fulfilled by the lead contact, Yun-Gon Kim
   (ygkim@ssu.ac.kr).

Materials availability

   This study did not generate new unique reagents.

Data and code availability

     * •
       The mass spectrometry proteomics data have been deposited at
       ProteomeXchange Consortium via the PRIDE partner repository. The
       mass spectrometry metabolomics data have been deposited at the
       MetaboLights database. Both datasets are listed in the [218]key
       resources table and publicly available as of the date of
       publication. (PRIDE: [219]PXD050884 and Metabolight: MTBLS12185).
     * •
       This paper does not report any original code.
     * •
       Any additional information required to reanalyze the data reported
       in this paper is available from the [220]lead contact upon request.

Acknowledgments