Increased Adipogenic Mesenchymal Stem Cell Programming Mediated by Impaired Wnt Signaling Due to Maternal Obesity, Cesarean Delivery and Infant Formula Feeding: The Perinatal March Toward Obesity

1. Introduction

The global epidemics of obesity and diabetes continue to advance unabatedly. In the United States, 70% of adults are overweight or obese. Childhood obesity has nearly doubled over the past two decades [1]. Children and adolescents with obesity are approximately five times more likely to have obesity in adulthood compared with their normal-weight peers [2].

Obesity, in turn, increases the risk of type 2 diabetes mellitus [3]. The perinatal period has emerged as a critical window for the development of obesity and diabetes and has therefore become a focus of scientific interest [4]. In particular, it is becoming increasingly evident that external and internal influences during the first 1000 days of life after conception have a substantial programming effect on the development of obesity [5,6,7,8,9].

Accumulating evidence indicates that dysregulated programming of mesenchymal stem cells (MSCs) toward adipocytes and, consequently, early adipogenesis, may already begin during fetal life [10,11,12,13,14,15,16,17,18]. Multipotent MSCs can self-renew or, in response to specific signals, further differentiate towards adipocytes, osteocytes, chondrocytes or myocytes [19,20].

A signaling pathway that is critical for MSC differentiation is canonical Wnt/β-catenin signaling. When activated, MSCs preferentially differentiate into osteocytes (osteogenesis) [21,22,23,24], whereas attenuation of Wnt/β-catenin signaling promotes their differentiation into adipocytes (adipogenesis) [25,26,27,28,29,30,31].

Boyle et al. [10] compared Wnt signaling of umbilical cord (UC)-MSCs from infants born to obese mothers (Ob-MSCs) versus mothers with normal weight (NW-MSCs) and found reduced Wnt/β-catenin signaling and a greater potential for adipogenesis in Ob-MSCs. This important observation in concert with literature evidence raised our research questions: (1) Is reduced Wnt/β-catenin signaling the fundamental mechanism driving adipogenic commitment of MSCs during the perinatal period of life? (2) Are maternal obesity, cesarean delivery and artifical infant formula feeding, which are all epidemilogically associated with an enhanced obesity risk, related to impaired Wnt signaling of MSCs? To follow these questions, the reader should be familiar with the mode of action of canonical Wnt/β-catenin signaling. The aim of this review is to synthesize mechanistic literature evidence and propose a unifying hypothesis that maternal obesity, cesarean delivery, and infant formula feeding converge on impaired Wnt/β-catenin signaling to drive adipogenic programming of MSCs.

2. Basic Principles of Canonical Wnt/β-Catenin Signaling

Among the molecular pathways reported to be altered in Ob-MSCs, including redox/forkhead box protein O1 (FOXO1) signaling, DNA methylation, and mitochondrial function, canonical Wnt/β-catenin is uniquely positioned at the intersection of intestinal stem cell homeostasis, enteroendocrine glucagon-like peptide 1 (GLP-1) secretion, and mesenchymal lineage choice, which motivates its selection as the organizing framework of this review. For a deeper understanding of canonical and non-canonical Wnt signaling we refer to recent excellent expert reviews [32,33,34,35,36]. Figure 1 illustrates basic principles of canonical Wnt/β-catenin signaling modified according to Rim et al. [32].

Among the best-characterized Wnt target genes are LGR5 (leucine-rich repeat-containing G protein-coupled receptor 5), a widely used marker of intestinal stem cells (ISCs), and TCF7L2 (transcription factor 7-like 2, also known as TCF4). LGR5 binds R-spondins and thereby potentiates Wnt signaling, supporting the maintenance and expansion of the ISC compartment [32,37]. The Wnt/β-catenin signaling pathway plays a critical role in increasing the activity of ISCs and promotes intestinal development [38,39,40,41,42]. Appropriate ISC activity plays also a key role for further differentiation of intestinal cells into nonsecretory and secretory intestinal cells. Knockdown of the nuclear Wnt effector TCF7L2 in mice negatively affected the ISC population and abrogated the development of enteroendocrine cells (EECs) [42]. Thus, Wnt signaling maintains intestinal homeostasis and the development of enteroendocrine cells.

2.1. Canonical Wnt/β-Catenin/TCF7L2 Signaling Enhances GLP-1 Signaling

Importantly, β-catenin/TCF7L2 binds the GCG promoter/enhancer and increases proglucagon expression. Experiments using the glycogen synthase kinase 3β (GSK3β) inhibitor lithium activates β-catenin signaling and constitutively active β-catenin increases both GCG mRNA and GLP-1 production in intestinal endocrine L-cells [43]. Shao et al. [44] confirmed that the Wnt signaling pathway effector TCF7L2 controls gut and brain proglucagon gene expression and glucose homeostasis. The TCF7L2 polymorphism rs7903146 is considered one of the strongest and most consistently replicated genetic risk factors for type 2 diabetes mellitus among common genetic variants. The T allele has been associated with reduced insulin content and secretion, and impaired incretin effects. GLP-1 is a proteolytic cleavage product of proglucagon [45]. Further evidence supports that GLP-1/GLP-1 receptor (GLP1R) signaling is upregulated by appropriate Wnt signaling [46,47,48,49,50,51]. As shown in pancreatic β-cells, Wnt/TCF7L2 signaling upregulates the LIM homeodomain protein ISL1, which enhances the expression of GLP1R [51].

2.2. Wnt/β-Catenin Signaling and C/EBPβ- and PPARγ-Mediated MSC Adipogenic Fate Determination

Wnt/β-catenin signaling inhibits MSC adipogenic differentiation but stimulates osteogenic differentiation [19,20]. The adipogenic transcription factor peroxisome proliferator-activated receptor-γ (PPARγ) plays a key role for adipogenic differentiation and promotes the expression of crucial genes driving adipogenesis [52,53], whereas Runt-related transcription factor 2 (RUNX2) is a pivotal transcription factor promoting osteogenesis [54,55] (Figure 2).

Two decades ago, an important crosstalk between Wnt/β-catenin signaling and PPARγ activity was established [56,57,58,59]. The differentiation of preadipocytes into adipocytes requires suppression of canonical Wnt signaling, which involves PPARγ-associated targeting of β-catenin to the proteasome [57]. Indeed, sustained activation of β-catenin through the expression of Wnt1 or Wnt10b in preadipocytes blocks adipogenesis by inhibiting PPARγ-associated gene expression. Liu et al. [57] showed that activation of PPARγ, by exposing Swiss mouse fibroblasts to troglitazone, stimulates GSK3β-dependent degradation of β-catenin, whereas a β-catenin mutant defective in GSK3β-mediated phosphorylation inhibits the expression of a subset of adipogenic genes. These pioneering data provide evidence that GSK3β is a key regulator of the balance between β-catenin and PPARγ activity. Activation of canonical Wnt signaling downstream of PPARγ blocks the expression of a select subset of adipogenic genes [57].

Furthermore, Liu et al. [58] identified a direct functional interaction between β-catenin and PPARγ involving the TCF/LEF-binding domain of β-catenin and a catenin-binding domain (CBD) within PPARγ, which enhances the proteasomal degradation of β-catenin. In addition, Wnt signaling induces the expression of WNT1-inducible signaling pathway protein 2 (WISP2), which inactivates the transcription factor zinc finger protein 423 (ZFP423) [60]. WISP2, a secreted protein encoded by a Wnt target gene, acts as a mesenchymal cell activator that enhances Wnt signaling while inhibiting PPARγ activation and adipogenic differentiation [60]. WISP2 is both secreted by mesenchymal precursor cells and highly expressed in their cytosol, where it binds the PPARγ transcriptional activator ZFP423 [61], thereby preventing its nuclear translocation [62].

Another Wnt target gene encodes WNT1-inducible signaling pathway protein 1 (WISP1), which inactivates PPARγ. Inactivation of PPARγ transcriptional activity was mediated, at least in part, by a direct physical association between WISP1 and PPARγ, followed by proteasome-dependent degradation of PPARγ [63]. Furthermore, Xu et al. [64] showed that Wnt5a represses PPARγ transactivation through H3K9 histone methyltransferase protein complexes.

CCAAT/enhancer-binding protein β (C/EBPβ) is rapidly induced during the early stages of adipogenesis and is responsible for the transcriptional induction of two major adipogenic transcription factors: PPARγ and CCAAT/enhancer-binding protein α (C/EBPα). Remarkably, overexpression of C/EBPβ substantially inhibited Wnt signaling. C/EBPβ is involved in the suppression of Wnt10b through direct binding to specific regions of the WNT10B promoter [65]. PPARγ also inhibits RUNX2 expression [66,67,68]. In particular, PPARγ physically interacts with RUNX2 and can interfere with RUNX2 binding and transactivation at osteoblast-specific element 2 (OSE2), thereby reducing RUNX2-driven gene transcription [68].

2.3. Molecular Crosstalk between GLP-1- and Wnt-Signaling Restricts Adipogenic Differentiation but Promotes Osteogenic Differentiation of MSCs

GLP-1 is another important regulator that restricts the adipogenic potential of MSCs, as recently reviewed by Weber et al. [69]. Table 1 represents observed GLP-1- and GLP1R agonist-mediated stem cell effects.

2.4. GLP-1 Augments Canonical Wnt Signaling

Wnt and GLP-1 signals exert synergistic effects on stem cell regulation. Wnt signaling not only enhances GLP-1 signaling [46,47,48,49,50,51], but GLP-1 also promotes Wnt/β-catenin signaling [73,76,79].

To understand the synergistic effects of GLP-1 and Wnt signaling, it is important to note that the activated GLP1R preferentially couples to Gαs, increasing cAMP levels and subsequently activating protein kinase A (PKA) [80,81,82]. Hino et al. [83] showed that phosphorylation of β-catenin by PKA inhibits β-catenin ubiquitination. Ser675 was identified as a PKA phosphorylation site. Substitution of this serine residue with alanine attenuated PKA-mediated inhibition of β-catenin ubiquitination, PKA-induced stabilization of β-catenin, and PKA-dependent activation of TCF. GLP-1/PKA activation enhances the nuclear translocation of β-catenin and the initiation of osteogenic gene expression. Activated PKA also phosphorylates phosphoinositide 3-kinase (PI3K), leading to AKT activation. AKT subsequently phosphorylates GSK3β, inhibiting its activity [79]. GSK3β-mediated degradation of β-catenin is thus inhibited, and nuclear accumulation of β-catenin is enhanced.

Notably, Zhao et al. [84] reported that AMP-activated protein kinase (AMPK) phosphorylates β-catenin at Ser552, thereby enhancing β-catenin/TCF-mediated transcription, in C3H10T1/2 mesenchymal cells. This murine embryonic fibroblast/mesenchymal progenitor cell line can differentiate into osteogenic and adipogenic lineages. Figure 3 illustrates the molecular interaction between Wnt and GLP-1 signals in MSC differentiation.

3. Maternal Obesity

The influence of maternal overweight and obesity on the perinatal differentiation of adipocyte precursor cells as a predisposing factor for subsequent overweight in offspring is currently a focus of scientific interest [85]. An important question is: How does maternal overweight or obesity during pregnancy influence the adipogenic potential of fetal MSCs? An elegant, noninvasive retrospective approach that imposes no burden on either mother or child during pregnancy is the isolation of MSCs from the umbilical cord (UC-MSCs) [10,16] or umbilical cord blood [86]. However, differences in MSC adipogenic potential have been observed depending on the tissue source, in the following order: adipose tissue-derived MSCs (AD-MSCs) > bone marrow-derived MSCs (BM-MSCs) > umbilical cord-derived MSCs (UC-MSCs) [87]. In particular, PPARγ mRNA expression was highest in AD-MSCs, lowest in UC-MSCs, and intermediate in BM-MSCs. These results suggest that differences in adipogenic potential among the three MSC types are associated with their basal PPARγ mRNA expression levels [87].

Bae et al. [86] investigated the adipogenic differentiation of MSCs isolated from umbilical cord blood and confirmed that increased Wnt5a/β-catenin signaling inhibits their adipogenic differentiation. The addition of Wnt5a inhibited adipogenic differentiation and reduced PPARγ expression and lipid storage. In contrast, depletion of β-catenin increased the adipogenic potential of umbilical cord blood MSCs. Interestingly, in a study involving 44 neonates, lower GLP-1 concentrations were measured in the umbilical cord blood of macrosomic neonates than in that of normal-weight neonates (6.9±2.96 ±2.9 vs. 10.3±3.7 pg/mL) [88]. Thus, MSCs in the umbilical cord and/or umbilical cord blood of neonates delivered by obese mothers may have been exposed to reduced GLP-1 signaling prenatally.

Undifferentiated UC-MSCs from infants of mothers with obesity (Ob-MSCs) exhibited higher protein levels of the transcription factor ZNF423 (zinc finger protein 423) than UC-MSCs from infants of normal-weight mothers (NW-MSCs), a difference associated with adipocyte hypertrophy following MSC differentiation [13]. ZNF423 is of major importance for adipogenesis because it induces the expression of the adipogenic transcription factor PPARγ [61]. Notably, ZNF423 activity is negatively regulated by Wnt signaling. Wnt signaling induces the expression of WISP1 and WISP2, both of which inhibit PPARγ activity [62,63].

In this context, the study by Boyle et al. [10] is highly informative. The authors investigated Wnt/β-catenin signaling and adipogenic differentiation in UC-MSCs from infants of normal-weight mothers (NW-MSCs) and infants of mothers with obesity (Ob-MSCs). Following differentiation, Ob-MSCs exhibited increased adipogenesis, with a 50% increase in PPARγ protein levels compared with NW-MSCs. Ob-MSCs showed 25% lower inhibitory phosphorylation of GSK3β, suggesting increased β-catenin degradation. However, this study did not investigate the effects of GLP-1 on PKA-mediated phosphorylation of β-catenin or AKT-mediated inhibitory phosphorylation of GSK3β.

Bellalta et al. [18] recently characterized the proteins in the MSC secretome following the isolation of UC-MSCs from Wharton’s jelly of infants born to normal-weight mothers (NW-MSCs) and infants born to mothers with obesity (Ob-MSCs). They found increased expression of the Wnt inhibitor SFRP1 (secreted frizzled-related protein 1) in Ob-MSCs compared with NW-MSCs. SFRP1 is a proadipogenic protein that inhibits Wnt signaling [89]. Similarly, the Wnt antagonist SFRP4 (secreted frizzled-related protein 4) increases adipogenic differentiation in human adipose tissue-derived MSCs [90].

In a further study, Bellalta et al. [17] showed that maternal overweight increases oxidative stress in MSCs, representing another potential mechanism associated with PPARγ activation. Compared with NW-MSCs, Ob-MSCs exhibited reduced stemness characteristics, including lower OCT4 (octamer-binding transcription factor 4, POU5F1) expression and decreased clonogenic capacity.

Remarkably, Wnt signaling and the transcription factor OCT4 are interconnected in determining early cell fate during differentiation and development [91]. OCT4 is a major regulator of the pluripotency network, controlling whether cells remain stem cells or begin differentiating into specialized cell types. Thus, OCT4 is associated with MSC stemness. Its overexpression increases MSC proliferation and colony-forming capacity [92,93]. Importantly, Li et al. [94] identified OCT4 as a novel target of Wnt signaling and demonstrated that the POU5F1 promoter is regulated by β-catenin/TCF. Indeed, OCT4 is a direct target of β-catenin/TCF-mediated transcription, and the binding site at positions −875/−881 of the POU5F1 promoter is critical for β-catenin/TCF-dependent gene expression.

In addition, Kelly et al. [95] provided evidence that β-catenin enhances OCT4 activity and reinforces pluripotency through a TCF-independent mechanism. They showed that stabilized β-catenin forms a complex with OCT4 and enhances its activity. OCT4 is a core component of the transcriptional network regulating pluripotency. Hade et al. [96] showed that, during the differentiation of preadipocytes into adipocytes, OCT4 levels declined significantly, whereas levels of PPARγ, the master transcription factor of mature adipocytes, increased. Intriguingly, Hannan et al. [97] observed that human embryonic stem cells (ESCs) transduced with an artificial OCT4-targeting short hairpin RNA displayed remarkably high lineage-specific spontaneous differentiation into adipocytes. Of note, the GLP1R agonist semaglutide stimulated the proliferation and maintenance of stemness in jaw bone marrow MSCs, accompanied by increased OCT4 and RUNX2 expression and enhanced osteogenesis [98].

Boyle et al. [99] observed reduced fatty acid oxidation (FAO) in UC-MSCs from infants of mothers with obesity (Ob-MSCs). In the Ob-MSC group with the most pronounced metabolic disturbances, correlations were found with neonatal adiposity and elevated umbilical cord blood insulin levels. Ob-MSCs with reduced FAO also showed lower AMPK activity. This finding again suggests a relationship with reduced Wnt signaling: AMPK, like GLP-1-activated PKA, phosphorylates β-catenin, thereby increasing its transcriptional activity, whereas AMPK inhibition resulted in increased PPARγ levels [84].

Erickson et al. [12] compared three metabolic profiles in UC-MSCs: (1) MSCs with normal fatty acid oxidation (FAO) in the setting of normal maternal weight, (2) MSCs with normal FAO in the setting of maternal obesity, and (3) MSCs with reduced FAO in the setting of maternal obesity. MSC FAO was negatively correlated with maternal fasting blood glucose and insulin levels, positively correlated with HDL cholesterol levels, and negatively associated with infant fat mass.

Notably, FAO is also regulated by Wnt/β-catenin signaling. In osteoblasts, activation of the Wnt signaling pathway increased oleic acid oxidation, whereas depletion of β-catenin reduced FAO and the expression of enzymes involved in β-oxidation [100]. In liver cells, loss of β-catenin likewise reduced FAO [101].

Iaffaldano et al. [102] observed that mitochondrial basal respiration (control: 46.9 ± 0.7; Ob: 32.6 ± 0.8 pmol/min), ATP-linked respiration (control: 29.3 ± 1.9; Ob: 20.1 ± 0.3 pmol/min), and maximal respiration (control: 75.2 ± 5.3; Ob: 50.5 ± 4.1 pmol/min) were significantly (P < 0.0001) lower in Ob-UC-MSCs versus control UC-MSCs. Paz et al. [103] reported also a trend toward lower maximal respiration and spare respiratory capacity of Ob-MSCs compared to NW-MSCs. Unexpectedly and in contrast to all other UC-MSC studies, Ob-MSCs exhibited an impaired adipogenic phenotype, characterized by reduced protein levels of C/EBPα and PPARγ during differentiation.

In summary, most UC-MSC studies reported enhanced adipogenic differentiation and decreased FAO in association with maternal obesity, findings that can be linked to reduced Wnt/β-catenin signaling.

4. Gestational Diabetes Mellitus

Overweight and obesity increase the risk of gestational diabetes mellitus (GDM) [104]. However, findings regarding the adipogenic potential of MSCs under hyperglycemic conditions are conflicting [105]. Cramer et al. [106] observed increased adipogenic potential and reduced osteogenic and chondrogenic differentiation in adipose tissue-derived MSCs exposed to high glucose concentrations. Incubation of adult bone marrow MSCs with sera from patients with type 2 diabetes mellitus increased adipogenic gene expression and lipid storage [107]. In contrast, Kim et al. [108] found reduced adipogenic and osteogenic differentiation in UC-MSCs from mothers with GDM compared with those from normoglycemic mothers. Bone marrow MSCs also showed reduced adipogenic differentiation following exposure to high glucose concentrations [109]. Cheng et al. [110] compared human adipose-derived stem cells obtained from the subcutaneous adipose tissue of donors with diabetes (dASCs) and donors without diabetes (nASCs). Although dASCs and nASCs treated with high glucose concentrations (4.5 g/L) exhibited decreased proliferation, increased expression of the pluripotency markers SOX2 (sex determining region Y)-box 2), OCT4, and homeobox protein NANOG was observed. Moreover, nASCs treated with high glucose concentrations exhibited decreased cell migration, enhanced senescence, and significantly higher intracellular reactive oxygen species levels. Their adipogenic and osteogenic differentiation capacities, however, remained comparable to those of cells treated with low glucose concentrations (1.0 g/L). Hyperglycemic conditions appear, in general, to reduce MSC differentiation potential and accelerate MSC senescence [105].

5. Cesarean Delivery

Mothers with overweight or obesity give statistically more births by cesarean delivery than normal-weight mothers [111,112]. The Federal Statistical Office of Germany had reported that cesarean delivery accounted for 33% of all births in Germany in 2024 [113], and a further global increase is expected. According to Betran et al. [114], the latest available data from 2010–2018, covering 154 countries and 94.5% of live births worldwide, indicate that 21.1% of women gave birth by cesarean delivery, with regional averages ranging from 5% in sub-Saharan Africa to 42.8% in Latin America and the Caribbean. Epidemiological data indicate that cesarean delivery is associated with a significantly increased risk of obesity compared with vaginal delivery. In a meta-analysis of 28 studies, Kuhle et al. [115] demonstrated that cesarean delivery was associated with an increased risk of childhood obesity compared with vaginal delivery (RR: 1.34; 95% CI: 1.18–1.51). A recently published meta-analysis of 20 studies likewise confirmed an increased risk of obesity in early childhood following cesarean delivery compared with vaginal delivery (RR: 1.23; 95% CI: 1.12–1.34) [116].

Since the influence of intrauterine factors on MSCs ends with the birth, the question arises whether, and if yes, how cesarean delivery might further promote adipogenic differentiation of MSCs. One possible explanation is the absence of the effects of vaginal delivery on the infant gut microbiota. In particular, cesarean delivery results in less efficient transmission of Bifidobacteria than vaginal delivery. Indeed, during the first postnatal months, the intestinal microbiome of infants delivered by cesarean delivery contains significantly reduced numbers of Bifidobacteria and Bacteroides species [117,118,119,120,121,122,123,124,125].

Acetate and lactate are the main products of human milk oligosaccharide (HMO) fermentation by Bifidobacterium species [126,127]. Beyond carbohydrate fermentation, Bifidobacterium species convert aromatic amino acids, such as tryptophan, phenylalanine, and tyrosine, into aromatic lactic acids—indole-3-lactic acid (ILA), phenyllactic acid, and 4-hydroxyphenyllactic acid—through an aromatic lactate dehydrogenase present in B. longum, B. breve, and B. bifidum [126,127]. Gastrointestinal Bacteroidota species produce acetate as one of their major short-chain fatty acid (SCFA) end products [128,129,130]. Notably, Bacteroides thetaiotaomicron can utilize human milk carbohydrates and release acetate [131].

Martin et al. [117] demonstrated that the delivery mode affected the organic acid profile produced by the intestinal microbiome. During the first and second months of life, the probability of detecting acetic and lactic acids was significantly lower in infants born by cesarean delivery than in those born vaginally. This is of critical importance because microbiota-derived lactate and acetate bind to their respective receptors, G protein-coupled receptor 81 (GPR81) and free fatty acid receptor 2 (FFAR2), and stimulate Wnt3 production by Paneth cells. Wnt3 then activates the Wnt3/β-catenin signaling pathway in neighboring ISCs [132,133]. Consistent with these findings, Wu et al. [134] recently reported that Lactobacillus amylovorus–lactate–GPR81 signaling stimulated β-catenin production and the differentiation of jejunal organoids.

Thus, impaired microbiota-derived lactate/acetate signaling due to reduced abundance of Bifidobacteria and Bacteroides species in the early intestinal microbiome after cesarean delivery may impair Wnt3-dependent ISC homeostasis (Figure 4).

Accumulating evidence indicates that gut microbiota can influence obesity. The gut microbiota contributes to host metabolism through several mechanisms, including increased dietary energy harvest, modulation of lipid metabolism, altered endocrine function, and increased inflammatory activity [135,136,137,138,139,140,141].

Yahagi et al. [142] investigated longitudinal changes in the gut microbiota and organic acid concentrations during the first two years of life in 12 Japanese infants. They confirmed that Bifidobacteria contribute significantly to organic acid production in healthy infants. In particular, acetate, lactate, and formiate, reported as rarely observed in adults in this context, are characteristic of breastfed infancy. Bifidobacterium longum subsp. infantis and the symbiotic activity of Bifidobacterium bifidum and Bifidobacterium breve efficiently produce these organic acids by metabolizing HMOs. Thus, HMO-utilizing Bifidobacteria play an important role in regulating infant gut organic acid profiles.

FFAR2, also known as G protein-coupled receptor 43 (GPR43), binds SCFAs produced by microbial carbohydrate fermentation [143]. FFAR2/GPR43 and FFAR3/GPR41 are expressed on enteroendocrine L cells and participate in GLP-1 secretion [144]. It is also important to note that lactate, through binding to its receptor GPR81, and acetate and propionate, through binding to FFAR2, increase GLP-1 secretion [145,146].

Pedersen et al. [147] observed that oral administration of 25 g of sodium lactate plus 6 g of sodium chloride to young adult men significantly increased plasma GLP-1 levels. Treatment with Bifidobacterium animalis subsp. lactis GCL2505 (BlaG) increased acetate and GLP-1 levels in both the gut and plasma [148].

Activation of FFAR2 by acetate and propionate has been identified as an important trigger for GLP-1 production and release by enteroendocrine L cells in the lower intestine. Notably, fructooligosaccharide supplementation in rats increased the density of FFAR2-positive EECs in the proximal colon by more than twofold compared with controls, alongside an increase in GLP-1-containing L cells [149]. In contrast, FFAR2-knockout (GPR43−/−) mice exhibit reduced SCFA-induced GLP-1 secretion [150].

Importantly, Bifidobacterium species associated with breastfeeding produce aromatic lactic acids in the infant gut [151]. Tryptophan metabolites, such as indole, stimulate GLP-1 release from enteroendocrine L cells through mechanisms involving the aryl hydrocarbon receptor (AhR) [152]. Vrzalová et al. [153] demonstrated that mixtures of intestinal microbial tryptophan metabolites influence AhR activity. Notably, Bacteroides thetaiotaomicron, B. eggerthii, B. ovatus, and B. fragilis produce indole-3-acetic acid (IAA) from tryptophan [154]. Another study involving 596 infants found that Bacteroides species were significantly less frequently detected, or were present at very low abundance, in infants born by cesarean delivery. Higher Bacteroides abundance was detected in 49.0% of vaginally delivered infants, compared with only 0.4% of infants born by cesarean delivery [119].

Indole-3-lactate has been shown to promote ISC differentiation into EECs [155]. In addition, IAA activates AhR and increases GLP-1 secretion by upregulating GCG expression and promoting stem cell differentiation into L cells [156]. Furthermore, supplementation with an AhR agonist or a Lactobacillus strain with a high capacity to produce AhR ligands improved intestinal barrier function and GLP-1 secretion [157].

Importantly, acetic and lactic acid levels are lower in infants born by cesarean delivery than in those born vaginally [117]. Compared with formula-fed infants, breastfed infants have higher concentrations of acetate, lactate, and indole-3-lactate in their stool [158,159,160].

Taken together, cesarean delivery may impair SCFA and indole/indole derivative-mediated enteroendocrine GLP-1 signaling. This may subsequently reduce GLP-1-stimulated Wnt/β-catenin signaling in MSCs, providing a further stimulus for their adipogenic differentiation [69,70,71,72,73,74,75,76,77,78,79].

6. Infant Formula Feeding

The metabolic relay race toward obesity continues with infant formula feeding. Cesarean delivery before the onset of labor is associated with delayed or less successful initiation of breastfeeding [161,162,163]. One possible explanation is that cesarean delivery, particularly emergency cesarean section performed before the onset of labor, reduces maternal endogenous release of oxytocin and prolactin [164,165].

A Canadian study by Rossiter et al. [166] concluded that exclusive breastfeeding without formula supplementation has a protective effect against the development of childhood obesity. The U.S. NHANES study showed that even short-term breastfeeding was associated with a lower risk of childhood obesity than formula feeding (OR: 0.31; 95% CI: 0.17–0.59). Children who were breastfed for three to six months had a lower risk of obesity at three to four years of age (OR: 0.56; 95% CI: 0.35–0.89) [167]. A recent meta-analysis including 26 publications (n=332,297) confirmed an inverse relationship between breastfeeding and obesity risk in children aged two to six years [168]. A dose–response relationship with breastfeeding duration was observed: one day to less than three months, OR: 1.07 (95% CI: 0.94–1.21); three to less than six months, OR: 0.96 (95% CI: 0.60–1.54); and six months or longer, OR: 0.67 (95% CI: 0.58–0.77). Each additional month of breastfeeding was associated with 4.0% lower odds of obesity (OR: 0.96 per month; 95% CI: 0.95–0.97).

Excessive postnatal protein intake, known as the “early protein hypothesis of obesity,” has been proposed as a possible explanation for the adipogenic effects of infant formula feeding [169]. Melnik et al. [170] recently provided a mechanistic explanation for this hypothesis based on impaired Wnt/β-catenin signaling in white adipose tissue MSCs. Excessive protein intake resulting from uncontrolled formula feeding or the use of high-protein formula increases serum levels of insulin, insulin-like growth factor 1 (IGF-1), and essential branched-chain amino acids (BCAAs), particularly leucine, in human infants [171,172] and newborn rhesus monkeys (Macaca mulatta) [173]. Notably, the abundance of fecal microbes known to vary in their ability to utilize complex carbohydrates, including Bifidobacterium and Ruminococcus from the Ruminococcaceae family, also increased with reduced protein intake [173].

High protein intake increases serum insulin, IGF-1, and BCAA levels, all of which stimulate the mammalian target of rapamycin complex 1-ribosomal protein S6 kinase beta-1 (mTORC1–S6K1) signaling pathway, a pathway linked to the pathogenesis of early childhood obesity associated with formula feeding [174]. After entering the nucleus, S6K1 can suppress the expression of key Wnt genes [175]. In particular, S6K1-mediated phosphorylation of histone H2B at Ser36 recruit enhancer of zeste homolog 2 (EZH2) to histone H3, where it mediates H3K27 trimethylation. This suppresses the expression of Wnt genes, including Wnt6, Wnt10a, and Wnt10b, while upregulating PPARγ and C/EBPα and increasing adipogenesis.

Scavo et al. [176] demonstrated that IGF-1 promotes cell growth and lipogenesis during the differentiation of human MSCs into adipocytes. Hu et al. [177] observed a critical attenuation of Wnt/β-catenin signaling in adipose tissue MSCs in response to elevated IGF-1 levels. They showed that IGF-1 increases PPARγ expression. PPARγ, in turn, increases the expression of Axin2, an inhibitor of Wnt/β-catenin signaling. Because circulating IGF-1 can reach the vascular stem cell niche of adipose tissue, excessive postnatal protein intake, accompanied by increased IGF-1/mTORC1/S6K1 signaling, may inhibit Wnt/β-catenin signaling in MSCs and thereby promote adipogenic differentiation [170,174,175,176,177,178].

Luk et al. [179] explored the role of the endothelial IGF-1 receptor (IGF1R) as a paracrine modulator of the white adipose tissue phenotype. They showed that reduced endothelial IGF1R expression in male mice receiving a high-fat diet leads to depot-specific, beneficial remodeling of white adipose tissue. Rajashekhar et al. [180] investigated the hypothesis that MSC fate within the perivascular niche could be modulated through interactions with endothelial cells (ECs), which regulate MSC adipogenic potential. An EC-specific effect was accompanied by increased expression of factors involved in Wnt signaling, most prominently Wnt1, Wnt4, and Wnt10a, which are well-known inhibitors of adipogenesis.

Consistent with the reported suppression of Wnt expression by excessive mTORC1–S6K1 signaling in MSCs [175], enhanced nuclear S6K1 signaling in ECs due to high protein intake and elevated plasma insulin, IGF-1, and BCAA levels may reduce EC-derived Wnt expression. This potential mechanism requires further experimental investigation.

Dysregulation of Wnt/β-catenin signaling in MSCs may also be promoted by additional factors associated with formula feeding. Bifidobacteria are not an inherent or mandatory component of infant formula. Some formulas are supplemented with specific Bifidobacterium strains as probiotics, while others contain no added Bifidobacteria. Infant formula does not reproduce the broad diversity and composition of HMOs naturally present in human milk. Human milk contains more than 100 structurally distinct HMOs, whereas HMO-supplemented infant formulas generally contain only a limited number of individual HMOs. Their HMO composition and diversity therefore differ substantially from those of human milk [181,182,183]. As shown in rhesus monkeys, high-protein infant formula reduced the abundance of intestinal Bifidobacteria [173].

Furthermore, infant formula contains no bioactive milk exosomes and exosomal microRNAs [184,185], which support intestinal development [186] and the growth and proliferation of Bifidobacteria [187]. Weil et al. [188] studied the gastrointestinal passage, intestinal epithelial cell uptake, and Argonaute 2 (AGO2) loading of milk microRNAs in neonates using xenomiRs as tracers. Of note, human milk extracellular vesicles (hmEVs) survive neonatal digestion and are taken up by neonatal intestinal enteroids [189].

Milk exosomes activate Wnt/β-catenin signaling and promote ISC activity and cell proliferation [190,191]. Treatment of ISCs with human milk-derived exosomes significantly upregulated the mRNA expression of Axin2, c-Myc, Cyclin D1, and LGR5 compared with exosome-free controls. In the presence of carnosic acid, a specific Wnt/β-catenin signaling inhibitor, cell viability was significantly reduced [190]. Consistent with these findings, Maghraby et al. [191] demonstrated in mice that bovine milk exosomes induced the proliferation of Lgr5+ ISCs in association with Wnt pathway activation.

Gross et al. [192] identified exosomes as carriers of Wnt proteins. The potential transfer of Wnts by milk-derived exosomes is currently under investigation.

Exosomal transfer of microRNA-148a, which suppresses the Wnt inhibitor DKK1 [194], may represent an indirect exosome-mediated mechanism that enhances intestinal Wnt/β-catenin signaling [195].

Unlike infant formula, human milk contains GLP-1 as a bioactive component. GLP-1 concentrations in milk collected at the end of a breastfeeding session, termed hindmilk, were negatively correlated with infant weight gain at six months of age [196]. Human milk also contains oxytocin, in contrast to infant formula [197,198]. Notably, oxytocin upregulates microRNA-148a-3p expression in milk exosomes [199]. Recent findings suggest that oxytocin may also contribute to increased ISC activity [200] and inhibit adipogenesis in adipose tissue and bone marrow stem cells while promoting osteogenesis [201].

Recently, Seitz et al. [202] investigated changes in the milk proteome from colostrum to mature milk. Wnt5a was identified as one of 50 proteins upregulated in colostrum compared with mature milk. Intriguingly, Wnt5a has been shown to promote osteogenic differentiation and inhibit adipogenic differentiation of human MSCs [203]. Bae et al. [86] used qPCR to examine the expression of Wnt family members (Wnt1, Wnt3a, Wnt4, Wnt5a, Wnt5b, and Wnt10b) in two lots of human umbilical cord blood MSCs with high (MSCs-H) and low (MSCs-L) adipogenic differentiation potential. Only Wnt5a mRNA levels were significantly decreased in MSCs-H following adipogenic induction. Under control conditions, β-catenin in MSCs-H was distributed throughout the cytoplasm and strongly localized to the plasma membrane. Adipogenic cells, by contrast, exhibited a reduced area of cytoplasmic β-catenin fluorescence, suggesting an active role for the canonical Wnt pathway during adipogenic differentiation. Western blotting showed higher Wnt5a and β-catenin protein levels in MSCs-L than in MSCs-H. Furthermore, Xu et al. [64] showed that Wnt5a represses PPARγ transactivation through H3K9 histone methyltransferase protein complexes. Intriguingly, Smith et al. [204] demonstrated in a human intestinal organoid model that the addition of human milk, compared with infant formula, increased β-catenin signaling, promoted organoid growth, and enhanced the proliferation of both ISCs and EECs.

Unlike infant formula, human milk contains stem cells with multilineage potential that are transferred to offsprings during breastfeeding [205,206,207]. Remarkably, Widjaja et al. [208] observed mesenchymal stem-like cells with high expression of the surface markers CD44, CD73, CD90, and CD105. They identified Wnt11 in the human milk stem cell secretome under hypoxic conditions and Wnt10b under nonhypoxic conditions. At present, it is uncertain whether human milk-derived MSCs (1) differentiate into adipocytes themselves, (2) promote or inhibit adipogenesis in other cells, or (3) regulate infant fat development during breastfeeding. These are three distinct possibilities that do not apply to MSC-free formula feeding.

Taken together, accumulating evidence suggests that infant formula feeding may also be associated with insufficient activation of Wnt signaling in MSCs, thereby contributing to adipogenic differentiation.

(A) Maternal obesity
Finding	References	Finding	References
In UC blood MSCs Wnt5a mRNA and β-catenin decreased during adipogenic differentiation	[86]	Inhibition of adipogenic differentiation after addition of Wnt5a	[86]
In Ob-MSCs higher protein levels of ZNF423	[13]	Inhibition of ZNF423 by Wnt1-induced WISP2	[62]
In Ob-MSCs increased increased levels of PPARγ protein	[10,17]	The Wnt1-induced proteins WISP1 and WISP2 inhibit PPARγ expression	[63]
Lower inhibitory phosphorylation of GSK3β, suggesting increased degradation of β-catenin.	[10]	Lithium chloride inhibition of GSK3β increased nuclear β-catenin content and normalized nuclear PPARγ in Ob-MSCs.	[10]
Increased expression of the Wnt inhibitor SFRP1 in Ob-MSC secretome	[18]	Proadipogenic SFRP1 inhibits Wnt signaling	[89]
Lower OCT4 (POU5F1) expression in Ob-MSCs	[17]	POU5F1 expression is upregulated by β-catenin/TCF	[94,95]
OCT4 shRNA-transduced human ESCs displayed high lineage-specific differentiation toward adipocytes	[97]	OCT4 levels declined during preadipocyte to adipocyte differentiation, whereas PPARγ increased.	[96]
Reduced fatty acid oxidation (FAO) in Ob-MSCs	[99]	Depletion of β-catenin reduced FAO in osteoblasts and hepatocytes	[100,101]
(B) Cesarean delivery
Finding	References	Finding	References
Lactate deficiency in the infant gut due to cesarean delivery with reduced transmission of lactate-producing Bifidobacteria	[117]	Lactate via GPR81 stimulates Wnt3 expression of Paneth cells that activate Wnt signaling of intestinal stem cells	[132]
Acetate deficiency in the infant gut due to cesarean delivery with reduced transmission of acetate-producing Bifidobacteria and Bacteroides	[117]	Acetate via FFAR2 stimulates Wnt3 expression of Paneth cells that activate Wnt signaling of intestinal stem cells	[133]
Acetate, lactate and indoles and indol-derivatives via FFAR2, GPR81 and AhR, respectively, stimulate GLP1-1 secretion of EECs	[145,146,147,152]	GLP-1 activates MSCs; GLP1R activates PKA, which phosphorylates β-catenin thereby enhancing β-catenin signaling	[69,79]
(C) Formula feeding
Finding	References	Finding	References
Increased IGF-1-TORC1-S6K1 signaling due to excessive formula protein intake	[171,174]	Nuclear S6K1-mediated Wnt gene silencing	[175]
IGF-1 promotes differentiation of human MSCs into adipocytes	[176]	Attenuation of Wnt/β-catenin signaling in adipose tissue MSCs in response to elevated IGF-1 levels due to PPARγ-mediated expression of Axin2, an inhibitor of Wnt/β-catenin signaling	[177]
Absence of milk exosomes in infant formula	[184,185,186]	Human and bovine milk-derived exosomes enhance Wnt/β-catenin signaling of rat and murine intestinal stem cells increasing LGR5 expression	[190,191]
Absent transfer of milk exosomal microRNAs including exosomal microRNA-148a	[184,185,186]	Absent milk microRNA-148a-mediated inhibition of the Wnt inhibitor DKK1	[194]
Fecal deficiencies of acetate, lactate and indole-lactate in formula-fed compared to breastfed infants	[158,159,160]	Yet uncharacterized effects on ISC Wnt signaling and enteroendocrine L cell GLP-1 secretion
Absence of MSCs in infant formula	[205,206,207]	human breastmilk stem cell secretome: Wnt11 secretion under hypoxic, Wnt10b under non-hypoxic conditions	[208]
Wnt5 has been detected in the human breast milk proteome, whereas standard formulas are not known to contain Wnt5a,	[202]	Wnt5a mRNA decreased in MSCs with high adipogenic potential by adipogenic induction. Wnt5a promotes osteogenic but inhibits adipogenic differentiation of human MSCs	[203]
Exposure of fetal small intestine-derived organoids to human breast milk or infant formula	[204]	Organoids supplemented with formula exhibited reduced proliferation and showed decreased expression of β-catenin, TCF3 and TCF7	[204]

(A) Maternal obesity
Finding	References	Finding	References
In UC blood MSCs Wnt5a mRNA and β-catenin decreased during adipogenic differentiation	[86]	Inhibition of adipogenic differentiation after addition of Wnt5a	[86]
In Ob-MSCs higher protein levels of ZNF423	[13]	Inhibition of ZNF423 by Wnt1-induced WISP2	[62]
In Ob-MSCs increased increased levels of PPARγ protein	[10,17]	The Wnt1-induced proteins WISP1 and WISP2 inhibit PPARγ expression	[63]
Lower inhibitory phosphorylation of GSK3β, suggesting increased degradation of β-catenin.	[10]	Lithium chloride inhibition of GSK3β increased nuclear β-catenin content and normalized nuclear PPARγ in Ob-MSCs.	[10]
Increased expression of the Wnt inhibitor SFRP1 in Ob-MSC secretome	[18]	Proadipogenic SFRP1 inhibits Wnt signaling	[89]
Lower OCT4 (POU5F1) expression in Ob-MSCs	[17]	POU5F1 expression is upregulated by β-catenin/TCF	[94,95]
OCT4 shRNA-transduced human ESCs displayed high lineage-specific differentiation toward adipocytes	[97]	OCT4 levels declined during preadipocyte to adipocyte differentiation, whereas PPARγ increased.	[96]
Reduced fatty acid oxidation (FAO) in Ob-MSCs	[99]	Depletion of β-catenin reduced FAO in osteoblasts and hepatocytes	[100,101]
(B) Cesarean delivery
Finding	References	Finding	References
Lactate deficiency in the infant gut due to cesarean delivery with reduced transmission of lactate-producing Bifidobacteria	[117]	Lactate via GPR81 stimulates Wnt3 expression of Paneth cells that activate Wnt signaling of intestinal stem cells	[132]
Acetate deficiency in the infant gut due to cesarean delivery with reduced transmission of acetate-producing Bifidobacteria and Bacteroides	[117]	Acetate via FFAR2 stimulates Wnt3 expression of Paneth cells that activate Wnt signaling of intestinal stem cells	[133]
Acetate, lactate and indoles and indol-derivatives via FFAR2, GPR81 and AhR, respectively, stimulate GLP1-1 secretion of EECs	[145,146,147,152]	GLP-1 activates MSCs; GLP1R activates PKA, which phosphorylates β-catenin thereby enhancing β-catenin signaling	[69,79]
(C) Formula feeding
Finding	References	Finding	References
Increased IGF-1-TORC1-S6K1 signaling due to excessive formula protein intake	[171,174]	Nuclear S6K1-mediated Wnt gene silencing	[175]
IGF-1 promotes differentiation of human MSCs into adipocytes	[176]	Attenuation of Wnt/β-catenin signaling in adipose tissue MSCs in response to elevated IGF-1 levels due to PPARγ-mediated expression of Axin2, an inhibitor of Wnt/β-catenin signaling	[177]
Absence of milk exosomes in infant formula	[184,185,186]	Human and bovine milk-derived exosomes enhance Wnt/β-catenin signaling of rat and murine intestinal stem cells increasing LGR5 expression	[190,191]
Absent transfer of milk exosomal microRNAs including exosomal microRNA-148a	[184,185,186]	Absent milk microRNA-148a-mediated inhibition of the Wnt inhibitor DKK1	[194]
Fecal deficiencies of acetate, lactate and indole-lactate in formula-fed compared to breastfed infants	[158,159,160]	Yet uncharacterized effects on ISC Wnt signaling and enteroendocrine L cell GLP-1 secretion
Absence of MSCs in infant formula	[205,206,207]	human breastmilk stem cell secretome: Wnt11 secretion under hypoxic, Wnt10b under non-hypoxic conditions	[208]
Wnt5 has been detected in the human breast milk proteome, whereas standard formulas are not known to contain Wnt5a,	[202]	Wnt5a mRNA decreased in MSCs with high adipogenic potential by adipogenic induction. Wnt5a promotes osteogenic but inhibits adipogenic differentiation of human MSCs	[203]
Exposure of fetal small intestine-derived organoids to human breast milk or infant formula	[204]	Organoids supplemented with formula exhibited reduced proliferation and showed decreased expression of β-catenin, TCF3 and TCF7	[204]

7. Combined Effects of Adipogenic Perinatal Factors

Maternal overweight and metabolic disturbances, together with impaired microbiome signaling associated with cesarean delivery and formula feeding rather than vaginal delivery and breastfeeding, may act synergistically to dysregulate Wnt signaling in MSCs during the perinatal period.

Which of these perinatal factors predominates in adipogenic dysregulation remains unclear and warrants further research. Two combinations represent contrasting scenarios: normal maternal weight, vaginal delivery, and breastfeeding versus maternal overweight or obesity, cesarean delivery, and formula feeding. Overall, there are eight possible combinations, whose effects on the endpoint of interest—childhood obesity—remain to be clarified (Table 3).

Studies examining only a single obesogenic factor, without accounting for all of the aforementioned perinatal influences, may therefore incompletely characterize the pathogenic molecular mechanisms underlying childhood obesity. Of particular importance, breastfeeding can partially compensate for the microbiome deficiencies associated with cesarean delivery [209,210,211,212]. Breastfeeding should therefore be promoted, particularly after cesarean delivery, as a potentially important measure for obesity prevention.

This presentation focuses on canonical Wnt/β-catenin signaling as a unifying node, but we acknowledge that redox-dependent FOXO1 signaling [15], epigenetic DNA methylation changes [99], and mitochondrial/AMPK alterations are reported in the same primary studies and may act in parallel to, rather than solely upstream or downstream of Wnt. A comparative pathway analysis across these frameworks remains to be performed.

8. Conclusions

Maternal obesity, cesarean delivery, and infant formula feeding are perinatal factors associated with an increased risk of childhood obesity and may influence the developmental trajectory during the first 1,000 days after conception. All three epidemiologically confirmed obesity risk constellations are associated with molecular alterations in MSCs that include impaired canonical Wnt/β-catenin signaling. We propose Wnt as a candidate unifying node, while recognizing that parallel pathways (redox, epigenetic, inflammatory) are also implicated and may co-operate in the perinatal programming of adipogenesis. Fetal MSCs derived from the umbilical cord tissue or umbilical cord blood of infants born to mothers with obesity exhibit increased adipogenic potential, which has been experimentally linked to impaired canonical Wnt/β-catenin signaling [10,86]. However, umbilical cord-derived MSCs may not fully reflect the adipogenic potential of adipose tissue-derived MSCs [87,213].

Cesarean delivery-associated disturbances in microbiota-derived signals in the postnatal gut, including lactate, acetate, and indole derivatives, can influence Wnt signaling in ISCs [132,133] and GLP-1 secretion by EECs [145,146]. Reduced EEC-derived GLP-1 signaling may, in turn, impair Wnt/β-catenin signaling in MSCs [69,79]. Similarly, compared with breastfeeding, formula feeding may provide a less favorable environment for microbiota-regulated signaling during early postnatal development. Supporting this concept, Smith et al. [204] demonstrated that exposure to infant formula, compared with human milk, was associated with reduced intestinal organoid growth, impaired β-catenin expression, and compromised EEC development.

At present, the strongest direct human evidence for altered MSC adipogenic differentiation during the perinatal period concerns maternal obesity. Cesarean delivery and formula feeding represent plausible additional contributing factors. However, their independent contributions and relative effect sizes remain uncertain. Maternal overweight and obesity are associated with an increased likelihood of cesarean delivery, which may be followed by a lower likelihood of breastfeeding. The co-occurrence of these three interconnected factors may therefore constitute a particularly relevant combination of exposures that could contribute to continued perinatal adipogenic programming, i.e., the perinatal march toward obesity. These interconnected risk factors should be addressed early through appropriate preventive strategies at the individual, clinical, and public health levels.