Microarray analysis revealed distinct expression patterns of many genes during adipogenesis in human SGBS cells and mouse 3T3-L1 cells. This may be attributed to species differences between humans and mice; however, there may also be other differences in their properties as preadipocytes. For instance, the profile of genes expressed by differentiated adipocytes derived from 3T3-L1 cells is markedly different from that expressed by mature adipocytes in mouse15. By contrast, the mRNA expression profile of adipocytes derived from SGBS cells is similar to that of primary human white subcutaneous adipocytes16–18. Nevertheless, numerous molecules implicated in adipogenesis have been identified and validated using 3T3-L1 cells. Thus, identifying genes induced during the adipogenic process in SGBS cells, but not in 3T3-L1 cells in this study, could serve as a viable strategy for identifying adipogenic factors as yet undiscovered in humans. In fact, among the 61 genes exhibiting more than 20-fold induction during the SGBS adipogenic process, but not in 3T3-L1 cells, some genes, except for GALNT15, whose involvement in adipogenesis is unknown to our knowledge, were included. Notably, with LMO3, a human-specific adipogenic gene8, being one of these 61 genes, there is a possibility that unidentified adipogenic factors may also be present within this group.
In this study, GALNT15 is identified as a gene induced during the adipogenesis of SGBS cells, consistent with reports indicating that GALNT15 is listed as one of upregulated during adipocyte differentiation from human adipose-derived stem cells19. Although the mRNA expression of Galnt15 was not induced, and protein expression was not detected during adipogenesis of 3T3-L1 cells, it cannot be ruled out that the induction was undetectable due to extremely low expression levels compared with those in SGBS cells. The regulation of Galnt15 gene expression is influenced by corticosterone and the stress response in the mouse hippocampus20, and the medium used to induce adipogenesis in 3T3-L1 cells also contains dexamethasone (DEX), a synthetic glucocorticoids. This study focused on the identification of novel human adipogenesis-related factors, therefore, the investigation in mice was limited to confirming Galnt15 expression in 3T3-L1 cells. However, it is necessary to carefully consider whether an transient increase in the expression of Galnt15 is observed during mouse adipogenesis or whether Galnt15 is also involved in mouse adipogenesis.
Although inhibition of GALNT15 induction clearly impeded adipogenesis in SGBS cells, overexpression of GALNT15 did not affect the accumulation of triglycerides or the proportion of cells containing lipid droplets, despite affecting the induction of PPARG mRNA 4 days after adipogenic induction in SGBS cells. This suggests that the induced expression levels of PPARG during the adipogenic process in control SGBS cells may be sufficient to affect adipogenesis under our experimental conditions, with further overexpression potentially having no additional effect on these aspects. Furthermore, overexpression of GALNT15 enhanced only the mRNA expression of CEBPA and its direct target gene LEP21, among the tested adipogenic marker genes. The role of CEBPA in adipogenesis is limited to the induction and maintenance of PPARG expression and the establishment of insulin sensitivity22. This suggests that the lack of adipogenic promotion in SGBS cells overexpressing GALNT15 could be attribute to sufficient levels of PPARG expression. Conversely, GALNT15 knockdown may affect adipogenesis by suppressing CEBPA expression, resulting in insufficient PPARG expression levels for adipogenic differentiation.
Abnormal O-GalNAc-glycosylation catalyzed by the GALNT family is associated with various human diseases, with particular attention focused on the link between GALNT2 and metabolic disorders, such as obesity, type 2 diabetes, and lipid abnormalities23. In vitro analysis has shown that a reduction in GALNT2 expression in HepG2 cells, a human hepatocarcinoma cell line, impairs insulin signaling and action24. Conversely, GALNT2 overexpression stimulates adipocyte maturation and enlargement in 3T3-L1 cells25. However, our microarray data revealed that, except for GALNT15, other GALNT family members did not exhibit a remarkable increase during the adipogenic process in SGBS cells (Supplementary table 2). This may suggest a more profound role for GALNT15 than for GALNT2 in human adipogenesis. GALNT15 does not show a significant relation with other GALNT family member11, and to our knowledge, its physiological function has not been thoroughly investigated. Our findings, combined with the fact that GALNT15 also serves as a marker gene candidate during osteocyte differentiation from canine adipose derived stem cells26, suggest that GALNT15 may play an important role in the differentiation of mesenchymal stem cells.
In conclusion, we have demonstrated that GALNT15 contributes to adipogenesis in SGBS cells by upregulating CEBPA and PPARG. However, the specific molecular mechanisms driving GALNT15 -induced adipogenesis, including the potential involvement of unidentified substrates or non-enzymatic functions of GalNAc-T15, remain unclear and require further investigation for comprehensive elucidation. Our findings suggest that GLANT15 is an attractive drug target for the treatment of obesity.