Adipose-derived stem cells (ADSCs) are becoming increasingly prominent in regenerative medicine due to their versatility and therapeutic potential (Ong et al., 2021; Feisst et al., 2015). ADSCs can differentiate into various cell types upon in vitro induction, including adipocytes, chondrocytes, and osteoblasts, and function as immunomodulators, aiding in migrating immune cells to damaged tissues (Andrzejewska et al., 2019; Ceccarelli et al., 2020). Their high proliferation rate, excellent differentiation capabilities, and ease of extraction with low donor site morbidity make ADSCs attractive for numerous treatments, especially for muscle and bone disorders.

ADSCs can be obtained from both white and brown adipose tissues, with subcutaneous fat being a particularly rich source accessible through minimally invasive procedures like power-assisted liposuction (Palumbo et al., 2018; Bajek et al., 2017). Methods to harvest and culture ADSCs typically involve enzymatic digestion and specific culture conditions to maximize yield. Lipoaspirate obtained via liposuction is a preferred method due to the high number of viable ADSCs it provides. Despite the lack of a standardized approach, these techniques ensure a robust supply of ADSCs for therapeutic applications. Compared to bone marrow-derived mesenchymal stem cells (BM-MSCs), which are limited in number and harder to retrieve, ADSCs offer a more abundant and easily accessible source with lower harvesting risks, positioning them as a viable alternative for mesenchymal stem cell (MSC) therapies (Romero et al., 2017; Francis et al., 2018). The differentiation potential of ADSCs into tissues is well-documented, but the molecular mechanisms underlying their osteogenesis are not fully understood (Jankowski et al., 2020).

Understanding the pathways and mechanisms involved in ADSC differentiation is crucial for optimizing their therapeutic applications. Investigating gene expression and related molecular processes offers deeper insights into how ADSCs differentiate and how these processes can be fine-tuned for regenerative medicine. To fully exploit ADSCs in clinical applications, it is imperative to understand the influence of ex vivo conditions and induced differentiations. Ex vivo manipulation of ADSCs, involving specific environmental cues and growth factors, plays a pivotal role in directing their differentiation pathways (Durandt et al., 2019; Mende et al., 2021). However, the complexity of these processes often introduces variability in the differentiation outcomes (Bunnell, 2021). Maximizing the therapeutic potential of ADSCs requires meticulous control and standardization of these ex vivo conditions, such as culture media composition, oxygen levels, mechanical forces, and the timing of gene expression changes during differentiation (Neri et al., 2013).

Canine ADSCs represent a promising model in stem cell research due to their biological similarities to human ADSCs, including comparable cellular characteristics (e.g., the expression of surface markers such as CD90, CD105 and CD73) and ability to differentiate towards osteogenic, chondrogenic and adipogenic lineages (de Bakker et al., 2013) In addition, these cells are widely accessible and canine models are increasingly recognized in translational research as they bridge the gap between rodent studies and human applications (Merlo and Iacono, 2023). Since dogs are companion animals, sharing the environment and living conditions with their owners, they tend to develop similar diseases to those occurring in humans, such as osteoarthritis or bone fractures (Hoffman and Dow, 2016). Therefore, canine models seem to be relevant in terms of possible therapeutic approaches not only in veterinary medicine, but also in humans.

A comprehensive understanding of the molecular processes governing ADSC osteogenesis, including gene expression and regulatory pathways, and their responses to varying ex vivo conditions and differentiation procedures, is critical for translating ADSC-based therapies from the laboratory to clinical practice. Ensuring the reproducibility and predictability of these conditions is vital to unlocking the full therapeutic potential of ADSCs in treating bone diseases and injuries. Thus, this study focused on functional gene expression analysis in canine ADSCs subjected to long-term culture and osteogenic differentiation. Bioinformatic analysis provided a broad view of the impact of ex vivo conditions and induced differentiation on ADSCs, offering valuable molecular insights. The knowledge gained serves as a reference for understanding ADSC-related processes and their application in bone regeneration therapies.

The PCA of the samples analyzed in the study demonstrated that the biological repetitions of the three studied time points form distinct clusters. This initially confirms the relatively similar gene expression among the studied sample groups (EC-early control, 1 day of culture; OC-osteoblast control, 14 days of culture without differentiating medium; OT-osteoblast test, 14 days of culture with differentiating medium), and major differences in gene expression between the groups. The results of the PCA are presented in Figure 1.

A total of 17793 genes were detected in the RNAseq analysis. Differential expression analysis, using a threshold of Log₂FC ≥ 1 and FDR<0.05, yielded 1736 DEGs between OC and EC, 1471 DEGs between OT and EC, and 722 genes between OT and OC. The results of the differential expression analysis were presented as volcano plots, compiled in Figure 2. Furthermore, the complete list of DEGs, including their LogFCs and p-values, were included in the manuscript as Supplementary Table 1.

GO enrichment analysis was performed on the differentially expressed genes to identify significantly over- and under-represented Biological Process (BP), Cellular Component (CC) and Molecular Function (MF) GO terms across each comparison. The BP category was chosen as the focus of this study, as it provides a comprehensive view of the functional dynamics underlying osteogenic differentiation. Nonetheless, the complete results of GO enrichment analysis were included in the manuscript as Supplementary Table 2. The analysis revealed 26 significantly over-represented and 11 under-represented GO BP terms in the early control vs osteoblast control comparison, 16 over- and 23 under-represented terms in the early control vs differentiated osteoblast comparison, and 9 over- and 43 under-represented terms in the osteoblast control vs differentiated osteoblast comparison. The results of the GO BP enrichment analysis are visualized as bar graphs in Figures 3–5.

Hierarchical clustering of GO terms revealed distinct groups of significantly over- and under-represented biological processes in each comparison. In the early control vs osteoblast control comparison, processes such as enzyme-linked cell assembly, collagen catabolic process, extracellular matrix organization, connective cartilage development, and hindlimb morphogenesis were notably positively enriched (Figure 6). Conversely, processes related to carboxylic acid and fatty acid metabolism, small molecule biosynthesis, blood circulation, cell adhesion, and angiogenesis were significantly negatively enriched (Figure 7). In early control vs differentiated osteoblasts comparison, positively enriched groups included apoptotic programmed process death, extracellular encapsulating structure organization, circulatory nephron system development, peptidyl-tyrosine modification phosphorylation, and negative population communication proliferation (Figure 8). In the same comparison, negatively enriched groups comprised prostate animal gland organ, angiogenesis tube vasculature vessel, tissue anatomical formation involved, adhesion cell cycle division, and vasoconstriction circulatory circulation process (Figure 9). In the osteoblast control vs differentiated osteoblast comparison, over-represented processes involved apoptotic programmed cell death, regulation of apoptotic programmed cell death, circulatory system, positive multicellular organismal process, regulation growth (Figure 10). Meanwhile, under-represented processes included positive developmental anatomical morphogenesis, neuron glial gliogenesis projection, locomotion behavior catabolic metabolic, cell leukocyte motility migration, and extracellular external encapsulation organization (Figure 11).

Finally, the KEGG pathway enrichment analysis revealed significantly enriched pathways associated with differentially expressed genes. In OC versus EC (Figure 12) comparison pathways such as “axon guidance” and “AMPK signaling pathway” were positively enriched, while “p53 signaling pathway”, “hypertrophic cardiomyopathy” and “regulation of actin cytoskeleton” were negatively enriched. Pathways such as “PI3K-Akt signaling pathway”, “focal adhesion” and ‘ECM-receptor interaction’ were both positively and negatively enriched.

In the OT versus EC comparison (Figure 13), a positive enrichment of pathways such as “Ras signaling pathway”, “PI3K-Akt signaling pathway”, “AMPK signaling pathway”, and “focal adhesion” were observed, while pathways like “axon guidance”, “hypertrophic cardiomyopathy”, “ECM-receptor interaction” and “tight junction” were negatively enriched. Notably, “Rap1 signaling pathway” was both positively and negatively enriched.

In the OT versus OC comparison (Figure 14) pathways such as “Rap1 signaling pathway”, “MAPK signaling pathway”, “PI3K-Akt signaling pathway”, “p53 signaling pathway”, “hypertrophic cardiomyopathy” and “focal adhesion” were positively enriched, while “calcium signaling pathway”, “axon guidance”, “cAMP signaling pathway” and “lysosome” were negatively enriched. “ECM-receptor interaction” was both positively and negatively enriched.

In the present study, the RNA-seq analysis was performed to shed light on biological processes involved in osteogenic differentiation of canine ADSCs. It is important to note that in vitro osteogenic differentiation typically requires at least 14 days, meaning the differentiated cells are exposed to prolonged culture conditions, which may impact their metabolism. To determine transcriptomic changes in differentiated ADSCs occurring only due to the differentiation process, a comparative was performed between ADSCs cultured for 14 days without differentiation medium (osteoblast control, OC) and ADSCs cultured for only 1 day (early control, EC). RNA-seq, a high-throughput technique that enables comprehensive transcriptome analysis, was used due to its reproducibility and effectiveness in identifying transcriptomic markers of osteogenic differentiation (Wang et al., 2009). Gaining further insight into the molecular changes in cultured and differentiated ADSCs is essential before considering their clinical applications. To provide a deeper understanding of the functional changes during osteogenic differentiation, we performed a pathway enrichment analysis, which allowed us to identify key molecular pathways driving the observed biological processes.

‘PI3K-Akt signaling pathway’ was both positively and negatively enriched in OC versus EC comparison, while its over-representation was observed in OT versus EC and OT versus OC comparisons. This pathway was already demonstrated to play crucial role in osteogenic differentiation of MSCs. In case of umbilical cord-derived MSCs (UC-MSCs) this pathway was positively enriched after the 3- and 5-day osteogenic differentiation period, while its inhibition led to the blockage of the differentiation (Gao et al., 2023). Additionally, Runx2, a key transcription factor for osteogenic differentiation was shown to work in conjunction with this pathway, promoting both osteoblast and chondrocyte differentiation and migration (Fujita et al., 2004). The current results are consistent with previous studies and suggest that ‘PI3K-Akt signaling pathway’ promotes osteoblast lineage commitment and maturation in the presence of osteogenic supplements after 14-day differentiation period in ADSCs.

The comparison of osteogenically differentiated ADSCs and both EC and OC revealed the upregulation of genes associated with apoptosis and its regulation. Notably, in the OC/OT comparison, five out of nine over-represented GO groups were related to cell death. It can be hypothesized that this was not due to the prolonged culture, as the osteoblast control, which was cultured for the same duration, did not show an increase in apoptosis-related gene expression compared to EC. Apoptosis plays a crucial role in maintaining bone homeostasis by balancing tissue renewal and degradation, which is essential for continuous bone remodelling. In addition, stem cells rely on robust quality control mechanisms to preserve their molecular integrity over extended durations. One critical aspect of this quality control involves initiating cell death damaged cells (Reya et al., 2001). Key regulators of apoptosis include members of the BCL2 protein family, which modulate cell survival depending on the BCL2/BAX ratio (Youle and Strasser, 2008). The roles of BCL2 and BCL2L1 (also known as BCL-XL) have been previously documented in the osteogenic, adipogenic, and neurogenic differentiation of human BM-MSCs. BCL2L1 was found to be expressed at consistent levels in both undifferentiated and differentiated BM-MSCs, where it exhibited anti-apoptotic activity. In contrast, BCL2 expression was dependent on the stage of differentiation and increased only in differentiated cells (Oliver et al., 2011). Similarly, in adipo- and osteo-induced dental pulp MSCs, BCL2 was also overexpressed (Lee et al., 2019). In this study, BCL2L1 was upregulated in osteo-induced canine ADSCs, suggesting the inhibition of apoptosis in these cells during osteogenic differentiation.

There are conflicting results regarding the expression of apoptosis-related genes in MSCs after differentiation. Similarly to our findings, Robert et al. (2018) observed an over-representation of apoptosis-regulation GO groups during the early stages of osteogenic differentiation in human ADSCs. Likewise, increases in apoptosis-related gene expression have been reported following osteogenic, chondrogenic, adipogenic, and neurogenic differentiation of human Wharton’s jelly-derived MSCs (WJ-MSCs) (Stefańska et al., 2023). Similarly, Oliver et al. (2013) reported an increased sensitivity towards apoptosis after adipogenic and osteogenic differentiation of human BM-MSCs, which they attributed to decreased DNA damage repair. On the contrary, other studies, such as by Lo Furno et al. (2013), have reported decreased apoptosis markers after adipogenic differentiation in human ADSCs, indicating that the expression of apoptosis-related genes may vary depending on the MSC source and differentiation stage, which should be carefully considered in future studies.

Additionally, ‘p53 signaling pathway’ was negatively enriched in OC versus EC comparison and positively enriched in OT versus OC comparison, reflecting its multifaceted role in cellular regulation. p53 encodes a transcription factor involved in major cellular processes, such as cell cycle arrest, cellular senescence and apoptosis (Hernández Borrero and El-Deiry, 2021). Its role in regulation of gene expression associated with mesenchymal differentiation programs, including osteogenic differentiation, has also been demonstrated (Jain and Barton, 2018). The negative enrichment of this pathway in OC versus EC suggests reduced stress response and apoptosis during the initial culture phase, while its over-representation in OT versus OC indicates the reactivation of this pathway during osteogenic differentiation. Although, studies utilizing murine MSCs revealed that p53 plays an inhibitory role in their osteogenic differentiation (Armesilla-Diaz et al., 2009; Tataria et al., 2006), the current results underscore the need for further investigation into its species-specific and context-depended functions.

Genes involved in “circulatory system development” were also upregulated in osteo-induced ADSCs as compared to EC and OC. However, terms related to “blood vessel morphogenesis” and “vasculature development” were under-represented, while “angiogenesis,” “tube morphogenesis,” “blood vessel development”, and “circulatory system development” were negatively enriched in EC/OT comparison. The connection between osteogenesis and the circulatory system is well-established, as both intramembranous and endochondral bone formation are closely linked to vascular ingrowth. In particular, capillary invasion during intramembranous ossification triggers MSC differentiation into osteoblasts (Kanczler and Oreffo, 2008). Moreover, the development of the circulatory system is crucial for maintaining the homeostasis of bone tissue, as blood vessels supply the nutrients required for osteogenic differentiation (Kanczler and Oreffo, 2008). Previous studies showed that ADSCs differentiate towards endothelial cells and enhance angiogenesis more than BM-MSCs (Cao et al., 2005; Brennan et al., 2017). Shaik et al. (2019) also observed the enrichment of genes associated with ECM organization and angiogenesis along with ERK1/2 and JNK pathways in human ADSCs following a 21-day osteogenic differentiation period, suggesting that angiogenic-related genes may promote vascular development and integration with bone tissue, a trend also observed in canine ADSCs. However, the under-representation of ‘blood vessel morphogenesis’ and ‘vasculature development’ groups might indicate that canine ADSCs during osteogenic differentiation tend to support the existing vessels rather than participate in the formation of new ones.

The pathway enrichment analysis provides a further insight into the dynamic changes of expression of genes associated with circulatory system during osteogenic differentiation. “Hypertrophic cardiomyopathy”, “dilated cardiomyopathy” and “arrhythmogenic right ventricular cardiomyopathy” were positively enriched in OT versus OC comparison, while negatively enriched in OT/EC and OC/EC comparisons, suggesting a complex role of these pathways in the progression form multipotent stated to osteogenic differentiation. The activation of cardiomyopathy-associated pathways in OT might indicate their involvement into osteogenic-specific processes, for instance in ECM remodeling or calcium singaling.

In osteoblast controls, genes associated with circulatory system development, as well as with blood vessel development and morphogenesis, angiogenesis, blood circulation, and vasculature development, were downregulated. This likely reflects the limitations of two-dimensional in vitro culture, which does not fully replicate the complexity of in vivo conditions, reducing the need for advanced developmental processes (Gibler et al., 2021). It is well established that telomere shortening occurs with cellular division, and prolonged culture of ADSCs has been linked to a decline in differentiation potential and replicative aging, ultimately leading to senescence (Bieback et al., 2012; Bonab et al., 2006). In the absence of specific differentiation cues in standard culture media, ADSCs may retain their MSC-like state rather than progressing toward non-native lineages. The current findings suggest that ADSCs maintained in standard culture conditions may be more inclined toward chondrogenic differentiation, as evidenced by the upregulation of genes associated with cartilage development and chondrocyte differentiation in the osteoblast control group. This propensity could be influenced by the source of the cells, their cellular memory, and the methylation status of key transcription factors that govern cell fate decisions (Xu et al., 2017; Jääger et al., 2012).

Although ADSCs have the potential to differentiate into neurogenic lineages, this differentiation is non-canonical and typically inefficient (Locke et al., 2011). However, both neuronal and glial markers, such as NSE, NeuN, or vimentin, were identified even in undifferentiated ADSCs and other MSCs (Safford and Rice, 2005). Human MSCs after osteogenic and adipogenic differentiation also expressed neuronal markers, namely, NeuN and βIII-tubulin (Foudah et al., 2013). In contrast, the present study revealed a decrease in the expression of genes related to neurogenesis and gliogenesis following osteogenic differentiation of canine ADSCs. Similarly, Shai et al. (2019) reported the downregulation of nervous system development-related genes and others associated with neurogenesis. A recent study suggests that osteogenic and neurogenic differentiation may be inversely regulated in ADSCs via the MAPK-mediated signaling pathway, with epiregulin promoting osteogenic differentiation while inhibiting neurogenic potential (Cao et al., 2020).

Consistently with these findings, the pathway enrichment analysis revealed over-representation of “axon guidance” pathway in OC versus EC comparison and under-representation in both OT versus EC and OT versus OC comparisons. Axon guidance factors comprise versatile proteins involved in regulation of both physiological and pathological processes, which are particularly active during nerve growth and regeneration. In addition, these proteins were demonstrated to be involved in osteogenic differentiation and bone formation, either by inhibitor of promotion of this processes (Liu et al., 2022). On the contrary to current results, “axon guidance” pathway was shown to be positively enriched in osteogenically differentiated BM-MSCs (Wan et al., 2021), indicating the need to perform further research regarding its role in MSCs osteogenic differentiation.

In osteo-induced ADSCs, a significant downregulation of genes associated with various morphogenic processes, such as “gland morphogenesis,” “animal organ morphogenesis,” “salivary gland morphogenesis,” “prostate gland morphogenesis,” and “tissue morphogenesis,” was observed compared to the early control group (EC). This suggests that ADSCs without differentiation cues remain in a more multipotent state, maintaining higher expression of genes involved in general morphogenesis (Wagner et al., 2005). In contrast, osteo-induced ADSCs become committed to the osteogenic lineage, leading to the upregulation of osteogenic differentiation-related genes and the corresponding downregulation of genes associated with other morphogenic pathways.

In summary, the current results provide valuable insights into the transcriptomic changes associated with key biological processes in canine ADSCs following osteogenic differentiation. The upregulation of genes involved in apoptosis regulation suggests a balance between cell survival and death, ensuring the quality of differentiating osteoblasts and the removal of damaged cells, which is critical for proper tissue development. Additionally, the upregulation of genes related to circulatory system development reinforces the role of differentiating ADSCs in supporting blood vessel formation and remodeling—an essential aspect of bone tissue growth and repair, with potential applications in bone tissue engineering. Furthermore, the observed downregulation of genes involved in neurogenesis and gliogenesis suggests a reciprocal regulation of these processes during osteogenic differentiation.

It is important to note that 2D in vitro cultures do not fully replicate the complexity of in vivo conditions, where cytokines, growth factors, and the cellular microenvironment play critical roles in determining the fate of ADSCs and guiding their differentiation. While transcriptomic analysis provides valuable insights into gene expression, it has limitations in capturing post-transcriptional modifications and protein activity. Therefore, complementary proteomic studies are required to fully understand the functional protein dynamics and their regulatory roles in osteogenic differentiation of ADSCs. Moreover, further functional and in vivo studies, as well as a deeper understanding of the signaling networks and transcription factors involved in ADSC osteogenic differentiation, are necessary to develop safe and effective stem cell therapies. However, in vitro models are invaluable for identifying key biological processes in osteo-induced ADSCs and highlighting potential targets for enhancing and optimizing differentiation. Extending this research to other MSC types will undoubtedly advance stem cell therapies and tissue engineering efforts.