Upregulated Mcp-1 mRNA expression in bone marrow cells from db/db mice
To investigate the alterations in gene expression profiles of bone marrow cells in type 2 diabetes and their pathological implications in the disease progression, transcriptome sequencing was performed on whole bone marrow cells obtained from spontaneously diabetic db/db mice.
The transcriptome sequencing results showed that, comparison to the Ctrl mice, there were 2915 DEGs identified in the bone marrow cells of db/db mice, of which 1025 were up-regulated and 1890 were down-regulated (Figure S1A). Subsequently, we conducted KEGG analysis on these DEGs, which revealed that the enriched pathways included PI3K-Akt signaling pathway, AGE-RAGE signaling pathway in diabetic complications, chemokine signaling pathway and Type Ⅱ diabetes mellitus, etc. (Figure S1B).
After analyzing the KEGG pathways mentioned above, we discovered that Mcp-1 was involved in two diabetes-related pathways, namely AGE-RAGE signaling pathway in diabetic complications and chemokine signaling pathway (Figure S1C). We hypothesize that persistent hyperglycemia results in the non-enzymatic glycosylation of various proteins in the body, leading to the accumulation of advanced glycation end products (AGEs). These AGEs then bind to their receptors (RAGE), which down-regulates PI3K/Akt levels and ultimately upregulates Mcp-1 mRNA expression mediated by NF-κB. These processes may play a crucial role in the development of chronic diabetes complications, including thrombosis, inflammation, and atherosclerosis (Figure S1D).
Moreover, the up-regulated chemokine in the chemokine signaling pathway included Mcp-1 (Ccl2), which combines with the corresponding receptor Ccr2 to downregulate PI3K/Akt through intracellular pathways. This subsequently inhibits the phosphorylation of FOXO and regulates cell proliferation, differentiation and apoptosis (Figure S1E).
Accumulated adipocytes in bone marrow cavity and markedly up-regulated expression of Mcp-1 of diabetic mice
To better explain the cause of abnormal gene expression in the bone marrow cells of db/db mice and its impact on glucose metabolism, we have successfully created a T2D mouse model by combining HFD feeding with MLD-STZ induction(10, 11). Moreover, we carefully selected spontaneously diabetic db/db mice to ensure that the experimental results were not affected by STZ-induced injuries to rodent islet cells (Figure S2).
Conventionally, BMAT is believed to merely act as a passive filler in the bone marrow cavity. However, due to its special anatomical location, BMAT has some distinct functions. In our study, we observed a significant increase of 1530.0% in the number of BMAs in T2D mice compared to the Ctrl group, along with a significant 158.2% increase in BMA diameter (Fig. 1A-C). Likewise, the number and diameter of BMAs in spontaneously diabetic db/db mice were markedly increased by 1482.9% and 203.4%, respectively, compared to the Ctrl group (Fig. 1D-F), indicating that the disorder of glucose metabolism in type 2 diabetes leads to an increase in the number and diameter of BMAs.
To clarify the pathological mechanism behind the accumulation of BMAT in type 2 diabetes, we examined the mRNA levels of osteogenesis- and adipogenesis-related genes in the metaphyseal bone tissue. The results showed that compared to the Ctrl group, the osteogenesis-related genes, Runx2, Col1a1, Ocn and Lcn2, in the metaphysis of the T2D group were significantly down-regulated, while the expression levels of adipogenesis-related genes, Pparγ and Adiponectin, were distinctly up-regulated (Fig. 1G). Similar results were obtained when we examined the same mRNA expression in the metaphyseal bone of db/db mice (Fig. 1H), suggesting that the disorder of glucose metabolism inhibits osteogenesis and promotes adipogenesis in BMSCs.
To further illustrate the pathological mechanisms underlying metabolic disorders in diabetic mice (as shown in Figure S2), we utilized Brdu/insulin and TUNEL/insulin double staining to represent the proliferation and apoptosis of islet cells, respectively. Our results demonstrated that the T2D group showed a 38.5% reduction in islet cell proliferation compared to the Ctrl group (Fig. 2A), while apoptosis increased by 17% (Fig. 2B). Consistent with these, beta-cells from db/db group exhibited a 39.5% reduction in proliferation (Fig. 2C) and a 25.1% increase in apoptosis (Fig. 2D). Furthermore, we measured the mRNA expression levels of key transcription factors, Pdx-1 and Nkx6.1, which determine the proliferation function of pancreatic beta-cells. Our findings revealed that the expression levels of Pdx-1 and Nkx6.1 in the T2D group were markedly down-regulated compared to the Ctrl group (Fig. 2E). Similarly, the mRNA expression levels of Pdx-1 and Nkx6.1 in db/db group were significantly down-regulated compared to the Ctrl mice (Fig. 2F).
We proceeded to isolate mouse islet beta-cells and collect supernatants from basal (2.8mM) and high glucose (25mM) KRBH buffer-stimulated islet cells to further explore the underlying reasons for the glucose disorder in type 2 diabetic mice. Compared to the Ctrl mice, the basal insulin secretion of the T2D mice decreased by 30.2%. Moreover, compared to their respective basal insulin secretion levels, the GSIS of the Ctrl mice increased by 126.1% (p < 0.01), while the GSIS of the T2D mice increased by 21.7% (p < 0.05), which means the GSIS of the T2D mice dramatically decreased by 62.5% compared to the Ctrl mice (Fig. 2G). Similarly, basal insulin secretion in db/db group mice decreased by 49.4%, and the GSIS decreased by 79.3% (p < 0.01) compared to the Ctrl group (Fig. 2H), indicating that the insulin secretion function of mouse islet cells stimulated by high glucose dramatically decreases in type 2 diabetes.
To clarify the source cells of the DEGs in bone marrow, we isolated BMA. Previous studies have established that BMA expresses adipose-specific genes, such as Adiponectin and Fabp4 (12). To investigate the similarities and differences between BMA and white adipose tissue (WAT) in type 2 diabetes, we examined the mRNA expression of epididymal white adipose tissue (eWAT) and BMA. Our results indicated that the expression of adipocyte-specific genes, Adiponectin and Pparγ, in BMA was distinctly lower than those in eWAT in Ctrl mice (Fig. 3A). What’s more, compared to the Ctrl group, the expression levels of Adiponectin and Pparγ in the peripheral eWAT of the T2D group were signally downregulated by 97.8% and 50.5%, respectively (Fig. 3B). The mRNA level of Adiponectin in BMA of T2D mice was dramatically decreased by 54.4%, while the expression of Pparγ was markedly increased by 138.0% (Fig. 3C). These results indicate that the adipogenesis ability of BMA was substantially lower than that of peripheral eWAT in Ctrl mice but was memorably upregulated in BMA compared to eWAT in type 2 diabetes, resulting in the dramatic accumulation of BMAT in type 2 diabetic mice.
Then, we compared the mRNA expression levels of inflammatory factors in eWAT and BMA. In the Ctrl group, the levels of Tnfα and Il1β were distinctively higher in BMA than in peripheral eWAT, while the mRNA level of Mcp-1 was much lower in BMA than in eWAT (Fig. 3D). In T2D mice, the mRNA expression levels of Tnfα and Il1β in eWAT were overtly up-regulated, whereas the expression of Mcp-1 was obviously down-regulated compared to the Ctrl group (Fig. 3E). However, compared to the Ctrl group, the mRNA expression levels of Tnfα and Il1β in the BMA of the T2D group were down-regulated, while the expression of Mcp-1 was significantly up-regulated (Fig. 3F). Similarly, we found that the expression level of Mcp-1 in the BMA of the db/db mice distinctly increased compared to the Ctrl group (Fig. 3G). These results suggest that T2D mice showed a local inflammatory phenotype in peripheral eWAT, and elevated inflammatory factors inhibit differentiation of peripheral eWAT, leading to the development of insulin resistance in adipose tissue. However, the BMA of T2D mice showed no obvious inflammatory phenotype compared to peripheral eWAT, but Mcp-1 is selectively upregulated in BMA.
In both diabetic patients and mouse models, serum Mcp-1 levels are significantly elevated(6). In the present study, we found that the mRNA expression level of Mcp-1 in BMA was dramatically up-regulated in type 2 diabetes (Fig. 3 and S1). To further investigate this, we detected the level of Mcp-1 in mouse serum and BMS. As depicted in Fig. 3H, compared to Ctrl group, the Mcp-1 levels in serum and BMS of the T2D group increased distinctly. Similarly, Mcp-1 levels in serum and BMS of db/db mice were also increased compared to the Ctrl group (Fig. 3I).
In order to investigate the potential relationship between Mcp-1 secretion by BMA and the changes in islet cells, we first examined the mRNA expression level of the Ccr2 receptor of Mcp-1 in mouse islet cells. Our findings revealed a dramatical upregulation of Ccr2 expression in the islet cells of both T2D and db/db groups compared to the Ctrl group (Fig. 3J-K), suggesting the elevated Mcp-1 may affect the islet cells in type 2 diabetic mice.
Inhibition of proliferation and insulin secretion function of MIN6 cells by Mcp-1
Previous studies have revealed that BMA may affect glucose metabolism both directly and indirectly. To explore the direct relationship between BMA and pancreatic islet cells, we utilized in vitro cell models. Initially, we assessed the level of Mcp-1 in different CM. As displayed in Fig. 4A, this analysis showed that Mcp-1 concentrations were lower in OP9-CM and high glucose (HG)-OP9-CM. Conversely, the level of Mcp-1 in OP9-derived adipocyte (OP9A)-CM was significantly higher than in OP9-CM, and the level of Mcp-1 in HG-OP9A-CM was further increased compared to OP9A-CM. Similarly, the Mcp-1 levels in primary bone marrow mesenchymal stem cells (BMSCs)-CM and HG-BMSCs-CM were lower. Whereas the level of Mcp-1 in BMA-CM was dramatically higher than in BMSCs-CM, and Mcp-1 level in HG-BMA-CM was further increased compared to BMA-CM (Fig. 4B). These results further confirmed the high expression of Mcp-1 in BMA, with its expression further amplified under high glucose stimulation. Subsequently, we utilized HG-OP9A-CM/HG-BMA-CM to investigate their effects on the proliferation, apoptosis and GSIS of mice insulinoma cell line-MIN6 cells.
After being co-cultured for 24 and 48 hours, it was observed that when compared to MIN6 cells cultured with OP9-CM, OP9A-CM caused a distinct inhibition in the proliferation of MIN6 cells by 17.0% and 18.6%. Moreover, HG-OP9A-CM further restrained the proliferation of MIN6 cells by 19.6% and 47.9% (Fig. 4C). Similarly, BMA-CM was found to be more effective than BMSCs-CM in inhibiting the proliferation of MIN6 cells. Notably, HG-BMA-CM was observed to significantly repress the proliferation of MIN6 cells (Fig. 4D), indicating that both bone marrow adipocyte-derived CM are effective in inhibiting the proliferation of MIN6 cells, with the inhibitory effect being further aggravated under high glucose stimulation.
When compared to OP9-CM, OP9A-CM was observed to increase the apoptosis of MIN6 cells. Additionally, HG-OP9A-CM further exacerbated the apoptosis of MIN6 (Fig. 4E). Likewise, BMA-CM was found to increase the apoptosis of MIN6 cells in comparison to BMSCs-CM. HG-BMA-CM further increased apoptosis of MIN6 cell line (Fig. 4F). These results indicate that both bone marrow adipocyte-derived CM promoted the apoptosis of MIN6 cells, with high glucose stimulation further promoting this effect.
Regarding the impact of CM on GSIS, it was observed that MIN6 cells in the Ctrl group exhibited a two-fold increase in GSIS. The OP9-CM intervention group showed a 1.6-fold increase, while the HG-OP9-CM, OP9A-CM and HG-OP9A-CM intervention groups exhibited increases of 1.3, 0.4, and 0.2 times, respectively (Fig. 4G). Notably, OP9A-CM treatment resulted in a significant 46.6% decrease in GSIS compared to OP9-CM, while HG-OP9A-CM caused a marked 51.6% decrease in GSIS. Similarly, BMA-CM intervention led to a significant 54.5% decrease in GSIS compared to BMSCs-CM, and HG-BMA-CM resulted in a clear 67.2% decrease in GSIS (Fig. 4H). These results indicate that both bone marrow adipocyte-derived CM significantly impede the insulin secretion capacity of MIN6 cells, with their inhibitory effects being further potentiated under high glucose conditions.
Proliferation inhibition of MIN6 cells through Mcp-1/PI3K/Akt pathway
Through transcriptome sequencing, we found that the mRNA expression level of Mcp-1 in bone marrow cells from db/db mice was significantly up-regulated compared to the Ctrl group (Figure S1). The up-regulated Mcp-1 is believed to regulate cell proliferation via the Mcp-1/Ccr2/PI3K/Akt signaling (Figure S1E). To further investigate this, we conducted western blot to detect Akt protein levels in MIN6 cells after intervention of different CM. The results showed no significant difference in the protein levels of total Akt (t-Akt) in MIN6 cells between different CM intervention groups. However, compared to the OP9-CM group, the protein level of phosphorylated Akt (p-Akt) in MIN6 cells was significantly down-regulated after OP9A-CM intervention, which was further dramatically down-regulated in the HG-OP9A-CM intervention group (as shown in Fig. 4I and the supplemental material).
Next, we examined the protein levels of p-Akt and t-Akt in MIN6 cells after exogenous Mcp-1 intervention. The results showed that the concentration of Mcp-1 increased, the protein levels of p-Akt in MIN6 cells were down-regulated in 10, 50 and 100ng/mL Mcp-1 intervention group compared to the Ctrl group (as shown in Fig. 4J and the supplemental material). Interestingly, the intervention of 200ng/mL Mcp-1 recombinant protein only slightly reduced the protein level of p-Akt in MIN6 cells compared to the Ctrl group, but significantly increased by 83.1% compared to the 100ng/mL group, suggesting that the effect of Mcp-1 on the p-Akt level in MIN6 cells is concentration-dependent.
Our results demonstrate that both bone marrow adipocyte-derived CM and exogenous Mcp-1 can directly reduce the phosphorylation level of Akt in MIN6 cells. We further investigated this issue by simulating or blocking Mcp-1 pathway. In the co-culture system between OP9-CM/HG-OP9-CM or OP9A-CM/HG-OP9A-CM and MIN6 cells, we intervened by adding Mcp-1 recombinant protein or its receptor antagonist RS. The results showed that adding 100ng/mL Mcp-1 recombinant protein to the co-culture system between OP9-CM/HG-OP9-CM and MIN6 cells had no significant effect on cell proliferation after 24 hours, but decreased cell proliferation after 48 hours compared to the Ctrl + PBS group (Fig. 4K). In contrast, blocking Mcp-1 signaling with RS restored the proliferation ability of MIN6 cells. Western blot showed that the protein levels of p-Akt in MIN6 cells after OP9-CM/HG-OP9-CM + 100ng/mL Mcp-1 intervention were distinctly decreased compared to the Ctrl group (as shown in Fig. 4L and the supplemental material). The above results suggest that BMA inhibit the proliferation of MIN6 cells by producing Mcp-1 to down-regulate Akt signaling.
Improvement glucose metabolism disorders in diabetic mice by blocking Mcp-1/Akt
Compared to the Ctrl group, the ab libitum fed glucose levels of the T2D + DMSO mice were significantly increased at one week (S1) and second week (S2). However, the T2D + RS group showed a dramatic 32.9% and 44.7% decrease in ab libitum fed glucose levels at S1 and S2, compared to T2D + DMSO mice (Fig. 5A). Additionally, the blood glucose levels of T2D + DMSO mice were significantly higher than those of Ctrl mice at each time point of IPGTT, whereas the blood glucose levels of the T2D + RS group were signally lower than those of the T2D + DMSO group, with a reduction of 57.8% at 120 minutes (Fig. 5B). Similarly, ab libitum fed glucose were markedly elevated in spontaneously diabetic db/db + DMSO mice compared to Ctrl mice but were obviously lower in db/db + RS mice (Fig. 5C). During IPGTT, the blood glucose levels of db/db + DMSO group mice were distinctively higher than those of Ctrl mice at each time point, with a significant increase of 388.3% at 120 minutes. In contrast, the blood glucose levels of db/db + RS mice were distinctly lower than those of db/db + DMSO mice at 0, 60 and 120 minutes, with a reduction of 63.9% at 120 minutes (Fig. 5D).
Furthermore, the serum insulin level of the T2D + DMSO group was significantly decreased compared to the Ctrl group, whereas the T2D + RS group showed a marked increase compared to the T2D + DMSO group (Fig. 5E). Similarly, the serum insulin level of the db/db + DMSO group was dramatically decreased compared to the Ctrl group, while the serum insulin level of the db/db + RS group was obviously elevated compared to the db/db + DMSO group (Fig. 5F).
After immunofluorescence staining of mice pancreas, it was observed that the proliferation of islet cells in the T2D + DMSO group was reduced by 25.6% compared to the Ctrl mice. Moreover, the T2D + RS group showed an overt increase of 55.2% in islet cell proliferation compared to the T2D + DMSO group (Fig. 5G). Blockade of Mcp-1 signaling in db/db mice yielded similar results to T2D mice (Fig. 5H).
In comparison to the Ctrl group, the T2D + DMSO group exhibited a 36.4% reduction in p-Akt protein levels in the islet cells, while the T2D + RS group showed an overt increase compared to the T2D + DMSO group (as shown in Fig. 5I and the supplemental material). Similarly, the blockade of the Mcp-1 pathway in db/db mice resulted in similar outcomes to those observed in T2D mice (as shown in Fig. 5J and the supplemental material). Our results indicate that inhibiting Mcp-1 signaling in type 2 diabetic mice can partially restore proliferation and insulin secretion of the islet beta-cells.