Search
Close this search box.

Molecular mechanism underlying miR-204-5p regulation of adipose-derived stem cells differentiation into cells from three germ layers – Cell Death Discovery

Detection of infection efficiency of miR-204-5p up lentiviral vector

Uninfected gADSCs were used for the Control group; gADSCs infected with the empty lentiviral vector were labeled as the LV_Con group, and gADSCs infected with the miR-204-5p lentiviral vector were labeled as the LV_miR group. Fluorescence microscopy showed no green fluorescence in the Control group, whereas a large amount of green fluorescence was observed in both the LV_Con and LV_miR groups (Fig. S1A). Flow cytometric analysis showed that cells expressing the green fluorescent protein in the LV_Con and LV_miR groups accounted for > 97% of selected cells (Fig. S1B). Real-time quantitative polymerase chain reaction (qPCR) revealed that miR-204-5p was highly expressed in the LV_miR group (Fig. S1C). These results indicate that lentiviral vectors have high infection efficiency and miR-204-5p can be stably overexpressed in gADSCs.

Effects of overexpression of miR-204-5p on the differentiation of gADSCs into adipocytes

Differentiation in the Control, LV_Con, and LV_miR groups into adipocytes was simultaneously induced. Oil Red O staining and lipid droplet content analysis showed that the volume and number of lipid droplets increased significantly with the induction time. In addition, compared with those in the Control and LV_Con groups, the number and content of lipid droplets in the LV_miR group were lower, and the volume was slightly smaller (Fig. 1A, B).

Fig. 1: Effects of overexpression of miR-204-5p on gADSCs differentiation into adipocytes and neurocytes.
figure 1

A Identification of adipocytes using oil red O staining (Bar: 100 μm). B Determination of lipid droplet content. C Transcription of adipocyte marker genes PPARG, ADIPOQ, PERILIPIN, LEP, and IRS1 detected using real-time qPCR. D Cell morphology during differentiation of Control group into neurocytes (Bar: 100 μm). E NGF in the supernatant of neurocytes detected using ELISA. F Cellular immunofluorescence showing the expression of neurocyte marker genes ENO2, TAU, MAP2, and RBFOX3 before and after differentiation in the Control group (Bar: 100 μm). G Fluorescence intensity analysis of F. H Cellular immunofluorescence showing the expression of ENO2, TAU, MAP2, and RBFOX3 after differentiation in the Control, LV_Con, and LV_miR groups (Bar: 100 μm). I Fluorescence intensity analysis of H. J Transcription of ENO2, TAU, MAP2, and RBFOX3 detected using real-time qPCR. Data were presented as mean ± SD; P values were determined by a two-tailed unpaired t test, *P > 0.05, **0.01 < P < 0.05, ***P < 0.01.

Transcription of the adipocyte marker genes peroxisome proliferator-activated receptor gamma (PPARG), adiponectin (ADIPOQ), PERILIPIN, leptin (LEP), and insulin receptor substrate 1 (IRS1) during differentiation was detected using real-time qPCR (Fig. 1C). Compared with their expression in the Control and LV_Con groups, PPARG and IRS1 were downregulated in the LV_miR group during differentiation. ADIPOQ and LEP were highly transcribed when not induced. With longer induction, although ADIPOQ remained highly transcribed, the gap gradually narrowed, and LEP showed a varying degree of decrease. The transcription level of PERILIPIN showed no statistically significant differences during the initial induction but significantly decreased on day 15.

These results indicate that gADSCs can differentiate into adipocytes in vitro and that overexpression of miR-204-5p has inhibitory effects on the differentiation of gADSCs into adipocytes.

Effects of overexpression of miR-204-5p on the differentiation of gADSCs into neurocytes

Differentiation of Control, LV_Con, and LV_miR groups into neurocytes was simultaneously induced. Morphological observations showed that after 2 h of induction, cell morphology began to change, cells retracted, and part of the cytoplasm extended outward. After 8 h of induction, cell protrusions were more obvious, cells were interlaced and connected to a network, and the morphology of the neurocytes was obvious (Fig. 1D). Detection of nerve growth factor (NGF) content showed that NGF secreted by the LV_miR group was higher than that in the Control and LV_Con groups after 8 h (Fig. 1E).

Cellular immunofluorescence was used to identify the expression of the neurocyte marker genes enolase 2 (ENO2), tubulin-associated unit (TAU), microtubule-associated protein 2 (MAP2), and RNA binding protein Fox-1 homolog 3 (RBFOX3) before and after differentiation. The fluorescence intensities of ENO2, TAU, MAP2, and RBFOX3 were significantly higher in cells induced for 8 h than in uninduced cells (Fig. 1F, G). In addition, the expression of TAU, MAP2, and RBFOX3 was significantly higher in the LV_miR group than in the Control and LV_Con groups after differentiation (Fig. 1H, I).

Transcription of the neurocyte marker genes ENO2, TAU, MAP2, and RBFOX3 before and after differentiation was detected using real-time qPCR (Fig. 1J). The transcription levels of ENO2, TAU, and RBFOX3 were higher in cells induced for 8 h than in uninduced cells; however, MAP2 expression did not show this pattern. In addition, compared with their expression in the Control and LV_Con groups, ENO2, TAU, and RBFOX3 were upregulated in the LV_miR group after 8 h, and the transcription level of MAP2 was higher than that in the LV_Con group.

These results indicate that gADSCs can differentiate into neurocytes in vitro and the overexpression of miR-204-5p can promote the differentiation of gADSCs into neurocytes.

Effects of overexpression of miR-204-5p on the differentiation of gADSCs into hepatocytes

Differentiation of Control, LV_Con, and LV_miR groups into hepatocytes was simultaneously induced. The results of glycogen PAS staining showed that the differentiated hepatocytes had varying shades of purple-red within the cells after 3 days of maturation. The nucleus was visible and tightly surrounded by a large number of vesicular structures (i.e., glycogen vesicles) (Fig. 2A). The detection of albumin (ALB) and urea content showed that ALB and urea secreted by the LV_miR group were higher than that of the Control and LV_Con groups following differentiation (Fig. 2B, C).

Fig. 2: Effects of overexpression of miR-204-5p on gADSCs differentiation into hepatocytes.
figure 2

A Glycogen vesicles in hepatocytes detected using glycogen PAS staining (Bar: 50 μm). B ALB in the supernatant of hepatocytes detected using ELISA. C Determination of urea in the supernatant of hepatocytes. D Cellular immunofluorescence identified the expression of hepatocyte marker genes AFP, ALB, HNF4A, and KRT18 before and after differentiation in the Control group (Bar: 100 μm). E Fluorescence intensity analysis of D. F Cellular immunofluorescence identified the expression of AFP, ALB, HNF4A, and KRT18 after differentiation in the Control, LV_Con, and LV_miR groups (Bar: 100 μm). G Fluorescence intensity analysis of F. H Transcription of AFP, ALB, HNF4A, and KRT18 detected using real-time qPCR. Data were presented as mean ± SD; P values were determined by a two-tailed unpaired t test, *P > 0.05, **0.01 < P < 0.05, ***P < 0.01.

Cellular immunofluorescence was used to identify the expression of hepatocyte marker genes alpha fetoprotein (AFP), ALB, hepatocyte nuclear factor 4-alpha (HNF4A), and keratin 18 (KRT18) before and after differentiation. The fluorescence intensities of AFP, ALB, HNF4A, and KRT18 were significantly higher in cells that matured by day 3 than in uninduced cells (Fig. 2D, E). In addition, compared with those in the Control and LV_Con groups, no statistically significant difference was observed in the expression levels of ALB or KRT18 in the LV miR-204-5p group after differentiation induction, whereas the expression levels of AFP and HNF4A were significantly increased (Fig. 2F, G).

Transcription of the hepatocyte marker genes ENO2, TAU, MAP2, and RBFOX3 before and after differentiation was detected using real-time qPCR (Fig. 2H). Compared with those in the uninduced cells, the transcription levels of AFP, ALB, HNF4A, and KRT18 increased in cells after 3 days of maturation, wherein that of KRT18 was more obvious. In addition, AFP, ALB, HNF4A, and KRT18 were upregulated in the LV_miR group, compared with their expressions in the Control and LV_Con groups, after 3 days of maturation.

These results indicate that gADSCs can differentiate into hepatocytes in vitro and overexpression of miR-204-5p promotes the differentiation of gADSCs into hepatocytes.

The molecular pathway of miR-204-5p regulating the differentiation of gADSCs into adipocytes

Transcriptome sequencing was performed on the Control, LV_Con, and LV_miR group cells. A total of 18,998 genes and 1,247 differential genes were obtained, of which the LV_Con_vs_LV_miR group had 255 differential genes. According to functional annotation, eight differentially expressed genes were associated with adipocyte differentiation in the LV_Con_vs_LV_miR group, namely ZBTB7C, ERO1A, CCND1, LEP, RNASEL, PTGS2, INHBB, and ADIPOQ. The corresponding line chart of the gene expression is shown in Fig. 3A, while the clustering heat map is in Fig. 3B. GO enrichment analysis revealed that these genes were involved in cell differentiation, cellular developmental process, and fat cell differentiation (Fig. 3C). KEGG enrichment analysis revealed that these genes were enriched in signaling pathways such as AMPK, Adipocytokine, and JAK-STAT (Fig. 3D).

Fig. 3: Analysis of differential genes and differential proteins associated with adipocyte differentiation.
figure 3

A Line chart of expression of differential genes associated with adipocyte differentiation. B Clustering heatmap of differential genes associated with adipocyte differentiation. C Chordal graph of GO enrichment of differential genes associated with adipocyte differentiation. D KEGG enrichment analysis of differential genes associated with adipocyte differentiation. E Venn analysis of the adipocyte_differentiation and LV_Con_vs_LV_miR groups. F Clustering heatmap of differential proteins associated with adipocyte differentiation. G GO enrichment analysis of differential proteins associated with adipocyte differentiation. H KEGG enrichment analysis of differential proteins associated with adipocyte differentiation. I Transcription of differential genes associated with adipocyte differentiation detected using real-time qPCR. J Expression of differential proteins associated with adipocyte differentiation detected using western blotting. Data were presented as mean ± SD; P values were determined by a two-tailed unpaired t test, *P > 0.05, **0.01 < P < 0.05, ***P < 0.01.

Proteome sequencing identified 6,569 proteins and 1,610 differential proteins, among which the LV_Con_vs_LV_miR group had 483 differential proteins. We identified 23 out of 6,569 proteins with functional annotations associated with adipocyte differentiation, named the adipocyte_differentiation group. Venn analysis of the adipocyte_differentiation and LV_Con_vs_LV_miR group revealed that only ADIPOQ and JAG1 were associated with adipocyte differentiation in the LV_Con_vs_LV_miR group (Fig. 3E) and were upregulated (Fig. 3F). GO enrichment analysis revealed that these two proteins are involved in adipocyte differentiation and play a negative regulatory role in this process (Fig. 3G). KEGG enrichment analysis revealed that these two proteins were enriched in signaling pathways such as those of AMPK, NOTCH, and PPAR (Fig. 3H).

ADIPOQ was detected in the differential expression analysis of the transcriptome and proteome and negatively regulated adipocyte differentiation. Real-time qPCR and western blotting showed that ADIPOQ expression was upregulated in the LV_miR group (Fig. 3I, J). ADIPOQ was a key factor in activating the AMPK signaling pathway, and AMPK, the core protein of this pathway, was also upregulated in the LV_miR group. Two important proteins associated with adipocyte differentiation, LEP and PPARG, were also involved in this pathway. LEP is a key factor in activating AMPK, whereas PPARG is located downstream of AMPK. Real-time qPCR showed that LEP was upregulated and PPARG was downregulated in the LV_miR group; no statistically significant difference in AMPK expression was observed (Fig. 3I). Western blotting showed that AMPK was upregulated and PPARG was downregulated in the LV_miR group; however, LEP expression was not detected (Fig. 3J). JAG1 is a key protein that activates the NOTCH signaling pathway and negatively regulates adipocyte differentiation. Real-time qPCR and western blotting showed that JAG1 and its downstream target NOTCH3 were upregulated in the LV_miR group (Fig. 3I, J). These results indicate that miR-204-5p may inhibit the differentiation of gADSCs into adipocytes by acting on the AMPK signaling pathway and the JAG1/NOTCH3 axis.

The molecular pathway of miR-204-5p regulating the differentiation of gADSCs into neurocytes

Functional annotation of the transcriptome showed that eight differentially expressed genes were associated with neurocyte differentiation in the LV_Con_vs_LV_miR group, namely CDON, EN1, CNTN4, CEND1, SPOCK1, SIX1, INHBA, and BRSK1. The corresponding line chart of gene expression is shown in Fig. 4A, while the clustering heat map in Fig. 4B. GO enrichment analysis revealed that these genes are involved in the regulation of nervous system development, neuron differentiation, and neurogenesis (Fig. 4C), while KEGG enrichment analysis showed that these genes were enriched in signaling pathways such as Hedgehog and TGF-beta (Fig. 4D).

Fig. 4: Analysis of differential genes and differential proteins associated with neurocyte differentiation.
figure 4

A Line chart of expression of differential genes associated with neurocyte differentiation. B Clustering heatmap of differential genes associated with neurocyte differentiation. C Chordal graph of GO enrichment of differential genes associated with neurocyte differentiation. D KEGG enrichment analysis of differential genes associated with neurocyte differentiation. E Venn analysis of the neurocyte_differentiation and LV_Con_vs_LV_miR groups. F Clustering heatmap of differential proteins associated with neurocyte differentiation. G GO enrichment analysis of differential proteins associated with neurocyte differentiation. H KEGG enrichment analysis of differential proteins associated with neurocyte differentiation. I Transcription of differential genes associated with neurocyte differentiation detected using real-time qPCR. J Expression of differential proteins associated with neurocyte differentiation detected using western blotting. Data were presented as mean ± SD; P values were determined by a two-tailed unpaired t test, *P > 0.05, **0.01 < P < 0.05, ***P < 0.01.

According to the functional annotation of the proteome, 91 proteins were associated with neurocyte differentiation, and these were named as the neurocyte_differentiation group. Venn analysis of the neurocyte_differentiation and LV_Con_vs_LV_miR groups revealed that seven differentially expressed proteins were associated with neurocyte differentiation in the LV_Con_vs_LV_miR group (Fig. 4E), namely NOTCH3, IMPACT, PRRX1, JAG1, EPHA4, VIM, and plexin-B2, of which four were upregulated and three were downregulated (Fig. 4F). GO enrichment analysis revealed that these proteins were mainly involved in the regulation of cell differentiation, neuron differentiation, and neurogenesis (Fig. 4G), while KEGG enrichment analysis revealed that these proteins were enriched in signaling pathways, such as NOTCH, Apelin, and Axon guidance (Fig. 4H).

Among the differential proteins identified using proteome analysis, JAG1 and NOTCH3, two key proteins in the NOTCH signaling pathway, have important regulatory effects on neurocyte differentiation and were upregulated in the LV_miR group (Fig. 3I, J). In addition, the differentially expressed proteins plexin-B2 and VIM, and genes CEND1 and BRSK1 also play important roles in neurocyte differentiation. Among these, plexin-B2, CEND1, and BRSK1 play positive regulatory roles in neurocyte differentiation, whereas VIM plays a negative regulatory role. Real-time qPCR showed that plexin-B2, CEND1, and BRSK1 were upregulated in the LV_miR group; however, there was no statistically significant difference in VIM (Fig. 4I). Western blotting showed that plexin-B2 was upregulated in the LV_miR group and the expression of VIM was significantly lower than that in the LV_Con group. In addition, CEND1 and BRSK1 expression was not detected (Fig. 4J). These results indicate that the promoting effect of miR-204-5p on the differentiation of gADSCs into neurocytes may be achieved by acting on the JAG1/NOTCH3 axis. In addition, miR-204-5p may promote the differentiation of gADSCs into neurocytes by upregulating the expression of plexin-B2, CEND1, and BRSK1 and downregulating that of VIM.

The molecular pathway of miR-204-5p regulating the differentiation of gADSCs into hepatocytes

The functional annotation of the transcriptome showed that four differentially expressed genes were associated with hepatocytes in the LV_Con_vs_LV_miR group, namely IL6, TGFA, E2F8, and INHBB. The corresponding line chart of gene expression is shown in Fig. 5A, while the clustering heat map in Fig. 5B. GO enrichment analysis revealed that these genes are involved in the regulation of cytokine and hepatocyte growth factor biosynthetic process (Fig. 5C), while KEGG enrichment analysis revealed that these genes were mainly enriched in signaling pathways such as FoxO, TGF-beta, and HIF-1 (Fig. 5D).

Fig. 5: Analysis of differential genes and differential proteins associated with hepatocytes.
figure 5

A Line chart of expression of differential genes associated with hepatocytes. B Clustering heatmap of differential genes associated with hepatocytes. C Chordal graph of GO enrichment of differential genes associated with hepatocytes. D KEGG enrichment analysis of differential genes associated with hepatocytes. E Venn analysis of the hepatocyte and LV_Con_vs_LV_miR groups. F Clustering heatmap of differential proteins associated with hepatocytes. G GO enrichment analysis of differential proteins associated with hepatocytes. H Transcription of differential genes associated with hepatocyte differentiation detected using real-time qPCR. I Expression of differential proteins associated with hepatocyte differentiation detected using western blotting. Data were presented as mean ± SD; P values were determined by a two-tailed unpaired t test, *P > 0.05, **0.01 < P < 0.05, ***P < 0.01.

Functional annotation of the proteome identified 13 proteins associated with hepatocytes, named the hepatocyte group. Venn analysis of hepatocyte and LV_Con_vs_LV_miR groups revealed that only two differentially expressed proteins were associated with hepatocytes in the LV_Con_vs_LV_miR group (Fig. 5E) and upregulated in the LV_miR group (Fig. 5F). GO enrichment analysis revealed that the two differentially expressed proteins were involved in the regulation of hepatocyte proliferation and integrin-mediated signaling pathway (Fig. 5G), none of which were associated with hepatocyte differentiation.

Only E2F8 was associated with hepatocyte differentiation. Real-time qPCR showed that E2F8 was upregulated in the LV_miR group compared with the LV_Con group (Fig. 5H). Western blotting showed that the expression of E2F8 in the LV_miR group was significantly higher than in the Control and LV_Con groups (Fig. 5I). These results indicate that the promoting effect of miR-204-5p on the differentiation of gADSCs into hepatocytes may be achieved by increasing the expression of E2F8.

The effects of AMPK and NOTCH signaling pathways on the differentiation of gADSCs into adipocytes

The gADSCs were treated with the AMPK activator (AMPK(+)) AICAR and inhibitor (AMPK(−)) dorsomorphin (Compound C) 2HCl and the NOTCH activator (NOTCH(+)) valproic acid (VPA) and inhibitor (NOTCH(−)) LY411575, respectively, using untreated gADSCs as control. The expression of key genes in the AMPK and NOTCH signaling pathways was analyzed using real-time qPCR and western blotting. Real-time qPCR showed that, compared with their levels in the control group, AMPK was upregulated and LEP and PPARG were downregulated in the AMPK(+) group, whereas no statistically significant difference in ADIPOQ expression was observed. In the AMPK(−) group, ADIPOQ, LEP, and PPARG were upregulated, whereas AMPK was downregulated (Fig. 6A). Western blotting showed that AMPK was upregulated, ADIPOQ and PPARG were downregulated, and no expression of LEP was detected in the AMPK (+) group. The results of the AMPK(−) group were opposite to those of the AMPK(+) group (Fig. 6B). In addition, real-time qPCR and western blotting showed that compared with their levels in the control group, JAG1 and NOTCH3 were upregulated in the NOTCH(+) group but downregulated in the NOTCH(−) group (Fig. 6C, D).

Fig. 6: Effects of AMPK and NOTCH signaling pathways on gADSCs differentiation into adipocytes.
figure 6

A, B Expression of key genes in the AMPK signaling pathway detected using real-time qPCR and western blotting. C, D Expression of key genes in the NOTCH signaling pathway detected using real-time qPCR and western blotting. E, F Effects of AMPK signaling pathway on gADSCs differentiation into adipocytes detected using oil red O staining and quantitative detection of lipid droplets (Bar: 50 μm). G, H Effects of NOTCH signaling pathway on gADSCs differentiation into adipocytes detected using oil red O staining and quantitative detection of lipid droplets (Bar: 50 μm). I Transcription of adipocyte marker genes before and after differentiation in AMPK(+) and AMPK(−) groups detected using real-time qPCR. J Transcription of adipocyte marker genes before and after differentiation in NOTCH(+) and NOTCH(−) groups detected using real-time qPCR. Data were presented as mean ± SD; P values were determined by a two-tailed unpaired t test, *P > 0.05, **0.01 < P < 0.05, ***P < 0.01.

Oil red O staining and quantitative detection of lipid droplets were performed on adipocytes differentiated from the Control, AMPK(+), AMPK(−), NOTCH(+), and NOTCH(−) groups. Compared with those in the Control group, the number and content of lipid droplets were lower in the AMPK(+) group but higher in the AMPK(−) group (Fig. 6E, F). In addition, the number and content of lipid droplets in the NOTCH(+) group were lower than those in the Control group; no statistically significant difference between the NOTCH(−) and Control groups was observed (Fig. 6G, H). Real-time qPCR showed that after differentiation, compared with their levels in the Control group, PPARG, ADIPOQ, and LEP were downregulated, and PERILIPIN and IRS1 were upregulated in the AMPK(+) group; however, their transcription levels were lower than those in the AMPK(−) group. PPARG, ADIPOQ, LEP, PERILIPIN, and IRS1 were upregulated in the AMPK(−) group (Fig. 6I). In addition, ADIPOQ and LEP were downregulated, while PPARG, PERILIPIN, and IRS1 were upregulated in the NOTCH(+) group. In the NOTCH(−) group, PPARG, ADIPOQ, LEP, and PERILIPIN were upregulated, whereas only IRS1 was downregulated (Fig. 6J). Therefore, the activation of the AMPK signaling pathway inhibited gADSCs differentiation into adipocytes, whereas inhibition of the AMPK signaling pathway had the opposite effect. In addition, activation of the NOTCH signaling pathway inhibited the differentiation of gADSCs into adipocytes, but the promotional effect of inhibiting this pathway on the differentiation of gADSCs into adipocytes was not obvious.

The effects of NOTCH signaling pathways on the differentiation of gADSCs into neurocytes

The gADSCs were treated with the NOTCH activator (NOTCH(+)) valproic acid (VPA) and inhibitor (NOTCH(−)) LY411575, respectively, using untreated gADSCs as a control. The expression of key genes in the NOTCH signaling pathway was detected using real-time qPCR and western blotting. JAG1 and NOTCH3 were upregulated in the NOTCH(+) group, while downregulated in the NOTCH(−) group (Fig. 6C, D).

Differentiation of the Control, NOTCH (+), and NOTCH(−) groups into neurocytes was simultaneously induced. Morphological observations showed that following an 8 h induction, the cell body retracted, part of the cytoplasm extended outward, the cells were interconnected into a network, and the morphology of the neurocytes was evident (Fig. 7A). NGF content was higher in the NOTCH (+) group, but lower in the NOTCH(−) group, than in the Control group (Fig. 7B). Cellular immunofluorescence showed that the fluorescence intensity of ENO2, MAP2, RBFOX3, and TAU in the NOTCH(+) group was higher than the corresponding intensities in the Control group, whereas they were lower in the NOTCH (−) group, except for RBFOX3 (Fig. 7C). Real-time qPCR showed that following an 8 h induction, compared with the levels in the Control group, ENO2, MAP2, RBFOX3, and TAU were upregulated in the NOTCH(+) group, while RBFOX3 was downregulated and ENO2 and MAP2 were upregulated in the NOTCH(−) group; there was no statistically significant difference in TAU, but their transcription levels were lower than those in the NOTCH(+) group (Fig. 7D). These results indicated that activation of the NOTCH signaling pathway promoted the differentiation of gADSCs into neurocytes, whereas inhibition of the NOTCH signaling pathway had the opposite effect.

Fig. 7: Effects of NOTCH signaling pathway on gADSCs differentiation into neurocytes.
figure 7

A Cell morphology before and after differentiation of the Control, NOTCH(+), and NOTCH(−) groups into neurocytes (Bar: 50 μm). B NGF in the supernatant of neurocytes detected using ELISA. C Cellular immunofluorescence identified the expression of ENO2, TAU, MAP2, and RBFOX3 after differentiation in the Control, NOTCH(+), and NOTCH(−) groups (Bar: 100 μm). D Transcription of ENO2, TAU, MAP2, and RBFOX3 before and after differentiation in the Control, NOTCH(+) and NOTCH(−) groups detected using real-time qPCR. Data were presented as mean ± SD; P values were determined by a two-tailed unpaired t test, *P > 0.05, **0.01 < P < 0.05, ***P < 0.01.