Role of Hyaluronan in Human Adipogenesis: Evidence from in-Vitro and in-Vivo Studies

Abstract

Hyaluronan (HA), an extra-cellular matrix glycosaminoglycan, may play a role in mesenchymal stem cell differentiation to fat but results using murine models and cell lines are conflicting. Our previous data, illustrating decreased HA production during human adipogenesis, suggested an inhibitory role. We have investigated the role of HA in adipogenesis and fat accumulation using human primary subcutaneous preadipocyte/fibroblasts (PFs, n = 12) and subjects of varying body mass index (BMI). The impact of HA on peroxisome proliferator-activated receptor gamma (PPARγ) expression was analysed following siRNA knockdown or HA synthase (HAS)1 and HAS2 overexpression. PFs were cultured in complete or adipogenic medium (ADM) with/without 4-methylumbelliferone (4-MU = HA synthesis inhibitor). Adipogenesis was evaluated using oil red O (ORO), counting adipogenic foci, and measurement of a terminal differentiation marker. Modulating HA production by HAS2 knockdown or overexpression increased (16%, p < 0.04) or decreased (30%, p = 0.01) PPARγ transcripts respectively. The inhibition of HA by 4-MU significantly enhanced ADM-induced adipogenesis with 1.52 ± 0.18- (ORO), 4.09 ± 0.63- (foci) and 2.6 ± 0.21-(marker)-fold increases compared with the controls, also increased PPARγ protein expression (40%, (p < 0.04)). In human subjects, circulating HA correlated negatively with BMI and triglycerides (r = -0.396 (p = 0.002), r = -0.269 (p = 0.038), respectively), confirming an inhibitory role of HA in human adipogenesis. Thus, enhancing HA action may provide a therapeutic target in obesity.

Keywords: 4-methylumbelliferone; Adipogenesis; BMI; Extracellular matrix; PPARγ; fat accumulation; hyaluronan; mesenchymal stem cell.

Figure 1

Hyaluronan (HA) from HA synthase (HAS)2 affects peroxisome proliferator-activated receptor gamma (PPARγ) transcripts. Subcutaneous preadipocytes/fibroblasts (PFs) (n = 4) were cultured until ~90% confluent, treated with HAS siRNA (A) or a hyaluronan-enriched supernatant (B) from HAS1 or HAS2 stable expressing HEK293 cell lines in a serum-free medium for 24 h, compared to scrambled siRNA or a supernatant from HEK293 cells with empty vector controls, respectively. PPARγ transcripts were measured by Q-PCR. Results are expressed as increased or decreased percentage changes in comparison to the controls (A) or transcript copy number (TCN) per 1000 copies of housekeeper gene (adenosine phosphoribosyl transferase, APRT) (B). Histograms = the mean ± SEM of all samples studied. * p < 0.04; ** p ≤ 0.01.

Figure 2

Enhanced adipogenesis by HA synthetic inhibitor, 4-MU. Confluent primary subcutaneous PFs were cultured in adipogenic medium (ADM) or complete medium (CM) with/without 4-MU for 22 days. Total RNA and nuclear protein were prepared. Fold effect (relative to the untreated control) of adipogenesis using foci counting (representative photos were shown with arrows indicating differentiating adipocytes), lipoprotein lipase (LPL) transcripts and oil red O (ORO) staining methods (n = 6). The table reports QPCR results together with foci numbers and ORO optical density values as the mean ± SEM. p-values are indicated in the figure above. ** p ≤ 0.01; *** p ≤ 0.006.

Figure 3

Enhanced PPARγ expression by 4-MU treatment. Confluent primary subcutaneous PFs were cultured in ADM or CM with/without 4-MU for 22 days. Total RNA and nuclear protein were prepared. (A) PPARγ transcripts (CM, n = 4; ADM, n = 8) and (B) nuclear protein (n = 4) were analysed by QPCR and Western blotting, respectively. Results are expressed as transcript copy number (TCN) per 1000 copies of housekeeper gene (adenosine phosphoribosyl transferase (APRT)) for QPCR, or PPARγ/actin for protein expression. Histograms = the mean ± SD (A) or SEM (B) of all samples studied. Two-tailed t-test has been used. * p < 0.04; ** p ≤ 0.01.

Figure 4

Decreased percentage of adiogenesis by LMW-HA treatment. LPL (terminal adipogenesis marker) was analysed by treatment of hyaluronan fragments (41–65 (65 kDa), 151–300 (300 kDa) and 1010–1800 (1 MDa)) in ADM compared to the no-HA control. Results are given as a percentage of the no-HA control (control as 100%) in LPL Transcript Copy Number (TCN)/1000 APRT. * p < 0.05.

Figure 5

Negative correlations of estimated circulating HA per adipose tissue and BMI/triglyceride. The scatterplot showing the relationship between blood volume-adjusted circulating HA (ng/kg) and BMI (kg/m²), or triglyceride (mmol/L). Statistically significant Spearman’s correlations have shown between baseline characteristics and circulating HA.