Identifying adipogenic chemicals: Disparate effects in 3T3-L1, OP9 and primary mesenchymal multipotent cell models

Abstract

3T3-L1 pre-adipocytes are used commonly to identify new adipogens, but this cell line has been shown to produce variable results. Here, potential adipogenic chemicals (identified in the ToxCast dataset using the Toxicological Priority Index) were tested for their ability to induce adipocyte differentiation in 3T3-L1 cells, OP9 cells and primary mouse bone marrow multipotent stromal cells (BM-MSC). Ten of the 36 potential adipogens stimulated lipid accumulation in at least one model (novel: fenthion, quinoxyfen, prallethrin, allethrin, pyrimethanil, tebuconzaole, 2,4,6-tris (tert-butyl)phenol; known: fentin, pioglitazone, 3,3',5,5'-tetrabromobisphenol A). Only prallethrin and pioglitazone enhanced lipid accumulation in all models. OP9 cells were significantly more sensitive to chemicals known to activate PPARγ through RXR than the other models. Coordinate effects on adipocyte and osteoblast differentiation were investigated further in BM-MSCs. Lipid accumulation was correlated with the ability to stimulate expression of the PPARγ target gene, Plin1. Induction of lipid accumulation also was associated with reduction in alkaline phosphatase activity. Allethrin, prallethrin, and quinoxyfen strongly suppressed osteogenic gene expression. BM-MSCs were useful in coordinately investigating pro-adipogenic and anti-osteogenic effects. Overall, the results show that additional models should be used in conjunction with 3T3-L1 cells to identify a broader spectrum of adipogens and their coordinate effects on osteogenesis.

Keywords: Adipogen; Adipogenesis; Multipotent mesenchymal stromal cells; OP9, 3T3-L1; Osteogenesis; PPARγ; RXR; ToxPi.

Figure 1.. Lipid accumulation induced by test compounds in 3T3 L1 and OP9 cell lines undergoing adipogenesis.

A standard hormonal protocol was used to induce adipogenesis (See Methods). At the initiation of differentiation, cells received no treatment (naïve) or were dosed with Vh (DMSO, 0.2% final concentration), positive controls (Rosi, 200 nM; LG268, 200 nM; TBT, 100 nM; TPhP, 10 μM), or test compounds as indicated in Table 1. Lipid accumulation was quantified by Nile Red staining on day 10. A) Positive controls in 3T3-L1. B) Test compounds in 3T3-L1. C) Positive controls in OP9. D) Test compounds in OP9. Hatched bars = Vh. Light gray bars = negative controls. Dark gray bars = experimental compounds. Data are presented as means ± SE from 3–5 independent biological replicates. Statistically different from Vh-treated (* p<0.05, ** p<0.01, ANOVA, Dunnett’s).

Figure 2.. Correlation of lipid accumulation induced in 3T3 L1 and OP9 cells.

Data are from Figure 1. Each data point represents the average lipid accumulation induced by a chemical (reported as “% Rosi Control) in each cell line. The green “X” is Vh. The red “X” is Rosi. Red circles indicate known RXR ligands. A) All compounds tested. B) Known RXR ligands removed. The dotted line represents the 90% prediction bands based on the linear regression.

Figure 3.. PPARγ reporter activation by test compounds.

Cos-7 cells transfected with a human PPARγ1 expression construct and a PPRE-luciferase reporter plasmid were dosed with Vh (DMSO, 0.1% final concentration) or the indicated test compounds. Reporter activity was measured after 24 hours. Data were normalized as described in the Methods. Data are presented as means ± SE from 4 independent transfections. Statistically different from Vh-treated (* p<0.05, ** p<0.01, *** p<0.001, ANOVA, Dunnett’s).

Figure 4.. Lipid accumulation induced by test compounds in primary mouse BM-MSCs undergoing osteogenesis.

After primary bone marrow cultures were established, differentiation was initiated with the addition of ascorbate, β-glycerol phosphate, insulin and dexamethasone. At the initiation of differentiation, cells received no treatment (naïve) or were dosed with Vh (DMSO, 0.2% final concentration), positive controls (Rosi, 200 nM; LG268, 200 nM; TBT, 100 nM; TPhP, 10 μM), or test compounds as indicated in Table 1. Lipid accumulation was quantified by Nile Red staining on day 7. A) Positive controls. B) Test compounds. Hatched bars = Vh. Light gray bars = negative controls. Dark gray bars = experimental compounds. Data are presented as means ± SE from 4 independent bone marrow preparations. Statistically different from Vh-treated (* p<0.05, ** p<0.01, ANOVA, Dunnett’s).

Figure 5.. Adipogenic gene expression induction by active compounds.

Primary bone marrow cultures were established and differentiation was initiated as described in Figure 4. Cultures received no treatment (Naïve) or were treated with Vh (DMSO, 0.1% final concentration), Rosi (positive control, 100 nM) or test compounds as indicated in Table 1. A-C) Gene expression (Pparg or its target genes Fabp4 and Plin1) was quantified after 7 days by qRT-PCR. D) Regression analysis of association between Fabp4 expression and lipid accumulation means (data are from Figure 4). The green “X” is Vh. The red “X” is Rosi. r = Pearson’s correlation coefficient. Data are presented as means ± SE from 25 (Vh) or 4–8 (test chemicals) independent bone marrow preparations. Statistically different from Vh-treated (* p<0.05, ** p<0.01, ***p<0.001, ANOVA, Dunnett’s). ATRA-treated was tested separately vs. Vh-treated (* p<0.05, Student’s t test).

Figure 6.. Suppression of alkaline phosphatase activity by test compounds.

Primary bone marrow cultures were established and differentiation was initiated as described in Figure 4. Alkaline phosphatase activity was quantified after 12 days. A) Positive controls. B) Test compounds. Hatched bars = Vh. Light gray bars = negative controls. Dark gray bars = experimental compounds. C) Regression analysis of association between alkaline phosphatase activity and lipid accumulation means. The green “X” is Vh. The red “X” is Rosi. r = Pearson’s correlation coefficient. Data are presented as means ± SE from 4 independent bone marrow preparations. Statistically different from Vh-treated (** p<0.01, *** p<0.001, ANOVA, Dunnett’s).

Figure 7.. Suppression of osteogenic gene expression by test compounds.

Primary bone marrow cultures were established and differentiation was initiated as described in Figure 5. A-C) Gene expression (Runx2 or its target genes Osx and Bglap) was quantified after 7 days by qRT-PCR. D) Regression analysis of association between Osx and Fapb4 (data are from Figure 5) expression means. The green “X” is Vh. The red “X” is Rosi. r = Pearson’s correlation coefficient. Data are presented as means ± SE from 24 (Vh) or 4–8 (test chemicals) independent bone marrow preparations. Statistically different from Vh-treated (* p<0.05, ** p<0.01, ***p<0.001, ANOVA, Dunnett’s).