TSH-SPP1/TRβ-TSH positive feedback loop mediates fat deposition of hepatocyte: Crosstalk between thyroid and liver

We conducted this study with two aims: (1) whether TRβ could be damaged by NAFLD, thereby represent thyroid hormone resistance-like manifestation and (2) to analyze the potential role of SPP1 in TH signaling pathway on the process of NAFLD. This study is expected to provide a new perspective on the therapeutic mechanism in the pathological course of NAFLD.

A total of 166 patients diagnosed with type 2 diabetes mellitus (T2DM) were enrolled in this study. All patients had a BMI above 24 kg/m2 and were stratified into two groups: NAFLD and Non-NAFLD groups. Ages, gender, BMI, duration of diabetes and biochemical markers were obtained from participants' records. We downloaded the dataset GSE48452 from GEO. The Pathview library was used to make the thyroid hormone signaling pathway visualization. The CIBERSORT algorithm was applied to calculate the infiltrated immune cells in obese NAFLD patients. C57BL/6 mice were randomly selected to constitute the normal control (NC) group and were fed a normal chow diet; the rest of the mice were fed a high-fat diet (HFD). After 12 weeks HFD feeding, the mice were sacrificed by cervical dislocation, and blood samples were collected. Mouse livers were also collected; one part of each liver was fixed in 10% formalin for histological analysis, and the other part was snap-frozen for subsequent molecular analyses. To explore the relationship between SPP1, TRβ and lipid deposition in hepatocytes, HepG2 cells were treated with 50 μ M concentration of PA and/or 20 ng/ml concentration of rh-SPP1 for 48h. In addition, the PC3.1-TRβ plasmid was constructed for further validation in HepG2 cells. We used THP-1 cells to construct an M1 macrophage model in vitro. We then analyzed THP-1 cells treated with various concentrations of PA or TSH.

(1) After adjusting for all factors that appeared P value less than 0.1 in the univariate analysis, BMI, TSH, and FT3 were significant independent risk factors of NAFLD (ORs were 1.218, 1.694, and 2.259, respectively); (2) A further analysis with BMI stratification indiacted that both FT3 and TSH had a significant change between individuals with NAFLD and Non-NAFLD in obesity subgroup; however, there was no statistic difference in over-weight group; (3) Bioinformatics analysis of GSE48452 had shown that several key molecular (including TRβ) of thyroid hormone pathway affected by NAFLD induced transcriptomic changes and the expression levels of SPP1, FABP4 and RPS4Y1 were significantly higher, while the expression levels of PZP and VIL1 were significantly decreased in NAFLD patients(adjusted p < 0.05, |logFC| > 1.0). The CIBERSORT algorithm showed increased M0 and M1, decreased M2 macrophage infiltration in NAFLD with comparison to healthy obese group; (4) After 12 weeks of HFD-feeding, the obesity mice had significantly higher serum TSH and In IHC-stained liver sections of obesity group, the intensity of SPP1 had a significantly increased, while TRβ reduced; (5) In vitro studies have shown SPP1 aggravated lipid deposition in hepatic cells dependent on down-regulating the expression of TRβ and TSH acts to promote secretion of SPP1 in M1 macrophage cells.

SPP1 secretion induced by M1 macrophage polarization, which may down-regulates TRβ in hepatocytes via paracrine manner, on the one hand, the lipid deposition aggravating in liver, on the other hand, a compensatory increase of TSH in serum. The increased TSH can further lead to the following SPP1 secretion of M1 macrophage. The positive feedback crosstalk between thyroid and liver, may be plays an important role in maintaining and amplifying pathological process of NAFLD.