Impaired mRNA splicing and proteostasis in preadipocytes in obesity-related metabolic disease

Preadipocytes are crucial for healthy adipose tissue expansion. Preadipocyte differentiation is altered in obese individuals, which has been proposed to contribute to obesity-associated metabolic disturbances. Here, we aimed at identifying the pathogenic processes underlying impaired adipocyte differentiation in obese individuals with insulin resistance (IR)/type 2 diabetes (T2D). We report that down-regulation of a key member of the major spliceosome, PRFP8/PRP8, as observed in IR/T2D preadipocytes from subcutaneous (SC) fat, prevented adipogenesis by altering both the expression and splicing patterns of adipogenic transcription factors and lipid droplet-related proteins, while adipocyte differentiation was restored upon recovery of PRFP8/PRP8 normal levels. Adipocyte differentiation was also compromised under conditions of endoplasmic reticulum (ER)-associated protein degradation (ERAD) hyperactivation, as occurs in SC and omental (OM) preadipocytes in IR/T2D obesity. Thus, targeting mRNA splicing and ER proteostasis in preadipocytes could improve adipose tissue function and thus contribute to metabolic health in obese individuals.


Introduction
phenotypes, NG vs. T2D. This study enabled the identification of a total of 1,758 143 proteins that were present in both SC and OM fat from NG and T2D obese 144 individuals, thus defining the human obese preadipocyte proteome. According to 145 GO Biological Process annotation, 55.5% of these proteins were related to 146 cellular and metabolic processes (Figure 1-Figure supplement 2A 189 One of the most highly expressed genes in SC preadipocytes from NG obese 190 individuals that was significantly down-regulated in both IR and T2D obese 191 subjects as compared to NG SC preadipocytes,was PRPF8/PRP8. Specifically,192 mRNA and protein levels of this key component of the major spliceosome were 193 reduced by 51% and 56%,respectively,in IR SC preadipocytes,and by 49% and 194 82% in T2D SC preadipocytes as compared to NG levels (Figures 2A-C). No 195 differences in PRPF8/PRP8 expression were observed among groups in OM 196 preadipocytes (Figures 2A-C). 197 Differentiation studies revealed that while PRPF8 expression levels remained  Similar to that observed for morbidly obese individuals from cohort 1 (Figure 2B), 208 significantly decreased PRPF8 transcript levels were detected in T2D vs. NG 209 individuals with simple obesity (Figure 2E). No differences were observed in OM 210 preadipocytes among groups ( Figure 2E). 212 We employed a siRNA strategy to down-regulate PRPF8 gene expression levels 213 in preadipocytes to mimic the conditions found in IR/T2D obese SC 214 preadipocytes as compared to NG obesity. These studies were carried out using 215 the human SC adipocyte cell line, SGBS cells. As observed for NG SC human 216 primary preadipocytes, PRPF8 mRNA levels reached a peak at early stages of 217 SGBS cell differentiation (D4) (Figure 3A). 218 siRNA treatment of SGBS preadipocytes (D4) decreased by 67% and 65% 219 PRPF8 mRNA and protein levels, respectively, at day 3 post-transfection (D7), 220 without changing cell viability (Figure supplements 6A-C). Morphometric 221 evaluation of Oil-Red O staining in confocal micrographs revealed that PRPF8-222 silenced preadipocytes accumulated more but smaller LDs than control cells, 223 which resulted in an increase in the total lipid content in cells exposed to PRPF8 224 siRNA ( Figure 3C). In silico analysis of CLIP_Seq data using ENCORI (The  Figure 3C). In particular, mRNA and protein levels of the fat specific PPARG 237 isoform, PPARG-2, were up-regulated upon PRPF8 silencing (Figures 3D-E). 238 Decreased levels and/or altered splicing patterns of BSCL2, CIDEB,and CIDEC,239 were also observed in silenced cells ( Figure 3F) 254 Notably, recovery of PRP8 protein levels by co-transfection of PRPF8-silenced 255 SGBS cells with an expression vector coding for this protein (PRPF8-pcDNA3.1) 256 reverted the effects induced by PRPF8 silencing on adipocyte and LD markers, 257 both at D7 (data not shown) and D10 (Figure supplements 6).  changes in both LD number and size were also reverted in rescue experiments 259 by PRPF8 re-expression ( Figure 3I). 260 Silencing experiments using human adipose-derived stem cells (hADSCs) The UPR is altered in preadipocytes of IR/T2D obese subjects 265 As mentioned earlier, pathway analysis of iTRAQ proteomic data indicated that 266 ER stress-related pathways were altered in both SC and OM preadipocytes from 267 T2D obese individuals when compared to NG obese subjects (Figures 1B and   268   1C). 269 Immunoblotting studies of additional human preadipocyte samples to those  Figure 4B). In addition, pPERK and pPERK/PERK ratio were up-regulated in 279 SC and OM preadipocytes in relation to IR and/or T2D ( Figure 4C). However,280 peIF2 and peIF2/eIF2 ratio as well as the eIF2-target, CHOP, were 281 significantly reduced in OM preadipocytes in relation to IR/T2D (Figures 4C and 282 4E). We observed enhanced IRE1 levels and decreased pIRE1/IRE1 ratio in 283 IR vs. NG OM preadipocytes ( Figure 4F). IR OM preadipocytes and IR/T2D SC 284 preadipocytes also exhibited higher levels of the spliced form of XBP1 than NG 285 preadipocytes ( Figure 4G). Finally, the ER chaperones, GRP94 and PDI, were 286 more abundant in T2D preadipocytes than in NG preadipocytes from both SC 287 and OM fat ( Figure 4H). Our proteomic study revealed enhanced levels of 11 288 additional ER chaperones in T2D SC preadipocytes (Figure 1-source data 2). 289

Dysregulation of ER-associated protein degradation (ERAD) in IR/T2D
290 obese preadipocytes 291 In association with the UPR, the ERAD represents a key quality-control 292 machinery that recruits unfolded/misfolded ER proteins via ER chaperones and  Figure 5D) in IR/T2D vs. NG obesity. These changes 301 occurred in both SC and OM preadipocytes, which split into two clusters (NG and 302 IR/T2D) when ERAD data was represented in a two-way hierarchical clustering 303 heatmap ( Figure 5E). Interestingly, as we observed for IR/T2D preadipocytes vs. 304 NG preadipocytes (either SC or OM), increasing BiP expression levels by   310 Finally, the ERAD process was also explored during adipocyte differentiation. 311 mRNA levels of ERAD genes increased during differentiation of human primary 312 adipocytes, especially in IR, which resulted in higher overall transcript contents 313 of these genes in IR preadipocytes than in NG preadipocytes (measured as AUC) 314 ( Figure 5-Figure supplement 9). 315 Hyperglycaemia/hyperinsulinemia alters preadipocyte ER proteostasis 316 In order to unveil the regulation of ERAD in response to obesogenic insults, 317 SGBS preadipocytes were exposed to high concentrations of glucose and insulin 318 (HGHI), TNFα, or fatty acids (palmitate or oleate), as in vitro models of 319 hyperglycaemia/hyperinsulinemia, inflammation, and hypertrophy due to lipid 320 overload, respectively (Díaz-Ruiz et al., 2015). None of the treatments 321 compromised cell viability (data not shown) and, except for oleate, they impaired 322 insulin-induced Akt phosphorylation ( Figure 5-Figure supplement 10C). 323 Exposure to HGHI increased the expression of BiP and nearly all the other ERAD 324 components tested ( Figure 5F). As shown in Figure 5G, expression levels of 325 ERAD genes enabled discrimination of HGHI-treated from control SGBS cells.

326
Given these observations, we next examined the activity of the protein 327 degradation machinery in the cytosol, the proteasome (Bard et al., 2018), in cells 328 exposed to HGHI. These studies showed that hyperglycaemic/hyperinsulinemic 329 conditions decreased the activity of the 26S proteasome while increasing the 330 amount of ubiquitinated proteins in SGBS preadipocytes (Figures 5H and 5K). 331 The naturally occurring bile acid, tauroursodeoxycholic acid (TUDCA), has 332 been shown to restore ER homeostasis in ER-stressed cells (Zhang et al., 2018). 333 We thus tested whether exposure of SGBS preadipocytes to TUDCA prior to 334 HGHI treatment could prevent HGHI-induced up-regulation of ERAD genes and 335 accumulation of ubiquitinated proteins in these cells, which was proven to be the 336 case (Figures 5I and 5J). TUDCA also reverted HGHI inhibitory effects on 337 PPAR and FABP4 expression levels ( Figure 5K). Notably, TUDCA also blocked 338 the effects induced by BiP overexpression on both ERAD genes and adipogenic 339 markers ( Figure 5K).

342
We found that alternative splicing in the SC adipose tissue, and the ER stress-343 UPR-ERAD system in both SC and OM fat represent essential components of 344 the adipogenic process that are altered in obese individuals with IR/T2D.

352
Our results are also indicative of IR/T2D-dependent ERAD hyperactivation in 353 preadipocytes of the two major fat depots, SC and OM, in obese individuals.

354
Maladaptative UPR and ERAD signalling upon ER stress induction may 355 contribute to impair adipocyte differentiation and to the progression of metabolic 356 disease in obesity.

357
Alternative splicing of mRNA enables cells to acquire protein diversity, which 358 is essential for cell differentiation and tissue growth and identity ( Two different yet functionally interrelated pathways were impaired in PRPF8-392 silenced SGBS and hADSCs cells in terms of both total gene expression levels 393 and isoform balance of the proteins involved, i.e., adipogenic transcriptional 394 program, and LD biogenesis and growth ( Figure 3J)        and LD biogenesis (data not shown). 649 We also employed an adipose-derived stem cell line (hADSCs) that was  siRNA-treated cells were exposed to 100 mM insulin for 15 min. Red-O was dissolved in 100 mL 100% isopropanol. Before use, the solution was 815 diluted 6:4 with distilled water, allowed to stand for 10 min, and then filtered 816 through Whatman no. 1 paper. Cells were exposed to this solution for 30 min at 817 RT in darkness and the unbound dye was rinsed with distilled water. Images were 818 captured with an inverted light microscope coupled to a camera. For    Co-funded by European Regional Development Fund/European Social Fund "Investing in your future"); and Consejería de Economía, Conocimiento, Empresas y Universidad/Junta de Andalucía/FEDER (BIO-0139). CIBEROBN is an initiative of the ISCIII, Spain. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.     were used for C-G and I; and two-way ANOVA was used for H. Normality      Wilk normality test) were used for A-G; and two-way ANOVA was used for I.