HILPDA Regulates Lipid Metabolism, Lipid Droplet Abundance, and Response to Microenvironmental Stress in Solid Tumors

Abstract

Accumulation of lipid droplets has been observed in an increasing range of tumors. However, the molecular determinants of this phenotype and the impact of the tumor microenvironment on lipid droplet dynamics are not well defined. The hypoxia-inducible and lipid droplet associated protein HILPDA is known to regulate lipid storage and physiologic responses to feeding conditions in mice, and was recently shown to promote hypoxic lipid droplet formation through inhibition of the rate-limiting lipase adipose triglyceride lipase (ATGL). Here, we identify fatty acid loading and nutrient deprivation-induced autophagy as stimuli of HILPDA-dependent lipid droplet growth. Using mouse embryonic fibroblasts and human tumor cells, we found that genetic ablation of HILPDA compromised hypoxia-fatty acid- and starvation-induced lipid droplet formation and triglyceride storage. Nutrient deprivation upregulated HILPDA protein posttranscriptionally by a mechanism requiring autophagic flux and lipid droplet turnover, independent of HIF1 transactivation. Mechanistically, loss of HILPDA led to elevated lipolysis, which could be corrected by inhibition of ATGL. Lipidomic analysis revealed not only quantitative but also qualitative differences in the glycerolipid and phospholipid profile of HILPDA wild-type and knockout cells, indicating additional HILPDA functions affecting lipid metabolism. Deletion studies of HILPDA mutants identified the N-terminal hydrophobic domain as sufficient for targeting to lipid droplets and restoration of triglyceride storage. In vivo, HILPDA-ablated cells showed decreased intratumoral triglyceride levels and impaired xenograft tumor growth associated with elevated levels of apoptosis. IMPLICATIONS: Tumor microenvironmental stresses induce changes in lipid droplet dynamics via HILPDA. Regulation of triglyceride hydrolysis is crucial for cell homeostasis and tumor growth.

Figure 1:

HILPDA dependent lipid droplet formation. A) Hilpda WT and KO MEFs were kept in complete media (CM), incubated in 1% O₂ for 72h or supplemented with 60μM oleate/linoleate/BSA complexes for 24h (FA), and lipid droplets were stained with Nile Red. B) HILPDA protein levels in WT and KO MEFs treated as in A. C) Quantification of LDs per cell in WT and KO MEFs in complete media or supplemented with FA for 24h. D) LD diameter distribution in WT and KO MEFs. Floating bars: minimum to maximum, vertical line at the median. E) Restoration of HILPDA expression in KO MEFs. Cells were transfected with either pLenti-Hilpda-C-Myc-DDK-IRES-neo (KO+Hilpda) or empty vector (KO) and stable transfectants were selected by G418 resistance. Black arrow: endogenous HILPDA, gray arrow: HILPDA-myc-Flag. Asterisk: non specific. F) Nile Red staining of KO MEFs and pool of reintroduced clones following treatment with oleate/linoleate for 24h. G) Genetic manipulations in HCT116 cells. Knockout lines were generated by the nickase CRISPR technology and four independent KO clones after puromycin resistance selection were pooled together. H) Quantification of LD per cell in complete media and after FA loading. I) TLC of hexane/isopropanol extracted lipids separated in cyclohexane:diethyl acetate:acetic acid and stained with primuline. Lipid class standards: CE cholesteryl esters, TAG: triglycerols, FA: fatty acid. ****p<0.0001, N.S. not significant by one way ANOVA and Sidak post hoc test.

Figure 2:

Lipid metabolism in HCT116 WT and HILPDA KO. Cells were kept in complete media (21%), incubated in 1% O₂ for 72h (1%) or supplemented with 60μM oleate/linolate/BSA (FA) for 24h. (A-F) Lipid extracts were analyzed by LC/MS and isobaric species measured against internal standards. (G-I) For phospholipid measurement the samples were spiked with internal standards. (J-K) Transcript levels of ELOVL1 and ELOLV7, respectively, measured by qRT-PCR after normoxic (21%) and hypoxic (1%) incubation. L) Western blotting detection of lipid metabolic proteins in HILPDA WT and KO HCT116. M) PGE2 secretion under the same conditions as in K. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, N.S. not significant by two-way ANOVA and Holm-Sidak test, N=3. Error bars: S.E.M.

Figure 3:

Decreased starvation-induced lipid droplet accumulation in the absence of HILPDA. A) HILPDA WT and KO MEFs and HCT116 were incubated in complete media (CM) or EBSS for 24h. Fixed cells were stained with LipidTox and Hoechst and imaged by confocal fluorescent microscopy. Scale bar: 10μm. B) HILPDA protein levels after similar treatement. C) Protein levels of autophagic flux markers LC3 and p62 in full media and EBSS in the absence or presence of 20 μM chloroquine (CQ) for 24h. Relative LC3II/LC3I ratios and p62 protein levels were calculated from 3 independent experiments. Top row: average values, bottom row: S.E.M. D) LD quantification in HILPDA WT and KO MEFs after 24h of EBSS treatment −/+ 20μM chloroquine (CQ) E) Quantification of glycerolipids after 24h in complete media (DMEM) or EBSS in MEF (left panel) and HCT116 (right panel), N=5–6. ***p<0.001, ****p<0.0001, N.S. not significant by one way ANOVA and Holm-Sidak multiple comparisons test, error bars: S.E.M.

Figure 4:

Post-transcriptional control of HILPDA upregulation. A) Relative levels of the HIF targets HILPDA and PDHK1 by RT-qPCR in MEF and HCT116. Samples were treated for 24h with: normoxia (21%), hypoxia (1%), loaded with 60μM oleate/linoleate/BSA (FA) or incubated in Earle’s balanced salt solution (EBSS). Shown are results from a representative of 2 independent experiments in 3 technical replicates. B) HILPDA protein levels after 24h treatments in Hif1ɑ WT and KO MEF. C,D) LipidTox staining of lipid droplets in MEF following the indicated incubations for either 72h (hypoxia), or 24h (all normoxic stresses). E) Impact of lipid turnover inhibitors on HILPDA protein in HCT116 cells. Samples were kept either in complete media (CM) or starved in EBSS in the absence or presence of: 20μM chloroquine (CQ), 20μM OOEPC, 5μM Triacsin C or 10μM A922500 for 24h.

Figure 5:

HILPDA truncation mutants restore lipid storage. A) Diagram of truncation mutations. B) Expression of truncation mutants in fatty-acid loaded cell lysates. Stable cell lines were generated by transfection of myc-Flag-HILPDA transgenes or empty vector (Neo) and selected by antibiotic resistance. Note that only the full lenght peptide is recognized by the rabbit anti-HILPDA antibody raised against the C-terminus. C) Lipid droplet localization of mutants in the HILPDA KO HCT116 background. The numbers refer to aminoacid length, with H63 being the full length peptide. Samples were loaded with 60 μM oleate/linoleate/BSA for 24h, fixed and stained with mouse anti-myc followed by anti-mouse-Alexa594 antibodies. Lipid droplets were stained with LipidTox and nuclei with Hoechst. Samples imaged with an Olympus FV3000 confocal microscope. Bar: 10μm. D) Triglyceride levels in the various mutants. *p<0.05, **p<0.01, ***p<0.001, N.S. not significant by 2way ANOVA and Dunnet test. N=3–6, error bars: S.E.M.

Figure 6:

ATGL inhibition restores lipid droplets in the absence of HILPDA. A) Glycerol release by MEFs incubated in EBSS for 24h. B) Glycerol release by MEFs after 24h starvation in the absence or presence of 25μM ATGListatin. C) Visualization of lipid droplets by LipidTox staining and fluorescence microscopy in samples treated as in B. D) LD quantification in MEFs treated as previously E) Size distribution of LDs in similarly treated MEFs. Boxes: min to max, line at the median. F) Accumulation of triglycerides in the presence of 25μM ATGListatin compared to identical stress conditions in the absence of inhibitor. Complete media (CM), starvation (EBSS) and 60μM oleate/linoleate/BSA loading (FA) for 24h. *p<0.05, **p<0.01, ***p<0.001, N.S. not significant by two way ANOVA and Holm-Sidak test. N=3–4 (A,B,F), error bars: S.E.M.

Figure 7:

HILPDA promotes tumor growth in vivo. A) In vitro cell growth of HCT116 WT and HILPDA KO cells grown in normoxia (21% O2) or hypoxia (1% O2). B) Volumes of HCT116 WT and KO subcutaneous tumors in female nude mice (WT N=8, KO N=12). C) Quantification of triglycerides in tumor explants after 3 weeks of in vivo growth (N=7). D) Staining of tumor cryosections with Nile red (top) and immunofluorescent detection of the proliferation marker Ki67 and apoptosis marker cleaved Caspase 3. E) Quantification of cleaved caspase 3 positive cells on HILPDA WT and KO tumor sections. (WT N=8 KO N=10, each point is the average of two non-consecutive tumor sections) *p<0.05, **p<0.01, in B by two way ANOVA and Sidak post-hoc test, in C, E by Student’s t-test, error bars: S.E.M.