Loss of HD-PTP function results in lipodystrophy, defective cellular signaling and altered lipid homeostasis

Abstract

His domain protein tyrosine phosphatase (HD-PTP; also known as PTPN23) facilitates function of the endosomal sorting complexes required for transport (ESCRTs) during multivesicular body (MVB) formation. To uncover its role in physiological homeostasis, embryonic lethality caused by a complete lack of HD-PTP was bypassed through generation of hypomorphic mice expressing reduced protein, resulting in animals that are viable into adulthood. These mice exhibited marked lipodystrophy and decreased receptor-mediated signaling within white adipose tissue (WAT), involving multiple prominent pathways including RAS/MAPK, phosphoinositide 3-kinase (PI3K)/AKT and receptor tyrosine kinases (RTKs), such as EGFR. EGFR signaling was dissected in vitro to assess the nature of defective signaling, revealing decreased trans-autophosphorylation and downstream effector activation, despite normal EGF binding. This corresponds to decreased plasma membrane cholesterol and increased lysosomal cholesterol, likely resulting from defective endosomal maturation necessary for cholesterol trafficking and homeostasis. The ESCRT components Vps4 and Hrs have previously been implicated in cholesterol homeostasis; thus, these findings expand knowledge on which ESCRT subunits are involved in cholesterol homeostasis and highlight a non-canonical role for HD-PTP in signal regulation and adipose tissue homeostasis.

Keywords: ESCRT; HD-PTP; Lipid homeostasis; PTPN23; Receptor signaling.

Fig. 1.

Ptpn23^H/H mice display lipodystrophy. (A) Schematic representation of the gene targeting strategy employed to generate hypomorph (Ptpn23^H), conditional flox (Ptpn23^fl) and knockout (Ptpn23⁻) alleles. A neo cassette between exons 6 and 7 serves as a moderate splice acceptor to reduce HD-PTP expression. Flp recombinase expression drives excision of the neo cassette via FRT sites to revert to a conditional flox. Cre recombinase expression results in excision of exons 5 and 6 through LoxP sites to generate a null allele through altering open reading frame. (B) Analysis of HD-PTP protein expression by western blotting in brain, brown adipose tissue (BAT), liver, lung, heart, muscle and skin from newborn Ptpn23^+/+ (n=6) and Ptpn23^H/H (n=3) pups. Analysis of HD-PTP expression in IAT from Ptpn23^+/+ (n=6) and Ptpn23^H/H (n=6) 14-day-old pups and Ptpn23^+/+ (n=6) and Ptpn23^H/H (n=6) adult male mice. *P<0.05; **P<0.01; ***P<0.005 (two-tailed unpaired t-test for equal n or Mann–Whitney for unequal n). (C) HD-PTP expression in the brain of Ptpn23^+/+ (n=8), Ptpn23^−/+ (n=6), Ptpn23^H/H (n=3), and Ptpn23^H/− (n=3) pups. *P<0.05; ***P<0.005; ns, not significant (Kruskal–Wallis followed by Mann–Whitney test). (D) Survival curves of female (n=118) and male (n=134) Ptpn23^H/H mice relative to Ptpn23^−/+ (n=34), Ptpn23^H/+ (n=63), Ptpn23^+/+ (n=38), and Ptpn23^fl/fl (n=58). P<0.0001 (log rank test). (E) Representative image of adult male Ptpn23^+/+ and Ptpn23^H/H mice. (F) Body mass of 10-day-old (n=11 female and n=10 male mice per genotype), 21-day-old (n=9 female and n=8 male mice per genotype), and adult (n=11 female and n=12 male mice per genotype) mice. **P<0.01; ****P<0.001; ns, not significant (two-way ANOVA corrected for multiple comparisons by Sidak's test). (G) Body fat composition measured by ECHO MRI of 10-day-old (n=11 female and n=10 male mice per genotype), 21-day-old (n=9 female and n=8 male mice per genotype), and adult (n=11 female and n=12 male mice per genotype) mice. **P<0.01; ****P<0.001; ns, not significant (two-way ANOVA corrected for multiple comparisons by Sidak's test). (H) Analysis of WAT mass relative to total body mass for age-matched adult male and female mice (6 weeks to 3 months; n=5 male and n=5 female mice per genotype). Adipose tissues examined: inguinal (IAT), mesenteric (MAT), perigonadal (GAT), perirenal (RAT), retroperitoneal (RPAT), subcapsular (SSAT) and brown (BAT). (I) Analysis of organ mass relative to total body mass for age-matched adult male and female mice (6 weeks to 3 months; n=5 male and n=5 female mice per genotype). Organs examined: liver, heart, lung, gastrocnemius. *P<0.05, **P<0.01, ***P<0.005, ****P<0.001; ns, not significant (Mann–Whitney test). (J) Quantification for adipocyte size (μm²) of Ptpn23^+/+ (n=5) and Ptpn23^H/H (n=3) male mice, *P<0.05 (Mann–Whitney test). (K) Representative images of IAT from adult Ptpn23^+/+ and Ptpn23^H/H male mice as used for quantification. Scale bars: 100 μm. All error bars presented as mean±s.e.m.

Fig. 2.

Ptpn23^H/H adipose tissue exhibits decreased insulin signaling. (A) Volcano plot of adult male IAT RPPA analysis (n=9 mice per genotype). 183 proteins display statistically significant alterations relative to Ptpn23^+/+. The dashed line represents the 0.05 P-value. (B) Heat map of significantly altered lipid metabolism factors from male IAT RPPA (two-tailed unpaired t-test). Factors elevated are colored red and factors decreased colored blue. (C) Pathway analysis of adult male IAT RPPA results indicates pathways significantly impacted in Ptpn23^H/H, colored blue for a decrease or red for an increase as assessed by z-score. (D) Heat map of significantly altered signaling components in adult male IAT RPPA (two-tailed unpaired t-test). Factors elevated colored red and factors decreased colored blue. (E) Representative western blot and quantification of IAT insulin receptor β protein expression. n=9 male mice per genotype. **P<0.01 (Welch's t-test). (F) Circulating insulin (mg/dl) in adult female (n=8–14 mice per genotype) and male (n=7 mice per genotype) mice. *P<0.05 (assessed within each sex by a Mann–Whitney test). (G) Quantification of levels of Akt phosphorylated T308 normalized to total Akt levels [p/t-Akt (T308)] following 100 ng/ml insulin stimulation of suspended IAT. Lysates generated at 1, 2 and 3 h post stimulation. n=3 experimental replicates. (+/+ insulin AUC 1303±204.9, H/H insulin AUC 955.3±107.3; mean±s.e.m.). (H) Quantification of levels of Akt phosphorylated T308 normalized to total Akt levels [p/t-Akt (T308)] following 100 ng/ml insulin stimulation of in vitro differentiated adipocytes. Lysates collected at 5, 10, 15, 30 and 60 min after stimulation (n=2). All error bars presented as mean±s.e.m.

Fig. 3.

EGFR activation is impacted at the level of receptor phosphorylation upon loss of HD-PTP. (A) iMEFs were serum starved 2 h and treated with 100 ng/ml EGF for 0 or 10 min. Cells were then stained for pERK1/2 (pERK) and assessed by flow cytometry as in the Materials and Methods. Each dot represents average of n=3 technical replicates, n=3 experimental replicates. *P<0.05 (two-tailed paired t-test). (B) Representative blots of pEGFR, EGFR, pERK and ERK following 0-, 5-, 10-, 15-, 30- and 45-min EGF stimulation. Blots were analyzed with FIJI software and the ratio of phospho/total (p/t) calculated for the proteins. (C) p/t-EGFR following EGF stimulation normalized to WT 0 min. Representative experimental replicate with n=3 technical replicates. (D) Following starvation as above, iMEFs were incubated on ice with an extracellular EGFR antibody and stained for flow cytometry as in the Materials and Methods. Histograms are an example of a single technical replicate, with ‘secondary only’ having no EGFR antibody. Each dot represents average of n=3 technical replicates, n=4 experimental replicates. ns, not significant (two-tailed paired t-test). (E) Following starvation, iMEFs were incubated with 100 ng/ml Alexa Fluor 488 (A488)–EGF on ice and fixed for assessment by flow cytometry. Saturated samples were pre-treated with 1000 ng/ml non-labelled EGF on ice. Histogram represents single technical replicate and bar graph shows fold increase in fluorescence from saturated to the A488-labelled samples. Each dot represents average of n=3 technical replicates, n=3 experimental replicates. ns, not significant (unpaired two-tailed t-test). All error bars presented as mean±s.e.m.

Fig. 4.

Cellular cholesterol is redistributed from the plasma membrane to intracellular compartments upon loss of HD-PTP. (A) iMEFs were incubated on ice with Alexa Fluor 488-conjugated CTxB and then assessed by flow cytometry as in the Materials and Methods. Histograms represent a single technical replicate, and depict gates used for Hi and Lo levels of plasma membrane CTxB staining. Quantification is given as percentage of cells within each gate. Each dot represents average of n=3 replicates. ***P<0.005 (multiple two-tailed unpaired t-tests). (B) iMEFs were incubated with Di-4-ANEPPDHQ for 45 min at 37°C, then immediately run on a flow cytometer. Shown is the Generalized Polarization (GP) value for membrane order calculated as in Waddington et al. (2019). Each dot represents average of n=3 technical replicates, n=4 experimental replicates. Significance assessed by paired t-test. (C) iMEFs were lysed in RPPA lysis buffer and samples tested for total cholesterol by Amplex Red cholesterol assay and normalized to lysate protein content. Each dot represents average of n=3 technical replicates, n=7 experimental replicates. *P<0.05 (unpaired two-tailed t-test). (D) Representative micrographs of iMEFs stained with Filipin (green) and LAMP1 (magenta). n=3 experimental replicates, with 20 images per replicate quantified in Fig. S4C,D. The line through a lysosome (in inset) was used to assess intensity of Filipin and LAMP1, with intensity values shown below the micrographs, in 4DI and 4DII. (E) iMEFs were lysed in ionic lysis buffer and run on Optiprep gradient. Each fraction was assessed by Amplex Red assay for total cholesterol and normalized to protein content. n=3 experimental replicates, denoted by distinct shapes. (F) Yeast containing an intracellular D4H–GFP probe were assessed for ergosterol distribution. Images processed in FIJI software for plasma membrane or intracellular ergosterol and quantified as percentage of cells within each group. n=398 bro1Δ and n=450 Wt, across n=4 experimental replicates. (G) Following overnight treatment with soluble cholesterol, iMEFs were serum starved for 2 h and incubated with MβCD during the final 30 min. Changes in pERK from 0–10 min EGF stimulation measured by flow cytometry. *P<0.05; **P<0.01; ****P<0.0001 (two-way ANOVA followed by Tukey test). (H) Representative micrographs of iMEFs incubated overnight with TopFluor-cholesterol (green) and stained for LAMP1 (magenta), with * denoting nucleus in zoom images. Cell outline (dashed white line) generated using phalloidin staining. n=60 micrographs across n=3 experimental replicates, quantified in Fig. S5C. All error bars presented as mean±s.e.m. Scale bars: 10 μm.

Fig. 5.

HD-PTP is necessary for localization of cholesterol transport proteins and lysosome–ER contact site formation. (A) Lysates were assessed by an Amplex Red cholesterol assay with or without a cholesterol esterase reagent. Esterified cholesterol was calculated by subtracting the without esterase value from the with esterase value. Each dot represents average of n=3 technical replicates, n=3 experimental replicates. *P<0.05 (two-tailed unpaired t-test). (B) Following plating on coverslips, iMEFs were loaded overnight with 500 μM oleate and then stained for lipid droplets utilizing Oil Red O staining (magenta), as described in the Materials and Methods. Shown are example micrographs. (C,D) Oil Red O micrographs quantified for size (C) and number of lipid droplets (D) in FIJI software. n=60 cells across n=3 experimental replicates, each experimental replicate matched by color, with experimental mean outlined in black. *P<0.05 (paired two-tailed t-test on experimental means). (E) iMEFs lysed with RPPA and shown as representative immunoblot of cholesterol transfer proteins ORP1, StARD3, NPC1, NPC2 and Actin. n=3 experimental replicates (quantified in Fig. S6B). (F) Example micrographs of cells processed for PLA (magenta) between Rab7 and VapA, as described in the Materials and Methods. (G) Quantification of PLA puncta/cell done in FIJI software. n=60 cells across n=3 experimental replicates, each experimental replicate matched by color, with experimental mean outlined in black. *P<0.05 (paired two-tailed t-test on experimental means). (H) Example micrographs of iMEFs plated on coverslips and processed for immunofluorescence, as described in the Materials and Methods, for NPC1 (green) and LAMP1 (magenta). (I) Quantification of NPC1 overlap with LAMP1 processed in FIJI software. Each dot represents a single cell z-stack, n=60 cells across n=3 experimental replicates, each experimental replicate matched by color, with experimental mean outlined in black. *P<0.05 (paired two-tailed t-test on experimental means). All error bars presented as mean±s.e.m. Scale bars: 10 μm (B,H); 5 μm (F).

Fig. 6.

Pharmacologic inhibition of Rab5 restores cellular phenotypes associated with loss of HD-PTP. Representative micrographs of WT and HD-PTP KO iMEFs stained for (A) ubiquitin and (B) Rab5 following the immunofluorescence protocol with digitonin permeabilization. Dashed line represents cell outline, generated using phalloidin staining. iMEFs in suspension, stained following same protocol as for the micrographs, were quantified by flow cytometry for (C) ubiquitin and (D) Rab5. Each dot represents average of n=3 technical replicates, n=3 experimental replicates. *P<0.05; **P<0.01 (two-tailed paired t-test). (E) Example micrographs of iMEFs stained by immunofluorescence for LAMP1 (magenta) and NPC1 (green) following overnight incubation with 40 μM NAP. Dashed line represents cell outline, generated using phalloidin staining. (F) Quantification of NPC1 overlap with LAMP1, done in FIJI software, following NAP incubation. n=60 cells across n=3 experimental replicates matched by color, with experimental mean outlined in black. A two-way ANOVA on experimental means revealed no significant differences. (G) iMEFs incubated with either 0 or 40 μM NAP overnight and serum starved for 2 h prior to treatment with 100 ng/ml EGF for 0 and 10 min. Cells then stained for pERK and assessed by flow cytometry as in the Materials and Methods. Bars represent fold increase in pERK from 0 to 10 min. Each dot represents average of n=3 technical replicates, n=3 experimental replicates. *P<0.05; **P<0.01; ***P<0.005; ****P<0.001; ns, not significant (two-way ANOVA followed by Tukey test). All error bars presented as mean±s.e.m. Scale bars: 10 μm.

Fig. 7.

Proposed model of HD-PTP in adipose tissue homeostasis. Ptpn23^H/H mice display lipodystrophy, defined as a decrease in mass of all WAT, and characterized by decreased RTK and insulin receptor signaling. In vitro, Ptpn23 KO iMEFs display decreased EGFR activation and an imbalance in cellular cholesterol distribution towards endosomal compartments. We propose this is due to a decrease in Lys–ER contact site formation and mislocalization of cholesterol transport proteins away from the lysosome, therefore limiting cholesterol transfer to the ER for redistribution through the cell. Accumulation of endosomal Rab5 may be upstream of the cholesterol transport defect, as inhibition of Rab5 by NAP rescues HD-PTP phenotypes. Figure created with BioRender.com.