Thyroidal expression of ER molecular chaperone GRP170 is required for efficient TSH-mediated thyroid hormone synthesis

Abstract

Intracellular trafficking of secretory and membrane proteins from the endoplasmic reticulum (ER) to the cell surface, via the secretory pathway, is crucial to the differentiated function of epithelial tissues. In the thyroid gland, a prerequisite for such trafficking is proper protein folding in the ER, assisted by an array of ER molecular chaperones. One of the most abundant of these chaperones, Glucose-Regulated-Protein-170 (GRP170, encoded by Hyou1), is a noncanonical hsp70-like family member. Thyroid follicular epithelial cells abundantly express GRP170, but the role of this abundant ER chaperone in thyrocytes remains unknown. Here, we have examined the effect of inducible Pax8-specific (thyroid and kidney) deficiency of GRP170 in mice, in parallel with siRNA-treated PCCL3 (rat) thyrocytes for knockdown of GRP170. Thyrocyte-specific loss of GRP170 in vivo triggers primary hypothyroidism with a deficient thyroidal response to Thyroid-Stimulating Hormone (TSH). In addition, knockdown of GRP170 in PCCL3 thyrocytes inhibits the folding and forward trafficking of TSH receptors to the cell surface. Taken together, our findings suggest that GRP170 contributes to the conformational maturation of TSH receptors and thyroid gland responsiveness to TSH, which is required for proper regulation of thyroid hormone synthesis.

Keywords: Cell biology; Endocrinology; Protein traffic; Thyroid disease.

Figure 1. GRP170 is expressed in the ER of thyrocytes.

(A) Immunofluorescence of control mouse thyroid costained for GRP170 (green), KDEL-containing proteins (red), DAPI (blue), and green/red merge (n = 5 animals examined). (B) Immunofluorescence of thyroid from mice with inducible Grp170 deletion, immunostained as in A (n = 5 animals examined). (C) PCCL3 (rat thyrocyte) cells transfected (as a control) with scrambled oligo, examined by immunofluorescence for the same antigens, with presentation as those shown in A. (D) PCCL3 cells transfected with siRNA for knockdown of Grp170, examined by immunofluorescence as in B. Both C and D were performed in 3 independent transfection experiments. Scale bars: 20 μm (A and B), 10 μm (C and D).

Figure 2. Genetic deletion of GRP170 in mouse thyroid follicular epithelial cells.

Mouse thyroid glands were harvested at 21 days after initiation of doxycycline treatment to induce Pax8-rtTA/LC1-Cre–mediated homozygous Hyou1^fl/fl deletion between introns 1 and 24 (25). (A) Thyroidal Grp170 mRNA levels measured by qPCR (normalized to Tbp; n = 4–5 animals per group; squares are males, circles are females; data are shown as mean ± SD, t test, *P < 0.05). (B) Immunoblotting of thyroid tissue lysates from the indicated genotypes with anti-GRP170 (actin is a loading control). (C) Quantitation of GRP170 protein levels from the indicated genotypes (n = 6–7 animals per group; squares are males, circles are females; data are shown as mean ± SD, t test, *P < 0.05).

Figure 3. Thyroid tissue response to induced thyrocyte-specific loss of GRP170.

(A) H&E staining of mouse thyroid tissue sections (n = 6–8 mice per group). Size bars: 20 μm. (B) Total thyroid gland size from the indicated genotypes (n = 3–5 per group; squares are males; data are shown as mean ± SD, t test). (C) Thyroidal BiP, spliced Xbp1, and Chop mRNA levels measured by qPCR from lysates of mice with thyroidal GRP170 deficiency, normalized to that of control animals (n = 5 per group; squares are males, circles are females; data are shown as mean ± SD, ANOVA). (D) Thyroidal BiP protein levels measured by immunoblotting in the indicated genotypes (n = 3–4 mice per group; actin is a loading control). (E) Immunoblotting of Tg protein from the indicated genotypes (n = 4 mice per group) after digestion with endoglycosidase H (+) or mock-digest (–). Endoglycosidase H sensitive and resistant bands are indicated.

Figure 4. Primary hypothyroidism in mice with deficiency of GRP170 in thyrocyes.

(A) Circulating TSH levels from control mice and those with induced deficiency of GRP170 at 21 days after initiation of doxycycline treatment (n = 5 animals per group; squares are males, circles are females). (B) Circulating levels of thyroxine (T₄) and triiodothyronine (T₃) from control mice and those with induced deficiency of GRP170 at 21 days after initiation of doxycycline treatment (n = 4–6 animals per group; squares are males, circles are females). (C) Immunoblotting of thyroid tissue lysates from the indicated genotypes with anti-TPO (actin is a loading control). (D) Quantitation of TPO protein levels from in lysates from mice with thyroidal GRP170 deficiency, normalized to that of control animals (n = 6–7 animals per group; squares are males, circles are females). (E) Immunoblotting of thyroidal TG protein levels from the indicated genotypes (these and the loading control actin come from the same samples as those shown in Figure 2B). (F) Quantitation of thyroidal TG protein levels from the indicated genotypes (n = 6–7 animals per group; squares are males, circles are females). (G) Immunoblotting of T₄ content of thyroidal TG protein from the indicated genotypes (immunoblotting from the same samples with anti-TG in lower panel). (H) Quantitation of T₄ content of TG protein in lysates from mice with thyroidal GRP170 deficiency, normalized to that of control animals (n = 4–6 animals per group; squares are males). Data are shown as mean ± SD, t test, *P < 0.05, ***P < 0.001.

Figure 5. Deficiency of TSH receptors and NIS in the thyroid glands of mice with thyrocyte deficiency of GRP170.

(A) Immunoblotting of NIS in thyroid tissue lysates from the inducible GRP170 deletion or controls (littermates lacking either Pax8-rtTA or LC-1). Animals of all genotypes were treated identically with doxycycline. Actin is a loading control. (B) Quantitation of either total NIS protein levels or only the upper NIS protein band in lysates from mice with thyroidal GRP170 deficiency, normalized to that of control animals (n = 6–7 animals per group; squares are males, circles are females; data are shown as mean ± SD, ANOVA, ***P < 0.001). (C) Thyroidal Nis, Tpo, Tg, Duox1, Duox2, Dio1, and Dio2 mRNA levels (normalized to Tbp; n = 5 per group; squares are males, circles are females; data are shown as mean ± SD, ANOVA, *P < 0.05) were measured by qPCR from lysates of mice with thyroidal GRP170 deficiency, normalized to that of control animals. (D) Immunoblotting of thyroidal TSH receptor protein levels from animals like those described in A. Actin is a loading control. (E) Quantitation of thyroidal TSH receptor protein levels from animals in D (n = 5–7 animals per group; squares are males, circles are females; data are shown as mean ± SD, t test, *P < 0.05).

Figure 6. Phenotype of thyrocyte-specific deficiency of GRP170 in PCCL3 cells.

(A) Immunoblotting of GRP170 and TG content of cells (last 6 lanes) and medium collected for 6 hours (first 6 lanes) at 48 hours after siRNA-mediated knockdown of GRP170, or scrambled oligo control. Actin is a loading control. (B) Quantitation of GRP170 protein levels and TG secretion from the indicated samples (n = 3 independent transfections in each group; data are shown as mean ± SD, t test, **P < 0.01). (C) Immunoblotting of NIS protein at 48 hours after siRNA-mediated knockdown of GRP170 (n = 3 transfections per group). Actin is a loading control. (D) Digestion of NIS protein with endoglycosidase H, or PNGase F in the samples as indicated (n = 4 transfections per group). (E) PCCL3 cells were transfected either with empty vector (–) or to express TSHR-GFP as indicated; after 24 hours, the cells were transfected with siRNA for knockdown of GRP170 or scrambled oligo control. At 48 hours after knockdown, the cells were lysed, digested with endoglycosidase H or PNGase F or were mock-digested, analyzed by 3%–8% Tris-Acetate NuPAGE, electrotransferred to nitrocellulose, and immunoblotted with anti-GFP antibodies. Endo H–sensitive forms are indicated with up-sloped red arrows; endo H–resistant forms are indicated with down-sloped green arrows. (F) Quantitation of the fraction of endo H–resistant TSHR-GFP from the indicated samples (n = 4 independent transfections in each group; data are shown as mean ± SD, t test, *P < 0.05). (G) Quantitation of surface-biotinylated TSHR-GFP in PCCL3 cells from the experiments shown in Supplemental Figure 4 (data are shown as mean ± SD, t test, *P < 0.05).

Figure 7. Trafficking of GFP-tagged TSH receptors by live cell imaging in PCCL3 cells.

PCCL3 cells were transfected to express TSHR-GFP and, 24 hours later, transfected with siRNA for knockdown of GRP170 or scrambled oligo control exactly as in Figure 6E. (A) At 48 hours after the second transfection, TSHR-GFP distribution was imaged in live control PCCL3 cells. (B) At 48 hours after the second transfection, TSHR-GFP distribution was imaged in live PCCL3 cells deficient for GRP170. The results in A and B were repeated and confirmed (3 independent transfection experiments).