Rare pLOF variants in SLC39A5 are associated with elevated serum zinc and nominal protection against type II diabetes (T2D).

(A) Serum zinc in carriers of SLC39A5 pLOF variants in the discovery cohort. Controls (Ref; SLC39A5+/+) and heterozygous carriers of pLOF variant alleles in SLC39A5 (Het; SLC39A5+/-). Subject numbers: Ref and Het respectively: n=5317 and n=15. (B) Trans-ancestry meta-analysis of association of SLC39A5 pLOF variants with T2D. (C-I) Serum zinc and insulin profile of age, sex and BMI-matched controls in serum call back study. Subject numbers: Ref and Het respectively: n=246-253 and n=86-91, **P < 0.01, unpaired t-test. Numeric data is summarized in Suppl. Table 1.

Loss of Slc39a5 results in elevated circulating and hepatic zinc levels in mice.

Serum zinc (A) and hepatic zinc (B) in Slc39a5+/+, Slc39a5-/- and Slc39a5+/- mice at 40 weeks of age, n=16-18. **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 improves glycemic traits in leptin-receptor deficient mice and in mice challenged with high fat high fructose diet (HFFD).

Female (A-D, I-L; ♀) and Male (E-H, M-P; ♂) mice. (A-H) Slc39a5-/-;Lepr-/- and corresponding control mice. (A, E) Body weight at 34 weeks. (B, F) Fasting blood glucose at 34 weeks. (C, G) Fasting insulin at 34 weeks. (D, H) Homeostatic model assessment for insulin resistance (HOMA-IR) at 34 weeks. Slc39a5+/+ and Slc39a5-/- (n=5-12), Lepr -/- and Slc39a5 -/-;Lepr -/- (n=10-15). *P < 0.05, **P < 0.01, ***P < 0.001, one-way ANOVA with post hoc Tukey’s test. (I-P) Slc39a5-/- and Slc39a5+/+ mice were fed HFFD or NC for 30 weeks. (I, M) Body weight at 30 weeks. (J, N) Fasting blood glucose at 30 weeks. (K, O) Fasting insulin at 30 weeks. (L, P) HOMA-IR at 30 weeks, n=11-15. *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test. Numeric data is summarized in Suppl. Table 4 and 5.

Loss of Slc39a5 improves liver function and steatosis in leptin-receptor deficient female mice and in female mice challenged with high fat high fructose diet (HFFD).

Slc39a5-/-;Lepr-/- and corresponding control mice (A-F) were sacrificed after 16 hour fasting at 34 weeks of age. (G-L) Slc39a5-/- and Slc39a5+/+ mice were fed HFFD or NC for 30 weeks and sacrificed after 16 hours fasting. (A, G) Representative images of livers stained with H&E. Scale bar, 200µm. (B, H) Hepatic triglyceride (TG) content in explanted liver samples at endpoint. (C, I) Serum ALT. (D, J) Serum AST. (E, K) NAFLD activity score, (F, L) Hepatic beta-hydroxybutyrate (BHOB). *P < 0.05, **P < 0.01, ***P < 0.001, Slc39a5-/-;Lepr-/-and corresponding control mice: one-way ANOVA with post hoc Tukey’s test, HFFD or NC: two-way ANOVA with post hoc Tukey’s test. Numeric data is summarized in Suppl. Table 4 and 5.

Loss of Slc39a5 results in elevated hepatic zinc and activation of hepatic AMPK signaling in leptin-receptor deficient female mice and female mice challenged with high fat high fructose diet (HFFD).

Analyses were done on explanted liver samples collected after 16 hours fasting at endpoint in Lepr-/- (A-C) and HFFD mice (D-F). (A, D) Immunoblot analysis of hepatic AMPK and AKT activation. AMPK and AKT signaling is activated in Lepr-/-;Slc39a5-/- mice and HFFD Slc39a5-/- mice (compared to their Scl39a5+/+ counterparts). (B, E) Hepatic zinc is elevated in Lepr-/-;Slc39a5-/- mice and HFFD Slc39a5-/- mice (n=10-21). (C, F) Elevated hepatic zinc results in increased Mt1 (zinc responsive gene) expression in both models. (G) Immunoblot analysis of primary human hepatocytes treated with zinc chloride (ZnCl2), magnesium chloride (MgCl2), okadaic acid (OA), metformin (Met) for 4 hours. Zinc activated AMPK and AKT signaling in primary human hepatocytes. (H) Densitometric analysis of immunoblots (compared to control). *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 improves hepatic inflammation and fibrosis in female mice challenged with diet-induced NASH.

Slc39a5-/- and Slc39a5+/+ mice were placed on a NASH inducing diet or NC for 40 weeks and sacrificed after 16 hours of fasting. (A, B) NASH Slc39a5-/- mice display reduced serum ALT and AST levels. (C-E) Histology scores for steatosis, hepatocyte hypertrophy, inflammation. (F) NAFLD activity score was reduced in NASH Slc39a5-/- mice. (G-I) NASH Slc39a5-/- mice display reduced fibrosis. (G) Representative images of explanted livers sample stained with picrosirius red indicative of collagen deposition. Scale bar, 300µm. (H, I) Fibrosis and steatosis-activity-fibrosis scores. n=6-7 (NC) and 8-11 (NASH), *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test. Numeric data is summarized in Suppl. Table 6.

SLC39A5 pLOF variants p.Y47*(c.141C>G), p.R311*(c.931C>T), and p.R322*(c.964C>T) encode for non-functional proteins.

HEK293 cell transfected with expression constructs encoding SLC39A5 wild-type (WT), Y47*, R311* and R322* variants. (A, B) Immunostaining and FACS analysis demonstrating WT SLC39A5 localization to the cell surface. (C) Overexpression of WT SLC39A5 results in Zn2+ mediated MRE activation in a dose dependent manner, n=4. (D) FACS analyses demonstrating that cell surface expression of SLC39A5 Y47*, R311* and R322* muteins is markedly reduced. (E) Variants Y47*, R311* and R322* did not mediate Zn2+ induction of MRE, n=8, Statistical comparison to Mock and WT, respectively: ***P < 0.001, ###P < 0.001, two-way ANOVA with post hoc Tukey’s test. Metal regulatory element (MRE), firefly luciferase (Fluc), renilla luciferase (Rluc), cytomegalovirus (CMV), gene of interest (GOI), internal ribosome entry site (IRES).

Generation and characterization of the Slc39a5-/- mice.

(A) Schematic representation of the Slc39a5 null allele. (B) Slc39a5 gene expression in liver and duodenum of Slc39a5-/- mice at 20 weeks of age, n=3-6. (C) Immunoblotting analyses demonstrating absence of SLC39A5 protein in liver of Slc39a5-/- mice at 34 weeks of age. (D-E) Heterozygous loss of Slc39a5 does not reduce fasting blood glucose in (D) Lepr-/- mice (at 20 weeks of age) and in (E) mice challenged with high fat high fructose diet (HFFD) for 18 weeks. *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 improves glycemic traits in Lepr-/- mice and in mice challenged with high fat high fructose diet (HFFD).

Female (A-D, I-L; ♀) and Male (E-H, M-P; ♂) mice. (A-H) Slc39a5-/-; Lepr-/- and corresponding control mice. (A-B, E-F) Oral glucose tolerance test (GTT) after 16 hour fasting, at 20 weeks. (C-D, G-H) Intraperitoneal insulin tolerance test (ITT), at 33 weeks. n=5-8. *P < 0.05, **P < 0.01, ***P < 0.001, one-way ANOVA with post hoc Tukey’s test. (I-P) HFFD mice. (I-J, M-N) GTT after 16 hours fasting, at 18 weeks. (K-L. O-P) ITT after 4 hour fasting, at 27 weeks. n=11-12. *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test. Area under curve (AUC). Numeric data is summarized in Suppl. Table 4 and 5.

Metabolic profiling of female Slc39a5-/-; Lepr-/- mice.

(A) Longitudinal body weight. (B-D) Analyses were done on explanted liver samples collected after 16 hour fasting at 34 weeks of age. (B) Hepatic expression of fatty acid synthase (Fasn) and glucose-6-phosphatase (G6pc). (C) Hepatic FASN and G6PC protein) Serum insulin profile upon fasting at 34 weeks of levels. (D) Densitometric analysis of hepatic FASN and G6PC. (F) Serum insulin profile in fed state at 32 weeks of age. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 reduces hepatic fatty acid synthase expression but does not change insulin profile of male Lepr-/- mice.

(A) Longitudinal body weight. (B-D) Analyses were done on explanted liver samples collected after 16 hour fasting at 34 weeks of age. (B) Hepatic expression of Fasn and G6pc. (C) Hepatic FASN and G6PC protein levels. (D) Densitometric analysis of hepatic FASN and G6PC. (E) Serum insulin profile upon fasting at 34 weeks of age. (F) Serum insulin profile in fed state at 32 weeks of age. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 reduces hepatic fatty acid synthase expression but does not change insulin profile in female mice challenged with high fat high fructose diet (HFFD).

(A) Longitudinal body weight during dietary intervention. (B-D) Analyses were done on explanted liver samples collected after 16 hour fasting in mice fed HFFD or NC for 30 weeks. (B) Hepatic expression of Fasn and G6pc. (C) Hepatic FASN and G6PC protein levels. (D) Densitometric analysis of hepatic FASN and G6PC. (E) Fasting serum insulin profile. (F) Fed serum insulin profile. (G) Representative images of Slc39a5+/- livers stained with H&E. Scale bar, 100um. (H) NAFLD activity score. *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 reduces hepatic fatty acid synthase expression but does not change insulin profile in male mice challenged with high fat high fructose diet (HFFD).

(A) Longitudinal body weight during dietary intervention. (B-D) Analyses were done on explanted liver samples collected after 16 hour fasting in mice fed HFFD or NC for 30 weeks. (B) Hepatic expression of Fasn and G6pc. (C) Hepatic FASN and G6PC protein levels. (D) Densitometric analysis of hepatic FASN and G6PC. (E) Fasting serum insulin profile. (F) Fed serum insulin profile. (G) Representative images of Slc39a5+/- livers stained with H&E. Scale bar, 100um. (H) NAFLD activity score. *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test.

Additional data of glucose stimulated insulin secretion, serum BHOB and pancreas histology in mouse models.

(A-H) Oral glucose tolerance test (GTT) was performed in Slc39a5-/- female (A-D) and male (E-H) mice after 16 hour fasting, at 15 weeks. Glucose (A-B, E-F) and insulin (C-D, G-H) levels were measured at 0, 15 and 60 mins. n=6-8. (I-J) Serum BHOB levels in female (I) and male (J) mice challenged with high fat high fructose diet (HFFD). *P < 0.05, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test. (K-L) Serum BHOB levels in Slc39a5-/-; Lepr-/- female (K) and male (L) mice. *P < 0.05, ***P < 0.001, one-way ANOVA with post hoc Tukey’s test. (M-N) No overt morphological deficits in pancreas resulting from Slc39a5 deficiency. Scale bar, 300µm

Loss of Slc39a5 improves liver function and steatosis in Lepr-/- male mice and reduces hepatic triglyceride in male mice challenged with high fat high fructose diet (HFFD).

Slc39a5-/-; Lepr-/- and corresponding control mice (A-F) were sacrificed after 16 hours fasting at 34 weeks of age. (G-L) Slc39a5-/- and corresponding control mice were fed HFFD or NC for 30 weeks and sacrificed after 16 hours fasting. (A, G) Representative images of livers stained with H&E. Scale bar, 200µm. (B, H) Hepatic triglyceride (TG) content in explanted liver samples at endpoint. (C, I) Serum ALT. (D, J) Serum AST. (E, K) NAFLD activity score, (F, L) Hepatic beta-hydroxybutyrate (BHOB). *P < 0.05, **P < 0.01, ***P < 0.001, Slc39a5-/-; Lepr-/- mice: one-way ANOVA with post hoc Tukey’s test, HFFD: two-way ANOVA with post hoc Tukey’s test. Numeric data is summarized in Suppl. Table 4 and 5.

Loss of Slc39a5 results in elevated hepatic zinc and activation of hepatic AMPK signaling in congenital and diet-induced obesity models.

Analyses were done on explanted liver samples collected from male mice after 16 hour fasting at endpoint of congenital (A-C) and diet-induced obesity (D-F). (A, D) Immunoblot analysis of hepatic AMPK and AKT activation. (B, E) Hepatic zinc measurements (n=10-21). (C, F) Hepatic Mt1 gene expression. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 does not alter hepatic magnesium, iron, copper, calcium and cobalt levels in Lepr-/- mice.

Female (A-C) and male (D-F) mice were examined at 34 weeks of age. (A, D) Densitometric analysis of hepatic AMPK and AKT signaling. (B, E) Hepatic expression of Slc39a5 and Mt2. (C, F) Hepatic ion quantification by flame atomic absorption spectrometry. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 does not alter hepatic magnesium, iron, copper, calcium and cobalt levels in mice challenged with high fat high fructose diet (HFFD).

Female (A-C) and male (D-F) mice were fed HFFD or NC for 30 weeks. (A, D) Densitometric analysis of hepatic AMPK and AKT. (B, E) Hepatic gene expression of Slc39a5 and Mt2. (C, F) Hepatic ion quantification by flame atomic absorption spectrometry. *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test.

Zinc activates AMPK and AKT signaling in time-dependent and dose-dependent manner.

(A) No differences in cell viability observed in human primary hepatocytes (after 4hr treatment) across different experimental groups. (B) Time-resolved (0-4hr) immunoblotting analyses of primary human hepatocytes treated with zinc chloride. (C) Immunoblots of HepG2 treated with zinc chloride (ZnCl2) and magnesium chloride (MgCl2). (D) Time-resolved (0-4hr) immunoblotting analyses of HepG2 treated with zinc chloride. Okadaic acid (OA), metformin (Met), N,N,N’,N’-Tetrakis(2-pyridylmethyl)ethylenediamine (TPEN).

Elevated hepatic zinc results in reduced protein phosphatase activity.

Analyses were done on explanted liver samples collected after 16 hour fast at endpoint of congenital obesity (A-B) and diet-induced obesity (C-D) challenges. Female (A, C) and Male (B, D) mice. (A-D) Ser/Thr and Tyr protein phosphatase activity. (E) Ser/Thr and Tyr protein phosphatase activity in primary human hepatocytes treated with zinc chloride (ZnCl2), magnesium chloride (MgCl2) and N,N,N’,N’-Tetrakis(2-pyridylmethyl)ethylenediamine (TPEN) for 4 hours. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA with post hoc Tukey’s test.

Loss of Slc39a5 reduces hepatic inflammation and fibrosis in male mice challenged with diet-induced NASH.

Mice were fed NASH diet or NC for 40 weeks and sacrificed after 16 hour fasting. (A-B) Loss of Slc39a5 reduces serum ALT and AST levels (biomarkers of liver damage). (C-E) Histology scores for steatosis, hepatocyte hypertrophy, inflammation. (F) NAFLD activity score. (G-I) Loss of Slc39a5 improves fibrosis in mice upon NASH dietary challenge. (G) Representative images of explanted livers sample stained with picrosirius red indicative of collagen deposition. Scale bar, 300µm. (H-I) Histology score for fibrosis and steatosis-activity-fibrosis score. n=7-10 (NC) and 15-17 (NASH), *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test. Numeric data is summarized in suppl. Table 6.

Loss of Slc39a5 improves liver function in mice challenged with diet-induced NASH.

Female (A-E) and Male (F-J) mice. (A, F) Body weight. (B, G) Fasting blood glucose. (C, H) Serum zinc. (D, I) Hepatic zinc. (E, J) Total hepatic SOD activity. n=6-10 (NC) and 8-17 (NASH), *P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with post hoc Tukey’s test.

Serum zinc and insulin profile assessment in the serum call back study.

Serum zinc levels in SLC39A5 heterozygous loss of function carriers are elevated by 12% as compared to age, sex, BMl-matched reference controls. Analyses of insulin production (insulin/c-peptide ratio), insulin clearance (proinsulin/insulin) and blood glucose in these samples demonstrated no differences based on genotype. Data represented in a graphical format in Fig. 1.

Tissue zinc content in human and mouse.

Human data adapted from Jackson et. al4. *p<0.05; **p<0.01, ***p<0.001, not significant (n.s.), upaired t-test. Values represent mean ± SD.

No differences in serum chemistry profile of Slc39a5+I+ and Slc39a5-/- mice.

Serum chemistry analysis in adult mice (40 weeks of age, both sexes) demonstrated no differences in pancreatic amylase, renal function parameters (blood urea nitrogen, creatinine, total protein and uric acid) and electrolytes (chloride, potassium and sodium) or liver enzymes (alanine aminotransferase; ALT and aspartate aminotransferase; AST).

Summary statistics for the congenital obesity model.

Loss of Slc39a5 improves glycemic traits and liver function in leptin-receptor (Lepr) deficient mice. Loss of Slc39a5 does not change insulin production (proinsulin/insulin), insulin clearance (insulin/c-peptide ratio). Data represented in a graphical format in Fig. 3, 4 and Suppl. Fig 5, 6.

Summary statistics for the diet-induced obesity model.

Loss of Slc39a5 improves glycemic traits and liver function in mice upon a high fat high fructose diet (HFFD) dietary challenge. Moreover, loss of Slc39a5 does not change insulin production (proinsulin/insulin), insulin clearance (insulin/c-peptide ratio). Data represented in a graphical format in Fig. 3, 4 and Suppl. Fig 7, 8.

Summary statistics for the diet-induced NASH model.

Loss of Slc39a5 improves hepatic inflammation and fibrosis in both female and male mice challenged with diet-induced NASH. Data represented in a graphical format in Fig. 6 and Suppl. Fig 14.