The zinc transporter ZIP12 regulates the pulmonary vascular response to chronic hypoxia

Abstract

The typical response of the adult mammalian pulmonary circulation to a low oxygen environment is vasoconstriction and structural remodelling of pulmonary arterioles, leading to chronic elevation of pulmonary artery pressure (pulmonary hypertension) and right ventricular hypertrophy. Some mammals, however, exhibit genetic resistance to hypoxia-induced pulmonary hypertension. We used a congenic breeding program and comparative genomics to exploit this variation in the rat and identified the gene Slc39a12 as a major regulator of hypoxia-induced pulmonary vascular remodelling. Slc39a12 encodes the zinc transporter ZIP12. Here we report that ZIP12 expression is increased in many cell types, including endothelial, smooth muscle and interstitial cells, in the remodelled pulmonary arterioles of rats, cows and humans susceptible to hypoxia-induced pulmonary hypertension. We show that ZIP12 expression in pulmonary vascular smooth muscle cells is hypoxia dependent and that targeted inhibition of ZIP12 inhibits the rise in intracellular labile zinc in hypoxia-exposed pulmonary vascular smooth muscle cells and their proliferation in culture. We demonstrate that genetic disruption of ZIP12 expression attenuates the development of pulmonary hypertension in rats housed in a hypoxic atmosphere. This new and unexpected insight into the fundamental role of a zinc transporter in mammalian pulmonary vascular homeostasis suggests a new drug target for the pharmacological management of pulmonary hypertension.

Extended Data Figure 1. Generation of congenic and sub-congenic strains

Congenic rat lines were produced by introgression of the F344 chromosome 17 QTL segment onto the WKY genetic background by repeated backcrossing. Congenic rat strain R47A (WKY.F344-D17Got91/D17Rat51) contains 15Mbp from the F344 donor region that maps to the distal end of the QTL on a WKY background. Three sub-congenic strains, SubA (WKY.F344-D17Got91/D17Rat47), SubB (WKY.F344-D17Rat47/D17Rat51) and SubC (WKY.F344-D17Rat131/D17Rat51), were produced containing separate fragments of the R47A donor region by backcrossing of (R47A × WKY) F1 with WKY parental rats. Three recombination events within the R47A congenic interval break the congenic interval into three smaller and overlapping sub-congenic intervals (Figure 1, main text).

Extended Data Figure 2. Dissection of QTL and Cardiovascular phenotype of rat strains

a. The hypoxia-resistant F344 phenotype tracks with the congenic R47A line. Rats were kept in 10%O2 for 2 weeks and right ventricular hypertrophy (RV/LV+Sep) was significantly attenuated in the congenic R47A strain (0.32±0.03, n=13, **P<0.01) compared to WKY rats (0.37±0.03, n=15), whereas congenic R42 rats (0.36±0.03, n = 17) were similar (NS) to WKY rats. b. An illustrative genetic map showing the relationship of the congenic strains (R42, R47A), subcongenic strains (SubA, SubB, SubC) and Slc39a12 to the original QTL (defined by a LOD score >3; Zhao et al Circulation 2001,103, 442-447) on a physical map of chromosome 17 (using Rat Genome Assembly v5.0). In normoxia, WKY, F344, R47A, SubA, SubB and SubC rats show no significant differences in c. mean pulmonary artery pressure (mPAP), d. right ventricular hypertrophy (RV/LV+Septum ratio) and e. vascular muscularisation (n=8 each group); f. Systemic blood pressure (SBP) is similar in all strains in both normoxia and hypoxic conditions. g. F344, R47A and SubB rats exhibit attenuated pulmonary vascular remodelling after 2 weeks exposure to a 10% O2 atmosphere compared to WKY, SubA and SubC rats (n=6 each group). Values are expressed as the mean ± standard error of the mean (SEM). *P<0.05, **P<0.01, ***P<0.001 compared to WKY (% of fully muscularised and partially muscularised vessels); ##P<0.01, ###P<0.001 compared to WKY (% of non-muscularised vessels) after One-Way ANOVA analysis followed by Bonferroni correction for multiple testing.

Extended Data Figure 3. Hypoxia-induced pulmonary vascular remodeling in parental strains

a. Upper panel sequence shows the WKY protein sequence (688aa) and lower panel shows the truncated F344 protein sequence (553aa). Stars (*) mark the mutated amino acids compared to WKY protein. Dotted line indicates the C-terminal truncated region in F344. The grey square highlights the metalloprotease motif. b. Prominent ZIP12 immunostaining is seen in remodelled pulmonary arterioles in the chronically hypoxic WKY rat alongside vessels with a double elastic lamina (stained with Van Gieson, EVG) but not F344 lungs exposed to hypoxia. (Red arrow: vessel with double elastic lamina; blue arrow: vessel with single elastic lamina)

Extended Data Figure 4. ZIP12 upregulation in response to hypoxia exposure and measurements of intracellular labile zinc concentration and proliferation of human pulmonary artery smooth muscle cells (HPASMCs) in normoxia conditions

a. Upregulation of ZIP12 in HPASMCs exposed to hypoxia, in contrast to other zinc transporters (n=6). b. Representative western blots demonstrating increased HIF-2α expression in HPASMCs after 24h hypoxia exposure. c. Confocal laser scanning images of HPASMC transfected with eCALWY-4 probe. Intracellular free zinc was not affected by transfection with ZIP12 siRNA in normoxia. d. Representative traces showing the changes in fluorescence ratio usng the eCALWY-4 probe. e. Quantification of intracellular zinc levels (n=10). f. ZIP12 siRNA did not affect proliferation of HPASMCs in normoxic conditions (n=5).

Extended Data Figure 5. Design of specific Slc39a12 ZFN and confirmation of mutant line

a. CompoZr^™ Custom Zinc Finger Nucleases (Sigma-Aldrich) for the rat Slc39a12 gene were designed to target exon 8 (a; Sigma-Aldrich). b-d. Cel-I surveyor assay and gene sequencing confirmed NHEJ-induced mutations in at least one pup (mutant 77). e. The 4bp (AGTT) deletion followed by 2bp insertion (TA) into mutant 77 caused a frame-shift in coding, introducing a stop codon leading to a truncated protein. Red star refers to stop codon. c. We subsequently genotyped next generation litters using SwaI (cutting point: 5′-ATTTAAAT-3′), showing 100% digestion for homozygous pups (−/−), 50% for heterozygous (+/−) and no DNA digestion for wild type rats (+/+).

Extended Data Figure 6. ZIP12 knockout attenuated hypoxia-induced pulmonary vascular remodelling

a. Representative lung sections from wild-type (WT) and ZIP12 −/− rats 2 weeks after hypoxia exposure. Elastic van Gieson (EVG) staining showing double elastic lamina (red arrow) in WT but single elastic laminae (blue arrow) in ZIP12−/− rats. b. Genetic disruption of ZIP12 in WKY rat attenuated pulmonary vascular remodelling after 2 weeks exposure to a 10% O2 atmosphere compared to wild-type (WT) rats (n=5 each group). *P<0.01 compared to WT (% of fully muscularised vessels); ##P<0.01, ###P<0.001 compared to WT (% of non-muscularised vessels) after One-Way ANOVA analysis followed by Bonferroni’s multiple comparison test. c. Ki67 staining showing reduced proliferation in hypoxic ZIP12−/− rat lungs compared to the WT strain. *P<0.01 compared to WT. d. Representative sections from hypoxic WT and ZIP12−/− rats lungs showing differences in staining with the proliferation marker, Ki67. e-g. Genetic disruption of ZIP12 in WKY rat did not influence e. systemic blood pressure (SBP) or f. cardiac output (CO) but attenuated hypoxia-induced increases in g. pulmonary vascular resistance (n=7 each group). Values are expressed as the mean ± standard error of the mean (SEM). *P<0.05, **P<0.01 compared to normoxic rats, #p<0.05 compared to wild-type (WT) hypoxic rats after One-Way ANOVA analysis followed by Bonferroni correction for multiple testing. h. ZIP12 targeted siRNA inhibition attenuates stress fibre formation in human pulmonary vascular smooth muscle cells (HPASMCs) in hypoxia (n=5 each group). **p<0.01 compared to normoxia control group, #p<0.05 compared to hypoxia control group. i. Representative pictures of actin stress fibre in HPASMCs. j. ex vivo angiogenesis studies demonstrated that vascular outgrowth from ZIP12−/− pulmonary vessels in response to hypoxia was attenuated compared to vessels from wild-type (WT) rats (n=12 each group, 2 rings/rat, 6 ZIP12 −/− and 6 WT rats). *P<0.05 compared to normoxia WT group; #P<0.05, ## P<0.01 and ### P<0.001 compared to hypoxia ZIP12−/− group. k. Representative pictures of pulmonary arteriole ring outgrowth at day 6.

Extended Data Figure 7. Carbonic anhydrase (CAIX) expression

a. Representative sections demonstrating increased CAIX expression in remodelled pulmonary arterioles in the lungs of rats exposed to alveolar hypoxia (2 weeks), monocrotaline (MCT, 3 weeks) or a chronic iron deficient diet (4 weeks). b-c. No CAIX staining was detected in pulmonary arteries of low altitude (normoxia control, CO calf) calves and sea level humans, but prominent CAIX immunostaining was observed in the remodelled pulmonary arteries of calves with severe pulmonary hypertension (Hx calf), in cattle with naturally occurring pulmonary hypertension (“Brisket disease”, BD) as well as patients with idiopathic pulmonary arterial hypertension (IPAH).

Extended Data Figure 8. Genetic disruption of ZIP12 in WKY rat attenuated monocrotaline-induced pulmonary hypertension

a. Mean pulmonary artery pressure (mPAP), b. right ventricular hypertrophy (RV/LV+Septum) and c. pulmonary arteriole muscularisation. (n=5 each group). Values are expressed as the mean ± standard error of the mean (SEM). *P<0.05, **P<0.01 compared to wild-type (WT) monocrotaline group after unpaired Student t-test. d. Representative lung sections from wild-type (WT) and ZIP12 −/− rats 3 weeks after monocrotaline injection. Elastic van Gieson (EVG) staining showing double elastic lamina (red arrow) in WT but single elastic laminae (blue arrow) in ZIP12−/− rats.

Figure 1. The pulmonary vascular response to hypoxia in the F344 rat is influenced by a region of chromosome 17 containing Slc39a12

a. A genetic map of 3 sub-congenic strains (SubA, SubB and SubC) derived from the R47A congenic strain (originally derived from a WKYxF344 cross) backcrossed with the WKY parental strain. The refined congenic region (orange) of 8.28Mb containing 65 genes is within the SubB strain. b-d. SubB exhibits attenuated pulmonary hypertension after 2 weeks exposure to a 10% O₂ atmosphere compared to WKY, SubA and SubC rats: b. mean pulmonary artery pressure (mPAP); c. right ventricular hypertrophy (RV/LV+Septum ratio) (n=17 WKY, 15 F344, 14 R47A, 8 SubA, 10 SubB, 10 SubC); d. vascular muscularisation (n=6 each group). Dotted line indicates mean measurements from all the rats in a normal oxygen atmosphere (21%O₂; mPAP = 14.7±0.3 mmHg; RVH=0.270±0.004; % muscularization=34.2±0.36; for actual values in rat strains see Extended Data Fig. 3). Values are expressed as mean ± standard error of the mean (SEM). *P<0.05, **P<0.01, ***P<0.001 compared to WKY after one-way ANOVA analysis followed by Bonferroni correction for multiple testing. e. The genes of interest (Slc39a12, St8sia6, Cubn, Nmt2, Dclre1c, Hspa14 and Cdnf) identified within the SubB congenic interval. The frameshift mutation in Slc39a12 introduces a stop-codon, resulting in a truncated protein.

Figure 2. Slc39a12 encodes a zinc transporter, ZIP12, which is up-regulated in pulmonary vascular tissue from mammals exposed to chronic hypoxia

a. ZIP12 mRNA levels in control and hypoxic WKY rat lungs (n=6 each group). b. Prominent ZIP12 immunostaining in remodelled pulmonary arterioles in WKY but not F344 rat lungs exposed to hypoxia. c. No ZIP12 staining was detected in pulmonary arteries of low altitude (normoxia control, CO calf) calves and sea-level humans, yet prominent ZIP12 immunostaining was observed in the remodelled pulmonary arteries of calves with severe pulmonary hypertension (Hx calf), in cattle naturally susceptible to pulmonary hypertension at altitude (“Brisket disease”, BD), as well as Kyrgyz highlanders residing above 2500m. d. Design of the luciferase reporter vector pGL4.10 containing a 5′ region of ZIP12 which includes a hypoxia response element (HRE) encoding for both HIF-1α and HIF-2α binding motifs or a mutant HRE sequence where the 5′-ACGTG-3′ motif has been replaced by 5′-AGCAG-3′(mHRE). e. Human pulmonary artery smooth muscle cells (HPASMCs) transfected with the ZIP12 HRE reporter vector demonstrated a significantly increased luciferase activity after exposure to hypoxia, but not in the cells transfected with the mutant HRE vector (n=6 per group). f. Increased levels of HIF-1α and HIF-2α bound to the ZIP12 HRE assayed by ChIP-qPCR of chromatin from HPASMCs cultured in normoxia and hypoxic conditions (n=3 per group). Data are calculated as percentage of input levels, with the dotted line marking percentages below mock immunoprecipitation (IP). Values are expressed as mean ± SEM. *P<0.05, **P<0.01, ***P<0.001 compared to normoxic control after One-Way ANOVA analysis followed by Bonferroni correction for multiple testing. NS, not significant.

Figure 3. ZIP12 knockdown inhibits hypoxia-induced increase in intracellular labile zinc concentration and proliferation of human pulmonary artery smooth muscle cells (HPASMCs)

a. Representative wide-field microscope images of HPASMC transfected with eCALWY-4 probe. Hypoxia exposure produced a striking increase in intracellular free zinc (resulting in decreased FRET). This was inhibited by transfection with ZIP12 siRNA. TPEN-mediated Zn²⁺ chelation was used to derive maximum fluorescence and 100 μM ZnCl₂ in the presence of the Zn²⁺ ionophore and pyrithione (ZnPyr) was used to derive the minimum fluorescence. b. Representative traces showing the changes in fluorescence ratio of the eCALWY-4 probe. Steady-state fluorescence intensity ratio citrine/cerulean (R) was measured, then maximum and minimum ratios were determined to calculate free Zn2+ concentration using the formula: [Zn2+] = Kd×(Rmax-R)/(R-Rmin), where the Kd for eCALWY-4 is 630 pM, the maximum ratio (Rmax) was obtained upon intracellular zinc chelation with 50 μM TPEN and the minimum ratio (Rmin) was obtain upon zinc saturation with 100 μM ZnCl2 in the presence of the Zn2+ ionophore, pyrithione (5 μM). c. Quantification of intracellular zinc levels (n=10 each group). d. Chronic hypoxia (48h) increases ZIP12 mRNA levels in HPASMCs, which is inhibited by Slc39a12 siRNA (n=5 each group). e. Representative immunoblot of ZIP12 demonstrating inhibition of hypoxia-stimulated ZIP12 protein expression by Slc39a12 siRNA in HPASMCs (n=3). f. ZIP12 siRNA inhibits hypoxia-induced proliferation in HPASMCs (n=5 each group). *P<0.05, ***P<0.001 compared to control group, #P<0.05 compared to hypoxia group. Scr, scramble siRNA control.

Figure 4. Genetic disruption of ZIP12 in WKY rat attenuates hypoxia-induced pulmonary hypertension

a-c. Zinc finger nucleases were used to disrupt ZIP12 in the WKY strain. Rats deficient in ZIP12 demonstrate allele dose-dependent attenuation of hypoxia-induced pulmonary hypertension compared to wild-type (WT) rats: a. mean pulmonary artery pressure (mPAP); b. right ventricular hypertrophy (RV/LV+Septum) (normoxia groups: n=10 WT, 8 ZIP12+/−, 12 ZIP12−/−; hypoxia groups: n=14 WT, 16 ZIP12+/−, 12 ZIP12−/−); c. pulmonary arteriole muscularisation (n=5 each group). ***p<0.001 compared to normoxia WT group, #p<0.05 compared to hypoxia WT group after one-way ANOVA analysis followed by Bonferroni correction for multiple testing. d. ZIP12 was undetectable by Western blot in hypoxic ZIP12−/− rats but increased in hypoxic wide-type (WKY) rats (n=3 each group). e. ZIP12 expression by immunohistochemistry of WT and ZIP12−/− rat lungs before and after hypoxia (2 weeks). f. ZIP12 expression in lungs from a chronic iron deficient rat, monocrotaline (MCT) rat and a patient with idiopathic pulmonary arterial hypertension (IPAH). g. Double immunofluorescence demonstrates co-localisation of ZIP12 and smooth muscle actin in the remodelled vessels from the IPAH patient.