Hyperglycemia-Induced Changes in ZIP7 and ZnT7 Expression Cause Zn²⁺ Release From the Sarco(endo)plasmic Reticulum and Mediate ER Stress in the Heart

Diabetes is an important risk factor for cardiovascular dysfunction through defective Ca²⁺ signaling (1,2). However, we have previously shown that Zn²⁺ release during the cardiac cycle results in cytosolic free Zn²⁺ increase (3), further triggering higher prooxidant species production that leads to oxidative damage (4). Conversely, oxidant exposure induces marked cytosolic free Zn²⁺ increases in cardiomyocytes (5), whereas hyperglycemia causes oxidative stress and increased cytosolic Zn²⁺ through underlying cardiac dysfunction (6). Similar to Ca²⁺, Zn²⁺ is essential for cellular functions in mammalian heart (7), serving as an important secondary messenger (8).

Excess Zn²⁺ can be detrimental to cells, particularly that of cardiomyocytes (3,5–7). Cytosolic Zn²⁺ plays an important role in excitation-contraction coupling in cardiomyocytes by shaping Ca²⁺ dynamics (3,4,9). Its level in cardiomyocytes is calculated as <1 nmol/L, but it is approximately fivefold higher in sarco(endo)plasmic reticulum [S(E)R] and less than the cytosolic level in mitochondria, as demonstrated previously with Förster resonance energy transfer (FRET) recombinant-targeted Zn²⁺ probes (10). Elevated cytosolic Zn²⁺ appears to contribute to deleterious changes in many cellular signaling pathways, including hyperglycemia-challenged cardiomyocytes (3–6,11).

Cellular Zn²⁺ fluxes are achieved and controlled by Zn²⁺ transporters (ZnTs) and Zn²⁺ importers (ZIPs) (12,13), although their distributions and functions are not yet well clarified in cardiomyocytes. The Zn²⁺-selective ion channel ZIP7 plays an important role in releasing Zn²⁺ from the S(E)R and Zn²⁺-associated induction of unfolded protein response in yeast (14) and is localized to the Golgi apparatus in CHO cells, allowing Zn²⁺ release from Golgi lumen into cytosol (15). ZIP7 facilitates the release of Zn²⁺ from the ER (16) and behaves as a critical component in subcellular redistribution of Zn²⁺ in other systems (17). In addition, protein kinase 2 (CK2) has been hypothesized to trigger cytosolic Zn²⁺ signaling pathways by phosphorylating ZIP7 (18), although some studies have also highlighted its important contribution to Zn²⁺ homeostasis under pathological conditions (19–21).

Although studies have shown the presence of weakly expressed ZIP7 and ZnT7 in mammalian heart (22,23), their subcellular localizations and functional roles are not yet known well. In that regard, previous studies have suggested that either an increase or an inhibition of one of them might contribute to cellular dysfunction (11), including defective insulin-mediated signaling pathways under hyperglycemia (24) or altered insulin secretion (25,26).

ER stress is one of underlying mechanisms of cardiac dysfunction, including diabetic cardiomyopathy (27–29). Studies suggest that a relationship exists between S(E)R function and cytosolic Zn²⁺ level in diabetic (DM) rat cardiomyocytes (29,30). In particular, the latter studies identified a close association among oxidative stress, cytosolic Zn²⁺ increase, ER stress, and cardiac dysfunction in diabetes. However, experimental evidence suggests a requirement for Zn²⁺ for proper ER function, with Zn²⁺ deficiency leading to ER stress (14,31) from decreased Zn²⁺ in the ER as a result of hypoxia or hypoglycemia (14). However, no clear data demonstrate which ZnTs play roles in controlling cytosolic Zn²⁺ increases in cardiomyocytes during hyperglycemia.

Thus, hypothesizing that disruption of ZnTs and the Zn²⁺ axis contributes to deleterious changes in diabetic cardiomyocytes is tempting. We aimed to first clarify their subcellular localizations and then to explore their functional roles in Zn²⁺ homeostasis. In addition, we tested their roles in cytosolic Zn²⁺ redistribution and development of ER stress under hyperglycemic conditions as a result of CK2α activation.

This study was approved by the Ankara University (Ankara, Turkey) ethics committee (115-449). Type 1 diabetes was mimicked by single injection of streptozotocin (50 mg/kg i.p.; Sigma-Aldrich) in 3-month-old male Wistar rats (n = 15), whereas control rats (n = 10) were injected with vehicle as previously described (2). After streptozotocin injection (7 days), rats with threefold higher blood glucose levels compared with preinjection levels were used as the DM group (n = 13 kept for 12 weeks).

Cardiomyocytes were isolated from the left ventricle by using the enzymatic method as previously described (5). Hearts were cannulated on a Langendorff apparatus, leaving preperfusion through the coronary artery with a Ca²⁺-free solution. Preperfusion was followed with 1 mg/mL of collagenase (Type 2; Worthington Biochemical Corporation, Lakewood, NJ)-containing solution for 30–35 min. Only Ca²⁺-tolerant rod-shaped cells were used to ensure cell viability and excitability as well to avoid SR Ca²⁺ overload and terminal contracture.

The embryonic rat heart–derived H9c2 cell line was purchased from ATCC (Manassas, VA) and was cultured in DMEM as previously described (10).

Cytosolic and S(E)R free Zn²⁺ levels ([Zn²⁺]_Cyt and [Zn²⁺]_ER, respectively) in H9c2 cells were measured with eCALWY sensors (Cyt-eCALWY4 and ER-eCALWY6, respectively) delivered with plasmids expressing Cyt-eCALWY4 and ER-eCALWY6. Measurements were performed as previously described (10). Images were captured at the 433-nm monochromatic excitation wavelength, and image analysis was performed with ImageJ software by using a custom-made macro. To calculate free Zn²⁺, maximum (R_max by using a heavy metal chelator, TPEN [N,N,N′,N′-tetrakis(2-pyridylmethyl) ethylenediamine] 50 μmol/L; Sigma-Aldrich) and minimum (R_min by using Zn²⁺ saturation with 100 μmol/L ZnCl₂ and Zn²⁺ ionophore pyrithione (Zn²⁺/Pyr 5 μmol/L) fluorescence ratios were used as previously described (13).

ZIP7 and ZnT7 localizations were determined by using anti-ZIP7 and anti-ZnT7 antibodies (PA5-21072 [Thermo Fisher Scientific] and sc-160946 [Santa Cruz] at 1:50, respectively) in confocal microscopy (LSM510; Zeiss). S(E)R localization was determined by transfection of H9c2 cells with ER-resident Discosoma species red fluorescent protein (DsRed-ER) (red) plasmid for 24 h. After performing general procedures, cells were incubated with specific antibodies to monitor their localizations. After overnight incubation, cells were incubated with appropriate secondary antibodies in the presence of BSA (5%; Alexa Fluor 488 donkey anti-rabbit and donkey anti-goat 1:1,000 for ZIP7 and ZnT7, respectively).

The Golgi apparatus was labeled with an anti-GM130 antibody cis-Golgi marker (ab52649 1:750; Abcam). Cells were then incubated either with anti-ZIP7 or with anti-ZnT7 and Golgi marker GM130 antibody for ZIP7 and ZnT7 (secondary antibodies: Alexa Fluor 488 donkey anti-rabbit 1:1,000 for ZIP7, Alexa Fluor 568 goat anti-mouse 1:1,000 for GM130, Alexa Fluor 488 donkey anti-goat 1:1,000 for ZnT7, and Alexa Fluor 647 rabbit anti-mouse 1:1,000 for GM130). Finally, cells were mounted with medium containing DAPI (blue). Images were deconvolved and analyzed for their colocalizations with use of Huygens software (https://svi.nl/HuygensProfessional) and processed with ImageJ software.

Left-ventricular S(E)R fractionation of hearts was performed by using an ER Isolation Kit (E0100; Sigma). Briefly, hearts were homogenized in isotonic extraction buffer and then centrifuged. Crude microsomal fraction was isolated from postmitochondrial fraction by using ultracentrifugation. For further purification and separation for rough ER and smooth ER, a self-generating density gradient procedure was performed according to the manufacturer’s instructions. Western blot analysis was carried out by using primary antibodies against ZIP7 and ZnT7. To confirm S(E)R-isolation, SERCA2 (sc-8094; Santa Cruz), Golgi 58K Protein (ab27043; Abcam), and cyclin E (sc-481; Santa Cruz) were used as S(E)R, Golgi, or nuclear markers, respectively.

H9c2 cells were stably transfected with four unique 29-mer short hairpin RNA (shRNA) constructs in lentiviral RFP (red fluorescent protein) vector (SR02938179A, B, C, and D; Origene) and noneffective 29-mer scrambled shRNA (sc-shRNA) cassette in pRFP-CB-shLenti vector as control (TR30033; Origene). Twenty-four hours before transfection, cells were seeded into separate dishes. Every shRNA construct with packaging and enveloping plasmids (psPAX [Addgene 12260] and pMD2G [Addgene 12259]) were mixed into a solution containing CaCl₂ (375 mmol/L) and then incubated for 30 min. DNA mixtures were added into HEPES buffer saline containing (in mmol/L) 12 dextrose, 50 HEPES, 10 KCl, 280 NaCl, and 1.5 Na₂HPO₄H₂O; every mixture was added into separate 15-cm dishes and incubated for 12 h. Viral supernatants were harvested after 24 h and 48 h, and cells were then infected with each lentivirus produced from every shRNA construct, including scrambled sequences at 3 plaque-forming units/cell (incubation with 10 μg/mL blasticidin for antibiotic selection). They were seeded into six-well plates and harvested for knockdown efficiency to measure ZIP7 mRNA levels (Supplementary Table 1). A mixture of four ZIP7 gene-specific shRNA constructs was used.

Total RNA was prepared by using an RNA Isolation Kit (740955.10; Macherey-Nagel), and purified total RNA was reverse transcribed with a ProtoScript First Strand cDNA Synthesis Kit (E6300S; New England Biolabs). First-strand cDNAs were quantified with GoTaq qPCR Master Mix (A6001; Promega). The amplified fragment size of PCR products for each primer and for all primers’ specificity were controlled for through the National Center for Biotechnology Information and Ensembl databases. Primer sequences for cyclophilin, ZIP7, and ZnT7 are listed in Supplementary Table 2. The fold changes in the genes were analyzed by comparative (2^−ΔΔCt) method.

The lysates were extracted with NP-40 lysis buffer (250 mmol/L NaCl, 1% NP-40, and 50 mmol/L Tris-HCl; pH 8.0 and 1 × protease inhibitor cocktail) from homogenized samples. Protein concentration of supernatants (centrifuged at 12,000g for 5 min at 4°C) was measured with a BCA Assay Kit (Pierce). Equal amounts of protein were separated on 12% SDS-PAGE Tris-Glycine or 4–12% NuPAGE Bis-Tris Protein Gels (Life Technologies). The membranes were probed with antibodies against GRP78 (sc-13968 1:200; Santa Cruz), Calregulin (CALR) (sc-11398 1:200; Santa Cruz), ZIP7 (sc-83858 1:200; Santa Cruz), ZnT7 (sc-160946 1:200; Santa Cruz), SERCA2 (sc-8094 1:200; Santa Cruz), Golgi 58K Protein (ab27043 1 μg/mL; Abcam), cyclin E (sc-481; Santa Cruz), GAPDH (sc-365062 1:1,000; Santa Cruz), pZIP7 (a mouse monoclonal antibody to pS275S276-ZIP7, which recognizes ZIP7 only when phosphorylated on residues at S275 and S276, from K.M.T.) and β-actin (sc-47778 1:500; Santa Cruz) in BSA, PBS, and Tween 20 solution. Specific bands were visualized with horseradish peroxidase–conjugated compatible secondary antibodies (anti-mouse 1:2,000, anti-goat 1:7,500, anti-rabbit 1:7,500) and detected by ImmunoCruz Western Blotting Luminol Reagent (sc-2048; Santa Cruz). The band densities were analyzed with ImageJ software.

Cells were treated as indicated and lysed in coimmunoprecipitation buffer containing (in mmol/L) 50 Tris, 100 NaCl, 1 EDTA, 1% Triton-X 100, and 10% glycerol (pH 7.4); protease inhibitors (1 mmol/L phenylmethylsulfonyl fluoride and 1 μg/mL each of leupeptin, aprotinin, and pepstatin); and phosphatase inhibitors (10 mmol/L sodium-fluoride and 1 mmol/L sodium orthovanadate) for 30 min at 4°C. The 600-μg protein lysates from aliquots (1 mL lysis buffer) were precleared through incubation with 30 μL Protein A/G Sepharose (Sigma) for 1 h at 4°C. The precleared samples were incubated with specific primary antibody (anti-CK2α 10 μg/mL, sc-6480; Santa Cruz) in lysis buffer for 2 h at 4°C and then 30 μL Protein A/G beads were added. Samples were then incubated overnight at 4°C, and beads were washed five times with lysis buffer, boiled, and separated by 10% SDS-PAGE.

CK2α (1 and 2) was silenced in H9c2 cells by using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s small interfering RNA (siRNA) transfection protocol. Briefly, cells were seeded into six-well plates and cultured for 48 or 72 h with a mixture of 25 nmol/L CK2α1 and CKα2 or nontargeted siRNAs (ON-TARGETplus SMARTpool Csnk2a1 and Csnk2a2-siRNA and ON-TARGETplus Non-targeting Control Pool, L-096197-02-0005, L-092756-02-0005, and D-001810-10-05, respectively; Dharmacon) with serum-free medium that included 5 μL/mL Lipofectamine. After incubation, the cells were extracted into buffers to examine protein and mRNA expression levels. The primer sequences designed for CK2α1 and CKα2 are listed in Supplementary Table 2.

Data are presented as mean ± SEM unless otherwise stated. Differences were determined using Student t tests with Bonferroni correction for multiple comparisons as required. GraphPad Prism 6.0 statistical software was used for the analyses. P < 0.05 was considered significant.

Protein and mRNA levels of ZIP7 (at 50 kDa) and ZnT7 (at 42 kDa) in isolated rat ventricular cardiomyocytes are shown in Fig. 1A and B. Both mRNA and protein levels of ZIP7 were significantly increased in DM rat cardiomyocytes, with significantly decreased ZnT7 levels. To test these changes that directly arose through hyperglycemia, we incubated isolated cardiomyocytes under high glucose (33 mmol/L) conditions for 3 h and then compared ZIP7 and ZnT7 with that of mannitol (23 mmol/L)-incubated cells. Similar changes were found in mRNAs to those of DM rat cells, except for their protein levels (∼10%) (Supplementary Fig. 1), which was probably due to shorter high-glucose exposure compared with diabetes.

For further validation, we used high-glucose–incubated (25 mmol/L for 24 h) H9c2 cells and measured protein and mRNA levels. As can be seen in Fig. 1C and D, both levels of ZIP7 increased significantly in high-glucose–incubated cells, whereas the ZnT7 mRNA level (Fig. 1D) was markedly (∼75%) decreased with a relatively small decrease (∼25%) in its protein level.

Next, we performed experiments to determine whether the expression of other ZnTs, such as ZIP1, ZIP6, ZnT1, and ZnT5, were affected in hyperglycemic cardiomyocytes. By using quantitative RT-PCR, we demonstrated that mRNA levels of these transporters were not significantly altered in high-glucose–incubated versus control cells (25 mmol/L for 24 h) (Supplementary Fig. 2A and B).

Since the use of a specific antibody developed by one of us (K.M.T.) in breast cancer cells (18), we examined the possible dependency of ZIP7 phosphorylation on hyperglycemia in both DM rat cardiomyocytes and high-glucose–incubated H9c2 cells (25 mmol/L for 24 h or 48 h). As shown in Fig. 1E, phospho-ZIP7 (pZIP7) levels measured at about 50 kDa in DM rat cells was about sevenfold higher compared with controls. Furthermore, the pZIP7 level in high-glucose–incubated H9c2 cells was higher compared with nonincubated cells in a time dependent manner (Fig. 1F). The pZIP7/ZIP7 ratio was similar in high-glucose–incubated H9c2 cells and higher for DM rat heart (0.96 ± 0.04 vs. 4.79 ± 0.39, respectively).

ZIP7 and ZnT7 have been reported in mammalian heart (22,23) and proposed to localize to the ER (17,21,25) and/or perinuclear vesicles associated with the Golgi apparatus (22,24) in cells but not in cardiomyocytes. Furthermore, ZIP6 and ZIP7 have been shown to colocalize to the ER regulating cytosolic Zn²⁺ homeostasis in dispersed pancreatic islet cells (19). In the current study, we postulated a similar endogenous subcellular localization for ZIP7 and ZnT7 to S(E)R in H9c2 cells. Therefore, after transfection with DsRed-ER (research design and methods), ZIP7 or ZnT7 localizations to S(E)R were visualized. Individual and merged confocal images are shown in Fig. 2A and B. The merged visual inspections show the presence of the majority of ZIP7 and ZnT7 in the S(E)R. For colocalization, Pearson correlation coefficients were calculated for ZIP7 and ZnT7 associated with S(E)R by using Huygens software (0.67 ± 0.07 [n = 3] and 0.72 ± 0.13 [n = 3], respectively).

To test whether ZIP7 and ZnT7 localizes to the Golgi apparatus, we used a combination of antibodies for either GM130 and ZIP7 or GM130 and ZnT7. Images are shown in Fig. 2C and D. The merged visual inspections show the presence of the minority of either ZIP7 or ZnT7 to the Golgi. For colocalization, Pearson correlation coefficients for ZIP7 and ZnT7 associated with the Golgi were 0.29 ± 0.02 (n = 3) and 0.51 ± 0.06 (n = 3), respectively.

To test whether hyperglycemia affects the location of ZIP7 and ZnT7 in cells, high-glucose–treated (25 mmol/L for 24 h) cells transfected with DsRed-ER were visualized. Confocal images are shown in Supplementary Fig. 2A–D. For colocalization of ZIP7 and ZnT7 in the S(E)R, Pearson correlation coefficients were calculated as 0.42 ± 0.09 (n = 3) and 0.54 ± 0.08 (n = 3). For colocalization of ZIP7 and ZnT7 in the Golgi apparatus, Pearson correlation coefficients were 0.11 ± 0.02 (n = 3) and 0.70 ± 0.08 (n = 3).

For a comparison of the effects of hyperglycemia on subcellular localizations of ZIP7 and ZnT7, we compared the calculated Pearson correlation coefficients in normal and hyperglycemic cells. Except for ZIP7 localization to the Golgi, localizations of other transporters was unaffected by hyperglycemia (Supplementary Fig. 3E).

The S(E)R localization of ZIP7 and ZnT7 was further investigated by Western blot in isolated S(E)R fractions (range 1–4) obtained from rat heart. The intensities of protein bands associated with either ZIP7 or ZnT7 showed both a gradually increasing strength from rough to smooth fractions of isolated S(E)R preparation (Fig. 2E). These data strongly demonstrate the presence of these two transporters in the smooth part of S(E)R. For further validation of their presence in S(E)R fractions, we assessed SERCA2a (110 kDa) as a positive control and Golgi marker (Golgi 58K Protein) and a nuclear marker cyclin E (53 kDa) as negative controls in the same S(E)R preparations (Fig. 2E). The last two markers were clearly observed in total cell lysate as well.

By using FRET Zn²⁺ probes (10,25), we measured free Zn²⁺ levels in cytosol and S(E)R in H9c2 cells after culture at normal (5.5 mmol/L) or high (25 mmol/L) glucose for 24 h. Representative free Zn²⁺ measurements (research design and methods) are shown in Fig. 3A and B (for cytosol and S(E)R, respectively). By using quantification from fluorescence intensity (10), Cyt-eCALWY4–infected and high-glucose–incubated H9c2 cells showed significantly high cytosolic free Zn²⁺ levels compared with controls (1.74 ± 0.15 vs. 0.97 ± 0.14 nmol/L), whereas S(E)R free Zn²⁺ levels compared with controls ranged from 4.88 ± 1.00 to 2.52 ± 0.36 nmol/L.

Of note, the total cellular free Zn²⁺ level in high-glucose–treated cells was lower than in control cells. Several reasons are possible for these differences. ZnTs located on sarcolemma might be activated by hyperglycemia. Alternatively, high cytosolic Zn²⁺ levels might result Zn²⁺ increases in other organelles, such as mitochondria, through the kinetic activation of uptake pathways (10).

Because the demonstration of ZIP7 and ZnT7 localizations to the S(E)R provides further evidence of the role of the S(E)R as an intracellular Zn²⁺ pool in cardiomyocytes (3), with high mRNA and protein levels of ZIP7 and high ZIP7 phosphorylation (rather than ZnT7) under hyperglycemia, we assessed the direct role of S(E)R in the redistribution of cellular free Zn²⁺ in hyperglycemic cardiomyocytes. We first silenced ZIP7 in H9c2 cells. The mRNA levels of ZIP7 and ZnT7 in ZIP7-silenced cells compared with sc-shRNA cells are shown in Fig. 4A and B (left). The ZIP7 mRNA level was reduced (>60%) compared with controls, whereas the ZnT7 mRNA level was not changed. As expected, ZIP7 immunoreactivity was reduced by >90% in silenced cells with no change in ZnT7 protein levels (Fig. 4A and B, right).

To test the role of ZIP7 in subcellular Zn²⁺ distribution further, we measured cytosolic and S(E)R free Zn²⁺ levels in nontargeted and ZIP7-targeted shRNA-infected cells by using previously mentioned protocols (research design and methods). As shown in Fig. 4C (left), cytosolic free Zn²⁺ was increased more than threefold in hyperglycemic cells compared with control cells (0.92 ± 0.14 vs. 3.28 ± 0.65 nmol/L; n = 36–45 cells). Furthermore, the cytosolic free Zn²⁺ in ZIP7-silenced cells under hyperglycemic conditions was not different from that of ZIP7-silenced cells under physiological conditions (Fig. 4C, right). However, the S(E)R free Zn²⁺ in nontargeted shRNA-infected hyperglycemic cells had trended less (36%) (Fig. 4D, left). Furthermore, the free Zn²⁺ level in S(E)R in ZIP7-silenced cells under hyperglycemic conditions was not different from that of ZIP7-silenced cells under physiological conditions (Fig. 4D, right). These results are consistent with an important role for S(E)R-localized ZIP7 in the redistribution of subcellular Zn²⁺ in cardiomyocytes under hyperglycemia.

Previously, marked ER stress and increased intracellular free Zn²⁺ in isolated left-ventricular cardiomyocytes from DM rats have been shown (4,29,30) and demonstrated an ER Zn²⁺ deficiency–associated disruption in its function and induction of ER stress in eukaryotes (14). Therefore, we hypothesized that similar changes in cardiomyocytes would lead to ER stress under hyperglycemia. To test this hypothesis, we first measured ER stress chaperone levels of GRP78 and CALR in high-glucose­–incubated cells (25 mmol/L for 24 h) compared with those of control cells. The GRP78 and CALR levels were increased significantly in hyperglycemic cells similar to those DM rat heart tissue (32) (Fig. 5A). We also examined GRP78 induction in response to a direct ER stress activator tunicamycin (TUN) (10 μmol/L for 18 h) and again showed increased GRP78 expression levels (Fig. 5B). To demonstrate a hypothesis related to hyperglycemia-associated changes in protein expression levels of these transporters can underlie the induction of ER stress in cardiomyocytes, we used direct ER stress–induced cardiomyocytes with TUN incubation and measured ZIP7 and ZnT7 expression levels, which were not significantly different from those of controls (Fig. 5C and D).

Because we observed a markedly higher ZIP7 phosphorylation level in DM rat cardiomyocytes (Fig. 1E) and high-glucose–incubated cells (Fig. 1F) versus their respective controls, we aimed to clarify whether these changes arose as a direct consequence of ER stress. To address this prediction, we treated cells with 10 μmol/L TUN for 24 h to induce ER stress directly and then measured pZIP7 levels. No significant differences were found between treated and untreated cells, implying no direct ER stress effect on ZIP7 phosphorylation (Fig. 5E). For further support associated with ZIP7-induced ER stress, we monitored expression levels of GRP78 and CALR in ZIP7-silenced cells under hyperglycemic conditions and found no difference from those of respective control conditions (Fig. 5F and G).

For further support of the association between activation of ZIP7-related increased cytosolic Zn²⁺ under hyperglycemia and ER stress induction, we increased cytosolic free Zn²⁺ through Zn²⁺/Pyr (100 nmol/L for 24 h). We then measured markedly increased GRP78 and CALR levels while these levels could be reversed to control levels with the ER stress inhibitor tauroursodeoxycholic acid (50 μmol/L for 24 h) (Supplementary Fig. 5A and B). The above results thus provide further support for the view that hyperglycemia-associated changes in pZIP7 and ZIP7 levels underlie an induction of ER stress.

CK2α is an unusual protein kinase, being constitutively active, and undergoes autophosphorylation (33). Both CK2α activity and expression are increased with low Zn²⁺ levels in a concentration-dependent manner (34). Because CK2α activation has been shown previously to prompt cytosolic Zn²⁺ signaling (18) and high-glucose (0–25 mmol/L) treatment (for 4 h) of cells induces significant CK2α activity in a concentration-dependent manner (35), we first examined CK2α expression in isolated cardiomyocytes from DM rat heart and then in Zn²⁺/Pyr-incubated (1 μmol/L for 20 min) rat cardiomyocytes compared with those of controls. Both CK2α expressions were significantly higher than in their respective controls (Fig. 6A). Furthermore, we repeated similar measurements in H9c2 cells maintained at high glucose (25 mmol/L for 24 h) or Zn²⁺/Pyr (1 μmol/L for 20 min). High-glucose incubation for 24 h induced a significant increase in CK2α expression (twofold), whereas an eightfold increase was observed in CK2α expression in the Zn²⁺/Pyr-incubated cells for 20 min (Fig. 6B). These data demonstrate that CK2α expression increases in response to hyperglycemia or elevated cytosolic free Zn²⁺ and might contribute in a feed-forward cycle to further Zn²⁺ release into cytoplasm. On the basis of data shown in Fig. 1F, the importance of time dependency of hyperglycemia-associated elevated cytosolic free Zn²⁺ effect on pZIP7 expression is emphasized.

Because of the association among CK2α activation/expression, ZIP7 phosphorylation, and Zn²⁺ influx into the cytoplasm from the ER previously reported in tamoxifen-resistant breast cancer cells (18), we performed coimmunoprecipitation measurements in H9c2 cells (Fig. 6C). These data reveal an association between CK2α activation (at most as a result of hyperglycemia) and increased cytosolic free Zn²⁺-dependent phosphorylation and activation of ZIP7 in response to hyperglycemia.

To test whether ZIP7 may be phosphorylated by other kinases, which also are activated by increased intracellular Zn²⁺ in cardiomyocytes (4,30,36), we performed further coimmunoprecipitation experiments by using various antibodies against them. We did not observe coimmunoprecipitation between ZIP7 and any of the other kinases.

To validate the essentiality of CK2α activation for ZIP7 phosphorylation, we silenced CK2α in normal and high-glucose–treated (25 mmol/L for 24 h) H9c2 cells and then measured pZIP7 and ZIP7 expression levels. CK2α protein level is markedly reduced in CK2α-silenced cells (Fig. 6D), whereas the mRNA levels of CK2α1 and CK2α2 are decreased (Supplementary Fig. 6). In addition, no ZIP7 phosphorylation was observed in CK2α-silenced cells under control or hyperglycemic conditions, whereas a sevenfold increased pZIP7 in nontarget siRNA–transfected hyperglycemic cells was observed (Fig. 6E and F). The ratio of pZIP7/ZIP7 is fivefold higher in hyperglycemia than in normoglycemia in nontargeted siRNA cells. However, silencing of CK2α has not changed this ratio under control and hyperglycemic conditions (Supplementary Fig. 7), suggesting the importance of increased ZIP7 phosphorylation in response to hyperglycemia.

Our overall aim was to examine roles of Zn²⁺ carriers ZIP7 and ZnT7 in cardiomyocytes under physiological conditions and as potential mediators of ER stress in hyperglycemia- and hyperglycemia-induced cardiac dysfunction. The results suggest that these Zn²⁺ carriers are likely to carry Zn²⁺ in opposite directions across the S(E)R membrane. Furthermore, we show that increased ZIP7 levels together with increased ZIP7 phosphorylation were likely to drive changes in S(E)R lumen and cytosol with deleterious consequences for cell function in both cases. Although not assessed directly, lowered ZnT7 expression under hyperglycemia, particularly under chronic conditions, is likely to further exacerbate these changes and contribute to deleterious consequences of Zn²⁺ redistribution between compartments. Of note, these transporters localize into the S(E)R membrane and may thus operate as functional proteins catalyzing Zn²⁺ release and uptake from the S(E)R.

To the best of our knowledge, this study is the first to assess both localization and functional roles of ZIP7 and ZnT7 as working in opposing directions in mammalian heart and, most importantly, their contributions to diabetic cardiac dysfunction through ER stress. In addition, we demonstrate markedly increased CK2α expression under hyperglycemia (at most through hyperglycemia activation of CK2α [35]) and provide evidence that this can lead to phosphorylation of the channel on residues previously implicated in control of Zn²⁺ release (18). In particular, our experiments that used CK2α-silenced cells provide further support for this view. Thus, a combination of transcriptional and posttranscriptional mechanisms may contribute to ZIP7 activation in hyperglycemia. Increases in both expression and phosphorylation of ZIP7 under hyperglycemia are time dependent, and the ZIP7 phosphorylation increase is significantly higher than that in chronic hyperglycemia, indicating a role of ZIP7 activation in cardiomyocytes in diabetes.

The data provide further evidence of a role for the S(E)R as an intracellular Zn²⁺ pool, contributing to Zn²⁺ regulation in cardiomyocytes, in a process controlled by ZIP7 and ZnT7. In this regard, S(E)R Zn²⁺ deficiency has been shown to induce unfolded protein response (14) and that Zn²⁺ homeostasis is involved through a putative ZnT localized to the S(E)R (37). Moreover, high glucose has been demonstrated to induce a stable increase in cytosolic Zn²⁺ in pancreatic β-cells and to lead to profound alterations in expression of genes important for Zn²⁺ homeostasis, such as a ZIP7 increase (25). Therefore, the current data are noteworthy when considering ZIP7 and ZnT7 roles located into the S(E)R and responsible for ER stress in part as a result of depressed S(E)R free Zn²⁺.

Cytosolic Zn²⁺ increase in cardiomyocytes can induce marked high phosphorylation in many kinases and oxidation of proteins responsible for heart function (3,4,30,38), which is consistent with previous studies on the role of ZIP7 in control of Zn²⁺ release from the S(E)R in cancer cells (17,18). In support of these earlier findings, we show that CK2α is activated in hyperglycemic cardiomyocytes and directly activated with increased cytosolic Zn²⁺ levels. More importantly, these observations support our present hypothesis. Indeed, CK2α plays a key role in the regulation of prosurvival as well as proapoptotic ER stress signaling by directly modulating the activities of ER stress–signaling actors by phosphorylation, regulating expression of the key factors of signaling pathways or binding to regulator proteins (39).

Little is known about cytosolic Zn²⁺, ZnTs, and cardiovascular complications. Early studies demonstrated that ZIP7 is involved in Zn²⁺ homeostasis of the Golgi apparatus (15). Although ZIP7 and ZnT7 have been shown to be weakly expressed in mammalian heart (22,23), we observed marked ZIP7 and ZnT7 protein levels in DM rat heart, although others have demonstrated a correlation among increased reactive oxygen species, ER stress, and CK2α activation (36,40) through cellular control of Zn²⁺ distribution with CK2α-mediated ZIP7 phosphorylation (18). Therefore, the current data represent a first investigation into the role of CK2α-mediated ZIP7 phosphorylation and Zn²⁺ homeostasis in cardiomyocytes under hyperglycemia and further propose a protective role for ZnT7 against oxidative stress or high susceptibility to diet-induced glucose intolerance/insulin resistance (24,41).

Hyperglycemia induces cardiac dysfunction as a result of increased oxidative and ER stress (42,43). The current data reinforce the hypothesis of an S(E)R role as an important Zn²⁺ pool in cardiomyocytes with ZIP7 and ZnT7 localization into the S(E)R and make an important contribution to Zn²⁺ homeostasis. Nonetheless, we also recognize a possible contribution of other ZnTs that contribute to ER stress in cardiomyocytes under hyperglycemia. The data, therefore, provide important information related to Zn²⁺ homeostasis through ZIP7 and ZnT7 functioning as opposing ZnTs localized to the S(E)R in cardiomyocytes. Perturbed expression of either or both transporters may lead to altered Zn²⁺ distribution across the S(E)R and lead to persistent ER stress.

Zn²⁺ homeostasis in cardiomyocytes under physiological and hyperglycemic conditions is summarized in Fig. 7. Regulation of subcellular free Zn²⁺ distribution is proposed to occur through the ZIP7/ZnT7 system localized to the S(E)R. When cardiomyocytes are exposed to high glucose, CK2α becomes activated, inducing a Zn²⁺ influx into cytosol from S(E)R through ZIP7 phosphorylation (18), although a downregulated ZnT7 may further contribute to this process. High glucose also stimulates a cytosolic free Zn²⁺ increase through mobilization of Zn²⁺ from metalloproteins (5,30,44,45), which in turn induce phosphorylation of several proteins (30,45,46). Of note, the latter might further stimulate ZIP7 activity and may lead to accelerated Zn²⁺-influx into cytosol from the S(E)R. Consequently, cytosolic free Zn²⁺ increases driven by multiple mechanisms lead to ER stress through and the overexpression of ER stress chaperones in cardiomyocytes during hyperglycemia. The current data provide novel insights into regulation of cellular Zn²⁺ and its role in hyperglycemia/diabetes-associated cardiac dysfunction. In addition, the findings may provide new targets, such as cellular Zn²⁺ regulation through mediation of ZnTs, and suggest that modulation of CK2α provides a novel means to correct diabetes-induced cardiac dysfunction.

Acknowledgments. The authors thank Yusuf Olgar (Department of Biophysics, Faculty of Medicine, Ankara University, Ankara, Turkey) for contributions to some of the Western blot analyses. The authors also thank Dr. Angeles Mondragon (Imperial College London) for assistance with plasmid amplification and virus production and Steve Rothery (Imperial College London) for assistance with imaging experiments.

Funding. E.T. acknowledges receipt of a European Cooperation in Science and Technology COST action STSM TD1304 grant and thanks the European Foundation for the Study of Diabetes for an Albert Renold Travel Fellowship. G.A.R. thanks the Medical Research Council (U.K.) for programme grant MR/J0003042/1, the Biotechnology and Biological Sciences Research Council (U.K.) for project grant BB/J015873/1, and the Royal Society for a Wolfson Research Merit Award and the Wellcome Trust for a Senior Investigator Award (WT098424AIA). B.T. thanks the Scientific and Technological Research Council of Turkey for grant SBAG-113S466 and COST action TD1304. K.M.T. thanks the Wellcome Trust University Research Award (grant number 091991/Z/10/Z).

Duality of Interest. No potential conflicts relevant to this article were reported.

Authors Contributions. E.T. performed the electrophysiological and imaging experiments, contributed to all biochemical and molecular biology experiments, and analyzed all experimental data and confocal images. V.C.B. and A.D. performed all biochemical and molecular experiments. G.R.J.C. contributed to the knockdown experiments. K.M.T. provided the antibody for ZIP7 and contributed to the discussion. G.A.R. supported the experimental data in his laboratory and reviewed and edited the manuscript. B.T. designed the study, wrote the manuscript, researched data, and provided funding for the study. B.T. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the Biophysical Society 60th Annual Meeting, Los Angeles, CA, 27 February–2 March 2016.