Altered K⁺ channel gene expression in diabetic rat ventricle: isoform switching between Kv4.2 and Kv1.4

Materials.

The affinity-purified Kv1.4 antibody, affinity-purified Kv4.2 antibody, and Kv4.2-specific antiserum have been previously characterized (6, 58). Immunological reagents were purchased from Jackson Immunoresearch Laboratories.

Induction of diabetes and tissue preparation.

All studies involving rats were done in accordance with the principles set forth in the Guide for the Care and Use of Laboratory Animals(National Institutes of Health, 1996). Male Sprague-Dawley rats (12 wk old) were randomly assigned to nondiabetic (control), diabetic (DM), and DM + IGF-II treatment (DM + IGF) groups. Each group contained four rats. After the rats were fasted overnight, diabetes was induced in rats with a single intraperitoneal injection of 50 mg/kg streptozotocin (STZ; Sigma) and maintained for 12 wk with free access to rat chow and water. At the end of 6 wk, the STZ-treated animals were implanted subcutaneously in the midback with osmotic minipumps (0.5 μl/h) that released either vehicle (1 mM acetate, pH 6) or IGF-II (5 μg · rat⁻¹ · day⁻¹) for the remaining 6 wk. Serum glucose concentrations were determined at the end of the first 6-wk period and on the day of euthanasia with the use of a glucose diagnostic kit (model 510A; Sigma). The cardiac ventricle and brain in each animal were removed, and small pieces of the left ventricular free wall near the left descending artery were embedded in Tissue Tek (Baxter) and slowly frozen at −80°C for immunohistochemical analysis. The remaining ventricular tissue was used for both membrane preparation and RNA extraction. One brain hemisphere (except for the olfactory bulb) was used for RNA extraction.

Northern blot analysis.

Total RNA was extracted from freshly isolated ventricles and brains using the guanidinium thiocyanate method (1). Ethidium bromide (40 μg/ml) was added to RNA samples before electrophoresis to enable visual confirmation of RNA integrity. After fractionation of 10 μg of total RNA through a 1% agarose-3% formaldehyde gel in 20 mM MOPS and 1 mM EDTA, pH 7.4, the RNA was transferred to a Nytran filter (Schleicher and Schuell) by capillary action in 20× saline sodium citrate. RNA was cross-linked to the filter by ultraviolet irradiation using a Stratalinker (Stratagene). The filter was hybridized with a rat α- or β-MHC oligonucleotide probe (Calbiochem). Oligonucleotide probes were 5′ end labeled with [γ-³²P]ATP (6,000 Ci/mmol; ICN) using T4 polynucleotide kinase. The hybridization and wash were performed according to standard protocols. After hybridization and wash, the filter was exposed to a PhosphorImager screen (Molecular Dynamics) for quantification and then were exposed to XAR-5 film (Kodak).

Ribonuclease protection assay.

The sequence of the cDNA templates used for the production of the radioactive antisense RNA probes for Kv1.4, Kv4.2, and Kv4.3 and the detailed procedure of ribonuclease protection assay (RPA) have been described previously (36). In brief, plasmids containing Kv channel cDNAs were linearized with the appropriate restriction enzyme and antisense cRNA probes were synthesized using the MAXIscript kit (Ambion) with [α-³²P]UTP (ICN Radiochemicals). The cyclophilin cRNA probe was synthesized from cDNA (pTRI-cyclophilin rat antisense control template) purchased from Ambion to detect cyclophilin mRNA as an internal control. RPA was performed using a RPAIII kit (Ambion) according to the manufacturer's protocol. Hybridization of the Kv channel and cyclophilin probes with 10 μg total sample RNA was carried out at 42°C overnight, followed by digestion with a RNase A and T1 mix at 37°C for 30 min. The reaction was terminated by the addition of SDS and proteinase K. After ethanol precipitation, the protected RNA fragments were subjected to electrophoresis in a 5% polyacrylamide-8 M urea gel. The gel was dried on 3M paper and subjected to quantitative analysis using a PhosphorImager. The mRNA levels were normalized to the levels of cyclophilin mRNA and expressed in relative arbitrary units.

Immunofluorescence.

Cryosections of the rat ventricle (10 μm) were incubated with primary rabbit antibody followed by biotin-conjugated goat anti-rabbit IgG (Jackson Immunoresearch) and CY3-conjugated streptavidin as previously described in detail (34). Samples were analyzed using a Nikon E800 microscope equipped with standard epifluorescence and a Princeton Instruments charge-coupled device camera. Antigen-blocking experiments were conducted to confirm Kv4.2 antiserum specificity. Here, the antibody staining was performed as described above except that cryosections were stained with antiserum that had been preincubated for 2 h at room temperature with 200 nmol of the Kv4.2 peptide per milliliter of diluted antiserum (34). Blocking experiments were not necessary with the other two antibodies because they were affinity purified.

Western blot analysis.

Approximately 0.5 g of each thawed rat ventricle was homogenized with a high-shear homogenizer (Tekmar) in 5 ml of 0.32 M sucrose-5 mM Na₂HPO₄ with the following protease inhibitors (Sigma, unless otherwise noted): 0.31 mg/ml benzamidine, 0.62 mg/mlN-ethylmaleamide, 1 mg/ml bacitracin, 1 μg/ml pepstatin, 1 μg/ml leupeptin, 0.31 mg/ml pefabloc (Boehringer Mannheim), and 0.8 μg/ml calpain inhibitor (Calbiochem). Tissue was homogenized on ice at high speed with a small (1.0 cm) diameter probe for 60 s. The homogenate was centrifuged at 4°C for 10 min at 3,000 rpm (Beckman JA 25.5 rotor) to remove large debris and nuclei. The supernatant was further homogenized by 10 strokes in a 7-ml Wheaton dounce tissue grinder on ice and centrifuged for 1 h at 4°C and 12,000 rpm (JA 25.5; Beckman Instruments). The resulting membrane pellet was resuspended in 100 μl of ice-cold PBS and stored at −80°C. Membrane proteins were fractionated by SDS-PAGE and transferred overnight to nitrocellulose membrane (Schleicher & Schuell) using the standard Laemmli method as previously described (33). Briefly, SDS sample buffer was added to the isolated membranes before electrophoretic separation on a 10% polyacrylamide gel using a minigel system (Bio-Rad). After overnight transfer onto nitrocellulose membrane, the samples were stained with 0.1% Ponceau S Solution (Sigma) (to visualize the quality of the transfer), rinsed with deionized water, and incubated overnight at 4°C in solution 1 (S1) [50 mM Tris (pH 7.5), 150 mM NaCl, and 0.1% Tween-20, and 5% nonfat dry milk] to block nonantigenic sites. The blots were then incubated in S1 at room temperature for 1 h with either Kv1.4 or Kv4.2 polyclonal antibodies (diluted 1:1,000). The blots were washed 3 times for 15 min each with solution 3 (S3) [50 mM Tris (pH 7.5), 500 mM NaCl, and 0.1% Tween 20] and then incubated for 1 h at room temperature in S1 containing a 1:3,000 dilution of horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG or 1:5,000 of HRP-conjugated goat anti-mouse IgG. The blots were washed three times for 15 min each with S3 and detection was achieved with the Renaissance enhanced chemiluminescence reagent (New England Nuclear Life Science Products) per the manufacturer's instructions.

Body weights, ventricular weights, and serum glucose levels in diabetic animals.

A single STZ treatment was used to experimentally induce the diabetic state over a 12-wk period. As summarized in Table 1, rats receiving STZ had significantly elevated serum glucose concentrations and reduced body weights compared with nondiabetic control animals. Low-dose IGF treatment (DM + IGF) affected neither serum glucose levels nor weight loss in the diabetic animals. There was no significant difference in ventricle weights among the experimental groups. The ventricle-to-body weight ratio increased in both DM groups relative to control because body weight was reduced by the diabetic condition, i.e., it was not indicative of cardiac hypertrophy. We also examined the ventricle and myocyte morphology (data not shown). There were no significant changes in ventricular wall thickness (control, DM, and DM ± IGF: 2.23 ± 0.14, 2.12 ± 0.16, and 2.26 ± 0.28 mm, respectively) or myocyte width (control, DM, and DM + IGF: 17.7 ± 2.2, 18.5 ± 2.7, and 18.3 ± 3.6 μm, respectively) among the groups (data not shown) (n = 40 myocytes total per group, each group consisting of four animals). These data indicate that diabetes, with or without IGF treatment, was not associated with cardiac hypertrophy using standard morphological measurements.

MHC isoform switching from the α- to β-isoform occurs in the hypertrophied or diseased rat heart (23, 46). Therefore, although gross ventricular hypertrophy was not detected, we tested for altered myosin gene expression. The ventricular α- and β-MHC mRNA levels as a function of treatment are shown by the Northern blot analysis of Fig. 1, A andB. Diabetic treatment resulted in a significant decrease of α-MHC mRNA to 32 ± 11% of control (P < 0.01), and IGF treatment restored the mRNA level to 51 ± 3% (P < 0.01 vs. DM) (Fig. 1 A). Diabetic treatment increased the β-MHC mRNA level to 259 ± 8% of control (P < 0.01), and IGF treatment reduced this change to 217 ± 31% of control (P < 0.05 vs. DM) (Fig. 1 B). Thus hypertrophy may have developed with a longer time period, for myosin isoform expression did change.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Effect of diabetes and insulin-like growth factor (IGF) treatment on myosin heavy chain (MHC) isoform mRNA levels in ventricle. The expression of cardiac α-MHC and β-MHC mRNA in rat ventricle was determined by Northern blot analysis as shown in A andB, top, respectively. Shown are representative autoradiographs from two rats from each experimental group. MHC isoform mRNAs were normalized to 28S ribosomal mRNA levels and expressed as arbitrary units (%control) in A and B,bottom. C, control; DM, diabetic. Values are means ± SD (n = 4 rats per group). *P < 0.05; **P < 0.01.

Effect of diabetes and IGF treatment on potassium channel α-subunit mRNA levels in heart and brain.

To examine the effect of diabetes and IGF treatment on Kv1.4, Kv4.2, and Kv4.3 gene expression in the heart, mRNA levels were measured in ventricles using an RNase protection assay with cyclophilin mRNA used as an internal control. As shown in Fig.2, A and B, Kv1.4 mRNA level increased to 179 ± 24% of control (P< 0.01), whereas Kv4.2 mRNA level decreased to 41 ± 5% of control (P < 0.01) in the diabetic rat ventricles. IGF treatment did not reverse these changes. In contrast, Kv4.3 mRNA levels showed no significant change with any treatment (Fig.2 C).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Change in K⁺ channel mRNA levels in diabetic ventricles. Kv1.4, Kv4.2, and Kv4.3 mRNA levels were measured in control, DM, and (DM + IGF) rat ventricles as shown inA, B, and C, respectively. Ventricular total RNA (10 μg) was subjected to RNase protection assay using³²P-labeled Kv and cyclophilin (cyclo) cRNAs as probes as described in materials and methods. A–C, left: representative autoradiographs obtained with each experimental group. Open arrowheads indicate the position of undigested Kv and cyclophilin (cyclo) cRNA probes, and closed arrowheads indicate the protected fragments. Kv mRNA was normalized to cyclophilin mRNA levels and expressed as relative arbitrary units (A-C, right). Presented are means ± SD (n = 4 rats per group). **P < 0.01.

To determine whether change in K⁺ channel mRNA levels in diabetes was restricted to the heart, mRNA levels of Kv1.4, Kv4.2, and Kv4.3 were examined also in the brain. As shown in Fig.3, A–C, diabetic animals had the same relative mRNA levels for all three channels in the brain. Two alternatively spliced variants of Kv4.3 are present in the rat brain, whereas only a single isoform is found in the ventricle (36, 51). Diabetes had no effect in relative levels of these two splice variants in the brain.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

K⁺ channel mRNA levels in diabetic brain. Kv1.4 (A) and Kv4.2 (B) mRNA levels were measured by RNase protection in C, DM, and (DM + IGF) rat whole brains.C: expression of Kv4.3 splice variants. Left:autoradiographs; right: Kv mRNA expressed as relative arbitrary units. With the RNase protection assay probe used in this study, the protected fragment for Kv4.3 long shows 312 bp and that for Kv4.3 short shows 165 and 90 bp (36). The 90-bp band is hidden by the robust cyclophilin mRNA band (103 bp). Presented are means ± SD (n = 4 rats per group).

Immunoreactive Kv4.2 and Kv1.4 protein levels in diabetic ventricle.

Kv α-subunit protein expression levels do not always correlate with mRNA levels in rat heart (61). In addition, altered expression could be occurring in cell types other than ventricular myocytes. Therefore, Kv4.2 immunostaining of ventricle cryosections was performed to address the location and amount of ventricular K⁺ channel protein levels in the diabetic state. In the control rat ventricle, Kv4.2 immunoreactivity was distributed evenly over myocytes, while being slightly concentrated at the ends of the cells [Fig. 4 A, consistent with previous reports (6)]. In the diabetic rat ventricle, the staining level was reduced compared with control (compare Fig. 4, A and B) as predicted from the changes in mRNA levels. The antibody staining of IGF- and vehicle-treated diabetic rat ventricles was similar (data not shown), again as predicted from the mRNA analysis. Figure 4, C andD, shows the staining observed with no primary antibody or with antiserum previously adsorbed with the peptide used for antiserum production. Only background staining was observed in the absence of primary antibody, and the peptide treatment reduced the staining to near-background levels. Together these data indicate antibody specificity. Also supporting the antibody specificity are the results shown in Fig. 4, E and F. Here, a different and affinity-purified anti-Kv4.2 antibody was used to stain control and diabetic myocardium, Fig. 4, E and F, respectively. Again, antibody staining was reduced in the diabetic tissue. Figure 4, G and H, illustrates the immunostaining observed with an affinity-purified anti-Kv1.4 antibody. As predicted from the changes in mRNA levels, the diabetic state caused an increase in Kv1.4 protein as detected by the immunofluorescence. Figure 4 G shows staining to control tissue, whereas Fig.4 H represents diabetic tissue. Although gradients of Kv1.4 and Kv4.2 expression exist across the ventricular wall (6, 8,58), a protein gradient was not observed for either channel in this study. Perhaps this is because the tissue sections employed in this study were from the left ventricular free wall base and the gradient is more apparent in the middle and apex of the ventricle (6). Together, these immunostaining data argue for altered expression in the left ventricular myocytes as opposed to cells derived from, for example, vascular beds.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Immunodetection of relative Kv4.2 and Kv1.4 protein levels in diabetic ventricle. A and B: staining of anti-Kv4.2 antiserum to C and DM rat heart, respectively.C and D: staining to control tissue with either no primary antibody or after preincubation of the primary antibody with the peptide immunogen. E and F: staining of affinity-purified anti-Kv4.2 antibody to C and DM rat heart, respectively. G and H: staining of affinity-purified anti-Kv1.4 antibody to C and DM rat heart, respectively.

Western blot analysis of Kv4.2 and Kv1.4 protein levels in diabetic ventricle.

We also confirmed via Western blot analysis that Kv4.2 and Kv1.4 protein levels in the ventricle changed in parallel to the observed changes in mRNA. As shown in Fig. 5, Kv1.4 protein was dramatically increased with the diabetic state, whereas Kv4.2 was markedly reduced. Because of the limited depth of film, quantitation was not attempted. Still, these data confirm the mRNA analysis and thus further support the idea that the change fromI _TOfast to I _TOslowaccompanying diabetes is due to the upregulation of the Kv1.4 protein expression at the expense of Kv4.2 channel expression.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Western blot analysis of Kv1.4 and Kv4.2 protein expression in DM and C ventricle. A: immunodetection of Kv1.4 protein in representative DM and C samples, respectively.B: levels of Kv4.2 protein. Equal amounts of membrane protein were loaded in each lane as confirmed by Ponceau staining and analysis with antitubulin antibodies (data not shown). Arrows, position of Kv1.4 and Kv4.2α subunits.

Abnormal Kv gene expression in diabetes.

APD is prolonged in the STZ-induced diabetic rat ventricle (31,38), and the attenuation of I _TO has been proposed as being responsible (48-50). Furthermore, an increased rate dependency (38) or prolongation of recovery kinetics of I _TO was reported (49,50). When heterologously expressed, Kv1.4 and Kv4.2 produce rapidly activating and inactivating outward currents and are considered to underlie cardiac I _TO (9, 51, 55,63). One of the greatest differences between these Kv channels is their recovery from inactivation. Kv1.4 recovers with a time constant of 3–8 s (2, 40, 41), whereas the Kv4.2 recovers much faster with a time constant of 200–1,000 ms (57, 58). Recent gene knockout and dominant negative approaches have confirmed that I _TOfast is encoded by Kv4.2, whereas I _TOslow is due to Kv1.4 (19, 62).

The data shown above indicate that Kv4.2 gene expression (Fig.2 B) and protein (Fig. 4, A and B, and Fig. 5) levels were reduced, whereas Kv1.4 gene expression (Fig.2 A) and protein levels (Fig. 4, G andH, and Fig. 5) were increased in diabetic heart ventricles. These changes are likely to represent the molecular mechanism responsible for the effects of diabetes on I _TOand the resulting increase in cardiac APD. Our findings support recently published work of Qin et al. (42), which shows that diabetes reduces ventricular Kv4.2 protein and mRNA levels. However, this study did not examine the levels of either Kv1.4 mRNA or protein. In contrast to our findings reported here, these investigators found Kv4.3 expression was also suppressed with diabetes.

Interestingly, Kv4.2 and Kv1.4 gene expression was changed without marked cardiac hypertrophy (Table 1, Fig.1) (42), albeit MHC gene expression was abnormal. This shows that the alteration in Kv4.2 and Kv1.4 mRNA levels was not secondary to ventricular hypertrophy. K⁺ channel expression in the heart is associated with cardiac hypertrophy (29), and it seems likely we would also have observed hypertrophy with longer duration or more deeply diabetic rats. The abnormal Kv1.4 and Kv4.2 gene expression in the ventricle was not observed in the brain (Fig. 3,A and B), showing that this abnormality could not be simply ascribed to hyperglycemia. Other possibilities include perhaps hemodynamic factors (12, 27, 56) or impaired autonomic neural control of the heart (17, 30, 53). In fact, the heart rate of the STZ-induced diabetic rats is reduced (12, 27, 56). It is interesting that, although both Kv4.2 and Kv4.3 contribute to cardiac I _TO, only the Kv4.2 expression was altered.

Our present data are in contrast to several other studies examining gene expression in hypertensive hearts or after infarction. Takimoto et al. (51) showed that in renovascular hypertensive rats, both Kv4.2 and Kv4.3 expression were decreased, whereas Kv1.4 expression was unaltered. In one study examining alteredI _TO expression after infarction both Kv4.2 and Kv4.3 were downregulated, whereas Kv1.4 expression increased (26).

Clearly, the cardiac K⁺ gene expression responds differently to hypertension, infarction, and the diabetic state studied in our present work. IGF treatment of diabetic rats did not affect the changes in Kv channel expression in the ventricle or brain. This treatment also had no effect on these tissues in nondiabetic rats.

Low-dose IGF treatment protects against MHC isoform switching.

α-MHC mRNA was reduced 32% (Fig. 1 A), whereas β-MHC was increased 259% (Fig. 1 B) in diabetic ventricles. This isoform switching is observed in cardiac but not skeletal muscle (35). It is noteworthy that MHC isoform switching was prevented by low-dose (∼15 μg · kg⁻¹ · day⁻¹) IGF-II treatment, despite continued hyperglycemia and body weight loss (Table1). This suggests that isoform switching may not be a consequence of hyperglycemia per se but may result from reduced IGF gene expression in the heart (5, 28). Likewise, low-dose IGF-I or IGF-II treatment can prevent diabetic neuropathy independently of hyperglycemia and weight loss (22, 64). Further study is needed to determine whether IGF protection results from indirect actions on autonomic function or direct actions on cardiac tissue. IGF treatment both prevents ultrastructural abnormalities in autonomic nerves (47) and attenuates myocardial injury and apoptosis after ischemia (7). Thus IGF-I improves cardiac output and decreases systemic vascular resistance after infarction (10). Both IGF-I and IGF-II act through the type I IGF receptor expressed on the surface of ventricular myocytes (18).

In summary, we have demonstrated that K⁺ channel genes regulating the I _TO switch from Kv4.2 to Kv1.4 in diabetic ventricles. This change may well explain the reduced rate of repolarization and abnormal I _TO observed in the diabetic heart. The switch in Kv channel isoform expression preceded observable cardiac hypertrophy, albeit MHC isoform gene expression was altered. IGF treatment blunted the diabetes-induced changes in MHC isoform gene expression independently of continued hyperglycemia but had no effect on Kv channel expression. These findings suggest IGF treatment may ameliorate some but not all aspects of cardiac contractile dysfunction in diabetes.