INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE NK CLASS OF HOMEODOMAIN proteins is essential in myogenic lineage development (19). In Drosophila, the NK2 homeodomain protein tinman is a transcription factor required for insect cardiac mesodermal determination (3, 5). In vertebrates, tinman is represented by a closely related family of NK2 genes including the murine cardiac homolog Nkx2-5 (21) (also called CSX, Ref. 20), the Xenopus homolog XNkx2.3 (15), and the human homolog CSX1, which has three isoforms (27). Nkx2-5 is detectable before both cardiac myogenic differentiation and the expression of cardiac-specific genes such as -cardiac actin (21) and atrial natriuretic factor (ANF) (14). Nkx2-5 is then continually expressed throughout cardiogenesis as well as in the adult heart of mice (20) and humans (27). The presence of Nkx2-5 in the adult myocardium suggests a role in maintaining the cardiac phenotype. Mice with an homologous knockout of Nkx2-5, like that of tinman in Drosophila (15), result in an embryonic lethal phenotype that occurs before cardiac looping (22). Nkx2-5 has been characterized predominantly as a cardiac transcription factor, and its target DNA binding element, NKE, resembles the serum response element (6, 7). Nkx2-5 has been recently found to provide specific transcriptional activation of two cardiac genes important in both the adult and embryonic heart: -cardiac actin (6, 8) and ANF (14).

Adult cardiac myocytes are terminally differentiated and respond to growth stimuli by hypertrophying (10). Likewise, the adult heart compensates for increased hemodynamic pressure load with a hypertrophic response. The degree of increase in cardiac mass is dependent on the severity and type of ventricular wall stress imposed (12). Sustained hemodynamic overload of the heart often leads clinically to congestive heart failure. Genes induced or upregulated during hypertrophic induction secondary to hemodynamic load include, but are not limited to, cardiac contractile proteins (10), certain protooncogenes (18), and the reexpression (or marked increase in expression) of a set of transcripts normally quiescent in the adult ventricle but that are predominantly expressed during embryonic life. Examples include ANF (28), skeletal -actin (26), -myosin heavy chain (17, 23, 25), and atrial light chain-1 (16). Because the exact mechanism(s) for transducing hemodynamic load into the cardiac hypertrophic response is unknown, it becomes important to examine genes that are potentially involved in the regulation of gene transcription during cardiac growth.

Because Nkx2-5 is expressed in the adult heart and is known to be capable of contributing to the activation of ANF (14) and -cardiac actin (6, 8) in vitro, the question is raised whether such a cardiac homeobox gene critical to the heart during development might also be capable of transcriptional activation in response to pressure overload in adult myocardium. A first step would be to examine Nkx2-5 transcript levels during hypertrophic stimulation. To address this, a feline right ventricular (RV) pressure overload (RVPO) model utilizing pulmonary artery banding was employed. This model allows study of the physiological effect of a doubling of RV hemodynamic load as an initiating and continuing stimulus. The effect of hemodynamic load on the RV is separated from systemic variables that would affect both ventricles. RV hypertrophy occurs over the first several days after pulmonary artery banding, and the RV wall growth response is largely completed by 14 days (29). Such an in vivo model offers several advantages including hemodynamic changes that resemble human disease (13), the ability to consistently examine different durations of pressure overload, and the use of the unloaded left ventricle (LV) as a same animal control (24). As such, specific cDNA probes for Nkx2-5 were used to compare changes in transcript levels during RVPO and to correlate with changes in ANF and -cardiac actin.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

RVPO models. Briefly, adult cats (Felis domesticus) between 2.5 and 4.1 kg were gently anesthetized with meperidine (10 mg/kg im), methohexital sodium (20 mg/kg ip), and -chloralose (60 mg/kg iv) followed by arterial cannulation for blood pressure monitoring, thoracotomy, and pericardiotomy. RVPO to 40-50 mmHg was induced with a partially occluding 3.5-mm pulmonary artery (PA) band. This is a level that produces RV hypertrophy that becomes maximal in terms of wall thickness within a 14-day period without evidence of ischemia or infarction (13). Controls consisted of sham-operated animals. At the time of death, animals received identical anesthesia with RV, and systemic pressure measurements were obtained before the isolation of the heart and perfusion with ice-cold heparinized saline via the coronary arteries. The RV and LV free walls were dissected and stored in liquid N₂ (24). An adequate response consisted of either a doubling of baseline RV systolic pressure, measurements above 42 mmHg (average baseline was 24 ± 1.2 mmHg), or a marked increase in RV-to-body weight ratio. By 14 days, RVPO hypertrophy reaches its maximum gross response as measured by serial echocardiography (29) and histology (13). All procedures and the care of the animals were in accordance with institutional guidelines and National Institutes of Health "Guide for the Care and Use of Laboratory Animals" [Department of Health and Human Services Publication No. (NIH) 85-23, Revised 1985].

RNA isolation and hybridization. Total RNA was isolated from frozen tissue pieces homogenized in a polytron in 4.0 M guanidinium thiocyanate using standard techniques (2). Poly(A)-enriched RNA was prepared using a Fast Track mRNA Isolation kit (Invitrogen). RNA samples were denatured, size separated on 1.0% agarose-formaldehyde gels, and transferred to Duralon nylon membranes (Stratagene). ³²P-labeled specific probes were hybridized in 50% formamide, 5× standard saline citrate (SSC), 2× Denhart's solution, 0.1% SDS, 0.5% dextran sulfate, and 100 µg/ml tRNA at 42°C overnight. After samples were serial washed to 0.2× SSC and 0.1% SDS at 42°C, signals were visualized with autoradiography at 70°C for up to 72 h. RNA loading was normalized for equivalency between each animal's RV and LV based on glyceraldehyde-3-phosphate dehydrogenase (GAPDH) signal. All autoradiography images were digitized and quantitated using National Institutes of Health image software. Statistical analysis was performed on all data with Super ANOVA software, and statistical significance was defined as a P value <0.05 by Dunnett's one-tailed (sample mean larger than control mean) test.

Feline -cardiac actin 3'-UTR cloning. The 3'-untranslated region (UTR) for the feline -cardiac actin gene was cloned using standard methods for rapid amplification of cDNA ends (3'-RACE). Briefly, first-strand cDNA synthesis was performed from feline heart total RNA using a Superscript preamplification system for first-strand cDNA synthesis kit (GIBCO-BRL). This served as a template for polymerase chain reaction (PCR) using a gene-specific 5'-primer, 5'-CTGTCCACCTTCCAGCA-3', and an oligo(dT)-specific 3'-primer, 5'-GACTC GAGTCGACATCG(T)₁₇-3'. The PCR reaction products were cloned into pCR2.1 (Invitrogen) according to the manufacturer's instructions. Several clones were sequenced and shown to be -actin. Comparison with the published feline -skeletal actin (24) and other species' -cardiac actin sequences led to the identification of a feline -cardiac actin clone.

³²P probe generation. A DNA fragment containing the GAPDH coding region from amino acids 142 to 207 (1) was PCR amplified from pBluescript using M13-forward and M13-reverse primers. A feline ANF-specific DNA fragment was PCR amplified using the 5'-primer 5'-GACGCCAGCATGAGCTCCTTC-3' and the 3'-primer 5'-CTCCAATCCTGTCCATCCTGC-3' from a feline cardiac myocyte library. A 1,300-bp DNA fragment containing the entire murine coding region for Nkx2-5 was obtained from the expression plasmid pCGNCSX (a generous gift of Dr. Timothy McQuinn, Ref. 6) by digestion with EcoR I. A 341-bp DNA fragment of Nkx2-5 containing the sequence between bases 299 and 640 (21) was PCR amplified using the 5'-primer 5'-GCCCACGCCYTTCTCAGTCA-3' and the 3'-primer 5'-TCCAGCTCCACYGCCTTCTG-3' and pCGNCSX as template DNA to generate an Nkx2-5-specific probe [5'-341(Nkx2-5)] that did not contain the NK family homeodomain. These cDNAs were isolated on 1.25% low-melting-point agarose gel by electrophoresis, purified using standard techniques, and labeled using ³²P (Du Pont-New England Nuclear) and a nick translation kit (GIBCO-BRL).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Nkx2-5 transcript levels are upregulated during cardiac hypertrophic growth. All PA-banded cats at death had an ending RV pressure of >40 mmHg (a pressure known to induce hypertrophic growth, Ref. 13) as well as a significant increase in their RV free wall-to-body weight ratios. These parameters for all the animals used in these experiments are listed in Table 1. Note that that although three animals were examined at each RVPO time point for each cDNA probe, because of tissue limitations not all animals were tested with every probe. To determine the relative level of Nkx2-5 mRNA during hemodynamic pressure overload, total RNA, and for one animal at each time point, poly(A) RNA, was isolated from each PA-banded and control animal's RV and LV. GAPDH hybridization was used to normalize the RNA loading between the RV and LV samples, since there is a significant increase in ribosomal RNA during cardiac hypertrophy (24) and GAPDH levels do not appreciably increase (32). Note that GAPDH normalization was between each of the RV and LV pairs for each animal, and as such, the RV-to-LV ratio is unity for each time point. Northern analysis utilizing the full-length murine Nkx2-5 probe demonstrated mRNA upregulation in triplicate RVPO samples after 2 days (mean, 5.1-fold; P < 0.05), 7 days (mean, 3.9-fold; P = NS), and to a lesser degree, 14 days (mean, 1.9 fold) of PA banding as compared with the same animal's LV as well as sham RV and LV controls (Figs. 1 and 2). No differences were found between results using total RNA and poly(A) RNA. Because the full-length Nkx2-5 cDNA probe contained the NK family homologous homeodomain, there is the possibility that hybridization was not specific for Nkx2-5. Additionally, all Nkx2-5 probes tested that contained the homeodomain sequence bound to 28S rRNA, thereby making the Northern analysis more technically difficult. Therefore, another cDNA probe was designed: 5'-341(Nkx2-5), containing the Nk family specific TN domain as well as the 5'-Pro/Ala-rich sequence specific for Nkx2-5 but not extending into the homologous homeodomain. Specific hybridization of Nkx2-5 to three animals at each time point showed statistically significant increased transcript levels at 2 days (mean, 1.9-fold; P < 0.05) of RVPO with only a 1.5-fold increase at 7 days (P = NS) and a return to baseline by 14 days (Figs. 1 and 2). No correlation between the degree of Nkx2-5 upregulation and any hemodynamic parameter has been established.

View larger version (113K): [in this window] [in a new window]   Fig. 1.   Upregulation of Nkx2-5 in right ventricular (RV) pressure overload (RVPO)-induced cardiac hypertrophy relative to -cardiac actin and atrial natriuretic factor (ANF). Northern analysis of feline heart total RNA isolated from RV and left ventricular (LV) tissue after 0 (sham control), 2, 7, and 14 days of RVPO. All RV and LV paired samples are from same animal, and same membrane was hybridized to each of 5 cDNA probes shown. Hybridization with full-length Nkx2-5 probe is seen at 1.5 kb (21) and is increased at 2 and 7 days and less so at 14 days in RV samples relative to control and LV samples. Samples were normalized for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA hybridization as shown [GAPDH does not change appreciably during development of cardiac hypertrophy (32)]. 5'-341(Nkx2-5) cDNA probe used was specific, since it does not contain NK family homeodomain and demonstrated increased hybridization at 2 and 7 days. Specific feline 3'-untranslated region (UTR) -cardiac actin and ANF cDNA hybridization were also increased in RVPO samples as shown.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   Quantification of upregulation of Nkx2-5 full-length and 5'-specific Nkx2-5 cDNA probes in RVPO-induced cardiac hypertrophy. Shown is Northern hybridization digital quantification with all data points being averages of triplicate experiments with RVPO durations of 0 (sham control), 2, 7, and 14 days (SE are shown). Each RV and LV digitized hybridization value was normalized by dividing it by digitized GAPDH signal. Then each RV-to-LV ratio was determined to quantitate Nkx2-5 upregulation (as such, GAPDH value would be 1.0 for each time point). Both Nkx2-5 probes demonstrated statistically significant increases in transcript levels with RVPO at 2 days, with a return toward baseline levels by 7 and 14 days. Because ANF levels are below level of detection in all control and LV samples, they are expressed as ratio of their RV signal normalized for GAPDH. SE is not shown for ANF, since there is no detectable sham RV level to normalize against and, once ANF becomes detectable, levels vary widely between RVPO cats. -Cardiac actin was significantly increased at all pulmonary artery-banded time points.

Both -cardiac actin and ANF transcript levels upregulate with Nkx2-5 during feline cardiac hypertrophic growth. Because Nkx2-5 activates the promoters of both -cardiac actin (8) and ANF (14), their respective transcript levels were investigated in the RVPO animal model. The RNA membranes used to examine for Nkx2-5 were hybridized with the feline -cardiac actin 3'-UTR probe (Figs. 1 and 2), with mRNA levels increasing at 2 days (3.2-fold), 7 days (3.9-fold), and 14 days (2.8-fold) of RVPO (all with statistically significant P values; P < 0.05) as compared with each animal's GAPDH normalized LV level. Similar results were seen when these membranes were hybridized with the ANF probe except that ANF was not detected in the control RV or LV or in any of the PA-banded LV samples (Figs. 1 and 2). The ANF signal for all the RVPO samples was variable in intensity, which, with no LV standardization, made formal statistics difficult other than to express the mean RV to GAPDH signal for each experimental group (Fig. 2).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

These studies demonstrate a link between the mechanical stimulus of hemodynamic load, which is a primary regulator of cardiac mass (13), and the upregulation of Nkx2-5 transcript levels specifically in the ventricle affected by the increased load. Nkx2-5 mRNA levels increased markedly in the pressure-overloaded RVs of PA-banded cats at 2 and 7 days, which was during the period of maximal RV wall growth (13, 29). Note that in this model 2 days is the earliest practical time of death, since a full clinical recovery of the animal from surgical effects may take up to 24 h. Because Nkx2-5 is present constitutively in the normal adult heart and transcript levels were found to increase markedly during ventricular hypertrophic growth, Nkx2-5 may be an important responder to adult ventricular wall stress. In support of this, overexpression of (Xenopus) XNkx2-5 and XNkx2-3 in Xenopus embryos resulted in myocardial thickening in otherwise morphologically normal hearts (11). Likewise, overexpression of Nkx2-5 in zebrafish produced disproportionately larger hearts in apparently otherwise normal embryos (9). Whatever other covariables may be involved, in our model they all would stem originally from hemodynamic load stimulus. The data showing Nkx2-5 levels returning toward baseline levels (i.e., that of the unloaded ventricles) after 2 wk (when maximal RV growth, in terms of wall thickness, has been achieved) suggest that Nkx2-5 transcript levels upregulate with pressure overload only during the period of most rapid growth. This would make detecting increases in Nkx2-5 difficult in animal models where cardiac hypertrophy is studied after its completion, rather than during its development. Supporting this is the recent finding that human Nkx2-5 (hCSX) mRNA levels are not altered in transplant recipient hearts with end-stage heart failure (30).

Nkx2-5 binds to a specific DNA site, the NK element (NKE) (5'-TNNAGTG-3') (7). Recently, an NKE site in the proximal ANF promoter was found sufficient for cardiac chamber-specific and developmental stage-specific activity (14). [ANF expression is considered a transcriptional marker of cardiac hypertrophy (10).] Similarly, the -cardiac actin promoter is activated by Nkx2-5 via NKE sites operating in tandem with serum response factor to maximally trans-activate when all four native serum response element binding sites are present (6). This suggests there may be a mechanistic link between ANF and -cardiac actin responses to Nkx2-5 levels in vitro. Although it does not prove any mechanistic link, this led us to correlate ANF and -cardiac actin expression with Nkx2-5 in the RVPO model. ANF and -cardiac actin mRNA levels increased in the RVPO cats during hypertrophic growth, paralleling early increases in Nkx2-5 (Figs. 1 and 2). It is interesting that -cardiac actin transcript levels increased in this model as prior studies in rats showed no increase after aortic banding (4). This may be a species difference, since the actin isoform composition of large mammals is different from rodents (31). The absence of detectable ANF in the normal adult ventricle even in the presence of Nkx2-5 suggests the role of multiple factors, of which Nkx2-5 may be but one trans-acting regulator.

Previous investigations have all included the NK homologous homeodomain in hybridization probes (20, 21, 27); therefore, it was felt important to examine specifically for Nkx2-5 independently of other NK homeobox-containing genes. A cDNA probe consisting of 341 nucleotides from the more divergent 5'-end of the coding region also demonstrated upregulation at 2 and 7 days of pressure overload (Figs. 1 and 2) in a similar pattern to the full-length Nkx2-5 hybridization. The possibility of other NK family members being present and/or upregulated is suggested by the differences in the increase between the full-length Nkx2-5 probe and the 5'-341(Nkx2-5) probe, but these cannot be directly compared and this was not formally addressed in these experiments. Because the full-length probe had greater specific activity, this may contribute to differences in signal levels. Also, the changes in Nkx2-5 transcript levels observed may be from either differences in transcription, mRNA stability, or some combination of the two.

Hemodynamic pressure overload in the heart has been shown to evoke certain changes in cardiac transcription. Several gene products are reexpressed that are only otherwise expressed during cardiac development. Examples include -myosin heavy chain (17, 23, 25), ANF (28), -skeletal actin (26), and atrial myosin light chain-1 (16). This suggests that part of an otherwise embryonic cardiac transcriptional program might reactivate in response to hemodynamic load (among other possible stimuli). The upregulation of Nkx2-5 found in this study is compatible with this hypothesis. Future studies will be required to determine if this cardiac homeobox gene product previously thought only active during cardiogenesis has a functional role in adult cardiac growth regulation.