INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

SEVERAL LINES OF EVIDENCE indicate that the cytokine tumor necrosis factor- (TNF-) may contribute to the development and progression of heart disease. It has been shown that the failing myocardium secretes TNF-, and both cardiomyocytes and noncardiomyocytes from the infarcted heart show elevated levels of TNF- gene expression and TNF- protein (9, 15, 25, 27). Moreover, TNF- can induce cardiomyocyte apoptosis and destabilization of atherosclerotic plaques (13, 17). Furthermore, transgenic mice with cardiac-specific overexpression of TNF- develop cardiomyopathy, and TNF- reportedly predicts right ventricular failure after human heart transplantation (1, 11).

Despite advances in both invasive and medical treatment, severe heart failure still carries a poor prognosis (2). The selective inhibition of TNF- using catalytic nucleotides, such as DNA enzymes, might be a potential therapeutic option. The cleavage specificities of these compounds are determined by their hybridizing antisense arms which anneal with the target mRNA in a complementary fashion. Although some studies have highlighted the in vitro activity of these molecules, their in vivo therapeutic potential is mainly unknown (20). Unfortunately, a full exploration of this technology has been impeded by the observation that native DNA is sensitive to nucleases (19, 23). By introducing a phosphorothioate modification into a DNA enzyme specific for TNF- mRNA, we have designed a stable DNA catalyst that retained its full cleavage activity. The therapeutic potential of this enzyme has now been tested in rats with severe congestive heart failure.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DNA enzymes and RNA substrates. The phosphodiester- and phosphorothioate-modified DNA enzymes were synthesized by Eurogentec (Seraing). The short RNA substrate corresponding to the target site was synthesized by Integrated DNA Technology (Coralville, IA). All molecules were polyacrylamide gel purified and 5'-labeled using T4 polynucleotide kinase (Promega; Madison, WI). The larger TNF- mRNA substrate was synthesized by in vitro transcription from a cloned PCR product. In brief, total RNA from lipopolysaccharide-stimulated rat monocytes was reverse transcribed with oligo dT as primer and the TNF- coding sequence was amplified using a TNF- specific forward primer located at 90 nucleotides upstream from the DNA enzyme cleavage site and a reverse primer located near the end of the open reading frame. After PCR, DNA was agarose gel purified and cloned into the pGEM-T vector (Promega). Positive clones were selected and one clone was used as a template for in vitro transcription using the T7 RNA polymerase.

In vitro cleavage activity, stability, and transfection of cells. The cleavage activities of the DNA enzymes were investigated under multiple turnover conditions. Cleavage products were separated by electrophoresis on polyacrylamide gels and visualized by a PhosphoImager (Molecular Dynamics; Sunnyvale, CA). The initial DNA enzyme cleavage rates and turnover numbers were determined from Eadie-Hofstee plots. The 5'-³²P-labeled test molecules were incubated in 50% freshly isolated human serum. Finally, the samples were separated on polyacrylamide gels and analyzed by the ImageQuant software (Molecular Dynamics). Rat monocytes were separated by density gradient centrifugation (Lymphoprep, Nycomed; Oslo, Norway) and plastic adherence. DNA enzymes were complexed to liposomes using N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methylsulfate (DOTAP, Boehringer Mannheim; Mannheim, Germany) to transfect lipopolysaccharide-stimulated cells.

Total RNA preparation and RT-PCR quantification. Total RNA was prepared from transfected and control monocytes/cardiomyocytes with the QuickPrep total RNA extraction kit (Amersham Pharmacia; Piscataway, NJ) and then reverse transcribed using the first-strand cDNA synthesis kit and the oligo-dT primer (Amersham Pharmacia). For PCR, the cDNA was amplified using the TNF- primers: forward 5' ACTCCCA- GAAAAGCAAGCAA 3', and reverse 5' TGGAAGACTCCTCCCAGGTA 3' - amplicon = 688 bp; and the  actin primers: forward 5' GCAAACATC-CCCCAAAGTTG 3', and reverse 5' CTCAACACCTCAAACCACTC 3' - amplicon = 381 bp. After 20 cycles of amplification (94°C 1/min, 56°C 1/min, and 72°C 1/min) a 5-min extension at 72°C followed. Five microliter samples were run on 1.2% agarose gels, visualized by ethidium bromide staining, transferred to nitrocellulose, and then probed with a [³²P]dCTP-labeled TNF- or actin probe and quantified using ImageQuant.

Animals, surgery, and protocol. The protocol was accepted by the local ethical committee. We used male Wistar rats (390-420 g) anesthetized with intraperitoneal barbiturate (50 mg/kg). An acute myocardial infarction was induced by ligating the left coronary artery through a thoracotomy in 40 rats, as previously described (26). In the 24 sham-operated animals, this artery was not ligated. Among the sham animals, perioperative mortality was zero. The perioperative mortality in rats with myocardial infarction was 46%, similar to other studies (18, 21). We also implanted catheters into one carotid and one femoral artery to be exteriorized and secured in the neck and sacrum, respectively. Postoperatively, all rats received bupremorphin (0.08 mg/kg sc).

1

1

Hemodynamic measurements and determination of cytokines. Mean arterial blood pressure (MABP) as well as LVEDP were measured via the indwelling carotid catheter and guided by pressure tracing. We employed the radiolabeled microsphere method to estimate cardiac output (CO) and myocardial blood flow, as described (5). Hemodynamic measurements were made in one set of rats 2 wk after myocardial infarction (Table 1) and in the remaining rats, after either 4 or 6 wk postinfarction.

Preparation of tissue samples and statistics. At the end of the observation period the animals were deeply anesthetized before they were killed with an intraperitoneal injection of barbiturate. We then carefully removed the heart and lungs in toto. The wet weights of these organs were recorded. We then determined the microsphere-related radioactivity and TNF- mRNA in the myocardium. The values are reported as means and SE. Differences were evaluated with the Kruskall-Wallis test, followed by the Bonferroni correction. Two-tailed tests were used and P < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In vitro cleavage activity and stability of the DNA enzyme. The 10-23 DNA enzyme core motif can recognize and cleave RNA sequences at a phosphodiester bond located between an unpaired purine and paired pyrimidine (19). We have found that an RNA site bridging the translation start AUG in the rat TNF- mRNA is accessible to DNA enzyme binding (data not shown). Therefore, a DNA enzyme targeting this site was designed. Figure 1A shows the binding of this enzyme to its target RNA site. To increase the stability, the hydroxy groups of the phosphate backbone within the DNA enzyme antisense arms were replaced with sulfur atoms to make it a phosphorothioate-modified DNA enzyme. The cleavage activities of the unmodified and modified DNA enzyme were tested under multiple turnover conditions. As illustrated in Fig. 1B, both enzymes cleaved the short substrate with similar efficacy. Under our cleavage conditions, the turnover numbers for the unmodified and modified DNA enzymes were similar: 0.80 ± 0.09 and 0.76 ± 0.12 min¹ (n = 4), respectively. Notably, the phosphorothioate-modified DNA enzyme effectively cleaved the in vitro transcribed TNF- mRNA (Fig. 1C) so that most of the mRNA substrate was cleaved within 30 min.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Sequence, cleavage activity, and stability of the DNA enzymes. A: base pairing of the active DNA enzyme with its RNA target site. The cleavage site is indicated by an arrowhead. The subscript s indicates the phosphorothioate linkages in the modified DNA enzyme. B: multiple turnover reaction kinetics of the unmodified and phosphorothioate-modified DNA enzymes. C: cleavage products (CP) of 32P internally labeled TNF- mRNA by the active DNA enzyme. D: DNA enzyme stability was examined by incubating 5'-32P-labeled gel purified enzymes in phosphate-buffered saline containing 50% human serum. Data are from one experiment and are representative of two other experiments.

Cell efficacy of the DNA enzyme. To examine the intracellular activity of the DNA enzymes, rat monocytes were prepared and transfected with the modified DNA enzyme (active). As an inactive control, a phosphorothioate-modified DNA enzyme with reversed antisense arms was used. To evaluate the cleavage activity of the enzymes, the relative amount of TNF- mRNA was determined with RT-PCR and analyzed with a PhosphoImager. As shown in Fig. 2A, the TNF- mRNA level in monocytes treated with the active DNA enzyme was markedly reduced compared with monocytes treated with the inactive DNA. Moreover, TNF-, but not IFN- production, was inhibited by the active DNA enzyme, whereas no apparent effect was obtained by the control enzyme (Fig. 2B), indicating the specificity of the active DNA enzyme for TNF- mRNA. This anti-TNF- activity in rat monocytes is most likely consistent with the DNA enzyme-mediated cleavage activity, because a DNA enzyme with 2'-O-methyl nucleotides within the antisense arms also inhibited TNF- gene expression in these cells (data not shown). We also determined the TNF- mRNA levels in noninfarcted myocardium from experimental rats given either the active or the inactive DNA enzyme. Figure 2C shows that a 4-wk treatment with the active DNA enzyme in vivo substantially reduced TNF- mRNA, hence supporting the data obtained with the monocytes in vitro.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   In vitro activity of the modified DNA enzyme. The active DNA enzyme and its inactive form were complexed with liposomes and delivered to rat monocytes. A: TNF- mRNA was quantitated by RT-PCR. B: TNF- and IFN- proteins present in the supernatant were determined with ELISA kits. The baseline concentrations of TNF- and interferon- (IFN-) were 7.8 ± 1.8 and 17.8 ± 3.3 ng/ml, respectively. C: TNF- mRNA was quantified from the myocardium of experimental rats treated with the active or inactive DNA enzyme. Values are means + SE based on three experiments and expressed as a percentage of controls (cells given liposomes only).

Induction of severe congestive heart failure. Data of hemodynamic parameters, weights, and TNF- concentrations obtained in sham and experimental rats 2 wk after heart surgery are shown in Table 1. In marked contrast to the sham animals, the experimental rats with myocardial infarctions developed after 2 wk a severe congestive heart failure evidenced by a substantial increase in LVEDP concomitant with a reduction in CO and MABP as well as augmented lung and heart weights. Paralleling these changes, myocardial blood flow declined in the experimental rats over this time period. Finally, we obtained elevated levels of TNF- in coronary sinus blood from the experimental rats. In eight rats, the concentration of TNF- in the general circulation increased from 2.7 ± 0.8 ng/ml at baseline to 6.7 ± 1.5 ng/ml (P < 0.05) 2 wk after heart surgery. Treatment with the active DNA enzyme reduced significantly the TNF- concentration to 2.4 ± 0.9 ng/ml after an additional 4 wk.

The DNA enzyme improves cardiac performance in rats with heart failure. To study the effect of the DNA enzyme in vivo, we implanted osmotic minipumps intraperitoneally 2 wk after heart surgery. Immediately before implantation of the peritoneal osmotic minipumps, LVEDP averaged 8.0 and 31.2 mmHg in the sham and experimental rats, respectively. Sham and experimental rats were then randomly allocated to receive either the active or inactive DNA enzyme or only liposomes before they were followed for an additional 2 or 4 wk before measurements were made. The LVEDP decreased markedly in experimental rats treated with the active DNA enzyme in contrast to rats treated with the inactive form (Fig. 3A), and after 4 wk of treatment LVEDP returned to normal values in each of the experimental rats treated with the active enzyme. We also detected increases in both CO and MABP in experimental rats treated with the active DNA enzyme, and by 2 wk neither CO nor MABP were significantly different from those of the sham animals receiving the active DNA enzyme (Fig. 3, B and C). While the lung weights declined toward baseline, the heart weights remained elevated despite treatment with the active DNA enzyme (Fig. 3, D and E). Importantly, blood flow to the myocardium increased on treatment with the active but not with the inactive enzyme (Fig. 3F). A 10-fold increase in the delivery of the active DNA enzyme (10 µg · kg¹ · h¹) did not cause further changes in any of the measured parameters. Moreover, the administration of only liposomes to experimental rats did not affect their hemodynamic performance (data not shown). Furthermore, hemodynamic and weight data obtained in sham-operated rats receiving the active enzyme (Fig. 3) did not differ from those of sham-operated rats receiving the inactive enzyme or only liposomes (data not shown).

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Improved hemodynamic performance in rats with congestive heart failure that were treated with the active DNA enzyme. Treatment started 2 wk after induction of acute myocardial infarction and continued for 2 or 4 wk. Values for left ventricular end-diastolic pressure (LVEDP) (A), cardiac output (B), mean arterial blood pressure (MABP) (C), and lung weight (D), improved on enzymatic treatment, whereas the heart weight remained elevated (E). Myocardial blood flow also increased in rats given this enzyme (F). Values are means + SE. Each column represents data from 8 rats. *P < 0.05; **P < 0.001 compared with sham animals.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   A: concentration of TNF- in coronary sinus blood decreased on treatment with the active DNA enzyme. Values are means + SE. Each column represents data from 8 rats. *P < 0.05; **P < 0.001 compared with sham animals given the active DNA enzyme. B: recovery of the DNA enzymatic cleavage activity from rat myocardial tissue. Samples of the myocardium were homogenized and the cleavage activities of the active and inactive enzymes were assessed using 32P-labeled substrates. Results are from 1 of 3 representative experiments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Severe congestive heart failure developed in the rats after induction of acute myocardial infarction. This was evidenced by 1) an almost tripling of LVEDP in experimental compared with sham-operated rats; 2) substantial reductions in MABP, CO, and myocardial blood flow in the experimental rats; and 3) development of a marked pulmonary edema. These hemodynamic changes compare favorably with a recent evaluation of this rat model (24). Notably, it was shown that a cut-off value of 15 mmHg for LVEDP predicted heart failure with 100% sensitivity and specificity when compared with echocardiographic measurements (24).

We have now designed a nuclease-resistant DNA enzyme that maintained its cleavage activity and inhibited TNF- gene expression in vitro as well as in vivo. Notably, a marked improvement of the cardiac performance was achieved in rats with congestive heart failure after a continuous supply of this DNA enzyme. Hence, LVEDP in the treated rats was normalized after 4 wk of treatment. In addition, the lung weight declined while the coronary perfusion increased, thus further indicating recovery of a failing myocardium. Concomitant with these hemodynamic improvements were the consistent decreases in the cardiac content of both TNF- mRNA and release of TNF- protein. In addition, the cleavage activity of the active DNA enzyme was present throughout the observation period. Care was taken to ensure that experimental rats with similar baseline LVEDP values were equally assigned to treatment with either the active or inactive DNA enzyme. In addition, all hearts were carefully examined postmortem and there was no apparent difference of the infarcted zones among experimental rats treated with either the active or inactive DNA enzyme. Hence, randomization criteria were fulfilled. Collectively, our findings suggest that improved cardiac function was ultimately due to the selective catalysis of TNF- mRNA by this active DNA enzyme.

Therapeutic modulation of specific gene expression using DNA enzymes requires that these molecules can access their RNA targets without interference with either their delivery to the cells or their inherent stability. We used a phosphorothioate-modified DNA enzyme that was complexed with cationic liposomes (22). A substantial amount of cleavage activity could be retrieved from the myocardium of the treated rats, indicating that neither a limited delivery nor enhanced degradation of the DNA enzyme had occurred. Moreover, addition of either the active or inactive DNA enzyme to either experimental or to sham-operated animals did not alter their hematological status or their body weights. Thus these molecules did not convey any apparent toxicity.

Independent observations support the notion that TNF- is involved in cardiac disease, but its exact role in the development of the failing myocardium has not been defined. Although it has been proposed that TNF- might confer a cytoprotective effect among myocardial cells exposed to environmental stress, the salutary effects are most likely contravened on prolonged exposure to excessive concentrations of TNF- within the heart (12). The inhibitory effects of TNF- on cardiac function could be due to its proinflammatory actions leading to increased inflammation and/or adverse remodeling of the myocardium (11, 15). Although the impaired cardiac function in the experimental rats improved on addition of the active DNA enzyme, the cardiomegaly was not reversed. Treatment with the DNA enzyme could have affected a remodeling process unrelated to the increased heart mass occurring in the early postinfarction period, possibly through stimulation of cardiac fibroblasts and metalloproteinases (3, 8). Alternatively, TNF- might affect postinfarction remodeling of the heart due to its negative ionotropic effects by uncoupling of -adrenergic receptors from adenyl cyclase or by downregulation of sarcoplasmic reticulum Ca²⁺-ATPase (4, 11). Furthermore, the suppressive activity on cardiac myocyte contractility and remodeling by TNF- might be related to upregulation of other proinflammatory cytokines such as interleukin-1 and -6 (15, 27). TNF- also activates the vasoactive cytokine endothelin which probably contributes to the development of heart failure and myocardial remodeling (6, 18, 26). Irrespective of the exact mechanism of action, inhibition of TNF- is a likely therapeutic option in advanced heart failure, because the myocardial content of both TNF- mRNA and TNF- protein are increased in this disease (9, 15, 25, 27). Although neutralization of preformed TNF- can reverse its negative inotropic effects, repression of TNF- synthesis has apparently not been previously attempted (10). Our data highlight the pivotal role of TNF- in the pathogenesis of advanced congestive heart failure. Because TNF- most likely plays a pathogenetic role in this and other debilitating diseases such as rheumatoid arthritis, inflammatory bowel disease, and leukemia (7, 14, 16), the advent of DNA enzymes might offer a novel strategy for selective inhibition of the detrimental effects of TNF-.