Targeted Gene Silencing of Tumor Necrosis Factor Attenuates the Negative Inotropic Effects of Lipopolysaccharide in Isolated Contracting Cardiac Myocytes

Abstract

Bacterial endotoxin (lipopolysaccharide) depresses cardiovascular function; however, the mediators and signaling pathways that are responsible for the negative inotropic effects of lipopolysaccharide are not fully known.

We used RNA interference to determine the relative role of tumor necrosis factor with respect to mediating the negative inotropic effects of lipopolysaccharide in isolated cardiac myocytes. Cardiac myocyte cultures were treated with lipopolysaccharide in the presence or absence of small interfering RNAs (siRNA) for tumor necrosis factor. We examined the effects of tumor necrosis factor siRNA on lipopolysaccharide-induced tumor necrosis factor messenger RNA (mRNA) and protein biosynthesis, as well as the negative inotropic effects of lipopolysaccharide in isolated contracting cardiac myocytes.

Treatment of adult cardiac myocyte cultures with tumor necrosis factor siRNA significantly attenuated lipopolysaccharide-induced tumor necrosis factor mRNA and protein biosynthesis, whereas transfection with a double-stranded RNA that does not target mammalian mRNA had no effect. Pretreatment with tumor necrosis factor siRNA significantly attenuated, but did not abrogate, the lipopolysaccharide-induced decrease in sarcomere shortening in isolated contracting cardiac myocytes. In contrast, tumor necrosis factor siRNA had a comparatively smaller effect on improving sarcomere shortening once the negative inotropic effects of lipopolysaccharide were fully established.

These results suggest that tumor necrosis factor plays an important upstream role in lipopolysaccharide-induced negative inotropic effects in isolated contracting cardiac myocytes and that other molecular mechanisms are responsible for the decrease in sarcomere shortening after sustained lipopolysaccharide signaling.

Bacterial endotoxin (lipopolysaccharide [LPS]) triggers the release of a cascade of endogenous mediators and induces hypotension, multiorgan failure, and death from sepsis and septic shock.^1,2 Although it has long been recognized that LPS depresses cardiovascular function, the precise signaling pathways that are responsible for mediating the negative inotropic effects of LPS have not been fully elucidated. Recent studies from this and other laboratories have shown that the deleterious effects of LPS are mediated, at least in part, through the toll-like receptor 4 (TLR4).^3–6 Given that activation of the TLR4 pathway leads to the increased expression of pro-inflammatory cytokines, which in turn is sufficient to produce negative inotropic effects, we sought to use the recently developed technique of RNA interference (RNAi)⁷ to determine what role, if any, the pro-inflammatory cytokine tumor necrosis factor (TNF) plays in mediating the negative inotropic effects of LPS in isolated contracting cardiac myocytes. Here, we show for the 1st time that RNAi can be used to selectively knock down TNF gene expression in LPS-stimulated myocytes and that selective TNF-gene knockdown attenuates the deleterious effects of LPS in isolated contracting adult cardiac myocytes.

Materials and Methods

Validation of siRNA Uptake in Cultured Cardiac Myocytes

To confirm transfection of siRNA into cardiac myocytes, we used siGLO RISC-Free siRNA (Dharmacon, Inc., part of Thermo Fisher Scientific; Lafayette, Colo), a stable, non-targeting control siRNA labeled with a DY-547 (Cy-3 equivalent) fluorescent label that provides a reliable visual display of transfection success. For siRNA transfection into cardiac myocytes, we used GeneSilencer siRNA Transfection Reagent (Genlantis, a division of Gene Therapy System, Inc.; San Diego, Calif), a proprietary cationic lipid formulation that provides high-efficiency siRNA transfer into mammalian cells. Briefly, freshly isolated feline cardiac myocytes were plated at a density of 8 × 10⁵ cells per well (∼50%–70% confluence) on laminin-coated 6-well plates that contained phenol-free M199 medium (GIBCO®, Invitrogen Corporation; Carlsbad, Calif) supplemented with insulin/transferrin/selenium (cellgro, Mediatech, Inc.; Manassas, Va) and 0.1% human serum albumin (Sigma-Aldrich; St. Louis, Mo). Cell cultures were incubated overnight at 37°C at 95% O₂, 5% CO₂. Two hours before transfection with siRNA, the cell culture medium was changed and replaced with fresh medium M199 that was free of human serum albumin. At the time of transfection, 5 μL of siRNA and/or 5 μL of GeneSilencer were added to each well, and the cells were incubated at 37°C at 95% O₂, 5% CO₂ for 30 minutes. After 30 minutes, the cells were examined for the presence or absence of intracellular DY-547-labeled siRNA by means of an Olympus IX71 inverted microscope (Olympus America Inc.; Center Valley, Pa) at x40 magnification, using a tetramethyl rhodamine iso-thiocyanate red filter set (540–565 nm).

Generation of Feline dsRNA and siRNA

Total RNA from feline spleen was first isolated and used as a template to obtain complementary DNA (cDNA) for feline TNF. The cDNA for TNF was synthesized by means of the real-time polymerase chain reaction, using the sequence information from GenBank (Acc No: NM-001009835) and validated by DNA sequencing. In order to generate a template for synthesizing double-stranded RNA (dsRNA) in vitro, a ∼399-base pair cDNA template for feline TNF was further PCR-synthesized by use of chimeric oligonucleotides, which contained a T7 promoter sequence in addition to the TNF sequence (5′ TTA ATA CGA CTC ACT ATA GGG AGA AGG ATC ATC TTC TCG AAC TCC GAG TGA CAA GCC 3′ and 5′ TTA ATA CGA CTC ACT ATA GGG AGA CCC TTC AGC TTC GGG GTT TGC TGG 3′). The size of the cDNA template for TNF with opposing T7 promoters at the 5′ end of each strand is ∼399 base pairs. The dsRNA for TNF was generated using the commercially available MEGAscript® RNAi kit (Ambion, an Applied Biosystems business; Austin, Tex). The resultant dsRNA was subsequently diced and column-purified to obtain TNF siRNA by means of the Dicer siRNA Generation Kit (Genlantis), and stored at −20°C.

Effect of siRNA on Lipopolysaccharide-Induced Tumor Necrosis Factor Protein and mRNA Biosynthesis

Previously, we have shown that treatment of cultured adult cardiac myocytes with 10 to 100 μg/mL of LPS results in a 3- to 5-fold increase in TNF levels that is maximal at 6 hours and declines thereafter.⁸ To determine the effects of TNF mRNA silencing on LPS-induced TNF biosynthesis, adult feline cardiac myocytes were isolated and cultured as described.⁹ On day 1 in culture, the myocytes were pretreated with TNF siRNA (500 ng/mL), nautilus dsRNA (25 μg/mL)—a nonmammalian dsRNA that does not target mammalian mRNA (negative control)—or diluent for 4 hours, followed by treatment with 25 μg/mL LPS (obtained from Escherichia coli Serotype 026:B6; Sigma-Aldrich) for 6 hours. The TNF biosynthesis was assayed in the cell culture media by use of an enzyme-linked immunosorbent assay (BioSource™ cytoscreen US Ultrasensitive™, catalog #KHC3013; Invitrogen), with a detection range of 0.5 to 32 pg/mL, as described.⁸ Centricon® micro-concentrators (Amicon®, Millipore; Billerica, Mass) were used to concentrate the cell culture medium by ultrafiltration through a low-adsorption, hydrophilic membrane with a cutoff value of 10,000 MW. For the analysis of LPS-induced TNF mRNA levels, cultured cardiac myocytes were treated for 4 hours either with TNF siRNA or nautilus dsRNA, as described above, or with diluent. The myocyte cultures were then stimulated with LPS for 90 minutes. The level of expression of TNF and L32 mRNA were determined using a custom-designed multiprobe riboquant ribonuclease protection assay (Pharmingen, BD Biosciences [San Jose, Calif], a segment of Becton, Dickinson and Company [Franklin Lakes, NJ]), as described.¹⁰

Effect of siRNA on Lipopolysaccharide-Induced Cardiac Myocyte Function

Sarcomere shortening of freshly isolated adult cardiac myocytes was evaluated by means of the IonOptix MyoCam™ system (IonOptix Corporation; Milton, Mass), as described.¹¹ Freshly isolated myocyte cultures were pretreated for 4 hours with TNF siRNA (500 ng/mL), nautilus dsRNA (25 μg/mL), or diluent, and were then stimulated with 25 μg/mL LPS for 6 hours. Suspensions of cells were then placed in a chamber mounted on the stage of an inverted microscope and superfused with a buffer, containing (in mM) the following: HEPES, 20 (pH, 7.4); calcium chloride, 1; sodium chloride, 137; potassium chloride, 5.4; dextrose, 15; magnesium sulfate, 1.3; and sodium dihydrogen-phosphate, 1.2. Experiments were conducted at 37°C. The myocytes were field-stimulated at a frequency of 0.5 Hz; changes in sarcomere length during contraction were captured using IonOptix software. In parallel control experiments, myocytes were pre-incubated with etanercept (30 μg/mL), a soluble TNF antagonist, for 4 hours, and were then treated with 25 μg/mL LPS for 6 hours. To determine whether the effects of TNF siRNA were secondary to disruption of TNF signaling, TNF siRNA-treated cells (500 ng/mL) were stimulated with TNF (200 U/mL) for 30 minutes, before the evaluation of cell motion. Finally, to determine whet her TNF siRNA was sufficient to reverse LPS-induced defects in sarcomere shortening, cardiac myocyte cultures were treated with LPS (25 μg /mL) for 6 hours, and this was followed by treatment with TNF siRNA (500 ng/mL) for 2 hours, before evaluation of cell motion.

Statistical Analysis

All values are expressed as mean ± SEM. One-way analysis of variance was used to test for differences in TNF concentrations and sarcomere shortening in the different experimental groups. Where appropriate, post hoc testing was performed using a Fisher exact test to detect differences between groups. Significant differences were said to exist at P < 0.05.

Results

Validation of siRNA Uptake in Isolated Cardiac Myocytes

To verify transfection of TNF siRNA into cardiac myocytes, we used DY-547-labeled siRNA in cultured adult cardiac myocytes. Figure 1C shows that DY-547-labeled siR NA did not enter isolated cardiac myocytes in the absence of a transfection reagent, whereas the same siRNA was readily detectable in the cardiac myocytes within 30 minutes in the presence of GeneSilencer, a cation-ic lipid transfection reagent (Fig. 1D). The diluent and GeneSilencer alone did not show any background fluorescence (Figs. 1A and 1B). This figure is representative of 3 separate experiments.

Fig. 1 Validation of small interfering RNA (siRNA) uptake in isolated cardiac myocytes. Cardiac myocytes were maintained in culture for 24 hours and were treated with A) diluent, B) “Gene-Silencer” alone, C) DY-547-labeled siRNA alone, and D) Gene-Silencer + DY-547-labeled siRNA. Myocytes were studied at ×400 magnification. In order to reveal the location of the myocytes, the cells were first photographed with the aid of low-level bright-field illumination and the tetramethyl rhodamine iso-thiocyanate red filter set (540–565 nm), and then by fluorescence microscopy in the absence of bright-field illumination. The arrows in panel D denote the presence of DY-547-labeled siRNA in the cultured feline cardiac myocytes.

Effect of siRNA on Lipopolysaccharide-Induced Tumor Necrosis Factor Protein and mRNA Biosynthesis

To determine whether siRNA interference was sufficient to disrupt LPS-induced TNF biosynthesis, we examined TNF mRNA and protein biosynthesis in the presence and absence of TNF siRNA. The ribonuclease protection assay depicted in Figure 2A shows that TNF mRNA was detectable in LPS-treated cardiac myocytes, whereas pretreatment with TNF siRNA knocked down LPS-induced TNF mRNA expression. In contrast, transfection of cardiac myocytes with nautilus dsRNA had no effect on LPS-induced m R NA biosynthesis; L32 mRNA levels were not affected by TNF siRNA. Figure 2B shows that the baseline levels of supernatant TNF were barely detectable in diluent-treated myocytes, whereas treatment with LPS resulted in a significant (P <0.0001) 6-fold increase in TNF production, consistent with our prior observations.⁸ Transfection of myocytes with TNF siRNA resulted in a significant (P <0.002) decrease in peak TNF levels after LPS stimulation. However, as shown, the levels of TNF in the supernatant of the LPS-treated myocytes were not reduced to baseline levels. In contrast, transfection with an irrelevant dsRNA had no significant effect (P = 0.20) on LPS-induced TNF biosynthesis in cardiac myocytes. Taken together, these results suggest that TNF siRNA is sufficient to attenuate LPS-induced TNF biosynthesis.

Fig. 2 Effect of tumor necrosis factor (TNF) small interfering RNA (siRNA) on lipopolysaccharide-induced TN F biosynthesis. A) Ribonuclease protection assay shows TNF gene expression knockdown after lipopolysaccharide (LPS) challenge in the presence and absence of siRNA (representation of 3 separate experiments); nautilus double-stranded RNA (dsRNA) was used as a negative control. B) Levels of immunoreactive TNF in the supernatants of LPS-stimulated myocyte cultures (6 hr) pretreated with diluent or TNF siRNA; nautilus dsRNA was used as a negative control (n=5 myocyte cultures per group). ^* = P < 0.01 compared with baseline; ^** P < 0.01 compared with diluent

Effect of siRNA on Lipopolysaccharide-Induced Cardiac Myocyte Dysfunction

To evaluate the potential functional significance of the siRNA TNF knockdown in LPS-treated cardiac myocytes, we examined the effects of TNF siRNA pre-treatment in LPS-treated isolated contracting cardiac myocytes. Figure 3A shows that treatment with LPS led to a significant (P <0.0001) decrease in sarcomere shortening, consistent with prior studies.^5,12–19 In contrast, pretreatment with TNF siRNA significantly (P <0.01) attenuated the LPS-induced decrease in sarcomere shortening, although it did not restore sarcomere shortening to baseline values (P <0.05 compared with baseline). As an additional positive control experiment to determine whether TNF was responsible for the LPS-induced decrease in sarcomere shortening, cardiac myocytes were pretreated with 30 μg/mL of etanercept before LPS stimulation. As shown, treatment with etanercept prevented the LPS-induced decrease in sarcomere shortening (P >0.05 compared with LPS alone). Next, to determine whether RNA interference disrupted the TNF-induced signaling pathway, we treated cardiac myocytes concurrently with TNF siRNA and exogenous TNF. The addition of exogenous T NF abrogated the effects of RNA interference, which suggested that the prevention of LPS-induced decrease in sarcomere shortening by TNF siRNA was a consequence of the decreased TNF biosynthesis, rather than nonspecific interference with TNF-mediated signaling after LPS provocation. Treatment with a nautilus dsRNA had no effect on the LPS-induced decrease in sarcomere shortening.

Fig. 3 Effect of tumor necrosis factor (TNF) small interfering RNA (siRNA) on lipopolysaccharide-induced negative inotropic effects in isolated contracting cardiac myocytes. A) Isolated cardiac myocytes were pretreated with diluent, TNF siRNA, etanercept, and TNF siRNA + exogenous TNF before stimulation with lipopolysaccharide (LPS) for 6 hours; nautilus double-stranded RNA (dsRNA) was used as a negative control (n ≥20 cells per group). B) Isolated cardiac myocytes were treated with lipopolysaccharide (LPS) for 6 hours and were then treated with diluent or TNF siRNA for an additional 2 hours; nautilus dsRNA was used as a negative control (n ≥20 cells per group). ^* = P <0.01 compared with baseline; ^** P <0.05 compared with diluent

To determine whether siRNA was capable of decreasing the negative inotropic effects of LPS once they were fully established, we treated the cells with siRNA after LPS treatment (“cell rescue”). Figure 3B shows that treating myocytes with TNF siRNA after the LPS-induced negative inotropic effects were fully established resulted in a partial (P <0.004 compared with LPS) reversal of the LPS-induced decrease in sarcomere shortening.

Discussion

In this study, we used RNA interference to demonstrate that TNF biosynthesis is necessary for the early negative inotropic effects of LPS in isolated contracting cardiac myocytes. Two lines of evidence support this. First, treatment of adult myocyte cultures with TNF siRNA significantly attenuated LPS-induced TNF mRNA and protein biosynthesis (Fig. 2). In contrast, transfection with a dsRNA that does not target mammalian mRNA had no effect on LPS-induced TNF mRNA or protein biosynthesis. Second, treatment of adult myocyte cultures with TNF siRNA significantly attenuated, but did not abrogate, the LPS-induced decrease in sarcomere shortening (Fig. 3A). In contrast, pretreating the cells with etanercept, a soluble TNF antagonist, completely abrogated the LPS-induced decrease in sarcomere shortening. This latter finding is consistent with the observation that RNA interference did not completely suppress TNF protein biosynthesis in isolated contracting cardiac myocytes (Fig. 1). Of more importance is the fact that the beneficial effects of TNF siRNA on the LPS-induced decrease in sarcomere shortening could be overcome by treating the cells concurrently with TNF, which suggested that TNF siRNA did not affect TNF signaling pathways. Although TNF siRNA was able to significantly improve sarcomere shortening in LPS-treated cardiac myocytes, this effect was relatively modest (Fig. 3B). Taken together, these observations suggest that TNF plays an important upstream role in LPS-induced negative inotropic effects on isolated contracting cardiac myocytes. However, our results also suggest that once the LPS-induced negative inotropic effects are established, other molecular mechanisms are responsible for the decrease in sarcomere shortening.

Lipopolysaccharide-Induced Negative Inotropic Effects

Table I summarizes the studies that have examined the effects of LPS in myocyte contractility in vitro. A sindicated, most of these studies have shown that LPS provokes negative inotropic effects in isolated contracting cardiac myocytes through a pathway that involves up-regulation of inducible nitric oxide synthase. More over, in most of these studies the isolated cardiac myocytes have been pretreated with LPS for 6 hours, which is sufficient to up-regulate inducible nitric oxide synthase (NOS₂) in cardiac myocytes. In contrast with these in vitro studies, a prior study from this laboratory showed that the superfusate (recirculating) from LPS-treated feline hearts perfused by the Langendorff technique was sufficient to provoke negative inotropic effects in isolated contracting cardiac myocytes, and that the negative inotropic effects of the superfusate were abrogated by pretreatment with a neutra lizing TNF antibody.²⁰ Similar findings were reported by Stamm and coworkers,²¹ who noted that a neutralizing TNF antibody inhibited the negative inotropic effects of LPS in Langendorff-perfused rat hearts. Therefore, the results of our current study both confirm and extend these prior studies a nd show t hat siRNA-mediated gene knockdown of TNF attenuates the negative inotropic effects of LPS in isolated contracting cardiac myocytes. Our observation that RNA interference had very modest effect on reversing the negative inotropic effects of LPS, once they were fully established (Fig. 3B), is consistent with the observations of others who have proposed that NOS₂ is responsible for the negative inotropic effects of lipopolysaccharide after 6 hours of treatment with LPS. Indeed, it is well recognized that TNF is sufficient to upregulate NOS₂ in isolated contracting cardiac myocytes, and that TNF-induced NOS₂ biosynthesis mediates the late negative inotropic effects of LPS.^22–24 Table I further shows that, in 2 studies in which the myocyte preparations were serum star ved, LPS provoked positive inotropic responses in isolated contracting myocytes.^18,19 Although the mechanism for this finding was not delineated in these studies, it is interesting to note that TNF can provoke positive inotropic effects in cardiac myocytes, depending on the intracellular glutathione status, which regulates neutral sphingomyelinase activity,²⁵ which in turn has been implicated in mediating the immediate negative inotropic effects of TNF.²⁶ Further studies will be necessary to determine whether the biphasic effects of LPS on myocyte contractility are dependent on the redox status of the cell.

TABLE I. Effects of Lipopolysaccharide on Isolated Contracting Cardiac Myocytes

Conclusion

In summary, this study shows for the 1st time that targeted gene silencing of TNF with RNA interference can be used to prevent negative inotropic effects of LPS in isolated contracting cardiac myocytes, and therefore suggests that TNF is an important upstream signaling molecule for mediating the negative inotropic effects of LPS.

Acknowledgment

The authors would like to acknowledge the technical assistance of Ping Yang.