Report

Dilated Cardiomyopathy and Heart Failure Caused by a Mutation in Phospholamban

See allHide authors and affiliations

Science  28 Feb 2003:
Vol. 299, Issue 5611, pp. 1410-1413
DOI: 10.1126/science.1081578

Abstract

Molecular etiologies of heart failure, an emerging cardiovascular epidemic affecting 4.7 million Americans and costing 17.8 billion health-care dollars annually, remain poorly understood. Here we report that an inherited human dilated cardiomyopathy with refractory congestive heart failure is caused by a dominant Arg → Cys missense mutation at residue 9 (R9C) in phospholamban (PLN), a transmembrane phosphoprotein that inhibits the cardiac sarcoplasmic reticular Ca2+–adenosine triphosphatase (SERCA2a) pump. Transgenic PLNR9C mice recapitulated human heart failure with premature death. Cellular and biochemical studies revealed that, unlike wild-type PLN, PLNR9C did not directly inhibit SERCA2a. Rather, PLNR9C trapped protein kinase A (PKA), which blocked PKA-mediated phosphorylation of wild-type PLN and in turn delayed decay of calcium transients in myocytes. These results indicate that myocellular calcium dysregulation can initiate human heart failure—a finding that may lead to therapeutic opportunities.

Heart failure is the leading cause of human morbidity and mortality (1). Reduced contractile function and pathological remodeling are recognized clinical hallmarks of heart failure, but the critical early events that impair myocyte performance are largely undefined (2). Intracellular Ca2+ handling is the central coordinator of cardiac contraction and relaxation (3). Contraction begins with sarcoplasmic reticulum (SR) release of Ca2+ into the cytosol via the ryanodine receptor; relaxation occurs with SR Ca2+ reuptake through the Ca2+ adenosine triphosphatase (ATPase) SERCA2a pump. Phospholamban (PLN), an abundant, 52–amino acid transmembrane SR phosphoprotein (4), regulates the Ca2+ATPase SERCA2a. SERCA2a activity is decreased in human heart failure (5, 6), but whether this is a primary or secondary process that reflects changes in SERCA2a, PLN, and/or other molecules is unknown. Studies in mice suggest that PLN has a fundamental role: PLN protein levels correlate with cardiac contractile parameters (7, 8), superinhibitory PLN molecules impair heart function (9–11), and PLN ablation (12) rescues a mouse heart failure model (13).

To investigate the role of PLN in human heart disease, we sequenced the gene in 20 unrelated individuals with inherited dilated cardiomyopathy and heart failure. In one sample, we identified a C → T missense mutation at nucleotide 25, which encodes an Arg → Cys substitution at codon 9 (R9C) (Fig. 1) in the cytosolic PLN domain. No other sequence variants were found. Clinical evaluations and medical records of this proband's family indicated autosomal dominant dilated cardiomyopathy (Fig. 1A). Affected individuals had increased chamber dimensions and decreased contractile function at age 20 to 30 years, with progression to heart failure within 5 to 10 years after symptom onset. Congestive heart failure was severe in 12 individuals, necessitating cardiac transplantation in four. The average age at death of affected individuals was 25.1 ± 12.7 years. We concluded that the R9C mutation caused disease because it cosegregated with affection status [maximal lodscore (logarithm of odds ratio for linkage) = 4.04, θ = 0], R9C altered a highly conserved residue (fig. S1), and R9C was absent from more than 200 normal chromosomes (14).

Figure 1

Cosegregation of PLNR9C and dilated cardiomyopathy in a large family. (A) Pedigree showing clinical status (circle, female; square, male; solid symbol, affected; clear symbol, unaffected; slashed symbol, death) and genetic status (+, PLNR9C/+; –, PLNwt). Clinical status was assessed without knowledge of genotype. Subjects III-2, III-5, III-8, and III-13 underwent heart transplantation. Individual III-11 (arrow) was the proband. (B) DNA sequence analysis of thePLN gene in the proband revealing a cytosine-to-thymidine substitution at nucleotide 25. The color of the trace reflects the presence of a specific nucleotide residue: C, blue; A, green; T, red; and G, black.

To study the functional consequences of this mutation, we created transgenic mice (Fig. 2) that express PLNR9C under the control of the α-cardiac myosin heavy chain promoter (fig. S2) (15). Two independent mouse lines (designated TgPLNR9C) selected for detailed analysis expressed transgene levels comparable to previously generated transgenic mice (16) that express wild-type PLN (designated TgPLNwt) (fig. S3). Cardiac phenotypes in TgPLNR9C lines were identical: Biventricular cardiac dilation began at age 4 months (Fig. 2, A and B), and as in humans with PLNR9C, dilated cardiomyopathy was rapidly progressive. As heart chamber size increased, TgPLNR9C mice developed diminished left ventricle contractility (Fig. 2B) and symptoms of terminal heart failure: dyspnea, lethargy, and peripheral cyanosis. Death occurred on average at age 20.9 ± 5.5 weeks (n = 24), with survival ranging between 12 and 35 weeks. Autopsy of TgPLNR9C mice with terminal heart failure revealed cardiomegaly, prominent left atria and massive left ventricular dilation (Fig. 2, A and B), pleural and pericardial effusions, and liver enlargement. Ventricular histopathology from humans and mice expressing PLNR9C (Fig. 2C) demonstrated myocyte enlargement (without disarray) and extensive interstitial fibrosis. These abnormalities were not observed in TgPLNwtmice (Fig. 2A) (14, 16, 17).

Figure 2

Characterization of hearts expressing PLNR9C. (A) Hearts of strain-matched wild-type (wt), TgPLNwt, and TgPLNR9C mice at 4 months of age, showing enlargement of TgPLNR9C hearts only. (B) Upper panels: In vivo assessment of left ventricular diameters and function by echocardiography (M-mode). End-systolic diameters (ESD) and end-diastolic diameters (EDD) were increased in the TgPLNR9C mice. Lower panels: Ex vivo cross sections show biventricular dilatation of the TgPLNR9C heart (right). (C) Histopathology of left ventricular heart tissue from 12-week-old, strain-matched TgPLNwt and TgPLNR9C mice and a human heterozygous for PLNR9C. Upper panels: Hematoxylin and eosin (H&E) staining reveals normal tissue in TgPLNwt mice and enlarged myocytes and nuclei in the TgPLNR9Cmouse and human PLNR9C/+ tissue. Lower panels: Collagen staining (Masson's trichrome) shows massive interstitial fibrosis in PLNR9C mouse and human hearts. All images are shown at equal magnification. Scale bar, 100 μm.

We also investigated the functional consequences of the R9C mutation in a human kidney cell (HEK-293) culture system (18,19). Calcium-dependent Ca2+ uptake, which provides a measure of SERCA2a affinity for Ca2+, was monitored in microsomal fractions of HEK-293 cells coexpressing SERCA2a and PLNR9C or PLNwt (18, 19). SERCA2a was significantly less inhibited by PLNR9C than by PLNwt, but when HEK-293 cells coexpressed PLNwtand PLNR9C, the normal inhibitory function of PLNwt dominated (table S1). Because these data did not indicate a dominant mechanism by which the PLNR9C mutation caused disease, we investigated whether PLNR9C altered phosphorylation of PLNwt, thereby affecting SERCA2a activity. The phosphorylation state of PLN determines its inhibitory activity in vivo; under physiologic conditions, ∼50% of PLN is phosphorylated and unable to inhibit SERCA2a (4, 20). Physiologic regulation of SERCA2a activity by PLN is principally mediated by β-adrenergic activation of protein kinase A (PKA), which phosphorylates Ser16(8). PLN phosphorylation was measured by incorporation of 32P into HEK-293 cells expressing PLNR9C or PLNwt. Because R9C disrupts the PLN epitope recognized by anti-PLN antibody 1D11 (14), we fused the FLAG epitope (17) to the NH2-terminus of PLN (NF-PLN) so as to identify expressed protein by antibody M2. NF-PLNwt, but not NF-PLNR9C, was readily phosphorylated by PKA (Fig. 3A).

Figure 3

PLNR9C traps PKA and prevents PLNwt phosphorylation. (A) PLNR9C is not phosphorylated by PKA. Autocardiography of 32P-phosphorylated microsomal fractions from HEK-293 cells expressing NF-PLNwt or NF-PLNR9C and SERCA2a. NF-PLNwt or NF-PLNR9C expression was monitored by Western blot. (B) PLNR9Cinhibits PLNwt phosphorylation by PKA. Western blots of phosphorylated microsomes from HEK-293 cells transfected with PLNwt and 0 to 8 μg of NF-PLNR9C were probed with phosPLNwtand FLAG (NF-PLNR9C) antibodies. (C) No phosphorylated PLN (phosPLN) in hearts with PLNR9C. Two-dimensional gel electrophoresis and Western blots of human (UA, unaffected; A, PLNR9C-affected) and 6-week-old mice (wt and TgPLNR9C) left ventricular extracts, probed with antibody 1D11, which detects only PLNwt. (D) PLN pentamers containing PLNR9C trap PKA. Microsomes from HEK-293 cells transfected with PLNwt and 0 to 8 μg of NF-PLNR9C were precipitated with antibody 1D11 and probed with antibody to PKA. PKA in pentamers increases with PLNR9C content, except when PLNR9Chomopentamers (PLNwt:PLNR9C ≤ 4) predominate. (E) Human myocardium with PLNR9C(A), but not PLNwt (UA), shows PKA trapping. Samples analyzed as in (D). (F) Total PKA activity in TgPLNR9C and wt myocyte lysates assayed by32P-labeling of a PKA-specific substrate. (G) Western blots for phospho-(Ser/Thr) PKA substrates in myofibrillar extracts of TgPLNR9C and wt hearts. [All methodologies detailed in (17).]

To determine whether PLNR9C altered phosphorylation of PLNwt, we transfected HEK-293 cells with varying amounts of NF-PLNR9C cDNA in the presence of a constant amount of PLNwt. Phosphorylation status was assessed using antibody 285 (21), which is specific for phosphorylated PLN (Fig. 3B). As NF-PLNR9C expression was increased and NF-PLNwt expression held constant, the level of phosphorylation of NF-PLNwt decreased, eventually to zero, indicating that PLNR9C prevented phosphorylation of PLNwt.

We next investigated PLN phosphorylation by two-dimensional gel electrophoresis and Western blot analysis (22) of left ventricular tissue extracts from an affected individual (individual III-2; Fig. 1) and a genetically unaffected individual, and from wild-type and TgPLNR9C mice studied before the onset of cardiac dysfunction. Because anti-PLN monoclonal antibody (mAb) 1D11 did not recognize PLNR9C, only PLNwt was identified. At least 50% of PLNwtwas phosphorylated in unaffected human and wild-type mouse hearts, whereas no phosphorylation was detected in samples from affected human or mouse hearts (Fig. 3C). We conclude that PLNR9C prevented phosphorylation of PLNwt.

The R9C mutation could disrupt both the PLN epitope site of mAb 1D11 and the binding site for PKA, because these involve overlapping amino acid residues (19). As a consequence, the R9C mutation might prevent dissociation of PKA from PLNR9C, thereby trapping PKA and effectively inactivating it. To evaluate this hypothesis, we immunoprecipitated PLN from extracts of HEK-293 cells expressing different ratios of PLNwt and PLNR9C; the extracts had been preincubated with the PKA catalytic subunit and ATP (Fig. 3D) (17). Under the conditions used (17), PLN forms a pentamer (14). All pentamers containing at least one PLNwt molecule were immunoprecipitated. As PLNR9C content increased in heteropentamers, there were increasing amounts of PKA in the immunoprecipitate; the level of PKA in the immunoprecipitate decreased when PLNR9C homopentamers became predominant. This pattern would occur if PKA forms a stable complex with PLNR9C, but not with PLNwt. Parallel human analyses revealed more PKA associated with PLN in cardiac tissues from patients with PLNR9C than in tissues from genetically unaffected individuals (Fig. 3E). However, total PKA activity levels and total PKA-dependent protein phosphorylation (17) were identical in protein lysates from TgPLNR9C and wild-type myocytes (Fig. 3, F and G), suggesting local subcellular effects of PKA trapping by PLNR9C.

We studied the consequences of PLNR9C on Ca2+ kinetics and sarcomere mechanics, after labeling isolated TgPLNR9C and wild-type myocytes with the fluorochrome Fura2 (17, 23). SR Ca2+ release velocities, time to 90% peak Ca2+ signal, and cell shortening velocities (+dL/dt) were similar in the mutant and wild-type cells (P = 0.86,P = 0.88, P = 0.53, respectively). However, the decay in Ca2+ transients was markedly prolonged in TgPLNR9C myocytes compared to wild-type cells [time for 80% decay of Ca2+ signal (T 80%) = 0.24 ± 0.13 s versus 0.17 ± 0.04 s; P = 0.001]. Myocyte relaxation was also delayed (time to 90% relaxation = 0.132 ± 0.10 s versus 0.098 ± 0.04 s; P = 0.03). A noteworthy finding was that depressed Ca2+kinetics could not be normalized by β-adrenergic stimulation of TgPLNR9C myocytes: Administration of 10−7 M isoproterenol accelerated SR Ca2+ reuptake in wild-type myocytes (T 50% = 0.094 ± 0.02 s versus isoproterenol T 50% = 0.075 ± 0.01 s; P = 0.00001) but did not alter Ca2+ reuptake in TgPLNR9C myocytes (P = 0.26). These functional data support the hypothesis that PLNR9C is constitutively active and produces chronic SERCA2a inhibition. Although PLNR9C is relatively inactive as a direct regulator of SERCA2a, its ability to block PLNwt phosphorylation explains how the mutant allele produces a dominant phenotype and demonstrates PLN dysfunction as a primary cause of human heart failure.

Previous human genetic studies have implicated eight different proteins in familial dilated cardiomyopathy: dystrophin, desmin, taffazin, lamin A/C, titin, actin, troponin T, and β-myosin heavy chain (24). The mechanism by which these sarcomere or cytoskeleton mutations cause disease is thought to be impairment of force production or transmission. In contrast, individuals with the PLNR9C mutation develop dilated cardiomyopathy and profound heart failure by a mechanism involving direct disturbance of myocellular Ca2+ metabolism due to constitutive SERCA2a inhibition. Although the possibility exists that mutations in cytoskeletal proteins alter Ca2+ homeostasis by indirect mechanisms, an equally likely conclusion is that there are at least two different pathways leading to heart failure, one disrupting structural elements and one affecting Ca2+ kinetics. Disturbances in myocellular Ca2+ may be the primary trigger of heart failure in other cardiovascular pathologies as well. These conditions and heart failure caused by PLNR9C should be ameliorated by therapeutic strategies that restore normal Ca2+ handling (8, 25), such as manipulation of SERCA2a/PLN activities.

Supporting Online Material

www.sciencemag.org/cgi/content/full/299/5611/1410/DC1

Materials and Methods

Figs. S1 to S3

Table S1

References

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: cseidman{at}rascal.med.harvard.edu

REFERENCES AND NOTES

View Abstract

Stay Connected to Science

Navigate This Article