A Caveolae Targeted L-type Ca²⁺ Channel Antagonist Inhibits Hypertrophic Signaling Without Reducing Cardiac Contractility

Abstract

Rationale

The source of Ca²⁺ to activate pathological cardiac hypertrophy is not clearly defined. Ca²⁺ influx through the L-type Ca²⁺ channels (LTCCs) determines “contractile” Ca²⁺, which is not thought to be the source of “hypertrophic” Ca²⁺. However, some LTCCs are housed in caveolin-3 (Cav-3) enriched signaling microdomains and are not directly involved in contraction. The function of these LTCCs is unknown.

Objective

To test the idea that LTCCs in Cav-3 containing signaling domains are a source of Ca²⁺ to activate the calcineurin-nuclear factor of activated T cells (Cn-NFAT) signaling cascade that promotes pathological hypertrophy.

Methods and Results

We developed reagents that targeted Ca²⁺ channel blocking Rem proteins to Cav-3 containing membranes, which house a small fraction of cardiac LTCCs. Blocking LTCCs within this Cav-3 membrane domain eliminated a small fraction of the LTCC current, almost all of the Ca²⁺ influx induced NFAT nuclear translocation, but did not reduce myocyte contractility.

Conclusions

We provide proof of concept that Ca²⁺ influx through LTCCs within caveolae signaling domains can activate “hypertrophic” signaling, and this Ca²⁺ influx can be selectively blocked without reducing cardiac contractility.

Introduction

Cardiovascular diseases increase cardiac systolic stress and cause myocyte hypertrophy. Increases in Ca²⁺ influx during pathological stress maintain pump function and activate signaling pathways that produce pathological hypertrophy and heart failure^1,2, but the source of this “hypertrophic” Ca²⁺ is still not clearly defined.

Increases in Ca²⁺ influx are essential for the activation of pathological hypertrophic signaling, including the calcineurin-nuclear factor of activated T-cells (Cn-NFAT) pathway^{2, 3}, the calmodulin-dependent protein kinase (CaMKII) pathway , and the protein kinase C (PKC) pathway⁵. The source of the Ca²⁺ that activates each of4 these signaling cascades is still not well known. Potential sources include L-^{6, 7} and T-^{8, 9} type Ca²⁺ channels and transient receptor potential (TRP) channels¹⁰.

In cardiac myocytes, voltage dependent L-type Ca²⁺ channels (LTCC, Cav1.2) are the predominant Ca²⁺ entry pathway and are essential for excitation-contraction coupling (ECC) and regulation of gene expression¹. Myocyte LTCCs are concentrated in T-tubular membranes where they are organized in junctional complexes with ryanodine receptors (RyR2) on the SR¹ to form a signaling microdomain involved in Ca²⁺-induced-SR Ca²⁺-release (CICR). This microdomain is the source of “contractile” [Ca²⁺].

Not all LTCCs are involved in CICR. A small fraction appears to be harbored in caveolae¹¹, which are highly specialized membrane regions that are stabilized by the scaffolding protein caveolin. Caveolin-3 is the major caveolin expressed in the heart and it functions to help organize local signaling microdomains¹¹. The function of caveolae based LTCCs are unknown and are the topic of this study.

Our studies test the hypothesis that “hypertrophic” Ca²⁺ enters the myocyte through LTCCs localized in Cav-3 based signaling microdomains and that this “signaling” Ca²⁺ is distinct from “contractile” Ca²⁺. This idea was examined with a novel LTCC blocker that selectively traffics to caveolae where it inhibits LTCCs within this signaling microdomain. Our results show that this Cav-3 localized LTCC antagonist can block “hypertrophic” Ca²⁺ without altering contractility. The caveolae targeted LTCC blocker was generated by molecular modification of Rem, a member of the RGK GTPase family that is known to potently inhibit LTCCs¹². Expressing wild type Rem in cardiac myocytes blocked I_Ca-L and this inhibition required a c-terminal membrane-docking domain¹³. Truncation of this membrane-docking domain (Rem^1-265) resulted in the inability of Rem to traffic to the membrane and to inhibit I_Ca-L. To specifically block caveolae-based LTCCs, the membrane-binding motif of native Rem was deleted and replaced with a canonical caveolin targeting domain sequence¹⁴, to create Rem^1-265-Cav. Rem^1-265-Cav targeted selectively to Cav-3 membranes, did not alter myocyte contractility, but blocked Ca²⁺-influx mediated activation of Cn-NFAT signaling. These data strongly support the hypothesis that LTCCs housed in Cav-3-containing microdomains is a source of “hypertrophic” Ca²⁺.

Methods

All methods have been described in detail in previously^7,15-17 and an in depth account is available in the Online Data Supplement.

Adult feline left ventricular myocytes (AFLVMs) were isolated¹⁵ and adenoviral gene transfer was used to introduce WT-Rem, Rem^1-265, Rem^1-265-Cav, and/or NFATc3-GFP¹⁷, all at a multiplicity of infection (MOI) of 100. Myocytes were cultured for a period of up to 4 days. Localization of AdRem constructs was confirmed using standard sucrose gradient fractionation and caveolin-3 immunoisolation techniques¹⁶.

Results

Rem^1-265-Cav Localizes to Caveolae

Native Rem was not detected in plasma membrane (PM) from AFLVMs. After adenoviral infection, WT-Rem was found in all AFLVM membrane fractions, consistent with the idea that it is in all PM domains. The Rem^1-265 peptide lacking the membrane association motif remained in the general cell homogenate (H). Rem^1-265-Cav localized to PM specifically within caveolin-containing lipid rafts (Figure 1A), verifying target specificity of the peptide containing the Cav binding motif. Rem^1-265-Cav did not alter the normal raft/caveolae targeting of LTCCs, eNOS and caveolin-3 which demonstrates that Rem^1-265-Cav does not displace molecules normally found in caveolae (Figure 1A). To prove that Rem^1-265-Cav specifically localizes to caveolae, rather than lipid rafts in general, caveolae organelles were immunoaffinity purified¹⁶. Similar to our observation in raft membranes, only Rem^1-265-Cav was seen in association with caveolin-3 (Figure 1B). These data show that Rem^1-265-Cav specifically traffics to Cav-3 membrane microdomains.

Figure 1. Rem^1-265-Cav and Cav1.2 distribution in plasma membranes.

(A) WT-Rem was in all plasma membrane (PM) populations. Rem^1-265-Cav localized specifically within caveolin-containing lipid rafts. Rem^1-265 was not in PMs and remained in the general cell homogenate (H). Rem^1-265-Cav did not alter the normal raft/caveolae targeting of Cav1.2, eNOS and caveolin-3. (B) Caveolae organelles were immunoaffinity purified with anti-Cav-3. Rem^1-265-Cav targeted specifically to caveolae whereas WT-Rem was both in caveolae (bound fraction) and in membrane samples depleted of caveolae (unbound fraction). Truncated Rem^1-265 was not found in PMs. (C) To estimate the fraction of Cav1.2 localized to caveolae versus non-caveolae PMs, a Cav-3 IP was performed (lane 3). The unbound fraction (lane 4) from the Cav-3 IP was re-immunoprecipitated for Cav1.2 (lane 5). To estimate the fraction of caveolae that contain Cav1.2, PM preparations were immunoprecipitated for Cav1.2 (lane 7). The resulting unbound fraction (lane 8) was then re-immunoprecipitated for Cav-3 (lane 9) and the intensity of the bands in both fractions was compared. (U=unbound fraction, B=bound fraction, “bead controls” test for non-specific binding to the solid support used for immunoisolation of caveolae vesicles)

To further assess what fraction of LTCCs localizes to caveolae membrane domains, PM samples were purified from AFLVMs and subjected to immunoaffinity isolation with anti-Cav-3 (Figure 1C). The unbound fraction from the Cav-3 IP was then re-immunoprecipitated for Cav1.2 (Figure 1C). Densitometric analysis (n=3) showed that 26.2+/-12.7% of LTCCs resides within caveolae. To estimate the fraction of caveolae that contain LTCCs, AFLVM PM preparations were immunoprecipitated with an antibody for Cav1.2 (LTCC). The resulting unbound fraction was then re-immunoprecipitated for Cav-3 and the intensity of the bands in both fractions were compared. Only 13.4 +/− 10.1% (n=3) of caveolae PM contained LTCCs.

Rem^1-265-Cav Blocks a Small Fraction of I_Ca,L and has no effect on Contractility

WT-Rem almost fully eliminated the L-type Ca²⁺ channel current (I_Ca,L) in AFLVMs at an MOI of 100 (Figure 2A). Rem^1-265 had no significant effect on I_Ca,L. Rem^1-265-Cav caused a small inhibition (about 15%) of I_Ca,L (Figure 2A). The smaller I_Ca,L block with Rem^1-265-Cav versus WT-Rem suggests that only a small fraction of LTCCs are localized to Cav-3 signaling microdomains.

Figure 2. Rem^1-265-Cav blocks a small fraction of I_Ca,L without reducing Contractility.

(A) WT-Rem (MOI of 100) almost completely abolished I_Ca,L while at equal MOI, Rem^1-265-Cav inhibited only a small fraction of the current. Rem^1-265 had no effect on I_Ca,L. Rem^1-265-Cav infected myocytes responded normally to Iso. Myocyte contractions (B) and Ca²⁺ transients (C) were measured 36-48 hours after infection. Rem^1-265-Cav had no significant effect on fractional shortening while WT-Rem almost completely abolished contraction. Rem^1-265 had no effect. Rem^1-265-Cav had no significant effect on Ca²⁺ transients (n=12-16). *** p<0.001

Inhibition of I_Ca,L with WT Rem eliminates LTCC regulation by catecholamines¹². AFLVMs infected with Rem^1-265-Cav responded normally to Isoproterenol (Figure 2A) while those infected with WT-Rem failed to respond (not shown). Myocytes infected with WT-Rem had markedly reduced fractional shortening (1.1+/-0.1% resting cell length) while those infected with truncated Rem^1-265 (5.0+/-0.5%) and Rem^1-265-Cav (4.4+/-0.4%) had contractions that were not significantly smaller than controls (4.7+/-0.5%) (Figure 2B). [Ca²⁺]_i transients were also unaffected by Rem^1-265-Cav (figure 2C). Collectively, these data support the idea that Rem^1-265-Cav blocks a small fraction of Cav-3 associated LTCCs without significantly effecting myocyte excitation contraction coupling (ECC).

Rem^1-265-Cav inhibits the β₂-Adrenergic Regulation of LTCCs

Myocyte Cav-3 membrane domains are thought to contain both LTCCs and β₂-adrenergic receptors¹¹. The β₂-adrenergic receptor specific agonist Zinterol in the presence of the β₁-adrenergic receptor specific antagonist CGP increased I_Ca,L and fractional shortening in AFLVMs. This effect was inhibited by Rem^1-265-Cav (Figure 3), suggesting that Rem^1-265-Cav can selectively block Ca²⁺ entry through β₂-adrenergic receptor regulated LTCCs within Cav-3 signaling microdomains.

Figure 3. β₂AR Signaling is Inhibited by Rem^1-265-Cav.

Zinterol in the presence of the β₁AR specific antagonist CGP increased I_Ca-L (A; n=7) and fractional shortening (B; n=8). Rem^1-265-Cav inhibited these effects. * p<0.05

Rem^1-265-Cav Inhibits NFAT Translocation to the Nucleus

AFLVMs infected with AdNFATc3-GFP or AdNFATc3-GFP and AdRem^1-265-Cav were electrically quiescent in culture. AFLVMs maintain low levels of cytosolic [Ca²⁺] with little or no spontaneous SR Ca²⁺ release⁹ and NFATc3-GFP is localized to the cytoplasm¹⁷. NFAT localization (cytoplasmic versus nuclear) was measured before (Figure 4A and C) and after (Figure 4B and D) 1 Hz pacing for 1 hour. Pacing caused NFAT to translocate from the cytoplasm to the nucleus in control AFLVMs (Figure 4B). AFLVMs infected with Rem^1-265-Cav had normal contractions and Ca²⁺ transients (shown above) but more than 90% of NFAT nuclear translocation was inhibited (Figure 4E and F). These results were confirmed in experiments in which bath Ca²⁺ was raised to 4 mM, which also induced NFAT to translocate to the nucleus in control cells. Rem^1-265-Cav inhibited NFAT translocation under these conditions (data not shown).

Figure 4. Rem^1-265-Cav Inhibits Pathological Hypertrophic Signaling.

AFLVMs were infected with AdNFAT-GFP (A, B, E) or AdNFAT-GFP and AdRem^1-265-Cav (C, D, F). NFAT-GFP translocation to the nucleus was measured before (A,C) and after (B,D) pacing at 1 Hz for 1 hour. Average data from 5 experiments is shown in E and F.

Discussion

Ca²⁺-Calmodulin (CaM) dependent activation of Cn-NFAT^{2, 3} produces pathological hypertrophy. Activation of this signaling cascade requires an increase in myocyte [Ca²⁺]. “Contractile” [Ca²⁺] does not appear to activate this hypertrophic process⁶.

The sources of “hypertrophic” [Ca²⁺] to activate Cn-NFAT signaling are not clearly defined. There is some evidence for a role for the LTCC^{6, 7}, however, there is also evidence for Ca²⁺ entry through T-type Ca²⁺ channels (TTCC)^{8, 9} and TRP channels¹⁰. The role of TTCCs is controversial since some studies suggest that this pathway is anti-hypertrophic⁸ rather than pro-hypertrophic⁹. The observation that excess Ca²⁺ influx through TTCCs channels is anti-hypertrophic⁸ suggests that Ca²⁺ influx per se is not pro- or anti-hypertrophic. Rather the pathway for Ca²⁺ influx and possibly the localization of this pathway appear to be critical.

The present experiments suggest that a small number of LTCCs are localized in a fraction of Cav-3 containing membranes to form a signaling microdomain that can activate Cn-NFAT signaling. Our results are consistent with a recent study suggesting that a subpopulation of LTCCs in Cav-3 containing membranes are housed with AKAP150, which binds Cn¹⁸. These Cav-3 localized LTCCs do not participate in ECC or the regulation of “contractile” Ca²⁺. Our Cav-3 targeting strategy was able to selectively traffic an LTCC-blocking Rem protein (Rem^1-265-Cav) to this signaling domain to block hypertrophic signaling without significantly reducing contraction or β1AR regulation of contractility. Our results suggest that any LTCCs associated with the small amount of Cav-3 others have found within T-tubules do not participate in ECC¹⁹.

Our results also show that not all Cav-3 containing membrane domains contain LTCCs, suggesting that hypertrophic signaling may take place in highly specialized signaling complexes. The signaling partners within these domains and their regulation in health and disease needs to be determined in future studies.

LTCCs are the major Ca²⁺ influx pathway in the adult mammalian heart and these channels are regulated biosensors that determine myocyte contractility and cardiac pump function¹. Our results suggest that nature uses a subpopulation of these channels, housed away from ECC signaling domains, as a source of Ca²⁺ to modulate myocyte size when the heart is subjected to stress. Selective block of this Ca²⁺ influx pathway by the novel reagents developed in this study might provide an approach to reduce pathological hypertrophy without reducing cardiac contractility.

Novelty and Significance.

What is Known?

• Pathological cardiac stressors such as hypertension and myocardial infarction activate signaling cascades that lead to cardiac myocyte hypertrophy.

• Increases in myocyte [Ca²⁺]_i are essential to initiate hypertrophic signaling through the cytoplasmic protein phosphatase Calcineurin (Cn).

• The cytoplasmic [Ca²⁺]_i transient is initiated by Ca²⁺ influx through L-type Ca²⁺ channels (LTCC) and drives cardiac contraction. This “contractile” [Ca²⁺]_idoes not appear to be the source of [Ca²⁺] to initiate pathological hypertrophy.

What New Information Does This Article Contribute?

• A small population of LTCCs is housed in Caveolin (Cav)-3 containing membranes (Caveolae) and is not involved in the regulation of contractile [Ca²⁺]_i.

• An LTCC-blocking protein can be trafficked to Cav-3 containing membranes to selectively block LTCCs in this signaling microdomain.

• Block of Ca²⁺ entry through LTCCs within caveolae abolished activation of the Cn signaling pathway that is known to cause pathological hypertrophy.

There is strong epidemiological evidence to suggest that pathological cardiac hypertrophy predisposes patients to adverse cardiac events and to heart failure. Increases in [Ca²⁺]_i are necessary to activate pathological hypertrophy, but the source of this Ca²⁺ appears to be distinct from the [Ca²⁺]_i that is required for cardiac contraction. This study shows that a small population of LTCCs in a specialized signaling microdomain can be a source of Ca²⁺ to activate hypertrophic signaling. We developed a novel reagent that selectively blocks “hypertrophic” Ca²⁺ entry through Cav-3-based channels without blocking the LTCCs that induce and regulate “contractile” Ca²⁺. These experiments suggest that selective block of LTCCs in Cav-3 based signaling microdomains could reduce pathological hypertrophy without causing adverse negative inotropic effects.

Non-Standard Abbreviations

LTCC

L-Type Calcium Channel

Cav-3

Caveolin-3

Cn

Calcineurin

NFAT

Nuclear Factor of Activated T cells

CaMKII

Calmodulin-Dependent Protein Kinase

PKC

Protein Kinase C

TRP

Transient Receptor Potential

ECC

Excitation Contraction Coupling

RyR2

Ryanodine Receptor

CICR

Ca²⁺-Induced-SR Ca²⁺-Release

AFLVM

Adult Feline Left Ventricular Myocyte

MOI

Multiplicity of Infection

CaM

Ca²⁺-Calmodulin

TTCC

T-Type Calcium Channel