TRPC1 channels are critical for hypertrophic signaling in the heart

Trpc1^−/− and cardiac hypertrophy

We first designed studies to test the hypothesis that TRPC1 mediated calcium entry in cardiomyocytes activates hypertrophic signaling. Trpc1^−/− mice and WT mice were subjected to transverse aortic constriction (TAC) to induce cardiac hypertrophy. Histological sections of hearts from WT mice demonstrated a marked increase in LV (left ventricle) size after eight weeks of TAC compared to sham operated mice (Figure 1A, top row). In contrast, the hearts of Trpc1^−/− mice did not demonstrate significant cardiac hypertrophy (Figure 1A, top row) nor was there an increase in collagen deposition as determined by Sirius red staining compared to the heart sections of WT mice (Figure. 1A, bottom row).

Serial echocardiograms (performed at 0, 4 and 8 weeks) demonstrated a progressive decline in the fractional shortening (% FS) of WT mice, whereas Trpc1^−/− mice maintained preserved % FS over a wide range of systolic pressure gradients (0–150mmHG) (Figure 1B and online table 1). A detailed statistical analysis for fractional shortening and LV/BW ratio has been shown in online table 3 & online figure 2. Invasive LV hemodynamic parameters measured from WT and Trpc1^−/− mice showed significant differences in the LV+dP/dt_max and LV-dP/dt_min only after the TAC operation. These data indicate that the loss of TRPC1 is associated with preserved contractility despite a marked increase in pressure load ( online table 1).

Consistent with the observations made by echocardiography and hemodynamics, we found marked difference in cardiac mass between hearts of Trpc1^−/− mice subjected to TAC compared to WT mice. The increase in cardiac mass following TAC was significantly reduced in the Trpc1^−/− mice as was evident by the change in left ventricular mass/body weight ratio (LV/BW), (3.1±0.23 to 4.5±0.175, before and after TAC respectively) as compared to WT mice (3.3 ±0.14 to 6.8±0.42) (fig. 1C & online table 1). Thus it is apparent from these data that Trpc1^−/− mice respond to pressure overload with a modest increase in cardiac mass and preserved cardiac function, whereas the same level of pressure overload in WT mice produced significant cardiac impairment. We also considered whether changes in TRPC2-7 channel expression occurred in the hearts of the Trpc1^−/− mice as mechanism for the protection seen in the I_NSC. In fact, TRPC channels (mRNA or protein) at baseline and after pressure overload did not differ between WT and Trpc1^−/− mice.(online figure 1A, B, C&D).

TRPC channels are known to be receptor operated cation channels activated downstream of G-protein coupled receptors, e.g. Angiotensin-II (Ang-II), raising the possibility that Ang-II mediated hypertrophic signaling may influence TRPC signaling (15),(11, 16, 17). WT and Trpc1^−/− mice infused with Ang-II (1000 ng/kg/min) for four weeks (28 days) to induce cardiac hypertrophy (Figure 1D and online Table 2).

I_NSC in adult cardiomyocytes

Next, we designed whole cell voltage clamp studies to measure TRPC currents from isolated adult cardiomyocytes taken from WT or Trpc1^−/− mice. Solutions were configured so as to limit voltage gated Ca²⁺ and K⁺ currents mainly by including inhibitors of L-type calcium channels and Cs⁺ to block K⁺ channels. We used an inverse ramp protocol from +100 mV to −100 mV to limit voltage gated Na⁺ channel currents and holding potential of 0 mV(see online methods)(18). Figure 2A displays current-time plots recorded from WT (brown traces) and Trpc1^−/− (blue traces) cardiomyocytes eight weeks after the TAC operation. WT cardiomyocytes displayed a greater current recorded at both +80mV and −80 mV membrane potentials compared to Trpc1^−/− cardiomyocytes. The non-selective currents were inhibited by gadolinium (Gd³⁺), a known TRPC channel blocker (Figure 2A). We noted that the I–V relationship recorded from sham and TAC operated mice was linear in shape with a reversal potential of 0 mV, features reminiscent of non-selective TRPC channels (Figure 2B). Interestingly non-selective currents recorded from cardiomyocytes taken from Trpc1^−/− mice displayed similar I–V relations with zero reversal potential (Figure 2C). However the current density of the non-selective currents recorded from Trpc1^−/− cardiomyocytes, both sham and TAC operated mice, was markedly reduced compared to the WT cardiomyocytes (Figure 2D).

To characterize the I_NSC recorded from cardiomyocytes, we sought evidence for the contributions of TRPC from cation selectivity, pharmacological profiling and gating mechanism (Figure 2E). Replacing Na⁺ in the external solution with N-methyl-D-glucamine (NMDG) dramatically reduced the non-selective current by greater than 50% from WT cardiomyocytes indicating the current is in part permeable to Na⁺. In addition, the I_NSC was rapidly blocked by the addition of gadolinium (Gd³⁺, 10µM) to the perfusion bath. Trivalent cation block of the I_NSC is a characteristic feature of TRPC currents (19). We also found that the putative TRPC1 current was equally permeable to calcium and barium which would further suggest this current is in part contributed by TRPC1 (Figure 2E). We found no difference in the Li³⁺ sensitive currents in cardiomyocytes isolated from WT and Trpc1^−/− mice indicating no change in NCX1 currents (not shown) (20). In addition, we found that immunolabeled TRPC1 only partially overlapped with that of NCX1 (online Figure 1E). When considered in total these results indicate that TRPC1 is likely to contribute to the non-selective cation background current recorded in adult cardiomyocytes.

Given the central role of the neurohormone angiotensin-II (Ang-II) in the stretch activated signaling in the cardiomyocyte (23), we tested whether Ang-II application influenced the non-selective cation current attributed to TRPC1. Ang-II stimulation of WT cardiomyocytes resulted in two-fold increase in the current density of the non-selective current (Figure 2F). In contrast, Ang-II failed to augment the TRPC current in cardiomyocytes isolated from Trpc1^−/− mice (Figure 2F). We also tested whether TRPC1 currents were activated by 1-oleoyl-2-acetyl-sn-glycerol (OAG), a stable cell permeable analog of DAG (Figure 2G). Here, perfusion of cardiomyocytes with OAG (10µM) activated a non-selective current similar to that seen with Ang-II. We found that the current density of OAG activated currents from Trpc1^−/− and WT cardiomyocytes were not different (Figure 2G). These results indicate that TRPC1 contributes to the Ang-II induced non-selective current, whereas TRPC1 does not influence the DAG-induced current. TRPC1 has also been implicated in store operated calcium entry as a SOC channel in many cell types including cardiomyocytes. We however did not find a difference in the rate of Ca²⁺ entry following store depletion in Trpc1^−/− neonatal cardiomyocytes compared to WT cells (online Figure 3). These findings support our model in which TRPC1 channels respond to G-protein signaling that is involved in the pressure overload response.

TRPC1 and the stretch response

We next determined whether cardiomyocytes from Trpc1^−/− mice responded differently to mechanical stretch than WT cardiomyocytes. Isolated neonatal cardiomyocytes subjected to a cyclical stretch at 15% strain at 1Hz frequency for 8–24 hours duration expressed augmented levels of BNP (brain natriuretic factor) and ANF (atrial natriuretic factor) in a time dependent manner (Figure 3A) peaking at 24 hours with a 7 fold increase in the BNP and ANF expression. Interestingly, treating WT cells with losartan to block angiotensin 1 (AT1) receptors prevented the induction of the stretch gene program. Moreover, Trpc1^−/− cardiomyocytes showed no change in the expression of BNP and ANF over any time point examined. Thus neonatal cardiomyocytes lacking TRPC1 displayed a blunted response to stretch in comparison to WT cardiomyocytes. The blocking of stretch response by losartan treatment suggests that TRPC1 mediated stretch signaling is downstream of the AT1 receptor.

It has been recently shown that mechanical stretch activates angiotensin receptors (AT1R) in an agonist independent manner that can result in activation of a non-selective cation current in vascular smooth muscle cells (24). These results may explain the recent findings demonstrating TRPC1 and TRPC6 channels as putative stretch activated channels (9). We therefore tested whether stretch influences the TRPC1 current in adult cardiomyocytes. Here, osmotic stress applied to the cardiomyocytes, by changing the osmolarity of the perfusion bath from 305 mOsm to 205 mOsm, increased the non-selective current (I_swell) in WT cardiomyocytes. The IV-relationship of the I_swell from adult cardiomyocytes strongly resembled that observed above for pressure overload and Ang-II. Osmotic stress did not activate the non-selective current in cardiomyocytes lacking TRPC1 (Figure 3B–C). We further characterized the I_swell in WT cardiomyocytes to determine if Ang-II signaling was involved in the osmotic activated currents. Inclusion of the phospholipase C (PLC) blocker, U-73122, in the pipette solution decreased the non-selective current at baseline and following cell swelling induced by osmotic stress (Figures 3D&F). We also demonstrate that the I_swell in WT cardiomyocytes signals through the Ang-II receptor as the current was significantly blocked by losartan (10µM) compared to vehicle treated WT cardiomyocytes (Figures 3E&F). These data indicate Ang-II signaling is fundamental to the current activated by cell swelling. Moreover, WT cardiomyocytes treated with tarantula toxin GsMTx-4, an inhibitor of stretch-activated channels also blocked the non-selective current induced by cell swelling, providing additional evidence that TRPC1 acts as a stretch activated channel (data not shown)(8, 25). Collectively, these studies show that Trpc1^−/− cardiomyocytes, in comparison to WT cells, fail to respond to different forms of stretch as indicated by the lack of stretch activated non-selective cation current (swelling and positive pressure) or changes in ANF and BNP mRNA expression (radial stretch). These results provide evidence in support of our model in which TRPC1 channels reside downstream of stretch-activated GPCR signaling to confer stretch dependent signaling associated with cardiac hypertrophy.

Loss of TRPC1 alters hypertrophic signaling

To this point our studies indicated that mice lacking TRPC1 are protected from the deleterious effects of pressure overload. Since it is now well established that calcium entry through TRPC channels influences calcineurin/NFAT signaling in many cell types including cardiomyocytes, we examined the phosphorylation state of NFATC3 in cardiac lysates prepared from WT and Trpc1^−/− mice as a marker of calcineurin/NFAT signaling. NFATC3 is dephosphorylated by calcineurin in pressure overloaded hearts as was detected in the WT mice in Figure 4A. We then examined whether the loss of TRPC1 influenced NFAT signaling. First, we found a significant increase in phospho-NFATC3 in the heart lysates prepared from sham operated Trpc1^−/− mice compared to WT mice indicating that more NFATC3 exists in the inactive state in the sham operated Trpc1^−/− mice (Figure 4A). No differences in total NFAT were seen among WT and Trpc1^−/− mice, but a greater fraction of NFATC3 existed in the phosphorylated state in TAC operated Trpc1^−/− mice compared to TAC operated WT mice (Figure 4A). These data suggest that calcineurin/NFAT activity in the hearts of mice lacking TRPC1 is less active at baseline and after pressure overload (Figure 4B).

We next designed studies using a transgenic NFAT indicator mice (26, 27) to further examine the relationship between TRPC1 and calcineurin/NFAT signaling. Neonatal cardiomyocytes from NFAT indicator mice were infected with adenoviruses carrying either siTRPC1 or siCTL (scrambled) constructs (Figure 4C). Cardiomyocytes were stimulated with Ang-II to alter the frequency of spontaneous calcium oscillations and activate the hypertrophic program (Figure 4D). In cells with TRPC1 silencing, we found a marked attenuation of the calcium oscillations compared to control silenced cells (Figure 4 D, E&F). This change in the oscillation frequency in the siTRPC1 neonatal cardiomyocytes (NCM) corresponded to a reduction in NFAT transactivation as measured by LacZ reporter gene compared to control NCM. These results support our in vivo studies suggesting that TRPC1 channels provide calcium entry needed for hypertrophic signaling through the calcineurin/NFAT signaling pathway in cardiomyocytes (Figure 4G).

TRPC1 and hypertrophic gene expression

Cardiac hypertrophy with diminished cardiac function has long been associated with changes in profiles of gene expression (28). We therefore measured the mRNA for brain naturetic peptide (BNP) and atrial naturetic (ANF) from the hearts of Trpc1^−/− and WT mice (sham and TAC operated). These neurohormones are released by cardiomyocytes subjected to increased wall stress and are important biomarkers for the activation of the fetal gene program in response to cardiac stress (29). WT mice subjected to pressure overload were found to have a nine fold increase in the expression of BNP and 16 fold increase in ANF; whereas Trpc1^−/− mice subjected to TAC showed no significant change in BNP and ANF expression (Figure 5A). We also found that the expression of the calcium pump, sarcoplasmic-endoplasmic calcium ATPase (SERCA2a), was decreased in the TAC-WT mice while there was no corresponding reduction in SERCA2a expression in Trpc1^−/− mice subjected to TAC. SERCA2a is known to be diminished in heart failure and may contribute to abnormal calcium signaling (30). It is clear from these studies that the attenuated cardiac hypertrophy and preserved cardiac function observed in mice lacking TRPC1 accompanied no changes in the gene expression profile associated with maladaptive cardiac hypertrophy.

To better understand the survival advantage offered to Trpc1^−/− mice in response to pressure overload, we next tested the hypothesis that Akt signaling pathways associated with cardiomyocyte survival would be enhanced in the Trpc1^−/− mice. In fact, we were unable to discern differences in the level of phospho-Akt between the sham and TAC operation in WT mice. In contrast, Akt phosphorylation was notably increased in heart lysates prepared from Trpc1^−/− mice following the TAC operation. Likewise phospho-mTOR was markedly increased in the heart lysates of Trpc1^−/− TAC operated mice compared to WT mice. No differences were noted in the total Akt or mTOR expression of WT or Trpc1^−/− hearts (Figure 5B & C). These results suggest that the preserved cardiac function and minimal changes to cardiac mass following pressure overload seen in mice lacking TRPC1 may be due in part to the preserved Akt and mTOR signaling which are associated with cell survival.