Trametinib alters contractility of paediatric Noonan syndrome-associated hypertrophic myocardial tissue slices

Myocardial slice preparation

Hypertrophic RV tissue was obtained via a resection procedure of the RV outflow tract (RVOT) in a 7-year-old male patient diagnosed with NS and severe RVOT obstruction warranting resection. Tissue was immediately transferred into a closed cardioplegia (4°C)-filled container and kept at 4°C. After transfer into a flow hood, four 0.5 × 0.5-cm-sized and 300-μm-thick tissue slices were produced using a vibratome according to the protocol developed by Fischer et al.²³ and described in detail by Hamers et al.²⁴ Briefly, tissue was embedded in 4% low-melt agarose and cut in 300-μm-thick slices using a vibratome (VT1200s, Leica AG, Wetzlar, Germany). Plastic laser-cut triangles (made in-house) were attached to the slices and placed in a commercially available BMCC on a rocker (60 r.p.m.) (InVitroSys GmbH, Gräfelfing, Germany) in a CO₂ incubator (5%; CB240, Binder GmbH, Tuttlingen, Germany) at 37°C. BMCCs contained medium M199 (31150-022, Thermo Fisher, Bremen, Germany) [supplemented with 5% Insulin-Transferrin-Selenium 100× (5150056, Thermo Fisher, Bremen, Germany), 5% Penicillin–Streptomycin (P0781-100ML, Sigma, Taufkirchen, Germany) and 50 mM β-mercaptoethanol (A1108.0100, AppliChem, Darmstadt, Germany)]. The tissue slices were continuously electrically stimulated to contract at 30 b.p.m. Additionally, tissue slices were also prepared, cultivated and treated (n = 2) from the RV of the explanted heart of an HTx patient, following the same protocol as for the paediatric tissue slices. The 51-year-old patient was diagnosed with dilated cardiomyopathy after a viral myocarditis (diagnosed in 2019) and transmural fibrosis.

Cultivation and treatment in BMCCs

Tissue slices were initially cultivated for 180 h to allow for the original trametinib treatment of the patient to wash out and to equilibrate to the ex vivo environment. After this period, four slices were treated with trametinib (GSK1120212, Selleckchem GmbH, Cologne, Germany), rapamycin (S1039, Selleckchem GmbH, Cologne, Germany), dasatinib (SML2589, Merck KGaA, Darmstadt, Germany) (all 100 nM) or dimethyl sulfoxide (DMSO) vehicle (0.1%) as control with medium change (every 2 days). Contraction force was analysed using LabChart Pro 8 (ADInstruments, Dunedin, New Zealand) and measured by an in-house developed macro, which averaged the contraction of the slice in a 5 min block for each hour of the cultivation. The average contraction force of the slices was normalized to the contraction force at the start of the cultivation.

Electrical stimulation protocols

Before and during treatment cultivation, distinct stimulation protocols were run to assess the contractile and electrophysiological status of the slices. Stimulation protocols consisted of post-pause potentiation (PPP) and force–frequency relationship (FFR). During PPP, electrical stimulation was halted for 120 s. This allows for Ca²⁺ loading of the SR. The contraction force induced by the first electrical stimulation after the pause was calculated as a percentage of the average twitch force of the five last cycles before the pause. This was done before, at 4 days and at 8 days after treatment. During FFR assessment, the electrical stimulation frequency is increased in a stepwise fashion, while the duration of the respective step is shortened to keep the number of induced contractions constant (20–40 beats/step). After 380 h of cultivation, of which 180 h with treatment, the slices were removed from the BMCCs.

Statistics

To compare the contractility over time after treatment start (Figures 4A and S2A), a two-way ANOVA with Dunnett correction for multiple comparisons was performed. For each contractility comparison before and after medium change (Figure 4B), a Mann–Whitney U test was performed. To compare the potentiation of the HTx slices (Figure S2B), an ordinary one-way ANOVA with Dunnett correction for multiple comparisons was performed. All statistics were performed, and graphs were made using GraphPad Prism (Version 9.3.1, Boston, USA).

Clinical patient characteristics

At 8 weeks of age, a male patient presented for the first time with hypertrophic obstructive cardiomyopathy, extensive left ventricular outflow tract obstruction and pulmonary hypertension resulting in heart failure with tachydyspnoea and the need for high-flow oxygen therapy. Electrocardiogram showed signs of ventricular hypertrophy without arrhythmias (Figure S1). The patient's mother and sister had previously been diagnosed with NS prior to the current patient and also showed HCM. NS was clinically diagnosed and confirmed by a genetic panel showing a pathogenic variant in the RAF1 gene [c.775T>A (p.Ser259Thr)]. Phosphorylation of residue Ser259 results in autoinhibition of RAF1, while the pathogenic variant Thr259 leads to a gain-of-function phenotype by facilitating binding of GTP-bound RAS.²⁵ The prenatal and immediate perinatal course of this patient had been uneventful. After initial treatment with high-dose beta-blocker and disopyramide (at the time of treatment, institution guidelines contraindicated the use of calcium channel antagonists to avoid hypotension and deteriorating outflow tract obstruction), the current patient underwent open heart surgery for left ventricular outflow tract resection at the age of 14 months.

At 7.5 years of age, the patient (22.2 kg) was re-admitted due to aggravation of hypertrophy and severe obstruction in the RVOT in the absence of significant left ventricular outflow tract obstruction. An off-label treatment with the MEKi trametinib (Novartis Germany AG, Nuremberg) was initiated (0.011 mg/kg/day at start; increased to 0.023 mg/kg/day after 7 days), which resulted in a decrease of the peak RVOT gradient from initial 151 to 124 mmHg (Table 1 and Figure 1A), a reduction of the cardiac insufficiency marker NT-proBNP z-scores from 6.5 to 5.2 (Figure 1B), improvement of diastolic dysfunction as depicted by reversal of the abnormal mitral valve E to A ratio on echocardiography (Figure 2), and an improvement of right and left ventricular strain parameters on cardiac magnetic resonance (CMR). The left ventricular ejection fraction remained unchanged and hyperdynamic throughout the course of trametinib (Table 1). Representative echocardiographic and CMR images from the time points prior and after initiation of MEKi treatment are presented in Figures 2 and 3.

Despite an initial RVOT gradient reduction under trametinib treatment, severe RVOT obstruction remained with gradients between 110 and 120 mmHg after 5 months of MEKi treatment (Figures 1B and 3K,L). This prompted cardiac surgery for the resection of the RVOT obstruction. The resected RV myocardial tissue was further processed in an experimental setting, where four RV myectomy slices were generated. Slices were cultured for 380 h in BMCCs on a mechanical rocker at 37°C (5% CO₂).

Contraction force development during cultivation

After placement of the prepared myocardial tissue slices in BMCCs on the rocking device, the contraction force of the slices stabilized and was normalized to its value at 6 days (i.e., 140 h) of culture (pre-treatment baseline) (Figure 4). In all slices, every medium change resulted in a temporary depression of contraction force (▴, Figure 4), and therefore, these data points are omitted from the timelines. After 190 h of cultivation, treatment with rapamycin, dasatinib or trametinib was initiated. When comparing the contractile force of each slice after treatment start (Figure 4), the trametinib-treated (P = 0.023) and rapamycin-treated (P ≤ 0.001) slices showed a significantly decreased contraction force compared with DMSO 0.1%-treated control. Intriguingly, treatment with trametinib, but not any other compound, tended to abrogate the temporary depression of contraction force induced by the medium change (P = 0.057; Figure 4B). In a separate RV myocardial slice preparation and cultivation from a 51-year-old female HTx patient, a similar significant effect of trametinib (P < 0.001) and rapamycin (P = 0.016) on contraction force was observed after an identical ex vivo treatment regimen compared with vehicle control (Figure S2A).

Assessment of SR function by PPP and FFR

PPP correlating to sarcoplasmic Ca²⁺ loading was assessed before and after treatment. Vehicle treatment (DMSO 0.1%) did not affect PPP. Trametinib and rapamycin treatment resulted in a respectively 31.2% (173.3% pre-treatment to 204.5% post-treatment) and 26% (181% pre-treatment to 207.4% post-treatment) increase of potentiation compared with pre-treatment (Figure 5), reflecting increased sarcoplasmic Ca²⁺ loading. Dasatinib treatment lowered the potentiation by 9% (202.6% pre-treatment to 193.6% post-treatment). In the separate RV HTx preparation, trametinib showed the strongest potentiation increase (P = 0.195) (Figure S2B). The FFRs of the four individual NS patient tissue slices [reflecting L-type Ca²⁺ channel (LTCC) function] were slightly different in shape prior to treatment (Figure 6). However, 4 and 8 days of vehicle and dasatinib treatment did not alter the shape of the FFR. Conversely, trametinib treatment lowered the contraction force at low frequencies and increased the contraction force at higher frequencies after 8 days of treatment. Rapamycin treatment increased the contraction force at both lower and higher frequencies. The RV tissue slices from the HTx patient did not show any noteworthy changes in FFR before and after treatment.

Downregulation of contractility by trametinib and rapamycin

In vivo treatment of a paediatric NS-HCM patient led to a reduction in RVOT gradient and ameliorated radial strain, which is suggestive of reduced contractility. Similarly, an ex vivo down-regulation of contractile force was observed in the trametinib- and rapamycin-treated slices of the same patient but not in the dasatinib-treated slice. The difference in contractile force regulation between these pharmacological treatments is possibly caused by the difference in their treatment targets. Here, we used compounds that modulate the MAPK pathway at different levels, namely, PTPN11, MEK1/2 and the mammalian target of rapamycin (mTOR) (Figure 7). It is important to note that dasatinib inhibits PTPN11, which is located upstream of the pathogenic RAF1 variant. Such upstream inhibition might not impact SR function in a patient with a gain-of-function RAF1 variant. Both trametinib and rapamycin act downstream of the pathogenic RAF1 variant at different levels. Trametinib is a highly selective MEKi. As RAF1 forms a complex with MEK1/2 and extracellular signal-regulated kinase 1/2 (ERK1/2),²⁶ inhibition of MEK1/2 prevents the formation of this complex. Rapamycin inhibits mTOR, which is located further downstream of the pathogenic RAF1 variant within the MAPK pathway. Because both trametinib and rapamycin act downstream of the pathogenic RAF1 variant, they may aid in restoring nominal Ca²⁺ handling and SR loading. In the HTx slices from a patient with dilated cardiomyopathy, trametinib reduced contractile force more than rapamycin. This supports our hypothesis that trametinib influences Ca²⁺ handling.

Trametinib enhances PPP

To further elucidate Ca²⁺ handling of each treated slice, PPP and FFR stimulation protocols were run and compared with DMSO 0.1%-treated control slices. As part of a predesigned electrical stimulation protocol, electrical stimulation was temporarily halted for 120 s to assess the relative contribution of SR Ca²⁺ release via the ryanodine receptor (RyR) in enhancing contractility. Halting the electrical stimulation of the slices leads to Ca²⁺ loading of the SR. With the first stimulated contraction, the increased Ca²⁺ load of the SR is released through the RyR, leading to a more powerful contraction than observed during continuous electrical stimulation. The PPP can therefore be used as a parameter to assess the relative contribution of Ca²⁺ released from the SR.²⁴ Analysis of the PPP revealed that trametinib and rapamycin increased the relative contribution of Ca²⁺ release from the SR via the RyR. Conversely, dasatinib had no effect on this parameter. Similarly, in the HTx procured slices, trametinib showed the strongest potentiation increase compared with the control-treated slice.

The increase in potentiation observed in the trametinib-treated slice could be an effect of the drug's inhibitory effect on hyperphosphorylation of the RyR located on the SR.²⁷ The myocardial tissue in this study was resected from a patient carrying a gain-of-function RAF1 variant. RAF1 activation leads to the reduced expression of sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase (SERCA),^{26, 28} and variants of this gene in particular are associated with the activation of the Ca²⁺/calcineurin.²⁸ This consequently will impair Ca²⁺ sequestration in the SR and cause diastolic accumulation of cytosolic Ca²⁺ ([Ca²⁺]_Cyt). This hypothesis is in line with a study generating iPSC-CM of an NS-HCM patient with a similar RAF1 variant (Ser257Leu). Here, a reduction in SERCA/phospholamban (PLN) ratios compared with non-affected iPSC-CM was observed, in combination with significantly reduced Ca²⁺ transients between cytosol and SR.²⁶ Furthermore, iPSC-CM generated from an NS patient with an LZTR1 variant showed a significant increased Ca²⁺ leak from the SR, which was attenuated with the addition of calcium channel blocker.⁵ An increased [Ca²⁺]_Cyt promotes the binding of Ca²⁺ to calmodulin, followed by the activation of Ca²⁺/calmodulin-dependent protein kinase II (CaMKII). CaMKII phosphorylates the RyR2,²⁹ also located in the SR membrane, stimulating the release of Ca²⁺ from the SR,³⁰ thereby possibly further aiding in the increase of [Ca²⁺]_Cyt. In addition, CaMKII inhibits SERCA²⁹ and therefore promotes Ca²⁺ leakage from the SR.

Trametinib increases contraction force at higher beating frequencies

The FFR assesses the entry of Ca²⁺ into the myocardial cells as a function of stimulation frequency and contractility and aids in the assessment of excitation–contraction coupling.³¹ [Ca²⁺]_Cyt is maintained by SR Ca²⁺ loading, LTCC and Ca²⁺/Na⁺ exchanger.³¹ With an increasing stimulation frequency, the [Ca²⁺]_Cyt increases.³¹ Hence, the contraction force increases directly to the increase in stimulation frequency. Literature suggests that healthy myocardial tissues show a positive FFR.³² A flat or negative FFR is characteristic for a failing heart.³² The FFRs of the four slices were already different at baseline, which may reflect a difference in composition and morphology as has previously been observed.³² However, prior to initiation of the treatment, repeated measurements of the FFR within one slice yielded highly similar results, as indicated by the small error bars (SEM). Previous studies have described increased intracellular Ca²⁺ levels and currents in HCM patients.³³ This is in line with the previously described observation in NS patients that a pathogenic RAF1 variant is associated with reduced SERCA expression and promoted activation of Ca²⁺/calcineurin. Increased [Ca²⁺]_Cyt is observed to activate Ca²⁺/calmodulin, which subsequently facilitates the phosphorylation of LTCCs. This leads to an increased flux of Ca²⁺ into the cytosol, increasing the [Ca²⁺]_Cyt further.³⁴ Dysregulation of Ca²⁺ handling leads to an altered contraction of the myocardium.³⁴

The baseline FFR of all four cultivated slices showed a negative or flat form. This observation was to be expected as the patient was diagnosed with CHD, which resulted in heart failure. Similar FFR forms are typical for heart failure patients and are associated with alterations in cardiac mineral handling.³⁵ Previous research using the identical scientific set-up observed a positive FFR or negative FFR when human diseased left ventricular tissue slices had been cultivated at lower [0.12 Hz (7 b.p.m.)] or higher [0.5 Hz (30 b.p.m.)] baseline stimulation frequencies.^{23, 24} The plateau phase of the FFR diminished after long-term cultivation of 4–16 weeks.²³ When increasing the [Ca²⁺]_Cyt by adding Bay K8644, the FFR transitioned into an exclusively negative FFR as well.²⁴ In the present study, trametinib treatment changed the FFR from negative to mixed positive/negative FFR. The molecular mechanism behind the observation seen after treatment of trametinib is similar in the FFR as in the PPP. Trametinib inhibits MEK1/2, counteracting the gain-of-function RAF1 activity. This lowers the Ca²⁺ leak from the SR, decreasing the [Ca²⁺]_Cyt and partly restoring the function of RyR and SERCA. In case there would be a full restoration of the SR-located calcium channels, a positive FFR was expected. This might be a first step towards an FFR resembling healthy myocardial tissue. Dasatinib and rapamycin treatment did not change FFR.

The cardiomyocytes of the slices used in this study presumably have high [Ca²⁺]_Cyt due to dysfunctioning Ca²⁺ handling. It is speculated that at higher stimulation frequencies, Ca²⁺ leaks out of the SR at higher rates, preventing the necessary calcium spikes needed for proper myocardial contraction. Alternatively, another study hypothesized that LTCC and RyR are located closely together to facilitate rapid intracellular calcium release. In the failing myocardium, T-tubule remodelling causes a greater distance between LTCC and RyR, resulting in decreased cardiac contractility.³⁶ However, literature on T-tubule composition regarding LTCC in NS hearts is lacking.

Limitations

The main limitation of our study is the limited number of samples available and the use of an adult HTx patient with dilated cardiomyopathy as a control. NS-HCM is a rare disease, and the amount of resection tissue is very limited (two 5 × 5 × 10 mm samples). With a typical tissue slice size in our culture system of 7 × 7 mm with a thickness of 300 μm, we were limited to four NS-HCM slices. As a control preparation, eight adult HTx slices were made. Compared with the four NS-HCM slices, these differed in size, biopsy location, and patient characteristics like age and gender. These slices were treated with trametinib, rapamycin or dasatinib. There is a limited availability of literature on the use of these drugs in the NS setting, and we are the first to compare the effects of these drugs on intact tissue from a patient with NS-HCM. Our results point towards aberrant Ca²⁺ handling in the tissue slices, but it is unfortunately not possible to perform Ca²⁺ measurements in our system. The study described here would have benefited from calcium measurements, supported by measurements of the activity and expression levels of the LTCC, RyR and/or SERCA. Performing those additional experiments was not possible due to the limited size of the tissue biopsy.

An alternative approach that has been used in literature is the use of induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). Various studies have looked into the interaction between trametinib and RAF1 variations in iPSC-CM.^{26, 37} These predominantly focus on the RAF1 Ser257Leu variation. Although this is a different genetic variant closely located to the Ser259 phosphorylation docking site, it resulted in a 44% reduction in phosphorylation.²⁶ In this conformation, RAF1 will remain in the open, activated conformation. Trametinib treatment in iPSC-CM was shown to normalize MEK1/2 and ERK1/2.³⁷ As discussed before, one study generating iPSC-CM observed lower SERCA/PLN ratios in RAF1 Ser257Leu variant-affected iPSC-CM in combination with significantly reduced Ca²⁺ transients between cytosol and SR.²⁶ SERCA/PLN ratios were rescued with the application of MEKi. This is in line with our finding that treatment of paediatric NS-HCM myocardial tissue slices with MEKi trametinib improves SR loading, as evidenced by the increased potentiation.

As with iPSC-CM, the cultivation of myocardial tissue slices in our set-up requires regular medium change. Medium change of the BMCC tissue cultures led to changes in contractile force (Figure 4B). Here, loss of CO₂ in the incubator increased the culture medium pH causing an initially positive inotropic effect.²⁴ We hypothesize that the responses of cardiomyocyte contractility to pH changes involve the Na⁺/H⁺-exchanger 1 (NHE1), which increases Na⁺ influx under alkaline conditions and may be inhibited by SRC-family tyrosine kinases.³⁸ Treatment with trametinib attenuated this effect, suggesting that this treatment protects the myocardium from adverse effects of temporarily increased pH. Even though contraction force is partly influenced by the method itself, the use of the BMCC to culture myocardial tissue slices allows for the tissue to remain intact and exposed to preload and afterload, mimicking the biophysical in vivo environment. Future studies should aim to incorporate Ca²⁺ measurements in cultivated myocardial tissue slices to further elucidate the role of altered SR Ca²⁺ handling and evaluate the effects of novel therapies on SR Ca²⁺ handling in NS-HCM.