An FDA-approved drug structurally and phenotypically corrects the K210del mutation in genetic cardiomyopathy models
美国食品药品监督管理局批准的药物在结构和表型上纠正了遗传性心肌病模型中的 K210del 突变

Dilated cardiomyopathy (DCM) is a leading cause of heart failure and sudden cardiac death (1, 2). DCM, which is characterized by left ventricular chamber enlargement and diminished systolic performance, is a result of a primary genetic disorder in the absence of other predisposing factors (2, 3). Point mutations in sarcomeric proteins can result in decreased calcium sensitivity, which is the underlying mechanism of many DCMs (4, 5). Conceptually, understanding the underlying mechanism(s) of DCMs could facilitate the development of targeted therapies. However, the precise mechanism of decreased calcium sensitivity in DCMs is poorly understood. Of the numerous sarcomeric protein mutations that result in DCM, troponin complex mutations account for the majority of all DCM-causing mutations. The troponin complex consists of troponin I (TnnI), troponin T (TnnT), and troponin C (TnnC), representing the only functional unit in the sarcomere that could directly bind Ca²⁺ (6). TnnC contains 2 EF-hand (structural) Ca²⁺-binding domains and an activation Ca²⁺-binding domain (7). Each structural Ca²⁺-binding domain is also believed to bind 2 Mg²⁺ during the relaxation phase (7, 8), and thus, both systolic and diastolic impairment are often hallmarks of DCM mutations affecting the troponin complex (9). In 2000, the first DCM-causing mutation was identified as a TnnT2 Lysine 210 deletion (ΔK210 hereafter) (10, 11). A recent study revealed that K210 deletion not only reduces contractility in mutant cardiomyocytes, but also causes cellular hypertrophy and impairs cardiomyocytes’ ability to adapt to changes in substrate stiffness (12). Importantly, a number of potential mechanisms of impaired contractility in ΔK210 cardiomyocytes have been identified; however, to date, the structural consequence of ΔK210 has not been determined (13). In fact, the structural consequences of any sarcomeric protein mutations are unknown.
扩张型心肌病（DCM）是心力衰竭和猝死的主要原因（1，2）。DCM 以左心室扩大和收缩功能减退为特征，是其他诱发因素不存在的情况下原发性遗传性疾病的结果（2，3）。肌节蛋白的点突变可能导致钙敏感性降低，这是许多 DCM 的潜在机制（4，5）。从概念上讲，了解 DCM 的潜在机制有助于开发针对性治疗。然而，DCM 中钙敏感性降低的确切机制尚不清楚。在导致 DCM 的众多肌节蛋白突变中，肌钙蛋白复合物突变占所有 DCM 致病突变的多数。肌钙蛋白复合物由肌钙蛋白 I（TnnI）、肌钙蛋白 T（TnnT）和肌钙蛋白 C（TnnC）组成，代表肌节中唯一能够直接结合 Ca ²⁺ 的功能单元（6）。TnnC 含有 2 个 EF 手（结构）Ca ²⁺ 结合域和一个激活 Ca ²⁺ 结合域（7）。 每个结构 Ca ²⁺ -结合域也认为在松弛阶段（7，8）结合 2 个 Mg ²⁺ ，因此，收缩和舒张功能障碍通常是影响肌球蛋白复合物的 DCM 突变的标志（9）。在 2000 年，第一个导致 DCM 的突变被鉴定为 TnnT2 赖氨酸 210 缺失（ΔK210 之后）（10，11）。最近的一项研究表明，K210 缺失不仅降低了突变心肌细胞的收缩力，还导致细胞肥大并损害心肌细胞适应底物刚度变化的能力（12）。重要的是，已经确定了ΔK210 心肌细胞收缩功能障碍的多种潜在机制；然而，迄今为止，ΔK210 的结构后果尚未确定（13）。事实上，任何肌节蛋白突变的结构后果都是未知的。

Single amino acid deletion or interchange can result in substantial alteration for the protein structure and/or function, leading to diseases associated with misfolded proteins which can be defined as “proteinopathies” (14, 15). From a therapeutic standpoint, identifying the structural consequences of a mutation can allow for the custom design of targeted therapies to “fix” the mutant structure. There are several examples of drugs that correct proteinopathies that have demonstrated phenotypic correction of mutant proteins and the resulting pathologies. One such success story has been reported in cystic fibrosis (CF). CF is a proteinopathy caused by loss-of-function mutations in the CF transmembrane conductance regulator (CFTR) protein, a cAMP-regulated chloride channel expressed primarily at the apical plasma membrane of secretory epithelia in the airways, pancreas, intestine, and other tissues (16). Lumacaftor (VX-809), tezacaftor, and FDL169 were identified as correctors targeting CF, where they bind to defective F508del-CFTR protein to normalize mutant folding functionally, although without direct structural correction evidence (17–19). In addition, several drugs have been proposed as therapeutics of apolipoprotein E4–associated (ApoE4-associated) neurological manifestations via reducing the ApoE4’s neurotoxicity in vitro and ApoE4 fragment levels in vivo, also without evidence of structural correction at the protein level (20, 21). Most recently, Chen et al. showed that arsenic trioxide, a known antineoplastic agent, is a true structural corrector, which corrects the protein structure of a p53 mutation expressed in different types of cancer through a cryptic allosteric site at both the phenotypic and structural levels (22). Therefore, structure-based corrector therapy could be a viable approach to target pathological conformational changes leading to proteinopathies. Herein, we present what we believe is the first crystal structure for a common genetic DCM mutation, ΔK210, on the troponin complex. We reveal the differential Ca²⁺ binding between WT and ΔK210. Comprehensive structure analysis uncovers the underlying mechanism for the allosteric regulation of the activation Ca²⁺-binding domain on TnnC upon K210 deletion. In silico structure-based drug repositioning identifies risedronate acid (risedronate), an FDA-approved drug for the treatment/prevention of osteoporosis and Paget’s disease (23), as a potential structural corrector to improve the ΔK210 DCM phenotype.
单个氨基酸的缺失或替换可能导致蛋白质结构和/或功能的显著改变，从而导致与错误折叠蛋白质相关的疾病，这些疾病可以定义为“蛋白质病”（14，15）。从治疗的角度来看，识别突变的结构后果可以允许针对突变结构进行定制化治疗设计以“修复”突变结构。有几种药物可以纠正蛋白质病，这些药物已经证明了突变蛋白质及其引起的病理学的表型纠正。其中一个成功案例是囊性纤维化（CF）。CF 是由 CF 跨膜传导调节蛋白（CFTR）的功能丧失突变引起的蛋白质病，CFTR 是一种 cAMP 调节的氯通道，主要表达在呼吸道、胰腺、肠道和其他组织的分泌上皮细胞的顶端质膜上（16）。Lumacaftor（VX-809）、Tezacaftor 和 FDL169 被鉴定为针对 CF 的校正剂，它们与缺陷的 F508del-CFTR 蛋白结合，从功能上使突变折叠正常化，尽管没有直接的结构校正证据（17-19）。 此外，有几类药物被提议作为与载脂蛋白 E4 相关（ApoE4 相关）神经表现的疗法，通过在体外降低 ApoE4 的神经毒性以及在体内降低 ApoE4 片段水平，而且没有蛋白质水平上结构校正的证据（20，21）。最近，陈等人表明，已知的三氧化二砷是一种真正的结构校正剂，它通过在表型和结构水平上的隐蔽性别构位点，校正了在不同类型癌症中表达的 p53 突变蛋白结构（22）。因此，基于结构的校正剂疗法可能是一种针对导致蛋白质病的病理构象变化的可行方法。在此，我们展示了我们认为的第一个常见遗传性 DCM 突变ΔK210 在肌球蛋白复合物上的晶体结构。我们揭示了 WT 和ΔK210 之间的差异 Ca ²⁺ 结合。综合结构分析揭示了在 K210 删除后，TnnC 上激活 Ca ²⁺ 结合域的别构调节的潜在机制。 基于计算机模拟的结构药物重新定位识别了利塞膦酸（利塞膦酸），这是一种经 FDA 批准用于治疗/预防骨质疏松症和佩吉特病（23）的药物，作为改善ΔK210 DCM 表型的潜在结构校正剂。

Calcium-binding dynamics of WT and ΔK210 troponin complex. DCM mutations are known to be associated with decreased cardiomyocyte calcium sensitivity (24, 25). However, the potential structural basis for the decreased calcium sensitivity in any DCM mutation has never been determined. We hypothesized that decreased calcium sensitivity for DCM mutation is secondary to impaired Ca²⁺ stimulus response at the purified troponin complex level. In order to determine the effect of K210 deletion on calcium-induced conformational changes of the purified troponin complex, we first assembled WT troponin complex (WT hereafter) or the ΔK210 troponin complex in vitro by coexpressing 3 components of troponin complex (TnnI, TnnT2, and TnnC) together (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/JCI174081DS1IF: NA NA NA). To investigate the effect of K210 deletion on the Ca²⁺ binding and exchange at the activation domain of TnnC in vitro, we utilized the F27W reporter to monitor the fluorescence change during the Ca²⁺ titration using the purified troponin complex (26). We introduced F27W to TnnC of either WT and ΔK210 to generate WT-F27W and ΔK210-F27W. Using this system, the endogenous fluorescence during calcium titration reflects a conformational change in response to Ca²⁺. We found that changes in fluorescence upon Ca²⁺ titration were significantly different in the WT-F27W compared withΔK210-F27W, where the conformational changes in the mutant troponin complex were significantly reduced compared with WT (Figure 1A). This result is important because it indicates that there is an endogenous difference in response to Ca²⁺ in the purified troponin complex between WT and ΔK210. Collectively, our results reveal that K210 deletion alters the troponin complex’s response to Ca²⁺ binding, impairs the sarcomere force generation, and decreases calcium sensitivity.
钙结合动力学研究 WT 和ΔK210 肌钙蛋白复合物。DCM 突变已知与心肌细胞钙敏感性降低相关（24，25）。然而，任何 DCM 突变中钙敏感性降低的结构基础从未被确定。我们假设 DCM 突变导致的钙敏感性降低是纯化肌钙蛋白复合物中 Ca ²⁺ 刺激反应受损的继发性结果。为了确定 K210 缺失对纯化肌钙蛋白复合物中钙诱导构象变化的影响，我们首先通过共表达肌钙蛋白复合物的 3 个组分（TnnI、TnnT2 和 TnnC）体外组装了 WT 肌钙蛋白复合物（以下简称为 WT）或ΔK210 肌钙蛋白复合物（补充图 1A；与本文相关的补充材料可在网上获取；https://doi.org/10.1172/JCI174081DS1）。为了研究 K210 缺失对 TnnC 激活域中 Ca ²⁺ 结合和交换的影响，我们利用 F27W 报告基因在纯化肌钙蛋白复合物中通过钙 ²⁺ 滴定监测荧光变化（26）。 我们引入 F27W 到 WT 和ΔK210 的 TnnC 中，以生成 WT-F27W 和ΔK210-F27W。使用该系统，钙滴定过程中的内源性荧光反映了 Ca ²⁺ 的构象变化。我们发现，与ΔK210-F27W 相比，Ca ²⁺ 滴定引起的荧光变化在 WT-F27W 中显著不同，突变肌球蛋白复合物的构象变化与 WT 相比显著减少（图 1A）。这一结果很重要，因为它表明在纯化的肌球蛋白复合物中，WT 和ΔK210 对 Ca ²⁺ 的反应存在内源性差异。总的来说，我们的结果表明 K210 缺失改变了肌球蛋白复合物对 Ca ²⁺ 结合的反应，损害了肌节力量生成，并降低了钙敏感性。

Figure 1  图 1

Structure characterization of WT and ΔK210 complex. (A) Fluorescence-based measurement of the binding affinity of Ca²⁺ for WT and K210del to troponin complex. (B) Representation of the overall structure of ΔK210 and WT complex. TnnC is shown in green, TnnT is shown in orange, and TnnI is shown in purple. Ca²⁺ is shown in pink sphere. This color scheme is consistent for the ΔK210 complex in all the following figures unless otherwise specified. (C) Representation showing superimposition of WT (gray) and ΔK210. (D) Surface representation colored by the vacuum electrostatic potential of WT in the same orientation as in C. (E) Surface representation colored by the vacuum electrostatic potential of the ΔK210 complex in the same orientation as in C. (F) Highlight of K210 deletion site (in green dash circle) and hinge region of TnnT and TnnI showing superimposition of WT (gray) and ΔK210 (TnnC in green, TnnT in orange, and TnnI in purple). Specific residues in the hinge region are shown in stick representation. (G) Representation showing superimposition of Ca²⁺-binding domains in WT (gray) and ΔK210 (TnnC in green, TnnT in orange and TnnI in purple). (H) Detailed interaction of Ca²⁺ in the activation Ca²⁺-binding pocket for TnnC in ΔK210 complex. Specific residues coordinating the Ca²⁺are shown in green stick representation. Hydrogen bonds are indicated with black dashed lines. (I) Detailed interaction of Ca²⁺ in the activation Ca²⁺-binding pocket for TnnC in WT complex. Specific residues coordinating the Ca²⁺are shown in gray stick representation. Hydrogen bonds are indicated with black dashed lines. (J) Representation showing superimposition of the detailed interaction of Ca²⁺ in the activation Ca²⁺-binding pocket for TnnC in WT complex (gray) and ΔK210 complex (green). Data are presented as mean ± SEM; unpaired 2-sided t test. ****P < 0.0001.
WT 和ΔK210 复合物的结构表征。（A）基于荧光的 Ca ²⁺ 与 WT 和 K210del 对肌球蛋白复合物的结合亲和力测量。（B）ΔK210 和 WT 复合物整体结构的表示。TnnC 用绿色表示，TnnT 用橙色表示，TnnI 用紫色表示。Ca ²⁺ 用粉色球体表示。此颜色方案在所有后续图中均保持一致，除非另有说明。（C）显示 WT（灰色）和ΔK210 叠加的表示。（D）以 C 中相同方向显示的 WT 的真空静电势表面表示。（E）以 C 中相同方向显示的ΔK210 复合物的真空静电势表面表示。（F）突出显示 K210 删除位点（绿色虚线圆圈）和 TnnT 与 TnnI 的铰链区域，显示 WT（灰色）和ΔK210（TnnC 为绿色，TnnT 为橙色，TnnI 为紫色）的叠加。铰链区域的特定残基以棒状表示。（G）显示 WT（灰色）和ΔK210（TnnC 为绿色，TnnT 为橙色，TnnI 为紫色）中 Ca ²⁺ 结合域叠加的表示。 (H) Ca ²⁺ 在ΔK210 复合物中 TnnC 激活 Ca ²⁺ -结合口袋的详细相互作用。与 Ca ²⁺ 配位的特定残基以绿色棒状表示。氢键用黑色虚线表示。(I) Ca ²⁺ 在 WT 复合物中 TnnC 激活 Ca ²⁺ -结合口袋的详细相互作用。与 Ca ²⁺ 配位的特定残基以灰色棒状表示。氢键用黑色虚线表示。(J) 显示 WT 复合物（灰色）和ΔK210 复合物（绿色）中 Ca ²⁺ 在激活 Ca ²⁺ -结合口袋的详细相互作用的叠加表示。数据以平均值±SEM 表示；非配对双尾 t 检验。****P < 0.0001。

Overall structure of ΔK210. To investigate the potential structural basis for the altered regulation of Ca²⁺ binding, we determined a 3D structure of ΔK210 bound with Ca²⁺ at 3.1 Å resolution (Figure 1B and Supplemental Table 1). In this study, the same constructs were selected but with different complex assembly strategies and crystallization conditions (27). To eliminate the bias from the purification process and crystallization condition, WT bound with 3 Ca²⁺ was also determined at 3.1 Å resolution with the same purification procedure and crystallization condition as ΔK210 and then used as a reference for structure analysis (Figure 1B and Supplemental Table 1). However, extensive trials to obtain the ligand-free or Mg²⁺ bound complex structure for either WT or ΔK210 failed. The overall structures of both WT and ΔK210 bound with 3 Ca²⁺ are helical dominated. The whole troponin complex consists of TnnI and TnnT, and TnnC contains 2 EF-handed Ca²⁺-binding domains, of which the activation Ca-binding domain binds 1 Ca²⁺ (designated as Ca1) while the structural Ca²⁺ binding domain binds 2 Ca²⁺ (designated as Ca2 and Ca3). Consistent with the published WT structure (27), TnnI/T/C engaged in extensive contact with each other to form a stable heterotrimer, where the binding of the switch peptide at the C-terminal of TnnI to the activation Ca²⁺-binding domain of TnnC upon Ca²⁺ association fully activates the whole complex (Figure 1, B and C).
ΔK210 的整体结构。为了研究 Ca ²⁺ 结合改变的潜在结构基础，我们确定了ΔK210 与 Ca ²⁺ 结合的 3D 结构，分辨率为 3.1 Å（图 1B 和补充表 1）。在本研究中，选择了相同的构建体，但采用了不同的复合组装策略和结晶条件（27）。为了消除纯化过程和结晶条件带来的偏差，使用与ΔK210 相同的纯化程序和结晶条件，也确定了 WT 与 3 Ca ²⁺ 结合的 3.1 Å分辨率结构，并将其用作结构分析参考（图 1B 和补充表 1）。然而，为了获得 WT 或ΔK210 的无配体或 Mg ²⁺ 结合的复合结构，进行了广泛的试验均未成功。WT 和ΔK210 与 3 Ca ²⁺ 结合的整体结构以螺旋为主。整个肌钙蛋白复合物由 TnnI 和 TnnT 组成，TnnC 包含 2 个 EF 手型 Ca ²⁺ 结合域，其中激活 Ca 结合域结合 1 个 Ca ²⁺ （指定为 Ca1），而结构 Ca ²⁺ 结合域结合 2 个 Ca ²⁺ （指定为 Ca2 和 Ca3）。 与已发表的 WT 结构（27）一致，TnnI/T/C 之间广泛接触，形成稳定的异源三聚体，其中 TnnI 的 C 端切换肽与 TnnC 的激活 Ca ²⁺ -结合域在 Ca ²⁺ 结合后结合，完全激活整个复合物（图 1，B 和 C）。

All the resolved structures for our WT, ΔK210, and the published WT (PDB: 1J1D) (27) shared the same dimensions of unit cell with the same space group and there are 2 protomers of troponin complex per asymmetric unit. The protomer A on the left is more stable than promoter B. Also, the alignment of the 2 protomers in the asymmetric unit showed that the 2 protomers in ΔK210 adopt similar conformation (root mean square deviation [r.m.s.d.] 0.853 Å for 304 Cα of 331 Cα) while they are more flexible in our WT and the published WT structures (r.m.s.d 1.295 Å for 271 Cα of 329 Cα in our WT and r.m.s.d 1.552 Å for 288 Cα of 330 Cα in the published WT) because the activation Ca²⁺-binding domain on both protomers have relative different orientations due to intrinsic flexibility (Supplemental Figure 1, B–D). Thus, the more stable protomer A with the same packing environment for all 3 structures was selected for further analysis. Meanwhile, a thermal-shift assay (TSA) was performed to show that ΔK210 is slightly more stable than WT at apo state and in the presence of either Ca²⁺ or Mg²⁺ to suggest further validation of our observations in the crystal structure, although this was not statistically significant (Supplemental Table 2).
所有我们 WT、ΔK210 和已发表的 WT（PDB：1J1D）（27）的已解析结构具有相同的晶胞尺寸和相同的空间群，每个非对称单元中有 2 个肌球蛋白复合物的原聚体。左侧的原聚体 A 比原聚体 B 更稳定。此外，非对称单元中 2 个原聚体的对齐显示，ΔK210 中的 2 个原聚体采用相似的构象（对于 331 个 Cα的 304 个 Cα的均方根偏差[r.m.s.d.]为 0.853 Å），而它们在我们 WT 和已发表的 WT 结构中更灵活（我们 WT 中的 r.m.s.d 为 1.295 Å，对于 271 个 Cα的 329 个 Cα；已发表的 WT 中的 r.m.s.d 为 1.552 Å，对于 288 个 Cα的 330 个 Cα），因为两个原聚体的激活 Ca ²⁺ -结合域由于内在的灵活性而具有相对不同的取向（补充图 1，B–D）。因此，选择了更稳定的原聚体 A，它在所有 3 个结构中具有相同的包装环境，以进行进一步分析。 同时，进行了一种热位移分析（TSA），以表明ΔK210 在无活性状态和存在 Ca ²⁺ 或 Mg ²⁺ 的情况下比 WT 略稳定，这为进一步验证我们在晶体结构中的观察结果提供了依据，尽管这并不具有统计学意义（补充表 2）。

Allosteric regulation of activation Ca²⁺-binding domain on TnnC by K210 deletion. Allosteric regulation is defined as cryptic or neighboring sites not related to the active sites, where their conformational changes induced potential changes in the protein dynamics (28, 29). Structural comparison between our WT and the published WT showed that they adopted almost the same conformation (r.m.s.d 0.568 Å for 321 Cα of 339 Cα, Supplemental Figure 2A). But compared with WT, ΔK210 has marked differences in 3 regions: K210 deletion site, hinge region (residues from 220–228 in TnnT2 and from 82–91 in TnnI) close to K210 deletion site, and the activation Ca²⁺-binding domain on TnnC far from K210 deletion site (Figure 1C). These 3 regions exhibited unique surface topography and electrostatic potential in ΔK210, which is not seen in either WT structure (Figure 1, D and E, and Supplemental Figure 2B). Specifically, the surface charges at the site of K210 were greatly altered upon K210 deletion, where there are 2 loops moving in the hinge region between WT and ΔK210 adopting the exact same conformation (Figure 1, C–E, and Supplemental Figure 2, B and C). The movement of these 2 loops resulted in a newly formed narrow cavity in ΔK210 (Figure 1E) and a slight shift of the whole ITarm ITarm, which is coiled-coil region of TnnI and TnnT (Figure 1F). Moreover, the activation Ca²⁺-binding domain on TnnC underwent a 4° rotation from WT to ΔK210 due to those alterations in hinge region and ITarm while the orientation of the whole domain remained the same in our WT and published WT (Figure 1G and Supplemental Figure 1D). Surprisingly, a striking difference was observed between the activation Ca²⁺-binding domain on TnnC of WT and ΔK210, where the Ca²⁺ coordination network of ΔK210 was distorted upon K210 deletion. In addition, S69 on TnnC of ΔK210, which is involved in the coordination with Ca²⁺ in WT, was flipped away from the calcium-binding pocket, which leads to the loss of one bond for Ca²⁺ coordination (Figure 1, H–J, Supplemental Figure 2E, and Supplemental Figure 3, A and B) while the structural Ca²⁺-binding domain has no obvious change between WT and ΔK210 (Supplemental Figure 3C). Moreover, the average length of bonds for the other coordinated residues with Ca²⁺ in ΔK210 (average of 2.9 Å for 5 hydrogen bonds) is longer than in WT (average of 2.6 Å for 6 hydrogen bonds) (Figure 1, H and I). Therefore, K210 deletion at TnnT induces local change on TnnT and TnnI and allosterically regulates the conformation of the activation Ca²⁺-binding domain and the hydrogen bond network of Ca²⁺ coordination on TnnC. These results provide the first structural evidence, to our knowledge, of the mechanistic basis of decreased calcium sensitivity in a DCM as a result of conformational changes in the activation binding domain of TnnC.
非共价调节 TnnC 激活 Ca ²⁺ -结合域上的 K210 删除。非共价调节定义为与活性位点无关的隐蔽或邻近位点，它们的构象变化诱导蛋白质动态的潜在变化（28，29）。我们 WT 与已发表 WT 之间的结构比较表明，它们采用了几乎相同的构象（321 个 Cα原子与 339 个 Cα原子的均方根偏差为 0.568 Å，补充图 2A）。但与 WT 相比，ΔK210 在 3 个区域有显著差异：K210 删除位点、铰链区域（TnnT2 中的 220-228 位残基和 TnnI 中的 82-91 位残基）靠近 K210 删除位点，以及远离 K210 删除位点的 TnnC 激活 Ca ²⁺ -结合域（图 1C）。这三个区域在ΔK210 中表现出独特的表面拓扑结构和静电势，这在 WT 结构中都没有看到（图 1D 和 E，以及补充图 2B）。具体来说，K210 位点的表面电荷在 K210 删除后发生了很大变化，其中在 WT 和ΔK210 之间铰链区域移动的 2 个环采取了完全相同的构象（图 1C-E，以及补充图 2B 和 C）。 这些 2 个环的运动导致ΔK210（图 1E）中形成了一个新的狭窄空腔，并且整个 ITarm（TnnI 和 TnnT 的螺旋-螺旋区域）（图 1F）发生轻微偏移。此外，由于铰链区域和 ITarm 的这些变化，TnnC 上的激活 Ca ²⁺ 结合域从 WT 到ΔK210 发生了 4°的旋转，而整个域的取向在我们 WT 和已发表的 WT 中保持相同（图 1G 和补充图 1D）。令人惊讶的是，在 WT 和ΔK210 的 TnnC 上的激活 Ca ²⁺ 结合域之间观察到显著的差异，由于 K210 的删除，ΔK210 的 Ca ²⁺ 配位网络发生扭曲。此外，ΔK210 上的 S69，在 WT 中参与与 Ca ²⁺ 的配位，从钙结合口袋翻转开来，导致 Ca ²⁺ 配位丢失一个键（图 1，H–J，补充图 2E 和补充图 3，A 和 B），而结构上的 Ca ²⁺ -结合域在 WT 和ΔK210 之间没有明显变化（补充图 3C）。 此外，ΔK210 中与 Ca ²⁺ 配位的其他残基的键长平均值（5 个氢键的平均长度为 2.9 Å）比 WT（6 个氢键的平均长度为 2.6 Å）长（图 1，H 和 I）。因此，TnnT 中 K210 的缺失在 TnnT 和 TnnI 上诱导了局部变化，并通过别构调节了激活 Ca ²⁺ -结合域的构象和 TnnC 上 Ca ²⁺ 配位的氢键网络。这些结果为我们提供了第一份结构证据，证明了 DCM 中钙敏感性降低的机制基础，这是由于 TnnC 激活结合域的构象变化所致。

Pharmacological correction of the ΔK210 structural defects. Herein, we adopted our previous structure-based drug repurposing approach to identify FDA-approved drugs targeting ΔK210 (30). The MMFF94 energy-minimized library of FDA-approved drugs was prepared and docked to the hinge region in ΔK210 (Figure 2A). The screening resulted in 5 top hits belonging to the bisphosphonates therapeutic class based on their binding energies and interacting profiles (Figure 2B). We found that the bisphosphonates family bind to the induced hinge motif at (HLNEDQLR) via hydrophobic interactions and hydrogen bond interactions, of which risedronic acid showed proper binding mode with respect to the Chemgauss scoring function (31) to the hinge region of ΔK210 (Figure 2C).
药理学纠正ΔK210 结构缺陷。在此，我们采用了我们之前基于结构的药物重新定位方法来识别针对ΔK210 的 FDA 批准药物（30）。制备了 FDA 批准药物的 MMFF94 能量最小化库并将其对接到ΔK210 的铰链区域（图 2A）。筛选产生了基于其结合能和相互作用轮廓的 5 个顶级命中，属于双磷酸盐治疗类别（图 2B）。我们发现双磷酸盐家族通过疏水相互作用和氢键相互作用与诱导的铰链基序（HLNEDQLR）结合，其中瑞斯德隆酸与ΔK210 铰链区域的结合模式与 Chemgauss 评分函数（31）相匹配（图 2C）。

Figure 2  图 2

Structure correction of ΔK210 complex by risedronate. (A) Schematic flow chart for the in silico molecular virtual screening, starting with energy minimized FDA-approved small molecules to be followed by semiflexible docking study targeting the hinge region in ΔK210 complex to end up with the top 5 drug candidates. (B) Chemgauss scores for bisphosphonates family members: zoledronic, risedronic, pamidronic, alendronic, and ibandronic acids. (C) Surface representation colored by the vacuum electrostatic potential of ΔK210 complex. The hinge region is highlighted in black box and risedronate is docked into the hinge region. (D) Representation of the overall structure of ΔK210 complex in the presence of risedronate. (E) Representation showing superimposition of WT (gray), ΔK210 (cyan), and ΔK210 complex in the presence of risedronate (purple). (F) Surface representation colored by the vacuum electrostatic potential of ΔK210 complex in the presence of risedronate. (G) Highlight of the hinge region of TnnT and TnnI showing superimposition of WT (gray), ΔK210 (cyan), and ΔK210 complex in the presence of risedronate (purple). Specific residues in the hinge region are shown in stick representation. (H) Detailed interaction of Ca²⁺ in the activation Ca²⁺-binding pocket for TnnC in ΔK210 complex in the presence of risedronate. Specific residues coordinating the Ca²⁺are shown in green stick representation. Hydrogen bonds are indicated with black dashed lines. (I) Representation showing superimposition of the detailed interaction of Ca²⁺ in the activation Ca²⁺-binding pocket for TnnC in ΔK210 complex (cyan) and ΔK210 complex in the presence of risedronate (green). (J and K) Fluorescence-based measurement of the binding affinity of Ca²⁺ to WT- and ΔK210-troponin complexes in the absence and presence of risedronate. Significance was assessed through Tukey’s multiple-comparison tests. *P < 0.05; ****P < 0.0001.
;; ΔK210 复合物通过利塞膦酸钠进行结构校正。（A）基于能量最小化的 FDA 批准的小分子进行计算机分子虚拟筛选的流程图，随后进行半柔性对接研究，针对ΔK210 复合物的铰链区域，最终筛选出前 5 名药物候选者。（B）双膦酸盐家族成员的 Chemgauss 评分：唑来膦酸、利塞膦酸、帕米膦酸、阿仑膦酸和伊班膦酸。（C）以ΔK210 复合物的真空静电势着色的表面表示。铰链区域用黑色方框突出显示，利塞膦酸钠对接到铰链区域。（D）利塞膦酸钠存在下ΔK210 复合物的整体结构表示。（E）显示 WT（灰色）、ΔK210（青色）和利塞膦酸钠存在下ΔK210 复合物（紫色）叠加的表示。（F）利塞膦酸钠存在下ΔK210 复合物的真空静电势着色的表面表示。（G）突出显示 TnnT 和 TnnI 的铰链区域，显示 WT（灰色）、ΔK210（青色）和利塞膦酸钠存在下ΔK210 复合物（紫色）的叠加。 特定铰链区域的残基以棒状表示。(H) 在有 risedronate 存在的情况下，ΔK210 复合物中 TnnC 激活 Ca ²⁺ -结合口袋的详细相互作用与 Ca ²⁺ 的详细相互作用。(I) 显示 Ca ²⁺ 在激活 Ca ²⁺ -结合口袋中 TnnC 的详细相互作用（青色）和ΔK210 复合物（绿色）在 risedronate 存在下的叠加表示。(J 和 K) 在有和无 risedronate 存在的情况下，通过荧光测量 Ca ²⁺ 与 WT-和ΔK210-肌球蛋白复合物的结合亲和力。通过 Tukey 多重比较测试评估了显著性。*P < 0.05; ****P < 0.0001。

Next, the crystal structure of ΔK210 bound with Ca²⁺ in the presence of risedronate (ΔK210-risedronate hereafter) was resolved at 2.6 Å to elucidate how risedronate might alter the mutant K210 structure. The more stable protomer A was selected for further analysis (Supplemental Figure 4, B and C). The overall structure of ΔK210-risedronate matched the active form of the troponin complex (Figure 2, D and E, and Supplemental Figure 4, B–D). Although we did not see the density of risedronate in our crystal structure, notable differences were observed in the structure of ΔK210-risedronate, compared with ΔK210 in the absence of risedronate. First, the surface topography and electrostatic potential were altered compared with both WT and the ΔK210 in the absence of risedronate, especially in the hinge region (Figure 2F). The 2 loops consisting of the hinge region in the structure of ΔK210-risedronate were more similar to the conformation of WT, although not identical to WT. This could potentially explain why we did not observe risedronate in the corrected crystal structure, since the predicted docking site in the ΔK210 hinge region no longer existed.
接下来，在存在利塞膦酸的情况下，ΔK210 与 Ca ²⁺ 结合的晶体结构（以下简称ΔK210-利塞膦酸）在 2.6 Å下被解析，以阐明利塞膦酸可能如何改变突变体 K210 结构。更稳定的原聚体 A 被选用于进一步分析（补充图 4，B 和 C）。ΔK210-利塞膦酸的整体结构与肌钙蛋白复合物的活性形式相匹配（图 2，D 和 E，以及补充图 4，B-D）。尽管我们没有在我们的晶体结构中看到利塞膦酸的密度，但与无利塞膦酸的 WT 和ΔK210 相比，ΔK210-利塞膦酸的结构中观察到明显的差异。首先，与 WT 和无利塞膦酸的ΔK210 相比，表面地形和静电势发生了改变，尤其是在铰链区域（图 2F）。ΔK210-利塞膦酸结构中的 2 个由铰链区域组成的环与 WT 的构象更相似，尽管不是与 WT 完全相同。这可以解释为什么我们没有在修正后的晶体结构中观察到利塞膦酸，因为预测的ΔK210 铰链区域的对接位点已不存在。

In addition, we found that the side chain of L224 in TnnT and F90 in TnnI were restored to the same conformation as WT (Figure 2G). The overall orientation of both Ca²⁺-binding domains remained unchanged compared with ΔK210 by itself (Supplemental Figure 4E). Intriguingly the activation Ca²⁺-binding pocket had a substantial conformational change due to the allosteric effect observed at the hinge region. Specifically, the loop containing S69 showed a shift toward Ca²⁺ and the side chain of S69 was flipped inward back to the WT orientation to reestablish the coordination with Ca²⁺ (Figure 2, H and I, and Supplemental Figure 4, F and G). The structure of ΔK210 in the presence of the other members of the bisphosphonate family, such as alendronate, ibandronate, neridronate, and pamidronate, was resolved in the same crystallization condition as ΔK210 in the presence of risedronate (Supplemental Table 1). No change was observed in the hinge region and activation Ca-binding domain in any of these structures compared with ΔK210 alone (Supplemental Figure 4H), and no ligand density was observed. Next, we examined the effect of risedronate on the dynamic conformational change in the troponin complex using WT-F27W and ΔK210-F27W by monitoring the endogenous fluorescence during calcium titration. We found that risedronate significantly changed the fluorescence intensity upon Ca²⁺ titration for WT-F27W compared with ΔK210-F27W (Figure 2, J and K). Collectively, the presence of risedronate could alter the Ca²⁺-binding affinity by affecting the conformation of the hinge region and allosterically correct the defect of the activation Ca²⁺-binding network caused by K210 deletion.
此外，我们发现 TnnT 中 L224 的侧链和 TnnI 中的 F90 被恢复到与 WT 相同的构象（图 2G）。与ΔK210 本身相比，两个 Ca ²⁺ -结合域的整体取向保持不变（补充图 4E）。有趣的是，由于在铰链区域观察到的别构效应，激活 Ca ²⁺ -结合口袋发生了显著的构象变化。具体来说，包含 S69 的环向 Ca ²⁺ 偏移，S69 的侧链翻转回 WT 取向，以重新与 Ca ²⁺ 建立配位（图 2，H 和 I，以及补充图 4，F 和 G）。在存在其他双磷酸盐家族成员（如阿仑膦酸钠、伊班膦酸钠、尼立膦酸钠和帕米膦酸钠）的情况下，ΔK210 的结构在相同的结晶条件下被解析，与存在利塞膦酸钠的ΔK210 相同（补充表 1）。与ΔK210 单独相比，这些结构中的铰链区域和激活 Ca 结合域没有观察到变化（补充图 4H），并且没有观察到配体密度。 接下来，我们通过监测钙滴定过程中的内源性荧光，使用 WT-F27W 和ΔK210-F27W 研究了利塞膦酸钠对肌球蛋白复合物动态构象变化的影响。我们发现，与ΔK210-F27W 相比，利塞膦酸钠在 Ca ²⁺ 滴定时显著改变了 WT-F27W 的荧光强度（图 2，J 和 K）。总的来说，利塞膦酸钠的存在可以通过影响铰链区域的构象和别构正确校正由 K210 缺失引起的激活 Ca ²⁺ -结合网络的缺陷。

In silico computational analysis for WT and ΔK210 after crystallization. In order to gain insights into the potential allosteric effects of K210 deletion on the troponin complex and the mechanism of correction by risedronate, the resolved crystal structures of the troponin complex for WT, ΔK210, and ΔK210+risedronate were analyzed at the levels of contacts between different chains, backbone bond angles, dihedral angles, and calcium-binding domain. Generally, slight changes were observed for the interacting hydrophobic, ionic, and hydrogen bond interactions at the levels of contacts between TnnT, TnnI, and TnnC (Supplemental Table 3) and within the same chain (Supplemental Table 4) of troponin complex for WT, ΔK210, and ΔK210+risedronate. With regards to the backbone bond angle, TnnT showed changes in the bond angle starting from H223 till Q228, especially at L224 and N225. TnnC showed remarkable changes in the backbone bond angles as well, especially in the calcium-binding amino acids. Bond angle change propagated across TnnI from the N-terminal domain to the C-terminal domain. With regards to the dihedral angle validated by ramachandran plots, TnnT showed a change in the planarity starting from H223 till Q228 (more planar), while the planarity of the rest of the protein was unchanged. TnnC showed crucial change in dihedral angles, especially in the calcium-binding amino acids (D67 and S69). In addition, we found that TnnI in the mutant form was less planar compared with the WT. Importantly, the crystal structure of the troponin complex for ΔK210+risedronate showed normalization of the planarity changes induced by the K210 mutation, especially from H223 to Q228. In addition, the observed changes in TnnC and TnnI in the ΔK210 were normalized in the presence of risedronate. These in silico results indicate that K210 deletion results in important changes in backbone bond angles and dihedral angles across all 3 troponins involved in the troponin complex, providing support for the observed allosteric effect, and that risedronate, at least in part, reverses these changes.
在硅计算分析 WT 和ΔK210 结晶后的情况。为了深入了解 K210 缺失对肌球蛋白复合体的潜在变构效应以及通过利塞膦酸钠的校正机制，对 WT、ΔK210 和ΔK210+利塞膦酸钠的肌球蛋白复合体的解析晶体结构在不同链之间的接触、骨架键角、二面角和钙结合域进行了分析。一般来说，在 WT、ΔK210 和ΔK210+利塞膦酸钠的肌球蛋白复合体中，TnnT、TnnI 和 TnnC 之间的接触以及同一链内的相互作用疏水性、离子性和氢键相互作用水平上观察到轻微的变化（补充表 3）和同一链内（补充表 4）。至于骨架键角，TnnT 从 H223 到 Q228 的键角发生了变化，特别是在 L224 和 N225 处。TnnC 的骨架键角也发生了显著变化，特别是在钙结合氨基酸中。键角变化从 N 端域传播到 C 端域的 TnnI。 关于由拉马钱德兰图验证的二面角，TnnT 从 H223 到 Q228（更平面）显示出平面性的变化，而蛋白质其余部分的平面性保持不变。TnnC 在二面角上显示出关键变化，特别是在钙结合氨基酸（D67 和 S69）中。此外，我们发现突变型 TnnI 的平面性比 WT 型低。重要的是，ΔK210+risedronate 的肌钙蛋白复合物的晶体结构显示出由 K210 突变引起的平面性变化的正常化，特别是从 H223 到 Q228。此外，在 risedronate 存在下，ΔK210 中 TnnC 和 TnnI 的观察到的变化得到了正常化。这些计算机模拟结果表明，K210 缺失导致所有 3 个参与肌钙蛋白复合物的肌钙蛋白的骨架键角和二面角发生重要变化，为观察到的别构效应提供了支持，并且 risedronate 至少部分地逆转了这些变化。

In vitro pharmacological evaluation for ΔK210 human-induced pluripotent stem cells-cardiomyocytes. Next, we generated human induced pluripotent stem cell–derived (hiPSC-derived) cardiomyocytes from a healthy (WT), DCM patient with K210 mutation, and genetically corrected ΔK210. PBMCs were collected from a ΔK210 patient and healthy control, underwent standard reprogramming protocol to iPSCs, and were assessed for their morphology, pluripotency, and trilineage differentiation across the endoderm, mesoderm, and ectoderm (Figure 3, A and B, and Supplemental Figures 5–7). Following differentiation to cardiomyocytes, the contraction velocity was examined, which showed impaired contractility in the ΔK210 iPSC–derived cardiomyocytes and normalization of contraction velocity in a genetically corrected cell line (Figure 3, C and D). Next, we tested the effect of the bisphosphonate family members (zoledronate, pamidronate, and risedronate) on the contraction velocity of the ΔK210 hiPSC–derived cardiomyocytes and the genetically corrected line. The 3 members of the bisphosphonate family did not show any change in the contraction velocity of the genetically corrected cell line (Figure 3E). Interestingly, only risedronate showed a significant increase in the contraction velocity of the ΔK210 hiPSC–derived cardiomyocytes, while zoledronate and pamidronate did not show any significant effect (Figure 3F). These results suggest that risedronate normalizes the contraction velocity of primary ΔK210 hiPSC–derived cardiomyocytes and may be a viable therapeutic option for the treatment of DCM resulting from this mutation.
体外药理学评估ΔK210 人诱导多能干细胞-心肌细胞。接下来，我们从健康（WT）、DCM 患者 K210 突变和基因校正的ΔK210 中生成人诱导多能干细胞来源（hiPSC 来源）的心肌细胞。从ΔK210 患者和健康对照者收集 PBMCs，经过标准重编程方案到 iPSCs，并对其形态、多能性和内胚层、中胚层和外胚层的三系分化进行评估（图 3A 和 B，补充图 5-7）。分化为心肌细胞后，检查了收缩速度，显示ΔK210 iPSC 来源的心肌细胞收缩力受损，而在基因校正的细胞系中收缩速度正常化（图 3C 和 D）。接下来，我们测试了双磷酸盐家族成员（唑来膦酸、帕米膦酸和利塞膦酸）对ΔK210 hiPSC 来源的心肌细胞和基因校正线的收缩速度的影响。双磷酸盐家族的 3 个成员在基因校正的细胞系中收缩速度没有变化（图 3E）。 有趣的是，只有利塞膦酸钠显示出ΔK210 hiPSC 来源的心肌细胞收缩速度的显著增加，而唑来膦酸钠和帕米膦酸钠没有显示出任何显著效果（图 3F）。这些结果表明，利塞膦酸钠可正常化原代ΔK210 hiPSC 来源的心肌细胞的收缩速度，可能是治疗由这种突变引起的扩张型心肌病的一种可行的治疗选择。

Figure 3  图 3

Risedronate corrects ΔK210 hiPSC-derived cardiomyocyte contraction. (A) Immunofluorescence of TnnT2^WT and TnnT2^K210+/– stained with DAPI (blue), cTnT (green), and α-actinin (red). (B) Morphology of iPSCs for WT, DCM, and genetically corrected DCM. (C) Contraction velocity for WT, DCM, and genetically corrected DCM iPSCs. (D) Contraction velocity for DCM and genetically corrected DCM iPSCs from day 0 till day 7. (E) Contraction velocity for genetically corrected DCM iPSCs treated with zoledronic acid (H, 300 ng/mL and L, 30 ng/mL), pamidronate (H, 1000 ng/mL and L, 100 ng/mL), and risedronate (H, 10 ng/mL and L, 2 ng/mL) with no difference at 2 dose levels. (F) Contraction velocity for DCM iPSCs treated with zoledronic acid (H, 300 ng/mL and L, 30 ng/mL), pamidronate (H, 1000 ng/mL and L, 100 ng/mL), and risedronate (H, 10 ng/mL and L, 2 ng/mL) at 2 dose levels, where risedronate showed significant increase in the contraction velocity of DCM-derived iPSCs. Data are presented as mean ± SEM; unpaired 2-sided t test. *P < 0.05; ****P < 0.0001. Scale bars: 100 μm (B).
瑞昔膦酸纠正ΔK210 hiPSC 来源的心肌细胞收缩。（A）用 DAPI（蓝色）、cTnT（绿色）和α-肌动蛋白（红色）染色的 TnnT2 ^WT 和 TnnT2 ^K210+/– 的免疫荧光。（B）WT、DCM 和基因校正的 DCM iPSCs 的形态。（C）WT、DCM 和基因校正的 DCM iPSCs 的收缩速度。（D）从第 0 天到第 7 天 DCM 和基因校正的 DCM iPSCs 的收缩速度。（E）用唑来膦酸（H，300 ng/mL 和 L，30 ng/mL）、帕米膦酸（H，1000 ng/mL 和 L，100 ng/mL）和瑞昔膦酸（H，10 ng/mL 和 L，2 ng/mL）处理的基因校正的 DCM iPSCs 的收缩速度，两种剂量水平无差异。（F）在两种剂量水平下，用唑来膦酸（H，300 ng/mL 和 L，30 ng/mL）、帕米膦酸（H，1000 ng/mL 和 L，100 ng/mL）和瑞昔膦酸（H，10 ng/mL 和 L，2 ng/mL）处理的 DCM iPSCs 的收缩速度，其中瑞昔膦酸显示出 DCM 来源的 iPSCs 收缩速度的显著增加。数据以平均值±SEM 表示；非配对双尾 t 检验。*P < 0.05；****P < 0.0001。刻度条：100 μm（B）。

In vivo pharmacological evaluation of risedronate ΔK210 DCM mouse model. To investigate whether risedronate could correct cardiac contraction of the ΔK210 in vivo, we developed a ΔK210 DCM mouse model using clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9), which led to an expected K210 deletion in TnnT2. Sequencing analysis confirmed the effective deletion of K210 in 1 allele, generating a heterozygote ΔK210 DCM mouse model (TnnT2^K210+/–) (Figure 4A). As expected, the isolated skinned papillary muscle fibers from this model exhibited dampened calcium-dependent force generation (Figure 4, B and C). At submaximal calcium concentration pCa 5.7, the force produced by TnnT2^K210+/– papillary muscle (5.42 ± 0.24 mN/mm²) was 50% less compared with WT (12.50 ± 1.21 mN/mm²) (Figure 4D). However, at maximal calcium concentration pCa 4.5, both TnnT2^K210+/– (36.96 ± 0.68 mN/mm²) and WT (38.14 ± 0.80 mN/mm²) papillary muscles produced similar and statistically insignificant force (Figure 4E). The calcium sensitivity pCa₅₀ in TnnT2^K210+/– (papillary fibers dropped to 5.482 ± 0.009 mN/mm² when compared with WT at 5.604 ± 0.012) (Figure 4F). Risedronate at 0.5 mM dosage was able to reverse the effect of TnnT2^K210+/– on force generation and calcium sensitivity (Figure 4, C–F). TnnT2^K210+/– developed a submaximal force of 11.91 ± 0.93 mN/mm² which was similar and statistically insignificant (P = 0.8626) compared with that of WT (12.50 ± 1.21 mN/mm²) and WT treated with 0.5 mM risedronate (12.86 ± 1.06 mN/mm²). Risedronate also improved the calcium sensitivity of K210 mutant to pCa 5.613 ± 0.006 mN/mm² and to that of WT 5.604 ± 0.012 mN/mm² and WT treated with 0.5 mM risedronate: 5.630 ± 0.021 mN/mm². The skinned papillary experiments demonstrated that risedronate at 0.5 mM can selectively improve and rescue the calcium-dependent force production in TnnT2^K210+/– and not in WT papillary muscle. Our in vivo pharmacokinetic (PK) studies involved s.c. injection of 150 μg/kg/d for female CD-1 WT mice to account for the parameters of PK to show correlated plasma concentration (C_max 213 ng/mL) along with the predicted human plasma concentrations (Supplemental Figure 8), where peak plasma concentrations in patients undergoing treatment with risedronate are approximately 24 ng/mL (corresponding to 84 nM concentration) (32).
体内磷酸盐钠ΔK210 DCM 小鼠模型药理学评估。为了研究磷酸盐钠是否可以纠正ΔK210 在体内的心脏收缩，我们利用成簇规律间隔短回文重复序列（CRISPR）/CRISPR 相关蛋白 9（Cas9）构建了ΔK210 DCM 小鼠模型，导致 TnnT2 中预期的 K210 缺失。测序分析证实了 K210 在 1 个等位基因中的有效缺失，生成了异合子ΔK210 DCM 小鼠模型（TnnT2 ^K210+/– ）（图 4A）。正如预期，从这个模型中分离出的去衣乳头肌纤维表现出减弱的钙依赖性力生成（图 4，B 和 C）。在亚最大钙浓度 pCa 5.7 时，TnnT2 ^K210+/– 乳头肌产生的力（5.42 ± 0.24 mN/mm ² ）比 WT（12.50 ± 1.21 mN/mm ² ）低 50%（图 4D）。然而，在最大钙浓度 pCa 4.5 时，TnnT2 ^K210+/– （36.96 ± 0.68 mN/mm ² ）和 WT（38.14 ± 0.80 mN/mm ² ）乳头肌产生的力和 WT 相似，且统计学上无显著性差异（图 4E）。 钙敏感性 pCa ₅₀ 在 TnnT2 ^K210+/– （与 WT 相比，乳头肌纤维下降至 5.482 ± 0.009 mN/mm ² ）时（图 4F）。0.5 mM 剂量的利塞膦酸能够逆转 TnnT2 ^K210+/– 对力生成和钙敏感性的影响（图 4，C–F）。TnnT2 ^K210+/– 发育出亚最大力量 11.91 ± 0.93 mN/mm ² ，与 WT（12.50 ± 1.21 mN/mm ² ）和 0.5 mM 利塞膦酸处理的 WT（12.86 ± 1.06 mN/mm ² ）相似且统计学上不显著（P = 0.8626）。利塞膦酸还改善了 K210 突变体的钙敏感性至 pCa 5.613 ± 0.006 mN/mm ² ，以及 WT 的 5.604 ± 0.012 mN/mm ² 和 0.5 mM 利塞膦酸处理的 WT：5.630 ± 0.021 mN/mm ² 。去衣乳头肌实验表明，0.5 mM 的利塞膦酸可以选择性改善和挽救 TnnT2 ^K210+/– 中的钙依赖性力生成，而不是 WT 乳头肌。 我们的体内药代动力学（PK）研究涉及对雌性 CD-1 WT 小鼠进行 150 μg/kg/d 的皮下注射，以考虑 PK 参数，以显示相关的血浆浓度（C _max 213 ng/mL）以及预测的人血浆浓度（补充图 8），其中接受利塞膦酸钠治疗的患者峰值血浆浓度约为 24 ng/mL（相当于 84 nM 浓度）（32）。

Figure 4  图 4

Administration of risedronate improves force generation in TnnT^K210+/– papillary muscles. (A) Sanger sequencing chromatogram obtained for the PCR amplicons generated from TnnT^K210+/– mice. (B) Calcium-dependent force generation in WT and TnnT^K210+/– papillary muscles. (C–F) Calcium-dependent force generation in WT, TnnT^K210+/–, and 0.5mM risedronate-treated WT or TnnT^K210+/– papillary muscles. (C) Force produced at submaximal pCa 5.7 (D) and pCa 4.5 (E), and calcium sensitivity pCa₅₀ (F). Data are represented as mean ± SEM. Statistical significance: 1-way ANOVA with Tukey’s test (B–F). ***P < 0.001. (B–F) n = 3; 1 fiber per animal.
双膦酸盐瑞斯德隆酸的应用可提高 TnnT ^K210+/– 扁桃体肌的力生成。（A）从 TnnT ^K210+/– 小鼠中生成的 PCR 扩增子获得的 Sanger 测序色谱图。（B）野生型（WT）和 TnnT ^K210+/– 扁桃体肌的钙依赖性力生成。（C–F）野生型、TnnT ^K210+/– 和 0.5mM 瑞斯德隆酸处理的野生型或 TnnT ^K210+/– 扁桃体肌的钙依赖性力生成。（C）亚最大 pCa 5.7 时产生的力（D）和 pCa 4.5 时产生的力（E），以及钙敏感性 pCa ₅₀ （F）。数据表示为平均值 ± SEM。统计显著性：1-way ANOVA 与 Tukey 检验（B–F）。***P < 0.001。（B–F）n = 3；每只动物 1 根纤维。

To determine the correction capacity of risedronate in TnnT2^K210+/– mice, we examined cardiac function following daily risedronate injection (150 μg/kg/d, s.c. – human dose equivalent for the treatment of osteoporosis) (33) by echocardiography every 2 weeks for 11 weeks, and a confirmation by cardiac MRI examination at 15 weeks, after which the animals were sacrificed, and the hearts processed (Figure 5, A–G). Echocardiography was first conducted at around the onset of puberty (at 5–6 weeks of age), which revealed a mildly depressed left ventricular ejection fraction (LVEF) on TnnT2^K210+/– mice even at this early developmental stage, compared with WT (Figure 5B). Interestingly, following a 2-week risedronate treatment, the LVEF of TnnT2^K210+/– mice was significantly increased, compared with vehicle-treated TnnT2^K210+/– mice (Figure 5B and Supplemental Figure 9A). We observed no significant difference in cardiac damage markers released in the bloodstream at this time point (Supplemental Figure 9, B–D). After 4 weeks, risedronate treatment demonstrated a notable increase in the LVEF of TnnT2^K210+/– mice, reaching values comparable to WT levels. This effect remained consistent throughout the study (Figure 5B) with a comparable heart rate among groups (Supplemental Figure 10A). These findings were corroborated by MRI examination following 15 weeks of risedronate treatment, which also indicated a significant rise in LVEF and lower circumferential strain (index of myocardial contractility) in the risedronate-treated TnnT2^K201+/– mice, compared with the vehicle-treated group (Figure 5, C–E). Risedronate-treated TnnT2^K201+/– hearts did not exhibit marked alterations in interstitial fibrosis (Figure 5F); however, they displayed a significant reduction in the cardiac cross-sectional area cell size (Figure 5G), with a trend of reduction in heart-to-body weight ratio (Supplemental Figure 10B), compared with vehicle-treated TnnT2^K201+/– mice.
为了确定利塞膦酸钠在 TnnT2 ^K210+/– 小鼠中的校正容量，我们通过每周两次的超声心动图检查（每两周一次，共 11 周）和在第 15 周进行的心脏 MRI 检查来评估每日利塞膦酸钠注射（150 μg/kg/d，s.c. – 骨质疏松症治疗的人体剂量等效物）后的心脏功能，并在之后处死动物，处理心脏（图 5，A–G）。超声心动图首次在青春期开始时进行（约 5-6 周大），即使在早期发育阶段，与 WT 小鼠相比，TnnT2 ^K210+/– 小鼠的左心室射血分数（LVEF）略有下降（图 5B）。有趣的是，经过两周的利塞膦酸钠治疗后，与接受溶剂处理的 TnnT2 ^K210+/– 小鼠相比，TnnT2 ^K210+/– 小鼠的 LVEF 显著增加（图 5B 和补充图 9A）。在此时间点，我们观察到血液中释放的心脏损伤标志物没有显著差异（补充图 9，B–D）。经过四周治疗后，利塞膦酸钠治疗显著提高了 TnnT2 ^K210+/– 小鼠的 LVEF，达到与 WT 小鼠相当的水平。 该效应在整个研究中保持一致（图 5B）且各组心率相当（补充图 10A）。这些发现通过 15 周利塞膦酸治疗后进行的 MRI 检查得到证实，这也表明与安慰剂处理的组相比，利塞膦酸处理的 TnnT2 ^K201+/– 小鼠的左心室射血分数（LVEF）显著升高，周向应变（心肌收缩力的指标）降低（图 5，C–E）。利塞膦酸处理的 TnnT2 ^K201+/– 心脏在间质纤维化方面没有明显变化（图 5F）；然而，与安慰剂处理的 TnnT2 ^K201+/– 小鼠相比，它们的心脏横截面积细胞大小显著减少（图 5G），心脏与体重比呈下降趋势（补充图 10B）。

Figure 5  图 5

Administration of risedronate improves cardiac contraction function in TnnTK210^+/– models. (A) Schematic for vehicle and risedronate administration to WT and TnnTK210^+/– mice at 150 μg/kg/d for 15 weeks with monitoring EF using echocardiography and MRI. (B) Serial echocardiography assessment of LVEF showing elevated LVEF on risedronate-treated TnnTK210^+/–. (C) MRI study showing representative images of end-diastole and end-systole for vehicle and risedronate treated TnnTK210^+/–. (D) MRI studies revealed elevated LVEF and (E) lower circumferential strain in risedronate treated TnnTK210^+/–. (F) Masson’s trichrome staining of heart sections showed nonsignificant change in the interstitial fibrosis. (G) WGA staining and cross-sectional area (CSA) quantification showed a significant decrease in cardiomyocyte cell size for risedronate-treated TnnTK210^+/–, compared with vehicle-treated TnnTK210^+/–. (H and I) Representative echocardiography images and serial echocardiography for cross-over study, where TnnTK210^+/– mice at 8 weeks of risedronate administration (150μg/kg/day) were switched to vehicle treatment for 4 more weeks and vice versa Data are represented as mean ± SEM. Statistical significance: 2-way ANOVA with Tukey’s test (B and I), and unpaired 2-sided t test (D–G). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (B) Vehicle-treated WT (n = 8), vehicle-treated TnnTK210^+/– (n = 17), risedronate-treated WT (n = 8), risedronate(150) treated TnnTK210^+/– (n = 11), and risedronate (75) treated TnnTK210^+/– (n = 9), where 150 and 75 show concentration of risedronate. (D and E) Vehicle-treated TnnTK210^+/– (n = 6), risedronate(150)-treated TnnTK210^+/– (n = 5), where 150 indicates concentration of risedronate. (F) Vehicle-treated TnnTK210^+/– (n = 15), risedronate-treated TnnTK210^+/– (n = 6). (G) Vehicle-treated TnnTK210^+/– (n = 6); risedronate-treated TnnTK210^+/– (n = 5). (I) Risedronate-to-vehicle (n = 5), vehicle-to-risedronate (n = 10).
双膦酸盐阿仑膦酸钠的应用可改善 TnnTK210 ^+/– 模型的心脏收缩功能。（A）示意图：以 150 μg/kg/d 的剂量对 WT 和 TnnTK210 ^+/– 小鼠进行阿仑膦酸钠和溶剂处理，持续 15 周，并使用超声心动图和 MRI 监测 EF。（B）连续超声心动图评估 LVEF，显示阿仑膦酸钠处理的 TnnTK210 ^+/– 的 LVEF 升高。（C）MRI 研究显示了溶剂和阿仑膦酸钠处理的 TnnTK210 ^+/– 的舒张末期和收缩末期代表性图像。（D）MRI 研究揭示了阿仑膦酸钠处理的 TnnTK210 ^+/– 的 LVEF 升高和（E）周向应变降低。（F）Masson 三色染色显示心肌切片的间质纤维化无显著变化。（G）WGA 染色和横截面积（CSA）量化显示，与溶剂处理的 TnnTK210 ^+/– 相比，阿仑膦酸钠处理的 TnnTK210 ^+/– 的心肌细胞大小显著减小。（H 和 I）代表性超声心动图图像和交叉研究中的连续超声心动图，其中 TnnTK210 ^+/– 小鼠在阿仑膦酸钠给药（150μg/kg/天）8 周后，改为溶剂治疗 4 周，反之亦然。数据以平均值±标准误表示。 统计显著性：双因素方差分析及 Tukey 检验（B 和 I），以及非配对双尾 t 检验（D–G）。*P < 0.05；**P < 0.01；***P < 0.001；****P < 0.0001。（B）载体处理野生型（n = 8），载体处理 TnnTK210 ^+/– （n = 17），利塞膦酸处理野生型（n = 8），利塞膦酸（150）处理 TnnTK210 ^+/– （n = 11），以及利塞膦酸（75）处理 TnnTK210 ^+/– （n = 9），其中 150 和 75 表示利塞膦酸的浓度。（D 和 E）载体处理 TnnTK210 ^+/– （n = 6），利塞膦酸（150）处理 TnnTK210 ^+/– （n = 5），其中 150 表示利塞膦酸的浓度。（F）载体处理 TnnTK210 ^+/– （n = 15），利塞膦酸处理 TnnTK210 ^+/– （n = 6）。（G）载体处理 TnnTK210 ^+/– （n = 6）；利塞膦酸处理 TnnTK210 ^+/– （n = 5）。（I）利塞膦酸至载体（n = 5），载体至利塞膦酸（n = 10）。

To ascertain the dependence of TnnT2^K201+/– cardiac function correction on the presence of risedronate, we conducted a crossover study aimed at assessing risedronate’s capacity to reverse the decline in LVEF. Following an 8-week period of risedronate treatment, risedronate-treated TnnT2^K201+/– mice exhibited an increase in LVEF similar to the earlier cohort, whereas the group treated with the vehicle displayed a progressive decline in LVEF. Subsequently, the 2 groups underwent a switch, with the vehicle-treated group transitioning to risedronate and vice versa. This transition led to a notable increase in LVEF upon initiation of risedronate and a gradual decline upon its discontinuation (Figure 5, H and I), thus reaffirming the initial findings.

Furthermore, we tested another dose level for risedronate to mimic the administered therapeutic dose for the treatment of osteoporosis in humans, where we administered risedronate (75 μg/kg/day, s.c.) for 5–6 weeks in TnnT^K210+/– mice (Supplemental Figure 10C). This study showed that risedronate treatment at this lower dose also resulted in a significant increase in LVEF (Supplemental Figure 10, D and E) without any changes in heart rate, heart-to-body weight ratio (Supplemental Figure 10, A and B), or interstitial fibrosis (Supplemental Figure 10F), compared with the vehicle-treated group. Finally, to address the correction ability of risedronate in later and more severe stages of heart failure, we conducted risedronate treatment in TnnT^K210+/– old mice (approximately a year of age). The results indicate that risedronate treatment rescued LVEF in TnnT^K210+/– mice (Supplemental Figure 11A), albeit at a slower rate than in young TnnT^K210+/– mice (significant improvement at 6 weeks versus 2 weeks of treatment in old versus young TnnT^K210+/– mice). In contrast, vehicle-treated TnnT^K210+/– mice exhibited a plateau EF value of around 70%–60%. While we did not observe a severe drop in EF of vehicle-treated TnnT^K210+/– mice at this age, we did note a 25% mortality rate (data not shown) during the treatment course, a sign of an enlarged left atrial appendage (Supplemental Figure 11B), and a modest increase in heart weight (HW)/body weight (BW) ratio (Supplemental Figure 11C), compared with risedronate-treated TnnT^K210+/– mice. Collectively, these results indicate that risedronate can improve LV systolic function for DCM resulting from K210del.

In this report we uncover the first, to our knowledge, structural insights into the mechanism of decreased calcium sensitivity of a DCM mutation. We show that ΔK210 results in allosteric structural changes in TnnC, resulting in calcium discoordination at S69. In addition, we provide multiple levels of evidence that identify an FDA-approved drug that corrects both the structure and function of the troponin complex and restores normal left ventricular function in heterozygous mice carrying the K210 mutation. Numerous genetic mutations affecting the troponin complex have been associated with hypertrophic cardiomyopathy (HCM) or DCM (34, 35); however, the mechanistic underpinning of how a specific mutation affects the structure and function of the troponin complex remains poorly understood. The troponin complex is highly dynamic with an extensive allosteric network, and any perturbation of amino acids could potentially change the overall dynamic state of the protein and, therefore, affect its function on calcium handling (36, 37). For example, the HCM mutation D145E on TnnC has been shown to allosterically transmit along a loosely structured backbone to allow TnnC to function only through 1 Ca²⁺ with high sensitivity (38). In addition, recent simulation studies have suggested that mutations could affect the structure and dynamics of the troponin complex and impact thin filament regulation (39, 40).

The current study offers several advances on both the mechanistic and translational fronts. First, our study demonstrates the first, to our knowledge, crystal structure of a mutant troponin complex, which revealed the mechanistic basis of decreased calcium sensitivity as a result of allosteric regulation of the activation Ca²⁺-binding domain of TnnC upon K210 deletion. We show that K210 deletion causes a local conformational change of TnnT close to the deletion site and an allosteric change of TnnC at S69, resulting in calcium discoordination. This was validated by in silico computational analysis showing that K210 deletion caused changes in the interacting contact between the 3 chains after refinement for the newly formed ionic bonds: ASP3 till LYS46, GLU213 till ARG98, and ARG147 till ASP 269 at the interacting chains TnnC with TnnI, TnnT with TnnI, and TnnT with TnnC, respectively (Supplemental Table 3). Also, K210 deletion resulted in changes in the interacting contact within the same chain after refinement for newly formed ionic bonds for LYS106 till GLU110 for TnnI and formation of new hydrogen bonds at TnnC (LYS21 till ASP25, SER35 till VAL72, and ASP141 till GLY146), TnnT (GLU234 till THR238), and TnnI (ARG111 till GLU115) (Supplemental Table 4).

Second, we provide the first evidence, to our knowledge, of a structure corrector of a sarcomeric protein mutation. Our in silico screen identified the bisphosphonate family, an FDA-approved class of drugs used for the treatment of osteoporosis, as candidates for binding to the hinge region of the mutant TnnT. Cocrystallization of the bisphosphonate candidates with the mutant troponin complex revealed that only risedronate corrected the mutant troponin structure, resulting in restoring S69 of TnnC to its original position. It is important to note here, however, that we did not visualize a risedronate density in any of the crystal structures obtained. A possible explanation for this observation is that the hinge region identified in the mutant TnnT (Figure 2, C and F), which was used for docking in our in silico screen, was no longer visualized in the corrected crystal in the presence of risedronate. This also suggests that risedronate binding to the troponin complex is likely weak, which is a desirable feature since strong binding might interfere with the sarcomere function.

Third, we provide the first evidence, to our knowledge, of pharmacological phenotypic correction of a cardiomyopathy mutation using both primary mutant human iPSC-derived cardiomyocytes and a new mutant mouse model that we generated. We show that risedronate normalizes contraction velocity in primary mutant human iPSC-derived cardiomyocytes, whereas other bisphosphonate family members (that did not correct the crystal structure) do not. We found that risedronate normalized LV systolic function when administered systemically in mice at FDA-approved doses. This occurred both at the maximum and lower dose range used for the treatment of osteoporosis. In addition, we also show that risedronate treatment rescued EF in ΔK210 old mice, which is a potential benefit for patients harboring long-term mutations. Furthermore, risedronate treatment prevented cardiac hypertrophy, as shown by the reduction in cross-sectional area measurement of cardiomyocytes. Intriguingly, risedronate also had a similar effect in a separate crossover study, where discontinuation of risedronate was associated with a rapid decline in LV function and initiation of risedronate in the cohort previously assigned to vehicle treatment resulted in rapid normalization of LV function.

It is important to note that the platform used in the current study may not be widely applicable to all other sarcomeric gene mutations. For example, we obtained the crystal structure of another TnnT mutation, R205L (Supplemental Figure 12); however, we did not observe any structural abnormalities, supporting the notions that mutations may affect a wide range of functions such as binding of regulatory proteins or association with other proteins not included in the crystal structure.

Collectively, these results demonstrate the validity of understanding the structural consequences of sarcomeric protein mutations and highlight a potential therapeutic strategy for genetic cardiomyopathies using existing drugs, which can facilitate rapid and affordable delivery of care to this vulnerable patient population.

Sex as a biological variable

Both male and female K210del mice were studied in the entire in vivo studies.

Protein expression and purification

hTnnT2 with residues 183–288 and hTnnI with residue 32–166 were subcloned into the first and second open reading frame (ORF) of pRSFDuet vector, respectively, and hTnnC with residue 1–161 was subcloned into the first ORF of pETDuet vector with N terminal 6xHis-tag followed by a protease from Tobacco Etch Virus protease recognition site, and both plasmids were transformed into Rosetta (DE3) pLysS cells (Novagen). According to the previously published structure, the cysteine-less variant TnnC (C35S/C84S) and the variant TnnI (T31M/C80A/C97A) were used. For the K210 deletion mutation (ΔK210), the cDNA was synthesized by Genescript and further subcloned into the first ORF of pRSFDuet vector for later cotransformation. Targeted protein was expressed in cultures grown in autoinduction media at 18°C overnight (41). The culture was harvested and sonicated in lysis buffer (50 mM Tris [pH 8.0], 150 mM NaCl, 1 mM DTT, 1 mM CaCl₂) and supplemented with protease inhibitors. The lysate was centrifuged, the supernatant was loaded onto a Ni-NTA affinity column (QIAGEN), and the beads were washed with wash buffer (20 mM Tris [pH 8.0], 150 mM NaCl, 1 mM DTT, and 20 mM imidazole [pH 8.0]) and eluted with elution buffer (20 mM Tris [pH 8.0], 150 mM NaCl, 1 mM DTT, and 250 mM Imidazole [pH 8.0]). The eluate was digested by TEV protease at 4°C overnight and then further purified by ion exchange chromatography followed by gel filtration chromatography. The peak fractions were collected and concentrated to about 25 mg/mL for crystallization screening with the gel filtration buffer (20 mM Tris [pH 8.0], 150 mM NaCl, 1 mM DTT, and 1 mM CaCl₂).

Crystallization and structure determination

The crystals of the TnnT2/TnnI/TnnC WT and ΔK210 mutant were obtained using the hanging drop, vapor-diffusion method by mixing 1 μL protein (25 mg/mL) with 1 μL reservoir solution containing 0.2 M sodium acetate and 20% PEG 3350 and incubating at 18°C. The crystals were observed after 3 days and reached the maximum size after 7 days. The crystals were flash-frozen in liquid nitrogen with 15% glycerol as the cryoprotectant. The datasets were collected at APS-19-ID at wavelengths of 0.97926 Å. Data were indexed, integrated, and scaled by the program HKL3000 (42). Phases were determined by molecular replacement program Phaser in CCP4 package (43) using the WT TnnT2/TnnI/TnnC structure (PDB ID: 1J1D) as a searching model. The model was further built manually with COOT (44) and iteratively refined using Phenix.refine (45). The PROCHECK program was used to check the quality of the final model, which shows good stereochemistry according to the Ramachandran plot (46). All structure figures were generated by using the PyMOL Molecular Graphics System, Schrödinger, LLC. Software used in this project was curated by SBGrid (47).

In silico molecular docking simulations

The US Food and Drug Administration–approved drug database was downloaded (http://www.drugbank.ca), and 3D structures were energy minimized using the MMFF94 force field. The whole library underwent multiconformer generation using Omega 2.5.1.4. The crystallized ΔK210 mutant model was used as a template structure via placing a box around the induced conformational changed site (LNEDQLR) to be docked with the energy-minimized FDA library via Fast Rigid Exhaustive Docking (FRED), version 3.0.1 (OpenEye Scientific Software, OEDOCKING 4.3.2.0. OpenEye, Cadence Molecular Sciences, Inc., http://www.eyesopen.com). The scoring assessment was conducted by validating the docked poses and estimating the Chemgauss score. 3D visualization was conducted using Vida and Pymol (31, 48, 49).

Both Tnnt^K210del and TnnT^K210del+risedronate underwent molecular dynamic (MD) simulations using NAMD (50) to predict the trajectories after MD simulations, visualized by VMD (51) or Pymol. This was typically started with energy minimization to remove steric clashes, followed by equilibrating the system under constant-temperature, constant-pressure (NPT) ensemble at 300 K and 1 bar to equilibrate the box size. Herein, we applied initially 20 ns NPT for both of Tnnt^K210del and TnnT^K210del+risedronate to be interpreted in terms of relative mean square deviation (RMSD), relative mean square fluctuations (RMSF), and radius of gyration (protein compactness).

In silico computational analysis for WT and ΔK210 after crystallization

Crystal structures of the Troponin complex for WT and K210del were analyzed at the levels of contacts between different chains, backbone bond angles, dihedral angles, and calcium-binding domain using the molecular operating environment (MOE), and z scores (cut-off values) were adjusted to satisfy the minor changes all over the whole complex.

Thermal shift assay

The samples for TSA were treated with both 3mM EGTA and 3mM EDTA during ion exchange purification and stored in the gel filtration buffer without CaCl₂. Solutions containing 5 μL of 1 mg/mL protein, 9 μL of TSA buffer (20 mM Tris [pH 8.0], 150 mM NaCl, 1 mM DTT, and 1mM CaCl₂), and 1 μL of SYPRO Orange (diluted 1/25 in water, purchased from MilliporeSigma) were added to PCR tubes and the final volume for the reaction was 15 μL. The PCR tubes were heated in an i-Cycler iQ5 Real-Time PCR Detection System (Bio-Rad) from 30°C to 95°C with an increment of 1°C. The fluorescence signals for each tube were monitored by a charge-couple camera (CCD). The melting temperature for the protein unfolding transition was analyzed by the built-in melt curve calculation mode.

Determination of Ca²⁺ binding induced fluorescence change

The steady-state fluorescence measurements were carried out for WT^F27W and ΔK210^F27W. Briefly, the fluorescence emission of WT^F27W and ΔK210^F27W were measured during Ca²⁺ titration by using an excitation wavelength of 276 nm and an emission wavelength of 340 nm. The fluorescence changes upon Ca²⁺ binding were determined by subtracting the fluorescence at maximal Ca²⁺ concentration from all other measurements and then expressing the resultant values as percentages of the maximum fluorescence. The data were plotted by GraphPad, version 8.

Reprogramming of iPSCs

Healthy and patient-specific iPSCs were collected and reprogrammed from human PBMCs using the CytoTune-iPS 2.0 Sendai Reprogramming Kit (Thermo Fisher Scientific) and cultured on feeder-free Matrigel-coated culture plates.

Cell culture

Human iPSCs were routinely grown on Matrigel-coated (Corning) 6-well plates using chemically defined E8 medium (Thermo Fisher Scientific) and passaged at a ratio of 1:12 every 4 days using Accutase solution (Sigma-Aldrich). hiPSC-CMs were maintained in a RPMI 1640 medium (Thermo Fisher Scientific) supplemented with B27 supplements (Thermo Fisher Scientific). The cells were maintained in the incubator with 5% CO₂ at 37°C.

Trilineage differentiation

The pluripotency of the hiPSCs was assessed by differentiating into 3 germ layers. The Human Pluripotent Stem Cell Functional Identification Kit (R&D Systems) was used to differentiate hiPSCs into mesoderm and ectoderm, and the STEMdiff Definitive Endoderm Kit (STEMCELL Technologies) was used to differentiate hiPSCs into endoderm.

Immunofluorescent staining

Cells were fixed in 4% paraformaldehyde (PFA) solution (Thermo Fisher Scientific) for 10 minutes and permeabilized with 0.1% Triton X-100 (Sigma Aldrich) in PBS (Thermo Fisher Scientific) for 10 minutes at room temperature. Cells were then blocked with 3% BSA (Sigma Aldrich) in PBS for 30 minutes, followed by overnight incubation at 4°C with 1% BSA solution containing a 1:200 dilution of primary antibodies. Cells were washed 3 times with 0.1% Tween-20 (Sigma Aldrich) in PBS, followed by incubation with 1% BSA solution containing secondary antibodies for 60 minutes at room temperature in the dark. Nuclei were counterstained with NucBlue Fixed Cell ReadyProbes Reagent (Thermo Fisher Scientific). All the antibodies used in the immunofluorescence staining are shown in Supplemental Table 5.

Karyotyping

hiPSCs chromosomal aberrations were detected using the KaryoStat assay (Thermo Fisher Scientific).

Sequencing

DNA was extracted from hiPSCs using DNeasy Blood & Tissue Kits (QIAGEN). PrimeSTAR GXL DNA Polymerase (Clontech) was used for PCR. The sequence of primers used for genotyping is as follows: forward primer, 5′-TGATCCTTCTTGCCCCTACCT; and reverse primer, 5′-TTCTTGCTGTGAGCCACCAGA.

Differentiation of hiPSCs into hiPSC-CMs

hiPSCs were split at 1:12 ratio using the Accutase solution (Sigma-Aldrich). When hiPSCs reached a confluency of 80%, the E8 medium was changed to RPMI supplemented with B27 without insulin (Life Technologies) and 6 μM of the glycogen synthase kinase 3-β inhibitor CHIR-99021 (Selleck Chemicals) for 2 days (D0–D1). On day 2, the medium was aspirated and replaced with RPMI+B27 minus. On day 3, the medium was changed to RPMI+B27 minus with 5 μM of the Wnt inhibitor IWR-1 (Selleck Chemicals) for 2 days. The medium was replaced with RPMI-B27 minus for 2 days on day 5, and then switched to RPMI-B27 for 3 days on day 7. On day 10, the medium was changed to RPMI-B27 without D-glucose (Life Technologies) for 3 days. This glucose starvation step further purified cardiomyocyte culture. After 2 days’ recovery using the medium RPMI-B27, cells were dissociated after 5 minutes incubation with TrypLE Select Enzyme (10×) (Thermo Fisher Scientific) at 37°C followed by seeding into 6-well plates cultured with RPMI-B27 with 10% Knock-Out Serum Replacement (KOSR) (Thermo Fisher Scientific). One day later, the medium was changed back to RPMI-B27. hiPSC-CMs were cultured in RPMI-B27 for experiments after the second purification as described above.

Drug treatment

Human iPSC-CMs were treated with FDA-approved candidate drugs for 2 and 7 days. DMSO (Sigma-Aldrich) was used as a control treatment unless noted. Cells were treated with zoledronic acid at 30 (low [L]) or 300 (high [H]) ng/mL, pamidronate at 100 (L) or 1000 (H) ng/mL, baclofen at 100 (L) or 1000 (H) ng/mL, fenoprofen at 10 (L) or 50 (H) μg/mL, or risedronate at 2 (L) or 10 (H) ng/mL. The medium containing fresh drugs was replaced every 2 days.

Contractility analysis

hiPSC-CMs were dissociated using TrypLE Select Enzyme (10×) (Thermo Fisher Scientific) for 5 minutes at 37°C. Afterward, cells were seeded on Matrigel-coated 96-well plates (50,000 cells per well) and cultured for 7 days to recover their synchronous beating before the assay. Phase-contrast cell motion movies of cardiomyocyte contraction were recorded using the Sony SI8000 Cell Motion Imaging System (37°C and 5% CO₂). Before data collection, the cells were equilibrated for 15 minutes. Video images of the hiPSC-CMs were recorded for 10 seconds, at a frame rate of 75 frames per second (fps), and a resolution of 1,024 × 1,024 pixels using a ×10 objective. Data analysis was performed using SI8000C Analyzer software (Sony).

Preparation of skinned papillary muscle fiber

Twelve- to fourteen-week-old mice were euthanized and hearts were isolated and placed in the chilled PBS buffer. Hearts were subjected to dissection along the aorta line, to divide it in left and right ventricles. A pair of papillary muscles was isolated and placed in the skinning buffer, to skin overnight at 4°C. Skinning buffer comprised 1% Triton X-100 in buffer containing 55.74 mM potassium propionate, 7 mM ethylene glycol bis(2-aminoethyl) tetra acetic acid, 100 mM N,N-bis(2-hydroxyethyl)-2-amino ethane sulfonic acid, 20 μM calcium chloride, 5.5 mM magnesium chloride, 5 mM dithiothreitol, 15 mM creatine phosphate, and 4.7 mM adenosine triphosphate with pH adjusted to 7.0 with 4 M potassium hydroxide and ionic strength maintained at 180 with potassium propionate. The calcium strength in the above buffer was minimal with pCa 9.0 also termed as relaxing buffer since the muscles don’t contract in the pCa 9.0 buffer (52–54). After overnight skinning, the fibers were ready for experimental use.

Calcium-dependent force generation in skinned papillary muscles

The skinned papillary muscle fibers were dissected in fibers of width of approximately 100 μM and then mounted between a high-speed length controller (Aurora Scientific 322C) and force transducer (Aurora Scientific 403A) using aluminum T-clips (55). The fibers were cycled through buffers containing increasing amounts of calcium concentration; pCa 6.0, 5.8, 5.7, 5.6, 5.4, and 4.5. All the pCa solutions were made by mixing the relaxing buffer (pCa 9.0) and activating buffer (pCa 4.5). The activating buffer (pCa 4.5) is the same as the relaxing buffer except for the addition of 7 mM calcium chloride. The fibers cycling through pCa solutions were allowed to achieve saturating maximal force before being subjected to the next pCa buffer. Risedronate was added to all the pCa solutions to assess their effect on calcium-dependent force generation in WT versus the TnnT2 K210 mutant mice skinned cardiac papillary fibers. 100× stock of risedronate was diluted to its experimental concentration to avoid excessive dilution of the pCa solution. For control experiments, PBS was added in the same amount as risedronate in pCa solution to offset any effects due to dilution. The comparison of calcium-dependent force generation and calcium sensitivity across control and test experiments was calculated by plotting the maximal force generated against the respective pCa. The data were fitted to a sigmoidal dose-response curve with a variable slope (GraphPad Prism, version 9).

KI-HDR using ssODN template insertion

Mice (BALB/cJ, JAX_000651) harboring the mutated Tnnt2 alleles were generated using CRISPR/Cas9 reagents at the Transgenic Technology Center of the University of Texas Southwestern Medical Center. The guide RNAs and donor ssODN were designed to delete lysine residue K210 and introduce a silent ClaI restriction site for genotyping purposes. The sgRNA sequence is AGAAGAAGATCCTGGCAGAG. The guide was selected using the CRISPR Design Tool (https://chopchop.cbu.uib.no/). crRNA and tracRNA were annealed and mixed with Cas9 protein to form a ribonucleotide protein complex. The ssODN (IDT) sequence (AGACAGAGCGGAAGAGTGGGAAGAGACAGACAGAGAGAGAGAAGAAGAAAATCCTGGCCGAGCGAAGGAAGGCGCTGGCAATCGATCATCTGAATG AAGACCAACTGAGGTGGGGACAGTTGGTTGGGTGGCCCCTGGCACTCTTCCTGA) was added to the mix and the cocktail was microinjected into the cytoplasm of fertilized one-cell eggs isolated from super ovulated females. The eggs were incubated in media containing cytochalasin-B immediately before and during microinjection to improve egg survival. The surviving eggs were transferred into the oviducts of day 0.5 pseudo pregnant recipient ICR females (Envigo, Inc.) to produce putative founder mice. Founder mice were identified via PCR using the primer set 5′-TGGGTCTTTCTCTCATGGTTTCC-3′ and 5′-GCTCAGATAAGAAAAGGGCCT-3′ and the amplicon was submitted for Sanger sequencing. F0 mice were bred with BALB/cJ mice to obtain F1 mice heterozygous for the mutated allele. The mouse’s age is indicated in the text of each experiment. Littermate controls were used whenever possible. No statistical methods were used to predetermine the sample size.

Drug dose(s) and administration

Our therapeutic regimen was designed to mimic the administered human dose ranges for risedronate at 2 levels, 75 and 150 μg/kg/day, s.c. for osteoporosis.

Transthoracic echocardiography

Assessment of in vivo heart function on conscious, nonsedated mice was performed using a Vevo2100 micro-ultrasound system, MS400C probe (VisualSonics), at baseline, 2 weeks after drug administration, and 4, 6, 8, and 11 weeks after drug administration. Echocardiographic M-mode images were obtained from a parasternal short-axis view at the level of the papillary muscles. Left ventricular internal diameters at end-diastole (LVIDd) and end-systole (LVIDs) were measured from M-mode recordings. Six representative contraction cycles were selected for analysis, and average indexes (LVIDd, LVIDs, and fractional shortening) were calculated for each mouse. All echocardiography measurements were performed in a blinded manner.

Cardiac MRI

MRI was performed on a 7T preclinical scanner (Bruker Biospec) using a 72 mm volume transmitter coil with a 2 × 2 phased array surface receiver coil. Mice were anesthetized with 1.5–2.5% isoflurane. The animal’s ambient temperature was maintained at 28°C using an MR Compatible Small Rodent Air Heater System (SA Instruments). Imaging was performed with prospective gating for ECG (SA Instruments) monitored using needle electrodes connected to front paws. Cine images in the short-axis plane were obtained using a gradient echo (FLASH) sequence. The following imaging parameters were used: echo time (TE)/repetition time (TR) = 3.9/10 ms; number of k-space lines per R-R = 1; slice thickness = 1 mm, number of averages = 3; flip = 15°; field of view (FOV) = 30 × 30 mm²; matrix = 192 × 192; in-plane resolution = 0.15 × 0.15 mm³. Five to 6 contiguous slices were obtained. MR tagging was performed on a short-axis slice at the mid-ventricular level using the SPAMM method. Imaging parameters were TE/TR = 4/15 ms; tag separation = 1 mm; tag thickness = 0.2 mm; number of k-space lines per R-R = 1; slice thickness = 1 mm, number of averages = 3; flip = 20°; FOV = 30 × 30 mm²; matrix = 192 × 192; in-plane resolution = 0.15 × 0.15mm³. Cardiac cine images were analyzed using the freely available software, Segment version 3.0 (http://segment.heiberg.se) (56). Myocardial strain analysis was performed using HARP (Diagnosoft Plus, Diagnosoft Inc.). Ventricular end-diastolic volumes, LV ejection fraction, and mean myocardial Eularian circumferential strain (Ecc) were calculated.

In vivo PK studies

Drug preparation and administration. Risedronate was dissolved in PBS, and 0.1 mL was administered to each mouse via the s.c. route to give a dose of 150 μg/kg. Mice were dosed daily for 4 days and after the fourth and final dose, animals were bled via the submandibular route at various time points into a K2EDTA tube, which was placed on ice. Blood was then centrifuged at 9,600g for 10 minutes to obtain plasma, which was stored at –80°C until analysis. Animals were then euthanized by inhalation overdose of CO₂ followed by a brief perfusion with PBS before isolation of tissues. Briefly, a butterfly needle attached to a P1 Peristaltic Pump (Pharmacia Fine Chemical) was inserted into the left ventricle, a nick was made in the right atrium, and PBS was infused at a rate of 2.25 mL/min until the liver had a blanched appearance (typically 20–25 mL PBS). The heart and kidneys were removed, weighed, and snap-frozen in liquid nitrogen. Tissues were homogenized in PBS (3× weight by volume) before analysis as described below.

Risedronate PK. Risedronate levels in mouse plasma and tissue homogenate were monitored by liquid chromatography/tandem mass spectrometry (LC-MS/MS) using an AB Sciex 6500 QTRAP mass spectrometer coupled to a Shimadzu Nexera LC. Risedronate was detected with the mass spectrometer in positive multiple reaction monitoring (MRM) mode by following the precursor to fragment ion transitions 340.2 to 214.2. A Phenomenex Kinetix C18 column (2.6 micron, 150 × 4.6 mm) was used for chromatography for PK studies with the following conditions: buffer A: 10% acetonitrile containing 10 mM ammonium acetate, buffer B: 90% acetonitrile containing 10 mM ammonium acetate; 0–0.5 min 12% B, 0.5–2.55 min gradient to 100% B, 2.55–3.5 min 100% B, 3.5–3.51 min gradient to 12% B, 3.51–5.0 12% B. Risedronate-d4 (transition 344.2-218.2) from Toronto Research Chemicals was used as an internal standard (IS). Samples of 50 μL were diluted with 500 μL of deuterated water (dH20), 100 μL of 2% BSA containing 250 ng/mL of the risdronate-d4 IS, and 10 μL of 1 M hydrochloric acid (HCl). UCT Inc. Clean-up Quaternary Amine w/Hydroxide acid (SPE), columns were sequentially conditioned with 0.5 mL methanol, 0.5 mL dH20, and 0.5 mL 20 mM HCl. Diluted samples were passed through these SPE columns by gravity, and columns were washed with 0.5 mL 20 mM HCL. Risedronate was eluted by gravity by sequentially adding 0.5 mL methanol, 0.5 mL 20% TMS-diazomethane in methanol, and 0.5 mL methanol. Samples were dried under nitrogen at 37°C, reconstituted with 100 μL 10 mM NH₄ acetate/acetonitrile 80:20 v:v, incubated at room temperature for 10 minutes, and spun at 812g. The supernatant was then analyzed as described above. Standard curves were generated using blank plasma (Bioreclamation) spiked with known concentrations of risedronate in DMSO and processed as described above. The concentration of drug in each time-point sample was quantified using Analyst software (Sciex), version 1.7.3. A value of 3-fold above the signal obtained from blank plasma or tissue homogenate was designated the limit of detection (LOD). The limit of quantitation (LOQ) was defined as the lowest concentration at which back calculation yielded a concentration within 20% of theoretical. PK parameters were determined using the noncompartmental analysis tool in Phoenix WinNonlin (Certara Corp).

Histology

Mouse hearts were collected and fixed in 4% PFA in PBS overnight at 4°C and then processed for paraffin embedding. Masson’s trichrome staining was performed according to standard procedures at the University of Texas Southwestern Medical Center core histology facility on paraffin sections.

Wheat germ agglutinin staining and cardiomyocyte size quantification

Wheat germ agglutinin (WGA) staining and quantification were performed as previously described (57, 58). In brief, the slides were incubated with WGA conjugated to Alexa Fluor 488 (50 mg/mL, Life Technologies) for 1 hour at room temperature following washing with 1× PBS. To quantify the cross-sectional cell size, 3 to 5 independent hearts per group with 3 different views and positions, each from left and right ventricles and septum, were captured at ×40 magnification. ImageJ (NIH) was used to quantify the size of round cardiomyocytes that contained a nucleus. At least 500 cells per sample were quantified.

Assessment of sera cardiac stress markers

Mice were intraperitoneally injected with risedronate (150 μg/kg/d) for 2 weeks. Submandibular blood was collected and clotted for 2 hours at room temperature before centrifugation for 15 minutes at 1000g at 4°C. The supernatant was collected for ELISA assay following the instrument’s protocol (proANP, Novus_68143; proBNP, Novus_76775; TnnI, Novus_00456). In brief, the standard samples and sera were incubated into atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), or TnnI coating 96-well plates for 90 minutes at 37°C. The liquid was subsequently removed before biotinylated detection antibody working solution incubation for 1 hour at 37°C. Next, the well was incubated with HRP conjugate working solution for 30 minutes at 37°C. After 30 minutes of incubation, the liquid in the well was discarded, and substrate reagent was added before stopping solution incubation to measure the concentration by a microplate reader. The optical density was measured by 450 nm and the result was analyzed by Prism, version 8.

Statistics

Data were analyzed and graphed using Prism (GraphPad), version 8. Data were presented as mean ± SEM. Comparisons were conducted via either 2-tailed Student’s t test, 1-way ANOVA, or 2-way ANOVA, followed by Turkey’s post hoc test with statistically significant differences.

Study approval

All mouse experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Texas Southwestern Medical Center and complied with the relevant ethical regulations regarding animal research. The study was carried out under Stanford Institutional Review Board and Stem Cell Research Oversight Committee guidelines.

Data availability

All data supporting the findings in this study are included in the main article and Supporting Data Values file. The crystal structure was deposited to the protein data bank with accession codes PDB 8FMM (WT), 8FMN (ΔK210), 8FMO (ΔK210-risedronate), 8FMP (ΔK210-pamidronate), 8FMQ (ΔK210-alendronate), 8FMR (ΔK210-ibandronate), 8FMS (ΔK210-neridronate), and R205L (8FMT).