High-throughput platform for depolarization-induced Ca2+ release for skeletal muscle disease mutation validation and drug discovery

In skeletal muscle excitation-contraction (E-C) coupling, depolarization of the plasma membrane triggers Ca2+ release from the sarcoplasmic reticulum (SR), referred to as depolarization-induced Ca2+ release (DICR). DICR occurs via the type 1 ryanodine receptor (RyR1), which physically interacts with the dihydropyridine receptor Cav1.1 subunit in specific machinery formed with additional components. Exome sequencing has accelerated the discovery of many novel mutations in DICR machinery genes in various skeletal muscle diseases. However, functional validation is time-consuming because it must be performed in a skeletal muscle environment. Here, we describe a high-throughput platform for DICR that is reconstituted in non-muscle HEK293 cells. We demonstrate that Cav1.1 mutations implicated in malignant hyperthermia and myopathy exhibit divergent effects on DICR. We also tested several RyR1 inhibitors and DICR modulators. This high-throughput DICR platform will accelerate validation of mutations and drug discovery for skeletal muscle diseases.


Introduction
important for the diagnosis and treatment of these diseases. This has been performed using 48 myotubes from mice that lack DICR machinery components, for example knockout mice of 49 Cav1.1 14 , β1a 15 and Stac3 16 . While this approach has evaluated several disease-causing 50 mutations in the DICR components 16-20 , it has certain limitations. For example, it is expensive 51 and labor intensive to maintain the mouse lines. Differentiation of cells into myotubes is time 52 consuming and is affected by various factors, such as the cell line used and experimental 53 conditions. Furthermore, transduction of foreign genes into myotubes is often inefficient. 54 There are no specific treatments for most diseases related to mutations in the DICR 55 machinery. A major cause for the slow pace of novel drug development is the lack of an 56 appropriate drug screening platform for these diseases. We have recently established an efficient 57 high-throughput platform for screening RyR1-targeted drugs using endoplasmic reticulum (ER) 58 Ca 2+ measurements in HEK293 cells 21,22 and using this platform we developed Compound 1 59 (Cpd1), a novel RyR1 inhibitor for treatment of MH 23,24 . Therefore, functional screening using 60 HEK293 cells is a promising platform for drug discovery. 61 membrane using the patch-clamp technique induced a rapid Ca 2+ transient without influx of 66 extracellular Ca 2+ . This provides a basis for developing reconstituted DICR. 67 In this study, we established a high-throughput platform for DICR that was 68 reconstituted in HEK293 cells. Baculovirus infection of essential components greatly improved 69 transduction efficiency. Chemical depolarization by expression of an inward-rectified 70 potassium channel (Kir2.1) and high concentration potassium [K + ] solutions enabled the 71 simultaneous stimulation of many cells. Furthermore, a genetically-encoded ER Ca 2+ indicator 72 quantitatively measured DICR without influence of extracellular Ca 2+ influx. Using this 73 platform, we successfully evaluated disease-causing mutations in Cav1.1 and tested several 74 RyR1 inhibitors and DICR modulators. Our high-throughput platform will accelerate the 75 validation of mutations and drug discovery for skeletal muscle diseases. 76
We next determined whether the reconstituted platform reproduces DICR in skeletal 129 muscle. Initially, we examined the role of the essential components. Removal of Cav1.1 resulted 130 in complete loss of high [K + ]-induced Ca 2+ release (Fig. 2a, b). Ca 2+ release was also lost in 131 cells lacking β1a or Stac3, and in cells expressing RyR2 instead of RyR1 (Fig. 2b). Removal 132 of JP2 significantly reduced the fluorescence change by 61 mM K + with a large rightward shift 133 in K + dependence (Fig. 2a, b). These results are consistent with previous findings of mice 134 lacking each component 9,14-16,34 . 135 Skeletal muscle Cav1.1 forms a complex with three auxiliary subunits, β1a, α2-δ and γ1 35 . 136 β1a is essential for DICR, while α2-δ and γ1 may play modulatory roles 36 . Additional expression 137 of α2-δ had no effect on high [K + ]-induced Ca 2+ release (Fig. 2c, d). Expression of γ1 did not 138 significantly affect high [K + ]-induced Ca 2+ release at 20 mM or lower [K + ] but it reduced the 139 fluorescence change at higher K + concentrations (Fig. 2c, d). In γ1 cells, R-CEPIA1er 140 fluorescence decreased but increased again with time, suggesting inactivation of the Ca 2+ 141 release (Supplementary Fig. 5). These findings are consistent with previous reports of the 142 effects of α2-δ 37 and γ1 38 on DICR. Therefore, our reconstituted platform can reproduce DICR 143 in skeletal muscle and is valid for evaluating disease-causing mutations or drugs. 144 We also tested the importance of Kir2.1. Removal of Kir2.1 completely abolished high 145 [K + ]-induced Ca 2+ release with substantial reduction in the R-CEPIA1er fluorescence intensity 146 ( Fig. 3a-c). We quantified resting ER [Ca 2+ ] by determining the maximum fluorescence 147 intensity of R-CEPIA1er using ionomycin/Ca 2+ (Supplementary Fig. 6). The resting ER [Ca 2+ ] 148 was severely reduced in cells without Kir2.1 (Fig. 3d). The resting membrane potential of 5 149

BV cells was polarized in normal Krebs solution and became depolarized by high [K + ] solution 150
containing 50 mM K + in the presence of ~145 mM intracellular K + (Fig. 3e) (Fig. 3d). Therefore, Kir2.1 is 155 indispensable for our reconstituted DICR platform. 156 Evaluation of disease-causing mutations in Cav1.1 157 Using the high-throughput DICR platform, we evaluated mutations in Cav1.1 which for R1086H (Fig. 4b) and that R1086H exhibited a smaller fluorescence change in response to 161 high [K + ] (Fig. 4c, d) compared with the wild type (WT) or the other mutants. This was partly 162 caused by low ER [Ca 2+ ], which reduces maximum fluorescence changes. In addition, R1086H 163 showed an enhanced K + -dependence, i.e., lower EC50 value for K + (Fig. 4c, e). In contrast, no 164 or only minor changes in ER Ca 2+ and K + dependence were observed for R174W and T1354S 165 cells (Fig. 4b- Fig. 7b). R1086H cells exhibited a clear 172 greater sensitivity to K + than WT cells (Supplementary Fig. 7c-e), supporting our hypothesis 173 of R1086H hyperactivation. 174 Previous functional characterization of these Cav1.1 mutants consistently showed 175 enhanced caffeine sensitivity, i.e., acceleration of Ca 2+ -induced Ca 2+ release (CICR) 17-19 . We 176 tested caffeine dependence of these mutants in our platform. Caffeine released Ca 2+ from WT 177 cells in a dose-dependent manner (Fig. 4f). R1086H cells also exhibited a caffeine dependence 178 but with reduced fluorescence change by 61 mM K + (Fig. 4g) and a significantly smaller EC50 179 ( Fig. 4h), indicating an enhanced caffeine sensitivity. In contrast, caffeine dependence was not 180 changed in R174W or T1354S cells ( Fig. 4f-h). 181 We next examined the effect of myopathy-related mutations in Cav1.1 on DICR 182 activity. We chose four mutations, E100K, F275L, P742Q, and L1367V 42 , none of which have 183 been functionally characterized (Fig. 5a). ER [Ca 2+ ] was not changed by any of these mutations 184 (Fig. 5b). DICR activity was completely lost in the F275L mutant (Fig. 5c, d) and there was a 185 substantial rightward shift in K + dependence for P742Q (Fig. 5c); The EC50 value was threefold 186 higher in P742Q (33.2 ± 1.0 mM) compared with WT (10.3 ± 0.9 mM) (Fig. 5e). The other two 187 mutations (E100K and L1367V) had no significant effects on K + dependence ( Fig. 5c-e). We 188 also tested the effects of these mutations on caffeine-induced Ca 2+ release. No substantial 189 changes were observed with any of the mutations (Fig. 5f-h).
The high-throughput DICR platform is expected to be useful for screening drugs for 192 muscle diseases. To test this possibility, we examined the effects of known DICR modulators. 193 We initially tested three RyR1 inhibitors (Fig. 6a). Dantrolene is a well-known RyR1 inhibitor 194 that is clinically used for MH 43 . Cpd1 is a novel potent RyR1-selective inhibitor that we recently 195 developed 23,24 . Procaine inhibits CICR but not DICR in frog skeletal muscle 44,45 . Dantrolene 196 (10 µM) and Cpd1 (3 µM) caused rightward shifts in K + dependence compared with the control, 197 indicating that they inhibit DICR (Fig. 6b, c). This is consistent with the suppression of twitch 198 and tetanic tension in isolated muscles and of reduced in vivo muscle weakness by 199 dantrolene 46,47 and Cpd1 24 . Procaine (5 mM), in contrast, did not inhibit DICR (Fig. 6b, c). All 200 three compounds shifted the caffeine dependence rightward, indicating inhibition of CICR ( Fig.  201 6d, e). These results indicate that dantrolene and Cpd1 inhibit both DICR and CICR, whereas 202 procaine is a CICR-selective inhibitor. 203 Lyotropic anions, such as perchlorate and thiocyanate, potentiate E-C coupling in 204 skeletal muscle 48-52 . Perchlorate (10 mM) and thiocyanate (10 mM) significantly shifted the K + 205 dependence leftward (Fig. 7a, b). In contrast, they did not affect caffeine dependence (Fig. 7c,  206   d). These results indicate that lyotropic anions potentiate DICR but not CICR. This is consistent 207 with previous findings showing that the potentiating effects of lyotropic anions are primarily 208 caused by shifting voltage dependence of the charge movement toward more negative 209 potentials 48-52 . Taken together, our high-throughput platform reproduces the effects of known 210 DICR modulators, indicating that it can be used to screen for and test drugs that affect DICR. 211

Discussion 212
In this study, we established a high-through platform for DICR that was reconstituted in 213 HEK293 cells expressing RyR1 and R-CEPIA1er. We made three key improvements to the 214 original method by Perni et al. 25 . First, the essential components were transduced using VSV-215 G pseudotyped baculovirus, which can effectively infect a wide variety of mammalian cells 216 without toxicity 28 . This greatly increased the transduction efficiency; almost all cells expressed 217 the essential components (Supplementary Fig. 1). Second, depolarization of the plasma 218 membrane was induced by high [K + ] solution. This enabled us to simultaneously stimulate 219 many cells (Fig. 1, Supplementary Fig. 2 and movie S1). For this purpose, we expressed 220 Kir2.1, an inward-rectifying potassium channel, which effectively hyperpolarized the 221 membrane potential 32 (Fig. 3). Third, ER [Ca 2+ ], instead of cytoplasmic [Ca 2+ ], was measured 222 essential components (Fig. 1a). In contrast substantial increases in cytoplasmic [Ca 2+ ] were 225 observed (Supplementary Fig. 4a). These improvements allow quantitative measurements of 226 DICR using a microplate reader. 227 In our reconstitution, five components, RyR1, Cav1.1, β1a, JP2 and Stac3, are essential for 228 DICR; removal of each component abolished or reduced the DICR activity ( Fig. 2a, b). This is 229 consistent with previous findings from mice lacking each component 9,14-16,34 . In addition, the 230 effects of α2-δ and γ1 auxiliary subunits on DICR activity corresponded with findings of 231 previous reports 37,38 (Fig. 2c, d). Therefore, our reconstituted platform successfully reproduces 232 DICR in skeletal muscle. We found that a substantial DICR still occurred without JP2 (Fig. 2a,  233 b). This is in contrast to reports by Perni et al. 25,53 , in which no voltage-gated Ca 2+ release was 234 observed in reconstituted cells without junctophilins. A possible reason for this is differences 235 in expression methods. Our VSV-G baculovirus is more effective at transduction than standard 236 lipofection; therefore, sufficient proteins were expressed that might spontaneously interact with 237 each other without junctophilins. 238 We quantitatively evaluated the effect of disease-casing mutations or drugs using the EC50 239 value for [K + ]. In our experimental conditions, the EC50 value for K + was about 10 mM ( Fig.  240 1d). The value is lower than that for skeletal muscle 54 or myotubes 55 in high [K + ]-induced Ca 2+ 241 transients. This may be partly explained by measurement differences. We calculated the EC50 242 from the fluorescence intensity for 125-150 seconds after stimulation (Fig. 1c). This reduced 243 the EC50 by ~7 mM compared with the value obtained just after stimulation (Supplementary 244  (Fig. 6, 7), strongly indicating that the EC50 value for K + is a valid parameter for 246 quantitative evaluation. 247 Formation of DICR machinery can cause peripheral localization of the components, e.g., 248 RyR1 and Cav1.1, in the reconstituted cells 25,53 . In our reconstituted cells, however, no such 249 peripheral-specific localization was observed; the components were localized throughout cells 250 except for in nuclei (Supplementary Fig. 1). This may be because of high levels of the 251 components expressed by baculovirus, which not only accelerates formation of the DICR 252 machinery (as discussed above), but also increases the number of uncoupled components. 253 Mutations in Cav1.1 are implicated in various muscle diseases, including MH and 254 consistent with previous findings using Cav1.1-deficient (dysgenic) myotubes 17-19 . All three 257 mutations show increased sensitivity to caffeine-induced Ca 2+ release compared with WT, 258 indicating enhanced CICR activity 17-19 . However, we showed that caffeine sensitivity was 259 increased only for R1086H ( Fig. 4f-h). A possible reason for this difference is the cells used. There are no specific treatments for most DICR-related diseases; therefore, there is an 277 urgent need to develop novel drugs for these diseases 11-13 . We here evaluated DICR modulators. 278 Among three known RyR1 inhibitors, dantrolene and Cpd1 suppressed DICR (Fig. 6). This is 279 consistent with the suppression of twitch and tetanic tensions of isolated muscles and of reduced 280 in vivo muscle weakness by dantrolene 46,47 and Cpd1 24 . Interestingly, procaine did not suppress 281 DICR (Fig. 6). Procaine inhibits CICR but not DICR in frog skeletal muscle 44 and reconstituted 282 frog α-RyR and β-RyR in RyR1-deficient myotubes 45 . Our data show similar procaine action in 283 a mammalian system. We also demonstrated that lyotropic anions (perchlorate and thiocyanate) 284 accelerate DICR (Fig. 7). This reproduces previous findings using skeletal muscle fibers [48][49][50][51][52] . 285 Interestingly, lyotropic anions did not affect caffeine dependence, suggesting no effect on CICR 286 (Fig. 7). Therefore, our high-throughput platform will also be useful for drug testing and 287 In summary, we established a high-throughput platform for skeletal muscle DICR. This 289 platform will be useful for both evaluation of disease-causing mutations and the development 290 of novel drugs for DICR-related diseases. 10% fetal calf serum, 2 mM L-glutamine, 15 µg/ml blasticidin, and 100 µg/ml hygromycin. 303

Western blotting 323
RyR1-HEK293 cells with or without baculovirus infection for the essential components were 324 lysed with Pro-Prep protein extraction solution (iNtRON Biotechnology). The extracted 325 proteins were separated on 3-15% polyacrylamide gels and transferred to PVDF membranes. 326 Membranes were probed with the primary antibodies described above, followed by HRP-327 Data are presented as means ± SD. Unpaired Student's t test was used for comparisons between 375 two groups. One-way analysis of variance (ANOVA), followed by Dunnett's test, was 376 performed to compare three or more groups. Two-tailed tests were used for all analyses. 377 Statistical analysis was performed using Prism v9 (GraphPad Software, Inc., La Jolla, USA).

Data availability 380
All data generated or analyzed in this study are available within the article and its 381 Supplementary Information. All raw data supporting the findings from this study are available 382 from the corresponding author upon request. Source data are provided with this paper.     (F/F0) in cells infected with 5 BV (black), 5 BV plus α2-δ (+α2-δ, purple) and 5 BV plus γ1 (+γ1, 597 orange). Data are means ± SD (n = 6). d Fluorescence change by 61 mM K + in 5 BV, +α2-δ, 598 and +γ1 cells. Data are means ± SD (n = 6) and were analyzed by one-way ANOVA with 599 Dunnett's multiple comparisons test. 600 Fluorescence change by 20 mM caffeine (g) and EC50 values for caffeine (h). Data are means 622 ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. 623 BV cells (red). F/F 0 was obtained by normalizing the averaged fluorescence of the last 25 seconds (F) to that of the first 25 seconds (F 0 ). Data are means ± SD (n = 6). e Fluorescence change by 61 mM K + or 10 mM caffeine. Fluorescence change is expressed as 1-F/F 0 (see panel d). Data are means ± SD (n = 6) and were analyzed by two-way ANOVA with Bonferroni's multiple comparisons test. Note that 5 BV cells show large fluorescence change with high [K + ] which is nearly 80% of that with caffeine.   Data are means ± SD (n = 9 for 5 BV, 6 for -Cav1.1 and -JP2). b Fluorescence change by 61 mM K + in the BVinfected cells lacking one of essential components (-Cav1.1, -β1a, -Stac3, or -JP2) or with RyR2 cells instead of RyR1. Data are means ± SD (n = 6) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. c [K + ] dependence of R-CEPIA1er fluorescence (F/F 0 ) in cells infected with 5 BV (black), 5 BV plus α 2 -δ (+α 2 -δ, purple) and 5 BV plus γ1 (+γ1, orange). Data are means ± SD (n = 6). d Fluorescence change by 61 mM K + in 5 BV, +α 2 -δ, and +γ1 cells. Data are means ± SD (n = 6) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test.   Data are means ± SD (n = 9). d, e Fluorescence change by 61 mM K + (d) and EC 50 values for K + (e). Data are means ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. f Caffeine dependence of v(F/F 0 ) in WT (black), R174W (red), R1086H (blue) and T1354S (green) Cav1.1 cells. Data are means ± SD (n = 9). g, h Fluorescence change by 20 mM caffeine (g) and EC 50 values for caffeine (h). Data are means ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test.   Resting ER [Ca 2+ ] in cells expressing WT and mutant Cav1.1. Data are means ± SD (n = 12) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. c [K + ] dependence of R-CEPIA1er fluorescence (F/F 0 ) in WT (black), E100K (red), F275L (blue), P742Q (green) and L1367V (magenta) Cav1.1 cells. Data are means ± SD (n = 9). d, e Fluorescence change by 61 mM K + (d) and EC 50 values for K + (e). Data are means ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. f Caffeine dependence of R-CEPIA1er fluorescence (F/F 0 ) in WT (black), E100K (red), F275L (blue), P742Q (green) and L1367V (magenta) Cav1.1 cells. Data are means ± SD (n = 9). g, h Fluorescence change by 20 mM caffeine (g) and EC 50 values for caffeine (h). Data are means ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test.   (Control, black) or with dantrolene (10 μM, red), Cpd1 (3 μM, blue) or procaine (5 mM, green). Data are means ± SD (n = 9). c EC 50 values for K + . Data are means ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. d Caffeine dependence of R-CEPIA1er fluorescence (F/F 0 ) in cells without (Control, black) or with dantrolene (10 μM, red), Cpd1 (3 μM, blue) or procaine (5 mM, green). Data are means ± SD (n = 9). e EC 50 values for caffeine. Data are means ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test.  Data are means ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. c Caffeine dependence of R-CEPIA1er fluorescence (F/F 0 ) in cells in the absence (Control, black) or presence of perchlorate (10 mM, red) or thiocyanate (10 mM, blue). Data are means ± SD (n = 9). d EC 50 values for caffeine. Data are means ± SD (n = 9) and were analyzed by one-way ANOVA with Dunnett's multiple comparisons test.