1. IntroductionThe sarco/endoplasmic reticulum (SR) Ca-ATPase (SERCA) plays a dominant role in clearing cytosolic Ca2+ in most cell types. Through its enzymatic cycling, SERCA transports Ca2+ across the ER/SR membrane (uptake activity) by Ca2+-dependent hydrolysis of ATP (Ca-ATPase activity) [1,2]. In muscle, this process maintains the Ca2+ gradient across the SR membrane, lowering the cytosolic Ca2+ concentration, as required for muscle relaxation, and maintaining the SR Ca2+ content necessary for Ca2+ release through the ryanodine receptor (RyR), as required for muscle contraction. Ca2+ also controls other key cellular functions, in both muscle and non-muscle cells, such as signaling, apoptosis, and metabolism [3]. Deficient control of Ca2+, often due to insufficient SERCA activity, is implicated in numerous disorders, including heart failure [4,5,6], cancer [7], diabetes [8], and muscular dystrophy [9].Diminished activity of the cardiac SERCA2a isoform in heart failure (HF), often accompanied by increased phospholamban (PLB)/SERCA ratio [10,11,12] and/or decreased PLB phosphorylation [12,13,14], results in slower and less complete relaxation after each beat, elevated diastolic Ca2+, and degraded cardiac function. Partial restoration of SERCA activity has been achieved by gene therapy, significantly improving HF symptoms in animal models [15,16], and SERCA2a gene therapy treatment for human HF patients advanced through phase IIb clinical trials [17], where it showed neutral results due to pre-existing rAAV antibodies [18]. Despite this surprising setback, SERCA activation remains a valid and attractive target for HF therapy.Dysregulation of the skeletal muscle isoform, SERCA1a, is also implicated in the pathology of numerous severe myopathies such as muscular dystrophy and sarcopenia [19,20,21]. In most of these cases, pathogenesis is fueled by excess cytoplasmic Ca2+ (in the absence of neural excitation), due to leak via RyR and/or insufficient SERCA1a activity. In many cases, the result is reduced excitation-contraction coupling due to decreased Ca2+ inside the SR (Ca2+ load). Recent studies indicate that many skeletal myopathies can be remedied by increasing SERCA1a expression [22,23] or decreasing sarcolipin (SLN) expression [24], both of which result in SERCA1a activation.Our goal is to accelerate therapeutic discovery for multiple myopathies that are characterized by elevated resting intracellular Ca2+ levels ([Ca2+]), ER stress, mitochondrial ROS production, excessive levels of phosphorylation, and Ca2+-promoted proteolysis. These defects are linked to an intracellular vicious cycle that calls for small-molecule therapeutics, which are targeted to restore normal physiology in the affected muscles [22,25,26,27,28,29]. We seek small molecules that increase SERCA activity in cardiac muscle, but also expect to find compounds relevant to skeletal muscle diseases. We have previously developed several high-throughput screening (HTS) strategies that were focused on fluorescence lifetime measurements of genetically encoded fluorescent biosensors, which were used in primary screens of chemical libraries [30,31,32,33]. In contrast, in the present study, we have developed a primary HTS assay directly targeting SERCA’s enzymatic activity. Our primary screen specifically assesses the skeletal SERCA1a isoform, due to its availability in nearly pure form from skeletal muscle tissue, but high homology of sequence, structure, and function, validated by decades of research, suggest a high likelihood of cross reactivity between isoforms. Indeed, most compounds previously found to activate SERCA1a also activate SERCA2a, with some isoform-specific variation in efficacy and potency.

Importantly, we have identified distinct classifications of compounds based on their chemical and physical properties and their functional effects on ATPase and Ca2+ uptake activities of SERCA. Such classifications will be essential regardless of the screening approach, and could facilitate mining of existing screening data, in lieu of de novo primary HTS efforts. Future studies will seek to (1) identify additional activating compounds from this ATPase-based screen or other fluorescent biosensor screens, (2) develop analogues optimized via medicinal chemistry, and (3) explore their physiological effects on muscle tissue and whole animals, to advance toward therapeutics.

4. Discussion 4.1. Robustness of HTS AssayOur overall goal is to identify SERCA-activating compounds for development to treat muscle diseases. Therefore, we adapted and miniaturized our well-established NADH-coupled SERCA activity assay [41] to screen a 46,000-compound library. This target-directed screening approach facilitated a high success rate in identifying SERCA-activating compounds. The use of a direct assay for SERCA enzymatic activity in the primary screen was slower (lower throughput) than our previous protein structure-based screening method [32], but it resulted in a pool of compounds that more reliably produced desired effects on SERCA activity. 4.2. Compound Classifications: ATPase vs. Ca Uptake ActivitiesUnder ideal conditions, SERCA transports two Ca2+ per molecule of ATP hydrolyzed [2,42,43,44]. Some experimental studies have directly observed this 2:1 coupling ratio [45], whereas others have also observed significantly lower enzymatic efficiencies (≤1:1 Ca2+:ATP) [46]. Nevertheless, changes in SERCA activity have typically been presumed to have equivalent, proportional effects on both the Ca2+ uptake and ATPase activities. The major known exception is regulation by sarcolipin (SLN), which partially uncouples these activities, allowing hydrolysis of ATP in the absence of Ca2+ transport, a function found to be critical for thermogenesis [47]. Through our analysis, we identified several classifications of compounds that differentially affect the Ca2+ uptake and ATPase functions of SERCA (detailed below). The critical variations were observed in the Ca2+ transport component, as all hit compounds strongly increased ATPase activities; this is presumably a selection bias due to the initial primary screen, which measured ATPase activity. Furthermore, all compounds have significantly larger effects on ATPase activity compared to Ca2+ uptake at the highest compound concentrations (>10 µM), but have more comparable and/or variable effects at lower concentrations (Table 1). Several compounds induced a 60% or more (as high as 85%) increase in ATPase activity, and all compounds increased ATPase function more than Ca2+ uptake, by a factor ranging from 0.5 to 8.

The ability to drive ATPase function much higher could be a result of an intrinsic thermodynamic threshold of SERCA enzymatic activity and/or an artifact of the assay systems. The rate limiting steps in SERCA’s transport process are a sequence of discrete structural transitions allowing for binding of cytosolic calcium and release into the ER/SR lumen, which physically constrains the maximal speed of Ca uptake. Breaking of a single chemical bond for ATP hydrolysis does not have the same thermodynamic limitations and thus can be driven faster than Ca uptake. Therefore, high compound concentrations can push rates of ATPase activity beyond the physical limits of Ca2+ transport. On the other hand, these inefficiencies, while important and interesting, appear to only occur at higher than therapeutically relevant concentrations (ideally < 1 mM). Medicinal chemistry may also be able to identify more efficient derivatives. Further physiological and metabolic testing is needed to determine if such inefficiencies extend to the cellular and organ level.

4.2.1. Linked Activating CompoundsWe identified several compounds where the initial increases in the CRCs for ATPase and Ca2+ transport activities appear closely matched at lower compound concentrations, which we have termed “Linked Activating Compounds” (Figure 4). The ATPase activity response for these compounds appears to rise in unison with Ca2+ uptake, up until the Ca2+ uptake effect begins to saturate, and ATPase continues to increase. This class of compounds increases SERCA activity while maintaining a consistent coupling ratio, at least at lower concentrations. 4.2.2. Improved Coupling CompoundsWe identified another class of compounds where the CRC response in Ca2+ uptake precedes that observed with ATPase function, which we have termed “Improved coupling compounds” (Figure 5). At lower concentrations, these compounds increase SERCA Ca2+ uptake through an increase in coupling efficiency rather than just driving overall kinetics of the enzyme. As the concentrations increase with these compounds, the ATPase effect catches and surpasses the Ca2+ uptake effect. This additional capacity to enhance the coupling efficiency of SERCA suggests that our baseline system and assay conditions are operating below the theoretical 2:1 Ca2+ to ATP ratio. 4.2.3. Uncoupling CompoundsWe also identified several compounds classified as uncouplers or uncoupling compounds (Figure 6). Similar to the protein inhibitor SLN, these compounds substantially increase SERCA ATPase function at the expense of Ca2+ uptake. The primary rate-limiting components of the SERCA reaction cycle are Ca2+ binding and subsequent transport. Thus, uncoupling allows for less constrained ATPase hydrolysis, resulting in higher observed activity rates. Interestingly, all compounds were contained in cluster 5. Compounds 15 and 17 have little to no Ca2+ uptake effect, while compound 16 has a sizeable inhibitor effect. 4.2.4. Tissue and Isoform SpecificityDual analysis of our Hit compounds in cardiac and skeletal SR further reveals unique targeting specificity of SERCA modulation. The primary difference between the tissue preparations is the isoform present, with SERCA2a in cardiac SR and SERCA1a in skeletal SR. These sample preparations also have different SERCA regulatory proteins, with cardiac SR containing PLB and skeletal SR containing SLN and myoregulin [48].While the relative increases in ATPase activity Vmax were generally consistent between the two sample preparations, the compound effects on Ca2+ transport tended to be higher in cardiac SR compared to skeletal (Table 1). In fact, nine compounds had a >50% increase in Ca2+ uptake Vmax effect in cardiac SR compared to skeletal. In addition, compound 16 (DS33804556) was observed to be an inhibitor of Ca2+ uptake in both cardiac and skeletal SR. These substantial similarities are likely due to the high degree of homology between SERCA1a and SERCA2a.The key exception was compound 12 (DS60405307), which shows a strong uncoupling effect in cardiac SR (decreasing Ca2+ uptake activity), with significant activation in Ca2+ uptake activity in skeletal SR (Figure 7). In addition, the two compounds in cluster 4 (compounds 13 and 14) were the only ones with a greater than 2-fold increase in Ca2+ uptake in cardiac SR vs. skeletal. These discrepant effects are probably due to a critical structural difference between SERCA isoforms, and/or the presence/absence of a tissue-specific SERCA regulatory partner. 4.3. Study Limitations

Our initial primary screen, along with all our follow-up ATPase and Ca2+ uptake assays, were performed only at a saturating Ca2+ concentration, pCa 5.4 (4 µM). While this facilitated data acquisition, we acknowledge that these compounds may differentially affect SERCA at lower Ca2+ concentrations. Future screens at different Ca2+ are expected to reveal additional unique classifications of SERCA modulating compounds.

In addition, since our primary screen and Hit selection process utilized ATPase activity in skeletal SR, our Hit compounds are biased toward ATPase activators; this might partially explain the higher maximal effects of compounds on ATPase versus Ca2+ uptake. In other words, we may have artificially selected compounds that have some uncoupling effect, thus facilitating the enhanced ATPase function.

4.4. Next Steps

While this study represents a substantial screening effort, especially from an academic laboratory, a 46,000-compound screen such as ours only represents an industrial pilot screen, with true HTS efforts sometimes utilizing libraries more than 100 times larger. Thus, we believe it would be premature to identify any lead compounds for a full battery of testing. The major takeaways from this study are not the compounds themselves but the strategic advances we have made toward an effective large-scale screening effort. Our primary goal was to optimize selection criteria for in-depth mining of existing screens and/or performance of a larger screen.

Moving forward, assessment of compound effects at lower [Ca2+] (e.g., pCa 6.2; i.e., 630 nM) will add another biochemical classification parameter, facilitating better selection criteria for advancement of compounds for further testing. Full Ca2+ response curves as is common in SERCA activity assays would also be informative in understanding Ca2+ affinity and cooperativity. However, such in-depth analysis will be better suited for the structure-activity response (SAR) phase of the drug discovery process, done in conjunction with medicinal chemistry in order to tease out more nuanced variations between similar compounds. We will also expand our pool of Hit compounds to be tested, from this screen or others, which will be prioritized based on physico-chemical criteria we have identified. Combinations of compounds from different classifications and scaffolds may yield an enhanced and synergistic response. Importantly, we will test the physiological (in situ, in vivo) effects of compounds to determine whether the in vitro activating effects on SERCA produce enhanced muscle function. Furthermore, we speculate that different compound classifications will differentially affect muscle function. Ideally, these avenues will be explored in parallel, as we discover which compound classifications are better suited as therapeutic leads. These additional experimental endeavors are critical before any lead selection. Above all, we are developing a systematic process for the selection of SERCA-activating compounds that could be advanced towards therapeutics of muscle disorders.