Image-Driven Modeling of Nanoscopic Cardiac Function: Where Have We Come From, and Where Are We Going?

Figures

• 

  FIGURE 1. Progress in unraveling cardiac EC coupling based on experimental imaging and mathematical modeling. Continual advancement of imaging technologies (left column) has provided the tools for critical experimental insight during the last decades. Complementary advancements in mathematical modeling have contributed to shifting the paradigm in our understanding of EC coupling (right column). In recent years, experimental and modeling developments have allowed a near merging of the spatial and temporal scales that each can be used to investigate. EM, electron microscopy; TEM, transmission EM; CICR, calcium-induced calcium release; SR, sarcoplasmic reticulum.

• 

  FIGURE 2. Evolving understanding of dyadic structure from electron microscopy (EM) studies. Visualization of dyads was first enabled by EM, as early as the 1960s. Even with 2D applications of this technique (A), dyadic junctions between the jSR and the cell surface [(i), chick myocardium, (Franzini-Armstrong et al., 1998)] or t-tubules [(ii), rat ventricular myocyte, (Novotová et al., 2020)] were readily apparent, and highlighted the tight geometry of the dyadic cleft. Electron dense “feet” visible in the cleft (arrows) were attributed to the cytosolic portion of the RyR. With the advent of tomography techniques and enhanced contrast agents, it became possible to assess the 3D orientation of individual RyRs within dyads (B, scale bars = 30 nm, Asghari et al., 2020). XY (i) and YZ (ii) orthagonal views are presented for a representative dyad (single arrow = jSR, double arrow = t-tubule). Indicated planes are positioned on a single RyR (red line = XY, blue line = XZ, and green line = YZ plane). (iii) A magnified view of the boxed region in ii illustrates selected sections (4 of 28 illustrated) taken across the dyadic cleft. Each of these four sections is presented in (iv), allowing…

• 

  FIGURE 3. Advances in super-resolution imaging provide novel insight into RyR organization. Although previous work had suggested that RyRs are present in large clusters, dSTORM super-resolution imaging of RyRs has shown that the channels are rather present in multiple small, neighboring clusters. On the cell surface, these arrangements are present in double rows, on either side of z-lines [(A, left panel), (Baddeley et al., 2009)]. The increased resolution of RyR imaging by dSTORM (green) is noted in comparison with diffraction-limited imaging (red). Maximal occupancy of imaged clusters was calculated based on an assumed grid-link arrangement of channels (A, right panel). Quantitative assessment of more complex RyR arrangements within the cell interior requires high-resolution in 3D. Using 3D dSTORM, internal Ca2+ release units (CRUs) containing multiple RyR clusters with diverse configurations were observed [(B), enlargement of indicated region at right; (Shen et al., 2019)]. To gain insight into the orientation of RyRs within dyads, correlative 3D dSTORM imaging of RyRs was combined with confocal t-tubule imaging, and the channels were localized at the interfaces of the two signals [(C),…

• 

  FIGURE 4. Image-derived approaches to geometric modeling of CRU and whole-cell cardiac ECC. (A) Image-driven models of single dyad (and CRU) structures have been implemented beginning with both super resolution light microscopy (dSTORM), (left) (Walker et al., 2014; Kolstad et al., 2018), and tilt-section tomography from high voltage TEM imaging (right) (Hake et al., 2012; Edwards et al., 2018). To date, no correlative light/EM datasets visualizing both protein and membrane localization have been employed for this purpose, although such approaches would be highly desirable. (B) High-volume EM methods, particularly SBF-SEM, have recently been used as a basis for constructing image-driven geometries for much larger subsections of a cardiac myocyte (Colman et al., 2017; Colman, 2019). Simulations in these geometries can be used to interrogate the interaction of local Ca2+ release with uncertain (and difficult-to-measure) physiological properties, such as intra-SR Ca2+ diffusion. At right, spatially distributed intra-SR Ca2+ concentration is shown at various times during the evoked macroscopic SR Ca2+ release. Models of this type will likely provide the basis for bridging between the…

• 

  FIGURE 5. Foundational models for idealized geometric modeling of cardiac ECC. (A) Two important early dyadic structure models involving stochastic RyR dynamics were provided by Sobie’s sticky cluster (Sobie et al., 2002) (A), and the 4-couplon model of Greenstein and Winslow (2002) (B). These two average dyadic models were each then extended to simulate whole-cell ECC by instantiating the 12,500–20,000 stochastically operating CRUs in the average ventricular myocyte Williams et al. (2011) and Wescott et al. (2016) (C). Implemented models that solved stochastic RyR dynamics across these many identical sticky-cluster-derived CRUs, whereas Hinch et al. (2004) and Hinch (2004) performed a clever timescale decomposition to markedly reduce the scale of the problem (D). Both families of model have been successful in replicating and/or interrogating a number of important macroscopic properties of ECC. However, neither is readily capable of assessing the impact of local RyR dispersion, and they do not implement known heterogeneities in CRU structure and function (see Section 2). Copyright permission to reproduce the figures.