Cell Biophysics–Physiological Contexts, from Organism to Cell, In Vivo to In Silico Models: One Collaboratory’s Perspective

Abstract

Here we present a retrospective, integrative review of the approaches and discoveries of our “collaboratory”, a meta-laboratory comprising cross-disciplinary collaborations across laboratories at fourteen different universities and clinics in seven different countries with shared lead investigators. By tying together insights from four decades of research and discovery, applied across cell types, as well as different tissues, organ systems, and organisms, we have aimed to elucidate the interplay between organisms’ movement and the physiology of their tissues, organs, and organ systems’ resident cells. We highlight the potential of increasing imaging and computing power, as well as machine learning/artificial intelligence approaches, to delineate the Laws of Biology. Codifying these laws will provide a foundation for the future, to promote not only the discovery of underpinning mechanisms but also the sustainability of our natural resources, from our brains to our bones, which serve as veritable “hard drives”, physically rendering a lifetime of cellular experiences and millennia of evolution.

1. Tissue Genesis as a Biological Scaling Algorithm

Remarkable spatial and temporal biological scaling algorithms are present throughout the life cycle of cells and the organisms they inhabit (typically, 80 days–400 years, depending on vertebrate species), as well as in the evolution of those organisms over millennia [1]. Cells do the work of postnatal healing, which itself recapitulates prenatal development; both cellular processes recapitulate evolutionary biological algorithms to create, pattern, and scale up tissue templates [1,2,3,4,5]. In a skeletal context, in utero bone formation and postnatal bone healing occur either indirectly via endochondral ossification or directly via intramembranous ossification. In endochondral ossification, a cartilage template (Anlage) forms and mineralizes to form skeletal elements, i.e., “inside-out”. In contrast, intramembranous ossification directly creates new skeletal tissue either at the bone periphery, i.e., “outside-in”, or through remodeling (osteoclastic resorption followed by osteoblastic infilling of the resulting cutting cones, Figure 1A,B) [3].

While postnatal healing largely recapitulates skeletal development algorithms, capacity is diminished. For example, tendons cut in utero retain the capacity to heal before but not after birth [6]. In contrast, the bridging of large bony defects due to trauma can occur both via postnatal endochondral ossification, where a soft callus forms and mineralizes over time, or via direct intramembranous bone formation (Figure 1C) [2,3,4,5,7,8,9,10,11]. The mode and spatiotemporal patterns of bone healing and ossification underpin the emergence of tissue architecture, from the molecular to the cellular and tissue length scales, in health and disease [1,2,3,4,5,7,8,9,10,11].

2. Life Cycle of the Organism and Its Inhabitant Cells

2.1. Interdependence of Movement, Growth, Remodeling, and Life

Throughout life, movement ties directly to cell viability and tissue growth/remodeling and vice versa [4,11,13]. Lack of appropriate biophysical cues can result in devastating congenital defects, such as hypoplastic left heart syndrome, where half of the heart does not develop, due to a putative lack of mechanical input (flow) at the earliest stages of heart and circulatory system development [14]. The mechanoadaptation feedback loop comprises force balances and mechanotransduction linking organismal-scale events, such as movement, to cell-scale events that detect matrix damage, tissue building, and tissue remodeling (Figure 2A) [13]. As organisms age, this feedback loop becomes less efficient, as do the cellular builders and adaptors of tissue and the cellular modulators of metabolic function. Hence, the effects of aging and disease become apparent at all physiological length scales (Figure 2B) [1,2,4,8,13].

2.2. Epidemiology of Cell Tissue, Organ, and Human Population Health

Ultimately, we aim to predict health changes and disease emergence at the earliest possible timepoints, both to prevent disease onset and to ensure the best possible therapeutic outcomes. At the start of our collaboratory’s work, our goal was to tie organism-level physical loading and physiology to cell-scale events, but many methods still needed to be developed, and at the time, computational and imaging power were in their nascency. It would take decades of technological developments, in parallel with novel experimental and computational approaches, to significantly advance our goal of establishing a new cell epidemiology approach to understand human health and disease.

3. Movement-Induced Deformation of Tissues and Their Resident Cells Drives Growth and Adaptation

3.1. Pioneering Studies on the Cellular Strain of Mechanical Loading

In the 1970s, Lanyon and his protégé Goodship ^‡ led the field by implementing experimental mechanics approaches in in vivo human (single-subject), porcine, and ovine models. Their early pioneering studies were designed to measure, for the first time, loading-induced bone surface deformations in situ and to probe the effects of mechanical loading in the functional adaptation of bone [17,18,19]. A decade later, Lanyon and Rubin developed the so-called “functionally isolated turkey ulna model”, the first in a series of experimental models designed to elucidate loading “doses” that could be prescribed to prevent disuse osteopenia or to accrue bone in a directed way. A compelling aspect of the turkey ulna model data was that as few as four cycles per day of a compressive load imbuing circa 2000 microstrain could maintain bone (prevent loss of bone due to disuse), with increasing cycles up to 36 or more per day increasing bone mass over time [20,21,22]. These studies inspired a generation of scientists to seek out the cellular mechanisms of bone density maintenance and bone growth, as well as to define exercise protocols to maintain bone and prevent disuse osteopenia.

^‡ Professor Allen Goodship passed in April 2025.

3.2. Mechanical Loading-Enhanced Perfusion of Bone and Its Resident Cells

Indeed, movement induces both volume- (dilatational) and shape-changing (deviatoric) stresses across length scales, effectively pumping fluids through poroelastic tissues while also straining the tissues themselves [4,22]. At the start of our collaboratory’s studies, little was known about the fluid component of bone or the role of mechanical loading in enhancing fluid flow, thereby inducing mechanical cues and (convective) molecular transport to and from osteocytes. Piekarski and Munro proposed that physiological loading would likely increase molecular transport to and from osteocytes embedded in mineralized matrix, using a highly idealized two-dimensional osteon schematic to demonstrate their hypothesis [23]. Only through data integration from different experimental, physical, and computational models could a more comprehensive picture of fluid flow in porous bone be discovered.

Hence, inspired by the work of Lanyon and colleagues, as well as Piekarski and Munro, our collaboratory developed and tested different, intersecting methods to link organism-to-organ-to-tissue-scale movement, forces, and induced strains to those experienced at the cell scale, essentially determining how higher length-scale stresses and strains are “felt” by the cells themselves. These methods included in vivo, ex vivo, and in vitro experimental models and computational approaches. While we highlight some of our collaboratory’s studies here, the field has grown into a robust and rich area of research that goes beyond the scope of this paper.

3.2.1. “Top-Down” Experimental Mechanics Approaches—In Vivo to Ex Vivo and In Vitro

Top-down approaches start with the entire system (“top” level, maximal complexity), e.g., the organism, working “down” to separate subsystems to address specific components of the whole. This approach allows for stepwise elucidation with the goal of deciphering complexity. Often, this approach offers a “first pass” to measure unknown parameters and to prioritize hypotheses in otherwise highly complex systems [24].

We used “top-down” experimental mechanics approaches, guided by Lanyon and collaborators’ fundamental studies [17,18,19,20,21], developing a series of ovine models to test how organ-level mechanical loading manifested at smaller length scales. These studies focused on the interdependent effects of mechanical loading-induced deformations at organ-tissue cellular length scales and mechanical loading-enhanced bone perfusion. Although quadrupedal (compared to bipedal humans), sheep (ovine model) share a similar anatomy and physiology to that of humans across organ systems. Several interlinked models connect mechanical loading effects on tissue deformation to load-induced fluid flow through tissues. First, to understand baseline loading conditions during normal gait, sheep were trained to walk on a treadmill while strains were measured on the anterior surface of the mid-diaphyseal metacarpus of the forelimb (Figure 3A,D) [25]. These measurements informed the development of two related ex vivo perfusion models. In these models, in vivo loading conditions could be replicated while controlling forelimb perfusion using microsurgical protocols akin to ex vivo large animal organ/limb transplant models [26,27].

The first model (Figure 3B) studied loading–perfusion coupling in a fairly isolated manner. We applied compressive loads to the explanted sheep metacarpus via Schanz screws inserted proximal and distal to the mid-diaphysis while perfusing the forelimb with Ringer’s solution and procion red, a fluorescent, small-molecular-weight tracer [26]. The model’s load application mode mimicked that used in Lanyon and Rubin’s turkey ulna model [20,21], with magnitudes tuned to mimic in vivo walking gaits. We showed, for the first time to our knowledge, that cyclic compressive loading during perfusion enhances molecular transport from the blood supply to osteocytes [26].

The second ex vivo compression model was designed to study mechanotransduction and transport through the synovial joint (radius–metacarpus, Figure 3C). We again applied load-mimicking physiological conditions, this time via the radius, with the hoof fixed in place. Perfusion was controlled in a manner analogous to a previous preparation [26]. We demonstrated that endoprosthetic implantation creates a cell-scale insurmountable gap in transport to the synovial joint and that mechanical loading increases transport beyond baseline diffusion [27].

Finally, we implemented a novel in vivo imaging model designed to image cell-scale remodeling events in situ (Figure 4). This model used a custom, hollow-threaded cylindrical titanium implant designed to offload the bone within while maintaining the blood supply. The implant design allowed for imaging from above using a long-working-distance lens and a confocal microscope. Osteoclastic resorption was observed over the course of four weeks, in areas predicted using finite element modeling to correspond to areas of bone offloaded by the surrounding titanium implant ring. Although the Institutional Review Board approved the use of contralateral, unoperated controls as appropriate, journal reviewers felt that a low-stiffness control implant (such as silicon or compliant plastic) was unnecessary. While pioneering, a lack of funding for a further series of experiments with a low-compliance implant prevented further development of the project. In retrospect, the study design aligned with essential animal ethics guidelines, including a reduction in animal numbers to a minimum required for reliable data [28,29,30]. No study or experimental model is without limitations, and consideration of these limitations is key to maximizing study statistical power and insights while minimizing animal numbers. Furthermore, the use of animal models has come under increasing scrutiny, though it is not yet obsolete; as noted by experts from the U.S. National Institutes of Health.

“Although advancements in alternative methods will revolutionize biomedical research, the continued use of animal models and support of their resource infrastructure remain crucial. Animal models provide a comprehensive, whole-organism perspective that alternative methods cannot yet fully replicate. They deliver critical insights into systemic interactions, long-term effects and multifactorial responses to treatments, which are essential for understanding complex diseases and developing effective therapies. Furthermore, animal models are essential in the regulatory approval process because they are used to ensure the safety and efficacy of new drugs and medical devices before progressing to human clinical trials.” [30].

3.2.2. “Bottom-Up” Approaches Using Physical, Virtual, and In Silico Models of Pericellular Flow Mechanics and Cell Health

In contrast to top-down approaches (see above), “bottom-up” approaches start with smaller, less complex components (subsystems) of the larger complex system. A new understanding of each subsystem can be pieced together to form a basis for understanding the entire complex system. The bottom-up approach lends itself to mechanistic hypothesis testing. System components and relevant independent and dependent variables can (typically) be directly measured and/or described, though they often require development and/or the application of innovative methods [24]. Indeed, the conceptual jump from organismal movement to tissue- to cell-scale elucidation of nano–microscale flow fields, associated cell-surface drag forces, and mechanically modulated convective transport necessitates the use of bottom-up approaches.

We developed and validated novel physical and in silico models to understand nano–microscale flow fields around osteocytes in situ. High-resolution imaging of osteocyte network volumes from healthy, osteoporotic, and osteoarthritic patients provided submicron resolution datasets. From these, we could rapid-prototype inverse physical models using photopolymerizable resins (note: this is now referred to as 3D printing) [31]. In fluid dynamics, similitude theory allows for the measurement of flow properties by scaling up system length scales if fluid viscosity is scaled up by the same amount. By creating inverse physical prototypes 1000 times the size of the microscopic osteocyte network volumes, we could apply a pressure gradient across the volume. By using a fluid with 1000 times the viscosity of bone fluid, we could measure the permeability of cellular networks, a critical system parameter that had not been measured previously (Figure 5(A1,A2)).

Using the same cellular network datasets, we predicted signal transmission efficiency through healthy and diseased cellular networks using both stochastic models (Figure 5(A3,B1–B3)) and virtual parametric models. Of note, these methods enabled assessment of pericellular matrix effects in addition to the effects of lacunae–canalicular morphologies in the presence/absence of cells (Figure 5(A4)). Diseased bone often contains empty lacunae, indicating widespread osteocyte death. By combining these results with live–dead cell assays in live sections of bone, we could predict network transport efficiency impairment with loss in cell viability (Figure 5(A5)) [31].

To understand the mechanical effects of flow, e.g., fluid drag on osteocyte processes, we needed ultra-high resolution (sub-nm) images of the flow channels within the pericellular network. Importing actual 3D geometries from electron microscopy proved ideal for the approach; using nano–microfluidic computational models, we simulated and predicted flow around osteocytes within these geometries. Interestingly, naturally occurring constrictions in pericellular spaces amplified predicted flow velocities and associated drag on osteocytes, like narrows in a stream (Figure 5B) [32].

More recently, we used episcopic correlative imaging modalities to image tissue volumes, from organ to sub-cell-length scales, in osteoarthritic guinea pig knees. By delivering a mixed molecular weight bolus of fluorescent tracers via the heart, we discovered that the respective tissues comprising the knee joint act as a molecular sieve. Further, we showed that co-delivering tight junction permeability-modulating cytokines (i.e., TGF-α and TGF-β) with fluorescent tracers disrupted barrier function at tissue interfaces and transport through the joint. This provided compelling evidence that systemic immune functions impact molecular transport between the circulatory and musculoskeletal systems, as well as within synovial joint tissue compartments. This likely occurs both in response to chronic cytokine elevation (e.g., with osteoarthritis) and in response to acute cytokine spikes typical of cytokine storms in acute illness (Figure 5C) [33,34,35].

Exposure to a spike in cytokines TNF-alpha and TGF-beta abrogated tissue barrier function—the 76 kDa fluorescent tracer permeated both the extracellular (ECM) and pericellular matrix (PCM), in stark contrast to previous studies in which barrier function remained intact and no tracer permeated PCM. This work was carried out in collaboration with Dr. Lucy Ngo [33,34,35].

Figure 5. Elucidating the interdependence of biomechanics, molecular transport, cell signaling efficiency, and cross-talk between organ systems at decreasing length scales using physical, digital, virtual, and experimental models. (A) First, we used novel physical and in silico models to understand nano–microscale flow fields around osteocytes in situ. Based on high-resolution confocal image stacks (volumes) of osteocyte networks from healthy, osteoporotic, and osteoarthritic human bone biopsies (A2), we 3D-printed inverse physical models, where the osteocyte network created voids in the resin blocks (A1), which were scaled up by 1000×. Using similitude theory, we could then measure the permeability of the blocks using a fluid with a viscosity 1000× that of bone fluid (also referred to as similitude theory in mechanical engineering). We predicted signal transmission efficiency through the respective cellular networks using stochastic models (A3,B1–B3) and virtual parametric models of cells and their networks (A4), where we could probe the respective effects of live cells in situ (by removing nodes from the network to simulate cell death), as well as pericellular matrix on signaling efficiency around cells [31]. (A5) Using nano–microfluidic computational models, we could import actual 3D geometries from electron micrographs to simulate and predict flow around osteocytes using computational fluid dynamics software for nano–microfluidics (CFD-Ace). This approach showed that natural constrictions in pericellular spaces would be predicted to amplify drag on osteocytes, like narrows in a stream [32]. (B) These could then be tied to in situ measures of cell viability in preserved slices of bone (B1), where the effects of cell death on signal transmission efficiency could be predicted (B2,B3) by removing network nodes to simulate cell death. (A,B) In collaboration with Professor Eric J. Anderson [31,32]. (C) State-of-the-art episcopic correlative imaging studies enable cell-to-organ-scale high-resolution imaging of tracer transport from the heart to tissue compartments of the osteoarthritic knee joint ((C1), sagittal cross section with patella out of plane) and their inhabitant cells ((C2–C5) osteocytes, (C7,C8) chondrocytes). A 70 kDa fluorescent tracer, was chosen and delivered via the heart as a bioinert surrogate for albumin (67 kDa, large transporter protein) and with co-delivered cytokine spikes (TNF-α and TGF-β). Micrographs from TNF-α and TGF-β groups are color coded (via red and blue, respective, outlines). (C1) Gross molecular separation of the tracer (red) is evident in autofluorescent musculoskeletal tissues (white), visualized using circa 5168 confocal microscopy fields of view. (C2–C5) At higher magnification, some regions exhibit tracer within the vasculature and osteocyte lacunae. Marrow is indicated by crosses (x) and vasculature by asterisks (*). Light and dark arrowheads point to the tracer within intracellular and/or pericellular regions of osteocytes, respectively. (C6) Comparison of mean tracer intensity in the PCM and ECM of cartilage in the TNF-α and TGF-β groups. No significant differences were noted between cytokine groups or regions of interest, indicating loss of barrier function with exposure to cytokine spikes. (C7,C8) Tracer distribution in the superficial cartilage (indicated by asterisks, *) versus subchondral bone (indicated by crosses, x), where light and dark arrowheads (insets) point to chondrons in the calcified joint cartilage layer and the tidemark, respectively. Animals from the TNF-α group (C7) exhibited lower levels of the tracer in the ECM, and even less of the tracer in the PCM compared to the TGF-β group (C8). The TGF-β group exhibited similar patterns of the tracer in the cartilage than the TNF-α group but was distinct in exhibiting elevated tracer permeation at the tidemark and within the chondron (not observed in the TNF-α group (C8). (A,B) Adapted from refs. [31,32], with permission from SPRINGER-VERLAG. (C) Adapted from ref. [35]. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.

Figure 5. Elucidating the interdependence of biomechanics, molecular transport, cell signaling efficiency, and cross-talk between organ systems at decreasing length scales using physical, digital, virtual, and experimental models. (A) First, we used novel physical and in silico models to understand nano–microscale flow fields around osteocytes in situ. Based on high-resolution confocal image stacks (volumes) of osteocyte networks from healthy, osteoporotic, and osteoarthritic human bone biopsies (A2), we 3D-printed inverse physical models, where the osteocyte network created voids in the resin blocks (A1), which were scaled up by 1000×. Using similitude theory, we could then measure the permeability of the blocks using a fluid with a viscosity 1000× that of bone fluid (also referred to as similitude theory in mechanical engineering). We predicted signal transmission efficiency through the respective cellular networks using stochastic models (A3,B1–B3) and virtual parametric models of cells and their networks (A4), where we could probe the respective effects of live cells in situ (by removing nodes from the network to simulate cell death), as well as pericellular matrix on signaling efficiency around cells [31]. (A5) Using nano–microfluidic computational models, we could import actual 3D geometries from electron micrographs to simulate and predict flow around osteocytes using computational fluid dynamics software for nano–microfluidics (CFD-Ace). This approach showed that natural constrictions in pericellular spaces would be predicted to amplify drag on osteocytes, like narrows in a stream [32]. (B) These could then be tied to in situ measures of cell viability in preserved slices of bone (B1), where the effects of cell death on signal transmission efficiency could be predicted (B2,B3) by removing network nodes to simulate cell death. (A,B) In collaboration with Professor Eric J. Anderson [31,32]. (C) State-of-the-art episcopic correlative imaging studies enable cell-to-organ-scale high-resolution imaging of tracer transport from the heart to tissue compartments of the osteoarthritic knee joint ((C1), sagittal cross section with patella out of plane) and their inhabitant cells ((C2–C5) osteocytes, (C7,C8) chondrocytes). A 70 kDa fluorescent tracer, was chosen and delivered via the heart as a bioinert surrogate for albumin (67 kDa, large transporter protein) and with co-delivered cytokine spikes (TNF-α and TGF-β). Micrographs from TNF-α and TGF-β groups are color coded (via red and blue, respective, outlines). (C1) Gross molecular separation of the tracer (red) is evident in autofluorescent musculoskeletal tissues (white), visualized using circa 5168 confocal microscopy fields of view. (C2–C5) At higher magnification, some regions exhibit tracer within the vasculature and osteocyte lacunae. Marrow is indicated by crosses (x) and vasculature by asterisks (*). Light and dark arrowheads point to the tracer within intracellular and/or pericellular regions of osteocytes, respectively. (C6) Comparison of mean tracer intensity in the PCM and ECM of cartilage in the TNF-α and TGF-β groups. No significant differences were noted between cytokine groups or regions of interest, indicating loss of barrier function with exposure to cytokine spikes. (C7,C8) Tracer distribution in the superficial cartilage (indicated by asterisks, *) versus subchondral bone (indicated by crosses, x), where light and dark arrowheads (insets) point to chondrons in the calcified joint cartilage layer and the tidemark, respectively. Animals from the TNF-α group (C7) exhibited lower levels of the tracer in the ECM, and even less of the tracer in the PCM compared to the TGF-β group (C8). The TGF-β group exhibited similar patterns of the tracer in the cartilage than the TNF-α group but was distinct in exhibiting elevated tracer permeation at the tidemark and within the chondron (not observed in the TNF-α group (C8). (A,B) Adapted from refs. [31,32], with permission from SPRINGER-VERLAG. (C) Adapted from ref. [35]. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.

3.3. Mechanical Triggers for Stem Cells to Home in on and Heal Injured Tissue After Trauma

Another of our collaboratory’s foci has been understanding similar mechanical signal translation in the context of bone regeneration. In a one-stage bone transport surgical procedure developed by Knothe, we tested the capacity of resident periosteal stem cells to generate new bone and bridge critically sized bone defects (2.54 cm, considered too large to bridge without surgical intervention) in an ovine femur model [9,10,11]. Such defects result from trauma, infection, and tumor resection. In this procedure, after bone fixation is implanted, the periosteum is carefully “lifted” from the underlying bone, preserving the microvasculature. Periosteal lifting is achieved by severing Sharpey’s fibers that anchor the periosteum to bone surfaces like Velcro. An osteotomy frees the underlying bone for translation into the original bone void, and the periosteal sleeve is secured around the “new” defect region. Sheep loaded their bones (stance shift loading) immediately upon recovery from anesthesia and surgery. We observed astonishing early stem cell activation within days of injury. Tissue neogenesis caused by stem cells moved from the periosteum to the defect, completely bridging the “new” defect within 16 weeks [9].

We then aimed to understand the interplay between early mechanical signals, associated with stance shift loading, and those transduced to the lifted periosteum. The goal was to decipher the periosteum’s resident stem cells’ remarkable intrinsic healing capacity and to determine if repair patterns correlated with localized strains. Again, we created an ex vivo ovine loading model to replicate stance shift loading after the one-stage bone transport procedure (Figure 6). To measure cell-scale strains in the lifted periosteum during stance shift loading, we rented ultra-high-resolution television lenses to record displacements of minuscule markers (spray paint). Digital image correlation was used to calculate local strain fields at cell-length scales, the first such study of its kind to our knowledge [36]. We found that early bone formation within the defect correlated better with periosteal regions experiencing large-magnitude changes from pre-injury strains rather than areas experiencing large strains [36]. Interestingly, the mode of that change (i.e., tensile or compressive) had no effect.

In parallel, we carried out mechanics of materials studies on an ovine femoral periosteum (Figure 6D). We measured its tensile modulus of elasticity in longitudinal (toe region, −1.93 MPa; beyond transition, −25.67 MPa) and circumferential (4.41 MPa) directions [37]. We also measured its shrinkage upon severing Sharpey’s fibers that attach the periosteum to all bone surfaces (longitudinally, −30% to 50%; circumferentially, −10% to 20%) [37,38,39]. With this data, we could calculate the endogenous prestress in the periosteum, where the periosteum from the anterior of the skeletally mature ovine femur exhibits 12.06–0.40 MPa longitudinal prestress and 0.77–0.43 MPa circumferential prestress [36,37,38]. To better understand the mechanistic interplay between cellular and biochemical modulating factors in periosteum-mediated tissue genesis and healing, we developed a mathematical model, coupling finite element methods (mechanics) with cell dynamics modeling to simulate the clinical scenario and to predict parameters that promote healing [40].

These cell tissue experimental and computational mechanics studies yielded unprecedented insights into the intrinsic pre-/residual stresses in the periosteum. This knowledge, together with a later live imaging study of the periosteum, was essential to understanding how the periosteum’s stress state could trigger resident, quiescent stem cells to detach from the cambium layer to home in on and heal critically sized bone defects (Figure 6). In the latter study, we used multiphoton imaging at high resolution to measure the crimp of the collagen to which quiescent periosteal mesenchymal stem cells adhere, before and after periosteal lifting. We prepared ovine periosteum samples analogous to the in vivo one-stage bone transport surgery and the ex vivo mechanical testing studies. Initially, Sharpey’s fibers were left intact, thus retaining the intrinsic periosteal prestress. After establishing baseline collagen crimping and periosteal cell morphology, those fibers were severed for periosteal lifting. These studies showed, for the first time to our knowledge, that the collagen crimp increased with the release of the Sharpey’s fibers (Figure 6(C1) intact, Figure 6(C3) severed), which was expected based on our prior materials testing. More interestingly, prestress release via Sharpey’s fiber severing resulted in a physical rounding up of the cells themselves, a trigger to detach and ingress into the defect to initiate regeneration and healing [41].

Figure 6. To understand the role of biophysical factors in the surprising rate and volume of new osteogenesis observed in the one-stage bone transport procedure (developed by Dr. U. Knothe), we needed to elucidate the mechanical milieux of mesenchymal stem cells residing in the periosteum ex vivo ((A,B) in collaboration with Professor Sara McBride-Gagy [36]), in vitro ((C) in collaboration with Dr. Nicole Yu and Dr. Renee Whan [41]; (D) in collaboration with Professor Sara McBride-Gagy [37]), and in silico (not shown—in collaboration with Dr. Shannon Moore [40]) models. (A) After surgical preparation for the one-stage bone transport procedure, the femur was resected and placed in a loading rig to mimic in vivo stance shift loading, the condition under which in vivo studies demonstrated robust intramembranous infilling of the defect within 16 weeks of surgery. To measure cell-scale microstrain at high resolution during stance shift loading, we videotaped the surface of the periosteum from different angles using ultra-high-definition television lenses ((A) upper left) and (B) carried out digital image correlation (DIC) to calculate strains in the vicinity of quiescent stem cells during loading. Statistical analysis of strain-to-osteogenesis patterns revealed that a change in baseline strains showed greater correlation to tissue genesis than strain magnitudes per se or particular thresholds [36]. (C) To determine whether periosteal lifting, which attaches the periosteum in a pre-stressed state to the surfaces of bones via collagenous Sharpey’s fibers, affects cell and nucleus shape (C1–C3), we carried out live imaging protocols on ex vivo periosteal preparations to measure the change in volume and shape of periosteal cells with (C1) and without (C3) prestress in situ. Imaging showed that cutting Sharpey’s fibers to lift the periosteum increases the crimping of the collagen to which periosteal cells attach and subsequent rounding of quiescent stem cell nuclei ((C1–C3), where cell outlines are indicated by white ellipses and cell nuclei are depicted as fuchsia ovals and crimped collagen is depicted in neon green); hence, cutting Sharpey’s fibers acts as a mechanical trigger for the cells to migrate to sites of injury and initiate tissue genesis [41]. (D) Mechanical testing and shrinkage (not shown) studies of ovine periosteum enable measurement of periosteum’s anisotropy and prestress [37,38,39]. Experiments depicted in this figure were conducted at the AO Research Institute in Davos, Switzerland (A,B); the Biomedical Imaging Facility of the University of New South Wales (C); and the Experimental and Computational Mechanobiology Labs of Case Western Reserve University (D), with IACUC approvals under the guidance of the relevant local committees. (A,B) Reproduced from ref. [36], with permission from Kluwer Academic Publishers (Dordrecht). (C) Reproduced from [41]. (D) Reproduced from [37].

Figure 6. To understand the role of biophysical factors in the surprising rate and volume of new osteogenesis observed in the one-stage bone transport procedure (developed by Dr. U. Knothe), we needed to elucidate the mechanical milieux of mesenchymal stem cells residing in the periosteum ex vivo ((A,B) in collaboration with Professor Sara McBride-Gagy [36]), in vitro ((C) in collaboration with Dr. Nicole Yu and Dr. Renee Whan [41]; (D) in collaboration with Professor Sara McBride-Gagy [37]), and in silico (not shown—in collaboration with Dr. Shannon Moore [40]) models. (A) After surgical preparation for the one-stage bone transport procedure, the femur was resected and placed in a loading rig to mimic in vivo stance shift loading, the condition under which in vivo studies demonstrated robust intramembranous infilling of the defect within 16 weeks of surgery. To measure cell-scale microstrain at high resolution during stance shift loading, we videotaped the surface of the periosteum from different angles using ultra-high-definition television lenses ((A) upper left) and (B) carried out digital image correlation (DIC) to calculate strains in the vicinity of quiescent stem cells during loading. Statistical analysis of strain-to-osteogenesis patterns revealed that a change in baseline strains showed greater correlation to tissue genesis than strain magnitudes per se or particular thresholds [36]. (C) To determine whether periosteal lifting, which attaches the periosteum in a pre-stressed state to the surfaces of bones via collagenous Sharpey’s fibers, affects cell and nucleus shape (C1–C3), we carried out live imaging protocols on ex vivo periosteal preparations to measure the change in volume and shape of periosteal cells with (C1) and without (C3) prestress in situ. Imaging showed that cutting Sharpey’s fibers to lift the periosteum increases the crimping of the collagen to which periosteal cells attach and subsequent rounding of quiescent stem cell nuclei ((C1–C3), where cell outlines are indicated by white ellipses and cell nuclei are depicted as fuchsia ovals and crimped collagen is depicted in neon green); hence, cutting Sharpey’s fibers acts as a mechanical trigger for the cells to migrate to sites of injury and initiate tissue genesis [41]. (D) Mechanical testing and shrinkage (not shown) studies of ovine periosteum enable measurement of periosteum’s anisotropy and prestress [37,38,39]. Experiments depicted in this figure were conducted at the AO Research Institute in Davos, Switzerland (A,B); the Biomedical Imaging Facility of the University of New South Wales (C); and the Experimental and Computational Mechanobiology Labs of Case Western Reserve University (D), with IACUC approvals under the guidance of the relevant local committees. (A,B) Reproduced from ref. [36], with permission from Kluwer Academic Publishers (Dordrecht). (C) Reproduced from [41]. (D) Reproduced from [37].

4. Work and Energy Transfer of Development and Healing

The Work of Living, from Organisms to Cells

Every movement of the human body and every process within the body recycles (transfers) energy in an imperfectly sustainable way, across length and time scales, until the end of the life cycle (Figure 2A) [2,13]. While work is defined as the energy transferred to or from an object as force is exerted over a distance, only recently has the concept been applied to cells and the cellular work of mechanoadaptation [2], where the term virtual work has been introduced.

During movement, internal and external forces are transferred across multiple length scales (left, Figure 2A) while tissues adapt to the dynamic mechanical environment (right, Figure 2A). Together, the transfer of force from the environment and the subsequent structure–function adaptation of the system constitute the dynamic process of functional adaptation, also referred to as mechanoadaptation. In an engineering context, a control volume delimits the system of interest to enable, e.g., the tracking of work performed on the control volume and energy crossing system boundaries. At each relevant length scale, i.e., cell to tissue to organ to system, energy is transferred. At the length scale of the organism and its organ systems, exercise regimes represent a form of work to train the cardiovascular and musculoskeletal systems. At the cell-length scale, e.g., a stem cell, work deforms the cell and entrains the cell over time, resulting in activity modulation and lineage commitment [2].

Interestingly, adaptive responses to decreased and increased loading, i.e., below and above baseline, also cross length scales and modalities. For example, offloading an area of tissue with a relatively stiffer implant (e.g., stainless steel has a higher elastic modulus than bone) results in not only a change to the mechanical milieux of cells residing in the tissue but also to the local perfusion of the tissue by multiple mechanisms The implant obstructs blood vessels, and/or reduced tissue strains limit load-induced fluid flow. In contrast, “overloading” over baseline, e.g., of the radius by resecting the ulna in adult sheep [18], has the opposite effect, resulting in increased density and girth in the radius. Recent studies demonstrate similar effects in stem cells, where “offloading of cells”, e.g., via stiffer local substrates or gels, results in mechanoadaptation at the matrix- (tissue genesis by cells), cell-, and subcellular-length scales. Similarly, hindering the cells’ capacity to polymerize tubulin, the compression-bearing elements of the cytoskeleton, drives mechanoadaptation at the length scale of the cell (Figure 7; see animated version in [42]) [42,43,44].

Cellular adaptation can occur via either passive (e.g., volume change due to dilatational stress) or active means in response to local mechanical cues (Figure 7). Early cell studies from our collaboratory demonstrated that seeding at increased density (as opposed to proliferating cells to the same density) exerts local compressive forces on cells [22,42,43,44,45,46,47,48]. Under such local compression, cells can use a range of mechanisms to balance these forces. Cell volume can decrease either passively or through activation of ion channels, releasing intracellular fluid to the extracellular matrix. Another alternative is active cell division. The total cellular surface area effectively increases to balance forces. Also, the cytoskeleton can stiffen, effectively increasing the cell’s capacity to resist compression, thereby balancing forces. If forces are balanced at the cell boundary, even in the presence of local compression, no adaptation is required (Figure 7). In contrast, under local tension, expansion of cell volume, either passively or through activation of ion channels, provides a mechanism to balance forces. According to the results of recent experiments [42,43,44,45] carried out in collaboration with protégé Dr. Vina Putra, this volume expansion leads to nuclear fragmentation due to tension transduced through the cell to the nuclear envelope. A further adaptation alternative is cell division, which, similarly to cells under compression, provides a means to increase total cell surface area and distribute the forces over a greater area. An additional alternative is for the cytoskeleton to soften, effectively increasing the cell’s capacity to expand under tension and balance forces.

5. Discussion

A series of studies conducted over several decades by a collaboratory including diverse global universities, medical clinics, and research institutes demonstrates the power of convergent, interdisciplinary research to understand the mechanoadaptation of organisms, organs, and tissues by their inhabitant cells. The remarkable conservation of adaptation across length scales points to the emergent codification of the Laws of Biology, which will enable future, prospective prescriptions of exercise regimes, as well as wearable and implantable devices to augment cell-to-organismal health throughout life.

The intrinsic limitations of each experimental model alone necessitate multipronged approaches, where even the “ideal” approach to imaging events across length scales in real time (Figure 5) shows limitations with respect to a lack of standard scientific controls for all independent variables. Nonetheless, integrating across related studies provides a nuanced perspective of adaptation under conditions significantly more controlled than “real life”, providing a lens through which to reduce complexity.

During the time period of the studies depicted in this paper, ethical considerations and the evolution of science itself have significantly reduced the use and societal acceptance of animal models. Indeed, technological advances may provide superior alternatives going forward, where organismal-to-cell-scale events will be observable in real time and throughout life. Certainly, with rapid developments in machine learning and artificial intelligence, more practical limitations of the scientific and experimental approaches are being recognized, which also will increase efficiency and decrease intrinsic (and until recently, underappreciated) biases going forward, e.g., of in vitro approaches [49].

6. Conclusions

Examination of a series of independent yet related studies carried out by a meta-laboratory with a shared lead investigator (MLKT) and lead surgeon (Dr. Ulf Knothe) demonstrates the interplay between the movement of organisms and the physiology of their tissues, organs, and organ systems’ resident cells. Taken together, this work highlights the promise of increasing imaging and computing power, as well as machine learning/artificial intelligence approaches, to predict emergent behavior and thereby delineate the Laws of Biology [50]. Codifying these laws will provide a foundation for the future to promote not only the discovery of underpinning mechanisms but also the sustainability of our natural resources, from our brains to our bones, which serve as veritable “hard drives”, physically rendering a lifetime of cellular experiences and millennia of evolution.

7. Patents

[A] Knothe Tate, M.L.; Lin, M. Priority 22 April 2021, Filed 22 April 2022, Published 27 June 2024. Dynamic body part simulator. US AU US20240212526A1.

[B] Knothe Tate, M.L. Priority 3 March 2020, Filed 3 March 2021, Published 11 January 2023. Mechanoactive materials and uses thereof. US AU US20240212526A1.

[C] Knothe Tate, M.L. Priority 4 May 2018, Filed 7 May 2019, Published 10 March 2021. Smart composite textiles and methods of forming. EP US AU EP3787893A1.

[D] Knothe Tate, M.L. Priority 15 June 2015, Filed 27 August 2021, Published 23 September 2021. Engineered materials and methods of forming.

[E] Knothe Tate, M.L. Priority 16 August 2013, Filed 15 August 2014, Published 24 August 2016. A substrate. WO EP US AU EP3033057A4.

[F] Knothe Tate, M.L.; Knothe, U. Priority 5 March 2010, Filed 7 March 2011, Published 12 January 2012. Granted 17 March 2015. Multilayer surgical membrane. US8979942.

[G] Knothe Tate, M.L.; Anderson, E.J. Priority 20 April 2007, Filed 21 April 2008, Granted 17 December 2013, Published 17 December 2013. Flow directing materials and systems. US8609132B2.

[H] Knothe Tate, M.L.; Knothe, U. Priority 19 May 2005, Filed 20 February 2003, Published 1 February 2011. Composition and method for inducing bone growth and healing. US7879107B2.