Cells in living organisms are constantly experiencing a variety of mechanical cues. From the stiffness of the extra cellular matrix to its topography, not to mention the presence of shear stress and tension, the physical characteristics of the microenvironment shape the cells’ fate. A rapidly growing body of work shows that cellular responses to these stimuli constitute regulatory mechanisms in many fundamental biological functions. Substrate strains were previously shown to be sensed by cells and activate diverse biochemical signaling pathways, leading to major remodeling and reorganization of cellular structures. The majority of studies had focused on the stretching avoidance response in near-uniform strain fields. Prior to this work, the cellular responses to complex planar strain fields were largely unknown. In this thesis, we uncover various aspects of strain sensing and response by first developing a tailored lab-on-a-chip platform that mimics the non-uniformity and complexity of physiological strains. These microfluidic cell stretchers allow independent biaxial control, generate cyclic stretching profiles with biologically relevant strain and strain gradient amplitudes, and enable high resolution imaging of on-chip cell cultures. Using these microdevices, we reveal that strain gradients are potent mechanical cues by uncovering the phenomenon of cell gradient avoidance. This work establishes that the cellular mechanosensing machinery can sense and localize changes in strain amplitude, which orchestrate a coordinated cellular response. Subsequently, we investigate the effect of multiple changes in stretching directions to further explore mechanosensing subtleties. The evolution of the cellular response shed light on the interplay of the strain avoidance and the newly demonstrated strain gradient avoidance, which were found to occur on two different time scales. Finally, we extend our work to study the influence of cyclic strains on the early stages of cancer development in epithelial tissues (using MDCK-RasV12 system), which was previously largely unexplored. This work reveals that external mechanical forces impede the healthy cells’ ability to eliminate newly transformed cells and greatly promote invasive protrusions, as a result of their different mechanoresponsiveness. Overall, not only does our work reveal new insights regarding the long-range organization in population of cells, but it may also contribute to paving the way towards new approaches in cancer prevention treatments.

Indiana University-Purdue University Indianapolis (IUPUI) / Tissue cells display mechanosensitivity in their ability to discern and respond to changes in the viscoelastic properties of their surroundings. By anchoring and pulling, cells are capable of translating mechanical stimuli into a biological response through a process known as mechanotransduction, a pathway believed to critically impact cell adhesion, morphology and multiple cellular processes from migration to differentiation. While previous studies on polymeric gels have revealed the influence of substrate elasticity on cellular shape and function, a lack of suitable substrates (i.e. with mobile cell-substrate linkers) has hindered research on the role of substrate viscosity. This work presents the successful design and characterization of lipid-bilayer based cell substrates of tunable viscosity affecting cell-substrate linker mobility through changes in viscous drag. Here, two complementary membrane systems were employed to span a wide range of viscosity. Single polymer-tethered lipid bilayers were used to generate subtle changes in substrate viscosity while multiple, polymer-interconnected lipid bilayer stacks were capable of producing dramatic changes in substrate viscosity. The homogeneity and integrity of these novel multibilayer systems in the presence of adherent cells was confirmed using optical microscopy techniques. Profound changes in cellular growth, phenotype and cytoskeletal organization confirm the ability of cells to sense changes in viscosity. Moreover, increased migration speeds coupled with rapid area fluctuations suggest a transition to a different migration mode in response to the dramatic changes in substrate viscosity.

Knochen ist ein lebendes Material, das seine Struktur an die mechanische Umgebung anpasst. Zur strukturellen Anpassung muss der Knochen die mechanische Belastung erfassen. Allerdings sind Knochen mechanisch so steif, dass die lokalen Verformungen zu klein sind um von den Knochenzellen direkt detektiert zu werden. Osteozyten sind Knochenzellen, die ein Zellnetzwerk in der mineralisierten Matrix bilden. Ihre Zellkörper sind in Lakunen untergebracht und ihre Zellprozesse in engen Kanälchen, den Canaliculi. Die Hypothese des Flüssigkeitsflusses besagt, dass der lastinduzierte Flüssigkeitsfluss durch dieses Lakunen-Canaliculi-Netzwerk (LCN) einen Verstärkungsmechanismus bereitstellt, der es den Osteozyten ermöglicht, die dynamische Belastung des Knochens zu erfassen. Wir stellen die Hypothese auf, dass die Architektur des LCN eine wesentliche Rolle in Bezug auf die Mechanosensitivität spielt, da sie den Flüssigkeitsfluss beeinflusst. Das zentrale Ziel dieser Arbeit ist es, diese Hypothese an realen LCN-Architekturen mit einem Modell des lastinduzierten Flüssigkeitsflusses zu testen und den resultierenden Fluss mit der Mechanoreaktion des Knochens zu vergleichen. Wir haben das LCN mithilfe konfokaler Laser-Scanning-Mikroskopie untersucht. Wir haben dann die auf den Kirchhoffschen Gesetzen basierende Schaltungstheorie verwendet, um die Geschwindigkeiten der Flüssigkeit in allen abgebildeten Canaliculi zu modellieren und darzustellen wie sich die verdrängte Flüssigkeit über das LCN verteilen würde. Basierend auf diesen Geschwindigkeiten wurde die Mechanoreaktion des Knochens vorhergesagt. In meiner Studie wurden die Knochen von Mäusen verwendet, wodurch kontrollierte in vivo Belastungsexperimente und die Messung der Mechanoreaktion in Bezug auf gebildeten bzw. resorbierten Knochen unter Verwendung von in vivo µCT möglich waren. Die Flüssigkeitsströmungsmuster durch das LCN korrelierten mit der gemessenen Mechanoreaktion. Das heißt, Knochenbildung wurde in Bereichen nahe höherem Fluss beobachtet, während Knochenabbau in Bereichen nahe geringem Fluss beobachtet wurde. Die Vorhersage der Mechanoreaktion unter Berücksichtigung der Architektur des LCN war quantitativ besser als eine Vorhersage, die nur auf mechanischer Belastung basiert. Qualitativ haben wir festgestellt, dass Gefäßkanäle im Kortex als lokale Senken des Flüssigkeitsflusses fungieren und daher den Fluss an der nahegelegenen Knochenoberfläche reduzieren. Im Gegensatz dazu nahmen die Strömungsgeschwindigkeiten für konvergente Netzwerkstrukturen zu, bei denen die Zahl der Kanäle zur Knochenoberfläche hin abnimmt. In einem zweiten Projekt konzentrierten wir uns auf gesunden, menschlichen osteonalen Knochen. Osteone sind zylindrische Strukturen um Gefäßkanäle, die praktisch vom umgebenden Knochen abgeschottet sind. Wir analysierten acht gewöhnliche Osteone mit einem nahezu homogenen LCN und neun Osteon-in-Osteonen, die durch eine ringartige Zone mit geringer Netzwerkkonnektivität zwischen dem inneren und dem äußeren Teil dieser Osteone gekennzeichnet sind. In Canaliculi, die die beiden Teile des Osteons in Osteonen überbrücken, wurde ein wesentlich höherer lastinduzierter Flüssigkeitsfluss beobachtet als in anderen Canaliculi. Dies führte dazu, dass der durchschnittliche Fluss 2,3-mal höher war als bei normalen Osteonen. Es ist daher wahrscheinlich, dass Osteon-in-Osteon-Konstruktionen besonders zur Mechanosensitivität des kortikalen Knochens beitragen. Die Untersuchungen in dieser Doktorarbeit legen nahe, dass die LCN-Architektur neben der mechanischen Belastung als Schlüsselfaktor für die Knochenanpassung dient. / Bone is a living material, which adapts its structure in response to the mechanical environment. For structural adaptation bone need to sense the mechanical loading. However, bone is so stiff that the local strains are too small to be directly sensed by bone cells. Osteocytes are bone cells that form a cell network located within the mineralized matrix. Their cell bodies are housed in lacunae and their cell processes in narrow canals, the canaliculi. According to the fluid flow hypothesis, load induced fluid flow through this lacunocanalicular network (LCN) provides an amplification mechanism which allows osteocytes to sense dynamic loading of the bone. We hypothesize that the network architecture of the LCN plays an essential role in bone’s mechanosensitivity, as it influences the fluid flow. We aimed to test these hypotheses by using real LCN architectures in a model of load induced fluid flow, and compare the resulting flow with the mechanoresponse of bone. We imaged the LCN using confocal laser scanning microscopy (CLSM). Image processing was then used to describe the LCN as a mathematical network consisting of edges and nodes, representing the canaliculi and their connections respectively. We then employed circuit theory, based on Kirchhoff’s laws, to model the velocities of the fluid in all the imaged canaliculi. Based on these velocities, the mechanoresponse of bone was predicted. Mice were used in my study, as this allowed a controlled in vivo loading and a measurement of the mechanoresponse in terms of formed/resorbed bone using in vivo µCT. Fluid flow patterns through the LCN of mice correlated with the measured mechanoresponse, i.e., bone formation was observed near surfaces of higher flow, while resorption was observed near surfaces with low flow. The prediction of the mechanoresponse considering the architecture of the LCN was quantitatively better than a prediction based on strains only. Qualitatively, we identified that vascular canals in the cortex act as local sinks of fluid flow and, therefore, reduce the flow at the nearby bone surface. In contrast, flow velocities increased in convergent network structures, where the flow is channeled into fewer canaliculi nearby the surface. In a second project we focused on healthy human osteonal bone. Osteons are cylindrical structures around vascular canals, which are practically sealed off from the surrounding bone. We analyzed 8 ordinary osteons with a rather homogeneous LCN, and 9 osteon-in-osteons, which are characterized by a ring-like zone of low network connectivity between the inner and the outer parts of these osteons. A substantially higher load-induced fluid flow was observed in canaliculi that bridge the two parts of the osteon-in-osteons. This resulted in an average flow, which was 2.3 times higher compared to ordinary osteons. It is therefore likely that osteon-in-osteons particularly contribute to the mechanosensitivity of cortical bone. Based on both studies in this PhD thesis we conclude that LCN architecture should be considered as a key determinant of bone adaptation besides mechanical loading.