Tumor emboli—aggregates of tumor cells within vessels—pose a clinical challenge as they are associated with increased metastasis and tumor recurrence. When growing within a vessel, tumor emboli are subject to a unique mechanical constraint provided by the tubular geometry of the vessel. Current models of tumor emboli use unconstrained multicellular tumor spheroids, which neglect this mechanical interplay. Here, we modeled a lymphatic vessel as a 200 μm-diameter channel in either a stiff or soft, bioinert agarose matrix to create a vessel-like constraint model (VLCM), and we modeled colon or breast cancer tumor emboli with aggregates of HCT116 or SUM149PT cells, respectively. The stiff matrix VLCM constrained the tumor emboli to the cylindrical channel, which led to continuous growth of the emboli, in contrast to the growth rate reduction that unconstrained spheroids exhibit. Emboli morphology in the soft matrix VLCM, however, was dependent on the magnitude of mechanical mismatch between the matrix and the cell aggregates. In general, when the elastic modulus of the matrix of the VLCM was greater than the emboli (EVLCM/Eemb > 1), the emboli were constrained to grow within the channel, and when the elastic modulus of the matrix was less than the emboli (0 < EVLCM/Eemb < 1), the emboli bulged into the matrix. Due to a large difference in myosin II expression between the cell lines, we hypothesized that tumor cell aggregate stiffness is an indicator of cellular force-generating capability. Inhibitors of myosin-related force generation decreased the elastic modulus and/or increased the stress relaxation of the tumor cell aggregates, effectively increasing the mechanical mismatch. The increased mechanical mismatch after drug treatment was correlated with increased confinement of tumor emboli growth along the channel, which may translate to increased tumor burden due to the increased tumor volume within the diffusion distance of nutrients and oxygen.

MK-8722, a systemic pan-AMPK activator improves glucose homeostasis in animal models of diabetes, but causes cardiac hypertrophy # . Two- and 3-dimensional tissue (2D,3D) constructs of human induced pluripotent stem cell derived cardiomyocytes (CM) (iCells, Fujifilm/CDI) were used to probe the translatability of cardiac hypertrophy and the role of AMPK in cardiomyocyte function. Chronic 10-day exposure (2 μM and 10 μM MK-8722) of iCell cardiomyocyte 2D monolayers, probed via monitoring of cellular impedance (CardioECR, Agilent), revealed a decrease in beat amplitude (-13%,2μM;-49%,10μM) as well as an increase in excitability, based on the ability to follow higher pacing frequencies (+81%,2uM;+34%,10uM). Direct force measurements in 3D iCell CM tissues (Biowire II, TARA Biosystems) reported loss of contractility when stimulated at 1Hz (-74%,2μM;-90%,10 μM), as well as changes in passive resting tension (+18%,2μM;-75%,10 μM) over chronic six-week exposure. MK-8722 caused a marked increase in Biowire excitability as revealed by a large drop in the voltage threshold of entrainment of paced tissues (-36%,2μM;-43%,10 μM). The increased excitability of both 2D and 3D CM models suggests a cellular correlate of the increased electrocardiographic QRS amplitude observed in rhesus monkeys upon chronic dosing with MK-8722 # . Hypertrophy of the Biowires could be clearly observed at the end of the 6-week incubation and presented as increased microtissue widths (+5%,2μM; +10%,10μM) and thickened regions of attachment to the chamber force sensors at each end of the Biowire. Volume estimation revealed increases (+22.3%, 2μM; +31.4%, 10 μM) in Biowire tissue volumes that were comparable to heart hypertrophic changes in preclinical test species # (mouse, rat, and rhesus monkey). The effect of 10μM MK-8722 was more pronounced when treatment began before tissues had completed a multi-week electrical stimulation maturation protocol, rather than after. The functional effects in both 2D and 3D CM models, and the contractile and structural readouts of the 3D Biowire model were highly informative in the interpretation of the clinical applicability of the pleiotropic cardiac effects of MK-8722 observed in preclinical testing. # Myers, et al. Science 357, p 507-511 (2017)

ABSTRACTTumor emboli – aggregates of tumor cell within vessels – pose a clinical challenge as they are associated with increased metastasis and tumor recurrence. When growing within a vessel, tumor emboli are subject to a unique mechanical constraint provided by the tubular geometry of the vessel. Current models of tumor emboli use unconstrained multicellular tumor spheroids, which neglect this mechanical interplay. Here, we modelled a lymphatic vessel as a 200 μm-diameter channel in either a stiff or soft, bioinert agarose matrix, and we modelled colon or breast cancer tumor emboli with aggregates of HCT116 or SUM149PT cells, respectively. The stiff vessel model constrained the tumor emboli to the cylindrical geometry, which led to continuous growth of the emboli, in contrast to the growth plateau that unconstrained spheroids exhibit. Emboli morphology in the soft vessel model, however, was dependent on the magnitude of mechanical mismatch between the vessel matrix and the cell aggregates. In general, when the elastic modulus of the vessel was greater than the emboli (Eves / Eemb >1), the emboli were constrained to grow within the vessel geometry, and when the elastic modulus of the vessel was less than the emboli (0 < Eves / Eemb < 1), the emboli bulged into the matrix. Inhibitors of myosin-related force generation decreased the elastic modulus and/or increased the stress relaxation of the tumor cell aggregates, effectively increasing the mechanical mismatch. The increased mechanical mismatch after drug treatment was correlated with increased confinement of tumor emboli growth along the vessel, which may translate to increased tumor burden due to the increased tumor volume within the diffusion distance of nutrients and oxygen.INSIGHT BOXThe growth of tumor emboli—aggregates of tumor cells within vessels—is associated with aggressive cancer progression and metastasis. Models of their growth have not taken into account their biomechanical context, where radial expansion is constrained, but lengthwise expansion is free in the vessel. Here, we modelled the vessel geometry with a cylindrical microchannel in a hydrogel. In contrast to unconstrained or fully embedded aggregates, vessel-like constraint promotes growth of emboli in our model. The growth advantage is increased when the matrix is stiffened or actomyosin contractility weakened, both of which effectively increase the magnitude of mechanical constraint. This study sheds light on increased tumor burden in vessel-based growth and indicates a need to study tumor progression in similar environments.