Decoupling diffusional from dimensional control of signaling in 3D culture reveals a role for myosin in tubulogenesis

Abstract

We present a novel microfabricated platform to culture cells within arrays of micrometer-scale three-dimensional (3D) extracellular matrix scaffolds (microgels). These microscale cultures eliminate diffusion barriers that are intrinsic to traditional 3D culture systems (macrogels) and enable uniform cytokine stimulation of the entire culture population, as well as allow immunolabeling, imaging and population-based biochemical assays across the relatively coplanar microgels. Examining early signaling associated with hepatocyte growth factor (HGF)-mediated scattering and tubulogenesis of MDCK cells revealed that 3D culture modulates cellular responses both through dimensionality and altered stimulation rates. Comparing responses in 2D culture, microgels and macrogels demonstrated that HGF-induced ERK signaling was driven by the dynamics of stimulation and not by whether cells were in a 2D or 3D environment, and that this ERK signaling was equally important for HGF-induced cell scattering on 2D substrates and tubulogenesis in 3D. By contrast, we discovered a specific HGF-induced increase in myosin expression leading to sustained downregulation of myosin activity that occurred only within 3D contexts and was required for 3D tubulogenesis but not 2D scattering. Interestingly, although absent in cells on collagen-coated plates, downregulation of myosin activity also occurred for cells on collagen gels, but was transient and mediated by a combination of myosin dephosphorylation and enhanced myosin expression. Furthermore, upregulating myosin activity via siRNA targeted to a myosin phosphatase did not attenuate scattering in 2D but did inhibit tubulogenesis in 3D. Together, these results demonstrate that cellular responses to soluble cues in 3D culture are regulated by both rates of stimulation and by matrix dimensionality, and highlight the importance of decoupling these effects to identify early signals relevant to cellular function in 3D environments.

Fig. 1.

A microfabricated approach to minimize diffusion barriers in 3D culture. (A) Schematic of fabrication of 3D microgels. PDMS, poly(dimethylsiloxane). (B) Phase contrast image of individual MCF-10A cells cultured in 42×42×75 μm microgels (top), and a brightfield cross-sectional image of BAMECs suspended within 50×50×60 μm microgels (bottom). Scale bars: 50 μm. (C) MDCK cells fixed 24 hours after culture in the indicated matrices. Scale bars: 20 μm. (D) Simulations of protein diffusion (molecular mass =66 kDa). Dashed lines indicate time to half-maximal concentration, and correspond to 2.33 minutes for microgels and 15.5 hours for macrogels. (E) Quantification of NFκB nuclear translocation in NIH 3T3 cells 30 minutes after stimulation with 25 ng/ml TNFα. Means ± s.e.m. are from three independent experiments. (F) NIH 3T3 cells fixed 30 minutes after stimulation with TNFα. Scale bars: 10 μm.

Fig. 2.

HGF-induced ERK activation underlying MDCK scattering and tubulogenesis is regulated by rate of stimulation, not matrix dimensionality. (A) MDCK cells before and after stimulation with 25 ng/ml HGF for 24 hours. (B) ERK phosphorylation in response to 25 ng/ml HGF stimulation in MDCK cells. Means ± s.e.m. are from at least five independent experiments. Colored asterisks in the graph represent P<0.05 relative to time 0, via Student's t-test with Bonferroni correction for multiple comparisons. Black asterisks represent significantly different (P<0.05) profiles between 2D, macrogel and microgel, via Mann-Whitney U-test. (C) MDCK cells treated with 10 μM U0126 and then stimulated with HGF. (D) Profiles of HGF stimulation applied to MDCK cells cultured in microgels. Ramp 1: 0.1 ng/ml/minute; Ramp 2: 0.05 ng/ml/minute. Cells were lysed 2 hours from start or end of stimulation (black, blue, green arrows). (E) ERK phosphorylation in response to dynamic stimulation with HGF. Means ± s.e.m. are from at least five independent experiments. *P<0.05; #, P<0.10, Student's t-test with Bonferroni correction for multiple comparisons. Scale bars: 20 μm.

Fig. 3.

HGF-induced MLC activity is determined by matrix dimensionality and regulates tubulogenesis. (A) Relative MLC phosphorylation in response to 25 ng/ml HGF in MDCK cells. Means ± s.e.m. are from at least five independent experiments. The phosphorylated myosin (ppMLC) blot for macrogels was exposed substantially longer to allow visualization of the relatively weak bands. *P<0.05 between data points at 240 minutes, via Student's t-test with Bonferroni correction for multiple comparisons. (B) Protein expression after knockdown of MYPT1 with siRNA relative to a non-transfected control. Means ± s.e.m. are from three independent experiments. *P<0.05; ***P<0.005, Student's t-test. (C) Control or MYPT1-knockdown MDCK cells stimulated with HGF. Scale bars: 20 μm.