Mapping cellular-scale internal mechanics in 3D tissues with thermally responsive hydrogel probes

Abstract

Local tissue mechanics play a critical role in cell function, but measuring these properties at cellular length scales in living 3D tissues can present considerable challenges. Here we present thermoresponsive, smart material microgels that can be dispersed or injected into tissues and optically assayed to measure residual tissue elasticity after creep over several weeks. We first develop and characterize the sensors, and demonstrate that internal mechanical profiles of live multicellular spheroids can be mapped at high resolutions to reveal broad ranges of rigidity within the tissues, which vary with subtle differences in spheroid aggregation method. We then show that small sites of unexpectedly high rigidity develop in invasive breast cancer spheroids, and in an in vivo mouse model of breast cancer progression. These focal sites of increased intratumoral rigidity suggest new possibilities for how early mechanical cues that drive cancer cells towards invasion might arise within the evolving tumor microenvironment.

Fig. 1. Conceptual overview of proposed technique to measure local residual elasticity.

a Poly N-isopropylacrylamide (PNiPAAM) hydrogel droplets reversibly expand and collapse based on temperature. PNiPAAM microgels can be compacted at tissue culture temperatures of 37 °C and embedded in tissues of interest, where they will keep their contracted state while the tissue is maintained in culture conditions. Reducing the temperature below the lowest critical solution temperature triggers the microgels to expand. The degree of the expansion permitted depends on the rigidity of the surrounding porous material. The expansion ratio of the sensor can hence be used to determine highly localized measurements of internal tissue residual elasticity after creep, at or near tissue culture conditions. b To fabricate the hydrogels, an oil/water vortex emulsion technique is used to produce polydisperse spherical microscale temperature-actuated mechanosensors (µTAMS). c Swelling transitions between expanded and compacted states occur at 34 °C, which can be d reproducibly observed over multiple temperature cycles. Different colors represent individual microgels in (c, d). Scale bar = 50 µm. Representative images are consistent over three batches of µTAMs.

Fig. 2. Modeling and characterization of µTAM expansion.

a µTAMs can be modeled as pre-strained springs when compacted, which then deform the surrounding matrix when the pre-strain constraint is removed. b Simulations using this conceptual approach indicate that µTAMs sensitivity to the stiffness of the surrounding matrix can be tuned based on stored strain energy in the µTAM, which depends on µTAM rigidity over the actuation stroke length and applied pre-strain. A characteristic sigmoidal curve is observed with maximized measurement sensitivities in distinct measurement regimes. c Empirical characterization data demonstrate similar sigmoidal behaviors base on µTAM polymer formulation (Supplementary Table 1; data presented as mean ± standard deviation (SD); n = 6–11 µTAMs). Dashed line shows simulated data from a sigmoidal data fit with iteratively optimized parameters (Eq. (2); Supplementary Table 3). d Multiple µTAM measurements of residual matrix elasticity are compared against rheological measurements of matrix stiffness to determine the precision of each measurement. A linear relationship between matrix stiffness (black dashed line) and measurement precision (yellow bounding lines) was observed, and used as a model to estimate the error in all subsequent measurements.

Fig. 3. Distinct internal spheroid mechanics arise based on tissue formation technique.

a HS-5 fibroblast spheroids can be formed using a printable aqueous two-phase system (ATPS), in which cells are confined within a small droplet of immiscible liquids or by confining cells within a small cavity in a hydrogel where they passively aggregate. b These techniques produce grossly similar spheroids, with subtle distinctions in internal architecture as assessed by tissue sectioning and staining (green = f-actin; blue = nuclei; scale bar = 100 µm). c PNiPAAM microgels can be randomly incorporated into 3D multicellular spheroid cultures during the tissue formation process. d Pooled µTAM measurements across multiple spheroids show no obvious patterns of internal residual elasticity based on spatial location within the spheroid. Data presented as measurement ± expected error. e Significant differences are observed in average internal rigidity between the two formation techniques. Data presented as mean ± SD for n = 48 and 56 individual µTAM readings across 30 ATPS and 40 micropocket spheroids respectively for (d, e) over three independent experiments. * denotes p = 0.0022 for an unpaired two-tailed t-test. Representative spheroid images from one of the three independent HS-5 spheroid generating experiments of each method.

Fig. 4. Residual elasticities within engineered tumors varies based on cell type.

a Spheroids generated from aggressively invasive MDA-MB-231 and less aggressive T47D metastatic breast cancer cell lines. Scale bar = 100 µm. b Spatial variation of internal residual elasticity in tumor spheroids of each cell type (data presented as measurement ± expected error, pooled from 25 to 27 spheroids with 1–2 µTAMs embedded in each). c Histogram of measurement data demonstrates that a significant fraction of µTAMs in the MDA-231 spheroids register high residual rigidities. d The average residual elasticity within spheroids are significantly different based on cell type. Box plots indicate the median and 25th to 7th percentiles, and the whiskers span the range. b–d *p = 0.0074 (unpaired two-tailed t-test) and n = 33 and 28 µTAM measurements for T47D and MDA-MB-231 spheroids, respectively).

Fig. 5. Measurements of residual internal tumor elasticity in a mouse cancer model.

a µTAMs were co-injected with mCherry-labeled T41 breast cancer cells into the mammary fat pads of female mice, and allowed to form tumors over several weeks. At various time points, tumors were excised and imaged in a temperature-controlled saline bath. b µTAMs are initially clustered together after injection, but disperse as the tumor develops over several weeks. c H&E stained tissue sections of excised fat pads at week 3 shows recovery of normal tissue architecture immediately around the needle injection site (sham, no cancer cells), and an absence of normal architecture in the tumor model. (Insets) Considerable variability in tissue cellularity is observed in distinct regions of the tumor after 3 weeks. d µTAMs are interspersed with mCherry-labeled 4T1 cells in the mammary fat pad, and change size when the temperature is decreased (Scale bar = 50 µm). e Comparison of residual elasticity within tumors indicates an increasing number of high-rigidity measurements as the cancer progresses towards metastasis, and a significant difference in measurement distributions between day 7 and day 21 of tumor progression. Box plots indicate the median and 25th to 75th percentiles, and the whiskers span the range. n = 28, 6, 20, and 33 individual µTAM stiffness measurements in sham, post-injection day 7, 14, and 21, respectively. (*p = 0.022 by non-parametric two-tailed Mann–Whitney test to compare the distribution of ranks between groups). Representative images derived from 6 animal replicates for each time point with both left and right intraductal mammary injections to generate separate tumors.