Hyaluronan-Derived Swelling of Solid Tumors, the Contribution of Collagen and Cancer Cells, and Implications for Cancer Therapy

Abstract

Despite the important role that mechanical forces play in tumor growth and therapy, the contribution of swelling to tumor mechanopathology remains unexplored. Tumors rich in hyaluronan exhibit a highly negative fixed charge density. Repulsive forces among these negative charges as well as swelling of cancer cells due to regulation of intracellular tonicity can cause tumor swelling and development of stress that might compress blood vessels, compromising tumor perfusion and drug delivery. Here, we designed an experimental strategy, using four orthotopic tumor models, to measure swelling stress and related swelling to extracellular matrix components, hyaluronan and collagen, as well as to tumor perfusion. Subsequently, interventions were performed to measure tumor swelling using matrix-modifying enzymes (hyaluronidase and collagenase) and by repurposing pirfenidone, an approved antifibrotic drug. Finally, in vitro experiments on cancer cell spheroids were performed to identify their contribution to tissue swelling. Swelling stress was measured in the range of 16 to 75 mm Hg, high enough to cause vessel collapse. Interestingly, while depletion of hyaluronan decreased swelling, collagen depletion had the opposite effect, whereas the contribution of cancer cells was negligible. Furthermore, histological analysis revealed the same linear correlation between tumor swelling and the ratio of hyaluronan to collagen content when data from all tumor models were combined. Our data further revealed an inverse relation between tumor perfusion and swelling, suggesting that reduction of swelling decompresses tumor vessels. These results provide guidelines for emerging therapeutic strategies that target the tumor microenvironment to alleviate intratumoral stresses and improve vessel functionality and drug delivery.

Figure 1

Measurement of tumor swelling. (A) δ_ref is defined as the thickness attained when NaCl concentrations are isotonic enough to shield electrostatic interactions; δ > δ_ref tumor swelling in hypotonic solutions (the thickness of the sample increases); the stress required to compress the sample from thickness δ to δ_ref defines the swelling stress, T_c. (B) The tumor specimen was initially compressed to 10% strain and allowed to relax. Subsequently, NaCl solution of specified concentration was added, causing the swelling of the tissue, which reached a new equilibrium. Finally, the specimen was compressed for another 20% strain. The difference in the stress between the two equilibriums was quantified as the swelling stress. Islet shows a schematic of the experimental set up.

Figure 2

Experimental data of tumor swelling. Swelling stress as a function of NaCl concentration for the four orthotopic tumor models employed in the study. The experimental data were fitted to Eq. (S14) (solid line) and values of the fitting parameters are shown in Supplementary Table S1.

Figure 3

Effect of ECM composition on swelling stress. (A) Representative immunofluorescence staining sections for hyaluronan (HA) and collagen (scale bar 100 μm), (B) Swelling stress as a function of HA area fraction, and (C) collagen area fraction showing no correlation. (D) Swelling stress is linearly proportional to the ratio of HA/collagen area fraction (y = 4.089x − 2.057, R² = 0.825). Five tumor specimens (n = 5) from each tumor type were used.

Figure 4

Model predictions of (A) solid stress applied externally to the tumor by the host tissue, (B) swelling stress, (C) IFP, and (D) Donnan osmotic fluid pressure as a function of the radial position from tumor center at different times. Model was specified for the MCF10CA1a tumors. External solid stress is compressive in the tumor interior and becomes tensile at the periphery, swelling stress remains spatially and temporally uniform, IFP increases with time owing to increased vessel permeability and drops to normal (zero) values at the periphery, whereas osmotic pressure is negligible.

Figure 5

Effect of tissue swelling on vessel perfusion. (A) Representative immunofluorescence staining sections for hyaluronan (HA) and collagen (scale bar 100 μm), (B) typical immunofluorescence staining sections for lectin and CD31 (scale bar 100 μm), (C) area fraction of collagen and HA, and (D) swelling stress for the control and treated tumors tested. Changes in stress between the control and each of the treated groups are statistically significant (P < .05) for both tumor types. (E) Fraction of perfused vessels as a function of tumor swelling showing their exponential decay relationship (dash line, y = −27.22e^(−21.49x), R² = 0.8913) and also the different mechanism of collagen reduction to improve perfusion (dash circle).

Figure 6

Cancer cell swelling (A) microscope image of a spheroid. Lines depict the perimeter of the spheroid before (0 hour) and an hour after addition of electrolyte solution. (B) Change in spheroid's diameter as a function of the tonicity for HT1080 and MiaPaCa2 cancer cells, (C) stress strain curve of 1% agarose gel, and (D) stress developed on the spheroids as a function of their elastic modulus.