Static compression of single chondrocytes catabolically modifies single-cell gene expression

Abstract

Previous work has established that mechanical forces can lead to quantifiable alterations in cell function. However, how forces change gene expression in a single cell and the mechanisms of force transmission to the nucleus are poorly understood. Here we demonstrate that the gene expression of proteins related to the extracellular matrix in single articular chondrocytes is modified by compressive forces in a dosage-dependent manner. Increasing force exposure catabolically shifts single-cell mRNA levels of aggrecan, collagen IIa, and tissue inhibitor of metalloproteinase-1. Cytohistochemistry reveals that the majority of strain experienced by the cell is also experienced by the nucleus, resulting in considerable changes in nuclear volume and structure. Transforming growth factor-beta1 and insulin-like growth factor-I offer mechanoprotection and recovery of gene expression of aggrecan and metalloproteinase-1. These results suggest that forces directly influence gene transcription and may do so by changing chromatin conformation.

FIGURE 1

CCA for studying single-cell gene response to unconfined compression. (A) This device exposes single adherent cells to unconfined creep compression at forces ranging from 10 nN to 200 nN or greater. Beam theory is used to calculate the force while measuring the resulting cellular deformation. (B) Unconfined compression of single cells modifies cell gene expression by mechanotransduction. For chondrocytes, this modification could occur in genes related to major processes involved in cartilage homeostasis and disease. Gene expression is measured in compressed chondrocytes through single-cell RT-PCR.

FIGURE 2

Distributions of GAPDH mRNA levels for all cells analyzed. (A) The distribution of GAPDH was found to be non-Gaussian. (B) Log₁₀ transformation of the data creates a Gaussian distribution, allowing linear statistics to be performed.

FIGURE 3

Abundance of mRNA in single chondrocytes normalized to GAPDH as determined by single-cell real-time RT-PCR. All values + 1, transformed with log₁₀ to obtain a Gaussian distribution. Only cells that had detectable GAPDH were analyzed. Sample sizes ranged from 31 to 11. All results presented as mean ± SD. Letters indicate significance determined by ANOVA with post hoc analysis (p < 0.05). (A) Aggrecan core protein and (B) collagen IIa mRNA levels normalized to GAPDH decreased with increasing force. Aggrecan expression experienced a sharp decrease after compression with 50 nN of force, whereas collagen IIa showed a dosage decrease from control cells to those exposed to 100 nN of force. (C) TIMP-1 mRNA levels increased in a dosage-dependent manner from control cells to those exposed to 100 nN of force. (D) Calculation of cell axial strain shows significant increases at each force level (p < 0.0001).

FIGURE 4

Microscopy of compressed chondrocytes. Single cells were compressed and fixed with paraformaldehyde at three levels of deformation (0.7, 2.25, and 3.1 μm) equal to mean cell deformations during single-cell gene expression analysis. F-actin was stained with Alexa Fluor 647 phalloidin, and the nucleus/chromatin with Hoechst 33342. For each cell, image z-stacks were acquired at 0.3-μm spacing. Two-dimensional deconvolution was performed for all fluorescent images, and three-dimensional reconstruction was performed for the F-actin axial view. The nuclear axial view is a single plane imaged near the center. Orthogonal yz planes from the center of the cell were created from z-stacks of F-actin and the nucleus. In the nuclear axial view, nuclei and openings in nucleus are larger with increasing deformation (arrows).

FIGURE 5

Strain analysis of cells and nuclei during compression. Letters indicate significance determined by ANOVA with post hoc analysis. (A) Axial nuclear strain versus axial cellular strain reveals a nearly one-to-one correlation. Mean cell strain values from Fig. 3 D are included for comparison. (B) Lateral nuclear strain versus lateral cellular strain shows that the nucleus experiences more lateral strain than the cell. (C) The cellular apparent Poisson's ratio suggests incompressibility at 0.7 and 2.25 μm. It also shows a significant increase and deviation from isotropy at 3.1 μm (p < 0.0001). (D) The nuclear apparent Poisson's ratio increased in each deformation group (p < 0.0001), also indicating deviation from isotropy at higher strains.

FIGURE 6

Gene expression and strain analysis of compressed single chondrocytes exposed to the growth factors TGF-β1 (5 ng/ml) and IGF-I (100 ng/ml). All results presented mean ± SD. Sample sizes ranged from 31 to 7. (A) With TGF-β1, aggrecan mRNA levels do not decrease significantly in response to static loads (*p < 0.001), are stimulated at 25 nN, and are significantly different from all control groups (⁺p < 0.05). (B) Post hoc analysis shows that TGF-β1 and IGF-I treatment resulted in collagen IIa mRNA levels that are less than those in controls (^#p < 0.01). (C) TIMP-1 mRNA increases from static compression are prevented with IGF-I (*p < 0.001). (D) Calculation of the axial strains of all chondrocytes tested with RT-PCR shows that both TGF-β1 and IGF-I significantly reduced the amount of axial strain experienced by cells (^‡p < 0.0001).

FIGURE 7

COMP gene expression of compressed single chondrocytes exposed to the growth factors TGF-β1 (5 ng/ml) and IGF-I (100 ng/ml). The mRNA abundance in single chondrocytes normalized to GAPDH as determined by single-cell real-time RT-PCR. COMP mRNA levels were not significantly affected by static compression. However, TGF-β1 and IGF-I significantly decreased the expression of COMP in single cells. Letters indicate significance determined by ANOVA with post hoc analysis (p < 0.05).

FIGURE 8

Correlation of gene relative abundance to axial cell strain. Multiple linear regression analyses show that (A) TGF-β1 significantly increases the slope parameter for aggrecan (*p < 0.001), and (B and C) IGF-I significantly decreases the slope parameter for collagen IIa and TIMP-1 (*p < 0.001). (D) Correlation of COMP relative abundance to axial cell strain showed no significant effect on abundance/axial cell strain slope parameter.