Cancerous and non-cancerous lung extracellular matrixFrom a microstructure-mechanical property study to the development of a 3D platform to unravel cell-ECM interactions

Abstract

Lung cancer is the leading cause of cancer death among both women and men. It causes more deaths than colon, breast and prostate cancers combined. It is also the second most common cancer in both men and women, about 13% of all new cancers are lung cancer. Approximately 228,150 new cases are expected for the year 2020, which will cause about 142,670 deaths in the United States as the American Cancer Society expects. The mechanical properties of the Extracellular matrix (ECM) of many tissues, and specifically the lung, have been proven to affect cell and tissue functions. Moreover, it is well known that there is a dynamic reciprocity between cells and ECM mechanics, and this communication is affected during pathologies. However, the mechanisms by which cells stiffen the matrix remain understudied. The aim of this thesis is to characterize the mechanical behavior at local scale of healthy and pathological lung ECM and correlate it to its local microstructure. To achieve this goal, an Atomic Force Microscopy head has been mounted on top of an epifluorescence microscope to measure at the same locations the mechanical properties of the ECM and the microstructure of the three main fibrillar proteins of the lung ECM: collagen I, collagen III and elastin. Cancerous and non-cancerous lung ECM samples from 7 patients were obtained. The samples were sliced in 7 µm thick samples and the collagen I, collagen III and elastin were immunostained following a primary/secondary antibody protocol. Then 400 AFM indentations of 500 nm were performed in a 100*100µm area while each protein map was imaged using the epifluorescence microscope. Considering all the patients, the mean value of the effective elastic modulus measured by AFM was of 6.33 ±1.13 kPa for non-cancerous lung ECM and of 15.65±4.04 kPa. Therefore, there is a 2.5 fold increase of stiffness in cancerous lung ECM compared to non-cancerous lung ECM. For all the samples, the Young’s modulus showed a Gaussian stiffness distribution. When all the indentation tests performed for each patient were plotted together, that is tests performed on the cancerous and non-cancerous regions of the same slice, the distribution obtained was a bimodal for all the patients. The first peak of the distribution was related to the non-cancerous ECM and the second peak to the cancerous ECM. The mean values obtained from the peaks of the bimodal distribution overestimated the measured mean of both cancerous and non-cancerous ECM. Then, the correlation between the composition and the stiffness of the ECM was studied. First, the volume fraction of the fibrillary proteins in the samples was calculated using two different references, one relative to the maximum intensity of all the samples and the other one relative to the maximum intensity of each sample. Both showed an increment of the collagen I between the non-cancerous and cancerous samples with a mean increase of 1.7 folds and 1.5 folds, respectively. A positive correlation between the Collagen I volume fraction and measured stiffness was found for each sample. When the comparison was made between samples, a higher correlation was found for the second volume fraction, with an R2=0.60. Then, a microstructure-mechanical property relationship was studied. For that, a model based on Eshelby´s inclusion problem was used to predict the mechanical behavior of the lung cancerous and non-cancerous ECM. This model can estimate the elastic modulus of a matrix with ellipsoidal inclusions inside, that would resemble the ECM with the Collagen I fibers as the inclusions. Two different fiber distributions were considered. The first one assumes that the Collagen I fibers are oriented in 3D. Using an elastic modulus for the collagen I of 100 MPa, in the range reported in literature, the values of the elastic modulus of the ECM were overestimated by two orders of magnitude. A new value of the elastic modulus of collagen I fibers was calculated using the model and the measures obtained at the 10 points with the highest volume fraction of collagen I in all the samples. This calculation was done separately for non-cancerous and cancerous samples obtaining an elastic modulus of the collagen I fibers of 390 kPa for non-cancerous samples and of 1050 kPa for cancerous samples, well below the values reported in literature. The model predicted the E of the non-cancerous and cancerous lung ECM with a mean absolute error of 25.08% and 32.74% respectively, and an R2=0.6155 was obtained when a linear regression was fitted for the predicted versus measured values. The second approach assumes that Collagen I fibers are oriented in 2D. In this case, the elastic modulus of collagen I fibers is assumed to be of 100 MPa, in the range reported in literature. The elastic modulus of the matrix was tuned in order to minimize the absolute average error between the measured and predicted elastic modulus of the ECM. This was done separately for the non-cancerous and cancerous samples, mainly because cross-linking was not measured in this work. The best results were obtained for an elastic modulus of the matrix of 0.12 kPa for the cancerous ECM and of 0.05 kPa for the non-cancerous ECM, and calculating the Collagen I volume fraction with the maximum intensity value of each sample as reference. The prediction showed a mean absolute error of 14.48% for the non-cancerous lung ECM and of 11.15% for the cancerous ECM, with a correlation of R2=0.944 when a linear regression is fitted for the predicted versus measured stiffness. Finally, a functional platform with tunable stiffness for the study of 3D single cell-ECM interactions based on Methacrylate Hyaluronic Acid hydrogels was developed. First, Hyaluronic Acid Methacrylate was synthesized, which when crosslinked with dithiothreitol gave a range of stiffnesses ranging from 0.2 to 19 kPa. This range comprehends both the mean values of the cancerous and non-cancerous ECM. Then, proof of concept 3D cell migration assays were performed for A549 and H1299 cells inside of three hydrogel with different stiffnesses.