Assessing the roles of collagen fiber morphology and matrix stiffness on ovarian cancer cell migration dynamics using multiphoton fabricated orthogonal image-based models

Abstract

Ovarian cancer remains the deadliest of the gynecological cancers, where this arises from poor screening and imaging tools that can detect early disease, and also limited understanding of the structural and functional aspects of the tumor microenvironment. To gain insight into the underlying cellular dynamics, we have used multiphoton excited fabrication to create Second Harmonic Generation (SHG) image-based orthogonal models from collagen/GelMA that represent both the collagen matrix morphology and stiffness (∼2-8 kPa) of normal ovarian stroma and high grade serous ovarian cancers (HGSOC). These scaffolds are used to study migration/cytoskeletal dynamics of normal (IOSE) and ovarian cancer (OVCA433) cell lines. We found that the highly aligned fiber morphology of HGSOC promotes aspects of motility (motility coefficient, motility, and focal adhesion expression) through a contact guidance mechanism and that stiffer matrix further promotes these same processes through a mechanosensitive mechanism, where these trends were similar for both normal and cancer cells. However, cell specific differences were found on these orthogonal models relative to those providing only morphology, showing the importance of presenting both morphology and stiffness cues. Moreover, we found increased cadherin expression and decreased cell alignment only for cancer cells on scaffolds of intermediate modulus suggesting different stiffness-dependent mechanotransduction mechanisms are engaged. This overall approach affords decoupling the roles of matrix morphology, stiffness and cell genotype and affords hypothesis testing of the factors giving rise to disease progression and metastasis. Further, more established fabrication techniques cannot simultaneously reproduce both the 3D collagen fiber morphology and stiffness. STATEMENT OF SIGNIFICANCE: Ovarian cancer metastasizes when lesions are small, where cells exfoliate from the surface of the ovary and reattach at distal sites in the peritoneum. The adhesion/migration dynamics are not well understood and there is a need for new 3D in vitro models of the extracellular matrix to study the biology. Here we use multiphoton excited crosslinking to fabricate ECM orthogonal models that represent the collagen morphology and stiffness in human ovarian tissues. These are then used to study ovarian cancer cell migration dynamics and we found that contact guidance and a mechanosensitive response and cell genotype all combine to affect the behavior. These models provide insight into disease etiology and progression not readily possible by other fabrication methods.

Keywords: Collagen; Extracellular matrix; Fabrication; Migration; Tumor microenvironment.

Figure 1.

Instrument diagram of the multiphoton fabrication microscope, showing the salient features of the scanning and laser control systems.

Figure 2.

Ovarian stromal images and corresponding fabricated scaffolds. (A) SHG optical sections of collagen from normal ovarian tissues. (B) Two-photon excited fluorescence images of the corresponding fabricated scaffolds. (C) SHG optical sections of collagen from high grade ovarian tumors. (D) Two-photon excited fluorescence images of the corresponding fabricated scaffolds. Each pattern is 200 x 200 μm in size with 50 μm in height. Scale bar =20 μm.

Figure 3.

Stiffness characteristics of the scaffolds. (A) Phase-contrast image showing the thickness of the scaffold. (B) Elastic modulus obtained for each laser exposure configuration using AFM. (C) Shrinking ratio of scaffolds exposed to ethanol.

Figure 4.

Representative trajectories of IOSE cells on image-based (A) normal and (B) high-grade stromal models over 72hrs across the three different stiffness groups. Representative trajectories of OVCA433 cells on image-based (C) normal and (D) high grade stromal models over 72hrs across the three different stiffness groups. There were approximately 20 cells tracked in each case.

Figure 5.

Migration dynamics for IOSE cells on normal and high-grade scaffolds with three different stiffnesses. (A) Instantaneous cell migration speed. (B) Motility coefficients (*p < 0.05). Migration dynamics for OVCA433 cells on normal and high-grade scaffolds with three different stiffnesses. (C) Instantaneous cell migration speed. (D) Motility coefficients (*p < 0.05).

Figure 6.

(A) Representative two-photon excited vinculin immunofluorescence images of IOSE cells on normal models with different stiffnesses. Scale bar = 20 μm. Focal adhesion density of IOSE (B) and OVCA433 (C) cells on image-based structures with different stiffnesses (*p < 0.05).

Figure 7.

IOSE cell alignment in response to ROCK inhibition on normal (A) and high-grade (B) stromal models with different stiffnesses. Analogous data for OVCA433 shown in (C) and (D), (*p < 0.05).

Figure 8.

Immunofluorescence intensity ratio for IOSE cells on/off patterns of both N-cadherin (A) and E-cadherin (B) on normal and high-grade stromal models with different stiffnesses and analogous data (C) and (D) for OVCA433 cells on/off patterns (*p < 0.05).