Injectable three-dimensional tumor microenvironments to study mechanobiology in ovarian cancer

Abstract

Epithelial ovarian cancers are among the most aggressive forms of gynecological malignancies. Despite the advent of poly adenosine diphosphate-ribose polymerase (PARP) and checkpoint inhibitors, improvement to patient survival has been modest. Limited in part by clinical translation, beneficial therapeutic strategies remain elusive in ovarian cancers. Although elevated levels of extracellular proteins, including collagens, proteoglycans, and glycoproteins, have been linked to chemoresistance, they are often missing from the processes of drug- development and screening. Biophysical and biochemical signaling from the extracellular matrix (ECM) determine cellular phenotype and affect both tumor progression and therapeutic response. However, many state-of-the-art tumor models fail to mimic the complexities of the tumor microenvironment (TME) and omit key signaling components. In this article, two interpenetrating network (IPN) hydrogel scaffold platforms, comprising of alginate-collagen or agarose-collagen, have been characterized for use as 3D in vitro models of epithelial ovarian cancer ECM. These highly tunable, injection mold compatible, and inexpensive IPNs replicate the critical governing physical and chemical signaling present within the ovarian TME. Additionally, an effective and cell-friendly live-cell retrieval method has been established to recover cells post-encapsulation. Lastly, functional mechanotransduction in ovarian cancers was demonstrated by increasing scaffold stiffness within the 3D in vitro ECM models. With these features, the agarose-collagen and alginate-collagen hydrogels provide a robust TME for the study of mechanobiology in epithelial cancers. STATEMENT OF SIGNIFICANCE: Ovarian cancer is the most lethal gynecologic cancer afflicting women today. Here we present the development, characterization, and validation of 3D interpenetrating platforms to shift the paradigm in standard in vitro modeling. These models help elucidate the roles of biophysical and biochemical cues in ovarian cancer progression. The agarose-collagen and alginate-collagen interpenetrating network (IPN) hydrogels are simple to fabricate, inexpensive, and can be modified to create custom mechanical stiffnesses and concentrations of bio-adhesive motifs. Given that investigations into the roles of biophysical characteristics in ovarian cancers have provided incongruent results, we believe that the IPN platforms will be critically important to uncovering molecular drivers. We also expect these platforms to be broadly applicable to studies involving mechanobiology in solid tumors.

Keywords: Agarose; Alginate; Cancers; Collagen; Extracellular matrix; Hydrogel; Injectable; Interpenetrating network; Mechanobiology; Ovarian cancers; Tumor microenvironment.

Fig. 1.. Collagen formation within semi-interpenetrating network hydrogels.

A. Xenograft, ovarian cancer patient derived tumors (top) were decellularized for downstream analysis (bottom). Bar segments = 1 mm. B. Injection molded 3% (w/v) agarose in the presence of 1 mg/mL collagen (top; dark pink) and 3% (w/v) alginate in the presence of 1 mg/mL collagen (bottom; tan). C, D. Histological stains (sirius red) of 3% (w/v) agarose and alginate hydrogels with increasing collagen content (1 mg/mL, 2 mg/mL, and 3 mg/mL) and decellularized tumors (slice thickness = 7.5 μm). E, F. Second harmonic generation (SHG) imaging of agarose-collagen and alginate-collagen IPNs plus decellularized tumors; reflecting conditions used for histology (sample thickness = 7.5 μm). G. Quantification of fiber area in SHG images; representative image of fiber area thresholding (top). H. Quantification of average fiber distance in SHG images; representative image threshold plus mesh overlay (top). All data points represent mean ± SDs with superimposed data points (alginate = tan, agarose = dark pink, decellularized tumors = grey); asterisk and hash-mark denotes significance P < 0.05 compared to decellularized tumors, determined by one-way analysis of variance (ANOVA) followed by post-hoc Mann-Whitney U-tests.

Fig. 2.. Comparison of nanoarchitecture observed in semi-IPN hydrogels and in the ECM of decellularized tumors.

A. SEM reveals ordered fibrous structures throughout decellularized patient derived ovarian cancer xenograft tumors. Left and center panels show two different representative images. Right panel shows an ovarian cancer cell within the matrix. B. Thick interpenetrating collagen fibers are observed throughout agarose-collagen and alginate-collagen (1 mg/mL) IPNs (1% to 3% of precursor alginate and agarose; w/v) with dense, well defined pore structure. C. Control agarose and alginate hydrogels ranging from 1% to 3% (w/v) without the presence of collagen, demonstrate modest variation of pore size with increased hydrogel content.

Fig. 3.. Rheological analysis of agarose and alginate IPN hydrogels.

Parallel-plate rheometric evaluation of: A. shear storage modulus (G’), B. shear loss modulus (G”), C. complex shear modulus (G*) for 1% (w/v) to 3% (w/v) agarose and alginate hydrogels with 1 mg/mL collagen, for n ≥ 4. All data points represent mean ± SDs with superimposed data points (alginate = tan, agarose = dark pink); asterisk and hash-mark (*, #) denotes significance P < 0.01 between conditions of either agarose-collagen or alginate-collagen respectively, while ampersand (&) denotes significance P < 0.01 between IPNs, determined by one-way analysis of variance (ANOVA) followed by post-hoc Mann-Whitney U test.

Fig. 4.. Optical accessibility, cell viability, and live-cell retrieval from IPN scaffolds.

A. OVCAR3 cells were encapsulated in 3% agarose + 1 mg/mL collagen and 3% alginate + 1 mg/mL collagen for 0, 24, and 48 hours respectively. Fluorescent images of calcein-AM and ethidium homodimer stained cells were captured using an inverted light microscope. B. In situ cell viability between each condition and time point did not show significance. C. Encapsulated OVCAR3 cells were imaged via multiphoton high resolution microscopy, displaying high degree of optical accessibility. D. OVCAR3 cells were recovered after encapsulation for 0, 24, and 48 hours in 3% agarose + 1 mg/mL collagen and 3% alginate + 1 mg/mL collagen. Recovered cells were analyzed for cell viability via flow cytometry. E. Flow-based quantification of recovered cell viability. All data points were plotted (n = 3); asterisk denotes significance P < 0.01 compared to each other condition, determined by ANOVA followed by post-hoc analysis.

Fig. 5.. Transduced stiffness utilizing IPN TME.

A. Fluorescent images of recovered GFP⁺ OVCAR3, cultured in 1.37% alginate and 1% agarose IPNs plus 3 mg/mL collagen (low stiffness, E₁), 3% alginate and 1.5% agarose IPNs plus 3 mg/mL collagen (high stiffness, E₂), and tissue culture plate (2D controls) for 48 hours. Cell migration was observed for an additional 48 hours after a vertical wound was inflicted. B. Corresponding image-based quantification of cell migration (all data points were normalized to initial wound distance, n ≥ 3 per condition). All data presented are means ± SEMs; single and double asterisks (*, **) denote significance of P < 0.05 and 0.01 with regards to agarose IPNs respectively, while a hash-mark (#) denotes significance of P < 0.05 with regards to alginate IPNs, determined by ANOVA within each corresponding time point, followed by post-hoc analysis.

Fig. 6.. Altered transcription in response to 3D IPN gels stiffness.

Transcriptional assessment via quantitative PCR (2^ΔΔCT, n ≥ 3 per condition) for GFP⁺ OVCAR3, cultured in 1.37% alginate and 1% agarose IPNs plus 3 mg/mL collagen (low stiffness, E₁), 3% alginate and 1.5% agarose IPNs plus 3 mg/mL collagen (high stiffness, E₂), and tissue culture plate (2D controls) for 48 hours. Trends display increased integrin, serpin, and MAPK related expression within the stiffer conditions. All data presented are means ± SDs; single and double asterisks (*, **) denote significance of P < 0.05 and 0.01, determined by Mann-Whitney U test when comparing responses in IPNs. The dotted black line represents gene expression in the 2D condition.