Fast-Relaxing Hydrogels Promote Pancreatic Adenocarcinoma Cell Aggressiveness through Integrin β1 Signaling

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is characterized by a dense extracellular matrix (ECM) exhibiting high stiffness and fast stress relaxation. In this work, gelatin-based viscoelastic hydrogels were developed to mimic the compositions, stiffness, and fast stress relaxation of PDAC tissues. The hydrogels were cross-linked by gelatin-norbornene-boronic acid (GelNB-BA), thiolated macromers, and a 1,2-diol-containing linear synthetic polymer PHD. Controlling the thiol-norbornene cross-linking afforded tunable stiffness, whereas increasing PHD content led to hydrogels with PDAC-mimicking fast stress relaxation. In vitro studies, including proliferation, morphology, and mRNA-sequencing, showed that fast-relaxing hydrogels supported PDAC cell proliferation, epithelial-mesenchymal transition (EMT), and integrin β1 activation. Blocking integrin β1 in vitro led to upregulating EMT markers in both slow and fast-relaxing hydrogels. However, this strategy profoundly impacted tumor growth rate and reduced tumor size but did not alter metastasis patterns in an orthotopic mouse model. This suggests a need to further evaluate the antitumor effect of integrin β1 blockade.

Figure 1

Cross-linking of fast-relaxing hydrogels. (A) Synthesis of GelNB-BA from reacting GelNB with CPBA. (B) Thiol–norbornene photoclick reaction for hydrogel cross-linking. (C) Synthesis of PHD from conjugating DOPAC to PHEAA. (D) Reversible boronate–ester bonding. (E) In situ photorheometry of hydrogel cross-linking. 365 nm light was turned on between 30 and 150 s. (F) Stress–relaxation profiles (at 10% strain) of GelNB or GelNB-BA gels containing PHD or PVA. (G) Effect of PEG4SH concentration on G′ of hydrogels. (H) Effect of PEG4SH concentration on stress relaxation of GelNB-BA/PHD/PEG4SH gels. All hydrogels were cross-linked by 6 wt % gelatin macromer (GelNB or GelNB-BA) with 2 mM LAP as the photoinitiator. Gel cross-linking was initiated by 365 nm light exposure at 5 mW/cm². PHD (or PVA in Figure 1F) was added at 2 wt %. Each test contained at least 3 independent gels.

Figure 2

Rheological properties of GelNB-BA/PHD/PEG4SH hydrogels. Effect of PHD concentration on (A) storage and loss moduli (G′ and G′′), (B) tan(δ), and (C) stress relaxation. GelNB-BA and PEG4SH were fixed at 6 and 1.5 wt %, respectively (R_thiol/NB = 1). Statistics were performed using One-way ANOVA, with *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001. (D) Chemical structures of PEG4SH. (E) Storage modulus and (F) stress–relaxation profiles of PEG4SH cross-linked GelNB-BA/PHD gels over 6 days in cell culture media. (G) Chemical structures of THA. (H) Storage modulus and (I) stress–relaxation profiles of GelNB-BA/PHD/THA gels over 7 days in cell culture media. Statistics were performed using Student’s t test, with *p ≤ 0.05, **p ≤ 0.01. Each test contained at least 3 independent gels.

Figure 3

Encapsulation of PCC spheroids in GelNB (elastic or slow-relaxing) or GelNB-BA (viscoelastic or fast-relaxing) hydrogels. (A) Day 1 and Day 6 images of RFP-COLO-357 in slow- and fast-relaxing hydrogels. (B) Size analysis of the COLO-357 cancer spheroids on Day 1 and 6. (C) EdU staining for proliferative cells at Day 0 and 6. (D) The percentage of EdU-positive cells in each spheroid was normalized to the total number of nuclei in the spheroid. Slow-relaxing gels were made with 6 wt % GelNB, 2 wt % PHD, and 1.5 wt % PEG4SH. Fast-relaxing gels were made with 6 wt % GelNB-BA, 2 wt % PHD, and 1.5 wt % PEG4SH. All gels were cross-linked with a 2 mM LAP initiator and under UV light (365 nm, 5 mW/cm²) for 2 min. The statistical significance of slow-relaxing and fast-relaxing gels was analyzed using Student’s t test analysis in Graphpad Prism software. For data that passed the normality test, a parametric t test with Welch’s correction was performed. For data that did not pass the normality test, a nonparametric Mann–Whitney test that does not assume Gaussian distribution was performed. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001.

Figure 4

Transcriptomic comparison of PCC spheroids encapsulated in slow- and fast-relaxing hydrogels. (A) PCA analysis of COLO-357 spheroids encapsulated in slow- and fast-relaxing gels. (B) Volcano plot for transcriptomic comparison between spheroids in fast-relaxing gels compared to those in slow-relaxing gels; determined by Deseq2 method, DEGs cutoff conditions are FDR < 0.05, |fold Change| ≥ 1.5. (C) Biological process GO term enrichment analysis of DEGs between fast-relaxing gels and slow-relaxing gels. (D, E) Selected pathways found significantly enriched or depleted (NES > 1.5 or < −1.5; FDR < 0.05) by GSEA in PDAC spheroids encapsulated in fast-relaxing gels compared to slow-relaxing gels. (D) NES > 0, the gene set is enriched in the fast-relaxing group. (E) NES < 0, the gene set is enriched in the slow-relaxing group.

Figure 5

KEGG pathway analysis. (A) ECM-receptor interaction pathway between slow- and fast-relaxing hydrogel groups. Rectangles indicate all of the genes involved in the pathway, and stars indicate genes related to the pathway that were differentially expressed between the slow-relaxing and fast-relaxing hydrogel groups. (B) Kaplan–Meier survival plots (obtained from TCGA database) showing reduced survival rate of patients over time with higher (red line) or lower (blue line) ITGB1 expression. (C) Representative images of Pa03C spheroids encapsulated in slow- and fast-relaxing gels treated with ITGB1 blocking antibody over a 6 day culture period. (D) Diameters of Pa03C and COLO-357 spheroids encapsulated in slow- and fast-relaxing gels. The encapsulated spheroids were all cultured in the presence of ITGB1 blocking antibody (except for the negative control group) over 6 days. Ordinary and nonparametric one-way ANOVA was performed on normal and non-normal data, respectively, p < 0.05. *, **, ***, and **** represent p < 0.05, 0.01, 0.0001.

Figure 6

The effect of matrix relaxation on tumor growth in vivo. (A) Images of primary tumor burden by BLI analysis over weeks 1, 3, and 5 from mice implanted with spheroids (∼65 per mouse) encapsulated in slow-relaxing, slow-relaxing + ITGB1 blocking antibody, or fast-relaxing + ITGB1 blocking antibody gels. Images are scaled to the Min and Max Radiance to each respective week and noted by the luminescence scale. (B) Regions of interest (ROI) are expressed as changes in the log area under the curve (AUC). Statistics are based on a comparison of log AUC values as described in Section 2, **p < 0.01, ****p < 0.0001 comparing log(AUC) values between weeks 1 to 5 in respective groups. The comparison of week 1 values was only significant, as noted by the dashed line, **p < 0.01. (C) Orthotopic tumor mass of mice at harvest. Each dot represents an individual mouse. * p < 0.05, One-way ANOVA.