Survival-Associated Cellular Response Maintained in Pancreatic Ductal Adenocarcinoma (PDAC) Switched Between Soft and Stiff 3D Microgel Culture

Abstract

Pancreatic ductal adenocarcinoma (PDAC) accounts for about 90% of all pancreatic cancer cases. Five-year survival rates have remained below 12% since the 1970s, in part due to the difficulty in detection prior to metastasis (migration and invasion into neighboring organs and glands). Mechanical memory is a concept that has emerged over the past decade that may provide a path toward understanding how invading PDAC cells "remember" the mechanical properties of their diseased ("stiff", elastic modulus, E ≈ 10 kPa) microenvironment even while invading a healthy ("soft", E ≈ 1 kPa) microenvironment. Here, we investigated the role of mechanical priming by culturing a dilute suspension of PDAC (FG) cells within a 3D, rheologically tunable microgel platform from hydrogels with tunable mechanical properties. We conducted a suite of acute (short-term) priming studies where we cultured PDAC cells in either a soft (E ≈ 1 kPa) or stiff (E ≈ 10 kPa) environment for 6 h, then removed and placed them into a new soft or stiff 3D environment for another 18 h. Following these steps, we conducted RNA-seq analyses to quantify gene expression. Initial priming in the 3D culture showed persistent gene expression for the duration of the study, regardless of the subsequent environments (stiff or soft). Stiff 3D culture was associated with the downregulation of tumor suppressors (LATS1, BCAR3, CDKN2C), as well as the upregulation of cancer-associated genes (RAC3). Immunofluorescence staining (BCAR3, RAC3) further supported the persistence of this cellular response, with BCAR3 upregulated in soft culture and RAC3 upregulated in stiff-primed culture. Stiff-primed genes were stratified against patient data found in The Cancer Genome Atlas (TCGA). Upregulated genes in stiff-primed 3D culture were associated with decreased survival in patient data, suggesting a link between patient survival and mechanical priming.

Keywords: 3D cell culture; confinement; mechanical memory; microgel; stiffness.

Figure 1

(a) The pancreas is a gland about 15 cm in length located in the abdomen and is involved with the secretion of digestive enzymes. (b) About 90% of cancers of the pancreas arise from the acinar cells that secrete these digestive enzymes and/or the ductal cells used in the transport of these enzymes. In healthy pancreatic tissue, ECM proteins are dispersed and unaligned, resulting in a soft microenvironment. (c) Pancreatic ductal adenocarcinomas develop desmoplasia, or the growth of fibrous tissue, around the tumor site. These fibrous tissue growths are associated with an increase in the stiffness surrounding and within the tumor microenvironment due to increased alignment and density of the ECM. Softer microenvironments have been associated with decreases in migration, better chemotherapeutic outcomes, and smaller tumor sizes, while stiffer microenvironments have been correlated with increases in inflammatory markers, cell proliferation, and recurrence.⁻

Figure 2

(a) Schematic of microindentation instrument prior to contact with bulk hydrogel. Displacement of the vertical piezoelectric nanopositioning stage, Z, the initial distance of the cantilever flexure from the vertical stage, L, and the spring constant of the flexure, K_F are shown. (b) During contact, the vertical stage displacement and flexure compression result in normal force (F_n) applied until the glass probe has reached some microindentation depth within the bulk hydrogel, d. (c) Force-displacement curves (approach and retraction) of soft 3 wt % (blue circles) and stiff 5 wt % (pink circles) polyacrylamide bulk hydrogels indented at v = 1 μm/s to a maximum normal force of F_n = 750 μN. The average and standard deviation of the reduced elastic modulus, E*, are calculated from the Hertzian contact mechanics fit (solid black lines) of the approach curve using 3 individual bulk hydrogels and 4 locations per sample, over a total of 12 indents (n = 12).

Figure 3

(a) Microgel rheology was performed between two aluminum parallel plates (25 mm diameter) with a 500 μm gap height. Frequency sweeps show the linear plateau of the storage modulus, G′. For the soft microgel (blue), the average storage modulus was G′ = 55 ± 10 Pa. For the stiff microgel (pink), the average storage modulus was G′ = 410 ± 60 Pa. (b) A strain sweep was performed at 1 Hz on jammed microgel (110 μm diameter particles) of either 3 wt % (blue) or 5 wt % polyacrylamide (pink). Both microgel formulations exhibit yielding at the storage modulus (G′) and loss modulus (G″) crossover point. (n = 2).

Figure 4

Experimental procedure for probing transcriptomic maintenance related to the mechanical environment by PDAC cells. Briefly, cells were dissociated from tissue culture plastic and a dilute suspension of cells in microgel was added to either soft (blue) or stiff (pink) 3D microgel culture. After 6 h, cells were isolated via centrifugation and added to 3D microgel culture as shown for an additional 18 h. Cells were extracted after 0, 6, and 24 h in 3D microgel culture, and libraries were prepared for RNA sequencing for each condition.

Figure 5

(a) Schematic showing cells removed from tissue culture plastic and cultured for 6 h in a 3D microgel microenvironment under either soft or stiff conditions. (b) Volcano plot of bulk RNA-sequencing results showing differentially expressed genes. Light gray points represent genes with nonsignificant p-values. Dark gray points represent genes with significant p-values. Points in red represent genes with significant p-values and absolute value log₂ fold-change >2.0. The p-values were adjusted using the Benjamini and Hochberg method. (c) Differentially expressed genes associated with pancreatic cancer comparing cells cultured in soft or stiff 3D microgel culture for 6 h. Different colors represent relative changes in the expression of each gene. Red bins represent upregulated genes, and blue bins represent downregulated genes. (d) Survival probability data are shown for n = 177 pancreatic cancer patient samples from The Cancer Genome Atlas (TCGA) using the top differentially expressed genes (log₂ fold-change >2.0) in PDAC cells after 6 h in stiff microgel culture. Patients with mRNA expression similar to the soft-cultured groups had a mean survival of 24 months, and patients with mRNA expression similar to the stiff-cultured groups had a mean survival of 18 months. Logrank (Mantel–Cox) p-value = 0.023.

Figure 6

(a) Schematic showing cells were moved from one 3D microgel (priming) environment after 6 h to another 3D microgel (final) environment for 18 h. (b) Differentially expressed genes associated with pancreatic cancer comparing cells cultured for 24 h in a 3D microgel culture system. Cells were switched after 6 h from either soft or stiff microgel to soft or stiff microgel (soft-to-soft, soft-to-stiff, stiff-to-soft, stiff-to-stiff). Hierarchical clustering shows the quality of replicates. (c) Survival probability data are shown for n = 177 patient samples from The Cancer Genome Atlas (TCGA) using the top upregulated genes (n = 28) in stiff-primed PDAC cells after 24 h in 3D microgel culture. Patients with mRNA expression similar to the soft-primed group had a mean survival of 30 months, and patients with mRNA expression similar to the stiff-primed groups had a mean survival of 19 months. Logrank (Mantel–Cox) p-value = 0.046.

Figure 7

(a) Representative sum-intensity z-projections of immunofluorescence imaging provide additional evidence of persistent mechanical conditioning for the tumor suppressor gene BCAR3. (b) Representative sum-intensity z-projections of the metastasis marker RAC3 are shown to further support protein conversion from RNA-based analyses. (c) Quantification of sum-intensity z-projection fluorescence is shown for the tumor suppressor gene BCAR3. Raw fluorescence intensity was normalized to cell area. Significance was calculated using the mean, standard deviation, and number of replicates with a Student’s t-test. (d) Quantification of sum-intensity z-projection fluorescence is shown for the metastasis-related gene, RAC3. Raw fluorescence intensity was normalized to the cell area, and means and standard deviations are shown. Significance was calculated using a Student’s t-test. (p-values are denoted as * < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001).