High Resolution Imaging of Primary Pancreatic Cell Attachment and Expansion on Cellulose Scaffolds
Due to the non-transparent nature of the cellulose scaffold, we examined cell attachment and extent of coverage on the scaffolds by SEM on days 1, 3, 5 and 10 after seeding (Figure 3). On day one, for all surface modifications, cells were distributed throughout the scaffolds. By day three, cells produced visible cytoplasmic extensions and were distributed more widely on the scaffolds. We observed a clear preference for non-coated unmodified (U) and amine-modified (A) and gelatin-coated, NaOH-modified (NG) surfaces compared with other conditions (Supplemental Figure 1). Significant cell expansion was observed by day five, along with cell migration into the scaffold interior pockets. Cell aggregates were also observed as well as extracellular matrix (ECM) on the scaffolds as shown in the images. These results were consistent with our MTT-assay quantification of viable cells over time under each condition. Cell aggregates were seen on uncoated, unmodified (U) and amine-modified (A) scaffolds more than on other surface conditions, which show a more extensive cell distribution (Supplemental Figure 1).
hiPSC-Derived NKX6-1+/PDX-1+ Pancreatic Cells Bound to Both NaOH- and Amine-modified Scaffolds
hiPSC-derived pancreatic cells were generated and matured to beta cells by combining two existing protocols (Kroon 2008, Millman 2016) and validated by immunocytochemical analysis of pancreatic markers (Figure 4) similar to previous experiments with primary pancreatic cells. On day 12 of differentiation, cells were detached and seeded onto cellulose scaffolds as single cells or as small aggregates. Based on the results of the MTT assay and SEM analysis using primary pancreatic cells, two conditions, uncoated amine-modified (A) scaffolds and gelatin-coated, NaOH-modified (NG) scaffolds were evaluated. Similar to previous experiments, 20,000 hiPSC-derived endocrine precursors were seeded on scaffolds and differentiation was continued in stage 5 and 6 differentiation media (Millman 2016) for 10-12 additional days (10 days for aggregates, 12 days for single cells), which should result in immature beta cells expressing markers such as insulin, NKX6-1, PDX-1 and MAF BZIP transcription factor A (MafA). The viable cell density was analyzed by MTT assay on days 1, 5 and 10/12, and cells successfully attached and proliferated on both surfaces (Figure 5A and Figure 5B). In comparison with single cells, the seeded aggregates had higher initial viable cell density; however, the viable cell density declined over time for cells on both surface conditions, with a very significant decrease on the gelatin-coated, NaOH-modified (NG) scaffolds on day 5, which reduced the number of viable cells to less than half of the initial number (Figure 5B). In contrast, the viable cell density for single cells on the uncoated amine-modified (A) scaffolds increased linearly over time. Even though single cells on the gelatin-coated, NaOH-modified (NG) scaffolds demonstrated slightly higher viable cell density than single cells on the amine-modified scaffolds, the viable cell number dropped suddenly on day 5, similarly to aggregates. To determine if cell viability was being affected or if cells were detaching from the scaffold surface, we used SEM to assess the attachment and morphology of the differentiating hiPSC-derived pancreatic cells on the scaffolds on days 1, 5 and 12. During differentiation, we observed a significant amount of ECM produced on uncoated amine-modified scaffolds loaded with single cells and aggregates (Figure 5E). In contrast, ECM deposition was lower on gelatin-coated, NaOH-modified (NG) scaffolds seeded with aggregates and single cells (Figure 5F). Images of aggregates on scaffolds from day 12, entering stage 6 (immature beta cells), reveal an almost empty scaffold, in agreement with MTT data showing a decrease in number of viable cells (Figure 5A). In both single cells and aggregates assays, uncoated amine-modified surfaces showed more uniform cell attachment and growth compared to gelatin-coated, NaOH-modified surfaces, and therefore, proved to be a better candidate for pancreatic cell culture.
Scaffold-Supported hiPSC-Derived NKX6-1+/PDX-1+ Pancreatic Cells Secrete Insulin
Insulin release by pancreatic cells on uncoated amine-modified (A) scaffolds and gelatin-coated, NaOH-modified (NG) scaffolds was quantified by ELISA on days 1, 5 and 10/12. Single cells on amine-modified (A) scaffolds produced the highest amount of insulin on day 1 (Figure 5C); however, insulin production declined on days 5 and 12, similar to single cells on gelatin-coated, NaOH-modified (NG) scaffolds (Figure 5D). The decrease in insulin production was consistent with loss of cells from the scaffolds. It is unclear if the remaining cells attached to the scaffolds exhibited any change in insulin production. Cell aggregates on both scaffold modifications had a very low insulin release at day 1. Even though insulin release increased by day 5 on both surfaces, it was still approximately half of the insulin released from single cells. On both surface modifications, insulin concentration surpassed that produced by a similar number of pancreatic cells cultured in a dish (data not shown). Based on ELISA results, single cells seeded on scaffolds secreted higher amounts of insulin than aggregates cultured on scaffolds, which in turn secreted more insulin than cells cultured on tissue culture dishes, with the highest amount produced by single cells on uncoated amine-modified (A) scaffolds.
Insulin Secretion by Scaffold-Attached hiPSC-Derived NKX6-1+/PDX-1+ Pancreatic Cells Increases with Integration into the Bioreactor
A miniature bioreactor was manufactured with the uncoated amine-modified (A) wicking matrix scaffold as that scaffold modification demonstrated the greatest potential for cell expansion and insulin production in our previous experiments. The bioreactor scaffold was pre-wetted with medium and then seeded with 5 x 106 or 1 x 107 hiPSC-derived pancreatic cells (Figure 6A). Insulin release and metabolite profiles were determined from assays of the culture medium. At the conclusion of experiments, MTT was added to the bioreactor to visualize viable cells (Figure 6B), and SEM visualization of cell attachment and distribution throughout the scaffold from the bioreactor was performed (Figure 6C).
Insulin ELISA of bioreactor samples from 1 x 107seeded cells showed an insulin production greater than 0.28 µg on day 2 of the bioreactor run. Insulin production was again evaluated at day 4, 6, 8, 10 and 13. Insulin production on day 6 increased to 0.3 µg and remained high through the last examined time point, day 13 (Figure 6D). When fewer cells were seeded (5 x 106) onto the bioreactor, an initial lag in insulin production was observed, but by days 7 to 13 higher insulin output was achieved, although slightly reduced versus seeding with 1 x 107 cells.
In addition to insulin production, we evaluated the metabolic profile of cells seeded at the 1 x 107 cell density for glucose and lactate production that are typically good indicators of process robustness. The metabolic profile of cells was determined using a YSI biochemical analyzer (Figure 6E). Glucose consumption in the bioreactor seeded with 1 x 107 cells increased ~2-fold between day 2 and day 6, after which the glucose consumption decreased ~15% between day 6 and 8 and another 10% by the end of experiment. Lactate production was much lower than glucose consumption and showed an opposite trend from glucose consumption, with a substantial drop in lactate production between day 2 and 8, followed by an increase until the end of the culture. The relatively high glucose consumption throughout the experiment is indicative of sustained metabolic activity of the cells in the bioreactor over 13 days. These results show cells in the bioreactor were able to attach to and grow on an amine-modified scaffold, maintain a steady metabolic activity, and produce high amounts of insulin for 13 days. A bioreactor inoculated with 5 x 106 cells was disassembled after 14 days and cells were harvested from the amine-modified scaffold by trypsinization and sonication; 50 x 106 cells were counted, demonstrating ~10-fold expansion of differentiating pancreatic cells in bioreactor.