Differentiating Human ES and iPS Cells to Pancreatic Progenitor Cells

The STEMdiff™ Pancreatic Progenitor Kit was optimized with reproducibility in mind. We designed both the medium formulation as well as the protocol to maximize the efficiency of differentiation across multiple human embryonic stem (ES) and induced pluripotent stem (iPS) cell lines (see Figure 2 here for the supporting data). At STEMCELL Technologies, we routinely use the Wisconsin H1 (WA1) and H9 (WA9) human ES cell lines during the development of our STEMdiff™ products. In addition, we have acquired or generated many iPS cell lines, derived from a variety of somatic cell types. Results generated using these cell lines indicate that the user of this kit can expect to see on average, greater than 65% of cells co-expressing PDX-1 and NKX6.1 at the end of Stage 4. Some experiments have generated greater than 90% pure populations of pancreatic progenitors. We do not see a significant difference in the efficiency of differentiation between human ES and iPS cell lines. Within a given experiment, well-to-well variability is also quite low, generally less than 5%.

While observing PDX-1 expression is a good sign, it is not sufficient to confirm successful generation of pancreatic progenitor cells. PDX-1 expression is also observed in neighbouring regions outside the developing pancreas, including the stomach and duodenum. It is critical to ensure that the PDX-1-expressing cells are of the pancreatic lineage. The best way to do this is by checking for co-expression of NKX6.1, as upregulation of both of these transcription factors is indicative of the transition to the pancreatic progenitor stage, and is observed in the human fetal pancreas at around 9 weeks of gestation. During human pluripotent stem cell differentiation to pancreatic progenitor cells, PDX-1 expression is upregulated during Stage 3, or between days 6 and 9 of the protocol. NKX6.1 expression is typically low at this stage, increasing significantly by the end of Stage 4. Additional markers that can also be used to ensure your cells are pancreatic progenitors include SOX9 and PTF1α. See Figure 3 on this page for a gene expression profile of these markers throughout the stages of human ES and iPS cell differentiation to pancreatic progenitor cells. To ensure efficient generation of PDX-1⁺/NKX6.1⁺ pancreatic progenitor cells, we recommend using our new pancreatic progenitor differentiation kit.

The recent publications by Rezania et al. and Pagliuca et al. describe culture methods for the generation of maturing pancreatic beta cells. These maturing beta cells express insulin and are able to secrete it in a glucose-dependent manner, although with altered secretion kinetics. The cells generated at the end of these two published protocols are both derived from stage 4 PDX-1⁺/NKX6.1⁺ pancreatic progenitor cells, which are similar to those generated using the STEMdiff™ Pancreatic Progenitor Kit. It is important to note that the cells generated at the end of the 4-stage protocol that accompanies the STEMdiff™ Pancreatic Progenitor Kit are truly pancreatic progenitor cells in that they retain the ability to generate both endocrine and exocrine cell lineages under appropriate conditions. Since the kit contains multiple basal media and supplements to support the different differentiation stages, it is also possible to isolate cells from earlier stages, thus allowing the user to study the development of pancreatic cell types from earlier precursors.

There is significant interest in many biological fields to move away from the use of serum and other undefined components in cell culture media. It is thought that removing undefined components will help reduce variability associated with lot-to-lot changes in the composition of these undefined components. In addition, stockpiling ‘good’ lots of serum can be a costly venture for smaller labs thus increasing the motivation to move towards more defined media. State-of-the-art human pluripotent stem cell maintenance media such as mTeSR™1 and TeSR™-E8™ are serum-free and defined formulations that promote robust feeder cell-independent expansion of human pluripotent stem cells in the undifferentiated state. The STEMdiff™ Pancreatic Progenitor Kit is likewise serum-free and defined. The most validated published protocols for the generation of pancreatic progenitor cells, such as those published by the ViaCyte group (Schulz et al., 2012), typically require the use of serum in early stages of the differentiation protocol. The STEMdiff™ Pancreatic Progenitor Kit was designed to mimic the performance of these established protocols but in the absence of serum. Together with our TeSR™ maintenance media, we are able to provide the user with a completely defined, serum-free workflow for the generation of pancreatic progenitor cells from human pluripotent stem cells.

The STEMdiff™ Pancreatic Progenitor Kit is designed and suitable for research use applications only. The medium is defined and serum-free but is not intended for clinical use.

The STEMdiff™ Pancreatic Progenitor Kit is comprised of 2 basal media and 6 supplements that are used in specific combinations during the 2 week differentiation. Accompanying the kit is a detailed step-by-step protocol, including a schematic of the differentiation process. Because the medium is supplied in this modular format, there is no need for complex medium manufacturing on the part of the user. Simply mix the appropriate stage-specific supplement and basal medium and add it to your cells each day to achieve robust and efficient differentiation. With our own in house testing, we find that variability is quite low between experiments and across cell lines. For example, we measured the percent PDX-1⁺/NKX6.1⁺ cells at the end of Stage 4 across 10 different experiments using the H1 cell line. The data show an average differentiation efficiency of 69.3 ± 11.9% (mean ± SD) with a range of 46.9 - 83.3%. Within a given experiment, duplicate wells showed an average difference of 1.9 ± 1.2% (mean ± SD). Finally, we show that the average efficiency of differentiation is similar in human embryonic stem cell lines and induced pluripotent stem cell lines. For these data, please refer to Figure 2B here. For the kit protocol, please view this product information sheet.

Achieving robust and high levels of NKX6.1 expression in human pluripotent stem cell-derived pancreatic progenitor cells has been a significant challenge in the field. As demonstrated in the 2013 publication from Dr. Tim Kieffer’s Lab at the University of British Columbia, in collaboration with BetaLogics (Rezania et al., 2013), enriching NKX6.1 expression in differentiating pancreatic progenitor cells significantly improves their ability to mature in vivo towards functional endocrine cells. Since this publication, more emphasis has been placed on characterizing PDX-1 and NKX6.1 co-expression during the differentiation process. We routinely assess the formation of pancreatic progenitor cells from human pluripotent stem cells by examining the expression of PDX-1 and NKX6.1 by qPCR as well as the co-localization of these markers by flow cytometry. Looking at a sample set of seven independent differentiation experiments, we find that the percentage of cells expressing PDX-1 as measured by flow cytometry is 80.7 ± 10.2% (mean ± SD) while the percentage of cells expressing NKX6.1 is 76.1 ± 6.3%. Within this data set, we see similarly low variation in the expression of NKX6.1 as with PDX-1. Nearly all PDX-1-positive cells also express NKX6.1, indicating a very efficient formation of pancreatic progenitor cells when using the STEMdiff™ Pancreatic Progenitor Kit. In addition, an examination of gene expression by qPCR indicates upregulation of PDX-1 and NKX6.1 to similar levels by the end of Stage 4. In five experiments, we found the upregulation of PDX-1 to be 2.7 ± 1.1-fold (mean ± SD) compared to housekeeping genes TBP and 18S ribosomal RNA and the upregulation of NKX6.1 to be 3.3 ± 1.2-fold. Both genes were upregulated to a similar extent with similarly low standard deviations, again demonstrating the reproducibility of generating pancreatic progenitor cells using this kit.

Reproducible differentiation across both human embryonic stem (ES) and induced pluripotent stem (iPS) cell lines is something we strive for when developing our STEMdiff™ kits. Demonstrating such reproducible differentiation across cell lines and across experiments is especially critical when studying human iPS cell lines from diseased patients or in studies where the genome of the cell line has been modified to mimic or correct a disease state. In these studies, changes in differentiation efficiency can be better attributed to the disease state or genetic modification, rather than the culture protocol, when a robust and reproducible system is used during the differentiation. We have presented data indicating similar efficiencies of differentiation in two human ES cell lines and two human iPS cell lines (see Figure 2 here) and we continue to add data for additional human iPS cell lines. Therefore, we have confidence that the STEMdiff™ Pancreatic Progenitor Kit is as likely to be compatible with newly derived human iPS cell lines as those protocols detailed in the literature.

The STEMdiff™ Pancreatic Progenitor Kit is supplied as two basal media and six pre-mixed supplements, all of which have been rigorously performance tested in differentiation assays to assure the user receives a high performing product. The user is simply required to combine the appropriate basal medium and supplement(s) and apply these complete media to the differentiating cells according to the detailed protocol provided with the kit. Along with our detailed protocols for the maintenance of undifferentiated human pluripotent stem cells using mTeSR™1, new and experienced users are provided with all of the reference material required to learn this workflow. In our experience, people using the kit for the first time are able to successfully generate pancreatic progenitor cells with the expected efficiency. In addition, our product and scientific support staff are available to assist with any questions that may arise regarding the use of our products.

To verify that human induced pluripotent stem cells are indeed pluripotent, there are several available methods. Each have their advantages and disadvantages. For example, the teratoma assay can provide information about the trilineage differentiation potential of human iPS cell lines in a single experiment, but has the disadvantage of being expensive (due to animal housing costs), laborious and requiring a trained pathologist to identify the different lineages within the heterogenous tumor. An alternative means of proving trilineage differentiation potential is to directly differentiate new iPS cell lines to each of the three germ layers in vitro, using specialized media for endoderm, mesoderm and ectoderm germ layer specification. Using defined and reproducible cell culture media provides a robust and consistent method of assessing differentiation potential and these protocols can be scaled down to smaller well formats to increase throughput. STEMCELL offers our our STEMdiff™ line of differentiation products, which include media to generate definitive endoderm (STEMdiff™ Definitive Endoderm Kit), early mesoderm (STEMdiff™ Mesoderm Induction Medium) or neural progenitor cells (STEMdiff™ Neural Induction Medium). Using these three products creates a robust tool for the assessment of trilineage differentiation potential in newly-derived human iPS cell lines.

STEMCELL Technologies has developed a comprehensive suite of tools to generate, characterize, expand and differentiate human induced pluripotent stem (iPS) cells. The STEMdiff™ Pancreatic Progenitor Kit is the latest tool within this workflow that provides robust and reproducible differentiation across different human iPS cell lines. The kit has been validated in-house on three independent human iPS cell lines, each achieving average efficiencies of greater than 60%, as assessed by the percentage of cells co-expressing PDX-1 and NKX6.1 by flow cytometry. Uncovering changes in the developmental potential of diseased iPS cell lines requires a robust and reproducible differentiation protocol, such that observations made between the diseased cell lines and normal controls can be attributed to the disease state, rather than inconsistencies in the differentiation process. For this reason, the STEMdiff™ Pancreatic Progenitor Kit is an ideal tool for this workflow.

Starting confluence can be a critical factor in the efficient generation of definitive endoderm. This is especially important when the differentiation procedure starts with the plating of a single cell suspension of undifferentiated pluripotent stem cells. Because confluence can be a relatively subjective measurement, we prefer to plate a specific number of cells into each well when starting a differentiation run. The optimal number of cells is 2.1 x 10⁵ cells/cm². This translates to 800,000 cells in a 12-well plate well, or 2 million cells in each well of a 6-well plate. When this number of cells is plated in mTeSR™1 with 10 µM Y-27632, the user can expect to see a monolayer of cells at greater than 80% confluence after overnight culture. There is more tolerance on the higher side of this cell number than on the lower side. Plating up to 1.2 million cells in a well of a 12-well plate will not adversely affect differentiation efficiency. Reducing the number of cells to 600,000 cells per well may result in reduced performance. Because confluence is a critical factor to achieving highly efficient definitive endoderm differentiation, plating single cells as a monolayer is preferred over plating of clumps or colonies. The size distribution of clumps or colonies tends to be significant with traditional scraping techniques. The size of these clumps can potentially influence the differentiation potential of the resulting colonies thus introducing variability into the system. Furthermore, it is difficult to achieve a target starting confluence when plating clumps, a problem that is largely avoided by working with single cells.

During the 14 days of differentiation, there is significant proliferation of cells. A confluent monolayer (single cell depth) is achieved very early in the protocol, certainly by day 2 when the cells are effectively definitive endoderm. Proliferation continues throughout the remaining stages such that after plating 800,000 cells to start, the final number of cells in the well is typically greater than 3 million. There is a significant decrease in the average size of the cells as they pack in, but the cells also begin to multilayer. We have estimated that at the end of Stage 4, the culture can be as much as 5-8 cell layers thick. This does present a challenge when attempting to image the cells in situ without the use of confocal microscopy. There are alternative options to performing successful immunocytochemistry in this scenario. One such option is to culture the cells on glass coverslips. Placing these coverslips on a microscope slide will significantly improve image quality but alone does not eliminate the issue of the cells being several layers thick. Adding the use of structured illumination using, for example, the ZEISS ApoTome technology can further improve image quality. The result is a clearer image than can be achieved with traditional fluorescence microscopy. A second option is to harvest the cells and either perform a cytospin, or embed in Optimum Cutting Temperature (OCT) Medium for cryosectioning or in agarose for subsequent paraformaldehyde (PFA) fixation and embedding in paraffin. Each of these techniques will help create clearer images for characterizing the differentiated cell population. All of these options are compatible with culturing the cells on standard tissue culture-treated plasticware. When using coverslips, these too should be coated in Corning® Matrigel® prior to use.

The media comprising the STEMdiff™ Pancreatic Progenitor Kit are compatible with standard plastic micropipette tips. We recommend that the user practice sterile technique and avoid the use of antibiotics.

1. We have observed that the cells between stage 3 and stage 4 show a low adherence efficiency in coverslips compared to the plate surface (both of them are treated with hES-qualified matrigel). Could you suggest how this could be improved?

In general, culturing cells on coverslips provides a significant advantage for imaging. We have found that placing coverslips in larger well formats increases the risk that the coverslip will move during medium changes, which in turn can cause cell loss. We find that when culturing on a coverslip is preferred, reducing the size of the culture well to restrict the possible movement of the coverslip improves the ability to perform successful differentiations. In addition, extra care should be taken to avoid damaging the cell layer during aspiration and medium exchanging. Damaging the cell layer by accidental pipette contact can lead to loss of the monolayer.

2. Are the cells at the end of stage 4 are comparable to the pancreatic endoderm (PE) or pancreatic endocrine precursors (PEP) cells, according to the characterization of the Rezania et al, 2014 paper?

The cells generated using the STEMdiff™ Pancreatic Progenitor Kit are best compared to the Pancreatic Endoderm (PE) as described in Rezania et al., 2014. As highlighted in this publication, a key difference between the PE stage (Stage 4) and the Pancreatic Endocrine Precursors (PEPs; Stage 5) is the expression of NEUROD1 and the restriction of the cells towards the endocrine lineage. While we observe an upregulation of NEUROD1 in Stage 4 compared to earlier stages of differentiation, this level of upregulation is not to the level demonstrated in Stage 5 cells as shown in Rezania et al., 2014. Furthermore, through in vivo maturation studies that were performed in collaboration with Dr. Timothy Kieffer at The University of British Columbia, we found that Stage 4 cells generated using the STEMdiff™ Pancreatic Progenitor Kit were capable of maturing towards both endocrine and exocrine (including ductal) lineages, suggesting that they had not yet been restricted solely to the endocrine pancreatic lineage.

3. Could you suggest a protocol for the in vitro differentiation of the progenitor cells produced by the kit into mature pancreatic endocrine β- cells?

There are three publications in the field that have thus far demonstrated in vitro protocols to generate maturing insulin-producing beta cells from pancreatic progenitor cells. These are Rezania et al., 2014; Pagliuca et al., 2014; and Russ et al., 2015. To test whether pancreatic progenitor cells generated using the STEMdiff™ Pancreatic Progenitor Kit are capable of maturation towards functional beta cells, we cultured Stage 4 cells in the Stage 5 and Stage 6 media as described in both the Rezania et al. and Pagluica et al. publications. In both experiments, we observed increases in insulin and glucagon gene expression by qPCR (see Figure 6 here). These data suggest that either protocol is compatible with pancreatic progenitor cells generated using this kit.