Abstract

   The clinical translation of induced pluripotent stem cells (iPSCs)
   holds great potential for personalized therapeutics. However, one of
   the main obstacles is that the current workflow to generate iPSCs is
   expensive, time‐consuming, and requires standardization. A simplified
   and cost‐effective microfluidic approach is presented for reprogramming
   fibroblasts into iPSCs and their subsequent differentiation into neural
   stem cells (NSCs). This method exploits microphysiological technology,
   providing a 100‐fold reduction in reagents for reprogramming and a
   ninefold reduction in number of input cells. The iPSCs generated from
   microfluidic reprogramming of fibroblasts show upregulation of
   pluripotency markers and downregulation of fibroblast markers, on par
   with those reprogrammed in standard well‐conditions. The NSCs
   differentiated in microfluidic chips show upregulation of
   neuroectodermal markers (ZIC1, PAX6, SOX1), highlighting their
   propensity for nervous system development. Cells obtained on
   conventional well plates and microfluidic chips are compared for
   reprogramming and neural induction by bulk RNA sequencing. Pathway
   enrichment analysis of NSCs from chip showed neural stem cell
   development enrichment and boosted commitment to neural stem cell
   lineage in initial phases of neural induction, attributed to a confined
   environment in a microfluidic chip. This method provides a
   cost‐effective pipeline to reprogram and differentiate iPSCs for
   therapeutics compliant with current good manufacturing practices.

   Keywords: differentiation, iPSC, microfluidic chip, microfluidics,
   neural stem cells, reprogramming, stem cell therapy
     __________________________________________________________________

   This study highlights the development of a microfluidic platform to
   reprogram somatic cells from donors into induced pluripotent stem cells
   and further differentiate them into neural stem cells. This confined
   microfluidic platform boosts neural stem cell generation commitment at
   an early stage, as denoted by the pathway enrichment analysis.

   graphic file with name ADVS-11-2401859-g006.jpg

1. Introduction

   The revolutionary discovery of reprogramming factors by Yamanaka et al.
   to revert somatic cells to their pluripotent stem cell state opened new
   avenues, especially in cell biology, diagnostics, and therapeutics,
   including patient‐specific disease modeling, testing of novel
   therapeutic modalities, and personalized genetic and therapeutic
   approaches.^[ [38]^1 , [39]^2 , [40]^3 ^] These reprogrammed cells,
   commonly referred to as iPSCs, are capable of self‐renewal and are
   pluripotent, that is, can be differentiated into any germ layer of
   interest. Non‐integrative reprogramming can show higher reprogramming
   efficiencies while bypassing virus‐induced genomic integration, thereby
   increasing the safety of such cell‐based therapeutics compared to the
   lentiviral induction methods.^[ [41]^4 ^] The development of
   non‐integrative reprogramming methods, such as synthetic and modified
   mRNA, has increased the interest in iPSC‐derived cell therapies.^[
   [42]^5 ^]

   The clinical application and approval of such iPSC‐based therapeutics
   encounter obstacles due to the inherent variability of the cells. This
   variability hinders standardization, resulting in increased risks and
   difficulty in delivering a robust product.^[ [43]^6 ^] Most current
   pluripotent stem cell‐derived therapies are allogenic, that is, not
   derived from the patient that will be treated, and are under
   development at centralized production facilities. Large‐scale and
   robust production in these centralized facilities often results in high
   costs and a large amount of manual workload.^[ [44]^7 , [45]^8 ^] A
   critical drawback with these allogeneic therapies is that they require
   the patients to be on a long‐term immunosuppressive drug regimen,
   limiting the potential of such therapies and adversely affecting the
   patient's quality of life.^[ [46]^7 , [47]^8 ^] In addition, achieving
   large‐scale and robust production in these centralized facilities poses
   a challenge, often resulting in high costs and manual workload. The
   alternative, autologous cell therapies—involve reprogramming and
   differentiation of the patient's own cells and are in less demand of
   immunosuppressive treatment.^[ [48]^9 , [49]^10 ^] However, these are
   considered too costly and not easily accessible due to the prohibitive
   costs involved with the current reprogramming methods. To make such
   autologous cell therapies accessible and affordable, an efficient and
   standardized pipeline is necessary to enable sustainable reprogramming
   of somatic cells from a patient and further differentiate them into the
   required therapeutic cell type.

   Currently, such cell therapy products are performed using standard
   systems such as well plates and flasks, enabling easy manipulation and
   regulation of the culture microenvironment. However, these culture
   methods cannot accurately control the cellular microenvironment and
   cell‐to‐cell signaling due to their large surface area and volume.^[
   [50]^11 , [51]^12 , [52]^13 ^] On the contrary, dedicated microfluidic
   solutions and developments in microfabrication, surface chemistry, and
   molecular biology enable the precise control of the cellular
   microenvironment by reducing the culture volume, making it easier to
   handle, improving cost‐effectiveness, and having the potential for
   automation and parallelization.^[ [53]^14 ^] These can be easily
   integrated into downstream analysis and quality control methods such as
   imaging and qPCR.^[ [54]^15 ^] Previous works have reported the
   differentiation of human embryonic stem cell (hESC) derived embryoid
   bodies and highlighted that the gene and protein expression is
   regulated by biophysical parameters such as confinement, 3D
   microenvironment, and stress, which influence cellular proliferation
   and differentiation.^[ [55]^16 , [56]^17 ^] The inter‐cellular factors
   and signaling molecules are concentrated in a confined chamber, thus
   enhancing cellular uptake.^[ [57]^18 ^] Thanks to the low volume to
   surface area, microfluidic cell culture can form a cell culture
   environment that, unlike traditional cultures, allows the extracellular
   accumulation of signaling molecules and factors released in the media
   by cells undergoing reprogramming or differentiation. These
   concentrated signaling molecules act as cues for the other surrounding
   cells to initiate cellular processes like proliferation and
   differentiation. This results in a more homogenous cellular population
   as an increased concentration of endogenous signaling molecules drives
   specific cell fate decisions.^[ [58]^19 , [59]^20 ^] The initial
   protocols for microfluidic reprogramming of human somatic cells (dermal
   fibroblasts) into iPSCs were first reported and optimized by the group
   led by Elvassore in 2015.^[ [60]^21 , [61]^22 , [62]^23 ^] Further work
   in this regard also highlighted an approach for maintaining and
   inducing the three germ layers in a microfluidic platform from hESC and
   iPSC cultures but could not derive a functional ectodermal cell type.^[
   [63]^21 , [64]^24 ^] Recent works involving culturing neural cells in a
   microfluidic setting usually require cells to be cultured in
   conventional well plates before being seeded into microfluidic devices
   for drug testing and disease model experiments.^[ [65]^25 , [66]^26 ,
   [67]^27 ^] On‐chip generation of NSCs could be used in the future to
   develop a microfluidic pipeline to investigate their potential for use
   in neural cell therapies, for example, for neurodegenerative diseases
   such as Parkinson's disease, spinal cord injury, and stroke.^[ [68]^10
   , [69]^28 , [70]^29 , [71]^30 , [72]^31 , [73]^32 , [74]^33 ^]

   Here, we present a microfluidic chip platform for simplified and faster
   reprogramming of fibroblasts into iPSCs, followed by their in‐chip
   differentiation into NSCs (Figure [75]1 ). This work advances and
   improves how iPSCs can be generated in microfluidic devices and
   demonstrates the first case of microfluidic generation of functional
   ectodermal cells toward the neural lineage as neural stem cells using
   the well‐known dual‐SMAD inhibition protocol. Our protocol results in
   reprogrammed iPSC colonies in about 10 days and NSCs in another 12
   days. The proposed microfluidic pipeline includes chips that are easy
   to fabricate using PDMS and reduce the cellular input, reagent
   requirement, and manual labor, leading to substantial cost savings. The
   platform can be easily modified to enable adaptability for
   differentiation into other cell types. This is also a significant
   improvement over previously reported works wherein the expansion and
   differentiation were performed in the conventional well plate format
   using a high concentration of factors, partially negating the benefits
   of miniaturization in a microfluidic platform.^[ [76]^21 , [77]^22 ,
   [78]^23 ^] We demonstrated that the chip‐reprogrammed and
   differentiated cell population had similar mRNA expression profiles to
   those in the conventional well‐plate format. We highlight the
   differences induced in a microfluidic system compared to a conventional
   cell culture platform using bulk RNA sequencing to study their clinical
   applicability and substantiate different biological functions and
   pathways in each culture format.

Figure 1.

   Figure 1
   [79]Open in a new tab

   Overview of the study, highlighting the device fabrication,
   reprogramming of somatic cells into iPSCs, and neural induction of
   iPSCs using dual‐SMAD inhibition protocol to yield neural stem cells.
   a) Process flow for fabrication of the microfluidic device with 0.4 and
   0.6 mm high channels used for somatic cell reprogramming (R) and neural
   induction (N), respectively, resulting in differences in channel volume
   and total volume indicated in table. b) Overview of the reprogramming
   process of somatic cells into iPSCs on both microfluidic devices and
   well plates using mRNA transfection. c) Overview of the neural
   induction process of iPSCs into neural stem cells on both microfluidic
   device and well plate using the dual‐SMAD inhibition protocol.

2. Results and Discussion

2.1. Design and Fabrication of the Microfluidic Device

   We developed a microfluidic polydimethylsiloxane (PDMS) device platform
   with dimensions of 75 mm x 25 mm, matching the footprint of a standard
   microscope glass slide (Figure [80]1a; see Experimental Section for
   detailed fabrication process). Each microfluidic device consists of six
   channels with a height of 0.4 and 0.6 mm for reprogramming and neural
   differentiation, respectively. Each channel had a surface area of 45
   mm^2. The dimensions of the channels, the surface area, and the volume
   are described in Table [81]S1 (Supporting Information). This platform
   was customizable and adaptable depending on cell and process‐specific
   requirements, as defined below. These devices had a substantially
   higher surface area (45 mm^2) to volume (18 or 27 µL) ratio as compared
   to the conventional 6‐well (960 mm^2 and 2000 µL) and 12‐well plates
   (350 mm^2 and 1000 µL) used for cellular reprogramming and
   differentiation, respectively. Therefore, we achieved a 100‐ and
   16‐fold reduction in reagent volumes used for reprogramming (channel
   height 0.4 mm) and differentiation (channel height 0.6 mm) compared to
   6‐ and 12‐well plates. Compared to previous reports that used
   photolithography for pattern‐definition of the master molds, our mold
   designs were initially prototyped using a fast commercial 3D printer to
   iterate design specifications, followed by industry‐standard
   CNC‐milling to implement an adaptable and scalable chip production
   method (see Experimental Section).^[ [82]^22 , [83]^23 ^] The same
   articles reported an approximately 60‐fold reduction in mRNA and
   reprogramming reagents needed for reprogramming. The devices also
   permit easy media exchange (18 µL/27 µL volume to replace all media).
   Our design facilitated pipetting to chip interfaces and minimized
   flow‐induced shear forces. The simple interface allows for further
   automation, potentially as a closed system that can be used to develop
   personalized cell‐based therapeutics.

2.2. Microfluidic Reprogramming of Somatic Fibroblast Cells

   To test our microfluidic platform initially, we used neonatal
   fibroblast cells and transfected them for four days. Upon colony growth
   and identification (Figure [84]2a), iPSCs were removed by Versene
   treatment and expanded for downstream analysis. Compared to manual
   colony picking, this novel isolation method for iPSC colonies does not
   require manual cutting of the channels. Hence, unlike the manual
   colony‐picking method, this Versene‐based method could be implemented
   in automated settings. Since iPSCs are known to be less adherent than
   fibroblasts, the use of an enzyme‐free method selectively dissociates
   iPSCs from a mixed population of cells, effectively removing the need
   for a trained professional to perform the clonal selection.^[ [85]^34
   ^] Isolated cells were passaged upon reaching 80% confluency on LN‐521
   coated plates with the addition of Y‐27632. The cells formed colonies
   upon removing Y‐27632 from the media, a well‐known behavior while
   culturing stem cells in‐vitro.^[ [86]^35 ^] The iPSCs were expanded in
   the iPSC media during subsequent passaging. The iPSC lines established
   using this method are a non‐clonal expansion of a bulk culture.

Figure 2.

   Figure 2
   [87]Open in a new tab

   Reprogramming of fibroblasts into induced pluripotent stem cells. a)
   Illustrative protocol employed for the reprogramming of fibroblasts and
   bulk expansion of isolated colonies. Scale bar 50 µm. b) qPCR analysis
   of the reprogrammed iPSCs shows downregulation of mesenchymal markers
   and upregulation of pluripotency markers in both well and chip formats.
   c) Flow cytometry analysis for standard intracellular pluripotency
   markers OCT4 and SOX2 and extracellular pluripotency markers SSEA4 and
   TRA‐1‐60 in iPSCs reprogrammed in wells (black) or chip (red
   intracellular, blue extracellular). Lines delineate quadrants (Q1–Q4)
   of negative and positive signals for the two fluorophores, and the
   numbers in Q2 indicate the percentage of cells that are double positive
   for both markers. Data generated after single cell expansion. The
   gating strategy is shown in Figure [88]S4 (Supporting Information).
   Panel (c) used the following cell lines KTH‐04: well p 19, chip p 20.
   KTH‐05: well p 14, chip p14. KTH‐06: well p 14, chip p 14.

   Adult fibroblast cells are an ideal cell type to test the capabilities
   of microfluidic platforms, as it is challenging to culture cells in
   confined environments.^[ [89]^22 ^] These are ideal candidates as
   primary fibroblasts are easily available and minimally invasive sources
   of cells in a practical scenario.^[ [90]^36 , [91]^37 ^] Furthermore,
   the easy commercial availability of such cells makes them ideal cell
   sources for prototyping. To further assess our microfluidic platform,
   we performed side‐by‐side reprogramming using mRNA in a conventional
   well plate and microfluidic platforms using 3 different donors of
   fibroblast cells that varied in age, sex, and pigmentation (Table
   [92]S2, Supporting Information). We achieved a 100% success rate in
   generating colonies in both well and chip culture conditions. In our
   protocol, following the transfection, more colonies appeared in the
   well conditions for two of the fibroblast cell lines used, the neonatal
   (KTH‐04) and one adult (KTH‐05), whereas the oldest fibroblast cell
   line had a low number of colonies for analysis (KTH‐06) in any of the
   culture conditions, which is attributed to the phenotypic, genetic and
   proliferative differences of the starting fibroblasts.^[ [93]^38 ,
   [94]^39 , [95]^40 , [96]^41 ^] Specifically, neonatal fibroblast cells
   had 38 colonies/10 mm^2 on average in the well conditions, whereas chip
   conditions had 20 colonies/10 mm^2 on average (Figure [97]S1,
   Supporting Information; well vs chip, p<0.0001). Adult fibroblast cells
   had, on average, 4 times fewer colonies compared to neonatal, 10
   colonies/10 mm^2. As with the neonatal ones, adult fibroblast cells
   reprogrammed in chips resulted in fewer colonies, 4 colonies/10 mm^2
   (Figure [98]S1, Supporting Information; well vs chip, p = 0.0226).
   Notably, due to a lower surface area, the chips require ninefold fewer
   cells than the seeding in wells, even when similar seeding densities
   are used. Furthermore, the downscaling in the microfluidic chip leads
   to a 100‐fold reduction in reagents needed. The cells seeded in the
   well and chip were imaged daily to ensure they looked healthy and
   showed no signs of cell death.

   After extraction and expansion, we further evaluated if there were any
   differences in expression levels of fibroblast markers and markers for
   undifferentiated stem cells related to pluripotency between the two
   culture conditions. We verified that the levels of the pluripotency
   markers were not significantly different between well and chip
   conditions (Figure [99]2b). The cells were characterized by qPCR at
   passages 6–8 to analyze the gene expression of reprogrammed cells
   relative to the parental fibroblast cells used for each specific
   reprogramming run performed in both well and chip formats in parallel.
   We validated the shift in the cellular state by a decrease in the
   relative expression of fibroblast‐specific genes, namely COL1A1, VIM,
   and CD44 (Figure [100]2b). In contrast, we observed significantly
   enriched pluripotency‐associated genes SOX2, NANOG, and OCT4
   (Figure [101]2b) compared to the starting cell lines. Monitoring the
   undifferentiated and pluripotent status of iPSCs is essential for
   quality controlling the reprogramming process and the iPSC lines.
   Additionally, for preserving experiment quality and reproducibility,
   flow cytometry was used to quality control iPSC lines pluripotency
   status at later passages (KTH‐04: well p 19, chip p 20. KTH‐05 and
   KTH‐06: well p 14, chip p 14) All iPSC lines showed high expressions of
   the well‐defined and commonly used markers for pluripotency OCT4, SOX2,
   SSEA4, and TRA‐1‐60 (Figure [102]2c; Figures [103]S2 and [104]S3,
   Supporting Information).^[ [105]^42 ^] OCT4 and SOX2, together with
   NANOG, belong to the pluripotency core regulatory network driving the
   identity of iPSCs.^[ [106]^43 ^] SSEA4 and TRA‐1‐60 are the two key
   extracellular pluripotent stem cell markers for iPSCs.^[ [107]^44 ,
   [108]^45 ^] Expression profiles are consistent with
   pluripotency‐associated protein expression levels in established clonal
   iPSC lines obtained from the iPS Core at Karolinska Institute: Control
   7‐II, Control 10‐V, and Control 14‐II, which were analyzed using the
   same multicolor flow cytometry panel and instrument settings (Figure
   [109]S4, Supporting Information).In addition, flow cytometry analysis
   of intracellular and extracellular markers confirms that well and chip
   conditions have very similar pluripotency‐related expression, with
   91.7%–98.8% cells double positive for OCT4 and SOX2 and 93.9%–98.0%
   double positive for SSEA4 and TRA‐1‐60 (Figure [110]S5, Supporting
   Information).

   To further assess the comparability of our generated iPSCs, we ventured
   into bulk RNAseq analysis to explore differences on a global scale. To
   compare the generated lines with known standard, we included three iPSC
   lines obtained from the iPS Core at Karolinska Institutet as a
   well‐established reference data set (Table [111]S4, Supporting
   Information).^[ [112]^46 , [113]^47 ^] The bulk RNA sequencing data was
   analyzed using the protocol listed in the Experimental Methods section.
   We assessed the quality of our generated lines (reprogrammed through
   conventional wells and microfluidic chips, KTH‐04, ‐05, and ‐06) by
   comparing them to the reference lines (Table [114]S4, Supporting
   Information). We observed that each cell line clusters close to each
   other and the reference lines (Figure [115]4 ). Further strengthening
   the validation of our proposal system, Principal Component Analysis
   (PCA) revealed that the reprogrammed cells, both in well and chip
   format arising from the same fibroblast cell line, cluster together
   (Figure [116]S6, Supporting Information), showing no significant
   differences between the iPSCs derived from the same cell source. There
   is, however, some difference between the iPSCs derived from different
   cell sources, which is a known effect arising due to the genetic
   variability, epigenetic factors, and the choice of the reprogramming
   method used.^[ [117]^39 , [118]^48 , [119]^49 ^] Although there are
   minor differences in the PCA plot, the biological variability between
   the reference lines is much higher than the differences between well
   and chip conditions per reprogrammed iPSC line.^[ [120]^19 , [121]^39
   ^]

Figure 4.

   Figure 4
   [122]Open in a new tab

   PCA of NSCs obtained by differentiation of iPSCs in microfluidic and
   conventional well plate format. NES cells are used as references. There