GelMA hydrogel as a scaffold to enhance the survival and differentiation of human induced lateral ganglionic eminence precursor cells

Cell reprogramming holds enormous potential to revolutionize our understanding of neurological and neuro-developmental disorders, as well as enhance drug discovery and regenerative medicine. We have developed a direct cell reprogramming technology that allows us to generate lineage-specific neural cells. To extend our technology, we have investigated the incorporation of directly reprogrammed human lateral ganglionic eminence precursor cells (hiLGEPs) in a 3-dimensional (3D) matrix. Hydrogels are one of the most promising bio-scaffolds for 3D cell culture, providing cells with a supportive environment to adhere, proliferate, and differentiate. In particular, gelatin methacryloyl (GelMA) hydrogels have been used for a variety of 3D biomedical applications due to their biocompatibility, enzymatic cleavage, cell adhesion and tunable physical characteristics. This study therefore investigated the effect of GelMA hydrogel encapsulation on the survival and differentiation of hiLGEPs, both in vitro and following ex vivo transplantation into a quinolinic acid (QA) lesion rat organotypic slice culture model. We demonstrate, for the first time, that the encapsulation of hiLGEPs in GelMA hydrogel significantly enhances the survival and generation of DARPP32 + striatal neurons both in vitro and following ex vivo transplant. Furthermore, GelMA-encapsulated hiLGEPs were predominantly located away from the reactive astrocyte network that forms following QA lesioning, suggesting GelMA provides a protective barrier for cells in regions of inflammatory activation. Overall, these results indicate that GelMA hydrogel has the potential to act as a 3D bio-scaffold to augment the viability and differentiation of hiLGEPs for research and translation of pharmaceutical development and regenerative medicine.


Introduction
The success of clinical translation of new therapies for the treatment of neurological disorders requires in vitro models to closely recapitulate in vivo biology and microenvironmental factors.It is well established that culturing cells in three-dimensional (3D) systems that mimic key factors of tissue, in particular the extracellular matrix (ECM), is more representative of the in vivo environment than simple two-dimensional (2D) cultures (Langhans, 2018).The emergence of bioactive materials has led to the development of bio-scaffolds that can mimic the native ECM environment and enhance cell adhesion, proliferation, migration, and differentiation.Hydrogels are one of the most promising bio-scaffolds for 3D cell culture.Hydrogels are water-swollen, hydrophilic polymers formed by physical and chemical crosslinking methods, which determine their mechanical and biomedical properties (Caliari and Burdick, 2016).Gelatin methacryloyl (GelMA) hydrogels have been used for a variety of 3D biomedical applications due to their biocompatibility, enzymatic cleavage, cell adhesion and tunable physical characteristics (Pepelanova et al., 2018;Yue et al., 2015).Gelatin has thermo-reversible cross-linking properties and retains arginine-glycine-aspartic acid peptide sequences which provide an aqueous microenvironment for cells and supports cell behaviours (Van Den Bulcke et al., 2000).The combination of UV light and a UV light-sensitive photo-initiator, Irgacure® 2959, can be used to initiate crosslinking of GelMA hydrogel due to the resultant low cytotoxicity (Lim et al., 2016).While Irgacure® 2959 has been used with UV light at a longer wavelength (365 nm) to minimize chromosomal aberrations and genetic instability in mammalian cells, this results in poor photo-crosslinking efficiency (Elkhoury et al., 2023;Lee et al., 2020).Therefore, selecting photo-initiators that can absorb in the visible-light range has been required for mammalian cell culture.Lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) is an alternative photo-initiator that allows photo-crosslinking in the visible light spectrum of 405 nm and enhances cell viability (Lee et al., 2020).Lim et al. (2016) developed and optimized a light-initiated polymerisation system containing two Vis-sensitive photo-initiators, ruthenium (Ru) and sodium persulfate (SPS), which enhanced cell viability and cytocompatibility compared to the conventional UV+ Irgacure® 2959 system (Lim et al., 2016).
In addition to bio-scaffolds offering a 3D environment more representative of the in vivo setting, it is crucial that in vitro models of neurodegenerative diseases recapitulate the pathological properties of the human disease including molecular age-associated features.We have previously established a direct cell reprogramming technology that enables us to generate neural precursor cells directly from adult human dermal fibroblasts (aHDFs) without first requiring the generation of pluripotent stem cells (Maucksch et al., 2012;Playne et al., 2018;Edwards et al., 2023;Connor et al., 2018;McCaughey-Chapman et al., 2023).Subsequent differentiation of our direct-to-neural precursor cells allows for the generation of a range of specific neural lineages with maintenance of age-associated signatures, critical for disease modelling (Maucksch et al., 2012;Playne et al., 2018;Edwards et al., 2023;Connor et al., 2018;McCaughey-Chapman et al., 2023).The concept of cell reprogramming highlights the role of the transcriptome in determining cell identity and lineages in which the fate of a specific cell type can be controlled by changing the transcriptome profile (Fang et al., 2018).Although previous studies have shown the potential of reprogrammed cells for disease modelling and cell replacement therapy, translating in vitro findings to the clinic remains a challenge due to limitations in biological reproducibility as well as reduced cell survival, differentiation and functional integration following 2-dimesional (2D) culture (Benchoua et al., 2021).Therefore, the provision of a supportive 3D matrix is crucial to ensure recapitulation of appropriate tissue structure and microenvironmental factors for human neural precursor cells both in vitro and following transplantation into the brain.Therefore, this study aims to investigate the effect of encapsulating human induced lateral ganglionic eminence precursor (hiLGEPs) cells in GelMA both in vitro and prior to transplantation into an ex vivo quinolinic acid (QA)-lesioned rat organotypic slice culture system.We have previously shown that hiLGEPs differentiate to functionally mature DARPP32+ striatal neurons following 2D culture (Connor et al., 2018).We propose the culture of hiLGEPs in a supportive 3D GelMA matrix will increase both cell survival and striatal differentiation when compared to 2D cultured cells.An overview of the in vitro and ex vivo protocols are shown in Fig. 1A and B, respectively.
Adult human dermal fibroblasts obtained from three individual donors [Cell Applications Inc] were transfected with chemically modified mRNA (cmRNA) encoding SOX2 and PAX6 to induce reprogramming to a neural fate and cultured in reprogramming medium as previously described (Connor et al., 2018;Monk et al., 2021).During the first week of reprogramming, the cells transformed from a flat elongated shape to a more rounded form (D0 & D3, Fig. 1C).After the first passage of reprogramming (P1 / Day 7, Fig. 1A&B), the cells aggregated into semi-adherent colonies (D10; Fig. 1C).By day 14 of reprogramming, most of the cells had adhered to the culture plate and displayed a polygonal shape (D14, Fig. 1C).
On Day 14 of reprogramming, hiLGEPs were collected and seeded onto either glass coverslips at a density of 320,000 cells/mL for 2D culture or encapsulated in GelMA hydrogel at a density of 5×10 6 cells/ mL (Fig. 1A&B).For 2D in vitro differentiation, glass coverslips were coated with 350μL GelTrex™ [Gibco], incubated for 1-hour at 37 • C followed by 30 minutes at room temperature.GelTrex was removed from the wells and 80,000 cells were immediately seeded onto each coverslip and cultured in differentiation medium in a 24-well plate.For 3D in vitro differentiation, GelMA hydrogel preparation involved dissolving GelMA macromer in sterile PBS by heating to 37 • C to make a 10 wt% GelMA solution.Two photoinitators, Ru and SPS, were also dissolved in PBS, then were further diluted with GelMA macromer and sterilised PBS, giving the final concentration of 5% GelMA, 0.5 mM Ru and 5 mM SPS. 150,000 hiLGEPs were resuspended in 30μL GelMA then dispensed into a sterilised silicone mould (6 mm width x 1 mm height) that was pressed down directly in a 24-well plate.GelMA solution was polymerized by exposure to a LED light source for 10 minutes before removing the mould, forming a 1 mm thick by 6 mm wide spherical GelMA construct.2D and GelMA encapsulated cells were cultured in 0.250 mL and 1 mL differentiation medium per well, respectively, as previously described (Connor et al., 2018;Monk et al., 2021).Cells were cultured at 37⁰C, 5% CO 2 and fed every second day for 14 days.
All three cell lines, derived from three individual donors, cultured on 2D coverslips exhibited a similar morphology throughout the differentiation period.Short neurites with distinct cell soma were detected at Day 3 of differentiation, with neurites extending gradually over two weeks of differentiation (Fig. 1D).All lines displayed multipolar processes by Day 10 of differentiation (Fig. 1D).By Day 14, all cell lines cultured in 2D had developed complex neuronal-like morphologies with multiple neurites and extensive branching from the cell soma (Fig. 1D).Similar morphologies were also observed for all three lines differentiated in GelMA hydrogel (Fig. 1D).GelMA-encapsulated hiLGEPs derived from all three cell lines developed processes by Day 3. Cells began to develop a basic network of neurites by Day 5.As differentiation continued, cells further matured with more complex neurite extensions and branching from the cell soma.The neurites were also seen to project across different z-planes of the hydrogel.At Day 14 of differentiation, a dense network of elongated processes spanning across the z-planes of the hydrogel was apparent in all cell lines (Fig. 1D).The colourization of the GelMA brightfield images is due to the presence of Ru/SPS (orange colour) in the GelMA hydrogel (Fig. 1D).
Cell viability was assessed using a live/dead assay and compared between hiLGEPs differentiation on glass coverslips (2D) and hiLGEPs encapsulated in GelMA hydrogel (3D).Cells were incubated with assay solution containing 2μM calcein AM [Invitrogen] and 4μM ethidium homodimer [EthD-1; Invitrogen] for one hour at room temperature.Green fluorescent calcein was used to indicate live cells and red fluorescent EthD-1 was used to indicate dead cells.We observed a small number of encapsulated hiLGEPs underwent cell death (EthD-1) at the edge of the hydrogel while live cells (Calcein AM) were visible towards the centre of the hydrogel (Fig. 1E).Ten random fields of view were imaged per coverslip or gel with three wells or gels (n=3) imaged in total (n=30 fields of view per cell line with n=3 independent lines).The number of Calcein + live cells and EthD-1 + dead cells was quantified from images taken at 10X magnification using a macro in ImageJ with appropriate threshold and particle size.The average percentage of viable cells across all three cell lines was determined as 82.11% ± 3.9% for GelMA encapsulated cells and 50.91% ± 4.09% for 2D-cultured cells.An independent samples T-test demonstrated a significant increase in viability for hiLGEPs encapsulated in GelMA hydrogel compared to hiLGEPs cultured in 2D (p=0.014) (Fig. 1E&F).
To characterise cell phenotypes, differentiated hiLGEPs were stained with TUJ1 [Abcam; 1:500] and co-labelled with either the medium spiny striatal neuron marker DARPP32 [Creative Diagnostic; 1:500] or the astrocytic marker S100β [Abcam; 1:250].Neurites were visible for cells cultured on both 2D coverslips and in GelMA hydrogel.While 2Dcultured cells exhibited restricted neurite extension, cells encapsulated in hydrogel developed elongated neurite projections across the z-planes and demonstrated extension to other cells in different planes.Representative fluorescent images of differentiated hiLGEPs expressing TUJ1 and DARPP32 in 2D culture and following encapsulation in GelMA hydrogel are shown in Fig. 2A.Differentiated 2D hiLGEP cultures and GelMA-encapsulated cells were imaged using a fluorescent inverted Nikon TE2000E microscope and Zeiss LSM 710 confocal microscope, respectively.Eight to 10 randomly selected fields of view were captured from each coverslip (42.5 ×42.5 cm 2 ) and each gel (91 ×91 cm 2 per z-stack), and the images were randomised before processing using ImageJ.Differentiation of (caption on next page) L. Nguyen et al.
In the differentiated hiLGEP cultures, we observed a population of TUJ1 -/DAPI + cells.We investigated the presence of astrocytes by staining for the astrocytic marker, S100β.Fluorescent images showed that S100β + astrocytes were present in both 2D and 3D GelMA hydrogel cultures.Interestingly, in both 2D and 3D-differentiated hiLGEPs, a subpopulation of S100β + /TUJ1 + co-labelled cells were seen (Fig. 2C) (Connor et al., 2018).Co-labelling of neuronal and astrocytic markers has previously been reported (Park et al., 2020) and may represent a population of radial glia or asterons.
To determine the ability of GelMA hydrogel to enhance hiLGEP survival and differentiation following transplantation, hiLGEPs, either encapsulated at a density of 90,000 cells in 10 μL GelMA or resuspended in differentiation medium (unencapsulation), were transplanted into the striatum of QA-lesioned rat organotypic slice cultures using a Hamilton syringe [26 s-gauge needle].The organotypic brain slice cultures were generated as described previously (McCaughey-Chapman and Connor, 2022).Briefly, the organotypic slices were derived from 8-10-day old male Sprague Dawley rats.Sagittal slices (300μm-thick) were cut using a vibratome at high frequency, then individual slices were mounted onto membrane inserts and cultured at the air-membrane interface in a 6-well plate [Thermo Fisher] in 1 mL of Minimum Essential Medium (MEM) containing 25% horse serum and 1% PS [Thermo Fisher] added below the membrane insert.Mitotic inhibitors, 4.4 mM uridine [Sigma], 4.4 mM 5-fluorodeoxyuridine [Sigma] and 4.4 mM cytosineβ-arabinofuranoside [Sigma], were added to the medium for the first 72 hours of slice culture to minimise glial scar formation.At Day 3 of culturing, the medium was switched to medium consisting of 2/3 horse serum medium and 1/3 serum free medium.Serum-free medium was made up of Advanced DMEM/F12 [Thermo Fisher], 2% B27 supplement [Thermo Fisher] and 1% N-2 supplement [Thermo Fisher] for 24 hours.For the lesioned model, slices were treated on Day 3 with 50μM QA diluted in culture medium for 24 hours then switched to fresh medium.At Day 5, medium was replaced with 1/3 horse serum medium and 2/3 serum free medium.Two days later, slices were provided with differentiation medium, and medium changes occurred every two days for a week before transplantation.Slices were incubated at 35 • C and 5% CO 2 .The human-specific biomarker, Ku80 [Cell Signalling; 1:400] was used to identify the transplanted hiLGEPs from the host rat tissue.Neural differentiation of the hiLGEPs ex vivo was assessed by quantification of the number of Ku80+ cells co-expressing the striatal neuron marker DARPP32 and the astrocyte marker GFAP.Antibody visualisation was achieved using the iDISCO tissue clearing method (Renier et al., 2014;McCaughey-Chapman and Connor, 2017).
GFAP staining was also observed in the QA-lesioned slices, the majority of which appeared to be rat-derived (Fig. 2G).Transplantation of both unencapsulated and encapsulated hiLGEPs gave rise to a population of GFAP + / KU80 + cells which were surrounded by GFAP + / Ku80 - rat astrocytes.We observed the number of GFAP + / Ku80 + cells in slices with GelMA-encapsulated cells (24.31% ±2.32%) was significantly reduced when compared to slices which received transplants of unencapsulated hiLGEPs (36.55% ±3.8%) (p =0.035; Fig. 2H).It is also noteworthy that in the slices transplanted with unencapsulated hiLGEPs, Ku80 + human cells appeared to be distributed within the astrocyte network, while the transplanted cells encapsulated with GelMA tended to be positioned away from the network (Fig. 2G).
This study incorporated our direct-to-hiLGEP cmRNA reprogramming protocol with GelMA hydrogel to characterise, for the first time, the survival and differentiation of hiLGEPs-derived neurons in a 3D matrix.Following reprogramming, hiLGEPs were differentiated for 14 days either on glass coverslips (2D) or encapsulated in GelMA hydrogel (3D).A live/dead cell assay revealed that GelMA hydrogel significantly increased cell survival compared to 2D culture with > 80% cell viability detected in GelMA encapsulated cultures.Interestingly, most of the dead cells were located around the edge of the hydrogel.This suggests that the GelMA hydrogel may not only provide a suitable microenvironment for cell survival but may also act as a barrier protecting the cells from potential hostile extrinsic factors.
While the TUJ1 neuronal yield was comparable between GelMA and 2D hiLGEP cultures, the yield of DARPP32 + cells was significantly greater in GelMA encapsulated cultures compared to 2D, suggesting the 3D GelMA microenvironment promotes lineage specification of TUJ1 + neurons to DARPP32 + medium spiny striatal neurons.Morphologically, TUJ1 + neurons differentiated in GelMA hydrogel displayed elongated neurite extensions throughout multiple z-planes indicating GelMA may augment the extension of neurites beyond the hydrogel and promote cell integration into host tissue in a transplant setting.
There is only one previous study that has investigated the in vitro generation of medium spiny striatal neurons using a 3D hydrogel system (Adil et al., 2018).Human pluripotent stem cell-derived lateral ganglionic eminence progenitor cells were cultured for 26 days on either Matrigel-coated polyesterine or in a scalable PNIPAAm-PEG hydrogel (Adil et al., 2018).The study showed cells encapsulated in PNIPAAm-PEG hydrogel generated a higher yield of DARPP32+ striatal neurons compared to 2D culture.At day 60 of culture 78% of the total cells expressed the neuronal marker MAP2, and of these, 55% expressed DARPP32 (Adil et al., 2018).With our protocol, ~80% of DAPI + cells expressed TUJ1, and ~90% of these generated DARPP32 + cells within 14 days of culturing.Based on the results from our in vitro studies, we investigated the ability for GelMA hydrogel encapsulated cells to survive and differentiate following transplantation into QA lesioned rat sagittal brain organotypic slice cultures.This culture system represents the pathology of the neurodegenerative disorder Huntington's disease (HD) (McCaughey-Chapman and Connor, 2022) and offers the unique opportunity to assess the viability and differentiation potential of cell replacement strategies prior to large in vivo animal studies.
We observed that hiLGEPs generated DARPP32 + neurons and GFAP + astrocytes following transplantation into QA-lesioned brain organotypic slice cultures.Most importantly, transplantation of hiLGEPs encapsulated in GelMA significantly increased the yield of DARPP32 + neurons co-labelled with the human nuclei marker Ku80.This indicates GelMA hydrogel supported the survival of hiLGEPs following transplantation and provided a suitable microenvironment to enabled transplanted cells to differentiate to medium spiny neurons in the QA-lesioned striatum.
A network of GFAP + / Ku80 -astrocytes was observed in the organotypic slice cultures in response to QA lesioning, which recapitulates the pathology of HD (McCaughey-Chapman and Connor, 2022).In the slices transplanted with unencapsulated hiLGEPs, Ku80 + cells were found to locate within the astrocyte network and a population of Ku80 + /GFAP + cells were observed.In contrast, GelMA encapsulated hiLGEPs gave rise to a population of Ku80 + cells which resided predominantly outside the endogenous GFAP + network with fewer GFAP + /Ku80 + cells generated compared to the unencapsulated hiL-GEPs.This further supports the potential use of GelMA hydrogel as a platform by which to enhance transplanted cell survival and differentiation, as well as provide protection from an inflammatory environment.
One of the major limitations of this study is that the source of the gelatin to prepare the GelMA hydrogel is of porcine origin.It also contains a small concentration of collagen type I, which is found mostly in dermal layers, whereas collagen type IV is the predominant collagen in the brain ECM (Lim et al., 2016).In conclusion, this study indicates that GelMA hydrogel provides a viable 3D bio-scaffold to promote the survival, differentiation, and functional integration of human neural precursor cells for future drug development and regenerative medicine.

Declaration of Competing Interest
Prof Bronwen Connor and Dr Amy McCaughey-Chapman have patent pending to the University of Auckland.There are no additional relationships or activities to disclose.

Fig. 2 .
Fig. 2. Encapsulation of reprogrammed hiLGEPs in GelMA enhances DARPP32 þ neuron yield in vitro and ex vivo following transplantation into QA-lesioned rat organotypic brain slices.(A) Expression of TUJ1 and DARPP32 in differentiated hiLGEPs cultured in vitro either on glass coverslips (2D) or encapsulated in GelMA hydrogel.Arrows indicate TUJ1 and DARPP32 co-labelling.Scale: 100 µm.Graphs demonstrating the average percentage of (B) TUJ1 + and DARPP32 + cells out of DAPI and (D) DARPP32 + co-labelled with TUJ1 + cells in differentiated hiLGEPs cultured either on 2D coverslips or encapsulated in GelMA hydrogel.Data represent mean ± SEM, n=3.Statistical significance was determined by an independent samples T-test with * for p<0.05.(C) TUJ1 and S100β expression in differentiated hiLGEPs cultured either on 2D coverslips or encapsulated in a GelMA hydrogel.White arrows and arrowheads indicate TUJ1 + and S100β + cells, respectively.Orange arrows designate TUJ1/ S100β co-labelling.Scale: 100 µm.The expression of (E) Ku80 and DARPP32 and (G) Ku80 and GFAP in differentiated hiLGEPs transplanted either unencapsulated or encapsulated in GelMA hydrogel into the striatum of ex vivo organotypic slice cultures after 14 days of differentiation.Arrows indicate Ku80/ DARPP32 co-labelling in (E) and Ku80/ GFAP co-labelling in (G).Scale: 400 µm.Graphs demonstrating the average percentage of DARPP32 + / Ku80 + cells in (F) and GFAP + /Ku80 + cells in (H).Data represents mean ± SEM, n=3.Statistical significance was determined by an independent samples T-test with * for p<0.05.