Microcarrier expansion of c‐MycERTAM‐modified human olfactory mucosa cells for neural regeneration

Human olfactory mucosa cells (hOMCs) have potential as a regenerative therapy for spinal cord injury. In our earlier work, we derived PA5 cells, a polyclonal population that retains functional attributes of primary human OMCs. Microcarrier suspension culture is an alternative to planar two‐dimensinal culture to produce cells in quantities that can meet the needs of clinical development. This study aimed to screen the effects of 10 microcarriers on PA5 hOMCs yield and phenotype. Studies performed in well plates led to a 2.9‐fold higher cell yield on plastic compared to plastic plus microcarriers with upregulation of neural markers β‐III tubulin and nestin for both conditions. Microcarrier suspension culture resulted in concentrations of 1.4 × 105 cells/ml and 4.9 × 104 cells/ml for plastic and plastic plus, respectively, after 7 days. p75NTR transcript was significantly upregulated for PA5 hOMCs grown on Plastic Plus compared to Plastic. Furthermore, coculture of PA5 hOMCs grown on Plastic Plus with a neuronal cell line (NG108‐15) led to increased neurite outgrowth. This study shows successful expansion of PA5 cells using suspension culture on microcarriers, and it reveals competing effects of microcarriers on cell expansion versus functional attributes, showing that designing scalable bioprocesses should not only be driven by cell yields.

Reported clinical trials have used an autologous approach that can lead to variable outcomes. Since hOMCs populations are highly variable between patients, these are challenging to expand and there is a lack of consistency between protocols used for tissue biopsy and preparation of cells for transplant (Féron et al., 1999). Therefore, it would be beneficial to develop an allogeneic or universal "off-theshelf" approach. We previously reported the generation of a candidate cell line from late-adherent hOMCs by genetic modification of primary cells with c-MycER TAM conditional immortalization technology, to advance a potential allogeneic therapeutic product for the treatment of SCI-denominated PA5 hOMCs (Santiago-Toledo et al., 2019). It is yet unclear whether a unique cell type within a hOMCs population is responsible for neural regeneration, or whether there is benefit in transplanting several types of cells within the population (Anna et al., 2017;Reshamwala et al., 2019). Clones from polyclonal populations such as PA5 hOMCs can be further derived, expanded, banked, and screened to generate an allogeneic cell therapy product for the treatment of SCI. The generation of a conditionally-immortalized hOMC population, such as PA5 hOMCs, enables a potentially extended life span, allowing the application of a cell-banking modelbased manufacturing process, necessary for allogeneic cell therapy.
The translation of such a therapeutic product to the market would require the development of a scalable bioprocess, able to yield large amounts of cells which can reach doses up to 1 × 10 7 cells/dose (Casarosa et al., 2014). Commonly, adherent or anchoragedependent cells have been grown in two-dimensional platforms such as tissue-culture flasks, cell factories, and automated systems. Even though these systems are reliable due to their widespread use in the industry, they are not easily scalable for producing large quantities of allogeneic cells, making them unsuitable for commercial stages (Brandenberger et al., 2011;Simaria et al., 2014). Among available technologies, microcarriers are a suitable alternative to perform expansion of adherent cells in a scalable manner in stirred tank bioreactors and offer a higher larger surface per unit volume of bioreactor. It may be feasible to manufacture early-stage clinical products using planar systems and then transition to stirred tank bioreactors for late-stage clinical trials and commercialization, yet this would require significant validation work to meet regulatory approval, and risk product specification failure later in development due to manufacturing process changes. Therefore, early adoption of market scale manufacturing technology is necessary.
Stirred tank bioreactors are well-explored systems that enable the monitoring and control of several culture parameters such as oxygen tension, pH, and agitation regimes. These platforms also offer a closed bioprocess environment with reduced operator interference and variability. Therefore, these are platforms that offer more robust and reproducible bioprocesses that theoretically can deliver consistent quality products, compliant with good manufacturing practice (GMP) and good clinical practice requirements.
In this study, we report the microcarrier expansion of a candidate cell population for the treatment of SCI, PA5 hOMCs, on microcarrier stirred culture using spinner flasks, with the aim to identify a microcarrier type that maximizes the expansion of neuroprotective cell types such as OECs, NSCs, and MSCs, which can lead to increased potency. First, a variety of commercially available microcarriers were selected based on a screening methodology using ultralow attachment 96-well plates. Besides the suitability of the microcarrier, considerations of compliance for GMPs and adaptability to xeno-free conditions were considered for microcarriers selection. Then, cells were grown on two types of microcarriers using spinner flasks, based on protocols previously reported for the expansion of hMSCs (dos Santos et al., 2011). To further understand growth kinetics, PA5 hOMCs were grown on microcarriers for 7 days. Cell phenotype was assessed through immunocytochemistry and real-time quantitative polymerase chain reaction (RT-qPCR). Potency of PA5 hOMCs was assessed through the capacity of the harvested cells to promote neurite outgrowth using a coculture assay of hOMCs and NG108-15 neurons.
We report the use of a microcarrier-based spinner flasks system for the expansion of the PA5 hOMCs population of cells, a candidate cell line for the treatment of spinal cord-injury.

| Microcarrier screening in well plates
Ten commercially available microcarriers were tested for expansion of PA5 hOMCs in ultralow attachment 96-well plates: Plastic, Plastic Laminin Coated (Plastic L), Plastic Plus, PronectinF®, Collagen, FACT III, Star-Plus, Hillex II, all from Pall. Synthemax II® (low and high concentration; Corning) and Cytodex I™ (GE Healthcare) were also used. Microcarriers specifications are shown in Table 1. Sterilization of Pall microcarriers and Cytodex I microcarriers was achieved by resuspension in distilled water (Gibco) and phosphate-buffered saline (PBS) Ca+ and Mg+ free (Lonza), respectively, followed by autoclaving at 121°C for 15 min. Synthemax II microcarriers were sterilized under a UV lamp for 90 min.
A monolayer of microcarriers with a total surface area of 1.28 cm 2 was used per well, to cover the bottom of a 96-well plate well. For six-well plates, a monolayer of microcarriers with a total surface area of 38 cm 2 was used per well. Total working volumes were of 200 µl and 6 ml for 96-well plates and six-well plates, respectively. Plastic L microcarriers were incubated with a solution of 20 µg/ml Cultrex® Mouse Laminin I (Trevigen) for 1 h in a standard incubator before cell seeding. All microcarriers were let to equilibrate for 1 h in a standard incubator with complete medium. A seeding density of 6,000 cells/cm 2 was used. Medium exchanges were performed for 50% of the total volume, every 2 days. Cells were culture in ultralow 96-well plates for a total of 7 days. For 96 well-plates, relative viable cell number was measured using the Cell Counting Kit-8 (CCK-8; Dojindo) every time medium exchanges were performed. Briefly, 10 µl of reagent was added per 100 µl of microcarrier suspension. After incubation for 1 h in a standard incubator, the supernatant was transferred to a 96-well plate and absorbance readings at 450 nm were performed in a microplate reader (Tecan). Quantification of the cells per well was performed based on a standard curve performed on the 1st day of culture.

| Spinner-flask culture
The protocol used for spinner flask culture was based on dos Santos et al. (2011). Suspension culture experiments were performed in 100-ml flat bottom spinner-flasks (Bellco) with 80 ml as working volume. Spinner-flasks were siliconized using Sigmacote (Sigma-Aldrich) to avoid cell attachment. After applying the reagent to all the glass surface and aspirating it, vessels were let to dry for 24 h in a fume hood. After, vessels were rinsed three times with distilled water (Gibco). Plastic and Plastic Plus microcarriers were used for spinner-flask culture with a total surface area of 518.4 cm 2 . Microcarriers were prepared by autoclaving in distilled water (Gibco) for 121°C for 15 min. After, the distilled water was aspirated as much as possible and complete medium was added. Microcarriers were let to equilibrate for 1 h in a standard incubator. Sixty-toeighty percent confluent PA5 hOMCs were detached as described in section 2.1 and an adequate amount of cells was added to the tube containing the microcarriers suspension in complete medium to a seeding density of 6,000 cells/cm 2 . After transferring the microcarrier and cell suspension to spinner-flasks containing adequate amounts of complete medium, an agitation of 40 rpm was started.
Spinner-flasks were maintained on a Bell-Ennium TM Compact 5 position magnetic stirrer platform (BellCo) in a standard incubator. To allow for air exchange, a side-arm of the spinner-flask was kept loosened (half a turn of the cap). Cell expansion was maintained for 7 days. Feeding was performed every 2 days by removing 50% of the expanded medium and adding the same | 331 amount of fresh complete medium. Sampling, also performed every 2 days, was achieved by retrieving 2 ml of a homogeneous microcarriers suspension for manual cell counting and immunocytochemistry. Cell counting and viability measurements were performed by trypan blue exclusion, using a hemocytometer.

| Harvest
Exhausted medium was retrieved from the vessel and the microcarrier-cell suspension was washed with 40 ml of HBSS (Gibco) twice. TryPLE Select (20 ml; Gibco) was incubated with the microcarrier-cell suspension for 15 min, at an agitation of 100 rpm in a standard incubator. To quench the dissociation reagent, 30 ml of complete medium was added. The suspension was passed through 40-µm cell strainers and centrifuged for 5 min at 400g. Harvested cells were counted using the trypan blue exclusion method and were either pelleted for RT-qPCR or plated at 6,000 cells/cm 2 to perform several assays such as immunocytochemistry and NG108-15 coculture assay.

| Metabolite analysis
On the days medium exchange occurred, metabolite analysis was performed using the CuBian HT270 analyzer (Optocell GmbH & Co, KG, to obtain the concentration of glucose, lactate, and ammonia.

| Immunocytochemistry
Samples were fixed using a solution of 4% paraformaldehyde in PBS for 20 min at RT. After washing with PBS (Lonza) for 5 min, twice, permeabilization was done through incubation of samples with a solution of 0.1% Triton X-100 (Sigma-Aldrich) in PBS for 20 min at room temperature. After washing samples twice with PBS for 5 min, samples were incubated with a blocking solution of 5% goat serum (Dako) in PBS for 1 h at room temperature. Primary antibodies against p75 NTR (Millipore), fibronectin (Sigma-Aldrich), CD90 (Millipore), nestin (Millipore), GFAP (Dako), S100β (Dako), and β-III tubulin (Sigma-Aldrich; Table 2)  Images of planar culture were acquired using the microscope system EVOS® FL (Thermo Fisher Scientific), while microcarrier culture 16-bit multicolor montage images were obtained with a Zeiss LSM 880 with Airscan confocal microscope system (Carl Zeiss) and Zen 2009 acquisition software (Carl Zeiss). The -∆∆ 2 C t method was used to analyze the data (Table 3).

| Coculture assay
A coculture of NG108-15 neurons with PA5 cells after expansion on microcarriers was used to assess neuron outgrowth. NG108-15 cells (ATCC HB-12317) are a hybrid rodent glioma-neuroblastoma cell line. As a positive control, a rat SCL 4.1/F7 (ECACC 93031294) Schwann cell line was used.
NG108-15 and F7 cells were grown for two passages on T-flasks coated with PLL (Sigma-Aldrich, 100 µg/ml) or uncoated, respectively. When 60%-80% confluency was reached, NG108-15 cells were passaged by hitting the flask and replated at 6,000 cells/cm 2 , while F7 cells were passaged following the same protocol used for PA5 hOMCs and re-plated at 6,000 cells/cm 2 . For the coculture assay, PA5 hOMCs were plated at 6,000 cells/cm 2 onto freshly coated 24-well plates with PLL (Sigma-Aldrich, 100 µg/ml). After 24 h, NG108-15 cells were plated on wells with PA5 hOMCs or F7 cells using a seeding density of 500 cells/well. Medium changes were performed every 2 days and samples fixed after 5 days.
4-OHT was not added in the medium for this assay. Immunocytochemistry was performed as previously described. Images were acquired with EVOS FL Imaging System (Thermo Fisher Scientific) at a ×100 total magnification. Fifteen frames were acquired per condition (five per technical repeat) and neurite quantification was performed manually using the NeuronJ (Meijering et al., 2004) plugin in ImageJ.

| Statistical methods
The Kolmogorov-Smirnov test was used for data normality and

Formulae
The specific growth rate, µ (day −1 ): where X is the concentration of cells, X 0 is the concentration of cells at t 0 . μ for the exponential phase, considered to be from day 3 to 7.
Doubling time, t d (days): where ( ) C X t is the cell concentration at the end of the culture and ( ) C X 0 is the initial cell concentration.
Specific nutrient consumption and metabolite production rate, reached by Day 7 (Figure 1a). Morphologically, PA5 hOMCs showed an elongated, fibroblastic like shape (Figure 1b). Cells stained positive for the glial markers p75 NTR , S100β, and GFAP which are commonly used to characterize the OEC phenotype (Au et al., 2002;Bianco et al., 2004;Hahn et al., 2005;Kawaja et al., 2009)   Representative images used to quantify neurite outgrowth are shown in Figure 5c. The average neurite length was 36 ± 1.3 µm for Plastic Plus microcarriers, significantly higher (p < .05) than the average neurite length obtained for Plastic (29.1 ± 1.5 µm) microcarriers and the positive control (27.9 ± 0.1 µm; Figure 5d). The maximum neurite length obtained for Plastic Plus was 96 ± 7.0 µm, significantly higher F I G U R E 3 PA5 hOMCs were grown on Plastic L, Synthemax II LC, Collagen, Plastic, PronectinF and Plastic Plus microcarriers for 7 days with 50% media change every 2 days and using an initial seeding density of 6,000 cells/cm 2 . Viable cell concentration (a) and cell viability (b) were measured. Metabolite analysis for glucose (c), lactate (d), and ammonium (e) were performed. (f) Expression of p75 NTR , S100β, GFAP, β-III tubulin, nestin, and fibronectin by PA5 hOMCs grown on Plastic L, Synthemax II LC, Collagen, Plastic, PronectinF, and Plastic Plus, assessed through RT-qPCR on Day 7. PA5 hOMCs before seeding were used as a reference. Data are represented as mean ± SEM, n = 3. Three independent repeats were performed for each condition. Significant differences were noted with (*) for p < .05. GFAP, glial fibrillary acidic protein; hOMC, human olfactory mucosa cell

| DISCUSSION
PA5 hOMCs are a potential candidate therapy for an allogeneic cell therapy for the treatment of spinal cord injury (Santiago-Toledo et al., 2019). High cell doses of up to 1 billion cells will be required to produce cell banks to be supplied to patients (Casarosa et al., 2014).
It has been shown that stirred tank bioreactor cell culture using microcarriers is a reliable, reproducible method to achieve high numbers of cells in a scalable manner (Qiu et al., 2016;Rafiq et al., 2013;Rodrigues et al., 2018;dos Santos et al., 2011). The first step to develop such a production method would be to select the right type  (Delorme et al., 2010;Murrell et al., 2005;Nash et al., 2001) and may therefore be regenerative constituents of the PA5 hOMC populations. The upregulation of nestin for cells grown on Plastic and Plastic Plus microcarriers could contribute to increased neural regeneration, and may have led to the survival of olfactory neurons, displaying the expression of β-III tubulin, an early neural differentiation marker. Furthermore, even though Plastic Plus microcarriers did not lead to such an increase in cell number, positive surfaces have been used in the past to promote attachment of neuronal cell types that may lead to neural regeneration (Kozak et al., 1978).
These results led to the selection of Plastic and Plastic Plus microcarriers for suspension culture in spinner flasks following a previously described method to grow hMSCs (dos Santos et al., 2011).
Plastic L microcarriers were not taken into further studies due to the impracticality of in-house coating of microcarriers which is an additional expense and manipulation which would have to be considered in the manufacturing process.  (Jones & Bianchi, 2015). These results can be explained by the upregulation of LDH activity promoted by the presence of the oncoprotein c-Myc (Shim et al., 1998 (Daud et al., 2012;Jonsson et al., 2013).
| 341 led to an enrichment of OECs, which were then responsible for the increase in neurite outgrowth on this condition. To fully elucidate the mechanism of action, it will be necessary to subclone the PA5 hOMCs population. These clones will have to be phenotypically characterized through immunocytochemistry or flow cytometry, expanded, and checked for potency using the NG108 coculture assay as well as animal studies.
In conclusion, we have demonstrated that expansion of PA5 hOMCs on Plastic and Plastic Plus microcarriers in suspension culture in agitated culture using spinner flasks is permissive. This study shows that expanding c-MycER TAM -derived hOMCs on Plastic Plus microcarriers leads to cells with increased potency and increased expression of p75 NTR , a marker expressed by OECs. The work pre-