Transplantation of IGF-1-induced BMSC-derived NPCs promotes tissue repair and motor recovery in a rat spinal cord injury model.

Bone marrow-derived mesenchymal stem cells (BMSCs) have therapeutic potential for spinal cord injury (SCI). We have shown that insulin-like growth factor 1 (IGF-1) enhances the cellular proliferation and survivability of BMSCs-derived neural progenitor cells (NPCs) by downregulating miR-22-3p. However, the functional application of BMSCs-derived NPCs has not been investigated fully. In this study, we demonstrate that knockdown of endogenous miR-22-3p in BMSCs-derived NPCs upregulates Akt1 expression, leading to enhanced cellular proliferation. RNASeq analysis reveals 3,513 differentially expressed genes in NPCs. The upregulated genes in NPCs enrich the gene ontology term associated with nervous system development. Terminally differentiated NPCs generate cells with neuronal-like morphology and phenotypes. Transplantation of NPCs in the SCI rat model results in better recovery in locomotor and sensory functions 4 weeks after transplantation. Altogether, the result of this study demonstrate that NPCs derived with IGF-1 supplementation could be differentiated into functional neural lineage cells and are optimal for stem cell therapy in SCI.

Introduction

Traumatic spinal cord injury (SCI) is a common and devastating nervous system disorder that often results from accidents. The pathophysiology of SCI involves a primary injury that directly disrupts nerve cells and surrounding blood vessels. Undermanaged primary injury leads to secondary injury cascades, such as inflammation, vascular dysfunction, ischemia, edema, and excitotoxicity. These symptoms perturb communication between the brain and the body, resulting in the loss of voluntary movements and sensation below the damaged plane (Hosseini et al., 2018; Tran and Silver, 2015).

Because of the slow regenerative capability of the central nervous system (CNS) cells upon injury, the treatment of SCI remains a therapeutic challenge. Surgical re-stabilization of the vertebral column and rehabilitation are the current medical practices to prevent secondary complexities and to provide support to the patient (Sandean, 2020). Despite the overall advancement in medical and surgical care, which has improved the outcomes for SCI patients, no effective treatment is currently available for neurological deficits after SCI.

Stem cell-based therapies using embryonic stem cells, neural stem cells, or mesenchymal stem cells (MSCs) are emerging approaches for spinal cord repair. The use of stem cells to treat neurodegenerative diseases, such as SCI, has been extensively studied (Kim et al., 2016; Qu and Zhang, 2017; Duncan and Valenzuela, 2017). In fact, stem cell therapy offers a wide range of medical benefits against SCI, such as axon re-myelination, restoration of neuronal circuitry, reduced inflammation, and promotion of angiogenesis (Xiong et al., 2010). Among MSCs from different sources, bone marrow-derived MSCs (BMSCs) are promising alternatives for the treatment of neurodegenerative disorders because of their easy accessibility, expandability in vitro, unique immunogenic properties, and capability to differentiate into neural cell types (Tanna and Sachan, 2014; Ankrum et al., 2014; Alexanian et al., 2008; Mezey et al., 2003).

BMSCs are multipotent stem cells capable of differentiation into many cell types, such as chondrocytes, adipocytes, osteocytes, and neural lineage cells (Takeda and Xu, 2015; Bai et al., 2004). It has been reported that BMSC transplantation enhanced axonal regeneration and promoted functional recovery in spinal cord injury animal models (Lee et al., 2003). Recently, a study has shown that intrathecal transplantation of allogeneic BMSCs did not produce neurological deficits or immune rejection against donor cells in a canine model (Benavides et al., 2021). Similarly, studies have reported that both autologous and allogeneic BMSCs can be safely administered in SCI patients (Pan et al., 2019; Karamouzian et al., 2012; Vaquero et al., 2017; Satti et al., 2016). All these pieces of evidence suggest that BMSCs could be a potential therapy for irreversible damage to the CNS.

In our earlier study, we reported a protocol for the differentiation of BMSCs into neural progenitor-like cells (NPCs) by the supplementation of insulin-like growth factor 1 (IGF-1) into the culture media along with epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF). The addition of IGF-1 generated better neurospheres than those produced using only a combination of EGF and bFGF in terms of cell survivability and proliferation (Huat et al., 2014). MicroRNA profiling of BMSC-derived NPCs delineated several key microRNAs associated with IGF-1 supplementation, with mir-22-3p being the most strikingly downregulated microRNA.

MiR-22-3p is a neuron-related microRNA expressed in the neurites (Fiumara et al., 2015). Overexpression of miR-22-3p induced cell senescence and decreased proliferation and migration of endothelial progenitor cells, whereas suppression of miR-22-3p showed reversed effects (Zheng and Xu, 2014). MiR-22-3p has also been reported to regulate the cell cycle during cerebellum development by targeting the Max and Myc genes (Berenguer et al., 2013). Despite vast information on miR-22-3p, how IGF-1 regulates miR-22-3p expression involved in the enhancement of BMSC differentiation toward NPCs remains unclear. The functional properties of BMSCs-derived NPCs also remain understudied. We hypothesize that NPCs derived under EGF, bFGF, and IGF-1 supplementation could better promote the recovery of the injured spinal cord in a rat model.

In this study, we first investigate the role of miR-22-3p in cellular proliferation and survivability of BMSC-derived NPCs via the loss-of-function approach. We also perform RNASeq of BMSCs-derived NPCs to elucidate the transcriptomic alterations during the differentiation of the BMSCs into a neural lineage. We validate the functional properties of BMSC-derived NPCs using both an electrophysiological approach and transplantation of NPCs into a rat SCI model. Overall, our experiment reveals that transplantation of NPCs derived under the influence of bFGF, EGF, and IGF-1 promotes better recovery of the injured spinal cord and improves sensory and motor functions in the rat SCI model.

Results

Primary Isolated BMSCs and Neural Induction

The characteristics of primary isolated BMSCs were determined using selected cell surface markers and the capability of BMSCs to differentiate into mesodermal lineage cells (Dominici et al., 2006). BMSCs exhibited fibroblast-like morphologies at 90% confluency in vitro; they expressed surface antigens CD90, CD44, fibronectin, vimentin and weakly expressed nestin (Figure 1A), whereas they were negative against the CD11b marker (Figure S1A). Moreover, BMSCs can differentiate into adipocytes, chondrocytes, and osteocytes under specific differentiation media. BMSC-derived adipocytes formed lipid droplets and were stained with oil red O staining. Dense chondrocytes were stained with alcian blue, whereas calcium deposits stained with alizarin red S were observed in the osteogenic differentiation (Figure S1B). Upon differentiation in the neural induction media, adherent BMSCs developed into floating neurosphere-like cells whose size increased with time (Figure 1B). Growth factor supplemented NPCs showed heterogeneous spherical shapes in the population of clonal neurospheres, whereas the neural basal only group appeared irregular in shape, with uneven surface morphology. Neurospheres derived with IGF-1 supplementation exhibited outstanding morphological features, and they had the largest colony size among the groups. Neurosphere-like cells from all groups expressed Sox2 and had low expression levels of fibronectin (Figure 1C). These data suggest that the BMSCs have differentiated into a neural lineage.

Figure 1

Supplementation of EGF, bFGF, and IGF-1 enhanced the cellular proliferation of BMSC-derived NPCs via downregulation of miR-22-3p. (A) Representative photomicrograph of BMSCs in culture. Immunocytochemical staining indicated that BMSCs expressed CD90, CD44, fibronectin, vimentin, and nestin. Nuclei were counterstained with Sytox Blue. Images were viewed under a confocal microscope. Scale bar: 100 μm. (B) Photomicrograph of free-floating neurospheres generated in neurobasal media with and without growth factor supplementation. Scale bar: 100 μm. The size of each neurosphere was measured using ImageJ based on the longest diameter. (C) Immunofluorescence staining of NPCs with Sox2 (scale bar: 50 μm) and fibronectin (scale bar: 20 μm). (D) Proliferation analysis of NPCs under different growth factors. Cells were incubated with an MTS reagent for 4 h, and changes in proliferation were studied at different time intervals. Data are represented as the mean optical density (OD) at 540 nm. (E) Expression level of miR-22-3p in BMSC-derived NPCs relative to the control (without growth factor). (F) Relative expression of the Pten, Akt1, and Tp53 genes in NPCs derived from different growth factor combinations compared with the control. The indicated fold-change values were normalized to the Actb and Gapdh controls. All experiments were repeated in three biological replicates. Data are presented as mean ± SD. Statistical analysis was performed using one-way ANOVA. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Involvement of MiR-22-3p in the cellular proliferation of BMSC-derived NPCs

The cell proliferation assay was used to evaluate the effects of growth factors on NPC proliferation. Both growth factor-treated groups showed significant differences (p < 0.05) compared with the control group throughout the experimental period (five time intervals: 24 h, 48 h, 72 h, 96 h, and 120 h) (Figure 1D). Multiple comparisons among groups demonstrated that those groups supplemented with growth factors had better proliferative activities than the neural basal only group. Notably, NPCs with IGF-1 supplementation showed the highest proliferation.

Our earlier microarray analysis reported the downregulation of miR-22-3p in NPCs derived under EGF + bFGF + IGF-1 supplementation (Huat et al., 2015b). Bioinformatic analysis revealed that Akt1, Pten, and Tp53, enriched in the negative regulation of apoptosis, GO:0043066, are potential targets of miR-22-3p. Therefore, the endogenous expression of miR-22-3p and the predicted targets were measured using quantitative real-time PCR (qPCR). Indeed, miR-22-3p expression in NPCs (EGF + bFGF + IGF-1) was significantly downregulated compared with both the control and NPCs (EGF + bFGF) (Figure 1E). Consistently, Akt1, Pten, and Tp53 were consistently upregulated in growth factors treated groups compared with the control (Figure 1F). Notably, Akt1 was significantly expressed in the presence of IGF-1. Thus, we postulated that the enhanced proliferation of NPCs in the presence of IGF-1 could be due to the suppression of miR-22-3p.

To further confirm the role of miR-22-3p, we performed a knockdown of endogenous miR-22-3p in BMSCs using the synthetic miR-22-3p inhibitor before neural induction. Co-transfected fluorescent-labelled miRNA was detected within the cytoplasm and nuclei of the BMSCs after 24 h of transfection (Figure S2A). More than 90% of the BMSCs were transfected and remained constant up to the third day in vitro (Figure S2B). The cells also remained viable throughout the experimental period (Figure S2C). Using the 24 h transfected BMSCs, we performed neural induction, as described earlier, and monitored the cells at 72 h. BMSCs transfected with scramble miRNA differentiated into larger neurospheres when supplemented with EGF + bFGF + IGF-1 compared with the condition without IGF-1. By contrast, NPCs differentiated from BMSCs transfected with the miR-22-3p inhibitor did not show significant differences in size (Figure 2A and B). Interestingly, NPCs derived from miR-22-3p knockdowned BMSCs generated more neurospheres without IGF-1 supplementation (Figure 2C). Of note, the expression of miR-22-3p remained downregulated after neural induction (Figure 2D). This finding suggests that miR-22-3p inhibition had a similar effect as IGF-1 supplementation.

Figure 2

MiR-22-3p negatively regulated Representative photomicrographs of neurosphere-like cells generated from transfected BMSCs. BMSCs were transfected with either a microRNA inhibitor negative control (scrambled) or an miR-22-3p inhibitor (miR22-Inh). Transfected BMSCs were induced into a neural lineage under serum-free conditions supplemented with respective growth factor combinations. Images were viewed under an inverted light microscope. Scale bar: 100 μm. (B) Size of neurospheres based on the longest diameter. (C) Number of neurospheres in each condition. (D) Validation of miR-22-3p expression in NPCs derived from miR-22-3p knockdowned BMSCs. The control group was derived from BMSCs transfected with a scrambled microRNA hairpin inhibitor. (E) Cellular proliferation of BMSC-derived NPCs treated with the miR-22-3p inhibitor, miR-22-3p mimic, or scramble control. (F) The mRNA expression of Akt1, Pten, and Tp53 was detected using qPCR. The expression was relative to the control group transfected with a scrambled inhibitor. (G) A luciferase reporter assay was conducted to demonstrate Akt1 as a direct target of miR-22-3p. Data are presented as the ratio of the luminescence intensities (relative light unit, RLU) of the GLuc/SEAP expression. Experiments were repeated in three biological replicates. Data are presented as mean ± SD. Statistical analysis was performed using either the Student's t-test or one-way ANOVA. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

The effect of miR-22-3p inhibition on NPC proliferation was quantified using the MTS assay. Cellular proliferation of NPCs derived from miR-22-3p knockdowned BMSCs supplemented with EGF + bFGF and EGF + bFGF + IGF-1 both showed enhanced proliferation compared with their respective controls (Figure 2E). Consistent with the observed morphology, miR-22-3p knockdowned NPCs (EGF + bFGF) exhibited a higher proliferative activity than miR-22-3p knockdowned NPCs (EGF + bFGF + IGF-1). The differences are most prominent on days 4 and 5. Next, to confirm the anti-proliferative effect of miR-22-3p, a gain-of-function experiment was performed by transfecting BMSCs with miR-22-3p mimic and scramble control followed by neural induction. Interestingly, NPCs transfected with the scramble control exhibited a higher proliferative activity than the miR-22-3p mimic transfected group.

To elucidate the interaction between the predicted genes and miR-22-3p knockdown, the effect of miR-22-3p inhibition on the expression of Akt1, Pten, and Tp53 was assessed using qPCR (Figure 2F). Knockdown of miR-22-3p in both groups showed upregulated expression of Akt1 and Tp53 genes but not Pten. The expression of Akt1 and Tp53 was decreased slightly in the presence of IGF-1. This finding supports our hypothesis that miR-22-3p is involved in fine-tuning cellular proliferation via the Akt1 signaling pathway.

To confirm that miR-22-3p is a direct regulator of Akt1, the 3′UTR sequence of Akt1 was cloned into a luciferase vector and transfected into BMSCs (Figure 2G). Co-transfection of the miR-22-3p mimic and plasmid containing the Akt1 3′UTR sequence resulted in a significant reduction of luciferase expression (2.452 ± 0.11 RLU, p < 0.05) compared with the control with the reversed sequence (3.154 ± 0.44 RLU) and scrambled microRNA (3.016 ± 0.16 RLU) (Figure 2G). These data confirm that Akt1 is a direct target gene of miR-22-3p.

Bulk RNA sequencing analysis of BMSC-derived NPCs

Next, we sought to understand how EGF, bFGF, and IGF-1 supplementation influenced the gene expression profile of BMSC-derived NPCs. We performed bulk RNASeq to compare the transcriptomic profile between BMSC-derived NPCs and undifferentiated BMSCs. To unveil the influence of IGF-1 supplementation, we also compared the transcriptomic profiles of NPCs derived with and without the presence of IGF-1. Assessment of similarity between replicates using principal component analysis (Figure 3A) and unsupervised hierarchical clustering using a distance matrix (Figure 3B) revealed that the BMSC-derived NPCs and BMSCs were well separated and that the samples clustered together, corresponding to the group. The pairwise transcriptomic comparison of NPCs derived with EGF, bFGF, and IGF-1 revealed 3,513 differentially regulated genes, with 1,917 genes being upregulated and 1,596 genes being downregulated compared with the BMSCs.

Figure 3

RNA sequencing comparing NPCs (EGF + bFGF + IGF-1) and BMSCs revealed a gene list associated with nervous system development (N = 3 per group). (A) Principal component analysis plot showing the separation of BMSC-derived NPCs from the BMSCs. (B) Unsupervised hierarchical cluster showing NPC replicates clustered together and separated from BMSCs. (C) Volcano plot illustrating the differentially expressed (DE) gene for each comparison between NPCs and BMSCs. Cut-off values of 0.05 and 2.0 are used for the p-value and fold change, respectively. Significant DE genes were annotated. p-values are capped at 50 for a maximized view. (D-E) Gene ontology (GO) terms enriched in genes upregulated (D) and downregulated (E) in NPCs. Significantly enriched (right-sided hypergeometric test) GO terms are shown in the y-axis, and the corresponding adjusted p-values (Bonferroni step-down) are reported in the x-axis. The sizes of the dots represents the percentages of genes associated with the GO term identified. (F) Gene set enrichment. Significantly enriched gene sets from MSigDB are reported for the gene expression analysis of NPCs compared with BMSCs. All the terms shown are upregulated in BMSCs. (G) UpSet plot of intersection across the three comparisons. The bar chart on the left indicates the total number of genes for each comparison separately for up and downregulated genes. The upper bar chart indicates the intersection size between sets of genes that are up or downregulated with one or more comparisons. The dark connected dots on the bottom panel indicate which comparisons are considered for each intersection.

The top 20 significantly differentially expressed protein-coding genes between NPCs (EGF + bFGF + IGF-1) and BMSCs are listed in Table 1. A complete list of the differentially expressed gene is provided in Table S3. The volcano plot demonstrates the significance and magnitude of expression changes, with some representative genes are labeled (Figure 3C).

Table 1

List of top 20 differentially expressed gene.

Ensembl ID	Gene symbol	Log2 Foldchange	Adjusted P-value
ENSRNOG00000021201	Txnip	7.15	0.00E + 00
ENSRNOG00000015505	Mfap5	−5.19	0.00E + 00
ENSRNOG00000011490	Angptl8	6.79	1.02E − 28
ENSRNOG00000007827	Cox4i2	6.39	1.16E − 127
ENSRNOG00000029342	Scn7a	7.09	1.40E − 08
ENSRNOG00000056151	Rasgrf2	7.86	1.90E − 07
ENSRNOG00000021260	Prnd	−9.21	2.05E − 10
ENSRNOG00000005906	LOC103690020	8.75	2.30E − 09
ENSRNOG00000033026	Dclk3	6.14	2.42E − 08
ENSRNOG00000008245	Ptgis	5.59	3.31E − 31
ENSRNOG00000014333	Vcam1	5.87	3.40E − 53
ENSRNOG00000017976	Slco2b1	11.17	5.09E − 16
ENSRNOG00000052070	Aldh1a3	5.33	5.53E − 42
ENSRNOG00000007865	Ephb1	7.92	5.65E − 07
ENSRNOG00000011334	Tmem63c	−5.50	5.96E − 11
ENSRNOG00000014751	Ret	8.45	6.74E − 15
ENSRNOG00000032178	Cenpa	−6.71	6.84E − 10
ENSRNOG00000011989	Vat1l	7.54	9.03E − 09
ENSRNOG00000010079	Ca3	−9.79	9.38E − 12
ENSRNOG00000047349	AABR07006269.1	−5.98	9.79E − 08

All genes differentially regulated 2-folds and above were analyzed for GO terminology (Table S4). Non-redundant GO enrichment analysis in the upregulated genes revealed that 30% of the genes were enriched in the GO term autonomic nervous system development, indicating that the genes related to nervous system development are upregulated during differentiation (Figure 3D). On the other hand, the GO terms enriched in the downregulated genes were related mainly to cytoskeletal proteins (Figure 3E). A complete list of the upregulated genes involved in the GO terms, such as the nervous system, neurogenesis, and neuronal projection, is presented in Table S4. Furthermore, gene set enrichment analysis indicated that processes related to cell cycle regulation and mitotic cytokinesis, which are crucial during cell proliferation, were enriched in BMSCs on the expression dataset of NPCs versus BMSCs (Table 2 and Figure 3F).

Table 2

List of pathways and biological processes. Reported are the terms significantly enriched in BMSCs.

MSigDB Gene Set	SIZE	ES	NES	FDR q-value
INTEGRIN PATHWAY	19	−0.77	−2.01	0.00
CELL CYCLE CHECKPOINTS	160	−0.49	−1.86	0.00
ACTIN FILAMENT BASED PROCESS	435	−0.40	−1.68	0.03
CELL CYCLE	343	−0.38	−1.58	0.04
MITOTIC CYTOKINESIS	32	−0.55	−1.56	0.04
MICROTUBULE CYTOSKELETON ORGANIZATION INVOLVED IN MITOSIS	66	−0.46	−1.55	0.04
CELL CYCLE G2/M PHASE TRANSITION	162	−0.41	−1.55	0.03

ES = Enrichment score; NES = Normalised enrichment score; FDR = False discovery rate.

Next, we compared the transcriptomic profiles of NPCs (EGF + bFGF + IGF-1) with those of NPCs (EGF + bFGF), which were reported earlier (Khan et al., 2020). Interestingly, only 11 differentially regulated protein-coding genes were identified between NPCs derived with and without IGF-1 (Table 3 and Table S5). Of note, WD repeat-containing protein 62 (Wdr62) and RET proto-oncogene were upregulated by 6.394-folds and 2.448-folds, respectively, in NPCs derived with IGF-1. Furthermore, assessment of the common and uniquely differently expressed genes across all the comparisons (Figure 3G) revealed some overlapping between differentially expressed genes (Jaccard score = 0.42), the upregulated genes of NPCs (EGF + bFGF) versus BMSCs and NPCs (EGF + bFGF + IGF1) versus BMSCs (Jaccard score = 0.41), and between the downregulated genes of NPCs (EGF + bFGF) versus BMSCs and NPCs (EGF + bFGF + IGF1) versus BMSCs (Jaccard score = 0.44). Despite the considerable overlaps, the transcriptomes of the two NPCs (derived with or without IGF-1) are distinct.

Table 3

Differentially expressed genes comparing between NPCs derived with and without IGF-1.

Ensembl	Genes symbol	Log2 FoldChange	Adjusted P-value
ENSRNOG00000013250	Pdcd5	8.825	8.4E − 07
ENSRNOG00000049708	Wdr62	6.394	1.3E − 02
ENSRNOG00000014751	Ret	2.448	3.2E − 06
ENSRNOG00000053272	Chi3l1	1.404	4.3E − 02
ENSRNOG00000007827	Cox4i2	1.253	9.9E − 06
ENSRNOG00000026605	Ifi27l2b	1.233	1.8E − 02
ENSRNOG00000003984	Apln	1.212	1.4E − 07
ENSRNOG00000061639	AABR07044404.1	1.203	1.0E − 02
ENSRNOG00000003244	Ltc4s	1.118	5.2E − 03
ENSRNOG00000020025	LOC108348052	−1.400	3.4E − 03
ENSRNOG00000046667	Fosb	−1.770	4.8E − 18

Terminal differentiation of NPCs on collagen hydrogel and electrophysiology

Next, we seeded NPCs generated under respective growth factor supplementations onto a collagen hydrogel matrix to mimic the 3D environment of the brain (Akcay and Luttge, 2021). Interestingly, NPCs from both groups differentiated into cells with neuronal features with elongated dendrites upon removal of growth factors and were cultured in the specialized NeuroCult™ NS-A differentiation media (Figure 4A). Differentiated cells exhibited pyramidal shapes with multiple neurite formations (Figure 2A Inset) and positively expressed beta-III-tubulin (TUJ1) and postsynaptic density protein 95 (PSD95) markers (Figure 4B). Notably, neurite protrusion and long neurite outgrowth were observed in both conditions but with more prominence on NPCs induced with EGF + bFGF + IGF-1. The formation of a synaptic junction-like structure was also observed between two differentiated cells (Figure 4C).

Figure 4

Terminally differentiated BMSC-derived NPCs exhibited neuronal-like phenotypes. (A) Representative photomicrograph of terminally differentiated BMSC-derived NPCs on a 3D collagen hydrogel. NPCs from both groups differentiated into neuronal-like cells. Images were viewed under an inverted light microscope. Scale bar: 100 μm. (B) Representative image of neuronal-like cells expressing neuronal markers, such as TUJ1 and PSD95. NPCs derived under the supplement of IGF-1 differentiated into neuronal-like cells with longer neurites. Scale bar: 50 μm. (C) Close-up image of neurite outgrowth from NPCs (EGF + bFGF + IGF-1). Neurite protrusions (white triangle) and elongated neurites (inset) were observed after 1 week in culture. Scale bar: 50 μm. (D) Representative phase-contrast image of neuronal-like cells used for patch clamp recording. (E) Representative traces of spontaneous miniature excitatory postsynaptic current recorded. Cells in both groups produced spikes with larger amplitudes than thoseof the control. Undifferentiated cells, assumed to be glial cells, served as controls during patch clamp recording. Experiments were repeated thrice independently.

Next, we investigated the synaptic characteristics of neuronal-like cells using voltage clamps. The miniature excitatory postsynaptic current revealed that neuronal-like cells derived from the NPCs (EGF + bFGF + IGF-1) exhibited higher amplitudes than the current traces recorded in the neuronal-like cells from NPCs derived only with EGF and bFGF (Figure 4D). These data suggested that neuronal-like cells differentiated from the MSC-derived NPCs were functional.

Recovery of SCI promoted by BMSC-derived NPCs

We studied next the effects of BMSC-derived NPC transplantation on the recovery of the spinal cord after a hemisection injury (Figure 5A). Both NPCs derived under growth factor supplementation were transplanted into the spinal cord lesion. The Basso, Beattie, and Bresnahan (BBB) locomotor rating scale was used to assess locomotor function after SCI in rats (Figure 5B). The results showed that the recovery of sequential hind limb motor recovery was elicited in all treated groups except the sham and vehicle control groups (Figure 5C). Rats transplanted with NPCs (EGF + bFGF + IGF-1) performed the best, recording the highest mean BBB scores of approximately 12 at day 28 post-transplantation. By contrast, the rats transplanted with NPCs derived without IGF-1 supplementation achieved stabilized mean BBB scores of roughly 9.

Figure 5

Transplantation of BMSC-derived NPCs in a spinal cord injury animal model promoted locomotor and sensory recoveries (N = 3 per group). (A) A 4-mm-long longitudinal cut along the midline of the spinal cord for lateral hemisection at the T9-T10 level. (B) Timeline of the entire in vivo transplantation study. (C) Open-field locomotor assessment. The hind limb function of all rats was assessed using the Basso, Beattie, and Bresnahan (BBB) locomotor scale. (D) Mechanical sensory assessment as conducted with the Von Frey filament test. Two-way ANOVA was performed, followed by a Tukey test to compare the mean difference among the groups and times. An asterisk (∗) indicates a significant difference of the treatment groups and the control group. A hashtag (#) indicates a significant difference between treatment days. (E) Relative expression of Map2, Gfap, Olig2, Mbp, Sox2, and nestin within the transplanted region of the spinal cord. (F) Western blot analysis and relative quantification of TUJ1, GFAP, APC, and MBP. Each protein band was normalized to GAPDH. Data are presented as mean ± SD. Statistical analysis was performed using one-way ANOVA to compare within groups. ∗p < 0.05, #p < 0.05, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Mechanical sensory function was assessed using Von Frey filaments. The starting filament was used at a force of 2 g. There were no differences in the paw withdrawal threshold before injury and at day 1 post-transplantation among all groups (Figure 5D). However, starting from day 7 post-transplantation, the average values of the paw withdrawal threshold in animals with NPCs (EGF + bFGF + IGF-1) were significantly reduced compared with those of the group receiving NPCs (EGF + bFGF), p < 0.05. These data suggested that transplantation of NPCs derived under IGF-1 supplementation promoted both locomotor and sensory recoveries of SCI.

We subsequently investigated the impact of stem cell transplantation on gene and protein expression within the transplanted region. Gene expression of neurons (Map2), astrocytes (Gfap), oligodendrocytes (Olig2), myelination (Mbp), and neural stem cells (nestin and Sox2) was quantified using qPCR (Figure 5E). Animals receiving NPCs (EGF + bFGF + IGF-1) exhibited significantly higher Map2, Mbp, and Olig2 than those animals receiving NPCs (EGF + bFGF). Moreover, Gfap expression was also increased in the animals transplanted with NPCs (EGF + bFGF + IGF-1), but they showed no significant differences compared with the animals receiving NPCs (EGF + bFGF). There was no significant difference in the level of stem cell markers, nestin and Sox2, in both NPC transplanted groups. These data suggested that transplanted NPCs have differentiated into neurons, astrocytes, and oligodendrocytes. Furthermore, a higher Mbp expression in animals receiving NPCs (EGF + bFGF + IGF-1) meant better re-myelination of the spinal cord in this group of animals. Western blot analysis showed that markers for neuronal cells (TUJ1), glial cells (GFAP), and Oligodendrocytes (APC) were expressed after NPC transplantation (Figure 5F). Animals transplanted with NPCs (EGF + bFGF + IGF-1) exhibited the highest expression of TUJ1 and APC consistent with the gene expression. These data suggested that NPCs derived under the influence of EGF + bFGF + IGF1 had a higher tendencies to differentiate into neuronal and glial cells compared with NPCs (EGF + bFGF).

To observe the effect of NPC transplantation in the cross-sectioned area of the injured spinal cords, tissues from the adjacent site of the lesion area were sectioned at 3 μm thickness at day 28 post-treatment. The recovery was minimal in the vehicle control group, in which the animals received only collagen hydrogel, with poorly developed neurons and some glial cells present in the gray matter. The white matter showed extensive microcyst formation and marked necrosis, as evidenced by perinuclear cytoplasmic vacuolation of the glial cells (Figure 6A). Furthermore, a noticeably higher number of immune cells infiltrated indicated a greater severity of inflammation. A few neurons and more astrocytes appeared in the gray matter in the vehicle group. Still, the white matter showed extensive microcyst formation and the presence of only a few astrocytes. In animals receiving NPCs (EGF + bFGF), microcysts and vacuolation were still present in the white matter. The profile of tissue healing was significantly improved in animals receiving NPCs (EGF + bFGF + IGF-1). Although vacuolation was still present, the severity was lesser than in animals receiving NPCs (EGF + bFGF), and the tissue structure was more compact and well formed.

Figure 6

Animals transplanted with NPCs (EGF + bFGF + IGF-1) exhibited less inflammation in the tissues adjacent to the lesioned site (N = 3 per group). (A) Hematoxylin and eosin (H&E) staining of the gray and white matter sections of the spinal cord. (B-E) Immunohistochemistry staining of spinal cord sections against GFAP (B) and APC (D) markers. The mean intensity of GFAP (C) and APC (E) staining was analyzed using the IHC toolbox/Deconvolution function plugin from ImageJ. Optical density (mean gray value) obtained using color deconvolution and computerized pixel profiling led to automated scoring. The final score was shown in the corrected OD value (background deducted), which represents the intensity of the brownish color; where the darker the brownish color, the higher the OD value. Scale bar: 100 μm. Data are presented as mean ± SD. Statistical analysis was performed using one-way ANOVA to compare within groups. ∗p < 0.05, ∗∗p < 0.01.

Microcyst formation and inflammatory cells were not observed in this group. We also found that animals receiving NPC transplantation exhibited a lower expression of GFAP than the vehicle control at the adjacent site of the lesion (Figure 6B and C). By contrast, spinal cord transplanted with NPCs showed more APC-positive cells (Figure 6D and E). These data suggested that NPC treatment could have an anti-inflammatory effect while promoting motor and sensory recovery through re-myelination.

Discussion

We have elucidated the molecular mechanism underlying the growth factors-induced neural differentiation of BMSCs. Supplementation of EGF, bFGF, and IGF-1 enhanced cellular proliferation and survivability through downregulation of miR-22-3p. The derived NPCs can differentiate terminally into cells exhibiting neuronal-like phenotypes. Moreover, RNASeq analysis unveiled a series of upregulated genes involved in neural differentiation and development, whereas genes involved in BMSC proliferation were downregulated. Intriguingly, transplantation of BMSC-derived NPCs promoted the recovery of the injured spinal cord in a rat model.

BMSCs have been reported to differentiate into NPCs under growth factor supplementation. The combination of EGF and bFGF is commonly used to differentiate BMSCs into free-floating aggregates called neurospheres (Chouw et al., 2020; Mung et al., 2016). In our earlier study, we reported that the addition of IGF-1 synergistically with EGF and bFGF further enhanced the cellular proliferation and survivability of BMSC-derived NPCs (Huat et al., 2014). IGF-1 is locally produced in the brain and is essential during development (Wrigley et al., 2017; Nieto-Estévez et al., 2016). A study reported that IGF-1-mediated neurogenesis involved the RIT1/Akt/Sox2 cascase (Mir et al., 2017). IGF-1 stimulated an RIT1-dependent increase in Sox2 levels, which subsequently enhanced pro-neural gene expression and promoted cellular proliferation. However, the expression of IGF-1 decreased significantly in the brain upon neuronal maturation (Song et al., 2016). Conversely, IGF-1 and IGF-1R expression levels in the brain were reported to increase following traumatic brain injury (Mangiola et al., 2015). Exogenous administration of IGF-1 was also neuroprotective in ischemic brain injury (Serhan et al., 2019). Increased hippocampal levels of IGF-1 promoted end-stage maturity of post-trauma-born of neurons and improved cognition following injury (Littlejohn et al., 2020). However, one study reported that long-term IGF-1 treatment could induce epileptic and neurotoxicity (Song et al., 2016), so the risk of immediate treatment of IGF-1 needs to be carefully considered. In the present study, NPCs were induced in vitro in the presence of IGF-1. Healthy neurospheres were subsequently used as cell therapy, thus avoiding the unnecessary risk of growth factor oversupply.

An in silico analysis of microRNA profiles delineated several important microRNAs associated with cell proliferation and programmed cell death that were downregulated upon IGF-1 supplementation (Huat et al., 2015a). Among the aberrantly regulated microRNAs, miR-22-3p was the most strikingly downregulated. MiR-22-3p is a highly conserved microRNA with multiple functions, such as epigenetic modification (Kim et al., 2015b), cell differentiation (Zhao et al., 2015), tumorigenesis (Zuo et al., 2015), and disease development (Huang et al., 2013). According to the earliest studies on miR-22-3p, the role of this microRNA in cellular proliferation and apoptosis is context dependent. Under normal physiological conditions, miR-22 is widely expressed in all tissues (brain, heart, liver, lung, kidney, smooth muscle, prostate, testis, ovary, placenta, and adipose) and is associated with cellular differentiation and senescence (Xiong, 2012; Xu et al., 2011; Jazbutyte et al., 2013). It has been shown that miR-22 regulated smooth muscle cell differentiation by targeting MECP2, whereas knockdown of miR-22 inhibited the differentiation (Zhao et al., 2015). Overexpression of miR-22-3p also decreased the cellular proliferation of cerebellar granular neuron precursors by targeting Max. By contrast, knockdown of miR-22-3p diminished the anti-proliferative activity of bone morphogenetic protein 2, causing the cell to proliferate (Berenguer et al., 2013).

On the other hand, miR-22 expression is commonly downregulated in cancer cell lines and is associated with uncontrolled metastasis (Xu et al., 2014; Yang et al., 2014; Xu et al., 2011). Suppression of miR-22 enhanced the cellular proliferation of tongue squamous cell carcinoma by regulating CD147 expression (Qiu et al., 2016). A similar effect of miR-22 inhibition was reported in hepatocellular carcinoma (Luo et al., 2017) and breast cancer cell lines (Kong et al., 2014). Moreover, overexpression of miR-22 has been reported to inhibit the cellular proliferation and migration of glioblastoma by targeting the SIRT1 expression (Chen et al., 2016). These studies proved that the miR-22-3p had various effects, which were context based.

BMSC-derived NPCs under IGF-1 supplementation showed higher Akt1 expression than the NPCs (EGF + bFGF) did, consistent with the enhanced cellular proliferation of NPCs observed in the presence of IGF-1. IGF-1 has been reported to activate the PI3K/AKT and MAP kinase signaling pathways, which confer neuroprotective effects (Wang et al., 2015). Bindings of IGF-1 to IGF-1R activated the receptor kinase, which phosphorylated various intracellular proteins, such as insulin receptor substrate-1 and Shc, leading to the activation of multiple pathways, including PI3K/AKT and MAPK (Laviola et al., 2007). The administration of IGF-1 has been reported to induce cellular proliferation of cultured myoblasts via the PI3K/AKT signaling pathway (Yu et al., 2015). All these evidences highlight the role of IGF-1 in promoting cell proliferation and survival.

Our study demonstrated the ability of miR-22-3p to inhibit Akt1 expression and proposed a mechanism of post-transcriptional regulation of the Akt1 gene. Co-transfection of the luciferase vector containing a wild-type 3′UTR sequence of Akt1 and miR-22-3p mimic resulted in a significant reduction in luciferase activity, proving the inhibition of Akt1 translation by miR-22-3p. Accumulating evidence reports that the Akt1 gene directly targeted by microRNAs, such as miR-143 and miR-302a, resulted in the growth inhibition of cancer cells (Zhang et al., 2015; Noguchi et al., 2013). By contrast, downregulation of the miR-99 family has been reported to release Akt1, which activates cell proliferation and migration via the PI3K/AKT pathway (Jin et al., 2013). Our study suggested that miR-22-3p could directly target and regulate the post-transcription of Akt1 in BMSC-derived NPCs upon IGF-1 supplementation.

As reported in our earlier study, NPCs derived with IGF-1 were better than those derived without IGF-1 (Huat et al., 2014). In the present study, we found 11 differentially expressed genes between the two NPCs (Table 3). Among these genes, Pdcd5 and Wdr62 were upregulated 256-folds and 64-folds, respectively. Programmed cell death 5 (Pdcd5) has been reported to regulate cell proliferation, cell cycle progression, and apoptosis (Li et al., 2018). A study reported that IGF-1 was inversely correlated with Pdcd5 expression (Yi et al., 2013). An increased level of programmed cell death during neural induction is expected when not all cells undergo neural differentiation (Pang et al., 2021). Nevertheless, apoptotic MSCs could be beneficial, warranting further investigation.

WD40-repeat protein 62 (Wdr62) is a spindle microtubule-associated phosphoprotein that is crucial for maintaining neural and glial cell populations during brain development (Shohayeb et al., 2020a; Alshawaf et al., 2017). Knockout of Wdr62 resulted in a significant reduction in the thickness of the hippocampal ventricular and dentate gyrus (Shohayeb et al., 2020a). By contrast, overexpression of Wdr62 has increased the cell proliferation and brain volume of the Drosophila larvae by activating pAKT signaling (Shohayeb et al., 2020b). The upregulation of the Wdr62 gene may signify the role of this gene in specifying the intermediate neural progenitor during BMSC differentiation into a neural lineage.

For the functional study, we transplanted BMSC-derived NPCs in the subacute stage of SCI. Two complementary locomotor function and Von Frey filament tests showed that the transplanted NPCs could significantly improve motor recovery and sensory function for the affected hind limb, respectively. The animals receiving NPCs (EGF + bFGF + IGF-1) displayed higher average of BBB score compared with those animals receiving NPCs (EGF + bFGF) and the vehicle alone. In addition, the average values of the paw withdrawal threshold in the animals receiving NPCs (EGF + bFGF + IGF-1) were significantly reduced compared with the vehicle control. All these pieces of evidence suggested that NPCs (EGF + bFGF + IGF-1) can improve the recovery of SCI in rat model.

The histology of spinal cord tissues after hemisection of SCI was examined using hematoxylin and eosin (H&E) staining to elucidate possible mechanisms leading to the observed functional recovery. The spinal cord section of the control group revealed marked areas of necrosis with vacuolation within the white matter as a result of chronic accumulation of edema and protein aggregates (Wang et al., 2021). By contrast, spinal cord transplanted with NPCs had minimal residual lesions that were barely visible in the cross-section of the spinal cord, and the profile of tissue damage was far more improved. Immunohistochemistry (IHC) staining of the spinal cord using antibodies was examined to detect stemness (SOX2), neural precursor cells (NESTIN), astrocytes (GFAP), oligodendrocytes (APC), and myelin (MBP) 4 weeks after transplantation. Our IHC data showed that NPCs derived under EGF, bFGF, and IGF-1 could differentiate into myelin-forming cells in complex niches after transplantation in the injured spinal cord. This development of myelination has the highest intensity among the groups. The data suggested that BMSC-derived NPC transplantation promoted spinal cord recovery by increasing re-myelination at the lesioned site. Nevertheless, our study did not exclude the possibility that endogenous stem cells or MSC-derived exosomes could also play a role in overall recovery (Mu et al., 2022).

To examine the differentiation potential of NPCs transplanted in the lesioned region, we have evaluated the gene expression for stemness (Sox2 and nestin), neurons (Map2), astrocytes (Gfap), oligodendrocytes (Olig2), and myelination (Mbp) using tissue within the transplanted region. Our results demonstrated a significant increase in Map2, Olig2, and Mbp gene expression in the spinal cords of animals receiving NPCs (EGF + bFGF + IGF-1) compared with those receiving NPCs (EGF + bFGF). Gfap expression was also increased in the NPC-treated groups compared with the vehicle control. It is well reported that oligodendrocytes are the myelinating cells of the CNS and play essential roles in the recovery of the injured spinal cord (Kuhn et al., 2019). Both stemness-related genes (Sox2 and nestin) were not expressed in the NPC transplanted groups. Migrating NSCs at the lesioned side of the spinal cord were reported to change their morphology and lose expression of SOX2 expression (Meletis et al., 2008). Similarly, nestin is commonly used as a marker for NSCs and is essential for NSC self-renewal (Park et al., 2010). Within 24 h of differentiation initiation, neural stem progenitor cells lose their nestin expression, followed by an increased Tuj1 expression (Kim et al., 2015a). Therefore, our data suggested that NPCs derived under IGF-1 supplementation could differentiate into neuronal or oligodendrocyte cell types.

We also evaluated the expression of TUJ1 (neuron), GFAP (astrocyte), APC (oligodendrocyte), and MBP (myelin) at the protein level. Animals receiving NPCs (EGF + bFGF + IGF-1) showed relatively higher expressions of TUJ1 and APC, consistent with the gene expression. The expression of APC indicated myelin sheath formation around the axon because oligodendrocytes are glial cells that support the CNS. They elongated into high numbers of branches and sub-branches, expanding into sheets of myelin membranes that wrapped around multiple neural axons. The myelin sheath accelerated rapid saltatory conduction and insulation of the nerve cells (Cohen et al., 2020). Furthermore, oligodendrocytes promoted neuronal and axonal survival by secreting different neurotrophic factors (Wilkins et al., 2003). The astrocyte marker (GFAP) was also expressed relatively higher in the NPC-transplanted groups. Astrocytes play important roles in sealing the lesion site in the early phase of neural damage (Okada et al., 2006). Studies have shown that astrocytes provide support and axonal guidance and aid in improving functional recovery after SCI (White and Jakeman, 2008). Intriguingly, a study reported that astrocytes can be reprogrammed into neurons in the injured spinal cord (Su et al., 2014). However, whether NPC-derived glial cells in this study contribute to the recovery warrant a further investigation.

In summary, our results suggested that IGF-1 supplementation, along with EGF and bFGF, enhanced the differentiation of BMSCs into NPCs in terms of cellular proliferation and survivability. We demonstrated the mechanism underlying the enhancement of NPC culture under the combination of EGF, bFGF, and IGF-1, which involved the miR-22-3p as a critical post-transcriptional regulator. This study presented incipient evidence that miR-22-3p was engaged in enhancing cellular proliferation and survivability of BMSC-derived NPCs under the influence of mitogens. Transplantation of NPCs in the SCI rat model promoted the recovery of sensory and motor functions. Our study has contributed significantly to the field and proposed that IGF-1 induced BMSC-derived NPCs are suitable candidates for stem cell therapy, especially in the treatment of SCI.

Limitations of the study

This study was limited by the inability to differentiate between transplanted NPCs and endogenous NPCs. Our initial attempt to tag the cells with fluorescence failed because of the high proliferation rate of stem cells, which resulted in the fading of the fluorescence signal. The instability of the fluorescence dye was another factor that contributed to the failure. A higher apoptotic rate was also observed in the tagged primary cells, significantly affecting the downstream experiments. Nevertheless, we believe that the therapeutic effects are the outcomes of transplantation because of the dormancy of endogenous neural stem cells. Still, we cannot rule out the contribution of endogenous neural stem cells in the overall recovery.

Declarations

Author contribution statement

Putri Nur Hidayah Al-Zikri: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Tee Jong Huat: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Amir Ali Khan: Conceived and designed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data.

Azim Patar, Mohammed Faruque Reza: Analyzed and interpreted the data.

Fauziah Mohamad Idris, Jafri Malin Abdullah, Hasnan Jaafar: Conceived and designed the experiments; Contributed reagents, materials, analysis tools or data.

Funding statement

Hasnan Jaafar was supported by the Universiti Sains Malaysia Research University Grant (grant no: 1001/PPSP/812170).

Fauziah Mohamad Idris was supported by USM Research Grant (grant number: RUI/PPSP/1001/812148).

Tee Jong Huat was supported by the Malaysia Toray Science Foundation (grant no: 304/PPSP/6150142/M126).

Jafri Malin Abdullah was supported by a grant from Malaysian National Cancer Council (MAKNA).

Amir Ali Khan was supported by a Competitive Research Grant from the University of Sharjah (grant no: 1602145036-P).

Data availability statement

Data associated with this study has been deposited at NCBI under the accession numbers GSE104548 and GSE60060.

Declaration of interest’s statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.