Caudalized human iPSC-derived neural progenitor cells produce neurons and glia but fail to restore function in an early chronic spinal cord injury model

Abstract

Neural progenitor cells (NPCs) have shown modest potential and some side effects (e.g. allodynia) for treatment of spinal cord injury (SCI). In only a few cases, however, have NPCs shown promise at the chronic stage. Given the 1.275 million people living with chronic paralysis, there is a significant need to rigorously evaluate the cell types and methods for safe and efficacious treatment of this devastating condition. For the first time, we examined the pre-clinical potential of NPCs derived from human induced pluripotent stem cells (hiPSCs) to repair chronic SCI. hiPSCs were differentiated into region-specific (i.e. caudal) NPCs, then transplanted into a new, clinically relevant model of early chronic cervical SCI. We established the conditions for successful transplantation of caudalized hiPSC-NPCs and demonstrate their remarkable ability to integrate and produce multiple neural lineages in the early chronic injury environment. In contrast to prior reports in acute and sub-acute injury models, survival and integration of hiPSC-derived neural cells in the early chronic cervical model did not lead to significant improvement in forelimb function or induce allodynia. These data indicate that while hiPSCs show promise, future work needs to focus on the specific hiPSC-derivatives or co-therapies that will restore function in the early chronic injury setting.

Keywords: Induced pluripotent stem cells; Neural progenitor cells; Spinal cord injury.

Figure 1. hiPSCs form region-specific neural progenitor cells

(A) Outline of differentiation protocol. (B-D) Morphological characteristics of differentiated IMR90 hiPSCs. At day 18 of differentiation, IMR90 hiPSCs formed neural rosettes (B), while at day 45, hiPSCs displayed a multipolar morphology and coalesced into neurobuds consistent with NPCs (C). Caudalized hiPSC-NPCs at day 66, just prior to transplant, tended toward a more complex morphology, as indicated in (D). (E-G) At day 45, differentiated hiPSCs expressed the pan-neuronal marker Tuj1 (E), and the astrocyte marker GFAP (F). (H-J) hiPSC-NPCs uniformly express HoxB4(H), and demonstrate expression of Sox9 (I). (K) Most hiPSC-NPCs express GFAP. (L) A small subset of hiPSC-NPCs express the oligodendrocyte marker GSTpi. (M) When co-cultured with RMT fibers, hiPSC-NPCs express synapsin and are observed in close association with muscle fibers (inset). Nuclei in (G, J, K-M) are counterstained with DAPI. Error bars are SEM. Scale bars = 50 μm.

Figure 2. Caudalized hiPSC-NPCs downregulate pluripotency genes and broadly upregulate neuron and glial specific genes

TaqMan low-density array data is arranged by pathway. (A) Pluripotent stem cell and neural stem cell markers. ESCs and hiPSCs expressed high levels of pluripotency markers, while hiPSC-NPCs downregulated these same genes and dramatically elevated expression of the neural stem cell markers PAX6 and SOX9. (B) Neuronal markers. hiPSC-NPCs expressed elevated levels of neuron-specific transcripts including DCX, MAP2, NCAM1 and NPY compared to pluripotent ESC or iPSC controls. (C) Astrocyte markers. hiPSC-NPCs expressed dramatically higher levels of GFAP and S100β and modest elevation of SLC1A2. (D) Oligodendrocyte markers. hiPSC-NPCs expressed many genes of the oligodendrocyte lineage, including CNPase and PDGRα. OLIG1 and OLIG2 were uniquely expressed in hiPSC-NPCs when compared to pluripotent stem cell controls.

Figure 3. A chronic cervical hemi-contusion model that recapitulates human pathology

(A) Fourth generation of the OSU injury device. (B) Contusion injury data: representative displacement and force tracings from one injury. (C) Four weeks after injury, GFAP staining reveals substantial cavitation and reactive gliosis on the side of contusion. (D) Eight weeks after transplant, hiPSC-NPCs have survived and integrated in the chronic contusion model. Most cells remain close to the site of injection (i.e. near the gray-white border of the dorsolateral funiculus and juxtaposed to the lesion cavity), but many migrated throughout the ipsilateral white and gray matter.

Figure 4. hiPSC neuronal subtypes and influence of hiPSC-NPCs on host neurons

(A) Projections from grafted hiPSC-NPCs surrounding host NeuN+/GFP- neurons. A number of hiPSC-NPCs formed NeuN+/GFP+ neurons. (B) hiPSC-NPCs did not form excitatory GluR1+ cells. (C) No hiPSC-NPC serotonergic cells were observed. (D) GABAergic hiPSC-derived neurons in the ipsilateral corticospinal tract. There was an increase in GABA signal compared to the contralateral region of the same section. (E) Increase in number and thickness of host ChAT+/GFP- neuronal projections surrounding GFP+ hiPSC-NPCs. Ipsilateral and contralateral panels are taken from the same coronal section. Scale bar = 200 μm, except in “Ipsilateral”/“Contralateral” panels, where scale bar = 400 μm.

Figure 5. Caudalized hiPSC-NPCs four weeks after transplant divide and express Sox9

(A) NPCs continue to divide after transplantation. A percentage of HuNu-positive cells also express the proliferation marker Ki-67. Arrowheads indicate HuNu/Ki-67 double-positive cells. (B) NPCs are positive for a human-specific GFAP antigen, but do not express MBP. Merge indicates little association of hGFAP-positive cells with MBP-positive regions of the injured cord. (C) Exogenous formation of immature neurons is rare after four weeks. HuNu-positive cells infrequently express doublecortin. Z-stack merge confirming colocalization of Dcx with HuNu. (D) A higher percentage of NPCs express Sox9 in vivo than in vitro. Scale bars = 50μm.

Figure 6. Caudalized hiPSC-NPCs eight weeks after transplant stop dividing and differentiate into neurons and glia

(A) HuNu-positive cells adjacent to the lesion cavity (asterisk) express the immature neuronal marker doublecortin and the pan-neuronal marker Tuj1. (B) Z-stack imaging of Dcx-positive hiPSC-NPCs (arrows) confirms neurogenesis in the injured spinal cord. (C) Phenotypic analysis of hiPSC-NPCs shows the percentages of Dcx-positive neurons, GFAP-positive astrocytes and GSTpi-positive oligodendrocytes at 4 and 8 weeks after transplant. (D) A portion of NPCs retain expression of GFAP. (E) Some NPCs have differentiated into GSTpi-positive oligodendrocytes. Error bars are SEM. Scale bars = 50 μm.

Figure 7. Transplants of caudalized hiPSC-NPCs do not significantly improve motor function compared to control conditions

(A) Weekly performance on the forelimb reaching task, an indicator of volitional functional recovery. No significant difference was observed among NPC and control groups. (B) Weekly performance on limb-use asymmetry test (LUAT), an indicator of paw preference and weight-bearing ability. (C) Comparison of pre-transplant to final LUAT scores, indicating improvement in both the NPC and sham treatment groups. PIx indicates x weeks after injury. PTx indicates x weeks after transplant. * = p < 0.01. Error bars are SEM.