For severe spinal cord injury (SCI) cases, poor prognosis usually is given due to the characteristics of obstacles after SCI. Stem cell transplantation is one of the most promising yet enigmatic treatments for SCI. In contrast to other types of stem cells, mesenchymal stem cells (MSCs), which can be easily harvested with less ethical and tumorgenic concerns, are recognized as a ready-to-use material for clinical trials at present. By using MSCs in treating SCI, many attributions are demonstrated: anti-inflammation, replacing host neural cells, secreting neurotrophic factors, and modulation the glial scar. Nevertheless, several recent studies showed that MSCs did not possess the ability to differentiate directly into neuronal cells following transplantation and pre-differentiate MSCs into neurospheres, a heterogeneous mixture of cellular aggregates, including neural stem and progenitor cells, prove to be more beneficial for treating SCI. Although some cell sources of MSCs including the bone marrow and adipose tissue have been studied, the suitable one for treating canine SCI is still not clarified. Therefore, the endogenous neuronal cells differentiation potential from canine bone marrow MSCs (cBMMSCs) with that of the adipose tissue-derived MSCs (cADMSCs), both of which are major sources of MSCs, and the levels of neurotrophic factors released from MSCs were compared here. Moreover, the potential of using generated neurospheres as a transplantation material was studied by the animal model.

In chaper 2, cBMMSCs and cADMSCs were isolated and expanded from canine bone marrow and subcutaneous adipose tissue. The proliferation assay was performed by counting the doubling time from first-passage to fourth-passage MSCs. Gene expressions of ectodermal marker, Nestin, βIII-tubulin, GFAP, NCAM, and stem cell maker, NANOG, OCT4, SOX2, were evaluated for cBMMSCs and cADMSCs by RT-PCR. Mesodermal differentiation assay of adipogenic, osteogenic, and chondrogenic lineages were evaluated for cBMMSCs and cADMSCs.

Results indicated that cADMSCs with a stable and shorter doubling interval proliferated faster than cBMMSCs. Both sources of MSCs were differentiated into mesodermal lineages of adipogenic, osteogenic, and chondrogenic successfully, and ectodermal makers of Nestin, βIII-tubulin, NCAM, and stem cells marker of OCT4 and SOX2, were detected by RT-PCR. According to the results, both cBMMSCs and cADMSCs expressed the properties of stem cells in proliferation and multipotent differentiation lineage, and cADMSCs could be harvested in a shorter time for transplantation. Besides, the expression of stem cells markers demonstrated the potential of multipotent differentiation and ectodermal makers expressed prior to any differentiation procedures indicated both MSCs have the potential in differentiating toward neuronal.

In chapter 3, Nestin-positive neurospheres were generated from cBMMSCs and cADMSCs and neuronal cells were differentiated from the generated neurospheres. The harvest rate of neurospheres was counted by cells harvested from generated neurospheres divided by pre-seeded cell numbers. Gene expression of neurospheres was evaluated by RT-PCR and markers were used as chapter 1. The levels of Nestin, OCT4, and SOX2 were compared for MSCs and generated neurospheres by real-time PCR analysis. Neural markers of βIII-tubulin, GFAP, NF200, S100, MAP2, MBP, and Nestin were evaluated by immunofluorescence analysis and the percentage of fluorescence-positive cells were counted and compared. Electrophysiological property of neuronal differentiated cell was evaluated by patch-clamp analysis.

The mRNA expressions of NANOG, Nestin, OCT4, and SOX2 were upregulated in neurospheres derived from both generated Nestin-positive neurospheres. Moreover, about 2 times of cells could be harvested from neuropsheres generated from cADMSCs than cBMMSCs. After neuronal differerntiation, neuron-like morphology was noted and notably, cBMMSC-derived neuronal cells expressed higher levels of βIII-tubulin. Immunofluorescence analysis detected the expression of neural markers of βIII-tubulin, GFAP, S100, NF200, and MAP2, in differentiated neuron-like cells. However, the electrophysiological properties of neuronal differentiated cell were not noted. According to the results, Although electrophysiological property was not verified for neuronal differentiated cells, the upregulation of the markers of neural stem cells and neural makers expressed in immunofluorescence analysis still indicated the generated neurospheres have the potential in differentiating toward functional neurons.

In chapter 4, NGF and BDNF were selected to evaluate their expression in cBMMSCs and cADMSCs at passage 1 and their generated neurospheres by RT-PCR. The levels of gene expression were compared between cBMMSCs and cADMSCs by semiquantative PCR. ELISA analysis was also performed to analysis the levels of NGF and BDNF released from cBMMSCs and cADMSCs at passage 1-3. Moreover, neural-progenitor like cell line, PC12, co-culture with cBMMSCs and cADMSCs was used to evaluate their effects on neurogenesis.

The results of RT-PCR demonstrated the expression of NGF in both MSCs and downregulated in generated neurospheres. The expression of BDNF was only noted in cADMSCs and also downregulated in generated neurospheres. cADMSCs have significant higher gene expression levels of NGF and BDNF than cBMMSCs. By using ELISA assay, the secretion of NGF was demonstrated in cBMMSCs and cADMSCs, and higher levels of NGF were secreted from cADMSCs at passage 1-2. In contrast, the levels of BDNF were not detected in both. After co-culture with cBMMSCs and cADMSCs for 8 days, compare to control group, more PC12 cells with neurite extension and longer neurite extension were noted, but there were no significant differences between the groups which co-culture with cBMMSCs and cADMSCs. According to the results, cADMSCs could secret higher levels of NGF. For the benefits of neurotrophic factors releasing, passage 1 and passage 2 cells of cADMSCs could be used for transplantation. However, both types of MSCs could encourage the neuronal differentiation of PC12 indicated that there are still other factors released from MSCs beneficial to the neuronal differentiation.

In chapter 5, the cells from generated neuropsheres were transplanted into canine's spinal cord and the differentiation fate of transplanted cells was studied. 3 dogs were used and divided into 3 groups: cells from cBMMSCs generated neurospheres (B-NS); cells from cADMSCs generated neurospheres (A-NS); and PBS injection (control). The spinal cord was exposed by laminectomy on L2 area and 1×106 cells pre-labeled with Hoechst 33342 for 1 hour were directly injected into spinal cord by using a Hamilton(R) syringe connected to a 30-gauge needle. After 14 days, the dogs were euthanatized and the transplanted site of spinal cord was excised for HE stain and immunofluorescence analysis. Mature neuron marker of MAP2, astrocyte maker of GFAP, and oligodendrocyte maker of MBP were used for evaluated the differentiation fate of transplanted cells.

Hoechst 33342 labeled cells were noted in immunofluorescence analysis, and MAP2 expressed in part of transplanted cells (B-NS: 15%, A-NS: 8%); GFAP was only slightly expressed in B-NS group (3%); and MBP was not expressed in both groups. The transplanted cells integrated well in spinal cord tissues and migrated for at least 4 mm. The results indicated the transplanted cells could survive for at least 14 days in vivo after transplantation and high migration activity was noted. Besides, the transplanted cells tend to differentiate into neuron lineage than other types of glial cells and the different results of differentiation between in vitro and in vivo demonstrated the fate of cell differentiation depend on surround environment. Consequently, the cells harvested from generated neurospheres could be used for transplantation for the purpose in replacing host neuron cells in SCI.

This study highlight that both cBMMSCs and cADMSCs could be differentiated into neurospheres and neuron-like cells in vitro and might replace host neuron cells by using the cells derived from neurospheres after transplantation, and therefore, these cells are suitable candidates for cell transplantation. Further, cADMSCs form a more suitable cell source as faster proliferation rate; larger number of cells harvested from cADMSC-derived neurospheres; releasing higher levels of neurotrophic factors.

近年、重度の脊髄損傷(SCI)に対する幹細胞移植療法が注目されている。用いる幹細胞として、間葉系幹細胞(MSCs)は採取が比較的容易で、倫理的問題や腫瘍原性も少なく、経済的側面からも獣医療で最も実用的な細胞と考えられる。SCIへのMSCs移植は神経細胞の補充や神経栄養因子の分泌などの効果が報告されているが、近年、MSCsを神経幹/前駆細胞の集合体であるニューロスフェア(NS)へ分化させると、より効率的に神経細胞を誘導できることが示された。一方、MSCsの細胞源として骨髄や脂肪組織が広く研究されているが、SCI治療に最適な細胞源は不明である。そこで、本研究では犬の骨髄由来間葉系幹細胞(cBMMSCs)と脂肪組織由来間葉系幹細胞(cADMSCs)、および両MSCsから誘導したNSを用い、神経細胞分化および神経栄養因子の分泌を評価・比較し、犬のSCI治療への有用性を検討した。

まず第2章では、幹細胞としての性質を評価するため、cBMMSCsとcADMSCsの増殖能およびNestin、βIII-tubulin、GFAP、NCAM(神経系マーカー)とNANOG、OCT4、SOX2(幹細胞マーカー)の発現を比較した。さらに脂肪・骨・軟骨への分化誘導を行った。その結果、両細胞でNestin、βIII-tubulin、NCAM、OCT4、SOX2の発現が確認され、脂肪・骨・軟骨へも分化した。以上より、両細胞は幹細胞特性を持ち、神経系細胞への分化能を内在することが示唆された。一方cADMSCsの細胞倍加時間はcBMMSCsより短く、短期間でより多くの細胞を得られる利点が示された。

次に第3章では、cBMMSCsとcADMSCsからNSを誘導し、その作製効率と第2章と同様のマーカー発現を比較した。Nestin、OCT4、SOX2は定量的RT-PCRにより誘導前後の発現量も比較した。さらに、NSから神経細胞を誘導し、βIII-tubulin、GFAP、NF200、S100、MAP2、MBP、Nestinの発現を蛍光免疫染色で評価すると同時に、パッチクランプ法で電気生理学的評価を行った。その結果、両細胞からNestin陽性のNSが誘導され、Nestin、NANOG、OCT4、SOX2の発現が上昇した。cADMSCsの作製効率はcBMMSCsの約2倍であった。神経細胞誘導により神経細胞様の形態を示す細胞が得られ、βIII-tubulin、GFAP、S100、NF200、MAP2の発現も確認されたが、電気生理学的活性は陰性であった。機能的神経細胞の誘導には至らなかったが、NSで幹細胞マーカー発現が上昇し、神経様の細胞が誘導されたことから、NS誘導による神経細胞分化能の上昇が示唆された。

第4章ではcBMMSCs、cADMSCsおよび両MSCsから誘導したNS(B-NSおよびA-NS)におけるNGF、BDNFの遺伝子発現と細胞外分泌を比較した。さらに、両MSCsとPC12の共培養系を用い、神経形成に対する効果を検討した。その結果、NS誘導により、NGFの発現が低下した。BDNF発現はcADMSCsのみで検出された。両MSCsからのBDNF分泌は検出されなかったがNGF分泌は検出され、cADMSCsはより高い分泌能を示した。また、両MSCsはPC12の神経形成を促進させたが、効果に差はなく、NGF以外の因子の存在も考えられた。

最後に第5章では、B-NSおよびA-NSのin vivoでの分化を評価した。健常ビーグル犬の第2腰椎レベルの脊髄にHoechst33342で標識したNSを移植後、14日目に移植部位を摘出し、H&E染色とMAP2、GFAP、MBPに対する蛍光免疫染色を行った。その結果、移植部位にHoechst陽性細胞が観察され、MAP2陽性細胞はB-NS群で15%、A-NS群で8%、GFAP陽性細胞はB-NS群のみでみられ、3%が陽性であった。MBP陽性細胞は観察されなかった。また、移植した細胞は4mm以上脊髄内を移動しており、高い運動能が示された。in vitroの結果と分化傾向は異なったが、犬SCIでもNS移植が有用である可能性が示された。

以上の結果から、犬の骨髄および脂肪由来間葉系幹細胞から誘導したNSでは神経細胞分化能が上昇し、脊髄損傷に対する移植治療に有用である可能性が示された。また、細胞の増殖能やNSの作製効率、神経栄養因子の分泌能からcADMSCsはより有用性が高い細胞源と考えられた。

以上本研究は、脊髄損傷に対する間葉系幹細胞移植療法の有用性を、神経細胞分化能および神経栄養因子分泌能の面から示したものであり、学術上、臨床応用上貢献するところが少なくない。よって審査委員一同は、本論文が博士(獣医学)の学位論文として価値あるものと認めた。