Therapeutic potential of in utero mesenchymal stem cell (MSCs) transplantation in rat foetuses with spina bifida aperta

Abstract Neural tube defects (NTDs) are complex congenital malformations resulting from incomplete neurulation in embryo. Despite surgical repair of the defect, most of the patients who survive with NTDs have a multiple system handicap due to neuron deficiency of the defective spinal cord. In this study, we successfully devised a prenatal surgical approach and transplanted mesenchymal stem cells (MSCs) to foetal rat spinal column to treat retinoic acid induced NTDs in rat. Transplanted MSCs survived, grew and expressed markers of neurons, glia and myoblasts in the defective spinal cord. MSCs expressed and perhaps induced the surrounding spinal tissue to express neurotrophic factors. In addition, MSC reduced spinal tissue apoptosis in NTD. Our results suggested that prenatal MSC transplantation could treat spinal neuron deficiency in NTDs by the regeneration of neurons and reduced spinal neuron death in the defective spinal cord.


Introduction
Neural tube defects (NTDs) are complex congenital malformations resulting from incomplete neurulation during the 4th week of gestation in human. Spina bifida and anencephaly are the most common and severe forms of NTD affecting about 1 in 2000 live births worldwide [1]. In China, approximately 80,000 to 100,000 newborns each year are diagnosed to have NTDs [2]. Nowadays, many babies could survive with spina bifida attributable to improvements in medical and surgical management and care. However, most of these patients who survive with NTDs have multiple system handicap and limited life expectancy [3][4][5]. Prenatal repair of NTDs in foetuses has been shown to reduce the incidence of shunt-dependent hydrocephalus, as well as hindbrain herniation in some patients [3,4,[6][7][8]. However, foetal surgical repair of the defects did not improve neurological outcome of patients [4]. Individuals with lumbosacral spina bifida continue to experience motor and sensory dysfunction of lower limbs, dysregulated anal and urethral sphincters after birth [9].
Foetal cellular therapy is a treatment option for a variety of birth defects. There are several perceived advantages of in utero cell transplantation: (i ) the rapid growth of the foetus provides opportunity for engraftment and expansion of grafted cells, (ii ) the immature immune system of the foetus would not reject the transplanted foreign cells and (iii ) early treatment of disease is beneficial or critical for effective treatment [10]. Foetal cellular therapy has been employed in the treatment of congenital haematologic disorders and immunodeficiency disease [11][12][13], and also in experimental myelomeningocele (MMC) foetal lamb model [14].
Our previous studies have revealed that sensory and motor neuron deficiency is a primary anomaly coexisting with the spinal # There authors contributed equally to this work. malformation in foetal rats with spina bifida aperta. The intrinsic neuron deficiency of the spinal cord could explain the observation that surgical repair of the unclosed neural tube fails to improve neural functions [15][16][17]. Taken together all these suggest that neuron replacement therapy basis on neuron regeneration or cell replacement of the defective spinal cord would be a promising approach to achieve better functional outcome in spina bifida patients.
Mesenchymal stem cells (MSCs) are multipotent stem cells and could differentiate into different cell types such as skeletal muscle, vascular, neuron, astrocyte, intestinal, bone and liver cells in response to different factors [18,19]. A number of studies indicated that MSCs trans-differentiated into neurons both in vitro and in vivo [20][21][22][23][24]. In particular, MSCs are advantageous over other types of stem cells such as neural stem cell (NSC) in neuroregenerative medicine due to their convenient isolation, lack of significant immunogenicity permitting allogenic transplantation without immunosuppressive drugs, lack of ethical controversy and their potentials to differentiate into tissue-specific cell types with trophic activity [18]. Spina bifida develops starting from the 4th week of gestation, and it can be diagnosed with prenatal ␣-fetoprotein screening and ultrasonography before the 12th week of gestation [4,25]. Prenatal diagnosis of spinal bifida allows early intervention, which could ameliorate the aggravation of the deficits [26]. In this study, we evaluated for the first time the therapeutic potentials of a combination of surgical repair and in utero MSCs transplantation in the treatment of spina bifida in a rat model.

Experiment animals
Outbred Wister rats of 10-12 weeks of age (250-300 gm) and 4 weeks (about 100 gm) of age were purchased from animal centre of China Medical University. The appearance of vaginal plugs in the female rat the morning after mating was timed as the embryonic day 0 (E0). Spina bifida aperta were induced with a single intragastric retinoic acid (140 mg/kg body weight; Sigma-Aldrich, St. Louis, MO, USA) administration on E10 as previously described [27]. All the animal experiments were performed with the approval obtained from the ethics committee of China Medical University.

Isolation, culture expansion, in vitro neuronal differentiation and transfection of bone marrow-derived MSCs
MSCs were isolated from 4-week-old Wister rat following previously published protocol [22]. Rat MSCs were cultured in DMEM/F12 (Gibco, Brooklyn, NY, USA) supplemented with 10% foetal bovine serum (FBS; Hyclone, Shrewsbury, NJ, USA) and 100 IU/ml penicillin to 100 g/ml streptomycin (Gibco) on 25 cm 2 tissue culture flasks (BD Biosciences, Franklin Lakes, NJ, USA). Primarily isolated MSCs were defined as P0, at confluency, cells were passaged (1 in 2 dilution) with fresh medium, and MSCs were cultured till passage 15, different passage MSCs were used for transplantation. Cultured MSCs expressed CD90, CD44, CD73 and CD29 but not CD34 and CD45 as revealed by flow cytometry using specific antibodies (551401, 550974, 551123, 562154, 560932, 554878; BD Biosciences) following previously published methods [28]. CD90 expression was used to estimate the purity of MSCs. For in vitro neural differentiation, P3-P5 MSCs were treated with BDNF alone (10 ng/ml) or BDNF (10 ng/ml) with retinoic acid (0.5 M) for 14 days [29]. Twenty-four hours before transplantation, MSCs were transfected with eGFP expression adeno-5 vector (SinoGenoMax.Co., Ltd, Beijing, China) (100 pfu/1 cell), for the visualization of MSCs after transplantation into rat foetus. Before transplantation, cells were trypsinized, centrifuged, resuspended in small aliquot of fresh medium, GFP positive cell numbers were counted, the final concentration were adjusted to give a serial concentration ranged from 7000 to 35,000 cells/l.

Foetal surgery and microinjection
We developed a new approach to transplant cells into foetuses with combined techniques of foetal surgery, microinjection and microsurgery. Retinoic acid treated pregnant rats of E16, E17 or E18 were anaesthetized with pentobarbitone sodium (40 mg/kg body weight). An incision was made in the abdominal wall, and the uterine horn was exteriorized. To relieve uterine spasm, the uterus was covered with wet gauze immersed with warm physiologic saline and atropine (0.1 mg/kg body weight) was given intraperitoneally. Under the microscope the position of lumbosacral spine of foetus was identified through the wall of uterus. Then 7-0-nylon purse-string suture and a small incision were made on the wall of uterus. The amniotic sac was opened and the defective region of spinal cord was exposed. GFP transfected MSCs suspension (0.2 l/injection site) was injected into the defective region of spinal cord with a micropipette (internal tip diameter 100 m) connected to a Hamilton syringe. Micropipettes for injection were made from borosilicate glass capillaries (model GD-1; Narishige Scientific Instruments, Tokyo, Japan) by using a micropipette puller (model PB-7; Narishige Scientific Instruments). After MSCs injection, the foetuses were returned to the uterus, and the wound of the uterus was closed (Fig. 1). In average, two to three foetuses could be injected in one dam without compromising the survival of the foetuses. The pregnant rats recovered from the anaesthesia within 1 hr and were returned to their home cage.

Survived cell observation and counting after transplantation
The pregnant rats on E20 were reanaesthetized with an overdose of pentobarbitone sodium, and the foetuses that had been transplanted were harvested by caesarean section. Foetuses were perfused transcardially with 15 ml physiologic saline followed by 25 ml 4% paraformaldehyde. The lumbosacral spinal column containing muscle, spinal cord and subcutaneous tissue was dissected and post-fixed in the same fixative for 24 hrs at 4ЊC, followed by cryopreservation in 20% sucrose for 24 hrs. The spinal column were then sectioned into 30 m serial transverse sections, and all GFP positive MSCs in the spinal column were observed and counted under fluorescence microscope (Nikon, Tokyo, Japan). All the sections were also stained with DAPI, and only those GFP positive cells that contained a nucleolus were counted to avoid the same GFP cell being counted more than once. The MSCs survival rates were calculated as the average of survived cell numbers/totally injected cell numbers/spinal cord. Image was taken with C1 confocal microscope (Nikon, Tokyo, Japan) or fluorescence microscope connected to a CCD camera (Nikon). Sections with GFP positive MSCs observed were marked and kept at -80ЊC in the dark for further immunofluorescence or TUNEL study.
For cell staining, microscopic slide with MSCs seeded 24 hrs before were fixed with 4% paraformaldehyde for 30 min., and were analysed by immunofluorescence. Ten (40ϫ magnification) non-contiguous and non-overlapping random chosen microscopic fields distributed evenly across the slide were examined, the number of MSCs expressing the specific neuron/glia marker was counted by observers who were not informed of the identities of the experiments. The percentages of MSCs expressing specific marker were determined and reported as the number of positive cells/total cells.

TUNEL analysis
TUNEL analysis and immunofluorescence for GFP was performed using In Situ Cell Death Detection Kit (Roche, Indianapolis, IN, USA) and with mouse anti-GFP antibody (Beyotime Institute of Biotechnology) according to manufacturer's protocol. Briefly, after antigen retrieval the sections were blocked with PBS containing 10% FBS and 0.1% Triton x-100, and incubated with anti-GFP antibody (1:200) at 4ЊC overnight. Then the sections were washed and incubated with TUNEL reaction mixture and TRITC-anti mouse IgG at 37ЊC for 1 hr. After washing, sections were stained with DAPI and mounted with anti-fade mounting medium. The images were taken with C1 confocal microscope (Nikon).

Real time PCR analysis
Eleven spinal cords transplanted with MSCs (MSCs transplantation group), five spinal cords injected with fresh medium (blank injection group) and 12 spinal cords without injection (no injection group) of E20 foetus with spina bifida aperta were dissected, and total RNA were isolated with RNeasy mini kit (Qiagen, Hilden, Germany) according to manufacturer's instruction. For cultured cells, RNA was isolated from at least 5 ϫ 10 5 cells. All RNA were quantified, and RNA with A260 nm/A280 nm ratio less than 1.8 was discarded. Total RNA (2 g) was reverse-transcribed at 37ЊC for 15 min., 85ЊC for 5 sec. and 4ЊC with random 6-mer oligos (50 pmol), oligo dT primer (25 pmol) and the PrimeScript RT reagent kit (Takara, Shiga, Japan) in 20 l of reaction solution. RT-PCR products were diluted 1:10. Diluted RT-PCR products (2 l) were subjected to quantitative real-time PCR RNase-free water, the reaction were performed in triplicate in LightCycler ® Capillaries (04929292001; Roche) with a Light Cycler 1.5 system (Roche). PCR was performed as follows: pre-denaturation at 95ЊC for 30 sec., 45 cycles of denaturation at 95ЊC for 5 sec., annealing at 55ЊC-63ЊC for 20 sec. Negative control (without reverse transcriptase) gave no signal. The relative level of gene expression for each sample was calculated using the 2 Ϫ⌬⌬ct method, and expressed as a fold induction compared with no injection control from the same pregnant rats after ß-actin normalization. GFP expression was analysed with Premix Ex Taq ™ (Perfect Real Time) kit (Takara). For GFAP and neurofilament expression, PCR products were also analysed with 3% agarose gel. Detailed information of primers and probe used in this study were listed in Table 1.

Statistical analysis
All analyses were performed in a double-blind manner. All values are expressed as mean Ϯ SEM. Student's t-test and Bonferroni's test were used for single and multiple comparisons, respectively. Chi-square test was used for comparisons of foetuses survival rates. P values less than 0.05 were considered to be statistically significant.

Culture, character and in vitro neuronal differentiation of MSCs
In culture, MSCs displayed fibroblast-like, spindle-shaped morphology. Cultured MSCs were CD90, CD44, CD29 and CD73 positive but negative for CD34 and CD45 (Fig. 2A). The purity of MSCs in culture was examined by CD90 expression. The purity of MSCs was enriched from 65% at P0 to 80% at P1, and Ͼ90% from P2 to P15 (Fig. 2B). No obvious morphological changes were observed within the culture period. We next examined nestin (neural precursor cell marker) and BDNF (brain-derived neurotrophic factor) expression in cultured MSCs (Fig. 2C). The percentages of P3-P6 MSCs expressing nestin and BDNF were around 3-4% (nestin, 3.01 Ϯ 0.17%; BDNF, 4.25 Ϯ 0.13%). The percentages of MSCs expressing these markers increased with prolonged culture (P10-P14) (nestin, 6.1 Ϯ 0.76%; BDNF, 6.64 Ϯ 0.69%) (Fig. 2D), which was indicative of increasing neuronal differentiation potential of cultured MSCs. Elevated BDNF expression in prolonged cultured MSCs was further confirmed by quantitative real time RT-PCR analysis (Fig. 2E). We also studied  (Fig. 2F). Real time PCR analysis showed weak GFAP and neurofilament expression in longer cultured MSCs, and the expression greatly increased in BDNF and retinoic acid treated MSCs compared with BDNF treatment only ( Fig. 2G and H).

Survival of transplanted MSCs in foetal rat spinal column
A total of 86 pregnant rats received foetal surgery and microinjection (see Fig. 1 and Material and methods section) on E16, E17 and E18, and 5 died before E20, the other 81 were killed on E20 for spine sample collection. From the 81 pregnant rats, a total of 195 transplanted foetuses were harvested, and 152 foetuses (77.9%) were alive. The overall survival rates in different operation groups were tabulated (Table 2), and better survival was achieved in the operation groups performed on later development stages (E17 or E18) as compared with E16 group, but it did not reach a statistical significance (P ϭ 0.67 E16 versus E17; P ϭ 0.25 E16 versus E18). Lumbosacral spinal columns of 136 foetuses were sectioned, and spinal cords of the other 16 foetuses (including 5 foetuses received blank injection) were processed for RNA isolation. Transplanted MSCs (GFP positive) mainly located in the defective region of spinal cord and displayed different cell shapes such as round, long and spindle, or with long processes (Fig. 3A). In average, 22.8 Ϯ 23.1% (n ϭ 96) of transplanted MSCs survived in the spinal column. In addition, GFP cells were occasionally detected in regions away from the injection site including the normal spinal cord, muscle, subcutaneous tissues and ganglion (Fig. 3B). These GFP cells could represent transplanted cells those migrated out of the injection site and/or those delivered outside of the spinal cord during the prenatal cell delivery procedure.
To improve the survival of MSCs in rat spinal column, we examined factors that may contribute to MSCs survival after transplantation. Firstly, we examined the impact of number of the injected cell. MSCs survival was found to be 17.82 Ϯ 9.   L versus H) (Fig. 3C) (Fig. 3D). Thirdly, we evaluated the effects of the embryonic stages of recipients on the survival of transplanted MSCs. Less MSCs (22.75 Ϯ 12.2%) survived in E16 foetal spinal column than in E17 (38.7 Ϯ 25.1%) and E18 (34.6 Ϯ 23.5%) group, again the differences did not reach a statistical significance (P ϭ 0.1, E16 versus E17 and P ϭ 0.18, E16 versus E18, respectively) (Fig. 3E).

Differentiation of transplanted MSCs in spinal column with spina bifida aperta
To ascertain if transplanted MSCs generated different types of cell in the recipient spinal column, we examined the expressions of specific cell markers in transplanted MSCs. During the 2-4 days of embryonic development, transplanted MSCs expressed markers for neural precursor cell (nestin), neuron (neurofilament), neurogliocyte (GFAP), motoneuron (P75NGFR) and myoblast (MyoD) (Fig. 4A). Systematic view of the images was shown in Figure S1. In line with the increasing neuronal differentiation potentials of MSCs in prolonged culture (Fig. 2D), prolonged cultured MSCs generated more neurons in the spinal column ( Fig. 4B and Table  3). Development stages of the recipients also affected neural differentiation of MSCs (Fig. 4C). MSCs in E16 spinal column displayed significantly higher neuronal marker expression as compared with E18 group (Table 3), there is a tendency of higher neural maker expression of MSCs survived in E16 spinal column compared with E17 group, but the differences are not significant (P ϭ 0.1, 0.16, 0.17 for nestin, GFAP and NF, respectively) ( Table 3). Our results indicated that transplantation of MSCs into rat foetus on E16 favoured MSCs differentiation.

Neurotrophic factors induction by MSCs transplantation
Besides neuron replacement, the neurotrophic factors induced by MSCs transplantation also play an important neuroprotective role in spina bifida aperta. After transplantation, 22.7% (69/304) and 21.6% (54/250) MSCs expressed BDNF and NGF, respectively, BDNF expression is much higher than before transplantation (Figs 5A and 2D). As shown in Figure 5B, immuno-reactivity of BDNF and NGF was localized to both the GFP immuno-positive transplanted cells, and in the adjacent host spinal cells. Taken all these suggested that transplanted MSCs not only expressed BDNF and NGF themselves, but also induced BDNF and NGF expression in the host spinal tissue. In MSCs transplantation group (successful transplantation was confirmed by real time analysis of GFP expression in dissected spinal cord), expression of neurotrophic factors in the spinal cord increased markedly (Fig. 5C). NGF and NT-3 expression were significantly higher in the transplanted groups (P ϭ 0.03 and 0.017 compared with blank injection group, respectively). BDNF expression was also elevated in MSCs transplantation group, although it was not significantly different from blank injection (P ϭ 0.26), which could be attributable to the relatively high basal BDNF expression in the spinal tissue [31].

Reduction of apoptosis after MSCs transplantation
Apoptosis plays an important role in the development of NTDs [32,33]. We studied the apoptosis of whole spinal column after transplantation by TUNEL analysis. Apoptotic cells were frequently localized at the region of defective spinal cord. In contrast, apoptotic cells were rarely found at the defective spinal cord region around the transplanted MSCs (Fig. 6A). Apoptosis was not observed in transplanted MSCs. In addition, the expression of caspase-3 and Bcl-2 in transplanted spinal cord was analysed by real time PCR analysis. Caspase-3 expression was significantly lower in MSCs transplantation group as compared with blank injection group (P ϭ 0.03) and no injection group while Bcl-2 expression was not significantly different in the three groups (Fig. 6B).

Discussion
Congenital malformations affect approximately 5% of all live births every year [34,35]. Several of these are associated with in utero foetal demise, significant neonatal morbidity and mortality. Despite early diagnosis of the condition in most of the cases, treatment options for the affected foetuses are still limiting and not satisfactory. Today, open IUMR (in utero MMC repair) is being offered to selected mothers in several centres [3,8,[36][37][38][39], and over 330 such interventions have been performed since 2000. It was reported that prenatal repair of the spinal cord defect greatly reduced the mortality of this birth defect, suggesting that attenuation of the progress of the damage of the defective spinal cord at early pregnancy is beneficial to the foetus survival. However, sensorimotor function could not be restored in most of the cases [4]. We have previously shown that sensory and motor neuron deficiency is a primary anomaly coexisting with the spinal malformation, which suggests that poor functional outcomes of surgical repair of the unclosed neural tube could be due to deficiency of sensory and motor neurons [15][16][17].
In this study, we examine the feasibilities and potentials of combination of surgical repair and in utero stem cell transplantation aiming at the attenuation of the progress of the malformation and possible neuron regeneration for the treatment of spina bifida. It is apparent that the survival of the foetus and the mother is the first challenge in foetal surgery especially in earlier embryonic development stages as E16 and E17. Our new microsurgical and microinjection approach achieves over 75% survival of the operated foetuses and 94% survival of the mothers. Transplanted MSCs expressed markers of neural precursor cells, neurons and    , injected with same amount of medium (blank injection, n ϭ 5) and without injection (no injection, n ϭ 12). Values presented fold expression differences compared with no injection group. P ϭ 0.03 compared with blank injection. neurogliocytes in the spinal column of rats with spina bifida aperta. The highest neural differentiation (19.2% MSCs differentiated into neural precursor cells, 16% into neurons and 33% into neurogliocytes) was obtained with the transplantation of P10-P14 MSCs into E16 foetuses after 4 days ( Fig. 4B and C, Table 3). MSCs have also been shown to give rise to neurons in the nervous system of adult or neonatal animals [20,22,[40][41][42]. However, the efficiency of MSCs differentiation into neurons was reported to be only 3% at 28 days post-transplantation in adult mice [40]. The higher neural differentiation of MSCs in the foetal rat spinal column suggests that the rapidly growing foetal nervous system favours the differentiation of transplanted cells into neuronal lineage. Expression of neurotrophic factors from transplanted MSCs and from the surrounding rat spinal tissue was found associated with reduced apoptosis of the rat spinal tissue, which is indicative of neuroprotective effect of MSCs on neural tissue. In line with our observations, neuroprotective effect of transplanted MSCs has also been previously reported [23,43,44]. In addition to neuron replacement potential of MSCs, neuroprotective functions of transplanted MSCs is also critical in the success of therapy in the context of spina bifida aperta since apoptosis plays an important role in the development of NTDs. Reduction of apoptosis of the spinal tissue could be partly explained by inhibition of caspase-3 activity by neurotrophic factors secreted from the MSCs and the surrounding spinal tissue [45,46]. In addition, Sadat et al. previously reported that adipose tissue-derived stem cells have a marked impact on anti-apoptosis through secreting IGF-I and VEGF, indicating there exits other stem cell sources than MSCs having anti-apoptotic effects [45]. Our data indicated that (i) MSCs differentiated into neurons in spinal cord, (ii) transplanted MSCs and the surrounding host spinal tissue secreted neurotrophic factors and (iii) apoptosis was reduced in the spinal tissues with transplanted MSCs. Functional outcome of prenatal MSCs transplantation especially in the recovery of sensorimotor functions has not been determined in present study due to the fact that those rat foetuses with spina bifida aperta died after birth. In future, we shall modify our surgical intervention so as to treat the foetuses at even earlier stages (E14 or E15) aiming to stop the damage progress of the unclosed neural tube at earlier stage of development so as to enhance the survival of MSCs implanted rats after birth, thus allowing assessment of the functional outcome of our treatment.
Foetal surgery and NSC delivery in experimental MMC foetal lamb model enhanced the survival and reduced the incidence of major paraparesis [14]. The transplanted NSCs remained as undifferentiated stem cells and produced neurotrophic factors in the recipient spinal cord. The beneficial effects of the NSC on the MMC foetal lamb was likely attributable to the production of neurotrophic factors from the transplanted NSC. It is still not known if transplanted NSC will differentiate into neurons in long term and replaced the missing neurons in the defective spinal cord.
In conclusion, we have established a new therapy for the potential treatment for spina bifida aperta, and this is the first attempt in the prenatal treatment of spina bifida aperta using MSCs transplantation. We provide evidence indicating that the transplanted MSCs could ameliorate sensory and motor neuron deficiency in spinal malformation by differentiation into neurons in situ and reduction of spinal neuron apoptosis. MSCs are multipotent and could differentiate into various cell types in addition to neurons. In future, the application of MSCs in cell replacement therapy for other congenital anomalies especially complicated malformations besides NTDs should be explored.  Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.