Prenatal transplantation of human amniotic fluid stem cell could improve clinical outcome of type III spinal muscular atrophy in mice

Spinal muscular atrophy (SMA) is a single gene disorder affecting motor function in uterus. Amniotic fluid is an alternative source of stem cell to ameliorate SMA. Therefore, this study aims to examine the therapeutic potential of Human amniotic fluid stem cell (hAFSC) for SMA. Our SMA model mice were generated by deletion of exon 7 of Smn gene and knock-in of human SMN2. A total of 16 SMA model mice were injected with 1 × 105 hAFSC in uterus, and the other 16 mice served as the negative control. Motor function was analyzed by three behavioral tests. Engraftment of hAFSC in organs were assessed by flow cytometry and RNA scope. Frequency of myocytes, neurons and innervated receptors were estimated by staining. With hAFSC transplantation, 15 fetuses survived (93.75% survival) and showed better performance in all motor function tests. Higher engraftment frequency were observed in muscle and liver. Besides, the muscle with hAFSC transplantation expressed much laminin α and PAX-7. Significantly higher frequency of myocytes, neurons and innervated receptors were observed. In our study, hAFSC engrafted on neuromuscular organs and improved cellular and behavioral outcomes of SMA model mice. This fetal therapy could preserve the time window and treat in the uterus.

Spinal muscular atrophy (SMA) affects about 8.5 to 10.3 of 100,000 newborns 1 . This is a life-threatening neurodegenerative disease with autosomal recessive inheritance. Congenital deletion or mutation of survival motor neuron (SMN) 1 gene on chromosome 5q13 leads to depletion of SMN protein 2 . Then, neurons from the anterior horn of spinal cord become too little to induce muscle contraction 3 . Besides, another similar gene, SMN2, is essential to the production of SMN protein as well. However, SMN2 could only compensate partially for the absence of SMN1 due to its incomplete version. The number of SMN2 copies is correlated to the severity of SMA. According to the gestational age of onset time and level of motor function, SMA was classified into three types. Type I SMA caused abnormal inactivity in the later stage of pregnancy and several challenges in fundamental function, including breathing and swallowing. Respiratory failure usually results in death at two years old. Type II SMA allowed normal development until 6 months old but no walking ability. These patients often lead to a longer life, about 25 years old. Type III SMA merely contributed to difficulty in climbing stairs and getting up from the floor, but the limitation of motion would progressively grow and require a wheelchair. However, their lifespan is approximately normal 4 . No therapy could treat SMA completely nowadays. However, type III SMA patients have milder symptoms and universally survive to an average age 5 . Therefore, several therapies were investigated to cure this type of disorder. www.nature.com/scientificreports/ Certain stem cells have shown their ability to treat neuromuscular diseases. For instance, neural stem cells could differentiate into motor neurons and decrease the death of neuron in SMA murine model 6 . Embryonic stem cells demonstrated their potential of forming neuromuscular junction and differentiation into neuron to repair a musculocutaneous nerve in SMA rat 7 . Induced pluripotent stem cells were proved to delay the atrophy of muscle and retain neuromuscular function in SMA murine model 8 . In brief, these stem cells had the common potential of neural differentiation for treating neuromuscular diseases. By delivering neuroprotective factors and replacing degenerated cells, these stem cells could delay the progression of the disease and restore neuromuscular function 9,10 . Human amniotic fluid stem sell (hAFSC) shared the capability of differentiation and regeneration 11,12 . In previous studies, hAFSC held the promise to cure several diseases of the nervous system, including white matter disease and myelomeningocele 13,14 . Also, hAFSC transplantation could be combined with fetal surgery to be applied in mice with spinal bifida 15 . SMA represented another candidate disease to be treated prenatally with hAFSC transplantation in uterus 16 . With less cell count, irreversible damage from SMA could be prevented as soon as possible 17 .
To our knowledge, hAFSC has never been tested in SMA model as a potential therapeutic treatment prenatally. Prenatal diagnosis of SMA could be completed by amniocentesis between 15 and 20th gestational week. Simultaneously, stem cells could be harvested without controversial ethical issues 18,19 . Compared with postnatal therapy with NSC and iPSC transplantation, much less cell dose was required in prenatal therapy. Additionally, transplantation of hAFSC in early gestation could cease progression of pathophysiology and prevent irreversible damage as soon as possible in comparison with the administration of Nusinersen and Onasemnogene abeparvovec in the childhood. After the injection of hAFSC in uterus, the cells were able to migrate to several host organs 20 . Consequently, the survival of neurons and muscle could be elevated to improve the outcome of SMA. Besides, SMA mice model provided a useful in vivo system for studying the potential of hAFSC therapy. The objective of this study was to investigate the therapeutic effectiveness of hAFSC transplantation for SMA.

Results
Stem cell positive characteristics of hAFSC. Amniotic fluid was sampled by amniocentesis from pregnant women during 15th to 20th gestational week. Then, stem cells expressing MSC markers were isolated from amniotic fluid and cultured in an amniotic medium. By flow cytometry, these hAFSC were positive for several surface markers of MSC, such as CD44, CD73, CD90, CD105 and CD117 (Fig. 1). However, hAFSC were negative for CD45, an immune and hematopoietic marker. By immunocytochemistry staining for DAPI, the hAFSCs expressed enormous stem cell markers, including CD44 and c-kit (Fig. 2).  Engraftment of hAFSC in organs of SMA mice after transplantation. All organs of SMA mice were extracted for post-mortem analysis 12 months after transplantation. By flow cytometry, the muscle and liver of SMA mice with hAFSC transplantation were positive for CD73, a surface marker of MSC (Fig. 5). However, just a little signal of CD73 was detectable in the spleen, bone marrow, brain and spinal cord of SMA mice with hAFSC transplantation.   (Fig. 6). Based on ACD scoring system, the semi-quantitative outcome of muscle and liver was showed (Fig. 7). The score of liver was 1.50 ± 0.88 and muscle was 0.67 ± 0.54 by RNA scope. The muscle and liver of SMA mice with hAFSC transplantation showed CD73 as an indicator of successful engraftment of the human cell. There was a higher level of engraftment in SMA mice with hAFSC transplantation compared with the negative control group. Nonetheless, little CD73 was observed in the spinal cord of SMA mice with hAFSC transplantation.
Comparison of muscle and neuron in SMA mice after hAFSC transplantation. Samples of muscle and spinal cord were made into frozen sections. The muscle of SMA mice with hAFSC transplantation expressed laminin α and PAX-7 by immunofluorescent staining (Fig. 8). PAX-7/Laminin positive cells per high power filed of SMA mice with hAFSC transplantation was significantly higher (5.83 ± 1.81 vs 0.67 ± 0.86, [P < 0.05]). Besides, hAFSC expressed IL-6 and VEGF, a cytokine and a growth factor involved in the regenerative pathway of satellite   (Fig. 9b). In tissue from the spinal cord, SMA mice with hAFSC transplantation have more neurons per anterior horn with a significant difference by both H&E and methylene blue staining (Fig. 10). In all three behavioral tests, SMA mice with hAFSC transplantation has better performance during whole study.   21 . Amniotic fluid represented not only a diagnostic specimen but also a great source of stem cells. Consequently, hAFSC were collected to establish the therapeutic potential for SMA.
To analyze the capability of differentiation and regeneration of hAFSC, the first step is to identify specific MSC markers. In our study, hAFSC showed several MSC positive markers, including CD44, CD73, CD90, CD105 and CD117. In addition, hAFSC demonstrated the pluripotency according to the expression of c-kit, a pluripotent-related marker. Identification of shared MSC markers indicated similar differentiative characteristics of other stem cells. This potential of differentiation into a trilineage-relative cell could regenerate and repair damaged tissue 22 .  www.nature.com/scientificreports/ Successful engraftment of hAFSC in SMA mice was the first step in transplantation. After the injection into the fetal peritoneal cavity, hAFSC would migrate widely to localize in multiple organs, such as the liver, spleen, bone marrow, brain, spinal cord and muscle. To confirm the engraftment, CD73, a surface marker of MSC, was utilized to identify the existence of a human cell in SMA mice 23 . With this stem-cell-shared characteristic, hAFSC was proved to possess the potential of differentiation and repairment. Our study showed a good distribution of hAFSC with no miscarriage related to the procedure. Flow cytometry analysis indicated a higher engraftment rate in muscle and liver. Newly generated cells with typical morphology were showed by RNA scope. However, less CD73 was detected in some hematopoietic and nervous organs, including the spleen, bone marrow and brain. Because not the entire organ contained positive cells, some samples of the same organ demonstrated a negative outcome. In addition, impaired organs, including muscle and spinal cord in SMA mice, could provide engraftment advantage to appeal more Hafsc 24 . Last but not least, cell differentiation might be incomplete due to the short interval between transplantation and post-mortem analysis. In short, the success of engraftment played a fundamental role in hAFSC transplantation.
To reduce degeneration and maintain the function of a neuron, the mechanism of hAFSC transplantation for SMA should be investigated. Declined expression of SMN protein leads to loss of motor neuron in spinal cord 25 . For example, laminin, a fundamental protein for cell differentiation, was involved in the growth of nerve axon 26 . Neural injury from SMA decreased laminin with proliferated integrin receptors to promote neural regeneration. Moreover, PAX-7, a transcription factor for neural crest development, a regulated proliferation of precursor cells in myogenesis 27 . In our study, hAFSC showed enormous PAX-7, which is also observed in SMA mice after transplantation of hAFSC. This expression of PAX-7 showed successful engraftment on organs, such as muscle and liver, in these SMA mice. Declined PAX-7 were detected in satellite cells of SMA murine model. Consistent with the previous study, SMA mice expressed scarce laminin and PAX-7 28,29 . After hAFSC transplantation, significant  www.nature.com/scientificreports/ differential expressions of both proteins were noted by immunofluorescent staining. This result indicated that hAFSC transplantation was able to modify the RNA metabolism of neurons and muscle to deal with SMA 30 . For further evaluation of hAFSC therapy for SMA, successful engraftment should be identified at the cellular level. Neural dysfunction of SMA was derived from the reduction of motor neurons and neuromuscular junctions 8 . In line with the previous study, SMA mice represented fewer myocytes of muscle and fewer neurons of anterior horn 31 . After hAFSC transplantation, the number of both cells increased significantly. Newly generated cells demonstrated typical morphology by H&E staining. Additionally, a lower percentage of non-innervated receptors in SMA mice with hAFSC transplantation suggested improved integrity and survival of the entire motor unit. This outcome indicated hAFSC provided recovery function for neural injury of SMA. From a clinical aspect, three behavioral tests were conducted to analyze practical effectiveness of hAFSC therapy. In the previous study, SMA mice performed poorly in all Rotarod maintenance test, tilting test and grasping test [32][33][34] . After hAFSC transplantation, muscle power of maintaining, holding and grasping improved significantly in our study. Though these behavioral tests might be affected by body weight, rostro-caudal axis, and etc., these favorable results still showed the therapeutic potential of hAFSC transplantation for SMA.
Our data represented the usefulness of SMA mice as a model to study hAFSC therapy. However, there were still a few limitations in this study. First, low level of tissue engraftment was noted due to the small number of hAFSC transplantation. Nevertheless, only 1% level of protein expression was needed to improve some congenital diseases 35 . Therefore, the level of engraftment we achieved might be sufficient to deal with SMA in our study. Increasing the number of hAFSC would be a choice to elevate the level of engraftment. Nonetheless, the expansion of hAFSC leads to the proliferation of committed MSC progenitor at an inappropriate location. Secondly, this study lacked data of maternal and fetal immune response because we focused on the presentation of engraftment. Data of immune response in comparing allogenic and maternal human-leukocyte-antigen-matched hAFSC transplantation was necessary to investigate the immune rejection of hAFSC. Thirdly, intraperitoneal injection of hAFSC is the optimal route for transplantation. Direct injection might do harm to the spinal cord. Based on the capability of migration, hAFSC could engraft on target lesion and repair damaged tissue. In addition, multiple injections of hAFSC were allowed due to its inherent safety. However, this intervention remained a danger for the fetus in the first trimester. Improvement of fetal imaging would make hAFSC transplantation more feasible.
SMA impaired motor neuron during embryogenesis and reduced acquisition of myelinated axon after birth. In-utero treatment was required to prevent irreversible degeneration 36 . Prenatal hAFSC therapy preserved the time window to treat SMA at an early stage. In our study, hAFSC transplantation showed the potential of neurogenesis. Vascular endothelial growth factor, an angiogenetic factor, from hAFSC was crucial for these two modifications of the nervous system 37 . Its paracrine effect enhanced the repairing in neural injury from SMA 38 . The role of VEGF in hAFSC transplantation and combined therapy with antisense oligonucleotide could be further studied in the future 39 .

Conclusion
SMA does irreversible harm to the fetus in early gestation. This study demonstrated the therapeutic potential of prenatal hAFSC transplantation for SMA. By amniocentesis in the regular prenatal examination, hAFSC of a small number are collected. hAFSC therapy preserves the time window to treat in the uterus. To treat SMA, this therapy could be applied with antisense oligonucleotide in the future. . The dishes were incubated in an incubator that was constantly under 37ºC, 95% air, and 5% carbon dioxide. Once 60-80% confluence was achieved, the adherent cells were detached from the plate by 0.05% trypsin and 0.02% sodium-EDTA (Life Technologies, UK). Then, they were transferred into a bigger dish or dispensed evenly into 2 dishes. The specific surface antigens of hAFSC were characterized by flow cytometry analysis. The cells were stained with phycoerythrin (PE)-conjugated antibodies against CD44, CD73, CD90, CD105, and CD45 (BD PharMingen, CA) and protected from light for 30 min at 4 ˚C. Following incubation, the cells were washed twice with Phosphate Buffered Saline (PBS). Thereafter, the cells were analyzed with the Calibur flow cytometer (Becton Dickinson, Heidelberg, Germany).

Immunocytochemistry staining.
To confirm the identity of stem cells, hFASC with specific surface markers would demonstrate green fluorescent protein by immunocytochemistry staining (ICC) 18 . 1 × 10 5 hAFSC were seeded on sterilized glass coverslips which were placed in six-well tissue culture plates and incubated at 37 °C in a humidified CO 2 incubator overnight. The culture medium was aspirated from each well and fixed by 4% paraformaldehyde for 20 min at room temperature (RT). Then, the cells were permeabilized by 0.2% Triton X-100 in PBS for 5 min at RT and washed three times with PBS. Blocked with 5% fetal bovine serum in PBS for 30 min at RT, the cells were incubated in a humidified chamber overnight at 4 °C after appropriate dilution of human PAX-7 (LifeSpan BioSciences) to the glass coverslips. Eventually, the cells were stained with 546 antirabbit IgG antibodies (Santa Cruz Biotechnology) and DAPI for one hour. All images were detected by confocal microscopes (Leica,Wetzlar, Germany).

Transplantation.
After the sixth passage, hAFSC were collected and prepared to a final concentration of 1 × 10 5 cells per 10 μL in PBS 42 . On embryonic day 14 (E14) after the timed mating, pregnant mice would be sedated by vaporized Isoflurane-Vet 5% in oxygen (VetTech Solutions Ltd, UK) and maintained by vaporized Isoflurane-Vet 2% in oxygen. The belly was sterilized by Povidone-Iodine surgical scrub (Vetasept, UK). Via longitudinal laparotomy, the abdominal cavity was exposed and the fetal mice within the bicornuate uterus were counted. 1 × 10 5 hAFSC (treatment group) or PBS (control group) will be injected through the maternal uterus and the surrounding AF into the peritoneal cavity of each pup via small-gauge RN needles (27)(28)(29)(30)(31)(32)(33)Hamilton). The pregnant mice will be checked by the caretaker in the animal center every day, and most of them delivered these recipient mice on E20 (six days after transplantation). The new-born mice were sacrificed by CO 2 inhalation 12 months after the delivery. The postmortem analysis was performed to see the evidence of human cell engraftment.

Motor function tests.
Three behavioral tests were used to evaluate the motor function of mice two months after birth 43 . In the Rotarod maintenance test, the mice were placed on a Rotarod, while the rolling rate of the transverse rod increased and the time each mouse remained on the rod was recorded. In the tilting test, the mice were placed on a progressively tilting wooden platform. The highest degree of inclination at which the mice could hold on for five seconds was recorded. For the grasping test, the grasping force of both forelimbs was tested using the grasping machine. Each mouse underwent ten studies every two months and the five best results were averaged.

Detection of human cells by DNA extraction and PCR analysis.
All organs including heart, liver, kidney, lung, spleen, muscle, bone marrow, brain and spinal cord of the SMA Type III mice were extracted 44  Fresh frozen sample. Liver, Muscle, Spinal Cord was embedded in optimum cutting temperature compound (OCT, Leica) and sectioned at − 25 °C with a thickness of 15 µm and mounted on slides 47 . Fix the sections by 4% PFA for 15 min at 4 °C. Dehydrate the sections with 50%, 70%, 100% EtOH at RT for five minutes. Use the Immedge hydrophobic barrier pen to draw a barrier 2-4 times around each section. Place the sections in pretreating 1 at RT for 10 min, followed by pretreating 4 at RT for 30 min. Target probes human CD73, PPIB (positive control) and DapB (negative control) were then hybridized for 2 h at 40 °C in HybEZ oven. Preamplifier, amplifier, and chromogenic detection were as described above for cultured cells.
H&E staining. Frozen tissue maximal sections from the spinal cord (10-μm thickness) and quadriceps muscle (5 μm) were stained in hematoxylin and eosin (H&E) 48 . Neurons with a size greater than 200 µm 2 located in the anterior horn of the spinal cord were analyzed. We counted the number of neurons (only neurons showing nuclei) according to their cell size: 200-400, 400-600 and > 600 μm 2 , respectively. The size of myocytes from the quadriceps was analyzed. We calculated the number of myocytes according to their cross-section size.
Immunohistochemistry staining. The hamstring muscle was removed at the age of 12 months. The maximal section of tissue was placed in 4% paraformaldehyde at RT for 3 h and transferred to 15% sucrose for 1 day and 30% sucrose at 4 °C for another 3 days 49 . Then, the specimen was rapidly frozen in liquid nitrogen-cooled isopentane. Frozen serial sections of muscles were cut into 10 μm thickness and blocked with 5% serum. The CO) was injected into bilateral gastrocnemius and tibialis anterior muscles of mice (2 μl for each muscle) at the age of 12 months with a 30-gauge needle 50 . Mice were sacrificed 5 days after the injection, and their spinal cord was removed. After paraformaldehyde fixation, frozen serial sections of the whole lumbar segment of the spinal cord were cut at 10-μm thickness and processed for histological analysis.