Introduction
The transplantation of neural stem progenitor cells (NSPCs) is considered a promising approach to the treatment of a range of central nervous system (CNS) disorders, including spinal cord injury [
1,
2], brain infarction [
3‐
5], amyotrophic lateral sclerosis (ALS) [
6,
7], and Parkinson’s disease [
8]. In countries in which the clinical use of fetal NSPCs is not permitted due to ethical limitations, induced pluripotent stem cells (iPSCs) represent a potential alternative source of NSPCs for use in cell therapy research and development. In Japan, iPSC stocks [
9,
10] are being established from peripheral blood mononuclear cells (PBMCs) of donors with a range of immunologically preferable genotypes. However, before such cells can be used in clinical applications, the safety of transplanted cells must be determined.
Transplant safety issues include infection, immunological problems such as rejection, complications resulting from drugs such as immunosuppressants, and complications resulting from unexpected migration or transplant behavior [
11,
12]. As for stem cells, which have the potential to develop into a variety of mature tissue types, efforts must be made to manage the risk of contamination by undifferentiated pluripotent cells, or the transformation of graft-derived intermediate progenitors into malignant tumor cells [
13‐
16].
Malignant tumors occasionally exhibit immature and embryonic-like structures, and normal embryonic cells share some characteristics with malignant tumor cells; i.e., normal developing tissues in which stem cell multiplication occurs exhibit significant mitotic activity, similar to that in malignant tumors. Under such conditions, the cellular mitotic index is not helpful in distinguishing stem cells from malignant tumors. NSPCs for transplantation therapies also exhibit characteristics of less developmentally mature cells, and thus it is necessary to classify the histology of transplanted cells by their developmental characteristics in order to distinguish them from the malignant transformation of transplants.
In the present study, we induced NSPCs from integration-free human peripheral blood mononuclear cell (PBMC)-derived iPSCs (iPSC-NSPCs) using two different protocols, compared their in vitro properties, and also compared their in vivo histology by transplanting them into intact striata or injured spinal cords of immunodeficient (NOD/Shi-
scid, IL-2R γ null (NOG) [
17] or NOD/
scid [
18]) mice. Our histological categorization may serve as a useful tool for predicting and describing the performance of NSPCs for future quality evaluations of cell products for future transplantation therapy.
Discussion
The risk of teratoma formation by transplanted iPSC-derivatives is widely recognized, and many attempts have been made to minimize such risks [
26,
27]. Although the transformation of iPSC-derived products has not been studied as extensively, a number of reports have showed the transformation of iPSC-derived intermediate progenitor cells as the result of genetic modification [
14,
28,
29], transgene activation [
30], or epigenetic events [
31]. We used integration-free iPSCs to minimize the risk of genetic modification and/or transgene re-activation, which were observed in a previous report by our group [
14]. However, some of the NSPCs from the three integration-free iPSCs used in this study retained proliferative characteristics in vitro and in vivo, which appeared to be attributable to karyotype abnormalities or
de novo CNVs that occurred during differentiation and culture processes. As the three iPSC-lines utilized in the present study were induced from blood samples from a single donor, the different genomic abnormalities observed in our iPSC lines may have occurred either during the reprogramming process or in cell culture. This suggests that even when integration-free iPSCs are used, it is not possible to completely eliminate the risk of genetic instability during NSPC production. This also indicates that the neural induction protocols that we used cannot eliminate all cells with genetic instability, and that genetic differences that emerged during the reprogramming process from the original iPSCs cannot be fully standardized. To control the proliferation of the derivatives in vivo, we thus should use iPSC lines without karyotype abnormalities or genomic instabilities.
Culture duration is also important in controlling the genomic and epigenomic character of cultured cells. Young iPSCs may be immature and unstable, but it has also been reported that culture-induced genomic and epigenomic aberrations can occur at any stage, as the undifferentiated state of iPSCs is inherently unstable and sensitive [
32]. We decided to use iPSCs as early as 11 passages for neural induction from feeder-free cultured iPSCs (1210B2 and 1231A3), in order to lower the risk of genetic abnormality associated with long-term culture process. Induced pluripotent stem cells established using the same method have been reported to express pluripotency-markers by passage 5 [
33]. We also confirmed expression of pluripotent markers in our cells by passage 11. Additionally, we needed to cultivate the cells for this length to obtain a sufficient number of cells for further use. With respect to on-feeder iPSCs (1201C1), we needed to culture these until passage 19 in order to obtain sufficient numbers of cells. The adequate passage numbers of iPSCs or NSPCs may thus differ depending on the culture method and the purpose of usage of cells, a condition that should be evaluated for each case.
It has been reported that iPSCs tend to acquire epigenetic and genetic modifications, including CNVs, during the reprogramming process and subsequent cell culture [
34‐
36], and that these genetic alterations may occur in genomic sites related to cancer development. It is impossible to eliminate the risk of genetic abnormality; however, our results with the 1210B2 EB- and NR- NSPCs indicate that the proliferative capacity of the NSPCs can be lowered by selecting cells with few genetic abnormalities. Careful assessment of cells, including their genomic stability, may thus be useful for reducing risks in transplantation.
Considering the differentiation of the induced NSPCs in vivo, the differentiation tendency was affected by the induction protocols used. It has also been reported that cellular migration or survival is affected by the injection site, timing of the transplant after injury, as well as by transplant dose [
37‐
39]. These properties are relevant to the normal mammalian CNS developmental process. During mammalian CNS development, NSPCs gain temporal and positional identities, and thus acquire defined and limited plasticity, such as neurons or glial cells [
40,
41]. We categorized the maturation of transplanted NSPCs by their formation of different developmental tissues, and found that the environment of the cells’ terminal destination site affected their developmental fate. When transplanted cells engrafted outside the CNS parenchyma, they nearly always underwent maturational arrest or delay of neural tissues, which were represented by BLT and UDNT. BLTs are the most primitive tissues, and when they were rarely found within the parenchyma, they exhibited extensive neural differentiation and maturation characteristics (Additional file
9: Figure S3A, S3B). In support of these developmental events, transplanted cells within the parenchyma of the CNS often formed neural tissues with adequate maturation (DNT).
These results indicate that the control of cellular migration is useful in controlling the fate of transplants. Considering the potential for clinical application of NSPC transplantation, the preferable histology would be different according to each purpose. Animal models with environments similar to those of recipient patients will be most informative with respect to the engraftment style and the migration potential of the NSPCs. However, in addition to NSPC transplantation using animal models that precisely reflect the disease environment in clinical settings, other conventional animal models may be used for specific purposes. For example, when assessing NSPCs to be used in the development of SCI treatments, we should select animal models of spinal cord injury to evaluate cellular maturation stages. Transplantation into intact animal brain, which is much easier and enables the assessment of larger numbers of cells at a given time, would also be informative in the assessment of the proliferative capacity of transplanted cells, as the proliferative tendencies of each cell line were similar in the intact brains and injured spinal cords evaluated in our study.
Interestingly, it appeared that neural tissues were regularly generated in developmental stages from particular regions (e.g., central canal, central canal-like cyst, and pia mater), particularly in injured spinal cords, whereas they frequently formed and grew in a round shape without associations with ventricles in non-pathologic brains. Other studies also indicate that cells proliferate better under pathologic conditions [
38,
42]. We speculate that under pathologic conditions, transplanted cells may favor migration into appropriate areas from which tissue or organ regeneration can be initiated, as is observed in fetal development, thus resulting in better engraftment.
The aim of the present study was to classify the histology of transplanted human iPSC-NSPCs in the mouse CNS by developmental potential to aid in the identification of malignant transformations of transplants during safety assessment of cells intended for use in transplantation therapy. Distinguishing between embryonic tissues and tumors by microscopy is difficult, as normal developing tissues in which stem cell multiplication occurs have significant mitotic activity similar to that of malignant tumors. Furthermore, in humans, especially in children, it is known that tumors often arise from remnants of embryonal tissues; however, the majority of such tissues involute or differentiate into mature tissues [
43].
The definition of a tumor is a mass of cells that proliferate without relation to pattern or rate of the growth of the part in which it is located. This unlimited cellular growth may lead to the development of malignant properties, including invasion of surrounding tissues and metastasis to other organs [
44]. In this study, we concluded from the following observations that none of the transplanted cells developed into malignant tumors. 1) We observed that the transplanted cells generated nervous tissues with differentiation regularity and limited growth, which we interpreted as a recapitulation of zonal formation in CNS development, as represented by characteristic DNT features. 2) Some UDNTs showed extensive growth along the meninges. This may suggest that these cells have unlimited proliferation potential and immature properties. However, the margin of this tissue was smooth and tended toward neural maturation, and no invasion of adjacent tissues has been observed so far. We thus classed this histologic subtype as tissue overgrowth accompanying cellular immaturity. 3) BLT induces the emergence of embryonal tumors, such as neuroblastomas, medulloblastomas, primitive neuroectodermal tumors, and Wilms tumors. The BLTs observed in the current study were a confined tissue that was occasionally found in the meninges. Additionally, BLT lacks mitotic figures and can differentiate into neural and mesenchymal lineages. These features may more closely resemble perilobar nephrogenic rests (PLNR), which are embryonal tissue remnants observed in kidneys of patients with overgrowth syndrome [
45]. BLT may differentiate into mature tissue or regress, as is observed in PLNRs. Only a few PLNR types develop into Wilms tumors [
46].
Considering the implications for these observations in human, we predict that if transplanted iPSC-derived cells can become cancer cells, they would be BLT-derived embryonal tumor cells that tend to localize in the meninges rather than an adult cancer type. We did not observe any cancerous lesions in the present study, but recognize that any cell has the potential to develop into cancer. Longer-term observation is thus required to determine ways in which malignant tumorigenesis can be inhibited.
Several approaches to the minimization of transplantation risk have been reported. In addition to eliminating pluripotent cell contamination, controlling the proliferative characteristics of cells by pre-treating them with MitomycinC [
47] or γ-secretase [
48] may be effective. Transplant ablation by transgenic
HSV-tk and gancyclovir administration [
49], or a caspase-based artificial cell death switch (iCaspase-9) [
50] may also be effective in cases of malignancy. In all cases, the correct diagnosis of acceptable or unacceptable histology must be made. From this perspective, our histological classification is also an effective approach for optimizing the safety of NSPC transplants.
These results provide important histological insights into the transplantation of human NSPCs into the CNS of animal models, with a focus on safety issues confronting future cell transplant therapeutics.
Methods
Additional details regarding several of the protocols used in this work are provided in the Additional files
12,
13,
14 and
15: Supplemental Experimental Procedure and Tables S7-S9.
Cell culture
Three lines of integration free human PBMC-derived iPSCs (1210B2, 1231A3, and 1201C1), which were established from ePBMCs® from the Cellular Technology Limited (OH, USA) at Center for iPS Cell Research and Application (CiRA: Kyoto, Japan) by an integration-free method [
51], were used. The 1210B2 and 1231A3 iPSCs were cultured with a feeder-free protocol [
33], and the 1201C1 iPSCs were cultured with an on-feeder protocol that uses SNL feeder cells. They were induced into NSPCs as previously described [
52,
53] with two slight modifications. Briefly, in the first protocol, the NSPCs were induced directly from embryoid bodies (EBs) by a protocol that consists only of a floating culture. In the second protocol, the EBs were adhered to laminin-coated culture dishes on day 7, and they subsequently formed neural rosettes (NRs), which were picked on day 14. We refer to NSPCs induced directly from EBs as EB-NSPCs, and those induced from the NR phase as NR-NSPCs. The NSPCs were expanded using the neurosphere culture technique [
23,
54].
Cellular analysis
Detailed experimental procedures were described in the supplemental materials. For the microarray analysis, total RNA was analyzed using Human Genome U133 Plus 2.0 Arrays (Affymetrix Inc., Santa Clara, CA) according to the manufacturer’s instructions. RT-PCR analysis was performed, as described previously [
54]. Cell surface marker expression was analyzed with a BD FACS Verse (BD Biosciences, San Jose, CA). For cell cycle analysis, the DNA contents of the cells were analyzed by propidium iodide staining with an EC800 Analyzer (Sony Biotechnology Inc., Tokyo, Japan). The proliferation assay was performed by measuring ATP with a CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI) on an ARVO X5 Multilabel Plate Reader (PerkinElmer, Waltham, MA). The cellular doubling time was calculated from the intensities of two sampling points in the logarithmic growth phase, as previously shown [
23]. The karyotype analysis was performed by conventional Giemsa staining and G-band analysis, and diagnosed as issued in the 2013 international system for human cytogenetic nomenclature (ISCN 2013) [
55]. CNV was analyzed using a CytoScan HD Array (Affymetrix Inc., Santa Clara, CA) according to the manufacturer’s instructions. For NSPC differentiation, cells were plated on Matrigel (Corning) and cultured in the same medium, supplemented with 1 % fetal bovine serum instead of growth factors. Phenotype analysis of differentiated cells was performed using immunocytochemistry on an IX81 microscopy with a fluorescence module (Olympus Corp., Tokyo Japan).
Animal model and cellular transplantation
All mouse studies were conducted in strict accordance with the Guide for Care and Use of Laboratory Animals of the Central Institute for Experimental Animals (CIEA, Kanagawa, Japan), and the experimental protocols were approved by the CIEA Animal Care Committee (Permit Number: 11029A) in accordance with Keio University School of Medicine (Tokyo, Japan) (Permit Number: 16-096-25).
For the brain transplant model, NSPCs were injected bilaterally into the striata of 9 week-old female NOG mice (2.0×10
6 cells per mouse) (Clea Japan, Tokyo, Japan). For the spinal cord injury models, the cell transplantations were performed on 9 week-old female NOD/
scid mice (Charles River Laboratories Japan, Inc., Tokyo, Japan), as previously described [
14]. Briefly, 5×10
5 NSPCs were transplanted to the epicenter of the injury 9 days after the moderate contusion injury (IH impactor, 60-70kdyn). After the transplantation, the brain transplant models were monitored for abnormal behavior, and the spinal cord injury models were monitored for lower limb motor function using the Basso Mouse Scale (BMS) [
56].
The mouse strain type differed between the transplantation models because of the capacity lamination of our facility to harvest immunodeficient animals.
Histological Analysis
Twelve to 26 weeks after transplantation, mice were sacrificed, and their brains or spinal cords were taken to make paraffin sections, which were evaluated via H&E staining or immunohistochemistry. In all the sagittal spinal cord sections, the left side indicates the rostral and the upper side indicates the dorsal side of the section.
The extent of the transplants in the CNS of the transplanted animals was analyzed by measuring the STEM121-positive and -negative areas using Adobe Photoshop (version 13.0; San Jose, CA, USA).
To produce pathological classifications, additional histology data from transplantation of human fetal NSPCs and other iPSC-derived NSPCs were evaluated.
Statistics
A significance criterion of p < 0.05 was used. A nonparametric Kruskal-Wallis test followed by the Mann-Whitney U test were used to analyze RT-PCR and graft volumes.
Acknowledgements
We thank Center for iPS Cell Research and Application (CiRA: Kyoto, Japan) for kindly providing the PBMC-iPSCs. We are also very grateful to the members of the spinal cord research team at the Department of Orthopaedic Surgery, Rehabilitation Medicine and Physiology and the members of the Department of Physiology, Keio University School of Medicine, as well as the members of the Institute for Clinical Research, Osaka National Hospital for their valuable assistance with the in vitro and in vivo experiments and animal care. We also thank Dr. Tohru Masui, Dr. Kenjiro Kosaki, and members of the SKIP Operating Committee for their assistance with making our database on the SKIP website, and also to Prof. Douglass Sipp (Keio University) for his invaluable comments on the manuscript.
Authors’ contributions
Conceptualization, KS, RF, and TS; methodology, KS, RF, YK, MN, and HO; investigation, KS, RF, TS, HF, SK, YN, YH, KK, MI, DK, THT, HS, and EI; writing the original draft, KS, RF, and TS; writing, review and editing, KS, RF, TS, JK, AI, YK, MN, and HO; funding acquisition, MN, YK, and OH; project administration, TM, KK, YK, MN, and HO; supervision, MM, MN, and HO. All authors read and approved the final manuscript.