Development, characterization, and hematopoietic differentiation of Griscelli syndrome type 2 induced pluripotent stem cells

Griscelli syndrome type 2 (GS-2) is a rare, autosomal recessive immune deficiency syndrome caused by a mutation in the RAB27A gene, which results in the absence of a protein involved in vesicle trafficking and consequent loss of function of in particular cytotoxic T and NK cells. Induced pluripotent stem cells (iPSC) express genes associated with pluripotency, have the capacity for infinite expansion, and can differentiate into cells from all three germ layers. They can be induced using integrative or non-integrative systems for transfer of the Oct4, Sox2, Klf4, and cMyc (OSKM) transcription factors. To better understand the pathophysiology of GS-2 and to test novel treatment options, there is a need for an in vitro model of GS-2. Here, we generated iPSCs from 3 different GS-2 patients using lentiviral vectors. The iPSCs were characterized using flow cytometry and RT-PCR and tested for the expression of pluripotency markers. In vivo differentiation to cells from all three germlines was tested using a teratoma assay. In vitro differentiation of GS-2 iPSCs into hematopoietic stem and progenitor cells was done using Op9 feeder layers and specified media. All GS-2 iPSC clones displayed a normal karyotype (46XX or 46XY) and were shown to express the same RAB27A gene mutation that was present in the original somatic donor cells. GS-2 iPSCs expressed SSEA1, SSEA4, TRA-1-60, TRA-1-81, and OCT4 proteins, and SOX2, NANOG, and OCT4 expression were confirmed by RT-PCR. Differentiation capacity into cells from all three germ layers was confirmed using the teratoma assay. GS-2 iPSCs showed the capacity to differentiate into cells of the hematopoietic lineage. Using the lentiviral transfer of OSKM, we were able to generate different iPSC clones from 3 GS-2 patients. These cells can be used in future studies for the development of novel treatment options and to study the pathophysiology of GS-2 disease.


Background
Griscelli syndrome type 2 (GS-2) is a rare, inherited, autosomal recessive disease with a frequency of < 1/1, 000,000 in neonates and has been shown to occur more often in Middle Eastern countries, such as Turkey, due to the increased rate of consanguineous marriages [1]. The disease is characterized by partial albinism (silvercolored hair, eyebrows, and eyelashes), hepatosplenomegaly, pancytopenia, immune deficiency, and neurological dysfunction [2]. In GS-2, RAB27A mutations cause the malfunction of a small GTPase, which plays an important role in vesicular fusion and cellular trafficking [2]. The RAB27A protein is responsible for the peripheral distribution of melanosomes in melanocytes and exocytosis of cytotoxic granules in the cytosol of cytotoxic T cells (CTL) and natural killer (NK) cells. In addition, RAB27A plays an important role in the binding of cytosolic granules to the cell membrane after T cell receptor signal activation [3][4][5]. GS-2 patients may also display a decrease in NK cell cytotoxicity resulting in impaired and late hypersensitivity [5]. Although in most cases granulocyte and lymphocyte counts are within normal limits, immunoglobulin levels may be decreased or increased. The only curative treatment for GS-2 is hematopoietic stem cell (HSC) transplantation [1,[6][7][8]. However, in the absence of human leukocyte antigen (HLA)-compatible donors, no other curative treatment options are available, and therefore, new treatment strategies, such as gene therapy, should be explored.
Via exogenous delivery of the transcription factors Oct4, Sox2, Klf4, and cMyc (OSKM), somatic cells can be reprogrammed into induced pluripotent stem cells (iPSCs) [9], which express typical pluripotency markers, such as SSEA-4, TRA-1-60, TRA-1-81, and alkaline phosphatase, and display three germ layer differentiation potential [9]. Although iPSC lines have been developed from various cell types [10], reprogramming efficiency and differentiation capacity of the iPSCs may vary according to cell source or type [11]. Especially in cases where sufficient cells cannot be obtained due to the rarity of the disorder, iPSC models have shown great potential in providing research material to develop new treatments or diagnostics, but also to study the pathophysiology of the diseases [10,12]. iPSC lines have been generated for a variety of diseases, including childhood hereditary diseases, and are now deposited in and made available through international iPSC biobanks [13][14][15]. In addition to disease modeling, iPSCs also enable the development and testing of new treatment modalities, such as gene editing.
Here, we aimed to generate iPSC lines from two different somatic cell types, i.e., mononuclear cells (MNCs) and multipotent mesenchymal stromal/stem cells (MSCs), that were obtained from 3 different GS-2 patients. We intended to generate an in vitro model reminiscent of GS-2 that (1) can be used as a platform to test new treatment strategies, such as gene therapy, and (2) will allow us to obtain a better understanding of the direct effects of the RAB27A mutation on differentiation and function and behavior of hematopoietic cells.

Materials and methods
Healthy donor and GS-2 bone marrow samples Bone marrow (BM) samples of 1-5 mL were aspirated from GS-2 patients (n=3) and healthy donors (n=7) for diagnostic and/or transplantation purposes, after obtaining written informed consent from each donor. Procedures and sampling were approved by the Hacettepe University Non-Interventional Ethical Committee (GO14/424-19).

Mutation analysis of patient cells before reprogramming
RAB27A gene mutations were determined in the material of all three GS-2 patients before and after reprogramming. Selected iPSC clones were expanded and all coding domains, as well as exon-intron intersection domains were assessed by next-generation sequencing with a Miseq device (Illumina) and analyzed using the Miseq Reporter (Illumina) software. Aligned "bam" files were analyzed with the IGV 2.3 (Broad Institute) software.

Karyotyping of iPSCs
For karyotyping, 3-4 drops of Colcemid were added to the culture and incubated for 3 h. iPSC cultures were trypsinized, and hypotonic salt solution was added to the cell pellets. Cells were prefixed with 5-6 drops of fixative, centrifuged, and resuspended in 5 mL fixative. Cells were washed twice, transferred to slides, and stained with Giemsa. A minimum of 10 metaphases was captured and analyzed.
In vivo teratoma formation and characterization 1 × 10 6 iPSCs from different clones derived from both healthy donors and GS-2 patients were resuspended in 100 μL Matrigel and injected directly in the left and right gastrocnemius muscle of Balb/c-Rag2 −/− mice (n= 10), kindly provided by Prof. Dr. Gerard Wagemaker, Erasmus Medical Center Rotterdam, The Netherlands [19], after pretreatment with 25 mg/kg Busulfan (Busilfex, Abdi İbrahim). All animal experiments were approved by the local Hacettepe University animal experiments ethical committee (2014/46-03 and 2015/ 68-10) and performed at the Hacettepe University Laboratory Animals Research and Application Centre. Teratomas were removed after 6-8 weeks of transplantation for histological analysis and infiltrated with paraffin following immersion fixation in 10% buffered formalin, dehydration in graded series of ethanol, and clearing in xylene. Three to five-micrometer-thick hematoxylin/eosin-stained sections were evaluated under a light microscope (Leica DM6B) with an attached digital camera in order to confirm the presence of three germ layer differentiation using the LASX software.

Statistical analysis
Differences in gene expression levels between somatic cells before reprogramming and generated iPSC clones were determined after normalizing genes of interest with the reference gene. The standard deviation of the mean was calculated, and T tests were performed using the Excel software program. P values <0.05 were considered significant.

Characteristics of GS-2 and healthy donor BM-MSCs
Microscopically, MSCs from healthy and GS-2 samples were similar and showed a spindle-shaped morphology (Fig. 1a). No significant differences were found in terms of differentiation capacity (Fig. 1a, Table S2). Both healthy and GS-2 MSCs were highly positive for specific MSC markers CD29, CD44, CD73, CD90, CD105, CD166, and HLA-DR and negative for the endothelial and hematopoietic markers CD31 and CD34 (Fig. 1b,  Table S3).

Lentiviral reprogramming of healthy and GS-2 MNCs and MSCs
Eighteen days after reprogramming of healthy donor and GS-2 MNCs, using LV-SF-OSKM, formation of idTOMpositive iPSC-like colonies was observed (Fig. 2a). Passage 3 BM-MSCs from healthy (n=1, GP) and GS-2 (n= 3; YF, IK, YKÇ) donors were reprogrammed using the bicistronic vector for YF and IK and the polycistronic vector for YKÇ, and colonies were collected between 18 and 21 days. At least 5 iPSC clones were generated for each sample, and the best clones were chosen for further use and characterization (Fig. 2b). iPSC clones derived from healthy and GS-2 samples maintained consistent morphology and growth rates for up to 12 passages.
Karyotyping of healthy and GS-2 iPSC clones showed normal karyotypes consisting of 46XX or 46XY chromosomes ( Figure S1). Teratoma slides revealed the presence of mesodermal striated and smooth muscle, adipose tissue, cartilage, blood vessel, endodermal respiratory and gastrointestinal epithelia, and ectodermal epidermis and axons histologically (Fig. 4), confirming the potential for differentiation into three germ layers of these iPSC clones.
Mutation analysis of healthy and GS-2 iPSC clones confirmed the absence of RAB27A mutations in healthy donor iPSCs and the presence of specific RAB27A gene mutations at the same locations in the GS-2 iPSC clones and the original GS-2 samples (Table 1).

In vitro hematopoietic differentiation of healthy and GS-2 iPSCs
iPSCs were co-cultured on Op9 cells with hematopoietic differentiation medium 1 (HDM1: αMEM, 10% FBS-HI, 1% penicillin/ streptomycin, 1% L-glutamine, 5 mM βmercaptoethanol, 4 mg/mL L-ascorbic acid, 10 μg/mL BMP-4) or hematopoietic differentiation medium 2 (HDM2: StemMACS HSC expansion medium, 1X STF, 10 μg/mL BMP-4). Within 4-5 days after co-culture, islands of hematopoiesis appeared in the cultures (Fig. 5a), and after 9 days of co-culture, cells were assessed for the presence of CD34, CD38, CD45, and/or CD43, a panhematopoietic marker, that has been shown to define the earliest hematopoietic progenitors in embryonic stem cell (ESC) cultures [21]. The results showed the presence of up to 14.3% CD34+ in serum-free HDM2 medium, despite considerable adipogenic differentiation of the stromal feeder layer Op9 cells. In contrast, significantly lower levels of CD34+ cells (4.3%) were detected in co-cultures maintained in HDM1 where, interestingly, the morphology of the Op9 cells remained unchanged (Fig. 5b). Not all clones were tested for hematopoietic differentiation potential, but from the 7 clones (2 healthy donor iPSC clones, 5 GS-2 iPSC clones, derived from MNCs and MSCs) tested, with the exception of one GS-2 clone (YF #A2C3), all clones showed expression of CD34+ after 9 days of co-culture with Op9 cells. This particular clone also showed much lower gene expression of NANOG and OCT4 (Fig. 3c) and was most likely only partially reprogrammed. Colony assays performed with iPSC-derived HSCs showed a preference for white colonies, which were confirmed to be mixed CFU-GM colonies expressing CD45 (91.6-92.3%), CD33 (47.1-95.0%), CD16 (9.0-9.1%), and CD14 (12.4-17.5%) markers (Fig. 5c). Optimization of the culture conditions by addition of ATRA and dexamethasone, in combination with further expansion of the iPSC-derived HSCs for an additional 21 days in presence of HSC expansion medium supplemented with 1X STF resulted in the expansion of considerable numbers of CD34+ cells, ranging from 40 to 80% CD34 after 14-21 days of expansion culture ( Figure S2). When these cells were transplanted into Rag2 −/− mice at a  dose of 5 × 10 5 -10 6 cells/mouse, engraftment was observed at 1 month after transplantation with the appearance of human cells in peripheral blood (PB) samples of mice (Fig. 6a). Maximal engraftment was observed at 2 months after transplantation, after which levels of human cells in murine PB steadily decreased. In addition to the presence of human leukocytes in PB, the presence of CD34+ and CD45+ human hematopoietic cells was also detected in BM and spleen at 3 and 6 months after transplantation (Fig. 6b).

Discussion
GS-2 syndrome is a rare autosomal recessive immunodeficiency that can only be treated with HSC transplantation. The rarity of the disease makes it difficult to obtain sufficient cell numbers for research and to allow the development of new treatment strategies. The two GS-2 mouse models, i.e., the C3H/ HeSn-Rab27a Ash /J or "Ashen mouse" [22] and the C57BL/6J-Rab27a Ash /J (Rab27a −/− ) mouse model created by backcrossing into C57BL/6J mice [23], both display an immune deficiency and HLH phenotype similar to GS-2 syndrome in humans. Nonetheless, an in vitro human GS-2 model is needed to develop and test new treatment strategies. Therefore, we aimed to generate several iPSC lines from GS-2 patients with different RAB27A mutations. Here, we show that iPSC clones derived from GS-2 MNCs and MSCs fulfill all the requirements of pluripotent stem cells, including the In addition to the detailed characterization of GS-2 iPSCs, we also provide a detailed characterization of GS-2 MSCs and compared the latter with healthy donor MSCs. In terms of "morphology" and immunophenotype, we found no differences between GS-2 and healthy donor MSCs. In addition, GS-2 MSC proliferation and differentiation capacity were similar to healthy donor MSCs. Thus, it appears that the RAB27A mutation does not negatively affect the overall proliferation or differentiation capacity of MSCs.
iPSC clones were successfully derived from both GS-2 MNCs and MSCs using lentiviral vectors, and no differences were observed in terms of morphology, immunophenotype, or expression of pluripotency genes in iPSCs derived from these two somatic cell sources. However, we did notice that using the same concentration of OSKM virus, MNCs showed a tendency for rapid reprogramming, whereas MSCs appeared to be more resistant to reprogramming, requiring a higher MOI. These data are in line with a study showing that cell type may affect reprogramming efficiency [24]. Similar to our results, Wang and colleagues showed that reprogramming of MNCs appeared to be more effective than reprogramming of MSCs [25].
Using different protocols to induce hematopoietic differentiation of the GS-2 iPSC clones, we found that MNC-derived iPSCs more rapidly differentiated towards HSCs, giving rise to up to 17% CD34+ in co-cultures with Op9 cells. In contrast, HSC differentiation from iPSCs derived from GS-2 MSCs proved much more difficult, resulting in much lower percentages of CD34+ cells (ranging from 2 to 4%). These data may indicate that the inherent, epigenetic memory of the iPSCs may influence both the efficacy of reprogramming, but also the differentiation capacity of these cells, at least in vitro. Accordingly, it has been shown previously that iPSCs may preserve their epigenetic memory and that tissue or somatic cell type-specific DNA methylation patterns may remain despite reprogramming, causing a relative preference of these cells to re-differentiate towards their cell type of origin rather than transdifferentiation into another lineage [26,27].
Choi et al. co-cultured iPSCs with Op9 cells and found similar levels of CD34+ cells [28]. However, these levels leave room for improvement, and many groups are currently working on protocols to increase the efficacy and efficiency of HSC differentiation. For example, ATRA has been shown to increase hematopoietic differentiation potential [29]. In addition, certain molecules, such as WNT pathway inhibitors (CHIR99021) [30,31], BMP4, and bFGF, have been reported to increase hematopoietic differentiation efficiency [32]. Using different combinations of these factors, we managed to increase the levels of CD34+ cells in vitro. Further culture modifications, including the use of gelatin-coated surfaces, different densities, and/or irradiation of Op9 cells, may further improve hematopoietic differentiation of iPSCs. Alternatively, Cerdan and colleagues performed HSC differentiation using embryonic body formation directly from ESCs and were able to obtain 90% CD45+ cells [33]. iPSC-derived HSCs were transplanted in immunedeficient Rag2 −/− mice, and engraftment was followed for up to 6 months. Despite high levels of CD34+ cells obtained in in vitro cultures, we found that human chimerism after the initial engraftment rapidly decreased after 2 months of transplantation. This lack of long-term engraftment may have been caused by many factors, including prolonged in vitro expansion of the CD34+ cells, which may have depleted long-term repopulating cells. Furthermore, the cells were not further sorted or selected for CD34 positivity, and infused cell populations may not have been sufficient to sustain engraftment. Also, in contrast to the Rag2 −/− /IL2R γc −/− mice, the Rag2 −/− mice that we used here are not the preferred humanized mouse model due to their remaining NK cell function. In addition, the low-dose preparatory regime used (25 mg/kg busulfan) may not have been sufficient to open up the bone marrow space and suppress remaining immunity. Therefore, engraftment might have been better or could have lasted longer in other humanized models, such as the NOD/SCID-IL2Rγ −/− (NSG) mice and NSG mice transgenic for human hematopoietic growth factors [34]. Lastly, iPSC-derived CD34+ cells may show defects in homing and/or cannot be sustained/maintained in the bone marrow due to the absence of certain surface antigens, including adhesion molecules. It is furthermore conceivable that iPSC-derived CD34+ cells may not have sufficiently long telomeres to support long-term hematopoiesis.
Since we cannot rule out that the observed shortterm engraftment was caused by the animal model, rather than by the transplanted iPSC-derived HSCs, all these factors should be further investigated. Nevertheless, the data are promising, and using modifications of the current protocols, both in vitro and in vivo, we believe that transplantation and long-term engraftment of iPSC-derived HSCs will be feasible.

Conclusions
Here, we show that the iPSC clones derived from GS-2 patients fulfill all requirements of true pluripotent stem cells, including the expression of pluripotency genes and the capacity for differentiation into cells of all three germlines. In addition, the ability of GS-2 iPSCs to differentiate in vitro into HSCs and allow, at least short-term, in vivo engraftment underlines the potential of these cells for the development of future treatment modalities. With further optimization of the protocols, these cells could serve as a model to investigate the effects of the RAB27A mutation on HSC differentiation and HSC potential and its role in the impairment of cytotoxic T cells and NK cells. Furthermore, this GS-2 iPSC model can be used to develop new lentiviral or CRISPR/Cas9-based gene therapy strategies for the treatment of Griscelli syndrome type 2. Fig. 6. Engraftment of human iPSC-derived HSCs in immune-deficient mice. Immune deficient Rag2 −/− mice were pre-treated with busulfan and transplanted with 5 × 10 5 -10 6 cells obtained from human iPSC-derived HSC cultures per mouse (n=15). a Peripheral blood samples were collected at 1-6 months after transplant and assessed for the presence of human hematopoietic cells. b Groups of mice were sacrificed at month 3 and 6 to assess the presence of human hematopoietic cells in the spleen (SPL) and bone marrow (BM) tissue