Teratoma Formation in Immunocompetent Mice After Syngeneic and Allogeneic Implantation of Germline Capable Mouse Embryonic Stem Cells

The use of embryonic stem cells (ESCs) in the treatment of human diseases has long been muted, because such a treatment must not only be efficacious but must also be shown to be completely safe (Baker et al., 2007; Lukovic et al., 2013). Indeed, in many previous studies, the use of injected ESCs was associated with the risk of teratoma development (Wakitani et al., 2003; Baharvand and Matthaei, 2004; Przyborski, 2005; Nussbaum et al., 2007; Dressel et al., 2008; Prokhorova et al., 2009; Zhang et al., 2012; Li et al., 2013). In a clinical setting, this is a critical limitation of such a treatment. Therefore, much


Introduction
The use of embryonic stem cells (ESCs) in the treatment of human diseases has long been muted, because such a treatment must not only be efficacious but must also be shown to be completely safe (Baker et al., 2007;Lukovic et al., 2013).Indeed, in many previous studies, the use of injected ESCs was associated with the risk of teratoma development (Wakitani et al., 2003;Baharvand and Matthaei, 2004;Przyborski, 2005;Nussbaum et al., 2007;Dressel et al., 2008;Prokhorova et al., 2009;Zhang et al., 2012;Li et al., 2013).In a clinical setting, this is a critical limitation of such a treatment.Therefore, much

Teratoma Formation in Immunocompetent Mice After Syngeneic and Allogeneic Implantation of Germline Capable Mouse Embryonic Stem Cells
Abdullah Aldahmash 1,2 *, Muhammad Atteya 1,3 , Mona Elsafadi 1 , May Al-Nbaheen 1 , Husain Adel Al-Mubarak 1 , Radhakrishnan Vishnubalaji 1 , Ali Al-Roalle 1 , Suzan Al-Harbi 4 , Muthurangan Manikandan 1 , Klaus Ingo Matthaei 1,5 , Amer Mahmood 1 * interest among the scientific community has been on the understanding of these processes.Most studies have used immunodeficient recipient mice into which these cells were implanted (Wakitani et al., 2003;Baharvand and Matthaei, 2004;Przyborski, 2005;Nussbaum et al., 2007;Dressel et al., 2008;Prokhorova et al., 2009;Li et al., 2013).The resultant teratomas consisted of all three germ layers (ectoderm, mesoderm, endoderm) representing all tissue types in a whole individual (Prokhorova et al., 2009).The use of immunodeficient mice appears logical since it could be expected that immunocompetent animals would reject these cells.However, one study (Dressel et al., 2008) has surprisingly reported that the syngeneic implantation of 129Sv ESCs into 129Sv immunocompetent mice also gave rise to teratomas consisting of all three germ layers, and at a high frequency.We wondered, therefore, if the state of undifferentiation, or pluripotency, of the injected cells may play a critical role in this process.The present study was, therefore, designed specifically to investigate this possibility by using three mouse ESC lines of early passage that have been proven to have germline transmission capacity at similar passage stages.Moreover, we were interested in determining whether these cells could form teratomas not only in syngeneic but also in allogeneic immunocompetent hosts.Here we report that ESCs can form teratomas in immunocompetent mice after allogeneic as well as syngeneic implantation.

Ethics statement
All animal experimental procedures of this study were approved by the Institutional Review Board (Ethical Approval ID 11/3215/IRB), College of Medicine, King Saud University, Riyadh, Kingdom of Saudi Arabia.

Cell culture
Three mouse ESC (mESC) lines were used in this study: 129Sv W9.5 passage (p) 19, BALB/c p11, and C57BL/6 p13.The cell lines were provided by the Stem Cell and Gene Targeting Laboratory, John Curtin School of Medical Research, the Australian National University, Canberra ACT 0200, Australia, and were transferred to King Saud University in liquid nitrogen.All three cell lines have been published previously and were shown to be germline competent at comparable passages; 129Sv (Szabó and Mann, 1994;Yang et al., 2006), BALB/c (Royan C1) (Baharvand and Matthaei, 2004), and C57BL/6 (Royan B1) (Baharvand and Matthaei, 2004;Sadraei et al., 2009).
Mouse ESCs were grown to near confluence.Medium was aspirated from the flask.The adhered cells were washed twice with Ca2 + Mg2 + -free phosphate-buffered saline (PBS) (Gibco, 20012), and incubated at room temperature for 2 minutes with PBS containing 0.5mM ethylene glycol tetra-acetic acid (EGTA) (Sigma-Aldrich E3889, St. Louis, USA), then for 1-2 minutes with 0.05% trypsin-ethylenediaminetetraacetic acid (EDTA) (Gibco 25300, Invitrogen) (Matthaei, 2009).To obtain a single cell suspension, the cells were triturated rapidly in the trypsin solution several times through a small-bore pipette.The trypsin was inactivated by the addition of complete ESC medium.The cell suspension was plated into a 10 cm tissue culture dish and incubated at 37°C for 15-30 minutes to allow mouse embryonic feeder (MEF) cells to attach.The ESCs were then carefully aspirated, the cell suspension counted using a hemocytometer and centrifuged at 1000 rpm for 5 minutes to obtain a cell pellet for appropriate dilution.

Teratoma formation
Forty-six BALB/c mice (a kind gift by the Stem Cell and Gene Targeting Laboratory, John Curtin School of Medical Research, the Australian National University, Canberra ACT 0200, Australia) 6-10 weeks old were used as host animals for the implantation of three mESC lines (129Sv, BALB/c and C57BL/6).Injections were performed under the kidney capsule in 21 mice and into the subcutaneous tissue in 25 mice.For renal subcapsular injection, 1510 6 mESCs/injection were resuspended thoroughly in 20µl cold complete ESC medium with 20mM HEPES (Gibco, 15630) in 1.5ml sterile Eppendorf tubes.For subcutaneous injection, 1510 6 mESCs/injection were resuspended in 150µl cold complete ESC medium in 1.5ml sterile Eppendorf tubes and kept on ice prior to injection.Subcutaneous injection was performed at 2 sites (flank and neck) in each animal.

Injection under the kidney capsule
The mice were anesthetized with Sevoflurane and the right kidney was exposed via a right paralumbar incision.A transverse tear was made in the ventral aspect of the kidney capsule using microforceps.Air was then injected using a 22G intravenous cannula through the tear to create a pocket.The same size cannula was used to inject the cells into the pocket.Then the tear in the renal capsule was closed with cautery.The abdominal wall musculature and overlying skin were closed with 6-0 vicryl sutures.

Subcutaneous injection
The mice were anesthetized with Sevoflurane.The cells were resuspended gently using a 25G syringe needle and then drawn into a 1ml syringe immediately before injection.The skin at the site of injection was lifted and the cells were injected into the subcutaneous space.

Histological analysis
Four weeks after the injection, the animals were euthanized and the sites of implantation were dissected for histological examination.Harvested tissues were fixed for 48hrs in 4% paraformaldehyde, paraffin sectioned at 5µm and stained with hematoxylin and eosin.Highresolution whole-slide digital scans of all glass slides were then created with a ScanScope scanner (Aperio Technologies, Inc.).The digital slide images were then viewed and analyzed using the viewing and image analysis tools of Aperio's ImageScope software.The size of each teratoma in terms of its cut-sectional area was measured in mm 2 .The composition of each teratoma was estimated semi-quantitatively by visual inspection of digital slide images.Differentiated elements in the teratomas were classified into ectodermal (skin, neural), mesodermal (cartilage, bone, muscle), and endodermal (glandular structures) components as a percentage of entire tissue.The percentages of undifferentiated tissue, cysts, necrosis and hemorrhage were also recorded.The remaining spindle-shaped stroma was included under mesoderm.

Statistical analysis
The incidence of teratoma was compared using Pearson Chi-square.For composition of teratoma, t-test for dependent variables was used for within group comparisons, while one-way analysis of variance was used for between group comparisons.Because the size varied widely among teratomas, it was compared between groups using the nonparametric Kruskal-Wallis analysis of variance and median test.Results were considered significant when p<0.05.

Results
The three mESC lines were cultured as undifferentiated ESCs in the presence of MEFs and LIF.Colonies from all cell lines exhibited positive staining for Oct3/4, Sox2, and SSEA1 (Figure 1).We then analyzed the tumor formation ability of ESCs (129Sv, BALB/c and C57BL/6) by syngeneic and allogeneic implantation into immunocompetent BALB/c mice.We found that the incidence of teratoma formation after syngeneic implantation of BALB/c ESCs into immunocompetent BALB/c mice was 16.67% (Table 1).Unexpectedly, however, we observed that when we implanted 129Sv and C57BL/6 ESCs into immunocompetent BALB/c mice (allogeneic implantation) teratomas were also formed in 33.3% and 30.43%, respectively (Table 1).
An encapsulated solid tumor was found at the implantation site in 30% and 40% of 129Sv-implanted mice, in 0% and 37.5% of BALB/c-implanted mice, and in 25% and 66.7% of C57BL/6-implanted mice, for subcutaneous and renal implantation, respectively (Table 1, Figure 2A).The overall teratoma development success rate was 20 out of 71 injections (28.17%).The overall incidence was highest with 129Sv (33.3%) followed by C57BL/6 (30.43%) and lowest with BALB/c (16.67%).For all cell lines combined, the incidence of teratoma was significantly higher under the kidney capsule (42.86%) than in the subcutaneous tissue (22%; p=0.0395).No teratoma development was observed in mice that had received the BALB/c cell line subcutaneously.
To investigate whether the site of injection plays a role in the size of teratoma formation we analyzed the size of teratomas based on the site of injection.The overall tumor size appeared largest when derived from 129Sv (66.07±99.43mm 2 ) followed by C57BL/6 (35.02±35.37mm 2 ) and smallest with BALB/c (5.30±4.77mm 2 ) ESCs.However, these differences did not reach statistical significance.Similarly the overall size appeared largest under the kidney capsule (53.80±109.84mm 2 ) followed by the subcutaneous tissue (39.78±29.79mm 2 ).The cut-sectional area of teratomas developed by 129Sv   2B).The teratomas that developed from the C57BL/6 cell line measured an average of 45.04±37.88mm 2 in the subcutaneous tissue and 9.97±2.60mm 2 under the kidney capsule (Table 1, Figure 2B).The largest sizes were observed with the 129Sv cell line injected under the kidney capsule and with the C57BL/6 cell line in the subcutaneous tissue (Figure 2B).Notably, the larger the teratoma had grown, the more tissue types could be identified histologically.We also found that the incidence and size of teratoma formation are site-dependent; injection under the kidney capsule having the highest success rate and producing the largest teratoma sizes (Figure 3A).Diverse ESC-derived tissues, belonging to the three embryonic germ layers, were encountered, reflecting the pluripotency of the ESCs.Most commonly present tissues were nervous tissue, keratinizing stratified squamous epithelium (ectoderm), smooth muscle, striated muscle, cartilage, bone (mesoderm), and glandular tissue in the form of gut-and respiratory-like epithelia (endoderm), in that order (Figure 3B).The overall average composition of teratomas (n=20) was toward significantly more ectoderm (61.70%±24.71)than mesoderm (17.25%±17.85,p=0.0001) or endoderm (6.80%±4.34,p=0.0000) and significantly more mesoderm than endoderm (p=0.0085)(Table 1, Figure 4D).Teratomas that developed from the 129Sv cell line in the subcutaneous tissue showed significantly more mesoderm than teratomas that developed under the kidney capsule (p=0.017)(Figure 4A).However, C57BL/6 and BALB/c did not show any significant difference in germ layer composition between injection sites (Figure 4B-C).Teratomas that developed  doi.org/10.7314/APJCP.2013.14.10.5705Teratoma Formation in Mice after Syngeneic and Allogeneic Implantation of Mouse Embryonic Stem Cells from the C57BL/6 cell line in the subcutaneous site contained significantly more ectoderm than teratomas that developed from the 129Sv cell line at the same site (p=0.0017)(Table 1, Figure 5A).Teratomas that developed under the kidney capsule did not show any significant difference in germ layer composition between the three cell lines; however they all showed a high tendency towards ectoderm (Figure 5B).
Clusters of undifferentiated cells with malignant nuclear criteria (including pleomorphism, hyperchromatism and frequent mitotic figures) were common in most of the tumors (Figure 3C).The overall percentage of undifferentiated cells was higher with 129Sv (8.7%±13.08)than BALB/c (5.67%±5.13)or C57BL/6 (5%±3.51)with no statistically significant differences (Table 1).Site comparison showed that, overall, renal subcapsular implantation was associated with a significantly higher percentage of undifferentiated cells than subcutaneous implantation.

Discussion
The use of stem cell therapy for the treatment of human diseases would dramatically improve our ability to alleviate debilitating conditions.The potential of using ESCs for this purpose has been muted since these cells have the ability to differentiate into every cell type in an entire individual.This would suggest that all organs could be treated once methods had been developed to direct the ESCs to differentiate into specific target tissues thereby eliminating the problem of limited donor tissue for treatments.However, concerns have been raised over the safety of this promising therapeutic approach, including whether the transplanted ESCs have the potential to form tumors after engraftment because of their unlimited selfrenewal and high differentiation potential (Arnhold et al., 2004;Amariglio et al., 2009;Lukovic et al., 2013).
A number of studies (Thomson et al., 1998;Wakitani et al., 2003;Stojkovic et al., 2005;Nussbaum et al., 2007;Dressel et al., 2008;Prokhorova et al., 2009;Zhang et al., 2012;Li et al., 2013;Lukovic et al., 2013) have shown the potential of ESCs to form teratomas by injecting undifferentiated ESCs into mice.Most studies used immunodeficient hosts such as severe combined immunodeficient (SCID) mice (Thomson et al., 1998;Wakitani et al., 2003;Stojkovic et al., 2005;Zhang et al., 2012).Dressel et al. (2008) surprisingly showed that teratoma development could also occur after injection of ESCs into syngeneic immunocompetent mice.This made us consider whether the state of differentiation of the ESCs could influence the outcome.The question, therefore, is whether the more undifferentiated ESCs are, the more they are prone to form teratomas.
We have for many years been using mouse ESCs from 3 different mouse strains to generate genetically modified mice.This process is completely reliant on the ability of the ESCs to remain totally pluripotent and integrate into the germline of chimera after injection into blastocysts; thereby producing ESC-derived offspring after mating with wild-type mice.Our cell lines had all been proven to be fully undifferentiated and germline competent using this technique (Campbell et al., 2002;Baharvand and Matthaei, 2004;Yang et al., 2006;Meyer et al., 2010).We therefore decided to test the ability of all 3 mouse ESC lines to form teratomas in immunocompetent mice, both in syngeneic as well as allogeneic combinations.
The present study shows that our ESCs had the capability to form teratomas after injection into syngeneic as well as allogeneic immunocompetent mice (Table 1).The incidence of teratomas was higher after implantation of ESCs from the 129Sv strain when compared to ESCs from the C57BL/6 and BALB/c strains.This might be expected since the 129Sv strain is most commonly used due to its high germline competence and stability during continued cell culture and expansion.Interestingly, in this study, syngeneic implantation of the BALB/c cells into BALB/c mice failed to produce teratomas in subcutaneous tissue while those that developed under the kidney capsule were only small.Previous studies reported teratomas developing after syngeneic implantation (129Sv ESCs into 129Sv mice (Arnhold et al., 2004;Swijnenburg et al., 2005;Dressel et al., 2008) or C57BL/6 ESCs into C57BL/6 mice (Nussbaum et al., 2007), but syngeneic implantation of BALB/c ESCs into BALB/c mice has not been described previously.In previous studies (Nussbaum et al., 2007;Xie et al., 2007) allogeneic ESCs also caused teratomas.However, in these previous studies, allogeneic teratomas initially detected were gradually rejected by an increasing inflammatory immune response.These findings suggested that immunogenicity of ESCs increases on their differentiation, i.e. once ESCs reach a more differentiated state, they can be more effectively recognized and rejected by the recipient immune system.In the present study, the reason why we did not encounter such immune rejection after allogeneic implantation is, at this stage, not known.
Histologic analysis of the teratomas showed that the most commonly present tissues were nervous tissue, keratinizing stratified squamous epithelium (ectoderm), smooth muscle, striated muscle, cartilage, bone (mesoderm), and glandular tissue in the form of gut-and respiratory-like epithelia (endoderm), in that order.Hence diverse ESC-derived tissues, belonging to the three embryonic germ layers, were encountered, thereby reflecting the pluripotency of the ESCs (Thomson et al., 1998;Oosterhuis and Looijenga, 2005;Mahmood et al., 2010;Bustamante-Marín et al., 2013;Pal et al., 2013).Importantly, this suggests that the differentiation potential of the 3 mouse ESC lines used is independent of microenvironmental factors that may be encountered at the injection site.These results are in agreement with previous studies that implanted a heterogeneous differentiated ESC population into brain (Riess et al., 2007), heart (Xie et al., 2007) and retina (Cui et al., 2013) and found that any remaining undifferentiated cells formed teratomas.Hence, our data also support that care should be taken that ESCs are fully differentiated before their use for cell-based therapy (Prokhorova et al., 2009), since any remaining undifferentiated cells, if irresponsive to microenvironmental differentiation signals, can form teratomas.Moreover, even after thorough selection of a highly purified population of ESC-derived neural (Arnhold et al., 2004;Lukovic et al., 2013) and islet (Fujikawa et al., 5710 2005) precursor cells, there remains a high risk of teratoma formation after engraftment.However, transplantation of FACS purified ESC-derived neural precursors expressing the neural precursor marker Sox1 into mouse brains (Fukuda et al., 2006) and targeting hESC-specific antiapoptotic factor(s) (i.e., survivin or Bcl10) represent an efficient strategy to selectively eliminate pluripotent cells with teratoma potential (Lee et al., 2013).
We found, in accordance with Cooke et al. (2006), that the incidence and size of teratoma formation are site-dependent; injection under the kidney capsule having the highest success rate and producing the largest teratoma sizes.Because of its rich vascular bed, the renal subcapsular space of mice is a suitable environment for growing and studying solid tumors (Press et al., 2008).This makes renal subcapsular implantation better to subcutaneous implantation for teratoma induction experiments, in terms of success rate and size of teratomas produced.
In this study, clusters of undifferentiated cells exhibiting malignant nuclear criteria (including pleomorphism, hyperchromatism and frequent mitotic figures) were common in most of the teratomas.These can be considered as malignant teratomas based on the presence of tissues of the three germ cell layers and the presence of both anaplastic foci and immature tissues (López and Múrcia, 2008).The overall percentage of undifferentiated cells (malignant-looking clusters) was higher with 129Sv than BALB/c or C57BL/6 ESCs.Renal subcapsular implantation was associated with a higher percentage of undifferentiated cells than subcutaneous implantation.This agrees with Cooke et al. (2006) who reported that subcutaneous implants were significantly slower growing and formed tumors composed of differentiated tissues, in contrast to cells grafted into the liver which rapidly produced large tumors containing predominantly immature cells.It is still largely unknown why ESCs, which lack overt chromosomal abnormalities, are tumorigenic.One hypothesis is based on the identification of the canonical WNT signaling as a critical determinant for the tumorigenicity and therapeutic function of ESC-derived retinal progenitors.The function of WNT signaling is primarily mediated by TCF7, which directly induces expression of Sox2 and Nestin (Cui et al., 2013).Also, similarity between ESCs and cancer cells has been proposed.The embryonic stem cell factor Sox2 is reported to be expressed in a variety of early stage postmenopausal breast carcinomas and metastatic lymph nodes (Lengerke et al., 2011).The study of Wong et al. (2008) demonstrated that embryonic stem cells and multiple types of human cancer cells share a genetic expression pattern that is repressed in normal differentiated cells.Recent discoveries also indicated that testicular carcinoma in situ (CIS) cells have a striking phenotypic and genotypic similarity to ESCs (Almstrup et al., 2006).The similarities between ESCs and cancer cells should be further investigated.
In the present study, implantation of mESCs caused teratomas in both syngeneic and allogeneic immunocompetent mice.It could be argued that our data are not consistent with a previously published report (Dressel et al., 2008) in which allogeneic transplants of mESC did not result in teratomas in immunocompetent hosts.However, the state of pluripotency of the ESCs used in that study (Dressel et al., 2008) was not clearly described, in contrast to our ESCs with proven germline competence.In addition, because of their early stage of development, ESCs are considered "immune privileged," i.e, unrecognizable by the recipient immune defenses (Magliocca et al., 2006;Ichiryu and Fairchild, 2013).Moreover, it has been reported that mESCs are resistant to antigen-specific cytotoxic T lymphocyte (CTL) due to the expression of serpin-6, an endogenous inhibitor of granzyme B (Abdullah et al., 2007).Dressel et al. (2008) stated that the MPI-II cells they used were susceptible to granule exocytosis-mediated killing by rat NK cells and also by peptide-specific mouse CTL.So their cells may not have been protected against cellular cytotoxicity and hence their allogeneic transplants into immunocompetent mice were rejected soon after injection.In contrast, our cells were not rejected after allogeneic implantation in immunocompetent mice since they may have been immune privileged and thereby escaped cytotoxic immune destruction and formed teratomas.Nevertheless, such argument warrants that the susceptibility of different ES cell lines to the cytotoxic immune response be further investigated.
In conclusion, we have confirmed that ESCs can form teratomas in immunocompetent mice.Furthermore, we showed that germline competent ESC can form teratomas not only in syngeneic but also in allogeneic hosts.It is clear, therefore, that in a clinical setting, it must be ensured that no undifferentiated cells remain which could form teratomas.Without this assurance, the use of ESCs should not be considered as a therapeutic option for patients with degenerative disease.

Figure 1 .
Figure 1.Characterization of the Three Different mESC Lines.Upper panel shows live cell image and colony formation of the three mESC lines grown under standard ESC conditions.Lower 3 panels show immunofluorescence cell staining of pluripotent colony cells with positivity for Oct4, Sox2 and SSEA1.Nuclear stain DAPI was used as a counter staining

Table 1 . Incidence, Mean Size and Percentages of Components of Teratomas Formed After Subcutaneous and Renal Subcapsular Injection of Three mESC Lines
*SC, subcutaneous; undiff., undifferentiated tissue; H/N, hemorrhage and necrosis received the BALB/c cell line, while those that developed under the kidney capsule were only small and measured an average of 5.3±4.77mm 2 (Figure