For clinically relevant studies on melanoma progression and invasiveness, in vivo experimental systems with a human cellular microenvironment would be advantageous. We have compared tumor formation from a human cutaneous malignant melanoma cell line (BL), after injection as conventional xenografts in the mouse, or when injected into a predominantly species-specific environment of human embryonic stem cell–derived teratoma induced in the mouse (the hEST model). The resulting melanoma histology was generally analogous, both systems showing delimited densely packed areas with pleomorphic cells of malignant appearance. A specificity of the integration process into the human embryonic teratoma tissues was indicated by the melanoma exclusively being found in areas compatible with condensed mesenchyme, similar to neural crest development. Here, also enhanced neovascularization was seen within the human mesenchymal tissues facing the BL melanoma growth. Furthermore, in the hEST model an additional melanoma cell phenotype occurred, located at the border of, or infiltrating into, the surrounding human loose mesenchymal fibrous stroma. This BL population had a desmoplastic spindle-like appearance, with markers indicative of dedifferentiation and migration. The appearance of this apparently more aggressive phenotype, as well as the induction of human angiogenesis, shows specific interactions with the human embryonic microenvironment in the hEST model. In conclusion, these data provide exciting options for using the hEST model in molecular in vivo studies on differentiation, invasiveness, and malignancy of human melanoma, while analyzing species-specific reactions in vivo. [Cancer Res 2009;69(9):3746–54]

Analysis of tumor tissue architecture, its cellular morphology, and the diagnostic use of cellular and molecular markers with empirical links to origin or malignancy, constitutes the basis for current tumor pathology. In experimental cancer research, tumor properties can be reproduced to varying extents in vitro, e.g., three-dimensional growth and matrix interaction (13). However, the temporal and spatial coordination of a cellular microenvironment is for obvious reasons more difficult to mimic in vitro and can only fully be accomplished in vivo. In vivo experimentation in humans is by necessity limited and as a result, animal experimentation, such as xenografting, has been the option. Extensive information has been achieved using this approach, but the lack of species-specific interaction between tumor and stroma is still a potential limitation. Indeed, specific reservations regarding the extent of xenotransplants recapitulating development of human tissue has been recognized (4). Although there are ample evidence for conservation of major gene expression response modes across the mammals and conserved expression patterns are often linked with specific functions, exceptions exist (5). It is reasonable to assume that limitations in cellular interactions in xenomodels could also be implied in the numerous cases where therapies toward human tumor xenografts have been shown efficient in animal studies but less efficient or even ineffective in human clinical trials. Hence, new experimental systems where human tumors can be studied in a species-specific cellular in vivo microenvironment would be advantageous.

One such system has recently been described by Tzukerman and colleagues (6, 7). In a human microenvironment, generated within a teratoma derived from human embryonic stem cells (hESC) in immunodeficient mice, growth of various human tumor cells (ovary, prostate, lung, breast, colorectum, neural, and skin) was supported. Intriguingly, the study included an anticancer therapy that caused complete tumor regression in a traditional xenograft model but did not eradicate the same tumor in the human teratoma microenvironment model.

We have previously made use of xenografting for assaying in vivo pluripotency of new hESC lines (8, 9), for studies of early differentiation after hESC xenotransplantation (1012) and links to tumor development (13, 14). Based on these studies, we hypothesize that an optimal support of tumor growth from the hESC teratoma microenviroment is supposedly at a point when the teratomas are rich in various tissues but before limited conditions leads to substantial necrosis. Injection of a low dose of karyotypically normal hESC under the testis capsule of SCID-Beige mice leads to a benign teratoma development with mainly mouse vascularization and a diminutive formation of immature human vessels (11, 15) starting from day 20 onwards, subsequently anastomizing with the host vascular system. Such neovascularization from hESC derivates coincides with a rapid expansion of the teratoma tissues. By day 30, this represents all three germ layers, and a more organized even organoid development occurs from day 45 and onwards (10).7

7

Gertow, Cedervall, Ahrlund-Richter, unpublished.

This time scale leaves an opportunity for further observations of up to an additional 45 days after injection of tumor cells. After this, the teratoma tissue grows too large to be adequately supported by the host, with extended necrotic areas developing. Using this experimental time frame, we have observed integration with ample tumor growth support after grafting human cell lines of several origins, such as neuroblastomas, medulloblastomas, and head and neck cancers.8
8

Cedervall et al., unpublished.

Here, we report detailed studies on the integration, growth, and differentiation of a human melanoma cell line in the hEST model. The BL cell line was derived from a lymph node metastasis of a cutaneous malignant melanoma in an adult male patient (16, 17).

Cutaneous malignant melanoma are highly aggressive plastic tumors showing a high degree of phenotypic heterogeneity and resembling multipotent dedifferentiated cells. They are able to mimic other types of cells (18, 19) and also have a molecular signature characteristic of a plastic dedifferentiated cell type (20). Communication between the melanoma cells and the adjacent host tissues in xeno-models has been shown to be critical for their progression, and of the utmost importance for the invasion and formation of metastases (e.g., ref. 21). Also, several lines of evidence have shown that metastatic melanoma cells are responsive to developmental cues in an embryonic microenvironment (18, 2224). Such plasticity of melanomas, proven both in vitro and in xeno-environments in vivo, provides further argument for studies of human melanomas in species-specific in vivo test systems, as described here for the hEST model.

As expected from earlier studies, we found the hESC-teratoma environment highly permissive for the integration and growth of injected BL melanoma cells. In addition, we found evidence of specificity for the process of integration because BL growth was exclusively found in areas compatible with condensed mesenchyme, similar to neural crest development. Specific interactions with the surrounding embryonic species–specific microenvironment were indicated by differences in melanocytic markers. In addition, a migratory BL melanoma phenotype, not observed in the xenografts, was observed in the hEST model. Spindle-like BL melanoma cells appeared in the surrounding loose mesenchymal or fibrous stroma, subsequently showing markers indicative of dedifferentiation and migration. Also, extensive human neovascularization was found in areas of human mesenchymal tissues facing the melanoma growth.

Cells. Two hESC lines, H9 (25) and HS181 (8), both 46XX, were cultured and verified for karyotype and markers as earlier described (26). The BL cell line was previously derived from a lymph node metastasis of a malignant melanoma in a male patient treated at the Karolinska University Hospital (16, 17). The BL cells were cultured in Iscove's modified Dulbecco's medium supplied with 10% fetal bovine serum and 1% Penicillin/Streptomycin, at 37°C, 6.8% CO2, and high humidity. The male origin of BL cells used was verified using sex chromosome specific probes and fluorescence in situ hybridization (FISH; below). All cell lines used were tested negative for Mycoplasma using PCR.

Animals and generation of hESC teratomas. The experiments were performed with permission by the Regional Committee for Animal Experimentation (Dnr N107/06). Six- to 8-wk-old SCID-Beige (C.B.-17/GbmsTac-scid-bgDF N7; M&B) males were used in the study. They were housed and maintained at 20°C to 24°C, 50% RH, a 14 to 10 h light-dark cycle with food and water ad libitum. The hESC teratomas were generated by the injection of 104 hESC under the testis capsule, as previously described (10, 11). The teratomas from the two hESC lines used, H9 and HS181, were similar in support of BL-melanoma tumor growth; thus, the data presented does not distinguish results from experiments with teratomas from H9 or HS181 cells, respectively.

Injection of BL cells. BL cells from log-phase growing cultures were trypsinized, washed, counted, and subsequently injected (106 cells in a 20 μL volume) either s.c. in the flank (sc; 2 mice), under the testis capsule (it; 3 mice), or into day 40 hESC teratomas induced under the testis capsule (hEST model; 4 mice). All BL-injected mice developed detectable BL tumor growth. Fourteen days later, the mice were sacrificed (cervical dislocation) and the teratoma/tumor growth removed and fixed in 4% neutral buffered paraformaldehyde overnight. The tissues were dehydrated through a graded series of alcohols to xylene, embedded in paraffin, serially sectioned into 5-μm-thick sections and stained using standard H&E.

FISH. Identification of the grafted tumor cells and hESC, respectively, was carried out by selecting opposite sex of grafted tumor cells and the hESC line used. A mixture of two Y and X chromosome-specific FISH probes (CEP XY; Vysis, Inc.) was used to distinguish between grafted cells of hESC origin (XX) and tumor origin (XY).

Briefly, slides were deparaffinized in xylene and rehydrated in a series of alcohol, followed by pretreatment with pepsin at 37°C. Genomic DNA and probe were denatured simultaneously by 5 min heating at 73°C. Hybridization was performed overnight at 42°C with a sex-specific FISH probe (CEP XY; Vysis, Inc.).

Immunocytochemistry. BL cells in culture were stained for the melanoma-associated markers HMB45, MelanA, or a combination of HMB45+MelanA+tyrosinase. Briefly, the cells were fixed for 5 to 10 min in 4% formaldehyde. Blocking was performed using 3% bovine serum albumin in PBS, primary antibodies were diluted (Table 1) and incubated for 1 h at room temperature. Incubation with secondary antibody was performed for 30 min at room temperature. Finally, the cells were counterstained using 4′,6-diamidino-2-phenylindole (DAPI).

Table 1.

Summary of marker studies; BL in vivo models

MarkerhEST-modelXenograft (sc)Xenograft (it)
Ki67 +* +* +* 
E-cadherin − − − 
MelanA − 
HMB45 − 
HMB45+MelanA+tyrosinase − 
NGFR (p75) 
Nestin 
Nodal + + + 
S100A4 + + + 
S100 (polyclonal antibody) + + + 
MarkerhEST-modelXenograft (sc)Xenograft (it)
Ki67 +* +* +* 
E-cadherin − − − 
MelanA − 
HMB45 − 
HMB45+MelanA+tyrosinase − 
NGFR (p75) 
Nestin 
Nodal + + + 
S100A4 + + + 
S100 (polyclonal antibody) + + + 

Abbreviation: NGFR, nerve growth factor receptor.

*

30-40% of cells positive.

Positive cells located in the border area between dense melanoma growth and the surrounding human loose mesenchyme.

Solitary positive cells, at random locations.

Immunohistochemistry. Slides with sections from teratoma/BL-tumor were deparaffinized in xylene and rehydrated in series of alcohol. Endogenous peroxidase activity was blocked by incubation with 3% H2O2 for 15 min, and antigen retrieval was performed as optimized for each antibody (Table 2). Blocking of unspecific binding was performed in 3% serum in PBS. Slides were incubated with primary antibody followed by incubation with secondary antibody (Table 1). Sections were dehydrated in ethanol and xylene, and mounted with Vectamount (Vector Laboratories, Inc.).

Table 2.

Antibodies and conditions used

AntigenSourceDilutionPretreatmentSec antibody/Visualisation
Ki67 Abcam, ab833 1:50 Citrate buffer Vectastain Universal Elite ABC kit 
E-cadherin Abcam, antibody 1416 1:50 Tris-EDTA buffer Vectastain Universal Elite ABC kit 
CD31 Dako, M0823 1:20 DakoCytomation Retrieval Solution DakoCytomation EnVision+System-HRP 
Melan A Dako, M7196 1:50 Tris-EDTA buffer Ventana ultraView 
    Universal DAB Detection kit 
HMB45 Dako, U7025 1:100 Tris-EDTA buffer Ventana ultraView 
    Universal DAB Detection kit 
Pan Melanoma Cocktail Biocare Medical, CM165C 1:400 Tris-EDTA buffer Ventana ultraView 
    Universal DAB Detection kit 
NGFR (p75) R&D, MAB367 1:100 Citrate buffer DakoCytomation EnVision+System-HRP 
Nestin Chemicon, MAB5326 1:100 Citrate buffer DakoCytomation EnVision+System-HRP 
Nodal Abcam, ab55676 1:100 Citrate buffer DakoCytomation EnVision+System-HRP 
S100 Dako, Z0311 1:500 Citrate buffer DakoCytomation EnVision+System-HRP 
S100A4 Mabs 20.1 (31) 1:50 Tris-EDTA buffer DakoCytomation EnVision+System-HRP 
AntigenSourceDilutionPretreatmentSec antibody/Visualisation
Ki67 Abcam, ab833 1:50 Citrate buffer Vectastain Universal Elite ABC kit 
E-cadherin Abcam, antibody 1416 1:50 Tris-EDTA buffer Vectastain Universal Elite ABC kit 
CD31 Dako, M0823 1:20 DakoCytomation Retrieval Solution DakoCytomation EnVision+System-HRP 
Melan A Dako, M7196 1:50 Tris-EDTA buffer Ventana ultraView 
    Universal DAB Detection kit 
HMB45 Dako, U7025 1:100 Tris-EDTA buffer Ventana ultraView 
    Universal DAB Detection kit 
Pan Melanoma Cocktail Biocare Medical, CM165C 1:400 Tris-EDTA buffer Ventana ultraView 
    Universal DAB Detection kit 
NGFR (p75) R&D, MAB367 1:100 Citrate buffer DakoCytomation EnVision+System-HRP 
Nestin Chemicon, MAB5326 1:100 Citrate buffer DakoCytomation EnVision+System-HRP 
Nodal Abcam, ab55676 1:100 Citrate buffer DakoCytomation EnVision+System-HRP 
S100 Dako, Z0311 1:500 Citrate buffer DakoCytomation EnVision+System-HRP 
S100A4 Mabs 20.1 (31) 1:50 Tris-EDTA buffer DakoCytomation EnVision+System-HRP 

Abbreviations: DAB, 3,3′-diaminobenzidine; HRP, horseradish peroxidase.

Masson Trichrome staining for collagenous matrix. The tissue sections were deparaffinized in xylene and rehydrated through graded series of alcohol. Counterstaining was performed using hematoxylin (8 min), and sections were thereafter continuously stained in Biebrich Scarlet-Acid Fuchsin Solution, Phosphomolybdic-Phosphotungstic Acid Solution, and finally Alinin Blue Solution for 5 min each. Sections were dehydrated in ethanol and xylene and mounted with Vectamount (Vector Laboratories, Inc.).

Image analysis. Imaging was performed using a Zeiss Axiovert 200M microscope and Openlab 5.0 software. Images in Figs. 1 to 3 were adjusted for auto levels in Photoshop.

Figure 1.

Histology of BL melanoma cells in vivo; sections showing densely packed pleomorphic cells with large irregular nuclei, distinct nucleoli, and a less prominent cytoplasm (H&E staining; magnification, ×40). A, xenograft after s.c. injection into SCID/Beige male mice. B, xenograft after injection under the testis capsule. C, injection into the hEST model. D, archival H&E staining of section from the original BL lymph node metastasis after surgery of the patient, showing tumor cells and extensive infiltration of smaller lymphoid cells (courtesy of Drs. G. Masucci and R. Kiessling, Department of Oncology and Pathology, Karolinska Institutet, Stockholm, Sweden). Images adjusted for auto levels in Photoshop.

Figure 1.

Histology of BL melanoma cells in vivo; sections showing densely packed pleomorphic cells with large irregular nuclei, distinct nucleoli, and a less prominent cytoplasm (H&E staining; magnification, ×40). A, xenograft after s.c. injection into SCID/Beige male mice. B, xenograft after injection under the testis capsule. C, injection into the hEST model. D, archival H&E staining of section from the original BL lymph node metastasis after surgery of the patient, showing tumor cells and extensive infiltration of smaller lymphoid cells (courtesy of Drs. G. Masucci and R. Kiessling, Department of Oncology and Pathology, Karolinska Institutet, Stockholm, Sweden). Images adjusted for auto levels in Photoshop.

Close modal

Histology of BL common to xenografts and the hEST model. The predominant histology in both xenografts and the hEST model (Fig. 1A–C) was a malignant, pleomorphic, appearance of densely packed BL melanoma cells, with large irregular nuclei, distinct nucleoli, and a less prominent cytoplasm. This cellular morphology was similar to tumor cells in the original BL patient metastasis, as seen from archival-fixed H&E-stained sections from the original tumor (exemplified in Fig. 1D).

Besides a high frequency of atypical mitotic figures (data not shown), all models showed a proliferation rate of ∼30% to 40% as calculated by Ki67 staining (1,000 cells counted in each section). Also, the degree of vascularization in the original metastasis and the experimental models did not apparently differ in that a similar occurrence of vessels was observed in ocular inspections of all sections (data not shown). However, a higher density of cells and lower abundance of stroma was observed in the experimental tumor models (Fig. 1A–C) compared with the original metastasis (Fig. 1D).

Features of BL melanoma cells in the hEST model. In the hEST model, the BL cell injection could not for practical reasons be guided into specific parts or structures of the teratoma. Therefore, the BL cells were injected randomly into the teratoma mass. In spite of this, a specificity of the integration was noticed. Using FISH analysis for human Y chromosome (red signal) and X chromosome (green signal), the BL cells were identifiable from hESC derivates (XX; double green signals). With this method BL cells were exclusively detected in areas compatible with mesenchyme. Typical areas of verified BL cell growth are illustrated in Fig. 2, and a further example also showing the corresponding FISH results are illustrated in Fig. 3. In four repeated experiments, melanoma cells were never detected within areas/tissues compatible with development into primitive neuroectoderm, secretory, or other epithelial structures, cartilage, etc., or in more organoid structures, such as primitive renal tissues or gut.

Figure 2.

Section of hESC teratoma 60 d after hESC injection—with inoculum of 106 BL melanoma cells injected on day 45 (H&E staining; magnification, ×2). An approximate outer border of three separate areas with densely packed BL cells are indicated by a dotted blue line (tumor cells were not detected in other parts of this section). The picture shows a typical hESC teratoma appearance with rich presence of tissues representing the three germ layers (areas compatible with mesenchyme (M), cartilage (C), smooth muscle (SM), neural development (N), adipose tissue (A), respiratory epithelial structures (E), etc. A centrally localized area with the beginnings of necrosis is also indicated). Images adjusted for auto levels in Photoshop.

Figure 2.

Section of hESC teratoma 60 d after hESC injection—with inoculum of 106 BL melanoma cells injected on day 45 (H&E staining; magnification, ×2). An approximate outer border of three separate areas with densely packed BL cells are indicated by a dotted blue line (tumor cells were not detected in other parts of this section). The picture shows a typical hESC teratoma appearance with rich presence of tissues representing the three germ layers (areas compatible with mesenchyme (M), cartilage (C), smooth muscle (SM), neural development (N), adipose tissue (A), respiratory epithelial structures (E), etc. A centrally localized area with the beginnings of necrosis is also indicated). Images adjusted for auto levels in Photoshop.

Close modal
Figure 3.

Illustration of BL melanoma growing in the hEST model. A, overview of section (H&E staining; magnification, ×4). B, higher magnification (×20) of the black boxed area in A, showing top left area dense BL growth and down right adjacent area with desmoplastic spindle-like cells compatible with loose mesenchyme. Two white boxes, areas also shown in C and D, respectively, with FISH analysis positively identifying BL-cells (red signal, Y chromosome; green signal, X chromosome) and hESC-derived cells (XX; double green signal). Nuclear counterstaining with DAPI (blue; magnification, ×100). C, dominating presence of BL cells at the outer border of the dense growth. D, presence of solitary BL cells (arrows) in the loose human mesenchyme. Images adjusted for auto levels in Photoshop.

Figure 3.

Illustration of BL melanoma growing in the hEST model. A, overview of section (H&E staining; magnification, ×4). B, higher magnification (×20) of the black boxed area in A, showing top left area dense BL growth and down right adjacent area with desmoplastic spindle-like cells compatible with loose mesenchyme. Two white boxes, areas also shown in C and D, respectively, with FISH analysis positively identifying BL-cells (red signal, Y chromosome; green signal, X chromosome) and hESC-derived cells (XX; double green signal). Nuclear counterstaining with DAPI (blue; magnification, ×100). C, dominating presence of BL cells at the outer border of the dense growth. D, presence of solitary BL cells (arrows) in the loose human mesenchyme. Images adjusted for auto levels in Photoshop.

Close modal

Albeit at a lower frequency compared with the predominant growth mode common to both models, the hEST model revealed an additional BL phenotype not present in the xenografts. This phenotype was characterized by a desmoplastic and spindle-like appearance and occurred adjacent to the areas of dense growth and within the surrounding human loose mesenchymal or fibrous stroma (Fig. 3A and B), as positively identified by FISH analysis (Fig. 3).

A notable human neovascularization was indicated in the surrounding stroma by staining for human CD31 (Fig. 4A). Vascularization within areas of dense tumor growth was negative for human CD31 indicating mouse origin (data not shown).

Figure 4.

Sections from BL melanoma growing in the hEST model (adjacent to Fig. 3B). A, IHC for human CD31, indicating presence of human neovascularization in the human stroma (magnification, ×20). B, Masson Trichrome staining showing presence of collagenous stroma (blue) in the loose human mesenchyme. Nuclear stain red (magnification, ×20). C, IHC for S100A4 showing nuclear staining of cells in the periphery of the densely packed area of BL cells (left) and also in the adjacent human stroma (right); magnification, ×20. D, IHC for S100A4 in the periphery of a BL s.c. xenograft. Single positive cells in the dense area of BL cells (left) and all cells negative in the loose mouse mesenchyme (right; magnification, ×20).

Figure 4.

Sections from BL melanoma growing in the hEST model (adjacent to Fig. 3B). A, IHC for human CD31, indicating presence of human neovascularization in the human stroma (magnification, ×20). B, Masson Trichrome staining showing presence of collagenous stroma (blue) in the loose human mesenchyme. Nuclear stain red (magnification, ×20). C, IHC for S100A4 showing nuclear staining of cells in the periphery of the densely packed area of BL cells (left) and also in the adjacent human stroma (right); magnification, ×20. D, IHC for S100A4 in the periphery of a BL s.c. xenograft. Single positive cells in the dense area of BL cells (left) and all cells negative in the loose mouse mesenchyme (right; magnification, ×20).

Close modal

Melanomas have been reported to be capable of vasculogenic mimicry to support their own growth and progression, i.e., by up-regulation of endothelial-associated genes such as VE-cadherin participating in the angiogenic process (2729). However, using FISH analysis, we failed to detect BL cells in the endothelial cells lining vessels. Instead, double signals for an X-specific probe suggested that the observed neovascularization was mainly of hESC origin (data not shown).

Using Masson Trichrome staining for collagenous matrix, an increased presence of collagenous stroma was seen along the front (Fig. 4B), compared with the more central areas of the BL tumor, which were generally poor in stroma.

Influence of the injected melanoma BL cells on the formation of the hESC-teratoma. The wide developmental pluripotency of the hESC lines used did not seem to be affected by the injections of BL melanoma cells into the developing teratomas. Similar to our previous observations, the hESC teratomas showed a progressive chaotic formation of tissues with marker expression and morphology indicative of all three germ layers, occurring with or without injections of BL cells (data not shown). Furthermore, in line with earlier findings of benign development in teratomas from karyotypically normal hESC (11, 12), no histological signs of malignancy in the hESC-derived tissues could be found in any of the preparations.

Cellular differentiation markers linked to malignant phenotype. S100A4 is a marker linked to migration and invasive capacity of malignant melanomas (30). Staining using monoclonal antibodies specific for the human S100A4 subtype (31) revealed a scattered expression in all models with single cells or small clusters in the dense areas. However, in the hEST model, a specific S100A4 pattern was observed with high expression in cells in periphery of the dense areas and also within the adjacent human stroma, including both melanoma cells and fibroblasts (Fig. 4C). This contrasted to the peripheral parts of the xenografted tumors, which showed more dispersed single S100A4+ cells. The adjacent mouse stroma did not show presence of human S100A4+ cells, as illustrated in Fig. 4D. Staining using a polyclonal antibody for the family of S100-proteins, often used in routine histopathology as a neuroectodermal indicator, revealed stained areas of varying intensity in all three models (data not shown).

HMB45 and MelanA and are markers used in the clinic (often in combination with tyrosinase) to support the diagnosis of malignant melanoma. Regarding the BL melanoma cell line, original material from the patient metastasis was diagnosed as only partly HMB45 positive (historical information from patient journal). In vitro cultured BL cells were negative for both HMB45 and MelanA when analyzed by immunohistochemistry (IHC; data not shown; ref. 17). In contrast, after injections into sc or it locations in vivo, an extensive cytoplasmic expression of HMB45, MelanA, and tyrosinase could be detected (data not shown; illustrated for MelanA in Fig. 5A). Interestingly, BL cells injected into the hEST model remained generally negative for all three markers, with the exception of few dispersed cells positive for MelanA (data not shown; Fig. 5B and C).

Figure 5.

A, IHC for MelanA of BL s.c. xenograft, showing strong positive reactivity (magnification, ×20). B and C, IHC for MelanA of BL cells growing in the hEST model; B, area of dense growth mainly negative; C, adjacent human stroma mainly negative (both ×20 magnification). D, IHC for Nodal of BL growing in the hEST model (magnification, ×20), showing extracellular granular staining in the border area between dense BL growth and the surrounding loose mesenchyme (box inset, ×100 magnification).

Figure 5.

A, IHC for MelanA of BL s.c. xenograft, showing strong positive reactivity (magnification, ×20). B and C, IHC for MelanA of BL cells growing in the hEST model; B, area of dense growth mainly negative; C, adjacent human stroma mainly negative (both ×20 magnification). D, IHC for Nodal of BL growing in the hEST model (magnification, ×20), showing extracellular granular staining in the border area between dense BL growth and the surrounding loose mesenchyme (box inset, ×100 magnification).

Close modal

Nodal, a marker linked to dedifferentiation and a neural crest-like phenotype, was rarely expressed in xenografted cells and then only in very few dispersed cells (data not shown). In the hEST model, Nodal-expressing cells were, almost exclusively, present focally at the border of invading fronts (Fig. 5D). Nestin and nerve growth factor receptor (NGFR, also denoted p75) are two other markers reportedly linked to dedifferentiation of melanoma cells (32, 33). Both Nestin and nerve growth factor receptor were generally positive in the BL melanoma cells in all models (data not shown).

A summary of the marker studies in the different models is shown in Table 1.

The presented findings verify the proof of concept for the hEST model as an alternative and novel in vivo platform for studies of human tumor cell lines, as previously suggested (6, 7) for several other human cell lines of different tumor origins. Importantly, however, we show the feasibility of molecular studies on in vivo tumor progression by comparing this model with conventional xenograft models (after sc or it injections) using markers linked to differentiation, malignancy, and invasiveness of melanomas.

The BL melanoma cell line used here is negative for the markers HMB45 and MelanA (this study and ref. 17). In vivo xenografts of the BL melanoma cells showed high expression of HMB45, MelanA, and Tyrosinase, a combination of three differentiation markers commonly used for the clinical diagnosis of malignant melanoma. When injected into the hEST model, the engrafted BL cells, however, remained negative for all three markers. Thus, the xeno-environment resulted in the expression of a set of differentiation markers that were not induced in the species-specific embryonic environment present in the hEST model.

Furthermore, Nodal, a marker linked to dedifferentiation and neural crest cells, showed the opposite occurrence. This marker was almost exclusively found expressed in the hEST model, at the border of invading BL fronts, and was generally negative in all other tested locations. There are several possible explanations for these differences and continued studies are now needed to distinguish the effect of species-specific interactions from influence of developmental cues (embryonic environment versus adult tissues). Potentially, the findings could also be of interest in the context of using these diagnostic markers for malignant melanoma.

In the xeno-models, although showing a malignant, pleomorphic, appearance, and also expression of the HMB45, MelanA, and tyrosinase markers, the BL growth was homogeneous with no signs of invasive behavior. A second type of growth was however observed in the hEST model, with areas of BL cells apparently migrating into the surrounding human stroma. A migrating phenotype of these cells was indicated by a positive staining for the S100A4 protein, a marker for malignant melanomas linked to migration and invasive capacity (30). This link is believed to be exerted by the ability of S100A4 to remodel the cytoskeleton or extracellular matrix via interactions with intermediate filaments such as actin. Several members of the S100 protein family also interact with p53, giving these proteins further ability to affect tumorigenesis in melanomas (23, 30, 3436). The exclusive finding of Nodal-expressing melanoma cells in the invading population is interesting because the embryonic morphogen Nodal has been found to be important for the invasiveness and tumorigenicity of melanomas (17). Specific inhibition of Nodal signaling could also promote reversion of a metastatic melanoma toward a less invasive and nontumorigenic phenotype (37). Considering this, it may be argued that the expression of genes linked to dedifferentiation, combined with the absence of markers linked to differentiation, suggests a more aggressive phenotype of the migrating BL melanoma cells in the hEST model. Thus, the hEST model allows us to compare melanoma cells in delimited dense focal areas, with melanoma cells migrating into surrounding human stroma in vivo. A second very important finding was the active human neovascularization occurring in the areas of human mesenchymal tissues facing the invasive melanoma growth in the hEST model. This finding is in line that of Tzukerman and colleagues (6) who reported extensive human neovascularization adjacent to and within growth of the A431 epidermoid carcinoma cell line, when using a similar experimental set-up. Contrary to Tzukerman and colleagues (6), we found vascularization within the tumor to be only of mouse origin. Because the development of mature human vessels in hESC teratomas is generally weak (11, 15), the integration of injected BL melanoma cells into the hESC teratoma in our system was initially prioritized close to mouse vessels. The observed strong human angiogenic response was probably a later event, occurring after integration and initial growth of melanoma cells. Because numerous anticancer therapies aim to interact with the angiogenic potential of tumors, this gives an opportunity both for exploring molecular signal systems regulating neovascularization and for the development and testing of new and more effective antiangiogenic therapies. The observed neovascularization also further stresses the importance of species-specific systems because the angiogenic response in this case was from the hESC-derived cells, not from the mouse endothelial cells or the BL melanoma.

From earlier studies, we know that the vast majority of cells in this hESC-teratoma microenvironment are of human origin but that host cells contribute to specific microstructures in the teratoma. This is predominantly, but not solely, restricted to the endothelial cells of the vessels, a few remaining and degenerate seminiferous tubules, and the supply of host blood cells (10). The human embryonic tissues show a remarkable degree of organization, even organoid development, and should be regarded as allogeneic in that undifferentiated hESC express not only low levels class I transplantation antigens (38) but also ABO antigens (39). Furthermore, the differentiated hESC derivatives show increased expression of histocompatibility antigens (40). The use of immunodefective SCID-BEIGE mutant recipients, which lack both T and B lymphocytes, as well as having a selective impairment of natural killer cell functions and macrophage defects (41) efficiently abrogates most immune recognition from the host. Thus far, we have never seen any signs of rejection in the hEST model.

The hEST model may provide a means for studies on tissue tropism of migrating cells. BL cells were injected randomly into the teratomas and therefore not intentionally guided into any specific tissue. They exclusively ended up in areas mostly compatible with condensed mesenchyme. Such areas are also compatible with neural crest development and in this respect our data resemble the results of Kulesa and colleagues (22). In a chicken model, they found that transplanted melanoma cells were distributed along host neural-crest-cell migratory pathways. Notably, no BL growth could be observed in tissues known to be targets of metastatic lesions for melanomas, such as lung epithelium. However, this might be expected considering the long in vitro adaptation of this cell line, as well as its origin from a regional lymph node metastasis and not a visceral metastatic tumor. Ongoing experiments with unmanipulated clinical melanoma tissue may shed further light on this.

The BL cell line carries one of the common activating oncogene mutations associated with cutaneous melanoma, the NRASQ61R mutation (41). We have previously shown that several characteristics of the malignant phenotype of BL cells seem to be associated with oncogenic NRAS signaling (42, 43). Knocking down the expression of NRASQ61R by specific siRNAs led to marked effects of cellular signaling downstream of oncogenic NRAS, as well as inhibition of proliferation, induction of apoptosis, and inhibition of migration and invasion of BL cells grown in vitro. These results indicated that oncogenic NRAS is essential for the malignant behavior of human melanoma cells carrying such a mutation. This oncogene and its effectors may be promising targets for future specific therapy in melanoma. We anticipate that the hEST system will now allow us to closely analyze the effects of interference with oncogenic NRAS signaling, as well as many other molecular aspects of human tumor progression in vivo.

No potential conflicts of interest were disclosed.

Grant support: Swedish Research Council (L. Ährlund-Richter), Petrus & Augusta Hedlunds Stiftelse (L. Ährlund-Richter), the Radiumhemmet Research Funds (L. Ährlund-Richter and J. Hansson) and Karolinska Institutet (L. Ährlund-Richter and J. Hansson), the Swedish Cancer Society, and the Norwegian Research Council (G.M. Gunhild). S. Jamil was the recipient of a stipend from the Higher Education Commission, Pakistan. Y.F. Cheng was the recipient of a scholarship from the Yen Tjing Ling Medical Foundation, Taiwan.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Drs. G. Masucci and R. Kiessling for supplying H&E sections of the original BL tumor.

1
Levenberg S, Huang NF, Lavik E, Rogers AB, Itskovitz-Eldor J, Langer R. Differentiation of human embryonic stem cells on three-dimensional polymer scaffolds.
PNAS
2003
;
100
:
12741
–6.
2
Smalley KS, Lioni M, Noma K, Haass NK, Herlyn M. In vitro three-dimentional tumor microenvironment models for anticancer drug discovery.
Expert Opin Drug Discov
2008
;
3
:
1
–20.
3
Seftor EA, Meltzer PS, Kirschmann DA, Margaryan NV, Seftor RE, Hendrix MJ. The epigenetic reprogramming of poorly aggressive melanoma cells by a metastatic microenvironment.
J Cell Mol Med
2006
;
10
:
174
–96.
4
Kuperwasser C, Chavarria T, Wu M, et al. Reconstruction of functionally normal and malignant human breast tissues in mice.
PNAS
2004
;
101
:
4966
–71.
5
Chen J, Blackwell TW, Fermin D, et al. Evolutionary-conserved gene expression response profiles across mammalian tissues.
OMICS
2007
;
11
:
96
–115.
6
Tzukerman M, Rosenberg T, Ravel Y, Reiter I, Coleman R, Skorecki K. An experimental platform for studying growth and invasiveness of tumor cells within teratomas derived from human embryonic stem cells.
PNAS
2003
;
100
:
13507
–12.
7
Tzukerman M, Rosenberg T, Reiter I, et al. The influence of a human embryonic stem cell-derived microenvironment on targeting of human solid tumor xenografts.
Cancer Res
2006
;
66
:
3792
–81.
8
Hovatta O, Mikkola M, Gertow K, et al. A culture system using human foreskin fibroblasts as feeder cells allows production of human embryonic stem cells.
Hum Reprod
2003
;
18
:
1404
–9.
9
Inzunza J, Gertow K, Strömberg AM, et al. Derivation of human embryonic stem cell lines in serum-free medium using postnatal human fibroblasts as feeder cells.
Stem Cells
2005
;
23
:
544
–9.
10
Gertow K, Przyborski S, Loring J, et al. Isolation of HESC-derived teratomas for the assessment of pluripotency. In: Current Protocols in Stem Cell Biology. Chapt 1B:4, 2007. p.1–24.
11
Gertow K, Wolbank S, Rozell B, et al. Organized development from human embryonic stem cells after injection into immunodeficient mice.
Stem Cells Dev
2004
;
13
:
421
–5.
12
Damjanov I, Ährlund-Richter L. Characterization of hESC-derived teratomas. In Human Stem Cell Manual: A Laboratory Guide, by Loring JF, Schwartz PH, and Wesselschmidt R. Elsevier Press. 2007.
13
Gertow K, Cedervall J, Unger C, et al. Trisomy 12 in HESC leads to no selective in vivo growth advantage in teratomas, but induces an increased abundance of renal development.
J Cell Biochem
2007
;
100
:
1518
–25.
14
Imreh M, Gertow K, Cedervall J, et al. In vitro culture conditions favoring selection of chromosomal abnormalities in human ES cells.
J Cell Biochem
2006
;
99
:
508
–16.
15
Gerecht-Nir S, Osenberg S, Nevo O, Ziskind A, Coleman R, Itskovitz-Eldor J. Vascular development in early human embryos and in teratomas derived from human embryonic stem cells.
Biol Reprod
2004
;
71
:
2029
–36.
16
Salazar-Onfray F, Nakazawa T, Chhajlani V, et al. Synthetic peptides derived from the melanocyte-stimulating hormone receptor MC1R can stimulate HLA-A2-restricted cytotoxic T lymphocytes that recognize naturally processed peptides on human melanoma cells.
Cancer Res
1997
;
57
:
4348
–55.
17
Salazar-Onfray F, Lopez M, Lundqvist A, et al. Tissue distribution and differential expression of melanocortin 1 receptor, a malignant melanoma marker.
Br J Cancer
2002
;
87
:
414
–22.
18
Topczewska JM, Postovit LM, Margaryan NV, et al. Embryonic and tumorigenic pathways converge via Nodal signaling: role in melanoma aggressiveness.
Nat Med
2006
;
12
:
925
–32.
19
Sell S. Stem cell origin of cancer and differentiation therapy.
Crit Rev Oncol Hematol
2004
;
51
:
1
–28.
20
Hendrix MJ, Seftor EA, Hess AR, Seftor RE. Vasculogenic mimicry and tumor-cell plasticity: lessons from melanoma.
Nat Rev Cancer
2003
;
3
:
411
–21.
21
Lee JT, Herlyn M. Microenvironmental influences in melanoma progression.
J Cell Biochem
2007
;
101
:
862
–72.
22
Kulesa PM, Kasemeier-Kulesa JC, Teddy JM, et al. Reprogramming metastatic melanoma cells to assume a neural crest cell-like phenotype in an embryonic microenvironment.
PNAS
2006
;
103
:
3752
–7.
23
Hendrix MJ, Seftor RE, Seftor EA, et al. Transendothelial function of human metastatic melanoma cells: role of the microenvironment in cell-fate determination.
Cancer Res
2002
;
62
:
665
–8.
24
Lee LM, Seftor EA, Bonde G, Cornell RA, Hendrix MJ.
Dev Dyn
2005
;
233
:
1560
–70.
25
Thomson JA, Itskovitz-Eldor J, Shapiro SS, et al. Embryonic stem cell lines derived from human blastocysts.
Science
1998
;
282
:
1145
–7.
26
Imreh MP, Wolbank S, Unger C, et al. Culture and expansion of the human embryonic stem cell line HS181, evaluated in a double color system.
Stem Cells Dev
2004
;
13
:
337
–43.
27
Maniotis AJ, Folberg R, Hess A, et al. Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry.
Am J Pathol
1999
;
155
:
739
–52.
28
Maniotis AJ, Chen X, Garcia C, et al. Control of melanoma morphogenesis, endothelial survival, and perfusion by extracellular matrix.
Lab Invest
2002
;
82
:
1031
–43.
29
Ruf W, Seftor EA, Petrovan RJ, et al. Differential role of tissue factor pathway inhibitors 1 and 2 in melanoma vasculogenic mimicry.
Cancer Res
2003
;
63
:
5381
–9.
30
Garrett SC, Varney KM, Weber DJ, Bresnick AR. S100A4, a mediator of metastasis.
J Biol Chem
2006
;
281
:
677
–80.
31
Flatmark K, Maelandsmo GM, Mikalsen SO, et al. Immunofluorometric assay for the metastasis-related protein S100A4: release of S100A4 from normal blood cells prohibits the use of S100A4 as a tumor marker in plasma and serum.
Tumour Biol
2004
;
25
:
31
–40.
32
Brychtova S, Fiuraskova M, Hlobilkova' A, Brychta T, Hirnak J. Nestin expression in cutaneous melanomas and melanocytic nevi.
J Cutan Pathol
2007
;
34
:
370
–5.
33
Ehrmann J, Kolár Z, Mokry J. Nestin as a diagnostic and prognostic marker: immunohistochemical analysis of its expression in different tumors.
J Clin Pathol
2005
;
58
:
222
–3.
34
Donato R. Intracellular and extracellular roles of S100 proteins.
Microsc Res Tech
2003
;
60
:
540
–51.
35
Mueller A, Schäfer BW, Ferrari S. The calcium-binding protein S100A2 interacts with p53 and modulates its transcriptional activity.
J Biol Chem
2005
;
280
:
29186
–93.
36
Kriajevska M, Fischer-Larsen M, Moertz E, et al. Liprin β 1, a member of the family of LAR transmembrane tyrosine phosphatase-interacting proteins, is a new target for the metastasis-associated protein S100A4 (Mts1).
J Biol Chem
2003
;
277
:
5229
–35.
37
Postovit LM, Margaryan NV, Seftor EA, et al. Human embryonic stem cell microenvironment suppresses the tumorigenic phenotype of aggressive cancer cells.
PNAS
2008
;
105
:
4329
–34.
38
Drukker M, Katz G, Urbach A, et al. Characterization of the expression of MHC proteins in human embryonic stem cells.
PNAS
2002
;
99
:
9864
–9.
39
Chen YT, Dejosez M, Zwaka TP, Behringer RR. H1 and H9 human embryonic stem cell lines are heterozygous for the ABO locus. Stem Cells Dev. Epub ahead of print 2008 Jun 18.
40
Bradley JA, Bolton EM, Pedersen RA. Stem cell medicine encounters the immune system.
Nat Rev Immunol
2002
;
2
:
859
–71.
41
Mosier DE, Stell KL, Gulizia RJ, Torbett BE, Gilmore GL. Homozygous scid/scid;beige/beige mice have low levels of spontaneous or neonatal T cell-induced B cell generation.
J Exp Med
1993
;
177
:
191
–4.
42
Eskandarpour M, Kiaii S, Zhu C, Castro J, Sakko AJ, Hansson J. Suppression of oncogenic NRAS by RNA interference induces apoptosis of human melanoma cells.
Int J Cancer
2005
;
115
:
65
–73.
43
Eskandarpour M, Huang F, Reeves KA, Clark E, Hansson J. Oncogenic NRAS has multiple effects on the malignant phenotype of human melanoma cells cultured in vitro.
Int J Cancer
2009
;
124
:
16
–26.