Uveal melanoma (UM) is uncommon among wild type mice. Efforts to develop transgenic mice to study this disease have resulted in pigmented tumors derived from the retinal pigment epithelium (RPE) or mixed tumors of RPE and UM complicating the study of UM specifically. Reported here are two early stage intraocular amelanotic melanomas discovered in 2 Tyr-HRAS+ Ink4a/Arf heterozygous (1 normal CKDN2A allele) transgenic FVB/n mice. These tumors were morphologically and immunohistochemically similar to spontaneous UM recently reported in the Ink4a/Arf homozygous (CKDN2A knockout) parent strain. The tumors originated in the posterior uveal tract. The neoplasms were comprised of bundles of spindle-shaped melanocytes admixed with some epithelioid cells. Tumors were immunohistochemically positive for neuron-specific enolase, S-100, pan-ras, but negative for cytokeratin and Melan-A. The development of early lenticular opacity and bilateral cataracts is a consistent phenotype of transgenic mice in which the retinoblastoma signaling pathway has been disrupted. Lenticular opacity and cataracts are rarely observed clinically in Tyr-HRAS+ Ink4a/Arf heterozygotes, rendering this strain suitable for ophthalmoscopy. Consequently, Tyr-HRAS+ Ink4a/Arf heterozygotes provide practical advantages, compared to the cataract-prone CKDN2A knockout strains, for real-time ophthalomoscopic detection and monitoring of UM while developing chemotherapeutic regimens and other research to understand the biology of UM.
Introduction
The morphology of amelanotic melanomas and certain other melanoma forms can mimic that of spindle cell sarcomas, carcinomas, and, in the case of desmoplastic melanoma, even scar tissue (McLaren and Ma, 2003; Wick and Patterson, 2003). Moreover, amelanotic melanomas in humans and other mammals are often heterogeneous in expression of typical melanocytic markers such as tyrosinase, Melan-A, HMB-45, S-100, and neuron-specific enolase (Koenig et al., 2001; Smith et al., 2002; Karimipour et al., 2004). In difficult cases where amelanotic melanoma is suspected, a battery of histochemical and immunohistochemical stains is often required for diagnosis, and even then the results may just be presumptive, based on the weight of evidence of the marker profile (McLaren and Ma, 2003; Wick and Patterson, 2003; Lim et al., 2006).
Uveal melanoma (UM) is a devastating disease that accounts for about 12% of all melanomas in human patients (Egan et al., 1988; Char 1997). The unfortunate outcomes are too often enucleation and/or metastasis to distant sites, most commonly liver, lung, bone, and brain (Woll et al.,1999; Shields et al., 2001). As a result, there is significant interest in developing new methods, particularly laboratory animal models, to aid in the understanding of the molecular pathways driving the aggressive nature of this disease. The two principal strategies used to develop animal models to study UM are: (a) implantation of cultured animal or human melanoma cells into the uvea of permissive animal hosts ( Mueller et al., 1999; Bonicel et al., 2000) and (b) use of transgenic mice with increased tendencies to develop UM in vivo (Bradl et al., 1991; Powell et al., 1995; Kramer et al., 1998; Syed et al., 1998). Implantation of cultured tumor cells introduces 2 disadvantages. Cultured tumor cells may develop adaptive changes selected by growth in vitro that result in important biological differences from tumor cells in vivo and the procedure may require immunocompromised hosts to prevent rejection of the xenografted tumor cells.
Genomic analyses of human uveal and cutaneous melanoma cases have identified mechanistic similarities (Hurst et al., 2003). Constitutive activation of the RAS-RAF-ERK signal transduction pathway is often detected in both forms of melanoma, notably via NRAS or BRAF mutations in cutaneous melanomas (Goel et al., 2006). Among uveal melanomas, however, activating mutations are rarely found in RAS or RAF oncogene family members (Zuidervaart et al., 2005), yet constitutive ERK activation is evident in what appears to be a cAMP-dependent protein kinase (PKA)-regulated mechanism (Calipel et al., 2006). In transgenic mice, forced expression of the human HRAS G12V oncogene in melanocytic cells using a recombinant tyrosinase promoter generates melanoma-prone strains that develop mixed pigmented ocular tumors derived from retinal pigment epithelium and melanocytes of the uveal tract (Kramer et al., 1998). Inactivation of the CKDN2A tumor suppressor gene (i.e., Ink4a/Arf gene) is also associated with cutaneous and uveal melanomas (de Snoo and Hayward, 2005; Hovland and Trempe, 2005). Via shared exons 2 and 3, but with separate promoters and first exons, the CKDN2A gene encodes 2 tumor suppressor proteins, p16Ink4a and p14Arf (p19Arf in mice) that regulate the retinoblastoma and P53 signaling pathways, respectively.
Although rare in human populations, inactivating germline mutations in CKDN2A lead to a 70% lifetime risk for developing cutaneous melanoma among affected families (de Snoo and Hayward, 2005). A similar increased risk for familial uveal melanomas has not been established for germline CKDN2A mutations (Soufir et al., 2000; Vajdic et al., 2003). Within tumors, however, acquired CKDN2A inactivation via promoter hypermethylation, mutation, or deletion is associated with cutaneous and uveal melanomas (Merbs and Sidransky, 1999; Edmunds et al., 2002; Straume et al., 2002). In transgenic mice, homozygous deletion of CKDN2A leads to a cancer-prone condition dominated by fibrosarcomas and lymphomas (Serrano et al., 1996). Offspring produced by crossing transgenic mice bearing the Tyr-HRAS oncogene with Ink4a/Arf null mice exhibit approximately 50% incidence of cutaneous melanoma within 6 months (Chin et al., 1997). We recently reported spontaneous amelanotic UM that developed in 20 (15.7%) of 127 male albino Try-HRAS+ Ink4a/Arf−/− transgenic FVB/N mice within 6 months of life and characterized the neoplasms histologically, immunohistochemically, and ultrastructurally (Tolleson et al., 2005). These tumors appeared to be derived from only uvea. One hundred percent of these mice had early bilateral lenticular opacity (LO) associated with cataract formation between the ages of 4 and 7 weeks—before the onset of intraocular tumors that precluded opthalmoscopic examination.
Transgenic Tyr-HRAS+ Ink4a/Arf heterozygous mice, like those used in the present study, have been reported to exhibit a lower incidence of cataract formation than the homozygous null strain (Ink4a/Arf −/−) (Chin et al., 1997), but, to our knowledge, UM has not been reported in the former. Moreover, early LO was not observed in the Ink4a/Arf heterozygous strain. In the current report we describe 2 cases of early UM in one-year-old mice arising in the posterior choroid of Tyr-HRAS+ Ink4a/Arf heterozygous mice free of any LO at necropsy and only very minimal lenticular degenerative changes detected histopathologically. Thus, melanoma-prone Tyr-HRAS+ Ink4a/Arf heterozygotes may provide practical advantages, compared to the cataract-prone CKDN2A knockout strains, for real-time ophthalomoscopic detection and monitoring of UM while developing and testing new chemotherapeutic regimes and in other research designed to better understand the biology of UM.
Case Report
The two mice detected with UM were part of a cutaneous photocarcinogenesis study designed to examine the effects of ultraviolet radiation on cutaneous melanoma formation. The Tyr-HRAS+ Ink4a/Arf heterozygous neonatal mice were exposed to 0, 1.0, or 2.0 J/cm2 of UVA + UVB once a day for 5 days on PND 3–8. The eyelids of these mice remained tightly closed during this period. This is consistent with previous reports that eyelids of mice remain sealed until 12–14 days after birth (Hoar 1982). The eyes from a total of 46 mice [17 controls, 16 low-dose (LD), and 13 high-dose (HD)] were examined. One case each occurred in the LD and HD groups, respectively. LO was observed grossly in only 1 of the 46 mice; neither of the 2 cases of UM described had LO.
The Tyr-HRAS+ Ink4a/Arf heterozygous animals used in our study were produced by crossing Tyr-HRAS+ Ink4a/Arf+/+ males with female Tyr-HRAS− Ink4a/Arf−/−breeders. The original parent Tyr-HRAS+ Ink4a/Arf+/−FVB/N breeders were obtained from the Mouse Models of Human Cancers Consortium, Frederick, MD (http://emice.nci.nih.gov). Animals were provided water and rodent chow (NIH31 feed) ad libitum. Animals were housed and treated according to the National Center for Toxicological Research, Institutional Animal Care and Use Committee, and American Association for Laboratory Animal Science guidelines. Affected animals were killed humanely with carbon dioxide for pathological evaluation.
At necropsy, the eyes were removed and preserved in modified Davidson’s fixative (30% formaldehyde, 15% ethanol, 5% glacial acetic acid, and 50% distilled water) for 24 hours prior to transfer to 70% ethanol before embedding in paraffin (Latendresse et al., 2002). Paraffin sections were stained with hematoxylin and eosin. Additional paraffin sections were de-waxed with xylene and rehydrated in graded ethanol solutions for immunohistochemical staining. Endogenous peroxidase was quenched with 3% H2O2 containing 0.1% sodium azide. The sections were heated for 20 minutes in antigen retrieval solution (10 mM sodium citrate, pH 6.0) using a 700-watt microwave oven on full power. Nonspecific binding was blocked using 0.5% casein in TBS (0.01 M Tris, 0.15 M NaCl, pH 7.2), followed by incubation with mouse monoclonal anti-melan-A antibodies (Vector Laboratories, Inc., Burlingame, CA), rabbit polyclonal anti-S-100 antibodies (Dako Corp., Carpinteria, CA), mouse monoclonal anti-neuron-specific enolase antibodies (Dako Corp), anti-cytokeratin 8 & 18 for retinal pigment epithelium (Pan Cytokeratin Plus, Biocare Medical, Walnut Creek, CA), or rabbit polyclonal antibodies to Ras reactive with human, rat, and mouse (Pan-Ras, Abcam Inc., Cambridge, MA).
Antigen-specific antibody binding was detected using a routine strepavidin staining system with biotinylated goat anti-mouse or anti-rabbit secondary antibodies (Jackson Immunoresearch Laboratories, West Grove, PA); streptavidin-conjugated horseradish peroxidase enzyme label (Jackson Immunoresearch Laboratories, West Grove, PA); and 3,3′-diaminobenzidine hydrochloride (DAB) chromogen. Tissue sections were lightly counterstained with Mayer’s hematoxylin. For negative controls for all specimens, PBS and/or IgG from the same species in which the primary antibody was grown were substituted for primary antibody. Furthermore, a battery of positive and negative control tissues were stained when indicated. The majority of our specimens had normal indigenous control tissues on the same slide as the tumor. In all cases, retina was positive for NSE, optic and peripheral nerves were positive for both S-100 and NSE. Neither of the 2 tumors was Melan-A positive. Both positive and negative controls consistently validated the efficacy of each immunohistochemical (IHC) marker examined.
Both of these early amelanotic uveal melanomas affected the posterior choroid. They were histologically identical to ones detected in the Tyr-HRAS+ Ink4a/Arf−/− parent strain (compare Figure 1 and Figure 2), which were characterized histologically, immunohistochemically, and ultrastructurally (Tolleson et al., 2005). One of the tumors in the current report had invaded the sclera and had extended into the extraocular muscle (Figure 1A). Both neoplasms were composed predominately of spindled melanocytes interspersed with a few cells manifesting plump epitheloid features (Figure 1B insert), and, both cases, the retinal pigment epithelium (RPE) appeared intact overlying the UM (Figure 1B and C). The tumors stained negatively for cytokeratin in contrast to the immunopositive RPE (Figure 3) suggesting that the neoplasms, like those reported by Tolleson et al. (2005), did not develop from the RPE. In this regard, both of these strains are apparently unlike earlier transgenic mice engineered for a predisposition to develop ocular melanoma (Syed et al., 1998; Beermann et al., 1999; Dithmar et al., 2000).
The expression of immunohistochemical melanocytic markers was also similar to those previously reported in the Tyr-HRAS+ Ink4a/Arf−/− parent strain (Tolleson et al., 2005). Both neoplasms were strongly immunopositive for NSE (Figure 4A and 4B1) and S-100 (Figure 4C1 and 4D1) and negative for cytokeratin (Figure 3). Additionally, melanoma cells in both tumors were immunopositive for Pan-Ras (Figure 4E), as were other normal appearing cells in various other ocular tissues. This ubiquitous expression of Ras in neoplastic and nonneoplastic tissues is consistent with previously reported data (Chesa et al., 1987) and highlights the non-specificity of this marker for UM. Melan-A is a melanocyte differentiation antigen localized in the inner membranes of premelanosomes (Wick and Patterson, 2003). Neither of the two tumors was Melan-A positive. The paucity of Melan-A-positive cells in amelanotic melanoma is a reasonable expectation.
In contrast to the 127 mice of the parent Tyr-HRAS+ Ink4a/Arf homozygous null strain that develop clinically detected cataracts between ages 4 and 7 weeks, only 1 cataract was detected at necropsy among the eyes of 46 Tyr-HRAS+ Ink4a/Arf heterozygotes. Impaired cellular differentiation results in transgenic mouse strains when signal pathways involving RB tumor suppressor proteins (pRB, p107, and p130) are disrupted (Gotz et al., 1991; Morgenbesser et al., 1994; Pan and Griep, 1994; Hyde and Griep, 2002;). Transgenic animals with defective RB function often exhibit bilateral cataracts with complete penetrance and early onset. The early cataracts observed by Tolleson et al. (2005) in the CKDN2A null strain were presumably due to significant disruption of either or both RB or P53 signaling pathways. Diminished expression of p16Ink4a and/or p19Arf by CKDN2A heterozygotes with haplosufficiency (one normal CKDN2A allele) appears to create a melanoma-prone phenotype, while still providing sufficient RB- and P53-dependent signaling to allow proper lens development. By exploiting this advantage over the Tyr-HRAS+ Ink4a/Arf−/− strain, which developed cataracts at a very early age before onset of UM, the CKDN2A heterozygote mouse model reported here should allow early clinical detection and monitoring.
The Tyr-HRAS+ Ink4a/Arf heterozygote strain also provides the opportunity to evaluate the roles of other genetic loci in the development and progression of UM via selective crossbreeding with other transgenic mouse strains. In hallmark studies of gene expression profiling using human UM, Onken and coworkers identified gene expression patterns that reliably segregated cases into high- and low-risk categories for metastasis and death and associated biological correlates to the gene expression signature of the high-risk category (Onken et al., 2004, 2006). Biological verification for the role of 1 gene, ID2, in the aggressive uveal melanoma phenotype was established via transgenic crosses. Additional candidate genes were revealed in these studies, and others are likely to exist, based on the results of cytogenetic changes associated with UM, including monosomy 3, 6p, deletions of 3p25.1–3p25.2, and gains at 8q (Hovland and Trempe, 2005).
In addition to heritable traits, such as CKDN2A mutations, that increase melanoma risk in humans, epidemiologic studies have identified potential environmental risk factors, including UV that may play a role in the development or progression of cutaneous and uveal melanomas. Furthermore, the status of cellular defense, detoxification, and repair pathways expressed by melanocytes affect the resistance to toxic endogenous or environmental agents associated with increased melanoma risk (reviewed in Tolleson, 2005). Tyr-HRAS+ Ink4a/Arf heterozygotes provide an animal system in which these factors can be evaluated systematically.
This model should be a valuable tool to determine the efficacy of new chemotherapeutics and other treatments for UM as well as advance our understanding of the biology of UM by being able to monitor progression or regression of the tumor in real-time.
