Portal de Periódicos da CAPES

Keratoacanthomas Frequently Show Chromosomal Aberrations as Assessed by Comparative Genomic Hybridization

2002; Elsevier BV; Volume: 119; Issue: 6 Linguagem: Inglês

10.1046/j.1523-1747.2002.19613.x

1523-1747

O. P. F. Clausen, Marzieh Beigi, Paula M. DeAngelis, Lars Bolund, Steen Kølvraa, Petter Gjersvik, Gro Mørk,

Plant Disease Resistance and Genetics

Keratoacanthomas are commonly occurring benign skin lesions localized to sun-exposed areas. They typically develop rapidly and may show cellular atypia and infiltration like cutaneous squamous cell carcinomas, but they finally regress spontaneously. This benign lesion shows a high degree of genetic instability as assessed by comparative genomic hybridization, with 35.7% (25 of 70) of the analyzed lesions harboring chromosomal aberrations. The same frequency of genetic imbalance was found in lesions from immunosuppressed organ transplant recipients (36.4%, 20 of 55) and in patients with keratoacanthomas without immunosuppression (33.3%, five of 15), indicating a common pathway in both situations. Recurrent aberrations, given as a fraction of lesions with aberrations, were gains on 8q (20.0%), 1p and 9q (each 16.0%), and deletions on 3p (20.0%), 9p (20.0%), 19p (20.0%), and 19q (16.0%). Many of the most frequently appearing aberrations in keratoacanthomas were not detected in any of the 10 squamous cell carcinomas analyzed, whereas some aberrations were shared by both types of lesions. Aberrations were found in early and late stages of keratoacanthoma development, indicating a role for genetic instability in the progression as well as involution of keratoacanthomas. There were no significant correlations between cytologic atypia and genetic imbalance, or between degree of infiltration and genetic aberrations, although there was a trend for keratoacanthomas with severe atypia to have aberrations. Thus malignant phenotypic development does not appear to be driven by the detected genetic aberrations. More detailed studies of chromosomal areas with recurrent aberrations are needed for the localization of putative genes that determine the biologic behavior of keratoacanthomas, and that may distinguish them from squamous cell carcinomas. Keratoacanthomas are commonly occurring benign skin lesions localized to sun-exposed areas. They typically develop rapidly and may show cellular atypia and infiltration like cutaneous squamous cell carcinomas, but they finally regress spontaneously. This benign lesion shows a high degree of genetic instability as assessed by comparative genomic hybridization, with 35.7% (25 of 70) of the analyzed lesions harboring chromosomal aberrations. The same frequency of genetic imbalance was found in lesions from immunosuppressed organ transplant recipients (36.4%, 20 of 55) and in patients with keratoacanthomas without immunosuppression (33.3%, five of 15), indicating a common pathway in both situations. Recurrent aberrations, given as a fraction of lesions with aberrations, were gains on 8q (20.0%), 1p and 9q (each 16.0%), and deletions on 3p (20.0%), 9p (20.0%), 19p (20.0%), and 19q (16.0%). Many of the most frequently appearing aberrations in keratoacanthomas were not detected in any of the 10 squamous cell carcinomas analyzed, whereas some aberrations were shared by both types of lesions. Aberrations were found in early and late stages of keratoacanthoma development, indicating a role for genetic instability in the progression as well as involution of keratoacanthomas. There were no significant correlations between cytologic atypia and genetic imbalance, or between degree of infiltration and genetic aberrations, although there was a trend for keratoacanthomas with severe atypia to have aberrations. Thus malignant phenotypic development does not appear to be driven by the detected genetic aberrations. More detailed studies of chromosomal areas with recurrent aberrations are needed for the localization of putative genes that determine the biologic behavior of keratoacanthomas, and that may distinguish them from squamous cell carcinomas. comparative genomic hybridization squamous cell carcinoma loss of heterozygosity Keratoacanthomas are common benign keratinocyte tumors mainly arising in sun-exposed skin of elderly persons. The lesion has long been recognized as an entity that can be differentiated from squamous cell carcinomas (SCC), which it may resemble clinically as well as histologically. It occurs mostly as a solitary lesion, although multiple lesions may appear. Keratoacanthomas typically appear as dome-shaped nodules with a centrally developing crater filled with keratin, and reach a size of 1.0–2.5 cm in diameter within 6–8 wk, whereafter they spontaneously involute, mostly within 6 mo (Schwartz, 1994Schwartz R.A. Keratoacanthoma.J Am Acad Dermatol. 1994; 30: 1-19Abstract Full Text PDF PubMed Scopus (250) Google Scholar;Elder et al., 1997Elder D. Elenitsas R. Jaworsky C. Johnson Jr, B. Lever's Histopathology of the Skin. 8th edn. Lippincott-Raven Publishers, Philadelphia1997: 731-734Google Scholar). In many instances the time of development is much longer, and lesions may persist for more than 1 y. Keratoacanthomas may appear as multiple lesions of the Ferguson-Smith type, and are then assumed to be inherited in an autosomal dominant manner (Ferguson-Smith et al., 1971Ferguson-Smith J.A. Wallace D.C. James Z.H. et al.Multiple self-healing squamous epithelioma.Birth Defects. 1971; 7: 157-163PubMed Google Scholar). They may also occur as the multiple eruptive keratoacanthomas of Grzybowski, the cause of which is unknown (Grzybowski, 1950Grzybowski M. A case of peculiar generalized epithelial tumors of the skin.Br J Dermatol Syphilol. 1950; 62: 310-313Crossref PubMed Scopus (80) Google Scholar). Cutaneous eruption of one or more keratoacanthomas may also appear in the Muir–Torre syndrome, characterized by low-grade visceral cancers usually of gastrointestinal or urologic origin (Muir et al., 1967Muir E.G. Bell A.J.Y. Barkow K.A. Multiple primary carcinomata of the colon, duodenum, and larynx associated with keratoacanthoma of the face.Br J Surg. 1967; 54: 191-195Crossref PubMed Scopus (318) Google Scholar). The genetic basis for this syndrome includes germline mutations in mismatch repair genes (Bapat et al., 1996Bapat B. Xia L. Madlensky L. et al.The genetic basis of Muir–Torre syndrome includes the hMLH1 locus.Am J Hum Genet. 1996; 59: 736-739PubMed Google Scholar). An increased incidence of keratoacanthoma is seen among immunosuppressed patients (Sullivan and Colditz, 1979Sullivan J.J. Colditz G.A. Keratoacanthoma in a subtropical climate.Australas J Dermatol. 1979; 20: 34Crossref PubMed Scopus (27) Google Scholar). The histopathologic diagnosis of keratoacanthomas is based on their architecture as well as cytologic features, and when their typical clinical history is known most keratoacanthomas can be distinguished from SCC (Elder et al., 1997Elder D. Elenitsas R. Jaworsky C. Johnson Jr, B. Lever's Histopathology of the Skin. 8th edn. Lippincott-Raven Publishers, Philadelphia1997: 731-734Google Scholar). Particularly in their early proliferative phase, keratoacanthomas share several features with SSC, such as infiltration and cytologic atypia. There is no single criterion that is sufficiently sensitive to distinguish reliably between keratoacanthoma and SCC (Cribier et al., 1999Cribier B. Asch P.-H. Grosshans E. Differentiating squamous cell carcinoma from keratoacanthoma using histopathological criteria.Dermatology. 1999; 199: 208-212Crossref PubMed Scopus (110) Google Scholar). There are several reports in the literature of lesions classified as keratoacanthomas that have metastasized, which have led to questions about keratoacanthomas being a distinct entity, or alternatively a variant of SCC (Hodak et al., 1993Hodak E. Jones R.E. Ackerman A.B. Solitary keratoacanthoma is a squamous cell carcinoma: three examples with metastases.Am J Dermatopathol. 1993; 15: 332-342Crossref PubMed Scopus (198) Google Scholar). To characterize better the biology of keratoacanthomas, we analyzed fresh frozen material from lesions clinically suspected to be keratoacanthomas. To assess genetic aberrations we used comparative genomic hybridization (CGH), a technique that allows simultaneous detection of amplified or deleted chromosomal regions of the total genome of neoplasms (Kallioniemi et al., 1992Kallioniemi A. Kallioniemi O.-P. Sudar D. Rutovitch D. Gray J.W. Waldman F. Pinkel D. Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors.Science. 1992; 258: 818-821Crossref PubMed Scopus (2774) Google Scholar). Such regions may harbor putative oncogenes or tumor suppressor genes, respectively, the localization of which will be of importance to understand better and investigate further the nature of neoplastic lesions. In particular, such information about keratoacanthomas will give us a better basis for understanding their ability to regress spontaneously, and eventually to distinguish them from SCC. Ninety lesions clinically suspected to be keratoacanthomas at the departments of dermatology, surgery, and plastic surgery at Rikshospitalet, Oslo, during 1995–2000 were excised and analyzed. The majority of patients were organ transplant recipients who received immunosuppressive treatment, but a significant number of patients presented lesions without being transplanted or receiving immunosuppression. Some patients with lesions expected to be SCC were included. The lesions were divided into two halves, one of which was fixed in formalin, embedded in paraffin, and processed for routine histopathologic evaluation; the other half was immediately stored at –80°C for further processing. Sections were cut at 5 μm thickness from the formalin-fixed material and stained with hematoxylin–eosin for routine histopathologic diagnosis. Criteria used for the diagnosis of keratoacanthoma and actual differential diagnoses were according to Elder et al., 1997Elder D. Elenitsas R. Jaworsky C. Johnson Jr, B. Lever's Histopathology of the Skin. 8th edn. Lippincott-Raven Publishers, Philadelphia1997: 731-734Google Scholar. Briefly, lesions were characterized by symmetrical appearance, with a centrally located, keratin-filled crater with overhanging epithelial lips or shoulders. From the center of the lesions epithelial strands, often with ground glass appearance, were penetrating the dermis. According to generally accepted criteria, we have included all such lesions, irrespective of the degree of cellular atypia or infiltrating growth. The lesions classified as keratoacanthomas were evaluated with respect to the following parameters: size of the lesion, degree of fibrosis and inflammation, cellular atypia, and infiltration. The age of the lesion as given by the patient at the time of excision was also recorded—if a realistic estimate was possible to make. The diameter of the lesion was measured in centimeters from the mid-section, the other parameters were scored as absent or mild (+), moderate (++), or severe (+++) by one of the authors (OPFC), who is an experienced dermatopathologist. Infiltration was judged to be absent when the growth was expansive, moderate, when it grew with finger-like, several cell layers thick extensions penetrating dermis, and severe, when small groups or single cells were invading dermis. For statistical purposes, the lesions were considered young when reported less than 4 wk old, and old when more than 6 mo. They were histologically old when moderate degree, or more, of fibrosis and inflammation, respectively, was seen, i.e., more than ++ in each category. Young lesions had either no, or mild fibrosis and inflammation, respectively, i.e., + or less in each category, respectively. The three degrees of atypia, infiltration, inflammation, and fibrosis were used in the correlation analysis. The samples stored at –80°C were analyzed by CGH as previously described (De et al., 1999De Angelis P.M. Clausen O.P.F. Schjølberg A. Stokke T. Chromosomal gains and losses in primary colorectal carcinomas detected by CGH and their associations with tumor DNA aneuploidy, genotypes and phenotypes.Br J Cancer. 1999; 80: 526-535Crossref PubMed Scopus (135) Google Scholar). Briefly, hybridization mixtures were prepared by mixing fluorescein isothiocyanate (FITC) labeled tumor DNA, Texas red labeled normal DNA, Cot-1 DNA, 3 M sodium acetate, and 100% ethanol. Slides were then incubated at 37°C in a humidified chamber for 3 d, washed in formamide solutions, and counterstained with 0.2 μM 4′-6′-diamidino-2-phenylindole. 4′-6′-diamidino-2-phenylindole, FITC, and Texas red fluorescence images from eight to 10 metaphases from each neoplasm were captured and digitized using the ISIS in situ imaging system (Metasystems, Altlussheim, Germany). Metaphase chromosomes were then segmented and green/red ratio profiles calculated using ISIS software. Amplifications and deletions were scored if the green/red ratio was above 1.15 and below 0.85, respectively, as described previously (De et al., 1999De Angelis P.M. Clausen O.P.F. Schjølberg A. Stokke T. Chromosomal gains and losses in primary colorectal carcinomas detected by CGH and their associations with tumor DNA aneuploidy, genotypes and phenotypes.Br J Cancer. 1999; 80: 526-535Crossref PubMed Scopus (135) Google Scholar). Most cases were hybridized at least twice, and all cases with gains of chromosomes or chromosome arms with many repetitive DNA sequences (e.g., chromosome 19) were subjected to reverse hybridizations to assess and avoid artifacts of staining. In these cases tumor DNA was labeled with Texas red instead of FITC, and control DNA with FITC instead of Texas red (Kallioniemi et al., 1994Kallioniemi O.-P. Kallioniemi A. Piper J. Isola J. Waldman F.M. Gray J.W. Pinkel D. Optimizing comparative genomic hybridization for analysis of DNA sequence copy number changes in solid tumors.Genes Chromosomes Cancer. 1994; 10: 231-243Crossref PubMed Scopus (936) Google Scholar). Statistical analysis was done using SPSS 9.0 software (SPSS, Chicago, IL). The Kendalls–Tau b-test and the χ2 test for correlations between parameter values was used. We also used paired sample statistics for testing the degree of intraobserver reproducibility of histopathologic scoring. The p-values (two-tailed) and κ values are given; p 0.610 as good. By using generally accepted criteria for histopathologic classification, 70 lesions were diagnosed as keratoacanthomas, 10 as verrucous/hyperplastic benign lesions without atypia, and 10 lesions as SCC. Fifteen of the 70 keratoacanthomas were found in nine patients without immunosuppression, whereas 55 keratoacanthomas were harvested from 18 immunosuppressed patients. Patients with multiple lesions showed heterogeneity, with lesions lacking as well as having aberrations, and the aberrations of multiple lesions from the same patient were different. The histopathologic classification of lesions into keratoacanthomas and SCC was the same on two occasions of blind classification, and the κ value for reproducibility of grading of atypia and fibrosis was 0.626 and 0.793, respectively, which is characterized as good. Utilization of reverse hybridization resulted in some of the chromosome 19 amplifications not being confirmed, consistent with previous reports (Kallioniemi et al., 1994Kallioniemi O.-P. Kallioniemi A. Piper J. Isola J. Waldman F.M. Gray J.W. Pinkel D. Optimizing comparative genomic hybridization for analysis of DNA sequence copy number changes in solid tumors.Genes Chromosomes Cancer. 1994; 10: 231-243Crossref PubMed Scopus (936) Google Scholar). Only confirmed aberrations were taken into account. Genetic aberrations detected by CGH were found in 35.7% (25 of 70) of keratoacanthomas, in 70% (seven of 10) of SCC, and in none of the 10 verrucous or hyperplastic benign lesions. Among keratoacanthomas in nonimmunosuppressed patients 33.3% (five of 15) showed genetic aberrations, whereas 36.4% (20 of 55) of lesions in immunosuppressed patients showed aberrations. The distribution of genetic aberrations among keratoacanthomas from immunocompetent and immunoincompetent patients, and from SCC are shown in Figure 1. The most frequent aberrations were gains on 8q (20.0%), 1p and 9q (each 16.0%), and losses on 3p (20.0%), 9p (20.0%), 19p (20.0%), and 19q (16.0%). Percentages indicate fractions of the particular chromosomal localization among lesions with aberrations. The total number of gains and losses was about the same within keratoacanthomas (i.e., 49 vs 45, respectively), and the mean number of detectable aberrations per lesion was 3.8 (range 1–13). Although the fraction of keratoacanthomas harboring aberrations were similar in individuals with and without immunosuppression, the number of aberrations per genome was higher in immunocompetent cases (7.6 aberrations vs 2.8). The localization of genetic aberrations in all keratoacanthomas and SCC are shown in Table I, a summary of the genetic aberrations of keratoacanthomas is shown in Figure 2, and a typical example of a karyogram from a keratoacanthoma is shown in Figure 3.Table ISummary of genetic aberrations of keratoacanthomas and SCCTumourGainsLossesKeratoacanthomas (transplanted) K3523A-9513q17p12-ter, 19p, 19q K3A-953p, 8p12-ter, 9p, 21q21-ter K1E-961q22-ter, 8q24.2-ter, 9q34.1-ter, 20q11.2-ter1pcen-31.3 K2649A-9619p13.2-ter, 19q13.2-ter K2A-961p34.1-ter, 9q33-ter, 15q22.3-25, 16p12-ter, 16q23-ter, 17p12-ter 19p13.2-ter, 19q K7464A-965p, 9qcen-22.1, 11qcen-14.13pcen-21.3, 5qcen-21, 7q31.1-ter, 9p13-ter, 11p, 11q23.3-ter, 17p13-ter, 19p, 19q, 21q22.1-ter K4A-967qcen-32, X K7A-9619p K3A-984p13.2-ter4q, 15q K3C-989p, 10pcen-22.2, 10q26.1-ter K3F-9819p K3G-9810q21.1-26.1 K3C-998q21.1-ter K5A-99X K14A-9917q22-ter K16A-993q24-ter, 9q31-ter3p, 11q22.1-ter K22B-9912q24.1-ter K1A-003p24.1-ter, 3q21-ter, 7q21.1-ter4q28-ter, 5q14-31.3, 9p, 18q9q21.1-ter, 18p, 19q13.2-ter K3C-004p14-ter, 22q12.2-ter K4B-003pcen-14.2, 5q12-terKeratoacanthomas (non-transplanted) K1A-951q32.3-ter, 4p14-ter, 8q22-ter, 17p10q, 13qcen-22 K2067A-951p34.2-ter, 1q, 16p, 19, 22q5q21-23.1, 9q21.3-34.2, 13q21.3-32 K3A-961p34.2-ter, 5p15.1-ter, 6pcen-22.2,2q21.3-32.1, 9p, 13q7p13-ter, 8p22-ter, 8q21.1-ter, 17 K2A-973p14.2-ter K2A-981p34.2-ter, 16p13.2-ter, 19Squamous cell carcinomas K2B-951p31.3-ter, 1qcen-22, 9q33-ter,4q22-31.1, 5q15-23.3, 6qcen-2111qcen-14.3, 12q24.1-ter, 15q22.1-25,13q14.3-3116, 17, 19, 20q, 21q21-ter, 22q, X K5A-981p32.2-ter, 1q32.1-ter, 8q24.1-ter, 194q31.2-ter K9A-991qcen-25, 20q K3B-0014q31-ter, 15q24-ter K3E-001p34.1-ter, 6q16.1-ter, 8q12-ter,4qcen-31.114q24.2-ter, 20 K5B-008q8p K5C-008q13-24.28p21.1-ter Open table in a new tab Figure 3A typical karyogram from one keratoacanthoma (lesion K1A00, Table II) with the detected aberrations. Gains are indicated as bars to the right of the chromosomes and losses as bars to the left.View Large Image Figure ViewerDownload (PPT) Correlation analyses were done by the Kendalls–Tau b-test as well as with the χ2 test, and the results were similar with respect to significance. The reported age of the lesions (estimated by the patients) correlated significantly with the histopathologic age estimated from fibrosis and inflammation (p=0.003), as well as with fibrosis alone (p=0.003). The degree of fibrosis was also positively correlated with the degree of inflammation (p=0.001). Cytologic atypia correlated significantly with the degree of fibrosis (p=0.001). Genetic aberrations were randomly distributed among young and old keratoacanthomas, irrespective of how age was estimated (i.e., either from histopathology or from the patients' reports) (data not shown). Genetic imbalance was also distributed among lesions with all degrees of atypia, as well as with all degrees of infiltration. This is now shown in Table II (a) and Table II (b), respectively. It can be seen from Table II (a) that there is a tendency for keratoacanthomas with severe atypia to have genetic aberrations, although most lesions with aberrations were seen in the other two categories. When considering separately the 15 keratoacanthomas from immunocompetent individuals, none of them harbored severe atypia, but two lesions had moderate atypia. Amplification of 4p correlated significantly with the presence of fibrosis (p=0.005), as well as with histopathologically estimated age of the lesion (p=0.027) (χ2).Table II(a) Correlations between different degrees of nuclear atypia and the presence of genetic aberrations assessed by CGH in keratoacanthomas. No statistical significance was found, although there was a trend for keratoacanthomas with severe atypia to have aberrations (χ2 test)Degree of atypiaGenetic aberrationsTotalAbsentPresentAbsent or mild31 (72.1%)12 (27.9%)43 (100%)Moderate10 (66.7%)5 (33.3%)15 (100%)Severe4 (33.3%)8 (66.7%)12 (100%)Total45 (64.3%)25 (35.7%)70 (100%)(b) Correlations between the degree of infiltration and the presence of genetic aberrations assessed by CGH in keratoacanthomas. No statistical significance was found (χ2 test)Degree of atypiaGenetic aberrationsTotalAbsentPresentAbsent or mild13 (72.2%)5 (27.8%)18 (100%)Moderate14 (63.6%)8 (36.4%)22 (100%)Severe18 (60.0%)12 (40.0%)30 (100%)Total45 (64.3%)25 (35.7%)70 (100%) Open table in a new tab To the best of our knowledge this is the first report of genomic screening for numerical sequence aberrations in a substantial number of keratoacanthomas by means of CGH. The major advantage of CGH is that the total genome of a lesion can be screened with respect to chromosomal losses and amplifications in one hybridization. Limitations are that balanced aberrations, such as translocations and inversions will not be detected, and that the hybridization target (the normal human metaphase) limits the resolution of the method. Deleted regions must therefore be in the magnitude of 5–10 Mb to be detected, whereas amplified regions down to 1–2 Mb may be detected if the gains are sufficiently large (Forozan et al., 1997Forozan F. Karhu R. Kononen J. Kallioniemi A. Kallioniemi O.-P. Genome screening by comparative genomic hybridization.Trends Genet. 1997; 13: 405-409Abstract Full Text PDF PubMed Scopus (238) Google Scholar). Thus, the finding of 35.7% cases with aberrations as assessed by CGH certainly indicates that keratoacanthomas have significant and even high levels of genomic imbalance, as many small lesions will remain undetected with this method. Admixture of inflammatory cells and other non-neoplastic cells will further decrease the sensitivity of the aberration detection (Clausen et al., 2001Clausen O.P.F. Andersen S.N. Strømkjaer H. Nielsen V. Rognum T.O. Bolund L. Kølvraa S. A strategy combining flow sorting and comparative genomic hybridization for studying genetic aberrations at different stages of colorectal carcinogenesis in ulcerative colitis.Cytometry. 2001; 43: 46-54Crossref PubMed Scopus (15) Google Scholar). The fact that the 10 benign hyperplastic/verrucous lesions did not show any detectable genetic aberrations suggests that the aberrations registered in keratoacanthomas are linked to their particular biology with rapid proliferation and subsequent regression. Seven of 10 SCC also showed aberrations by CGH, and some of the detected aberrations in keratoacanthomas and SCC overlap. Thus, one may conclude that keratoacanthoma is a benign lesion with a surprisingly high degree of genetic imbalance. There are few reports on genetic aberrations in solitary keratoacanthomas. Waring et al., 1996Waring A.J. Takata M. Rehman I. Rees J.L. Loss of heterozygosity analysis of keratoacanthoma reveals multiple differences from cutaneous squamous cell carcinoma.Br J Cancer. 1996; 73: 649-653Crossref PubMed Scopus (47) Google Scholar studied fractional allelic losses in 11 lesions judged to be definite keratoacanthomas. Loss of heterozygosity (LOH) was examined in loci that were known frequently to be lost in SCC and in actinic keratoses, and altogether 26 chromosome arms were analyzed with one marker for each arm. Isolated losses were found at 9q, 9p, and at 10q, with a fractional allelic loss of only 1.3% for each. These frequencies are considerably lower than those reported for SCC and actinic keratoses, being in the order of 30% and 45%, respectively (Quinn et al., 1994Quinn A.G. Sikkink S.A. Rees J.L. Basal cell carcinomas and squamous cell carcinomas show distinct patterns of chromosome loss.Cancer Res. 1994; 54: 4556-4559PubMed Google Scholar). Microsatellite instability and LOH were recently analyzed in two series of sporadic keratoacanthomas (Peris et al., 1997Peris K. Magrini F. Keller G. et al.Analysis of microsatellite instability and loss of heterozygosity in keratoacanthoma.Arch Dermatol Res. 1997; 289: 185-188Crossref PubMed Scopus (21) Google Scholar;Langenbach et al., 1999Langenbach N. Kroiss M.M. Rüschoff J. Schlegel J. Landthaler M. Stolz W. Assessment of microsatellite instability and loss of heterozygosity in sporadic keratoacanthomas.Arch Dermatol Res. 1999; 291: 1-5Crossref PubMed Scopus (24) Google Scholar). Both studies indicated that microsatellite instability and LOH do not play a significant part in the development of keratoacanthomas, unless they develop in the setting of Muir–Torre syndrome. Genetic aberrations measured by LOH depend on the selection and number of chromosomal regions that are chosen for the investigation. The relatively low levels of fractional allelic loss found in keratoacanthomas thus may reflect the markers chosen. These were often selected on the basis of what was frequently found in SCC, which is the most important differential diagnosis of keratoacanthoma. The CGH results presented here, with aberrations in 35.7% of the analyzed lesions, indicate that the frequency of fractional allelic imbalance in keratoacanthomas probably is much higher than hitherto reported. The frequency of keratoacanthomas with genetic aberrations in patients with and without immunosuppressive treatment were similar (Figure 1). There were, however, more than twice as many aberrations per keratoacanthoma with aberrations among immunocompetent individuals compared with immunosuppressed organ transplant recipients. This is consistent with the data from Rehmann et al., 1997Rehmann I. Quinn A.G. Takata M. Taylor A.E.M. Rees J.L. Low frequency of allelic loss in skin tumors from immunosuppressed individuals.Br J Cancer. 1997; 76: 757-759Crossref PubMed Scopus (14) Google Scholar, who reported a lower frequency of allelic loss in skin tumors of immunosuppressed individuals compared with immunocompetent ones. This may be related to the observation that neoplastic keratinocytic lesions occur more frequently in organ transplant recipients (Jensen et al., 1999Jensen P. Hansen S. Møller B. et al.Skin cancer in kidney and heart transplant recipients and different long-term immunosuppressive therapy regimens.J Am Acad Dermatol. 1999; 40: 177-186Abstract Full Text Full Text PDF PubMed Scopus (674) Google Scholar), and does not necessarily indicate differences in developmental pathways of lesions. Among young as well as old lesions a substantial fraction of keratoacanthomas showed genetic aberrations. This indicates that affected genes may be involved and responsible for early neoplastic development as well as for the spontaneous regression of the lesions. Amplification of 4p was significantly correlated with old lesions. As only three lesions had this aberration, however, this finding has to be confirmed. Cellular atypia was also significantly correlated with old lesions in this study. This is in contrast to what is generally assumed, namely that atypia is related to the early proliferative state (Schwartz, 1994Schwartz R.A. Keratoacanthoma.J Am Acad Dermatol. 1994; 30: 1-19Abstract Full Text PDF PubMed Scopus (250) Google Scholar). A separate analysis of keratoacanthomas from the immunocompetent individuals showed that no lesions with severe atypia were found in this group. An explanation for this, however, may be that the majority of these lesions were young. Although keratoacanthomas have clinical, architectural, and cytologic features that make the differential diagnosis from SCC possible in most cases, these two lesions have malignant features in common that may make a reliable distinction between them very difficult in some cases, especially when only a small biopsy is available. Furthermore, there are many reports on lesions classified as keratoacanthomas that have metastasized (Hodak et al., 1993Hodak E. Jones R.E. Ackerman A.B. Solitary keratoacanthoma is a squamous cell carcinoma: three examples with metastases.Am J Dermatopathol. 1993; 15: 332-342Crossref PubMed Scopus (198) Google Scholar). As there are no reliable histopathologic criteria to distinguish these types of lesions in atypical and difficult cases (Cribier et al., 1999Cribier B. Asch P.-H. Grosshans E. Differentiating squamous cell carcinoma from keratoacanthoma using histopathological criteria.Dermatology. 1999; 199: 208-212Crossref PubMed Scopus (110) Google Scholar), the question of whether keratoacanthoma is a variant of SCC has been raised. In this study we have included all lesions with characteristic clinical history that have the typical architectural appearance of keratoacanthomas, irrespective of the degree of cellular atypia and invasion, in order to correlate genotypic aberrations with all phenotypic features. We did not find statistical significant correlations between the presence of genetic aberrations and the degree of infiltration, or the degree of cellular atypia, respectively (Table II a,b). Based on this, it may be argued that lesions with malignant phenotype, combined with typical characteristics of keratoacanthomas, biologically belong to the same group of lesions. There was, however, a tendency for keratoacanthomas with severe atypia to have genetic aberrations (Table II a). This may be explained by an association between nuclear atypia and DNA aneuploidy, as aneuploid tumors have been shown to have more aberrations detected by CGH than diploid ones (De et al., 1999De Angelis P.M. Clausen O.P.F. Schjølberg A. Stokke T. Chromosomal gains and losses in primary colorectal carcinomas detected by CGH and their associations with tumor DNA aneuploidy, genotypes and phenotypes.Br J Cancer. 1999; 80: 526-535Crossref PubMed Scopus (135) Google Scholar). LOH at 3p, 9p, 9q, 13q, 17p, and 17q have been reported in actinic keratosis and cutaneous SCC (Quinn et al., 1994Quinn A.G. Sikkink S.A. Rees J.L. Basal cell carcinomas and squamous cell carcinomas show distinct patterns of chromosome loss.Cancer Res. 1994; 54: 4556-4559PubMed Google Scholar;Kushida et al., 1999Kushida Y. Miki H. Ohmori M. Loss of heterozygosity in actinic keratosis, squamous cell carcinoma and sun-exposed normal-appearing skin in Japanese: difference between Japanese and caucasians.Cancer Lett. 1999; 140: 169-175Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar;O'Connor et al., 2001O'Connor D.P. Kay E.W. Leader M. Murphy G.M. Atkins G.J. Mabruk M.J.E.M.F. A high degree of chromosomal instability at 13q 14 in cutaneous squamous cell carcinomas: indication for a role of a tumor suppressor gene other than Rb.J Clin Pathol Mol Pathol. 2001; 54: 165-169Crossref Scopus (22) Google Scholar). Recurrent deletions of regions at 3p and 9p were also seen in the present study by CGH, both representing 20.0% of lesions with aberrations. Two known tumor-suppressor genes located on 3p may be involved in the tumorigenesis of cervical carcinomas, namely transforming growth factor-β receptor II and fragile histidine triad (Herzog et al., 2001Herzog C.R. Crist K.A. Sabourin C.L.K. Kelloff G.J. Boone C.W. Stoner G.D. You M. Chromosome 3p tumor-suppressor gene alterations in cervical carcinomas.Mol Carcinog. 2001; 30: 159-168Crossref PubMed Scopus (15) Google Scholar). Chromosome arm 3p also harbors a common fragile site (FRA3B), which contains a spontaneous human papillomavirus 16 integration site (Wilke et al., 1996Wilke C.M. Hall B.K. Hoge A. Paradee W. Smith D.I. Glover T.W. FRA3B extends over a broad region and contains a spontaneous HPV 16 integration site: direct evidence for the coincidence of viral integration sites and fragile sites.Hum Mol Genet. 1996; 5: 187-195Crossref PubMed Scopus (220) Google Scholar). It is known that the CDKN2A (p16INK4A) gene coding for p16, which is a growth inhibitor involved in cell cycle traverse, is located on 9p21. Furthermore, the oncogene c-myc is located on 8q, which we have shown is amplified in keratoacanthomas. To what extent aberrations of these genes are involved in the development of keratoacanthomas remains to be determined. We have not found any reports in the literature on genetic aberrations in cutaneous SCC assessed by CGH. We have so far analyzed 10 SCC with CGH, seven of which showed genetic aberrations that partly overlap with those seen in the keratoacanthomas (i.e., amplifications on 1p, 1q, 8q, and chromosome 19, and deletion of 4q); however, those aberrations seen most frequently in keratoacanthomas in this study, i.e., losses at 3p, 9p, and 19p were not seen in any of the 10 SCC analyzed by us. Thus the present data may indicate a genetic basis to differentiate between keratoacanthomas and SCC. More cases of SCC need to be analyzed, however, and further studies with more detailed screening of genetic imbalance and of gene expression at the single gene level, e.g., with microarray methods, will be performed to answer this question. This work was supported by the Norwegian Cancer Society and Medinnova SF.