Tumor cell proliferation and microsatellite alterations in human ameloblastoma

Oral Oncology (2008) 44, 50–60

available at www.sciencedirect.com

journal homepage: http://intl.elsevierhealth.com/journals/oron/

Tumor cell proliferation and microsatellite alterations in human ameloblastoma M. Migaldi a, G. Sartori a, G. Rossi a, A. Cittadini b, A. Sgambato

b,*

a Dipartimento Misto di Anatomia Patologica e di Medicina Legale, Sezione di Anatomia Patologica, University of Modena and Reggio Emilia, Modena, Italy b ` Cattolica del Sacro Istituto di Patologia Generale, Centro di Ricerche Oncologiche ‘‘Giovanni XXIII’’, Universita Cuore, Largo Francesco Vito 1, 00168 Rome, Italy

Received 12 September 2006; accepted 6 December 2006 Available online 16 February 2007

KEYWORDS

Summary Ameloblastoma is the most common odontogenic tumor. It can exhibit a variety of histological patterns, a great infiltrative potential and a high recurrence rate. Mutations in microsatellite sequences are a hallmark of neoplastic transformation but little is known about their role in ameloblastoma development. In this study DNA was extracted from laser-microdissected samples of 24 ameloblastomas and was analyzed for the status of 22 microsatellite loci. The occurrence and the pattern of microsatellite alterations, in form of loss or length variation, was evaluated and correlated with the Ki67 labeling index and with other clinicopathologic parameters. The prognostic significance of these alterations was also evaluated. High Ki67 expression was significantly associated with a shorter disease-free survival (p = 0.003 by logrank test). Alterations of at least one of the selected loci was observed in all (100%) the ameloblastomas analyzed with a mean of 4 altered microsatellites for each tumor. The microsatellites most frequently altered were D9S747 and D11S488 (42%). All the other loci analyzed were altered in less than 40% of cases and some of them (D3S1312, D3S1300, IFNA, D9S164, D13S176 and TP53) did not show alterations in any of the ameloblastomas analyzed. No relationship was observed between the occurrence of microsatellite alterations and other parameters, such as patients age and gender, tumor size, localization and histotype. The occurrence of microsatellite alterations was more frequent in tumors displaying a high Ki67 labeling index (p = 0.03) and in a univariate analysis was predictor of an increased risk of disease recurrence (p = 0.039 by log-rank test). These findings demonstrate that microsatellite alterations are frequent event in ameloblastomas. They also suggest that evaluation of tumor cells proliferative activity and microsatellite alterations may be helpful to stratify ameloblastomas prognostically and to predict the clinical behavior of these tumors. c 2006 Elsevier Ltd. All rights reserved.

Ameloblastoma; Microsatellite instability; Allelic loss; Proliferation index; Ki67



* Corresponding author. Tel.: +39 06 3016619; fax: +39 06 3012753. E-mail address: [email protected] (A. Sgambato).



1368-8375/$ - see front matter c 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.oraloncology.2006.12.004

Tumor cells proliferation and microsatellite alterations in human ameloblastoma

Introduction Ameloblastoma is a rare neoplasm that accounts for about 1% of all oral tumors1 and arises from the epithelium of the dental lamina mainly in the mandible (80%) but also in the maxilla (20%). It usually affects young adults in the fourth–fifth decades of life, causing local discomfort.2 Ameloblastoma is generally benign, grows slowly and is not associated with symptoms until it becomes large. However, it is locally aggressive and displays a strong tendency to recur.3,4 Four distinct types of ameloblastoma are described from a clinical point of view: solid, unicystic, peripheral and desmoplastic.5 All are essentially formed by islands of epithelium resembling an organ epithelium, embedded in a mature, fibrous, connective tissue.6 The different histopathological patterns do not correlate with a different clinical behavior, especially in term of local recurrence, and at the present the major prognostic factor appears to be the surgical treatment (i.e., radical or conservative) that strongly affects the recurrence rate.7 However, despite the risk of recurrence is significantly reduced by a radical (range 18–27%) compared to a conservative (range 20–90%) surgical treatment, the available data suggest a high variability in the biological and clinical behavior of the tumors whose molecular determinants remain unknown.8 Microsatellites are short (1–10 base pairs) repetitive sequences that occur primarily in noncoding regions of DNA.9 The number of microsatellite repeat units located at a given locus is genetically determined and has been very useful as marker for genetic mapping.10 However, because of their repetitive structure, microsatellites are susceptible to slippage errors during DNA replication, resulting in insertion or deletion mutations. Loss of microsatellites has been used as a marker of loss of heterozygosity (LOH) and has been shown to involve several loci throughout the genome. The accumulation of mutations in microsatellite sequences throughout the genome is known as microsatellite instability (MSI). MSI is a form of genomic instability characterized by variations in the length of microsatellite repeats and is a hallmark of neoplastic transformation. Two different types of MSI have been described in human cancers. The ‘‘classical’’ MSI is typical of hereditary nonpolyposis colorectal cancers (HNPCC) and is associated with germline mutations in genes involved in DNA mismatch repair (MMR) mechanisms.11 More recently, a second variety of instability has been reported that appears to be not related to MMR deficiency and is best seen at tetranucleotide repeats.12 This form of instability has been termed EMAST (elevated microsatellite alterations at selected tetranucleotides) and appears to occur with high frequency in several human cancers.12,13 Limited data have been reported so far on the occurrence of MSI in ameloblastoma. In this study we have analyzed a large number of microsatellite loci to detect microsatellite alterations in a series of ameloblastomas and compared the results obtained with the clinical evolution of tumors and with the expression of a traditional marker of cell proliferation such as Ki67. We found that microsatellite alterations are a frequent event and, as well as tumor cells proliferative activity, might have

51

prognostic significance in ameloblastomas. The implications of these findings are discussed.

Materials and methods Patients selection All the cases of ameloblastomas recorded between September 1991 and April 2003, and for which detailed clinicopathologic data were available, were retrieved from the computerized files of the Department of Pathologic Anatomy of the University of Modena and Reggio Emilia. After excluding cases (n = 9) undergoing decalcification for the presence of bone, the remaining 24 cases were reviewed by two pathologists (MM and GR) by examination of several hematoxylin and eosin-stained tumor slides (mean, 2.5 slides for tumor; range, 2–8) and a consensus diagnosis was reached. They included biopsies or surgical specimens from 11 males and 13 females with age ranging between 15 and 87 years. The selection did not require approval by an Institutional Review Board because the samples were coded and the names of the patients were not revealed. Tumors were classified according to 2005 WHO classification14 and displayed different histological features. Detailed description of the cases under study is presented in Table 1.

Immunohistochemical analyses Immunohistochemical analyses were performed on routinely processed, formalin-fixed, paraffin-embedded tissues, employing an avidin–biotin complex immunoperoxidase technique (Vectastain ABC kit; Vector Laboratories, Burlingame, Ca) as previously described.15,16 Briefly, 5 lm sections were cut from representative blocks and subsequently dewaxed, rehydrated and then microwave pretreated (10 min at 750 W in 10 mM citrate buffer at pH 6.0). Sections were then incubated with 0.3% hydrogen peroxide in methanol for 30 min to block endogenous peroxidase. After blocking for 1 h at room temperature, a monoclonal anti-Ki67 antibody (clone MIB-1; from Dako, Glostrup, Denmark) was applied for 1 h at room temperature (22–24 C) diluted 1:200 in the blocking solution. Binding was visualized using the Vectastain diaminobenzidine kit (Vector Laboratories) and counterstaining was performed with 1% modified Meyer’s hematoxylin. Controls for specificity of staining were performed by immunostaining duplicate sections in the absence of the primary antibody. Sections of breast carcinomas known to be positive for Ki67 served as a positive control. Positive and negative control slides were included within each batch of slides. Immunostained nuclei were considered positive when they showed a distinct brown color in the absence of background staining. The fraction of positive-stained nuclei was scored by examining at least 10 random high-power field (400·) for each sample and the percentage of cells with positive reaction was calculated semiautomatically by means of a computer-assisted cellular image analyzer on a total of at least 1000 tumor nuclei per case. The labeling index (LI) was calculated as the number of positive cells · 100/total number of cells (positive + negative). All scoring and interpretations of the results were performed by 2 of the authors independently

52

M. Migaldi et al.

Table 1 Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 a b c d

Features of 24 ameloblastomas analyzed for tumor cells proliferation activity and microsatellite alterations Age (y)

Gendera

Siteb

Diameter (cm)

Histotype

Recurrencec

Surgeryd

16 58 22 39 29 58 15 70 22 87 82 29 66 69 17 21 75 35 18 61 69 39 75 21

F F F F F F M F M F M F M M M F M M M F M F M F

Mx-r Md-r Md-r Mx-l Md-r Md-r Md-r Md-l Md-l Md-l Mx-l Md-l Md-l Md-r Md-l Md-l Md-l Md-l Md-l Md-r Md-l Md-r Mx-l Md-r

1.7 1 4 2.7 2 5 0.6 1.5 3 0.7 1.5 0.7 1.5 1.2 1.2 1.5 2 4 3 2.8 2.5 2.5 1.8 1.6

Peripheral Solid Solid Solid Solid Solid Solid Peripheral Peripheral Solid Solid Peripheral Solid Solid Solid Solid Solid Solid Solid Peripheral Unicystic Solid Solid Solid

N Y Y N Y N N N N N Y N Y N Y N N N N N N N Y N

C R C C C R R R R R R R C R C R C R R R C R R R

F = female; M = male. Mx-r/l = maxilla-right/left; Md-r/l = mandible-right/left. Y (yes) and N (no) refers to the occurrence of recurrence during the period of follow-up. C = conservative; R = radical.

(MM and GR) without knowledge of clinical outcome or other clinicopathologic variables. To assess interobserver variation, the results of the two measurements were compared by paired t-test and no statistical differences were found (data not shown). The few cases with discrepant scoring were reevaluated jointly on a second occasion, and agreement was reached.

Microsatellite analysis Twenty-two microsatellite markers located on 9 chromosomes, including 11 chromosomal arms were selected (Table 2). Genomic DNA was prepared from 5 lm thick paraffin-embedded tissue sections as previously reported.16 In brief, histological sections were deparaffinized by a onehour incubation at 65 C followed by incubation in xylene. They were then rehydrated in a series of decreasing ethanol concentrations and stained with hematoxylin/eosin. By laser microdissection, approximately 0.5 cm2 of tumor and normal tissues (mainly lymphocytes) were prepared from each section, respectively. Microdissection was performed under direct observation with an inverted microscope using a laser capture microdissection apparatus (Olympus IX70, Laser Scissors Pro 300, Olympus Italia, Segrate, Milan) controlled by a joystickoperated electronic manipulator. The microdissected samples were transferred to an eppendorf tube containing digestion buffer (2 mg/ml proteinase K in 50 mM Tris (pH

8.5), 1 mM EDTA, 0.5% Tween 20). The tubes were then incubated for 24 h at 37 C, followed by 10 min at 95 C to eliminate any remaining proteinase K activity. PCR was performed in 10 ll reaction mixture containing 25 pmol each of primers, 1 ll of extracted DNA from at least 100 cells, 200 lM each of deoxynucleoside triphosphate, 2 mM magnesium chloride, 1 ll of commercial PCR buffer (Applied Biosystem, Applera Italy, Monza, Italy) and 1 unit of AmpliTaq Gold polymerase (Applied Biosystem). The PCR conditions were as follows: after an initial 2 min denaturation step at 95 C, 40 amplification cycles were performed, each consisting of denaturation (30 s at 95 C), annealing (30 s at 55–60 C, depending on the primers), and elongation (30 s at 72 C) steps followed by a final step at 72 C for 10 min. The amplified PCR products were then run on a 4% denaturing polyacrylamide gel for the ABI-Prism 310 automatic sequencer (Applied Biosystem) with a labelled marker (TAMRA 500) as an internal size standard. The run was performed in accordance with the supplier’s protocol. Twenty-two primer sequences (MWG-Biotech, Florence, Italy) flanking 22 microsatellite repeat polymorphisms located at 21 chromosomal regions of 9 different chromosomes were used, as follows: D1S2883, D3S3050, D3S1300, D3S1312, D3S2418, D4S243 (MH34), D4S243 (SHGC), D5S346, D9S54, D9S747, D9S162, IFNA, D9S171, D9S164, D11S488, D13S153, D13S176, D13S802, D16S310, D16S476, D17S654, and TP53 (Table 2). The data were analyzed with a Genescan 3.0 software (Applied Biosystem). Microsatellite

Tumor cells proliferation and microsatellite alterations in human ameloblastoma

53

Table 2 Microsatellite loci, primer sequences, type of repetitive sequence and corresponding genes analyzed in a series of ameloblastomas Locus (chromosomal arm)

Primer sequences

Type of repetition

D1S2883 (1p)

AAATCTGGTCTTCTGTTTTCACTAT TTCCAAATGTTGACTCTGC TGGTGGTATGCATTTGTCAG ATTCCCTGACTTCAAGTGCA AGCTCACATTCTAGTCAGCCT GCCAATTCCCCAGATG TGGGTTCTGCCTCCAA GGCTCCCCAGGGTAAG GTACATTCTGCACACGTACCC AGGACTGTAGGATTCCAAAGG TCAGTCTCTCTTTCTCCTTGCA TAGGAGCCTGTGGTCCTGTT AATCCCTTTTCTACCTTTCTATCAC GAGAGGAGAGATAAAAGATGTAAATG ACTCACTCTAGTGATAAATCGGG AGCAGATAAGACAGTATTACTAGTT GAAAGTCCAGAACTAAGTAG TGTGGATAGGTATATATAGC GCAATGACCAGTTAAGGTTC AATTCCCACAACAAATCTCC CTATGATCGTGCCACTGCAC GTTTGCCTGGGGATTGATTT AGCTAAGTGAACCTCATCTCTGTCT ACCCTAGCACTGATGGTATAGTCT GCCATTATTGACTCTGGAAAAGAC CAGGCTCTCAAAATATGAACAAAAT TGCGCGTTAAGTTAATTGGTT GTAAGGTGGAAACCCCCACT AGGCAATAGAGACCCTGTG GATGATGAATTGTTACTGAGAG AGCATTGTTTCATGTTGGTG CAGCAGTGAAGGTCTAAGCC CTGTGGGATTCCTTAGTGATAC ATATTCAGACAAAAGCCAAGTTA CACAGTGAGACTCTATCTCAAAAA TCAGACTGGCTTAGACTGTGG GGGCAACAAGGAGAGACTCT AAAAAAGGACCTGCCTTTATCC TTGCACTCCACTCTGGGCA TTGCCTTGGCTTTCTGTTGG GACCTAGGCCATGTTCACAGCC GACATCCATTGGCACCACCCCAA AGGGATACTATTCAGCCCGAGGTG ACTGCCACTCCTTGCCCCATT

Dinucleotide

D3S3050 (3p) D3S1300 (3p) D3S1312 (3p) D3S2418 (3q) D4S243 (MH34) (4p) D4S243 (SHGC) (4p) D5S346 (5q) D9S54 (9p) D9S162 (9p) D9S164 (9q) D9S171 (9p) D9S747 (9q) IFNA (9p) D11S488 (11q) D13S153 (13q) D13S176 (13q) D13S802 (13q) D16S310 (16q) D16S476 (16q) D17S654 (17p) TP53 (17p)

analysis was performed by comparing the positions of the bands corresponding to the tumor and the normal DNAs according to the manufacturer’s manual. Briefly, peak height of each microsatellite locus for independent injections of each normal and each tumor sample was obtained. Length variation was determined when the electropherogram showed the presence of novel peaks, with an evident shift, in the tumor DNA, that were not present in normal DNA (Fig. 1A). Loss of heterozygosity (LOH) was also detected and was defined as the loss of a wild type allele (peak) in tumor compared with normal DNA (Fig. 1B). The

Gene involved

Tetranucleotide Dinucleotide

FHIT

Dinucleotide

FHIT

Trinucleotide Tetranucleotide Tetranucleotide Dinucleotide Dinucleotide Dinucleotide

p16

Dinucleotide

P16

Dinucleotide

p16

Tetranucleotide Dinucleotide Tetranucleotide Dinucleotide

pRb

Dinucleotide

pRb

Tetranucleotide Tetranucleotide Pentanucleotide Dinucleotide Dinucleotide

p53

mathematic model of LOH determination used is the following: height of normal allele two=height of normal allele one height of tumor allele two=height of tumor allele one An LOH value 6 0.5 indicated that the tumor sample showed significant loss of the longer allele whereas an LOH value P 1.5 indicated a significant loss of the shorter allele. Results were confirmed in all cases in duplicate experiments using independently extracted DNA samples.

54

M. Migaldi et al.

Figure 1 Electropherograms of microsatellite analysis using the automated single-capillary ABI PRISM sequencer model 310 Genetic Analyzer. Shown are the electropherograms of normal and tumor DNA samples. (a) An alteration of peaks is observed in tumor DNA. (b) Loss an allele is observed in tumor compared to normal DNA.

Statistical analysis The association between molecular and clinicopathological parameters were calculated using contingency table methods and tested for significance using the Fisher’s exact test. Bonferroni correction was used to correct for multiple comparisons. Survival curves were calculated using the Kaplan– Meier method and the log-rank test was used for the analysis. Time to recurrence was calculated from operation to the first documented recurrence. In the absence of recurrence, follow up time was calculated from operation to

the date of the last documented follow up (January 2005). All calculations were performed by the SPSS rel. 8.00 (Statistical Package for Social Science) and the results were considered statistically significant when the p value was 60.05.

Results Twenty-four patients with ameloblastoma were included in this study. They included 11 (45.8%) men and 13 (54.2%) women with mean age at diagnosis of 46 years (range 15–87

Tumor cells proliferation and microsatellite alterations in human ameloblastoma years) which was slightly lower in women (mean 43; range 16–87) than in men (men 49; range 15–82) but the difference was not significant. The mandible was affected in 20 (83.3%) and the maxilla in 4 (16.7%) cases with a maxilla to mandible ratio of 1–5. The left mandible was the most frequent localization (n = 11; 45.8%), followed by the right one (9; 37.5%), the left (3; 12.5%) and the right (1; 4.2%) maxilla. The left mandible was most frequently affected in men (7/11; 63.6%) whereas the right one was more frequently affected in women (7/13; 54%). From a histological point of view, the series included 5 (20.8%) peripheral, 18 (75%) solid and one (4.2%) unicystic (Table 1) types. The solid type was more frequent in men (82%) than in females (69%), whereas the peripheral type was more frequent in females (31%) than in men (9%), but these differences were not significant. The only unicystic case was observed in a male patient (Table 1). Primary tumor size was measured to be on average 2.1 cm. Median size was 1.75 cm, with a range between 0.6 and 5.0 cm. The unicystic case displayed the largest size with a 2.5 cm diameter followed by solid and peripheral cases with a mean size of 2.1 and 2.0 cm, respectively. The mean follow-up was 57 months with a range of 2– 146 months. All of the patients were alive at the date of the last documented follow up. Sixteen patients underwent radical surgery (67%) while the remaining ones underwent a conservative surgery (33%). No relationship was observed between the type of surgery and other variables, such as gender, histotype, localization and recurrence (data not shown). Seven (29%) patients recurred during the period of follow-up, including four men (36%) and three (23%) women. Recurrent cases were all solid type tumors (Table 1) with a mean disease-free survival of 16 months (range 2-31). None of the peripheral cases nor the unicystic one displayed a recurrence during the period of follow-up (Table 1). Six of the seven recurrences affected the mandible (3 left and 3 right). The size of recurrences was smaller than primary tumors with a mean of 1.9 cm (range 1.0–4.0). No patient developed metastases, thus excluding malignancy for all of the tumors under study.

Proliferative activity The expression of Ki67, a traditional marker of cell proliferation, was evaluated by immunostaining in the series of ameloblastomas analyzed in this study. Only cells with a clear nuclear staining were regarded as positive. In the neoplastic tissue, overall staining was very low, mirroring the slow growth rates of the tumors, and was frequently heterogeneous in different areas with the percentage of reactive cells ranging from 0.4% to 5.8% and a median value of 1.05 (mean = 1.6%; SE = 1.3) (Fig. 2). When tumors were stratified according with histological type, mean percentage of positive cells was 1.7 (range 0.4–4.2), 2.2 (range 0.5–5.8) and 1.9 in peripheral, solid and unicystic cases, respectively, but these differences were not statistically significant. Ki67 labeling index (LI) was slightly higher in tumors treated with conservative (mean 1.8; range 0.5–4.2) compared to cases treated radically (mean 1.5; range 0.4–5.8) but this

55

Figure 2 Representative examples of Ki67 immunostaining in two cases of solid (A) and peripheral (B) ameloblastomas (original magnifications: 100·).

difference was not significant. When LI for Ki67 was correlated with the clinical behavior, mean percentage of positive cells was 3.2 (range 2.0–5.8) in the seven recurrent cases and 0.9 (range 0.4–2.3) in the nonrecurrent cases, and this difference was statistically significant (p < 0.001). We then used the median value to stratify tumors in cases with high (>1.0% positive cells) and low (=1.0%) Ki67 expression. No relationship was observed between Ki67 expression and tumor localization and histotype. However, it is noteworthy that 7 of the 12 (58%) high expressor tumors recurred during the period of follow-up whereas no recurrences occurred amongst the 12 low expressor cases and this difference was highly significant (p = 0.002). In an univariate analysis high Ki67 expression was significantly associated with a shorter disease-free survival in our series of patients (p = 0.003 by log-rank test) (Fig. 3A).

Microsatellite analysis LOH at one or more loci was detected in 11 (46%) tumors analyzed with a mean of 1 deleted microsatellites for case (range 0–5; median 0). It is noteworthy that LOH can be evaluated by the method used only when normal DNA shows heterogeneity of the length of tandem repeats and in this

56

M. Migaldi et al.

Figure 3 Kaplan–Meier curves for disease free survival in a series of ameloblastomas. Patients were stratified according to Ki67 labelling index (a) and to the frequency of microsatellite alterations (b) (see text for details).

study the mean number of noninformative cases was 4 (range 2–6; median = 3) for each locus. Length variation was observed in 23 (96%) cases with a mean of 2.7 altered microsatellites for each case (range 0–6; median 2). Overall, alterations of at least one of the selected loci, in the form of deletion or alteration of peak(s), was observed in all (100%) of the 24 ameloblastomas analyzed in this study with a mean of 4 (18%) altered microsatellites of the 22 examined for each tumor (range 1–9; median 4). No tumors displayed microsatellite loci showing both MSI and LOH. Based on these results, we used the median values to classify tumors as displaying a high or a low frequency of microsatellite alterations, as previously suggested.17 Thus, tumors were classified as displaying high frequency of microsatellite alterations if they showed more than two shifts and/or at least one LOH in the loci analyzed. Seventeen tumors (71%) fell in this category (13 tumors with more than 2 shifts, 11 displaying at least one LOH and 5 displaying both types of alterations) with the remaining seven (29%) being classified as displaying low frequency of microsatellite alterations. No relationship was observed between the occurrence of microsatellite alterations and other parameters, such as patients age and gender, tumor size, localization and histotype. On the other hand, high expression of Ki67 was detected in 11 (65%) cases displaying a high frequency and only in one (14%) with low frequency of microsatellite alterations and this difference was significant (p = 0.03). Seven (41%) of the 17 tumors with high frequency and none (0%) of the 7 with low frequency of microsatellite alterations recurred during the period of follow up but this difference was not significant (p = 0.056). However, in an univariate analysis high frequency of microsatellite alterations was significantly associated with a shorter disease-free

survival in our series of patients (p = 0.039 by log-rank test) (Fig. 3B). No other significant relationships were observed considering the occurrence of length variations or LOH separately.

Microsatellite alteration at each locus The microsatellites most frequently altered in our series of ameloblastoma were D9S747 and D11S488 (42%). All the other loci analyzed were altered in less than 40% of cases (Table 3). When considered separately, MSI was most frequently detected for the loci D11S488 (29%), D17S654 (29%), D3S2418 (25%) and D9S747 (25%), whereas microsatellites most frequently deleted were D4S243-MH34, D9S747, D9S171 e D13S153 (each of them deleted in 17% of cases). It is noteworthy that some of the loci (D3S1312, D3S1300, IFNA, D9S164, D13S176 and TP53) did not show alterations, in the form of deletion or alteration of peaks, in any of the ameloblastomas analyzed (Table 3). No relationship was observed between alterations in any of the loci analyzed and patients age, gender, tumor localization and histotype (data not shown). We also could not detect a significant relationship between alterations at a specific microsatellite and tumor recurrence for any of the loci analyzed (data not shown). However the Kaplan–Meier curves of disease-free survival showed that the occurrence of microsatellite alterations (in term of MSI and/or LOH) at the D3S2418 (p = 0.03 by log-rank test), D9S162 (p = 0.04) and D4S243-MH34 (p = 0.04) loci was associated with an increased risk of recurrence in a univariate analysis in our series of ameloblastomas (Fig. 4). A significant separation was also observed for the Kaplan–Meier curves of disease free survival when patients were stratified according

Tumor cells proliferation and microsatellite alterations in human ameloblastoma Table 3

57

MIB-1 labelling index (LI) according to the status of 22 genetic loci in a series of 24 ameloblastomas

Locus

MSI n (%) (LI)

LOH n (%) (LI)

Negative n (%) (LI)

D1S2883 D3S3050 D3S1300 D3S1312 D3S2418* D4S243 (MH34)** D4S243 (SHGC) D5S346 D9S54*** D9S162 D9S164 D9S171 D9S747 IFNA D11S488 D13S176 D13S153 D13S802 D16S310 D16S476 D17S654 TP53

5 5 0 0 6 5 1 1 5 5 0 5 6 0 7 0 4 1 1 2 7 0

1 2 0 0 3 4 0 0 2 2 0 4 4 0 3 0 4 0 0 1 2 0

18 17 24 24 15 15 23 23 17 17 24 15 14 24 14 24 16 23 23 21 15 24

(21) (1.9) (21) (1.3) (0) (0) (0) (0) (25) (2.7) (21) (2.6) (4) (1.2) (4) (0.6) (21) (3.0) (21) (1.9) (0) (0) (21) (1.7) (25) (1.2) (0) (0) (29) (1.7) (0) (0) (17) (2.1) (4) (0.5) (4) (1.2) (8) (2.4) (29) (1.8) (0) (0)

(4) (3.8) (8) (1.6) (0) (0) (0) (0) (12) (2.1) (17) (2.2) (0) (0) (0) (0) (8) (2.4) (8) (2.9) (0) (0) (17) (1.8) (17) (1.4) (0) (0) (12) (1.6) (0) (0) (17) (2.3) (0) (0) (0) (10) (4) (1.2) (8) (3.1) (0) (0)

(75) (1.4) (71) (1.7) (100) (0) (100) (0) (62) (1.0) (62) (1.1) (96) (1.6) (96) (1.6) (71) (1.1) (71) (1.3) (100) (0) (62) (1.5) (58) (1.8) (100) (0) (58) (1.5) (100) (0) (67) (1.4) (96) (1.6) (96) (1.6) (87) (1.5) (62) (1.3) (100) (0)

MSI: indicates the occurrence of shifts in the DNA peak. LOH: indicates the deletion of the corresponding locus. * p < 0.017; ** p < 0.046; *** p < 0.01 considering the LI in cases with MSI in the specific locus compared to tumors in which the same locus was not altered (see text for details).

with alterations in the DS954 locus, but the difference was not significant (p = 0.05 by log-rank test) (data not shown). We then analyzed whether a correlation exist between microsatellite alterations and the Ki67 LI. As shown in Table 3 the percentage of Ki67 positive cells was significantly higher in tumors displaying MSI at the loci D3S2418 (p < 0.017), D4S243-MH34 (p < 0.046) and D9S54 (p < 0.01) compared with tumors in which the same loci were not altered. Similarly, a higher Ki67 LI was observed in tumors displaying LOH at the loci D1S2883, D9S162, and D17S654 but the differences did not reach significance (data not shown).

Discussion Ameloblastoma is the most common epithelial odontogenic tumor and, while some authors contend that it is a lowgrade malignancy, the general consensus is that this tumor is distinctly benign.18 However, this designation is potentially misleading since it imparts a simultaneous designation of ‘‘nonaggressive’’ whereas these tumors often display a locally aggressive, infiltrating phenotype and are characterized by a high percentage of recurrences.8 The type of treatment (i.e., conservative vs radical surgery) is an important prognostic factor with conservative treatment being associated with a higher percentage of recurrences, up to 93%19 compared to radical treatment. Our study may have lacked the power to detect a significant correlation between type of treatment and risk of recurrence because of the relatively small number of cases (Table 1). However, despite the literature clearly supports the superiority of

radical resection compared with ‘‘conservative’’ therapy in permanently eliminating the ameloblastoma, it is also clear that not all patients who undergo such aggressive surgical therapy are cured and the classical prognostic parameters (such as tumor localisation, size, morphology and histology) are inadequate to identify high risk patients. Proliferative activity of tumor cells has been reported as an important prognostic marker and indicator of aggressiveness. Several studies have investigated cellular proliferative activity in ameloblastomas but limited data are available about its relationship with the clinical behavior of the disease.20–26 In this study, we evaluated tumor cells proliferative activity in a series of ameloblastomas as assessed by Ki67 immunostaining. Ki67 is a highly specific marker of proliferating cells, maximally expressed during S phase and rapidly degraded after mitosis. Half-life of Ki67 is 60–90 min and it has been suggested that Ki67 staining is more accurate than the counting of mitoses or proliferating cell nuclear antigen (PCNA) staining.27 We found that ameloblastomas displayed a generally low, although variable, Ki67 labeling index with the percentage of tumor positive cells ranging from 0.4% to 5.8% with a median value of 1.05. These results are in agreement with previous data in the literature which reported similar results obtained evaluating the expression of Ki67 in ameloblastomas.22–25 We did not observe any statistically significant difference in the Ki67 LI among the histological subtypes of ameloblastomas analyzed in this study. These findings are in agreement with previous studies in which no differences were found in the proliferative activity among different histological types of solid ameloblas-

58

M. Migaldi et al.

Figure 4 Kaplan–Meier curves for disease free survival in a series of ameloblastomas. Patients were stratified according to the presence of microsatellite alterations (both MSI and/or LOH) at the indicated loci (see text for details).

toma.23,26,28,29 On the other hand, it is noteworthy that a statistically significant difference was observed between recurrent (mean LI = 3.2%) and nonrecurrent (mean LI = 0.9%) cases (p < 0.001). Moreover, when the median value was used to identify high and low proliferating tumors, a statistically significant association was observed between high Ki67 expression and recurrences (p = 0.002). In an univariate analysis high Ki67 expression was significantly associated with a shorter disease-free survival in our series of patients (p = 0.003 by log-rank test) (Fig. 3A). Two types of microsatellite alterations have been reported in human tumors: deletion, which causes loss of heterozygosity (LOH) at the corresponding loci and modification, also known as microsatellite instability (MSI). MSI was initially demonstrated in a subset of hereditary nonpolyposis colorectal cancers (HNPCC) and was associated with germline mutations in genes involved in DNA mismatch repair (MMR) mechanisms.11 Although MSI is primarily a feature of HNPCC-related tumors, it also occurs in a subset of most sporadic human cancers.30 Conflicting results have been reported on the frequency of MSI in many

tumors and these discrepancies can be attributed to different factors, including the source of tumor DNAs used for the analysis and the numbers and types of microsatellites included in each study. Thus, to standardize detection of MSI a panel of repeat loci (the Bethesda consensus panel) was established to define MSI.17 The panel only included mono- and dinucleotides repeats which are mainly affected by MMR deficiency. A second variety of instability has been more recently reported that appears to be not related to MMR deficiency and is best seen at tetranucleotide repeats.12 This form of instability has been termed EMAST (elevated microsatellite alterations at selected tetranucleotides) and appears to occur with high frequency in a variety of human tumors.12 No data are available on the type and the frequency of microsatellite alterations in ameloblastomas. However, the choice of informative microsatellite markers is imperative in this type of studies and we preferred to not use the Bethesda consensus panel of microsatellites since ameloblastoma is not found at a higher incidence in HNPCC kindred30 and mutations in the human DNA mismatch repair system do not appear to be involved

Tumor cells proliferation and microsatellite alterations in human ameloblastoma in ameloblastomas.31 Thus, we selected a series of loci which have been previously reported to be altered in human tumorigenesis and choose to not include mononucleotides but only dinucleotide and larger repeat sequences, including tetranucleotides. Moreover, to override problems due to the source of tumor DNA, we used laser microdissection that allows the preparation of high quality microdissected tissues as source of DNA suitable for subsequent molecular analysis. This is very important since the contamination due to normal cells is a potentially relevant confusing factor and can compromise the ability of detecting microsatellite alterations using PCR-based assays. We also choose to analyze a large (n = 22) number of microsatellites in order to obtain information as much as possible informative on the occurrence of microsatellite alterations in ameloblastomas. We found that microsatellite alterations, in the form of deletion (LOH) or modification (MSI) of repetitive sequences, was a frequent event in ameloblastomas and was detectable, at one or more loci, in all (100%) of the tumors analyzed (Table 3). An interesting result of this study was that the occurrence of microsatellite alterations was more frequent in tumors displaying a high Ki67 LI (p = 0.03) and in a univariate analysis was predictor of an increased risk of disease recurrence (Fig. 3B). Moreover, an interesting result of this study was that the occurrence of alterations in specific loci was predictor of an increased risk of recurrence (Fig. 4). The observation that those loci were the same whose alterations were associated with a high Ki67 LI, further supports the relationship between tumor cell proliferation and risk of recurrence in these tumors. The occurrence of microsatellite alteration is expression of a loss of DNA repair efficiency. It is possible that this deficiency contribute to tumorigenesis by increasing the mutation rate of genes directly involved in the process or that, on the other hand, it is simply an epiphenomenon of genetic and molecular alterations occurring in tumor cells. However, the finding that some of the loci were not altered in any of the tumors analyzed suggests that alterations at specific loci preferentially occur and/or are selected during ameloblastoma development. We observed a significant association between the occurrence of microsatellite alterations, especially at specific loci, and a high tumor cell proliferation activity, as assessed by the proliferation marker Ki67, in our series of tumors (Table 3). The significance of this finding remains to be defined: it might suggest that an increased tumor cell proliferation might favor microsatellite alterations or, on the other hand, that the occurrence of microsatellite alterations could determine a deregulation of cell growth. Regardless of the underlying molecular mechanisms, determining whether particular cancers show microsatellite alterations may have important biological and clinical implications. In fact, their identification can help to understand tumor pathogenesis and might have prognostic importance, as suggested by the results of the present study (Fig. 3B and 4). Few studies have investigated genetic alterations in ameloblastomas. Stenman et al. analyzed cultured benign odontogenic tumors and reported a rare occurrence of chromosomal alterations in ameloblastoma cells.32 More recently, evaluation of chromosomal alterations by compara-

59

tive genomic hybridization was able to identify alterations in a limited number (about 11%) of ameloblastomas.24,33 In the most recent study, fluorescence in situ hybridization was also performed and revealed several aberrations of chromosome 22, especially loss at 22q, even in cases with a normal CGH profile, thus suggesting that they might be a common event in ameloblastomas.33 Our study, which was initiated before the publication of this paper, did not include microsatellites localized on chromosome 22. However, taken together with the data in the literature, our findings definitively demonstrate that microsatellite alterations, in the form of instability or deletion, is a frequent event in ameloblastomas suggesting a possible role in the process of tumor development. Moreover, it might have a prognostic significance being a predictor of high risk of recurrence, if the results of the present study will be confirmed on a large cohort of patients. Our findings are in agreement with the only available paper investigating allelic losses in ameloblastomas.34 This study analyzed 5 tumor suppressor genes in a series of 12 ameloblastomas and found a high frequency (average 46%) of allelic loss.34 Although a direct comparison is not possible since different loci were analyzed in the two studies, it is noteworthy that our findings diverge from the study by Nodit et al.34 since it reported a high frequency of allelic losses (between 33% and 50%) at the p53 locus whereas the TP53 locus, corresponding to the p53 tumor suppressor gene, and the loci D3S1312 and D3S1300, corresponding to the FHIT gene, were not altered in any of the tumors analyzed in the present study. The reason for this discrepancy is not clear. However, it is possible that deletions involve small regions within the p53 locus and thus they might be not always evident, depending on the pair of primers utilized for the analysis. On the other hand, our data are in agreement with previous studies reporting a very low frequency or absence of p53 gene alterations in ameloblastomas.35,36 Further studies will be needed to clarify these discrepancies. In conclusion, we demonstrated that microsatellite alterations occur at a very high frequency in ameloblastomas suggesting that they might play a role in the process of tumor development. Moreover, our findings also demonstrate that proliferative activity of tumor cells (as assessed by Ki67 labeling index) and microsatellite alterations may be important factors in predicting, if not in determining, the likelihood of recurrence or aggressive behavior of these tumors. These results are of great importance considering that the traditional pathologic variables (including histology) are unable to stratify ameloblastomas prognostically and cannot explain differences in the clinical behavior of these tumors.

Conflict of interest statement None declared.

References 1. Melrose R. Benign epithelial odontogenic tumors. Semin Diagn Pathol 2002;16:271–87. 2. Philipsen H, Reichart P, Nikai H, Takata T, Kudo Y. Peripheral ameloblastoma: biological profile based on 160 cases from the literature. Oral Oncol 2001;37:17–27.

60 3. Regezi J. Odontogenic cysts, odontogenic tumors, fibroosseous, and giant cell lesions of the jaws. Mod Pathol 2002;15:331–41. 4. Pinheiro J, Freitas V, Moretti A, Jorge A, Jaeger R. Local invasiveness of ameloblastoma. Role played by matrix metalloproteinases and proliferative activity. Histopathology 2004;45:65–72. 5. Modolo F, Martins M, Loducca S, de Araujo V. Expression of integrin subunits alpha2, alpha3, alpha5, alphav, beta1, beta3 and beta4 in different histological types of ameloblastoma compared with dental germ, dental lamina and adult lining epithelium. Oral Dis 2004;10:277–82. 6. Waldron C. Odontogenic cysts and tumours. In: Neville B, Dam C, Allen C, Bouquot J, editors. Oral and maxillofacial pathology. Philadelphia: WB Saunders Company; 2002. p. 589–642. 7. Ghandhi D, Ayoub A, Pogrel M, Macdonald G, Brocklebank L, Moos K. Ameloblastoma: a surgeon’s dilemma. J Oral Maxillofac Surg 2006;64:1010–4. 8. Muller H, Slootweg P. The ameloblastoma, the controversial approach to therapy. J Maxillofac Surg 1985;13:79–84. 9. Hearn CM, Ghosh S, Todd JA. Microsatellites for linkage analysis of genetic traits. Trends Genet 1992;8:288–94. 10. Weissenbach J, Giyapay G, Dib C, Vaginal A, Morisette J, Millaseau P, et al. A second generation linkage map of the human genome. Nature (London) 1992;359:794–801. 11. Loeb LA. Microsatellite instability: marker of a mutator phenotype in cancer. Cancer Res 1994;54:5059–63. 12. Mao L, Lee DJ, Tockman MS, Erozan YS, Askin F, Sidransky D. Microsatellite alterations as clonal markers for the detection of human cancer. Proc Natl Acad Sci USA 1994;91:9871–5. 13. Danaee H, Nelson H, Karagas M, Schned A, Ashok T, Hirao T, et al. Microsatellite instability at tetranucleotide repeats in skin and bladder cancer. Oncogene 2002;21:4894–9. 14. Barnes L, Eveson JW, Reichart P, Sidrasky D, editors. Head and neck tumors – pathology and genetics, WHO class of tumours. Lyon: IARC Press; 2005. 15. Sgambato A, Migaldi M, Faraglia B, De Aloysio G, Ferrari P, Ardito R, et al. Cyclin D1 expression in papillary superficial bladder cancer: its association with other cell cycle-associated proteins, cell proliferation and clinical outcome. Int J Cancer 2002;97:671–8. 16. Migaldi M, Sartori G, Rossi G, Garagnani L, faraglia B, De Gaetani C, et al. Prevalence and prognostic significance of microsatellite alterations in young patients with bladder cancer. Mod Pathol 2005;18:1176–86. 17. Boland C, Thibodeau S, Hamilton S, Sidransky D, Eshleman J, Burt R, et al. A National Cancer Institute Workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 1998;58:5248–57. 18. Gold L. Biologic behavior of ameloblastoma. Oral Maxillofac Surg Clin North Am 1991;3:21. 19. Sehdev M, Huvos A, Strong E, Gerold F, Willis G. Proceedings: ameloblastoma of maxilla and mandible. Cancer 1974;33: 324–33.

M. Migaldi et al. 20. Li T, Browne R, Matthews J. Expression of proliferating cell nuclear antigen (PCNA) and Ki-67 in unicystic ameloblastoma. Histopathology 1995;26:219–28. 21. Funaoka K, Arisue M, Kobayashi I, Iizuka T, Kohgo T, Amemiya A, et al. Immunohistochemical detection of proliferating cell nuclear antigen (PCNA) in 23 cases of ameloblastoma. Eur J Cancer B Oral Oncol 1996;32B:328–32. 22. Ong’uti M, Cruchley A, Howells G, Williams D. Ki-67 antigen in ameloblastomas: correlation with clinical and histological parameters in 54 cases from Kenya. Int J Oral Maxillofacial Surg 1997;26:376–9. 23. Sandra F, Mitsuyasu T, Nakamura N, Shiratsuchi Y, Ohishi M. Immunohistochemical evaluation of PCNA and Ki-67 in ameloblastoma. Oral Oncol 2001;37:193–8. 24. Ja ¨¨ askela ¨inen K, KJ J, Leivo I, Saloniemi I, Knuutila S, Heikinheimo K. Cell proliferation and chromosomal changes in human ameloblastoma. Cancer Genet Cytogenet 2002;136: 31–7. 25. Meer S, Galpin J, Altini M, Coleman H, Ali H. Proliferating cell nuclear antigen and Ki67 immunoreactivity in ameloblastomas. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2003;95: 213–21. 26. Barboza C, Pereira Pinto L, Freitas RA, Costa AL, Souza L. Proliferating cell nuclear antigen (PCNA) and p53 protein expression in ameloblastoma and adenomatoid odontogenic tumor. Braz Dent J 2005;16:56–61. 27. Brown D, Gatter K. Ki67 protein: the immaculate deception? Histopathology 2002;40:2–11. 28. Kim J, Yook J. Immunohistochemical study on proliferating cell nuclear antigen expression in ameloblastomas. Eur J Cancer Oral Oncol 1994;30B:126–31. 29. Carvalhais J, Aguiar M, Arau ´jo V, Arau ´jo N, Gomez R. p53 and MDM2 expression in odontogenic cysts and tumours. Oral Dis 1999;5:218–22. 30. Arzimanoglou I, Gilber F, Barber H. Microsatellity instability in human solid tumors. Cancer 1998;82:1802–20. 31. Castrilli G, Piantelli M, Artese L, Perfetti G, Rubini C, Fioroni M, et al. Expression of hMSH2 and hMLH1 proteins of the human DNA mismatch repair system in ameloblastoma. J Oral Pathol Med 2001;30:305–8. 32. Stenman G, Sandros J, Mark J, Happonen R. Observations by Gbanding in benign odontogenic tumours. Cancer Genet Cytogenet 1986;19:253–9. 33. Toida M, Balazs M, Treszl A, Rakosy Z, Kato K, Yamazaki Y, et al. Analysis of ameloblastomas by comparative genomic hybridization and fluorescence in situ hybridization. Cancer Genet Cytogenet. 2005;159:99–104. 34. Nodit L, Barnes L, Childers E, Finkelstein S, Swalsky P, Hunt J. Allelic loss of tumor suppressor genes in ameloblastic tumors. Mod Pathol 2004;17:1062–7. 35. Shibata T, Nakata D, Chiba I, Yamashita T, Abiko Y, Tada M, et al. Detection of TP53 mutation in ameloblastoma by the use of a yeast functional assay. J Oral Pathol Med 2002;31:534–8. 36. Kumamoto H, Izutsu T, Ohki K, Takahashi N, Ooya K. p53 gene status and expression of p53, MDM2, and p14 proteins in ameloblastomas. J Oral Pathol Med 2004;33:292–9.