Elias Campo, Laboratory of Pathology, Hospital Clínic, Villarroel 170, 08036 Barcelona, Spain. E-mail: email@example.com
Anaplastic large cell lymphoma (ALCL) is a T/null-cell neoplasm characterized by chromosomal translocations involving the anaplastic lymphoma kinase (ALK) gene (ALK). Tumours with similar morphology and phenotype but negative for ALK have been also recognized. The secondary chromosomal imbalances of these lymphomas are not well known. We have examined 74 ALCL, 43 ALK-positive and 31 ALK-negative, cases by comparative genomic hybridization (CGH), and locus-specific alterations for TP53 and ATM were examined by fluorescence in situ hybridization and real-time quantitative polymerase chain reaction. Chromosomal imbalances were detected in 25 (58%) ALK-positive and 20 (65%) ALK-negative ALCL. ALK-positive ALCL with NPM-ALK or other ALK variant translocations showed a similar profile of secondary genetic alterations. Gains of 17p and 17q24-qter and losses of 4q13-q21, and 11q14 were associated with ALK-positive cases (P = 0·05), whereas gains of 1q and 6p21 were more frequent in ALK-negative tumours (P = 0·03). Gains of chromosome 7 and 6q and 13q losses were seen in both types of tumours. ALCL-negative tumours had a significantly worse prognosis than ALK-positive. However no specific chromosomal alterations were associated with survival. In conclusion, ALK-positive and negative ALCL have different secondary genomic aberrations, suggesting they correspond to different genetic entities.
Anaplastic large cell lymphoma (ALCL) is a T/null-cell neoplasm genetically characterized by recurrent chromosomal translocations involving the anaplastic lymphoma kinase (ALK) gene (ALK) at 2p23, resulting in the constant expression of the ALK protein. These tumours constitute a well-defined entity presenting predominantly in young patients with a relatively favourable prognosis, although some patients will die of progressive disease. Morphologically, the tumours show a broad cytological spectrum from common type with large atypical cells to a small cell variant but constantly express the cytokine receptor CD30 (Shiota et al, 1995; Benharroch et al, 1998; Falini et al, 1999). The most common translocation involving ALK is the t(2;5)(p23;q35) which fuses ALK to the nucleophosmin gene (NPM) (Morris et al, 1994). In about 20% of cases alternative translocations have been described fusing ALK to various partners, such as TPM3 (Lamant et al, 1999), TFG (Hernandez et al, 1999, 2002; Siebert et al, 1999), ATIC (Wlodarska et al, 1998; Colleoni et al, 2000; Ma et al, 2000; Trinei et al, 2000), TSPYL2 (Touriol et al, 2000), MSN (Tort et al, 2001, 2004), KIAA1618 (Cools et al, 2002) and MYH9 (Lamant et al, 2003). These translocations have transforming potential but seem to induce, in vitro, certain differences in cell growth, migration and invasiveness of the transformed cells that may influence the biological behaviour of the tumours (Armstrong et al, 2004). Despite the well-defined primary chromosomal translocations of these lymphomas, the profile of secondary chromosomal aberrations in ALK-positive ALCL is not well known.
In addition to ALK-positive ALCL, several studies have recognized a large T-cell lymphoma with similar morphological and phenotypic features but lacking the genetic alterations and expression of ALK. These tumours have also a different clinical presentation involving predominantly adults with advanced age and following a more aggressive clinical course (Falini et al, 1999; Gascoyne et al, 1999; Suzuki et al, 2000; ten Berge et al, 2003a). Similar to ALK-positive ALCL, the profile of genetic alterations of this group of lymphomas has not been well defined. Malignant lymphomas are frequently characterized by a distinct recurrent primary chromosomal alteration, such as the 2p23 rearrangements, in ALK-positive ALCL. In addition, the profile of secondary chromosomal alterations is also relatively specific of each disease entity (Bea et al, 1999, 2002; Soulier et al, 2001; Tsukasaki et al, 2001). Therefore, the analysis of chromosomal aberrations in ALK-positive and negative ALCL may provide useful information to define the genetic relationship between these two types of tumours.
This study examined a large series of ALK-positive and negative ALCL using comparative genomic hybridization (CGH) to define the profile of chromosomal imbalances of these tumours, determine their potential clinical relevance and elucidate whether these types of lymphomas are genetically distinct.
Material and methods
Forty-three ALK-positive ALCL were selected from the files of the Departments of Pathology of the Centre Hospitalier Universitaire du Purpan, Toulouse, France, Hospital Clinic of Barcelona, Barcelona, Spain, British Columbia Cancer Agency, Vancouver, Canada, University of Würzburg, Würzburg, Germany and Institute of Human Genetics University, Hospital Schleswig-Holstein, Kiel, Germany. All samples were obtained at diagnosis before any therapy was given. In one case an additional sequential sample could be examined at relapse 7 months after diagnosis. All cases were selected based on the presence of a large number (>60%) of tumour cells. Nine tumours were small cell variants. Immunohistochemical analysis was performed in all cases. Thirty-five cases had a T-cell and eight had a null phenotype. Twenty-eight tumours had the NPM-ALK translocation whereas a variant translocation was observed in eight tumours (TPM3-ALK variant, three cases; TSPYL2-ALK, two cases; ATIC/ALK, two cases; and TFG-ALK, one case). Four additional cases had only cytoplasmic restricted distribution of ALK protein, indicating the presence of a variant translocation, but molecular studies failed in these cases and in three cases no information regarding the pattern of ALK protein expression could be retrieved although they had an ALK break detected by Fluorescence in situ hybridization (FISH). Twenty-five patients were male and 18 female with a mean age of 21·4 years (range 4 months to 75 years). The Ann Arbor stage distribution was I 11%, II 33%, III 19% and IV 37%. Response to therapy was assessed according to standard criteria (Cheson et al, 1999). Among the 29 patients in whom the response was assessable, 25 (86%) achieved complete response (CR) to therapy.
To compare the genomic profile of ALK-positive and ALK-negative ALCL, thirty-one ALK-negative ALCL were selected from the same institutions. Thirteen of these cases had been included in a previous study (Zettl et al, 2004). The diagnosis of these ALK-negative cases was established according to the World Health Organization (WHO) classification (Delsol et al, 2001). Twenty-three tumours were of T-cell phenotype and eight of null phenotype and all of them showed strong and diffuse CD30 positivity. All samples were obtained at diagnosis before any treatment. Twenty patients were men and 11 women with a mean age of 54·2 years (range 7–92 years). The Ann Arbor stage distribution was: stage I 19%, II 29%, III 19% and IV 33%. Among the 16 patients in whom the response was assessable, 11 (69%) achieved CR and five (31%) failed to treatment.
Patients were treated according to the local protocols with different adriamycin-containing regimens in all but two cases. Complete follow-up information was available in 33 ALK-positive and 21 ALK-negative ALCL. The median follow-up of surviving patients was 45 months (range, 1–189) (ALK-positive ALCL, 44 months and ALK-negative ALCL, 79 months).
Comparative genomic hybridization
High-molecular-weight DNA was obtained from fresh-frozen material in 61 cases, from cell suspension frozen tissue in six cases, and from formalin-fixed paraffin-embedded tissue blocks in seven ALK-negative cases. Standard proteinase K/RNAse treatment and phenol–chloroform extraction was performed. CGH analysis was performed as previously described (Bea et al, 1999; Zettl et al, 2004). All CGH data of individual cases are available at http://www.ncbi.nlm.nih.gov/sky.
Cytogenetic studies were performed in six patients at diagnosis prior to any treatment using routine protocols (Schlegelberger et al, 1999). Karyotypes were described according to the International System for Human Cytogenetic Nomenclature, ISCN 2005 (Shaffer & Tommerup, 2005).
Fluorescence in situ hybridization and molecular studies
Fluorescence in situ hybridization and DNA molecular studies of TP53 and ATM, located at 17p13 and 11q22, two of the most commonly gained and lost regions respectively, in ALK positive ALCL were performed in selected cases. TP53 was examined by FISH using the LSI TP53 (17p13) Spectrum Orange-labelled Probe (Vysis, Drowners Grove, IL, USA) and the centromeric CEP17 (17p11.1-q11.1) Spectrum Green-labelled probe (Vysis). Two different observers scored 50 nuclei for each case. A true gain was considered when >5% of nuclei showed three hybridization signals as previously described (Bea et al, 2002).
Gene copy number changes of TP53 and ATM genes were performed by real-time quantitative polymerase chain reaction (RQ-PCR) with genomic DNA using the ABI Prism 7700 Sequence Detector System (Applied Biosystems, Foster City, CA, USA) and ß2-microglobulin gene (B2M) and albumin gene (ALB) as control genes as previously described (Bea et al, 2005). For each gene, the cut-off ratio was set as the mean ratio of the control samples ±3 SD units. Cut-offs for gains and losses of TP53 and ATM genes were 1 ± 0·22 and 1 ± 0·24, respectively. The primers and probes used are listed in Table I.
Table I. Primers and Taqman probes used in the real-time quantitative polymerase chain reaction.
We performed loss of heterozygosity (LOH) analysis involving the ATM locus for cases with losses of 11q21-q23 using three microsatellite markers located within and near the ATM gene. These markers were D11S2179, located between exons 62 and 63 of ATM, D11S1294, located 200 kb telomeric from ATM, and D11S1778, located between the ATM gene and the previous marker. Primers used and PCR conditions for the amplification and evaluation of these markers were described previously (Lespinet et al, 2005).
Differences among groups of tumours according to the parameters evaluated were compared by chi-square test. Overall survival (OS) was defined as the time from diagnosis to the time of death or last follow-up. Patients still alive were censored at the last known date of contact. The actuarial survival analysis was performed according to the method described by Kaplan and Meier, and differences were analysed by the log-rank test. P-values ≤0·05 were considered to indicate statistical significance.
Chromosomal imbalances in ALK-positive ALCL
Twenty-five of the 43 ALK-positive tumours (58%) showed chromosome gains (n = 51) or losses (n = 47) with a mean number of total alterations of 4 ± 3·8 (median: 3; range: 0–21) per sample. Irrespective of the morphological variant of the lymphoma, the most frequent imbalances in the aberrant cases were gains of 2q(12%), 7p(12%), 17p(28%) and 17q(28%), and losses of 4q(28%), 11q(28%) and 13q(28%) (Fig 1, Table II). No high-level DNA amplification was detected in any case. Single CGH imbalances were detected in three cases and consisted of losses of chromosome 4, 11q14-q23 and gain of 18p. Minimal common overrepresented regions were limited at 17q24-qter (seven cases) and 7p21-pter (three cases), whereas the regions with common loss of genetic material were delineated to 4q13-q28 (seven cases), 11q21-q23 (seven cases) and 13q21-q31 (six cases) (Table III). Twelve cases showed evidence for an ALK variant translocation not targeting NPM and chromosomal imbalances were observed in five (42%) of them. These alterations included gains of 2p and 2q and losses of chromosome 4, 5q, 6q, 9q, 11q and 13q (Fig 1, Table III). Therefore, these cases showed a similar profile of CGH alterations than ALCL with the NPM-ALK translocation. Similarly, no differences in genomic imbalances were detected between cases with common morphology and small cell morphology. Two sequential samples of one case (case 59) showed no alterations at diagnosis and a gain of 18q22-q23 region in the subsequent sample obtained at relapse 7 months after the diagnosis.
Table II. Comparative genomic hybridization (CGH) results in 43 ALK positive and 31 ALK negative ALCL.
Gains of more than one copy are indicated in bold.
Twenty of the 31 ALK-negative tumours (65%) showed chromosomal imbalances (total number of losses 56 and total gains 47) with a mean number of alterations of 5·4 ± 6·2 per case (median: 4; range: 0–25). Recurrent amplifications were detected in 9p and 17q (two cases each). Three cases presented losses of 13q31-qter, 9p24-pter and 6q as single alterations. The most frequent alterations in the altered cases were gains of 1q (35%), 5q (19%), 6p (20%), 7p (15%), 7q (15%), 8q (20%),12q (20%) and 17q (15%), and losses of 4q (15%), 6q (30%), 11q (15%) and 13q (35%) (Fig 2, Table II). Minimal common overrepresented regions were located at 1q41-qter (seven cases), 6p21 (four cases), 7p21-pter (three cases), 7q31-qter (three cases), 8q24-qter (four cases) and 12q24-qter (three cases), whereas the common lost regions were limited to 6q13-q22 (six cases), 11q21-q23 (three cases) and 13q32-q33 (six cases) (Table III).
Comparison between ALK-positive and ALK-negative cases
Chromosomal alterations that occur more frequently in ALK-positive ALCL were gains of 17p and 17q24-qter and losses of 4q13-q21 and 11q14 (P = 0·05). Similarly, gains of the 1q41-qter (P = 0·02) and 6p21 (P = 0·03) were significantly associated with ALK-negative cases. Several frequent imbalances, such as gains of chromosome 7 and losses of 6q, 11q and 13q, were observed in both types (Table III).
FISH and molecular studies
To determine whether the common 17p gains observed in ALK-positive ALCL included extra copies of TP53 FISH and RQ-PCR studies were performed in three and two cases respectively, with gains of this region. FISH detected an extra copy of this gene in the three cases (cases 52, 44 and 30) with 17p gains by CGH but not in six cases with a normal profile. Similarly, RQ-PCR showed gains of TP53 in two cases (cases 24 and 52) with 17p gains but in none of the 10 cases with a normal profile of chromosome 17p. These findings indicate a good concordance between the genetic and molecular studies confirming that 17p gains in ALK-positive ALCL include a gain of TP53. TP53 protein overexpression is a common phenomenon in ALCL that is not related to mutations of the gene (Cesarman et al, 1993; Rassidakis et al, 2005). To determine whether TP53 overexpression could be associated with the increased gene copy number observed in this study, we stained 12 ALK-positive ALCL for TP53 protein, four of them had gains of TP53 and eight showed a normal 17p profile. TP53 overexpression with >30% of positive cells was observed in five (42%) tumours. The remaining cases were negative (two cases) or showed only occasional positive cells (five cases). No relationship between the expression pattern and the gene copy number was observed because only one of the tumours with increased TP53 copy number had protein overexpression whereas the other three ALCL were either negative (one case) or had only occasional positive cells (two tumours).
One of the most common lost regions in ALK-positive ALCL was 11q14-q23. To determine whether these losses included the ATM locus, we performed a RQ-PCR analysis of the ATM gene in five tumours with losses of this region and 10 cases with a normal chromosome 11 profile. Deletions of ATM were detected in four of the five cases with losses in the 11q21-q23 region (cases 1, 21, 28 and 31) but in none of the tumours with a normal chromosome 11 profile.
Loss of heterozygosity analysis was performed in three microsatellite markers near/within the ATM locus in three tumours with 11q14-q23 losses and three cases with a normal profile in which DNA from normal lymphocytes of the same patients could be obtained. The analysis of these markers showed a LOH in the region of ATM in the three cases (cases 18, 21 and 67) with chromosomal losses but in none with a normal profile. In case 18, LOH was demonstrated in the more telomeric microsatellite D11S1294, whereas the microsatellite D11S1778 showed a normal bi-allelic profile and the intragenic D11S2179 was non-informative, suggesting that the deletion involved a region telomeric to ATM. Case 21 had a LOH in the microsatellite D11S1294 but the microsatellites D11S1778 and D11S2179 were non-informative and, therefore, the exact limits of the LOH could not be determined. In case 67, LOH was demonstrated in the D11S2179 microsatellite whereas the microsatellite D11S1294 showed a shift 244 to 248 and the microsatellite D11S1778 showed a normal bi-allelic profile.
Conventional cytogenetics and CGH comparison
Conventional cytogenetics results from six cases, four ALK-positive and two ALCL ALK-negative ALCL, were compared with CGH data. All the results were in complete accordance except for trisomy 7 in case 74 that was observed by conventional cytogenetics but not by CGH (Tables II and IV). The low number of metaphases in which this alteration was observed (3/22) is a possible explanation for this discrepancy, indicating that is present only in a subpopulation of the tumour.
Table IV. Conventional cytogenetics results of six cases.
The ALK-positive ALCL patients were significantly younger than ALK-negative ALCL (21·4 ± 16·7 vs. 54·2 ± 23·7; P ≤ 0·001) but showed no differences in terms of gender, stage and International Prognostic Index. The number and profile of genetic aberrations did not correlate with the clinical characteristics in either ALK-positive or ALK-negative patients. Eighteen of 54 patients with available follow-up died during this time, with a 5-year OS of 67% (95% confidence interval: 53–81%). Patients with ALK-positive ALCL showed a longer OS than ALK-negative ALCL (5-year OS: 79% vs. 51% respectively; P = 0·03) (Fig 3). A detailed analysis regarding survival was performed in ALK-positive and ALK-negative ALCL cases separately; it showed that no altered region, or number of alterations per case was associated with prognosis in either group.
Previous genetic and molecular studies in ALCL have been very successful in identifying the t(2;5) translocation and a series of variant translocations, always involving ALK at 2p23, as the primary genetic event in this subtype of ALCL. These findings have been instrumental in defining ALK-positive ALCL as a specific disease entity. However, the secondary genetic alterations that may be important in the evolution of the disease are not well known. This is due in part to the low number of recurrent genetic alterations detected in the relatively few cases that have been examined by classical genetic studies (Chou et al, 1996; Pittaluga et al, 1997; Ott et al, 1998; Rosenwald et al, 1999; Colleoni et al, 2000; Cools et al, 2002; Onciu et al, 2003; Liang et al, 2004).
Our CGH analysis shows that ALK-positive ALCL carry frequent secondary chromosomal imbalances, including losses of chromosome 4, 11q and 13q, and gains of 7, 17p and 17q. Previous cytogenetic studies of these tumours have recognized only recurrent gains of chromosome 7 (Ott et al, 1998). Gains of this chromosome have been observed also in 9–31% of other T-cell neoplasms, including peripheral T-cell lymphoma, unspecified (PTCL-U) (Lepretre et al, 2000; Zettl et al, 2004). Similarly, losses of 13q and gains of 17q are also common alterations observed in PTCL-U and other T-cell neoplasms (Schlegelberger et al, 1994; Renedo et al, 2001; Tsukasaki et al, 2001; Melendez et al, 2004; Zettl et al, 2004). However, losses of chromosome 4 and gains of 17p are uncommon among other types of T-cell lymphomas, including ALK-negative ALCL, suggesting that they may be relatively characteristic of ALK-positive ALCL (Tsukasaki et al, 2001; Zettl et al, 2004). ALK-positive and negative ALCL are considered to share morphological and phenotypic features. However, these tumours differ in their clinical presentation, response to therapy, survival and the global expression profile, suggesting that they must have different pathogenetic mechanisms (ten Berge et al, 1999, 2003b; Falini et al, 1999; Gascoyne et al, 1999; Suzuki et al, 2000; Lamant et al, 2007). The differences in the profile of certain chromosomal aberrations observed in this study would support the idea that these lymphomas are also different genetic diseases.
The chromosomal imbalances observed in ALK-negative ALCL also differ from those identified in PTCL-U (Table V). Losses of 5q (26%) and 9p (31%) were commonly detected in a previous study of PTCL-U with aberrant karyotypes (Zettl et al, 2004) but were absent in our ALK-negative ALCL population (P = 0·01 and P < 0·01, respectively). In addition, losses of 12q were present in 28% PTCL-U (Zettl et al, 2004) but only in one ALK-negative ALCL (5%) (P = 0·04). These tumours had instead gains of 12q in 15% of the cases. In the previous study of PTCL-U (Zettl et al, 2004), 11 of the 35 alterations had variable expression of CD30, these cases showed losses of 5q (9%), 9p(27%) and 12q(27%), a genetic profile more similar to the alterations of the remaining PTCL-U than to those observed in ALK-negative ALCL.
Table V. Comparison of genetic alterations in ALK-negative ALCL and PTCL-U.
One of the most common alterations in ALK-positive ALCL was gain of 17p. This alteration is unusual in other types of lymphomas. Several studies have indicated that TP53 protein was overexpressed in these tumours, independently of mutations of the gene (Cesarman et al, 1993; Rassidakis et al, 2005). We did not find any correlation between the presence of 17p gains and the overexpression of the protein, suggesting that other mechanisms regulate TP53 expression in these tumours. Curiously, a recent study that generated transgenic mice carrying supernumerary copies of TP53 showed that these animals had an enhanced response to DNA damage and were more resistant to develop induced cancer, indicating that this genetic modification was protective against cancer development (Garcia-Cao et al, 2002). The extra copies of TP53 observed in the relatively indolent ALK-positive ALCL suggest that these tumours may be an interesting model to examine the role of supernumerary TP53 in the pathogenesis of tumours.
Losses of 11q21-q23 were also a common alteration in ALCL. This region is frequently lost in T-prolymphocytic leukaemia associated with ATM mutations. The present study confirmed that losses of this region included ATM in some tumours but not always, suggesting that other genes may be the target of these losses in ALK-positive ALCL.
The clinical outcome of patients with ALK-positive ALCL is usually favourable and the survival significantly better than that of ALK negative ALCL patients (Shiota et al, 1995; Falini et al, 1999; Suzuki et al, 2000; ten Berge et al, 2003b). The present results are concordant with these previous observations. However, we did not identify specific genetic alterations with prognostic significance in either of the two types of ALCL.
In summary, the present study demonstrated that ALK-positive and negative ALCL have a different representation of secondary genetic alterations, supporting the concept that they are different biological entities. In addition, ALK-negative ALCL seem to carry chromosomal aberrations that also differ from the profile observed in other T-cell neoplasms, suggesting that it may correspond to a distinctive genetic subtype of these lymphomas.
This work was supported by the Spanish Ministry of Education and Science (SAF 05/5855), and Instituto de Salud Carlos III, Red Temática de Investigación Cooperativa de Cáncer, Grant ‘Distinció per a la Promoció a la Recerca de la Generalitat de Catalunya’ (60BA200406008), Grant ‘Ajut de suport a grups de recerca consolidats de la Generalitat de Catalunya’ (1-2005-SGR00870) and the Kinder-Krebs-Initiative (KKI) Buchholz, Holm-Seppensen.