Professor Daniel Catovsky, Academic Department of Haematology and Cytogenetics, The Royal Marsden Hospital, 203 Fulham Road, London SW3 6JJ, UK. E-mail: firstname.lastname@example.org
We reviewed eight cases that were diagnosed before 1995 with B-prolymphocytic leukaemia (B-PLL) harbouring t(11;14)(q13;q32) and/or cyclin D1 staining. Thirteen B-PLL patients without t(11;14) were selected as controls. Peripheral blood, bone marrow and histological sections were re-examined without cytogenetic information. Final diagnosis was made using morphology, cytogenetics, immunophenotype and immunohistochemistry. Clinical characteristics were similar for both groups except for younger age, male predominance and extranodal involvement in cases with t(11;14). CD5 was more frequently positive in the t(11;14)+ group (80%) than in the t(11;14)− group (31%). Surface membrane immunoglobulin was strongly expressed by all t(11;14)+ cases, but only 45% of t(11;14)− cases. Histopathological and cytological review of cases with t(11;14) showed an infiltrate with a mixture of cells, some resembling prolymphocytes and others with mantle cell lymphoma (MCL) morphology. Blood films of cases with t(11;14) showed features suggestive of B-PLL in three, and in others, a mixture of cells resembling MCL and nucleolated ones; none corresponded to the blastoid form of MCL. We suggest that ‘B-PLL’ with t(11;14) may represent a splenomegalic form of MCL evolving with leukaemia. These cases illustrate the importance of tissue diagnosis with cyclin D1 staining and fluorescence in situ hybridization analysis in B-cell leukaemia with prolymphocytic features.
Prolymphocytic leukaemia (PLL) was first described in 1974 as a rare variant of chronic lymphocytic leukaemia (CLL) (Galton et al, 1974). Characteristically, patients presented with massive splenomegaly, minimal lymphadenopathy, very high lymphocyte counts, resistance to chemotherapy and had a poor prognosis (Galton et al, 1974; Bearman et al, 1978; Catovsky, 1982). The diagnosis is based mainly on morphology, with prolymphocytes comprising a high proportion of the circulating lymphocytes.
Mantle cell lymphoma was accepted as a distinct clinicopathological entity in 1994, but clinical and morphological variants such as MCL with leukaemia and nucleolated variants have been more recently recognized (Dunphy & Perkins, 2001; Schlette et al, 2001; Wong et al, 2002). Because of the clinical and morphological similarities between MCL with leukaemia and B-PLL, we re-examined diagnostic material from patients with B-PLL known to be t(11;14) positive, and compared them with those without t(11;14) in order characterize the relationship between these two disease processes.
Patients and methods
Patients with a diagnosis of B-PLL who had the t(11;14)(q13;q32) translocation were identified from fluorescence in situ hybridization (FISH) and cytogenetic records. Only cases with histological samples of spleen, lymph nodes or other tissues were selected. Eight patients diagnosed between 1977 and 1995 were found, all of whom had undergone splenectomy and had tissue available for examination including one case that was positive for cyclin D1 staining but without cytogenetic studies available. A second group of 13 patients diagnosed with B-PLL, who were negative for the t(11;14) translocation, were selected as controls. Three of the latter had spleen samples available for histological review. One patient included in the control group was negative for t(11;14) and had weak, equivocal cyclin D1 positivity. Histology of tongue, stomach and colon were also available in three patients. Five t(11;14)(q13;q32) positive patients and 12 controls (including the case with borderline cyclin D1) were evaluated by FISH analysis. Three patients were tested by conventional cytogenetics alone and all had additional cytogenetic abnormalities.
Case 1. 35-46XY del(6)q16, t(11;14)(q13;q32), +3 markers.
Case 3. 46XX −6, −21, t(6:9)(p21;q34) t?, 8p−, 16q+.
Clinical and laboratory data were retrieved from the hospital database and from treating physicians.
The peripheral blood films and bone marrow aspirates were reviewed by four of the authors, and the histological sections were reviewed by another. Information as to the cytogenetic status of the patients was withheld until a final consensus on the diagnosis was made.
Immunophenotyping was performed by flow cytometry (FACScan) on mononuclear cells as previously described (Matutes et al, 1994). A cut-off of 30% was used for CD5, CD23 and FMC7 positivity.
A final diagnosis was reached using in combination morphology, immunophenotype, cytogenetics/FISH, histology and immunohistochemistry.
Cytogenetics and FISH analysis
Mononuclear cells were isolated and cultured with mitogens for conventional cytogenetics and G-banded according to standard procedures, and processed for FISH by treating with hypotonic KCl and fixing in Carnoy's solution (one part glacial acetic acid and three parts methanol). The probes used for FISH are as follows. (i) LSI IGH/CCND1 dual colour (IGH-green/CCND1–red), dual fusion DNA probe (Vysis UIC, London, UK). This probe gives two fusion signals in the presence of t(11;14)(q13;q32), representing the derivative chromosomes, along with one green and one red signal from the normal homologues. (ii) Cosmids: Cos 6.7 and Cos H1.5 (Dr E. Schuuring, Leiden, the Netherlands) (Brookes et al, 1992) labelled by nick translation with biotin and digoxigenin. Both are localized on 11q13 encompassing a 750 kb region around the BCL-1 major translocation cluster. Thus the red and green cosmid probes are co-localized in the normal 11 homologues, and are split if a translocation breakpoint falls between them. (iii) Whole chromosome 11 probe (green) combined with centromere 11 (red), producing three green signals if t(11;14)(q13;q32) is present, and two red signals for the centromere. (iv) LSI TP53 (17p13.1) (Vysis UIC).
Histology and immunocytochemistry for cyclin D1 on paraffin sections
Routinely, fixed paraffin wax-embedded material from the splenectomy specimens were retrieved from the pathology files. Haematoxylin and eosin stained sections were examined and immunohistochemistry was performed to analyse cyclin D1 staining. For immunocytochemistry, sections were cut at 3–4 μm and mounted on SuperFrost Plus microscope slides (Menzel-Glaser, Braunschweig, Germany). The sections were immersed in preheated high pH Tris–EDTA–citrate buffer (pH 9) and boiled under pressure in a microwave for 4 min. The sections were stained with a monoclonal antibody against cyclin D1 (Novocastra Laboratories, Newcastle-Upon-Tyne, UK) with a streptavidin biotin technique on an automated Ventana ES Immunostainer (Ventana, Strasbourg, France) according to the manufacturer's protocol using a diaminobenzamine detection system. Suitable control tissue (including a confirmed case of MCL) was used. Appropriate staining was also confirmed by detection of positive staining of endothelial cells within each test and tissue section.
Based on the cytogenetic/FISH studies, the cases were divided into two groups. The first group comprised eight cases: seven cases were t(11;14) positive and one case was cyclin D1+ but did not have cytogenetic studies. In six of the eight cases cyclin D1 was positive in spleen and bone marrow samples, and in two cyclin D1 was not available.
Among the second group of 13 cases, t(11;14) was not detected by FISH or conventional cytogenetic analysis, and cyclin D1 staining was negative in two of them. In one case, the cyclin D1 staining was weak and equivocal. In the remaining 10 cases cyclin D1 was not available.
Peripheral blood films were available for review in 20 cases and comprised seven of eight with t(11;14), and 13 without t(11;14). Cells from three of these seven cases with t(11;14) had morphological features suggesting B-PLL and, in another, there was a mixture of small lymphocytes and larger nucleolated cells resembling CLL with an increased number of prolymphocytes. In all of these four cases, the predominant cell population was of medium size with relatively condensed but not clumped chromatin, regular nuclear outline and a single nucleolus (Fig 1). The morphology in the three other cases within this group was as follows: typical of MCL in one case as shown by a pleomorphic picture, speckled nuclear chromatin, nuclear indentations and absent or very small nucleolus; the other two cases were difficult to classify on morphological grounds as there was a mixture of cells with features of B-PLL and others with nuclear clefts suggestive of MCL (Fig 2). None of these cases had morphological features suggestive of the blastoid form of MCL.
In the group without t(11;14), nine of the 13 cases had morphological characteristics of B-PLL, namely monomorphic lymphoid cells with less condensed chromatin, a prominent, central and vesicular nucleolus without nuclear irregularities and slightly basophilic cytoplasm (Fig 3). In the remaining four cases, nucleoli were less prominent, or there was a mixture of lymphocytes with nuclear indentations and prolymphocytes.
Cases with t(11;14) showed a characteristic cellular composition. In two cases, there was diffuse effacement of both red and white pulp with a mixture of cells that included some with morphology typical of MCL. Intimately admixed with these cells was a more dominant population of intermediate sized cells that were slightly larger than those seen in MCL. These cells had nuclei showing coarse, clumped chromatin and a single, central, small but prominent eosinophilic nucleolus (Fig 4). In six cases, the infiltrate was mainly nodular with a less pronounced dispersed infiltration of the red pulp (Fig 5). In the latter group, the nodules contained intermediate-sized nucleolated cells similar to those seen in the first group, but a variable proportion of the nodules had a central core of cells with MCL morphology. The dispersed cells in the red pulp infiltration showed more typical MCL-like morphology.
In B-PLL cases that lacked t(11;14) there was a predominantly nodular infiltration of the spleen by a uniform population of intermediate-to-large cells with abundant cytoplasm (Fig 6). The nuclei were more pleomorphic in size and shape than those seen in cases with t(11;14). The chromatin was less clumped but contained a similar single, mostly central, eosinophilic nucleolus.
Sections of colon and tongue from two t(11;14) positive cases showed, respectively, solid nodules of small MCL type cells and a dense subepithelial infiltrate of large mantle cells admixed with nucleolated cells. Bone marrow trephine biopsies were available for review in five t(11;14) positive and four t(11;14) negative cases. The t(11;14) positive cases showed a mild to dense nodular and interstitial pattern of infiltration. In some, nucleolated (PLL-like) cells were seen with mantle-like cells; in two cases the latter cells predominated. Three of the cases that lacked the t(11;14) showed nodular infiltrates of nucleolated cells with an inter-trabecular distribution. The fourth case showed nucleolated cells with a diffuse, packed pattern. A gastric biopsy from a third case showed an infiltrate in the lamina propria and muscularis mucosa with non-nucleolated angular cells consistent with MCL.
Clinical and laboratory data
The median age was 56 and 73 years for the t(11;14) positive and negative patients, respectively, with a slight male preponderance in the former. Splenomegaly, absence of lymphadenopathy and central nervous system involvement were similar in both groups, but extra-nodal sites of involvement (tongue, stomach, intestine and skin) were seen only in the t(11;14) positive group. In both groups high leucocyte counts and anaemia were common at diagnosis, and thrombocytopenia was less frequent (Table I). The leukaemic cells of the majority of the patients in both groups were FMC7+. CD5 expression was more frequently positive in the t(11;14) positive group than in the t(11;14) negative group. Surface membrane immunoglobulin (SmIg) was more strongly expressed in t(11;14) positive cases and CD23 expression was more frequently absent in t(11;14) negative cases, but these differences were not statistically significant (Table I). TP53 deletions by FISH were frequent in both groups.
Median survival was 56 months in the t(11;14) negative group and was not reached in the t(11;14) positive group, but the difference was not statistically significant (Fig 7). One patient in the t(11;14) positive group survived for 14 years after diagnosis. All patients in the t(11;14) positive group underwent splenectomy, and in addition, received up to five other lines of therapy. Partial responses to CHOP (cyclophosphamide, hydroxydaunomycin, Oncovin, prednisone) occurred in three patients, but these were short lived (<4 months). One patient had a partial response to fludarabine lasting 11 months, and later achieved a complete remission lasting 5 months with Campath 1H. Two patients received chlorambucil with minimal response. There was limited clinical information available among the t(11;14) negative cases. One patient remained well and has not required treatment 25 months after diagnosis. Two patients were treated with chlorambucil with minimal response, and three patients underwent splenectomy.
Following the review of histopathology, immunophenotype, immunohistochemistry and cytogenetics, all seven cases previously diagnosed as t(11;14) positive B-PLL and one with cyclin D1 positive B-PLL were re-classified as MCL. The clinical characteristics of the patients in both groups were similar and non-contributory to the diagnosis. CD5, CD23 and strength of immunoglobulin expression differed but were not statistically significant. Peripheral blood and bone marrow morphology alone was insufficient for an accurate diagnosis; in most cases, the differences were slight and challenging, even for an expert panel.
The spleen histology, however, appeared different between the t(11;14) positive and negative cases. In particular, in all cases in which t(11;14) was confirmed, a population of typical mantle cells could be found. These mantle-type cells were absent in the t(11;14) negative cases. We compared these cases with splenectomy specimens from patients with typical MCL without circulating nucleolated cells (data not shown). In the latter cases, the infiltrate was predominantly nodular with the nodules composed of small-to-intermediate sized cells with irregular nuclei. The nuclei had coarsely granular chromatin but no nucleoli, in contrast to the t(11;14) positive cases reviewed in the present series.
Wong et al (2002) reported four patients with a nucleolated variant of MCL with leukaemic manifestations mimicking B-PLL. All four had splenomegaly, generalized lymphadenopathy and prolymphocytoid cells in the peripheral blood. The diagnosis was based on an immunophenotype of CD5+, CD19+, CD23−, overexpression of cyclin D1 and supplemented by lymph node biopsy which showed MCL, pleomorphic variant in all four, and the cytogenetic findings of t(11;14)(q13;q32) in three (Wong et al, 2002). Our t(11;14) positive cases showed certain histological similarities to the cases of nucleolated MCL described by Wong et al (2002). However, in the majority of our cases, particularly those with a nodular pattern, the number of nucleolated cells were in marked excess compared with classical mantle cell morphology.
The final diagnosis of the eight patients was based on spleen histology, cyclin D1 staining on tissue sections and cytogenetics, particularly FISH analysis on peripheral blood samples. One patient was included in the t(11;14) negative group based on the absence of the translocation by FISH, typical peripheral blood morphology of B-PLL and spleen histology, with equivocal cyclin D1 staining.
The World Health Organization classification defines B-PLL as a malignancy of B prolymphocytes affecting the blood, bone marrow and spleen; these cells exceeding 55% of the lymphoid cells in the blood, and excludes transformed CLL (Catovsky et al, 2001). The difficulty in distinguishing B-PLL from the leukaemic blastoid variant of MCL by morphology or immunophenotype is acknowledged.
We have reviewed here a large number of patients previously designated as B-PLL, including eight who were t(11;14) and/or cyclin D1 positive. These cases were diagnosed in a period before and during which MCL was first accepted as a separate entity, and the diagnostic importance of immunophenotype and cytogenetics was not apparent.
These cases highlight the importance of tissue diagnosis with cyclin D1 staining and cytogenetics in leukaemias with prolymphocytic features. We suggest that B-PLL with t(11;14) may represent, in fact, a splenomegalic form of MCL often associated with leukaemia, and that in future, this investigation should be performed in all cases in which a diagnosis of B-PLL is suspected. Studies involving VH mutational analysis and gene expression profiling may provide a definitive answer as there is information pertaining to MCL (Orchard et al, 2003; Staudt, 2003), but not yet to B-PLL. In the meantime, we propose that t(11;14) and/or cyclin D1 positivity in this context may define a subset of MCL with leukaemia, a condition not uncommon in our experience (Orchard et al, 2003).
We are grateful to Dr Ann Watson, Dr Phil Bevan, Dr M. D. Nightingale, Dr Richard Lush and Dr Ruth Spearing for contributing clinical data and histological sections, to R. Morilla and K. Owusu-Ankomah for cell marker studies and to J. Swansbury for the Kaplan–Meier survival analysis. This work was supported by a scholarship from the Barclay Family Cancer Research Foundation (RR) and by a Clinical Fellowship from the Leukaemia Research Fund, UK (NPJ).