In 1982, Weisenburger et al. first introduced the concept of mantle-zone lymphoma.1 According to the terminology used at that time, this was regarded as a variant of intermediate lymphocytic lymphoma that proliferated as wide mantles around non-neoplastic appearing germinal centers (GC). One year later, Swerdlow et al. found that the pattern described above was part of the spectrum of the centrocytic lymphoma of the Kiel Classification and might correspond to initial lymph node involvement by the tumor.2,3 In 1985, Pileri et al. reported on 18 cases of small B-cell lymphomas displaying a mantle-fashion growth around reactive GC, which turned out to be quite heterogeneous on closer examination.4 In fact, at disease presentation 13 of the 18 cases displayed cytological and immunological findings consistent with centrocytic lymphoma, while the remaining ones corresponded to neoplasms that nowadays would be diagnosed as marginal zone lymphoma. Interestingly, the centrocytic lymphomas were CD5+, FMC7+ and CD10− and showed progression to a diffuse growth pattern in follow-up biopsies. Two important conclusions were reached: (i) the mantle-fashion growth was produced by different types of B-cell lymphoma and could not be used as a diagnostic criterion, and (ii) conversely to what is reported in the Kiel Classification,3 centrocytic lymphomas are derived from a normal counterpart other than GC, possibly related to mantle cells. In 1992, the International Lymphoma Study Group (ILSG) made its first experiment to overcome the discrepancies in terms of lymphoma classification that had hampered communication across the Atlantic for some decades.5 In particular, the American and European ILSG members agreed on the existence of a tumor derived from mantle B cells that was consequently termed mantle-cell lymphoma (MCL). The criteria for its recognition were drafted and included 2 years later in the Revised European-American Lymphoma Classification that led the way to the latest edition of the World Health Organization (WHO) Classification of Tumours of Haematopoietic and Lymphoid Tissues.6,7 The present article will focus on the diagnosis and bio-pathology of MCL in the light of the criteria of the WHO Classification7 and recent reports in the literature including those of Quintinilla-Martinez et al., Mozos et al. and Dictor et al. in this issue of the journal.8–10 It will not include the still ongoing debate on the multiple options proposed for the treatment of MCL patients, which often have limited efficacy (comprehensively reviewed by Ghielmini et al.).11
MCL is a distinct entity, representing from 3–10% of non-Hodgkin’s lymphomas and more frequently affecting middle-aged to older males (male:female ratio = 2–7:1).7,12 It is listed among peripheral B-cell lymphomas, is usually composed of small to medium-sized elements with irregular nuclear contours and CCDN1 translocation, and is thought to stem from peripheral B cells of the inner mantle-zone of secondary follicles, mostly of naïve pre-GC type.7,12 MCL is generally regarded as an aggressive, incurable disease with the median survival of affected patients being 3–4 years.12
MCL more often presents in stage III–IV with lymphadenopathy, hepatosplenomegaly, bone-marrow involvement, and leukemic spread.7,12 The latter two findings can be missed by morphological studies alone and are shown by, respectively, immunohistochemistry and FACS analysis in most if not all instances.7,12 Waldayer’s ring and the gastro-intestinal tract are frequently affected.7,12 Most cases of multiple lymphomatous polyposis correspond to MCL.
MCL usually consists of small to medium-sized lymphoid elements with irregular nuclear contours, somewhat dispersed chromatin, inconspicuous nucleoli, and scant cytoplasm.7,12,13 Within this context, centroblasts, immunoblasts, prolymphocytes and para-immunoblasts are not encountered. This is relevant for the differential diagnosis from follicular lymphoma, lymphoplasmacytic lymphoma and chronic lymphocytic leukemia/small lymphocytic lymphoma, which by definition, contain centroblasts, immunoblasts and prolymphocytes/para-immunoblasts, respectively. Comprised within the neoplastic population, hyaline small vessels and/or epithelioid histiocytes are frequently found. Clusters of plasma cells can be seen: they are usually reactive (polytypic at immunohistochemistry), only exceptionally representing a feature of tumor differentiation with obvious monoclonality and tendency to accumulate in the center of lymphomatous nodules.14 The number of mitotic figures varies, although it is usually low to moderate (see below).
In the lymph node, spleen and Waldayer’s ring, neoplastic cells give rise to a mantle-fashion, nodular or –more frequently – diffuse growth pattern.7,12,13 Notably, repeated biopsies demonstrate that such patterns correspond to disease progression. In fact, in the early phases the tumor substitutes the mantle-zone; then, it invades reactive GC producing a nodular pattern; finally, the nodules merge together causing diffuse effacement of the organ’s structure. The bone marrow can be variably involved: paratrabecular infiltrates are the commonest finding.
Besides the classical type described above, there are some cytological and/or morphological variants of MCL that can cause diagnostic problems. These are termed blastoid, pleomorphic, small cell, and marginal zone-like.7,12,13 The first consists of slightly larger cells resembling lymphoblasts with dispersed chromatin and a high mitotic rate. Blastoid MCL is clinically aggressive, as is the pleomorphic form, which is characterized by a cell population that is highly variable in size and shape, with frequent mitotic figures and a proportion of rather large elements showing oval to irregular nuclear contours, pale cytoplasm and at times prominent nucleoli. Both the blastoid and pleomorphic forms occur de novo, rarely representing morphological progression of the disease. The small cell variant is composed of small, round lymphocytes with more clumped chromatin, mimicking chronic lymphocytic leukemia/small lymphocytic lymphoma. Occasionally, the small lymphomatous elements can be almost exclusively restricted to the inner mantle zones or to narrow mantles, a pattern termed in situ MCL and thought to have a more indolent behavior.15 Finally, the marginal zone-like subtype displays prominent foci of cells with abundant pale cytoplasm resembling marginal zone or monocytoid B-elements.7,12 Notably, these cases are at times characterized by splenomegaly, intrasinusoidal bone-marrow infiltration and a leukemic picture in the absence of lymph node involvement. They can be easily confused with splenic marginal zone lymphoma and seem to have a very indolent course. As for in situ MCL, the diagnosis of this rare form of the tumor is feasible only upon immunohistochemical detection of cyclin D1 (see below).
MCL is characterized by expression of the B-cell markers CD19, CD20, CD22, CD79a and BSAP/PAX5.7,12,13 Notably, CD20 is strongly expressed on the surface of neoplastic cells, a feature that can be useful for the differential diagnosis from chronic lymphocytic leukemia/small lymphocytic lymphoma, which shows weak CD20 positivity in the small cell component.13 The search for immunoglobulin (Ig) heavy and light chains usually reveals IgM/D positivity with more frequent lambda restriction.7,12 Neoplastic cells are usually CD5+, FMC-7+, BCL-2+, CD10−, BCL-6−, and IFR4−, with occasional weak expression of CD23.7,12,13 In the cases with nodular growth, residual GC cells (CD10+, BCL-6+) can be detected admixed with the neoplastic population. Furthermore, aberrant phenotypes have been reported in MCL (such as negativity for CD5 and positivity for CD10 and/or BCL-6), sometimes associated with the blastoid/pleomorphic variants or occurring in conjunction with BCL6/3q27 translocations.7 Stains for CD21 and CD35 usually show loose follicular dendritic cell meshworks. In the light of this, although the phenotypic profile of MCL is distinctive, it is not entirely pathognomonic and as such cannot always assist in the differentiation of MCL from follicular lymphoma, lymphoplasmacytic lymphoma, marginal zone lymphoma and chronic lymphocytic leukemia/small lymphocytic lymphoma. For this purpose, the most useful examination is the strong expression of cyclin D1, which is never found in the above mentioned categories of B-type non-Hodgkin’s lymphomas.7,12,13 This attribute is sustained by the t(11;14)(q13;q32) translocation (see below).7,12,13 Notably, cyclin D1 expression is lacking in a small percentage of cases of MCL: because of practical and conceptual implications, these cases will be treated separately (see below). Last but not least, cyclin D1 expression is found in hairy cell leukemia (although weak) and in about 16% of plasma cell myelomas, also carrying t(11;14)(q13;q32).13 When commercially available, the antibody produced by one of the authors (Falini)16 and raised against the IRTA-1 receptor physiologically expressed by marginal zone B-cells, could represent a further tool to aid the differential diagnosis between MCL (IRTA-1−) and nodal marginal zone lymphoma (IRTA-1+ in about 80% of cases).17
The t(11;14)(q13;q32) translocation
The t(11;14)(q13;q32) translocation is regarded as the primary genetic event in MCL.7,12 It can easily be demonstrated by fluorescent in situ hybridization (FISH) and involves the Ig heavy chain gene (IGH@) at the 14q32 locus and a region at 11q13 designated BCL1.12 The expression of CCND1, located in the latter and encoding for cyclin D1, is deregulated by the translocation.7,12 This causes over-expression of cyclin D1 both at the mRNA and protein levels in MCL. Cyclin D1 plays an important role in the cell cycle regulation of G1-S transition following mitotic growth factor signaling.7 Unlike in MCL, in normal cells cyclin D1 is transiently expressed and binds to CDK4 and CDK6 to form a CDK/cyclin complex able to phosphorylate the tumor suppressor gene retinoblastoma (RB1) facilitating cell cycle progression.12 Deregulated expression of cyclin D1 is thought to overcome the effect of RB1 and p27/kip, leading to the development of MCL.7,12 Some tumors over-express shorter cyclin D1 transcripts, usually generated by secondary 3′ rearrangements in the CCD1 locus.7,12 These truncated transcripts cause very strong protein expression, high proliferative activity and a more aggressive clinical course.7,12
Notably, it has been reported that inactivating mutations of the ATM and/or CHK2 genes (recorded in 40–75% of MCL) are at times found in the germline of some patients.7,12 This might be relevant to the development of the tumor: in fact, such mutations cause deregulation of the DNA damage response pathway and increased genomic instability.7,12 In particular, t(11;14) –which should affect immature/naïve B cells as suggested by its detection in the peripheral blood of 1–2% of healthy individuals – would occur in the setting of a population carrying these mutations and lead to constitutive deregulation of cyclin D1 and early expansions of tumor B cells in the mantle zone of lymphoid follicles. This might also explain why MCL is one of the malignant neoplasms with the highest level of genomic instability, which might cause additional oncogenic events needed for the expansion of MCL cells with classical morphology, as well as alterations of genes involved in cell cycle regulation and senescence regulatory pathways that would lead to more aggressive variants of MCL.7,12,18 In particular, there is a high number of non-random secondary chromosomal aberrations in MCL, including gains of 3q26, 7p21 and 8p24 (MYC), and losses of 1p13-p31, 6q23q27, 9p21, 11q22-q23, 13q11-q13, 13q14-q34, and 17p13-ter.7,18 Trisomy 12 is detected in about 25% of cases. Tetraploid clones are more common in the pleomorphic and blastoid variants than in the classical type (80%, 36% and 8%, respectively).7,12 t(8;14)(q24;q32) involving MYC is rare, but associated with a more aggressive clinical course. Finally, MCL carries frequent mutations of TP53, homozygous deletions of the INK4a/ARF, amplifications of the BMI1 polycomb and CDK4 genes, and occasional micro-deletions of the RB1 gene, all producing high proliferative activity.7,12
The possible steps of the molecular pathogenesis of MCL are outlined in Figure 1 in relation to immuno-morphological findings.
Antigen receptor genes
Ig genes are rearranged. IGH@ genes are unmutated in most cases, although mutations are found in a minority of MCL.7,12 The load of the latter is lower than in mutated chronic lymphocytic leukemia.7,12 In addition, in MCL IGH@ somatic mutations do not correlate with ZAP70 expression and survival.7,12 A biased usage of IGH@ genes has been reported in some cases, suggesting that MCL may originate from specific B-cell subsets. Unlike in chronic lymphocytic leukemia, in MCL VH3-21 usage occurs in unmutated cases and bears a lower amount of genomic imbalances and a better prognosis.7
Gene expression profile
In 2003, the Lymphoma/Leukemia Molecular Profiling Project group showed that MCL has a specific gene signature that differs from the signatures of small lymphocytic lymphoma and diffuse large B-cell lymphoma of both the GC B cell-like and activated B cell-like types.19 Interestingly, this signature encompassed the usual cyclin D1-positive cases and a small subgroup of cyclin D1-negative MCL. Besides elucidating the pathogenesis of the tumor, the gene expression profile (GEP) revealed 20 proliferation-associated genes whose expression measurement identified subsets of patients whose median survival differed by more than 5 years. Differences in cyclin D1 mRNA abundance synergized with INK4a/ARF locus deletions to dictate tumor proliferation rate and survival. Two years later, the same group reported on six cases of MCL lacking cyclin D1 that exhibited the same clinical, morphological and GEP features as classical MCL but did not carry the t(11;18)(q12;q32), as determined by FISH.20 These cases did, however, express either cyclin D2 or cyclin D3 at both the mRNA and protein levels, suggesting that up-regulation of these cyclins may substitute for cyclin D1 in the pathogenesis of MCL.
Cyclin D1-negative mantle cell lymphoma
Since cyclins D2 and D3 are also expressed by normal B cells and most, if not all, B-cell non-Hodgkin’s lymphoma,8 although at lower levels,12 these markers cannot be confidently used for differentiating cyclin D1-negative MCL from other small B-cell tumors.8,12 In 2008, this led Campo et al. to conclude that “the only reliable criteria to establish the diagnosis of cyclin D1-negative MCL seemed to be the microarray profile that is limited to research environments”.12 In this issue of the journal, three reports focus on tools that can allow the easy identification of such cases and can be applied in most histopathology labotatories.8–10 First of all, Quintinilla-Martinez et al. propose FISH analysis and/or quantitative reverse transcriptase polymerase chain reaction as a means for making the differential diagnosis of cyclin D1-negative MCL.8 This approach is based on the one hand on previous observations of bone fide cyclins D1−/D2+ MCL carrying translocations involving the CCND2 locus with either the IGK@ [t(2;12)(p12;p13)] or IGH@ locus [t12;14](p13;q32)], not recorded in either chronic lymphocytic leukemia or CD5+ mantle zone lymphoma, and, on the other hand, on the at least 10-times higher mRNA levels of cyclin D2 in D1−/D2+ MCL in comparison with normal lymphoid tissue, follicular lymphoma, mantle zone lymphoma and chronic lymphoma and chronic lymphocytuc leukemia.8 Mozos et al. candidate the expression of the neural transcription factor SOX11 as a powerful bio-marker for the identification of cyclin D1-negative MCL.9 In fact, GEP showed significantly higher expression of the SOX11 gene in classical MCL than in Burkitt’s lymphoma, diffuse large B-cell lymphoma, primary mediastinal B-cell lymphoma and follicular lymphoma.9 However, one third of cases of Burkitt’s lymphoma revealed SOX11 mRNA levels similar to those in MCL, thus indicating that SOX11 over-expression is not restricted to MCL.9 These findings were corroborated by immunohistochemical studies that showed nuclear SOX11-positivity in most, if not all, MCL tested, including 12 cyclin D1-negative cases.9 Interestingly, 25% of Burkitt’s lymphomas, all examples of B- and T-lymphoblastic leukemia/lymphoma, one classical Hodgkin’s lymphoma and 2/3 T-prolymphocytic leukemias also turned out to be positive.9 Equivalent results were reported by Dictor et al., who immunostained 172 specimens for the SOX11 N and C termini.10 Their findings indicate that SOX11 can represent a useful tool for the differentiation of MCL from other small B-cell lymphomas in general but is certainly of pivotal importance for identifying the cyclin D1-negative cases. Of course, it should be integrated in a panel of markers in order to avoid any confusion between a potential cyclin D1-negative blastoid MCL and lymphoblastic leukemia/lymphoma, which may be confused morphologically but are easily differentiated by the global phenotypic profile.
Apart from the rare forms with an indolent behavior reported above, MCL has a particularly poor outcome7,12 Several attempts have been made to subdivide patients into different risk groups. In particular, the European MCL Network has shown that the number of mitotic figures and Ki-67 marking can allow the identification of three subgroups with different proliferation rates (mitotic indices: <25, 25–49, and >50/mm2 and Ki-67 indices: <10%, 10–40%, and <40%, respectively) significantly different outcomes.21 Such differences are conserved in advanced-stage MCL patients treated with anti-CD20 immuno-chemotherapy.22 Interestingly, these findings are in keeping with the above reported GEP data19 and point to the prognostic value of proliferation in MCL. Additional bio-pathological adverse prognosticators – which may be independent of proliferative activity – have been reported, such as blastoid/pleomorphic morphology, trisomy 12, karyotype complexity, TP53 mutation/over-expression/loss, gains of chromosome 3q, deletions of 9p, and a series of clinical parameters including overt peripheral blood involvement.7,12 Recently, a clinical MCL index (MIPI), based on age, ECOG, performance status, lactate dehydrogenase values and leukocyte count, and a five-gene model (RAN, MYC, TNFRSF10B, POLE2, and SLC29A2) applicable to both frozen and routine samples and based on quantitative reverse transcriptase polymerase chain reaction analysis, have been proposed as further prognostic indicators.23,24
This manuscript was supported by grants from BolognAIL, Centro Interdipartimentale di Ricerche sul Cancro “Giorgio Prodi” and Fondazione Cassa di Risparmio in Bologna (Bologna, Italy).
Dr. Stefano A. Pileri is Full Professor of Pathology and Director of the Hematopathology Unit at Bologna University. Dr. Brunangelo Falini is Full Professor of Hematology at Perugia University.
- Copyright© Ferrata Storti Foundation