Molecular basis of mantle cell lymphoma


  • Francesco Bertoni,

    1. Experimental Oncology, Oncology Institute of Southern Switzerland, Bellinzona, Switzerland
    2. Experimental Haematology, Barts and The London – Queen Mary's School of Medicine and Dentistry, London, UK
    Search for more papers by this author
  • Emanuele Zucca,

    1. Experimental Oncology, Oncology Institute of Southern Switzerland, Bellinzona, Switzerland
    2. Medical Oncology, Oncology Institute of Southern Switzerland, Bellinzona, Switzerland
    Search for more papers by this author
  • Finbarr E. Cotter

    1. Experimental Haematology, Barts and The London – Queen Mary's School of Medicine and Dentistry, London, UK
    Search for more papers by this author

Francesco Bertoni MD, Experimental Oncology, Oncology Institute of Southern Switzerland, Via Vincenzo Vela 6 - Stabile IRB, 6500 Bellinzona, Switzerland.

The term mantle cell lymphoma (MCL) represents a group of lymphoma subtypes previously classified as centrocytic lymphoma, lymphocytic lymphoma of intermediate differentiation, intermediate cell lymphoma or diffuse small-cleaved cell lymphoma (Zucca et al, 1994; Swerdlow et al, 2001). MCL account for approximately 6% of all non-Hodgkin's lymphomas (The Non-Hodgkin's Lymphoma Classification Project, 1997). Despite being previously considered a low-grade and indolent lymphoma, MCL appears to have the worst characteristics of both low- and high-grade lymphomas: incurable and aggressive (Weisenburger & Armitage, 2000). MCL patients have a median age of over 60 years, a male predominance, and usually disseminated disease at diagnosis. MCL median time to progression and survival are the shortest among all lymphoma subtypes (The Non-Hodgkin's Lymphoma Classification Project, 1997). The blastoid variant has a much worse prognosis than the classic form, with a median survival of <2 years. To date, there are no standard treatments for MCL (Zucca et al, 1998; Weisenburger & Armitage, 2000; Dreyling et al, 2001; Cabanillas, 2002; Press, 2002).

The immunophenotype (Campo et al, 1999; Lai & Medeiros, 2000; Pileri et al, 2000; Dreyling et al, 2001; Swerdlow et al, 2001) of MCL cells is characteristically that of a mature B cell, expressing CD19, CD20, CD22 and CD79A, moderate to strong immunoglobulin (Ig)M and/or IgD surface immunoglobulins (sIg). Unique among all non-Hodgkin's lymphomas, MCL express lambda more commonly than kappa Ig light chains (IgL). Similar to chronic lymphocytic leukaemias (CLL), MCL are CD5+ and CD43+, but CD23 (conversely, CLL are CD23+), as well as CD10. MCL characteristically overexpress cyclin D1, due to the t(11;14)(q13;q32) translocation (see below).

The new World Health Organization classification (Swerdlow et al, 2001) defines MCL more clearly, comprising two main morphological subtypes, typical and a blastoid (blastic) variant. Typical MCL is characterized by a monotonous proliferation of small- to medium-sized lymphocytes, with irregular nuclei, condensed chromatin, inconspicuous nucleoli and scant, pale cytoplasm (Harris et al, 1994; Campo et al, 1999; Lai & Medeiros, 2000; Weisenburger & Armitage, 2000; Dreyling et al, 2001). In the blastoid variant the cells are medium-sized lymphocytes, resembling lymphoblasts, with scant cytoplasm, rounded nuclei with finely dispersed chromatin and inconspicuous nucleoli. The mitotic index is low in the classic and high in the blastic variant. It is often assumed that the blastoid variant represents the transformation of classic MCL.

In addition to the malignant cell type, histologically there are three different ‘architectural’ patterns within MCL and these may represent different stages of MCL lymph node infiltration. ‘Early’ lesion, the mantle-zone pattern, has an initial infiltration of the mantle-zone surrounding normal germinal centres. A nodular pattern resembling pseudofollicles may be considered a more extensive infiltration and the third, diffuse pattern, shows loss of the germinal centres due to neoplastic cells.

The MCL has now become a clearly defined subtype of B-cell lymphoma. A number of distinct genetic and biological alterations are associated with the disease including the t(11;14)(q13;q32) translocation, ATM (ataxia telangiectasia mutated) alterations and 11q deletion. Understanding the biological and genetic mechanisms underlying the bad prognostic disease, as well as providing insight into its pathogenesis may also provide pointers towards potentially curable therapy for MCL. This review aims at bringing together current thinking on the genetic basis of this disease.

Genomic abnormalities

Karyotype studies of MCL have shown the t(11;14)(q13;q32) as its hallmark, often associated with other recurrent chromosomal aberrations (Leroux et al, 1991; Knuutila et al, 1998, 1999, 2000; Bea et al, 1999; Bentz et al, 2000; Bigoni et al, 2001; Martinez-Climent et al, 2001).

The t(11;14)(q13;q32) and the cell cycle

In the t(11;14)(q13;q32) translocation, present virtually in all cases of MCL (Leroux et al, 1991; Rimokh et al, 1994; Zucca et al, 1995; Dreyling et al, 1997; Ott et al, 1997a; Chibbar et al, 1998; Fan et al, 1998; Bertoni et al, 1999a; Li et al, 1999; Stamatopoulos et al, 1999; Katz et al, 2000; Welzel et al, 2001; Belaud-Rotureau et al, 2002), the cyclin D1 gene, on 11q13, is juxtaposed to the IgH ‘Joining’ (J) region, on 14q32 (Fig 1). The t(11;14) determines the ectopic and deregulated expression of cyclin D1 in lymphoid cells, due to the juxtaposition of the gene to the strong B-cell IgH transcription enhancers. The expression of cyclin D1 is characteristic of virtually all MCL cases.

Figure 1.

Diagram of the breakpoints on chromosomes 14q32 and 11q13 determining the t(11;14)(q13;q32) translocation.

The breakpoints on 11q13 occur in the region termed BCL1, about 120 kb from the gene coding for cyclin D1 (CCND1, PRAD1) (de Boer et al, 1993; Williams et al, 1993) (Fig 1). Detailed analysis of the 11q13 and 14q32 breakpoints show that the translocations occur mainly during an attempted primary rearrangement between one ‘Diversity’ (D) and one J segment of the Ig locus (see below for Ig rearrangement) (Stamatopoulos et al, 1999; Welzel et al, 2001). Thirty-fifty per cent of the breakpoints on 11q13 occur within a restricted region, the major translocation cluster (MTC). Rearrangements involving this area can be amplified from genomic DNA by polymerase chain reaction (PCR) using primers directed to the Ig JH region and to the MTC region. The PCR assay for the t(11;14) translocation represents a molecular marker diagnostic for MCL. However, as it is directed exclusively to the MTC, it can detect the rearrangement in only 35% of the cases. In contrast, fluorescence in situ hybridization (FISH) can show the t(11;14) translocation in almost all cases: in a series of 35 MCL cases a positivity rate of 97% for FISH compared with 37% by PCR was reported (Belaud-Rotureau et al, 2002).

Polymerase chain reaction is the first method of choice in molecular follow-up studies of MCL (Ghielmini et al, 2000; Magni et al, 2000). In the bcl1/JH PCR negative cases, PCR assays targeting VDJ and Kde rearrangements may be used as molecular tools for detecting B-cell monoclonality at diagnosis or during follow-up studies (see below).

The D-type cyclin, in combination with cyclin-dependent kinase  4 and 6 (cdk4 and cdk6), regulates the cell cycle transition between the G1 to the S phase, phosphorylating the retinoblastoma protein (pRb) (Sherr, 2000; Blagosklonny & Pardee, 2001). There are three types of tissue-specific cyclin D. Myelomocytic and lymphoid cells use D2 and D3, while epithelial and mesenchymal cells mainly use D1. The passage from G1 to S phase is the most crucial point, after which the cell is irreversibly committed to completing the cycle. During the G1 phase, extracellular stimuli determine whether the cells undergo DNA replication, followed by the cell replication process or, if they exit the cell cycle, entering the quiescent G0 phase. Holoenzymes, containing regulatory (cyclin) and catalytic (cdk) subunits, control the passing from one to another cell cycle phase (Fig 2). G1 progression requires a sustained expression of the D-type cyclins, which, in turn, need a continuous mitogenic stimulation. This determines the effect of exogenous factors on the cell commitment to replication. Cyclin  E and cdk2 quickly follow cyclin  D/cdk4-6, and act on pRb and on other target proteins that are still unknown.

Figure 2.

Diagram of the interactions between cell cycle proteins at the G1-S phase. PCNA, proliferating cell nuclear antigen.

In MCL, cyclin D1 is expressed continuously and independently of exogenous factors, and it replaces cyclins D2 and D3 (Ott et al, 1997b; Campo et al, 1999). Cyclin D1 expression by itself is not sufficient to keep the cells in cycle and transgenic mice that overexpress cyclin D1 in their B cells do not have an increased incidence of lymphoma (Lovec et al, 1994; Adams et al, 1999). The mitotic index is low in classic MCL, and high in the blastoid variant. In normal cells, two classes of inhibitors regulate the cyclin-cdk complexes: the Cip/Kip family and the INK4 family. The Cip/Kip class comprises p21WAF1/CIP1 (chromosome 6p21.2), p27KIP1 (12p13) and p57KIP2 (11p15.5), which inhibit both cyclins D and E. The INK4 family includes p16INK4A, p15INK4B (all 9p21), p18INK4C (1p32) and p19INK4D (19p13), which selectively inhibit the cyclin D-associated kinases. The gene encoding for p16INK4A encodes a second protein, p14ARF, which links the Rb pathway to p53. Expression of p14ARF is induced by abnormal mitogenic signals induced by overexpression of other proteins such as Myc and E2F1, and it antagonizes the activity of Mdm2, a p53-negative regulator. P53 (17p13.1) acts as a genome guardian. Activated by DNA damages, p53 can stop cell cycle, through p21 activation, to allow DNA repair. If this fails, p53 can trigger the apoptotic process. Figure 2 shows a simplified diagram of the proteins involved in cell cycle regulation. Besides cyclin D deregulation, other regulatory molecules can be defective in MCL, and can contribute to lymphoma pathogenesis. As many defects are more prevalent among blastic variants, this might explain its higher mitotic index and worse prognosis in comparison with classic MCL. Three of seven (43%) blastoids, and only one of 21 (5%) classic variants presented loss of p16 expression (Pinyol et al, 1998). Deletion of p16 locus also usually involved the loss of p15 (Gronbaek et al, 1998; Pinyol et al, 1998), as well the loss of p14ARF (Pinyol et al, 2000). P53 inactivation, as a consequence of deletion or mutations, is more frequent in the blastic variant than in classic type (Greiner et al, 1996; Hernandez et al, 1996; Zoldan et al, 1996). However, as p53 is activated by ATM following DNA damages, and ATM function can be reduced in MCL patients (see below), it is likely that similar to CLL (Pettitt et al, 2001), p53 function might also be reduced due to the loss of ATM. MDM2 can be overexpressed in MCL (Moller et al, 1999). Decreased levels of p27 expression, as a result of an enhanced proteasome-mediated degradation of the protein, were shown in more than 50% of MCL cases (Chiarle et al, 2000, 2002; Izban et al, 2000).

A series of cyclin-dependent kinase inhibitors, such as flavopiridol and UCN-01, have been developed and are undergoing preclinical and clinical evaluation as single agents or in combination with chemotherapy (Shapiro, 2001; Lin et al, 2002; Kouroukis et al, 2003). The deregulation of the cell cycle in MCL and the lack of any adequate conventional treatments make them a possible tool for the treatment of this lymphoma. The increased degradation of p27 protein also suggests a role for proteasome inhibitors (Adams et al, 2000; Pham et al, 2003).

ATM and 11q deletion

Between 20% and 40% of MCL cases show a deletion of chromosome 11q22-23 (Monni et al, 1998, 1999; Stilgenbauer et al, 1999; Bentz et al, 2000; Zhu et al, 2000; Bertoni et al, 2004). A similar deletion can be found in CLL, a lymphoid neoplasm that shares some genetic similarities with MCL (Dohner et al, 1997; Zhu et al, 1999, 2000; Bentz et al, 2000; Struski et al, 2002). Two distinct minimal deleted regions have been reported in MCL: one at 11q22.2, comprised within the yeast artificial chromosome (YAC) 801e11 (Stilgenbauer et al, 1999), and another at 11q23.3, covered by YAC 755b11 (Monni et al, 1999). The ATM gene is located within the first minimal deleted region. It was shown to be completely inactivated by deletions or mutations in nine of 12 (75%) MCL patients: in all seven cases with 11q loss and in two of five (40%) with normal 11q (Schaffner et al, 2000). In another series, ATM alterations were detected in 12 of 28 (42%) (Fang et al, 2003) and in eight of 20 (40%) MCL cases (Camacho et al, 2002).

The ATM gene was first cloned and found to be mutated in patients with ataxia teleangectasia, an autosomal recessive disease characterized by cerebella ataxia, immunodeficiency, increased sensitivity to ionizing radiations and a predisposition to lymphoid tumours. T-cell neoplasms outnumber B-cell tumours in patients with ataxia teleangectasia. A similar pattern can be observed in ATM −/− knock-out mice (Liao & Van Dyke, 1999; Petiniot et al, 2000; Spring et al, 2001). ATM is an ubiquitously expressed phosphoprotein, whose main known function is to respond to DNA damage represented by double-strand breaks (Lavin et al, 1999; Khanna, 2000; Boultwood, 2001; Taylor & Stark, 2001). Both exogenous (ionizing radiations, chemotherapeutics, chemicals) and endogenous (oxidative damages, replication, programmed rearrangements, meiosis) factors can induce DNA double-strand breaks. ATM signals the presence of double-strand breaks to the cell cycle checkpoints to avoid the cell undergoing division in the presence of important DNA damages (Fig 3). While chk1 and chk2 mediate the arrest at the G2/M checkpoint, p53 links ATM to the G1/S checkpoint.

Figure 3.

Diagram of the ATM pathway.

As ATM is linked to the repair of DNA double-strand breaks due to programmed rearrangements, its loss is associated with an increased risk of chromosomal translocations during VDJ rearrangement (Liao & Van Dyke, 1999; Petiniot et al, 2000; Stankovic et al, 2002). The somatic mutation process also requires double-strand breaks (Jacobs & Bross, 2001; Kong & Maizels, 2001) and it targets not only the VDJ region but other genes as well (Kelsoe, 1999; Peng et al, 1999; Capello et al, 2000; Pasqualucci et al, 2001; Storb et al, 2001). The actual percentage of cases of MCL that bear somatic mutated Igs is still not completely defined (see below). The blastoid variant seems to bear mutated Igs more often than common MCL (Pittaluga et al, 1998; Laszlo et al, 2000), and often has a complex karyotype and a worse prognosis. It might be hypothesized that cases with MCL undergoing somatic mutations with no or reduced ATM function could present a higher risk of chromosome breakages. Of interest, cases with ATM inactivation have a higher number of chromosomal aberrations (Camacho et al, 2002).

Karyotype abnormalities

Various recurrent secondary karyotypic abnormalities, some similar to the ones occurring in CLL, have been reported in MCL (Table I) (Knuutila et al, 1998, 1999, 2000; Monni et al, 1998, 1999; Bea et al, 1999; Rosenwald et al, 1999; Stilgenbauer et al, 1999; Bentz et al, 2000; Dreyling et al, 2000; Wlodarska et al, 2000; Mabuchi et al, 2001; Martinez-Climent et al, 2001; Migliazza et al, 2001; Onciu et al, 2001; Wolf et al, 2001; Allen et al, 2002). A gain of genomic material suggests the presence of an oncogene. Conversely, DNA losses can underline the presence of tumour suppressor genes at a particular locus. The vast majority of the targeted genes are still unknown. DNA losses and, even more, gains can involve large DNA regions. Thus it is likely that more than one gene is involved in different MCL cases, according to the size of DNA deletion or losses.

Table I.  Regions of genomic DNA gains and losses in mantle cell lymphoma.
 Percentage of casesKnown and putative involved genes
 9p2115–30P16, P15, P14

The above-mentioned P16/P15/P14 and P53 are the strongest candidates for 9p21 and 17p13 respectively. A second gene might be involved with P53, as two different minimal deleted regions can be detected (Dreyling et al, 2000).

The c-MYC oncogene might be the target of 8q22-24 amplification. It is overexpressed in a subset of MCL cases (Bea et al, 1999; Hernandez et al, 1999; Hofmann et al, 2001; Martinez-Climent et al, 2001), and 8q24 rearrangements have been reported (Tirier et al, 1996; Au et al, 2000). Mouse models showed a strong synergistic transforming effect of overexpressed cyclin D1 and N- and L-MYC (Lovec et al, 1994).

BCL2 is highly expressed in MCL. However, the gene does not seem to be amplified at DNA level, even if the 18q21 region, where BCL2 is mapped, is amplified by comparative genomic hybridization (CGH) (Bea et al, 1999).

BCL6, at 3q27, does not play a relevant role in MCL pathogenesis. It is translocated in <5% of cases (Chaganti et al, 1998; Bea et al, 1999) and is the target of the Ig somatic mutation process in a low percentage of cases (Capello et al, 2000; Pasqualucci et al, 2001).

The polycomb gene BMI-1, on 10p12, is amplified in four of 36 (11%) MCL cases (Bea et al, 1999, 2001). The polycomb transcriptional repressor genes represent a highly conserved family of proteins, which, by creating heterogeneous multimeric complexes, regulate lymphoid development and contribute to cell cycle regulation (van Lohuizen, 1999). Polycomb targets are the HOX gene clusters, and probably many others, such as the P16 INK4A/P14 ARF locus. Two groups of polycomb complexes have been identified in humans: BMI-1-containing polycombs and ENX- and EED-containing polycombs. The expression of BMI-1- and EED-containing complexes is mutually exclusive in lymphoid cells (Raaphorst et al, 2000). The EED/ENX-1 is expressed in proliferating germinal centroblasts and BMI-1/RING1 in resting mantle cells and in centrocytes. Resting MCL cells express BMI-1/RING-1 (van Kemenade et al, 2001; Visser et al, 2001). Proliferating MCL cells overexpress only ENX-1 (EZH2, EZH1), not EED, and they do not down-regulate the BMI-1 complex (van Kemenade et al, 2001; Visser et al, 2001). Transgenic mice models suggest an oncogenic capacity for BMI-1 and RING1 (van Lohuizen, 1999), and EED seems to play an important role in the negative regulation of cell proliferation. Despite apparently normal levels of p16 in MCL cases with BMI-1 amplification (Bea et al, 2001), the polycomb pathway appears deeply deregulated in MCL.

Down-regulation of P16 and P14, as well as overexpression of BMI-1, were detected by a large gene expression study performed on 101 MCL RNA samples using the Lymphochip microarray (Rosenwald et al, 2003). This study also defined 42 genes that are more highly expressed in MCL than in diffuse-large B-cell lymphomas, follicular lymphomas or in CLL. An index based upon the level of expression of 20 genes involved in cell proliferation is apparently able to discriminate four groups of patients with different overall survival.

The immunoglobulin gene rearrangement

B lymphocytes undergo an extensive genomic rearrangement within their Ig loci to express functional Igs (Küppers et al, 1999; Vanasse et al, 1999; Delves & Roitt, 2000). At the early pro-B stage one of approximately 30 D segments is first joined to one of six J regions. At the late pro-B stage, the D–J unit is joined to one of approximately 200 ‘Variable’ (V) genes. A rearranged region with a DNA sequence that is unique in each mature B lymphocyte is formed. The VDJ rearrangement is a fundamental mechanism in the generation of antibody diversity. This apparently random recombination of the multiple V, D and J segments, in addition to excision of bases and insertion of non-germline ‘N’ nucleotides at the V–D and D–J junctions, generates an unlimited antibody repertoire. If the VDJ rearrangement is not functional, the cell undergoes the rearrangement of the second IgH allele. Otherwise, the heavy chain is expressed on the surface of the large pre-B cell, together with a surrogate light chain (μψL complex). The expression of μψL induces the cell to rearrange one of the Ig light chain (IgL) genes, usually starting at the κ locus, followed by κ deletion, and then the λ locus. The expression of a complete IgM molecule on the cell surface excludes further IgL rearrangements. The immature B cell leaves the bone marrow, and can migrate to the secondary lymphoid organs. In the germinal centres, under the influence of antigen-specific CD4+ T-cells and follicular dendritic cells, B cells undergo proliferation, somatic hypermutation within V regions, and isotype class switch. The accumulation of somatic mutation within VH and VL genes greatly increases the affinity of antibodies for antigens. In the isotype class switch, B cell change their constant (C) chain, replacing the Cμ or Cδ fused to the VDJ segment, with one of the Cγ, Cα or Cε isotypes, depending on the provided T-cell co-stimulatory factors.

Sequence analysis of the IgH V genes expressed by MCL report a low rate of somatic mutations, with cases of MCL with highly mutated IgH genes being uncommon (Hummel et al, 1994; Küppers et al, 1997; Camacho et al, 2003; Orchard et al, 2003; Walsh et al, 2003; Bertoni et al, 2004). MCL is thought to be derived from naïve B cells, i.e. pregerminal centre B cells not yet exposed to antigen (Fig 4). Na B cells are characterized by a lack of somatic mutations within their IgH VH genes. The presence of cases with IgH somatic mutations suggests that there is either a germinal centre phase for a subset of MCL or, as discussed by Walsh et al (2003), an extra-germinal centre IgH hypermutation process. In terms of prognosis, while CLL cases with mutated IgH genes have a significantly better prognosis than unmutated ones (Krober et al, 2002; Oscier et al, 2002; Crespo et al, 2003), mutated MCL cases, however, do not differ in outcome from IgH unmutated cases, an observation confirmed by other recent reports (Camacho et al, 2003; Orchard et al, 2003; Walsh et al, 2003; Bertoni et al, 2004).

Figure 4.

The origin of lymphoma subtypes relative to the presence of somatic hypermutations in the immunoglobulin heavy chain genes.

MCL cases show a biased usage of individual VH genes, with VH3-21 being the most common followed by VH3-23 and VH4-34 (Camacho et al, 2003; Walsh et al, 2003; Bertoni et al, 2004); VH3-21 is usually combined with a light chain containing the Vλ 3-19 (Walsh et al, 2003). VH3-21 positive MCL cases are usually unmutated, and are associated with a better prognosis. This is in contrast to CLL. Indeed, VH3-21-positive CLL cases are mutated and have a worse prognosis.

The rearranged VDJ segment is almost unique for each B-cell clone, and it is therefore a potential marker of clonality in MCL (Kurokawa et al, 1997; Bertoni et al, 1999a; Pittaluga et al, 1999; Hoeve et al, 2000; Theriault et al, 2000). PCR analysis with consensus upper primers specific for the highly conserved framework region 3 (FR3A) at the 3′-end of the VH segment and consensus lower primers to the JH is the most commonly used. The resulting PCR product has a length of approximately 60–120 bp, and is thus feasible even using DNA samples extracted from paraffin-embedded samples. As MCL Ig genes are not heavily mutated, the rate of false negativity by PCR is low.

The peculiarity that MCL more often express the IgL λ than the κ offers another useful molecular marker of clonality. IgL λ + B cells often undergo a deletional rearrangement of the IgL κ genes, due to a site-specific rearrangement of the κ deletion element (κde) with either the recombinant signal sequence (RSS) intron of the constant gene segment (Cκ), or with the RSS in the 3′-site of the V segments (Vκ) (Hieter et al, 1981; Beishuizen et al, 1997; van der Burg et al, 2001). We analysed a series of 12 cases of MCL by PCR, and a monoclonal κde rearrangement was detected in nine of ten IgLλ + cases (Bertoni et al, 1999b). Intron RSS/κde and Vκ/κ de rearrangements were present in 67% (6/9) and 33% of the cases respectively.

The expression of Igs on the cell surface of MCL cells and the absence of ongoing mutation within the cells make them good targets for vaccines. Clinical trials with autologous tumour-derived Ig idiotypes (Kwak, 2000) or with DNA vaccines encoding the tumour IgL and IgH idiotypes (Stevenson, 1999) are both running.

The MCL is an aggressive and incurable disease. Understanding the biological and genetic mechanisms underlying the bad prognosis, as well as providing insight into its pathogenesis, may also provide pointers towards potentially curable therapy for MCL. The cell cycle machinery, heavily deregulated in MCL, is already the target of molecular-based therapeutic approaches. The biased usage of particular Ig heavy-chain VH genes requires further studies to evaluate the possible therapeutic or prognostic implications.


This work was partially supported by the Swiss Cancer League, Swiss Group for Clinical Cancer Research (SAKK), Leukaemia Research Fund and Fondazione San Salvatore.