1. Introduction

Studies of the epidemiology and etiology of non-Hodgkin lymphoma (NHL) are complicated by the vast clinical and biological heterogeneity among the numerous subtypes constituting this group of malignancies. There is also evidence of etiologic heterogeneity among lymphoma subtypes, underscoring the need for subtype-specific studies in this area. Continuous changes in the classification of hematolymphoproliferative malignancies over time add to the difficulty of interpreting and comparing studies of subtype-specific incidence patterns and trends. With regard to mantle cell lymphoma (MCL), an equivalent histological entity was reported by Rappaport et al. in 1956 [1], but it was not until 1992 that MCL was recognized as a distinct clinical B-cell NHL subtype [2] that was subsequently adopted into the REAL [3] and WHO classifications [4]. Therefore, descriptive epidemiological data of MCL from the period before 1992 mainly rely on retrospective re-classifications of case series, while population-based studies are limited to the last 15–20 years. In this review, we give an overview of published data concerning MCL epidemiology and etiology available through Medline searches using combinations of the following key words: “non-Hodgkin lymphoma,” “mantle cell lymphoma,” “epidemiology,” “incidence” and “risk”.

2. Epidemiology

2.1. Incidence

NHL ranks among the ten most common malignancies and accounts for approximately 3–4% of all cancer cases worldwide [5], although it should be recognized that incidence data primarily stem from developed countries. MCL is reported to constitute between 2 and 10% of all NHL [6], [7] and [8]. Based on national registry data from the 1990s and early 2000s, the annual incidence of MCL in the US was on average 0.51–0.55/100 000 persons, age-standardized to the US population in the year 2000 (Table 1). MCL incidence is thereby in the same range as the incidence of other NHL subtypes such as marginal zone lymphoma, lymphoplasmacytic lymphoma and Burkitt lymphoma [8] and [9]. The incidence of MCL was slightly higher among whites (0.6/100 000) than among blacks and Asians (∼0.3/100 000) [7] and [9]. In Europe, the annual incidence of MCL was similarly estimated as on average 0.45/100 000 persons based on cancer registry data from 20 countries in the beginning of the 21st century [8] (Table 1).

2.2. Distribution by age and sex

MCL is typically diagnosed at ages well above 60 years. In the US from 1992 to 2004, the median age at MCL diagnosis was 67 years among men and 70 years among women [7], and very few cases were diagnosed below the age of 30 years [7] and [9]. While all NHL is more common among men than women, the magnitude of this sex difference varies by NHL subtype [10]. MCL is more than twice as common among men compared with women both in Europe (male-to-female ratio 2.3:1 [8]) and in the US (male-to-female ratio 2.5:1 [7]Fig. 1). This sex difference is among the most pronounced for B-cell NHL subtypes, exceeded mainly by hairy cell leukemia (male to female ratio ∼4:1) and Burkitt lymphoma (male to female ratio ∼3.5:1) [4], whereas the more common follicular lymphomas occur with close to similar frequency in the two sexes [8] and [9].

2.3. Time trends

Analyses of registry data have shown that the incidence of all NHL increased dramatically in many Western countries in the second half of the 20th century. In the 1970s and 1980s, the rate of the increase averaged 2–4% annually and was exceeded only by lung cancer in women and malignant melanoma in both sexes [11]. From the late 1990s onwards the rate of the increase has slowed and even leveled off in several countries [12], [13] and [14]. It has been estimated that phenomena such as diagnostic misclassification and changes in diagnostic practices and tumor classification systems account for less than half of the observed increase [15]. Despite intense research efforts [16] and [17], the remainder of the observed increase in NHL incidence is still largely unexplained and continues to mystify.

Reports of time trends in incidence of MCL are scarce. Existing US data describe a steady increase in MCL incidence from 1992 to 2004 (Fig. 1). The increase was most prominent in the first half of the studied period and was primarily observed among white men, among the elderly (above 60 years of age) and among cases diagnosed with generalized stage IV disease (Fig. 2) [7]. However, a reciprocal decrease in the incidence of chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) and unspecified lymphoid neoplasms was noted in the US from 1992 to 2001 [9]. Because the reliability of early histological coding for MCL was low in the US registry before 2001 [18], it is possible that rather than a true increase, the observed time trends of MCL incidence reflect changes in diagnostic practice, particularly with the introduction of immunohistochemical staining for cyclin D1 [9].

3. Etiology

Although much of the etiology of NHL remains to be elucidated, including the reasons behind the recent increase in incidence, new environmental and genetic factors are continuously suggested to be associated with all NHL or with one or several of its more common subtypes. More etiologic clues have been revealed for some NHL subtypes than for others. Well-established risk factors such as severe primary or acquired immune suppression (HIV/AIDS, organ transplantation, rare hereditary immunodeficiency syndromes) are primarily but not exclusively linked with the development of EBV-positive diffuse large B-cell lymphoma and Burkitt lymphoma [19] and [20], whereas the occurrence of follicular lymphoma is unrelated to these factors. Extranodal mucosa-associated marginal zone lymphomas (MALT) are strongly linked with local organ-specific inflammatory or infectious processes, as exemplified by the development of parotid gland lymphoma in patients with Sjögren's syndrome [21] and [22] and of gastric MALT in individuals with chronic Helicobacter pylori infection [23] and [24]. MCL belongs to the group of NHL subtypes for which the etiology is yet to be explained, and specific efforts to this end have so far been limited. To our knowledge, available published studies of MCL etiology have all been conducted within the framework of investigations of all NHL and major subtypes. As a consequence, these studies have not been driven by MCL-specific hypotheses, and have often been hampered by low statistical power to test MCL-specific associations.

4. Environmental factors

4.1. Immune competence and infectious agents

Returning to the strong and well-described link between immune suppression and lymphoma development, lymphomagenesis in this context is still complex and multifactorial [19] and [25]. In HIV/AIDS, risk of lymphoma is clearly connected with the level of immune dysfunction, and the introduction of anti-retroviral therapy in HIV/AIDS leading to restoration of CD4 counts has lowered the incidence of lymphoma among these patients [26]. However, the increased occurrence of lymphoma in HIV/AIDS is also related to lymphocyte-transforming properties of the retrovirus itself, and to opportunistic infections with lymphotrophic herpesviruses such as the Epstein–Barr virus (EBV) and Kaposi's sarcoma-associated herpesvirus/human herpes virus 8 (HHV-8) [27]. In the organ transplantation setting, heart and lung recipients are at a higher risk of NHL compared with kidney recipients, presumably due to the administration of immunosuppressive drugs in higher doses and with higher immunosuppressive capacity in these patients [28]. There is hitherto no data to suggest an increased risk of MCL in severely immunosuppressed individuals.

In immune-competent individuals, some viruses are also consistently associated with occurrence of specific NHL subtypes, with substantial geographic variation in their impact relating to differing prevalences of infections or of co-factors in the general population [29]. These viruses include EBV (associated with endemic Burkitt lymphoma in Africa and, to a lesser extent, sporadic Burkitt lymphoma in other places), human T-cell leukemia/lymphoma virus 1 (HTLV 1; cause of adult T-cell leukemia/lymphoma in, e.g., Japan and the Caribbean basin), hepatitis C virus (HCV; associated with low-grade B-cell NHL, in particular Waldenström macroglobulinemia [30]) and HHV-8 (associated with HHV-8-positive plasmablastic lymphoma and primary effusion lymphoma) [29]. More recently, evidence is also accumulating of a role for hepatitis B virus (HBV) in NHL, particularly diffuse large B-cell lymphoma [31]. So far, none of these viral agents have been linked with risk of MCL [32].

Bacteria involved in local lymphoma development most notably include, as mentioned above, H. pylori in gastric MALT lymphoma, but there are also reports of a role for Chlamydia psittaci in orbital lymphomas [33] and Campylobacter jejuni in immunoproliferative small intestinal disease [34]. Interestingly, infection with European strains of the spirochete Borrelia burgdorferi, in particular when manifesting as a chronic inflammation of the skin known as Acrodermatitis atrophicans, has been linked with the occurrence of cutaneous lymphoma [35], and also with nodal lymphomas, including a case of MCL [36]. In a large population-based Scandinavian case–control study, self-reported history of Borrelia and serologic evidence of past Borrelia infection were each associated with a 2–3-fold increased risk of MCL, and no association was observed with any other NHL subtype [37] (Table 2). The association with MCL remained among patients who did not recall Borrelia infection, yet tested positive for anti-Borrelia antibodies [36]. If the etiologic association is real, the chronic inflammatory state of the skin and/or other organs induced by the spirochete, may resemble the setting in which H. pylori infections may induce gastric MALT lymphoma. This putative specific association with MCL has so far not been confirmed or refuted in other analytical studies. Infection with strains of B. burgdorferi present in the US has not been associated with risk of cutaneous NHL or MCL [38].

Irrespective of the scarce evidence for associations between specific infectious agents and MCL, a role for antigenic drive (by exogenous or endogenous antigens) in the etiology of at least a subset of MCL cases is mounting and cannot be dismissed. Specifically, studies have shown that the immunoglobulin gene repertoire in MCL is restricted and features precisely targeted and probably functionally driven somatic hypermutation [39], [40], [41] and [42]. Accounting for these findings in future epidemiologic studies of the potential role for infectious agents as risk factors for MCL, may provide vital clues to MCL etiology. A clinical feature of MCL that may also have etiologic relevance is the frequent involvement of the gastrointestinal tract [43] and [44]. It is therefore tempting to speculate that risk of MCL may be related to infectious agents affecting the gastrointestinal tract, or variation in the microbial gut flora.

4.2. Autoimmune and chronic inflammatory disorders

Several autoimmune and chronic inflammatory disorders are clearly associated with an increased risk of NHL, and thus represent another important model of lymphomagenesis, namely that of local and/or systemic chronic immune stimulation without evidence of the influence of exogenous infectious agents. Sjögren's syndrome, rheumatoid arthritis and systemic lupus erythematosus (SLE) are all associated with diffuse large B-cell lymphoma, and Sjögren and SLE also with the occurrence of MALT lymphoma [45]. Celiac disease is associated with intestinal T-cell lymphoma (enteropathy-type T-cell lymphoma) and B-cell NHL. In psoriasis, there is some evidence to support a specific association with T-cell NHL [45]. In a large Scandinavian study of autoimmunity/inflammation in NHL, a statistically significantly increased risk of MCL was noted among patients with type 1 diabetes [46] (Table 2). However, diabetes was self-reported, the finding was based on small numbers, and the association was not confirmed in a larger pooled study [47].

4.3. Lifestyle and occupational factors

Several lifestyle factors have been evaluated as risk factors for NHL and subtypes in pooled analyses of case–control studies within the International Lymphoma Epidemiology (InterLymph) Consortium (http://epi.grants.cancer.gov/InterLymph), including between 150 and 400 cases MCL and considerably larger numbers of controls (Table 2). Body mass index, cigarette smoking and alcohol intake were not implicated as risk factors for MCL, but the statistical power to detect any association was limited due to the relatively small sample size of MCL, in spite of the fact that several studies were pooled for these analyses [48], [49] and [50]. Recreational exposure to ultraviolet radiation exposure has been associated with a reduced risk of all NHL in some but not all studies [51]. A similar picture of decreasing risks with increasing frequency of sun exposure was observed for MCL, but the estimates did not reach formal statistical significance (p for trend = 0.08) [51]. Atopic disease has also been observed to be associated with a reduced risk of all NHL, but not specifically of MCL [52]. Occupational exposure to certain pesticides and solvents is repeatedly implicated in the pathogenesis of B-cell NHL and evidence of such associations is considered to have strengthened in recent years [53] and [54]. In a recent large pooled case–control study within Europe with thorough assessment of occupational histories [55], exposure to several solvents, including benzene, toluene and xylene, was found to increase the risk of follicular lymphoma and CLL in particular. Risk of MCL was not assessed. However, occupational exposure to gasoline, including benzene and a large number of other chemicals, was unrelated to risk of NHL overall in a recent meta-analysis, although potential associations with particular NHL subtypes could not be evaluated [56].

5. Family history and genetic susceptibility

Family history of hematopoietic malignancies has been linked with a two-fold increased risk of MCL [57], a magnitude similar to that for several other NHL subtypes, including diffuse large B-cell and follicular lymphoma [57] and [58], but lower than that for CLL [59]. The large pooled study by Wang et al. [57] was however based on self-reports of familial malignancies among MCL cases and controls, a data source that lends some uncertainty to the true strength of the association [60]. This evidence of familial aggregation has often been taken to support the presence of inherited genetic susceptibility to lymphoma, but could also be due in part to shared environmental exposures among family members, a research area that has received little attention.

In the context of genetic susceptibility to NHL, there is little evidence of highly penetrant genetic traits in association with the disease [61] and [62]. Instead, candidate gene studies focusing on low-penetrance polymorphic variants in the risk of NHL and its subtypes have consistently revealed associations with variants in genes encoding the proinflammatory cytokines tumor necrosis factor (TNF), lymphotoxin-alpha (LTA) and interleukin-10 (IL10) [63], [64] and [65]. In particular, the TNF rs1800629 (−308G > A), LTA rs909253 (−252A > G) and IL10 rs1800890 (−3575T > A) gene variants have been associated with increased risk of diffuse large B-cell lymphoma [63] and [65]. TNF rs1800629 was associated with a 2.8-fold increased risk of MCL in a recent northern European analysis that included 120 MCL cases (OR 2.8, 95% CI 1.4–5.9 [64]), but the association was not confirmed in a large international pooled study [50]. In the Nordic study [64], IL10 rs1800890 was also associated with increased MCL risk, whereas the pooled study noted an MCL association with another IL10 variant, rs1800896 (−1082A > G) [65]. Laboratory studies have shown that the TNF rs1800629 minor allele is directly linked with elevated TNF-alpha protein levels [66], while IL10 rs1800890 indirectly affects TNF-alpha levels through decreased downregulation of TNF via IL10 [67]. Since TNF is a key activator of inflammation partly through the nuclear factor-κB pathway, and promotes cell proliferation, survival, invasion and angiogenesis [68], these findings could imply that a proinflammatory milieu is involved the pathogenesis of some NHL subtypes, including MCL.

The more agnostic and large-scale genome-wide association study approach for evaluation of genetic susceptibility to cancer has been successfully applied in CLL [61], [69] and [70] and follicular lymphoma [61] and [71]. Interestingly, strong susceptibility variants in the human leukocyte antigen (HLA) class I and II regions have been identified and replicated in follicular lymphoma [61] and [71] and familial CLL [69], and were also recently associated, although more moderately so, with risk of diffuse large B-cell lymphoma in a validation study [71]. These findings suggest an important role of antigen-presenting capacity in response to exogenous or endogenous antigens in lymphomagenesis. No statistically significant results have so far been reported for associations between gene variants in the HLA class I or II region and risk of MCL [71].

6. Conclusions and future directions

In summary, MCL is a mature B-cell neoplasm with a clear male preponderance occurring primarily in elderly patients and with a similar incidence in Europe and the US. Although there are some data to suggest that MCL incidence has been increasing over the last 20 years, this observation may also reflect an improved diagnostic awareness. The causes of MCL are unknown and MCL has not been associated with strong immunosuppression as have several other NHL subtypes. If anything, exploration of potential environmental and genetic risk factors of MCL in large studies of NHL suggests possible links with family history, inflammation/infection and antigen presentation, in line with recent immunogenetic tumor characterization studies, although many factors remain to be investigated in detail. Given the rarity of MCL, systematic international efforts are needed to assemble large enough sample sizes of patients with MCL and permit robust evaluation of potential etiologic factors in future studies.

Conflict of interest statement

The authors declare that there are no conflicts of interest.