Burkitt's Lymphoma - A Pathologist's Perspective

Cytological appearance of Burkitt's lymphoma. These tumor cells were obtained from a simple fine-needled aspirate of a jaw tumor in an African child. The white spots are tiny fat globules, or "lipid vacuoles," which are characteristic of this tumor.

Denis Burkitt wrote his first paper on the lymphoma that bears his name in 1958. The feature that drew his attention to the tumor was the characteristic, and often dramatic, involvement of the jaws. The first of his many contributions to the study of this tumor was to show that patients with jaw tumor also had visceral tumors and that similar visceral tumors were seen in patients without jaw involvement.

Burkitt's lymphoma (BL) has a characteristic cytomorphology. Labeling tumor cells in vitro with tritiated thymidine showed a proliferation rate of 100%, a figure subsequently confirmed by labeling with proliferation markers such as Ki67. An international meeting of hematopathologists in 1969, sponsored by the World Health Organization and the International Agency for Research on Cancer, expressed a majority view that BL should be defined on the basis of its cytomorphology. Two members of the group, however, were of the opinion that the tumor should be defined on the basis of its clinico-pathological features and they rejected the view that BL exists in a specific histological and cytological sense. More than 30 years later, it is apparent that the minority group were at least partly correct, insofar that BL has different clinical features in different world regions, is also predisposed by immuodeficiency syndromes and HIV infection, and may, in fact, be a family of closely related tumors.

BL, as defined by cytology and histology, falls into at least three categories. Endemic BL, as described by Burkitt, shows age-related jaw tumors and visceral tumors involving the kidneys, liver, endocrine organs, gonads, breasts and gastro-intestinal system. Involvement of peripheral lymph nodes is unusual. The tumor occurs in children, with a peak age incidence of seven years, and occasionally in young adults.

Sporadic BL, as seen in Europe and the USA, typically involves the abdomen, and has a particular penchant for the terminal ileum and ascending colon. A number of patients present with massive intra-abdominal tumors and the exact origin of these tumors may be difficult to ascertain. Peripheral lymph node involvement is more common, and bone marrow involvement not infrequent, particularly at relapse. This is often referred to as acute B cell leukemia, but patients with marrow involvement respond well to the same therapy used for BL and poorly to acute lymphoblastic leukemia therapy. BL is the most common non-Hodgkins lymphoma of children in the developed world.

A third category of BL is AIDS related BL, a tumor that occurs in HIV-positive individuals, usually early in the course of their disease prior to significant immuo-suppression. It presents more often with peripheral lymphadenopathy and often involves unusual sites. Interestingly, in spite of the high prevalence of HIV in Africa, there is no evidence that this has influenced the incidence of BL. In contrast, the incidence of Kaposi's sarcoma has been dramatically increased.

It can be seen that the eponymous title "BL" has been applied to two tumors that are clinically different from the tumors described by Denis Burkitt in Africa. The reason why these tumors are morphologically identical is presumably because they have translocations involving one of the immunoglobulin genes and the c-myc oncogene. This results in c-myc deregulation and consequent unrestrained cell proliferation without differentiation. The resulting B-blasts therefore look identical.

The Epstein-Barr virus status of the three varieties of BL is also different. Endemic BL is 100% EBV positive; rare EBV negative cases have been described from Africa but it is not clear whether these are endemic cases. The EBV status of sporadic BL varies with the childhood prevalence of EBV infection in the community from which the cases are derived. In Europe and North America, where EBV infection occurs later in childhood, it is in the region of 20%, in north Africa and Iran, where EBV infection occurs earlier, it is between 70-90%. Paradoxically, only 20 to 30% of AIDS-related BL are EBV associated, although almost all patients have antibodies against EBV.

The terms endemic and sporadic BL are not entirely appropriate since rare cases with the typical clinical features of endemic BL are seen in Europe and the USA, and it is possible that a small fraction of African cases correspond to sporadic BL. In addition, it is not clear how these terms should be applied in other world regions which sometimes have clinical patterns intermediate between the two (e.g. some parts of South America and Turkey). Basing distinctions on clinical features alone is imprecise at best.

The different features of the subtypes of BL probably reflect origins from different types of B-cells and possibly differences in the pathogenesis of the tumors. Studies of cases of BL from the American BL Registry, mainly of sporadic type, showed a transition between reactive follicles and tumor, suggesting an origin from follicle centre cells. This author has suggested that endemic BL may be derived from marginal zone B-cell and therefore is a type of MALT lymphoma. The evidence for this is circumstantial in that the tumor frequently involves mucosal sites, does not usually affect peripheral lymph nodes and shows relative sparing of the bone marrow and spleen. Mucosal B-lymphocytes migrate to the breasts during late pregnancy and lactation, which might provide an explanation for why massive breast involvement by tumor is seen in young women who present with endemic BL during pregnancy and lactation. Denis Burkitt recorded cases in which there was spontaneous regression of breast tumors on cessation of lactation.

If endemic BL is derived from mucosal lymphocytes, how can we account for the jaw tumors that are such a characteristic feature of this tumor? Involvement of the jaws greatly exceeds involvement of any other part of the skeleton. Jaw tumors are age- related. All children aged three with BL seen in Uganda in the 1960s had jaw tumors, whereas by the age of 15 only 10% had jaw involvement. The age incidence of jaw tumors coincides with the period of maximum dental development. It also coincides with the presence of cellular bone marrow in the mandible. Recent studies using magnetic resonance imaging have shown that cellular marrow in the mandible retreats toward the condyle and is largely replaced by fatty marrow in late childhood.

Jaw tumors in endemic BL usually involve more than one quadrant of the jaw and involvement of all four quadrants is not uncommon. The tumors are clonally related, precluding independent origin in each quadrant, and suggesting that the tumor cells specifically home to and proliferate at this site. A few years ago Dr Jayola Thomas and I conducted a histological study on segments of mandible obtained at postmortem examination on children from Ibadan, Nigeria. Most of these children had died of infection or trauma. We identified lymphoid aggregates in relation to the unerupted teeth, at the site where the earliest jaw tumors are detected by radiology. This finding raises many questions. Are these aggregates of mucosal lymphoid tissue? The dental epithelium is a down-growth from the oral epithelium and therefore of mucosal origin. Is this lymphoid tissue physiological or is it induced by an infectious agent, possibly EBV? Unfortunately the jaw marrow is not the easiest tissue to study but it would be of interest to know whether this lymphoid tissue is found in the jaws of children from other areas of the world. It is intriguing to consider that if these lymphoid aggregates are caused by an infectious agent much more common in African children, and if they also account for the high frequency of jaw involvement in Africa, then environmental factors, like host factors (e.g. the lactating breast), may have a role in determining the clinical sites of disease.

submitted by Dennis Wright
Professor of Pathology Emeritus
The University of Southampton