—  SYMPOSIUM #20  —

Hodgkin Lymphoma: Diagnostic and Biological Insights
Moderators: Dr. Philippe Gaulard and Dr. Nancy Lee Harris

Section 4 - Mixed Cellularity and Lymphocyte Depleted Hodgkin Lymphoma: Athology, Epidemiology, Clinical Features, and the Role of Epstein-Barr Virus

Pierre Brousset, MD, PhD
Department of Pathology
Purpan Hospital
Toulouse, France


1) Morphology and Immunophenotype of the Histopathologic Subtypes of Classical Hodgkin Lymphoma
Definition: A neoplasm composed of mononuclear Hodgkin cells and multinucleated Reed-Sternberg cells in a background containing a variable mixture of reactive small lymphocytes, eosinophils, neutrophils, histiocytes, plasma cells, fibroblasts and collagen fibres. Some Hodgkin and Reed-Sternberg cells may have a condensed cytoplasm and pyknotic nuclei (mummified cells). The Hodgkin and Reed-Sternberg cells represent only a minority of the cellular infiltrate with a frequency ranging from 1 % to 10 %.

Four histological subtypes have been distinguished: nodular sclerosing HL, mixed cellular HL, lymphocyte rich classical HL, and lymphocyte depleted HL. The immunophenotype of the mononuclear Hodgkin cells and multinucleated Reed-Sternberg cells is identical in these histological subcategories. They are positive for CD30 in nearly all cases, and for CD15 in the majority of cases (80% des cas) (Chittal et al.). They are usually negative for CD45 but the staining is difficult to interpret because of the surrounding small lymphocytes. In approximately 30 % of cases, B cell antigens such CD79a and/or CD20 are detected. However, their expression is extremely variable with some Hodgkin and Reed-Sternberg cells being clearly positive and others being completely negative. Transcription factors involved in B-cell differentiation can be used to phenotype HL (Browne et al.). Thus, PAX5+/Oct-2-/BOB.1- phenotype is in favor of classical HL whereas PAX5+/Oct-2+/BOB.1+ is predictive of Nodular LPHL. Expression of T cell antigen by a minority of Hodgkin and Reed-Sternberg cells may be encountered in some cases. Expression of epithelial membrane (EMA) is rare (less than 5 % of the cases) and is usually weak. In 50% to 70% of cases, neoplastic cells are latently infected by EBV and positive for latent membrane protein 1 (LMP-1) and negative for EBNA2 (latency type 2). Most Hodgkin and Reed-Sternberg cells express the proliferation-associated nuclear antigen Ki-67. Reed-Sternberg cells express several types of cytokines and interleukins (IL1a, IL5, IL6, TGFβ…) which influence greatly the microenvironment of the tumor cells. For example, in cases of HL with eosinophilia, Reed-Sternberg cells strongly express IL5.

Hodgkin and Reed-Sternberg cells contain monoclonal immunoglobulin (Ig) gene rearrangements in most cases. There is a high load of somatic mutations in the variable region of the Ig heavy chain genes (VH) and Ig mRNA transcripts are absent (Kuppers et al).

Clinical features: Classical Hodgkin lymphoma most often presents with involved lymph nodes of the cervical region (75 % of cases) followed by the axillary and inguinal regions. Primary extranodal involvement is rare. 55 % of patients have localized disease (stage I and II). Approx. 60 % of patients (the majority of them with nodular sclerosing Hodgkin lymphoma) have large mediastinal masses. Systemic symptoms, present in approx. 25 % of patients, consist of fever, drenching night sweats, and significant body weight loss. There is a bimodal age specific incidence curve with a peak at 15 to 40 years and a second peak late in life. Without therapy the clinical course is moderately aggressive. With modern therapy, 70 to 80 % of patients show long term survival. Before the progresses of polychemotherapy there was a strong association between histologic type and overall survival. An excellent response to modern therapy has greatly diminished the differences.

A) Mixed Cellularity Hodgkin Lymphoma (MCHL)
This entity corresponds to 20-30% of all cases of classical HL and predominates in male (SR:3:1) and in HIV-positive patients. MCHL is diagnosed frequently in an advanced stage with clinical and biological B and b symptoms respectively. Contrary to NS, MCHL does not involve the mediastinum but more frequently infiltrates the spleen (25%) and the bone marrow (5-10%). The incidence of bone marrow involvement reaches 15-30% in HIV+ patients (WHO).

Morphologically, MCHL is more often diagnosed on a resected lymph node the architecture of which is destroyed by a granulomatous proliferation. In some cases, residual lymphoid follicles may persist but in general, the whole lymph node is involved. The background granuloma is composed of an admixture of different cell types is various proportions (lymphocytes, plasma cells, neutrophils, eosinophils, histiocytes nore or less epithelioid). One of this cell types may predominate, for instance eosinophils in case of interleukin-5 production by the tumor cells. In other cases, epithelioid cells may predominate, giving rise to tuberculoid granulomas. Contrary to NSHL, MCHL does not present sclero-hyalinosis and lacunar cells but rather diagnostic Reed-Sternberg cells and Hodgkin variants. The latter cells may be associated to areas of necrosis or apoptosis. In some cases, mummified (mummy) cells are frequent. . In some cases, there is a nodular pattern due to the presence of B-cell follicles and a preferential lymphoma involvement in the interfollicular regions (so-called interfollicular classic HL initially described by Dorfman). The phenotype of RS and H cells is that of lacunar cells (pan B+/-, pan T-, CD30+, CD15+, EMA-/+, ALK-, PAX5+) (WHO, Zukerberg et al.). In case of expression of CD20 (L26), one important sign in keeping with the diagnostic of HL is the heterogeneity of the staining. There is a kind of gradient from cells clearly positive to cells completely unstained.

B) Lymphocyte Depleted Hodgkin Lymphoma (LDHL)
Synonyms: Jackson and Parker:Sarcoma; Lukes and Butler: Diffuse Fibrosis and Reticular; Rye : Lymphocytic Depletion; REAL: Lymphocyte Depletion.

The appearance is highly variable, including the degree of lymphocyte depletion. One pattern may resemble mixed cellularity with increased numbers of Hodgkin and Reed-Sternberg cells. In another pattern, pleomorphic mononuclear Hodgkin cells predominate, producing a sarcomatous appearance. These cases may be difficult to differentiate from anaplastic forms of large cell non-Hodgkin's lymphoma. Another pattern is characterized by a diffuse fibrosis or proliferation of fibroblasts and/or myofibroblasts with only a few Hodgkin and Reed-Sternberg cells. The morphologic diagnosis of lymphocyte depleted (LD) subtype is now rarely made (1-5% of HL). There is no difference in phenotype with NS or MC HL.

Some cases previously diagnosed as lymphocyte depletion are now recognized as anaplastic large cell lymphoma. This form of HL tends to affect elderly patients with systemic symptoms and/or with involvement retroperitoneal lymph nodes, and bone marrow. The clinical course is aggressive.

2) Detection of Epstein-Barr Virus in HL
The incidence of EBV infection in HL varies with histological subtypes. Virtually all cases of LPHL are negative for EBV. The most frequent incidence is observed in MC subtypes with 60-70% of the cases being infected. The incidence in NS subtypes varies significantly with the different reports between 10 to 40% of the cases (Delsol et al.). Classically, the mediastinal forms of NS HL are EBV-negative. The incidence of EBV infection in LRHL looks weak. Of note, virtually all cases of HL in HIV-positive patients are associated with EBV. This finding is also noted in childhood cases from developing countries. In all cases of EBV+ HL, the detection of a clonal population of episomes indicates that only one type of EBV is present within RS cells.

Eleven EBV genes are expressed in latent infection. Some proteins encoded by these genes can be detected by monoclonal antibodies such as those directed against EBV LMP1 and EBNA2. There is usually a close correlation between the results obtained with in situ hybridization using EBER oligonucleotides and with immunohistochemistry with anti-LMP1 antibody but the former method is more sensitive. In addition to LMP1, six nuclear proteins (EBNA-1, 2, 3A, 3B, 3C or LP) are usually expressed in latently infected B-cells but not in RS cells. Latency type II infection (EBNA1+, EBNA2-, LMP1-) is the usual finding in RS cells, similar to that observed in nasopharyngeal carcinoma and nasal T/NK cell lymphoma (Delsol et al).

On the basis of EBV findings, three groups of HL can be distinguished. In the first (50% to 60%), EBV is present in H-RS cells and commonly found in scarce infiltrating small lymphocytes («reservoir lymphocytes» whichare present in all individual who have been infected by EBV. In the second (10 %), only EBV-positive small lymphocytes are detectable (Meggetto et al.). In the third group (40 %), EBV is detected neither in H-RS cells nor in small lymphocytes. Comparative analysis of genome sequences between RS cells and bystander lymphocytes indicates that the strains are distinct, in particular when looking at the polymorphism of the carboxy terminal region of LMP1. However, several experiments indicate that, compared to EBV strains observed in lymphocytes from non-neoplastic lesions, the genome sequence of EBV is strongly divergent. In fact, viral strains detected in HL tissues including RS cells and bystander B-lymphocytes, are infected by different although related EBV strains, but are four times more polymorphic than EBV strains infecting bystander B-lymphocytes of reactive lymph nodes (Faumont et al.). The biological significance of these observations remains open, in particular regarding the origin and chronology of multiple infections in the same patient.

3) Clinical Implications of EBV Findings in HL
So far, the several studies dealing with the clinical and prognostic implications of EBV findings have failed to yield clear definitive data. Recently, two conflicting studies have been published in Blood. The first study by Jarrett et al. conducting a Scottish study, demonstrated an impact of the presence of EBV in HL on presenting features and outcome in age-defined subgroups of patients. In response to this study, Herling et al from the MD Anderson in Houston reported negative results i.e; that EBV has no clear impact in the outcome, even after 10 years of follow up. One of the main arguments developed in the latter report is the heterogeneity of the treatments received by the patients.

Genomic alterations (deletions and point mutations near the 3'end) of the LMP1 gene were found to be clustered in about 10% of cases (Knecht et al.). Such molecular abnormalities were suspected to be associated with an aggressive behaviour but to date larger series of cases based on this parameter have not been investigated.

There are some points which are now accepted, in particular that the EBV genome persists throughout the evolution of the disease. In 12 HL cases with relapses (range 14-126 months), there were seven EBV positive cases (MC= 6; NS=1) and five EBV-negative cases (MC=2; NS=2). Close agreement was found between the detection of EBV in Reed-Sternberg cells at the time of the first diagnosis and at relapses in all cases tested. These results suggest that the virus persists throughout the course of the disease. In addition, the molecular investigations performed in two EBV-positive cases, both at the onset and at relapses showed that the size of terminal repeat fragments were the same. LMP1 sequence analysis further confirmed the persistence of a distinctive viral strain in each of the 2 cases with point mutation and/or 30bp deletion of the LMP1 gene. The persistence of a given strain in early and late relapses supports the view that in Hodgkin's disease, relapses are related to a single residual tumor cell clone.

4) Further Readings
  1. Browne P. Am J Clin Pathol. 2003;120:767-77

  2. Butler JJ. Semin Diagn Pathol 1992;9:252-256

  3. Chittal S, et al. Am J Surg Pathol 1988;12: 9-21.

  4. Delsol G et al. Am J Pathol 1992;140:247-253.

  5. Faumont N, et al. Virus Res. 2004;101:163-73.

  6. Herling M, et al. Blood. 2006 ;107:1240.

  7. Jarrett RF, et al. Blood. 2005;106:2444-51.

  8. Knecht H. Oncology. 2001;60:289-302

  9. Kuppers R., Adv. Cancer Res. 2002; 84: 277–312.

  10. Meggetto F, et al. J Virol. 1997;71:2547-9.

  11. Vassallo J, et al. Am J Surg Pathol. 2006;30:223-9.

  12. World Hearth Organization classification of tumours. Pathology and genetics. Tumours of haematopoietic and lymphoid tissues. Jaffe ES, Harris NL, Stein H, Vardiman JW eds. IARC Press, Lyon 2001.

  13. Zukerberg LR et al. Am J Pathol 1991;139:475-483.