We report the case of relapsed AITL, which, in addition to the “classical” follicular helper T-cell (TFH) phenotype , displayed a striking expression of CD20. Whilst expression of CD20 has been reported in T-cell lymphomas , this case is one of only very few published observations of CD20 expression in AITL and is unique in that the CD20 expression was acquired upon relapse. The first case of CD20-positive AITL was reported by Yokose et al. as peripheral T-cell lymphoma with clinical characteristics resembling angioimmunoblastic lymphadenopathy . The second report was by Tachibana et al. in 2011  with features very similar to the herein presented case, in particular also the acquisition of CD20 expression upon disease progression. Only recently another case was observed by Foukas et al. .
Remarkable in our case is the clear-cut positivity of the neoplastic T-cells for CD20, a B-cell marker, and the progression from classical AITL in 2004 to tumour-cell rich AITL in 2011. “Lineage infidelity” of phenotypic markers is a well-documented phenomenon in lymphomas [10–13]. Although the biological mechanisms and significance of this characteristic are poorly understood, it has the potential to cause diagnostic havoc. Therefore a multimodal approach that considers morphology, pheno- and genotypic characteristics is essential to achieve a final correct diagnosis. Hypothetically, CD20 positivity in T-cell lymphomas may include derivation from subsets of CD20 positive T-cells undergoing neoplastic transformation or CD20-acquisition following neoplastic transformation of the T-cells; the latter applying to our case, where CD20 was acquired upon relapse. Indeed, CD20 expression may be acquired in T-cell lymphomas following activation of the T-cells, as has been demonstrated in stimulated lymph nodes from monkeys with simian immunodeficiency virus . This experimental evidence is supported by the current case as well as the series reported by Tachibana, Rahemtullah and colleagues [5, 8] where the proportion of CD20/CD30 co-expressing T-cells increased over time. CD20 may therefore represent an “activation marker” acquired after neoplastic T-cell transformation. Another possible explanation for the observed CD20 acquisition may be trogocytosis . Upon interaction with surrounding cells, lymphoid cells, especially CD8 positive cytotoxic lymphocytes and NK-cells, can initiate membrane bridges with target cells, thereby capturing small membrane patches from their interaction partners and in the process potentially acquiring “lineage-foreign” antigens . Since at least in our case there were no morphological hallmarks of cell cannibalism or signs of hemophagocytosis, the AITL was of CD4 lineage and CD20 was the single “lineage-improper” antigen expressed by the malignant cells, we speculate that trogocytosis might not explain our observations.
Of clinical relevance is the possibility of targeting such cases with Rituximab, a chimeric murine/human monoclonal antibody directed against the CD20 antigen. Currently, our patient was treated (off-study) with high dose chemotherapy combining cisplatin with cytosine arabinoside and dexamethasone with addition of Rituximab (R-DHAP). This treatment regimen lead to a partial remission and the patient subsequently underwent autologous stem cell transplantation. However, the disease relapsed only two months later and he succumbed to sepsis in neutropenia upon salvage immunotherapy with Lenalidomide in preparation for allogeneic stem cell transplantation.
Treatment strategies incorporating the anti-CD20 antibody Rituximab have significantly improved results in patients with mature B-cell lymphomas . And whilst Rituximab has been shown to suppress EBV-positive B-immunoblasts in the microenvironment of AITL , a recent small clinical trial revealed no clear survival benefit of adding Rituximab to conventional CHOP chemotherapy to target the intratumoral B-cells .
Potential diagnostic difficulties and pitfalls in this case were not only apparent because of the aberrant CD20 expression but also because of the unusual density of the tumour cell infiltrate, resulting in morphological overlap with peripheral T-cell lymphoma not otherwise specified (PTCL, NOS). An important clue to the correct diagnosis of AITL – also in the initial biopsy – was the expanded mesh works of follicular dendritic cells, highlighted by the CD23 stain and the characteristic phenotype with expression of TFH markers. The differentiation between relapsed AITL with aberrant CD20 expression and AITL relapse as frank B-cell lymphoma, a particular possibility in such instances, can be accomplished by testing for and comparing T- and B-cell clonality both in the initial sample and upon relapse as well as by evaluation of the immunohistochemical profile of the neoplastic cells in serial sections.
Taken together this case highlights the spectrum by which AITL can present. From the “classical” form readily suspected in haematoxylin and eosin stains with a dispersed tumour-cell infiltrate sparing peripheral cortical sinuses, with expanded follicular dendritic cell mesh works and abundance of high endothelial vascular structures to the more tricky form mimicking PTCL, NOS with dense tumour-cell infiltrates, aberrant marker expression and low numbers of EBV-positive B-cells. Correct classification of peripheral T-cell lymphomas may become increasingly important and in particular the diagnosis of PTCL, NOS should not be used as a “waste basket” for tricky cases. Increasingly molecular profiling provides evidence that different pathways are primarily involved in different subsets of T-cell lymphomas. Indeed, as is the case with AITL, in which activation of the NF-κB pathway appears to be of importance , such new insights may even lead to more tailored treatment options in the future, including the use of inhibitors of the NF-κB pathway in AITL.