HTLV-1 Epitope (BA21) Reactivity in Rare Bone Marrow Failure Diseases

Abstract 1724

LGL leukemia is thought to be an antigen-driven disease. Extensive data support the antigen activated nature of leukemic LGL cells. Leukemic LGL constitutively express perforin and other markers of activated killer cells. T-LGL leukemia has been characterized as an accumulation of apoptosis resistant effector memory cytotoxic T lymphocytes (CTL). The actual target recognized by these activated CTL is still not characterized. LGL leukemia is at one end of a spectrum of hematologic disorders characterized by aberrant CTL responses. Conditions at the other end of this spectrum also feature CTL attacks on bone marrow, leading to cytopenia. These conditions are known as bone marrow failure (BMF) diseases and include pure red cell aplasia (PRCA), myelodysplastic syndromes (MDS), aplastic anemia (AA), and paroxysmal nocturnal hemoglobinuria (PNH). LGL leukemia is characterized by an easily detectable dominant clone whereas the CTL clones in these marrow failure diseases are less obvious. It has been postulated that exposure to infectious agents contributes to the onset of the full spectrum of LGL-driven diseases. One approach to determine exposure to infectious agents is to study the antibody patterns of infected and non-infected populations. In clinical screening, sera are usually assayed by ELISA to check for reactivity against recombinant viral proteins, followed by Western blot and PCR. In the case of human T cell leukemia viruses 1 and 2 (HTLV-1/2), less than 1% of the general U.S. population demonstrates cross-reactive antibodies. Conversely, at least 21% of LGL leukemia patients exhibit cross-reactive serology to HTLV1/2. The pattern of reactivity is very different between LGL leukemia patients and other cross-reactive groups. Comprehensive testing reveals that most LGL leukemia patients are not infected with HTLV-1/2, or with HTLV-3/4. We have demonstrated that HTLV Env reactivity in LGL leukemia was directed at the transmembrane-associated BA21 region/p21e. The BA21 epitope overlaps the immunogenic p21e region of the HTLV-1 envelope. Small scale studies indicated that at least 25% of sera from LGL leukemia patients were reactive to BA21, and that reactivity coincided with elevated overall HTLV-1 serum antibody expression. Since the presumptive antigen involved in the pathogenesis of LGL leukemia is not known, the identification of disease-specific antigens that are not cross-reactive with normal specimens is critical. To achieve this goal, a different way to look at epitope specificity is needed. We addressed this need by performing a B-cell epitope analysis for the BA21 sequence using multiple prediction databases. Antigenicity was predicted for three regions of BA21, namely the amino terminus containing EQCR, and two regions near the carboxyl terminus (PPLE, WGLN). Of these, PPLE-containing sequences were predicted to be the most immunogenic. We then used array-adapted alanine screening followed by ELISA to select disease-specific BA21 epitopes. Overall, 80 sequences were tested for specificity against alanine substitutions, using up to 10 mcL of serum per patient. The results of the array-adapted alanine screen clearly demonstrated that PPLE-containing sequences were universally recognized by LGL leukemia and normal serum groups. Conversely, peptides that included the amino terminus with sequence EQCR were better recognized against their alanine substituted sequences. ELISA testing was therefore performed for peptides from these regions. The resulting BA21 epitope was recognized by at least 40% of LGL leukemia sera but not by normal donor sera. Since a common pathogenesis is postulated for other CTL-mediated hematologic diseases, the levels of BMF-BA21 IgG were determined for participants in the Bone Marrow Failure Diseases Consortium of the Rare Diseases Clinical Research Network. We found that a substantial number of sera from participants with MDS, AA, PNH and PRCA also recognized the epitope. Reactivity with the BMF-BA21 epitope was independently associated with CTL-driven BMF diseases. These data provide further support for the hypothesis that a variety of hematologic diseases characterized by antigen-activated CTL result from a common pathogenetic mechanism. This project was supported by NIH Grant Number U54RR019397. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.