Severe combined immunodeficient (SCID) mice were injected intravenously with 5 x 10(6) primary bone marrow (BM) blasts from newly diagnosed patients with E2A-PBX1 fusion transcript positive t(1;19)(q23;p13) pre- B acute lymphoblastic leukemia (ALL). A marked variation existed in the pattern and extent of leukemic cell engraftment in SCID mice challenged with t(1;19) pre-B ALL blasts. Blasts from some patients caused disseminated leukemia that was detected by histopathology and/or flow cytometry, whereas blasts from other patients produced occult leukemia that was only detected by flow cytometry and/or polymerase-chain reaction. Notably, the ability of primary t(1;19) pre-B ALL blasts to cause disseminated leukemia in SCID mice was associated with poor prognosis. Six of six patients whose blasts caused disseminated leukemia in SCID mice relapsed at a median of 7.8 months (range: 5.7 to 25.2 months). In contrast, the remaining four patients whose blasts did not engraft or only partially engrafted remain in complete remission at 28 to 47 months. A new E2A-PBX-1 fusion transcript positive t(1;19) pre- B ALL cell line (designated LC1;19) with the composite immunophenotype CD7-CD10+CD19+CD45-HLA-DR+C mu+ was established by expanding BM blasts from a SCID mouse, which died of human t(1;19) ALL at 7 weeks after inoculation of primary leukemic blasts from a t(1;19) ALL patient. This cell line caused disseminated and invariably fatal leukemia when greater than 10(4) cells were injected intravenously into SCID mice. Total body irradiation followed by syngeneic BM transplantation (BMT) showed limited efficacy against LC1;19 leukemia in SCID mice. To our knowledge, this study is the first to (1) examine the in vivo growth of primary t(1;19) pre-B ALL blasts in SCID mice and (2) show that leukemic blasts from a majority of newly diagnosed t(1;19) pre-B ALL patients cause disseminated human leukemia in SCID mice. Our results indicate that t(1;19) pre-B ALL is biologically heterogeneous with regard to its in vivo growth pattern in SCID mice, a feature that may be predictive of prognosis. The described LC1;19 SCID mouse model may prove particularly useful for designing more effective treatment strategies against poor-prognosis t(1;19) ALL.