The relationships between dihydrofolate reductase (DHFR) levels or methotrexate membrane transport and acute lymphoblastic leukemia (ALL) immunophenotype were evaluated on 51 T-cell and 44 B-precursor ALL specimens from 90 pediatric ALL patients at diagnosis and relapse, using a fluorescent methotrexate analog (PT430) and flow cytometry assay (Matherly et al, Blood 85:500, 1995). For T-cell ALL, 35 of 45 (78%) of newly diagnosed patients' specimens exhibited elevated DHFR relative to DHFR levels in ALL blasts from methotrexate-responsive patients. For 30 of 45 diagnostic T-ALL specimens, DHFR expression was heterogeneous, with up to 3 separate subpopulations covering a 44-fold range; the DHFR-overproducing fractions comprised 10% to 88% of the total blasts. Elevated DHFR was less common in B-precursor ALL at diagnosis, being detected in only 17 of 36 specimens (47%); 11 of these samples exhibited DHFR heterogeneity. Median maximal DHFR levels were fourfold higher in T-cell than B-precursor ALL at diagnosis. Within a particular phenotypic group, there were no correlations between DHFR levels and patient prognostic features, including age, sex, chromosomal abnormalities, white blood cell counts (WBCs), and percentage of S-phase. However, for B-precursor ALL, there was a notably higher fraction of African-American than white patients with elevated DHFR. For patients (both phenotypes) with low WBCs (<50,000/ microL), event-free survival times were significantly shorter for those expressing DHFR above a threshold level than for patients expressing DHFR below this level (P < .016); this relationship was not seen for patients with high WBCs (>50,000/microL). Elevated DHFR was detected in 11 of 14 relapse specimens (5 of 6 T-cell and 6 of 8 B-precursor). Two of five paired relapse specimens (both T-cell) from patients treated with methotrexate exhibited increased DHFR levels over those at diagnosis (2.5- to 5-fold); the fraction of DHFR-overproducing blasts was also increased in 4 of 5 paired relapse specimens (2 B-precursor and 2 T-cell). In contrast to the variations in DHFR, highly impaired methotrexate transport was detected in only 6 of 95 ALL specimens, including both diagnosis and relapse. There was no correlation between phenotype and impaired transport. These data provide further rationale for the use of mechanistically based prognostic factors to complement known biologic or disease-based prognostic indicators in the design of ALL therapy including methotrexate, particularly with patients presenting with low WBCs. The finding of a markedly increased frequency of DHFR overexpression in T-cell over B-precursor ALL suggests that this difference is associated with the poorer prognosis of patients with T-cell ALL treated with standard-dose antimetabolite therapy and implies that higher-dose methotrexate (> or = 1 g/m2) during consolidation therapy may be useful in the treatment of this disease.