PA was noted to have erythrodermatitis and alopecia soon after birth. These symptoms persisted and by age 2 months, lymphadenopathy, heptosplenomegaly, chronic neutropenia (0-100 cells/μl), and lymphopenia were present. Multiple hospitalizations due to infections ensued. Treatment with G-CSF(10-20 μg/kg/day) gave no definitive improvement in PMN counts (0-200 cells/μl), but precipitated thrombocytopenia. CID was diagnosed because of poor lymphocyte enumeration and proliferation studies, and absence of Ig production. Biopsy of skin, lymph node, and bone marrow (BM) revealed excessive numbers of histiocytes. On repeated studies, malignancy could not be detected and 4 separate chromosomal analyses were normal (nl). BM exhibited an apparent myeloid arrest. It was hypothesized that a serum inhibitory factor was suppressing PMN, and possibly lymphocyte, development. PA's serum enhanced patient and control lymphocyte proliferation to mitogens above that of nl serum. PA's serum inhibited both patient and control BM cell growth in soft agar progenitor assays, whereas patient colony formation was near-normal in the presence of nl serum. Serum TNF-α was 59 pg/ml (nl < 20 pg/ml) by ELISA assay. RT-PCR of RNA from BM mononuclear cells disclosed elevated levels of TNF-α mRNA. Normalizing the TNF-α levels with concurrently amplified β-actin by calculating ratios of densitometric readings gave 0.45±0.07 and 0.13±0.03 (p < 0.01) for PA and control samples, respectively. Comparisons with 20 other cytokines and growth factors revealed no differences between PA and controls. Previous studies have implicated that TNF-α is a potent inhibitor of myeloid development. We believe that this is the first report demonstrating that elevated TNF-α in the bone marrow is associated with severe neutropenia in a patient. The process leading to dysregulation of TNF-α production (from histiocytes?) could be responsible for both CID and neutropenia in PA, and perhaps plays a role in other conditions where neutropenia is a prominent feature.