Human natural killer cell expansion is regulated by thrombospondin-mediated activation of transforming growth factor-beta 1 and independent accessory cell-derived contact and soluble factors.

Natural killer cells (NK) were studied to determine factors important in their expansion. Flourescence-activated cell sorter (FACS) purified CD56+/CD3- NK cells cultured alone for 18 days in rIL-2 containing medium (1,000 U/mL) showed enhanced cytotoxicity but only minimal expansion. NK expansion was increased (12.5 +/- 1.6-fold) by coculturing NK with soluble factors produced by irradiated peripheral blood mononuclear cells (PBMNC) in which the two populations were separated by a microporous membrane. However, maximal NK expansion was always observed when NK were cocultured in direct contact with irradiated PBMNC (49.4 +/- 5.9-fold). To determine if marrow stroma, which supports differentiation of primitive NK progenitors, was a better accessory cell population than irradiated PBMNC, NK were cocultured in direct contact with primary marrow stromal layers. NK expansion with marrow stroma was similar to PBMNC. Fibroblast cell lines (M2-10B4, NRK-49F, NIH-3T3) and human umbilical vein endothelial cells (HUVEC), all homogeneous populations and devoid of monocytes, also exhibited a similar contact-dependent increase in NK expansion. Experiments were designed using fixed M2-10B4 stromal cells to separate the contact-induced proliferative stimuli from soluble factors. NK plated directly on ethanol/acetic acid-fixed M2-10B4, which leaves stromal ligands (cell membrane components and ECM) intact, resulted in increased NK expansion compared with medium alone. We further show that the combination of independent contact and soluble factors is responsible for maximal late NK expansion (days 28 through 40) but paradoxically inhibits early NK expansion (day 7). The proliferation inhibitory effects were verified by 3H-thymidine uptake and could be detected at days 2 through 6 but no longer 14 days after the initiation of the culture. We show that both laminin and thrombospondin inhibit early NK proliferation, whereas only thrombospondin was capable of also stimulating late NK expansion. The effect of thrombospondin on early NK proliferation is related to activation of transforming growth factor-beta 1 (TGF-beta) because anti-TGF-beta neutralizing antibody completely abrogated thrombospondin-mediated inhibition of early NK proliferation. Although inhibitory early in culture, active TGF-beta added only at culture initiation increases late NK expansion similar to thrombospondin. TGF-beta was not present in the thrombospondin preparation but latent TGF-beta in serum, or TGF-beta transcripts identified in IL-2-activated NK could explain paracrine or autocrine mechanisms for the regulation of NK proliferation. Finally, anti-TGF-beta neutralizing antibody only minimally affects stroma-mediated inhibition of early NK proliferation suggesting that aside from thrombospondin/TGF-beta, additional contact factors are important for the regulation of NK proliferation.