Evidence for molecular recognition of cells in mitosis by human NK cells. (A-H) The percent of 721.221 or Daudi cells in mitosis or other cell-cycle stages bound by NK cells after 10 minutes of coincubation, determined by confocal microscopy of cells stained for α-tubulin. (A) Target cells (721.221) were pretreated with different glycosaminoglycan-degrading enzymes: S hyalurolyticus hyaluronidase, which catalyzes degradation of hyaluronan; chondroitinase ABC, which degrades chondroitin sulfate; or heparinase I, which degrades heparin and to a lesser extent heparan sulfate. (B) NK cells were incubated with the different glycosaminoglycans indicated to test their ability to block specific recognition of target cells in mitosis. (C) NK cells were preincubated with an anti-CD44 mAb, which blocks the interaction between CD44 and hyaluronan or with a control isotype-matched anti-CD44 antibody that does not block this interaction. (D-E) NK cells were preincubated with monoclonal or polyclonal antibodies recognizing NKp46 or NKG2D, or isotype-matched control antibodies (i.c.), prior to coincubation with (D) 721.221 or (E) Daudi target cells. (F) NK-cell clones characterized by flow cytometry as having low (NKp46^low; inset, dotted line) or high (NKp46^high; inset, solid line) surface expression of NKp46 (inset: shaded histogram represents isotype-matched control staining) were compared for their capacity to bind 721.221 cells in mitosis or 721.221 in other stages of the cell cycle. (G) The binding of several NKp46^high and NKp46^low NK-cell clones to 721.221 cells is shown as the percentage increase in binding to cells in mitosis compared to cells in other stages of the cell cycle. (H) NK cells were preincubated with anti-NKp44 or control mAb. Data from at least 3 independent experiments are shown, and the total numbers of cells are indicated. *P <.05, **P <.005; ns indicates not significant.