Human leukocyte antigens (HLA) class I molecules restrict the interaction between cytotoxic T cells and target cells. Abnormalities in HLA class I antigen expression and/or function may provide tumor cells with a mechanism for escaping immune surveillance and resisting T cell-based immunotherapies. The potential for applying T cell-based immunotherapy in the treatment of acute myeloid leukemia (AML) has stimulated interest in analyzing HLA class I antigen expression on leukemic blasts in this disease. Little information is available in the literature. We have analyzed HLA class I antigen expression on bone marrow samples from 25 newly diagnosed AML patients by indirect immunofluorescence staining with monoclonal antibodies. Five of these patients were also studied at relapse. Leukemic blasts were resolved from normal lymphocytes by staining with anti-CD45 antibody; CD45 expression is dim on leukemia cells, but bright on lymphocytes. HLA class I antigen expression was higher on leukemic blasts than on autologous lymphocytes in all but one case. Moreover, there was no significant change in HLA class I antigen expression at relapse. These results suggest that abnormalities in HLA class I antigens are infrequent in AML and should not represent a major obstacle to the application of T cell-based immunotherapies in this disease. Leukemia (2001) 15, 128-133.

These data suggest that c-mpl expression is an adverse prognostic factor for treatment outcome in adult AML that must be considered in the analysis of clinical studies using thrombopoietin in AML.

Preparation of membrane pellet Tumour specimens removed from -70'C were allowed to thaw on ice, while those from sucrose/glycerol were re-hydrated in homogenising buffer (see below). The tumour specimen was dried with tissue paper to remove any excess water or buffer. Specimens were washed in ice cold saline. The tumour was bisected and two separate samples of tumour were cut (2-3 mm minimum) from either half, one piece placed in formal saline for pathological analysis and the other in sucrose/glycerol buffer for later immunohistochemical analysis (to be reported later). The remainder of the tumour was used for the biochemical assay. Fresh homogenising buffer was prepared (20 mM Hepes, 2 mM EDTA, 0.5 mM PMSF to pH 7.4). Tumour sections were then cut into small 1 mm blocks and weighed (usually 1 gram), then placed in a centrifuge tube to which was added 5 ml g-' (wet weight) of ice cold homogenising buffer. Tumour was homogenised on ice with an ultra turrax (Janke and Kunkel) with 2 x 15 s bursts at maximum speed but allowing the homogenate to cool between bursts. The homogenate was centrifuged at 1,000 g for 10 min. The resulting supernatant was centrifuged at 12,000 g for 1 h. The nuclear pellet was resuspended in homogenising buffer and stored at -20C until required for DNA analysis (Modified Burton). The pellet from the high speed spin was resuspended in 1-2 ml of radioimmunoassay buffer (RIA buffer: 0.2 M Na2HPO4, 0.2 M NaH2PO4, 0.1% sodium azide, 0.15 M sodium chloride, 0.1 M EDTA and 0.5% bovine serum albumin to pH 7.4) depending on the initial wet weight (1 ml 500 mg-') and stored at -20C in 1.5 ml polystyrene tubes (Eppendorfs). Prior to storage each suspension was submitted to glass teflon homogenisation to ensure an even suspension.

Topical CTC-96 applications were at least as effective as Viroptic in diminishing disease signs and corneal surface virus at concentrations less than one-thousandth that of Viroptic.

Flow cytometric cell sorting is commonly used to obtain purified subpopulations of cells for use in in vitro and in vivo assays. This can be time-consuming if the subpopulations of interest represent very low percentages of the cell suspension under study. Often the desired subpopulations are identified by twocolor immunofluorescence staining. Generally, cell sorting is performed with a flow cytometer configured to trigger on light scatter signals, then sort windows are set based upon the signals from both fluorescent markers. We demonstrate that triggering the cytometer with the fluorescence signal from antibody staining common to both of the desired subpopulations, then sorting the subpopulations based upon staining of a second marker, substantially increases the speed of cell sorting vis-a-vis traditional methods. This is because undesired events are not analysed, allowing an increase in the throughput rate. While desired subpopulations of cells can be obtained by this method, undesired (i.e., nonstaining) cell "contaminants" increase and may require a second sort. The combined time for the initial enrichment sort and a second sort can be less than sorting once using standard methodology. Alternatively, the degree of contamination may be controlled by adjusting the concentration of the cell suspension and by the sample flow rate.