Factors associated with response to platelet transfusion following hematopoietic stem cell transplantation.

We studied 526 consecutive post-transplant platelet transfusions to determine the factors associated with response to platelet transfusion in the BMT setting. Poor responses to platelet transfusions occurred frequently, with 310 of the 484 evaluable transfusions (64%) resulting in post-infusion corrected count increments of less than 7500. Factors associated with poor response to platelet transfusion by both univariate and multivariate analysis included, (1) presence of serum lymphocytotoxic antibodies; (2) male sex; (3) body surface area greater than 1.7 m2; (4) transfusion of red cells on the day of the platelet infusion; (5) concurrent administration of steroids; (6) major ABO mismatch between the recipient and the platelet product; and (7) (among women) a history of one or more pregnancies prior to transplant. Paradoxically, a history of greater than 25 blood product exposures prior to transplant, and evidence of prior CMV infection in either the bone marrow donor or recipient were associated with higher CCIs by both univariate and multivariate analysis. Factors that showed little correlation with response to platelet transfusion included, (1) age of the infused platelet product; (2) concurrent fever; (3) recent administration of intravenous immunoglobulin; and (4) absolute neutrophil count at the time of the infusion. The factors associated with response to platelet transfusion in BMT patients appear to be different from those observed in the non-transplant setting.

Sensitization to epithelial antigens in chronic mucosal inflammatory disease: immunochemical similarity of macromolecules in human colonic epithelium with murine intestinal epithelial cell-associated components (ECAC).

In order to explore whether human intestinal epithelium contains cell-associated components that are organ-specific and therefore potentially autoantigenic, macromolecules were isolated in aqueous-soluble form from histologically normal human colonic epithelial cells, partially purified under nondenaturing conditions, then characterized biochemically and immunologically. Preparative non-SDS-polyacrylamide gel electrophoresis followed by electro-elution into a Tris-glycine buffer separated five human colonic fractions (yield 86%), all of which possessed an acidic isoelectric point (4.0-4.8) and a modest carbohydrate content (0.9-15%), resembling fractions of corresponding Rf similarly prepared from murine colonic epithelial cells. Immunological characterization suggested that two of the gel-purified fractions (B1, B2) isolated from colonic epithelium contained organ-specific determinants: (i) reactivity of B1/B2-specific immunoglobulin was not diminished by prior absorption with proteins from other epithelial surfaces (tracheal mucosal components) nor from normal serum; (ii) epitopes on ECAC B1 and B2 fractions were shared across species lines, with reactivity for the cross-reactive determinant being as much as 20-fold that of control protein; and (iii) localization studies with specific antibody showed an organ- and cell-restricted immunofluorescence pattern on freshly fixed human and murine mucosal epithelial tissue, using B1-specific monoclonal antibody. Given the reported binding of inflammatory bowel disease serum immunoglobulin and lamina propria mononuclear cells for murine ECAC determinants, our studies suggest these same antigens are present in human colonic epithelium, and are available as potential binding sites for antibody or lymphocytes participating in an autoaggressive immune response.