To produce monoclonal antibody (mAb) specifically against human thrombomodulin (hTM), an immune-tolerizing procedure was employed to generate monoclonal antibodies specific to hTM. Female BALB/c mice were first immunized with CHO cells following at 10 min, 24 h, 48 h by intraperitoneal injection of different doses of cyclophosphamide (CP) 2 times at an interval of 2 weeks, thereby tolerizing the mice to common epitopes shared between CHO and CHO-TM5 cells. Subsequently the selected mice with the lowest titer of serum polyclonal antibody by cellular enzyme-linked immunoabsorbent assay (CELISA) were immunized with CHO-TM5 cells, which have stable high level expression of hTM, to produce antibodies specific to hTM 3 times at an interval of 2 weeks. On the third day after the third immunization, mouse with the highest titer of serum polyclonal antibody was sacrificed and spleen cells were harvested to prepare hybridoma cells with SP2/0 cells at the ratio of 10 to 1. Hybridoma cells were then cultured at 96 well plates for screening with CELISA. To improve probability to obtain specific mAb, CELISA was applied twice. The first CELISA was done with polyethylene ELISA plate with a mono layer of CHO-TM5 cells. The positive clones from the first screen were then selected by reacting with similar screening ELISA plate but having CHO cells monolayer instead. Only clones that were positive for the first screening and negative for the second screening were kept, and called as CHO-TM5(+)CHO-hybridoma cells. BALB/c mice were intraperitoneally injected with the selected hybridoma cells. Ascites were collected and monoclonal antibodies were purified using FPLC, and its Ig class, subclass, and titer were then determined respectively. The specificity of the yielded mAb was identified with CELISA, flow cytometry, ABC immunohistochemistry and immunoblotting. Detection of CELISA showed that 100 mg/kg dose of CP could tolerize the mouse to common epitopes shared between CHO and CHO-TM5 cells. And CELISA also discovered that all hybridomas positive for CHO-TM5 cells were negative for CHO cells. Five lines of positive hybridoma cells had been obtained altogether and 2177 was selected randomly for next investigation. The Ig subclass of the mAb 2F7 was IgG1 and the titer of ascitic mAb was 1 x 10(-6). Furthermore, the content of ascitic mAb was 19.56 g/L and chromosome numbers is 98. Flow cytometry, CELISA and Western blotting assays demonstrated that mAb 2F7 could specifically recognize hTM expressed on CHO-TM5 and human umbilical vascular endothelial cells (HUVEC). Meanwhile, the tissue specificity of mAb 2F7 was also identified by immunohistochemical ABC staining. On the other hand, Western blotting assays indicated that mAb 2F7 could recognize the antigen protein with 105 ku molecular mass under reduction condition. Moreover, the dissociation constant of mAb 2177, 1.22 x 10(-9) mol/L, indicated the affinity higher than some others. The results suggest that the immunotolerizing protocol provides a convenient general method for producing antibodies specific to desired protein isoforms. mAb 2F7 can specifically recognize the natural hTM expressed mainly on vascular endothelial cells, which will potentially useful for investigating the functions and clinic values of hTM.