Deficiency of Either Carboxypeptidase B2 or Carboxypeptidase N Exacerbated Disease Activity in a Mouse Model of Hemolytic Uremic Syndrome

Two different basic carboxypeptidases circulate in blood - carboxypeptidase N (CPN) and proCPB2. CPN is constitutively active, while proCPB2 is a zymogen, (also termed thrombin activatable fibrinolysis inhibitor, TAFI), and is activated by the endothelial thrombin/thrombomodulin complex to CPB2. Their kinetics of substrate cleavage are distinct but both can efficiently inactivate the complement anaphylatoxins, C3a and C5a. Hemolytic uremic syndrome (HUS) is caused by Shiga toxin (stx) producing strains of E. coli and is characterized by the triad of hemolysis, thrombocytopenia and uremia. We hypothesized that in a mouse model of HUS, Cpb2^-/- and Cpn^-/- mice would have prolonged C5a anaphylatoxin activity thus causing disease exacerbation.

HUS was induced by stx2 and LPS administration in WT, Cpn^-/- and Cpb2^-/- mice. In Cpb2^-/- mice, median time to death was earlier (60 hours, n=15) than in WT mice (96 hours: n=42, p<0.0001), and had greater kidney and liver damage shown by increases in ALT, AST, BUN and creatinine levels at 48 hours (creatinine Cpb2^-/-: 1.01 mg/dL, WT: 0.25 mg/dL; Cpb2^-/- control: 0.20 mg/dL and WT control: 0.19 mg/dL; Cpb2^-/- p<0.0001 vs. all other groups, n>9). An increase in hemolysis was demonstrated by reduced RBC count and hemoglobin level plus an increase in total bilirubin and LDH. Profound thrombocytopenia (Cpb2^-/-: 121,000/μL vs. WT: 217,000/μL; p=ns) developed in both genotypes (control Cpb2^-/-: 1,001,000/μL vs. control WT: 1,141,000/μL; p=ns but vs. either Cpb2^-/- or WT with HUS, p<0.0001) and thus the HUS clinical triad was present. Histology showed tubular epithelial necrosis in the kidney ante-mortem. Administration of either toxin separately caused milder disease without the characteristics of HUS and with no observed mortality. Induction of the disease depended on co-administration of both toxins. Treatment with anti-murine C5 antibody (0.75 mg every 24 hours from 3 hours before disease initiation) improved survival of both WT and Cpb2^-/- mice with a median survival time of 168 hours for both genotypes (n=11, p=0.003 and <0.0001 respectively) and normalized the outcomes between the genotypes.

Cpn^-/- mice also died sooner (median time to survival 81.5 hours, n=28) than WT mice (96 hours, n=42, p=0.0002). The median survival time between Cpb2^-/- and Cpn^-/- mice was also significantly different (60 vs. 81.5 hours, p=0.0083). This is a first direct comparison of the role of CPN vs. CPB2 in regulating C5a activity in a disease relevant mouse model. Our study suggests that both CPB2 and CPN protect against HUS by inactivation of C5a with CPB2 having a greater effect than CPN. When Cpb2^+/-/Cpn^+/- mice are crossed, all expected genotypes are recovered in the expected Mendelian ratios including double deficient Cpb2^-/-/Cpn^-/- mice. Thus absence of both plasma basic carboxypeptidases is not essential for murine life. We are currently evaluating the Cpb2^-/-/Cpn^-/- mice in our HUS model.