Cheryl A. Johanns

The recombinant 23-kDa N-terminal fragment of bactericidal/permeability-increasing protein (rBPI23) decreases Escherichia coli-induced mortality and organ injury during immunosuppression-related neutropenia.

Cyclophosphamide-induced neutropenia exacerbates septic shock and multiple organ injury in conscious rats during Escherichia coli (EC) bacteremia despite antibiotics and fluid administration. We hypothesized that such shock and inflammatory organ injury would be mitigated by rBPI23s microbicidal activity and/or binding of EC endotoxins. Four days after 100 mg cyclophosphamide/kg, catheterized rats with < 300 PMNs/μL were pretreated with rBPI23 or the irrelevant 22 kDa protein thaumatin [3.3–6.6 mg/kg, i.v. in 0.9% NaCI (NS)] 5 min before graded i.v. infection with 5 ± 109 or 1 ± 1010 cfu of EC serotype 055:B5 ending at t = 0. Posttreatment with each protein continued (3.3–6.6 mg/kg in 1 ml NS/h) through 8 h, in addition to penicillin plus amikacin sulfate at t = 1.5 and 8 h. Arterial samples were obtained before pretreatment and at t = 1.5, 4.5, 8, and 24 h when animals were necropsied. One of eight thaumatin + 5 ± 109 EC rats and none of six thaumatin + 1010 EC rats survived 24 h. In contrast, rBPI23 significantly reduced mortality after either inoculum, improved bacterial clearance, and led to renormalization of early EC-induced hypotension, hypothermia, tachypnea, hyperoxemia, and hypocarbia. Compared with thaumatin, however, rBPI23 did not reduce circulating endotoxin or bioactive and antigenic tumor necrosis factor-α. Sepsis-induced severe neutropenia (< 50 PMNs/μL) evident in all EC rats by t = 1.5 h was reversed with rBPI23 by t = 8 h, but thrombocytopenia (<5 ± 104 platelets/μL) evident in all groups by t = 4.5 h was not altered. Although rBPI23 did not prevent EC-induced increases in microvascular permeability, it reduced pulmonary interstitial edema, microvascular congestion, and hepatocytic steatosis. Thus rBPI23 enhances host defense during neutropenic bacteremia, independent of circulating levels of endotoxin or of proinflammatory cytokines including TNF.

A recombinant tumor necrosis factor-α p80 receptor:Fc fusion protein decreases circulating bioactive tumor necrosis factor-α but not lung injury or mortality during immunosuppression-related gram-negative bacteremia☆

PURPOSE During gram-negative bacteremia (GNB), tumor necrosis factor-alpha (TNF-alpha) is a critical early mediator of host defense, whose overexpression can initiate acute lung injury, multiple organ failure, and death. In this study we evaluated the ability of a chimeric fusion protein containing two extracellular domains of the human p80 TNF-alpha receptor and the Fc region of human IgG1 (TNFR:Fc) to reduce circulating TNF-alpha, and to ameliorate organ injury and improve survival in a rodent model of GNB during immunosuppression-related neutropenia. MATERIALS AND METHODS Conscious catheterized male rats (n = 37) with stable cyclophosphamide-induced neutropenia were infected intravenously (i.v.) with 5 x 10(9) live Escherichia coli (EC, serotype O55:B5) ending at t = 0. All animals received antibiotics (penicillin/ amikacin sulfate) at t = 0.5 and t = 8 hours, and 0.9% sodium chloride (normal saline solution (NS), 1 mL/h) from t = 0 to 8 hours. Subgroups were post-treated at t = 0.5 hours with a 1.0 mL i.v. dose of TNFR:Fc (60, 600, or 1,200 micrograms; Immunex), 600 micrograms of human IgG1-kappa or IgG1-lambda (Sigma), or NS alone (controls). A separate TNFR:Fc pretreatment subgroup received 600 micrograms/rat of the fusion protein 5 minutes before starting EC infusion. Hemodynamics were monitored continuously through t = 24 hours, and arterial samples were collected at baseline and at t = 1.5, 4.5, 8, and 24 hours after EC were analyzed for blood gases, quantitative culture, serum endotoxin, bioactive and antigenic TNF-alpha, and formed elements. Postmortem tissues were examined for histopathologic changes. RESULTS Compared with antibiotic-treated and fluid-supported controls, TNFR:Fc dose-dependently reduced bioactive but not antigenic TNF-alpha without altering bacterial clearance, serum endotoxin, or 24-hour survival. Of note, 600 micrograms of IgG1-kappa or IgG1-lambda attenuated peak bioactive TNF-alpha to a similar degree as 1,200 micrograms TNFR:Fc, and also significantly reduced serum endotoxin levels. Nevertheless, by t = 8 hours all bacteremic rats were hypothermic with tachypnea-related hypocarbia and hyperoxemia and were thrombocytopenic. At death, all subgroups showed similar hepatic glycogen depletion and pulmonary congestion with perivascular edema and alveolar hemorrhage. CONCLUSIONS Although TNFR:Fc and its idiotypic control IgG1 reduced circulating bioactive TNF-alpha, neither treatment prevented progression of lethal shock with attendant organ injury in this conscious, antibiotic-treated and fluid-resuscitated model of immunosuppression-related GNB.