Potential Therapeutic Intervention Suggested by Profound Deficiency of Neutrophil-Derived Bactericidal/Permeability-Increasing Protein (BPI) and Other Alterations in LPS-Directed Innate Immunity after Myeloablative Hematopoietic Stem Cell Transplantation (HSCT).

Myeloablative conditioning used for allogeneic HSCT damages the gut resulting in leakage of lipopolysaccharide (LPS, or endotoxin) into the circulation. LPS activates antigen-presenting cells to secrete inflammatory cytokines, including TNF, that enhance donor T cell responses to alloantigens. The relationship of TNF production and responsiveness to aGVHD and HSCT mortality has been demonstrated in animal models and with human correlative data. Use of either TNF or LPS antagonists in animal models or mice deficient in TNF or LPS responsiveness has led to profound amelioration of aGVHD. To determine whether LPS antagonists might have a therapeutic role in allogeneic HSCT, we measured plasma levels of 3 soluble LPS-modulating proteins in 30 sibling donor:recipient pairs undergoing myeloablative HSCT. Measurements included: 1) BPI, a constituent of neutrophil primary (azurophilic) granules with potent LPS-neutralizing activity at nanomolar concentrations in all body fluids, 2) LPS-binding protein (LBP) and 3) soluble CD14 (sCD14), which are both acute phase reactants that at basal levels enhance LPS activity by delivering LPS to its cell surface receptor but that at high levels detoxify LPS by delivering it to plasma lipoproteins. Spontaneous TNF production by whole blood and monocyte surface expression of TLR4, a key component of the LPS receptor, were also measured. Peripheral blood was obtained from donors (single time point) and patients pre-HSCT (baseline), Days 0, 7, 14, 21, 28 and on day of aGVHD. Plasma BPI levels fall dramatically from baseline, concurrent with severe neutropenia, while LBP (D7) and then sCD14 (D21–28) rise. Monocyte TLR4 expression was elevated at D7 and at onset of aGVHD. Patients with high spontaneous TNF production were at higher risk of aGVHD (p<.05) while those demonstrating greater concentrations of BPI at baseline appeared at lower risk for aGVHD (p<.01). Trends were observed suggesting lower aGVHD risk in patients with higher (i.e. potentially LPS inhibitory) levels of LBP and sCD14. Our results are consistent with a pathophysiologic pathway in which LPS, acting via TLR4, induces TNF production that contributes to aGVHD. The demonstration that patients are deficient in BPI and that deficiency of BPI and perhaps other LPS detoxifying proteins is associated with increased aGVHD risk suggest that replenishing BPI may represent a novel approach to shielding patients during the periHSCT period from LPS-induced pathophysiology, including cytokine driven events leading to aGVHD. Therefore, we are initiating a clinical trial of a recombinant N-terminal human BPI fragment (rBPI21 or Opebecan; XOMA (U.S.) L.L.C.), an investigational agent with an excellent safety profile and in vivo LPS-neutralizing activity, in myeloablative HSCT.