Prognostic Significance of Complement System Activation After Allogeneic Hematopoietic Stem Cell Transplantation.

The complement system consists of several serum proteins and cell membrane receptors that can be activated by three possible pathways leading to the production of pro-inflammatory factors such as the anaphylatoxins C3a and C5a and the formation of the membrane attack complex. The complement system has been implicated in the pathophysiology of several immune diseases and we previously reported that it was activated after allogeneic haematopoietic stem cell transplantation (HSCT) both in humans and mice. It's activation after allogeneic HSCT following a myeloablative conditioning could predict the development of gastrointestinal (GI) GVHD in humans. We now present updated data with long term follow up on an enlarged series of patients who did or did not activate complement after allogeneic HSCT. We also discuss the preliminary results of the inhibition of complement activation on experimental models of GVHD.

Complement activation was determined by measurment of complement proteins before and once a week up to 3 months after allogeneic HSCT in 34 patients allografted for diverse haematological malignancies in our institution following conventional myeloablative conditioning. Results were correlated to the clinical evolution of allogeneic HSCT. Preclinically, inhibition of mouse complement activation by C1 inhibitor esterase or an anti-C5 monoclonal antibody was tested in a parent (C57BL/6, H-2^b) to F1 [(C57BL/6xDBA2), H-^bd] GVHD mouse model. Lethally irradiated (9.75 to 11 Gy) mice reconstituted with either syngeneic or allogeneic bone marrow cells and splenocytes (10⁷ and 13 to 20 ×10⁶ cells/recipient, respectively) were treated with complement inhibitors between day-7 to Day 21 or day 1 to Day 28 post-HSCT.

Fifteen (44%) patients showed an activation of the classical pathway, as defined by a decrease of C3 and C4 proteins below normal values and of at least 50% of their pre-HSCT levels. Activation occurred during the 4 first weeks following HSCT in 11 patients, between 6 and 8 weeks in 2 cases and later, after withdrawal of immunosuppression, in 2 patients. Pre-transplant characteristics (age at transplant, sex, underlying haematological disease, conditioning regimen, use of ATG, type of donor and CMV risk) of patients who showed an activation of complement were comparable to those who did not. We confirmed that activation of complement was significantly associated with acute GI GVHD (80 % in the activated group versus 5.3% in the non activated, p=0.0028) and occurred from 1 to 3 weeks before the appearance of clinical GVHD. Patients in the activated group had a significantly increased incidence of capillary leak syndrome concomitant to the conditioning toxicity or GVHD (66.7% in the activated group vs 10.5% in the non activated, p=0.019). There was no significant difference between the groups in terms of skin or hepatic acute GVHD, TMA, VOD and chronic GVHD. In a landmark analysis by day 100 post-HSCT, overall survival was significantly impaired in the group of patients with complement activation (p=0.008) because of increased toxicity related mortality (p=0.02) and relapse (p=0.01) risks in comparison to the patients without complement activation.

Complement inhibitors capable of blocking the first pathway (C1 inhibitor esterase) or terminal complement activation (anti-C5 mAb) were used in a mouse model of GVHD. These inhibitors were administrated before and/or during the phase of activation of complement which starts on day 4 post-HSCT. However, GVHD clinical signs and mortality did not appear to be improved following complement inhibitor administration in this mouse model.

Activation of complement after allogeneic HSCT following a myeloablative conditioning appears as a predictive marker of acute GI GVHD and a pejorative prognostic factor of transplant outcome. However, inhibition of complement activation in a mouse experimental model of allogeneic HSCT did not allow the control of GVHD. It is not clear, therefore, that this mouse model is useful as a model of human GVHD.

No relevant conflicts of interest to declare.