Inflammatory cytokines and acute graft-versus-host disease after reduced-intensity conditioning allogeneic stem cell transplantation

In standard myeloablative allogeneic stem cell transplantation (allo-SCT), several lines of evidence have suggested that inflammatory cytokines act as mediators of acute graft-versus-host disease (aGVHD). Perturbation of the cytokine network may function as a final common pathway of target organ damage, and the rapid onset of severe aGVHD can be considered a “cytokine storm.”^1-4  Nonmyeloablative, or reduced-intensity conditioning (RIC), regimens before allo-SCT have emerged as an attractive modality to decrease transplant-related toxicity while preserving the graft-versus-tumor effect.^5-7  The use of such less cytotoxic regimens has modified the natural history of transplant-related complications, especially of aGVHD.^8,9  Our knowledge of the pathophysiology of aGVHD is based primarily on results of analyses performed in patients who underwent myeloablative allo-SCT. The aim of this study was to investigate the role of inflammatory cytokines on aGVHD incidence and severity in patients who underwent RIC allo-SCT with an HLA-identical sibling.

One hundred thirteen consecutive patients treated with RIC allo-SCT in different trials at the Institut Paoli-Calmettes (Marseille, France) were included in this study. Written informed consent was obtained from each patient and donor. The studies were approved by the local ethics committee and were performed according to institutional guidelines. Eligibility criteria for RIC allo-SCT are detailed elsewhere.⁹  Patient and donor characteristics are summarized in Table 1.

The RIC regimen included either fludarabine, oral busulfan, and antithymocyte globulin (ATG) (Thymoglobuline; Genzyme, Lyon, France)^6,9  or fludarabine and low-dose total body irradiation.⁷  Supportive care has been previously reported and was similar during the entire study period.¹⁰  GVHD prophylaxis was performed with cyclosporin A (CSA) alone⁹  or with CSA and mycophenolate mofetil.⁷  Most (96%) patients received a peripheral blood stem cell allograft mobilized with granulocyte-colony-stimulating factor (G-CSF) (10 μg/kg per day).

Plasma samples were collected and stored at -80°C according to a prespecified schedule applied to all patients during the entire study period. Times for cytokine measurements were as follows: (1) before allo-SCT: median, day -7 (range, -23 to -3); (2) day 0: all procedures performed on day 0; (3) 1 month after allo-SCT: median, day 28 (range, days 20-41); (4) 2 months after allo-SCT: median, day 59 (range, days 49-75); and (5) 3 months after allo-SCT: median, day 91 (range, days 76-114). The levels of 10 different cytokines (interleukin-1β [IL-1β], IL-6, IL-8, IL-10, IL-12p70, IL-18, tumor necrosis factor alpha [TNF-α], interferon-alpha [IFN-α], IFN-γ, and Fas-ligand) were measured on these collected samples using commercially available enzyme-linked immunosorbent assay (ELISA) kits according to the manufacturer's instructions. Naive CD4⁺ T lymphocytes were identified by 4-color flow cytometry using the CD3, CD4, CD45RA, and CD27 mAbs (Beckman-Coulter, Marseille, France) as previously described.¹¹  Monocytes were counted by an automated counter.

Acute GVHD was evaluated according to standard criteria. On diagnosis of grades II to IV aGVHD, all patients were primarily treated with CSA and methylprednisolone (2 mg/kg per day). Data were computed using SEM software (Silex, Mirefleurs, France) and SPSS 9.0 for Windows (SPSS, Chicago, IL). The Mann-Whitney U test was used for comparison of continuous variables. Categorical variables were compared using the χ² test. The association of time to aGVHD with cytokine levels and other relevant variables was evaluated in a multivariate analysis with the use of a Cox proportional hazards regression model.¹² 

Table 2 shows the results and kinetics of the different cytokines measured in the entire study population. Except for IL-12p70, all other cytokines showed little variation in the blood in the first 3 months after allo-SCT. When looking at the kinetics of IL-12p70 secretion before and after allo-SCT, there was a dramatically different pattern between patients with grades II to IV aGVHD and patients with grades 0 to I aGVHD (Figure 1A). With a median follow-up of 470 (range, 125-2175) days, the incidence of grades II to IV aGVHD was 45% (95% confidence interval [CI], 36%-54%; median onset, 32 days after allo-SCT). Forty-seven (42%) patients did not experience aGVHD, and 15 (13%) patients had grade I aGVHD. In univariate analysis performed in the subgroup of 75 patients for whom all tested cytokines were measured closely and rigorously before aGVHD clinical onset, advanced disease stage (P = .04) and high IL-12p70 level (P < .001) measured around the first month after allo-SCT were significantly associated with the development of clinically significant grades II to IV aGVHD. Levels of all other cytokines did not show any significant correlation with the risk for aGVHD (Table 3). IL-12p70 levels were significantly correlated with the severity of aGVHD: grades 0 to I, median 468 pg/mL; grade II, median 2538 pg/mL; and grades III to IV, median 4615 pg/mL (P < .001; Figure 1B). In patients experiencing grades II to IV aGVHD, IL-12p70 levels decreased after aGVHD therapy (Figure 1A).

Given that IL-12p70 is mainly produced by myeloid dendritic cells (DCs) and that peripheral blood monocytes are considered the main pool of DC precursors,¹³  we investigated the recovery of monocytes in the groups of patients with and without clinically significant aGVHD. Interestingly, we found more rapid recovery of monocytes before aGVHD clinical onset in patients with grades II to IV aGVHD than in patients with grades 0 to I aGVHD (median, 829/μL vs 552/μL; P = .005; Figure 1C).

At the effector level, naive CD4⁺ lymphocytes are the principal target for IL-12p70, which induces their differentiation into T_H1 effectors.^13,14  Here, we observed a significantly more robust recovery of genuine naive CD3⁺CD4⁺CD45RA⁺CD27⁺ T cells before aGVHD clinical onset in patients with grades II to IV aGVHD compared with patients with grades 0 to I aGVHD (median, 50/μL vs 16/μL; P = .006; Figure 1D). Finally, in multivariate analysis, IL-12p70 level measured before aGVHD clinical onset was the strongest predictive factor for aGVHD development and severity (P < .001; relative risk [RR], 10.7; 95% CI, 3.8-30.6).

Previous studies established that aGVHD pathophysiology after myeloablative conditioning can be summarized as a 3-step process: phase 1, during which inflammatory mediators are produced and a proinflammatory environment is established; phase 2, during which donor T lymphocytes are activated and T-cell target interactions can take place; and phase 3, during which T cells specifically kill target host cells through various mechanisms. Other effectors, such as natural killer (NK) cells, monocytes, granulocytes, and DCs, may cause specific and nonspecific tissue lesions.¹⁵  Our study showed a specific kinetic profile and a marked increase in circulating IL-12p70 levels that significantly correlated with the incidence and severity of aGVHD, highlighting its major role in aGVHD pathogenesis. Indeed, direct and indirect evidence suggests that aGVHD is mediated by the action of donor-derived T lymphocytes thought to polarize into type 1 T cells after stimulation with IL-12 from DCs.^16,17  Most important, in our study, peak secretion of IL-12p70 was concomitant with the expansion of naive CD4⁺ T cells, but not with other CD4⁺ T-cell subsets (data not shown). Type 1 lymphocytes composed of CD4⁺ T_H1 and CD8⁺ T cytotoxic (Tc1) cells produce IFN-γ and Fas-ligand that severely injure multiple organs, leading to aGVHD.^16,18  The role of IL-12p70 in aGVHD initiation is supported by the observation that IL-12 administration stimulates aGVHD development in a murine strain combination that normally develops a T_H 2-type, autoimmune chronic GVHD syndrome, further emphasizing the implication of an afferent T_H1-type response in the generation of aGVHD.¹⁹  The IL-12-driven T_H1-type processes involve not only CD4⁺ and CD8⁺ T-cell responses but also a phase of monocyte/macrophage activation with the production of inflammatory cytokines such as TNF-α, IL-1, and IL-6, leading to the classical “cytokine storm” of aGVHD. In our study, IL-12p70 was the only cytokine that showed a significant correlation with aGVHD. This apparent discrepancy might be explained by the fact that some studies demonstrating the implication of other proinflammatory cytokines have been performed at the level of mRNA cytokine expression in GVHD target organs^20-22  and by the difference in conditioning-related inflammation²  that is less broad and less nonspecific in our study, allowing detection of the true T_H1 response.

Although triggering signals for aGVHD induction, such as the role of toll-like receptors, are yet to be established, our findings are in line with previous data showing a role for IL-18, another IFN-γ-inducing cytokine, in aGVHD pathophysiology.²³  Here, we could reconstitute a genuine T_H1 loop, supporting a model in which aGVHD primarily reflects a type 1 alloreaction (rapid monocytes/DC recovery, IL-12p70 secretion, naive CD4⁺ T-cell expansion, T_H1 and Tc1 cell differentiation) in the context of RIC allo-SCT, consistent with data from the myeloablative setting indicating a major role of the antigen-presenting cell (APC) network (monocyte/macrophage/DC) in the pathophysiology of aGVHD.²⁴  However, this study did not address the role of APC origin in aGVHD initiation. Indeed, given the specific role of host-derived APCs compared with donor-derived APCs in the initiation of aGVHD,²⁵  monitoring of the chimerism status of APCs, especially at the site of aGVHD target organs, would allow further insight into aGVHD mechanisms in the RIC setting.

Obviously, IL-12p70 is not the exclusive cytokine involved in aGVHD pathogenesis after RIC allo-SCT.²⁶  However, the fine functions of immune effectors tend to be more evident in less toxic regimens,^11,27  offering new opportunities for a better understanding of aGVHD pathophysiology. From a practical point of view, close and early monitoring of IL-12p70 can be a useful indicator of aGVHD, and anti-IL-12-based therapy can represent a potential tool for controlling alloreactivity.

We thank N. Baratier for technical assistance, the nursing staff for providing excellent care for our patients, and the physicians of the Hematology and Medical Oncology Departments at the Institut Paoli-Calmettes for their important study contributions and dedicated patient care. We thank D. Maraninchi and C. Mawas for their continuous support. We also thank the “Association pour la Recherche sur le Cancer (ARC),” the “Ligue Nationale contre le Cancer,” the “Fondation de France,” the “Fondation contre la Leucémie,” the “Etablissement Français des Greffes (EFG),” the “Association Cent pour Sang la Vie,” the “Association Laurette Fuguain,” and the GEFLUC, for their generous and continuous support for our clinical work and basic research.