Late-onset acute GVHD: clues for endothelial GVHD

Late acute (LA) graft-versus-host disease (GVHD) has recently been added to the clinical spectrum of GVHD. In this issue of Blood, Holtan et al (and the Chronic GVHD Consortium) studied LA GVHD in a large prospective cohort and describe its link with circulating angiogenic factors.¹ 

Thanks to the 2005 National Institutes of Health Consensus Conference diagnostic criteria of GVHD,²  it is now widely accepted that acute GVHD can occur well beyond the classical landmark of day 100. LA GVHD can be considered as de novo if signs develop after day 100 or recurrent if signs occur beyond day 100 in a patient with a previous history of acute GVHD.

In the study by Holtan and coworkers, the authors took advantage of the Chronic GVHD Consortium Network to study, for the first time prospectively, the incidence, response to therapy, and outcome of patient with LA GVHD. Among 909 patients, 83 developed LA GVHD (2-year cumulative incidence, 11%) at a relatively early time posttransplant (around day 100). As expected, organ involvement was mostly represented by gut and skin. Systemic steroids were newly started or increased in nearly two-thirds of the case, more than half of the patients needed second-line treatment (mainly for progressive disease), and only one-fourth discontinued immunosuppressive therapy. The median failure-free survival was only 7 months and the 2-year overall survival was 56%. Finally, in analyzing the overall cohort of 909 patients, authors found that LA GVHD was associated with a 1.7-fold increased risk of overall mortality and nonrelapse mortality (NRM). All of these prospectively collected data are of significant clinical interest and will be useful for patient counseling and treatment decisions.

The strength of this study comes from ancillary studies which analyzed circulating angiogenic factors. The authors analyzed prospectively collected samples from cases (n = 55) and controls (n = 50) and data from an independent validation cohort (n = 37). They analyzed 4 epidermal growth factor (EGF) receptor (EGFR) ligands, 2 EGFR ligand sheddases, and 3 regulators of angiogenesis (see their article for details). The main result of this biological analysis was that plasma amphiregulin (AREG), an EGFR ligand, and elevated AREG-to-EGF ratio were associated with increased NRM and decreased probability of survival.

The relationship between acute GVHD and endothelium has long been discussed, with evidence of increased circulating levels of molecules like thrombomodulin, plasminogen activator inhibitor 1, pathological evidence of endothelial lesion³  and donor-derived endothelial cells⁴  in severe intestinal GVHD, or, more recently, experimental evidence of neovascularization.⁵  The link between endothelial activation and the allogeneic reaction has been reviewed relatively recently.⁶  Here, the investigators focused on angiogenic factors they previously studied in classical acute GVHD (before day 100).⁷  Among 9 tested molecules, only AREG and an elevated AREG-to-EGF ratio correlated with clinical outcomes. It is of note, however, that these biological alterations were found not only in patients with LA GVHD but also in patients with classical acute GVHD; they were absent in patients with chronic GVHD.

The interaction between the EGFR with its ligands turns out to be extremely complex nowadays because at least 7 ligands have been described. Among them, EGF is a high-affinity ligand whereas others like AREG act as a low-affinity ligand⁸  (of note only 4 have been studied by Holtan et al).

In our opinion, the most intriguing (and interesting) aspects of the study by Holtan et al are those on AREG. Emerging data suggest that AREG might be a critical component of type 2–mediated resistance and tolerance. Notably, numerous studies demonstrated that in addition to the established role of epithelial- and mesenchymal-derived AREG, multiple cell hematopoietic-derived subsets can express AREG, including mast cells, basophils, and innate lymphoid cells type 2 (reviewed in Zaiss et al⁹ ). Last but not least, a fascinating article by Rudensky’s group recently described that, in addition to their suppressor function, regulatory T cells (Tregs) have a major, direct, and nonredundant role in tissue repair and maintenance.¹⁰ 

Thus, as is usual in good science, the study by Holtan and coworkers raises more questions than answers: what are the sources of AREG (epithelial, endothelia, Treg, …)? Is that a general phenomenon in acute GVHD that the host uses for tissue repair? And, last but not least, are endothelial cells truly targeted by direct or indirect allogeneic recognition or simply damaged by inflammatory mediators?