Alexandre et al. were the first to report the survival of ABO incompatible kidney allografts despite the return of anti-blood group antibody in the recipient after transplantation (1). In his original series of patients, plasmapharesis and splenectomy were performed prior to transplantation to remove anti-graft antibody and prevent hyperacute rejection. Detectable titers of anti-blood group antibody returned in some individuals but without evidence of humoral rejection. This resistance to antibody-mediated rejection, called ‘accommodation’, has since been observed by others after vascularized transplantation across ABO barriers, into HLA-pre-sensitized recipients and also after experimental xenotransplantation in both concordant and discordant species combinations, including after pig-to-primate transplants.
Accommodation, although a desirable outcome in patients prone to humoral rejection, is nevertheless an unusual event; the development of anti-graft antibodies is most usually associated with graft dysfunction and eventual graft failure. Accordingly, the focus of several research groups over the past few years has been to understand the basis of the protection in accommodated grafts, with the long-term aim of devising strategies to induce it in the clinic.
Over a decade ago, Platt and Bach (2) proposed three potential mechanisms: The first two dealt with the possibility that antibody deposition (or the consequences of deposition) in the graft might be altered, by a change in either the titer or the specificity of the antibodies or, alternatively in the level of expression or the nature of antigenic epitopes on graft endothelial cells. Data from one small animal model suggests that antigen down-modulation may underpin accommodation in some circumstances (3). Their third proposal was that the graft underwent an active process resulting in a protected state against antibody deposition and complement activation.
There is much evidence from small animal studies to support this third hypothesis. Most convincingly, accommodated grafts can be distinguished from non-accommodated grafts by the specific up-regulation, in the graft vasculature, of anti-apoptotic and anti-inflammatory genes such as Bcl-xL, Bcl-2 and A20 (4). The important work of Soares et al. has demonstrated that rapid neo-expression of heme-oxygenase-1 in the hours after transplantation (5) and subsequent generation of carbon monoxide (6) is essential for the induction of accommodation in rodent models.
Few studies have considered the mechanisms of accommodation in human grafts. Salama et al. (7) defined glomerular and peritubular capillary endothelial cell expression of Bcl-xL as a specific feature of the accommodated state in 3 out of 4 grafts transplanted into HLA-sensitized recipients who had been rendered cross-match negative at the time of transplantation by pre-transplantation immunoadsorption and in whom antibody recurred within the first 8 days.
A paper in this issue makes an important contribution to this field. Park et al. have collected a series of 13 patients with living-donor ABO incompatible kidney allografts surviving longer than a year with good graft function, despite persistently positive anti-blood group antibody titers and continuing expression of blood group antigens in the graft. Using microarray analysis, they describe a phenotype acquired by these accommodated grafts, characterized by expression of a distinct pattern of genes, including TNF-α, TGF-β1, SMAD5, protein kinase GFRA1 and MUC1, with changes visible at their first analysis, 3 months post-transplantation. Their results are confirmed by RT-PCR and immunohistological staining for some of the proteins. These are novel gene associations with accommodation. Interestingly, no increased expression of Bcl-2, Bcl-xL or HO-1 was found, raising the possibility that these genes, identified as playing an important role in protection early after transplantation, may not be as important in human organs beyond the first 3 months.
This paper by Park et al. is important for two reasons. First, if accommodation is to be studied in other stable transplant populations, it is these late markers of protection that will be most useful indicators of the process. Second, it contains a description of a single patient who developed accommodation rather than progressive graft dysfunction after acute humoral rejection, which had been treated successfully with corticosteroids and plasmapharesis. This important observation, albeit in one patient, has never been described in animal models and it illustrates that antibody-mediated rejection does not necessarily prevent the subsequent development of a protective phenotype.
Significant questions about accommodation still remain, such as why it develops in some grafts in the presence of antigraft antibodies but not others, what factors promote its spontaneous development and how well accommodated grafts resist other potentially damaging influences such as ischemia-reperfusion injury and T cell-mediated pathology. Understanding how to recognize accommodated from nonaccommodated organs takes us a step nearer to being able to address these questions in the future.