- Top of page
To elucidate the mechanism of self-protection against anti-donor blood-group antibody known as accommodation, we studied 16 human ABO-incompatible living-donor kidney transplant recipients at 3 and 12 months post transplantation. Both circulating anti-blood-group antibody and the target blood-group antigen in the graft were demonstrable in all patients after transplantation. Thirteen of 16 grafts had normal renal function and histology, while three grafts with prior humoral rejection demonstrated significant glomerulopathy and thus did not meet the criterion for accommodation. Using microarrays, we compared five 1-year protocol ABO-compatible renal graft biopsies to four accommodated ABO-incompatible graft biopsies. Significant alterations in gene expression in 440 probe sets, including SMADs, protein tyrosine kinases, TNF-α and Mucin 1 were identified. We verified these changes in gene expression using RT-PCR and immunohistochemistry. Heme oxygenase-1, Bcl-2 and Bcl-xl were not increased in ABO-incompatible grafts at any time-point. We conclude that accommodation is always present in well-functioning, long-surviving ABO-incompatible kidney transplants. This self-protection against antibody-mediated damage may involve several novel mechanisms including the disruption of normal signal transduction, attenuation of cellular adhesion and the prevention of apoptosis.
- Top of page
Accommodation was first observed in ABO-incompatible renal allografts (7). The term was later applied to xenografts (14) and in ABO-compatible grafts in which alloantibody is present (15). The current study defines several unique features of accommodation in ABO-incompatible kidney transplants including: 1) detectable anti-donor antibody; 2) normal graft histology; 3) presence of A or B antigen in the graft; and 4) renal function similar to ABO-compatible patients. In addition, these data demonstrate that anti-AB antibodies return to detectable levels in all patients (accommodated and non-accommodated), even in the absence of humoral rejection or chronic graft damage. Furthermore, blood-group antigen is detectable in all recipients months after transplantation, making it likely that some form of accommodation must occur for enduring graft survival in ABO-incompatible kidney transplants.
Three patients had low GFR values and protocol biopsies showing chronic changes including interstitial fibrosis and glomerulopathy. It is unknown if these observations were secondary to the prior humoral rejection or due to a lack of protection from persistent antibody-mediated damage, i.e. lack of accommodation. One patient who had an episode of humoral rejection now has an accommodated graft at 1 year. It remains questionable whether or not the presence of subsequent accommodation will protect the graft from future chronic damage.
In contrast to previous studies suggesting that high-titer patients do not develop accommodation easily, the current data show that accommodation can occur over a wide range of anti-blood-group antibody titers. Ishida et al. found that all patients with anti-AB antibody titers <1 : 8 maintained excellent graft function, but 11 of 16 patients (69%) with anti-AB antibody titers persisting >1 : 32 lost their grafts within 12 months after transplantation (16). In a study of ABO-compatible kidney transplants the presence of anti-MHC alloantibody appeared to be associated with chronic allograft nephropathy, even when many criteria for accommodation were met (17). In our experience, 6 of 13 patients with anti-AB antibody titers ≥1 : 32 demonstrated excellent graft function at 1 year. Protocol biopsies of these grafts did not show chronic changes, suggesting that accommodation is protective in these patients. Given these issues, continued monitoring of graft histology for patients with high antibody titers will be important to determine whether chronic allograft nephropathy will become a problem in long-term graft survival.
The mechanism by which circulating antibody damages the graft at different time-points after transplantation remains unclear. Possible mechanisms include complement-mediated vascular thrombosis, antibody-dependent cellular cytotoxicity or direct signaling via binding to endothelial cell antigens. This binding might deliver a signal that causes endothelial cell activation leading to apoptosis, possibly in the absence of complement (18). Recent data suggest that some forms of accommodation may occur as the result of the increased intragraft expression of anti-apoptotic ‘protective’ genes by the graft which blunt the response to antibody. Genes such as HO-1, Bcl-xl and A20 have been associated with accommodation in rodent xenograft models (19–22). Similarly, Bcl-xl expression was increased in ABO-compatible kidney-transplant recipients with apparent accommodation to circulating anti-graft alloantibody (15). No microarray studies of these different models of accommodation have been described.
Data from this study suggest that HO-1, Bcl-2, Bcl-xl and Bax are not associated with accommodation in ABO-incompatible kidney transplants. However, it is possible that these molecules may be important in protection against antibody-mediated injury early after transplantation – a time-point not assessed in these studies. Interestingly, although HO-1 expression has been associated with humoral rejection, we did not find its expression in grafts experiencing humoral rejection (data not shown). Our data indicate that a significant alteration in intragraft physiology occurs in accommodated kidneys and some of the pathways appear to have significant implications for protection against antibody-mediated injury. Using microarrays to perform an unbiased survey of changes in gene expression, we found that 440 transcripts related to numerous biological pathways were significantly up-/down-regulated. Specifically, our data suggest that the accommodated graft is in a pro-survival environment through modifications in the expression of pathways involving SMAD, protein tyrosine kinases, TNF-α and Muc1.
The SMAD proteins form a complex pathway regulating the signaling effects of Transforming Growth Factor-β (TGF-β) and Bone Morphogenetic Protein (BMP), both of which are involved in regulating cell growth, differentiation and apoptosis (23). Mediating the effects of these intracellular signals appears very important in the accommodated graft, as three significant transcripts observed in this study were related to this pathway: EGFR (ΔHI + 0.63, p-value 0.010), SMAD5 (ΔHI − 1.68, p-value 0.014) and SMAD4 (ΔHI − 1.06, p-value 0.022). The SMADs are a family of receptor substrates that, with the aid of coactivators/repressors, translocate to the nucleus where they act as transcription factors (23). SMAD2 is a receptor-regulated SMAD (R-SMAD) whose corepressor is TGIF. When activated, EGFR triggers a cascade of events involving the Ras-Mek signaling pathways leading to the phosphorylation and stabilization of TGIF (24). The presence of TGIF within the cell will promote the formation of SMAD2/TGIF corepressor complexes in response to TGF-β (23), thereby prohibiting the binding of SMAD2 coactivators and repressing the effects of TGF-β. Similarly SMAD5, an R-SMAD whose ligand is BMP (23), was significantly down-regulated, thereby promoting a cell survival environment. After binding to coactivator molecules, each R-SMAD couples with a Co-SMAD and translocates to the nucleus as a functional transcriptional complex (23). SMAD4 is a Co-SMAD whose deletion has been shown experimentally to abrogate/abolish the signaling effects of TGF-β (25). Taken together, these effects may attenuate intracellular signaling via TGF-β and BMP, promoting cell survival.
Protein tyrosine kinases (PTK) represent a large class of molecules that we found significantly altered in accommodation. Several molecules, including GFRA1, HEK2 and PRKB, are associated with PTK and are found among our most statistically significant transcripts. GFRA1 (ΔHI + 1.45, p-value 0.018) is a receptor which mediates binding and activation of the PTK RET, leading to a decrease in apoptosis and an increase in cell survival (26). In contrast to GFRA1, PRKB was very down-regulated (ΔHI − 2.04, p-value 0.003). PRKB binds cAMP, a potent inhibitor of T-cell activation (27). In vitro studies have shown that mutation of PRKB markedly decreases expression of leptin (27), a known regulator of proliferation of naïve and memory T cells (28). A reduction in leptin will decrease pro-inflammatory Th1 cells and increase Th2 cells. These changes in various PTK molecules provide insights into the mechanism by which accommodation affects signal transduction.
The microarray also identified changes in gene expression in the TNF-α family of molecules. The accommodated patients had significantly decreased levels of TNF-α (ΔHI − 0.82, p-value 0.033) and the enzyme that cleaves precursor TNF-α to its mature form, TACE (ΔHI − 1.16, p-value 0.044). Mature TNF-α is a cytokine that causes leukocyte recruitment to the site of inflammation during early immune response (29). TNF-α binds the TNF-R1 receptor, leading to activation of NF-kB and c-Jun N-terminal protein kinase (29). NF-kB is a transcription factor that mediates gene expression for cytokines and other genes related to inflammation and leads to cell death via apoptosis (30). Interestingly, the TNF-α receptor-associated factor TRAF6, a signal transducer in the NF-kB pathway (31), is also significantly down-regulated (ΔHI − 0.97, p-value 0.030). Taken together, these changes implicate the TNF-α family in accommodation.
Finally, three significantly up-regulated probe sets were related to Muc1. Muc1 is a large transmembrane protein normally expressed on the apical surface of ductal epithelia (32). It contains a very large extracellular domain consisting mostly of tandem repeats, a membrane-spanning domain and a small phosphorylated cytoplasmic tail (33). Several in vitro studies have characterized the protein, revealing several functions, including cell adhesion, cell signaling and immunoregulation (33–37). Cell adhesion has been explored through the interactions of Muc1 with E-selectin and intercellular adhesion molecule 1 (ICAM-1), two well-characterized mediators of leukocyte adhesion to endothelial cells during the inflammatory process. The interaction between ICAM-1 and E-selectin with Muc1 has been localized to the tandem repeat region of the Muc1 extracellular domain. Overexpression of Muc1 can promote cell–cell adhesion through binding of exogenous ICAM-1 (38). Conversely, overexpression of Muc1 can abrogate adhesion by blocking the binding sites of endogenous E-selectin, thereby protecting the cell against immune surveillance mechanisms (33). Further, when phosphorylated, the cytoplasmic tail of Muc1 can bind β-catenin, causing a reduction in the amount available for binding with E-cadherin and promoting anti-adhesion between endothelial cells (39). Adding to the functional complexities of Muc1, the cytoplasmic tail has also been shown to adversely affect signal transduction via Grb2/SOS (37) which mediate the signals for several receptor tyrosine kinases, including EGFR (35).
In total, the observed changes in gene expression provide insight into novel mechanisms of self-protection against antibody-mediated injury via alterations in signal transduction, cell–cell adhesion, T-cell activation pathways and the prevention of apoptosis (pro-survival). It should be noted that when using microarrays to analyze such elaborate processes, it is possible that some important transcripts/pathways may not be easily resolved. This is due in part to the decreased sensitivity of microarrays which prohibit the identification of cytokines and other potentially important markers that are expressed at undetectable levels. In addition, mRNA derived from heterogenous cell populations (i.e. biopsies) may mask relevant changes to markers within an individual cell type. These limitations add to the complexities of working with microarrays and necessitate the need to study important identified markers by some other means such as RT-PCR or immunohistochemistry.
The use of ABO-incompatible living-donor kidneys provides a new source of kidneys for transplantation. However, the unique situation involving the presence and persistence of anti-donor blood-group antibodies presents a complex set of immunologic problems that are poorly understood. In this study, we investigated antibody titers, presence of antigen, GFR and graft histology to determine which patients had accommodation. In addition, we monitored intragraft changes in gene expression and found dysregulation of signal transduction machinery and immune surveillance that is consistent with the promotion of cell survival. These results provide evidence that the graft plays a role in its own survival in accommodation. Interestingly, the situation in which circulating antibody and its target antigen coexist in an organ without requiring clinical intervention occurs in numerous human disease states, both renal and nonrenal including Goodpasture's syndrome (40), p-ANCA positive vasculitis (41) and autoimmune diabetes (42). Thus, it is likely that some form of self-protection against antibody-mediated injury exists normally in humans and is not just a phenomenon observed in transplantation.