"INVASIVE AND NON-INVASIVE DIAGNOSIS OF SUBCLINICAL RENAL TRANSPLANT REJECTION. PROTOCOL BIOPSIES, URINE NMR AND PROTEOMICS"
David Rush for the Winnipeg Transplant Group
Department of Medicine, University of Manitoba Winnipeg, Manitoba, Canada
In 1994, our group first reported that Banff criteria for acute rejection could be found in patients with stable graft function (i.e. a change in serum creatinine <10% from baseline), an entity that we called "subclinical rejection".1 We subsequently reported that Banff criteria for Type I rejection were present in approximately 30% of protocol biopsies performed in the first 3 months post-transplant.2 The early immunosuppressive regimen used in our studies consisted of cyclosporine, azathioprine and prednisone. However, an increase in the baseline immunosuppression to include cyclosporine microemulsion, mycophenolate mofetil and prednisone, while decreasing the incidence of early clinical rejections, failed to lower the prevalence of subclinical rejection.3
Other groups have also performed protocol biopsies in the early post-transplant months. Both Legendre et al4 and Nankivell et al5 have demonstrated that the prevalence of subclinical rejection observed in a 3-month protocol biopsy in cyclosporine-treated renal allograft recipients was 29% and 23% respectively, similar to our centre.
The use of tacrolimus instead of cyclosporine may be associated with a lower prevalence of subclinical rejection. Jurewicz reported a prevalence of 18% of subclinical rejection at 3 months in cadaveric renal transplant recipients on tacrolimus, azathioprine and prednisone.6 More recently, Gloor et al reported a prevalence of subclinical rejection of only 2.6% at 3 months in patients on tacrolimus, mycophenolate mofetil and prednisone.7 In this latter study, however, over 60% of patients were recipients of living-related grafts, and greater than 50% of patients had received induction therapy with anti-lymphocyte agents.7 In very early (mean time 8 days post-transplant) protocol biopsies in patients receiving tacrolimus and steroids, two thirds of whom also received mycophenolate mofetil, Shapiro et al showed that 25% of patients showed Banff Type I or Type II rejection, despite stable or improving renal function.8 The long-term significance of this very early inflammation is unknown.
Subclinical rejection is an alloimmune response
The interstitial infiltrate and tubulitis that characterize subclinical rejection is the result of an alloimmune response. A correlation between subclinical rejection and HLA mismatching has been reported by our group,9 as well as others.5,8 By multivariate analysis, only HLA matching between donor and recipient correlated with the presence of subclinical rejection. The prevalence of subclinical rejection in the 1, 2, and 3-month protocol biopsies was 20%, 25% and 0%; 30%, 32% and 32%; and 63%, 37% and 30% in zero, 1 and 2 HLA-DR mismatched patients, respectively.9
More recently our group has reported that subclinical rejection is more prevalent in patients presensitized to Class I and Class II HLA antigens as detected by flow-cytometry crossmatching.10 Early, clinically silent inflammation in the graft appears therefore to be largely an alloimmune response triggered by either mismatching or presensitization to the major histocompatibility antigens.
Is subclinical rejection pathogenic or is it beneficial to the graft?
Results from animal models have suggested that some allograft infiltrates may be immunoregulatory and therefore beneficial to the long-term survival of the graft. The data in humans, however, strongly favors a pathogenic role for subclinical inflammation.
Using immunohistochemistry techniques, our group found an increasing frequency of expression of pro-inflammatory phenotypic and activation markers from normal, through subclinical, to clinical rejection biopsies for cells of both the lymphocyte and monocyte lineage.11,12 In our studies only the macrophage activation marker allograft inflammatory factor-1 (AIF-1) (but not CD25, CD69 or perforin) discriminated between subclinical and clinical rejection.11 Our group has also studied the presence of gene transcripts for chemokine, cytokine and cytotoxic lymphocyte products in protocol biopsies using PCR techniques.13,14 Gene transcripts for RANTES, MIP-1 , IL-2, IL-4, IL-10, IL-15, TNF- , and IFN- were present in biopsies from patients with subclinical rejection and largely absent in normal protocol biopsies. Transcripts for perforin, Fas ligand, and granzyme B were also present in patients with subclinical rejection, although in reduced amounts when compared to biopsies from patients with clinical rejection episodes. There were no differences in IL-10 and IL-15 transcripts in clinical and subclinical rejection biopsies.
Additional, albeit indirect, data in favor of a pathogenic role for subclinical rejection comes from the early observation of Isoniemi et al, who reported that in patients who had not experienced clinical acute rejection episodes, the development of chronic histological changes occurred in inverse relation to the amount of immunosuppression they had received.15 Similarly, Legendre et al reported a patient cohort that never experienced clinical rejection episodes, in whom the development of chronic rejection at 2 years was preceded by subclinical rejection at three months.4 Moreover, Kirk et al showed that whereas lymphocyte infiltration and pro-inflammatory gene transcripts were commonly found in stable grafts with mild chronic histological changes, persistent lymphocyte infiltration was associated with histological worsening and graft functional deterioration at later time points.16 More recently, both Shishido et al17 and Birk et al18 demonstrated that acute subclinical rejection leads to progression of fibrosis in sequential protocol biopsies performed in children. Finally, the data from our randomized and non-randomized studies suggest that beneficial outcomes (better function and a lower chronic histological score) result from treatment of subclinical rejection.2,3
Non-invasive diagnosis: Nuclear magnetic resonance (NMR) and proteomics
There is thus increasing evidence in both adults and children that clinically overt rejection is only a portion of the burden of alloimmune injury sustained by the graft during its lifetime. Subclinical rejection may represent a substantial proportion of that burden, the pathogenicity of which has been difficult to prove conclusively because of the lack of sensitive methods for its detection. There are obvious difficulties in having to rely on the protocol biopsy for the diagnosis of subclinical rejection. Patient risk and procedural cost are obvious drawbacks. In addition, the prevalence of subclinical rejection is likely to decrease with modern immunosuppression, and its detection by protocol biopsy may become more subject to sampling error due to the "patchiness" of early inflammation in the graft.
Confirmation that subclinical alloreactivity contributes to adverse long-term transplantation outcomes will require the development of novel diagnostic tests that detect clinically silent inflammation in the graft. Such tests would ideally be non-invasive, thus allowing for safe, frequent sampling of the graft, and would be capable of distinguishing alloimmune from other forms of graft inflammation. Such tests may include (a) the demonstration of rejection-specific genes in blood and urine, as has been done in tissues using microarray techniques,19 (b) the finding of alloactivated or cytotoxic cells or their products in the blood or urine,20-24 (c) the pattern of urinary metabolites identified by magnetic resonance (MR) or infrared (IR) spectroscopy,25 or (d) by the identification of protein patterns by the technique of urine proteomics.26,27
Our group has had experience with the latter two techniques. In both our MR/IR and proteomic studies, we have used the protocol biopsy as the "gold standard" to which the spectra are compared. In the case of urine MR and IR, a report of our early experience has been published.26 Our preliminary findings with urine proteomics are in press.27 Both techniques show great promise. Using a combination of urine MR and IR spectra, as previously reported,26 the distinction between rejection (clinical or subclinical) and normal graft histology can be made with approximately 90% sensitivity/specificity and positive/negative predictive value. It should be emphasized however that the test requires validation in a clinical trial; and this is currently underway. We have examined sequential urines in our patients and found that the characteristic spectra of rejection can precede the protocol biopsy diagnosis of subclinical rejection or the development of clinical rejection by about two weeks. Similarly, persistence of the spectra is correlated with non-resolution of the rejection by biopsy and vice versa. The spectra of rejection are not seen with non-inflammatory entities – e.g. acute tubular necrosis. One disadvantage of the MR/IR technique is that approximately 100 cases of each discrete pathological entity (e.g. acute rejection) and their respective spectra are required to generate a robust "biomarker". So far the MR/IR biomarker is the same subclinical and clinical rejection. However, when ~100 cases of each entity are observed, differences in their spectral patterns will be sought.
The urine proteomic studies have thus far been centered on the protein patterns observed in clinical rejection episodes.27 Nevertheless, a characteristic pattern of urinary proteins is seen in such patients, as compared to the urine proteins seen in transplant patients with acute tubular necrosis, recurrent glomerular disease, normal grafts, and in normal controls or in normal patients with urinary tract infections. The identification of these proteins is now being pursued.
In conclusion, detection of allograft is now possible using non-invasive monitoring techniques. These techniques may allow for tailoring of the intensity of immunosuppression to the inflammatory status of the graft, and result in the reduction of both the incidence of chronic rejection and the unwanted side effects of immunosuppressive therapy. Implementation of these techniques may further improve the outcomes currently observed in renal transplant patients.