SEARCH

SEARCH BY CITATION

Keywords:

  • Antibody-testing;
  • Banff;
  • precision diagnostics;
  • transplantation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Antibody Testing—Paradigms Lost
  5. Laboratory Support Strategies for the Incompatible Recipient
  6. Diagnostics Challenges in Transplantation Pathology
  7. Disclosure
  8. References

The Canadian and American Societies of Transplantation held a symposium on February 22, 2012 in Quebec City focused on discovery, validation and translation of new diagnostic tools into clinical transplantation. The symposium focused on antibody testing, transplantation pathology, molecular diagnostics and laboratory support for the incompatible patient. There is an unmet need for more precise diagnostic approaches in transplantation. Significant potential for increasing the diagnostic precision in transplantation was recognized through the integration of conventional histopathology, molecular technologies and sensitive antibody testing into one enhanced diagnostic system.


Abbreviations
ABMR

antibody-mediated rejection

ABO-I

ABO-incompatible

AST

American Society of Transplantation

BOS

bronchiolitis obliterans syndrome

CST

Canadian Society of Transplantation

DSA

donor specific, anti-HLA antibodies

IVIG

intravenous immune globulin

MRI

mean fluorescence intensity

PCA

principal components analysis

TCMR

T cell-mediated rejection

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Antibody Testing—Paradigms Lost
  5. Laboratory Support Strategies for the Incompatible Recipient
  6. Diagnostics Challenges in Transplantation Pathology
  7. Disclosure
  8. References

Accurate diagnoses are the prerequisite for precision medicine in transplantation. It was only recently that precision for classifying ‘rejection' into T cell-mediated rejection (TCMR) and antibody-mediated rejection (ABMR) considerably increased [1]. This was primarily driven by the advent of new diagnostic tools (C4d staining and solid-phase HLA antibody testing) allowing for more accurate diagnosis of these two differently treated rejection entities [2, 3]. However, current diagnostic criteria for biopsy assessment are empirically derived, prone to subjective interpretation and thus of limited accuracy [4]. In addition, methods and platforms used for assessment and monitoring of alloantibodies are subject to interinstitutional variation [5]. It is likely that this lack of precision in diagnostics impacts patient care and causes significant variation in patient management and choice of therapeutic strategies. This may impact clinical trials by contaminating treatment arms with false positive and false negative cases, representing a significant bias for the true potential of new drugs [6, 7].

The unmet need in transplantation is to utilize more precise diagnostics. To this end the Transplant Diagnostics Community of Practice of the American Society of Transplantation (AST) and the Canadian Society of Transplantation (CST) held a joint symposia during the 2012 annual CST meeting. The symposium focused on recent advances in applying “omics” technologies to transplant biopsies and the use of solid-phase HLA antibody testing platforms to improve transplant diagnostics.

Antibody Testing—Paradigms Lost

  1. Top of page
  2. Abstract
  3. Introduction
  4. Antibody Testing—Paradigms Lost
  5. Laboratory Support Strategies for the Incompatible Recipient
  6. Diagnostics Challenges in Transplantation Pathology
  7. Disclosure
  8. References

Current technologies for HLA testing

HLA testing has evolved substantially over the last 40 years, and comprehensive molecular HLA typing and sensitive solid-phase HLA antibody screening and identification methodologies now facilitate donor–recipient humoral risk assessment [8]. The donor-specific cross-match has evolved from complement-dependent cytotoxicity methods, with increasing sensitivity due to AHG-augmentation and now to flow-cytometry-based methods [3, 9-11]. The introduction of solid-phase antibody testing platforms has improved both sensitivity and specificity of HLA and non-HLA antibody identification. Limitations of these assays must be recognized for the accurate interpretation of test results. The MFI (mean fluorescence intensity) output of the assay is variable between laboratories and is not intended to be a quantifiable metric [12]. Clinically, although higher MFI may be correlated with worse outcomes, there is still a wide range of MFI over which clinical consequences are not clearly differentiated [13-15]. Single HLA antigen bead assays have been modified to distinguish complement fixing antibodies from noncomplement fixing [16, 17]. Although HLA antibodies that fix complement are associated with higher rates of rejection and allograft loss, noncomplement binding antibodies still have documented clinical consequences [18]. Development of rapid, high-throughput sequencing and PCR assays for HLA typing have allowed for a more accurate definition of recipient and donor HLA genotypes. High-resolution HLA typing allows for differentiation of unique epitopes to which a recipient may make specific antibodies. These enhanced methods of HLA typing and antibody identification provide greater insights than ever before in determining recipient humoral risk profiles.

Significance of antibody detection posttransplantation

While the detrimental effect of donor-specific, anti-HLA antibodies (DSA) after kidney transplantation is well established and most insight into pathomechanisms of ABMR were obtained from kidney transplants [19], the role of DSA (anti-HLA and non-HLA) in other organ transplants is still not understood to the same extent. Therefore the meeting focused on the role of DSA in other organ transplants than kidney. Due to the lack of defined pathological criteria for ABMR in some organ transplants such as liver and lung, diagnosis of ABMR largely depends upon finding donor-specific antibodies in the serum. Using solid-phase assays this has become more sensitive and specific and hopefully will now allow for better understanding of the clinical and pathological features of ABMR in these organs.

Previous dogma led us to believe incorrectly that ABMR does not affect the liver allograft. However, increased graft loss rates have been reported in recipients with DSA [20]. A recent study of 19 cross-match positive patients found that 15 spontaneously converted to a negative cross-match, one received empiric Rituximab and converted, and 3 developed ABMR (defined as HLA DSA in serum, C4d liver tissue staining, and graft dysfunction) [21]. In a separate analysis of 43 patients with abnormal liver tests prompting a biopsy, DSA and C4d were found more commonly in patients with rejection (especially with ductopenia), implying a role of DSA in both acute and chronic rejection of liver allografts [22]. In the analysis of O'Leary et al., 92% of recipients with chronic rejection had HLA DSA compared with 61% of control patients without rejection (p = 0.003) [23]. An analysis of 760 liver allograft recipients demonstrated de novo class II DSA in 7% and documented it as an independent predictor of graft loss. Thus, there is an urgent need to create unified diagnostic criteria for ABMR in liver allografts so that systematic testing and treatment trials can be undertaken.

Similarly diagnostic criteria for ABMR in lung transplants are still not well defined [2]. Obliterative bronchiolitis is the histologic hallmark of chronic rejection, and bronchiolitis obliterans syndrome (BOS) is the clinical surrogate. An association between HLA antibodies, lymphocytic bronchiolitis and BOS has now been recognized [24-27]. Hachem and colleagues instituted a clinical protocol to screen for the development of DSA after lung transplantation. Sixty five of 116 (56%) recipients developed DSA postlung transplantation [28]. Forty four of these were treated with a single dose of rituximab and monthly intravenous immune globulin (IVIG) and 17 were treated with monthly IVIG alone. Twenty seven of the 44 (61%) cleared the DSA, and 11 of the 17 (65%) treated with IVIG alone cleared the DSA. Regardless of treatment regimen, recipients who cleared DSA were significantly less likely to develop BOS and had better survival than those with persisting DSA. However, DSA per se was not associated with an increased risk for BOS (BOS developed in 38% of patients who developed DSA and in 46% who did not; p = 0.34), underscoring that in addition to DSA other factors can cause BOS.

Although HLA antigens are the major targets of the alloimmune response causing ABMR, DSA is found in only 60% of heart transplant recipients with biopsy-proven ABMR [29]. This suggest that antibodies against non-HLA antigens may contribute to the development of ABMR, not only in heart but also in other transplants [30-32]. In the absence of HLA antibodies, ABMR in heart transplants was associated with antibodies to donor MICA or endothelial cells. Since endothelial cell proliferation and migration contribute to development of transplant vasculopathy, the Reed lab examined the changes in the cytoskeleton of the endothelial cells caused by anti-HLA antibodies [33]. Analysis by tandem mass spectrometry revealed a unique cytoskeleton proteome for HLA class I antibody-stimulated cells. Using annotation tools, 12 proteins were identified, which were unique to the HLA class I stimulated group. Eleven of the candidate proteins were phosphoproteins and exploration of their predicted kinases provided clues as to how these proteins may contribute to the understanding of HLA class I-induced ABMR. One of the candidates, eukaryotic initiation factor 4A1 (eif4A1), is involved in protein synthesis and proliferation downstream of mTOR complex 1 suggesting it a potential therapeutic target.

Laboratory Support Strategies for the Incompatible Recipient

  1. Top of page
  2. Abstract
  3. Introduction
  4. Antibody Testing—Paradigms Lost
  5. Laboratory Support Strategies for the Incompatible Recipient
  6. Diagnostics Challenges in Transplantation Pathology
  7. Disclosure
  8. References

ABO incompatible transplantation

ABO-incompatible (ABO-I) transplantation is a strategy to increase the number of donors. In addition to infant heart transplants, adult and pediatric kidney, liver and lung ABO-I transplants have been performed [34]. Strategies to decrease the isohemagglutinin (anti-A or anti-B) titers are generally necessary to reduce the risk of hyperacute rejection and future episodes of ABMR. Current assays use standard agglutination testing to determine isohemagglutinin titers, but show marked variability across laboratories. Consequently, the strategies for ABO-I transplants differ across the organ groups with respect to the importance of the isohemagglutinin subtype on long-term outcomes and at what titer a transplant can be performed safely. In 2008, the College of American Pathologists introduced standardized protocols for isohemagglutinin determination [35]. However despite this effort the variation between laboratories has not been eliminated. Therefore, it is essential to understand what testing is being performed at an individual's institution, what the variation within the laboratory is, and to locally determine a titer that is safe to transplant across. Future development of alternative methods for measuring isohemagglutinins, such as flow cytometry or ELISA, may increase the accuracy and decrease variation [36].

How the HLA laboratories can facilitate access to transplantation in paired exchange programs. National and single centre experience

Accurate and precise testing for recipient HLA antibodies and donor HLA typing is critical to the success of paired exchange programs. The variances in local versus multicenter or national program were highlighted. In Canada the small population of most transplant programs limits the number of paired exchanges that can be done locally. Therefore a National program was developed with the initial goal to maximize transplants but minimize immunological risk. This required extensive collaboration between the HLA laboratories across the country. To minimize unexpected positive cross-matches the laboratories agreed upon minimum requirements for HLA typing, antibody testing and interpretation. These efforts facilitated successful transplantation of 122 transplants. The outcome in these patients is very good and slightly superior to those transplanted from the waitlist.

The transplant center in San Antonio Texas used a different approach to maximize transplants in highly sensitized patients. A single-center algorithm was developed empirically. No strict cut-off for unacceptable antibody burden is part of this algorithm. It encompasses an individual risk assessment based on HLA locus and sensitization status, i.e., for a highly sensitized individual the most favorable match is sought with a lower risk antibody if a DSA negative donor is unlikely to be found. This is achieved by performing cross-matches with potential donors. This contrasts with many multicenter paired exchange programs where assignment of unacceptability is often more restrictive to reduce the likelihood of a positive cross-match leading to the breakdown of a proposed chain [37]. Inclusion of compatible pairs in the programs has expanded the center registry and improved the options for matching. Further expansion of the donor pool will be possible by combining paired donation with desensitization [38]. This local paired exchange program increased the live donation rate by ∼30%, with the most prominent increment in highly sensitized patients. Regular communication between the transplant programs and the HLA laboratories is essential to determine the threshold for determining unacceptable antibodies and the programs willingness to process transplanting across low-level DSA.

Diagnostics Challenges in Transplantation Pathology

  1. Top of page
  2. Abstract
  3. Introduction
  4. Antibody Testing—Paradigms Lost
  5. Laboratory Support Strategies for the Incompatible Recipient
  6. Diagnostics Challenges in Transplantation Pathology
  7. Disclosure
  8. References

SWOT analysis of the Banff system

The Banff classification represents international consensus for histopathological diagnosis in allograft biopsies. The Banff process established consensus around how lesions in biopsies are scored and translated into diagnoses and had significant impact on improving patient care [39]. However such self-organized consensus process is associated with strengths, weaknesses, opportunities and threats [40]. The major strength is the inclusiveness and cohesiveness of the international community involved in the Banff process coupled with continuous refinement of the classification based on emerging data. An example of such a self-correction is the elimination of the term chronic allograft nephropathy [41] and the introduction of specific diagnosis into the classification, in particular ABMR [42]. This represented a first step toward precision diagnostics in transplantation and allowed for more tailored treatment of ‘rejection'. However, a major weakness from the beginning was and still is that the lesions in the biopsies are empirically derived and thus not specific for any disease entities. Furthermore, their assessment is prone to subjective interpretation and limited reproducibility [4]. The opportunity is to incorporate new diagnostic tools into the Banff classification, which add value to histology and improve diagnostic precision. Currently several Banff working groups focus on a data-driven, evidence-based refinement of the classification [2].

Diagnosing antibody-mediated rejection—the significance of C4d

The diagnosis of ABMR in renal allografts requires three components: antibody-mediated tissue pathology, evidence of complement activation, presence of DSA [42]. C4d deposition in peritubular capillaries is considered a biomarker for complement activation in the allograft. But there are caveats: the methods for detecting C4d are of variable sensitivity and reproducibility, cases of complement-independent, i.e. C4d-negative ABMR are described, complement activation can happen outside the allograft, and C4d can wax and wean, in particular after treatment [2, 43-45]. Transplant vasculopathy after heart transplantation is observed in wild-type mice infused with noncomplement fixing DSA and also in complement-deficient mice [46], providing evidence that C4d is not only of limited sensitivity but also of limited specificity. Furthermore none of the histological lesions associated with antibody-mediated injury is absolutely specific. Thus the unmet need is for more specific diagnostic tools for antibody-mediated injury in the tissue. Recent molecular studies revealed sets of transcripts associated with endothelial cell activation and NK cell infiltration to be associated with the presence of DSA and histological features of microcirculation inflammation in renal transplants [47-49]. It is expected that a refined diagnostic approach for ABMR will integrate serological, morphological, immunopathological and molecular features.

Infection masquerading as rejection

Infection and rejection share common mechanisms of allograft injury and represent a frequent diagnostic challenge. Infection may trigger TCMR by causing upregulation of MHC antigens and releasing proinflammatory cytokines. Conversely, tissue injury due to ABMR or TCMR can create a milieu that is favorable for the replication of microbes [50]. Infectious mimics of ABMR include sepsis-related acute tubular necrosis, glomerulitis and peritubular capillaritis secondary to cytomegalovirus infection, neutrophilic tubulitis occurring in the setting of acute pyelonephritis, hepatitis C virus associated thrombotic microangiopathy and transplant glomerulopathy and hepatitis B virus associated membranous nephropathy [51, 52]. A misdiagnosis of TCMR is possible if it is not recognized that tubulitis may reflect infection with viruses [53, 54]. Immunohistochemistry, HLA antibody testing, serology, urine cytology, electron microscopy and quantitative PCR can help establish the correct diagnosis. Infectious organisms to keep in mind in allograft organs other than the kidney include viral hepatitis after liver transplantation, Toxoplasmosis and Chaga's disease in the heart, Giardia and Cryptosporidium in the intestine and Pneumocystis in the transplanted lung. Granulomatous T cell infiltrates in any organ should prompt diagnostic work up for mycobacterial and fungal infections. In recent years, effective screening and chemoprophylaxis has significantly reduced the incidence of posttransplant infections.

To biopsy or not. Value added by protocol biopsy in heart transplantation?

The rationale for protocol biopsies in heart transplantation is detection of rejection to permit early intervention. However, under current immunosuppression clinical TCMR after heart transplantation is rare, challenging the need for protocol biopsies. Today's biopsy practice in most centers is therefore to stop protocol biopsies after 1–2 years. But there is still no prospective evidence indicating that patients not being biopsied have the same outcome than those undergoing protocol biopsies. It is possible that new molecular biomarkers in the peripheral blood (e.g. gene expression profiles) may replace protocol biopsies, but comprehensive validation between biopsy findings and blood assessment is still lacking. Currently available molecular tests can detect low rejection grades, but these tests cannot differentiate high-grade rejection from infection and cannot detect ABMR. Robust genotype–phenotype correlation requires the biopsy to be part of an iterative validation process for improved diagnostics in heart transplantation.

Molecular diagnostics in transplantation

After a decade of experience, opinions regarding the usefulness of microarray studies vary widely. Unfortunately, many published analyses are plagued by poor analysis strategies [55]. Some of the common mistakes related to microarray data analysis are [56]: (1) insufficient exploratory data analysis. Visual/graphical exploration can help guide analyses, e.g. do not calculate correlations between variables that are not linearly related [57]. Principal components analysis (PCA) is very useful in this regard. (2) Improper use and interpretation of clustering. Clustering samples based on the genes differentially expressed between phenotypes is not supportive evidence for the clinical significance of the genes. Clustering is not suitable for predicting [55]. Hierarchic clustering generates tree-like relationships within the data even when none truly exists, e.g. when the data are randomly generated. (3) Improper validation. Do not allow any information from the test set (e.g. differentially expressed genes) to leak into the validation procedure. The use of a single data split is highly inefficient, especially for small studies. Cross-validation methods are almost always to be preferred [58, 59]. (4) Bad experimental design, e.g. limited challenge bias—evaluating a diagnostic test using nonchallenging samples, e.g. normals versus rejecting. By leaving out clinically relevant but diagnostically ‘difficult' samples (e.g. borderline cases), reported accuracies, sensitivities, etc. are inflated [60]. Limited challenges bias represents by far the most serious and common problem seen in transplant microarray studies today.

The major challenge in discovering and validating new molecular diagnostics is in the lack of a true gold standard. Histological lesions assessed by the Banff systems are not specific and thus consensus diagnoses based on these lesions are of limited diagnostic accuracy. Therefore, first the diagnostic precision in the biopsy needs to be increased before noninvasive biomarkers can be established based on biopsy diagnoses. Dr. Halloran's group approaches this challenge by creating molecular classifiers based on the best available knowledge for assigning a diagnostic label to the biopsy. Building cross-validated molecular classifiers revealed that variance with the classifier is also biggest with cases at the interface between normal versus abnormal. For example the molecules used by a recently build ABMR classifier are endothelial and NK cell associated thus are derived from the biology underlying antibody-mediated injury.

Another significant unmet need is to reliably assess the donor organ quality and identify those which can be utilized or which have the potential to be improved. Recently a molecular signature which predicts outcome after lung transplantation is being developed [61]. Lungs for donation after cardiac death have a different transcriptional signature compared to donation after brain death [62], they show a strong activation of innate immunity and inflammatory response, which could be downregulated using gene therapy (e.g. AdIL10 gene therapy) [63], thus representing a promising potential therapeutic intervention to recover inferior organs. Using a novel ex vivo perfusion and treatment of inferior lungs, these organs were brought up to performance of transplantable, “good” lungs [64]. Furthermore, the molecular assessment of injured kidneys has been demonstrated to superior to morphology in correlating with function, since light microscopy is essentially unable to detect and quantify acute epithelial injury [65]. This will be very useful in the assessment of potential donor organs. Furthermore, the extent of parenchymal injury also is the best predictor of allograft survival in kidney transplants [66].

Future prospects in transplantation diagnostics

Opportunities lie ahead for improving transplant diagnostics by integrating new molecular technologies to the assessment of submicroscopic pathology and recipient humoral risk. The challenge lies in developing ways to refine the high-dimensional data coming off these “omics” platforms in a way that they can be used for patient care. This requires validation of recent molecular discoveries in prospective studies, including the development of guidelines for quality assurance and reporting standards. The evidence-based integration of the various diagnostic platforms will require close communication between the clinician and laboratory. Applied and integrated biostatistics will be crucial to this end. It is through this interdisciplinary collaboration that diagnostic precision will be accomplished in transplantation.

Disclosure

  1. Top of page
  2. Abstract
  3. Introduction
  4. Antibody Testing—Paradigms Lost
  5. Laboratory Support Strategies for the Incompatible Recipient
  6. Diagnostics Challenges in Transplantation Pathology
  7. Disclosure
  8. References

The following author of this manuscript has conflicts of interest to disclose as described by the American Journal of Transplantation. Dr. P.H. has shares in TSI (Transcriptome Sciences Inc.), a university company with an interest in molecular diagnostics.

The other authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Antibody Testing—Paradigms Lost
  5. Laboratory Support Strategies for the Incompatible Recipient
  6. Diagnostics Challenges in Transplantation Pathology
  7. Disclosure
  8. References