EXPERIMENTAL AND CLINICAL TRANSPLANTATION

Objectives: For decades, the detection of anti-HLA antibodies in candidates for solid-organ transplant has been performed with the traditional complement-dependent cytotoxicity method; this assay has been then integrated with the introduction of solid-phase assays. Over the past 20 years, the Luminex assay has become the most widely used in clinical laboratories due to both increased sensitivity and specificity versus enzyme-linked immunosorbent assay. However, even the Luminex technique has shown some critical issues, and choosing the most reliable method still remains challenging. In this study, we verified the concordance of the results obtained in detecting anti-HLA antibodies with 2 kit vendors that provide reagents for the Luminex platform.

Materials and Methods: We used 314 serum samples from patients on wait lists for solid-organ transplant. Sera were tested with LABScreen Mixed-LSM12 (One Lambda-Thermo Fisher, Canoga Park, CA, USA) and LIFECODES LifeScreen Deluxe-LMX (Gen-Probe-Immucor, Stanford, CT, USA),which we indicated as vendor A and vendor B, respectively. Anti-HLA class I and class II antibody analyses were conducted by verifying the concordance of the results with Cohen kappa coefficient statistics and confidence interval.

Results: The kappa coefficient statistics showed “substantial” reliability for class I (0.61; confidence interval, 0.50-0.73) and “moderate” reliability for class II (0.56; confidence interval, 0.43-0.69). There were no considerable differences in results between the 2 kits regarding overall assignment of negativity or positivity of a sample. Discordant data between positive values for a test and negative for the other were found for samples with weak antibody positivity.

Conclusions: Some discordant data were probably attributable to several factors such as the composition of the kits, the antibody titer in the serum, whether sera were diluted, different washing methods, and type of plate used.

Key words : Concordance, HLA antibody detection, Luminex technology, Transplantation

The presence of alloantibodies directed against human leukocyte antigens (HLA) represents a barrier to transplant and affects a favorable long-term outcome in solid-organ transplant.^1,2 The main immunizing events that trigger alloantibody production are pregnancies, miscarriages, blood transfusions, and previous transplant procedures. Considerable efforts have been made so far for the most correct identification of these antibodies both in the pretransplant and in the posttransplant period.³ For almost 40 years, the technique based on the complement-dependent cytotoxicity (CDC) has been the mainstay in detection of anti-HLA antibodies; over recent years, it has been integrated but not replaced by more sensitive solid-phase assays, such as the enzyme-linked immunosorbent assay and bead-based technology (for example, flow-panel reactivity antibody and the Luminex platform).^4-7

In our study, we used the Luminex system (Luminex Corporation, Austin, TX, USA). It is based on the use of 2 lasers that excite the fluorochromes on beads coated with HLA molecules. The antigen-antibody complex formed binds a second antibody labeled with phycoerythrin. When this complex passes through the detector, a red laser excites the beads and a green laser excites the phycoerythrin. The combination of these 2 signals defines the specificity of the antibodies present in a serum sample.

The composition of the HLA antigen panel on beads varies among kits according to the increase of sensitivity. That is, for screening tests, the beads are coated with HLA-A, HLA-B, and HLA-C molecules for class I and HLA-DR, HLA-DQ, and HLA-DP molecules for class II, belonging to multiple individuals. For the qualitative kit of panel reactivity antibody tests, each bead is coated with an HLA class I and class II phenotype from a single individual. Finally, in single antigen kits, the beads are coated with single recombinant HLA alleles for class I and II. For the Luminex platform, the screening tests for anti-HLA antibodies allow a large number of samples to be verified simultaneously and provide information on the presence or absence of these antibodies in a short runtime compared with the CDC method.

There are unquestionable advantages with the Luminex anti-HLA-antibody assay in organ transplant; however, the regular use of solid-phase multiplex bead arrays could highlight some drawbacks, leaving unresolved the problem about choice of the optimal method for antibody detection.^8-10 Therefore, the aim of our study was to verify the concordance of the results obtained in anti-HLA antibody detection in a group of serum samples with 2 vendor kits that provide reagents for the Luminex platform.

Materials and Methods

Samples
At our center, patients on wait lists for heart or kidney transplant are routinely investigated every 3 months for anti-HLA antibodies with Luminex solid-phase assays. Some patients are also monitored during posttransplant follow-up.

From wait list patients, we randomly selected 314 serum samples that had extra aliquots available for comparisons between the 2 kits. The patient group tested for this study included 218 men and 96 women with mean ± standard deviation age of 55 ± 10.98 years. All patients were white, except for 2 with Asian heritage. Seventeen of 314 patients received deceased donor transplants, 72 women had pregnancies/miscarriages, 93 patients had blood transfusion, and 173 patients were without sensitization events. All patients had negative CDC auto-crossmatch with T and B cells, separately considered. Table 1 shows details on demographics and characteristics of our study’s patient population.

Aliquots of the selected samples were frozen at -80°C, and all used samples were subjected to a single freeze-thaw cycle. Sera were tested with LABScreen Mixed-LSM12 (One Lambda-Thermo Fisher, Canoga Park, CA, USA) and LIFECODES LifeScreen Deluxe-LMX (Gen-Probe-Immucor, Stanford, CT, USA), referred to as vendor A and vendor B, respectively. To minimize the run-to-run variability, the same serum was tested by a single operator with both kits, following manufacturer’s instruction; moreover, discordant samples were retested in duplicate. Briefly, serum obtained by centrifugation of whole blood was incubated with the mixture of beads for 30 minutes to allow antigen-antibody binding. After samples were washed with phosphate-buffered saline to remove everything unbound, a secondary antibody-immunoglobulin G conjugated with a fluorochrome-phycoerythrin was added to allow detection of antigen-antibody complex by the Luminex laser system. The main characteristics of the 2 Luminex kits are summarized in Table 2.

No ethics committee approval was requested since the standard procedures of enrollment to deceased-donor transplant wait lists are covered by the written consent form signed by patients that were obtained at recruitment. All protocols conformed to the ethical guidelines of the 1975 Declaration of Helsinki.

Data analyses
Data from LABScreen Mixed assays were analyzed with Fusion version 2.0 software (One Lambda) on the Luminex 100/200 flow analyzer (Luminex Corporation). To determine the normalized background ratio, the reactivity of each bead was calculated according to the “raw” fluorescence values using the following formula: [(sample N bead - sample negative control bead)/(negative control N bead -negative control negative control bead)].The positive cut-off was calculated according to the relative ratio between the patient sample and negative control; it was fixed at 2.0 for both HLA class I and HLA class II antibodies.

Data from the LIFECODES LifeScreen Deluxe assays were analyzed with its Match IT! version 1.0 software. Three adjusted ratios were calculated for each bead of the HLA class I and class II molecules. Mean fluorescence intensity (MFI) for each sample bead was divided for each of the 3 negative control bead (CON1, CON2, CON3), taking into account the background noise due to bead variation (background adjustment factor; BAF), using the following formula: [(individual bead MFI/CON1 MFI) - BAF] = adjusted ratio 1; [(individual bead MFI/CON2 MFI) - BAF] = adjusted ratio 2; [(individual bead MFI/CON3 MFI) - BAF] = adjusted ratio 3.

A sample was considered positive for a class I and/or class II HLA antibody if any one of the 3 adjusted ratios showed a positive bead reaction for the LIFECODES LifeScreen DeluxeProbe I-01 and Probe II-01 beads (score ≥ 1). A positive value for any 2 of the adjusted ratios indicated a positive bead reaction for all remaining beads (score ≥ 2).

Statistical analyses
Data were analyzed to evaluate the concordance categorically (positive or negative). The results were collected into a contingency table, and the validity of the results obtained with the kit from vendors A and B was evaluated by correcting the estimates due to chance. Indeed, we calculated reliability by calculating the Cohen kappa coefficient (Kc).11 We categorized reliability of the 2 kits based on Kc point estimates,^11,12 with slight reliability having Kc = 0.0-0.20, fair reliability having Kc = 0.21-0.40, moderate reliability having Kc = 0.41-0.60, substantial reliability having Kc = 0.61-0.80, and almost perfect reliability having Kc = 0.81-1.0. We used 95% confidence intervals (CI) to evaluate the meaning of the obtained Kc result.

Figure 1 compares results of each kit in the screening test for detection of anti-HLA class I and class II antibodies. Of the 314 samples tested with the screening test, we found that 222 samples (70.70%) with vendor A (One Lambda) and 235 samples (74.84%) with vendor B (Gen Probe) were negative for both class I and class II HLA antibodies. Regarding sera positive against only class I, we observed 51/314 samples (16.24%) with vendor A and 21/314 samples (6.69%) with vendor B. Conversely, regarding sera positive against only class II, we observed 16/314 samples (5.09%) with vendor A and 37/314 samples (11.78%) with vendor B. Regarding sera positive for both classes, there were 25/314 samples (7.96%) with vendor A and 21/314 samples (6.69%) with vendor B. Therefore, of the 51 positive samples only for class I with vendor A, there were 30 negative with vendor B; of the 37 sera positive only for class II with vendor B, there were 21 sera negative with vendor A.

For data analyses, when we compared the 2 kits, reliability of class I and class II comparisons using Kc results was substantial for class I (Kc = 0.61; 95% CI,0.50-0.73) and moderate for class II (Kc = 0.56; 95% CI, 0.43-0.69) (Table 3). Details on concordance and heterogeneity of the results between the 2 kits (for class I and class II anti-HLA antibodies) are shown in Figure 2. In particular for class I, 36 samples gave positive results with vendor A but were negative with vendor B, whereas only 2 samples were positive with vendor B but negative with vendor A. Moreover, for class II, 10 samples were positive with vendor A but negative with vendor B, and 27 samples were positive with vendor B but negative with vendor A (Figure 2).

The main findings of our study were that we found no considerable differences in results between the 2 kits regarding the attribution of negativity of a sample with screening tests. Conversely, regarding the positive data, we had some interesting observations. As shown in Figure 1 and Table 3, although the positivity of a serum sample for both classes is almost in agreement, when we compared the positivity only regarding 1 of the 2 antibody classes, we observed some differences: the screening kit of vendor A showed more positive cases than the screening kit of vendor B for class I. Conversely, there were more positive cases for class II for vendor B than for vendor A.

Anti-HLA antibody screening is crucial for patients awaiting a transplant and during posttransplant follow-up.¹³ However, there are still some unclear issues, and many reports have discussed limitations of the techniques and other reports state concerns about the significance of these antibodies and their clinical implications in organ transplant.^14-16 In addition, each laboratory has adopted its own cut-off to determine the threshold of positivity, with no standardization to date. It follows that there may be discrepancies between laboratories on the evaluation of positive results of a serum.¹⁷

Traditionally, the status of sensitization of patients has been defined by the CDC method; however, this method has shown some limitations due to the panel composition, the quality of the fresh cells used, and the difficulty to distinguish class II antibodies in the presence of antibodies of class I.^18,19 Conversely, new methodologies such as the Luminex platform have demonstrated both a greater sensitivity and an unambiguous recognition of antibodies against one or both classes.²⁰ The use of bead-based Luminex screening in laboratories has become widespread worldwide and has improved the detection of anti-HLA antibodies. In this context, to improve our understanding of the limitations but also of the potential of Luminex-based technology, we focused our study on interpretative and technical aspects regarding its use. We assessed the results obtained by 2 commercial reagent kits (vendors A and B) for assignment of positivity or negativity.

We recognize that our results are limited by the small number of samples tested and that we could deepen our study by using a single antigen kit for finding antibody specificities. Nevertheless, our findings fully agree with the above-mentioned issues; we should not underestimate the factors related to the composition of the kits (representativeness of the specific HLA allelic level, origin, nature, and density of HLA molecules coated on beads), antibody titer present in the serum, the use of serum diluted or not diluted, different washing methods, and type of plate used (spin/flick vs filter tray). Overall, these are the likely causes of some data not fully concordant.¹⁵ We found that samples with weak antibody positivity were affected mainly by the critical issue of heterogeneity of results between the 2 kits; however, strongly positive samples revealed the same result. Several variables are involved in the preparation of products intended for use in detection of HLA antibodies and not least is the subjectivity regarding data interpretation. Increasing efforts are now directed to developing guidelines for the standardization of methods and interpretation of these data, for reliability results to be as consistent as possible.^5,17

In conclusion, our data reflect the still open debate on the choice of the most reliable techniques to use in the study of anti-HLA antibodies. Furthermore, by preserving the reproducibility of the results in a sample serum through the use of kits of different composition, and when possible from different vendors, a more accurate and reliable study of each antibody pattern could be achieved.

Volume: 15 Issue: 6 December 2017

FULL TEXT