Secondary Logo

Journal Logo

HLA Diagnostics: Evaluating DSA Strength by Titration

Tambur, Anat, R., DMD, PhD1; Wiebe, Chris, MD2,3

doi: 10.1097/TP.0000000000001817

HLA antibodies, and specifically donor-specific-HLA antibodies, play a key role in transplant-related diagnostics and decision-making. It is now clear that the simple differentiation between absence and presence of HLA donor-specific antibodies does not provide sufficient granularity in all clinical circumstances. It addition, knowledge of HLA antibody strength has potential utility at different stages of recipient evaluation along the transplant timeline from initial pretransplant evaluation, evaluation of a specific potential donor, and posttransplant monitoring for de novo donor-specific antibodies. Here we compare data evaluating HLA antibody strength using the conventional IgG-mean fluorescence intensity approach with serial dilution studies (titration) and of C1q binding (C1q-mean fluorescence intensity). The added value of titration studies along the 3 milestones of the transplant cycle is emphasized.

1 Transplant Immunology Laboratory, Comprehensive Transplant Center, Northwestern University, Chicago, IL.

2 Department of Medicine, University of Manitoba, Winnipeg, Manitoba, CA.

3 Diagnostic Services of Manitoba, Winnipeg, Manitoba, CA.

Received 30 October 2016. Revision received 16 January 2017.

Accepted 12 February 2017.

The authors declare no funding or conflicts of interest.

Correspondence: Anat R. Tambur, DMD, PhD, Transplant Immunology Laboratory Comprehensive Transplant Center, Fienberg School of Medicine, Northwestern University, Chicago, IL. (

HLA antibody testing serves to provide several diagnostic purposes, depending on the status of the patient along the transplant pathway: (I) initially, when a patient is evaluated as a transplant candidate—HLA antibody testing provides an estimate of the patient’s alloimmune status; (II) during evaluation of a potential donor—testing is done to assess compatibility and risk stratification for a potential transplant; and, (III) posttransplantation, evaluating humoral responses against the allograft.

Originally, HLA antibody assessment attempted to provide “transplantibility index” based on a measure of panel-reactive antibody (PRA, or the % of HLA antigens against which a patient had antibodies, within the panel tested). A refinement of PRA calculation was introduced when centers had to report only those antigens considered as unacceptable (calculated PRA), based on a uniform panel of antigens reflecting HLA frequencies of previous United Network for Organ Sharing-listed donors.1,2 This latter approach enforced accountability within the transplant community, whereby antigens listed as unacceptable precluded patients from being considered as potential recipients for donors carrying these alleles. A center-specific strategy had to be put in place to ensure more due diligence in registration of unacceptable antigens. Although PRA and calculated PRA approaches take into account only the breadth of sensitization, antibody load/strength, a significant additional dimension beyond the presence or absence of HLA antibodies, has received less attention.

The most common tool to evaluate the presence of HLA antibodies is the Luminex based single-antigen bead (SAB) assay. Two vendors provide rather similar platforms, and the readout provided is mean fluorescence intensity (MFI) units. This scale was supposed to provide a reflection of how much antibody is bound to the antigen coating the bead, although both vendors refrained from claiming their assay is quantitative. Notwithstanding, given that the fluorescence is emitted as a result of a secondary antibody binding to the HLA-specific antibody, the common expectation was that the assay is at least semiquantitative and the emission should be proportional to the amount of bound antibody. Indeed, in many cases, there is a correlation between the amount of HLA antibody present and the MFI units. However, antibody binding to the beads, or secondary antibody binding to the HLA antibody can be affected by multiple factors.3-6 A few confounding factors, especially those that are serum/patient specific, were reported recently. For example, multiple publications reported on the role of complement in contributing to an inhibitory effect (colloquially addressed as “prozone effect”). By treating specific serum samples with ethylenediaminetetraacetic acid (EDTA), dithiothreitol, or preheating the serum before performing the assay, most of the inhibitory effect can be eliminated or reduced substantially.7-10 It is important to note that not all antibodies within a single serum sample will be affected by inhibition to the same extent, and some antibody specificities may not be affected at all.11 As a result, the magnitude of inhibition cannot be appreciated unless dilution studies are performed. Previously we reported that about 70% of highly sensitized patients have beads demonstrating inhibition.11 Importantly, most articles addressing inhibition have not performed serial dilution and thus it is impossible to judge whether all inhibition was removed.

A poor correlation between MFI values and antibody load is also observed in cases where the antibody tested recognizes a target shared by many of the multiplexed antigens. In this case, each antigen binds a fraction of the total antibody-load resulting in MFI values that underrepresent antibody strength. In this case, recording the MFI value alone is likely to under represent the strength of the antibody.12,13

An additional serum-specific confounding factor is that some serum contain antibody levels that saturate the antigen on the beads. Our center has been using titration/serial-dilution studies for over a decade to detect and quantify this phenomenon. Although we do not commonly dilute the serum above a titer of 1:1024, we have observed patients with antibody loads of 1:16 384, 1:32 768, and even 1:65 536 titers. These are inconceivable loads of antibody that are not likely to respond to any treatment or desensitization strategy. However, these extreme levels of antibody cannot be fully appreciated by using the routine SAB assay because the antigen coating the beads becomes saturated at much lower levels of antibody.

Back to Top | Article Outline


To better determine the correlation between IgG-MFI, C1q-MFI and dilution-studies/titer units we studied a cohort of 48 highly sensitized patients awaiting transplantation and under evaluation for potential desensitization protocols—43 of 48 with class I antibodies, 45 of 48 with class II antibodies. All serum samples were treated to reduce potential interfering factors (“prozone”). The total number of SAB data points was 8491 (4171 class I and 4320 class II). For this study, a positive SAB response was arbitrarily chosen as 1000 IgG-MFI units. Only positive data points are presented in the following figures. Titer assignment was based on the last dilution before MFI dropped below 1000 units.

Figure 1A depicts the association between IgG-MFI and titer at the individual HLA-loci. The mean IgG-MFI value per titer for each of the HLA-Loci is shown in Figure 1B and Table 1. Several observations can be clearly documented. All beads, regardless of the HLA-locus, reach saturation around the titer of 1:1024. Higher titers result in only minimal mean increase in IgG-MFI increases and the variability between the different data points is high. Importantly, a linear correlation between IgG-MFI and titer values is observed up to an IgG-MFI of about 10 000, or a titer of 1:32. This observation is critical as it helps to explain the perception that IgG-MFI values can provide an accurate indication of antibody titer. However, beyond a titer of 1:32 the correlation is poor. Thus, it seems that a linear correlation between IgG-MFI and titer values does exist but it is limited only to the low titer antibodies.





Another important observation is the fact that HLA-DQ antibodies show the highest titer (up to 1:65 516) and variability, followed by antibodies in HLA-A and HLA-DRB3/4/5 loci. The significance of these observations is not clear. The best correlation between IgG-MFI and titer is seen for HLA-DRB1 antibodies. None of the patients in this cohort had HLA-DP antibodies with titer > 1:512, thus limiting our ability to comment on HLA-DP antibodies. The data presented here clearly demonstrate that high MFI values do not equate high titer, a misconception that is frequently held in our field. Titer can be determined only by performing dilution studies.

The same serum samples were run in parallel using the C1q assay and the correlation between C1q-MFI values and titer data, per locus, is presented in Figure 2A. The mean C1q-MFI value is plotted against the titer, per locus, in Figure 2B and presented in Table 2. Consistent with previous reports, activation of the complement cascade and C1q binding only occurs when there is sufficient antibody—typically a titer of 1:32 to 1:64 (Figure 2B).11,14 For titers, 1:64 or greater variations in the association between C1q and titer may be explained by saturation of beads or Ig-isotype of these antibodies but this was not tested. The number of data point for HLA-DP antibodies is rather small, but it looks like all higher titer DP antibodies support binding of C1q. The highest ranges of C1q-MFI variability in relation to C1q binding is seen for HLA-B and HLA-DQ antibodies. Most loci show a decrease in C1q-MFI as titer increase above a certain point, usually more than 1:1024. We can only speculate that high levels of C1q component bound may inhibit the binding of the secondary/reported antibody.





Antigen/antibody binding is affected by different properties leading to changes in binding affinity and avidity. Those specific characteristics cannot be measured by the Luminex assays. However, as we have previously reported,11,15 and as demonstrated in Figure 3, antibodies with similar MFI units can have significantly different titer values. HLA locus specific comparisons of IgG-MFI units, C1q-MFI units, and IgG-titer values plotted in sets of 5000 IgG-MFI units, are presented in Figure 3 and Table 3 to demonstrate this point. Looking at the spread of titer values for each MFI range it is evident that there is some correlation between the mean values in each cluster—on the overall population level. However, when faced with a specific patient scenario, having a particular donor-specific antibody, the range of potential titers per MFI cluster can be quite wide. For example, the range of HLA-A antibody titers within the cluster of 1000 to 5000 IgG-MFI units contain antibodies that are 10 fold different in titer (1:1-1:1024). Similar ranges are observed for the IgG-MFI clusters up to greater than 20 000 MFI. Thus, although statistically there is a linear relationship between IgG-MFI and titers at least at the lower MFI values (roughly up to 10 000 MFI units), titer may be a better guide for clinical decisions in individual patients.





In a recent study, we compared IgG-MFI, C1q-MFI and dilution studies to measure efficacy of desensitization using Rituximab as induction followed by multiple cycles of plasmapheresis/low-dose intravenous immunoglobulin.15 Antibody levels were measured by the three approaches immediately before antibody removal therapy and immediately at the completion of treatment to determine “delta-reduction” of antibody load. Neither IgG-MFI nor C1q-MFI showed consistent reduction of antibody load, leading to the enigmatic observation that different antibodies within the same serum sample respond differently to treatment. Using dilution studies we demonstrated a very tight and consistent reduction of antibody titer/load in line with the expected response to treatment. Hence we concluded that for treatment purposes, measuring antibody load by titration is more accurate and predicable than using MFI units through either the IgG or C1q assays.

Back to Top | Article Outline


De novo donor-specific antibodies (dnDSA) can develop at any point posttransplantation and have been associated with a greater risk of graft dysfunction and loss.16,17 Serial serum monitoring in low-risk renal transplant recipients has documented that 2 of 3 dnDSA can be detected in the serum before the development of clinical dysfunction.16 After dnDSA development heterogeneity exists in the progression to clinical dysfunction and graft loss that is partially, but not entirely captured by clinical and pathologic risk factors available when dnDSA develops.18 Using dnDSA titers or MFI to aid risk assessment is intriguing because it is noninvasive, rapid, and relatively inexpensive. In addition, accurate risk stratification would allow physicians to target patients with high-risk dnDSA using more aggressive immunosuppression while sparing those with low-risk dnDSA the toxicity and cost associated with potent immunosuppression.

As mentioned above, MFI has often been used as a surrogate of antibody strength in studies dealing with HLA antibodies, including dnDSA. Unfortunately, studies correlating MFI, C1q, or C3d status with clinical outcomes are almost universally a mixture of pretransplant DSA and dnDSA, making it difficult to understand the significance of dnDSA MFI at the time of its development.19-21

A recent was the first to titrate dnDSA at the time of first development and correlate dnDSA titer with histology, clinical dysfunction, and graft loss.18 dnDSA C1q status was evaluated and EDTA was used to prevent complement inhibition. Testing serum with or without EDTA revealed that 1.2% of all beads (17.2% of dnDSA beads) were affected by complement inhibition with the median MFI increasing from 3659 (range 7–16 475) to 20 275 (range, 9666-24 463) after EDTA treatment. There was a strong correlation (R2 = 0.8) between the mean MFI of EDTA-treated serum and dnDSA titer after serial dilutions of 1:16 (mean MFI 385 ± 512), 1:64 (mean MFI 2037 ± 2595), and 1:024 (mean MFI 14 762 ± 6101). The correlation between MFI and dnDSA titer was lower (R2 = 0.6) when sera was not pretreated with EDTA. Recipients in this study had a progressively higher risk of graft loss (17%, 29%, and 63%) with increasing dnDSA titer of 1:16, 1:64, 1:1024 respectively (P < 0.01). dnDSA titer also correlated with antibody-mediated rejection and T cell mediated rejection in biopsies done at the time of dnDSA development. However, once the time from dnDSA development to graft loss was considered, a Kapan-Meier graft survival analysis showed no correlation between allograft loss and dnDSA titer (P = 0.14). Furthermore, multivariate Cox models of post-dnDSA survival revealed that the clinical phenotype of the recipient (subclinical vs. clinical at the time of dnDSA development) and their medication adherence were the only independent multivariate predictors of graft loss. After adjusting for clinical phenotype and adherence dnDSA titer (ie, titer ≥1:64, or ≥1:1024) was not correlated with graft loss or estimated glomerular filtration rate (eGFR) decline. Interestingly, there were no patients who had both a subclinical phenotype and dnDSA titer less than 1:64 that progressed to graft loss (8/70 dnDSA recipients). Thus, a combination of low titer and subclinical phenotype may predict low risk for graft loss; however, larger cohorts will be required to evaluate this using multivariate models.

An important observation from the above data is that the IgG MFI correlated better with titer and C1q status in the posttransplant dnDSA study compared with the pretransplant and desensitization studies mentioned above. There are multiple reasons that may explain the discrepancy: (a) the study done on posttransplant dnDSA generally deals with much lower titer antibodies (patients were already immunosuppressed) in MFI ranges where the correlation is generally better; (b) in the posttransplant dnDSA study all sera were treated with EDTA to remove complement inhibition whereas the heat inactivation was used in some of the pretransplant sera; (c) the EDTA concentration used in the posttransplant dnDSA study was significantly greater (25 mM vs. 60 mM) which may have resulted in more complete elimination of the complement inhibition.

Although there are no other studies of dnDSA titer correlated with clinical outcomes, MFI has been correlated with outcome in previous reports. DeVos et al.22 studied a cohort of 48 patients with HLA-DQ dnDSA and found no correlation between antibody-mediated rejection or graft loss when dnDSA were categorized as weak (MFI, 2000-4000); moderate (MFI, 4000-8000); strong (MFI, 8000-15 000); or very strong (MFI ≥ 15 000). Similarly, Dieplinger et al23 studied a small cohort (n = 24) of renal transplant recipients and found no significant association with the median MFI at first detection and 25% or greater eGFR decline. Although peak MFI in the first year after dnDSA detection and delta MFI slope over time correlated with eGFR decline, no attempts were made to adjust for the type or duration of treatment recipients may have received post-dnDSA. A common issue in these and other studies is that complement inhibition was not prevented routinely nor titration used to capture the full picture of dnDSA strength.

Although some have argued that the C1q assay (covered in more detail in other reports in this issue) can uncover recipients whose dnDSA are high risk despite a low MFI, recent data indicates patients with low IgG MFI C1q-positive DSA are in fact the beads affected by complement inhibition. In EDTA-treated serum all C1q-positive beads had an MFI of 10 126 or greater (range, 10 126-24 462) and EDTA MFI predicted C1q status with an area under the curve 0.99.18 If sera were not pretreated with EDTA, then the same C1q-positive beads had an MFI range of 7 to 20 853 (median, 7859). Schaub et al24 also used EDTA-treated sera and found MFI to be an excellent predictor C1q status (area under the curve, 0.98). Because most published reports have not used EDTA, dilution, or other mechanisms to prevent complement inhibition it is not surprising that MFI cutoffs that predict C1q positivity have varied from 6237 to 14 154 with a broad range of sensitivities and specificities.24-27 Thus, it is unclear if any meaningful gain in risk prediction is obtained from the C1q assay beyond knowledge of the MFI in EDTA-treated sera or knowledge of the dnDSA titer.

Back to Top | Article Outline


The data presented here emphasize the inaccuracy of MFI as the sole measure of antibody strength and highlight the utility of dilution studies as the best measure of antibody load. As long as complement inhibition has been reduced or prevented, low-level IgG-MFI values provide a reasonable assessment of antibody strength. In the presence of complement inhibition, epitope spread, or bead saturation, the correlation between antibody strength and MFI is significantly jeopardized making it difficult to interpret studies which have tried to correlate MFI and clinical outcomes without addressing these limitations. This study adds to previous evidence suggesting that much of the C1q binding capabilities are associated with increased antibody load (quantity) rather than an inherent quality of the antibody.11,14,28

In the pretransplant assessment, many patients do not require precise antibody titration because the antibody in question is deemed unacceptable. Alternatively, if desensitization is contemplated (eg, highly sensitized patient with a potential living donor) – dilution studies are the only method to capture the true strength of antibodies, providing guidance for the likelihood of effectively reducing antibodies to acceptable levels.15 Furthermore, dilution studies can monitor the efficacy of desensitization or response to treatment in antibody-mediated rejection.

In the posttransplant assessment, it is important to consider dnDSA strength in the context of clinical phenotype, medication adherence, and allograft histology. Recipients presenting with clinical dysfunction and nonadherence at the time of dnDSA development are known to be high risk of progression to graft loss and neither dnDSA MFI nor titer provide additional risk stratification after adjustment for these risk factors.18 Importantly, for both dnDSA and pretransplant DSA, serial measurements of antibody strength to assess therapeutic success necessitate accurate evaluation after taking into account epitope spread, bead saturation, and complement inhibition.

Back to Top | Article Outline


1. Cecka JM, Kucheryavaya AY, Reinsmoen NL, et al. Calculated PRA: initial results show benefits for sensitized patients and a reduction in positive crossmatches. Am J Transplant. 2011;11:719–724.
2. Cecka JM. Calculated PRA (CPRA): the new measure of sensitization for transplant candidates. Am J Transplant. 2010;10:26–29.
3. Tait BD, Süsal C, Gebel HM, et al. Consensus guidelines on the testing and clinical management issues associated with HLA and non-HLA antibodies in transplantation. Transplantation. 2013;95:19–47.
4. Zachary AA, Lucas DP, Detrick B, et al. Naturally occurring interference in luminex assays for hla-specific antibodies: characteristics and resolution. Hum Immunol. 2009;70:496–501.
5. Gebel HM, Bray RA. HLA antibody detection with solid phase assays: great expectations or expectations too great?. Am J Transplant. 2014;14:1964–1975.
6. Reed EF, Rao P, Zhang Z, et al. Comprehensive assessment and standardization of solid phase multiplex-bead arrays for the detection of antibodies to HLA. Am J Transplant. 2013;13:1859–1870.
7. Schnaidt M, Weinstock C, Jurisic M, et al. HLA antibody specification using single-antigen beads—a technical solution for the prozone effect. Transplantation. 2011;92:510–515.
8. Visentin J, Vigata M, Daburon S, et al. Deciphering complement interference in anti-human leukocyte antigen antibody detection with flow beads assays. Transplantation. 2014;98:625–631.
9. Kosmoliaptsis V, O'Rourke C, Bradley JA, et al. Improved Luminex-based human leukocyte antigen-specific antibody screening using dithiothreitol-treated sera. Hum Immunol. 2010;71:45–49.
10. Schwaiger E, Wahrmann M, Bond G, et al. Complement component C3 activation: the leading cause of the prozone phenomenon affecting HLA antibody detection on single-antigen beads. Transplantation. 2014;97:1279–1285.
11. Tambur AR, Herrera ND, Haarberg KM, et al. Assessing antibody strength: comparison of MFI, c1q, and titer information. Am J Transplant. 2015;15:2421–2430.
12. Konvalinka A, Tinckam K. Utility of HLA antibody testing in kidney transplantation. J Am Soc Nephrol. 2015;26:1489–1502.
13. Tambur AR, Lavee J. Incorporating human leukocyte antibody results into clinical practice. J Heart Lung Transplant. 2016;35:851–856.
14. Diebolder CA, Beurskens FJ, de Jong RN, et al. Complement is activated by IgG hexamers assembled at the cell surface. Science. 2014;343:1260–1263.
15. Tambur AR, Glotz D, Herrera ND, et al. Can solid phase assays be better utilized to measure efficacy of antibody removal therapies?. Hum Immunol. 2016;77:624–630.
16. Wiebe C, Gibson IW, Blydt-Hansen TD, et al. Rates and determinants of progression to graft failure in kidney allograft recipients with de novo donor-specific antibody. Am J Transplant. 2015;15:2921–2930.
17. Wiebe C, Gibson IW, Blydt-Hansen TD, et al. Evolution and clinical pathologic correlations of de novo donor-specific HLA antibody post kidney transplant. Am J Transplant. 2012;12:1157–1167.
18. Wiebe C, Gareau AJ, Pochinco D, et al. Evaluation of C1q status and titer of de novo donor-specific antibodies as predictors of allograft survival. Am J Transplant. 17:703–711, [published online August 19 2016].
19. Fichtner A, Süsal C, Höcker B, et al. Association of c1q-fixing DSA with late graft failure in pediatric renal transplant recipients. Pediatr Nephrol. 2016;31:1157–1166.
20. Lefaucheur C, Viglietti D, Bentlejewski C, et al. IgG donor-specific anti-human HLA antibody subclasses and kidney allograft antibody-mediated injury. J Am Soc Nephrol. 2015;27:293–304.
21. Loupy A, Lefaucheur C, Vernerey D, et al. Complement-Binding anti-HLA antibodies and kidney-allograft survival. N Engl J Med. 2013;369:1215–1226.
22. DeVos JM, Gaber AO, Knight RJ, et al. Donor-specific HLA-DQ antibodies may contribute to poor graft outcome after renal transplantation. Kidney Int. 2012;82:598–604.
23. Dieplinger G, Everly MJ, Rebellato LM, et al. Changes in successive measures of de novo donor-specific anti-human leukocyte antigen antibodies intensity and the development of allograft dysfunction. Transplantation. 2014;98:1097–1104.
24. Schaub S, Hönger G, Koller MT, et al. Determinants of c1q binding in the single antigen bead assay. Transplantation. 2014;98:387–393.
25. Comoli P, Cioni M, Tagliamacco A, et al. Acquisition of C3d-binding activity by de novo donor-specific HLA antibodies correlates with graft loss in nonsensitized pediatric kidney recipients. Am J Transplant. 2016;16:2106–2116.
26. Guidicelli G, Guerville F, Lepreux S, et al. Non-complement-binding de novo donor-specific anti-HLA antibodies and kidney allograft survival. J Am Soc Nephrol. 2015;27:615–625.
27. Lachmann N, Todorova K, Schulze H, et al. Systematic comparison of four cell- and luminex-based methods for assessment of complement-activating HLA antibodies. Transplantation. 2013;95:694–700.
28. Yell M, Muth BL, Kaufman DB, et al. C1q binding activity of de novo donor-specific HLA antibodies in renal transplant recipients with and without antibody-mediated rejection. Transplantation. 2015;99:1151–1155.
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.