Report on the Fourth International Granulocyte Immunology Workshop: progress toward quality assessment

I M M U N O H E M A T O L O G Y

Report on the Fourth International GranulocyteImmunology Workshop: progress toward quality assessment

Geoffrey Lucas, Susan Rogers, Masja de Haas, Leendert Porcelijn, and Juergen Bux

BACKGROUND: A formal quality assurance (QA)scheme has been established to facilitate proficiencytesting for granulocyte antibodies and antigens.STUDY DESIGN AND METHODS: Fifteen laboratoriesparticipated in the Fourth International Granulocyte Im-munology Workshop. The main objective of the work-shop was to establish a formal QA scheme for granu-locyte serology and molecular typing methods. Asecondary objective was to determine the relative sensi-tivities of the granulocyte immunofluorescence test,granulocyte agglutination test, and MoAb immobilizationassays using defined antisera and protocols.RESULTS: Laboratories scored between 16.7 and 100percent (mean, 57.5%) of the maximum available in theserologic part of this QA exercise. There were particularproblems in detecting granulocyte-specific human neu-trophil antigen-1 (HNA-1a) IgM antibodies and HNA-2aantibodies in the presence of HNA-1b antibodies. Thegranulocyte immunofluorescence test was more sensi-tive than the granulocyte agglutination test in titrationstudies, but the latter method more readily identified thepresence of HNA-3a antibodies. HNA genotyping wasgenerally well performed, with nine laboratories obtain-ing 100-percent correct results for HNA-1a, HNA-1b,and HNA-1c.CONCLUSIONS: There is a need to standardize thedetection of granulocyte-specific antibodies. Laborato-ries with good performance tended to use two methodsfor detecting granulocyte-specific antibodies and anHNA-typed panel of granulocytes. The use of a methodfor elucidating mixtures of granulocyte- and lymphocyte-reactive antibodies (e.g., MoAb immobilization assay)and the use of methods for detecting both cytotoxic andnoncytotoxic HLA class I antibodies were also associ-ated with a higher than average performance.

Granulocyte-specific antibodies have estab-lished clinical significance in disorders suchas neonatal alloimmune neutropenia, febriletransfusion reactions, transfusion-related acute

lung injury (TRALI), drug-induced neutropenias, immuneneutropenia following bone marrow transplantation, andboth childhood and adult autoimmune neutropenias.However, the detection of these antibodies remains tech-nically difficult, and only a small number of laboratoriesworldwide undertake this work. Three granulocyte im-munology international workshops have previously beenorganized to improve standardization and to establish aforum for discussion. All three previous workshops in-cluded proficiency testing as part of the studies, but thenumber of sera and/or DNA samples, the method of scor-ing, and the interpretation of results varied in each study.In the first workshop, despite the inclusion of definedhuman sera to enable participants to prepare a typedpanel, the main problem encountered by the laboratorieswas the correct identification of antibody specificity.1 Thesecond workshop2 enabled the interlaboratory assess-ment of both the MoAb immobilization of granulocyte

ABBREVIATIONS: GAT = granulocyte agglutination test; GCLT

= granulocyte chemiluminescence test; GIFT = granulocyte im-

munofluorescence test; GIFT-F = GIFT with a flow cytometric

endpoint; GIFT-M = GIFT with a microscopic endpoint; HNA =

human neutrophil antigen; IBGRL = International Blood Group

Reference Laboratory; LIFT = lymphocyte immunofluorescence

test; MAIGA = MoAb immobilization of granulocyte antigens;

MAIPA = MoAb immobilization of platelet antigens; MPHA =

mixed passive hemagglutination assay; QA = quality assurance;

TRALI = transfusion-related acute lung injury.

From the International Blood Group Reference Laboratory,

Bristol, UK; Central Laboratory of the Red Cross, Amsterdam,

Netherlands; Institute for Clinical Immunology and Transfu-

sion Medicine, Justus Liebig University, Giessen, Germany.

Address reprint requests to: Geoffrey Lucas, PhD, Interna-

tional Blood Group Reference Laboratory, Southmead Road, Bris-

tol BS10 5ND, United Kingdom; e-mail: [email protected].

Received for publication June 20, 2001; revision received

October 17, 2001, and accepted November 8, 2001.

TRANSFUSION 2002;42:462-468.

462 TRANSFUSION Volume 42, April 2002

antigen (MAIGA) assay3,4 and human neutrophil antigen(HNA)5 genotyping for HNA-1a and HNA-1b.6 The con-clusion from the results of the second workshop2 was thatgranulocyte antibodies should be investigated using aminimum of two methods: the granulocyte immunoflu-orescence test (GIFT)7 and the granulocyte agglutinationtest (GAT).8 The results confirmed that although the GAThas low sensitivity compared with the GIFT, it was theonly reliable method for detecting HNA-3a antibodies.The third workshop enabled the first interlaboratory as-sessment of HNA-1c genotyping and focused on theevaluation of MoAbs for use as HNA typing reagents andas capture antibodies in the MAIGA assay. The results ofthis workshop were presented at the 5th European Sym-posium on Platelet and Granulocyte Immunobiology inS’Agaro, Spain in May 1998.

The primary objective of this workshop was to estab-lish a formal quality assurance (QA) scheme using adocumented scoring system, which would be applied tosubsequent workshops and would enable progress to bemonitored. Furthermore, participation in a formal QAscheme would satisfy a regulatory requirement for granu-locyte immunology laboratories wishing to become ac-credited. There are no existing QA schemes for granulo-cyte immunology because the small number oflaboratories has precluded the establishment of nationalschemes; hence, it was appropriate to establish an inter-national QA scheme. An additional objective of this work-shop was to determine the relative sensitivity of differentassays for the detection of granulocyte-specific antibod-ies. The results of this workshop were discussed in detailat the 6th European Symposium on Platelet, Granulocyte,and Red Cell Immunobiology in Amsterdam, Nether-lands, in September 2000.

MATERIALS AND METHODSParticipantsFifteen laboratories participated in the workshop (Table1). The order of the laboratories in Table 1 does not cor-

respond to the numeric code used to identify the labora-tories in subsequent tables.

The workshop was organized by the InternationalBlood Group Reference Laboratory (IBGRL) in consulta-tion with the Granulocyte Immunology Steering Commit-tee, which consisted of Juergen Bux, MD (Giessen), Masjade Haas, MD, PhD (Amsterdam), Leendert Porcelijn, MD(Amsterdam), and Geoff Lucas, PhD (Bristol). The threelaboratories associated with the steering committeemembers have considerable experience in granulocyteimmunology and were regarded as reference laboratoriesin cases where there was no clear-cut consensus regard-ing results.

The staff of the granulocyte immunology laboratoryat the IBGRL was divided into two groups: an organizingteam and a participating team. The members of the par-ticipating team were not involved in archiving or process-ing of submitted samples, preworkshop testing, and dis-cussions about the possible inclusion of sera, nor werethey allowed access to other participant’s results prior tosubmitting their own results or the workshop deadline. Inthis way, the IBGRL was able to participate in the work-shop on the same basis as other laboratories.

Sample processing and preworkshop testingParticipants were invited to submit serum and/or plasmasamples containing granulocyte-reactive antibodies forinclusion in the workshop. Plasma samples were recalci-fied, treated with thrombin, and filtered. Sera 1 and 5were diluted with serum from untransfused, male groupO blood donors. The serum from these blood donors hadbeen found to be unreactive in the GIFT, the lymphocyteimmunofluorescence test (LIFT),9 and the granulocytechemiluminescence test (GCLT).10 The serum used as adiluent for Serum 1 was from an HNA-1b/HNA-1b donor,thereby avoiding the possibility that HNA-1a antibodiesmight be neutralized by soluble CD16 carrying HNA-1a.

All sera and/or plasma samples were tested at theIBGRL using the previously mentioned methods to con-

TABLE 1. Participants in the Fourth International Granulocyte Immunology WorkshopName Institute City Country

Professor Bierling Hopital Henri Mondor Creteil FranceDr. Bux Justus-Liebig University Giessen GermanyDr. Curtis The Blood Center of Southeastern Wisconsin Milwaukee, WI United StatesMs. Eiber Red Cross North Central Blood Services St. Paul, MN United StatesDr. Flesch Institute of Transfusion Medicine Kiel GermanyDr. Kekomaki Red Cross Blood Transfusion Service Helsinki FinlandDr. Lucas/Ms. Rogers International Blood Group Reference Laboratory Bristol United KingdomDr. Macek Zavod RS za transfuzijo krvi Ljubljana SloveniaDr. Minchinton Australian Red Cross Blood Service Brisbane AustraliaDr. Muniz-Diaz Hospital Sant Pau Barcelona SpainDr. Porcelijn/Dr. de Haas Central Laboratories of the Red Cross Amsterdam The NetherlandsDr. Ribera/Dr. Parra Banc de Sang de I’ICS Barcelona SpainProfessor Shibata/Dr. Miyamoto University of Tokyo Tokyo JapanDr. Valentin-Bouquet Institut de Biologie Nantes FranceProfessor Zupanska Institute of Haematology and Blood Transfusion Warsaw Poland

GRANULOCYTE IMMUNOLOGY WORKSHOP

Volume 42, April 2002 TRANSFUSION 463

firm the antibody specificities described by the submit-ting laboratory. Each sample was coded and divided intoapproximately 1.5-mL aliquots.

The DNA samples were isolated and tested at theInstitute for Clinical Immunology and Transfusion Medi-cine (Giessen, Germany).

Design of the workshop and distributionof specimensEach participant laboratory was supplied with a detailedset of instructions and response sheets for each part ofthe workshop to ensure that results were reported in auniform manner. The serologic part of the QA exerciseconsisted of six serum and/or plasma samples, and themolecular typing part of the QA exercise consisted of twoDNA samples. All workshop samples were shipped on dryice to ensure uniform sample quality.

QA schemeSerologic investigations. Laboratories were invited

to detect and identify any granulocyte-reactive antibod-ies present in the six samples by using their own granu-locyte panel and “in-house” laboratory methods. Thepresence of “granulocyte-specific,” “lymphocyte-reactive,” and “HLA class I antibodies” was recorded as“Yes,” “No,” or “Not determinable” using the responsesheets. Participants were advised that it was their inter-pretation of their results that would be used to evaluatetheir performance. When there was not a clear-cut con-sensus among the laboratory results regarding the iden-tity or presence of granulocyte-specific antibodies in asample, the results obtained by the reference laboratorieswere taken into account. Each participant’s score wascalculated as a percentage of the maximum attainablemark.

Molecular typing. Laboratories were invited to usemolecular typing methods to determine the HNA-1a,HNA-1b, and HNA-1c status of two DNA samples. Fifty �Lof DNA (concentration, 0.2 �g/�L) was sent to each par-ticipant.

Relative sensitivities of different assaysThe sensitivity of different assays was determined by ti-trating three known HNA antisera (Serum A, anti-HNA-1a; Serum B, anti-HNA-2a; and Serum C, anti-HNA-1b).Participants were instructed to prepare doubling dilu-tions of Sera A, B, and C in PBS and/or 0.1-percent BSA(+EDTA if this was normally included) from neat to 1 in4096 and to add one volume of the neat or diluted serumto an equal volume containing the granulocyte suspen-sion. A negative control serum (Serum D) and a positivecontrol serum containing multispecific HLA class I anti-bodies (Serum E) were also provided. The participantsused the following scoring system to grade the results: 0

= negative, 0.5 = doubtful positive, 1 = positive, and 2 =strong positive.

In all methods, except MAIGA assays, Serum A andSerum B were titrated against HNA-1a/HNA-1a and HNA-2a+ granulocytes, respectively, whereas in the MAIGA as-says, Serum A and Serum C were titrated against HNA-1a/HNA-1a and HNA-1b/HNA-1b granulocytes using theCD16 (clone 3G8) MoAb provided. The CD16 MoAb 3G8was supplied at a concentration of 1 mg per mL. A work-ing dilution of 1 in 100 (volume, 10 �L) was found to beoptimal in preworkshop studies, but participants wereasked to perform titration experiments to find the opti-mum dilution for use in their own assay. Three MAIGAassay protocols with different detection endpoints (col-orimetric microplate method, chemiluminescence tubemethod, chemiluminescence microplate method) wereprovided to aid those laboratories that had not estab-lished this method.

Typed granulocyte panel cellsAll laboratories had access to HNA-typed cells, but insome laboratories, the panel was very restricted in termsof both the range of antigens and the number of panelcells (Table 2).

Tests for granulocyte antibodies and antigensThe GIFT and GAT were the most commonly used sero-logic methods in the QA exercise. Fourteen of 15 labora-tories used the GIFT. Ten laboratories consistently used acombination of GIFT and GAT, as recommended follow-ing the second international workshop.2 A further labo-ratory only used the GAT to test selected sera. One labo-ratory used the GIFT and the leukocyte agglutination test.A total of 12 laboratories used the MAIGA assay to detecteither CD16-reactive (n = 11), CD177-reactive11 (n = 6), orCD11/18-reactive (n = 5) antibodies. Two laboratoriessupplemented their investigations by using either theGCLT or the mixed passive hemagglutination assay(MPHA). The following tests were used to detect antilym-phocyte or HLA class I antibodies: the lymphocytotoxicitytest, LIFT, HLA class I ELISA, immune phagocytosis inhi-bition assay, and MAIPA/MAIGA assays. Some laborato-ries used platelet-absorbed (n = 2) or DTT-treated (n = 1)sera (Table 2).

Eleven laboratories used the PCR with site-specificprimers for HNA genotyping. One laboratory used a PCR-RFLP method, and one used a PCR–enzyme-linked mini-sequence assay.

RESULTS

Serologic investigationsAll 15 laboratories reported results for this part of the QAexercise. The conclusions of the laboratories regarding

LUCAS ET AL.


the presence or absence of antibodies and the specific-ity(s) detected are presented in Table 3. The total scoresof the 15 laboratories for the serologic part of the QAexercise ranged from 16.7 to 100.0 percent; the meanscore was 57.5 percent. There were five laboratories thatobtained less than 50 percent of the attainable marks.The results obtained for each sample were as follows:

Sample 1: anti-HNA-1a. Eight out of 15 laboratoriesdetected HNA-1a antibodies in this sample. These anti-bodies were only consistently detected in the GIFT andwere generally unreactive in the GAT. One laboratory re-ported CD16 and CD11b antibodies, whereas the remain-ing six laboratories failed to detect any antibodies.

Sample 2: anti-HNA-3a. Ten out of 15 laboratoriesdetected HNA-3a antibodies in this sample. The GAT wasthe most consistent method in demonstrating the specific-ity of these antibodies, but some laboratories were able todetermine this specificity by using the GIFT or MPHA. Fourlaboratories reported HLA class I antibodies in this sample.

Sample 3: anti-HNA-1b and anti-HNA-2a. Seven outof 15 laboratories detected HNA-1b antibodies in thissample, and an additional three laboratories identifiedthe presence of CD16-reactive antibodies. The majority oflaboratories successfully identifying the HNA-1b antibod-ies used the MAIGA assay. HNA-5a antibodies were re-ported by one laboratory. Six laboratories reported thepresence of HLA class I or lymphocyte-reactive antibod-ies. Two laboratories also reported the presence of HNA-2a antibodies by using a MAIGA assay with the CD177(NB1) MoAb 7D8. A further laboratory reported the pres-ence of granulocyte-specific antibodies in addition to theHNA-1b antibodies. Postworkshop studies in this latterlaboratory confirmed that these antibodies were specificfor HNA-2a, as determined in the MAIGA assay by usingthe 7D8 capture MoAb. The use of the CD177 (NB1)MoAb MEM166 in the MAIGA assay did not facilitate thedetection of the HNA-2b antibodies in this sample. Allthree laboratories identifying the HNA-2a antibodies

TABLE 2. Panel cells and methods used by participants

Laboratorynumber

Phenotype/genotypeof panel cells

Number ofpanel cells

Tests for granulocyte-reactive antibodies Tests forantilymphocytic/

HLA class I antibodiesGIFT-M* GIFT-F† GAT MAIGA Others

1 HNA-1a, 1b, 1c, 2a+, 2a–, 3a+,3a–

5 IgG ✓ LCT, IPI‡

2 HNA-1a, 1b, 2a+, 2a–, 3a+, 3a– 8 IgG MPHA§ LCT�3 HNA-1a, 1b, 1c, 2a+, 2a–, 3a+,

3a–, 4a+, 9a+6 to 10 IgGAM ✓ CD11/18 Platelet

absorption4 HNA-1a, 1b, 2a+, 3a+, 3a– 6 IgGAM ✓ DTT-

treatedsera

LCT

5 HNA-1a, 1b, 1c, 2a+, 2a–,3a+, 3a–

10 IgG IgG ✓ CD16CD11/18

LCT, MAIPA/MAIGA¶for HLA class 1

6 HNA-1a, 1b, 1c, 2a+, 2a–,3a+, 3a–, 4a+, 5a+

9 IgGAM ✓ CD16 LCT

7 HNA-1a, 1b, 1c 4 IgG, IgM ✓ CD16 LIFT8 HNA-1a, 1b, 1c, 2a+, 2a–, 3a+ 8 to 10 IgGAM CD16

CD177LCT, HLA class I

ELISA9 HNA-1a, 1b, 1c, 2a+, 2a–,

3a+, 3a–20 to 22 IgG, IgM ✓ CD16 MAIPA/MAIGA for HLA

class I10 HNA-1a, 1b, 1c, 2a+, 2a–,

3a+, 3a–8 IgG, IgM ✓ CD16

CD177MAIPA/MAIGA for HLA

class I11 HNA-1a, 1b, 1c, 2a+, 2a–,

3a+, 3a–, 4a+9 to 16 IgGAM ✓ CD16

CD177CD11/18

HLA class I ELISA

12 HNA-1a, 1b, 1c, 2a+, 2a–,3a+, 3a–

9 to 11 IgG, IgM ✓ CD16CD177CD11/18

MAIPA/MAIGA for HLAclass I

13 HNA-1a, 1b 2 CD16 LIFT14 HNA-1a, 1b, 2a+, 2a–, 3a+,

3a–6 to 11 IgG, IgM ✓ CD16

CD177GCLT LIFT, LCT, MAIPA for

HLA class I, HLAclass 1 ELISA

15 HNA-1a, 1b, 1c, 2a+, 2a–,3a+, 3a–, 4a+, 5a+

7 to 9 IgGAM CD16CD177CD11/18

LAT** LIFT, MAIGA for HLAclass I

* GIFT with microscopic endpoint.† GIFT with flow cytometric endpoint.‡ LCT = lymphocytotoxicity test; IPI = immune phagocytosis inhibition test.§ Mixed passive hemagglutination.� Lymphocytotoxicity test.¶ MoAb immobilization of platelet/granulocyte antigen.** Leukocyte agglutination test.



were reference laboratories, and this specificity wastherefore included in the QA.

Sample 4: negative control serum. Twelve out of 15laboratories confirmed the absence of granulocyte- orlymphocyte-reactive antibodies in this sample. Threelaboratories detected granulocyte-specific antibodies;two laboratories further identified these as HNA-1b andHNA-1c antibodies, respectively.

Sample 5: anti-HLA Bw4. Fourteen out of 15 labo-ratories detected lymphocyte antibodies; 13 out of theseidentified HLA class I antibodies. Five laboratories iden-tified HNA-1b (n = 2), HNA-1a (n = 1), HNA-3a (n = 1), orCD16 (n = 1) antibodies.

Sample 6: anti-HNA-2a + HLA class I. Nine out of 15laboratories identified the presence of HNA-2a antibod-ies, and 14 laboratories identified the presence of lym-phocyte antibodies; 11 laboratories identified these anti-

bodies as HLA class I specific. All six laboratories using aMAIGA assay for CD177 (NB1) successfully identified theHNA-2a antibodies. Two laboratories reported HNA-1aand HNA-4a antibodies, respectively.

Molecular typingEleven laboratories correctly typed the two DNA samplesfor HNA-1a, HNA-1b, and HNA-1c, as judged by a com-parison with the preworkshop test results. A further twolaboratories correctly genotyped these samples for HNA-1a and HNA-1b but did not attempt genotyping for HNA-1c. Two laboratories did not attempt this part of theworkshop.

Relative sensitivities of different assaysA comparison of the sensitivity of different methods wasmade using defined HNA sera and incubation conditions.

TABLE 3. Summary of participants’ interpretation of results and associated scores*

Laboratoryno.

Specificity identified by submitting laboratory

Totalscore

Overallmaximumscore (%)

Sample 1Anti-HNA-1a (IgM)

Sample 2Anti-HNA-3a

Sample 3Anti-HNA-1b

Sample 4Negative

Sample 5Anti-HLA-Bw4

Sample 6Anti-HNA-2a+ HLA

1 HNA-1a HNA-3a CD16 HNA-1c HLA HNA-2a, HLA 14 58.32 Negative HNA-3a CD16 Negative HLA HNA-2a, HLA 13 54.2

HLA HNA-1b3 Negative HNA-3a Unknown GS† HNA-1b HLA HNA-4a 4 16.7

HLA4 HNA-1a HNA-3a HNA-1b‡ Negative HLA HNA 16 66.7

HNA-3a Lym§5 Negative Lym CD16 Negative HLA Unknown GS 11 45.8

HLA6 HNA-1a HNA-3a HNA-1b Negative Lym Unknown GS 15 62.5

Lym‡ CD16 Lym7 Negative HLA HNA-1b‡ Negative ND ND‡ 8 33.3

8 Negative Lym CD16 Unknown GS HLA HNA-2a, HLA 10 41.79 HNA-1a HLA Unknown GS Negative HLA-Bw4 HNA-2a, HLA 13 50.0

10 Negative HNA-3a HNA-1b Negative HLA HNA-2a, HLA 17 70.8HNA-2a� HNA-1b

11 HNA-1a HNA-3a HNA-5a Negative HLA HNA-2a, HLA 17 70.8HNA-1b

12 CD16 HNA-3a HNA-1b Negative HLA-B12 HNA-2a, HLA 17 70.8CD11b, HLA HLA HLA‡

13 HNA-1a Negative HLA Negative HLA HLA‡ 8 33.314 HNA-1a HNA-3a HNA-1b Negative HLA HNA-2a, HLA 24 100.0

HNA-2a�15 HNA-1a HNA-3a HNA-1b Negative HLA HNA-2a, HLA 21 87.5

HLA Unknown GS,HLA¶

Expected HNA-1a HNA-3a HNA-1b Negative HLA (Bw4) HNA-2a, HLA 24 100.0result HNA-2a

* The boxes indicate an error, for example, failure to detect an antibody, detection of antibody where there was none, determination of in-correct specificity, and failure to detect all antibodies. The score obtained varies with the number and type of error, for example, reportinganti-CD16 or unknown granulocyte-specific antibodies when an anti-HNA-1a was present were considered lesser errors than a failure todetect any antibodies.

† Unknown granulocyte-specific antibody.‡ Failure to detect HNA-2a antibodies.§ Lymphocyte reactive, not HLA class I antibodies.� Anti-HNA-2a identified by two reference laboratories before the workshop deadline.¶ Anti-HNA-2a identified by one reference laboratory after the workshop deadline.

LUCAS ET AL.


The endpoint of the titrations was determined as the lastdilution at which a positive result (scored 1 or 2 by theparticipant) was observed. Doubtful positive reactions(score 0.5) were ignored.

Fourteen laboratories titrated anti-HNA-1a (SerumA) and anti-HNA-2a (Serum B) using the GIFT. Eightlaboratories used a microscopic endpoint (GIFT-M), andseven laboratories used a flow cytometric endpoint(GIFT-F; one laboratory used both GIFT-M and GIFT-F).All but two laboratories used a serum:cell volume ratio ofunity, as requested in the instructions. Seven laboratoriestitrated these sera using the GAT, and one laboratoryeach used the GCLT and MPHA.

Laboratories using the GIFT-M achieved higher end-point titers (range, 16-512 for HNA-1a and 8-2048 forHNA-2a) than laboratories using the GIFT-F (range, 2-128for HNA-1a and 8-128 for HNA-2a). The GAT was lesssensitive than the GIFT-M or GIFT-F for both sera (end-point titers ranged from 2-32 for HNA-1a and from neatto 4 for HNA-2a), even though most laboratories used aserum:cell volume ratio greater than unity. In the GCLTand MPHA, endpoint titers of 512 were obtained for anti-HNA-1a (Serum A), and titers of 512 and 128, respectively,were obtained for anti-HNA-2a (Serum B).

Twelve laboratories titrated Sera A and C using theMAIGA assay and MoAb 3G8. The titers for the anti-HNA-1a (Serum A) ranged from undetectable to 32, whereasthe observed titer for the anti-HNA-1b (Serum C) rangedfrom 4 to 256. Eleven laboratories used a colorimetricendpoint in the MAIGA assay, and one laboratory used achemiluminescence endpoint assay (tube method). A di-rect comparison between these results is difficult becauseof differences in the serum:cell volume ratios used andthe differences in the number of target granulocytes em-ployed by different laboratories. There was also a widevariation in the final concentration (200-5 �g/mL) andvolume (10-100 �L) of MoAb (3G8) used.

DISCUSSION

Recognition of the clinical significance of granulocyte an-tibodies is generally low, partly because of the smallnumber of laboratories able to perform these tests reli-ably. The number of laboratories has been restricted bythe availability of HNA typing sera and the technicalproblems posed because granulocytes are labile cells,which cannot be stored. The importance of granulocyteantibodies in TRALI has been recently highlighted in theSerious hazards of transfusion reports,12 and there maybe an increased demand for granulocyte antibody detec-tion in the future because of the increased number ofgranulocyte transfusions being performed. The detectionof granulocyte antibodies is also of value in the autoim-mune neutropenias. In infants, granulocyte antibody in-vestigation aids the differential diagnosis in cases in

which childhood leukemias are also suspected withoutthe need for bone marrow biopsies. In adults, the dem-onstration of granulocyte antibodies can prove beneficialbefore commencing expensive therapies, such as IVIG orG–CSF administration.

The basic requirements for the successful detectionof granulocyte-specific antibodies include the use of ap-propriate methods and access to a well-defined panel ofHNA-typed granulocytes. Bux and Chapman2 recom-mended the use of both the GIFT and the GAT for thedetection of granulocyte antibodies. The use of MAIGAassays is also indicated in cases in which there are com-plex mixtures of antibodies or in which confirmatory testsare required. The granulocyte immunology laboratoryshould also have access to methods for detecting bothcytotoxic and noncytotoxic HLA class I antibodies. Thesetests are especially useful in cases where HNA-3 antibod-ies are suspected, as the target antigens are expressed onboth granulocytes and lymphocytes.

Previous workshops1,2 established that the standardof granulocyte immunology was generally high but thatthere were problems in the identification of granulocyte-specific antibodies, especially HNA-3a and CD16 anti-bodies. In this workshop, there was wide variation in theproficiency of antibody detection, and again, certaintypes of granulocyte-specific antibodies caused particu-lar problems. Granulocyte-specific IgM antibodies are ofestablished clinical importance, but only eight of 15 labo-ratories successfully identified the HNA-1a IgM antibod-ies in Serum 1, which was obtained from a donor impli-cated in a case of TRALI.13 Failure to detect theseantibodies occurred either because laboratories failed touse an anti-IgM antiglobulin reagent or because an anti-IgGAM reagent was used, which may have containedrelatively small amounts of anti-IgM. However, somelaboratories failed to detect these antibodies despite us-ing an IgM-specific reagent. The GAT, generally, failed todetect the predominantly HNA-1a IgM antibodies in Se-rum 1 and was shown to be insensitive compared withthe GIFT for the detection of HNA-1a, HNA-1b, and HNA-2a antibodies in titration studies. However, the GAT wasthe most reliable method for detecting HNA-3a antibod-ies (Serum 2). Several laboratories reported the presenceof HLA class I antibodies in Serum 2, apparently confus-ing lymphocyte-reactive antibodies with HLA class I an-tibodies.

The HNA-2a antibodies in Serum 3 could be detectedin the MAIGA assay using MoAb 7D8 but not when MoAbMEM166 was used, which suggests that these two MoAbsrecognize different epitopes of the CD177 (NB1) mol-ecule. In contrast, the HNA-2a antibodies in Serum 6were detected using both of these MoAbs. The use of bothMoAbs in parallel would appear to be an appropriatestrategy for maximizing the detection of CD177 (NB1)-reactive antibodies by MAIGA assay. The results of the



third workshop showed that two MoAbs are also requiredto maximize HNA-1a and HNA-1b antibody detection inthe CD16 MAIGA assay.

In conclusion, there was considerable variation inthe sensitivities of the GAT, GIFT, and MAIGA assays per-formed in different laboratories, even when steps weretaken to harmonize serum:cell volume ratios, test sera,and MoAbs. The variation in endpoint, the erroneous de-tection of HNA antibodies in negative control sera (Se-rum 4), and the failure to detect and identify clinicallysignificant granulocyte-specific antibodies suggest thatthere are many technical factors that need to be resolvedand emphasize the need for standardization in these as-says. In contrast, the molecular typing for HNA-1a, HNA-1b, and HNA-1c was performed to a high standard. Ad-vances in technology and increased cooperation betweenlaboratories fostered by international workshops have ledto increased awareness of the significance of granulocyte-specific antibodies. Further progress is required to ensurethat these antibodies and antigens are consistently andreliably detected. As in platelet immunology,14 the estab-lishment of a formal QA scheme may be instrumental inthis process.

ACKNOWLEDGMENTS

We are indebted to all staff of the participating laboratories for

their co-operation in this workshop. We are especially in-

debted to Gail Eiber (American Red Cross, St. Paul, MN), Brig-

itte Flesch, PhD (Institute of Transfusion Medicine, Kiel, Ger-

many), and Philippe Bierling, MD, PhD (Hopital Henri

Mondor, Creteil, France) for supplying additional sera used in

this workshop. We would also like to thank the Medarex Inc.

(Princeton, NJ) for generously supplying the MoAb 3G8 used

in this workshop and Paul Metcalfe, PhD (National Institute of

Biological Standards and Control, Herts) for advice regarding

workshop organization and presentation.

REFERENCES1. Lucas GF, Carrington PA. Results of the First International

Granulocyte Serology Workshop. Vox Sang 1990;59:251-6.

2. Bux J, Chapman J. Report on the Second International

Granulocyte Serology Workshop. Transfusion 1997;37:977-

83.

3. Metcalfe P, Waters AH. Location of the granulocyte-

specific antigen LAN on the Fc� receptor III. Transfus

Med 1992;2:283-7.

4. Bux J, Kober B, Kiefel V, Mueller-Eckhardt C. Analysis of

granulocyte-reactive antibodies using an immunoassay

based upon monoclonal-antibody-specific immobilization

of granulocyte antigens. Transfus Med 1993;3:157-62.

5. Bux J. Nomenclature of granulocyte alloantigens. Transfu-

sion 1999;39:662-3.

6. Bux J, Stein EL, Santoso S, Mueller-Eckhardt C. NA gene

frequencies in the German population, determined by

polymerase chain reaction with sequence-specific prim-

ers. Transfusion 1995;35:54-7.

7. Verheught FW, dem Borne AE, Decary F, Engelfriet CP.

The detection of granulocyte alloantibodies with an indi-

rect immunofluorescence test. Br J Haematol 1977;36:533-

44.

8. Jiang AF, Lalezari P. A micro-technique for the detection

of leukocyte agglutinins. J Immunol Methods 1975;7:

103-8.

9. Decary F, Vermeulen A, Engelfriet C. A look at HL-A anti-

sera in the indirect immunofluorescence test (IIFT). In:

Amos DB, Van Rood JJ, eds. Histocompatibility testing

1975. Copenhagen: Munksgaard, 1975:380-90.

10. Lucas GF. Prospective evaluation of the chemilumines-

cence test for the detection of granulocyte antibodies:

comparison with the granulocyte immunofluorescence

test. Vox Sang 1994;66:141-7.

11. Kissel K, Santoso S, Hofmann C, et al. Molecular basis of

the neutrophil glycoprotein NB1 (CD177) involved in the

pathogenesis of immune neutropenias and transfusion

reactions. Eur J Immunol 2001;31:1301-9.

12. Williamson LM, Lowe S, Love EM, et al. Serious hazards

of transfusion (SHOT) initiative: analysis of the first two

annual reports. Br Med J 1999;319:16-9.

13. Lucas G, Rogers S, Evans R, et al. Transfusion-related

acute lung injury associated with interdonor incompat-

ibility for the neutrophil-specific antigen HNA-1a. Vox

Sang 2000;79:112-5.

14. Metcalfe P, Allen D, Chapman J, Ouwehand WH. Inter-

laboratory variation in the detection of clinically signifi-

cant alloantibodies against human platelet alloantigens.

Br J Haematol 1997;97:204-7.

LUCAS ET AL.


Documents

Report on the Fourth International Granulocyte Immunology Workshop: progress toward quality assessment