An Isologous Antigen-Antibody Reaction Human Neutrophiles ... · Leukocyte antigen-antibody reactions were studied.,by a previously described agglutination method (10), in which serial

Journal of Clinical InvestigationVol. 45, No. 11, 1966

An Isologous Antigen-Antibody Reaction with HumanNeutrophiles, Related to Neonatal Neutropenia *

PARVIZ LALEZARI t ANDGEORGETTEE. BERNARD(From the Department of Hematology, Laboratory Division, Montefiore Hospital and

Medical Center, New York, N. Y.)

Development of methods for detection of leu-kocyte antigens has made possible the classifica-tion of human leukocytes into different antigenictypes, distinct from those of erythrocytes. Todate several such antigen systems have been de-scribed (1-4). Studies on the distribution ofsome of these antigens in various blood cells haveindicated that they are present in both granulo-cytic and nongranulocytic cells, as well as inplatelets (5, 6). It has been shown that theseantigens are also present in certain nonhemato-poietic tissues, such as placenta, kidney, and lung(5, 7).

This study concerns a new leukocyte antigen,the first that appears to be specific for neutro-philes. In 2 unrelated families, this antigen wasresponsible for maternal immunization duringpregnancy, and the subsequent transplacentaltransfer of the antibody appeared to be respon-sible for neonatal neutropenia.

MethodsAntibody sources. Test plasmas used in this study

were obtained from 3 multiparous female donors. Ma-ternal plasma DeR was obtained 7 years after the lastpregnancy. The first, the third, and the fourth childrenof this donor had documented neonatal neutropenia andwere subjects of a previous publication (8). St plasmawas obtained from the mother of a pair of dizygotictwins. The twins were the third and the fourth childrenof the family, and both had neonatal neutropenia (9).The third plasma, containing a monospecific antibody

* Submitted for publication November 26, 1965; ac-cepted August 1, 1966.

Supported by U. S. Public Health Service grant HE10036-01.

Presented in part.at the Annual Meeting of the Ameri-can Federation for Clinical Research, May 1965, AtlanticCity, N. J., and at the Second International Conferenceon Histocompatibility Testing, August 1965, Leiden,Holland.

tAddress requests for reprints to Dr. Parviz Lalezari,Dept. of Hematology, Montefiore Hospital and MedicalCenter, NewYork, N. Y. 10467.

designated Ke (10), was also obtained from a multiparousfemale, none of whose children ever presented any evi-dence of neutropenia. This antibody had previously beenshown to detect an antigen common to leukocytes,platelets, and several solid tissues. Of these donors onlyMrs. Ke had received a tranfusion-1 U of blood 6years before the present study, 30 years after her lastpregnancy.

Blood was anticoagulated with 1/100th vol of 10% di-sodium ethylenediaminetetraacetate (EDTA), and cell-poor plasma was collected after centrifugation at 1,350 gfor 30 minutes. Plasma specimens were stored in 1-mlaliquots at - 200 C.

Cell preparations for agglutination reaction. Leuko-cyte suspensions were prepared by a method in whichhexadimethrine bromide (Polybrene) is used for sedi-menting red cells in blood samples anticoagulated byEDTA (11). The leukocyte-rich supernatants wereseparated from healthy normal blood donors, and the leu-kocyte counts were adjusted to about 4,000 per mm' bydilution with autologous cell-poor plasma. Leukocyteswith a high proportion of eosinophiles were obtainedfrom patients with various parasitic infestations or theLoffler syndrome. Preparations with a high proportionof lymphocytes were obtained from healthy children underthe age of 8.

Test system. Leukocyte antigen-antibody reactionswere studied., by a previously described agglutinationmethod (10), in which serial twofold dilutions of all thetest plasmas -were prepared in 0.1-ml vol, with isotonicsaline as the diluent. To each dilution, 0.05 ml of leuko-cyte suspension was added. The results were read mi-croscopically after 5 and 18 hours of incubation at 370 C.The strength of the agglutination reaction was gradedas previously described (10), and the antibody titer wasdetermined as the highest dilution of the test plasmathat produced agglutination.

At the end of the incubation period, the incubationmixtures with high lymphocytes or eosinophiles wereplaced on microscope slides, dried, and subjected toWright and Giemsa staining for microscopic examina-tion (12).

Cell preparations for absorption. Blood cells wereprepared from normal healthy donors as described byGreen and Solomon (13). In this method leukocyte-platelet-rich plasma is obtained after dextran sedi-mentation of red cells. The cells in the supernatant arethen sedimented by contrifugation in the narrow part ofan "oil bottle," yielding distinct layers of the various

1741

PARVIZ LALEZARI AND GEORGETTEE. BERNARD

cell types. Each layer may thus be separately aspiratedand resuspended in the donor's own cell-poor plasma.Relatively pure platelet, mononuclear, and granulocytesuspensions can thus be obtained. Platelet preparationswere virtually free from other contaminating cells. Sepa-rated mononuclear cells, which usually had 5 to 10%(by volume) platelet contamination, were centrifuged at350 g for 15 minutes. This centrifugation, which al-lowed sedimentation of mononuclear cells while keepingthe platelets in suspension, was usually repeated onceor twice until the packed cells contained no more than2%o (by volume) platelets. This was measured by ap-plying the microcapillary method used for the "leukocrit"determination (10) whereby leukocytes and platelets arepacked into two macroscopically distinct layers. Thus,it was possible to determine the size of each layer.Differential counts were performed on stained smears ofconcentrated cell preparations. Mononuclear preparations,which were contaminated by neutrophiles in excess of12%, as determined by differential counts on 300 cells,were not used. Monocytes and large lymphocytes con-stituted about one-third of the mononuclear cells, theremaining being small lymphocytes. In the granulocytelayer, red cells represented about 2%o of the packed cellvolume, mononuclear cell contamination did not exceed5%,o and platelets were virtually absent.

Although eosinophiles could not be prepared with thesame degree of purity as other cells, it was possible toobtain concentrated preparations by use of blood fromtwo nonleukemic patients with eosinophilia. One pa-tient had LUffler's pneumonia; the other had chroniceosinophilia of undetermined origin. In these patientsthe leukocyte counts were between 20,000 and 40,000mm3, 80%o of which were eosinophiles. The techniquesused for cell separation were those described above forpreparation of leukocytes, and the final suspensions con-tained about 80% eosinophiles.

Solid tissues were processed as follows: Placenta wasobtained from the delivery room; liver, kidney, and lungwere from autopsies performed not later than 4 hoursafter death. Tissues were cut into pieces approximately10 X 10 X 2 mmand were rinsed repeatedly in isotonicsaline until the wash solution was grossly clear. Thefragments were passed through a 100-mesh stainless steelscreen, and the separated cells were suspended and washedthree times in 12 ml of normal plasma obtained from maledonors who had not received transfusions. For eachwashing, the cell suspensions were centrifuged at 350 gfor 15 minutes. Leukocyte typing of all tissue donorswas determined ante-mortem; with placenta, the agglu-tination reactions of both maternal and cord leukocyteswere ascertained.

Method for absorption and elution. In preliminary ex-periments-, the separated blood cells were adjusted to a"leukocrit" or "thrombocrit" of 15%o, a concentrationthat had been satisfactory in the study of Ke antibody(10). It soon became evident that with this antigenconcentration, DeR antibody was only partially absorbed;for complete absorption, it was necessary to use packedcells. In each experiment, volumes of packed blood ortissue cells, from 0.2 to 1 ml, were resuspended in 3 vol

of undiluted plasma containing the antibody, and themixtures were incubated at 37° C for 30 minutes. Theabsorbed plasmas were then separated by 15 minutes ofcentrifugation at 1,350 g.

For antibody elution, the packed cells obtained afterabsorption experiments were resuspended and washedfive times in 12 ml of isotonic saline. The washed,packed cells were finally resuspended in twice their vol-ume of saline. The cell suspensions were then heated at560 C for 10 minutes and centrifuged at 1,350 g for 10minutes in a clinical centrifuge kept at 56° C in a heatedbox. The absorbed plasmas and the eluates were storedat - 20° C until used.

DEAE cellulose column chromatography was per-formed on immune plasmas as previously described (10).

Results

As determined by DEAE cellulose columnchromatography, DeR antibody was mainly inthe plasma fractions eluted by 0.02 M phosphatebuffer pH 6.3 and corresponding to yG-globulins.The fraction eluted by 1 M NaCl, which con-tained yM-globulins, had little activity. Thesefindings are illustrated in Figure 1.

Demonstration that DeR antibody detected asingle antigen

Antibody absorption. Aliquots of DeR plasmawere absorbed with 14 different DeR+ granulo-

De R ANTIBODY

12-

N.z

0

a.

E

Protein-- - Antibody titer

ctI--

0

0ao32 F

z16 S

U-880

-i4<

U02 °

I '

0.02M IMFRACTION NUMBERS

FIG. 1. DEAECELLULOSECOLUMNCHROMATOGRAPHYONDERPLASMA.

1742

NEUTROPHILEANTIGEN-ANTIBODY REACTION

TABLE I

Absorption of antibody with granulocytes

Reaction of absorbed plasmawith "positive" leukocytes

Totalno. of

No. of leukocyte No.Absorbing cell samples agglu- No. not

cell type donors tested tinated agglutinated

DeR+ neutrophiles 14 113 0 113DeR- neutrophiles 3 31 31 0DeR+ eosinophiles 1 11 11 0DeR- eosinophiles 1 5 5 0Ke + eosinophiles 2 15 0 is

cyte samples, and each absorbed plasma speci-men was tested with 4 to 12 different leukocytesamples of the donor panel, making a total of 113tests. The results of these experiments are givenin Table I, which shows that each of the DeR+granulocyte samples used for absorption com-pletely absorbed the antibody; the absorbed plas-mas were nonreactive with all panel memberstested. Specificity of antibody absorption wasdemonstrated by the failure of three differentDeR- cells to absorb the antibody.

Antibody elution. Table II shows that anti-DeR antibody was eluted only from sensitizedDeR+ granulocytes and that the eluates reactedwith all DeR+ leukocyte samples tested. Theeluted antibody failed to react with DeR- cells.

64 r

W 32

4.

2..4'4.51

7

EN-

DeR AHrIGEN -

TABLE II

Antibody elution from granulocytes sensitizedby DeR antibody

No. of samplesGranulocyte samples agglutinated by

used for elution eluates

12 DeR+ 63/633 DeR- 0/13

These data strongly suggest that DeR antibodydetects a single antigen.

Incidence of the antigen

Of 140 normal and unrelated Caucasian donorstested with DeR plasma 79 (56%c) gaxe positivereactions with titers of 1: 4 to 1: 128. Antibodytiters were reproducible with +l-tube differenceswhen tested 3 or 4 times.

Distribution of DeR antigen among blood cells

Blood lymphocytes and inonocytes. Antibodytiters of DeR plasma determined before andafter absorption with blood mononuclear cells areshown in Figure 2. Seven experiments wereperformed with absorption by the monocyte-lym-phocyte preparations from DeR+ donors, inwhich antibody titrations were performed on thedonors' own leukocytes. In one experiment aDeR- donor was used as a control, and the

BEFORE ABSORPTION

AFTER ABSORPTION

2 3 4 5 6 7+ + + + + +

LEUKOCYTE DONORS

I 8

FIG. 2. ANTIBODY ABSORPTIONBY MONONUCLEARCELLS.

1743

__j

PARVIZ LALEZARI AND GEORGETTEE. BERNARD

titration was performed on DeR+ cells. In eachcase absorption was followed by insignificantreduction of antibody titer, and the results ob-tained with DeR+ donors were identical to thosewith the DeR- donor. Our previously publishedstudies have shown that Ke antibody was com-pletely absorbed by mononuclear cells (10). Noantibody could be eluted from the mononuclearcells used in absorption of anti-DeR plasma.Microscopic examination of the stained slidesprepared from agglutinated leukocytes of 5 dif-ferent DeR+ and Ke+ normal children withhigh lymphocyte counts revealed that with DeRantibody only neutrophiles participated in theagglutination, whereas lymphocytes remainedfree. In contrast, with Ke antibody, all cellsparticipated in the agglutination. These data in-dicate that blood mononuclear cells do not con-tain DeR antigen.

A cytotoxicity test (14) 1 using DeR antiserumand lymphocytes prepared from 50 random donorsresulted in no positive reactions. Similarly, dyeexclusion tests (15) 2 failed to reveal cytotoxicityagainst lymphocytes obtained from 5 DeR+donors.

Blood eosinophiles. Aliquots of DeR and Keplasmas were absorbed with eosinophile prepara-tions obtained from a DeR+ and a DeR- donor,both of whom were Ke+. The absorbed plasmaswere tested against a panel of DeR+ and Ke+cells. As shown in Table I, Ke antibody wasremoved by the eosinophiles, but DeR antibodywas not.

Microscopic examination of stained slides pre-pared from agglutinated leukocytes of 5 differentDeR+ and Ke+ patients with high eosinophilecounts revealed that with DeR antibody, onlyneutrophiles participated in the agglutination;eosinophiles remained free. With Ke antibody,indiscriminate participation of all cells was evi-dent. The data thus indicate that blood eosino-philes possess Ke but not DeR antigen.

Platelets. Platelets prepared from 9 DeR+and 4 DeR- donors failed to absorb antibody,and no antibody was eluted from the plateletsused for absorption. Ke antibody, as previously

1 Kindly performed by Dr. Paul Terasaki, Los Angeles,Calif.

2 Kindly performed by Dr. Paul Engelfriet, Nether-lands.

shown (10), could be absorbed by and elutedfrom the platelets. It thus appears that DeRantigen is not present on platelets.

Erythrocytes. Erythrocytes of 5 donors withDeR+ leukocytes were used in antibody absorp-tion and elution experiments. The absorptionprocedure produced no more than a 1-tube changein antibody titers, and no antibody activity couldbe shown in the eluates.

Distribution of DeR antigen in nonhemopoietictissues

Placenta. The results of DeR and Ke antibodyabsorptions are shown in Figure 3. In the first5 experiments, both maternal and cord blood leu-kocytes were DeR+. The sixth experiment rep-resents a case in which the cord cells were re-active, but the mother was negative. In theseventh and eighth cases, both maternal and cordblood leukocytes were DeR-. The plasmas, afterabsorption with placental cells, were tested withseveral DeR+ donors. The change in antibodytiters was minimal and identical for both DeR+and DeR- placenta donors. The same placentapreparations were used for absorption of Keantibody, and in all cases an antibody titer changeof greater than 4 tubes was achieved.

Other body tissues. Similar experiments wereperformed with kidney, liver, and lung tissuesobtained from 2 donors whose leukocytes wereboth DeR+ and Ke+. Antibody titer changesafter absorption by these tissues are shown inFigure 4. The results indicate that liver, kidney,or lung cells obtained from DeR+ donors failedto absorb the antibody. In contrast, Ke antibodywas absorbed by both kidney and lung cells andpartially by liver cells. Elution studies showedthat only Ke antibody could be recovered fromthe absorbing cells, and all cells from Ke+ donorshad this property. These data suggest that noneof the tissues tested contained DeR antigen.

Since DeR antigen might have been present invarious tissue cells in quantities too small to ab-sorb a significant amount of the antibody, ab-sorption studies with the following modificationswere undertaken: Serial dilutions of DeR plasma,from 1: 2 to 1: 16 were prepared, normal plasmabeing used as the diluent. An aliquot from eachdilution was absorbed with J vol of packed cells

1744

NEUTROPHILEANTIGEN-ANTIBODY REACTION

64 -

32 _16 _ @

2 1

De R ANTIODY 0 BEFORE ABSORPTION

I 0. AFTER ABSORPTION I

a I I JeI. o*II I I

Ke ANmT/ODY V BEFORE ABSORPTION

, , V AFTER ABSORPTION

FIG. 3. COMPARISONOF DERWITH KE ANTIBODYAFTER ABSORPTIONBY

PLACENTALCELLS.

that were prepared from kidney, lung, and theliver of a DeR+ tissue donor. No changesgreater than 1 tube in the antibody titers were

observed in any of the plasma dilutions used.In addition, aliquots of undiluted DeR plasmawere subjected to 3 consecutive absorptions, eachtime with i vol of packed liver, kidney, and lungcells. This procedure did not produce any changein the antibody titer, and no antibody was elutedfrom the cells used for absorption. Similar ob-servations were made when packed platelets were

used.In an experiment in which a blood mononuclear

preparation with granulocyte contamination ofabout 17o was used, no antibody titer changeswere produced by absorption when DeR plasmawas diluted 1: 2 and 1: 8 before absorption. In2 other experiments, a 1:8 dilution resulted incomplete removal of the antibody, presumablydue to contaminating neutrophiles that were

present in excess of 7%o.

Properties of St antibodyThe titer of St antibody ranged from 1 :2 to

1: 16 with leukocytes obtained from various do-

nors. As described for DeR plasma, St plasmawas also subjected to DEAE cellulose columnchromatography, and antibody activity was mainlyfound in -yG-globulin fractions. Distribution ofthe antigen in various blood cells and body tissueswas also studied. As shown in Figure 5, Stantibody was absorbed only by neutrophiles. Byelution techniques antibody activity was recov-

ered only from absorbing neutrophiles. It was

also demonstrated that eosinophiles and lympho-cytes did not participate in the agglutination ofthe neutrophiles produced by St antibody.

To test whether both DeR and St antibodiesdetected the same antigen, we determined thereaction patterns of each in 100 unrelated and ran-

domly selected leukocyte donors. With the ex-

ception of the strength and the titer of reaction,for both antibodies identical results were ob-tained. The reaction patterns produced by DeRand St antibodies were similar when leukocytesof the members of both families were tested.Cross absorption of DeR and St plasmas showedthat the antibodies for both were completely re-moved by the leukocytes obtained from the fatherof each family. Similarly, DeR and St anti-

14.

1%

1745

PARVIZ LALEZARI AND GEORGETTEE. BERNARD

De R ANTIODY

IJ4 L. I II.1..III

2

M-16

C256t12

qh. 12164%1 32

16

84

Ia2

BEFORE ABSORPTIONAFTER ABSORPTION

I'l II I@

Ke ANTIBODY V BEFORE ABSORPTIONv AFTER ABSORPTION

II1 11KIDNEY LUNG LIVER

rISSUE DONOR/KIDNEY LUNG LIVER

rISSUE DONOR2

FIG. 4. DER AND KE ANTIBODY ABSORPTION BY TIS-SUES OF TWODIFFERENT DONORS. Although Ke antibodyabsorption by the liver cells appears insignificant, theantibody was eluted from all the absorbing tissue cells.In contrast, DeR antibody was not eluted from any ofthe tissues used for absorption.

bodies eluted from each of these cell preparationsagglutinated the leukocytes of both fathers.

These data suggest that similar to DeR, Stantibody reacts only with neutrophiles and thatboth detect the same antigen.

Variation in the antibody titers found withdifferent leukocyte donors raised the possibility

that the observed phenomenon might represent adose effect. Family members of donors whoseleukocytes gave negative reactions, or those whichstrongly reacted with one of the antibodies, werestudied. The leukocyte preparations were simul-taneously tested with DeR and St plasmas. Theresults are shown in Figure 6. It was found thatboth antibodies invariably showed a parallel re-activity with each donor, but the titer with Stantibody was 2 to 3 tubes less than that obtainedfor DeR.

DeR and other leukocyte antigens

During the Second International conferenceand workshop on tissue typing held in Leiden,Holland, in August 1965, DeR plasma and 76other antileukocyte antisera were tested againsta panel of 40 leukocyte donors. The antisera usedin this study were provided from different labora-tories and were presumed to be monospecific.These antisera identified most of the previouslyknown leukocyte antigens. To evaluate possiblerelationships between the antigens identified bythese sera, we compared the reaction patterns pro-duced by each with the aid of a computer. DeRplasma revealed a positive correlation to onlyone antiserum, B/14.3 This antibody had beenproduced by active immunization in a hematologi-cally normal subject, by repeated injections of

3 Kindly supplied by Dr. Ruggero Ceppellini.

St ANTIBODYJ BEFOREABSORPTION

U AFTER ABSORPTION

MER- PLATE- MONO-ROCYrES LErS NUCLEAR

CELLS

NEUTRO- KIDNEY LUNG LI VERAHILES

FIG. 5. ANTIBODYABSORPTIONSTUDIES ON ST PLASMA.

8S..

N2

1746

NEUTROPHILEANTIGEN-ANTIBODY REACTION

D. FAMILY M. FAMILY Htp. FAMILY

Hs. FAMILY St. FAMILY

OTT

DeR. FAMILY

FIG. 6. FAMILY STUDIES USING BOTH DER AND STANTIBODIES. Solid figures indicate "strong" reactors;half-shaded figures indicate "weak" reactors; unshadedfigures represent nonreactors. t indicates the deceasedmember.

leukocytes obtained from selected donors ( 16).Ceppellini, Mattiuz, and Curtoni had shown withthe EDTA agglutination method that, similar tomost of the earlier serum samples obtained fromthe immunized subject (Bu), B/14 detected a singleantigen with a frequency of 58% in the Italianpopulation (16). By a cytotoxicity test, how-ever, B/14 contained a mixture of antibodies,different from the agglutinin, with prevalent anti-4a activity (17).

Further agglutination studies performed in thislaboratory revealed similar specificity for DeRand B/14 antisera with 20 additional leukocyte

donors. The only difference was in the muchhigher titer obtained with B/14.3 The distribu-tion of antigen detected by B/14 was also studied,and as shown in Figure 7, it was found that theantibody was absorbed only by blood neutro-philes; no antibody absorption was achieved byred cells, platelets, mononuclear cells, or kidney,lung, or liver cells. These observations were re-produced by the use of platelets from 3 differentsources and tissues of 2 different donors. It wasalso found that blood lymphocytes and eosino-philes did not participate in neutrophile aggluti-nation produced by B/14 antibody. Mattiuz andCeppellini have recently confirmed these find-ings; they have shown that lymphocyte suspen-sions obtained from BU+ donors did not absorbEDTA agglutinins (17).

DiscussionThe data presented above indicate the identi-

fication of an antibody directed against an antigenthat is unique to neutrophilic leukocytes. Asingle antigen was evidently detected, as illus-trated by two different test systems: Absorptionstudies with DeR+ leukocytes caused completeremoval of DeR antibody by each of a panel ofleukocyte donors; plasma so absorbed failed toreact with leukocytes from numerous unrelatedDeR+ subjects. Eluted antibody from each in-dividual leukocyte preparation reacted with allthe leukocyte samples with which the unabsorbedplasma reacted, and the previously negative sam-

EJ BEFORE ABSORPTION

U AFTER ABSORPTION

ERYTH- PLATE- MONO-ROCYrTES LErS NUCLEAR

CELLS

LEurRo-PHILES

KIDNEY LUIlG LIVER

FIG. 7. ANTIBODYABSORPTIONSTUDIES ONB/14 ANTIBODY.

512

1%4.i 128

II.

° 8

2

1747

PARVIZ LALEZARI AND GEORGETTEE. BERNARD

ples remained negative. These findings are incontrast to those obtained with plasmas frommultitransfused patients, which contained mul-tiple antibodies against several different leuko-cyte antigens. When these "complex" antibodieswere completely absorbed with leukocytes froma single donor, residual antibody activity remainedthat was active against cells from most otherunrelated donors (10).

The finding of concordant reactions producedby both DeR and St plasmas in 100 differentdonors indicates that both plasmas detect thesame antigen (p <0.0001 by the x2 method).This interpretation is further supported by DeRand St family studies, in which the two maternalleukocyte preparations were nonreactive and thepaternal cells strongly reacted with both anti-bodies. Furthermore, when the two antiserawere absorbed with the paternal cells of one fam-ily, they both became nonreactive with the pa-ternal cells of the other family, and the antibodyeluted from these absorbing cells had identicalspecificity.

The antigen demonstrated by DeR and Stantibodies appears to be identical to the antigendetected by the leukocyte agglutinin present inB/14 antiserum, an antibody experimentally pro-duced by Ceppellini and his associates. This neu-trophile-specific isoantibody produced by an im-munization route other than transplacental is ahelpful adjunct to the present observations.

The DeR antigen system may be composed oftwo or more alleles. Since 56%o of the populationreacts with the antibody, DeR and its allele wouldhave gene frequencies of 0.56 and 0.44, respec-tively, were the alleles limited to two. Twelveper cent of the population would be DeR homo-zygous and 44%o heterozygous. Confirmation ofthese values awaits discovery of the allelic anti-gen. The results of the family studies suggest agenetic pattern, and although our effort to findan antibody that may detect the postulated allelehas thus far been unsuccessful, the family dataobtained are consistent with the hypothesis thatthe strength of the reaction is related to the anti-gen dose. Leukocytes of presumed heterozygoteswere less reactive than cells of presumed homo-zygotes. This phenomenon, however, may alsobe due to augmentation by an unrelated gene.Although observation of a dose effect from leuko-

cyte antigens has not been previously reported,this phenomenon is well known to occur withthe red cells (18, 19).

The most convincing criterion for establishingantigenic determinants is demonstration of spe-cific absorption and elution of the correspondingantibody. By these techniques, we have foundno evidence of DeR antigen activity in any ofthe studied cells except polymorphonuclear neu-trophiles. The possibility remains, however, thatthe antigen is present in these tissues in a modi-fied form not readily reactive with the antibodyor in a concentration not detected by the meth-ods used. The latter possibility seems unlikely,since no progressive fall in the antibody titeroccurred after repeated absorptions. Neither wasthe antibody activity removed when it was di-luted before absorption. In contrast to DeR, Keantibody was removed by the absorption proce-dures. Although there has been no opportunityto use Ke- tissue cells as a control, it has beendemonstrated that Ke antibody absorption ishighly specific in that granulocytes, mononuclearcells, and platelets obtained from Ke- donorsdo not absorb antibody (10). The demonstra-tion of Ke antigen on the same cell preparationsused in this study indicates that at least somecell antigens were not destroyed by the proceduresemployed, and in the preparations with higheosinophiles no evidence of antigen alteration dueto the underlying disease was found. Althoughit has been shown that agglutinating antibodiesare not necessarily cytotoxic (14, 17), the failureof DeR antibody to exert cytotoxicity on lympho-cytes is an observation in agreement with theconclusion that DeR antigen is not present onlymphocytes. Distribution of leukocyte antigensin granulocytes and lymphocytes has been studied(5, 15, 20). According to these reports, sug-gestive evidence indicates that granulocytes andlymphocytes may have antigenic differences. Withthe discovery of specific leukocyte groups, it hasbeen shown that antigens of groups 4 and 5 arepresent on granulocytes, lymphocytes, platelets,and several tissues including kidney, placenta,skin, lung, and spleen (1, 5). An additionalsystem of leukocyte antigens, designated LA, hasalso been found in lymphocytes and platelets (6).Shulman, Marder, Aledort, and Hiller have de-scribed a lymphocyte-specific antigen (21). Thus,

1748

NEUTROPHILEANTIGEN-ANTIBODY REACTION

DeR antigen is unique to date in its specificlocalization to neutrophiles. The present studydemonstrates that within a group of closely relatedcells such as eosinophiles and neutrophiles, an-tigenic differences may exist. This finding doesnot imply different origins for these cells, but mayrather be related to specialized cell functions.

The observation that DeR antigen is confinedto neutrophiles corresponds to laboratory findingsin the affected children with neonatal neutropenia,who showed transient but specific absence ofneutrophiles in their peripheral blood and bonemarrows, with preservation of other blood ele-ments (8, 9, 22). Identification of the same an-tigen, responsible for the development of ma-ternal antibodies and the subsequent occurrenceof neonatal neutropenia in two unrelated families,supports the concept that a similar immunemechanism was involved in both cases. It wasalso shown that both antibodies were primarily7 S globulins, proteins capable of crossing theplacenta. These findings may offer an explana-tion for the rarity of this entity: Significantneutropenia occurred when the involved antigenwas limited to neutrophiles. In contrast, anti-bodies to leukocyte antigens with wide bloodand tissue distribution do not produce leukopenia.If maternal antibodies to the latter type of anti-gens cross the placenta, they may be absorbedby various tissues to such a degree that theireffects on neutrophiles are negligible. Thus, in-vestigators (22-24) who failed to find neonatalneutropenia in babies born to mothers who hadleukocyte antibodies may have been working withantibodies having a broad spectrum of cellreactivity.

The incidence of maternal immunization againstvarious leukocyte antigens may be related to theirtissue distribution. Mothers are more likely tobecome immunized against ubiquitous antigenswhen antigenic stimuli arise from fetal cells otherthan leukocytes, and even placental cells. On theother hand, for immunization against fetal-specificneutrophile antigens, the antigenic stimulus canderive only from fetal neutrophiles. In view ofthese considerations, many maternal leukocyteantibodies may be eliminated as potential sourcesof neonatal neutropenia.

Whether specific cell antigens are significant inhomograft rejection is unknown, but their pres-

ence suggests that this potential source of incom-patibility should be explored.

Summary

A leukocyte antigen is described that is uniqueto blood neutrophiles and appears to be involvedin the pathogenesis of neonatal neutropenia. Theantigen was identified by a leukocyte agglutininfound in plasmas of 2 unrelated mothers, eachof whom had several newborns with isoimmuneneonatal neutropenia. The antibodies were dem-onstrated to be primarily yG-globulins. Evidencewas obtained to indicate that antibodies of bothmothers identified the same antigen. Cross ab-sorption and antibody elution studies demon-strated that a single antigenic determinant witha gene frequency of 0.56 was involved. Antibodyabsorption and elution were achieved only byneutrophiles. The antigen was not found on redcells, platelets, mononuclear cells, eosinophiles,placenta, or lung, liver, or kidney cells. Thisantigen distribution correlates with the selectiveneutropenia observed in the affected newborns.We suggest that this specific localization of theantigen is essential for the occurrence of neo-natal neutropenia and that leukocyte antibodiesdirected against widely distributed antigens arenot likely to affect the neutrophiles of the new-born.

Acknowledgments

The authors wish to thank Drs. Theodore H. Spaet,Richard Rosenfield, and Evelyn Gaynor for their as-sistance in preparation of this manuscript.

References1. Van Rood, J. J., and A. Van Leeuwen. Leukocyte

grouping. A method and its application. J. clin.Invest. 1963, 42, 1382.

2. Van Leeuwen, A., J. G. Eernisse, and J. J. Van Rood.A new leukocyte group with two alleles: leuko-cyte group five. Vox Sang. (Basel) 1964, 9, 431.

3. Payne, R., M. Tripp, J. Weigle, W. Bodmer, and J.Bodmer. A new leukocyte isoantigen in man.Cold Spr. Harb. Symp. quant. Biol. 1964, 29, 285.

4. Bruning, J. W., A. Van Leeuwen, and J. J. Van Rood.Leukocyte antigens in Histocompatibility Testing1965. Copenhagen, Munksgaard, 1966, pp. 275-84.

5. Van Rood, J. J. Leukocyte grouping. A methodand its application. Thesis, University of Leiden,1962.

6. Bialek, J. W., W. Bodmer, J. Bodmer, and R. Payne.Distribution and quantity of leukocyte antigens in

1749

PARVIZ LALEZARI AND GEORGETTEE. BERNARD

the formed elements of the blood. Transfusion(Philad.) 1966, 6, 193.

7. Amos, D. B., and N. Peacocke. Leukoagglutination.A modified technique and preliminary resultsof absorption with tissues. Proceedings of the9th Congress of the European Society of Haema-tology, Lisbon, 1963. Basel/New York, S. Karger,1963, pp. 1132-40.

8. Lalezari, P., M. Nussbaum, S. Gelman, and T. H.Spaet. Neonatal neutropenia due to maternal iso-immunization. Blood 1960, 15, 236.

9. Hanna, M. Personal communication.10. Lalezari, P., and G. E. Bernard. Identification of a

specific leukocyte antigen: another presumed ex-

ample of 5b. Transfusion 1965, 5, 135.11. Lalezari, P. A new technic for separation of hu-

man leukocytes. Blood 1962, 19, 109.12. Lalezari, P., and T. H. Spaet. Studies on the genetics

of leukocyte antigens. Blood 1959, 14, 748.13. Green, I., and W. Solomon. Separation of human

lymphocytes and monocytes using an "oil bottle."J. clin. Path. 1963, 16, 180.

14. Terasaki, P. I., and J. D. McClelland. Microdropletassay of human serum cytotoxins. Nature (Lond.)1964, 204, 998.

15. Engelfriet, C. P., and V. P. Eijsvoogel. Cytotoxiciso-antibodies against leukocytes. Vox Sang.(Basel) 1965, 10, 228.

16. Ceppellini, R., P. L. Mattiuz, and M. Curtoni. Char-acteristics and evolution of leukoagglutinins de-veloped in a patient receiving multiple transfusionsfrom selected donors in Histocompatibility Testing.Washington, D. C., National Academy of Science-National Research Council, 1965, pp. 71-82.

17. Mattiuz, P. L., and R. Ceppellini. Personal com-munication.

18. Stratton, F., and P. H. Renton. Practical BloodGrouping. Oxford, Blackwell, 1958, p. 175.

19. Race, R. R., and R. Sanger. Blood Groups in Man,4th ed. Philadelphia, F. A. Davis, 1962, p. 4.

20. Walford, R. L. Leukocyte Antigens and Antibodies.New York, Grune & Stratton, 1960, p. 66.

21. Shulman, N. R., V. J. Marder, L. M. Aledort, andM. C. Hiller. Complement-fixing isoantibodiesagainst antigens common to platelets and leuko-cytes. Trans. Ass. Amer. Phycns 1962, 75, 89.

22. Dausset, J., and G. Malinvaud. Les agranulocytoseset granulopenies neo-natales par iso-immunisationfoetomaternelle. Nouv. Rev. franc. Hemat. 1964,4, 154.

23. Payne, R., M. R. Rolfs, M. Tripp, and J. Weigle.Neonatal neutropenia and leukoagglutinins. Pedi-atrics 1964, 33, 194.

24. Jensen, K. G. Leukocyte antibodies in serums ofpregnant women. Vox Sang. (Basel) 1962, 7, 454.

1750