Gene frequencies of the five major human platelet antigens in African American, white, and Korean populations

H.O. KIM, Y. JIN, T.S. KICKLER, K. BLAKEMORE, O.H. KWON, AND P.F. BRAY

Background: The study of the immunogenetics of the human platelet antigens is important to the improvement of diagnosis and genetic counseling and to the devel- o ment of eening programs for women at risk of having babies with neonatal arl ,mmunq rombocytopenia. Description of the immunogenetics of the human platelet ant ens in some racial groups has been incomplete. Study Design and Methods: A reverse dot blot technique employing polymerase chain reaction-amplified genomic DNA was applied in genotyping the five major human platelet antigens in the following populations: 100 African American and 100 white .women admitted to the obstetric unit at Johns Hopkins Hospital (Baltimore, MD) and 100 inpatients at Yonsei Universi (Seoul, Korea). Results: The gene frequencies of HPA-2b?kF) and HPA-5b (SP) in African Ameri- cans were twice those in whites (African Americans: 0.18 and 0.21, respectively; whites: 0.09 and 0.11, respectively). There is a very low gene frequency of the HPA- Ib (PP') allele in Koreans (0.005). No significant differences were found in the gene frequencies of the human platelet antigens in whites in this series and those in pub- lished European studies. Conclusion: These studies indicate a higher potential risk for alloimmunization to HPA-2 (KO) and HPA-5 (Br) antigens in African Americans than in whites. In addi- tion, the low gene frequency of HPA-Ib ( P P ) in African Americans and Koreans suggests that alloimmunization to HPA-1 a (PIA1) would be very unusual in these populations. These data may provide the basis for planning neonatal alloimmune thrombocytopenia screening programs in certain ethnic populations.

Abbreviations: HPA = human platelet antigen; PCR = polymerasechain mctlon; SDS =sodium dodecyl sulfate; SSC = salinesodium cltrate.

HUMAN PLATELET alloantigens are important in the pathogenesis of neonatal alloimmune thrombocytopenia and posttransfusion purpura and in some cases of plate- let transfusion~fi-actoriness.~-3 Platelet membrane gly- coproteins codbin amino acid polymorphisms that give rise to the human platelet antigenic phenotype. Single- nucleotide substitutions in the genes encoding the plate- let membrane glycoproteins are responsible for the amino acid changes that determine the antigenic ~tructure.4.~ To date, eight human platelet antigen (HPA) systems have been de~cribed.~ Traditionally, these antigens have been referred to as platelet-specific antigens. However, a work- ing party on platelet serology has recommended the term human platelet antigen and has proposed a numbering nomenclature primarily based upon the chronological

From the Departments of Medicine, Obstetrics, and Pathology, The Johns Hopkins University School of Medicine. Baltimore, Maryland, and thc Depemnent of Clinical Pathology, Yonsei University College of Medi- cine, Seoul. Korea.

Supported by grant ROI HL49748 from the National Institutes of Health and by the National Blood Foundation.

Received for publication February 16, 1995; revision received June 2,1995. and accepted June 5,1995.

TRANSFUSION 1995;35:863-867.

order in which the antigens were serologically described? But it is now clear that there are at least six known gly- coprotein IIIa alleles that encode five different glycopro- tein IIIa polymorphic and antigenic variants? and this numbering system is somewhat misleading in implying that platelet antigens are diallelic. For this reason, we will refer to the human platelet antigens by both the HPA no- menclature and the traditional, widely recognized nomen- clature.

In general, immunophenotyping procedures have been used to perform immunogenetic studies on the human platelet antigens. Von dem Borne et al? summarized the reported HPA phenotype frequencies in whites, Japanese, Koreans, and South American Indians. The only docu- mented study of African Americans reported solely on typing for HPA- 1 .' Immunophenotyping is limited by the general lack of quality typing antisera, especially for the less commonly encountered antigens. Consequently, a variety of DNA-based methods have been developed.*-I2 To circumvent the limitations of immunophenotyping, we have used reverse dot blot hybridization to determine the gene frequencies of the five major human platelet anti- gens in 100 African Americans, 100 whites, and 100 Koreans. This is the first large genotyping study per-

863

864 KIM ET AL. TRANSFUSION Vol.35.No. 1 0 4 9 9 5

formed on these human platelet antigens in populations in the United States. These studies have implications for the development of strategies for identifying patients at risk of being alloimmunized to human platelet antigens.

Materials and Methods We collected into EDTA 3 mL of blood from 100 African

American and 100 white women admitted to the obstetric unit at Johns Hopkins Hospital and from 100 Koreans at Yonsei UniversiQ Hospital in Seoul, Korea. All of these studies have been conducted according to the principles expressed in the Declaration of Helsinki.

Genomic DNA was prepared by a rapid method of DNA Briefly, whole blood (0.5 mL) was added to an equal

volume of polymerase chain reaction (PCR) lysis buffer (PLB; 0.32 m o m sucrose, 10 mmoVL Tris [pH 7.5],1% Triton X-100, 5 mmoYL MgClJ. Samples were centrifuged for 20 seconds at 13,000 x g, the pellet was resuspended in 1 mL of PCR lysis buffer, and the process was repeated twice. The pellet was re- suspended in 0.5 mL of PCR incubation buffer (50 mmolL KCl, 10 mmovL Tris [pH 8.3],2.5 mmoVL MgC4,O.l mg/mL gela- tin, and two nonionic detergents: 0.45% Nonidet P-40 "p-401 and 0.45% 'heen 20 [both, Sigma Chemical Co., St. Louis, MO]), plus 4 mg per mL of proteinase K. and incubated at 65°C for 1 hour. We then boiled the sample for 20 minutes to inacti- vate the proteinase K and stored it at -20°C. Twenty-five pL of lysate was approximately equivalent to 1 pg of genomic DNA.

We applied the reverse dot blot technique as previously described,1o with slight modifications, to genotype the 10 ma- jor human platelet antigens. Oligonucleotides of 15 to 18 bps that were specific for each of the 10 major human platelet an- tigen alleles were applied to nylon membranes (Biodyne C, Pall Biosupport, East Hills, NY). W o hundred ng of genomic DNA was amplified in a 50-pL PCR product with 10 pmol of the biotinylated primers that flank the DNA polymorphi~m.'~ In general, we typed the major human platelet antigens with just two PCR reactions (multiplex PCR): one PCR reaction con-

tained two pairs of primers (for HPA-2 [KO] and HPA-3 [Bak]) and the other contained three pairs (for HPA-1 [PIA], HPA-4 [Pen], and HPA-5 [Br]). Amplification conditions for multiplex PCR using primers for HPA-2 (KO) and HPA-3 (Bak) were 95°C for 4 minutes, followed by 35 cycles of 94°C for 1 minute, 56°C for 45 seconds, and 72°C for 2 minutes. Amplification condi- tions for multiplex PCR using primers for HPA-1 (PIA), HPA-4 (Pen), and HPA-5 (Br) were 95°C for 4 minutes, followed by 30 cycles of 94°C for 45 seconds, 52°C for 45 seconds, and 72°C for 1 minute. In all cases, we used positive control DNAs for each antigen system and a negative control with no template DNA.

Twenty-five pL of each PCR reaction from a single indi- vidual was hybridized to each filter for 3 to 12 hours at 45°C. The strip was washed in 0 . 7 5 ~ saline-sodium-citrate (SSC) (Ix SSC = 150 mmoVL NaCl, 15 mmoVL trisodium citrate, pH 7.0) and 0.1-percent sodium dodecyl sulfate (SDS) and incubated for 30 minutes at mom temperature with streptavidin horserad- ish peroxidase conjugate (Boehringer Mannheim, Indianapo- lis, IN) diluted 1-in-4000 in 2x SSC and 0.1-percent SDS. We then washed the strips three times in 2x SSC and 0.1-percent SDS for 2 minutes and twice in 0.1 M sodium citrate, pH 5.0, for 2 minutes. The color was detected with either 3,3'-5.5' tetramethylbenzidine chromogenic substrate (TMB, Sigma) as describedI0 or with an enhanced chemiluminescence kit (ECL Western Blotting Detection System, Amersham, Oakville, ON, Canada). In the latter case, after washing, the strip was incu- bated for 1 minute with ECL detection reagents and exposed to film (Xrp-1, Kodak, Rochester, NY) for 15 to 30 minutes.

Gene frequencies were derived from the Hardy-Weinberg equilibrium for diallelic systems,14 and the z test was used for statistical analysis.

Results In our previous study,I0 we did not use multiplex PCR to

perform reverse dot blot analysis. Figure 1(A) demonstrates the ability to produce PCR fragments of the appropriate size when more than one set of biotinylated primers is used in PCR. These

FLG. 1. Reverse dot blot analysis of the platelet-specific antigens using multiplex PCR. A) Ethidium-stained 2-percent agarose gel containing products from FCR. Tests were performed with primers that amplify the HPA-I (PP) andHPA-4 (Pen) alleles (Lanes 2 and 3), HPA-2 (KO) and HPA-5 (Er) (Lanes 4 and 5). HPA-3 ( E d ) (Lanes 6 and 7). and HPA-I (PP), HPA-4 (Pen), and HPA-5 (Er) (Lanes 8 and 9). Controls with no template DNA are shown in Lanes 2.4,6, and 8. PCR conditions were as described in the Methods. Lane 1 ,+x174DNA digested with HueIII standard. B) Reverse dot blot analysis. Roducts from two multiplex PCR assays in a single individual (Panel A, Lam 9; note that the HPA-2 [KO] and HPA-3 [Bak] combination is not shown in Panel A) were hybridized to the strip and detected with the enhanced chemiluminescence system. Letters written on the template refer to the HPA nomencla- tlw.

TRANSFUSION 1995-Vol. 35. No. 10 FREQUENCY OF GENES ENCODING PLATELET ANTIGENS 865

PCR products were able to undergo hybridization with the ap- propriate membrane-linked allele-specific oligonucleotide (Fig. l[BI).

We next used the reverse dot blot technique to genotype the human platelet antigens from 100 African Americans, 100 whites, and 100 Koreans. Table 1 shows the frequencies for each genotype. Homozygosity for HPA-lb (Pw, HPA-26 (KO"), HPA- 3b (Bukb), and HPA-5b (BP) was 2, 3, 12, and 2 percent, re- spectively, in whites; 0,3, 15, and 4 percent in African Ameri- cans; and 0, 1,9, and 0 percent in Koreans. The HPA-4b (Penb) allele was not detected in any of the populations studied. The gene frequencies in the 200 chromosomes from each ethnic group studied, together with the reported frequency in the Dutch population, are shown in Table 2.

Discussion We have performed an extensive genotyping study on

human platelet antigens from African Americans and white Americans and from Koreans (living in Seoul, Korea) by using a reverse hybridization technique on genomic DNA. Genotyping to determine the frequency of human platelet antigens was first performed by McFarland et a1.,8 and its utility has since been demon- strated by other^.^-'^ Genotyping is useful and comple- mentary to immunophenotyping, especially when simul- taneously occurring antibodies to multiple human platelet antigens are present. It also circumvents the major difi- culty in the serologic analysis of thrombocytopenia due to antibodies against alloantigens: the unavailability of certain rare antisera, such as anti-HPA- 1 b (-PIAz) and anti- HPA-4b (-Penb). We have previously shown that the re- verse dot blot method is a specific, sensitive, and rapid technique.'O Determination of the genotype of the five major human platelet antigens can be accomplished in 1 day and with just two PCR reactions. This represents a considerable savings in time, effort, and expense.

Table 1 . Genotvpe fequencies of the five major human platelet antigens

Genome freauencv' I%) Genotype Whiis African Americans Koreans HPA-ldla (PPWP) 80 84 99 HPA- l d l b (P P1/Pv 18 16 1 HPA-lWlb(PP/P* 2 0 0 HPA-2aQa (KO%* 85 67 75

H P A - m b (Koe/Kd) 3 3 1 HPA-W3a(BaPBak.) 46 40 43 HPA-W3b (BaPkh") 42 45 48 HPA-3W3b (BaWBaP) 12 15 9 HPAW4a (Pen"/perp) 100 100 100 HPA&4b(Perp/Perp) 0 0 0 HPA-4b/46 ( P enb/perp) 0 0 0 HPA-Sd5a (BPBP) 79 62 94 HPA-W' (6PBP) 19 34 6 HPA-W' (6PBP) 2 4 0

HPA-2&2b (Kob/Kd) 12 30 24

Genotyping was performed on DNA from 100 unrelatedyaidu- als in each group.

We have identified several differences in the genotype frequencies of the human platelet antigens in African Americans and whites. First, the HPA-Zb (PIA2) allele is found slightly less commonly in African Americans than in whites: 16 percent versus 20 percent of the individu- als in this study. Ramsey and coworkers' immuno- phenotyped 243 African Americans in Pittsburgh, esti- mating gene frequencies for HPA-la (PIA') and HPA-lb (PIA2) as 0.936 and 0.064, respectively. These data are quite similar to our gene frequency determination (HPA- la, 0.92; HPA-Zb, 0.08). In the two studies, only a single HPA-lb (PPZ)-homozygous individual was found. There- fore, it seems likely that neonatal alloimmune thrombo- cytopenia due to incompatibility at the HPA-I (PIA) lo- cus is quite unusual in the African American population. Second, we found the gene frequencies of HPA-2b (KO") and HPA-Sb (BP) to be twice as high in African Ameri- cans as in whites (African Americans: 0.18 and 0.21, respectively; whites: 0.09 and 0.11, respectively). This suggests that the rate of alloimmunization to these anti- gens may differ in African Americans and whites. Larger population studies are required to determine if this sug- gestion will be borne out.

The HPA-la (PIA') and HPA-lb (PlA2) gene frequen- cies in whites in our study were 0.89 and 0.1 1. We found two HPA-16 (PP2)-homozygous individuals. There were no significant differences in the frequency of HPA-la (PIA') and HPA-lb (PIA2) alleles in whites in Europe and the United state^?^.'^ although we found a slightly higher HPA-la (PIA') and slightly lower HPA-Zb (PIA2) gene fre- quency in whites in Europe.

Because of the scarcity of HPA-2 (KO) antisera, there are few reports estimating the frequency of the alleles of the HPA-2 (KO) system. The genotype frequencies of the HPA-2u (Kob) and HPA-26 (KO") alleles (97% and 15%, respectively) among whites in our study are not signifi- cantly different from those previously reported by immunophenotyping (99.3% and 13.6%)'' We previ- ously reported the immunophenotype frequency of the HPA-3 (Bak) system in 91 unrelated whites.13 The geno-

Table 2. Gene frequencies ofplatelet-specific antigens calculated from genotype frequencies

US population HwwPe whites African Americans Koreans Dutch' HPA-la(PP3 0.89 0.92 0.995 0.846 HPA-lb(PP) 0.11 0.08 0.005 0.154 HPA-2a (Kob) 0.92 0.82 0.87 0.934 HPA-2b ( KC?') 0.09 0.18 0.13 0.066 HPA-3a (Bakd) 0.67 0.63 0.67 0.555 HPA-3b (Balp) 0.33 0.37 0.33 0.445 HPA-4a (Pee) 1.00 1 .00 1.00 1.OOo HPA-4b(Perp) 0.00 0.00 0.00 o.Oo0

HPA-S(BP) 0.11 0.21 0.03 0.098 HPA-Sa (6P) 0.89 0.79 0.97 0.902

Dutch gene frequency as reported by Simsek et aLB

866 KIM ET AL. TRANSFUSION Vol. 35. No. 10-1995

type frequency of HPA-3a and HPA-3b in the current study in the different racial groups is similar to our previous immunophenotyping results13 and to the genotype fre- quency reported by Simsek et aL9 in the Dutch popula- tion using a combination of immunophenotyping and allele-specific restriction enzyme digests.

The Pen antigen system was first detected in the Japa- nese popu1ati0n.I~ The HPA-4b (Penb) allele was not de- tected in any of the 300 individuals we studied. The ab- sence of this allele provides further evidence that it is extremely rare in whites. Our data on HPA-5 (Br) allelic frequencies constitute the first such report in the US white population, and are similar to those detected in European population^.^*^^

It is well known that the low-frequency alleles for each of the human platelet antigens are rare in A~ians . '~ .~ ' Among 1 100 Chinese in Taiwan, no HPA-la (PIA1)-nega- tive individuals were found." Shibata et al.I9 studied the HPA- 1 a (PIA') phenotype frequency in 300 Japanese and found that none were HPA-la (PIA1) negative. Among 100 Korean samples in our study, we found no HPA-lb / lb (PP2/PlA2) individuals and only a single heterozygote. Be- cause the HPA- 1 (PIA) system seems to be responsible for most cases of alloimmune-mediated thrombocytopenia, the rarity of homozygous HPA- 1 b (PlA2) may be the prin- cipal reason for the lower incidence of neonatal allo- immune thrombocytopenia and posttransfusion purpura in Koreans, and in Asians in general. Han et al.I5 reported a phenotype frequency of HPA-la, -3a, -4a, and -4b in Koreans that was similar to ours, but they observed a considerably higher frequency of HPA-lb. Our data are consistent with the low frequency of HPA-lb in other Asian populations, and the discrepancy between our study and that of Han et al. is presumably due to the different techniques used.

Our study reports the first gene frequencies of HPA- 2a, -2b, -3b, -5a, and -5b in Koreans. The genotype fre- quency of the HPA-26 (KO") allele (25%) in Koreans was the same as that in Japanese (24.5%)5 and higher than that in Chinese (9%).22 We identified no HPA-4b (Penb) alleles in the Korean samples. The HPA-4b (Penb) phe- notype is estimated to be below 1 percent in Asian coun- t r i e ~ . ' ~ . ~ ' . ~ ~ Presumably, the testing of a larger group would detect some of these rare alleles. The genotype frequency of the HPA-Sb (BP) allele was 6 percent in Ko- reans, which was similar to that in the Japanese popula- tion (8.67%) but lower than that reported by immuno- phenotyping in the Chinese population (17.73%)?'

Finally, the frequencies of some platelet-specific anti- gens have been studied in South American Indiansz3 and Ind~nes ians .~~ The genotype frequency of HPA-lb in Indonesians is 1.8 percent, which is similar to that in the Chinese, Japanese, and Korean populations and lower than that in the white population. Indonesians and South American Indians have HPA-56 genotype frequencies of

9 percent and 4.9 percent, respectively, which are inter- mediate between those of the Korean and white or Afri- can American populations in this study.

The antigen system most commonly responsible for the development of neonatal alloimmune thrombocytopenia and posttransfusion p q u r a is the HPA- 1 (PIA) ~ystem. '*~*~ When there is no family history of neonatal alloimmune thrombocytopenia, antenatal screening for detection of fetuses at risk for neonatal alloimmune thrombocytopenia currently is not performed. It has been suggested that all pregnant women should be screened for HPA-1 ( P r ) gen~type.~' However, before comprehensive screening programs can be initiated, it is necessary to know the gene frequency, and hence the risk for neonatal alloimmune thrombocytopenia, in the population to be studied. This will allow 1) the most appropriate therapy to be admin- istered, 2) genetic counseling to be given, 3) factors in- volved in their pathogenesis to be studied, and 4) com- prehensive screening programs for obstetric patients to be developed. Our results suggest that a screening pro- gram for maternofetal incompatibility for HPA-1 (PIA) in African American and Asian populations would have an extremely low yield.

References 1. Mueller-Eckhardt C, Kiefel V, Grubert A, et al. 348 cases of

suspected neonatal alloimmune thrombocytopenia. Lancet

2. Von dem Borne AE, Kuijpers RW, Goldschmeding R. Platelet antigens in alloimmune thrombocytopenia. In: Nance SJ, ed. Clinical and basic science aspects of immunohematology. Arlington: American Association of Blood Banks, 1991:115-35.

3. Kickler TS. Platelet immunology. In: Anderson KL. Ness PM, eds. Scientific basis of transfusion medicine. Philadelphia: WB Saunders, 1994:304-15.

4. Kunicki TJ, Newman PJ. The molecular immunology of human platelet proteins. Blood 1992;80 1386-404.

5. Newman PJ. Nomenclature of human platelet alloantigens: a problem with the HPA system? Blood 1994;83: 1447-5 I .

6. Von dem Borne AE, Wary F. ICSWISBT Working Party on platelet serology. Nomenclature of platelet-specific antigens. Vox Sang 1990;58:176.

7. Ramsey G, Salamon DJ. Frequency of PIA' in blacks. Transfu- sion 1986;26531-2.

8. McFarland JG, Aster RH, Bussel JB, Gianopoulos JG, Derbes RS, Newman PJ. Prenatal diagnosis of neonatal alloimmune thrombocytopenia using allele-specific oligonucleotide probes. Blood 1991;78:2276-82.

9. Simsek S, Faber NM, Bleeker PM, et al. Determination of hu- man platelet antigen frequencies in the Dutch population by immunophenotyping and DNA (allele-specific restriction en- zyme) analysis. Blood 1993;81:835-40.

10. Bray PF, Jin Y, Kickler T. Rapid genotyping of the five major platelet alloantigens by reverse dot-blot hybridization. Blood

11 . Skogen B, Bellissimo D, Hessner MJ, et al. Rapid determina- tion of platelet alloantigen genotypes by polymerase chain reac- tion using allele-specific primers. Transfusion 1994;34955-60.

12. Unkelbach K, Kalb R, Santoso S, Kroll H, Mueller-Eckhardt C, Kiefel V. Genomic RFLP typing of human platelet alloantigens ZW(PIA'), KO, Bak, and Br (HPA-1.2.3.5). Br J Haematol

13. Jin Y, Diek HC, Nurden A, Bray PE Single-strand conforma- tion polymorphism analysis is a rapid and effective method for

1989;18:363-6.

1994;84:4361-7.

1995;89: 169-76.

TRANSFUSION 1995-Vol. 35. No. 10 FREQUENCY OF GENES ENCODING PLATELET ANTIGENS 867

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

the identification of mutations and polymorphisms in the gene for glycoprotein IIIa. Blood 1993;822281-8. Bodmer WF, Cavalli-Sforza LL. Genetics, evolution, and man. San Francisco: WH Freeman. 1976232-58. Han KS. Cho HI, Kim SI. Fquency of platelet-specific antigens among Komns determined by a simplified inmunofluorescence test. Transfusion 1989;29:708-10. Shulman NR, Marder VJ, Hiller MC, Collier EM. Platelet and leukocyte isoantigens and their antibodies: serological, physio- logic and clinical studies. Prog Hematol 1964;4:222-304. Mueller-Eckhardt C, Kiefel V, Santoso S. Review and update of platelet alloantigen systems. Transfus Med Rev 1990,498- 109. Kickla TS, Herman JH, Furihata K, Kunicki TJ, Aster RH. Iden- tification of Bakb, a new platelet-specific antigen associated with posttransfusion purpura. Blood 1988;7 1:894-8. Shibata Y, Matsuda I. Miyaji T, Ichikawa Y. Yuk'. A new plate- let antigen involved in two cases of neonatal alloimmune thrombocytopenia. Vox Sang 1986;50 177-80. Kiefel V, Santoso S, Katzmann B. Mueller-Eckhardt C. The Bf / BP alloantigen system on human platelets. Blood 1989;73:2219- 23. Mueller-Eckhardt C, Santoso S. Kiefel V. Platelet alloanti- gens-molecular, genetic, and clinical aspects. Vox Sang

Lin M, Shieh SH, Yang TF. Frequency of platelet-specific anti- gens among Chinese in Taiwan. Transfusion 1993;33:155-7. Inostroza J. Kiefel V, Mueller-Eckhardt C. Frequency of plate- let-specific antigens PI'", Bak', Yuk', Yukb, and B f in South American (Mapuches) Indians. Transfusion 1988;28:586-7.

1994,67(S~ppl):89-93.

24. Santoso S, Santoso S, Kiefel V, Masri R, Mueller-Ekkhardt C. Frequency of platelet-specific antigens among Indonesians. Transfusion 1993;33:739-41.

25. Flug F, Karpatkin M. Karpatkin S. Should all pregnant women be tested for their platelet PLA (Zw, HPA- 1 ) phenotype? Br J Haematol1994.86 1 -5.

Hyun 0. Kim, MD, Fellow, Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, MD; current ad- dress: Assistant Professor, Department of Clinical Pathology, Yonsei University College of Medicine, Seoul, Korea.

Ying Jin, MD, Fellow, Department of Medicine, Johns Hopkins University School of Medicine.

Thomas S. Kickler, MD, Associate Professor, Division of Tms- fusion Medicine, Department of Pathology, Johns Hopkins Univer- sity School of Medicine.

Karin Blakemore, MD. Associate Professor, Fetal-Maternal Medi- cine and Genetics, Department of Obstetrics, Johns Hopkins Univer- sity School of Medicine.

Oh Hun Kwon, MD. Associate Professor, Department of Clini- cal Pathology, Yonsei University College of Medicine.

Paul F. Bray, MD, Associate Professor, Division of Hematology, Departments of Medicine and Pathology, Johns Hopkins University School of Medicine, 720 Rutland Avenue, Ross 1015, Baltimore, MD 21205. [Reprint requests]