Soluble type A substance in fresh-frozen plasma as a function of ABO and Secretor genotypes and Lewis phenotype

Transfusion and Apheresis Science 32 (2005) 255–262 intl.elsevierhealth.com/journals/tras

Soluble type A substance in fresh-frozen plasma as a function of ABO and Secretor genotypes and Lewis phenotype F.J. Achermann a, F. Julmy a, L.G. Gilliver b, T.P. Carrel a, U.E. Nydegger a

a,*

University Clinic of Cardiovascular Surgery, HGEK Inselspital, CH-3010 Bern, Switzerland b Auckland University of Technology, Auckland, New Zealand Received 18 December 2003; accepted 16 May 2004

Abstract Soluble ABO blood group substance (SAS) in fresh-frozen plasma (FFP) and its cognate alloantibody titer reduction capacity (TRC) are not considered when prescribing this product for plasma exchange (PEX) therapy of ABO incompatible transplant recipients. SAS was quantified in 250 single FFPs using ELISA. Total and IgG class-specific anti-A TRCs of FFPs were measured using a microhemagglutination inhibition assay. SAS level depended not only on the A subtype (p < 0.0001) and the Secretor status (p < 0.0001), but also on the expression of ALeb in A1 secretors (p < 0.0001). The variation was as great as 137.6 arbitrary units (aU) for 14 A1 Le(a
b
) secretors and 1.2 aU for 6 A2 non-secretors. Homozygous expression of the A1, A2 and Secretor alleles did not increase SAS levels. Only total anti-A TRC, but not IgG class-specific TRC depended on the detected SAS level (r = 0.566, p = 0.0003).
2005 Elsevier Ltd. All rights reserved. Keywords: Plasma exchange; ABO incompatible transplantation; Hemagglutinin neutralisation

1. Introduction In addition to the HLA system, the importance of the ABO histo-blood group system is increas-

*

Corresponding author. Tel.: +41 31 632 23 29; fax: +41 31 632 44 43. E-mail address: [email protected] (U.E. Nydegger).

ingly recognised in the light of ABO-compatible and, more recently, ABO-incompatible solid organ transplantation [1,2]. Transplantations over this barrier are contraindicated except for bone, cornea and skin allografts [3]. However, clinicians wish to overcome the ABO barrier because of the poor availability of ABO matched organs in urgent situations or interdiction of cadaver transplantation. Immunosuppression [4,5], extracorporal

1473-0502/$ - see front matter
2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.transci.2004.05.007

256

F.J. Achermann et al. / Transfusion and Apheresis Science 32 (2005) 255–262

immunoabsorption (EIA) [6,7] plasma exchange (PEX) [8] and i.v. infusion of synthetic A/B saccharides [9,10] are the present approaches to reduce the isohemagglutinin titers of anti-A/B to prevent humoral rejection of ABO incompatible transplants. In experimental xenotransplantation the infusion of a polymeric form of aGal trisaccharide (GAS914) was recently reported to significantly reduce xenoreactive antibodies but not improve survival [11]. Usually AB plasma or plasma of the organ donors ABO type is used for PEX therapy of ABO incompatible transplanted patients without regard to the amount of soluble A/B substance [8]. The aim of this study was to estimate the amount of SAS in FFP as a function of the ABO and Secretor genotypes and Lewis phenotype and to quantify the total and IgG class-specific anti-A titer reduction capacity (TRC) of different plasma types using a microhemagglutination inhibition assay.

2. Materials and methods 2.1. Blood collection and antibodies All 250 tested blood samples were collected with written informed consent of the donors at the Regional Red Cross Blood Transfusion Centre of Bern (Switzerland) from normal volunteers. Plasma was separated and stored in aliquots at
20 C until analysis. We used the following antibody reagents: BRIC 145 anti-A (International Blood Group Reference Laboratory, Bristol, UK); Seraclone anti-A, clone: A003 (Biotest AG, Dreieich, Germany); S2 anti-A, clone: 9113D10 (Diagast GmbH, Aachen, Germany); Goat anti-mouse IgG1—biotin conjugate (Southern Biotechnology Associates, Inc., Birmingham AL, USA) and Goat anti-mouse IgM—biotin conjugate (Sigma Chemical Co., St. Louis, MO, USA); DiaMed-ID Microtyping System ABO/D (DiaMed, Cressier, Switzerland); DiaMedID anti-A1 absorbed; DiaMed-ID monoclonal anti-H; DiaClon anti-Lea and anti-Leb. The following two MoAbs anti-A were tested against an extensive glycolipid panel to map their analytical specificity using thin layer chromato-

graphy (TLC) [12]: The BRIC 145 reagent reacted strongly with A type 2 (A-6-2), A type 1 (A-6-1), ALeb (A-7-1), A type 4 (A-7-4) plus some reactivity with extended structures. The Seraclone anti-A reagent reacted strongly with A-4-6 (four sugar A on a type 6 core chain), A-6-2, A-6-1, A-7-1 and a whole range of extended structures. The reactivity pattern of these two antibodies is similar to that expected for human polyclonal anti-A. 2.2. ABO and Lewis blood group phenotyping Serological confirmation of donor ABO and Lewis status was performed using DiaMed ID Micro Typing System ABO/D and DiaCell test red blood cells according to the manufacturerÕs directions. 2.3. ABO blood group genotyping Four millilitres of blood were collected in EDTA; genomic DNA was extracted from 200 ll EDTA blood using a commercial kit (High pure PCR Template Preparation Kit Roche, Basel, Switzerland), and blood group ABO polymorphism was analysed by sequence-specific primer polymerase chain reaction (PCR-SSP) investigating exon VI and VII on chromosome 9 [13]. We used eight specific PCRs in parallel to detect specific nucleotide sequence differences between the A1 allele and the O1, O2, B, and A2 alleles as described earlier [14]. All primers were prepared from Microsynth (Balgach, Switzerland) according to Gassner et al. [14] with the exception of the O1 primers [15]. Concentration of detection primers was 0.4 lM, the one of control primers 0.24 lM. PCR was performed in a UNO II thermal cycler (Biometra, Tampa, FL, USA) under following thermocycling parameters: denaturation of DNA at 94 C for 5 min, 35 cycles were performed consisting of 94 C for 1 min, 55 C for 1 min, 72 C for 1 min followed by a final extension at 72 C for 5 min and stored at 4 C until analysis. Gene amplification was confirmed running a 2% agarose gel containing 0.8 ll/ml ethidium bromide for UV transillumination at 302 nm.

F.J. Achermann et al. / Transfusion and Apheresis Science 32 (2005) 255–262

2.4. Secretor genotyping PCR-SSP typing of the Secretor status was done according to Procter et al. [16]. Thermocycling parameters were: denaturation of DNA at 94 C for 5 min, 28 cycles consisting of 94 C for 1 min, 55 C for 1 min, 72 C for 1 min followed by a final extension at 72 C for 5 min and stored 4 C until analysis. 2.5. Quantification of SAS in FFP by indirect antigen capture ELISA BRIC 145 capture antibody was diluted 1:50 and coated onto 96-well microtiter plates (Nunc MaxisorpTM, Nunc A/S, Rosklide, Denmark) overnight at 4 C. FFPs were diluted 1:10 with PBS pH 7.4 containing 2% BSA and 5% Tween-20. One type A plasma containing SAS at high level was defined as standard and serially diluted on each plate from 1:10 to 2560. One type O plasma was used as negative control. Plates were washed 3· with washing buffer (PBS, 0.05% Tween-20, pH 7.4) and 100 ll/well of diluted FFPs were added in duplicate and incubated for 90 min at 37 C. After washing the plates were incubated for 2 h at room temperature with Seraclone anti-A diluted 1:5. Plates were then washed and biotinylated goat anti-mouse IgM was used in a dilution of 1:1500 and allowed to bind for 1 h at 37 C. After a further washing step alkaline phosphatase labelled streptavidin (Amersham, Du¨bendorf ZH, Switzerland) was added in a dilution of 10
3 and incubated for 30 min at 37 C. The plates were finally washed 3· with 0.9% NaCl containing 0.05% Tween-20. 4-Nitrophenyl-phosphate (1 mg/ml) in 1 M diethanolamine pH 9.6 was subsequently used as substrate. Colour development was measured photometrically after 10 min respectively when the highest standard concentration showed an absorbance of 1.0 ± 0.1 at 405 nm. Level of SAS in individual samples was expressed as arbitrary units (aU) and calculated by linear regression using the standard curve. Tests to check the assay performance characteristics resulted in an inter-assay-variance of v@ = 0.148 by testing each of 35 SAS reactive samples with different levels on 9 different days.

257

The intra-assay-variance was v@ = 0.065 when testing 7 SAS reactive samples with different levels 15·. The lower limit of detection (LLD) was elected as the mean value + 2 SD of all tested type O and type B plasma samples and was set as 2.3 aU. 2.6. Quantification of the total and IgG classspecific anti-A TRCs of different FFP blood types by a microhemagglutination inhibition assay To test the total anti-A TRC, 6 plasma samples of each group: A1 Le(a
b
) secretors, A1 Le(a
b+), A1 non-secretors, A1B, A2, and A2B were randomly mixed undiluted in equal parts with 36 type O samples. Similarly, type O plasma samples were mixed 3· as negative controls. To visualise total anti-A TRC of the different FFPs all single type O samples were diluted 1:2 with dilution buffer (PBS pH 7.4 containing 1% BSA), which allowed us to refer anti-A titers of each plasma mixture to its corresponding titer of the type O plasma. To test the IgG class-specific anti-A TRC the specificity mapped murine class IgG BRIC 145 anti-A was similarly mixed with the same type A and AB samples and with dilution buffer as well. Mixtures were then incubated for 1 h at 37 C to allow SAS to neutralise anti-A. Reference samples and the corresponding mixtures were serially diluted up to 1:512 on the same Ubottom microtiterplate (Nunc A/S). Additionally, the S2 anti-A reagent was run on each plate to monitor the inter-assay stability of the test system. 50 ll/well of a 5% RBC suspension (Medion Diagnostics, Du¨dingen, Switzerland) was then added to a final volume of 100 ll. After mixing, plates were incubated for 1 h at 37 C. Agglutination titers were immediately determined by reading the streaking patterns against a well-lit background. Reduction by one or several titer steps was referred to as TRC. 2.7. Statistical analyses All statistical analyses were done using StatView software version 5.0.1 (SAS Institute Inc., Cary, NC, USA). Normal distribution of data was qualitatively tested through box-plot as well

258

F.J. Achermann et al. / Transfusion and Apheresis Science 32 (2005) 255–262

as percentile-plot analyses for the compared groups. For almost all data of all groups no normal distribution could be expected, which is why non-parametric power analyses were put in place to evaluate differences and dependencies. Therefore, Mann–Whitney U-test was used to test the null hypotheses, if the SAS level depends on the ABO type, the Secretor status, the expression of Lea and/or the Leb antigen. The dependency of TRC on the SAS level contained in FFP was evaluated through Spearman rank correlation. The appropriate null hypothesis was formulated single shaped. Furthermore, the same test was used to monitor the dependency of the TRC on individual anti-A titers of the randomly used type O samples, the appropriate null hypothesis was formulated double shaped. p-Values <0.05 were considered as being significant.

3. Results 3.1. ABO, Secretor and Lewis, pheno- and genotypes of the tested population Blood samples of 250 donors were pheno- and genotyped with respect to ABO; there was complete agreement between both procedures. 150 donors were blood type A, 50 donors were type AB, 20 were type B and 30 type O. Such distribution does not correspond to the one found in the Caucasian population, because we commissioned these ABO types for the purpose of this study.

The A1, Secretor and Lewis types of the A and AB donor samples are shown in Table 1. 3.2. Level of SAS in FFP of different blood types All A1 and A1B plasmas contained SAS. Four of 26 A2 samples lacked detectable SAS; three were from non-secretors. Only 3/12 A2B samples had detectable SAS; two were from non-secretors. The diagnostic specificity (D-SP) of the SAS-ELISA was 98% as there was only one type B sample, which was tested reactive for SAS, its aU was 3.8, close to the LLD. The diagnostic sensitivity (DSN) of the SAS-ELISA was 97% for type A plasma and 82% for type AB. D-SP and D-SN were determined using all 250 ABO typed plasma samples. Level of SAS was compared to the ABO and Secretor genotypes as well as to the phenotypic Lewis status. It depended primarily on the ABO type, whereby large differences were found between type A1 and type A2 donors (p < 0.0001) (Table 2). Co-expression of the B locus clearly led to a decrease in the level of SAS (p < 0.0001). A1, A1 B and A2 secretors had significantly higher SAS levels than their ABO analogues non-secretors (p < 0.02) (see also Fig. 1). SAS levels were higher in plasma from A1 Le(a
b+) secretors than A1 Le(a
b+) individuals. In contrast, expression of the Lea antigen did not affect the detected level of SAS in non-secretors. A1B non-secretors had significantly more SAS than A2 secretors (p = 0.0022). A1B secretors had SAS to equal extent than A1 non-secretors.

Table 1 Distribution of the different blood types with respect to ABO, Secretor and Lewis status

Se Le(a
b+) Le(a
b
) Non-Se Le(a+b
) Le(a
b
) Total Frequency (%)

A

A1

A2

AB

A1B

A2B

Total

Frequency (%)

120 101 19 30 25 5 150 100

100 86 14 24 20 4 124 82.7

20 15 5 6 5 1 26 17.3

35 33 2 15 10 5 50 100

25 23 2 13 8 5 38 76

10 10 0 2 2 0 12 24

155 134 21 45 35 10 200

77.5 67.0 10.5 22.5 17.5 5.0 100.0

81% of all type A and type AB donors expressed the A1 allele and 84.5% the Lewis b allele.

F.J. Achermann et al. / Transfusion and Apheresis Science 32 (2005) 255–262

259

Table 2 Overview of the SAS levels found in plasma of different blood types Blood type A AB A1 A2 A1B A2B O B A1 Le(a
b
) secretor A1 Le(a
b+) secretor A1 non-secretor A1B secretor A1B non-secretor A2 secretor A2 non-secretor A2B

SAS median

n

Minimum

Maximum

58.7 24.7 65.9 8.3 28.3 0 0 0

150 50 124 26 38 12 30 20

0 0 8.5 0 5.2 0 0 0

281.4 113.5 281.4 59.3 113.5 8.8 0 3.3

137.6 71.8 22.4 29.8 21.9 9.3 1.2 0

14 86 24 25 13 20 6 12

63.3 13 8.5 12.4 5.2 0 0 0

186.6 281.4 91.7 113.5 42.9 59.3 16.8 8.8

In contrast to Fig. 1 showing the mean SAS aU ± 1 SEM, this table lists Median, Minimum and Maximum SAS aU, which are the basis of the applied non-parametric statistical analyses (Mann–Whitney U-test). The upper part combines the SAS levels of the main ABO blood types irrespective of Secretor and Lewis status. The lower part lists the ABO types split for Secretor; A1 secretors are further split into Le(a
b
) and Le(a
b+).

Fig. 1. SAS levels in FFP as a function of ABO/Secretor genotypes and Lewis phenotypes. SAS levels are expressed as mean ± 1 SEM. p-Values were calculated by Mann–Whitney U-test. Not all tested blood types are represented as horizontal single bars: only A1 secretors are split into Le(a
b+) and Le(a
b
), since only for these individuals a statistically significant difference was shown; nonsecretors are not split into Le(a+b
) and Le(a
b
), since there were no statistical significant differences between these two groups; A2B are not split into secretors and non-secretors, since they showed hardly any SAS in our test system and only 2 of them were nonsecretors. As can be seen, the statistical difference between A1 non-secretors and A1B secretors was not significant. No significant difference between A2 non-secretors and A2B could be shown.

There were no statistically significant different SAS levels found between the genotypes A1/A1,

A1/A2 and A1/O as well as between A2/A2 and A2/O. Homozygous expression of the Secretor

260

F.J. Achermann et al. / Transfusion and Apheresis Science 32 (2005) 255–262

locus did not increase SAS levels compared to individuals who express Secretor heterozygously. 3.3. Anti-A isoagglutinin neutralisation by SAS evaluated by microhemagglutination inhibition assay

4. Discussion

Total anti-A TRC depended on the SAS level as shown by Spearman rank correlation test (r = 0.566; see Fig. 2). Despite this significant correlation (p = 0.0003) some single samples with high SAS levels had a low TRC. Conversely some single samples with low or even undetectable SAS levels substantially reduced the agglutination titers. In fact TRC depended also on the agglutination titer of the randomly used type O plasma samples (r = 0.470, p = 0.0037). To ensure that the used type O samples did not have significantly different agglutination titers among the tested groups, the Kruskal–Wallis test was applied revealing that this was not the case. IgG class-specific anti-A TRC was undetectable: the agglutination titer of the

TRC in titer steps 2.5

SAS aU

160 140

2.0

120 100

1.5

80 1.0

60 40

0.5

20 0

0

Fig. 2. TRC upon anti-A isoagglutinin neutralisation by SAS. Left panel shows the TRC (mean ± 1 SEM) of plasma samples with different blood types. Each symbol combines the estimated TRCs of 6 type A and AB plasma samples. The right panel displays the SAS levels of the same plasma samples. There was a strong dependency of the TRC on the SAS level (p = 0.0005). A1 Le(a
b
) secretors ( ) had the strongest TRC as well as the highest SAS amount. A1 Le(a
b+) ( · ) plasma samples of E-29: also had a strong TRC, according to their appropriate SAS level. No statistically significant difference between A1 non-secretor (h) and A1B ( ) individuals could been shown. A2 (n) plasma samples had a low SAS level and a moderate TRC. A2B (,) samples showed hardly any SAS but a low TRC.





reference sample and all plasma mixtures were 1:32.

Quantitative estimation of soluble carbohydrates such as SAS may be achieved using different procedures such as colorimetry, thin layer-, column-, gas- or high power liquid-chromatography. SAS occurs under different glycosylation conditions, molecular weights and chemical class substances (glyocolipids, glycosphingolipids, ceramid hexasacchairde; reviewed in [17]), association thus being extremely heterogeneous in terms of structure and chemical property [18]. The developed indirect antigen capture ELISA to quantify SAS as a function of ABO, Secretor and Lewis status revealed a plausible concentration pattern (Fig. 1); this is consistent with previous reports, where ABO and Secretor genotypes were not determined [19,20]. Since capture and detection antibodies were derived from different cell clones, the binding characteristics to the glycotopes they detect were further verified using TLC. This analysis revealed the specificity of these reagents against A types 1, 2 and 4, i.e. those recognised by human polyclonal anti-A (data not shown). Different SAS levels in human plasma may be explained by the encoding gene constellation ABO/Se/Le. As expected, secretors had much higher SAS amounts than non-secretors, since the Se-fucosyltransferase is expressed in epithelial tissues concerned with exocrine secretion [21]. Expression of Leb did significantly decrease SAS levels in plasma of A1 secretors. This reduction may conceptually be explained by the Le-fucosyltransferase conversion of A type 1 into ALeb, a glycotope structurally different from A type 1. It seems to be likely that the MoAbs being used for the SAS-ELISA are failing to capture or recognise the ALeb glycotopes, even though the MoAbs used react strongly with ALeb on TLC plates; probably this is due to the different antigen presentations. Such antigen presentation differences to MoAbs are well established between TLC and cell membranes. As shown by microhemagglutination inhibition assays, ALeb is either low or not recognised

F.J. Achermann et al. / Transfusion and Apheresis Science 32 (2005) 255–262

by human polyclonal anti-A, confirming the analytical specificity of the SAS-ELISA. In contrast to Leb there was no dependency of SAS on the expression of the Lea antigen in non-secretors. Type AB donors had less SAS in plasma than type A donors, because A and B glycosyltransferases attach their substrates on the same acceptor molecules thus entering competition. That A2B plasma contained almost no detectable SAS in ELISA, yet a moderate total anti-A TRC was seen, may indicate some lack in diagnostic sensitivity in contrast to the high diagnostic specificity. In this study, no type B or type H soluble blood group substances were quantified. It is probable that the amount of soluble B and H substance is regulated through the same genetic mechanism as SAS. So far PEX therapy for ABO incompatible transplant recipients is performed without regard to the amount of soluble ABO blood group substance in replacement plasma. As shown in this study, plasma contains widely different amounts of SAS, and the total Ig-class anti-A TRC clearly depended on the SAS level in FFP. In contrast no IgG class-specific TRC was detected using the BRIC 145 reagent. It is expected, that IgM does much more precipitate soluble ABH substance than IgG. It remains still unknown if in vivo adsorption of isoagglutinins has a beneficial effect for ABO incompatible transplanted patients. Such attempt to eliminate isoagglutinins might cause stimulation of B lymphocytes to produce immune, IgG class-specific antibodies to ABH structures at higher level several days after PEX and might favour antibody-dependent graft rejection. At least, PEX or double filtration apheresis performed with albumin replacement would not provide for uncertainty in extent of A-anti-A immune complex formation but potentially not contribute to more efficient neutralisation of antiA either.

Acknowledgments The authors acknowledge T. Schulzki, MD, Ch. Heiniger and her team at the Bern Blood Bank as well as the blood donors for consenting research

261

with their samples, A. Baumann, chief-technician, Inselspital Bern Hospital Blood Bank, for providing Seraclone, Steve Henry, Ph.D., and Henrik Clausen, Ph.D., for scientific advice. The study was supported by a grant from Octapharma GmbH, Seidenstrasse 2, 8853 Lachen SZ, Switzerland. References [1] Rydberg L. ABO-incompatibility in solid organ transplantation. Transfus Med 2001;11(4):325–42. [2] Wu A, Buehler LH, Cooper DKF. ABO-incompatible organ and bone marrow transplantation: current status. Transpl Int 2003;16:291–9. [3] Eastlund T. The histo-blood group ABO system and tissue transplantation. Transfusion 1998;38(10):975–88. [4] Mohacsi P, Rieben R, Nydegger UE. Immunosuppression in ABO-incompatible transplantation. Transplant Proc 2001;33(3):2223–4. [5] Ishida H, Tanabe K, Furusawa M, Ishizuka T, Shimmura H, Tokumoto T, et al. Mycophenolate mofetil suppresses the production of anti-blood type antibodies after renal transplantation across the abo blood barrier: ELISA to detect humoral activity. Transplantation 2002;74(8): 1187–1189. [6] Pierson 3rd RN, Loyd JE, Goodwin A, Majors D, Dummer JS, Mohacsi P, et al. Successful management of an ABO-mismatched lung allograft using antigen-specific immunoadsorption, complement inhibition and immunomodulatory therapy. Transplantation 2002;74(1):79–84. [7] Rieben R, Korchagina EY, von Allmen E, Hovinga JK, Laemmle B, Jungi TW, et al. In vitro evaluation of the efficacy and biocompatibility of new, synthetic ABO immunoabsorbents. Transplantation 1995;60(5):425–30. [8] Hanto DW, Fecteau AH, Alonso MH, Valente JF, Whiting JF. ABO-incompatible liver transplantation with no immunological graft losses using total plasma exchange, splenectomy, and quadruple immunosuppression: evidence for accommodation. Liver Transpl 2003; 9(1):22–30. [9] Yandza T, Lambert T, Alvarez F, Gauthier F, Jacolot D, Huault G, et al. Outcome of ABO-incompatible liver transplantation in children with no specific alloantibodies at the time of transplantation. Transplantation 1994;58(1): 46–50. [10] Cooper DK, Good AH, Ye Y, Koren E, Oriol R, Ippolito RM, et al. Specific intravenous carbohydrate therapy: a new approach to the inhibition of antibody-mediated rejection following ABO-incompatible allografting and discordant xenografting. Transplant Proc 1993;25(1 Pt 1): 377–8. [11] Zhong R, Luo Y, Yang H, Garcia B, Ghanekar A, Luke P, et al. Improvement in human decay accelerating factor transgenic porcine kidney xenograft rejection with

262

[12]

[13] [14]

[15]

[16]

F.J. Achermann et al. / Transfusion and Apheresis Science 32 (2005) 255–262 intravenous administration of gas914, a polymeric form of alphaGAL. Transplantation 2003;75(1):10–9. Campbell JE, Henry SM. Specificity mapping of monoclonal antibodies using glycolipids and immunostained thin layer chromatograms. Transfus Clin Biol 2001; 8(Suppl 1):251s. Yamamoto F. Molecular genetics of ABO. Vox Sang 2000;78(Suppl 2):91–103. Gassner C, Schmarda A, Nussbaumer W, Schonitzer D. ABO glycosyltransferase genotyping by polymerase chain reaction using sequence-specific primers. Blood 1996;88(5): 1852–6. Mifsud NA, Haddad AP, Condon JA, Sparrow RL. ABO genotyping-identification of O1, O1*, and O2 alleles using the polymerase chain reaction-sequence specific oligonucletide (PCR-SSP) technique. Immunohematology 1996;12: 149–53. Procter J, Crawford J, Bunce M, Welsh KI. A rapid molecular method (polymerase chain reaction with

[17]

[18]

[19]

[20]

[21]

sequence-specific primers) to genotype for ABO blood group and secretor status and its potential for organ transplants. Tissue Antigens 1997;50(5):475–83. Schenkel-Brunner H. Human blood groups. Chemical and biochemical basis of antigen specificity. Wien, New York: Springer; 2000. Clausen H, Hakomori S. ABH and related histo-blood group antigens; immunochemical differences in carrier isotypes and their distribution. Vox Sang 1989;56(1):1–20. Bernoco M, Danilovs J, Terasaki PI, Cartron JP, Mollicone R, Le Pendu J, et al. Detection of combined ABH and Lewis glycosphingolipids in sera of H-deficient donors. Vox Sang 1985;49(1):58–66. Tilley CA, Crookston MC, Brown BL, Wherrett JR. A and B and A1Leb substances in glycosphingolipid fractions of human serum. Vox Sang 1975;28(1):25–33. Oriol R, Le Pendu J, Mollicone R. Genetics of ABO, H, Lewis, X and related antigens. Vox Sang 1986;51(3): 161–71.