UvA-DARE (Digital Academic Repository)

Genetic basis of rare group variants

Wigman, L.

Publication date 2013 Document Version Final published version

Link to publication

Citation for published version (APA): Wigman, L. (2013). Genetic basis of rare blood group variants.

General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

Download date:06 Oct 2021

Genetic basis of rare blood group variants

Financial support was granted by: Joop en Annelies Wigman, Philippine Sanquin Blood Supply, Amsterdam Amsterdam Medical Center, Amsterdam MRC-Holland b.v., Amsterdam

ISBN: 978-90-5335-759-0

Cover: Wouter Wigman and Lonneke Haer-Wigman Lay-out: Simone Vinke, Ridderprint B.V., Ridderkerk, the Netherlands Printing: Ridderprint B.V., Ridderkerk, the Netherlands Genetic basis of rare blood group variants

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Agnietenkapel op dinsdag 3 december 2013, te 12:00 uur

door

Lonneke Wigman

geboren te Terneuzen Promotor: Prof. dr. C.E. van der Schoot

Copromotor: Dr. M. de Haas

Overige leden: Prof. dr. F. Baas Prof. dr. E. Bakker Prof. dr. M.H.J. van Oers Prof. dr. W.H. Ouwehand Prof. dr. J.J. Zwaginga

Faculteit der Geneeskunde Contents

Chapter 1 General introduction and scope of this thesis 7

Chapter 2 Comprehensive genotyping for 18 blood group systems using a 27 multiplex ligation-dependent probe amplification assay shows a high degree of accuracy

Chapter 3 RHD and RHCE variant and zygosity genotyping via multiplex 51 ligation-dependent probe amplification

Chapter 4 Characterization of known and novel RHD variant alleles in 37.764 79 Dutch D- pregnant women

Chapter 5 SMIM1 underlies the and influences traits 103

Chapter 6 Genetic screening for the Vel- phenotype circumvents difficult 123 serological screening due to variable Vel expression levels

Chapter 7 Molecular analysis of immunized Lan- or Jr(a-) patients and validation of 143 genotyping assays to screen blood donors for Lan- and Jr(a- )phenotype

Chapter 8 Familial azotemia is caused by a duplication of the UT-B transporter 161

Chapter 9 General discussion 173

Appendix Nederlandse samenvatting 183

Dankwoord 191

Curriculum Vitae 195

Lijst met publicaties 197

Portfolio 199

Chapter 1

General introduction and scope of this thesis

membrane of red blood cells (Figure 1). Blood group antigensare structures, , carbohydrates orlipids, present onthe blood group system 1). (Table blood group, ofwhichthegeneticbasiswasrecently elucidated, willbeassigned asthe34 are bloodgroup . M . group antigens, it took more than 25 years to discover another blood group antigen, the pressure pathogens. ofvarious ofblood group The hugevariety antigensisthoughtto have dueto evolutionary arisen Origin ofbloodgroup antigens assumed to originate from pressure ofmicrobial pathogens. malaria aremalaria responsible for differential expression ofblood on proteingroup antigenscarried serologically by characterized humanantibodiesandtheantigenmustbeinheritable. completely thesame. 1901,Landsteiner discoveredIn thatred bloodcellsofdifferent healthy humans were not Blood group antigens cells to animmunized recipient. Consequently,reaction. to transfuse compatible important itisvery donorred blood can destroy alldonorred bloodcells, whichresults inasometimesfatal transfusion present onthetransfused donorred bloodcells, theimmunesystem oftherecipient a recipient ofared bloodcell transfusion hasanantibody to abloodgroup antigen ofred bloodcellstransfusion. bloodgroup thesurface On antigens are present. When be transfused. Donor blood cannot be simply transfused to patient every in need of anemia in patients disorder, with an erythropoiesis red blood cells from a donor can cause constitutional anemia. To relieve massive trauma bloodlossafter orto relieve some individualshave (for instance inerythropoiesis adefect thalassaemia)whichcan to andbrain, vitalorgans, resultingtransport indeath. suchastheheart Furthermore, out ofthebody. Amassive lossofred bloodcells willleadto thefailure ofoxygen essential for humanlife. trauma During hugeamounts ofred bloodcells canflow anddeliver bloodcellsoxygenRed carry to allcells ofourbodyandare therefore sensitivity ofserologicalsensitivity bloodgroup typing. of theindirectantiglobulintest by Coombs etal. in1945,whichdramaticallyincreases the group red 1). andtwo (Table bloodcellantigenseries collections group bloodgroup whichare antigens are divided intosystems, known, thirty-three sixblood and named them A, B and O. 2-4 ofbloodgroupThe majority antigenswere discovered thedevelopment after 1 Hedetected three differences onthemembraneof red bloodcells 6,8,9 To beentitledasabloodgroup antigen,theantigenmustbe 1 10,11 Although intensive research was done to more detect blood The different antigensof thecarbohydrate blood groups are 6,8,9 Not all structureson the red blood cellmembrane 2,5 At thismomentmore than 300different blood 12,13 Parasitic pathogens suchas 6,7 It is expected thatthe isexpected It Vel General introduction General 6,8,9

9

Chapter 1 Chapter 1 structures.14,15 The effect of malaria on blood group antigen expression can still be observed via the geographic distribution of specific blood group phenotypes.14-16 For instance, in West- Africa almost 100% of the population lacks the complete expression of the Duffy glycoprotein, which makes them resistant for malaria infection by Plasmodium vivax.14,16 Nevertheless, for many blood group antigens the evolutionary pressure is not known and possibly even non existing. Differential blood group antigen expression might also be due to genetic drift, for many blood group antigens the geographic distribution can simply be explained by founder effects.17

Figure 1. Model of structures that carry blood group antigens on the red blood cell membrane. (modified from Reid et al. 2013; ref6).

Molecular basis of blood group antigens The first blood group antigens for which the molecular basis was clarified were the M and N antigen, when Siebert and Fukuda cloned the GYPA in 1986.18 Due to the rapid evolvement of DNA techniques, the genetic basis of the majority of blood group antigens was elucidated in less than ten years.10,11,19 At this moment there are, however, still blood group antigens of which the genetic basis is not elucidated, for instance the Ata and Emm blood group antigens.6,9 Differential expression of blood group antigens can be due to single

10 Table 1. Overview of the 34 blood group systems, 6 blood group collections and 2 red blood cell antigen series

ISBT ISBT name Number Chromo- Gene Anti- Frequency Genetic difference(s) between antigen 1 and antigen 2 Antigen 2 Frequency No of some loca- gen 1 of antigen of antigen antigens tion 1* 2* Blood group systems 1 ABO 4 9q34.1-q34.2 ABO A 47% c.297A>G† c.526C>G c.657C>T† c.703G>A c.796 c.803 c.930 B 13% C>A G>C G>A† 2 MNS 46 4q31.21 GYPA M 78% c.59C>T c.71G>A c.72T>G N 72% GYPB S 55% c.143C>T S 89%

3 P1PK 3 22q13.2 A4GALT P1 79% c.42C>T P2 21% 4 Rh 52 1p36.11 RHD D 83% Deletion of RHD‡ Null 17% phenotype RHCE C 68% c.48C>G c.178A>C c.201G>A c.203G>A c.307 c 80% T>C E 29% c.676C>G e 98% 5 Lutheran 20 19q13.32 LU Lua 8% c.230G>A Lub 99.8% 6 Kell 34 7q34 KEL K 9% c.578C>T k 99.8% Kpa 2% c.841C>T Kpb 100% 7 Lewis 6 19p13.3 FUT3 Lea 22% § Leb 72% 8 Duffy 5 1q23.2 DARC Fya 66% c.125G>A Fyb 83% 9 Kidd 3 18q12.3 SLC14A1 Jka 77% c.838G>A Jkb 74% 10 Diego 22 17q21.31 SLC4A1 Dia 0.01% c.2561C>T Dib 100% Wra <0.01% c.1972G>A Wrb 100% 11 Yt 2 7q22.1 ACHE Yta >99.8% c.1057C>A Ytb 8% 12 Xg 2 Xp22.33 PBDX Xga 89% / Not determined General introduction General 66%** MIC2 CD99 100% Not determined Null Rare phenotype 13 Scianna 7 1p34.2 ERMAP Sc1 >99% c.169G>A Sc2 1%

a b 11 14 Dombrock 8 12p12.3 ART4 Do 67% c.378C>T c.624T>C† c.793G>A Do 82%

Chapter 1 Chapter 1 2* 10% 99% Rare Rare Rare Rare Rare Rare Rare Rare Rare <1% <1% 3.5% <0.1% <0.01% >99.9% of antigen of antigen Frequency Frequency b b b a- b a- R In Cr Kn Co LW Ok Null Null Null Null Null Null Mer- Ge-2 Fors1- JMHK- Ge-2-3 Duclos- Antigen 2 Antigen phenotype phenotype phenotype phenotype phenotype phenotype c.3670C>G c.3630G>C c.3669T>G c.3629G>T c.3660T>C c.422G>A‡ Deletion of exon 2 of GYPC Deletion of exon Deletion of exon 3 of GYPC Deletion of exon c.3620C>T c.3567A>G Genetic difference(s) between antigen 1 and antigen 2 1 and antigen antigen between Genetic difference(s) c.299A>G c.679G>C c.134C>T c.4681G>A c.137G>C c.376C>T‡ Deletion of XK ‡ c.274G>A c.511C>T‡ c.619C>T c.1049G>A c.202C>T‡ c.316G>A c.IVS5+1G>A c.887G>A‡ 1* 96% 100% 100% 100% 100% 100% 100% 100% >98% >99% 99.5% 94.5% 0.10% 99.9% <0.1% 92%†† >99.9% >99.9% >99.9% >99.9% of antigen of antigen Frequency Frequency - a a a a a a a x I P H K Jr In Cr GIL Kn Ok Co LW Rg1 Ch1 Ge2 Ge3 JMH Anti Fors1 MER2 gen 1 Duclos XK C4B CR1 BSG C4A DAF FUT1 GYPC CD44 AQP1 AQP3 RHAG Gene ICAM4 CD151 CD108 GBGT1 ABCG2 GCNT2 B3GALNT1 - tion 11p13 1q32.2 2q14.3 1q32.2 3q26.1 9q34.2 4q22.1 7p14.3 6p24.2 6p21.3 9p13.3 Xp21.1 15q24.1 11p15.5 19p13.2 19p13.3 6p21.32 19q13.33 Chromo some loca - 1 1 9 1 3 4 3 9 4 6 1 1 4 1 1 1 of 16 11 Number antigens I JR Kx Ok Hh Gill JMH FORS GLOB RAPH RHAG Knops Indian Colton Chido- Wiener Cromer Gerbich Rodgers ISBT name ISBT Landsteiner- 19 21 18 20 22 25 16 17 23 24 15 26 27 28 30 29 31 32 No ISBT ISBT

Table 1. (Continued) Table

12 Table 1. (Continued)

ISBT ISBT name Number Chromo- Gene Anti- Frequency Genetic difference(s) between antigen 1 and antigen 2 Antigen 2 Frequency No of some loca- gen 1 of antigen of antigen antigens tion 1* 2* 33 LAN 1 2q36 ABCB6 Lan >99.9% c.197_198insC‡ Null <0.01% phenotype 34w Vel 2 1p36.32 SMIM1 Vel >99.9% c.64_80del Null <0.1% phenotype Blood group collections 205 Cost 2 Csa >98% Csb 34% 207 Ii 1 i 208 Er 3 Era 100% Erb <0.01% 209 Globoside 2 LKE 98% 210 unnamed 2 Lec 1% 213 MN CHO 6 Hu 1% Red blood cell antigen series 700 Low- 18 By Rare Incidence antigens 900 High- 6 Ata 100% Incidence antigens * When the frequency differs between populations, the frequency of the Caucasian population is stated † Silent mutations General introduction General ‡ Multiple mutation can each cause the phenotype of antigen 2, the most frequently occuring mutation is stated § The presense or absence of the Lea and Leb antigens is determined by the functionality of multiple fucosyltransferases encoded by the FUT gene family ** Differential expression between females and males, respectively †† Most likely the frequency of the RAPH antigen is higher, the expression is, however, in many persons to low to detect with standard serological typing 13

Chapter 1 Chapter 1 nucleotide polymorphisms, multiple nucleotide polymorphisms or genetic rearrangements (Table 1).10,19 Described below are three blood group systems, the ABO, Rh and Kell blood group systems, that provide an example for each of the three genetic mechanisms.

The ABO blood group system The antigens of the ABO systems are carbohydrate structures.20,21 The difference between the A and B antigen is the presence of respectively a N-acetyl-D-galactosamine or a D-galactose at the end of a carbohydrate carrier structure.22,23 Individuals who have the O antigen lack the presence of any carbohydrate at this position.24 The ABO gene that is responsible for the expression of the ABO antigens, codes for an enzyme of the glycosyltransferase cluster. Glycosyltransferases attach carbohydrates to an oligosaccharide acceptor chain.25 Individuals with blood group A have an allele of ABO coding for a1,3-N-acetylgalactosaminyltransferase [A transferase] that can only attach the N-acetyl-galactosamine carbohydrate.26 Individuals with blood group B have an allele that codes for the highly similar a1,3-galactosyltransferase [B transferase], but this enzyme can only attach the galactose carbohydrate.27 The difference in enzymatic specificity between the A and B transferases is due to four amino acid changes caused by four nucleotide polymorphisms in ABO (Table 1).28 Individuals who have the O blood group have a mutated allele that codes for a non-functional and therefore lack the expression of A or B transferase.21,29 Many different alleles of the ABO gene are known that give rise to the O antigen, the most frequently occurring O allele has a deletion of one single nucleotide.21,29

The The Rh blood group system is encoded by two very homologous , RHD and RHCE, that code for the RhD and RhCE protein respectively.30-32 Individuals who are D negative [D-] lack the expression of the RhD protein.32,33 In the Caucasian population 17% of the population is D-, which is almost always caused by the complete deletion of RHD (Table 1).6,32,33 In the Negroid population 8% of the population is D-, in this population the D- phenotype is in 19% due to a complete deletion of RHD, in 66% due to the RHD*Ψ allele (RHD*Pseudogene) with a 37 insertion and a premature stopcodon and in 15% due to the hybrid RHD*03N.01 allele.6,34-36 In the Asian population the D- phenotype is very rare less than 0.5% of the population is D-.6,37 Next to the normal D positive [D+] and D- expression, three types of variant RhD expression exist: weak D, Del and partial D expression.6,9,36 Individuals with weak D expression express the complete RhD protein, however, in low quantities; between 100 to 5000 RhD antigens per red blood cell compared to the 10,000 till 30,000 antigens in normal D+ individuals.6,9,36 Weak D expression is most often caused by single mutations in RHD that cause an amino acid change in the transmembrane or intracellular parts of the protein.38 The Del phenotype is also characterized by weak expression of the complete RhD protein. However, individuals with the

14 frame. time antigensitisdifficult toobtaincompatibledonorfrequency redblood cellsinashort premature stopcodon ordisruptionofasplicesite inKEL. that causes a single amino acid difference (Tablethat causesasingleaminoaciddifference 1). the K and its antithetical k antigen is due to the difference of a single nucleotide polymorphism D variant express an RhD protein that lacks one or several of the thirty D-epitopes. express proteinD variant anRhD thatlacksoneorseveral ofthethirty antigen. itdifficult toobtainlargemakes numbersofdonorswho are negative for a high frequency antigen, for instance0.04%ofthepopulationis Vel- andonly0.005%ofthepopulationisLan-, rare. sites region ormutations intheC-terminal protein. oftheRhD D instance theKp the bloodgroupThe antigensoftheKell system onaprotein, are alsocarried whichisencoded by The bloodgroup Kell system the red bloodcellsneedto bethawed, washedandshippedto thepatient. beforeamount oftimeittakes thered bloodcells can betransfusedinto apatient,because blood are thehighcostsof freezing andmaintainingfrozen red bloodcellsand thelarge K/k antigens (Table 1). K/k antigens(Table a singleaminoacidchange, butatadifferentKell protein positiononthe compared to the also weakened compared expression.also weakened RhD to normal than 90% of the individuals of most populations. blood group frequency High antigens are blood group antigens that are expressed in more bloodgroup frequency High antigens refers: D-elution. technique, to sensitive whichthenameofthisphenotype absorption-elution with thevery Blood intheNetherlands, untila patient isinneedoftheblood. antigen thisdonorbloodisfrozen andstored, for instance, in theSanquinBankofFrozen antigen. high frequency To of overcomedonor blood negative for shortage a high frequency the donor population is not routinely lacking, to identify donors that typed are negative for a (an) extracellular loop(s) of the RhD protein. loop(s)oftheRhD (an) extracellular dueto singleormultiplemutationsthatcauseaminoacidchangesin canalsoarise variants certain D-epitopes because the CE counterparts are present atthesepositions. becausetheCEcounterparts D-epitopes certain homologous ofRHDare exchanged from arise withthevery hybrid variants allelesinwhichparts partial el phenotype have phenotype suchlow expression protein oftheRhD thatexpression isonlydetectable KEL gene. 56-58 9,62 iscausedby aheterogeneousThis phenotype setofmutationsresulting ineithera 6,9 The low occurrence ofindividualswhoare negative for bloodgroup ahighfrequency Furthermore, because for antigens reliablereagents most high frequency typing are 51-53 RHCE a anditsantitheticalKp However, incontrastto RhD, the differential expression between for instance 39 The D The gene. 55 protein,The complete isvery nullphenotype, lackoftheKell theKell el 46-48 phenotype is most often caused by mutations that disturb splice causedby ismost often mutationsthat disturb phenotype These folded hybrid allelesencodeacorrectly protein thatlacks b antigen, thedifferential expression isalsocaused by 48-50 In many partial D variants theexpression Dvariants is many partial In 6,8,9 54 9 For bloodgroup otherKell antigens, for

For patients with to high 6,59-61

37,40-42 63 Disadvantagesoffrozen Individuals with a partial withapartial Individuals 64

General introduction General 46-48 Partial D 43-45 Most Most 15

Chapter 1 Chapter 1

Little was known about the high frequency antigens Vel, Jra and Lan at the start of this project.9 All three are not expressed on granulocytes, monocytes or lymphocyte and all three antigens are resistant to enzymatic and chemical treatments of red blood cells using for instance bromelain, chymotrypsin, DTT or acid.6,65,66

The Vel antigen The Vel antigen was first described in 1952 in a patient who suffered from an acute hemolytic transfusion reaction after a .67 The identified was directed against a high frequency antigen, subsequently named “Vel”.67 Multiple severe hemolytic transfusion reactions in patients with anti-Vel have since been described.68-71 A large individual variation in the level of Vel expression is present and in some people the Vel expression is very weak, making it difficult to correctly type for the presence and absence of the Vel antigen.72

The Jra and Lan antigens The first examples of respectively the Jr(a-) and the Lan- phenotype were described at congresses in 1970 and 1961.73,74 Both anti-Jra and anti-Lan are able to induce a severe hemolytic transfusion reaction to Jr(a+) or Lan+ red blood cells, respectively.73,75 The capacity of anti-Jra to destroy red blood cells seems to increase after each encounter with the Jra antigen.75,76 Moreover, hemolytic transfusion reactions due to anti-Jra seem to get more severe after each incompatible blood transfusion and also hemolytic disease of the fetus and newborn was getting more severe in subsequent of a woman with anti-Jra.75,76 Recently, the genetic basis of the Jr(a-) phenotype and the Lan- phenotype were elucidated.77-79 The Jra and Lan antigens are both carried on a protein structure, the ABCG2 or ABCB6 protein, respectively, and negativity is caused by the complete absence of the protein.77-79 Multiple mutations have been described in ABCG2 or ABCB6 that can each cause negativity for the Jra or Lan antigen, respectively.77-82 The Jr(a-) phenotype is most often caused by the c.376C>T, c.706C>T or c.736C>T mutation in ABCG2 which all introduce a premature stopcodon.77,78,80 The molecular basis of Lan- shows large heterogeneity, in 49 Lan- persons a total of 24 different mutations are described.79,82 Only three mutations, c.459del, c.574C>T and c.2256+1G>A in ABCB6, have been detected in five or more Lan- individuals.82 Before the genetic basis of the Jra and Lan antigens were linked to the ABCG2 and ABCB6 protein, respectively, these two proteins were already studied for their putative porphyrin transporter function.83 In individuals who are Jr(a-) or Lan- porphyrin levels are elevated, confirming the function of ABCG2 and ABCB6 as porphyrin transporter.77,79 Heme is a porphyrin and ABCG2 and ABCB6 may play a role in heme uptake into the red blood cell. No aberrant red cell parameters were, however, detected in Jr(a-) and Lan- individuals.77,79

Allo-antibodies to blood group antigens

16 to the activation ofthecompleteto complementcascade. theactivation ofthe red byhemolysis is characterized the destruction blood cells in the bloodstream due positive for anantigen to which he/she has made antibodies. transfusion hazard. safe, Although transfusionofred bloodcellsisgenerallyvery HTRsremain animportant Hemolytic transfusion reaction red bloodcellare destroyed . viaintravascular orextravascular of red blood cells via phagocytosis byof red macrophages bloodcellsviaphagocytosis inthespleenandliver. carrying foreign antigens,carrying willbecleared they by thespleen. normally that an individual lacks on his/her own red blood cells. antibodies prime theredantibodies prime bloodcellsfor enhancedclearance. antibodies thatare notableto completely thecomplementsystem, activate buttheallo- positive for thecorresponding antigen. are becausethey ableto destroy are clinicallyimportant, Allo-antibodies red cellsthatare Clinical implications ofbloodgroup allo-antibodies cases canbefatal. destroyed andthe cytotoxic hemoglobinisreleased into thebloodstream, whichinsevere intravascular hemolysis. antibodies to theABObloodgroup antigensandthe Vel offenders antigenare of known antigen. are antibodiesinindividualswhothemselvesAllo-antibodies lackthecorresponding HDFN is a disease in which the red blood cells of the fetus are and/or newborn destroyed Hemolytic diseaseofthefetusand newborn [HDFN]. and newborn hemolytic reactions:transfusionreaction[HTR]anddiseaseofthefetus antigen an immune response occurs and allo-antibodies areantigen animmuneresponse produced occursandallo-antibodies to theantigen. in a pregnant woman delivery. or during encountered a transfusion of red during the fetal-maternal blood cells or during blood contact the symptoms are, however, generallylesssevere. hemolysis aremain complications of extravascular the same as in intravascular hemolysis, lack. with, for instance, bloodgroup Awilldevelop antibodies to bloodgroup themselves Bthey immune system viaHLAclassII. cleared red bloodcellsare processed andpeptidesfrom theantigensare presented to the blood group allo-antibodies. antigenshave naturallyoccurring 6,9,21 Immune allo-antibodies only arise after anencounter after withablood group onlyarise allo-antibodies antigen Immune 9,64 Allo-antibodies can be naturally occurring or immune antibodies. can be naturally occurring Allo-antibodies 97-99 95,96 Not all allo-antibodies are ableto Notallallo-antibodies causeintravascular hemolysis, allo- 64,93,94 AnHTRoccurswhenanindividualistransfusedwithred bloodcells 93 Extravascular hemolysis is characterized by hemolysisischaracterized theenhanced clearance Extravascular

90,91 When theimmunesystem apeptidefrom detects aforeign 84-86 64,93,94 When an individual encounters red blood cells Allo-antibodies play a role in two kinds of play arole kinds Allo-antibodies intwo 94 Extravascular hemolysisinvolves Extravascular allo- 97,98 6,9 64,93,94 These “foreign antigens” can be The red bloodcellsare instantly 93 6,9,21 During anHTR the transfused During Soon after birth everybody everybody birth after Soon 87-89 Antigensfrom the 64,93,94 General introduction General Intravascular Intravascular 6,9,64 The ABO The 97,98 89,92 The

17

Chapter 1 Chapter 1 by allo-antibodies from the mother.86,100,101 Antibodies of a pregnant woman including allo- antibodies to red blood cell antigens are able to cross the maternal/fetal placental barrier and when the red blood cells of the fetus are positive for the corresponding antigen they will be destroyed due to the maternal antibodies.102 In severe cases of HDFN the fetus becomes anemic due to the large destruction of his/her red blood cells.86 Anemia in the fetus can only be treated with intra-uterine transfusions.86 Furthermore, mild cases of HDFN, that do not need intra-uterine transfusion, can become life-threatening after birth.103 Babies with HDFN have higher bilirubin concentrations, because bilirubin is released into the blood stream when red blood cells are destroyed.104 Before birth, bilirubin is transferred to the mother who excretes it.104-106 After birth the newborn’s liver is too immature to clear the high levels of bilirubin and deposits of bilirubin may occur in the skin or in the brain nuclei, which can cause brain damage.104-106 The removal of bilirubin can be stimulated by phototherapy or bilirubin concentrations can be lowered via .104 HDFN is most often caused by allo-antibodies to the RhD blood group antigen, moreover anti-D often causes severe HDFN.101 Since the introduction of the administration of anti-D prophylaxis in D- pregnant woman in 1960 the amount of cases of HDFN due to anti-D is substantially decreased.107,108

Blood group antigen typing For safe it is important to correctly characterize blood group antigens in blood donors, recipients of red blood cells and pregnant women.64,93,94 At this moment blood group typing is most often done via serological typing.6,9,64 Serological typing of red blood cell recipients and/or donors is, however, not always possible.109 Firstly, after transfusion or in case of the presence of auto-antibodies or multiple allo-antibodies, serological typing is very difficult and less reliable.110,111 Secondly, for most blood group antigens, serological typing relies on the availability of polyclonal human antisera, hence for some antigens it is impossible to screen large cohorts of donors due to the scarcity of the sera.112,113 Furthermore, the available sera are not always able to correctly detect variant antigen expression.48,112,114 It is important to correctly detect the variant expression of a blood group antigen, because red blood cells with weak antigen expression can immunize recipients and can cause a HTR.115,116 Moreover, recipients with a partial variant of a blood group antigen also need to be correctly identified, because they are able to produce allo-antibodies to the epitopes they miss.48,114 In cases in which is not feasible, it is possible to determine the blood group antigen status via genotyping.110,111,117 Several blood group genotyping assays have been developed that are able to determine the phenotype of multiple blood group antigens in one test.118-123 Genotyping has already been used for screening of rare blood group phenotypes in donors and some blood banks even have started to implement genotyping of blood donors as an alternative to serological typing.111,113,119,124

18 the variable levels of the variable Vel expression. We intended to theeffect characterize ofmutations next As stated above ofthe characterization Vel- difficultdue to viaserology phenotype isvery background ofthe Vel thatleadsto the antigenandthegeneticvariation Vel- phenotype. was notyet elucidated. how chapter we 5itisdescribed have In thegenetic determined could be verified, however, of this thesis the genetic basis of the at the start Vel- phenotype expressionvariable levels ofthe Vel antigen. Via for genotyping, negativity the Vel antigen rare Vel- red blood cells.of Characterization Vel- difficult duethe donors via serology is very reaction whentransfusedwith Vel+ red bloodcells. Therefore shouldonly receive they very Persons immunized to the Vel asevere antigenmightexperience hemolytictransfusion to concludetheclinicalrelevance ofthese rare (chapter 4). variants willbedetermined,alleles are RhD phenotype detected, alleleonthe theeffect ofthevariant expression pregnant lowextremely orabsentRhD inDutch D- women. D variant When new serologically women peryear, D- with to rare determine variants Rh gave ustheopportunity will berecognized by thisgenetictest. The available datafrom thescreening of~25.000 for thepresence ofaD+fetus. pregnant Dallele, avariant When aD- this woman carries To prevent pregnant Dimmunization,allDutch D- women are offered agenetic test to screen alleles(chapter 3). RHvariant frequently occurring We for technique aimedto thatcantype develop thepresence agenotyping ofthemost ofthealleles viaSangersequencingdifficult. characterization RHD andRHCE,whichmakes alleleshavepopulation. Many dueto ofthesevariant geneticrearrangements arisen between the allelesoccurfrequentlyalleles. oftheseRHvariant inspecificpopulations, Some for instance bloodgroupThe Rh resulting system haslarge inmore geneticvariety than250RHvariant probe amplification[MLPA]using themultiplexligation-dependent technique (chapter 2). antigens. We therefore aimedto assay developfor anew to genotype bloodgroup antigens expensive dedicated alimited equipmentandcanonlydetect numberofbloodgroup multiplebloodgroupbeen developed to genotype antigensinonetest. These tests require common bloodgroup recent antigens. years In assays several dedicated have genotyping To examinerare bloodgroup antigens, we firsthad to beable thepresenceto detect of expression. of the effect of these rareand characterization variants blood group on antigen variants The research presented inthis thesisfocuses onthegeneticbackground ofrare bloodgroup Scope ofthisthesis RHD*06.01 alleleintheCaucasian populationortheRHCE*01.04 allele intheBlack Scope ofthisthesis Scope 19

Chapter 1 Chapter 1 and polymorphisms in SMIM1 on the level of Vel expression. Furthermore, a high-throughput genotyping assay was developed to overcome all technical difficulties of serological identification of Vel- donors (chapter 6).

For the high frequency antigens Jra and Lan antigens it is also very difficult to obtain large numbers of negative donors. Due to the rarity of the Jr(a-) and Lan- phenotype and serological impairments, Jr(a-) and Lan- red blood cells are only available in frozen stocks. In chapter 7 we investigated the genetic basis of the Jr(a-) and Lan- phenotype in the Dutch population and we aimed to develop a high throughput screening assay for the detection of Jr(a-) and Lan- donors.

Dominant familial azotemia is a very rare disease, which has only been described in two families. The Kidd blood group antigen is carried on a urea transporter, which is also expressed on kidney cells. The Kidd null phenotype is linked with impaired urea homeostasis. In chapter 8 we investigated the genetic background and aberrant urea transport in a Dutch family with dominant familial azotemia.

Finally, in chapter 9 a discussion of all results described in this thesis and of blood group genotyping in general is given.

20 26. 25. 24. 23. 22. 21. 20. 19. 18. 17. 16. 15. 14. 13. 12. 11. 10. 9. 8. 7. 6. 5. 4. 3. 2. 1. References blood-group Acharacter.blood-group Biochem.J. 1968;109(2):315-7. Hearn VM, SmithZG, Watkins associated withthehuman WM. An a-N-acetyl-D-galactosaminyltransferase system. Nature 1990;345(6272):229-33. Yamamoto F, ClausenH, White groupT, geneticbasisofthehisto-blood S.Molecular J, ABO Hakomori Marken Vox Sang. 1959;4(2):97-119. Watkins WM, Morgan WT. Possible geneticalpathways for thebiosynthesisofbloodgroup mucopolysaccharides. substances. Nature 1955;175(4459):676-7. Watkins whichdecompose theblood-group WM, Morgan by simplesugarsofenzymes WT. Inhibition substances andsimplesugars. Br.J.Exp.Pathol. 1953;34(1):94-103. Morgan WT, Watkins WM. The inhibitionofthehaemagglutininsinplantseedsby humanbloodgroup 2009;25(2):48-59. JR,OlssonML. Storry The ABObloodgroup system revisited: areview andupdate. Immunohematology. glycosyltransferases, andABOgenes. Immunohematology. 2004;20(1):3-22. Yamamoto F. ABObloodgroup antigens, system—ABH oligosaccharide anti-Aandanti-B, AandB Review: Daniels G. The moleculargeneticsofbloodgroup polymorphism. Transpl.Immunol. 2005;14(3-4):143-53. Fr.Transfus.Immunohematol. 1986;29(4):251-66. PD,Siebert Fukuda biological M.Molecular studyofthestructureandexpression ofhumanglycophorinA.Rev. M. Kimura ofmolecularevolution: areview ofrecent evidence. Jpn.J.Genet.The neutraltheory 1991;66(4):367-86. Gebremedhin A,etal. oftheDuffyblood The globaldistribution group. Nat.Commun. 2011;2:266. Howes RE,Patil AP, PielCW, OA, FB, PW, Nyangiri Kabaria Gething PA, Zimmerman C,BeallCM, Barnadas to Plasmodium vivax.Adv.Parasitol. 2013;81:27-76. PA,Zimmerman Ferreira MU, Howes RE,Mercereau-Puijalon O. andsusceptibility bloodcellpolymorphism Red FyFy. genotype, blood-group N.Engl.J.Med. 1976;295(6):302-4. SJ, LH,Mason Clyde DF, MH. McGinniss The resistance to factor Plasmodium vivaxinblacks. The Duffy- Glycobiology 1999;9(8):747-55. Gagneux P, considerations in relating to diversity biological oligosaccharide A. Varki Evolutionary function. 2010;16(3):295-301. expressing bloodgroup bacteria Rivera-Marrero antigen.Nat.Med. C,XiaB, kill etal. immunelectins Innate Rodrigues CM,Dias-Baruffi M, Stowell Gourdine SR,Arthur LC, JP, Heimburg-Molinaro J,Ju Molinaro T, RJ, basis ofbloodgroup GA.Molecular Denomme expression. Transfus.Apher.Sci. 2011;44(1):53-63. basisofbloodgroup JR,OlssonML.Genetic diversity.Storry Br.J.Haematol. 2004;126(6):759-71. Daniels G.HumanBloodGroups. 2ed. Oxford: 2002. Science; Blackwell Vox Sang. 1990;58(2):152-69. CP, et al. Blood group terminology 1990. The ISBT on Party Working Terminology Antigens. for Cell Red Surface Lewis M,Anstee DJ, Bird BrodheimJP, GWG, E,Cartron Contreras M,Crookston MC,Dahr W, DanielsG,Engelfriet report. Berlin Voxterminology: Sang. 2011;101(1):77-82. ofBlood Society et al. International Transfusion onred Party Working cellimmunogeneticsandbloodgroup JR,Castilho L,DanielsG,Flegel Storry WA, G,Francis JJ, JM,Moulds Garratty Olsson ML,Poole CL,Moulds J, 2012. ME,Lomas-FrancisReid C,OlssonML. The BloodGroup AntigenFacts 3ed. SanDiego:Academic Book. Press; Pathol. 1945;26:255-66. test RR.Anew for ofweak AE,Race agglutinins.Coombs thedetection andincomplete RR,Mourant Rh Br.J.Exp. 1927;24:941-2. Levine P.Landsteiner K, onindividualdifferences Further ofhumanblood. observations Proc.Soc.Exp.Biol. 1927;24:600-2. Levine P.Landsteiner K, differentiating agglutinablefactor Anew individualhumanbloods. Proc.Soc.Exp.Biol. ME.Bloodgroups:Daniels G,Reid thepast50years. Transfusion 2010;50(2):281-9. humanblood. Onagglutinationofnormal Landsteiner K. Transfusion 1961;1:5-8. General introduction General 21

Chapter 1 Chapter 1

27. Race C, Ziderman D, Watkins WM. An alpha-d-galactosyltransferase associated with the blood-group B character. Biochem.J. 1968;107(5):733-5. 28. Procter J, Crawford J, Bunce M, Welsh KI. A rapid molecular method (polymerase chain reaction with sequence- specific primers) to genotype for ABO blood group and and its potential for organ transplants. Tissue Antigens 1997;50(5):475-83. 29. Yamamoto F, McNeill PD, Hakomori S. Genomic organization of human histo-blood group ABO genes. Glycobiology 1995;5(1):51-8. 30. Cherif-Zahar B, Bloy C, Le Van Kim C, Blanchard D, Bailly P, Hermand P, Salmon C, Cartron JP, Colin Y. Molecular cloning and protein structure of a human blood group Rh polypeptide. Proc.Natl.Acad.Sci.U.S.A 1990;87(16):6243- 7. 31. Le Van KC, Mouro I, Cherif-Zahar B, Raynal V, Cherrier C, Cartron JP, Colin Y. Molecular cloning and primary structure of the human blood group RhD polypeptide. Proc.Natl.Acad.Sci.U.S.A 1992;89(22):10925-9. 32. Arce MA, Thompson ES, Wagner S, Coyne KE, Ferdman BA, Lublin DM. Molecular cloning of RhD cDNA derived from a gene present in RhD-positive, but not RhD-negative individuals. Blood 1993;82(2):651-5. 33. Colin Y, Cherif-Zahar B, Le Van Kim C, Raynal V, van Huffel V, Cartron JP. Genetic basis of the RhD-positive and RhD-negative blood group polymorphism as determined by Southern analysis. Blood 1991;78(10):2747-52. 34. Singleton BK, Green C, Avent ND, Martin PG, Smart E, Daka A, Narter-Olaga EG, Hawthorne LM, Daniels G. The presence of an RHD pseudogene containing a 37 base pair duplication and a nonsense mutation in africans with the Rh D-negative blood group phenotype. Blood 2000;95(1):12-8. 35. Daniels G, Faas BH, Green C, Smart E, Maaskant-van Wijk PA, Avent ND, Zondervan HA, von dem Borne AE, van der Schoot CE. The VS and V blood group polymorphisms in Africans: a serologic and molecular analysis. Transfusion 1998;38(10):951-8. 36. Chou ST, Westhoff CM. The Rh and RhAG blood group systems. Immunohematology. 2010;26(4):178-86. 37. Shao CP, Maas JH, Su YQ, Kohler M, Legler TJ. Molecular background of Rh D-positive, D-negative, D(el) and weak D phenotypes in Chinese. Vox Sang. 2002;83(2):156-61. 38. Wagner FF, Gassner C, TH, Schonitzer D, Schunter F, Flegel WA. Molecular basis of weak D phenotypes. Blood 1999;93(1):385-93. 39. Okubo Y, Yamaguchi H, Tomita T, Nagao N. A D variant, Del? Transfusion 1984;24(6):542. 40. Mak KH, Yan KF, Cheng SS, Yuen MY. Rh phenotypes of Chinese blood donors in Hong Kong, with special reference to weak D antigens. Transfusion 1993;33(4):348-51. 41. Sun CF, Chou CS, Lai NC, Wang WT. RHD gene polymorphisms among RhD-negative Chinese in Taiwan. Vox Sang. 1998;75(1):52-7. 42. Flegel WA, von Zabern I, Wagner FF. Six years’ experience performing RHD genotyping to confirm D- red blood cell units in Germany for preventing anti-D immunizations. Transfusion 2009;49(3):465-71. 43. Tippett P, Sanger R. Observations on subdivisions of the Rh antigen D. Vox Sang. 1962;7:9-13. 44. Scott M. Section 1A: Rh serology. Coordinator’s report. Transfus.Clin.Biol. 2002;9(1):23-9. 45. Tippett P, Lomas-Francis C, Wallace M. The Rh antigen D: partial D antigens and associated low incidence antigens. Vox Sang. 1996;70(3):123-31. 46. Cartron JP. Defining the Rh blood group antigens. Biochemistry and molecular genetics. Blood Rev. 1994;8(4):199- 212. 47. Mouro I, Le Van Kim C, Rouillac C, van Rhenen DJ, Le Pennec PY, Bailly P, Cartron JP, Colin Y. Rearrangements of the blood group RhD gene associated with the DVI category phenotype. Blood 1994;83(4):1129-35. 48. Westhoff CM. Rh complexities: serology and DNA genotyping. Transfusion 2007;47(1 Suppl):17S-22S. 49. Jones JW, Finning K, Mattock R, Williams M, Voak D, Scott ML, Avent ND. The serological profile and molecular basis of a new partial D phenotype, DHR. Vox Sang. 1997;73(4):252-6. 50. Avent ND, Jones JW, Liu W, Scott ML, Voak D, Flegel WA, Wagner FF, Green C. Molecular basis of the D variant phenotypes DNU and DII allows localization of critical amino acids required for expression of Rh D epitopes epD3, 4 and 9 to the sixth external domain of the Rh D protein. Br.J.Haematol. 1997;97(2):366-71. 51. Redman C, Marsh WL, KA, Avellino GP, Johnson CL. Isolation of Kell-active protein from the red cell membrane. Transfusion 1984;24(2):176-8. 52. Marsh WL, Redman CM. Recent developments in the Kell blood group system. Transfus.Med.Rev. 1987;1(1):4-20.

22 79. 78. 77. 76. 75. 74. 73. 72. 71. 70. 69. 68. 67. 66. 65. 64. 63. 62. 61. 60. 59. 58. 57. 56. 55. 54. 53. 2012;44(2):170-3. blood andspecifies thenew ABCB6 isdispensableforgroup system erythropoiesis Langereis. Nat.Genet. Helias V, SaisonC, BallifBA,Peyrard T, Takahashi J, Takahashi H, Tanaka JC,Puy M,Deybach H,Le M,etal. Gall 2012;44(2):131-2. Zelinski T, Coghlan G,LiuXQ, ME.ABCG2 Reid nullalleles definetheJr(a-)blood group Nat.Genet. phenotype. Junior. 2012;44(2):174-7. Nat.Genet. et al. NullallelesofABCG2 encodingthebreast cancerresistance protein blood definethenew group system Saison C,Helias V, BallifBA,Peyrard T, Puy H,Miyazaki T, Perrot S, Vayssier-Taussat M, Waldner M,Le Pennec PY, hemolytic diseaseofthefetus associated withanti-Jr. andnewborn Transfusion 2008;48(9):1906-11. Peyrard T, Pham L, Fleutiaux BN, Arnaud S, Brossard Y, B, P, Guerin I, Rouger Le Pennec PY. Fatal of theliterature. Transfusion 2004;44(2):197-201. Kwon MY, PA, SuL,Arndt DP. G,Blackall Garratty Clinicalsignificance ofanti-Jra:two casesand of review report population. Prog.23 Stroup M. Jr. M, MacIlroy Five examples of an inantibody thedefining an antigen of highCaucasian frequency M,vander M,Moes van derHart Veer M,vanLoghem JJ. bloodgroup new HoandLan:two 1961. antigens. In of Vel system antibodies. Vox Sang. 1968;15(2):125-32. Adebahr ME,AllenFH,Jr., JK, Issitt PD, R,Reihart Kuhns Oyen WJ. Anti-Vel antibodyshowing 2,anew heterogeneity Vox Sang. 1986;51(2):108-11. DL,Kinney Becton TR. Aninfantgirlwithsevere autoimmune hemolyticanemia:apparent anti-Vel specificity. 1998;75(1):70-1. ofaninvivo detection haemolyticanti-vel by thegeltest. L,ClasenC.Unsatisfactory J, Bartz VoxNeppert Sang. 1961;1:111-5. Levine P, White JA,Stroup negative M.Seven membersinthree generationsofafamily.Ve-a (Vel) Transfusion apropos ofacaseanti-VEL immunization].Bibl.Haematol. 1965;23:309-11. Battaglini PF, RM.[Studyofthe J, C,SalmonNicoli Ranque Bridonneau population intheMarseilles VEL factor EB. bloodfactor: [New Sussman LN,Miller Vel.]. Rev.Hematol. 1952;7(3):368-71. Immunohematology. 1992;8(3):53-7. onandchemicalmodificationsofhigh-frequency redcellantigens. Daniels G.Effectofenzymes Br.J.Haematol. 1986;62(2):301-9. Dunstan RA.Statusofmajorred cellbloodgroup antigensonneutrophils, andmonocytes. lymphocytes HG,Anstee DJ.Klein Mollisons’ Blood Transfusion inClinicalMedicine. 11ed. Oxford: Publishing; Blackwell 2013. bloedproducten/sanquin-bloedbank-ingevroren-bloed/. Sanquin BloedbankIngevroren Bloed2013Apr26Available from http://www.sanquin.nl/producten-diensten/ Immunohematology. 2011;27(4):131-5. ME. Reid ofhigh-incidencebloodgroup oflow-incidenceThe and901series ISBT 700series antigens. bloodgroupLee basisofKell S.Molecular phenotypes. Vox Sang. 1997;73(1):1-11. human KELgeneabolishestheexpressionbloodgroup ofKell antigens. J.Biol.Chem. 2001;276(13):10247-52. Yu LC, Twu YC, ChangCY, amutationatthesplicesite of phenotype: basisoftheKell-null LinM.Molecular J.Biol.Chem. nullphenotype. theKell defects underlying 2001;276(29):27281-9. Lee AP, S,RussoDC,Reiner Lee JH,SyMY, Telen MJ, Judd WJ, SimonP, MJ, Rodrigues Chabert T, etal. Molecular K-,phenytype k-, Kp(a-b-). Vox Sang. 1961;6:620-3. HC. Chown HR,Soltain H,Nevanlinna B, Lewis M,Kaita The pedigrees peoplealready asof oftwo reported 1959;183(4675):1586. bloodgroup H,Lewis Nature K-,k-,Kp(a-b-). exampleoftheKell M,Chown phenotype B,Kaita Gard E.Afurther Nature phenotype. blood-group 1957;180(4588):711. Chown Kell H.Anew B, Lewis M,Kaita KEL3, KEL4,andKEL21alleles, andtheKEL17KEL11alleles. Transfusion 1996;36(6):490-4. Lee S, Wu S,NaimeD, X,Son M,Okubo Reid Y, Sistonen P, C.Point Redman KEL10,the mutationscharacterize Lee S, Wu M,Zelinski X,Reid T, Blood1995;85(4):912-6. (K1)phenotype. basisoftheKell C.Molecular Redman Proc.Natl.Acad.Sci.U.S.A 1991;88(14):6353-7. Lee bloodgroup S,ZambasED, structureofKell protein. Marsh cloningandprimary C.Molecular WL, Redman rd Ann.Mtg.Am.Ass.Blood Banks,86.1970. General introduction General 23

Chapter 1 Chapter 1

80. Hue-Roye K, Lomas-Francis C, Coghlan G, Zelinski T, Reid ME. The JR blood group system (ISBT 032): molecular characterization of three new null alleles. Transfusion 2012. 81. Hue-Roye K, Zelinski T, Cobaugh A, Lomas-Francis C, Miyazaki T, Tani Y, Westhoff CM, Reid ME. The JR blood group system: identification of alleles that alter expression. Transfusion 2013. 82. Saison C, Helias V, Peyrard T, Merad L, Cartron JP, Arnaud L. The ABCB6 mutation p.Arg192Trp is a recessive mutation causing the Lan- . Vox Sang. 2013;104(2):159-65. 83. Krishnamurthy P, Schuetz JD. The role of ABCG2 and ABCB6 in porphyrin metabolism and cell survival. Curr. Pharm.Biotechnol. 2011;12(4):647-55. 84. Schonewille H, van de Watering LM, Loomans DS, Brand A. Red blood cell alloantibodies after transfusion: factors influencing incidence and specificity. Transfusion 2006;46(2):250-6. 85. Bauer MP, Wiersum-Osselton J, Schipperus M, Vandenbroucke JP, Briet E. Clinical predictors of alloimmunization after red blood cell transfusion. Transfusion 2007;47(11):2066-71. 86. Moise KJ. Red blood cell alloimmunization in . Semin.Hematol. 2005;42(3):169-78. 87. Crosby WH. Normal functions of the spleen relative to red blood cells: a review. Blood 1959;14(4):399-408. 88. Callender ST, Powell EO, Itts LJ. The life span of red cell in man. J.Pathol.Bacteriol. 1945;57:129. 89. Schonewille H. Review of the literature on red cell alloimmunization. In Red Blood Cell Alloimmunization after Blood Transfusion. Leiden: Leiden University Press; 2008. p. 17-27. 90. Pieters J. MHC class II-restricted antigen processing and presentation. Adv.Immunol. 2000;75:159-208. 91. Watts C. Capture and processing of exogenous antigens for presentation on MHC molecules. Annu.Rev. Immunol. 1997;15:821-50. 92. Urbaniak SJ. to human red blood cell antigens. Vox Sang. 2002;83 Suppl 1:293-7. 93. Daniels G, Poole J, de Silva M, Callaghan T, MacLennan S, Smith N. The clinical significance of blood group antibodies. Transfus.Med. 2002;12(5):287-95. 94. Poole J, Daniels G. Blood group antibodies and their significance in transfusion medicine. Transfus.Med.Rev. 2007;21(1):58-71. 95. Transfusie- en Transplantatiereacties in Patiënten (TRIP) 2013 Apr 30 Available from http://www.tripnet.nl/ pages/nl/. 96. Serious Hazards of Transfusion (SHOT) 2013 Apr 30 Available from http://www.shotuk.org. 97. Capon SM, Goldfinger D. Acute hemolytic transfusion reaction, a paradigm of the systemic inflammatory response: new insights into pathophysiology and treatment. Transfusion 1995;35(6):513-20. 98. Strobel E. Hemolytic Transfusion Reactions. Transfus.Med.Hemother. 2008;35(5):346-53. 99. Kumar S, Bandyopadhyay U. Free heme toxicity and its detoxification systems in human. Toxicol.Lett. 2005;157(3):175-88. 100. Moise KJ. Fetal anemia due to non-Rhesus-D red-cell alloimmunization. Semin.Fetal Neonatal Med. 2008;13(4):207-14. 101. Urbaniak SJ, Greiss MA. RhD haemolytic disease of the fetus and the newborn. Blood Rev. 2000;14(1):44-61. 102. Chown B. Anaemia from bleeding of the fetus into the mother’s circulation. Lancet 1954;266(6824):1213-5. 103. Mollison PL, Cutbush M. Haemolytic disease of the newborn; criteria of severity. Br.Med.J. 1949;1(4594):123-30. 104. Schwartz HP, Haberman BE, Ruddy RM. Hyperbilirubinemia: current guidelines and emerging therapies. Pediatr. Emerg.Care 2011;27(9):884-9. 105. Davidson LT, Merritt KT, Weech AA. Hyperbilirubinemia in the newborn. Am.J.Dis.Child 1941;61:958. 106. Mollison PL, Cutbush M. A method of measuring the severity of a series of cases of hemolytic disease of the newborn. Blood 1951;6(9):777-88. 107. Crowther C, Middleton P. Anti-D administration after childbirth for preventing Rhesus alloimmunisation. Cochrane.Database.Syst.Rev. 2000;(2):CD000021. 108. Koelewijn JM, de Haas M, Vrijkotte TG, Bonsel GJ, van der Schoot CE. One single dose of 200 microg of antenatal RhIG halves the risk of anti-D immunization and hemolytic disease of the fetus and newborn in the next pregnancy. Transfusion 2008;48(8):1721-9. 109. Westhoff CM. The potential of blood group genotyping for transfusion medicine practice. Immunohematology. 2008;24(4):190-5.

24 124. 123. 122. 121. 120. 119. 118. 117. 116. 115. 114. 113. 112. 111. 110. polymerase chainreactionscreening for blooddonorswithrare phenotypes. Transfusion 2008;48(6):1169-73. Wagner FF, BittnerR,Petershofen priming- A,Muller Doescher EK, sequence-specific TH. Cost-efficient bloodgroup alternative approach. genotyping—the spectrometry-based Transfus.Med.Rev. 2013;27(1):2-9. S,Frey C,Meyer Gassner mass BM, time-of-flight laserdesorption/ionisation, Vollmert C. Matrix-assisted nanofluidic real-time polymerasechainplatform.reaction Transfusion 2010;50(1):40-6. Hopp K, Weber BellissimoD, K, ST, Johnson Pietz B. redusinga High-throughput bloodcellantigengenotyping donors for minorred bloodcellandplatelet antigens. Transfusion 2006;46(5):841-8. A,PhillipsMontpetit MS,MongrainI,Lemieux R,St-Louis molecularprofiling M.High-throughput ofblood TE, etal. The BloodgenProject oftheEuropean Union,2003-2009. Transfus.Med.Hemother. 2009;36(3):162-7. Avent A,Flegel ND, Martinez WA, E,Madgett M,DanielsGL,Muniz-Diaz ML,NoguesN,Pisacka OlssonML,Scott DNA analysis. Transfusion 2007;47(4):736-47. et al. of24minorred Determination bloodcellantigensfor more than2000blooddonorsby high-throughput Hashmi G,Shariff T, Zhang Y, Cristobal Seul M,J, ChauC, Vissavajjhala K, Charles-Pierre P, BaldwinC,Hue-Roye D, 2010;12(4):453-60. group optimization,validation,andoneyear nucleotidepolymorphisms: ofroutine clinicaluse. J.Mol.Diagn. Di CJ, M,Chiaroni Silvy J, BaillyP. SinglePCRmultiplexSNaPshot reactionfor ofeleven blood detection celldiseasepatients. the managementofmultiply-transfusedsickle Transfusion 2002;42(2):232-8. M,BiancoC,PellegrinoCastilho L,Rios J, Jr., FL,SaadST, Alberto Costa FF. ofbloodgroups DNA-basedtyping for cells.2005;45(4):520-6. Transfusion Wagner T, GF, Kormoczi BuchtaC, Vadon M,Lanzer G,Mayr WR, Legler TJ. Anti-Dimmunizationby DELred blood Beattie KM,Sigmundcellpatients. J, KE,McGraw ShurafaM.U-variant bloodinsickle Transfusion 1982;22(3):257. 35. Dahr W. subsystem Miltenberger oftheMNSsbloodgroup andoutlook. system. Review Vox Sang. 1992;62(3):129- 2011;44(1):93-9. ST, GA, Johnson Denomme Pietz B. red of blood donors. Mass-scale cell genotyping Transfus.Apher.Sci. screening. Vox Sang. 2009;97(3):198-206. Veldhuisen B, CE,deHaasM.Bloodgroup vanderSchoot from genotyping: patientto high-throughput donor compatible bloodcomponentsfor alloimmunized patients. Vox Sang. 2009;97(1):61-8. et al. Set-up androutineblooddonorsto facilitate useofadatabase10,555genotyped thescreening of Perreault J, Lavoie J, Painchaud P, Cote M,Constanzo-Yanez J, Cote G,Gendron R,Delage F, DubucS,Caron B, patients whohave recently received atransfusion. Transfusion 2000;40(1):48-53. M,Powell ME,Rios Reid VI, Charles-Pierre D, Malavade V. DNAfrom bloodsamplescanbeusedto genotype General introduction General 25

Chapter 1

Chapter 2

Comprehensive genotyping for 18 blood group systems using a multiplex ligation-dependent probe amplification assay shows a high degree of accuracy

Lonneke Haer-Wigman1 Yanli Ji2 Martin Lodén3 Masja de Haas1 C. Ellen van der Schoot1* Barbera Veldhuisen1*

1 Sanquin Research, Amsterdam and Landsteiner Laboratory, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands 2 Institute of Clinical Blood Transfusion, Guangzhou Blood Center, Guangzhou, China 3 MRC-Holland b.v., Amsterdam, the Netherlands * Equal contribution

Accepted in Transfusion Chapter 2

Abstract Background: In recent years genotyping methods have been implemented in blood banks as alternative to comprehensive serologic typing. We evaluated a newly developed assay for convenient and comprehensive genotyping of blood group alleles based on multiplex ligation-dependent probe amplification [MLPA] technology.

Methods: We analyzed 103 random and 150 selected samples to validate the specificity of the blood-MLPA assay that is able to determine the presence, absence, and copy number of 48 blood group and 112 variant alleles of 18 blood group systems. A total of 4038 serologic typing results, including results of 52 different antigens, were available for these samples.

Results: In 4018 (99.5%) of the 4038 serologic typing results the predicted phenotypes by the blood-MLPA were in concordance with serologic typing. Twenty discordant results were due to false-positive serologic results (n = 2), false-negative serologic results (n = 1), inability of routine serologic typing to detect variant antigens (n = 14), or false-positive prediction from the blood-MLPA due to the presence of a (n = 3).

Conclusion: The blood-MLPA reliably predicts the presence or absence of blood group antigens, including almost all clinically relevant blood group antigens, except ABO, in patients and donors. Furthermore, it is the first assay that determines copy numbers of blood group alleles in the same test. It even provides more detailed and accurate information than serologic typing, because most variant alleles are immediately recognized. Because only standard laboratory equipment is needed, this assay finally offers the possibility to comprehensively type recipients and makes extensive matching for selected patients groups more feasible.

28 auto-antibodies, serologic is severely typing hampered and less reliable. equipment is often required.equipment isoften Via genotyping it is possible to correctly predicttheexpression itispossibletoVia correctly ofbloodgroup genotyping antigens. or pregnancyapersonmay develop alloantibodiesagainst “foreign” bloodgroup antigens. multiple bloodgroup antigensintheirdonorsasanalternative for serologic typing. years several blood banks implemented high-throughput methods to determine genotyping patient isinthesecircumstances helpfulandincreasingly applied. antisera. most bloodgroup antigensserologic relies ofpolyclonal typing ontheavailability human of bloodgroup antigensmonoclonalantibodiesare commercially available, while for status inblooddonorsandrecipients to prevent immunizationand/oranHTR. relevant bloodgroup antigensinonetest. group systems, seven bloodgroup RBCantigenseries. collections,and two More than300bloodgroup andhave antigensare beenclassifiedinto known 33blood Blood group antigensare structuresexpressed membrane. onthered bloodcell[RBC] Introduction relevant bloodgroup antigens. the blood-MLPA assay of all clinically is able to genotyping predict the phenotype correctly are commercially available atthe moment. The aimofthisstudy wasto evaluate whether group allelesthan the highormedium-throughput bloodgroup assays that genotyping 50 targets inasingletube, canbedetermined thatisableto farmore determine blood equipment. Furthermore, theblood-MLPA assay allows ahighlevel ofmultiplexing, upto electrophoresis method,to usegenotyping which requires andcapillary onlyathermocycler probe amplification [MLPA]the multiplexligation-dependent technique. The MLPA isaneasy Therefore, assay basedon genotyping ofanew we developed andanalyzed theperformance expression. Currently, serologic isstillthestandard typing methodto bloodgroup determine antigen still very limited inthecommercially availablestill very bloodgroup assays. genotyping HTR and/orhemolyticdiseaseofthefetus andnewborn. against blood group antigens are clinically relevant, because not all antibodies cause severe which serology atalimited candetect level. the fetus andnewborn. These antibodiescangive hemolytictransfusionreactions[HTRs]and/or disease of GPA-GPB-GPA (e.g., polymorphisms by ABOantigens),or multiple-nucleotide geneticrecombination (e.g., Antigenic differences (e.g., polymorphisms are causedKandkantigens), by single-nucleotide known. For almostall blood group the different antigens the genetic basis that underlies antigens is 16 8,9 Multiple genotyping methods have genotyping Multiple been developed aset that determine ofclinically For someblood group are andMNS)ahighnumberofvariants systemspresent, (Rh 5,6 Serologic is, typing however, notalways possible. hybrid allelesorRHD 3,4 Therefore bloodgroup itiscrucialto thecorrect determine antigen 13,17-20 Furthermore, thenumberofantigensthatcanbetested is deletion causingDnegativity). 8,9 13,17-21 After transfusionorincaseofthepresence of For thesetests investment in dedicated 3

7 11 1-3 Onlyfor alimited number Furthermore, intherecent Uponabloodtransfusion Genotyping ofbloodgroupGenotyping antigens 10 Genotyping of the Genotyping 5 Notallantibodies 13,16,22-24 9,12,13

3

12,14,15 29 2 1

Chapter 2 Chapter 2

Material and methods Blood samples and DNA isolation All cases were included after informed consent was given. Samples from 63 random Dutch donors, 40 random Chinese donors, and 98 Dutch donor or patient cases with a rare blood group phenotype were included. Twelve DNA samples, selected for the presence of a rare blood group phenotype, were provided by C. Westhoff of the New York Blood Center. An additional 40 cases were selected for the presence of a rare blood group allele based on the screening of MLPA results of more than 750 cases. Ethylenediaminetetraacetic acid– anticoagulated blood was collected and genomic DNA was isolated from white blood cells using a DNA extraction kit (QIAamp DNA blood mini kit, Qiagen).

Serologic typing Samples were serologically typed, using standard serologic gel column card or tube agglutination, for multiple blood group antigens: M, N, S, s, U, Mur, MINY, Mia, C, c, E, e, D, Cw, VS, Lua, Lub, K, k, Kpa, Kpb, Jsa, Jsb, Lea, Leb, Fya, Fyb, Jka, Jkb, Dia, Dib, Wra, Yta, Ytb, Sc1, Sc2, Doa, Dob, Joa, Hy, Coa, Cob, LWa, LWb, Ge2, Ge3, Ge4, Cra, Kna, and Inb. Absorption-elution was performed according to the manufacturers’ protocol (Gamma ELU-KIT II, Immucor-Gamma).

MLPA genotyping An MLPA assay specific for the analysis of blood group alleles was developed. A schematic overview of the MLPA reaction is shown in Supplementary Figure S1. The blood-MLPA assay contains 100 probe combinations that are able to determine the presence, absence, and copy number of 48 blood group alleles and 112 variant alleles (Table 1). The probe combinations of the blood-MLPA are divided into three pools, p401, p402, and p403. Because probe combinations targeting antithetical alleles interfere with each other, probe combinations that target antithetical alleles are dived over probe mix p401 and p402. Furthermore, probes combinations that hybridize to nearby regions also interfere with each other. To be able to cover all probe combinations that target variant alleles of the Rh blood group system a third mix, p403, was necessary. Mix p403 also contains probe combinations that detect the presence of variant alleles, resulting in weakened or loss of expression of the MNS, Duffy, Dombrock, and Gerbich blood group systems. The MLPA reaction was performed according to the manufacturer’s protocol (MRC Holland) on a thermocycler (Biometra T1, Westerburg BV; or Veriti, Applied Biosystems). In short, 5 μL containing 50 ng of DNA was denatured and 1.5 μL probe mix and 1.5 μL SALSA MLPA dilution buffer were added. After 16 to 20 hours of hybridization of the probe combinations to genomic DNA at 60°C, 1 μL of SALSA ligase 65, 1.5 μL of SALSA ligase buffer A, and 1.5 μL of SALSA ligase buffer B were added and incubated for 15 minutes at 54°C. A polymerase chain reaction [PCR] was performed on the complete ligation sample by adding 2 μL of universal primers and 0.5 μL of SALSA polymerase. PCR conditions were as follows: 5 minutes at 72°C;

30 Table 1. Performance of the probe combinations of the blood-MLPA tested in 253 cases

Bloodgroup Name of detecting probe Detecting Number of cases Number of cases Present in mix system combination Allele name Antigen positive in blood-MLPA negative in blood-MLPA 002 MNS GYPA*01_59C p401 GYPA*01 M 197 56 GYPA*02_59T p402 GYPA*02 N 187 66 GYPB*03_143T p401 GYPB*03 S 114 139 GYPB*04_143C p402 GYPB*04 s 224 29 GYPB*05_170T p403 GYPB*01N absence of U 250 3 GYP*502, GYP*503 and GYP*504 GYPB_psIII_ivs2 p403 250 3 (Presence of pseudoexon III) GYPB#03N.01/02 p403 GYPB*03N.01 or GYPB*03N.02 U weak 8 245 GYPB#03N.03/04 p402 GYPB*03N.03 or GYPB*03N.04 U weak 0 253 GYP#301/501 p401 GYP*301/GYP*501 Mur 5 248 004 Rh Probe cominations are stated RHD*01, RHD*01N.01 and 94 RH RhD and 94 Rh p401, p402, p403 in previous article variant alleles variants RH02_203G, RHCE02_insC RhC p401, p403 RHCE*C RhC 137 116 RHCE02_307C Rhc p402 RHCE*c Rhc 190 63 RHCE05_676C RhE p401 RHCE*E RhE 76 177 RHCE05_676G Rhe p401 RHCE*e Rhe 225 28 RHCE#05_733G p402 RHCE*01.20 VS 19 234 RHCE#01_122G p402 RHCE*02.08.01 Cw 7 246 005 Lutheran LU*01_230A p401 LU*01 Lua 16 237 LU*02_230G p402 LU*02 Lub 250 3 006 Kell KEL*01_578T p401 KEL*01 K 31 222 KEL*02_578C p402 KEL*02 k 239 14 a KEL*02.03_841T p401 KEL*02.03 Kp 21 232 ofbloodgroupGenotyping antigens KEL*02.-03_841C p402 KEL*02.-03 Kpb 251 2 KEL*02.06_1790C p401 KEL*02.06 Jsa 12 241 KEL*02.-06_1790T p402 KEL*02.-06 Jsb 250 3 007 Lewis LE_59G (wt) p401 FUT3 wt allele 238 15 LE_59T (null) p402 FUT3 silent allele Le(a-b-) 77 176 008 Duffy FY*01_125G p401 FY*01 Fya 147 106 FY*02_125A p402 FY*02 Fyb 180 73 FY#02N.01_-67C p403 FY*01N.01 or FY*02N.01 Fy(a-) or Fy(b-) 44 209 31 FY#02M.01_265T p403 FY02M.01 Fy(b+ weak) 9 244

Chapter 2 Chapter 2 0 4 0 1 3 0 1 0 2 6 2 0 1 33 49 91 110 242 243 253 244 245 248 253 234 251 228 233 244 251 Number of cases in blood-MLPA negative 0 9 8 5 0 2 9 2 11 10 19 25 20 253 143 220 249 253 204 162 252 250 253 252 253 251 247 251 253 252 Number of cases in blood-MLPA positive a b a b a b a b a b a b a b a b a b Antigen Jo(a-) Do Hy- Gy(a-) Yt Co Co LW Wr Jk Wr Yt Ge2 Ge3 Ge4 Cr(a+) Cr(a-) Kn Kn Ok(a+) Ok(a-) Di Sc1 Sc2 Do LW In Di In Jk Detecting Allele name Allele DO*01.-05 DO*02 DO*02.-04 DO*02N.01 YT*01 CO*01 CO*02 LW*07 DI*04 JK*02 DI*03 YT*02 GE*02 GE*03 GE*04 CROM*01 CROM*-01 KN*01 KN*02 OK*01 OK*-01 DI*01 SC*01 SC*02 DO*01 LW*05 IN*01 DI*02 IN*02 JK*01 p403 p402 p403 p402 p401 p401 p402 p402 p402 p402 p401 p402 p403 p403 p403 p401 p402 p401 p402 p401 p402 p402 p401 p402 p401 p401 p401 p402 p402 Present in mix Present p401 DO#01.-05_350T DO*02_793G DO#02.-04_323T DO#02N.01 YT*01_1057C CO*01_134C CO*02_134T LW*07_299G DI*04_1972G JK*02_838A DI*03_1972A YT*02_1057A GE*02_67G GE*03_131G GE*04_243G CROM*01_679G CROM*01-_697C KN*01_4681G KN*02_4681A OK*01_274G OK*-01_274A DI*01_2561T SC*01_169G SC*02_169A DO*01_793A LW*05_299A IN*01_137C DI*02_2561C IN*02_137G Name of detecting probe combination JK*01_839G 011 Yt 015 Colton 020 Gerbich 021 Cromer 022 Knops 024 Ok 010 Diego 013 Scianna 014 Dombrock 016 LW 023 Indian Bloodgroup Bloodgroup system 009 Kidd Table 1. (Continued) Table

32 15 minutes at 95°C, 40 cycles 15 of minutes30 seconds at 95°C, 30 seconds at at 60°C, 95°C, and 40 80 cycles seconds at and 3.4μLofIDCore Primer AorIDCore Set Primer B. Set PCRconditions were asfollows: 12.5 μLcontaining25to IDPCRmix,0.4μLofbiotindCTP, 200ngofDNA 6.3μLofMaster on a PCR systemwere (GeneAmp 9700, Applied Biosystems) in a totalperformed volume of targeted regions ofgenomicDNAweretwo multiplexPCR.BothPCRprocedures amplifiedin accordingThe IDCore+ to themanufacturer’s reactionwasperformed protocol. short, In analysis of33bloodgroup allelesofninebloodgroup allelesandeleven variant systems. ID Core+ IDCore+ (Progenika Biopharma). isaLuminex-based assay thatisspecific for the serologic data,122DNAsampleswere platform, alsoanalyzed the withanothergenotyping resultsBecause theobtainedgenotyping were more comprehensive thantheavailable ID Core+ genotyping single runwhenallthree blood-MLPA poolsare tested. on time. Per 1and30samplescanbeanalyzed between simultaneouslyina thermocycler hybridization, amaximumof5hoursruntimeongeneticanalyzer and2hoursofhands- The complete blood-MLPA approximately assay takes 25hours, including16hoursof FigureSupplementary S2. andbloodgroupSoftware predictionoutcome oftheblood-MLPA Excel fileare shown in can betransferred information system. Datadisplay into oftheGenemarker alaboratory phenotypes. Subsequently, conclusions theExcelandphenotype filecontaininggenotype applicationforsoftware automatic conversion ofraw blood-MLPA and datainto geno- in-house use. MRC-Holland, thecommercial supplieroftheblood-MLPA, isdeveloping a independent ofthenumbersamples. The Excel isdeveloped by matrix theauthorsfor withinminutes,(Excel, andphenotype whichtranslates theseratiosinto ageno- Microsoft), and thecontrol sample. These ratiosare transferred into theblood-MLPA computer matrix (Genemarker, Version to thetested ratiosbetween determine 1.85,Softgenetics) samples For dataanalysis, into raw computer dataofthegeneticanalyzer software isimported used inparallelwiththetested thesameprotocol. samplesusingexactly the bloodgroup allelesinthetested DNAsamples. Eachcontrol sampleisprepared to be probe target (see Web Resources) and are used as a reference to determine copy number of covered in the cell line. The control samples contain an established number of copies of each of humancelllineandplasmidDNA,containingtargets for probe combinationsthatare not previously. of theMLPA asdescribed reaction,whichhasbeenreplaced by MRC-Holland, wasperformed on a geneticanalyzer3130, Applied Biosystems). (Model For some samples the old protocol Biosystems), and0.5μLofSize Standard 500-Liz,AppliedBiosystems) (GeneScan wasanalyzed minutes of1.5μLMLPA at 72°C.Amixture sample, Applied 8.5μLofformamide (Hi-Di, of 30 seconds at 95°C, 30 seconds at 60°C, and 1 minute35 cycles at 72°C; followed by 20 17 Eachblood-MLPA control samplethatconsistsofamixture poolhasanartificial Genotyping ofbloodgroupGenotyping antigens 33

Chapter 2 Chapter 2

72°C followed by 7 minutes at 72°C. The PCR product is hybridized to Luminex beads using 5 μL of PCR product, 60 μL of hybridization ID buffer, 29 μL of control ID buffer, and 6.0 μL of ID Core A Beads or ID Core B beads. Hybridization conditions were as follows: 5 minutes at 95°C followed by 15 minutes at 52°C. Samples were labeled by adding 0.7 μL of SAPE and 74.3 μL of SAPE dilution buffer to the hybridization mix and incubation for 15 minutes. The samples were analyzed on an analyzer using the accompanying software (Luminex 100 and IS2.3WB, respectively, Luminex Corp.). Data analysis was performed using the PGKWS software ID Core+ analysis mode.

Allelic discrimination Assays detecting the FY*01 and FY*02 or FY*01N.01 and FY*02N.01 alleles were performed with TaqMan technology. The assays were performed on a real-time PCR system (StepOnePlus, Applied Biosystems) in a total volume of 25 μL, containing 2.5 to 250 ng of DNA, 12.5 μL of 2× Taqman Universal PCR Mastermix (Applied Biosystems), 0.25 μM forward and reverse primers, and 0.125 μM probes. PCR conditions were as follows: 10 minutes at 95°C, 50 cycles of 10 seconds at 95°C, and 1 minute at 60°C; at the end of each cycle the fluorescence signal was measured. Data analysis was performed using the StepOnePlus software allelic discrimination mode.

PCR and sequencing Exon-specific sequencing was performed using specific primers flanking all exons of AQP1, GYPA, GYPB, LU, KEL, EKLF, and RHD. The PCR was performed on a thermocycler (Veriti, Applied Biosystems) in a total volume of 20 μL, containing 50 to 150 ng DNA, 10 μL of 2× GeneAmp Fast PCR Master Mix (Applied Biosystems), 0.5 μM forward and reverse primer. PCR conditions were as follows: 10 seconds at 95°C, 35 cycles of 10 seconds at 95°C, and a specific annealing- elongation temperature and time for each primer set (ranging from 60°C to 70°C), followed by 1 minute at 72°C. PCR products were sequenced with a cycle sequencing kit on a sequencer (ABI BigDye Terminator v3.1 kit and ABI 3130XL, respectively, Applied Biosystems).

Results Performance of the blood-MLPA probes with plasmid DNA The blood-MLPA contains 66 probe combinations that are designed to detect alleles coding for 48 blood group antigens of 17 blood group systems (Table 1). The blood-MLPA contains another 34 probe combinations that are designed to detect the presence of 112 alleles that cause weak, partial, or absent expression of a blood group antigen (Table 1). The performance of 25 of these 34 probe combinations which are able to detect 79 RHD variant and 17 RHCE variant alleles have been described previously.17 The remaining nine probe combinations detect variant alleles of the MNS, Lewis, Duffy, and Dombrock blood group systems (Table 1).

34 (Table 1). In cases heterozygous for cases heterozygous (Table 1). In Wr samples, because DNA samples negative for these blood group antigens were not available previously validated theblood-MLPA for alleles. ofRHDvariant thedetection blood-MLPA detected anRHD blood-MLPA astheprimary the samegenotype ninecasesinwhichthe result 3).In (Table (Table 3). sequencing, was performed testThis determined in all second cases genotyping method,DNA sampleandasecondgenotyping IDCore+, and/or allelicdiscrimination, ninecasestheblood-MLPAcases for 3).In antigens(Table two wasrepeated usingthesame bydetermined serologic were typing casesfor detected insixteen oneantigenandintwo predictedDiscrepancies by thephenotype theblood-MLPA between andthephenotype Analysis ofdiscrepant samples based onadifferent(Table approach, were concordant 2). bydetermined theblood-MLPA platform theIDCore+, andasecondgenotyping whichis by aswasdetermined serologicthe samephenotype typing. All1717(100%)genotypes 2). For 4018of the4038(99.5%)serologic theblood-MLPA available phenotypes predicted of theblood-MLPA wascompared by determined serologic to thephenotype (Table typing MLPA test results,all100probe combinationsinto account. taking The predicted phenotype wasdesigned tofromAn Excel thepredicted determine phenotype theblood- workbook of theseprobe combinations. for theseantigens,was halfthatofsampleshomozygous whichdemonstrates thespecificity DI*04 OK*-01 for theprobe combinationsspecific rarefor theextremely GYPB*03N.03/04 , DO*02N.01and positivetwo andonenegative 1).Nopositive sample(Table DNAsampleswere available ofprobe combinations(n=67)wastested themajority withatleast samples; furthermore, alleles)probe variant combinationsonatleastthree Rh detect positive andthree negative were ableto test 61ofthe75(allprobe combinationsminustheprobe combinationsthat sample”) orwithoutthespecifictargets (“negative sample”) oftheprobe combinations. We results,or genotyping to obtaingenomicDNAsampleswiththespecifictargets (“positive samples. basedonpreviously Asubsetofthesesampleswasselected obtainedserologic oftheblood-MLPAThe specificity assay wasdemonstrated usingasetof253genomic DNA Performance ofblood-MLPA assay products(data notshown).the expected plasmid thatcontainedthegenomicsequenceofallprobes, asatemplate. Allprobes yielded First, ofallprobes oftheblood-MLPA theperformance wasanalyzed usingaclonedDNA (Wr null alleles 1). (Table Inversely, the probe combinations specific for the high-frequency b ), LW*05 (LW a ), GE*02, variant allele,variant nogeneticfollow-up becausewe wasperformed, GE*04, andOK*01 b (n = 9) or LW a (n = 8) the obtained fluorescence signal (Ok a ) alleleswere detected inallDNA Genotyping ofbloodgroupGenotyping antigens 17 35

Chapter 2 Chapter 2 U- Mur-

RhD- RhD+ (7) or RhD- (3)

M+N+ Serological typing Serological

Lu(a-b-)

0 0 1 0 0 0 1 0 0 0 0 0 2 0 0 0 0 0 0 0 1 0 0 0 1 10 Number with serology discordant

6 5 2 3 1 2 6 3 2 9 26 48 22 69 99 47 42 20 63 75 54 163 119 168 105 158 Number typed with serology x x x x x x x x 2 1 1 0 7 5 4 82 69 33 14 37 33 26 47 42 62 34 120 122 ID Core+ 6 2 3 1 0 4 7 3 27 49 84 23 45 23 56 63 83 19 66 13 248 176 140 185 106 131 253 237 Blood-MLPA + w C S-s- U+ U weak U- Gp.Dane GP.Hop/GP.Bun/GP.HF E-e+ S-s+ RhD+ GP.Mur E+e- E+e+ S+s+ S+s- RhD- RhD variant C-c+ M-N+ C+c- C+c+ Lu(a+b-) VS+ M+N+ M+N- Predicted phenotype Predicted Lu(a+b+) Lu(a-b+) 004 Rh 005 Lutheran Total cases tested Total 002 MNS Bloodgroup system Bloodgroup Performance of Blood-MLPA compared with the IDCore+ genotypingon 253 samples method and serology with the IDCore+ compared of Blood-MLPA 2. Performance Table

36 Table 2. (Continued) Number typed Number with Bloodgroup system Predicted phenotype Blood-MLPA ID Core+ Serological typing with serology discordant serology 006 Kell K+k- 14 0 14 0 K+k+ 17 10 16 2 K+k- K-k+ 222 112 196 0 Kp(a+b-) 2 0 2 0 Kp(a+b+) 19 3 17 0 Kp(a-b+) 232 119 177 0 Js(a+b-) 3 0 3 0 Js(a+b+) 9 3 4 0 Js(a-b+) 241 119 20 0 007 Lewis Le(a-b-) 15 x 3 0 008 Duffy Fy(a+b-) 80 49 70 0 Fy(a+b+) 65 31 60 0 Fy(a+b+ weak) 1 0 1 0 Fy(a-b+) 75 33 61 1 Fy(a-b-) Fy(a-b+ weak) 3 2 3 0 Fy(a-b-) 28 7 20 0 009 Kidd Jk(a+b-) 91 38 78 0 Jk(a+b+) 113 59 99 0 Jk(a-b+) 49 25 44 0 010 Diego Di(a+b-) 1 0 1 0 Di(a+b+) 8 7 2 0 Di(a-b+) 244 115 20 0

Wr(a+b-) 0 x ofbloodgroupGenotyping antigens Wr(a+b+) 9 x 8 0 Wr(a-b+) 244 x 157 0 011 Yt Yt(a+b-) 234 112 5 0 Yt(a+b+) 16 10 4 0 Yt(a-b+) 3 0 3 0 013 Scianna Sc:1,-2 251 x 14 0 Sc:1,2 1 x 1 0

37 Sc:-1,2 1 x 1 0

Chapter 2 Chapter 2

Co(a+b+)

Serological typing Serological

0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Number with serology discordant

4 0 0 4 0 2 3 7 5 1 5 1 5 2 5 2 6 2 5 55 10 10 Number typed with serology x x x x x x x x x x x x x x x x x 1 0 0 0 0 2 0 11 58 12 50 110 ID Core+ 4 0 3 3 8 0 4 3 5 0 1 0 2 6 2 0 0 21 99 31 14 245 228 252 107 251 233 251 253 Blood-MLPA Co(a-b+) LW(a+b-) Gy(a-) Co(a+b-) Heterozygous Hy(-) Heterozygous Gy(a-) Co(a+b+) LW(a+b+) LW(a-b+) Jo(a-) Hy(-) Heterozygous Jo(a-) Ge:2,3,4 Do(a-b+) Ge:-2,3,4 Ge:-2,-3,-4 Ge:-2,-3,4 Do(a+b+) Do(a+b-) Cr(a+) Predicted phenotype Predicted Cr(a-) Kn(a+b-) Kn(a+b+) Kn(a-b+) In(a+b+) In(a+b-) In(a-b+) Ok(a+) Ok(a-) Continued) 016 LW 015 Colton 020 Gerbich 014 Dombrock 021 Cromer Bloodgroup system Bloodgroup 022 Knops 023 Indian 024 Ok Table 2. ( Table

38 Table 3. Discrepancies between predicted phenotype by the Blood-MLPA and phenotype determined by serology

predicted phenotype via genotyping method† Phenotype determined by serology Blood Second genotyping Absorbtion- Phenotype Blood-MLPA Routine serology UCN group method‡ Elution Sequence analysis was correctly system Primary Repeated determined by Primary result Repeated result result result 212 MNS M+N- M+N- M+N+ M+N+ M+N- blood-MLPA 235 Duffy Fy(a-b+) [1x FY*02.N01] Fy(a-b+) [1x FY*02.N01] Fy(a-b+) [1x FY*02.N01] Fy(a-b-) Fy(a-b+) blood-MLPA No mutationst in 70 Colton Co(a+b-) Co(a+b-) Co(a+b-) Co(a+b+) blood-MLPA§ AQP1 73 Rh Weak D type 2 [RHD*01W.02] RhD- RhD- Del blood-MLPA 243 Rh Del [RHD*01EL.01] RhD- RhD- Del blood-MLPA 245 Rh Partial D [RHD*11] RhD- RhD- Del blood-MLPA 65 MNS U weak U weak U weak U- blood-MLPA DAU-0, DAU-1, DAU-2, DAU- 65 Rh RhD+ blood-MLPA 3, DAU-6 of DAU-7 [RHD*10] 75 Rh Partial D [RHD*10.05] RhD+ RhD+ blood-MLPA 104 Rh Partial D [RHD*10.05] RhD+ RhD+ blood-MLPA 233 Rh Partial D [RHD*10.05] RhD+ RhD+ blood-MLPA 87 Rh Weak D [RHD*01W.01] RhD+ RhD+ blood-MLPA 240 Rh Partial D DVII [RHD*07.01] RhD+ RhD+ blood-MLPA 12 Rh Partial D DVII [RHD*07.01] RhD+ RhD+ blood-MLPA MUR-, Mi(a- New variant GYPB 235 MNS Mur+ [GYP*301] Mur+ [GYP*301] Mur+ MUR- Serology ), MINY- allele 205 Kell K+k+ K+k+ K+k+ K+k- K+k- k- KEL*02N.06 Serology Genotyping ofbloodgroupGenotyping antigens 218 Kell K+k+ K+k+ K+k+ K+k- K+k- k- KEL*02N.17 Serology New In(Lu) allele 141 Lutheran Lu(a-b+) Lu(a-b+) Lu(a-b+) Lu(a-b-) Lu(a-b-) Lu(b) weak blood-MLPA EKLF*525dup519_525 New variant RHD* 239 Rh RhD+ RhD+ RhD+ RhD- RhD- Del blood-MLPA 712C>A 244 Rh RhD+ RhD+ RhD+ RhD- RhD- Del RHD*01W.38 blood-MLPA††

† In the case with a RHD, FY or GYPB variant allele, the detected variant allele is given between brackets [ ] ‡ The second genotyping method was; IDCore+, allelelic discrimination or sequencing § We hypothesize that incorrect serological typing is the cause of the discrepancy; no red blood cells were available to confirm this 39 †† The positive phenotype determined by the blood-MLPA is more correct to the actual weak phenotype compared to the negative phenotype determined by serology

Chapter 2 Chapter 2

Serologic typing was repeated in sixteen of the eighteen cases (Table 3). In two cases the repeated serologic results were discrepant compared to the first registered serologic result, but in agreement with the blood-MLPA: a false-positive serologic result for the N antigen (Unique Case Number 212 [UCN212]) and a false-negative result for the Fyb antigen (UCN235), respectively. For UCN70 unfortunately no RBCs were available to repeat serologic typing. In this case we hypothesize that the discrepancy is caused due to a false-positive Cob serologic typing result. Because the blood-MLPA determined a copy number of 2 for the antithetic CO*01 allele (Coa), indicating that the MLPA detected both alleles and sequencing of the coding sequence of AQP1 (coding for the Colton blood group system) confirmed a homozygous CO*01 allele without mutations (Table 3). In eleven cases the blood-MLPA determined the presence of a blood group allele causing partial or weak antigen expression, while routine serologic typing determined the absence (n = 4) or normal presence (n = 7) of the antigen (Table 3). In three cases in which routine serology detected the absence of an antigen, the sensitive absorption-elution technique determined a weak expression concordant with blood-MLPA results. In UCN65 the presence of the GYPB*03N.03 allele was confirmed by the ID Core+ genotyping method (Table 3). Because this allele gives a very low expression of the U antigen routine serologic methods might failed to detect the low expression. Unfortunately, no RBCs were available to confirm this by absorption-elution. The seven RhD variants predicted by the blood-MLPA (one weak D Type 1, four DAU variants, and two DVII variants) with normal RhD expression in routine RhD serologic typing can only be detected when a more extensive panel of anti-D is used for typing. For the remaining six antigens (1x Mur, 1x Lub, 2x k, and 2x RhD) the blood-MLPA and the second genotyping method predicted the presence of the antigen, while repeated serologic typing determined the absence of these antigens (Table 3).We hypothesized that the discrepancy for these antigens was due to the presence of a variant allele for which no detecting probes are present in the blood-MLPA. Indeed, in all cases a variant allele was identified by sequencing: a novel GYPB allele (GYPB*[136+731G>A; 136+773C>G]), a novel In(Lu) variant allele (EKLF*525dup519_525), the KEL*02N.06 variant allele (KEL*223+1G>A),25 the KEL*02N.17 variant allele (KEL*1546C>T),25 the RHD*01W.38 variant allele (RHD*833G>A),5 and a novel RhD variant allele (RHD*721C>A; Table 3). For UCN41 with the EKLF*525dup519_525 allele and UCN293 with the novel RHD*721C>A allele a weak expression of Lub and RhD, respectively, was determined via absorption-elution. The mutations of the novel GYPB*[136+731G>A; 136+773C>G] allele, located in pseudoexon III of GYPB, create the nucleotide sequence encoding the Mur antigen. The nucleotide at the splice site of the pseudoexon III is, however, not mutated (c.136+803T in this new GYPB variant allele). During splicing the pseudoexon III is removed, causing the absence of the Mur antigen in this new variant allele. In the GYP*501 variant allele the splice site of pseudoexon III is mutated (c.136+803T>G), during splicing the pseudoexon III is implemented into the mRNA, causing the expression of the Mur antigen.

40 hybrid allelesabolishtheexpression oftheSors. Dueto thepresence oftheGYPB*03N.01 Table 4.GYPA by allelesdetermined theblood-MLPA andGYPB assay cases thehybrid to allelewaspresent theGYPB*03N.01 next any GYP of theprobe of theknown combinationpatterns in Table GYP two 4.In forcombination patterns theGYP once by serology. far, So agreat ofGYP diversity detected, alleleontheantigenexpression theeffect atleast ofthevariant wasconfirmed DO Dombrock bloodgroup systems. this study, In atotal of 20GYP The blood-MLPA Duffy, allelesoftheMNS,Rh, isableto variant determine Lewis, and ofvariantDetection allelesby the blood-MLPA to callthisallele † When “1” probe two isindicated underneath combinations, probe oneofthetwo combinationsshouldbepresent GYPA-01 Allele GYPA*02 GYPB*03 GYPB*04 GYPA*01N GYPB*01N GYPB*03N.01/GYPB*03N.02 GYPB*03N.03/GYPB*03N.04 GYP*301.01/GYP*301.2 GYP*401 GYP*501 GYP*502 GYP*503 GYP*504 GYP*new allele1 GYP*new allele2 variant alleles werevariant determined by the blood-MLPA assay. For alleles that were all variant variant alleles of two cases, allelesoftwo variant the probe didnot fit combinationpattern M+ Phenotype N+ S+ s+ M- N-S+/s+U+ M- M+/N+ S-s-U- M+/N+ S-Uweak M+/N+ S-Uweak M+/N+ S+/s+Dane+Mur+ M+/N+ S-s-St(a+) MUT+ MINY+ Hil+ Mur+ M+/N+ s+Mi(a+) MINY+ MUT+ Hop+ Mur+ M+/N+ S+Mi(a+) TSEN+ MUT+ MINY+ Hil+ Mur+ M+/N+ s+Mi(a+) M+/N+ s+ Mi(a+) Hil+ MINY+MUT+ Hil+ M+/N+ s+Mi(a+) S- s-Uweak orUnegative S- s-Uweak orUnegative alleles thatcanbedetected by theblood-MLPA are shown variant alleleshasbeenidentified.variant The probe specific variant allele† Copy numberofprobe combination in 1 0 0 GYPA*01_59C GYPA 1 1 1 1 1 1 1 1 1 1 1 1 1 variant alleles (Table 4). In these 4).In alleles(Table variant or 0 1 0 GYPA*02_59T GYPB*03N.02 , 74RH 77 LE 1 0 0 1 1 0 0 1 0 0 0 0 GYPB*03_143T 1 1 1 1 Genotyping ofbloodgroupGenotyping antigens 0 1 0 0 0 0 1 0 1 1 0 0 GYPB*04_143C 1 1 1 1 1 0 1 1 1 0 0 0 0 0 1 0

GYPB_psIII_ivs2 GYPB 1 1 1 1 1 0 1 1 1 0 1 1 1 1 1 1 allele. Both new GYPB*05_170T 52 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0

FY, and21 GYPB*03N.01/02 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 GYPB*03N.03/04 GYP and 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0

41 GYP#301/501

Chapter 2 Chapter 2

GYPB*03N.02 alleles in these cases, we were unable to determine whether the hybrid alleles cause either a weak expression or complete absence of the U antigen.

Copy number variation in blood group alleles The blood-MLPA assay is able to determine whether a blood group allele has a copy number of 0, 1, 2, or more than 2. All genes encoding blood group antigens, except for the RhD and MNS variant antigens, showed a copy number of 2 (data not shown).

Discussion In this study we showed that the blood-MLPA assay is a robust and easy method to correctly predict the phenotype and copy number for clinically relevant blood group antigens, except ABO antigens. Of all presently available genotyping assays the blood-MLPA can determine by far the largest set of blood group antigens in one test.13,17,18,20,21 The intended use for the blood-MLPA is a single test for a more extensive typing of a patient than can be achieved by serology. It can also replace serology if serologic typing is impossible (e.g., after transfusion). The blood-MLPA consists of 100 probe combinations that detect 48 alleles encoding blood group antigens and 112 variant alleles of 18 blood group systems. Twenty-five probe combinations of the blood-MLPA that detect 96 RH variant alleles were validated in a previous article.17 Because the blood-MLPA assay is a quantitative assay, it also determines copy number variation. We report for the first time that, except for the Rh and MNS blood group systems in which it is known that copy number variation frequently occurs, copy number variation is not detected in any of the remaining blood group systems. The specificity of the 75 non–RH-variant probe combinations was shown with a set of carefully selected samples that were already known to be positive or negative for the tested antigens. Only for three rare genotypes (GYPB*03N.03/04, DO*02N.01, and OK*-01) we were not able to obtain a positive sample and for five high frequency antigens (DI*04, LW*05, GE*02, GE*04, and OK*01) no negative samples were available. All 1717 blood group alleles determined by both the blood-MLPA and the ID Core+ genotyping method were concordant. Furthermore, 4018 antigens (99.5%) of the 4038 blood group antigens determined by both the blood-MLPA and serologic typing were concordant. Sequencing of the relevant regions in these discrepant cases revealed that none of the discrepancies was due to failure of a blood-MLPA probe combination, showing a technical specificity of the blood-MLPA assay of 100%. For 14 antigens of 13 cases the discrepancy was due to (incorrect) results recorded upon serologic analysis: false-negative serologic results (n = 1), false-positive serologic results (n = 2), or the inability of routine serologic typing to detect variant antigen expression (n = 11). For another three antigens of three cases the blood-MLPA predicted a normal expression, while serologic typing showed the absence of the antigen. Sequencing demonstrated the presence of variant alleles for the Lutheran or Rh blood group system. For these three cases, routine serology missed the

42 and sequencing equipment. It is the mostcomprehensive blood groupand sequencingequipment. It assay genotyping blood-MLPA inmostlaboratories, becauseitonly requires canbeperformed athermocycler antigens and the amount of samples that can be typed per run still varies hugely. perrunstillvaries antigens andtheamountofsamples thatcanbetyped of blood group genotyping. Furthermore, forthese assays can type a limited number of implementation of patients, costswithholdslarge-scale butalsoinbloodbanks, start-up are capable to induce a primary alloimmunization. are capableto induceaprimary should not patients receivewho are D– RBCsexpression,with weak RhD because these RBCs perform theassay andare thereforeperform expensive to implement. commercially available bloodgroup methodsall require genotyping dedicated equipmentto thepastyears multiplebloodgroupOver techniques have genotyping beendeveloped. The analyses. on ageneticanalyzerandtheMLPA isrelatively short other inbetween canbeperformed coststo implementtheblood-MLPAstart-up assay are low, becausetheruntimeofan MLPA costs are relatively low. Furthermore, whenageneticanalyzer is already present atalaboratory, costs related to coupling ofoligonucleotidesto asolidphasesuchasbeadsorchips, theassay Because, incontrastto mostotherbloodgroupenzymes. assays there genotyping are no blood-MLPA isthatthereagents costsconsistsonlyofrelated to oligonucleotidesand with weak RhD with or weakLu RhD concordant withtheweak expression found uponfollow-up serologic investigation. Persons expression whereas oftheblood-MLPA thepredicted phenotype “antigen positive” ismore to apatientwithanti-Uantibodydelayed HTRmay occur. blood. Furthermore, whenU-weak RBCs, serologically incorrectly asU-,are typed transfused it ispossibleto transfusethemore rare commonU-weak U- donorbloodinstead ofthevery asU-negative [U-]whilethe blood-MLPAtyped predicted theU-weak correctly phenotype, the expression antigens. ofvariant For instance, inthecasethatwasserologically incorrectly Dombrock bloodgroup systems by theblood-MLPA. isclinicallyrelevant detect to correctly It isthedirectrecognition Duffy,typing ofthemostprevalent allelesoftheMNS,Rh, variant and can containupto 150probe combinations. Anadvantageoftheblood-MLPA over serologic of 131 probe combinations (including control probe combinations) and three MLPA mixes additional probe combinationsto theblood-MLPA, becausethecurrent MLPA assay consist (e.g., the Jra,Lan,and/or Vel antigens)are addedto theblood-MLPA. ispossibleto add It combinations thattarget antigensofwhichthegeneticbasiswasrecently discovered target nullalleles(e.g., frequently occurring for bloodgroup theKell system) andprobe of the blood-MLPA increased when probe could even be further combinations that accurate as the determination of blood group antigens by serologic typing. The accuracy The predictionofbloodgroup antigensby theblood-MLPA isasaccurate oreven more for theblood-MLPA assay to predictthepresence orabsenceofbloodgroup antigens. probe of99.9%wasobtained has nodetecting 3).Basedonthese results (Table anaccuracy to the presence of anullantigen) allele for(2x kantigens and 1x Mur which the blood-MLPA b expression can receive or Lu(b+) RhD+ blood, respectively. 18,19 The finalthree discrepancies were due 20 Anadditionaladvantageofthe 13,21-24 Especiallyfor genotyping Genotyping ofbloodgroupGenotyping antigens 13,17,18,20,21 6 Moreover The 43

Chapter 2 Chapter 2 available at the moment and the only assay that reliably can determine the RhD phenotype and gene copy number. The blood-MLPA is a medium-throughput genotyping method: 30 samples can be run in approximately 25 hours including 2 hours of hands-on time. Because the limiting factor of the blood-MLPA assay is the amount of samples that can be tested in a single thermocycler, throughput can easily be increased by adding extra PCR equipment.

In conclusion, the blood-MLPA genotyping assay correctly determines clinically relevant blood group alleles. Similar to HLA typing in which genotyping has replaced serologic typing, the safety of blood transfusion can be increased by using genotyping. Although rare null alleles will be a drawback of each genotyping method, except for complete sequencing, the use of the blood-MLPA will result in better matched blood transfusions and lower levels of immunization. Especially in patients with a high frequency of variant blood group alleles, such as sickle cell patients, who have a recurrent blood transfusion need. The blood-MLPA is the first commercially available genotyping method that provides in a relatively simple test more accurate information than serologic typing. If at the same time blood donors are more typed for all relevant antigens by an assay that is more adapted for automated high-throughput like the ID Core+ assay the revolutionary changes as once foreseen9,25,26, awaiting the first high- throughput blood group genotyping assays, may finally become true.

Acknowledgements We thank Peter Ligthart (Sanquin Diagnostic Services, Amsterdam, the Netherlands) for his technical assistance; Lianne Schuitemaker and Margreet Berkhout (Sanquin Diagnostic Services, Amsterdam, the Netherlands) for the extensive serologic typing; Karel de Groot (MRC Holland, Amsterdam, the Netherlands) for assistance in designing the MLPA probe mixes; Progenika Biopharma (Vizcaya, Spain) for providing the IDCore+ assays; and Connie Westhoff and Sunitha Vege (Laboratories of Immunohematology and Immunochemistry, New York Blood Center, New York, NY) for providing DNA samples.

Web Resources Product description of blood-MLPA mix p401, p402, and p403, including copy number of each probe present in control sample. Download pdf file at http://www.mrc-holland.com/WebForms/WebFormProductDetails. aspx?Tag=tz2fAPIAupKyMjaDF\E\t9bmuxqlhe/Lgqfk8Hkjuss|&ProductOID=FTLzRGltR9Y| Accessed at 07-07-2013.

44 5. 4. 3. 2. 1. References 6. 27. 26. 25. 24. 23. 22. 21. 20. 19. 18. 17. 16. 15. 14. 13. 12. 11. 10. 9. 8. 7. Reid ME,Lomas-FrancisReid C,OlssonML. 3 The bloodgroup book. antigenfacts bloodcellalloimmunizationinpregnancy. Red Hematol KJ. Semin 2005;42:169-78. Moise antibodies. Transfus 2002;12:287-95. Med Daniels G,Poole J, deSilvaM,Callaghan T, MacLennan S,SmithN. The clinicalsignificance ofblood group antenatal care. Vox Sang1974;26:551-9. Spielmann W, S. Prevalence Seidl of irregular red cell antibodies and their significance inblood transfusion and ME.Bloodgroups:Daniels G,Reid thepast50years. Transfusion 2010;50:281-9. Daniels G.Humanbloodgroups. 2 Press; 2012. Anstee DJ. Goodbye to agglutinationandallthat? Transfusion 2005;45:652-3. serology finished? Transfusion 2007;47(1 Suppl):1S-2S. G. Garratty Where are we, andwhere are we going, withDNA-basedapproaches inimmunohematology? Is nanofluidic real-time polymerasechainplatform.reaction Transfusion 2010;50:40-6. Hopp K, Weber BellissimoD, K, ST, Johnson Pietz B. red usinga Highthroughput bloodcellantigengenotyping donors for minorred bloodcell andplatelet antigens. Transfusion 2006;46:841-8. A,PhillipsMontpetit MS,MongrainI,Lemieux R, St-Louis molecular profiling M.High-throughput ofblood 2003-2009. Transfus Hemother2009;36:162-7. Med A,JinochP,G, Hacker Svobodova E,deHaasM. I,vanderSchoot The BloodgenProject ofthe European Union, PM, JR,BeiboerS,Maaskant-vanWijk vonI,JiménezE, Zabern TE, Storry Tejedor D, López M,Camacho E,Cheroutre Avent A,Flegel ND, Martinez WA, E,Madgett M,DanielsGL,Muniz-Diaz ML,NoguesN,Pisacka OlssonML,Scott blood group alternative approach. genotyping–the spectrometry-based Transfus 2013;27:2-9. Rev Med S,Frey C,Meyer Gassner mass BM, time-of-flight laser desorption/ionisation, Vollmert C.Matrix-assisted Beattie KM,Sigmundcell patients. J, KE,McGraw ShurafaM.U-variant bloodinsickle Transfusion 1982;22:257. Quantitation ofDsites onselected D”“weak and D”“partial red cells. Vox Sang1993;65:136-40. DC,Ouwehand B, MA, McDougall Gorick WH, Overbeeke Tippett P, Hughes-Jones DJ. NC,vanRhenen cells.2005;45:520-6. Transfusion Wagner T, GF, Kormoczi BuchtaC, Vadon M,Lanzer G,Mayr WR, Legler TJ. Anti-Dimmunizationby DELred blood 2013;53:1559-74. probe amplification. viamultiplexligation-dependent genotyping andzygosity and RHCEvariant Transfusion Haer-Wigman L, Veldhuisen R,Loden B, Jonkers M,Madgett TE, Avent ND, CE.RHD deHaasM,vanderSchoot one year ofroutine clinicaluse. Diagn JMol 2010;12:453-60. optimization,validation,and nucleotide polymorphisms: Di CJ, M,Chiaroni Silvy J, BaillyP. SinglePCRmultiplexSNaPshot reactionfor ofeleven bloodgroup detection A flexiblearray format rapidblood for large-scale, grouptyping. DNA Transfusion 2005;45:680-8. Hashmi G,Shariff Seul M,T, Vissavajjhala K, Charles-Pierre P, Hue-Roye D, Lomas-FrancisReid ME. A, C,Chaudhuri microarray hybridization. Transfusion 2005;45:667-79. JT, of bloodgroup genotyping antigens de byHaas M. Rapid multiplex polymerase chain reaction and DNA Beiboer SH, Wieringa-Jelsma T, Maaskant-VanWijk PA, CE,vanZwieten vanderSchoot D, R,Roos denDunnen by high-throughput DNAanalysis. Transfusion 2007;47:736-47. Lomas-Francis of 24 minor red ME. Determination C, Reid blood cell antigens for more than 2000 blood donors Hashmi G,Shariff T, Zhang Y, Cristobal Seul M,J, ChauC, Vissavajjhala K, Charles-Pierre P, BaldwinC,Hue-Roye D, and antigengenotypes.Transfusion 2005;45:660-6. GA, Denomme analysisfor polymorphism Van red cell multiplexsingle-nucleotide M.High-throughput Oene celldiseasepatients. the managementofmultiply-transfusedsickle Transfusion 2002;42:232-8. M,BiancoC,PellegrinoCastilho L,Rios FL,SaadST, JJ, Alberto Costa FF. ofblood groups DNA-based typing for patients whohave recently received atransfusion.Transfusion 2000;40:48-53. M,Powell ME,Rios Reid VI, Charles-Pierre D, Malavade V. DNAfrom bloodsamplescanbeusedto genotype WesthoffRh complexities:serology CM. andDNAgenotyping. Transfusion 2007;47(1Suppl):17S-22S. learned? Transfus 2011;25:111-24. Rev Med Heathcote DJ, Carroll TE, Flower years ofantibodiesto MNSsystem hybrid glycophorins:whathave RL.Sixty we 2008;24:190-5. Westhoff CM. The potential ofblood Immunohematologygroup genotyping for transfusionmedicinepractice. nd ed. Oxford: 2002. Science; Blackwell Genotyping ofbloodgroupGenotyping antigens rd ed. SanDiego(CA): Academic 45

Chapter 2 Chapter 2

46 Chapter 2 information Supporting Genotyping ofbloodgroupGenotyping antigens 47

Chapter 2 Chapter 2

Figure S1. Schematic overview of the MLPA reaction. a) An MLPA reaction consists of up to 50 dedicated probe combinations. Each probe combination consists of two or three pieces. Supplementary Figure S1 shows an example of probe combinations consisting of three probe pieces. Each probe combination has two primer tags (light green), that are identical for all tag carrying probe pieces, a target specific region (light blue) and a stuffer region (purple). b) After denaturation of genomic DNA the different probe pieces hybridize to their specific target sequence. A mutation positioned at the target of probe 2 prevents the complete hybridization of probe piece 2a. c) Probe pieces 1a, 1b and 1c hybridize exactly adjacent to each other and are, upon addition of ligase, ligated to generate a PCR template. Ligation of probe piece 2a to probe piece 2b will fail, due to the incomplete hybridization. d) The PCR template is amplified using a universal primer set, which is fluorescently labeled (dark green with yellowstar). e) Each generated product has a unique length (due to the stuffer region). A Genetic Analyzer is used for size separation of the PCR products. After size separation a fluorescent signal is detected for probe 1, because the specific target of probe 1 was present. No signal is obtained for probe combination 2, because the specific target of probe combination 2 was not present.

48 by determined theblood-MLPAand predicted phenotype Excel matrix. are andgenotype automatically determined.Excel-matrix thephenotype The tableshows thegenotype the results is the same as in A. c) When data of all three MLPA probe mixes is copied into the blood-MLPA of theanalyzed samplecompared to thereference ofeachprobe isstated inatable. of Interpretation the reference CROM*01 astwo reference, therefore thecopy-number of CROM*01 arrow) shows half the amount offluorescence compared intensity to thefluorescence ofthe intensity CO*01 therefore thecopy-number ofCO*01 combination (red arrow) theamountoffluorescence showscompared twice intensity to the reference, FY*01 fluorescence asthesignals intensity ofthe reference, indicatingthatthe tested samplehasthesame signal of the probe combination targeting probe combinations target chromosomal regions whichare supposedto bechromosomally stable. The probes (green arrows) are used for the analyzed between normalization sample and the reference. Control sample andred peaks indicate signals from thecontrol samplethatisused asareference. The control Figure S2. copy number asthereference; oneFY*01 alleles. In contrast,the signalalleles. obtained for In the Result of MLPAResult mix p401. a) The blue peaks indicate the probe signals from the analyzed alleles therefore ouranalyzed sample has oneCROM*01 of theanalyzed sampleisdoublecompared to thereference: two FY*01 allele. The signal obtainedfor theCO*01 of the analyzed sample (blue arrow) shows the same in theanalyzed sampleis half thatofthereference, CROM*01 CROM*01 specific-probe combination (orange Genotyping ofbloodgroupGenotyping antigens allele. b) The ratio specific-probe 49

Chapter 2

Chapter 3

RHD and RHCE variant and zygosity genotyping via multiplex ligation–dependent probe amplification

Lonneke Haer-Wigman1 Barbera Veldhuisen1 Remco Jonkers1 Martin Lodén2 Tracey E. Madgett3 Neil D. Avent3 Masja de Haas1 C. Ellen van der Schoot1

1 Sanquin Research, Amsterdam and Landsteiner Laboratory, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands 2 MRC-Holland b.v., Amsterdam, the Netherlands 3 School of Biomedical and Biological Sciences, Plymouth University, Plymouth, United Kingdom.

Transfusion 2013;53(7):1559-74 Chapter 3

Abstract Background: The presence of a D variant may hamper correct serologic D typing, which may result in D immunization. D variants can be determined via RHD genotyping. However, a convenient single assay to identify D variants is still lacking. We developed and evaluated a multiplex ligation-dependent probe amplification [MLPA] assay to determine clinically relevant RHD and RHCE variant alleles and RHD zygosity.

Methods: We analyzed 236 cases (73 normal and 163 selected samples) with the RH-MLPA assay, which is able to determine 79 RHD and 17 RHCE variant alleles and RHD zygosity. To confirm the results, mutations were verified by RHD and/or RHCE exon–specific sequencing and RHD zygosity was verified by quantitative real-time polymerase chain reaction for 18 cases.

Results: In 99% of the cases, the RH-MLPA assay correctly determined whether a person carried only wild-type RHD and RHCE alleles (n = 69) or (a) variant RHD allele(s) and/or (a) variant RHCE allele(s) (n = 164). In only three cases, including two new RHD variant alleles, the variant allele was not identified, due to lack of detecting probes. These were RHD*16.02, a new partial RHD allele, RHD*525A (p.Phe175Leu), and a new D- null allele, RHD*443G (p.Thr148Arg). All RHD (n = 175) and RHCE variant alleles (n = 79) indicated by the RH-MLPA assay were confirmed by sequencing. RHD zygosity was confirmed by quantitative PCR. Two hematopoietic chimeras were recognized.

Conclusion: The RH-MLPA genotyping assay is a fast, easy, and reliable method to determine almost all clinically relevant RHD and RHCE variant alleles, RHD zygosity, and RHD+/RHD- chimeras in blood donors, blood recipients, and pregnant women.

52 of theRHD reactions and/orhemolyticdiseaseofthefetus and/ornewborn. an immune response in D- recipients.an immuneresponse inD- against theepitopes miss. they carrying a partial RHD apartial carrying RHCE is expanding. cases with discrepant serologic results and genotyping molecular that analysis require further applied for andfor blooddonortyping noninvasive fetal RHD alleles, alleles encoding the weak (D D, andD-elution A wide genetic diversity existsin RHD A widegeneticdiversity epitopes; hence they will not make anti-D.epitopes; willnotmake hencethey convenient andcheapassay isstilllacking. alloanti-D) isin whereable a to and all variants make variants frequently the carrier occurring alloanti-D formation inindividualswithweak Dvariants. systems in humans. bloodgroupThe Rh system isoneofthemostcomplexandimmunogenicbloodgroup Introduction the fetus and/ornewborn. are for atrisk hemolyticdiseaseof anti-D formation andsubsequentlyofanti-D–mediated aD+fetusOnly pregnant andwhoare women orexpress variant whoare carrying anRhD D- protein, butinlow amounts. RhD variant ispresent. variant RhD which determine to correctly it isimportant are nature variant, basedontheexact ofRhD present. is whichvariant determine only reveal butcannotexactly thepresence variant, ofanRhD of theRHD*Ψallele(RHD*Pseudogene) orthehybridRHD*03N.01 deletion oftheRHD Type population. 1intheblackAfrican expression protein. oftheRhD prevent antigens. immunizationto theRh statusinblooddonors, theRh determine bloodrecipients,correctly andpregnant women to carries the D antigen and the RhCE protein theC,c, theDantigenandRhCE E,andeantigens. carries carries of blooddonors, recipients, andpregnant women. Serology with monoclonal anti-Discurrently thestandard status method to determine theRh RHCE bloodgroupThe antigensoftheRh system are genes, encodedby closelylinked RHD two variant alleles.variant that encode the homologous RhD and RhCE proteins, andRhCE that encodethehomologousRhD respectively. variant alleles,expressionvariant basedontheirphenotypic (see Web resources). Individuals 2,5 Becausetransfusionpolicy, aswell aspreventive measurements pregnancy, during and 5,25 14 alleleshaveThese variant beenclassified into RHD four groups: thepartial Amethodto identifyabroad (all variants rangeofclinicallyrelevant Rh RHCE 1,2 gene (theRHD*01N.01 Antibodies against the Rh antigens can give Antibodiesagainst the Rh severe hemolytictransfusion allele lackoneormore Depitopes andare prone alloantibodies to make 2,5,6,14 genes. 3,6

RHD 2,15 Currently, thepresence ofaD+fetus with canbedetermined 5,25,26 In theory, individualswithaweak Dor In 2,15 14,15 and Because molecular genotyping methodsare Becausemoleculargenotyping increasingly Individuals carrying aweak Dor carrying Individuals causedby ismostoften theentire phenotype The D- RHCE 19,20 21-23 and Individuals with a D- nullallelelackthecomplete withaD- Individuals

5-7 14,25 allele) intheCaucasian populationorthepresence 14,15 variant alleles can be determined via genotyping viagenotyping allelescanbedetermined variant

RHCE, resulting inmore than250RHD There have been,however, describing reports 2,15,24 el ) phenotypes, the D- null alleles and the ) phenotypes,the D- Serologic will, typing inmostcases, 16-18 Bothweak Dand variants andzygosity ofRHDvariants Genotyping genotyping, thenumberof allele, formerlycalled(C)ce 3,4 Therefore, itiscrucialto el alleleexpress theRhD el variant express variant allD 8 proteinThe RhD 9-13

el and can elicit RHCE and 53 S

Chapter 3 Chapter 3 noninvasive fetal RHD typing.27,28 If the father is homozygous for the RHD*01 allele (wild- type RhD) this test can be omitted.6,14 Upon serologic typing of the C, c, E, and e antigens of the father, a prediction of the paternal zygosity can be made based on the most frequently occurring RHD/RHCE haplotypes.6,29 Because the haplotype frequencies differ between ethnic populations, these predictions are not reliable without knowledge of the ethnic background.6,29 The exact paternal zygosity can be determined by genotype-based methods; however, the methods used at the moment are cumbersome and can be confounded by genetic alterations, especially in the black African population.30-33 We therefore developed a multiplex ligation–dependent probe amplification [MLPA] genotyping assay. This single assay detects mutations and copy number variation of the RHD and RHCE genes and is easy to use, as it only requires a thermocycler and capillary electrophoresis equipment. The aim of this study was to evaluate the performance of the RH-MLPA genotyping assay in the determination of clinically relevant RHD and RHCE variant alleles and RHD zygosity.

Material and methods Materials DNA samples from 73 random blood donors with normal Rh serology were included after informed consent was given (including 11 D- and 62 D+ samples; 44 C+, 60 c+, 27 E+, and 68 e+ samples). Sanquin Diagnostics is the national reference laboratory for the analysis of the presence of Rh variants in blood recipients, blood donors, and pregnant women. For this study 97 cases were selected, which were analyzed between 2003 and 2011. This selection included preferentially a minimum of three samples for each known RHD variant allele and all samples of which the specific RHD variant was not yet determined. Twenty DNA samples were provided by other blood group centers and were selected for the presence of a specific RHD variant allele. Forty-six cases positive for the RHD*Ψ allele were obtained from the screening program of D- pregnant women.

D serology Serologic D typing was routinely preformed with two anti-D reagents. A monoclonal anti-D reagent (immunoglobulin [Ig]M clone MS201, Sanquin Reagents) and a monoclonal blend reagent (IgM clone TH28 and IgG clone MS26, Sanquin Reagents) were used in a method with an immediate spin at room temperature. All D- samples were also tested with the monoclonal blend reagent in the indirect antiglobulin test. In case of a discrepancy between the results of the two reagents, the D-epitope pattern was determined with an in-house panel of anti-D reagents and/or the extended partial RhD typing set from Bio-Rad Laboratories B.V..

54 developed. This An MLPA assay specific for theanalysisofRHD MLPA DNAbloodminikit). (QIAamp kit isolated from white bloodcellsusingaDNAextraction Ethylenediaminetetraacetic acid–anticoagulated andgenomicDNAwas bloodwascollected DNA isolation the Version For 1.85,Softgenetics). positive for control eachruntwo sampleshemizygously (Genemarker, usingcomputer software Applied Biosystems). Dataanalysiswasperformed 500-Liz, AppliedBiosystems)(GeneScan wasanalyzed onageneticanalyzer 3130, (Model of MLPA sample, 8.5μLofFormamide AppliedBiosystems), and0.5μLofsize standard (Hi-Di, 30 secondsat60°Cand1minute at72°C,followed by 20minutes of1.5μL at72°C.Amixture sample. PCRconditionswere asfollows: of30secondsat95°C, 5minutes at95°C,35cycles dilution buffer, 2μLofuniversal primers, 0.5μLofSALSApolymerase, and10μLofligation in atotal volume of50μLcontaining4SALSAdilutionbuffer, 2μLofSALSAenzyme and incubated for 15minutes at54°C.A polymerase chainreaction[PCR]wasperformed 1.5μLofSALSAligaseBuffer A,and1.5μLSALSA ligaseBufferSALSA ligase-65, B were added buffer were added at 25°C. After at60°C hybridization 16and20 hours,for between 1μLof containing 100 ng of DNA was denatured and 1.5 μL probe mix and 1.5 μL SALSA MLPA dilution (Biometra on athermocycler T1, Westerburg B.V. or 5μL Veriti, short, AppliedBiosystems). In The MLPA according to themanufacturers’ reactionwasperformed protocol Holland) (MRC of theRH-MLPA hadto bedividedinto three pools, 1). p401,p402, andp403(Table Because probes hybridizing to thesameregions willhamper each others’ binding, theprobes several subtypes,for between discriminate example, RHD*12.01 combinations are RHD usedto thecopy determine numberofwild-type mutation, and two RHCE mutation, andtwo samples canbeanalyzed simultaneouslyinasinglerunconsistingofallthree RH- MLPA mixes. time.hours, 1and 16 Between including16hoursof hybridization and2hoursofhands-on example oftheresults ofanMLPA reaction. The total RH-MLPA approximately assay takes 20 were usedasreference Figure samplesto Supplementary zygosity. determine S1shows an another 28RHD The designthe additional20probe 1). waschosen(Table combinationsathree-piece probe combinationswere conventionally designed andconsisted probe oftwo piecesandfor while mutationprobe thepresence combinationsdetect ofamutated sequence. Twenty-five RHD*01 RH-MLPA assay isableto 51RHD distinguishbetween allele, viaRHD asdetermined and four RHCE RH-MLPA assay contains17RHD mutation probe combinations (Table 1). So-called wild type probe wildtype mutation probe 1).So-called combinations(Table variant alleles the assayvariant canidentifybutcannot the maintype, Exons 5-and7-specific quantitative real-time PCR and wild type, fivewild type, RHCE , their variants and zygosity was andzygosity RHCE, theirvariants and 13RHCE or variants andzygosity ofRHDvariants Genotyping RHD*12.02 wild type, 21RHD wild type, or variant alleles.variant For RHCE alleles (Table 2). alleles (Table sequence, 55

Chapter 3 Chapter 3

3 3 4 5 0 6 9 3 4 1 5 Mutation verfiedMutation sequencing by 3 3 4 0 6 4 4 1 14 13 46 222 191 220 208 221 123 220 207 168 196 159 230 Signal Signal detected in cases tested P401 P401 P402 P403 P402 P401 P403 P403 P402 P403 P402 P401 P402 P401 P403 P402 P401 P402 P401 P402 P401 P402 P401 Present in Present mix Use of probe/detecting Use RHD*01 RHD*01W.03 RHCE RHD*01N.08 RHCE*02.08.01 RHD*03.01, RHD*03.04, RHD*04.01, RHD*03N.01 RHD*01 and RHCE*C RHD*01N.10 RHCE*c RHD*07.01 RHD*01 and RHCE*C RHCE*C RHD*01 RHD*03.01, RHD*03.04, RHD*03.06, RHD*04.01, RHD*12.03, RHD*03N.01 RHD*01W.05 RHD*01 RHD*01EL.02 RHD*12 RHD*01 RHD*01 RHD*ψ RHD*01 RHCE*e 217 149 143 144 156 241 224 214 218 158 180 156 167 192 188 175 247 163 187 224 130 211 205 Probe Probe size c.419T c.225C c.405C c.383A c.639C c.712A c.269G c.-173C c.487-5T c.486+4T Secondary site‡ ligation c.8G c.48C c.48A c.186T c.380T c.410T c.667T c.307C c.329C c.270A c.509C c.602C c.446A c.455A c.514A c.609A c.122G c.676G c.-132A insertion c.486+1A c.336-20G c.[201G; 203G] c.[201G; RHCE*C specific Primary site† ligation Exon or Exon Intron 5’ UTR 5’ Exon 1 Exon Exon 1 Exon Exon 1 Exon Exon 1 Exon Exon 2 Exon Exon 2 Exon Exon 2 Exon Exon 2 Exon Exon 2 Exon Intron 2 Intron 2 Exon 3 Exon Exon 3 Exon Exon 3 Exon Exon 3 Exon Intron 3 Exon 4 Exon Exon 4 Exon Exon 4 Exon Exon 4 Exon Exon 5 Exon Exon 5 Exon Gene RHD RHD RHD/RHCE RHD RHCE RHD RHD/RHCE RHD RHCE RHD RHD/RHCE RHCE RHD RHD RHD RHD RHD RHD RHD RHD RHD RHD RHCE Probe name Probe D00_-132A MD01_008G DCE01_048C MD01_048A MCE01_122G MD02_186T DCE02_203G MD02_270A CE02_307C MD02_329C DCE02_IVS1-20G CE02_ins_C D03_380T MD03_410T MD03_446A D03_455A MD03_IVS3+1A MD04_509C D04_514A D04_602C MD04_609A D05_667T CE05_676G Performance of RHD and RHCE-specific-probes from the RH-MLPA tested in 236 cases tested the RH-MLPA from of RHD and RHCE-specific-probes 1. Performance Table

56 Table 1. (Continued) Probe name Gene Exon or Primary Secondary Probe Use of probe/detecting Present in Signal Mutation verfied Intron ligation site† ligation site‡ size mix detected in by sequencing tested cases CE05_676C RHCE Exon 5 c.676C c.712A 211 RHCE*E P401 51 D05_697G RHD Exon 5 c.697G 192 RHD*01 P403 195 MCE05_733G RHCE Exon 5 c.733G c.787A 204 RHCE*01.20, RHD*03N.01 P402 19 10 D05_787G RHD Exon 5 c.787G c.801+17C 181 RHD*01 P401 198 D05_800A RHD Exon 5 c.800A 148 RHD*01 P403 202 MD06_807G RHD Exon 6 c.807G 167 RHD*01N.18, RHD*ψ P403 46 4 MD06_809G RHD Exon 6 c.809G 192 RHD*01W.01 P402 24 5 RHD*03.01, RHD*03.06, MD06_819A RHD Exon 6 c.819A c.802-9T 169 RHD*09.03, RHD*09.04, P401 5 3 RHD*09.05 MD06_845A RHD Exon 6 c.845A 164 RHD*15 P403 5 4 MD06_872G RHD Exon 6 c.872G 132 RHD*09.05 P402 0 0 MD06_885T RHD Exon 6 c.885T c.906G 174 RHD*11 P401 5 4 D06_932A RHD Exon 6 c.932A 237 RHD*01 P402 203 D07_941G RHD Exon 7 c.941G 174 RHD*01 P403 211 D07_989A RHD Exon 7 c.989A 143 RHD*01 P401 211 RHD*04.03, RHD*09.01, MD07_1025C RHD Exon 7 c.1025C c.992A 151 P402 7 4 RHD*09.02, RHD*01W.29 andzygosity ofRHDvariants Genotyping D07_1061T RHD Exon 7 c.1061T 181 RHD*01 P403 206 MD07_1063A RHD Exon 7 c.1063A c.1073+25T 222 RHD*25 P401 3 3 MD08_1136T RHD Exon 8 c.1136T 211 RHD*10 P403 5 5 MD09_1154C RHD Exon 9 c.1154C c.1999A 236 RHD*01W.02 P401 8 5 D09_1193A RHD Exon 9 c.1193A c.1227+2T 187 RHD*01 P402 219 MD09_1227A RHD Exon 9 c.1227A c.1227+46C 204 RHD*01EL.01 P403 3 3 D10_1375C RHD Exon 10 c.1375C 162 RHD*01 P401 223

57 † Position as counted from ATG translation start site and nucleotide of the ligation site that is specific for the nucleotide detected by the probe ‡ Position as counted from ATG translation start site and nucleotide of a second ligation site which is used to make the probe specific for RHD or RHCE

Chapter 3 Chapter 3

MD06_819A

MD06_819A

MD03_410T

MD06_819A

MD03_410T

MD02_329C MD07_1025C

MD03_410T

MD02_186T

MD03_410T

MD02_186T

Gain of mutation-probe combination(s) in combination(s) Gain of mutation-probe allele variant specific P401 D10_1357C P401

P402 D09_1193A P402

- - -

P403 D07_1061C P403

- - - -

P401 D07_989A P401

- -

P403 D07_941G P403

- -

P402 D06_932A P402

- - -

P403 D05_800A P403

- - - - -

P401 D05_787G P401

------

P403 D05_697G P403 ------

RHD P402 D05_667T P402

------

P402 D04_602C P402

------

P401 D04_514A P401

- - - -

P402 D03_455A P402

------

P402 D03_380T P402

- - -

P403 DCE02_203G P403

-

P402 DCE02_-20C P402

-

P401 D00_-132A P401

P401 CE05_676G P401

+ + + + + + P401 CE05_676C P401

+ + P402 CE02_307C P402

+

RHCE P401 CE02_ins_C P401

P402 DCE01_048C P402

Loss or gain of wild-type-probe allele variant Loss in specific combination(s) RHD*05.07 RHD*05.06/RHD*05.08 RHD*05.04/RHD*05.05/ RHD*05.09 RHD*05.03/RHD*13.02 RHD*05.02 RHD*05.01 RHD*04.06 RHD*04.05 RHD*07.01 RHD*04.03 RHD*06.04 RHD*03.06 RHD*06.03 RHD*03.04/RHD*04.01 RHD*03.03 RHD*06.02 RHD*03.02 RHD*03.01 RHD*06.11 Variant allele Variant RHD*02/RHD*04.04 -MLPA assay alleles that can be determined 2. RHD and RHCE -variant using the RH -MLPA Table

58 Table 2. (Continued)

Loss or gain of wild-type-probe combination(s) in specific variant allele

RHCE RHD

Gain of mutation-probe combination(s) in Variant allele specific variant allele P402 D05_667T P402 D03_380T P402 D04_602C P403 D05_800A P402 D06_932A P401 D07_989A P402 D03_455A P401 D04_514A P403 D05_697G P401 D05_787G P403 D07_941G P401 D00_-132A P402 CE02_307C P401 CE05_676C P401 CE05_676G P403 D07_1061C P401 D10_1357C P402 D09_1193A P401 CE02_ins_C P402 DCE02_-20C P402 DCE01_048C P403 DCE02_203G RHD*07.02 + MD02_329C RHD*09.01 - - MD07_1025C RHD*09.02 - - - MD07_1025C RHD*09.03/ RHD*09.04 - - MD06_819A RHD*09.05 - - MD06_819A MD06_872G RHD*10.00/RHD*10.01/ RHD*10.02/RHD*10.03/ MD08_1136T RHD*10.06/RHD*10.07 RHD*10.04 - MD08_1136T RHD*10.05 - - MD08_1136T RHD*11 MD06_885T RHD*12.01/RHD*12.02 - MD04_509C RHD*12.03 - MD03_410T MD04_509C variants andzygosity ofRHDvariants Genotyping RHD*13.01 + - - - - RHD*14.01 ------RHD*14.02 ------RHD*15 MD06_845A RHD*16.01/RHD*30 - RHD*17.01/RHD*17.03 - RHD*17.02 - - RHD*25 - MD07_1063A

59 RHD*01W.01 MD06_809G

Chapter 3 Chapter 3

MCE05_733G

MD06_807G

MD03_410T

MD06_807G MD02_270A MD01_048A

MD04_609A

MD_IVS3+1A MD09_1227A

MD07_1025C

MD09_1154C MD02_186T Gain of mutation-probe combination(s) in combination(s) Gain of mutation-probe allele variant specific P401 D10_1357C P401

P402 D09_1193A P402

-

P403 D07_1061C P403

P401 D07_989A P401

P403 D07_941G P403

P402 D06_932A P402

P403 D05_800A P403

+ P401 D05_787G P401

+ P403 D05_697G P403

RHD P402 D05_667T P402

-

P402 D04_602C P402

-

P401 D04_514A P401

P402 D03_455A P402

P402 D03_380T P402

P403 DCE02_203G P403

-

P402 DCE02_-20C P402

P401 D00_-132A P401

P401 CE05_676G P401

+ + + + + + P401 CE05_676C P401

P402 CE02_307C P402

+ +

RHCE P401 CE02_ins_C P401

P402 DCE01_048C P402

+ Loss or gain of wild-type-probe allele variant Loss in specific combination(s) Continued) RHD*01N.20 RHD*01N.18 RHD*01N.10 RHD*01N.08 RHD*01N.07 RHD*01N.06 RHD*01N.05 RHD*01N.04 RHD*01N.03 RHD*01N.02 RHD*ψ RHD*01N.01 RHD*01EL.02 RHD*01EL.01 RHD*01W.41 RHD*01W.29 RHD*01W.14/ RHD*01W.40/RHD*01W.51 RHCE*01.04/RHCE*01.05 RHD*01W.02 RHD*03N.01 Variant allele Variant Table 2. ( Table

60 Table 2. (Continued)

Loss or gain of wild-type-probe combination(s) in specific variant allele RHCE RHD

Gain of mutation-probe combination(s) in Variant allele specific variant allele P402 D05_667T P402 D03_380T P402 D04_602C P403 D05_800A P402 D06_932A P401 D07_989A P402 D03_455A P401 D04_514A P403 D05_697G P401 D05_787G P403 D07_941G P401 D00_-132A P402 CE02_307C P401 CE05_676C P401 CE05_676G P403 D07_1061C P401 D10_1357C P402 D09_1193A P401 CE02_ins_C P402 DCE02_-20C P402 DCE01_048C P403 DCE02_203G RHCE*01.07 + + RHCE*01.08/RHCE*01.09 + RHCE*01.20 MCE05_733G RHCE*01.20.06 + + MCE05_733G RHCE*01.22 ≥1 + + + + RHCE*ce-D(9)-ce + RHCE*02.02 + + RHCE*02.04 ≥1 + + + +

RHCE*02.08.01 MCE01_122G andzygosity ofRHDvariants Genotyping RHCE*02.10.01 + + RHCE*02.10.02 + + + RHCE*02.17 + RHCE*03.03 + + RHCE*03.04 + 61

Chapter 3 Chapter 3

RHD multiplex PCR In our institution all samples that show unexpected weak or negative reactions with D-typing reagents are currently analyzed using a multiplex PCR, simultaneously detecting RHD Exons 3, 4, 5, 6, 7, and 9 and an internal control.34 The RHD multiplex [RHD-MPX] PCR was performed on a DNA engine thermocycler (Dyad, Bio-Rad Laboratories B.V.) in a total volume of 50 μL, containing 10 to 1000 ng DNA, 5 μL of 10x FastStart Taq DNA polymerase buffer without MgCl2 (Roche), 3 μL of MgCl2 25 μM stock solution (Roche), 1 μL of PCR-grade nucleotide mix (Roche), 15 μL of primer mix (primer dilution between 0.1 and 1.2 μM; Eurogentec). PCR conditions were as follows: 5 minutes at 95°C, 33 cycles of 1 minute at 95°C, 1 minute at 56°C, and 1 minute at 72°C, followed by 5 minutes at 72°C. The presence or absence of the RHD exons and internal control was analyzed by gel electrophoresis on a 1.5% (wt/vol) agarose gel (Invitrogen).

PCR and sequencing RHD and RHCE exon–specific sequencing was performed using RHD- and RHCE-specific primers flanking each exon. The PCR was performed on a thermocycler (Veriti, Applied Biosystems) in a total volume of 20 μL, containing 50 to 150 ng DNA, 10 μL of 2x PCR master mix (GeneAmp Fast, Applied Biosystems), 0.5 μM forward and reverse primer. PCR conditions were as follows: 10 seconds at 95°C, 35 cycles of 10 seconds at 95°C, and a specific annealing- elongation temperature and time for each primer set ranging from 61 to 68ºC, followed by 1 minute at 72°C. PCR products were purified using PCR product cleanup (ExoSAP-IT, GE Healthcare), according to manufacturers’ protocol. The sequence reaction was performed on a thermocycler (Veriti, Applied Biosystems) in a total volume of 20 μL, containing 1 μL of purified PCR product, 1 μL of 2.5x polymerase mix (BigDye Terminator v1.1, Applied Biosystems), 3.5 μL of 5x buffer (BigDye Terminator, Applied Biosystems), and 0.25 μM forward or reverse primer. Sequence conditions were as follows: 25 cycles of 15 seconds at 95°C, 10 seconds at 50°C, and 4 minutes at 60°C. Sequence products were analyzed on a genetic analyzer (3130, Applied Biosystems).

RHD Exons 5- and 7-specific quantitative real-time PCR Zygosity was determined with an RHD Exons 5- and 7-specific quantitative real-time PCR. Primers were developed to detect ALB35 (quantification of input DNA), SRY36, and RHD Exons 527 and 737. The RHD Exons 5 and 7 primers and probes were synthesized by TIB MOLBIOL and the primers for SRY and ALB were synthesized by Invitrogen. The real-time PCR was performed on a sequence detection system (ABI PRISM 7000, Applied Biosystems) in a total volume of 25 μL, containing 25 ng of DNA, 12.5 μL of universal primer PCR master mix (Taqman, Applied Biosystems), 0.1 μM of the RHD Exon 5 and SRY probe or RHD Exon 7 and ALB probe and 0.3 and 0.9 μM of the RHD Exon 5 and SRY forward and reverse primers, respectively, or 0.3 μM of the

62 RHCE*C, becausetheRHD 8 iscompletely identical. RHD The probe wild-type combinationsdetecting RHCE*03 two casesshowedtwo the presence oftheRHD*07.02 fournot shown). In casesoneormore mutation probe combinationsignals were detected: serology. All possiblecommonRHD/RHCE 2 andc.203G inExon 2.RHCE*c oftheRHCE*C isbasedonthedetection genotyping probe combinationfor RHD combinations are ableto thecopy detect numberofallRHD serology, Rh with normal representing phenotypes. allcommonRh The RHD oftheRH-MLPA theperformance Next, assay wasanalyzed withDNAfrom 73healthy donors probe combinationsWild-type Performance ofRH-MLPA assay withgenomicDNA oftheRH-MLPAperformance wasevaluated withasetofgenomicDNAsamples. template. For all probes signals correct were detected (data not shown) and subsequently the using aclonedDNAplasmidsample, containingthegenomicsequenceofallprobes, as a First, all Performance ofRH-MLPA assay withplasmidDNA Results manufacturers’ protocol. multiplex (Powerplex 16 HS system, Promega). according toThe the reaction was performed The presence was determined with genomic DNA using of short-tandem-repeat chimerism multiplexShort-tandem-repeat PCR Excel 2003). Office (Microsoft usingcomputer software Data analysiswasperformed minutes of15secondsat95°Cand1minute at 50°C,10minutesat60°C. at95°C,and50cycles RHD presence allelesdetected by theRH-MLPA. ofthevariant type probetype combinationsignals were absent. The RH- MLPA RHCE*02.08.01 one case had an Similarly RhD). probewild-type combination signals were present, cases (n = 11) all while in the D- second ligationsite for thec.712AIn all D+ RHCE-specificcases nucleotide. (n = 62) all RHCE RHCE*e Exon 7andALB specificity,two the latter probe combinations consist of three probe pieces, creating a RH-MLPA (n = 22) and mutation probe wild-type combinations (n = 23) were analyzed genotyping is based on the detection ofc.676C isbasedonthedetection andc.676G,genotyping respectively. To ensure , (RhcE) allele (CeC and forward andreverseforward PCRconditionswere primers. Real-time asfollows: 2 RHCE*04 w ). Sequencing oftheinvolved). Sequencing exons the inthesecasesconfirmed and Exon 8wasincluded, becausethesequenceofRHD (RhCE) genotyping results genotyping were(RhCE) completely concordant with allele next to an RHD*Ψ allele next RHCE*C genotyping is based on the detection ofc.307C. isbasedonthedetection RHCE*E genotyping sequence ofExon 2iscompletely identical. RHCE*C haplotypes werehaplotypes atleastonce(data determined allele next to anRHD*01 allele next specific 19-nucleotide insertion in Intron in specific 19-nucleotideinsertion RHD*01 RHCE*01 exons, except thatofExon 8.No allele and one case carried an allele and one case carried variants andzygosity ofRHDvariants Genotyping (RhCe), , RHCE*02(RhCe), (Rhce) Exon 2alsodetect wild-type probewild-type allele (wild-type allele (wild-type and RHCE RHD wild- Exon RHD and 63

Chapter 3 Chapter 3

Mutation probe combinations The mutation probe combinations of the RH-MLPA were selected to recognize, in combination with the wild-type probe combinations, the majority of clinically relevant RHD variant alleles and most frequently occurring RHCE variant alleles in the Caucasian and black African populations. The performance of the mutation probe combinations was evaluated with DNA from 97 cases, for whom serology indicated the presence of an Rh variant and with DNA from 66 cases, for whom the presence and/or expressing of an Rh variant was already determined. With this set of samples 19 RHD mutation probe combination signals and two RHCE mutation probe combination signals were obtained at least once. To confirm the presence of a mutation indicated by one of the mutation probe combinations, the exons containing the respective mutations were sequenced for at least three independent samples (if available). All sequencing results were concordant with the results of the RH-MLPA (Table 1). For two RHD mutation probe combinations MD02_270A specific for the RHD*01N.10 allele and MD06_872G specific for the RHD*09.05 allele no genomic DNA was available. The RHCE mutation probe combination MCE05_733G was designed to determine whether an individual is positive (c.733G) or negative (c.733C) for the VS antigen. Of note, individuals carrying an RHCE*01.04 allele (ceAR) have the c.733G mutation, but are VS-.38 We designed the MCE05_733G with a second ligation site specific for the c.787A RHCE wild-type single- nucleotide polymorphism [SNP], which is present in all VS+ alleles and absent in the RHCE*01.04 allele. Six of the nine cases carrying a RHCE*01.04 allele were VS- and indeed these samples showed the absence of signal for MCE05_733G. In the three other cases the second RHCE allele was a VS+ allele and as expected the MCE05_733G signal was obtained.

Determination of RHD and RHCE variant alleles using the RH-MLPA assay In 160 of the above-described 163 individuals the RH-MLPA indicated the presence of one or more variant alleles. In these 163 cases a total of 172 RHD and 78 RHCE variant alleles or variant allele groups were determined as shown in Table 3. The assigned RHD and RHCE variant allele(s) were in all cases compatible with the initial serologic results obtained for the cases. In addition, in 99 of the 163 selected samples, we performed an RHD-MPX PCR, in which RHD Exons 3, 4, 5, 6, 7, and 9 are RHD specifically amplified. The results of the RHD-MPX combined with the results of the serology were concordant with the RHD variants concluded from the RH-MLPA, except for 11 cases. In these cases, carrying the RHD*05.05, RHD*10.02, RHD*10.03, RHD*11, or RHD*15 allele, the RHD-MPX results showed no aberrant pattern, suggesting the presence of weak D variants, while the RH-MLPA indicated the presence of partial RhD variants. Sequencing of the exons containing the mutations showed that in these cases the RH-MLPA assigned the correct RHD variant alleles (Table 3). In one case the RH-MLPA indicated the presence of a new partial RHD allele, containing the mutations of both RHD*03.02 and RHD*25 in one allele (c.150T>C, c.178A>C, c.201G>A, c.203G>A, c.307T>C, and c.1063G>A).

64 Table 3. RHD and RHCE variant alleles detected by the RH-MLPA assay in 163 DNA samples

Selected donors RH-MLPA conclusion Additional genotyping§ Variant allele

Serology and detected in Sequence‡ RHD RHCE RHD RHCE RHD-MPX PCR† tested cases Partial D alleles RHD*03.01 RHD*03.01 RHCE*01.20 1 New RHD variant (c.150T>C; RHD*03.02 c.178A>C; c.201G>A; c.203G>A; 1 c.307T>C; c.1063G>A) RHD*03.03 RHD*03.03 1 DIVa RHD*03.04 or RHD*04.02 RHD*04.02 3 RHD*03.04 RHD*02 or RHD*04.04 RHD*04.04 1 DVI t1 RHD*06.01 2 DVI t2 RHD*06.02 7 DVII RHD*07.01 4 DAR†† RHD*09.01 RHCE*01.04 or RHCE*01.05 RHCE*01.04 2 RHCE*01.04 or RHCE*01.05 and RHD*09.01 RHCE*01.04 2 RHCE*01.20 RHCE*01.04 or RHCE*01.05 RHCE*01.04 RHD*09.01 (homozygous) 1 (homozygous) (homozygous) RHD*09.01 and RHD*03N.01 RHCE*01.04 or RHCE*01.05 RHCE*01.04 1 RHD*09.02 RHCE*01.04 or RHCE*01.05 RHCE*01.04 2 RHD*01W.29 1

RHD*11 RHD*11 3 andzygosity ofRHDvariants Genotyping RHD*14.02 RHD*14.02 1 RHD*16.01 RHD*16.01 1 DFR RHD*17.01 1 RHD*17.02 2 RHD*12.01 RHD*12.01 or RHD*12.02 RHCE*01.08 or RHCE*01.09 RHD*12.01 RHCE*01.08 1 RHD*25 RHD*25 2 Partial D variant RHD*01 RHD*16.01 1 New variant 1 RHD*525A 65 RHD*05.07 3

Chapter 3 Chapter 3 7 1 1 1 2 1 1 1 1 1 2 1 5 1 1 1 2 3 3 2 2 4 22 30 Variant allele Variant detected in cases tested RHCE

Additional genotyping§ Additional RHD

RHD*09.03 RHD*09.03 RHD*09.03 RHD*05.05 RHD*10.02 RHD*10.03 RHD*10.03

RHCE

RHCE*01.20 RHCE*01.20 RHCE*ce-D(9)-ce

RHCE*ce-D(9)-ce an d RHCE*01.20 RHCE*ce-.D(9)-ce -MLPA conclusion RH -MLPA or RHD*05.09 RHD*05.05 RHD RHD*01W.02 and RHD*03N.01 RHD*01W.02 RHD*10.05 RHD*09.03 or RHD*09.04 RHD*09.03 or RHD*09.04 and RHD*03N.01 RHD*09.03 or RHD*09.04 and RHD*03N.01 RHD*10.00 , RHD*10.01 RHD*10.02 RHD*10.06 or RHD*10.07 RHD*10.03 , RHD*10.00 , RHD*10.01 RHD*10.02 RHD*10.06 or RHD*10.07 RHD*10.03 , and RHD*Ψ RHD*11 RHD*01W.03 RHD*05.04, RHD*15 RHD*01W.05 RHD*01W.01 and RHD*03N.01 RHD*01W.01 RHD*01W.03 RHD*01W.05 RHD*01W.01 and RHD*06.02 RHD*01W.01 RHD*01EL.01 RHD*01EL.02 RHD*01EL.02 RHD*ψ RHD*ψ Sequence‡

RHD*01W.03

RHD*01W.05 RHD*01EL.01 RHD*01EL.02

RHD*ψ Selected donors Serology and Serology PCR† RHD -MPX Weak D type 2 Weak Weak D variant Weak Weak D variant Weak Del variant

el Continued)

alleles

Weak D and Weak D- null alleles Table 3. ( Table

66 Table 3. (Continued)

Selected donors RH-MLPA conclusion Additional genotyping§ Variant allele

Serology and detected in Sequence‡ RHD RHCE RHD RHCE RHD-MPX PCR† tested cases RHCE*ce-D(9)-ce and RHCE*01.04 or RHD*ψ RHCE*01.04 1 RHCE*01.05 RHD*ψ (homozygous) RHCE*ce-D(9)-ce (homozygous) 3 RHD*ψ and RHD*03N.01 RHCE*ce-D(9)-ce 5 RHD*ψ and RHD*03N.01 RHCE*ce-D(9)-ce and RHCE*01.20.06 1 RHD*01N.08 RHD*01N.08 3 New variant D negative‡‡ RHD*01 1 RHD*443G RHD chimera Chimera Chimera 50% RHD 1 Weak D variant Chimera 30% RHD 1 RHCE variant RHCE*02.08.01 RHCE*02.08.01 3 alleles ceHAR RHCE*01.22 1

RHD*10.00, RHD*10.01, RHD*10.02, RhCE variant RHCE*03.04 RHD*10.00 1

RHD*10.03, RHD*10.06 or RHD*10.07 andzygosity ofRHDvariants Genotyping RHCE*02.10.01 1

† RHD and RHCE variant allele determination after serology and/or RHD-MPX-PCR results ‡ RHD and RHCE variant alleles that were chosen for their specific genotype § Sequencing was performed when the RH-MLPA was not able to determine the specific subtype of the variant allele †† Serology and the RHD-MPX-PCR are not able to distinguish between the DAR1, DAR2 and Weak D type 29 variant alleles ‡‡ This patient was included in our series, because her current D serology (D-) was discordant with a previously reported D+ phenotype by the of Azerbaijan 67

Chapter 3 Chapter 3

Zygosity analysis showed the presence of one RHD allele and sequencing confirmed the presence of the mutations. Unfortunately, no red blood cells [RBCs] were available for further serologic typing. In 23 cases the RH-MLPA detected the presence of a variant allele, but could not discriminate between different subtypes (e.g., RHD*10 alleles and RHD*09.03 or RHD*09.04) and the samples were sequenced to assign the specific RHD or RHCE variant subtype (Table 3). In three cases in which serology indicated the presence of an Rh variant, the RH-MLPA indicated the presence of a single normal RHD*01 allele. Sequencing of all RHD exons revealed that one case concerned a hemizygously present RHD*16.02 allele (c.676G>C), for which no RHD-specific probe combination is included. In the two other cases two newRHD variant alleles were detected. One new RHD variant allele has the c.525C>A (p.Phe175Leu) mutation and the other RHD variant allele has the c.443C>G (p.Thr148Arg) mutation. The latter was determined as D- via the absorption-elution technique using a polyclonal anti-D (anti-D bromelain, Sanquin Reagentia). This patient was included in our series, because her current D serology (D-) was discordant with a previously reported D+ phenotype by the blood bank of Azerbaijan. Probably the donor was regarded as D+ because of the CE phenotype (RhCcee). The new RHD*525A (p.Phe175Leu) variant allele results in absence of expression of Epitope 1.2 (determined with monoclonal antibodies [MoAb] LHM174/102 and LHM70/45 of the extended partial RhD typing set of Bio-Rad Laboratories) and Epitope 2.2 (determined with MoAb 5C839). The LHM169/81 MoAb (extended partial RhD typing set), which detects Epitope 1.1 showed the same strength of reactivity for both the D+ control and the RBCs expressing the RHD*525A (p.Phe175Leu) allele.40 Similarly, all other D epitopes tested were normally present. The patient was typed positive for the C, c, and e antigens. In two other cases no RHD variant allele was detected, but an aberrant RHD copy number of 0.3 and 0.5, respectively, was detected. These samples are further discussed below; see “RHD and RHCE zygosity” results. For three variant alleles a loss of signal of a wild-type probe combination was obtained, whereas presence of the RHD-specific nucleotide was shown by sequencing. In three cases carrying the RHD*25 or the new partial RHD allele (c.150T>C, c.178A>C, c.201G>A, c.203G>A, c.307T>C, c.1063G>A), the D07_1061T signal was absent. Apparently, the c.1063G>A mutation present in these alleles impaired the complete hybridization and ligation of D07_1061T. In the 46 cases carrying the RHD*Ψ allele, the D04_514A signal was absent. The 37-bp duplication in intron 3 (c.487-19) present in the RHD*Ψ allele prevents the hybridization of the D04_514A probe combination, because the insertion results in a shortening of the hybridization site to only 13 nucleotides instead of the normal 34 nucleotides. All 46 cases carrying the RHD*Ψ allele were also associated with an unexpected gain of the D09_1193A signal. The RHD and RHCE Exon 9 were therefore PCR amplified in three cases that were homozygous for the RHD*Ψ allele. In these cases RHCE Exon 9 could not be amplified, while RHCE Exons 8 and 10 were normally amplified and sequencing of these exons and surrounding introns showed RHCE wild-type sequence. We therefore hypothesize that in this RHCE variant allele, linked to

68 the 0, 1, or 3 could be explained in all cases. All cases with a RHCE*01.04, For cases the two were concordant bothassays. between usingRHD typing with zygosity by determined theRH-MLPA,zygosity 18cases(seven DD for theRHD cases, 163 cases were for found the to be hemizygous RHD RHD andRHCEzygosity shown). that thesepersonswere positive indeedheterozygous for theRHCE*ce-D(9)-ce alltested samplestheCE09_1193Tshowedcombination. In acopy numberof1,indicating forRHD*Ψ alleleandonecaseheterozygous thisallelewere tested withtheCE09_1193Tprobe MLPA probe combinationfor theRHCE one RHCE break pointsofthemutated RHCE forblackAfricans.been described As D- noproductwasamplifiedintheRHCE serologically positive for have reactionpatterns thecandeantigensalsonoabnormal This hybrid RHCE detect detect probe combinationsexcept for theCE05_676C andCE05_676Gprobe combinationsthat probe theothercases, combinations. acopy In numberof2wasobtainedfor allRHCE-specific 207ofthe236cases(88%)anRHCE In RHD*01DNAispresent. ispossiblewhen8%ormore chimerism hemizygous of theRhD MLPA quantificationoftheamount canreliably andcorrect ahematopoietic detect chimerism probewild-type combinations. RHD*01 When 2%ormore hemizygous RHD*01 probe combinationsandthepresence of4%hemizygous in Figure 1 the presence of 0.25% hemizygous (CcDdEe) wasmixed for withDNAfrom RHD*01N.01 apersonhomozygous the samples were from individuals withahematopoietic chimerism. To of evaluate thesensitivity D.” Further multiplexPCRshowed geneticanalysisusingshort-tandem-repeat thatthese number of0.5andthecasewithacopy numberof0.3wasserologically as determined “weak RHD*01 specific probe combinations, indicatingthepresence of, respectively, 50and30%hemizygously RH-MLPA for to hematopoietic DNA from detect chimerism, a person hemizygous RHD*Ψ allele, RHCE RHD*Ψ allele and hence are for heterozygous the and Exon 9primers. To thepresence confirm ofthis alleleincaseswhichhavehybrid only RHCE*E (RhD+). Dserology(RhD+). showedcopy amixed fieldinthecasewithanRHD- reaction RHCE gene (DD),andin12casesbothRHD and zygosity were for determined all236DNAsamplesby RH-MLPA.zygosity these Of allele is normally expressed caseswereallele isnormally becausethesethree all homozygous RH-MLPA detected an abnormal RHCE*e, respectively. RHCE*E/RHCE*e The abnormal Exon 9isreplaced by RHD Exons 5-and7-specificquantitative real-time PCR. The results allele are located outsidetherangeofatleastoneour specific c.1193T SNP. for the Seven caseshemizygous copy numberof2wasobtainedfor allRHCE-specific RHD*01 RHCE*ce-D(9)-ce Exon 9resulting inaRHCE*ce-D(9)-ce RHD genes were deleted (dd). To the confirm RHCE*ce-D(9)-ce RHCE*ce-D(9)-ce DNA is detected by two copy number of 0.5 and 0.3 for all RHD gene (Dd and eleven Dd variants andzygosity ofRHDvariants Genotyping DNA isdetected by allRHD RHCE*01.08, ), 58 cases homozygous ), 58 cases homozygous allele, we developed an Exon 5copy number of DNA ispresent theRH- (ccddee). As shown ) were compared Exon 9PCR,the allele (data not RHD or RHCE*01.22 wild-type wild-type RHD*01 allele. RHD- 69

Chapter 3 Chapter 3

1.00 Legend 5'UTR exon 2 exon 3 exon 4 exon 5 0.75 exon 6 exon 7 exon 9 exon 10

0.50 Probe signal (comparedto CcDdEesample)

0.25

0.00 % % % 1% 2% 4% 8% 16 32 64 0.25% 0.50% ccddee CcDdEe Percentage of CcDdEe DNA added to ccddee DNA Figure 1. RH-MLPA results obtained with a mixture of hemizygous RHD*01 D+ DNA from a donor pheno- and genotyped as (CcDdEe) mixed with homozygous D- RHD*01N.01 DNA typed as (ccddee). DNA was added in a dilution range from 64% to 0.25% RHD*01 in RHD*01N. The ratio of the RHD wild-type probe combinations is calculated using the hemizygous RHD*01 (CcDdEe) sample as a reference sample. For exons of which more than one set of RHD wild-type probe combinations is detected, the mean of the probes is given. Error bars indicate the standard deviation. Two RHD wild-type probe combinations detected the presence of 0.25% RHD*01. The presence of 2% RHD*01 is reliably detected by the RH-MLPA and the presence of 4% RHD*01 is detected by all RHD wild-type probe combinations. allele have an RHCE*E/RHCE*e Exon 5 copy number of 1 (heterozygous variant allele) or 0 (homozygous variant allele), because they all contain the c.712A>G mutation, which impairs ligation of the second ligation site specific for the c.712A RHCE wild-type SNP of CE05_676C or CE05_676G. In 20 cases an elevated RHCE*E/RHCE*e Exon 5 copy number of 3 was obtained. All these cases carried next to their RHCE alleles an RHD-CE-D hybrid allele that contained a complete RHCE Exon 5. Therefore, the CE05_676G (including the RHD*05.07, RHD*06.02, and RHD*03N.01 allele) or CE05_676C (RHD*06.01) was able to bind to the hybrid allele, causing an elevated RHCE*E/ RHCE*e Exon 5 copy number.

70 to ahybrid RHCE*ce For the first time it was recognized that the as ofRHD andto themajority determine the Rh-phenotype thisstudywe showedIn thattheRH-MLPA assay isareliable genotyping methodto predict Discussion alleles were found:nullalleleRHD*443G aD- probe for oftheRHD*16.02 detection determined withtheRH-MLPA,determined theRHD because itdetermines First, of the the specificity electrophoresis equipment. capillary (p.Phe175Leu). We therefore conclude that the performed. protein. To profound thisunexpected confirm RhDexpression effect an modelshouldbe with apositively charged aminoacid(arginine) transmembrane inthefifth RhD region ofthe allele, causedby amutationresulting inthesubstitutionofapolaraminoacid(threonine) A new fromdiscriminated theotherRHD*10 alleles. is clinicallyrelevant, becausethe RHD*10.00 the specificsubgroup. OnlyintheRHD*10 alleles (10%)detected by theRH-MLPA, to determine additionalsequencingwasnecessary (c.150T>C, c.178A>C, c.201G>A, c.203G>A, c.307T>C, 26ofthe254variant andc.1063G>A). In group wasassigned. onecasetheRH-MLPA In RHD partial couldeven anew determine allelesdetected by theRH-MLPA all254variant allele.variant In orvariant variant thecorrect few samplesfollow-up rare analysisisneededto show oranew thepresence ofeitheravery and/or in the RHD revealed thepresence ofapartial allele(s)was(were) and/orvariant onecasesequencing ifawild-type present.correctly In RHD-MPX PCRorby RHD cases thatwere previously serologically for andinsomecasesalreadywithan typed Rh typed To oftheRH-MLPA evaluate theaccuracy wastested withasetof163 assay itsperformance c.1063G>A mutationandabsenceofD04_514AinRHD*Ψalleles. due to amutationcloseby alleleswith theligationsite: absenceofD07_1061Tinvariant combinations inwhichthelossofsignal wasnotcaused by mutationattheligationsite, but mutated sequenceaccording to theirdesign. We showed exceptionsprobe two ofwild-type All withplasmidDNAandsubsequentlygenomicDNA. combinations wasdetermined RHD RHD RH-MLPA because rare.it isextremely RHD*443G RHCE zygosity. zygosity. The MLPA technique is easy to use, as itonlyrequires and athermocycler and 42

variant in the vast majority ofsamplesanalyzed ina reference inthevastmajority variant laboratory. Onlyin RHCE (p.Thr148Arg) allele was detected in this study. This null seems to be a D- wild-type andmutationprobe or wild-type combinationsdetected thewild-type variant allelecontainingRHD variant exon–specificIn 160cases(98%)theRH-MLPA sequencing. assigned RHD and RHCE allele. In the other two cases, two new variant RHD cases, variant theothertwo new allele. two In allele (RHD*16.02),for whichnoprobe wasincluded variant groupvariant ofthesubtype (n=4)determination wild-type and wild-type (p.Thr148Arg) RHD andapartial RHD*Ψ allele, with normal RhD expression,allele, RhD shouldbe withnormal 41 It is still possible to add an extra mutation is still possible to add an extra It RH-MLPA can be used to determine an . Exon 9(RHCE*ce-D(9)-ce) This couldbe allele is linked (in all 46 persons tested)allele is linked RHD and copy numberofExon 9. This variants andzygosity ofRHDvariants Genotyping RHCE and RHCE variant allelesaswellvariant mutation probe allele RHD*525A allele RHD 71

Chapter 3 Chapter 3 hybrid allele contains two mutations c.1170C>T and c.1193T>A, encoding for one amino acid change p.Glu398Val at the intracellular C-terminal tail of the Rhce protein. Serologic D and CE typing in persons homozygous positive for the RHD*Ψ RHCE*ce-D(9)-ce haplotype showed the normal presence of the c and e antigens and the absence of all D epitopes. The amino acid change is present in the intracellular tail of the protein and thus far no effect on the expression of the RHCE epitopes has been shown.21 The RH-MLPA assay is very accurate in determining gene copy number variation. In the RH-MLPA the RHD copy number is based on the signals derived from 17 RHD wild-type probe combinations. This makes the RH-MLPA more suitable for RHD and RHCE zygosity determination than real-time quantitative PCR30 or amplification of the hybrid Rh box.31-33 Clinically, it is important to determine RHD zygosity in fathers, to assess the need for fetal RHD typing in RhD alloimmunized D- women. Furthermore, because the RH-MLPA is highly accurate in the determination of exon copy number, this is the first high-throughput assay that is able to recognize hybrid RHD/RHCE alleles next to a normal RHD gene. Next to the “normal” zygosity scores of 0, 1, and 2, the RH-MLPA is also able to reliably determine the presence of hematopoietic chimerism of just 2% RHD*01 (RhD+) in an RHD*01N.01 (RhD–) background. It is important to detect RHD+/RHD- chimeras, because it has been shown that transfusion of RBCs from a donor with a hematopoietic chimerism of 6% RHD+ DNA was able to immunize two D- recipients.43 The RH-MLPA can also reliably genotype patients who received multiple non-leukoreduced RBC units. After massive transfusions a (weak) signal derived from donor white blood cells might be detected with the RH-MLPA.44 Because the MLPA is a quantitative method, it will be readily recognized if a signal is obtained from a minor population of transfused white blood cells. Some variants were correlated with a loss or gain of the RHCE*E/RHCE*e Exon 5 copy number. In the RHCE*01.04, RHCE*01.05, RHCE*01.08, and RHCE*01.22 alleles the correct binding of the CE05_676G and CE05_676C probe combination is inhibited and therefore it is not possible to determine whether these variant alleles contain the RHCE*E or RHCE*e variation. Because the RHCE*01.04, RHCE*01.05, RHCE*01.08, and RHCE*01.22 alleles result in partial e antigen expression, an individual carrying one of these alleles should always be determined as RHCE*e+, hence Rhe+.45-47 For the RHD*06 and RHD*03N.01 hybrid alleles in which the RH- MLPA suggest an RHCE*E/RHCE*e Exon 5 copy number of three it is known that the hybrid alleles result in the expression of the e antigen (RHD*06.02 and RHD*03N.01 allele) or the E antigen (RHD*06.01 allele).48 For these cases the presence of the RHCE*e and/or RHCE*E SNP present in the hybrid allele should be taken into account when determining the RhEe status. For the precise determination of variant RHCE alleles and the identification of rare RHCE variant alleles an additional MLPA has to be developed, but as listed in Table 2 the most frequently occurring RHCE variant alleles are recognized with the current RH-MLPA.

72 molecular analysiswillincrease. the number of cases with discrepant serologic results andgenotyping that require further still growing, methodsare becausemoleculargenotyping increasingly appliedandtherefore by for instancetheRH-MLPA assay, ofbloodrecipients, blooddonors, andpregnant women is thereby providing asafe basisfor bloodtransfusion. The needfor extensiveRHD comprehensive ofRHD typing and conclusion, the In Accessed at05-10-2012. blood-group-terminology/bloodgroup-allele-terminology/ isbtweb.org/working-parties/red-cell-immunogenetics-and-bloodgroup-terminology/ ISBT andBloodGroup Cell onRed Immunogenetics Party Working Terminology http://www. Web Resources samples withspecificRHDgenotypes. and Yanli Ji(Guangzhou BloodCenter, GuangzhouGuangdong, China)for providing DNA ofHaematology andBlood (Institute Pisacka Transfusion,UK),Martin Prague, Czech Republic), for Institute (Bristol Transfusion Bloodand Sciences, NationalHealthService Transplant, Bristol, for Clinical Transfusion Daniels Geoff Ulm,Germany), andImmunogenetics Medicine Transfusion von (Institute Zabern Inge Ulm, Germany), andImmunogenetics Medicine assistance indesigning theMLPA probe mixes, Christof Weinstock for (Institute Clinical his technical deGroot Holland, assistance, (MRC Karel Amsterdam, for theNetherlands) We Amsterdam, for thankPeter (SanquinDiagnostic theNetherlands) Services, Ligthart Acknowledgements RHCE variant alleles.variant Therefore, itcanbeusedasasingleandrapidassay to facilitate RH-MLPA assay determines clinically relevant correctly genotyping and RHCE variants inbloodrecipientsvariants andblooddonors, variants andzygosity ofRHDvariants Genotyping genotyping, RHD 73

Chapter 3 Chapter 3

References 1. Daniels G, Reid ME. Blood groups: the past 50 years. Transfusion 2010;50:281-9. 2. Westhoff CM. Rh complexities: serology and DNA genotyping. Transfusion 2007;47 Suppl:17S-22S. 3. Urbaniak SJ, Greiss MA. RhD haemolytic disease of the fetus and the newborn. Blood Rev 2000;14:44-61. 4. Daniels G, Poole J, de Silva M, Callaghan T, MacLennan S, Smith N. The clinical significance of blood group antibodies. Transfus Med 2002;12:287-95. 5. Avent ND, Martinez A, Flegel WA, Olsson ML, Scott ML, Nogues N, Pisacka M, Daniels GL, Muniz-Diaz E, Madgett TE, Storry JR, Beiboer S, Maaskant-vanWijk PM, von Zabern I, Jiménez E, Tejedor D, López M, Camacho E, Cheroutre G, Hacker A, Jinoch P, Svobodova I, van der Schoot E, de Haas M. The Bloodgen Project of the European Union, 2003-2009. Transfus Med Hemother 2009;36: 162-7. 6. Moise KJ. Red blood cell alloimmunization in pregnancy. Semin Hematol 2005;42:169-78. 7. Wiener AS, Unger LJ. Rh factors related to the Rho factor as a source of clinical problems; diagrammatic representation of their reactions and prediction of still undiscovered Rh factors. J Am Med Assoc 1959;169:696-9. 8. Cherif-Zahar B, Mattei MG, Le Van Kim C, Bailly P, Cartron JP, Colin Y. Localization of the human Rh blood group gene structure to region 1p34.3-1p36.1 by in situ hybridization. Hum Genet 1991;86:398-400. 9. Arce MA, Thompson ES, Wagner S, Coyne KE, Ferdman BA, Lublin DM. Molecular cloning of RhD cDNA derived from a gene present in RhD-positive, but not RhD-negative individuals. Blood 1993;82:651-5. 10. Cherif-Zahar B, Bloy C, Le Van Kim C, Blanchard D, Bailly P, Hermand P, Salmon C, Cartron JP, Colin Y. Molecular cloning and protein structure of a human blood group Rh polypeptide. Proc Natl Acad Sci U S A 1990;87:6243-7. 11. Le Van Kim C, Mouro I, Cherif-Zahar B, Raynal V, Cherrier C, Cartron JP, Colin Y. Molecular cloning and primary structure of the human blood group RhD polypeptide. Proc Natl Acad Sci U S A 1992;89:10925-9. 12. Simsek S, de Jong CA, Cuijpers HT, Bleeker PM, Westers TM, Overbeeke MA, Goldschmeding R, van der Schoot CE, von dem Borne AE. Sequence analysis of cDNA derived from reticulocyte mRNAs coding for Rh polypeptides and demonstration of E/e and C/c polymorphisms. Vox Sang 1994;67:203-9. 13. Avent ND, Ridgwell K, Tanner MJ, Anstee DJ. cDNA cloning of a 30 kDa erythrocyte membrane protein associated with Rh (Rhesus)-blood-group-antigen expression. Biochem J 1990;271:821-5. 14. Flegel WA. Molecular genetics and clinical applications for RH. Transfus Apher 2011;44:81-91. 15. Daniels G. Human blood groups. 2nd ed. Oxford: Blackwell Science; 2002. 16. Wagner FF, Frohmajer A, Ladewig B, Eicher NI, Lonicer CB, Muller TH, Siegel MH, Flegel WA. Weak D alleles express distinct phenotypes. Blood 2000;95:2699-708. 17. Wagner FF, Gassner C, Muller TH, Schonitzer D, Schunter F, Flegel WA. Molecular basis of weak D phenotypes. Blood 1999;93:385-93. 18. McGann H, Wenk RE. Alloimmunization to the D antigen by a patient with weak D type 21. Immunohematology 2010;26:27-9. 19. Denomme GA, Wagner FF, Fernandes BJ, Li W, Flegel WA. Partial D, weak D types, and novel RHD alleles among 33,864 multiethnic patients: implications for anti-D alloimmunization and prevention. Transfusion 2005;45:1554- 60. 20. Wagner T, Kormoczi GF, Buchta C, Vadon M, Lanzer G, Mayr WR, Legler TJ. Anti-D immunization by DEL red blood cells. Transfusion 2005;45:520-6. 21. Singleton BK, Green CA, Avent ND, Martin PG, Smart E, Daka A, Narter-Olaga EG, Hawthorne LM, Daniels G. The presence of an RHD pseudogene containing a 37 base pair duplication and a nonsense mutation in Africans with the Rh D-negative blood group phenotype. Blood 2000;95:12-8. 22. Colin Y, Cherif-Zahar B, Le Van Kim C, Raynal V, Van Huffel V, Cartron JP. Genetic basis of the RhD-positive and RhD-negative blood group polymorphism as determined by Southern analysis. Blood 1991;78:2747-52. 23. Faas BH, Beckers EA, Wildoer P, Ligthart PC, Overbeeke MA, Zondervan HA, von dem Borne AE, van der Schoot CE. Molecular background of VS and weak C expression in blacks. Transfusion 1997;37:38-44. 24. Reid ME, Lomas-Francis C. The blood group antigen facts book. 2nd ed. San Diego (CA): Academic Press; 2003. 25. Veldhuisen B, van der Schoot CE, de Haas M. Blood Group genotyping: from patient to high-throughput donor screening. Vox Sang 2009;97:198-206. 26. Avent ND. Large scale blood group genotyping. Transfus Clin Biol 2007;14:10-5.

74 48. 47. 46. 45. 44. 43. 42. 41. 40. 39. 38. 37. 36. 35. 34. 33. 32. 31. 30. 29. 28. 27. Avent ND, Finning KM,Liu W, Dphenotypes. biology ofpartial ML.Molecular Scott Transfus ClinBiol1996;3:511-6. Br JHaematol 1996;92:751-7. resultsphenotype from substitutionofexon 5 of theRHCEgeneby thecorresponding exon oftheRHD gene. EA,FaasBeckers MA, DJ, vanRhenen CE.The R0HarRH:33 vanderSchoot BH,vonAE,Overbeeke demBorne origin: identificationandtransfusionsafety. Blood2002;100:4223-31. P,C, Rouger JP, R,Cartron Kotb Ansart-PirenneinblackindividualsofAfro-Caribbean RHCEphenotypes H.Rare Noizat-Pirenne F, Pennec Lee K, PY, SimonP, P, Kazup BachirD, AM, M,Juszczak Rouzaud Roussel G,Menanteau 1999;94:4337-42. frequently variant found Rhce withceAR, anew blacks. inAfrican Blood exons inlinkage 4,5,and7,often DJ, PC,Berger L,vanRhenen CE, vander Schoot MB,Hemker Ligthart Wijk PA. involving variant DAR, RhD anew Transfusion 2000;40:936-42. bloodsamplesofpatientswhohave by analysisofperipheral genotypes recently received multipletransfusions. Rozman P, Dovc T, C.Differentiation Gassner ofautologous ABO, RHD, andFY blood group RHCE,KEL,JK, Wagner FF, Frohmajer A,Flegel WA. inDnegative RHDpositive Europeans. haplotypes 2001;2:10. BMCGenet 1996;87:2968-73. using retroviral blood group ofK562cellsestablishesthemolecularbasisRh antigens. transduction Blood Smythe JS, Avent ND, Judson PA, Parsons SF, PG, Anstee DJ. Expression Martin of RHD and RHCE gene products protein RhD vestibule.acidsubstitutions attheextracellular Transfusion 2008;48:25-33. Flegel WA, A, von I, Doescher Zabern Wagner FF, J,Vytiskova M. DCS-1, DCS-2, and Pisacka DFV share amino serology. 1A:Rh M.Section Scott Coordinator’s report. Transfus ClinBiol2002;9:23-9. monoclonal anti-D. Transfus 1993;3:67-9. Med Lomas K, C,McColl Tippett P. DIIcellswith Further antigenDdisclosedby testing complexitiesoftheRh category Transfusion 1998;38:951-8. CE.The van derSchoot and VS V bloodgroup aserologic in Africans: polymorphisms andmolecularanalysis. Daniels GL,Faas E,Maaskant-Van BH,Green CA,Smart Wijk PA, Avent ND, HA,vonAE, demBorne Zondervan cell-free fetal DNAfrom plasma.Obstet maternal 2004;103:157-64. Gynecol RJ, GC,BossersB,Rijnders Christiaens vanderSmagtJJ, CE,deHaasM.Clinicalapplicationsof vanderSchoot 1998;62:768-75. Am JHumGenet Quantitative analysis of fetal DNA in maternal plasma and : implications for noninvasive prenatal diagnosis. Lo YM, Tein MS,Lau HainesCJ, LeungTK, TN, Poon PM, Wainscoat PJ, JS,Johnson NM. ChangAM, Hjelm regionusing junctional specific probes.TaqMan 1998;12:2006-14. Leukemia JJ. quantitative Real-time PCRfor ofminimalresidual thedetection diseaseinacute lymphoblasticleukemia Pongers-Willemse MJ, Verhagen OJ, Tibbe GJ, deHaas AJ, Wijkhuijs V, Roovers CE,vanDongen E,vanderSchoot 1998;38:1015-21. exons. ofRHDby multiplex polymerase chain reactionanalysisofsixRHD-specific CE. Genotyping Transfusion Maaskant-VanWijk PA, Faas MA, DJ, vonAE,vanRhenen van derSchoot demBorne BH,deRuijter JA,Overbeeke Wagner FF, JM,Flegel Moulds WA. box Transfusion variation. mechanismsofRhesus Genetic 2005;45:338-44. Wagner FF, Flegel WA. box. RHDgenedeletionoccurred Blood2000;95:3662-8. in theRhesus events.2005;45:327-37. Transfusion butyieldsmore determination insightinto RH-relatednonwhite evolutionary personshampersRHDzygosity Grootkerk-Tax MG,Maaskant-VanWijk PA, vanDrunenJ, RHlocusin CE.The highlyvariable vanderSchoot 2007;47:715-22. of four an evaluation and RHD mosaicism/chimerism: quantitative methods. Transfusionof RHD zygosity GR,ClausenFB,Krog Dziegiel MH.Quantitation of RHDby real-time polymerasechainreactionfor determination of hemolyticdiseasethefetus related andnewborn to anti-D. Prenat Diagn 2010;30:1207-12. Pirelli Pietz predictingrisk ST, BC,Johnson KJ, Pinder HL,BellissimoDB. ofRHDzygosity: determination Molecular 2011;118:1340-8. D,of rhesus c, EandofKinalloimmunisedpregnant women: ofa7-year evaluation clinicalexperience. BJOG GC, deHaas M.NoninvasiveSchefferSchoot CE, PG,van der fetal blood group Page-Christiaens genotyping noninvasive Transfusiona new fetal service. RHDgenotyping 2002;42:1079-85. PW, PG,Soothill Finning Avent KM,Martin ND. Prediction offetal Dstatusfrom plasma:introduction of maternal variants andzygosity ofRHDvariants Genotyping 75

Chapter 3 Chapter 3

76 Chapter 3 information Supporting variants andzygosity ofRHDvariants Genotyping 77

Chapter 3 Chapter 3

Supplementary Figure S1. Example of analysis of mix P401. a) The blue peaks indicate the signals from the analyzed sample and red peaks indicate signals from a reference sample (CcDdEe genotype). The black arrows indicate the control-probe-combination signals and are used for normalization of both the analyzed and reference sample. The signals of the RHD-specific-probe combinations of the analyzed sample (green arrows) show the same fluorescence intensity as the signals of the reference samples, indicating that the tested sample has the same RHD copy number as the reference sample (Dd). In contrast, the signal obtained for the RHCE*e specific-probe combination (light blue arrow) shows twice the amount of fluorescence intensity compared to the reference sample, therefore the copy number of RHCE*e is double of that obtained for the reference sample, therefore the copy number of RHCE*e is 2n in the analyzed sample. The RHCE*C and RHCE*E specific-probe-combination signals (dark blue arrows) are absent in the analyzed sample, while the reference sample shows signals, indicating that these alleles are absent in the analyzed sample. The red arrow indicates the presence of the MD_819A signal in the analyzed sample and thus the presence of the RHD*09.03, RHD*09.04 or RHD*09.05 allele. b) The results are shown in the fourth column as ratios of the signals obtained for the analyzed sample compared to signals obtained for the reference sample. The Bin Size indicates the length of the probe. Interpretation of the results is the same as in a).

78 Chapter 4

Characterization of known and novel RHD variant alleles in 37,764 Dutch D- pregnant women

Lonneke Haer-Wigman1* Tamara C. Stegmann1* Florentine F. Thurik1 Renate Bijman1 Bernadette Bossers2 Goedele Cheroutre2 Remco Jonkers2 Peter Ligthart2 Barbera Veldhuisen1,2 Masja de Haas1,2 C. Ellen van der Schoot1

1 Sanquin Research, Amsterdam and Landsteiner Laboratory, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands 2 Sanquin Diagnostic Services, Amsterdam, The Netherlands

Manuscript in preparation Chapter 4

Abstract Background: In the Netherlands D- pregnant women are offered a quantitative fetal RHD genotyping assay to determine the RHD status of their fetus to guide anti-D prophylaxis. We characterized maternal RHD variant alleles recognized in this genotyping assay.

Methods: In 1.3% of the 37,764 D- pregnant women who were tested with the quantitative fetal RHD genotyping assay, targeting exon 5 and 7, a maternal variant allele was suspected and serological and molecular analysis was performed.

Results: In 0.96% (95% CI 0.86% - 1.05%) of the Dutch D- pregnant women a variant allele is present and 47% of these women carry the RHD*Ψ allele. Forty-five different RHD variant alleles were detected, including fourteen novel alleles. Eleven novel D- null alleles were identified, including one allele with a single missense mutation (RHD*443C>G; p.Thr148Arg) and one allele with a single amino acid deletion (RHD*424_426del; p.Met142del), both located in the transmembrane region of the RhD protein. The RHD*721A>C (found in seven individuals) and RHD*884T>C variant alleles cause Del expression and for the novel RHD*[178A>C; 689G>T] allele we postulate a partial weak D expression. In a single pregnant woman with an initial weak D typing an additional novel RHD variant allele, RHD*492C>A, was detected, causing partial weak D expression. The phenotypes of the RHD*443C>G, RHD*492C>A and RHD*721A>C mutation were confirmed with a heterologous expression study.

Conclusions: Fifteen new RHD variant alleles were identified and we determine for the first time that a single amino acid change or deletion can cause the D- phenotype.

80 caused by hybrid alleles or due to mutations in the extracellular parts of the RhD protein. oftheRhD caused by parts hybrid allelesordueto mutationsintheextracellular prophylaxis pregnant isadministrated to D- women. red blood cells are individualscompatibleD- anti-D formationtransfusedand inD- transfusion reactionsand/orsevere hemolyticdiseaseofthefetus andnewborn. a splicesite. the RhD protein.the RhD in low caused by quantitiesthatismostoften mutationsinthetransmembraneregions of a quantitative fetal RHD pregnantamong Dutch D- pregnant women. Since July2011,Dutch D- women are offered andnovel) RHD The of(known aimofourstudywasto thefrequency determine blood group antigens. bloodgroupThe DantigenoftheRh system isoneofthemostimmunogenicandcomplex Introduction transfusion orpregnancy, anti-Dcanbeproduced. The Rh is highly polymorphic andmany alleleshave locusishighlypolymorphic RHDvariant The beendescribed. Rh D different that individuals with because it is unlikely weak D or alleles is of importance, variant other D- nullalleles are rare.other D- adsorption-elution technique. adsorption-elution protein sensitive ontheirred bloodcellmembrane, whichcanonlybedetected withthevery high immunogenicity. occur in the Black population; in 66 % and 15% of D- Blackindividualsrespectively.occur intheBlackpopulation;66%and15%ofD- nullallelesthatfrequentlya nonsensemutationandthehybrid RHD*03N.01alleleare D- Resources). that disruptasplicesite orto large hybrid alleles(Web The RDH*Ψallelewith dueto nonsensemutations,phenotype to to mutationsleadingto mutations aframeshift, To date, more nullalleleshave than50,socalled, thatcausetheD- beendescribed D- at risk ofDimmunization. at risk RHD gene. counterparts. arose oftheRHDgeneare dueto replaced geneticrecombination by inwhichpart RHCE configuration onchromosome 1. locusconsistsofthehighlyhomologousRHD The Rh the D antigen, the so called partial weakalleles. RHDvariant the Dantigen,socalledpartial Often or more Depitopes. expression protein inwhichtheRhD expression lacksone oftheDantigenorsocalledpartial fetus. RHD el expression produce alloanti-D, Dexpression who are incontrastto individualswithpartial genotyping assay is performed on maternal plasma,whichcontainscellfree onmaternal DNAofthe assay isperformed genotyping 22,23 variant alleles causing partial Dexpression expression alsocauseweakened of allelescausingpartial RHD variant ofcellfree origin DNA isofmaternal andthereforeThe farmajority thepresence of 13

11 RHD Partial Dexpression, inwhichoneormore Depitopes are ismostoften lacking, 13,14 variant alleles carry oneormultiplemutationsinthe Anothergroup allelescarry ofRHDvariant variant alleles can cause the complete absence of the RhD protein, alleles can causethecompletevariant absenceoftheRhD weakened 18 Individuals withD Individuals 10,11 3

1,2 withtheDantigenvia individualcomesinto contact When aD- genotyping assay togenotyping guide anti-D prophylaxis. This quantitative D- individuals lack protein,the complete RhD D- which explains its 21

17 Individuals with a weak D phenotype express proteinwith aweak D phenotype the RhD Individuals 19 D el 12 expression causedby ismostoften mutations thatdisrupt Onegroup alleles, variant theRHDhybrid alleles, ofRh el expression have aneven lower amountoftheRhD 8,9 4,5 and

Anti-Dcancausesevere hemolytic genes located in a tail-to-tail RHCE geneslocated inatail-to-tail 10,11 the between The distinction variant alleles in RhD- individuals allelesinRhD- RHD variant variant alleles variant 6,7 Toprevent 15-17 fetal All 10,11 81 20

Chapter 4 Chapter 4 a maternal RHD variant allele will result in much stronger signals in the quantitative PCR assay than expected to arise from fetal DNA. In this paper we present the serological and genetic follow up of cases identified among almost 38,000 screened Dutch D- pregnant women.

Material and methods Samples and fetal RHD genotyping assay Between July 2011 and December 2012 37,764 Dutch D- pregnant women were tested in the 27th week of pregnancy for the presence of a D+ fetus using a quantitative fetal RHD genotyping assay based on TaqMan chemistry. DNA was isolated from 1 mL maternal plasma using a DNA isolation kit (DNA and Viral NA Large Volume Kit; Roche Holding AG) on a MagNa Pure 96 Instrument (Roche) via manufacturer’s protocol. The quantitative fetal RHD genotyping assay consist of two TaqMan tests, one targeting RHD exon 5 and one targeting RHD exon 7, which are performed in triplicate. This quantitative fetal RHD genotyping assay is extensively described by Scheffer et al.22 When at least two of the three Ct values of both assays were below 32 a maternal variant allele was suspected. When at least two of the three Ct values of exon 7 was below 32, while Ct values of exon 5 were above 32 a RHD*Ψ or RHD*06 maternal variant allele was suspected. In both cases material was stored and additional genotyping and extended serology was performed to determine whether a variant allele and which variant allele was present. Statistical analysis was performed using a 95% confidence interval [95% CI].

Serology Standard serology to determine the D phenotype in a pregnant woman was performed using two anti-D antibodies. A monoclonal anti-D reagent (immunoglobulin [Ig]M clone RUM-1, Sanquin Reagents) and a monoclonal blend reagent (IgM clone TM28 and IgG clone MS26, Sanquin Reagents) were used in a method with an immediate spin at room temperature. In all samples in which no agglutination was detected an indirect antiglobulin using the monoclonal blend reagent was performed. Plasmas of all women negative in this serological assay were tested in the quantitative fetal RHD genotyping assay. In women in whom a maternal variant allele was suspected, because Ct values were below 32 in this assay, a second serological test was performed. This second serological test consists of three monoclonal IgM anti-D (RUM-1, BS226 (Bio-Rad Laboratories B.V.) and ESD1-M (ALBA Bioscience)) that are tested in a method with an immediate spin at room temperature and in a method with a spin after 15 minutes incubation at room temperature and three monoclonal blend reagents (IgM clone TH28 and IgG clone MS26 (Sanquin reagents and ImmucorGamma) and IgM clone D7B8, IgG clone H112196 and IgG clone LORIFA (Ortho)), one monoclonal IgG (5C8) and a polyclonal IgG reagent (Bio-Rad Laboratories B.V.) that were tested in a method with an immediate spin at room temperature, in a method with 15 minutes incubation at 37ºC and a spin and/or in the indirect antiglobulin test. When also in this second test the D- phenotype was determined

82 followed by 20 minutes at 72°C. of1.5 A mixture μL MLPA sample, 8.5 μL Hi-Di of30secondsat95°C,60Cand1minute5 minutesat72°C, at72°C,35cycles sample by adding2μLuniversaland0.5SALSApolymerase. primers PCRconditionswere: onthecomplete ligation minutes at54°C.Apolymerasechainreaction[PCR]wasperformed μL SALSAligasebuffer A and1.5μLSALSAligasebuffer B were addedandincubated for15 hybridization 1.5 of the probe combinations to genomic DNA at 60°C, 1 μL SALSA ligase-65, 1.5 μLprobe mixand1.5μLSALSAMLPA dilutionbuffer were added. 16-20hoursof After 5 μL containing 50-100 ng of DNA was denatured and (Appliedshort, Biosystems). In The MLPA viamanufacturer’s protocol reactionwasperformed ona Veriti Thermocycler RHCE copy numbersofthe5’ UTR,exon 1,2and8intron 1and2. and c.1112G positionsofRHDandRHCE,developed to thecombinedRHDand determine combinations, targeting thec.-698T, c.123A, c.149-4875A, c.149-882G, c.244T, c.335+2838C each primer setrangingeach primer from 61to 68ºC,followed by 1minute at72°C.PCRproductswere of10 secondsat95°Candaspecificannealing/elongation temperaturecycles andtime for Biosystems), andreverse 0.5μMforward primer. PCRconditions were: 10secondsat 95°C,35 Fastof 20μL,containing50-150 ng DNA,10μLof2xGeneAmp (Applied Mix PCRMaster synthesized by Eurogentec. ona The PCRwasperformed Veriti inatotal volume thermocycler promoter region ofRHDwassequenced(hg19,chr.1:g.25597899_25598887). wereThe primers somecasesallexons andintronIn ofRHDwere boundaries indicated sequenced. alsothe If DNA sequencing version 1.85(Softgenetics). software Genemarker using analyzed Analyzer ona3130Genetic (Applied Biosystems). Dataanalysiswasperformed (Applied Biosystems)500-Liz Size and 0.5 μL GeneScan Standard (Applied Biosystems) was RHCE-MLPA was performed. Probe Amplification(RH-MLPA) indicated, If assay the (mixp401,p402andp403,MRC-Holland). the pregnant women, DNAsampleswere analyzed Ligation-dependent withtheRH-Multiplex Blood Mini Kit. Kit. Blood Mini The copy numberofthe DNA (QIAamp kit DNAwasisolated fromMaternal white bloodcellsusingaDNA extraction RH-MLPA Partial RhD Typing Laboratories. ofBio-Rad Set monoclonal IgGantibodiesthattarget protein allepitopes oftheRhD and/ortheExtended consistingofeleven kit third typing serological usinganin-houseRhD test wasperformed a variant, otherthantheDVI wassuspected Dvariant in thesecondserological test apartial and whenthepresence oftheRHD*Ψ anti-D (Bio-Rad Laboratories B.V.)anti-D (Bio-Rad weak expression to of the a very D detect ELU-KITusing theGamma viamanufactures’ II(ImmucorGamma) protocol usingapolyclonal 24 Furthermore, one cases was tested with seven MLPA new probe allele was excluded, absorption-elution was performed wasperformed allele wasexcluded, absorption-elution RHD gene and the presence of RHD variant alleles in RhD- individuals allelesinRhD- RHD variant el variant alleles in variant variant allele. If allele. variant If T Formamide 83

Chapter 4 Chapter 4 purified using ExoSAP-IT (GE Healthcare), according to manufacturer’s protocol. The sequence reaction was performed on a Veriti thermocycler (Applied Biosystems) in a total volume of 20 μL, containing 1 μL of purified PCR product, 1 μL 2,5x BigDye Terminator v1.1 Cycle (Applied Biosystems), 3.5 μl 5x BigDye Terminator Buffer (Applied Biosystems) and 0.25 μM forward or reverse primer. Sequence conditions: 25 cycles of 15 seconds at 95°C, 10 seconds at 50°C and 4 minutes at 60°C. Sequence products were analyzed on a 3130 Genetic Analyzer (Applied Biosystems).

Heterologous expression system As a wild-type RHD construct the RHD coding sequence flanked by a BAMHI and NOTI digestion site was ordered at Invitrogen (Breda, The Netherlands) and cloned into the pHeftig vector, encoding GFP as a transduction control. The RHD*443G, RHD*492A, RHD*721C or RHD*1154C mutations were mutated into the wild-type RHD construct using QuickChange II XL Site-Directed Mutagenesis Kit (Agilent) via manufactures’ protocol. K562 cells were lentivirally transfected with the RHD wild type construct, different variant constructs and a mock construct. Forty-eight hours after transduction, cells were harvested and screened for RhD expression by flow cytometry using nine monoclonal IgG anti-D of the ALBAclone Advanced partial RhD typing kit (ALBA Bioscience), namely LHM169/81, LHM76/59, LHM76/55, LHM50/28, LHM169/80, LHM57/17, LHM76/58, LHM59/19 and LHM77/664 and an additional IgG anti-Rh29 that can detect very weak D expression levels (BRIC69). Data analysis was performed with FlowJo Version 8 software (TreeStar).

Results In 0.96% of the Dutch D- pregnant women a variant RHD allele is present All Dutch pregnant women are typed for their RhD status using a monoclonal IgM and monoclonal blend anti-D. Since July 2011, women who are typed D- in this initial test are offered a quantitative fetal RHD genotyping assay to test the RHD status of their fetus to determine whether they need to receive anti-D prophylaxis. The quantitative fetal RHD genotyping assay is performed on DNA isolated from maternal plasma and targets exon 5 and 7 of RHD. Because cell free fetal DNA is present in an overwhelming background of maternal cell free DNA, maternal variant alleles disturb the quantitative fetal RHD genotyping assay. Of note, the frequently occurring RHD*03N.01 D- null allele is not amplified in this RHD genotyping assay and is therefore not detected. To determine the range of Ct values that corresponds with a maternal allele, the fetal RHD genotyping assay was performed on 1000 D+ pregnant women, none of the women had Ct values of both exon 5 and 7 above 32 (data not shown). Hence, the cut-off value in which a maternal variant allele was suspected was set at a Ct value of below 32.

84 performed, becausethe D performed, identified a specific were theRH-MLPA 227ofthe275casesonwhomgenotyping performed, In assay genotyping wasdetermined. phenotype 667T>C; 819G>A;919G>A] allele that causes the D therecentlywas detected andthese three RHD*[602C>G; described individualscarried mutations) thatwasidentified by theRH-MLPA. allcasesanadditionalc.919G>A mutation In oftheRHD*09.03allele(withc.[602C>G; Dphenotype 667T>C;819G>A] weak partial 42 (16%)casesaD serological wasconfirmed, follow-up. however, phenotype 145(55%)casestheinitialD- In in allelered the262casesof275withavariant blood cellswereIn available for extensive Women withstandard determined serology D- carried45different variant alleles theRHD*Ψallele. women allelecarries withavariant pregnantD- womenallelethathasRHD avariant present (Supplementary Table conclusion,in0.96%(95%CI0.86% -1.05%)oftheDutch S1).In alleleis estimate aRHDvariant thatin~86ofthe183caseswhichnofollow-up isperformed of RHD allelesin cases in which follow-up was used to calculate was performed, the distribution theremainingconcentrations. 275womenallelewasidentified. In avariant of The distribution thesecasestheCtvaluesarosepresence from oftheRHD*01N.01allele).In highfetal RHD allelewasidentified,variant butthedeletionofRHDgenewasdetected (homozygous pregnant 39(12.4%)ofthe314D- women no inwhichgeneticfollow-upIn wasperformed RHD*01N.01 allele. the presence ofeitheronethesealleles. Except for for the onecasethatwashomozygous confirmed in whichthepresence genotyping ofaRHD*ΨorRHD*06allelewassuspected amplification of only exon 7) no because in all other casesfollow-up (n = 169) was performed, Furthermore, (maternal in54caseswhichaRHD*ΨorRHD*06allelewassuspected 129ofthe497women wasstored nomaterial In andnofollow-up couldbeperformed. onDNAisolated fromfollow-up white bloodcells. alleleswasperformed variant ofmaternal because Ct values of one or both exons were 314 of the 497 women below genetic 32. In 497 (1.3%)women allele was suspected, variant womena maternal 2011 and 2012. In between The quantitative fetal RHD the D of thisnovel alleleandnoadditionalmutationswere variant three detected. caseswith In thepresencemutations ofboththeRHD*03.03 andRHD*09.03allele. confirmed Sequencing allele, onecasetheRH-MLPA In wasperformed. additional genotyping identifiedanovel variant alleles in cases in which no follow-up was performed (Supplementary (Supplementary RHD allelesincaseswhichnofollow-upTable wasperformed S1). We el RHD*[361T>A; 380T>C; 383A>G;455A>C;602C>G; 667T>G; 819G>A],thatconsistsofthe phenotype the phenotype el phenotype, in 68 (26%) cases a partial Dandinseven (3%)casesaweak D in68(26%)casesapartial phenotype, RHD RH-MLPA detected an allele in which variant allele. In three allele.variant of the In 227 cases, additional sequencing was el genotyping assay was performed on a total of 37,764 D- pregnant onatotal of37,764D- assay wasperformed genotyping phenotype ofthese cases didnot correspond withthe described phenotype exon 5and/or7ispresent and47%ofthe el phenotype. RHD exon 10 was absent. In 25 In the remaining 48 In cases variant alleles in RhD- individuals allelesinRhD- RHD variant 85

Chapter 4 Chapter 4 these cases the RHD*(1-9) or the RHD*(1-9)-CE(10) allele can be present, because both alleles have been described to cause the Del phenotype.25-27 The RHCE-MLPA detected two copies of RHCE exon 10 and we therefore conclude that these cases carried the RHD*(1-9) allele. In another case in which a RHD-RHCE-RHD allele was suspected the RHCE-MLPA and seven newly developed MLPA probe combinations (targeting the combined RHD and RHCE copy number of the 5’UTR, exon 1, 2 and 8 and intron 1 and 2) were performed. Taking the results of all MLPA probe combinations together in combination with the D- phenotype and RhCcee phenotype of this case, we postulate that this case most likely carries the known RHD*D-CE(2-9)-D allele (which is described to be linked with RhCe expression30) and a novel allele in which RHD exon 1 is deleted (RHD*(2-10)) (Supplementary Figure 1). In 43 cases additional sequencing was performed because the RH-MLPA was either indicating the presence of a normal RHD gene (n = 38) or was not able to identify the specific variant allele (n = 5). In twenty cases twelve novel RHD variant alleles were identified (Table 1). In one of the cases with a novel variant allele we detected the heterozygous presence of the c.178C and c.1136T mutation and the homozygous presence of the c.689T mutation. We presume that in this case, with partial D expression, the known RHD*10.01 and a novel RHD*[178A>C; 689G>T] allele are present, instead of two novel alleles (RHD*689C>T and RHD*[178A>C; 689G>T; 1136C>T]). All novel alleles were detected in single cases, except for the RHD*721A>C, allele, which was detected in seven cases and the RHD*1074-1G>A allele, which was detected in three cases (Table 1). A total of 45 different RHD variant alleles, including fourteen novel RHD variant alleles, were identified in the 275 cases, as listed in Table 1. Remarkably, in two cases, both with the Del and normal RhCE expression (RhCcee phenotype), an apparently normal wild-type RHD allele without mutations was detected. Also no mutations were detected in the promoter region of the RHD gene.

Characterization of fifteen novel RHD variant alleles Extended serological typing was performed in the cases carrying one of the fourteen novel RHD variant alleles. Furthermore, we analyzed another case carrying a novel variant allele (RHD*492C>A encoding p.Asp164Glu) discovered in a pregnant woman who had weak D expression in the initial D typing (Table 2). Eleven of the fifteen novel RHD variant alleles were determined to cause the D- phenotype by absorption-elution: the RHD*1084C>T allele (encoding p.Gln362Ter) with a nonsense mutation, two alleles (RHD*125_125delAA and RHD*1174delA) with frame shift mutations and the allele (RHD*(2-10)) in which RHD exon 1 was deleted and four alleles with mutations that disrupt a splice site (RHD*335G>T, RHD*[634+1G>T, 1136C>T], RHD*1073+1G>T or RHD*1074-1G>A) (Table 2). The novel allele that contained mutations of both the RHD*03.03 and RHD*09.03 variant alleles also causes the D- phenotype (Table 2).

86 Table withstandard allelesdetected in267women serology 1.RHDvariant D- determined D- phenotype RhD Partial RhD RHD*1174del§ RHD*1084C>T§ RHD*1074-1G>A§ RHD*1073+1G>T§ § RHD*[634+1T;1136T] 602C>G; 667T>G;819G>A]§ RHD*[361T>A; 380T>C;383A>G;455A>C; RHD*443C>G§ RHD*424_426del§ RHD*335G>T§ RHD*125_126del§ RHD*(2-10)§ RHD*01EL.09† RHD*01EL.08† RHD*01EL.05† RHD*952C>T RHD*922G>T RHD*660delG RHD*Ψ RHD*Ψ RHD*Ψ RHD*Ψ RHD allele1 RHD*04.02 § RHD*[178C; 689T] RHD*17.02 RHD*15 RHD*11 RHD*10.02 RHD*09.01 RHD*09.01 RHD*06.02 RHD*06.02 RHD*06.01 RHD*05.07 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01-CE(2-9)-D RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*Ψ RHD*03N.01 RHD*01N.03 RHD*01N.01 RHD allele2 RHD*01N.01 RHD*10.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 RHD*03N.01 RHD*01N.01 RHD*03N.01 RHD*01N.01 RHD*01N.01 RHD*01N.01 variant alleles in RhD- individuals allelesinRhD- RHD variant for genotype cases positive Number of 102 13 11 12 40 1 1 3 1 1 1 1 1 1 1 1 1 3 1 3 2 1 6 1 1 1 9 2 1 2 1 6 1 87

Chapter 4 Chapter 4

Table 1. (Continued)

Number of RhD phenotype RHD allele 1 RHD allele 2 cases positive for genotype Del RHD*01EL.01 RHD*01N.01 9 RHD*01N.22‡ RHD*01N.01 1 RHD*[602C>G; 667T>G; 819G>A; RHD*01N.01 2 919G>A] RHD*[602C>G; 667T>G; 819G>A; RHD*[602C>G; 667T>G; 819G>A; 1 919G>A] 919G>A] RHD*93_94insT RHD*01N.01 6 RHD*(1-9) RHD*01N.01 2 RHD*(1-9) RHD*Ψ 1 RHD*1252_1253insT RHD*01N.01 1 RHD*721A>C§ RHD*01N.01 7 RHD*884T>C§ RHD*01N.01 1 RHD*01 RHD*01N.01 2 Weak RhD RHD*01W.01 RHD*01N.01 3

RHD*01W.02 RHD*01N.01 3 RHD*01W.05 RHD*01N.01 1 RHD*01W.22 RHD*01N.01 1 RHD*01W.38 RHD*01N.01 1

† In literature it has been described that the RHD*01EL.05, RHD*01EL.08 and RHD*01EL.09 alleles cause the Del phenotype, we detected, however, in all cases carrying either one of these alleles the D- phenotype ‡ In literature it has been described that the RHD*01N.22 allele cause the D- phenotype, we detected, however, the Del phenotype in the case positive for this allele § Novel variant allele

All mutations of this variant allele (c.[361T>A; 380T>C; 383A>G; 455A>C; 602C>G; 667T>G; 819G>A]) are located in the putative transmembrane or intra-cellular parts of the RhD protein (Figure 1). Unexpectedly, two novel variant alleles that caused the D- phenotype contained mutations that resulted in a single amino acid substitution or a single amino acid deletion in the transmembrane region of the RhD protein (Figure 1): the RHD*443C>G (encoding p.Thr148Arg) and RHD*424_426del (encoding p.Met142del) alleles (Table 2). The RHD*443C>G allele was already described by us in a single individual in a previous study24 and in this study we confirm the D- phenotype of the RHD*443C>G allele in another individual. Two novel alleles cause partial D expression, RHD*[178A>C; 689G>T] (encoding p.[Ile60Leu; Ser230Ile]) and RHD*492C>A (encoding p.Asp164Glu) with mutations in the fourth or third extracellular loop of the RhD protein, respectively (Figure 1). We were unable to determine the exact phenotype for the RHD*[178A>C; 689G>T] allele, because the case positive for this allele carried another variant allele, the RHD*10.01 allele. The epitope pattern in this case,

88 Table 2. Fifteen novel variant alleles

Linked Nucleotide Exon Position in RhD Initial Extended Adsorption- RhD Allele name Protein changes‡ RHCE Changes† (Intron) protein serology serology Elution phenotype genotype RHD*1084C>T c.1084C>T 8 p.Gln362Ter Negative Negative Negative D- RHCE*02 RHD*125_126del c.125_126delAA 1 p.Lys42fs Negative Negative Negative D- RHCE*02 RHD*1174del c.1174delA 9 p.Ile392fs Negative Negative Negative D- RHCE*02 RHD*(2-10) c.1-?_148+?del 1 p.0? Negative Negative Negative D- RHCE*01 RHD*335G>T c.335G>T 2 r.spl? Negative Negative Negative D- RHCE*01 RHD*[634+1G>T; c.[634+1G>T; (4) and 8 r.spl? Negative Negative Negative D- RHCE*01 1136T] 1136C>T] RHD*1073+1G>T c.1073+1G>T (7) r.spl? Negative Negative Negative D- RHCE*03 RHD*1074-1G>A c.1074-1G>A (7) r.spl? Negative Negative Negative D- RHCE*02 RHD*[361T>A; c.[361T>A; 380T>C; p.[Leu121Met; Val127Ala; 6 mutations in 380T>C; 383A>G; 383A>G; 455A>C; 3, 4, 5 Asp128Gly; Asn152Thr; transmembrane region Negative Negative Negative D- RHCE*01 455A>C; 602C>G; 602C>G; 667T>G; and 6 Thr201Arg, Phe223Leu; and one mutation in 667T>G; 819G>A] 819G>A] Ala273Ala] intracellulair region RHD*443C>G c.443C>G 4 p.Thr148Arg Transmembrane Negative Negative Negative D- RHCE*02 RHD*424_426del c.424_426delATG 3 p.Met142del Transmembrane Negative Negative Negative D- RHCE*02 RHD*721A>C c.721A>C 5 p.Thr241Pro Transmembrane Negative Negative Positive Del RHCE*01 variant alleles in RhD- individuals allelesinRhD- RHD variant RHD*884T>C c.884T>C 6 p.Met295Thr Transmembrane Negative Negative Positive Del RHCE*02 Transmembrane and RHD*[178A>C; Partial c.[178A>C; 689G>T] 2 and 5 p.[Ile60Leu; Ser230Ile] fourth extracellulair Negative Partial§ - RHCE*01 689G>T] weak D loop Partial RHD*492C>A c.492C>A 4 p.Asp164Glu Third extracellulair loop Weak Partial - RHCE*02 weak D

† Position as counted from ATG translation start site; homo = homozygous, hetero = heterozygous ‡ Position as counted from Met translation start site § The case positive for this variant allele carried a second variant allele RHD*10.01 that causes partial D expression, therefore 89 we cannot exclude that this variant allele can also cause the D- phenotype

Chapter 4 Chapter 4

Legend Single amino acid changes of the RHD*424_426del allele encoding p.Met142del, the RHD*443C>G allele encoding p.Thr148Arg or the RHD*1084C>T allele encoding p.Gln362Ter Amino acid changes of the RHD*[361T>A; 380T>C; 383A>G; 455A>C; 602C>G; 667T>G; 819G>A] allele encoding p.[Leu121Met; Val127Ala; Asp128Gly; Asn152Thr; Thr201Arg, Phe223Leu; Ala273Ala] Single amino acid changes of the RHD*721A>C allele encoding p.Thr241Proor the RHD*884T>C allele encoding p.Met295Thr Single amino acid change of the RHD*492C>A allele encoding p.Asp164Glu Amino acid changes of the RHD*[178C; 689T] allele encoding p.[Ile60Leu; Ser230Ile]

Extracellular region 323232404040 989898 109109109 154154154 167167167 230230230239239239 285285285293293293 347347347 358358358

Transmembrane region

777 747474777777 133133133136136136 188188188204204204 261261261 266266266 318318318324324324 391391391 Intracellular region

CCC III III NNN

Figure 1. The position of amino acid changes in the RhD protein present in seven different RHD variant alleles. Schematic representation of a two dimensional model of the RhD protein. Each dot represent an amino acid, dots within the black lines indicate putative membrane regions of RhD with residue numbers. Colored dots represent position of amino acids that are mutated in the RHD*424_426del, RHD*443C>G, RHD*1084C>T, RHD*[361T>A; 380T>C; 383A>G; 455A>C; 602C>G; 667T>G; 819G>A], RHD*721A>C, RHD*884T>C, RHD*492C>A or RHD*[178A>C; 689G>T] allele. loss of epitopes 1 (LHM70), 5 (rD7C2) and 8 (HIMA-36), corresponds to the epitope pattern described for the RHD*10.01 allele.28 We, however, postulate that for both variant alleles of this case the phenotype is caused by the c.689G>T mutation. Because, the additional mutation in the RHD*10.01 allele (c.1136C>T) has no effect on RhD expression in carriers of theRHD*10.00 allele28 and we assume that the additional mutation of the novel allele (c.178A>C) present in the transmembrane region of the RhD protein (Figure 1) also has no effect on RhD expression. The RHD*492C>A allele gave partial D expression, epitope 5 (rD7C2) and part of epitope 8 (HIMA-36 was negative, while LHM76/58 was positive) were not detected, while all other epitopes were normal positive (Table 3). The three tested IgM anti-D were unable to agglutinate red blood cells with c.492C>A mutation (Table 3) indicating that this variant allele has next to partial D also weakened D expression. Two novel alleles caused the Del phenotype, the RHD*721A>C allele (encoding p.Thr241Pro) and RHD*884T>C allele (encoding p.Met295Thr), both with a single mutation in the putative transmembrane region of the RhD protein (Figure 1).

90 Table 3. D-epitope expression in the RHD*443G, RHD*492A, RHD*[178C;689T], RHD10.01, RHD*721A, RHD*1154A and RHD*01 alleles

RHD*[178C; 689T] and RHD*10.01† RHD*443G RHD*492A RHD*721A RHD*1154A RHD*01 RHD*10.01 Monoclonal RhD Red blood Red blood Transduced Transduced Transduced Transduced Red blood Transduced Red blood cells Red blood cells Red blood cells antibody epitope cells cells K562 K562 K562 K562 cells K562 RT 37ºC IAT IAT RT 37ºC IAT RT 37ºC IAT RT 37ºC IAT RT 37ºC IAT LHM70 1 - - - + 3 LHM169/81 1 + - - 4 + - +/- 4 + 5C8 2 + - 3 - 3 LHM76/59 3 + - + +/- +/- + LHM76/55 3 + + ? 4 + +/- +/- 4 + AUB-2F7/ 5 - Fiss rD7C2 5 - - - 2 + LHM50/28 6/7 + - 4 + - +/- 4 + LHM169/80 6/7 + + - 4 + +/- +/- 4 + LHM57/17 6/7 + - 4 + - +/- 4 + LOS1 6/7 + - 4 4 HIRO-5 6/7 + - 4 4 LHM76/58 8 + + - 4 + - +/- 4 + HIMA-36 8 - - - 3 LHM59/19 8.2 + - - - +/- + LHM77/64 9 + + - 4 + - +/- +

MS26 9 + + 3 individuals allelesinRhD- RHD variant Blend anti-D (MS26 + 6/7, 9 ------3 - - - 4 4 4 TM28) RUM-1 ------4 4 ESD1-M ------4 4 BRIC69 - + +/- + + Polyclonal + - - - 2 2 3 - - - 2 3 3 anti-D † epitope pattern described by Wagner et al.28 91

Chapter 4 Chapter 4

Negative non-transduced control Positive wild type RHD*01 control Specific RHD variant allele

RHD*443C>G RHD*492C>A Count Count Count Count RHD*721A>C RHD*01W.02

Fluorescence intensityintensity (arbitrary units) Figure 2. RhD expression levels of the RHD*443C>G, RHD*492C>A and RHD*721A>C variant alleles in a heterologous expression assay. Overlay plots of the fluorescence intensity representative for the RhD expression levels, of K562 cells transfected with constructs containing the RHD*443C>G, RHD*492C>A or RHD*721A>C cDNA (dark gray line with tinted area) with the RHD*01W.02 (light gray dotted line) or wild type RHD*01 cDNA (black line). The level of expression of the transporter (GFP) of RHD*443C>G, RHD*492C>A, RHD*721A>C and RHD*01W.02 compared to the wild-type RHD*01 control were, respectively, 31%, 96%, 91% and 115%. The RHD*01W.02 sensitivity control showed weakened RhD expression levels compared to the RHD*01 wild type allele. The RHD*443C>G allele had completely no RhD expression, the RHD*492C>A allele had exactly the same RhD expression level compared to the wild-type RHD*01 allele and the RHD*721A>C showed weakened RhD expression, even weaker compared to the RHD*01W.02 allele.

Expression of three variant alleles in a heterologous expression system It is relevant to correctly determine the phenotype produced by of a novel RHD variant allele, because carriers of alleles causing the partial D and D- phenotype can be immunized to the D-antigen, while this is unlikely for carriers of alleles causing the Del or weak D phenotype. Out of the fifteen novel variant alleles the phenotype could be assured for eight D- null alleles,

92 the RHD*01EL.08andRHD*01EL.09,have to giveandD beendescribed boththeD- D in certain D in certain from the phenotype as reported for alleles.from asreported thesevariant D Several thephenotype thatwasdetected inthisstudydeviated (n =1)andtheRHD*01N.22phenotype four allelesRHD*01EL.05 variant sixcasescarrying In membrane expression protein oftheRhD ismutated. (Table (Table 3 and Figure 2). The cells. The blood the expression on the transduced K562 was determined onredsimilar to the phenotype allele (Figure 2).For allthree alleles, RHDvariant RHD*443C>G, 3). The tested ascontrols andindeedbothwere positive withallninemonoclonalIgGanti-D(Table onreddetermined bloodcells. RHD*01alleleandtheRHD*01W.02The wild-type allelewere allele were examinedinaheterologous expression system thephenotype inK562to confirm splice site. far, So oftheremaining seven allelestheRHD*443C>G, a premature stopcodon, missedthecomplete exon 1orcontainedmutationsthataffected a mutationwith because theseallelescontainedanonsensemutation,frameshift D expression, in16%ofthewomen D follow-up in26%ofthewomen phenotype, partial detected in55%ofthe women theD- 45 different alleles, variant novel including fourteen theRHD*Ψ half ofthewomenallelecarried withavariant present thatcontainsRHDexon 5and/orexon follow-up 7.Genetic thatalmost determined thepresent studywe pregnantIn show thatin~1%oftheDutch D- womenalleleis avariant Discussion anti-D except LHM59/19targeting epitope 8.2,whichwascompletely negative 3). (Table expressionallele showed RhD levels, normal thesameasRHD*01allele(Figure 2),for all positive 3). (Table weakly LHM76/55 andLHM169/80)were detected very The RHD*492C>A compared to theRHD*01W.02 Yet, theRHD*01EL.08 hasoccurred inindividualscarrying allo-immunization despite themutationatsplicesite consensus. mutation intheintron orinthepromoter boundaries region oftheRHDgene, theD expression. Interestingly, cases, intwo with awild-type with theD et al. was detected. phenotype RhCcee The expressioncases. RhCe inthesetwo wasnormal Flegel phenotype el phenotype 29 also described asinglecasewithoutmutationsintheRHDexons alsodescribed andintron boundaries RHD*01W.02 alleleshowed, however, lower expression levels compared to theRHD*01 el 11,29,30 and a normal RhCe phenotype. Possibly, phenotype. RhCe andanormal inthesecasesagenethatisrequired for el variant alleles is too low to be detected by the adsorption-elution technique. allelesistoo low variant to bedetected by theadsorption-elution 31 . It isassumedthatD . It allele was originally described insinglecasewith the andtheRHD*01EL.05allelewasoriginally described RHD*443C>G RHD*721A>C allele was determined to give the D- phenotype withallanti-D toallele wasdetermined givephenotype theD- allele (Figure 2).Furthermore, onlyafew anti-D(LHM76/59, el expression from arises asmalldegree splicing ofnormal el allele showed very weak allele D showedexpression levels,very lower expression and in 3% of the women very weakexpression D andin3%ofthewomen very 31,32 (n =1),RHD*01EL.083),RHD*01EL.09 It may bethatthelevel ofexpression It variant alleles. serologicalRHD variant Overall variant allele. In total allele. wevariant identified In RHD*01 allele and without any RHD*721A>C andRHD*492C>A, RHD*721A>C andRHD*492C>A variant alleles in RhD- individuals allelesinRhD- RHD variant el variant alleles, variant including allele 33 orthe el and 93 el

Chapter 4 Chapter 4

RHD*01EL.05 allele10, hence individuals positive for the RHD*01.05 and RHD*01EL.08 allele are considered to be at risk of D immunization. Finally, in a single case carrying the RHD*01N.22 allele we detected the Del phenotype, the same phenotype as was detected by Flegel et al.29, while in the initial case with this genotype the D- phenotype was discribed34. The RHD*01N.22 allele encodes a truncated RhD protein that lacks a small part of the C-terminal intracellular tail. Because the nonsense mutation is at the end of the RHD protein, possibly some of this truncated RhD protein is able to incorporate into the membrane. For instance, the Del phenotype is also detected in individuals carrying the RHD*(1-9) allele, including three cases carrying the RHD*(1-9) allele of this study, that encodes a truncated RhD protein only nine amino acids longer than the truncated RhD protein encoded by the RHD*01N.22 allele.25 We identified fourteen novel alleles (eleven D- null, one partial weak and two elD alleles) in our cohort of D- pregnant women and one additional novel allele was identified in a pregnant woman with weak D expression in the initial serological typing. The D- phenotype was determined for a novel allele with a nonsense mutation (RHD*1084C>T), two novel alleles with frame shift mutations (RHD*125_125delAA or RHD*1174delA), four alleles with mutations that disrupt a splice site (RHD*335G>T, RHD*[634+1G>T, 1136C>T], RHD*1073+1G>T or RHD*1074- 1G>A) and one allele with the entire deletion of exon 1 (RHD*(2-10)). The presence of this last allele could not be unambiguously proven, because of the presence of another RHD variant allele in this case. The D- phenotype was also determined in a case positive for a variant allele that contained mutations of both the RHD*03.03 and RHD*09.03 variant alleles (RHD*[361T>A; 380T>C; 383A>G; 455A>C; 602C>G; 667T>G; 819G>A]). Because, the RHD*09.03 allele causes only a moderate weakening of the RhD expression and the RHD*03.03 allele has completely no effect on RhD expression levels, it is unexpected that the combination of these mutations completely abolishes RhD expression levels. Interestingly, one allele with a single missense mutation (RHD*443C>G encoding p.Thr148Arg) and one allele with the deletion of a single amino acid (RHD*424_426del encoding p.Met142del) also cause the D- phenotype. For the RHD*443G allele the D- phenotype was confirmed in a heterologous expression study in K562. Both alleles have mutations in the fifth putative transmembrane region of the RhD protein. Other RHD variant alleles with mutations in this region, drastically diminish D expression, for instance the RHD*01EL.07 (encoding p.Ala137Glu) and RHD*01EL.12 (encoding p.Leu153Pro) alleles that cause Del expression27,29, these alleles, however, do not completely abolish D expression as was detected in the RHD*443G and RHD*424_426del allele. Two novel variant alleles (RHD*[178A>C; 689G>T] encoding p.[Ile60Leu;Ser230Ile] and RHD*492A encoding p.Asp164Glu) caused partial weak D expression. Because the case positive for the RHD*[178A>C; 689G>T] allele also carried the RHD*10.01 allele, we were unable to determine the exact phenotype of this variant allele neither to unambiguously demonstrate the presence of this novel allele. We postulate that the c.689T mutation (p.Ser230Ile) which is present in both variant alleles of this case is responsible for the phenotype, hence a weak

94 protein causetheD p.Met295Thr) with a single missense mutation in the transmembrane region of the RhD The novel RHD*721A>Callele(encoding p.Thr241Pro) andRHD*884T>C(encoding ofepitope 8werewhich epitope absent. 5andpart D expression on redpartial blood cells and using the heterologous expression system in in onlyasinglecase. This allelewasnotdetected inthestudiesofFlegel etal. seven cases, whereas allothernovel alleles, except theRHD*1074-1G>A using the heterologous expression system. Interestingly, allele was detected in this variant partial expression in which epitope 1, 5 and 8 partial are absent. The http://www.uni-ulm.de/~fwagner/RH/RB2/P_RHDDnegative.htm. Accessed 20-08-2013 nullalleles. ofD- Inventory Web Resources D- nullallelesare identified.D- assay itisessentialthatthemostfrequently viaagenotyping occurring of theDphenotype women canbeusedto assays, optimizeD- RHDgenotyping becausefor prediction correct pregnancies next andasblooddonorD+.Furthermore, ofextensivelytyped thiscohort 1,2,3or5areweak Dtype recognized andcanberegarded inthecurrent and, possible ofthisgroupimmunization. Genotyping ofwomen hastheadvantagethatwoman with Dalleleandneedadministrationofanti-Dprophylaxis tonull alleleorpartial prevent anti-D exon 5 and/or exon 7. a D- of pregnantThe far majority women allele carry with a variant RHD alleleharboring aDvariant pregnant conclusion,0.96%oftheDutch D- womenIn carry had commonDutch surnames this alleleisspecific for theDutch population.All women positive for theRHD*721A>Callele and Orzinska et al. and Orzinska 36 performed inGermany, performed Austria andPoland, respectively, indicating that el phenotype. phenotype. oftheRHD*721A>Callelewasconfirmed The phenotype RHD*492C>A variant alleles in RhD- individuals allelesinRhD- RHD variant allele, were detected allele causes weak 29 , Polin etal 95 35

Chapter 4 Chapter 4

References 1. Daniels G, Reid ME. Blood groups: the past 50 years. Transfusion 2010;50(2):281-9. 2. Westhoff CM. Rh complexities: serology and DNA genotyping. Transfusion 2007;47(1 Suppl):17S-22S. 3. Colin Y, Cherif-Zahar B, Le Van Kim C, Raynal V, Van Huffel V, Cartron JP. Genetic basis of the RhD-positive and RhD-negative blood group polymorphism as determined by Southern analysis. Blood 1991;78(10):2747-52. 4. Urbaniak SJ. Alloimmunity to RhD in humans. Transfus.Clin.Biol. 2006;13(1-2):19-22. 5. Frohn C, Dumbgen L, Brand JM, Gorg S, Luhm J, Kirchner H. Probability of anti-D development in D- patients receiving D+ RBCs. Transfusion 2003;43(7):893-8. 6. Daniels G, Poole J, de SM, Callaghan T, MacLennan S, Smith N. The clinical significance of blood group antibodies. Transfus.Med. 2002;12(5):287-95. 7. Urbaniak SJ, Greiss MA. RhD haemolytic disease of the fetus and the newborn. Blood Rev. 2000;14(1):44-61. 8. Klein HG, Anstee DJ. Mollisons’ Blood Transfusion in Clinical Medicine. 11 ed. Oxford: Blackwell Publishing; 2013. 9. Kumpel BM. On the immunologic basis of Rh immune globulin (anti-D) prophylaxis. Transfusion 2006;46(8):1271- 5. 10. Daniels G. Human Blood Groups. 3 ed. Oxford: John Wiley & Sons; 2013. 11. Reid ME, Lomas-Francis C. The Blood Group Antigen Facts Book. 2 ed. San Diego: Academic Press; 2003. 12. Cherif-Zahar B, Mattei MG, Le Van Kim C, Bailly P, Cartron JP, Colin Y. Localization of the human Rh blood group gene structure to chromosome region 1p34.3-1p36.1 by in situ hybridization. Hum.Genet. 1991;86(4):398-400. 13. Cartron JP. Defining the Rh blood group antigens. Biochemistry and molecular genetics. Blood Rev. 1994;8(4):199- 212. 14. Mouro I, Le Van Kim C, Rouillac C, van Rhenen DJ, Le Pennec PY, Bailly P, Cartron JP, Colin Y. Rearrangements of the blood group RhD gene associated with the DVI category phenotype. Blood 1994;83(4):1129-35. 15. Singleton BK, Green CA, Avent ND, Martin PG, Smart E, Daka A, Narter-Olaga EG, Hawthorne LM, Daniels G. The presence of an RHD pseudogene containing a 37 base pair duplication and a nonsense mutation in africans with the Rh D-negative blood group phenotype. Blood 2000;95(1):12-8. 16. Daniels GL, Faas BH, Green CA, Smart E, Maaskant-Van Wijk PA, Avent ND, Zondervan HA, von dem Borne AE, van der Schoot CE. The VS and V blood group polymorphisms in Africans: a serologic and molecular analysis. Transfusion 1998;38(10):951-8. 17. Chou ST, Westhoff CM. The Rh and RhAG blood group systems. Immunohematology. 2010;26(4):178-86. 18. Wagner FF, Gassner C, Muller TH, Schonitzer D, Schunter F, Flegel WA. Molecular basis of weak D phenotypes. Blood 1999;93(1):385-93. 19. Okubo Y, Yamaguchi H, Tomita T, Nagao N. A D variant, Del? Transfusion 1984;24(6):542. 20. Avent ND, Finning KM, Liu W, Scott ML. Molecular biology of partial D phenotypes. Transfus.Clin.Biol. 1996;3(6):511-6. 21. Flegel WA. Molecular genetics and clinical applications for RH. Transfus.Apher.Sci. 2011;44(1):81-91. 22. Scheffer PG, van der Schoot CE, Page-Christiaens GC, de HM. Noninvasive fetal blood group genotyping of rhesus D, c, E and of K in alloimmunised pregnant women: evaluation of a 7-year clinical experience. BJOG. 2011;118(11):1340-8. 23. Lo YM, Corbetta N, Chamberlain PF, Rai V, Sargent IL, Redman CW, Wainscoat JS. Presence of fetal DNA in maternal plasma and serum. Lancet 1997;350(9076):485-7. 24. Haer-Wigman L, Veldhuisen B, Jonkers R, Loden M, Madgett TE, Avent ND, de HM, van der Schoot CE. RHD and RHCE variant and zygosity genotyping via multiplex ligation-dependent probe amplification. Transfusion 2012. 25. Fichou Y, Chen JM, Le MC, Jamet D, Dupont I, Chuteau C, Durousseau C, Loirat MJ, Bailly P, Ferec C. Weak D caused by a founder deletion in the RHD gene. Transfusion 2012;52(11):2348-55. 26. Silvy M, Chapel-Fernandes S, Callebaut I, Beley S, Durousseau C, Simon S, Lauroua P, Dubosc-Marchenay N, Babault C, Mouchet C, et al. Characterization of novel RHD alleles: relationship between phenotype, genotype, and trimeric architecture. Transfusion 2012;52(9):2020-9. 27. Li Q, Hou L, Guo ZH, Ye LY, Yue DQ, Zhu ZY. Molecular basis of the RHD gene in blood donors with DEL phenotypes in Shanghai. Vox Sang. 2009;97(2):139-46.

96 36. 35. 34. 33. 32. 31. 30. 29. 28.

Wlodarczyk B, etal. inPolish RHDvariants Wlodarczyk blooddonorsroutinelyasD-. typed Transfusion 2013. M,Pawlowska E, A,SliwaB, Kowalewska Polin J, M,Bednarz A,GuzK, H,Pelc-Klopotowska Gielezynska Orzinska 2009;49(4):676-81. blood donors inUpper Austria. with thepotential of anti-Dimmunizationamong seemingly D- Transfusion Polin H,Danzer M,Gaszner W, Broda D, St-Louis M,Proll C.IdentificationofRHDalleles J, Gabriel Hofer K, 38. Presence ofRHDinserologically C/E+individuals:aEuropean D-, multicenter study. Transfusion 2005;45(4):527- A,Drnovsek C,Doescher Gassner TD, Rozman P, EicherNI,Legler TJ, H,Kleinrath S,Garritsen T, Lukin Egger B, etal. individuals are prone to anti-Dalloimmunization. Transfusion 2005;45(10):1561-7. GF,Kormoczi C, ShaoCP, Gassner M,Legler Uchikawa TJ. DEL Acomprehensive partial analysisofDELtypes: individuals in Taiwan: amechanismstudy. Biochim.Biophys.Acta 2010;1800(6):565-73. Liu HC, Eng HL, Yang YF, Wang YH, Lin KT, Wu HL, Lin RNA splicing in RHD 7-9 TM. exonsAberrant of DEL . Transf.Clin.Biol. phenotype Del MINAMI A,Okubo Green CA,KIMURAK, Singleton BK, Y, DanielsGL. Two RHDmutationsassociated new withthe Wagner FF, Frohmajer A,Flegel WA. inDnegative RHDpositive Europeans. haplotypes 2001;2:10. BMC.Genet. forunits inGermany preventing anti-Dimmunizations. Transfusion 2009;49(3):465-71. Flegel WA, von Z,I, Wagner FF. Sixyears’D- red bloodcell to confirm RHDgenotyping performing experience Blood 2002;100(1):306-11. Wagner FF, GA,EicherNI,Flegel KS,Heymann B, Ladewig Angert WA. The DAU allelecluster oftheRHDgene. 8[Suppl 1],9S.2001. variant alleles in RhD- individuals allelesinRhD- RHD variant 97

Chapter 4 Chapter 4

98 Chapter 4 information Supporting variant alleles in RhD- individuals allelesinRhD- RHD variant 99

Chapter 4 Chapter 4

18 2.6 14.9

RHD*06 allele

49 7.4 42.1 Number of cases be positive predicted to the for RHD*Ψ allele 97 0.4 0.9 3.5 0.0 0.4 0.0 1.3 28.1 62.0 absent 86 9.6 2.1 3.5 1.0 1.9 0.0 1.7 56.6 10.0 Number of cases in which a RHD variant be allel is predicted to present 0 3 0 0 7 0 1 0 3 10 30 62 57 10 183 Number of cases in which the genotype not was determined

1 9 3 33 46 RHD*06 allele

1 2 1 93 25 122 Number of cases positive Number of cases positive the for RHD*Ψ allele 3 0 0 1 3 4 0 1 0 8 1 0 3 15 39 absent 7 1 3 9 4 1 0 3 1 0 4 81 35 126 275 Number of cases in which allel is a RHD variant present

1 3 8 1 1 3 8 7 10 10 84 16 35 127 314 Number of cases in which the genotype was determined Calculation of number cases with a RHD variant cohortCalculation allele in the total women of 37,764 D- pregnant 1 3 0 2 1 3 13 10 15 94 46 70 45 10 184 497 Total >32 <30 <30 >32 <30 >32 <30 30-31 30-31 31-32 31-32 30-31 31-32 30-31 31-32 Exon 7 Exon

<30 <30 <30 <30 >32 >32 >32 30-31 30-31 30-31 31-32 30-31 31-32 31-32 31-32 Exon 5 Exon Total Supplementary Table S1. Supplementary Table

100 variant alleles.variant The ratiooftheMLPA, representing thecopy numberoftheRHD, FigureSupplementary S1. case, with the D copy numberof3indicatingthepresence ofaRHD-RHCE -RHD hybrid allele. We postulate thatinthis lacks theexpression ofthe5’ UTRandexon 1.Furthermore, theRHCE exon 3,4,5,6,7andninehave a number ofRHDandRHCE hasblackboxesdepicted andexons of the RHCEgeneare asopenboxes. depicted The combinedcopy RHD andRHCEgenes, istranslated inamodel oftheRHDandRHCEalleles. Exons oftheRHD RHCE*02 alleleare present. with the allelelinked RHD*01-CE(2-9)-D(10) withthe RHCE*01alleleandtheknown is deleted linked el and RhCcee phenotype, a noveland RhCcee allele variant is 3for the5’ UTR andexon 1indicating thatoneoftheRHDorRHCEalleles Schematic representationSchematic of MLPA results RHD inasinglecasewithtwo

Nucleotide position c.-698T c.-132A or c.-132G RhD) and RhC (targetting c.48C c.123A c.149-882G c.149-4875A c.149-20G (targetting RhC and RhD) c.149-109nt insertion (targetting RhC) RhD) and RhC (targetting c.203G c.244T c.307C (targetting Rhc) c.335+2838C c.380T or c.380C c.455A c.514A c.602C or c.602G c.667T c.676G or c.676C c.697G c.787G or c.787A c.800A c.932A or c.932G c.941G c.989A or c.989C c.1061C c.1112G c.1193A or 1193T c.1357C Exon or (intron) 5'UTR 1 (1) 2 (2)3 (3) 4 (4) 5 (5) 6 (6) 7 (7) 8 (8) 9 (9) 10 MLPA ratio for RHD 0,8 1,1 1,0 1,0 1,0 1,0 0,9 1,0 0,9 1,2 0,8 0,9 0,8 1,0 1,9 MLPA ratio RHCE 2,1 2,3 0,9 2,9 2,8 3,1 2,9 2,9 2,8 2,8 2,1 Combined MLPA ratio of RHD and RHCE 3,0 1,8 3,2 3,2 3,6 2,9 3,0 4,1 3,5 3,6

RHD*(2-10) RHD*(2-10) is present in which exon 1 RHD*01-CE(2-9)-D

RHCE*01

RHCE*02 individuals allelesinRhD- RHD variant RHCE or combined gene are 101

Chapter 4

Chapter 5

SMIM1 underlies the Vel blood group and influences red blood cell traits

Lonneke Haer-Wigman1* Ana Cvejic2,3* Jonathan C. Stephens2,4* Myrto Kostadima5, Peter A. Smethurst2,4, Mattia Frontini2,4, Emile van den Akker1, Paul Bertone5, Ewa Bielczyk-Maczyńska2,3,4, Samantha Farrow2,4, Rudolf S. Fehrmann6, Alan Gray7, Masja de Haas1, Vincent G Haver6, Gregory Jordan8, Juha Karjalainen6, Hindrik H.D. Kerstens9, Graham Kiddle2,4, Heather Lloyd-Jones2,4, Malcolm Needs7, Joyce Poole10, Aicha Ait Soussan1, Augusto Rendon2,4,11, Klaus Rieneck12, Jennifer G. Sambrook2,4, Hein Schepers6, Herman H. Silljé6, Botond Sipos5, Dorine Swinkels13, Asif U. Tamuri5, Niek Verweij6, Nicholas A. Watkins4, Harm-Jan Westra6, Derek Stemple2, Lude Franke6, Nicole Soranzo3, Hendrik G. Stunnenberg9, Nick Goldman5, Pim van der Harst6* C. Ellen van der Schoot1* Willem H. Ouwehand2,3,4* Cornelis A. Albers2,3,4,13*

1 Sanquin Research, Amsterdam and Landsteiner Laboratory, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands 2 University of Cambridge, Cambridge, UK 3 Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, UK 4 National Institute for Health Research (NIHR) and National Health Service (NHS) Blood and Transplant, Cambridge, UK 5 European Molecular Biology Laboratory –European Bioinformatics Institute, Wellcome Trust Genome Campus, Hinxton, UK, 6 University of Groningen, University Medical Center Groningen, Groningen, The Netherlands, 7 NHS Blood and Transplant, Tooting, London, UK, 8 Somerville, Massachusetts, USA, 9 Faculty of Science, Nijmegen Centre for Molecular Life Sciences, Radboud University, Nijmegen, The Netherlands, 10 NHS Blood and Transplant, Filton, UK. 11 Medical Research Council Biostatistics Unit, Institute of Public Health, Cambridge, UK. 12 Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark. 13 Radboud University Medical Centre, Nijmegen, The Netherlands * Equal contribution

Nat Genet. 2013;45(5):542-5 Chapter 5

Abstract & Introduction The blood group Vel was discovered 60 years ago1, but the underlying gene is unknown. Individuals negative for the Vel antigen are rare and are required for the safe transfusion of patients with antibodies to Vel.2 To identify the responsible gene, we sequenced the exomes of five individuals negative for the Vel antigen and found that four were homozygous and one was heterozygous for a low-frequency 17-nucleotide frameshift deletion in the gene encoding the 78 amino-acid transmembrane protein SMIM1. A follow-up study showing that 59 of 64 Vel- individuals were homozygous for the same deletion and expression of the Vel antigen on SMIM1-transfected cells confirm SMIM1 as the gene underlying the Vel blood group. An expression quantitative trait locus [eQTL], the common SNP rs1175550 contributes to variable expression of the Vel antigen (P = 0.003) and influences the mean hemoglobin concentration of red blood cells [RBCs] (P = 8.6 × 10-15)3. In vivo, zebrafish with smim1 knockdown showed a mild reduction in the number of RBCs, identifying SMIM1 as a new regulator of RBC formation. Our findings are of immediate relevance, as the homozygous presence of the deletion allows the unequivocal identification of Vel- blood donors.

104 were investigated by test 2. Two caseswere found to have weak Vel [Vel+ was available. RBCsfrom seven siblingsfrom seven different Vel- caseswithantibody to Vel to Vel. addition,anotherfive clinicalcaseswithantibody to In Vel were enrolled, butnoDNA definitively Vel-, we enrolled 16 individuals (Supplementary Figure S2) who antibody formed not more seraagainst thantwo Vel. To have accessto samplesfrom individualswhowere 1 (Uniquesamplenumber[USN]45–USN50) were by confirmed test 2butgenerallywith to Vel thantheantiserausedin andUK. The Netherlands The sixsamplesnegative intest the numberstested were maintained)were screened by test 1usingdifferent lotsofantibody addition, the RBCs from tens(no pre of thousands of blood donors in ­ Denmark Figure RBCswere available, S1),for were whichcryopreserved investigated by test 3. In samples withdiscordance SMIM1 between hemagglutination test on Vel+ cells(test 3)hasaforensic level ofsensitivity. Anumberof Adsorption andelutionofantibodyto Vel followed by titrationoftheeluted antibodyinthe ofsampleswithnegative testthat afraction 2resultshave actually weak Vel expression. low amountsofthe Fig.Vel istherefore antigen(Supplementary S1).It reasonable to assume three different potent antibodies to Vel; however, this test doesnotalways extremely detect therefore tested test (test by 2)withatleast amore sensitive hemagglutinationconfirmatory 1 isprone to highlevels offalsepositive andfalsenegative results. Negative sampleswere polyclonal humanantiserumcontainingantibodiesto FigureVel (Supplementary tested for the Vel bloodgroup by ahemagglutination screening test (test 1)usingasingle RBCs from 350,000 non-remunerated nearly donors in and England Thewere Netherlands ofRBCsforPhenotyping the Vel bloodgroup Material andmethods to human IgG (Molecular Probes,to humanIgG(Molecular A-11013). immunoglobulin [Ig]Gto Vel antibody detected by Alexa Fluor 488–labeledgoatsecondary of Velcytometry abundanceontheRBCmembranewith themostpotent immune-purified population. in thenormal by the measurementThis by assumptionwassupported flow with thenotionthatabundance ofthe Vel substantially antigenonRBCsislow and varies The results obtainedwiththehemagglutinationtests withantibodyto Vel are compatible expression onRBCs. anti­ immunopurified some or all ofthefollowing antisera: NL04-1, NL04-2, NL05, NL07,NL14, NL15 andNL22. The Dutch sampleswere for typed with Veltyping wasperformed withNL25,andconfirmation IBGRL_anti-Vel_002, and, inanumberofcases, athird, NHSBT-IBGRL_anti-Vel_003, wasused. antisera, NHSBT-IBGRL_anti-Vel_001 further with two performed was typically and NHSBT- antisera, NHSBT-Tooting_anti-Vel_001 and NHSBT-Tooting_anti-Vel_002. Confirmation testing pregnancy. NHSBloodand Transplant used fortwo different initialidentification of donors antigen were obtained from Vel- individuals immunized against Vel by transfusion or during the remaining five were found to be Vel- (Supplementary Figure S2).Allantibodies to the Vel body to Vel measurement of NL25wasusedforcytometry flow Vel genotype and genotype Vel (Supplementary phenotype underlies the SMIM1 underlies Vel bloodgroup w ] expression, and cise records on S1). Test 105

Chapter 5 Chapter 5

Genetic analysis DNA was extracted from blood or saliva samples using stand­ard laboratory procedures. The DNA samples from 5 individuals (USN1–USN5) were analyzed by exome sequencing, and all 96 samples were used for genetic analysis by Sanger sequencing. The SNP rs1175550 was typed either by Sanger sequencing or with a variant-specific TaqMan probe using standard conditions. Sanger sequencing was carried out on the coding fraction and intronic flanking regions of the SMIM1 gene. In England, consent was obtained under the Cambridge BioResource stage 1 enrolment protocol, which has been approved by the Cambridgeshire 1 Research Ethics Committee. Samples from the Dutch donors and patients were obtained with informed consent, and use of the samples of Vel- and Vel+w donors was approved by the Sanquin Clinical Consultative Service. Samples in Denmark were obtained according to national regulations for consenting individuals for biomedical research.

Exome sequencing and data analysis Libraries were prepared and indexed using the Illumina TruSeq library prep kit. Sequence capture was performed using Roche Nimblegen SeqCap EZ Human Exome v3.0. The five initial samples­ were sequenced on the Illumina HiSeq 2000 platform. Reads were aligned with Burrows-Wheeler Aligner [BWA] v 0.6.1 (ref.4), duplicates were marked with Picard (see Web Resources), realignment was performed around indels, base quality was recalibrated and variant calling was performed with GATK5. On average, 5.9 Gb of sequence was generated per sample, and 92% (range of 90–95%) of the 64 Mb of captured target sequence was covered by at least tenfold read depth. Variants were annotated using the Ensembl Variant Effect Predictor6, and allele frequencies were obtained from the European population in the Phase 1 release of 1000 Genomes Project data.7 We considered all variants predicted to disrupt protein-coding sequence with an allele frequency below 5% as potential causative variants underlying the Vel- phenotype.

Transcript profiling of blood precursor cells A compendium of transcripts was generated as part of the BluePrint project8 (see Web Resources) by sequencing RNA samples obtained from highly pure preparations (>95%) of five differ­ent primary progenitor and precursor blood cells. Precursor cells for RBCs and —erythroblasts and megakaryocytes, respectively were generated by the culture of CD34+ hematopoietic stem cells [HSCs] for 10 to 12 days, as described previously.9 Both the HSCs and other pro­genitor and precursor cells were isolated from the mononuclear cell fraction of donations of human cord blood by fluorescence-activated cell sorting using monoclonal antibodies against CD markers specific for certain stages of differentiation and lineage commitment. Sorted cells, ranging in number from 20,000 to 100,000, were directly collected into TRIzol (Invitrogen) and stored at −80 °C. For paired-end sequencing,

106 Genomics Viewer [IGV]. ViewerGenomics coverage Sequencing multimap correction. over genomiclociwasvisualized inIntegrative Sequence readsSequence were aligned to assembly the using FebruaryGSNAP 2009 high-coverage specifications. with TruSeq Allreagents reagents were (Illumina). usedaccording to themanufacturer’s 2000system fragments. onaHiSeq Paired-endlibrary wasperformed sequencingoflibraries input. For multiplexingpurposes, Bioscientific’s ChIP NEXTflex Adapters were ligated tothe (Clontech) using100pgoftotalfor sequencingandAdvantage RNAas 2PCRkit Illumina wereanalysis libraries prepared and amplified withthe SMARTer Ultra Input RNA Low (Qiagen). usingtheRNeasyminikit and concentrationstep wasperformed Transcriptome total RNA wasisolated following themanufacturer’s protocol, andanadditionalpurification HEK293T cells (7.5 × 10 SMIM1 transfections Odyssey Infrared Imaging System (Li-Cor Biosciences). System (Li-Cor Imaging Infrared Odyssey gel. Fluorescence outat25°Con a4.5%polyacrylamide wasvisualizedwas carried withthe volume of15μl. The wasincubated mixture for 30minutes atroom temperature, andPAGE 2.5% glycerol in a final mM and 75 ng/μl poly(dIdC) Tris (pH 7.5), 50 mM KCl and 1 mM DTT), with 0.2pMannealedoligonucleotidesinbindingbuffer (10 2 to10 μgofnuclearextract products usingastandard protocol. by mixing DNA-protein bindingreactionwas performed 700tagsonthe5’with IRDye end. Labeledprobes were annealed to excess unlabeledPCR rs1175550. variant genomic sequenceflanking probesThe wereforward fluorescently labeled reagents (Pierce).extraction Oligonucleotides (Biolegio) were designed on the basis of the were preNuclear extracts ­ assays shift Electrophoretic mobility [EMSAs] H4). expression wasassessedby antibodyto CD271conjugated to APC(MiltenyiBiotec, ME20.4-1. growth antibodyto humanIgG(A-11013), receptor andnerve factor labeled secondary andelutionfrom by adsorption wasperformed purification Vel+ RBCs)andAlexa Fluor 488– IgG antibody tothe transfectants was assessed using human immune-purified Vel (immune Coulter). usinga instrument(Beckman screenedcytometry byFortessa flow Vel expression in transfection, hoursafter cellsweremLwpRRLsSMIM1itNGFR. and harvested Forty-eight Quantification of was performed using Cufflinks Quantification ofgeneexpression wasperformed of novel from splicesites eachotherwere atgenomicdistancesupto 100kb allowed. version disabled. 2012-07-20 with trimming A maximum of five mismatches andidentification 12 5 ) were plated in 6 cm pared from K562cellsusingNE-PERNuclearandCytoplasmic 2 plates and lentivirally transduced with underlies the SMIM1 underlies Vel bloodgroup 11 v.1.3.0 allowing for 107 10

Chapter 5 Chapter 5

Zebrafish studies General maintenance, collection and staging of the wild-type zebrafish were carried out according to the Zebrafish Book.13 Embryos were maintained Staining with o-Dianisidine for hemoglobin was performed as previously described.14 Photomicrographs were taken with a Zeiss AxioCam HRC camera attached to a LeicaMZ16 FA dissecting microscope. Morpholinos targeting zebrafish smim1 and standard control oligonucleotide were obtained from GeneTools. Oligonucleotides were resuspended in sterile water, and approximately 1 nl was injected into zebrafish embryos at the one-cell stage. For both smim1 splice-blocking morpholinos and the standard control oligonucleotide, a concentration of 6.4 μg/μl was used. The efficiency of splice-site morpholino-mediated gene knockdown was determined by RT- PCR amplification of template using gene-specific for­ward and reverse primers. The following program of cycling was used for KOD Hot Start PCR: 95 °C for 2 min, 95 °C for 20 s, 61 °C for 10 s and 70 °C for 12 s (40 cycles).

Evolutionary analysis We retrieved genomic regions with flanking sequences of 300 bp in length for orthologs of the human SMIM1 (ENSG00000235169), human SMIM2 (ENSG00000139656) and human SMIM3 (ENSG00000256235) genes from Ensembl.15 In addition to the 16 orthologs listed in Ensembl for SMIM1, we added 1 identified manually by synteny in zebrafish (ENSDARG00000075500). For this particular ortholog, we found seven SMIM3 zebrafish paralogs. We used webPRANK16 (genomic model) to create three sequence align­ments: (i) SMIM1 sequences were aligned using the species tree available from Ensembl as a guide, (ii) SMIM1 and SMIM2 sequences were grouped for the second alignment, and (iii) SMIM1, SMIM2 and SMIM3 sequences were grouped for the third alignment. We fetched the protein translation products of the canonical transcripts of the SMIM genes mentioned above (if existing) from Ensembl. We used StatAlign 1.1 (ref.17) to perform statistical alignments of the protein sequences (WAG substitution model, 50,000 burn-in cycles, 500,000 cycles after burn-in, sampling rate of 1,000, 3 replications). We projected the canonical transcript of exons in human SMIM1 across all sequences in the genomic alignments produced above to generate protein-coding sequence alignments. We analyzed the SMIM1 alignment using the M0 (one-ratio) and M3 (discrete) site models available in the codeml program of PAML.18

Gene coexpression network analysis We used a ‘guilt-by-association’ approach to predict likely functions for genes on the basis of gene coexpression.­ However, it is important to realize that some phenomena exert very strong transcriptomic effects and therefore will overshadow more subtle effects. To be able to identify such subtle relationships as well, we conducted principal-component analysis [PCA]

108 correlation ofeachindividualmicroarray withPC ­ vari probe-specific which explainsaround 80–90%ofthetotal variance. [PC PCAonthesamplecorrelation matrix. subsequently conducted component The firstprincipal subse­ 8 batches, owing to its size, by randomly assigning the samples to 1of these batches. We 230 2.0 Array. For U133 Plusthe 43,278 2.0 Array samples, we ran RMA in U133AArray, Human Genome and6,123Rat 4302.0Array Genome 18,639Mouse measure [RMA] for normalization. We couldrunRMA onallsamplesatoncefor the20,108 18,639 and6,124arrays, geneexpres respectively) Multi-Array andusedtheRobust ­ For230 2.0Array). eachoftheseplatforms, we downloaded theraw CELfiles(20,108,43,278, U133 Plus Genome Rat 2.0Array, and Genome 4302.0Array Affymetrix Mouse Affymetrix expression platformsGenome U133AArray, (Affymetrix Human Genome Human Affymetrix from Expression the Gene Omnibus [GEO]. We confined analyses to four dif­f expression data for three different species (Homosapiens, R.S.N.F.,on an unprecedented scale (J.K., H.-J.W. and L.F., unpublished data). We gene collected different platforms andspecies by collapsingtheprobe identifiers tohumanEnsemblgenes biological phenomena. We therefore used theindividualcomponentsandintegrated the nents are potentially equally biologically relevant, certain as eachof the components describes profound effects on expression (many pathways enriched and GO the other compo terms), diverseoften phenomena. Although,for firstcomponentsdescribe eachspecies, thevery pathway,or Reactome biologically indicatingthatthesecomponentsdescribe relevant but are significantly (FDR<0.05) enriched for atleast1geneontology or KEGG,BioCarta [GO] term analysis[GSEA] showed enrichment thateachofthese 2,206components Subsequent gene-set three species. data expressionper Affymetrix platform, and many ofacross these are the well conserved component. Jointly, inthe 79%and90%ofthevariance thesecomponentsexplainbetween 2302.0)by375 for Genome requiring Cronbach’s Rat α>0.70for eachindi­ U133A, 777forGenome U133Plus HumanGenome 2.0,677for 430 2.0and Genome Mouse the identificationofa total of2,206 components(377 robustly estimated principal for Human genes.between Therefore, we PCAontheprobe conducted resulting correlation matrix, in on many itdifficult geneswould toidentifythemore make subtle relationshipsthatexist guilt-by-association coexpression analysis, we reasoned thatthepresence ofprofound effects canbeusedfor control theaforementionedthis datasetonwhichquality wasperformed analysis (54,736humansamples, 17,081mousesamplesand6,023ratsamples).Although R lesser quality willhavelesser quality lower correlation withPC < 0.75. After quality control,< 0.75.After quality atotal of77,840different samples remained for downstream qc ] on such a matrix nearly always describes a constant pattern (dominating aconstantpattern always thedata), describes nearly ] onsuchamatrix quently conducted quality control quality onthedata. quently conducted We first removed duplicate samples, and ance, independentofthebiological samplehybridized to thearray. The qc qc . We removed samplesthathadcorrelation canbeusedto outliers, detect of asarrays 19,20 Mus musculusand This canberegarded pattern as underlies the SMIM1 underlies Vel bloodgroup ) norvegicus Rattus erent Affymetrix vidual principal vidual principal sion 109 ­

Chapter 5 Chapter 5 and used orthology information­ from Ensembl for the mouse and rat platforms, resulting in a harmonized matrix of 19,997 unique Ensembl genes × 2,206 principal components. We subsequently predicted the most likely GO biological process using the following strategy. We first ascertained each individual GO term and assessed for each principal component whether the genes that were explicitly annotated with this GO term showed a significant difference from the genes that were not annotated with this GO term using a t test. We converted the resulting P value into an enrichment z score (to ensure normality). We subsequently investigated SMIM1 and correlated the 2,206 principal-component eigenvector coefficients of SMIM1 with each GO term by taking the 2,206 enrichment z scores as the expression profile for that GO term. A significant positive cor­relation means that SMIM1 has an expression profile that is comparable to that of genes annotated with the GO term. We visualized this method at the Gene Network website (see Web resources). To correct for multiple testing, we permuted Ensembl gene identifiers. Using permuted data, we repeated enrichment z-score calculation and investigated how strongly SMIM1 correlated with the permuted pathway. We repeated this analysis 100 times and observed that a P-value cutoff of 1.18 × 10−5 corresponded to an FDR of 0.05. This resulted in significant prediction of 14 GO biological process functions.

Results & Discussion We screened nearly 350,000 blood donors for the absence of expression­ of the Vel antigen, showing less than 1 in 4,000 to be negative (Supplementary Figure S1). In all cases, we used antibodies derived from Vel- individuals immunized against Vel by transfusion or during pregnancy. Establishing the absence of the Vel antigen is challenging owing to its low and variable abundance (Figure 1a). To identify the gene underlying the Vel blood group, we sequenced five individuals negative for the Vel antigen (Supplementary Figure S1) on the Illumina HiSeq 2000 plat­form after targeted enrichment with the Roche Nimblegen SeqCap EZ Human Exome v3.0 protocol. All but one were homozygous for a 17-nucleotide frameshift deletion (hg19, chr. 1: g.3691998_3692014delGTCAGCCTAGGGGCTGT) in SMIM1, and the remaining individual was heterozygous for this deletion (Figure 1b and Supplementary Figure S3). The deletion has a low allele frequency of 1.6% (119/7,562) in the UK10K cohort, on which whole-genome sequencing has been carried out at mean coverage of 6×. In follow-up analysis, we performed Sanger sequencing on an additional 84 unrelated individuals who were either Vel- (n = 64) or showed weak expression of the antigen (n = 20). A total of 63 out of 69 Vel- individuals were found to be homozygous for the deletion (Figure 1b and Supplementary Figure S1). Notably, all 16 Vel- clinical cases (Supplementary Figure S2) who carried antibodies against Vel, confirming­ their Vel- status with extremely high confidence, were homozygous for the deletion. Given the population frequency of the deletion, this replication is highly significant (P = 1 × 10-58).

110 as the gene underlying the SMIM1asthegeneunderlying confirming cytometry, Vel bloodgroup. AU, units. arbitrary expression ofthe Vel withhumanpolyclonal antigen,asdetermined antibodiesagainst Vel andflow c) Overexpression ofhumanSMIM1cDNAinHEK293Tcells. allofthetransfected Nearly cellsshowed or asecond, different mis­ of the Vel antigen.Finally, six individuals classifiedas Vel- were heterozygous for thedeletion alteration. This findingshows deletionofSMIM1 thatheterozygous remaining for individualwasheterozygous missense mutationencodingap.Met51Lys anew RBC hemoglobinconcentration(P=8×10 rs1175550 wasassociated withlower expression ofSMIM1inwholeblood(Figure 2a)andlower mean lGTCAGCCTAGGGGCTGT) the inSMIM1underlies Vel- phenotype. The majoralleleofthecommonSNP deletion(hg19, chr. ofa17-nucleotideframeshift inheritance b) Homozygous 1: g.3691998_3692014de Further, 19outof20 Vel+ Vel for, antigenmeasuredcytometry by flow respectively, a Vel- individual, an Vel+ Figure 1.The weak expression andelution,an by onlydetectable absorption Vel+ SMIM1 geneencodesthe Vel blood group antigen.a)RBCmembraneexpression ofthe sense mutationreplacing with anarginine. Met51 Heterozygosity w individualswere for heterozygous thedeletion(Figure 1b);the −15 ) inalarge meta-analysis ofGWAS ofRBCparameters3. w individualanda Vel+ individual. underlies weak expressionunderlies underlies the SMIM1 underlies Vel bloodgroup w individual with very individualwithvery 111

Chapter 5 Chapter 5 for the null allele in these individuals is most likely explained by misclassification­ of extremely weak Vel expression as Vel-. For one case (USN48), we were able to retrieve RBCs, and this case was indeed weakly positive for the Vel antigen (Supplementary Figure S1). From the extremely weak Vel expression in the individuals with missense mutations, we infer that these mutations may lead to inability of the SMIM1 protein to incorporate in the membrane21,22, or, alternatively, it is possible that these mutations only modify the epitope, leading to greatly reduced binding of the polyclonal antibody to Vel. Further experiments are needed to determine the functional consequences of these missense mutations. We next validated this finding in vitro by overexpressing human SMIM1 cDNA in HEK293T cells (Figure 1c).

Figure 2. Common SNP rs1175550 is an eQTL for SMIM1 and is associated with RBC traits. a) Gene expression levels (y axis) as a function of imputed allele dosage (x axis) for 1,420 individuals6. The number of individuals with the indicated genotype is given in parentheses. AU, arbitrary units. a) Open chromatin determined by formaldehyde-assisted investigation of regulatory elements and sequencing (FAIRE-seq) in erythroblasts, the precursor cell of RBCs, indicates that the rs1175550 SNP is located in a regulatory element. The y axis shows the number of sequencing reads. This is further supported by binding in the myeloid-erythroid cell line K562 of the subset of transcription factors assayed by the Encyclopedia of DNA Elements [ENCODE] Project26 that are differentially expressed in erythroblasts27. Bars encompass peaks of transcription factor occupancy. Darkness of shading is proportional to the maximum signal strength in any cell line. a) Expression of the SMIM1 transcript based on RNA sequencing [RNA-seq] in the hematological lineage. SMIM1 is not transcribed in the lymphoid progenitor but is highly expressed in RBC precursor cells, erythroblasts. FPKM, fragments per kilobase per million reads. d) rs1175550 was also associated with the mean corpuscular hemoglobin concentration [MCHC] in RBCs in the large meta- analysis at P = 8.6 × 10−15 (ref.3). Each SNP has both a SMIM1 eQTL association P value (left y axis) and an MCHC association P value from the meta-analysis of GWAS for RBC parameters (right y axis).

112 [LD] withrs1175550(r colocalized (Figure 2d).AsecondSNP, disequilibirum rs1184341, whichisinstrong linkage for lower all correlated parameters. 0.005), hemoglobinconcentration(P The individualsclassifiedashaving Vel+ our geneticfindings, the thisidentifiesSMIM1asthegeneunderlying Vel bloodgroup. Vel. alloftheSMIM1-transfected Nearly cellsexpressed the Vel antigen. Taken together with antibodyThe to usinghumanimmune-purified Vel cytometry antigenwasidentified by flow in RBCs(P rs1175550isassociated withdecreasedcommon variant meanhemoglobinconcentra­ [GWAS] 72,000individuals ofsixRBCparameters. innearly role to SMIM1, we considered associationstudies arecent meta-analysisofgenome-wide which isofsuf­fi The SMIM1protein containsasinglestretch of22hydrophobic residues (from Val53 to Val74), SMIM1 individuals. are haplotype normal, theminoralleleofrs1175550onnon-deletion Vel+but carrying of the Vel antigenSMIM1;we hypothesize for thatindividualsheterozygous thedeletion This, therefore, suggeststhattheSMIM1 chromosome in24individualsby chanceisP allele ofrs1175550onthenon-deletion the major the probability of observing in 1b). the On UK10K the cohort, basis of its frequency withoutthemissensemutation(Figure haplotype the Aalleleofrs1175550onwild-type rs1175550; the Vel- for individualheterozygous rs1175550andthemis­ were for heterozygous for thedeletionwere reduce themajorallelefactors homozygous of level ofSMIM1 Figureto (Supplementary theminor(G) alleleofthisvariant S4)suggeststhatrepressive the precursors (Figure 2c). Greater binding of nuclear proteins to the major (A) allele compared humanbloodcellprogenitors and byas determined thesequencingofRNAfrom primary (Figure 2b)butnotinalymphoidcellline, ofexpression whichiscompatiblewithitspattern cellline bindingatthispositioninamyeloid-erythroid RBCs, factor withcleartranscription population. regionThis SNPislocated theprecursor inaregulatory inerythroblasts, cellsof (Figure 2a)andexplainsmore inSMIM1 than60% ofthevariation with decreased SMIM1 rs1175550, located inthesecondintron ofSMIM1,wasstrongly associated (P (the Areference ofthecommonvariant allele, of77%intheUK10Kcohort) allelefrequency lowers ofSMIM1onthecopy notdisrupted thetranscription by thedeletion. The majorallele the Vel antigen. We therefore hypoth­ SMIM1 SMIM1 was only recently annotated as a protein-coding gene and has no known function. function. was onlyrecently annotated geneandhasnoknown asaprotein-coding = 8.6×10 in theeQTL studyandtheassociationsignal for RBChemoglobinconcentration transcript levels,transcript asnoted above (Figure 2a). We found thattheassociationsignal cient length to act asatransmembranedomain. cient lengthto act To assign apossiblebiological transcription. All24 transcription. Vel- individualsorthosewithweak Vel expression who -15 2 =0.92)butwhichwasnotdirectly tested inthegene expression study ) andisassociated atnominalsignifi­ transcript levels inaneQTLtranscript studyinwholeblood1,240individuals 3 The majorallelewasalsostrongly associated (P = 0.001)andmeanRBCvolume (P esized that these individuals carry amodifierallelethat esized thattheseindividualscarry w expression low generallyhadextremely amountsof eQTL, expression to rs1175550,contributes variable cance levels withRBCcount(P 3 Indeed, themajoralleleof Indeed, underlies the SMIM1 underlies Vel bloodgroup transcript levels inthe transcript sense mutation carried sense mutationcarried = 4.5×10 = 1×10 = 1×10 -6 ), whichare -250 = 0.003. ) with tion -250 113 = 23 )

Chapter 5 Chapter 5 and the GWAS, also showed strong evidence for colocalization (Supplementary Figure S5). Additionally, we did not identify other target genes for the rs1175550 eQTL (at false discovery rate [FDR] < 0.05). These results strongly suggest that SMIM1 mediates the effect of the GWAS signal. On the basis of this finding, we conclude that SMIM1 affects the mean hemoglobin­ concentration of RBCs, although the effect size is likely to be small according to the GWAS result, and the precise causative variant(s) remain to be identified. The role for SMIM1 in the forma­tion and hemoglobinization of RBCs was further supported by a large-scale gene coexpression analysis (Supplementary Table 1). To establish the biological relevance of SMIM1 in RBC forma­tion in vivo, we performed morpholino-mediated knockdown in zebrafish. Over the years, zebrafish has proven its suitability as a model system for furthering the understanding of hematopoiesis in humans.24,25 SMIM1 has evolved under strong purifying selection (Supplementary Figure S6). The second protein-coding exon of human SMIM1, which encodes the transmembrane domain, is well conserved with the zebrafish ortholog smim1 we identified (Supplementary Figure S6). Knockdown of smim1 resulted in mild but consistent reduc­tion in the number of RBCs compared with control embryos (Figure 3).

Figure 3. Zebrafish knockdown of smim1. Whole-mount o-Dianisidine staining for hemoglobin at 3 d post-fertilization [3 d.p.f.] showed mild reduction in the total number of mature primitive erythrocytes (black arrow) in the zebrafish smim1-depleted embryos compared to control embryos. MO, morpholino. Scale bars, ~100 μm.

This observation is consistent with the effect of reduced SMIM1 transcript levels on the average hemoglobin content of human RBCs, as hemoglobin availability is a critical deter­ minant of the number of RBCs. We investigated RBC and iron homeostasis parameters in twelve Vel- blood donors and observed a weak but nonsignificant trend suggestive of depletion in iron reserves (data not shown). Low iron concentration is the most common cause of low hemoglobin levels, and it should be taken into account that females and males with hemoglobin levels below 12.5 and 13.5 g/dl, respectively, are removed from the donor pool. This selection may have reduced

114 In summary, we haveIn identifiedSMIM1 fully char­ occursupstreamframeshift oftheregion encodingthetransmembranedomain)andto more deletionindeedresultsto (the establishwhethertheframeshift incomplete lossoffunction SMIM1 protein intheregulation ofRBCparameters islimited. Further studiesare necessary inzebrafish suggestthatthe hemoglobin concentrationandthemildphenotype role for the deletion,thesmalleffect size ofthecommonSNPrs1175550onmeanRBC homozygous the formation of RBCs. However, the apparently consequences of limited phenotypic deletionon frameshift to obtain anunbiasedestimate oftheeffect ofthehomozygous concentration; anindependent follow-up studyinalarge, populationisrequired unselected thetrueeffect oftheSMIM1 the power to observe BluePrint Project, http://www.blueprint-epigenome.eu/ coexpressionGene net Picard, http://picard.sourceforge.net/ Web Resources funded by BiomedicalResearch theNIHRCambridge Centre. Jointly, theseresources have inexcess of10,000research volunteers, andtheresources are were enrolled BioResource viatheCambridge andtheNIHRBioResource for Diseases. Rare http://www.UK10K.org/. Vel-negative donorsandthosewithweak Vel expression inEngland A fulllistoftheinvestigators to whocontributed thegeneration ofthedataisavailable from derived fromdata generated by samples from the UK10K Consortium, the TwinsUK cohort. cord bloodandP. for helpwith immune hemagglutination. Ligthart useof This studymakes R.J. vanderLei for J.J. helpwithcellsorting, A.vanLoon Erich, andcolleaguesfor collecting andmorphology,Moes,Mesander and cytometry H. G. withbloodcellflow forErber support for enrichment theexomefor sequencing. performing We thankS.Garner, and Downes K. W. We inthisstudy. thanktheindividualswhoparticipated We andI.Simeoni thankA.Rogers Acknowledgements in theregulation ofRBChemoglobinparameters. ofsmim1 knockdown inzebrafish with formation observed and that expression tocontributes variable ofthe Vel ofthedele­ carriers antigeninheterozygous studies indicates thattheSMIM1 Integrative analysisofageneexpression studyandameta-analysisofgeneticassociation ofincompatibledonorRBCs byand sometimeslife-threateningantibodiesto destruction Vel. of Vel- blooddonors, thereby preventing erroneous Vel andreducing ofsevere typing therisk underlies Vel- statusisofdirectclinicalrelevance, asitallows theunequivocal identifica­ 17-nucleotidedeletionpolymorphism for alow-frequency finding thathomozygosity SMIM1 acterize the biological and biochemical function ofSMIM1. thebiologicalacterize andbiochemicalfunction affects the mean hemoglobin concentration ofRBCs. The mildly reduced RBC work analysis,­work http://genenetwork.nl:8080/GeneNetwork/ eQTL instrong rs1175550(or avariant LDwithrs1175550) as the gene underlying the as thegeneunderlying Vel bloodgroup. Our deletion on iron reserves andhemoglobin deletion oniron reserves further supports arole for supports SMIM1 further underlies the SMIM1 underlies Vel bloodgroup tion tion 115

Chapter 5 Chapter 5

References 1. Sussman, L.N. & Miller, E. Un nouveau facteur sanguine “Vel”. Rev Hémat 1952;7:368–71. 2. Daniels G. Human blood groups. 2nd ed. Oxford: Blackwell Science; 2002. 3. van der Harst, P. et al. Seventy-five genetic loci influencing the human red blood cell. Nature 2013;492:369–75. 4. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 2009;25:1754–60. 5. DePristo, M.A. et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet 2011;43: 491–8. 6. McLaren, W. et al. Deriving the consequences of genomic variants with the Ensembl API and SNP Effect Predictor. Bioinformatics 2010;26:2069–70. 7. The 1000 Genomes Project Consortium. An integrated map of genetic variation from 1,092 human genomes. Nature 2012;491:56–65. 8. Adams, D. et al. BLUEPRINT to decode the epigenetic signature written in blood. Nat Biotechnol 2012;30:224–6. 9. Gieger, C. et al. New gene functions in megakaryopoiesis and platelet formation. Nature 2011;480: 201–8. 10. Wu, T.D. & Nacu, S. Fast and SNP-tolerant detection of complex variants and splicing in short reads. Bioinformatics 2010;26:873–81. 11. Roberts, A., Pimentel, H., Trapnell, C. & Pachter, L. Identification of novel transcripts in annotated genomes using RNA-Seq. Bioinformatics 2011;27:2325–9. 12. Robinson, J.T. et al. Integrative genomics viewer. Nat Biotechnol 2001;29:24–6. 13. Westerfield, M. The Zebrafish Book. Eugene University of Oregon Press;1994. 14. Detrich, H.W. et al. Intraembryonic hematopoietic cell migration during vertebrate development. Proc Natl Acad Sci USA 1995;92:10713–7. 15. Flicek, P. et al. Ensembl 2012. Nucleic Acids Res 2012;40:D84–D90. 16. Löytynoja, A. & Goldman, N. webPRANK: a phylogeny-aware multiple sequence aligner with interactive alignment browser. BMC Bioinformatics 2010;11:579. 17. Novák, Á., Miklós, I., Lyngsø, R. & Hein, J. StatAlign: an extendable software package for joint Bayesian estimation of alignments and evolutionary trees. Bioinformatics 2008;24:2403–4. 18. Yang, Z. PAML 4: Phylogenetic Analysis by Maximum Likelihood. Mol Biol Evol 2007;24:1586–91. 19. Sherlock, G. Analysis of large-scale gene expression data. Curr Opin Immunol 2000;12:201–5. 20. Alter, O., Brown, P.O. & Botstein, D. Singular value decomposition for genome-wide expression data processing and modeling. Proc Natl Acad Sci USA 2000;97:10101–6. 21. Körmöczi, G.F. et al. Genetic diversity of KELnull and KELel: a nationwide Austrian survey. Transfusion 2007;47:703– 14. 22. Wester, E.S. et al. KEL*02 alleles with alterations in and around exon 8 in individuals with apparent KEL:1,−2 phenotypes. Vox Sang 2010;99:150–7. 23. Fehrmann, R.S.N. et al. Trans-eQTLs reveal that independent genetic variants associated with a complex phenotype converge on intermediate genes, with a major role for the HLA. PLoS Genet 2011; 7:e1002197. 24. Hsia, N. & Zon, L.I. Transcriptional regulation of hematopoietic stem cell development in zebrafish. Exp Hematol 2005;33:1007–14. 25. ee Jong, J.L. & Zon, L.I. Use of the zebrafish system to study primitive and definitive hematopoiesis. Annu Rev Genet 2005;39:481–501. 26. The ENCODE Consortium. An integrated encyclopedia of DNA elements in the human genome. Nature 2012;489:57–74. 27. Watkins, N.A. et al. A HaemAtlas: characterizing gene expression in differentiated human blood cells. Blood 2009;113:e1–e9.

116 Chapter 5 information Supporting underlies the SMIM1 underlies Vel bloodgroup 117

Chapter 5 Chapter 5

Supplementary Table S1. Gene coexpression network analysis

Predicted GO Biological Process Function P-Value Z-Score Hemoglobin metabolic process 1.34E-16 8.27 Erythrocyte development 1.15E-09 6.09 Heme metabolic process 1.19E-09 6.08 Heme biosynthetic process 1.46E-09 6.05 Erythrocyte differentiation 1.04E-07 5.32 Porphyrin5containing compound biosynthetic process 2.10E-07 5.19 Tetrapyrrole biosynthetic process 2.10E-07 5.19 Tetrapyrrole metabolic process 1.55E-06 4.81 Porphyrin5containing compound metabolic process 1.55E-06 4.81 Erythrocyte homeostasis 1.78E-06 4.78 Gas transport 3.99E-06 4.61 Iron5sulfur cluster assembly 5.75E-06 4.54 Metallo5sulfur cluster assembly 5.75E-06 4.54 Granulocyte differentiation 1.18E-05 4.38

Blood donors screened for Vel status by HA test 1 n = ~350,000 (100%)

Negative HA test 1 result n = 269 (0.08%)

Confirmation HA test 2 Vel weak Negative HA test 2 result n = 19 n = 76 (0.02%)

Exome sequencing Sanger sequencing n = 5 n = 43(+6)

Heterozygous Heterozygous Homozygous Homozygous Heterozygous Heterozygous SMIM*∆64_80 SMIM*∆64_80 SMIM*∆64_80 SMIM*∆64_80 nsSNP-M51R (USN14) SMIM*∆64_80 n = 3(+1) n = 1, USN3 n = 4, USN1,2,4,5 n = 38(+5) nsSNP-M51K (USN76) n = 19 USN16,34,37,(48)

USN48: Vel weak status Vel weak status USN14: Vel weak status postulated HA test 3 positive postulated postulated USN76: HA test 3 positive Vel weak Vel weak

Supplementary Figure S1. Vel- and Vel(+w) blood donors. Vel-typed donors were identified by phenotyping their red blood cells by a haemagglutination with an anti-Vel (HA test 1) and the Vel status of apparent Vel- donors was confirmed by HA test 2. This identified 76 Vel- and 19 Vel(+w) individuals; 48 of the Dutch and English Vel- donors were available for this study and another 6 Danish Vel- blood donors were included (numbers for Danish donors and related genotyping data are between brackets). The observed SMIM1 genotypes for the 73 samples are presented.

118 seven siblingswere notincludedthere. USN[UniqueStudyNumber]. caseswithanti-Vel ofthesixteen SMIM1 genotypes andDNA(top offigure) are presented the inthemain text; DNA sampleswere obtainedfrom clinicalcaseswithanti-Vel sixteen andfrom allseven siblings. The observed clinical caseswithanti-Vel were identifiedandseven siblingsfrom seven ofthe21cases were alsoenrolled. FigureSupplementary S2. Detected during pregnancy n = 3 = n Vel- clinicalcaseswithanti-Vel immuneantibodiesandtheirsiblings. Twenty-one Cases with anti-Vel anti-Vel Cases with Detected before SMIM* Homozygous Homozygous surgery n = 16 = n n = 16 16 = n n = 5,USN55,66,67,68,71 n = 4 = n ∆ 64_80 SMIM* Homozygous Homozygous n = 5 = n Vel- ∆ 64_80 No No clinical Siblings record n = 7 = n n = 9 = n n = 2,USN95,96 Heterozygous Heterozygous SMIM* Vel+ n = 2 = n underlies the SMIM1 underlies Vel bloodgroup anti-Vel without DNA ∆ 64_80 w Cases Cases with n = 5 = n 119

Chapter 5 Chapter 5

Supplementary Figure S3. Exome sequencing of 5 Vel- individuals The top four individuals are homozygous for the 17 nucleotide framehift deletion (hg19, chr1:Δ3691998A 3692014:GTCAGCCTAGGGGCTGT/A) in SMIM1, the bottom individual is heterozygous. DNA was enriched using the Roche Nimblegen SeqCap EZ Human Exome v3.0 protocol. Data was visualized using IGV.

Supplementary Figure S4.Differential binding of nuclear proteins at the intronic SNP rs1175550. Electrophoretic mobility shift assays [EMSA] in nuclear protein extracts from the myeloid erythroidK562 cell line cells showed higher protein affinity of the probe containing the A allele (major allele of the SMIM1 intronic SNP rs1175550, lane 1) than the G

120 rs1184341 isinstrong LD(r The positionofaSNPonthehorizontalby axisisdetermined theLDr2withcommonSNPrs1184341. the meta-analysis of GWAS of red blood cell parameters (red,right axis) (van der Harst et al, Nature, 2012). from axis)andanMCHCassociationp-value SNP hasbothanSMIM1eQTL (black,left associationp-value associationsignal. Eachcircle concentration[MCHC] representsmean corpuscular ortriangle aSNP. Each Supplementary Figure 1inmaintext). upstream there ofrs1175550,andisalsonucleosomedepleted inerythroblasts, bloodcellprecursor (see Figure S5.Colocalization oftheSMIM1eQTL association signal andthe red bloodcell 2 = 0.92)withrs1175550,theleadeQTL andGWAS SNP. rs1184341is288bp underlies the SMIM1 underlies Vel bloodgroup 121

Chapter 5 Chapter 5

Supplementary Figure S6. Alignment of human SMIM1 and its vertebrate orthologs. Genomic DNA alignment, including Ensembl orthologs of human SMIM1 and its manually identified zebrafish homolog. Only the alignment region surrounding the two exons with protein coding DNA are shown (two upstream noncoding exons are not shown). Four species (turtle, coelacanth, rat and tenrec) with large intronic insertions were included in the multiple alignment calculation, but are excluded from the plot for clarity. Sequence features were mapped via the human sequence and are shown above the alignment (GWAS SNP rs1175550, Deletion nucleotide 64<80, Transmembrane Region, SMIM1 Exons). b)Coding sequence DNA alignment, extracted from the genomic DNA alignment by keeping only alignment columns present in the human protein coding transcript. Ambiguous or missing characters are colored gray. Codon specific estimates from the PAMLM3 model analysis are shown below the alignment (Codon Selective Pressure) with avertical axis ranging from ω = 0 to 1; blocks are scaled according to the ω estimate, with sites under purifying selection colored blue, neutral selection colored gray, and positive selection colored red. The human SMIM1 CDS and protein sequence are shown directly above the alignment, with amino acid residue numbers shown below the sequence and a horizontal line marking the location of the cytoplasmic KCK protein motif. The estimated site- wise selective pressures across SMIM1 (bottom track) showed that most sites (~95%) are under strong purifying selection. The second exon shows especially strong conservation, with most sites showing evidence of strong purifying selection and no sites under apparent positive selection.

122 Chapter 6

Genetic screening for the Vel- phenotype circumvents difficult serological screening due to variable Vel expression levels

Lonneke Haer-Wigman1 Shabnam Solati1 Aïcha Ait Soussan1 Erik Beckers2 Pim van der Harst3 Marga van Hulst-Sundermeijer4 Peter Ligthart1 Dick van Rhenen5 Hein Schepers3 Tamara Stegmann1 Masja de Haas1 C. Ellen van der Schoot1

1 Sanquin Research, Amsterdam and Landsteiner Laboratory, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands 2 Maastricht University Medical Centre+, Maastricht, The Netherlands 3 University Medical Center Groningen, University of Groningen, Groningen, The Netherlands 4 Sanquin Diagnostic Services, Amsterdam, The Netherlands 5 Sanquin Blood Bank, Rotterdam, Netherlands

Manuscript under preparation Chapter 6

Abstract Background: Serological determination of the Vel- phenotype is challenging due to variable Vel expression levels. In this study we investigated the genetic basis for variable Vel expression levels and developed a high-throughput genotyping assay to detect Vel- donors.

Methods: In 107 Vel+w and 548 random Caucasian donors genetic variation in the SMIM1 gene was studied and correlated to the level of Vel expression. A total of 3,366 Caucasian, 621 Black and 333 Chinese donors were screened with a high-throughput genotyping assay targeting the SMIM1*64_80del allele.

Results: The Vel+w phenotype is caused by the presence of one SMIM1 allele carrying the major allele of the rs1175550 SNP in combination with in most cases a SMIM1*64_80del allele or in few cases the SMIM1*152T>A or SMIM1*152T>G allele. In ~6% of Vel+w donors genetic factors in SMIM1 could not explain the weak expression. We excluded the possibility that lack of expression of another blood group system was correlated with variable Vel expression levels. Furthermore, using a high-throughput Vel genotyping assay we detected two Caucasian Vel- donors.

Conclusions: Variable Vel expression levels are influenced by multiple genetic factors in SMIM1 and also by genetic or environmental factors. Due to this variable Vel expression serological typing of the Vel- phenotype is difficult and a genotyping assay targeting the c.64_80del deletion in SMIM1 should be considered to screen donors for the Vel- phenotype.

124 of rs1175550 on the non-deletion haplotype. of rs1175550onthenon-deletion weakened Vel thec.64_80del expression andthemajorallele deletionheterozygous carried effect ofrs1175550onSMIM1/Vel thatallindividualswith regulatory by thefact issupported reaction canoccurdueto intravascular hemolysisof Vel+ red bloodcells. with anti-Vel istransfusedwith Vel+ red bloodcellsasevere immediate hemolytictransfusion genetic variation thatleadsto the genetic variation Vel- phenotype. Recently, we othergroups andtwo identifiedthegeneticbasisof Vel antigenandthe phenotype. inthedonorpopulationwillcircumventphenotype thedifficultserological typingofthe Vel- human anti-Vel sera show heterogeneous levels of reactivity. due to large in individual variation Vel expression levels and because the scarcely available are notroutinelyfor typed Vel. Serological screening for the Vel- iscumbersome, phenotype of Variation in Vel expression levels may also berelated inintron to thers1175550variation 2 transcript levelsis associated compared withhigherSMIM1transcript to themajor Aallele. a bloodtransfusionhave atransfusionhazard situation. inanemergency Patients withantibodiesdirected to the Vel bloodgroup antigenandwhoare inneedof Introduction The geneticbasisofweak aimsofthecurrent studyisto theunderlying characterize Vel [Vel+ rare. Vel cantherefore onlybetransfusedwith Vel- red bloodcells, the Vel- is, phenotype however, SMIM1 (c.64_80del) thatabolishes thecomplete expression oftheSMIM1protein. the underlies Vel antigen and the Vel- is causedby a17-basepair deletion in phenotype molecular typing assay tomolecular typing identifydonorswiththe Vel- phenotype. levels ofthe Vel antigen.Furthermore, we developed andvalidated ahigh-throughput expression ofSMIM1ontheexpression andto theeffect characterize ofthegeneticvariation p.Met51Arg). acid substitution, were detected (c.152T>A encoding p.Met51Lys and c.152T>G encoding mutations, atthesamenucleotidepositionofSMIM1,butresulting inadifferent amino Furthermore, individualswithweak intwo Vel expression missense singleheterozygous two expression ofthe Vel antigeniscorrelated presence withtheheterozygous ofthisdeletion. variation is a regulatory region isaregulatory forvariation levels. SMIM1transcript Chilcotin Indians inCanada. Chilcotin Indians and the Vel- occasionallyintheAsian hasbeendescribed population andthe phenotype SMIM1. 1,6 The overall ofthe frequency Vel- intheCaucasian populationis~0.025% phenotype 11,14 Anexpression quantitative traitlocus[eQTL] studyshowed thatthers1175550 11

6-9 The availability of The availability Vel- donorsiseven lower, becausedonors 11 11-13 The SMIM1protein encodedby SMIM1 11,14 The minorGalleleofrs1175550 10 Genetic screening Genetic for the Vel- Molecular basisofweak Molecular Vel expression 3-5 Patients withanti- 1,2 When apatient 11-13 14 Weak This 11,12 125 w ]

Chapter 6 Chapter 6

Material and methods Samples and DNA isolation All cases were included after written informed consent was provided. In 2005, 141 donors with weak Vel expression and five Vel- donors were identified upon serological screening of 10,500 donors with a human polyclonal anti-Vel serum. Nineteen of the 141 Vel+w donors have also been reported in previous work.11 For 107 of the 141 donors material collected in 2005 was available. A total of 3,366 Caucasian, 621 Black and 333 Black Asian random donor samples were included. Samples of Dutch Caucasian blood donor samples were collected by Sanquin Blood Supply, The Netherlands. Asian and Black donor samples were collected by The South African Blood Transfusion Service in Johannesburg South-Africa. Additional Black donor samples were collected by The Ethiopian Red Cross Society Transfusion Service in Addis Ababa, Ethiopia and The Red Cross Blood Bank Curaçao in Willemstad, Curaçao. DNA was extracted from white blood cells using a DNA extraction kit (QIAamp DNA Blood Mini Kit).

Serology Screening for the Vel- phenotype on 10,500 donors with blood group O was performed using undiluted human serum (NL15) in the Cellbind card technique. All donors with a negative agglutination reaction were tested in a second round using standard tube agglutination technique with NL15 and one to three additional human anti-Vel sera (NL25, NLVzO and/or NLVzA). A donor was determined to have weak Vel expression levels when at least one of the Vel sera gave a positive reaction. Vel antigen expression was examined via agglutination using Liss/Coombs cards (Bio-Rad, Veenendaal, The Netherlands) with anti-Vel serum of NL25 on En(a-), S-s-U-, Rh- (regulator type), In(Lu), Lu(a-b-), K0, Kx-, Fy(a-b-), Jk(a-b-), Gy(a-), Co(a-b-), JMH-, p and Ge:-2,-3 red blood cells and on red blood cells of a patient with paroxysmal nocturnal hemoglobinuria. Agglutination strength was assessed after serum and red blood cells were incubated for 15 minutes at 37ºC with a subsequent spin for 10 minutes.

Flow cytometry Vel expression on red blood cells was measured using flow cytometry with immunopurified human polyclonal anti-Vel (NL25 and/or NL04-01). The purified human polyclonal anti-Vel was obtained after incubation of human serum containing anti-Vel with group O Vel+ red blood cells. After washing, the antibodies were eluted from the red blood cells using the Gamma ELU-KIT II (ImmucorGamma). The purified anti-Vel was incubated for 30 minutes at room temperature in a 3% (vol/vol) red blood cell suspension. Subsequently, red blood cells were washed and Alexa-488 labelled goat anti-human-IgG (Molecular Probes) was added in a 1:500 dilution and incubated for 30 minutes at room temperature. Red blood cells were washed and analyzed by flow cytometry (LSRFortessa cell analyzer, Becton Dickinson). Statistical

126 (M1613, Pelicluster), CD47(427603, B.V.,CD44 (G4426Dickinson 559942,Becton Breda, (LA18.18),CD36 RhAG theNetherlands), (BS56),GPD(BRIC4),CD35(E11), Duffy(64-4A8),Band3(BRIC71andBRIC90),ICAM-4 (Bric18), [GP]A (AME1, Pelicluster), (Lu-4F2), (BRIC69)B-CAM Kell (703.7),Rh GPB(OSK29-4), RhCE red cellmembraneproteins; specific monoclonalantibodies[Moabs] murine forGlycophorin Flow to above, measure asdescribed expression cytometry, wasperformed levels ofother were repeated andthesecondmeasurement wasusedfor analysis. Samples withameanfluorescence oflessthan400(n=36)or intensity over 3500(n=17) with the average per experiment intensity mean fluorescence of all experiments. intensity theaverage by correcting of thesampleswasnormalized perexperiment meanfluorescence batch wise;themeanfluorescence red intensity bloodcellsofthe548donorswasperformed t-test. usinganunpaired two-tailed analysis wasperformed The Vel expression levels ofthe Genetic typing of the minor G and major A allele of the typing rs1175550 SNP of Genetic TaqMan assay antibody.secondary Alexa-488 labelledgoatanti-human-IgGorPEanti-mouseIgGwasusedasa (AB5) wereagainst RhD and human MoAbs used. When antibodies were not directly labeled, B.V.), B.V.), Dickinson CD59(p282,Becton CD71(M3162,Pelicluster) (67A4) andE-Cadherin allele confirmation sequencingwasperformed. SMIM1*64_80del alleleconfirmation TaqMan assay. for the minorGalleleofrs1775550orpositive allcaseshomozygous for In the pool was positive all samples of the pool were tested separately using the of apoolwasinthesamerange astheDCtvalueofpositive control sample. When a (average Ct valueofthec.64_80del TaqMan assay –average CtoftheALBTaqManassay) StepOnePlus software. Apoolwas “positive” for theSMIM1*64_80delallelewhenDCtvalue for themutation(negative control) usingthe were along. taken Dataanalysiswasperformed negative for the mutation (positive control) and a sample containing 150 ng DNA negative themutationinabackground of150ngDNA containing 1.6ngDNAfrom adonorcarrying fluorescence signal wasmeasured. Pools were tested induplo.In eachrunapositive control the of15secondsat95°Cand1minute at60 °C attheendofeachcycle 95°C, 50cycles Biosystems) andreverse and0.3μMforward primers. PCRconditionswere: 10minutes at volume (Applied of25mL,containing100-200ngDNA,12.5 mL SYBR®Green Mix PCRMaster Biosystems) according to manufacturers’ inatotal reactions were protocol. performed short In a finalconcentration of150 ng/mL.AllassaysonaStepOnePlus were (Applied performed These ofDNA of 45Caucasian poolscontainedamixture or90Asian orBlackdonorswith and theALB TaqMan andtheSMIM1*64_80del TaqMan onpoolsofDNA. were alsoperformed to measure total DNA concentration. All TaqMan assays on single samples were performed for theSMIM1*64_80del allelewas developed. A TaqMan assay for theALBgenewasused A allele. For high-throughput of the genotyping Vel- a phenotype TaqMan assay specific usingtwo performed TaqMan assays containing, respectively, specific primers for theGor R&D Systems Europe Ltd.), Dickinson CD55(1A10,Becton Molecular basisofweak Molecular Vel expression SMIM1*64_80del SMIM1 was 127

Chapter 6 Chapter 6

DNA sequencing Primers were developed to sequence all exons and introns of SMIM1. The polymerase chain reaction [PCR] was performed on a Veriti thermocycler (Applied Biosystems,) in a total volume of 20 mL, containing 50-150 ng DNA, 10 mL of 2x GeneAmp Fast PCR Master Mix (Applied Biosystems), 0.5 mM forward and reverse primer. PCR conditions were: 10 seconds at 95°C, 35 cycles of 10 seconds at 95°C and 15 seconds at 64°C, followed by 1 minute at 72°C. PCR products were purified using illustra ExoProStar (GE Healthcare) according to manufacturer’s protocol. The sequence reaction was performed on a Veriti thermocycler in a total volume of 20 mL, containing 1 mL of purified PCR product, 1 mL 2.5x BigDye Terminator v1.1 Cycle (Applied Biosystems), 3.5 ml 5x BigDye Terminator Buffer (Applied Biosystems) and 0.25m M forward or reverse primer. Sequence conditions were: 25 cycles of 15 seconds at 95°C, 10 seconds at 50°C and 4 minutes at 60°C. Sequence products were analyzed on a 3130 Genetic Analyzer (Applied Biosystems).

SMIM1 transduction The mLwpRRLsSMIM1itNGFR vector was used as a wild-type SMIM1 construct. Vectors containing the c.152T>A and c.152T>G were made from the wild-type SMIM1 construct using site directed mutagenesis. HEK293T cells (7.5 × 105) were plated in 6-cm2 plates and lentivirally transduced with the three different constructs and a mock construct. Forty-eight hours after transduction, cells were harvested and screened by flow cytometry using purified human polyclonal anti-Vel (NL25 and NL04-01) (see flow cytometry) and an antibody to CD271 (Milteny Biotech B.V.) was taken along as a transduction control.

Results The major allele of rs1175550 is correlated with reduced Vel expression levels To determine the effect of the rs1175550 SNP on Vel expression levels, 548 random Caucasian blood donors were genotyped for the rs1175550 SNP and SMIM1*64_80del allele. The frequency of the major A allele of rs1175550 (78.6%) and of the minor G allele (21.4%) of this cohort was the same as the European population frequency (see Web Resources). Twenty of the 548 donors were heterozygous for the SMIM1*64_80del allele. Vel expression levels in the 548 donors were measured via flow cytometry using purified anti-Vel (NL25). Donors homozygous for the minor G allele had significantly higher Vel expression levels compared to donors heterozygous for rs1175550 and Vel expression of these donors was significantly higher compared to donors homozygous of the major A allele of rs1175550 (Figure 1). Furthermore, donors who were heterozygous for the SMIM1*64_80del allele had significantly lower Vel expression levels compared to donors with the same rs1175550 haplotype, but without the c.64_80del deletion (Figure 1). Interestingly, donors who were homozygous for the major A allele of rs1175550 had roughly the same Vel expression levels as donors who had one

128 SMIM1*64_80del alleleon Vel expression levels, aswe already postulated. thatthepresence oftheminorGallelers1175550neutralizesIndicating theeffect ofthe theminorGalleleofrs1175550. haplotype SMIM1*64_80del alleleandonthenon-deletion 80del allele and the minor G allele of rs1175550 on the non-deletion haplotype (p=0.67). haplotype 80del alleleandtheminorG ofrs1175550onthenon-deletion for the A allele of homozygous rs1175550 and compared to SMIM1*64- donors with a heterozygous rs1175550). No difference in Vel expression level was determinedin donors without the c.64_80del and for forin donorsheterozygous majorAalleleof rs1175550andp=<0.0001 indonorshomozygous levels compared to for donorsheterozygous c.64_80del (p=0.0467 withthesamers1175550haplotype withoutthec.64_80delallele ofrs1175550(p=<0.0001).Donors hadsignificantly higher Vel expression these donorshadsignificantly higher Vel expression levels compared for theA to donorshomozygous for theGallelecompareddonors homozygous to for donorsheterozygous rs1175550(p=0.0004), 80del allele)were alongasnegative controls. taken Vel expression levels were significantly higherin unpaired two-sided T-test. Four donorsdetermined Vel for negative theSMIM1*64- (homozygous with an SMIM*152T>Aallele. heterozygous Statisticalanalysiswasperformed donorswhocarry two fluorescence intensity, error barsindicate the2.5%and97.5%percentile. two blackdotsdepict The levels werecytometry. Box usingflow determined plotshows 25%and75%percentile and average in SMIM1. via inthedonorswasdetermined The SMIM1genotype TaqMan assays and Vel expression Figure 1.Vel expression for levels andc.64_80del in552donorsgenotyped thers1175550variation

n = 35 162 c.64_80del wt/wt wt/wt wt/wt wt/ rs1175550 Mean Fluorsence Intensity (arbitrary units) 10000 1000 100 GA AA AAA AA AG AA AG GG *** 331 3 *** * p=0.67 genotype *** *** wt/ 17 4 *** Molecular basisofweak Molecular Vel expression / 11 129

Chapter 6 Chapter 6

The heterozygous expression of SMIM1*64_80del allele and the major A allele of rs1175550 on the non-deletion haplotype causes the Vel+w phenotype In 2005, 10,500 Dutch donors were serologically typed for the Vel antigen. In five (0.047%) donors the Vel- phenotype was detected and in 141 (1.3%) donors weak Vel expression was detected. In 107 of the 141 Vel+w donors leukocytes were available for DNA isolation. To investigate whether a genetic variation in SMIM1 underlies the Vel+w phenotype, all exons and part of intron 2 (hg19, chr. 1: g.3691316-3691832) in SMIM1 were sequenced. In 101 of the 107 donors the Vel+w phenotype could be assigned to the heterozygous presence of the SMIM1*64_80del allele and the presence of the major A allele of rs117550 on the non-deletion haplotype. The remaining six Vel+w donors were all homozygous for the major A allele of rs1175550, but were negative for the c.64_80del deletion. None of the Vel+w donors was positive for the minor G allele of rs1175550. The serological typing results from the series in 2005 showed that the weak agglutination reactions in five cases without the deletion were of comparable strength to the average levels of agglutination observed with the donors heterozygous for the SMIM1*64_80del allele (Table 1). The original data of UCN13 could not be retrieved anymore. Furthermore, we verified the weak Vel expression levels of these six donors by flow cytometry analysis with immunopurified NL25. Red blood cells of the donation in 2005 were available of the six donors without the c.64_80del deletion in SMIM1 and of 91 Vel+w donors heterozygous for the SMIM1*64_80del allele, which were taken along as controls. As an additional control, red blood cells of ten random donors homozygous positive for the major allele of rs1175550 were taken along. The Vel expression of donors without the c.64_80del deletion did not significantly differ from that of the 91 Vel+w donors heterozygous for the SMIM1*64_80del allele, while expression levels were significantly lower compared to ten control samples homozygous for the major allele of rs1175550 allele (Supplementary Figure S1).

Table 1. Vel expression levels of six donors with weak Vel expression levels without the c.64-80del in SMIM1 PEG technique PEG technique PEG technique PEG technique NL15 anti-Vel in NL15 anti-Vel in NL15 anti-Vel in NL25 anti-Vel NLVzA anti-Vel in anti-Vel NLVzA NLVzO anti-Vel in anti-Vel NLVzO

UCN technique Cellbind 13 0 42 0 0 0.5 0 1 57 0 0 2 0 0 30 0 0 2 0.5 1 40 0 0.5 1 2 2 26 0 0.5 2 1 2 Vel+w donors heterozygous for 0 0.3 1.8 1.1 0.9 SMIM1*64_80del Agglutination in Vel+ donors 1-2 2-3

130 these donorshasalready previously beendescribed c.152T>A missense mutation (encoding p.Met51Lys) in these ninedonorswere sequencedfor allexons donorsthe two andintrons ofSMIM1.In investigate inSMIM1are whetherothergeneticfactors responsible for weak Vel expression, to levelpresence the heterozygous could not be linked of the Genetic analysisGenetic ofnine Vel+ (Table 2).Except(Table inone Vel+ ofintronicand thedistribution SNPs wasnotdifferent compared to the European population (Figure theremaining ninedonorsnomutationswere 1).In present inthecodingsequence donorsislowerin thesetwo compared to for donorsheterozygous theSMIM1*64_80del of rs1175550. for for theminorallele in expression the majororhomozygous donorshomozygous between on expression levels oftheothermembrane proteins, becausewe difference didnotdetect Vel+ and Vel- red bloodcells(Figure 2).Moreover, thelevel of Vel expression hadalsonoeffect red cellmembraneproteins CD36, CD47,CD55,CD71and E-cadherin wasdetected between levels of the GPA, GPB, RhD, Kell,Duffy, B-CAM, RhCE, Band GPD, 3,ICAM-4, RhAG, CD35, CD44, membrane proteins are influenced by thelevel of Vel expression. Nodifference inexpression proteins. addition,we investigated In whethertheexpression levels ofotherred blood cell whichlackexpression hemoglobinuria, ofallGPI-linked a patientwithparoxysmal nocturnal or JMH. CD55,RhAG Aquaporin 1,Xk, Vel expression onthered wasalsonormal bloodcellsof GPB, RhD, Kell,Duffy, B-CAM, RhCE, UreaB, GPC,GPD, transporter Do,Acetylcholinesterase, that themembraneexpression ofthe Vel antigenisnot dependentontheexpression ofGPA, of redwas detected onallthese types blood cells(data notshown) andwe therefore conclude Colton, Kx, Cromer,Dombrock, RHAG or the JMH blood group systems. Normal Vel expression complete expression Lu,blood cellsthatlacked oftheMNS,Rh, Kell,Duffy,Kidd,Gerbich, Yt, expression ofotherred bloodcellmembraneproteins, Vel expression onred wasdetermined withotherproteins.interact To investigate whether Vel expression is dependent on the The ofthe protein thatcould Vel part residues attheextracellular cysteine antigenhastwo membrane proteins Vel expression levels are not correlated withexpression levels ofother red bloodcell elementasrs1175550. is present inthesameregulatory Dutch donorsorin279European inEnsembl(See donorsdescribed Web Resources). This SNP intronic SNPrs191041962. The minoralleleofrs191041962wasnotdetected in91random with the same haplotype. Furthermore, in six of the 107 Vel+ hadreproducibly haplotype) lowernon-deletion Vel expression levels compared to donors heterozygous forheterozygous the for theminoralleleofrs1175550andonedonor allele ofrs1175550,onedonorhomozygous of548randomdonors, for three the major donors(one thecohort donorhomozygous In SMIM1*64_80del w donor, positive whowasheterozygous for theminoralleleof w donorswithunexplained weak expression allele and carrying the minor allele of rs1175550 on the allele and carrying 11 . Interestingly, the Vel expression levels SMIM1 was detected (Table 2), one of w donors the weak expression Molecular basisofweak Molecular Vel expression SMIM1*64_80del allele. To allele 131

Chapter 6

Chapter 6 c.*58G>A

A GA GA GG GG GG GG GG GG GG

3’ UTR 3’ c.152 T>A c.152

A TT TT TT TT TT TT TT T A T A

Exon 4 Exon rs71634364

A

GA GA GA 8% GG GG GG GG GG GG 0.06 rs2282455 Intron 3 Intron

C TT TC TC TC TC TC

CC CC CC 0.25 42% c.64_80del

D

wt/ D wt/wt wt/wt wt/wt wt/wt wt/wt wt/wt wt/wt wt/wt

Exon 3 Exon rs1175549

C

CC AA AA AA AA AA AA AA AA 0.54 21% rs9424296

A

CC CC CC CC CC CC CC CC CA 7% 0.30 10% rs1175550

G

AA AA AA AA AA AA AA AA GG 0.54 19% 21% rs6673829

A

AA GA GA GA GG GG GG GG GG 0.78 29% 32% rs191041962

T

C T CC CC CC CC CC CC CC CC 0% 0% 0.00 rs70940313

ins ins ins wt/ ins/ 0.97 21%

Intron 2 Intron wt/wt wt/wt wt/wt wt/wt wt/wt wt/wt wt/wt rs2794732

A

AA AA GA GA GG GG GG GG GG 0.87 45% rs143872648

D

D /

wt/ D wt/ D wt/wt wt/wt wt/wt wt/wt wt/wt wt/wt donors rs1184341 w

T TT

CT CT CC CC CC CC CC CC 0.96 22% rs2840327

A

AA GA GG GG GG GG GG GG GG 0.72 10% rs34961462

T TT

CC CC CC CC CC CC CC CC 0.79 12% rs2275819 A AA GA GG GG GG GG GG GG GG 0.21 10% Intron 1 Intron test) 2 Minor allele nucleotide 26-566 UCN 42 Dutch frequencyDutch of minor allele† population European frequency of minor allele 26-561 26-175 13 30 40 26 57 p value (Chi genotype determined in nine Vel+ 2. SMIM1 genotype determined in nine Table † Determined donors on 91 random Dutch

132 has previously been described heterozygous in a donor with very weak inadonorwithvery has previouslyheterozygous beendescribed Vel expression levels. p.Met51Lys SMIM1allele. to awildtype next The c.152T>G mutationencodingp.Met51Arg In this study we identified two thisstudywe identified In Vel+ The c.152T>A andc.152T>G mutations inSMIM1causeweak Vel expression red cellmembraneproteinsCD36, CD47,CD55,CD71andE-cadherin wasdetected. expression ofthe GPA, GPB, RhD, Kell,Duffy, B-CAM, RhCE, GPD, Band3,ICAM-4, RhAG, CD35,CD44, allele of rs1175550 (p = <0.001) and between Vel+ donors and Vel- donors (p = 0.035). No differential Vel+ for theminorGalleleofrs1175550and donorshomozygous Vel+ for themajor donorshomozygous with anunpaired two-sided T-test. The expression ofthe Vel antigenwassignificantly different between donors. Statistical analysis of the expression levels between Vel+ and Vel- red blood cells was performed - dotted black line).Expression levels were oneighttill24 determined Vel+ donorsandfour tilleight Vel- - grey line)andon Vel- for red theSMIM1*64_80del allele(open squares; bloodcellshomozygous mean - blackline),on Vel+ for red themajorAalleleofrs1175550(grey bloodcellshomozygous rounds; mean proteins on Vel+ for red theminorGalleleofrs1175550(blackdiamonds;mean bloodcellshomozygous Scatter plots showcytometry. the mean fluorescence of the different intensity red blood cell membrane Figure 2.Expression levels ofnineteen red bloodcellmembraneproteins were viaflow determined specific mutations were expressed inHEK T293cells. The c152T>G(p.Met51Arg) variant containing the assay. constructs andtwo amockconstruct construct, ASMIM1wild-type in the transmembrane region (Figure 3a). The effect of the c.152T>A and c.152T>G mutations regionThese missensemutations are located oftheprotein inaconserved nearorjustwithin Mean Fluoresence Intensity (arbitrary units) 100000 SMIM1 ontheexpression level of Vel studiedusingaheterologous expression wasfurther 10000 1000

Vel (25 100 10

0 GpA 646 ) (AME-1 GpB (OSK29-4

)

RhD (AB5))

RhC E (7 03-7) Rh B-CAM (Bri c6 9) (LU-4F2)

Kell (B

R Duffy ( IC 1 8)

w 6 Band3 (B4-4A8) donorswiththec.152T>A missensemutationencoding

R Band3 (B IC 7

Antigen (Antibody) 1 ) R IC IC AM-4 (BS 956 0 )

GpD (Br )

ic4) CD35 (

CD44 ( E1 1 ) G RhA 4 4 G 26) ( LA18.18) CD36 (

M1 Molecular basisofweak Molecular Vel expression CD47 ( 613 ) 4 27603) CD55 (

1 A1 CD59 ( 0 ) p CD71 ( 2 8 2) Vel- Vel+(rs1175550 AA) GG) Vel+ (rs1175550 M3 162 E Ca ) d he rin 133 11

Chapter 6 Chapter 6 caused almost the complete absence of the Vel antigen, while the c.152T>A (p.Met51Lys) variant caused a reduction of ~70% of the Vel antigen (Figure 3b). The PolyPhen software predicts that the p.Met51Arg substitution has a damaging score of 0.799 (the closer to 1, the more predicted damaging effect of the mutation) and the p.Met51Lys substitution has a damaging score of 0.5.15 a)

b) Negative non-transduced control Positive wild type SMIM1 control SMIM1*152G allele SMIM1*152A allele Count

Fluorescence Intensity (Arbitrary Units) Figure 3. Effect of missense mutations in SMIM1. a) Conservation of the SMIM1 gene via cross-species sequence alignment of protein homologues in ten different species. Conserved amino acids are colored black, amino acids with the same properties are colored grey, non-conserved amino acids are colored white. Bars below indicate the degree of conservation. The amino acid location bar indicates the putative position of the amino acid in the

Genetic screening for Vel- blood donors Because the Vel- phenotype is caused by a single genetic defect, genetic screening for donors with the Vel- phenotype can be easily performed. A TaqMan assay, sensitive enough to detect one positive sample in a pool of 90 samples (data not shown), was developed for the SMIM1*64_80del allele. A total of 3,366 Caucasian (including the cohort of 548 donors mentioned in the first paragraph of the results), 621 African black and 333 Asian donors were screened with this TaqMan assay in pools containing 45 or 90 donors. In total 94 Caucasian,

134 gene (due to presence heterozygous oftheSMIM64_80delallele)have roughly halfthe Vel gene dosageeffect for theSMIM1gene. RBCofdonorswhohave onlyonefunctional Vel forheterozygous rs1175550 hadintermediate expression addition,we levels. seeaclear In correlated with increased Vel expression levels compared to themajor A allele, anddonors responsible for variable Vel expression levels onred bloodcells;theminorGalleleisstrongly thisstudy, levels.and higherSMIM1transcript In we show thatthers1175550SNPisalso Furthermore, Fermann region, has lower withnuclear binding affinity proteins compared to the major A allele. previously region,contains aregulatory thatmay influence the and Vel expressionwe inerythroblasts positiveand Chinese donors heterozygous for the their race, theSNPs surrounding thec.64_80del were sequencedinCaucasian, black African hence allpersonspositive forthesameancestralallele, regardless thec.64_80del, of carry To decidewhetherthe17nucleotidedeletioninSMIM1occurred inasinglegeneticevent, phenotype. but larger population studiesare ofthe frequency neededto thecorrect determine Vel- donors is Vel-. ofthe The frequency Vel- population in Black and Asian populations is lower, (95% CI0.01%–1.19%).Usingthisallelefrequency, we calculate that~1in5000Caucasian (95% CIblackpopulation 0.15% –0.98%)and0.60% in theAfrican Chinese population SMIM1*64_80del inatube agglutination test withNL25serum.allele wasconfirmed ofthe The allelefrequency allele. The expected forVel- theSMIM1*64_80del donorshomozygous ofthetwo phenotype in allcases, except Caucasian for two samplesthatwere theSMIM1*64_80del homozygous presence ofthesecasesdetected theheterozygous Sequencing oftheSMIM1*64_80del seven blackandfour African Asian donorswere positive for theSMIM1*64_80del genetic variations in SMIM1 genetic variations difficult.phenotype The Velantigeniscoded bytheSMIM1 expressionThe highlyvariable levels ofthe Vel serological ofthe antigen makes typing Vel used to screen donorsfor the Vel- phenotype. Vel expression levels. Furthermore, thedeveloped highthroughputassay canbe genotyping Thehaplotype. deletion presence of the this study we showIn that weak Vel expression is caused most byoften the heterozygous Discussion to thec.64_80dellinked deletioninSMIM1. rs9424296 andtheminorallelesofrs2282455rs71634364(data notshown), thatare themajorallelesofrs1181893,rs6673829,rs1175550and orhomozygous), (heterozygous 11 showed thattheminorGalleleofrs1175550SNP, located inthis regulatory SMIM1*64_80del allele and the major A allele of the rs1175550 SNP on the non- allele was1.46%(95%CI1.17%-1.74%)intheCaucasian population,0.56% et al. mutations also cause weakened SMIM1*152T>A andSMIM1*152T>Gmutationsalsocauseweakened 14 are responsible for expression. this variable 2 of Intron determined a determined strong the minor G allele associationbetween SMIM1*64_80del gene andwe investigated which Molecular basisofweak Molecular Vel expression allele. All donors carried carried SMIM1 allele. allele 135

Chapter 6 Chapter 6 expression compared to RBC of donors with two functional SMIM1 alleles with the same rs1175550 haplotype. The lowest Vel expression levels were found in donors carrying one SMIM1 allele with major A allele of rs1175550 and one SMIM1*64_80del allele. Interestingly, the minor G allele of rs1175550 on the non-deletion haplotype overcame low expression caused by the presence of the SMIM64_80del allele. In a cohort of >10,000 donors screened for the Vel phenotype we identified the Vel+w phenotype in 1.3% of the donors. The frequency of donors with a SMIM1*64_80del allele and the major A allele of rs1175550 on the non-deletion haplotype (which gives the lowest Vel expression levels) is also ~1.3% (taken in account an allele frequency of 1.6% for the SMIM64_80del allele and 79% for the major A allele of rs1175550). Indeed, 94% of the Vel+w donors was heterozygous for the c.64_80del deletion and homozygous for the major A allele of rs1175550. Hence, in most individuals the Vel+w phenotype is caused by one SMIM1 allele carrying the major A allele of rs1175550 in combination with a SMIM1*64_80del allele. The remaining six Vel+w donors were homozygous for the major A allele of rs1175550, but were negative for the c.64_80del deletion. Furthermore, in 548 random Caucasian donors another three Vel+w donors were detected of whom the low Vel expression could not be accounted to presence of the c.64_80del deletion. Two of these nine Vel+w donors were heterozygous positive for the SMIM1*152T>A allele, that is already described to weaken Vel expression levels. In the other seven donors no mutations were present in the coding sequence and no common intronic variation was detected. In one donor the minor allele of rs191041962 was heterozygous present, while the minor allele of rs191041962 was not detected in 367 Caucasian control samples. The rs191041962 is located in the same regulatory element as rs1175550 and it is possible that the minor allele of rs191041962 has an additional enhancing effect on transcription binding, and thereby further reduces Vel expression levels. Nevertheless, in at least six Vel+w donors no mutations and variations in SMIM1 could explain the weak Vel expression levels. We therefore conclude that other genetic factors outside the introns and exons of SMIM1 or environmental factors also influence Vel expression levels. For that reason, we investigated whether other blood group antigens have an effect on Vel expression levels. No differential Vel expression levels were discovered on red blood cells that lacked the complete expression of the MNS, Rh, Lu, Kell, Duffy, Kidd, Gerbich, Yt, Dombrock, Colton, Kx, Cromer, RHAG or the JMH blood group systems. We therefore conclude that Vel expression is not dependent on the expression of these other blood group antigens. The other way around, no difference was detected in expression levels of GPA, GPB, RhD, RhCE, B-CAM, Kell, Duffy, Band 3, ICAM-4, GPD, CD35, CD44, RhAG, CD36, CD47, CD55, CD71 and E-cadherin red cell membrane proteins between Vel+ and Vel- red blood cells. Two Vel+w donors were identified to be heterozygous positive for the SMIM*152T>A allele encoding p.Met51Lys and we previously11 described one donor heterozygous positive for the SMIM*152T>G allele encoding p.Met51Arg. In a heterologous expression assay the c.152T>G

136 Storry Storry blackandAsianthe African seemsto belower (~1in30,000donorshasthe Vel- phenotype). c.64_80del in deletion,aswasdetermined Vel- donors. deletion were positive for allSNPs inintron 2 andintron 3ofSMIM1thatareto the linked blackand Chinesedonorsthedeletionwasidenticaland withthe as intheAfrican as a nucleotide single deletion genetic occurredevent, because most likely upon reduction. [kD] protein31 kiloDalton complex,whereas thisbandmigrates to asinglebandof18 kD Vel protein. oftheSMIM1/ motif ataminoacidposition67till71indicates apossiblehomodimerization allele, whichcanbeexplainedwhenSMIM1formstype ahomodimer. theGXXXG Indeed, that themutationhasadominantnegative effect on Vel expression from the remaining wild- one functional Vel presence allele (heterozygous of the with a p.Met51Lys mutation had far lower expression levels compared than donors with only amino acid, respectively, atthisposition.Unexpectedly, thered donors blood cellsofthetwo while thec.152T>A andc.152T>G mutationsencodethepositively charged lysineorarginine SMIM1/Velwild type protein anon-polarmethionineaminoacidispresent atposition51, that thep.Met51Lys substitution is lessdamaging thanthep.Met51Arg the substitution.In reduced the Vel expression levels with ~70%. Moreover, the PolyPhen also predicts software mutation almost completely abolished Vel expression levels, whilethec.152T>A mutation serology; ~1in4,000 serology; Vel- individualsintheCaucasian population. in theCaucasian population, whichisroughly by determined thesameasfrequency SMIM1*64_80del allelewe thattheoccurrence determine ofthe Vel- is~1in5,000 phenotype Caucasianand detected two Vel- donors. Furthermore, ofthe usingtheallelefrequency and validated ahighthroughputassay to genotyping screen donorsfor the Vel- phenotype assay canmore donorswiththe easilydetect Vel- thanserology. phenotype We developed Because onlytheSMIM1*64_80delalleleiscorrelated withthe Vel- agenotyping phenotype reduction of Vel expression levels. the c.152T>A mutation is present on the wild-type any inEnsembl, oftheotherpopulationsdescribed SMIM1 (c.*58G>A), which is not detected in the 279 Caucasian individuals or individuals of of SMIM1/Vel. Althoughwe cannotformally mutationinthe3’UTRof exclude thattheextra oftheSMIM1protein,dimerization red henceinhibitcorrect cellmembraneincorporation (0.56%, 95%CI0.15%-0.98%)ofthisstudy. Furthermore, aswasalready proposed (0.09%,95%CI0.02%-0.17%)compared Americans African to blackpopulation theAfrican population. The American [NHLBI] Exome Project, Sequencing thattheSMIM1*64_80delalleleispresent intheAfrican et al. 12 12 already Lung, demonstrated, usingdataoftheNationalHeart, andBloodInstitute Moreover, undernon-reducing circumstances the Vel by a antigeniscarried 13 The missensemutationscausing thep.Met51 alteration mightinhibit SMIM1*64_80del allelehad, however, alower inthe frequency SMIM1 allele and responsible for additional observed inbothindividualspositive for observed SMIM64_80del allele). This suggests Molecular basisofweak Molecular Vel expression 6,11 within The frequency in the Caucasian 12,13 , the17 137

Chapter 6 Chapter 6

In conclusion, weakened expression of the Vel antigen is most often caused by the heterozygous SMIM1*64_80del allele in combination with the minor A allele of rs117555 on the non-deletion haplotype. This weakened Vel expression makes serological screening for the Vel- phenotype difficult. An easy to perform high throughput genotyping assay can correctly identify Vel- donors and using this assay the availability of Vel- donors and therefore Vel- red blood cells can be increased, making safe transfusion possible for individuals who have anti-Vel.

Acknowledgements The authors like to thank R. Bijman for her assistance with flow cytometry experiments, J. Hooydonk (The South African Blood Transfusion Service (Johannesburg, South Africa) for collection of samples from Black and Asian donors, G. Tesfaye (The Ethiopian Red Cross Society Transfusion Service Addis Ababa, Ethiopia) and A.J. Duits (The Red Cross Blood bank Curaçao, Willemstad, Curaçao) for the collection of samples from black donors.

Web Resources Population frequency of the rs1175550 and rs191041962 SNPs, respectively. http://www.ensembl.org/Homo_sapiens/Variation/Population?db=core;r=1:3691028- 3692028;v=rs1175550;vdb=variation;vf=928760, http://www.ensembl.org/Homo_sapiens/ Variation/Population?db=core;r=1:3690982-3691982;v=rs191041962;vdb=variation; vf=51904433. Accessed 160813.

138 15. 14. 13. 12. 11. 10. 9. 8. 7. 6. 5. 4. 3. 2. 1. References

Mol.Genet. 2001;10(6):591-7. Mol.Genet. Sunyaev S,Ramensky V, I,Lathe Koch W, P. AS,Bork Kondrashov Prediction humanalleles. of deleterious Hum. intermediate genes, withamajorrole for theHLA.PLoS.Genet. 2011;7(8):e1002197. al. Trans-eQTLs reveal converge associated withacomplexphenotype that independent geneticvariants on Fehrmann RS, Jansen RC, Veldink JH, Westra HJ, Arends D, Bonder MJ, Fu J, P, Deelen Groen A, et HJ, Smolonska Disruption ofSMIM1causesthe Vel- EMBOMol.Med. bloodtype. 2013;5(5):751-61. Ballif BA,Helias V, Peyrard T, C,SaisonLucien Menanteau N,Bourgouin JP, S,Le M,Cartron Gall L. Arnaud for anullalleleofSMIM1definesthe Vel-negative blood group 2013;45(5):537-41. Nat.Genet. phenotype. MK, M,Christophersen JR,Joud B,Thuresson OlssonML.Homozygosity BN,Nilsson B,Storry AkerstromB, Sojka 2013;45(5):542-5. Nat.Genet. E,Farrow S,etal. the SMIM1underlies Bielczyk-Maczynska Vel bloodgroup andinfluences red bloodcelltraits. Cvejic A,Haer-Wigman L,Stephens J, M,SmethurstPA, Kostadima P, Frontini E,Bertone M,vandenAkker of Vel system antibodies. Vox Sang. 1968;15(2):125-32. Adebahr ME,AllenFH,Jr., JK, Issitt PD, R,Reihart Kuhns Oyen WJ. Anti-Vel antibodyshowing 2,anew heterogeneity Cleghorn TE. rareThe bloodgroup occurence 1961. inBritain ofcertain factors the Anaham(Chilcotin) Am.J.Phys.Anthropol. Indians. 1970;32(3):329-37. Alfred BM,Stout TD, J, Lee Petrakis NL.Bloodgroups, M,Birkbeck of phosphoglucomutase, andcerumentypes Chandanayingyong D, Sasaki GreenwaltTT, TJ. Bloodgroups of the Thais. Transfusion 1967;7(4):269-76. 2012. ME,Lomas-FrancisReid C,OlssonML. The BloodGroup AntigenFacts 3ed. SanDiego:Academic Book. Press; 1998;75(1):70-1. ofaninvivo detection haemolyticanti-vel by thegeltest. L,ClasenC.Unsatisfactory J, Bartz VoxNeppert Sang. 1961;1:111-5. Levine P, White Ja,Stroup negative M.Seven membersinthree generationsofafamily.Ve-a (Vel) Transfusion apropos ofacaseanti-VEL immunization].Bibl.Haematol. 1965;23:309-11. [Studyofthe Battaglini Pf, Rm. J, population C,SalmonNicoli intheMarseilles Ranque Bridonneau VEL factor EB. bloodfactor: [New Sussman LN,Miller Vel.]. Rev.Hematol. 1952;7(3):368-71. Daniels G.HumanBloodGroupsoud. 2ed. Oxford: 2002. Science; Blackwell Molecular basisofweak Molecular Vel expression 139

Chapter 6 Chapter 6

140 Chapter 6 information Supporting Molecular basisofweak Molecular Vel expression 141

Chapter 6 Chapter 6

10000 * ***

p = 0.45

1000 Mode Fluoresence Intensity (arbitrary units) (arbitrary ModeIntensity Fluoresence

100 rs1175550 AA AA AA c.64_80del wt/wt wt/ wt/wt SMIM1 genotype Supplementary Figure 1. Vel expression levels in 97 Vel+w donors and ten Vel+ donors. The SMIM1 genotype was determined via sequencing of SMIM1 and Vel expression levels were determined using flow cytometry. Box plot shows 25% and 75% percentile and average fluorescence intensity, error bars indicate the 2.5% and 97.5% percentile. Statistical analysis was performed with an unpaired two-sided T-test. The six donors in which agglutination determined weakened Vel expression, while in these cases the SMIM1*64-80del allele was absent (white boxplot), had significantly lower Vel expression levels compared to control samples with same genotype (dark grey box plot) (p = 0.0114). Moreover, no significant difference was detected between these six donors and 91 donors heterozygous positive for the SMIM1*64_80del allele (light grey boxplot) (p = 0.4503).

142 Chapter 7

Molecular analysis of immunized Lan- or Jr(a-) patients and validation of genotyping assays to screen blood donors for Lan- and Jr(a-) phenotype

Lonneke Haer-Wigman1 Aïcha Ait Soussan1 Peter Ligthart2 Masja de Haas1,2 C. Ellen van der Schoot1

1 Sanquin Research, Amsterdam and Landsteiner Laboratory, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands 2 Sanquin Diagnostic Services, Amsterdam, The Netherlands

Submitted to Transfusion Chapter 7

Abstract Background: Individuals with anti-Jra or anti-Lan can only be transfused with rare Jr(a-) and Lan- red blood cells. We therefore characterized mutations in Dutch Jr(a-) and Lan- individuals and developed a high throughput genotyping assay to detect donors with the Jr(a-) or Lan- phenotype.

Methods: Six Jr(a-) and seven Lan- persons, who all made anti-Jra or anti-Lan, were sequenced for ABCG2 or ABCB6 and copy-number of ABCG2 and ABCB6 was determined. A total of 3,366 Caucasian, 621 Black and 333 Chinese donors were screened with high-throughput screening assays targeting frequently occurring mutations causing the Jr(a-) or Lan- phenotype.

Results: In the six tested Jr(a-) individuals the previously described c.376C>T, c.706C>T and c.736C>T nonsense mutations in ABCG2 were detected. In the seven Lan- individuals twelve different mutations were detected. No copy-number variation was detected for ABCG2 and ABCB6, respectively. In the donors that were screened, we detected five Caucasian donors heterozygous for the c.706C>T or 736C>T mutation in ABCG2 and nine Caucasian donors heterozygous for the 574C>T mutation in ABCB6. No Black or Chinese donors were found positive for a mutation.

Conclusions: We describe eight new mutations in ABCB6 of which seven, including three missense mutations, underlie the Lan- phenotype and we determine that a partial or complete gene deletion of ABCG2 or ABCB6 is not responsible for the Jr(a-) or Lan- phenotype, respectively. Due to the large heterogeneity of mutations causing the Jr(a-) or Lan- phenotype, genetic screening for the Jr(a-) and Lan- phenotype is not efficient.

144 0.03% and has also been detected inthe Caucasian population. be selected forbe selected transfusion. present inindividualswiththe Jr(a-)orLan-phenotype. by founder effects, deletionsofABCG2 itcannotbeexcluded orABCB6are thatalso(partial) is the most frequently detected mutation in Jr(a-) individuals. the Jr(a-) or Lan- phenotype, respectively.the Jr(a-)orLan-phenotype, expression oftheJr becausecommercialreagentsLan- phenotype, typing are Furthermore, lacking. variable cassette [ABC] transporters andare bothputative transporters. porphyrin transporters cassette [ABC] phenotype in the Japanese population. phenotype mutation intheJapanesepopulation(1.7%)isresponsible for the “high” incidenceoftheJr(a-) frozen units. lowexcept onlypresent andoften foras ofJr(a-)orLan-unitsisvery Japan,theavailability ABCG2 are more frequently detected inCaucasian Jr(a-)individuals. respectively. are Jr(a-)orLan-were shown to lackthecomplete expression oftheABCG2 orABCB6protein, ABCG2. p.Gln141Lys, c.1714A>C encodingp.Ser572Arg andc.1858G>A encodingp.Asp620Asn) in Lan- individuals. and onlythec.459delC andc.574C>T mutationsinABCB6 The Jr Introduction and Lan- individuals is homozygous for amutation. and Lan-individualsishomozygous Despite the large heterogeneity in of Lan- Jr(a-) individuals and Jr(a-) individuals, the majority ABCB6. by theABCG2is carried protein encodedby ABCG2 andLanby theABCB6protein encodedby reaction canoccur. is transfusedwithJr(a+)orLan+red bloodcells, respectively, asevere hemolytictransfusion rare. very (c.826C>T, c.1028G>A, c.1762G>A andc.2216G>A). expression ofmissensemutations inABCB6 of onenullalleleand/orthe(homozygous) Blacks andCaucasians. hasbeendetected inAsians,is ~0.005%intheCaucasian populationandthisphenotype Blacks. inAfrican oftheJr(a-)phenotype frequency oftheLan-phenotype The frequency than 99.9% individuals of all populations. of theJr Recently, thegeneticbasisofbothJr screened for theJr(a-)andLan-phenotype. a andLan blood group antigens are antigens and are high frequency expressed in more 16-18 21 Already 20different mutationshave thatcausetheLan-phenotype beendescribed a 3 antigen[Jr(a+ inJapanwithanincidenceof ismostfrequentlyThe Jr(a-)phenotype occurring The ABCG2 protein andtheABCB6protein, belongto thelarge familyofATP-binding 14,15 16-18 To increase ofJr(a-)andLan-unitsallblooddonors shouldbe theavailability Multiple mutationsinABCG2 Multiple 16,24-26 a 9-11 and Lan antigen makes serologicaland Lanantigenmakes screening lessreliable. Weak Lan[Lan+ Therefore, rare Jr(a-)orLan-blooddonorsshould bloodfrom very 6-8 to theJr When anindividualwithallo-antibodies w )] is linked withthree)] islinked missensemutations(c.421C>A encoding 9,11 Onlyafew routinely countries screen donors for theJr(a-)or w ] expression iscorrelated expression withheterozygous 20 a andLanbloodgroup antigenswaselucidated. The c.706C>T and c.736C>T nonsense mutations in 16-18,20-26 1-3 As a consequence, for negativity Jr or ABCB6 have thateachcause beendescribed The c.376C>T nonsensemutationinABCG2 24,25 16-18,20-26 have beendetected infive ormore Althoughthismightbeexplained 20,22,23 Genotyping oftheJr(a-)andLan-phenotype Genotyping 4,5 No data is available on the The “high” incidence of this 17,18 Weakened expression Weakened 19 Individuals who Individuals a orLanantigen 12,13 a and Lan is In general, In 16-18 Jr 145 a

Chapter 7 Chapter 7

Since the molecular basis of the Jra and Lan antigen has been determined, it became possible to predict the Jra and Lan phenotype according to genetic results. The aims of our study were to characterize mutations present in Dutch Jr(a-) and Lan- individuals, to determine the copy number of ABCG2 and ABCB6 in these individuals and to develop a high-throughput genotyping assay to screen Dutch, Chinese and Black blood donors for the Jr(a-) and Lan- phenotype.

Material and methods Samples and DNA isolation Six Jr(a-) and seven Lan- individuals from the Netherlands, who all had made anti-Jra or anti-Lan due to a pregnancy, respectively, were included. For screening 3,366 Caucasian blood donors, 621 Black blood donors and 333 African Chinese blood donors were included. Caucasian donor samples were collected by Sanquin Blood Supply, The Netherlands. African Chinese and African Black donor samples were collected by The South African Blood Transfusion Service in Johannesburg South-Africa. Additional African Black blood donor samples were collected by The Ethiopian Red Cross Society Transfusion Service in Addis Ababa Ethiopia and The Red Cross Blood bank Curaçao in Willemstad Curaçao. All cases were included after informed consent was given. In all cases ethylenediaminetetraacetic acid [EDTA] anticoagulated blood was collected and genomic DNA was isolated from white blood cells using a DNA extraction kit (QIAamp DNA Blood Mini Kit).

Serology The Jr(a-) and Lan- phenotype were serologically typed using standard serological tube agglutination with at least two anti-Jra or anti-Lan, respectively. An individual was determined Jr(a-) or Lan- when all tested anti-Jra or anti-Lan sera, respectively, did not agglutinate his/her red blood cells. To determine the Jr(a-) phenotype sera named J02, J21 (serum of Unique Case Number 2 [UCN]2), J22, J23 (serum of UCN6), J29 (serum of UCN4), Jjb, Jai, Jsf and/or Jto were used and to determine the Lan- phenotype sera named L01, L02, L14, L16 (serum of UCN12), L28, L33, Lbm, Lga and/or Lmw were used. The presence of anti-Jra or anti-Lan in the serum of a Jr(a-) or Lan- individual, respectively, was defined when all tested, but at least two Jr(a-) or Lan- red blood cells did not agglutinate with the serum. Absorption-elution, to detect weak expression of the Lan antigen, was performed using serum L01. Serum was added in a 4 to 1 dilution to red blood cells and incubated for 30 minutes at 37ºC, elution was performed according to manufacturers’ protocol (Gamma ELU-KIT II, ImmucorGamma). Flow cytometry was performed to test the presence of anti-Lan in the eluates. Eluates were added 1:1 to a 3% red blood cell suspension and incubated for 30 minutes at room temperature. Subsequently, red blood cells were washed and goat anti-human-IgG (Alexa-488, Molecular Probes) antibody in a 1:250 dilution was added and incubated for 30 minutes at room temperature. Red blood cells were washed and analyzed by flow cytometry (LSRFortessa cell analyzer).

146 In order to mutations in ABCG2In detect DNA-Sequencing developed to thecopy determine numberoftheABCG2 andABCB6 Probe Ligation-dependent Multiplex Amplification [MLPA] probe combinations were Multiplex Probe Ligation-dependent Amplification BigDye Terminator onanABI3130XLsequencer(Applied Biosystems). v3.1kit Healthcare) according to manufacturer’s protocol. PCRproductswere sequencedwithABI 64°C, followed by 1minute at72°C.PCRproductswere usingillustraExoProStar purified (GE PCR conditions were: of 10 seconds at 95°C and 10 15 seconds at 95°C, 35 cycles Fastof 2xGeneAmp (Applied Biosystems), andreverse 0.5μMforward Mix PCRMaster primer. (Applied Biosystems) inatotal volumethermocycler of20μL,containing50-150ngDNA,10mL ABCB6 ABCB6*1867delinsAACAGGTGA, assays targeted ABCB6*459delC, mutationscausingtheLan-phenotype: ABCG2*376Tcausing theJr(a-)phenotype: , that eithercausetheJr(a-)orLan-phenotype. Three TaqMan assay targeted mutations For high-throughput genotyping TaqMan assays were developed specific for mutations TaqMan assay settings. version 1.85(Softgenetics, software State College, USA)usingstandardGenemarker MLPA using analyzed Analyzer ona3130Genetic (Applied Biosystems). Dataanalysiswasperformed (Applied Biosystems)500-Liz Size and 0.5 μL GeneScan Standard (Applied Biosystems) was (Applied Biosystems). of1.5μLMLPA Amixture Thermocycler Formamide sample, 8.5μLHi-Di The MLPA according to manufacturers’ was performed protocol in a (MRC-Holland) Veriti exon 1,8and 17 ofABCB6 combinations targeting exon 3,7and15ofABCG2, three Lanprobe combinationstargeting a StepOnePlus (Applied Biosystems), theABCG2*376T, the were performed using thesameprotocol,were performed however, (Applied SYBR Green Mix PCRMaster thefluorescence signaleach cycle wasmeasured. assays withSYBRGreenTaqMan reagentia were: 10 of minutes15 seconds at 95 °C and 1 minuteat 95°C, 50 cycles at 60°C at the end of (Applied Biosystems), andreverse 0.3μMforward and0.1μMprobe. primers PCRconditions volume of25μL,containing2.5-250 ngDNA,12.5μL2x TaqMan Universal PCRMastermix using SYBRGreen reagentia. TaqMan assays with TaqMan reagentia inatotal were performed ABCB6*1867delinsAACAGGTGA, assays madeuseofaprobe and TaqMan reagentia andtheABCB6*459delC, ALB genewasusedto DNAconcentrations. determine The assays on were performed were developed. ona The polymerasechainreaction[PCR]wasperformed Veriti were applied in combination with four control probe combinations. assays wereABCB6*1942T andABCB6*IVS2256+2G performed . A ABCB6*1942T andABCB6*IVS2256+2G TaqMan assay for and , primers flanking allexons ofABCG2 flanking ABCB6, primers ABCG2*706TFive andABCG2*736T TaqMan. ABCG2*706T, Genotyping oftheJr(a-)andLan-phenotype Genotyping gene. Three Jr gene. Three ABCG2*736T andALB ABCB6*574T, ABCB6*574T, a probe and 147

Chapter 7 Chapter 7

Biosystems) was used instead of 2x TaqMan Universal PCR Mastermix and no probe was added. Data analysis was performed using the StepOnePlus software. For screening with the TaqMan assays, pools of 90 donors were made with a final concentration of 150 ng DNA. Each pool was tested in duplicate in all TaqMan assays targeting mutations causing the Jr(a-) and Lan- phenotype and in duplicate in the ALB TaqMan assay for calibration. In each run a positive control was taken along that contained 1.6 ng positive DNA in a background of 150 ng DNA and a negative control containing 150 ng DNA that did not contain the targeted mutation. When no positive genomic DNA was available a plasmid containing the specific target was used as a positive control. The outcome of the pools was calculated by subtracting the average Ct value of the ALB assay of a pool from the average Ct value for each assay targeting a mutation causing the Jr(a-) or Lan- phenotype. When this DCt value was in the same range as the positive control a pool was determined positive. When a pool of 90 donors was determined to be positive, two sub pools containing 45 donors were tested using the same TaqMan assay. All samples of the positive sub pool where than separately tested in the same TaqMan assay to identify the donor that was positive for the targeted mutation. Donors determined positive for a mutation causing the Jr(a-) or Lan– phenotype were sequenced for all exons of ABCG2 or ABCB6, respectively.

Results Genetic basis of Dutch Jr(a-) and Lan- individuals To identify the genetic basis of Dutch Jr(a-) and Lan- individuals, all exons of ABCG2 or ABCB6 were sequenced in six individuals who were Jr(a-) and in seven individuals who were Lan- . All individuals had made allo-anti-Jra or allo-anti-Lan, respectively. In the six Dutch Jr(a-) individuals negativity was caused due to three different nonsense mutations (c.376C>T, c.706C>T or c.736C>T) in ABCG2 (Table 1). Five individuals were homozygous for one of these mutations (Table 2). UCN5 was heterozygous for two of these nonsense mutations, c.376C>T and c.706C>T, and carried an additional heterozygous missense mutation c.421C>A encoding p.Gln121Lys (Table 2). All detected mutations have previously been described (Table 1). In the seven Dutch Lan- individuals twelve different mutations were detected in ABCB6 (Table 1). Three of these mutations have already been described to be related with the Lan- phenotype (c.574C>T, c.1867delins and c.1942C>T) and one mutation was described to be related with the Lan+w phenotype (c.1762G>A). These mutations were either homozygous present (UCN8, UCN12) or heterozygous next to another not yet described mutation: in UCN7 this concerned a mutation disrupting the start codon (c.1A>C), in UCN9 a splice site mutation (c.971-1G>A) and a missense mutation (c.827G>A encoding p.Arg276Glu) and in UCN11 a nonsense mutation (c.2155C>T) and a missense mutation (c.55A>T encoding Met19Leu) (Table 2). In UCN10 two new heterozygous missense mutations were detected (c.1825G>A encoding p.Val609Met and c.1912C>T encoding p.Arg638Cys) (Table 2). UCN13 was found to be homozygous for a new splice site mutation (c.2351+1G>A) (Table 2).

148 Table 1. Mutations described by Helais et al.16, Saison et al.17,24, Zelinski et al.18, Hue-Roye et al.20,21, Tanaka et al.22, Tobita et al.23, Reid et al.25 and Yamamuro et al.26 and mutations reported in this article that alter Jra or Lan expression

Number of persons in Exon or Studies in which the Phenotype† Gene Nucleotide change‡ rs number Aminco Acid change§ this study carrying the intron mutation was detected mutation Jr(a-) ABCG2 Exon 2 c.2T>C p.0 22 Jr(a-) ABCG2 Exon 2 c.187_197del p.Ile63fs 17 Jr(a-) ABCG2 Exon 3 c.244_245insC p.Thr82fs 20 Jr(a-) ABCG2 Intron 3 c.263+1G>A r.spl? 23 Jr(a-) ABCG2 Exon 4 c.289A>T p.Lys97Ter 23 Jr(a-) ABCG2 Exon 4 c.337C>T p.Arg113Ter 20, 22, 23 Jr(a-) ABCG2 Exon 4 c.376C>T rs72552713 p.Gln126Ter 3 17, 18, 20, 21, 22, 22 Jr(a-) ABCG2 Exon 6 c.542_543insA p.Phe182fs 17 Jr(a-) ABCG2 Exon 6 c.565_566del p.Gly189fs 23 Jr(a-) ABCG2 Exon 7 c.706C>T rs140207606 p.Arg236Ter 5 17, 18, 20, 21, 23 Jr(a-) ABCG2 Exon 7 c.730C>T p.Gln244Ter 17 Jr(a-) ABCG2 Exon 7 c.736C>T rs200190472 p.Arg246Ter 4 18, 20 Jr(a-) ABCG2 Exon 7 c.784G>T rs200473953 p.Gly262Ter 20, 21 Jr(a-) ABCG2 Exon 7 c.791_792del p.Leu264fs 17 17

Jr(a-) ABCG2 Exon 8 c.875_878dup p.Phe293fs oftheJr(a-)andLan-phenotype Genotyping Jr(a-) ABCG2 Exon 9 c.1111_1112del p.Thr371fs 17 Jr(a-) ABCG2 Exon 13 c.1515del p.Ala506fs 23 Exon 5 en Jr(a-) ABCG2 c.[421C>A; 1515del] p.[Gln141Lys;Ala506fs] 22 13 Jr(a-) ABCG2 Exon 13 c.1591C>T rs201584210 p.Gln531Ter 20 Jr(a-) ABCG2 Exon 14 c.1723C>T p.Arg575Ter 23 Jr(a-) ABCG2 Exon 15 c.1789_1790insT p.Ala597fs 23 Jr(a+w) ABCG2 Exon 5 c.383A>T rs149106245 p.Asp128Val 23 Jr(a+w) ABCG2 Exon 5 c.421C>A rs2231142 p.Gln141Lys 21, 23 149 Jr(a+w) ABCG2 Exon 16 c.1858G>A rs34783571 p.Asp620Asn 21

Chapter 7 Chapter 7

25 16 26 16 16 26 23 23 23 23 21 23 21 23 23 23 24 16 26 25 26 24, 25 Studies in which the Studies detected was mutation 1 1 1 10 Number of persons in this study carrying the mutation p.Trp412Ter p.Leu515fs p.Thr294fs p.Gly318fs r.spl? p.Gln239Ter p.Arg240ter p.Arg276Glu Aminco Acid change§ Acid Aminco p.Asp620Gly p.[Gln141Lys;Arg147Gln] p.[Gln141Lys;Thr153Met] p.[Met152Thr;Cys608Arg] p.Ser340del p.Gly462Arg p.Ser572Arg r.spl? p.Cys608Arg p.Leu614Trp p.0 p.Phe29del p.Ala66fs p.Ala101fs p.Val126fs p.Leu154fs p.Arg192Trp rs148458820 rs200125320 rs number rs200894058 rs149202834 c.1236G>A c.1533_1543dup c.881_884del c.953_956del c.971-1G>A c.717G>A c.718C>T c.827G>A Nucleotide change‡ c.85_87del c.1859G>A c.[421C>A;440G>A] c.[421C>A;458C>T] c.[455T>C;1819T>C] c.1017_1019del c.1384G>A c.1714A>C c.1820+1G>A c.1819T>C c.1841T>G c.1A>C c.197_198insG G c.301_302ins c.376delG c.459del c.574C>T Exon 6 Exon 9 Exon Exon 4 Exon 4 Exon Intron 4 Exon 3 Exon 3 Exon 3 Exon Exon or Exon intron Exon 1 Exon Exon 16 Exon 5 Exon 5 Exon 5 and Exon 16 9 Exon 12 Exon 14 Exon Intron 15 16 Exon 16 Exon 1 Exon 1 Exon 1 Exon 1 Exon 1 Exon 2 Exon ABCB6 ABCB6 ABCB6 ABCB6 ABCB6 ABCB6 ABCB6 ABCB6 Gene ABCB6 ABCG2 ABCG2 ABCG2 ABCG2 ABCG2 ABCG2 ABCG2 ABCG2 ABCG2 ABCG2 ABCB6 ABCB6 ABCB6 ABCB6 ABCB6 ABCB6 Continued) ) w Lan- Lan- Lan- Lan- Lan- Lan- Lan- Lan- Jr(a+ Phenotype† Lan- Unclear Unclear Unclear Unclear Unclear Unclear Unclear Unclear Unclear Lan- Lan- Lan- Lan- Lan- Lan- Table 1. ( Table

150 Table 1. (Continued)

Number of persons in Exon or Studies in which the Phenotype† Gene Nucleotide change‡ rs number Aminco Acid change§ this study carrying the intron mutation was detected mutation Lan- ABCB6 Exon 9 c.1558_1559insT p.Val520fs 25 Lan- ABCB6 Exon 10 c.1617delG p.Gly539del 26 Lan- ABCB6 Exon 11 c.1690_1691del p.Met564fs 16 Lan- ABCB6 Exon 11 c.1709_1710del p.Glu570fs 16 Lan- ABCB6 Exon 13 c.1825G>A p.Val609Met 1 Lan- ABCB6 Exon 14 c.1867delins p.Gly623fs 2 16 Lan- ABCB6 Exon 14 c.1912C>T p.Arg638Cys 1 Lan- ABCB6 Exon 14 c.1942C>T p.Arg648Ter 1 16 Lan- ABCB6 Exon 15 c.1985_1986del p.Leu662fs 16, 25 Lan- ABCB6 Exon 16 c.2155C>T p.Glu719Ter 1 Lan- ABCB6 Intron 16 c.2256+2T>G r.spl? 16 Lan- ABCB6 Intron 17 c.2351+1G>A rs150574070 r.spl? 1 Lan+w ABCB6 Exon 3 c.826C>T rs57467915 p.Arg276Trp 24, 25 Lan+w ABCB6 Exon 5 c.1028G>A p.Arg343Gln 25 Lan+w ABCB6 Exon 12 c.1762G>A rs145526996 p.Gly588Ser 1 24, 25 w Lan+ ABCB6 Exon 16 c.2216G>A p.Arg739His 25 oftheJr(a-)andLan-phenotype Genotyping Unclear ABCB6 Exon 1 c.20A>G p.Tyr7Cys 26 Unclear ABCB6 Exon 1 c.55A>T p.Met19Leu 1 Unclear ABCB6 Exon 1 c.403C>A rs201932534 p.Arg135Ser 26 Unclear ABCB6 Intron 3 c.869-2A>G r.spl? 26 Unclear ABCB6 Exon 6 c.1199_1210del p.Ile400_Tyr404del 26 Unclear ABCB6 Intron 16 c.2256+1G>A r.spl? 26 Unclear ABCB6 Exon18 c.2383_2385del p.Leu795del 26

† Unclear is stated when more experiments need to be performed to determine the exact phenotype caused by this allele

151 ‡ Position as counted from ATG translation start site § Position as counted from Met translation start site

Chapter 7 Chapter 7

Table 2. Sequencing and MLPA results of six Jr(a-) persons and seven Lan- persons and donors screened positive for a mutation causing the Jr(a-) or Lan- phenotype

Patient/ Jra or Lan anti-Jra or Nucleotide Predicted Amino UCN Ethnicity Alleles Donor phenotype anti-Lan change(s)† Acid change(s)‡ 1 Patient Caucasian Jr(a-) Yes 2 c.376C>T (homo) p.Gln126Ter 2 Patient Asian Jr(a-) Yes 2 c.376C>T (homo) p.Gln126Ter 3 Patient Asian Jr(a-) Yes 2 c.376C>T (hetero) p.Gln126Ter c.706C>T (hetero) p.Arg236Ter c.421C>A (hetero) p.Gln141Lys 4 Patient Caucasian Jr(a-) Yes 2 c.706C>T (homo) p.Arg236Ter 5 Patient Caucasian Jr(a-) Yes 2 c.706C>T (homo) p.Arg236Ter 6 Patient Caucasian Jr(a-) Yes 2 c.736C>T (homo) p.Arg246Ter 7 Patient Caucasian Lan- Yes 2 c.1A>C (hetero) p.0 c.1867delins (hetero) p.Gly623fs 8 Patient Caucasian Lan- Yes 2 c.574C>T (homo) p.Arg192Trp 9 Patient Caucasian Lan- Yes 2 c.827G>A (hetero) p.Arg276Glu c.971-1G>A (hetero) r.spl? c.1762G>A (hetero) p.Gly588Ser 10 Patient Caucasian Lan- Yes 2 c.1825G>A (hetero) p.Val609Met c.1912C>T (hetero) p.Arg638Cys 11 Patient Unknown Lan- Yes 2 c.1867delins (hetero) p.Gly623fs c.2155C>T (hetero) p.Glu719Ter c.55A>T (hetero) p.Met19Leu 12 Patient Caucasian Lan- Yes 2 c.1942C>T (homo) p.Arg648Ter 13 Patient Caucasian Lan- Yes 2 c.2351+1G>A (homo) r.spl? 14 Donor Caucasian No c.706C>T (hetero) p.Arg236Ter c.34G>A (hetero) p.Val12Met§ 15 Donor Caucasian Jr(a+) No c.706C>T (hetero) p.Arg236Ter 16 till 18 Donor Caucasian Jr(a+) No c.736C>T (hetero) p.Arg246Ter 19 till 25 Donor Caucasian Lan+ No c.574C>T (hetero) p.Arg192Trp

26 and 27 Donor Caucasian Lan+ No c.574C>T (hetero) p.Arg192Trp c.2196G>A (hetero) p.Lys732Lys

† Position as counted from ATG translation start site; homo = homozygous, hetero = heterozygous ‡ Position as counted from Met translation start site § This missense mutation has no effect on Jra expression

152 of ABCB6, according to the two-dimensional structure predicted by the of ABCB6,according to thetwo-dimensional Tmhmm (see Server p.Gly588Ser, p.Met609Val and p.Arg638Cys domain are in the intracellular nucleotide-binding the ABCB6protein, thep.Arg192Trp mutationislocated inthetransmembraneregion andthe more detail. The p.Met19Leu andp.Arg276Glu loops of mutations are situated inextra-cellular p.Met609Val andp.Arg638Cys alterations, respectively) ontheABCB6protein were analyzed in c.1825G>A andc.1912C>T encodingthep.Met19Leu, p.Arg192Trp, p.Arg276Glu, p.Gly588Ser, The effect ofthesixmissensemutations (c.55A>T, c.574C>T, c.827G>A, c.1762G>A, blood cells mutations inABCB6areMissense ableto abolishcomplete Lanexpression on red regionMet19Leu ofABCB6. mutation isinanonconserved p.Gly588Ser, p.Met609Val andp.Arg638Cys areas oftheABCB6protein, are inhighlyconserved whilethe acids are colored white. Barsbelow indicate thedegree ofconservation. The p.Arg192Trp, p.Arg276Glu, areacids are amino acids with the coloredsame properties colored grey, black, amino non-conserved cross-species sequence alignment of protein homologues in 10 different amino species. Conserved ofregionsof theABCB6protein. containingthesemissensemutationsinABCB6via b)Conservation p.Gly588Ser, p.Met609Val andp.Arg638Cys missensemutationsare located intheintracellularpart the protein, thep.Arg192Trp mutationislocated inthetransmembraneregion oftheprotein andthe and p.Arg638Cys. of The Met19Leu andArg276Glu part mutationare positionedintheextracellular dots represent sixmissensemutations Met19Leu, p.Arg192Trp, Arg276Glu, p.Gly588Ser, p.Met609Val putative membraneregions ofABCB6predicted by the withresidue numbers. Black TMHMM Server model of the ABCB6 protein. Each dot represent an amino acid, dots within the black lines indicate Figure representation mutationsdetected inABCB6.a)Schematic dimensional 1.Missense ofatwo Genotyping oftheJr(a-)andLan-phenotype Genotyping 153

Chapter 7 Chapter 7

Web Resources) (Figure 1a). Furthermore, the effect of the mutations was analyzed using the Sift and PolyPhen software.27,28 All missense mutations, except the p.Met19Leu mutation, were predicted to be extremely damaging, because they are present in highly conserved regions of the ABCB6 protein (Figure 1b). The p.Met19Leu was determined to be benign/tolerated, this mutation was, however, detected in an allo-immunized Lan- individual that already carried two null-alleles (UCN11).

Figure 2. ABCB6 protein expression in UCN8, UCN9 and UCN10. To detect Lan expression in UCN8, UCN9 and UCN10 an absorption-elution was performed using serum of L01. As a negative control red blood cells of UCN13 with the Lan- phenotype were taken along. As a positive control normal Lan+ red blood cells and a sensitivity control: a mixture 95% UCN13 Lan- red blood cells and 5% Lan+ red blood cells, were taken along. The presence of anti-Lan in the eluate, hence expression of the Lan antigen, was tested using flow cytometry on red blood cells of five Lan+ individuals (grey bars; error bars indicate standard deviation) and red blood cells of one Lan- person (UCN13, open bars). No Lan expression was detected in UCN8, UCN9 and UCN10 and the negative control UCN13, while in the positive control and sensitivity control Lan expression was detected.

To determine the effect of the missense mutations on the expression of the ABCB6 protein/ Lan antigen the sensitive absorption-elution technique was performed with red blood cells from UCN8, UCN9 and UCN10. As expected, since these cases all produced anti-Lan, no Lan expression was detected on red blood cells of UCN8, UCN9, UCN10 and of the negative control UCN13. In the positive control and a sensitivity control, containing a mixture of 95% Lan- (UCN13) and 5% Lan+ red blood cells, expression of the Lan antigen was detected (Figure 2).

154 Screening for theJr(a-) wasdetected ineitherABCG2number oftwo 2). orABCB6(Table mutations (UCN1, 2, 4, 5, 6, 8, 12 and 13) a copy- all cases with homozygous 17 ofABCB6. In detected. MLPA probes were designed to target exons 3,7and15ofABCG2 us to investigate oftheABCG2 whetherlossofheterozygosity andABCB6genescouldbe mutation intheliterature for thecausative ofJr(a-)andLan-individualsbeinghomozygous The highfrequency two allelesofABCG2Jr(a-) andLan-individualscarry andABCB6 were positive for eithertheJr (UCN15-17, UCN19-24andUCN26).As expected, serological showed typing thatallcases c.2196G>A bloodcellswere 2).Red mutation(Table available inten cases ofthefourteen were detected. screened to oneLan-donor. detect theBlackandChinesedonorsnoneofmutations In donors needto bescreened to oneJr(a-)donorand~500,000donorsneed to detect be Caucasian donorsare screened for c.706T andc.736T inABCG2 andc.574TABCB6 ~1,000,000 in to have no affect on the expression of the Jr mutation wasdetected, inUCN14thec.34G>A missensemutation,thathasbeendescribed three presence heterozygous caseanadditionalheterozygous 2).In ofthemutations(Table of all exonsUCN27). Sequencing of UCN18) andninedonorswere positive for thec.574C>T missensemutationinABCB6 UCN15), three donorswere positive for thec.736C>T nonsensemutationsinABCG2 donors, donorswere two positive for thec.706C>T nonsense mutationinABCG2 (UCN14and donors were tested inpoolscontainingamaximumof90donors. Amongthe3,361Caucasian of DNA of 90 samples (data not shown). A total of 3,366 Caucasian, 621 Black and 333 Chinese were developed. Allassays are sensitive enoughto onepositive detect DNAsampleinapool phenotype (c.459delC,phenotype c.574T, c.1867delinsAACAGGTGA, c.1942T andc.2256+2G (c.376T,the Jr(a-) phenotype c.706T and c.736T in mutations were developed. TaqMan assays targeting genotyping three mutations causing assays for genotyping each causetheJr(a-)orLan-phenotype, themostfrequently occurring developed to screen donorsfor Becausemultiplemutationscan theJr(a-)andLan-phenotype. To increase thenumbersofJr(a-)andLan- donors, ahighthroughputassay was genotyping

and Lan-phenotype 16-18,20,25 a orLanantigen.Basedontheseresults we calculated thatwhen despite the highnumberofcausative mutationsprovoked ABCG2 or a antigen ABCB6, respectively, determined in all cases the ABCG2) and five mutations causing theLan- 17,20 and in UCN26 and UCN27 the silent Genotyping oftheJr(a-)andLan-phenotype Genotyping and exons 1,8and in (UCN19- (UCN14- ABCB6) 155

Chapter 7 Chapter 7

Discussion In this study we determined the genetic basis of Dutch Jr(a-) and Lan- individuals, that all made alloanti-Jra or alloanti-Lan, respectively. We show that the Jr(a-) or Lan- phenotype is not caused by the entire deletion or partial deletion of ABCG2 or ABCB6, respectively. Furthermore, the developed high-throughput genotyping assays targeting frequently occurring mutations in ABCG2 and ABCB6 can detect Jr(a-) and Lan- individuals. Genotyping of six Dutch Jr(a-) individuals identified three previously described nonsense mutations in the Dutch population that cause the Jr(a-) phenotype, c.376C>T, C.706C>T and c.736C>T in ABCG2 (Table 1 and 2). In one case (UCN5) an additional missense mutation c.421C>A was detected, next to two heterozygous nonsense mutations. Interestingly, this polymorphism has been previously described in the Japanese population with an allele frequency of 0.45 and it was shown that this mutation results in a lower expression level of ABCG2.29 We therefore conclude that the coincidence of this mutation with two heterozygous nonsense mutations probably occurred in an allele carrying a nonsense mutation, most likely the 376C>T allele, because this mutation is also linked to Japanese ethnicity. Genotyping of seven Dutch Lan- individuals, who were all identified because they produced allo-anti-Lan, revealed that also in the Dutch population the Lan- phenotype shows a high genetic heterogeneity. Eight new mutations, of whom seven underlie the Lan- phenotype, and four previously described mutations were detected (Table 1). For five newly detected mutations we have strong indications that they cause the Lan- phenotype. Because they were present in a Lan- individual who made anti-Lan and they were a nonsense mutations (c.2155C>T), a mutation that disrupt the start (codon 1A>C), a splice site mutation that was homozygous present (c.2351+1G>A) or a missense mutation that was present next to another missense mutation (c.1825G>A encoding p.Val609Met and c.1912C>T encoding p.Arg638Cys). In two Lan- individuals three heterozygous mutations were detected in ABCB6, making it more difficult to assign the causative mutation. Saison et al. showed with a heterologous expression system that one of the three mutations, the c.1762G>A missense mutation encoding p.Gly588Ser, in UCN9 was to be unable to completely abolish Lan expression24. Moreover, Reid et al. determined that this mutation (determined in a single donor homozygous for c.1762G>A) results in the Lan+w phenotype using five human anti- Lan sera and one monoclonal anti-Lan (OSK43).25 We therefore assume that the two other mutations in UCN9 (the missense mutation c.827G>A encoding p.Arg276Glu and the c.971- 1G>A mutation disrupting a splice site) are responsible for the Lan- phenotype, although in the article of Saison et al.24 the c.1762G>A mutation was homozygous present in a patient who made anti-Lan. In the other individual with three heterozygous mutations, one of the heterozygous missense mutations (c.55A>T coding for p.Met19Leu) was predicted to have no effect on the ABCB6 protein and was detected next to two mutations resulting in null alleles, it is therefore highly unlikely that the c.55A>T missense mutation contributes to the

156 no copy number variation wasdetected inABCG2no copy numbervariation ABCB6 genes in Jr(a-) and Lanindividuals, respectively. the Dutch Jr(a-) and Lan- individuals In mutation results ABCG2 in weakened expression causative mutation for ofJr(a-)andLan-individualstoa behomozygous highfrequency with thereported The high number of causative mutations for in combination the Jr(a-) and Lan- phenotype Lan- phenotype. phenotype inastudyof phenotype Tobita etal. mutation (c.421C>A encoding p.Gln141Lys) with the Jr(a-) linked has been incorrectly mutations in the Black or Chinese cohorts, probablymutations in the Black or due Chinese tocohorts, the small size of these cohorts. in the Caucasian population, c.574C>Tphenotype in todetermined positive be heterozygous for the mostprevalent mutationcausing theLan- population, respectively c.706C>T and736C>TinABCG2 , for most prevalent oneinthe of the Caucasiantwo mutations causing the Jr(a-) phenotype and 333 Chinese donors. Five Caucasian donors were determined to positive be heterozygous c.2256+2G in targeting mutations frequentlythat cause the Jr(a-) (c.376T. occurring c.706T and c.376T To screen donorsfor we developed high-throughput theJr(a-)andLan-phenotype assays red cellmembrane. p.Arg638Cys missensemutationscausethecomplete absenceoftheABCB6protein onthe the already p.Arg192Trp described detected p.Arg276Glu, andthenewly p.Met609Val and Lan- donorswere notreactive witheachothersred bloodcells. We therefore concludethat regions atthetransmembraneorintracellularregion oftheprotein andtheseraofthese unlikely,seems very becauseallfour mutationsresulted inaminoacidchangesconserved technique. Furthermore, expressionsera andalsowiththesensitive absorption-elution partial anti-Lan andtheirred bloodcellswere Lan-withatleastfour determined different anti-Lan p.Arg276Glu, p.Met609Val andp.Arg638Cys) 1). (Table These three individualsdeveloped the ABCB6protein, provide because they onlyasingleaminoacidchange(p.Arg192Trp, detected inthree Lan-individualswillnotcompletely abolishtranslationofmRNAencoding ourstudyfour (personalcommunication).In causative mutationsinABCB6 Jr(a-) phenotype with thec.421C>A mutationadditionalmutationsinABCG2 thatwere responsible for the missense mutation in to thec.1762G>Ain theprevious thatwasinitiallylinked paragraph, theLan-phenotype expression oftheJr isrelevant identifytheeffect to ofmutationsdetected inABCG2 correctly It andABCB6 or ABCB6,butdueto thepresence affected oftwo allelesofABCG2 orABCB6. the of Dutch is not due to Jr(a-) and Lan- individuals negativity loss of heterozygosity ABCG2 (c.459delC,) orLan-phenotype c.574T, c.1867delinsAACAGGTGA, c.1942T and in ABCB6) were developed andvalidated onatotal of3,366Caucasian, 621Black 16-18,20,25 a andLanantigen,to As predictthephenotype. beableto correctly stated ABCB6 prompted usto investigate oftheABCG2 lossofheterozygosity and is most likely incorrect. is most likely 23 It has been experimentally shown thatthec.421C>A has beenexperimentally It 21,29 and Indeed, TobitaIndeed, etal. detected inallcases 24 ABCB6. We any did not detect of the ABCB6. We therefore conclude that in Also for the Jr and nineCaucasian donorswere Genotyping oftheJr(a-)andLan-phenotype Genotyping a antigen, a missense on the ABCG2 157

Chapter 7 Chapter 7

Interestingly, the c.376C>T nonsense mutation that has an allele frequency of 0.017 in the Japanese population was not detected in our Chinese cohort, therefore the “high” expression of c.376C>T nonsense mutation seems to be restricted to Japan. Furthermore, we did not detect donors positive for four of the five targeted mutations that cause the Lan- phenotype. Only one of these mutations, the c.459delC mutation, has been detected in more than five Lan- donors, this mutation is, however, only detected in Japanese donors and might be restricted to this population.26 Due to the large heterogeneity of mutations that cause the Jr(a-) or Lan- phenotype, at least in the Dutch population, it is not efficient to screen donors by genotyping. For instance, ~25 Lan- donors would be detected when 500,000 donors are serologically screened for the Lan- phenotype (taking into account a frequency of 0.005% for the Lan- phenotype), wile our genotyping assay would detect only one Lan- donor. We therefore propose, that screening for Jr(a-) and Lan- donors is more efficiently using large-scale serological typing and genotypic confirmation of mutations in ABCG2 or ABCB6, respecively. Nevertheless, specific ethnicities in which a specific mutation that causes the Jr(a-) or Lan- phenotype occurs frequently can be genetically screened to increase the numbers of Jr(a-) or Lan- donors, for instance testing Japanese donors for the c.376T mutation in ABCG2.

Acknowledgements The authors like to thank J. Hooydonk (The South African Blood Transfusion Service (Johannesburg, South Africa) for collection of samples from Black and Chinese donors, G. Tesfaye (The Ethiopian Red Cross Society Transfusion Service Addis Ababa, Ethiopia) and A.J. Duits (The Red Cross Blood bank Curaçao, Willemstad, Curaçao) for the collection of samples from Black donors.

Web Resources TMHMM Server v. 2.0; prediction of transmembrane helices of the ABCB6 protein. http://www. cbs.dtu.dk/services/TMHMM/ Accessed 12-06-2013. The protein sequence of ABCB6 that was used in the TMHMM Server was retrieved from http://www.ensembl.org/Homo_sapiens/Transcript/Sequence_Protein?db=core;g=ENSG00 000115657;r=2:220074490-220083712;t=ENST00000265316 Accessed 12-06-2013

158 24. 23. 22. 21. 20. 19. 18. 17. 16. 15. 14. 13. 12. 11. 10. 9. 8. 7. 6. 5. 4. 3. 2. 1. References

mutation causing the Lan- blood type. mutation causingtheLan-bloodtype. Vox Sang. 2013;104(2):159-65. Saison C, Helias V, Peyrard T, JP, L,Cartron Merad L. Arnaud The ABCB6 mutationp.Arg192Trp isa recessive analysis oftheJr(a-)inJapanesepeople. Vox Sang. 105[Sup1],230.2013. Tobita S,Osabe R,Kaito T, Saitou M, IsaK, Tsuneyama H, M.Genetic Yabe M,Minami Uchikawa R,Ogasawara K, population. Transfusion [Epubaheadofprint]. 2013;doi:10.1111/trf.12277 Tanaka I, M,Kamada H, Takahashi Matsukura J,K, Tani Kimura Y. intheJapanese theJr(a-)phenotype Defining print]. system: identification of alleles that alter expression. Transfusion 2013;doi:[Epub 10.1111/trf.12118. ahead of Zelinski K, T,Hue-Roye Cobaugh A,Lomas-Francis C,Miyazaki T, Tani Y, WesthoffReid ME. CM, The JRblood group ofthree nullalleles. new characterization Transfusion 2012;53(7):1575-9. Lomas-Francis K, Hue-Roye C,Coghlan G,Zelinski T, ME. Reid The JRbloodgroup system (ISBT032):molecular Pharm.Biotechnol. 2011;12(4):647-55. P,Krishnamurthy JD. Schuetz The role Curr. ofABCG2 metabolismandcellsurvival. andABCB6inporphyrin 2012;44(2):131-2. Zelinski T, Coghlan G,LiuXQ, ME.ABCG2 Reid nullallelesdefinetheJr(a-)blood group Nat.Genet. phenotype. Junior. 2012;44(2):174-7. Nat.Genet. et al. NullallelesofABCG2 encodingthebreast cancerresistance protein blood definethenew group system Saison C,Helias V, BallifBA,Peyrard T, Puy H,Miyazaki T, Perrot S, Vayssier-Taussat M, Waldner M,Le Pennec PY, 2012;44(2):170-3. blood andspecifiesthenew ABCB6 isdispensableforgroup system erythropoiesis Langereis. Nat.Genet. Helias V, SaisonC,BallifBA,Peyrard T, Takahashi J, Takahashi H, Tanaka JC,Puy M,Deybach H,Le M,etal. Gall J,Matilainen Peyrard T, etal. witharare Donors pheno(geno)type. Vox Sang. 2008;95(3):236-53. HW,Reesink CP, Engelfriet C, H, Gassner Schennach Wendel S, Fontao-Wendel R, de Brito MA, Sistonen P, Nance ST. How to find, recruit andmaintainrare blooddonors. Curr.Opin.Hematol. 2009;16(6):503-8. R. JR,Oyen VariationStorry inLanexpression. Transfusion 1999;39(1):109-10. redto high-frequency cellantigenJra. Vox Sang. 1994;66(1):51-4. Miyazaki T, Kwon KW, Yamamoto K, Tone Y, H,Kato Ihara S.Ahumanmonoclonalantibody T, H,Sekiguchi Ikeda of theliterature. Transfusion 2004;44(2):197-201. Kwon MY, PA, SuL,Arndt DP. G,Blackall Garratty Clinicalsignificance ofanti-Jra:two cases and of review report 1994. Jowitt S,Powell H,Shwe KH,Love EM. Transfusion reactiondueto anti-Jra. Transfusion Med. 4[Suppl1], 473a. 6. AllenFH.Asecondexampleof anti-Lan. JK, MH,IssittPD, McGinniss Vox AJ, Grindon Reihart Sang. 1968;15(4):293- 1998. EA, Reddy Smart V, Fogg P. Anti-Lanandtherare Africa. inSouth Lan-negative phenotype Vox Sang. 74[Suppl1]. negative inJapanese. Transfusion 1984;24(6):534-5. Okubo Y, Yamaguchi H,Seno T, Araki Y, NoguchiM,ShiodaK, Takai M,DanielsG. The rare red Lan cellphenotype 1962.p. 493. Karger; M,vander M,Moes van derHart Veer M,vanLoghem JJ. bloodgroup new HoandLan:two antigens. In Vienna: population. Proceedings Association ofthe23rd oftheAmerican ofBloodBanks.1970. annualMeeting Stroup M.Five M,MacIlroy inthe Caucasian examplesofanantibodydefiningantigenhighfrequency in theJapanesepopulation. Vox Sang. 1978;35(4):265-7. Anexampleofanti-Jra ofJraantigen andfrequency K. causinghemolyticdiseaseofthenewborn H,Ito Nakajima 2012. ME,Lomas-FrancisReid C,OlssonML. The BloodGroup AntigenFacts 3ed. SanDiego:Academic Book. Press; Daniels G.HumanBloodGroups. 3ed. Oxford: 2013. Science; Blackwell Vox Sang. 1990;58(2):152-69. CP, et al. Blood group terminology 1990. The ISBT on Party Working Terminology Antigens. for Cell Red Surface Lewis M,Anstee DJ, Bird BrodheimJP, GWG, E,Cartron Contreras M,Crookston MC,Dahr W, DanielsG,Engelfriet Genotyping oftheJr(a-)andLan-phenotype Genotyping 159

Chapter 7 Chapter 7

25. Reid ME, Hue-Roye K, Huang A, Velliquette RW, Tani Y, Westhoff CM, Lomas-Francis C, Zelinski T. Alleles of the LAN blood group system: molecular and serologic investigations. Transfusion 2013;doi: 10.1111/trf.12285. [Epub ahead of print]. 26. Yamamuro Y, Isa K, Ogasawara K, Osabe T, Tsuneyama H, Yabe R, Okazaki H, Tadokoro K, Enomoto T, Watanabe S, et al. The new mutations of ABCB6 gene in Lan- Japanese. Vox Sang. 105[Sup 1], 230-231. 2013. 27. Sim NL, Kumar P, Hu J, Henikoff S, Schneider G, Ng PC. SIFT web server: predicting effects of amino acid substitutions on proteins. Nucleic Acids Res. 2012;40(Web Server issue):W452-W457. 28. Sunyaev S, Ramensky V, Koch I, Lathe W, Kondrashov AS, Bork P. Prediction of deleterious human alleles. Hum. Mol.Genet. 2001;10(6):591-7. 29. Imai Y, Nakane M, Kage K, Tsukahara S, Ishikawa E, Tsuruo T, Miki Y, Sugimoto Y. C421A polymorphism in the human breast cancer resistance protein gene is associated with low expression of Q141K protein and low-level drug resistance. Mol.Cancer Ther. 2002;1(8):611-6.

160 Chapter 8

Familial azotemia is caused by a duplication of the UT-B transporter

Gabor E. Linthorst1 Lonneke Haer-Wigman2 Jeff M. Sands3 Janet D. Klein3, Tiffany L. Thai3, Arthur J. Verhoeven1, Rob van Zwieten2, Maaike C. Jansweijer4, Lia C. Knegt1, Minke H. de Ru 1, Jaap W. Groothoff 4 Michael Ludwig5 Anita T. Layton6 Arend Bökenkamp7

1 Academic Medical Centre, Amsterdam, The Netherlands. 2 Sanquin Research Amsterdam and the Landsteiner Laboratory, Academic Medical Centre, Amsterdam, The Netherlands. 3 Emory University, Atlanta, GA, USA. 4 Emma Children’s Hospital / Academic Medical Centre, Amsterdam, The Netherlands. 5 University of Bonn, Bonn, Germany. 6 Duke University. Durham. NC. USA. 7 VU University Medical Centre, Amsterdam. The Netherlands.

Submitted to Journal of the American Society of Nephrology Chapter 8

Abstract We describe a family with autosomal-dominant azotemia, characterized by high serum urea levels due to strongly diminished fractional urea excretion (2 - 9%; normal > 30%), while all other kidney functions are normal. The phenotype segregates with a duplication of the SLC14A1 gene resulting in overexpression of urea transporter B on erythrocytes and in the descending vasa recta. The mechanism by which over-expression of UT-B leads to increased urea re-absorption is unclear.

162 the Kidd bloodgroupthe Kidd withtheJkantigens. vasa recta. UT-B where it formsis expressed the basis of most abundantly in the erythrocyte while UT-B duct, collecting Henle and the inner is medullary expressed in the descending was approved by theinstitutional review board oftheAcademic Centre Medical Amsterdam. consent was obtained fromInformed all family-members (or legal representatives). The study intact. and calciumdepositsinthetubulointerstitium whichwasotherwise recovered. At theageof5years, arenal biopsydemonstrated sclerosis of6out14glomeruli, dialysis, whichhisrenalpartially after function necessitatingtemporary glomerulonephritis of 18monthshedeveloped acute renal failure from histologically proven poststreptococcal hypercalciuria, stature nephrocalcinosis, andmoderate mentalretardation. short At theage respectively. Besidesazotemia, III.1hasfeatures of Williams syndrome withhypercalcemia, Presenting symptoms inII.1andII.2were failure to thriveattheagesof6and2years, equation duplication, overexpression ofUT-B andincreased urea flux. azotemiaHere afamilywithautosomal-dominant whichsegregates we report withSLC14A1 function as facilitative pathwaysfunction independent of water orions. for transmembrane transport suggestive azotemia. ofpre-renal Two have humanurea been identified, transporters which depends onthecirculating 27-70%;levels below volumebetween 30%are andvaries reabsorption predominantly inHenle’s Fractional duct. urea loopandthecollecting excretion involvesof renalUrea glomerularfiltrationandtubular function. handlinginthekidney Urea andtheexcretion isvitalfor ofnitrogen water conservation waste andisusedasamarker Introduction Estimated GFR[eGFR] wascalculated usingthe MDRDequationinadults concentrations from childhooddueto impaired early urea excretion (pedigree inFigure 1). Three Dutch familydemonstrated membersofa non-consanginous highserumurea Family characteristics Material andmethods creatinine levels. in maximalrenal concentrationcapacity, are butthey asymptomatic andhave urea normal withnonsensemutationsinSLC14A1urea yet. Individuals transporters, show somereduction are located intandemonchromosome 18. Urea transporter A [UT-A]Urea transporter coded by glomerular filtrationrate andotherarerenal tubularfunctions normal. isolated defect inrenal urea highserumurea excretion concentrations, while andvery Familial azotemia rare isanextremely autosomal by dominantdiseasecharacterized an 3

7 inIII.1,respectively. Fractional excretion [FE]ofurea (FEurea %=urea serum x100/urea serum xcreatinine SLC14A2 B [UT-B] and urea transporter coded by urine 2 No human disease has been linked to defects in Nohumandiseasehasbeenlinked 1 ). UT-A isfound inthethindescendinglimbof Familial azotemia iscausedby aduplicationoftheUT-B transporter 4,5

6 and the Schwartz andtheSchwartz SLC14A1 urine 163 x 1

Chapter 8 Chapter 8

I.1 I.2

II.1 II.2 II.3

III.1 Figure 1. Pedigree. Family members with azotemia are indicated in black

CGH array A comparative genomics hybridization [CGH] array was performed using a 4 x 180K oligo array (custom design ID: 023363. Agilent technologies Inc) as described before.8 All genome coordinates mentioned in this study are according to human genome build 19 (GRCh37). Fluorescence In-situ hybridization [FISH] and GTG banding was performed according to standard methods.

MLPA Multiplex ligation-dependent probe amplification [MLPA] was used to determine the copy number of the Jka and Jkb alleles in II.2, III.1 and four random donors. MLPA using Blood Group mixes p401 and p402 (MRC-Holland) was performed and analysed according to the manufacturer’s instructions. An individual with Jka+b- phenotype has two Jka alleles. An individual with Jka+b+ has one Jka and one Jkb allele.

DNA sequencing Sequence analysis of SLC14A1 and SLC14A2 was performed on genomic DNA of II.2 and III.1 to determine the presence of mutations in these genes.

Kidd blood group status and UT-B expression on erytrocytes Kidd blood group status was determined with routine serological tests. FACS analysis was used to quantify UT-B (Kidd) expression on erythrocytes from II.2, III.1 and 31 random Jka+b- and Jka+b+ donors. Anti-Jka or anti-Jkb (Immucor) antibody was added in a 1:8 dilution to a 3% (vol/ vol) erythrocyte suspension and incubated for 45 min at room temperature. Subsequently, erythrocytes were washed and goat anti-human-IgG (Alexa-488, Molecular Probes) antibody was added in a 1:500 dilution and incubated for 45 min at room temperature. Erythrocytes were washed and diluted in PBS and analyzed by flow cytometry and software (LSRII Becton Dickinson). Results are expressed as Mode Fluoresence Intensity [MFI].

164 previously. anti-rabbit IgG and visualized with diaminobenzidine (brown)peroxidase-linked as described Biopsy slideswere stained with1:20.000anti-UT-B antibodyfollowed by horseradish blot analysiswere snap-frozen. child diagnosed asacontrol. withminimalchangenephropathy Samplesfor served Western biopsyfrom biopsyfrom III.1wasobtainedattheageof5years. a6-yearA kidney Akidney old UT-B expression inthekidney area includestheSLC14A1 (UT-B) gene, butnotSLC14A2 Figurex1dn, 18q12.3-21.1(41570396 -42018874)x3mat(Supplementary S1). The duplicated 7q11.23(72338350 -72933599) at 18q12.3-21.1:arr II.2 andIII.1demonstrated againof448 kb for allindividuals. analysisshowedStandard karyotypes normal cytogenetic The CGH ofII.1, array Affected familymembershave a duplication ofSLC14A1 Table data 1.Laboratory urea raised. concentrationsare disproportionally accordance withfamilialazotemia. While child III.1hasmild to moderate renal failure, serum members have highplasmaurea levels andastrongly diminishedurea clearancein investigations offamilymembersareThe laboratory shown in Table 1. Three ofthefamily Pedigree analysis Results visualized usinginfrared withtheLICOR detection protein Odyssey analysissystem (Lycor). overnight followed anti-rabbitIgG. Bandswere by 2hourincubationwithAlexa680-linked previously. Biopsy samples were prepared for SDS-PAGE andproteins were separated asdescribed using aSPOT cameraandsoftware. visualized withanOlympusmicroscope andimagescollected usinga10xor40xobjective Reference range Individual III.1 II.1 II.3 II.2 I.1 I.2 9 9 After transfer oftheproteins themembranewasincubated to withanti-UT-B PVDF Cell nuclei were counter-stained were withhemotoxylin (blue).Stainedsections (years) Age 65 41 65 38 35 6 Gender m m m m

f f Creatinine (mg/dL) 0.48 0.86 0.64 0.74 1.03 0.63

Familial azotemia iscausedby aduplicationoftheUT-B transporter min/1.73m2) GFR (mL/ > 90 131 107 118 (UT-A) Figure (Supplementary S1). 89 84 66 Urea (mg/dL) 275.8 225.3 135.7 7-43 43.2 33.1 27 FE urea (%) > 30 1.4 1.5 8.9 28 32 59 165

Chapter 8 Chapter 8

FISH analysis with probes in the 18q12.3 - 21.1 region confirmed a tandem duplication in III.1. The duplication was not detected in I.1 and I.2, suggesting parental germline mosaicism. Furthermore, in III.1 a loss of 595 kb at 7q11.23 (atypical William’s syndrome microdeletion) was present. A cryptic balanced structural chromosome aberration of the 7q11.23 (72338350- 72933599) region was excluded in the parents of III.1 by FISH analysis using the probe RP11- 101D2. The Kidd serological phenotype of II.2 was Jk(a+b-), MLPA showed a copy number of 3.0 for Jka, indicating a genotype of Jkaaa that corresponds to three copies of SLC14A1. Her son III.1 had a serological phenotype of Jk(a+b+). MLPA showed 2.1 copies of the Jka allele and 1.1 copy of the Jkb allele indicating a genotype of Jkaab again demonstrating the presence of three copies of SLC14A1. As expected, the Jk(a+b+) controls had one Jka and one Jkb allele and the Jk(a+b-) controls had two Jka alleles and no Jkb allele. To determine whether mutations were present in the two known human urea transporters in the affected family members all exons and intron boundaries of SLC14A1 and SLC14A2 were sequenced in II.2 and III.1. II.2 and III.1 had no mutations in SLC14A1 (UT-B). The variation in SLC14A1 that codes for the Jk blood group antigens (c.838G or c.838A) was concordant with serology and MLPA in both II.2 and III.1. II.2 had no mutations in SLC14A2 (UT-A). One nonsynonymous variation in SLC14A2 was detected in III.1 (c.1529G>A, rs9960464). This variation is common in the population (frequency of the minor A allele is 17%) and we assume that the variation has no effect on the function of the UT-A transporter.

UT-B expression levels are increased on erythrocytes and renal cells in the affected family members To determine the effect of the duplication of the SLC14A1 gene, Jk expression levels were determined on red blood cells of two affected family member; II.2 and III.1. Jka expression, determined by flow cytomtery, was significantly higher on red blood cells of II.2 compared to red blood cells of Jk(a+b-) controls (Figure 2a). The Jka expression on red blood cells of III.1 was significantly higher compared to red blood cells of Jk(a+b+) controls, while no difference in aJk expression was determined compared to Jk(a+b-) controls. No difference of Jkb expression on red blood cells of III.1 was detected compared to red blood cells of Jk(a+b+) controls and also no difference was detected on red blood cells of II.2 compared to red blood cells of Jk(a+b-) controls (Figure 2b). This indicates a clear gene dosage effect of the SLC14A1 gene, because III.1 has the same Jka expression level as the Jk(a+b-) control samples that have the same copy number of 2 for the Jka allele, while the Jkb expression level was the same as Jk(a+b+) controls that have the same copy number of 1 for Jkb. Westernblot analysis using anti-UT-B on red cell membranes of II.2 and III.3 confirmed the overexpression of UT-B on red blood cells (data not shown).

166 of II.2andIII.1have significantly higherJk withanunpaired97.5% percentile. two-sided Statisticalanalysiswasperformed T-test. bloodcells a)Red shows 25%and75% percentile andaverage fluorescence intensity, error barsindicate the2.5%and the same Jk phenotype (p<0.0001andp=0.0003,respectively). Jk the sameJkphenotype difference inJk of III.1were comparablewithexpression levels controls onred (p=0.8498).b)No bloodcellsof Jk(a+b-) cells ofcontrols(p=0.4896andp 0.0512, respectively). with thesameJkphenotype in theIMCDhumans(data notshown). cells. As expected, theIMCDstained positive for UT-A1, indicating thatUT-B isnotexpressed [IMCD] showing UT-B duct collecting staining inthevasarecta,butnotinnermedullary (Figure tissuefrom 3cand3d).Renal anothercontrol containedinnermedullasections Western blot analysis demonstrated that expression of UT-B was roughly doubled in III.1 expression ofUT-B indicating orthotopic sections but notincortical (Figure 3aand3b). both,staining waspositive formedulla sections.In UT-B in thedescendingvasarecta, biopsiesofpatientIII.1andthecontrol were andouter The kidney to restricted cortex (open boxes). The Jk Figure 2. Expression oftheJkantigensonred bloodcellsofII.2andIII.1(grey boxes) andcontrol samples a) a) Mode Fluoresence Intensity (arbitrary units)

Controls Jk(a+b-) b expression levels wasdetected inred bloodcellsofII.2andIII.1compared to red blood II. 2 Jk( a+b- a (a)andJk Con )

trols Jk(a+b+) b b) (b)expression levels werecytometry. by determined Box flow plot III. 1 J k( a+ b+ ) Con a expression levels compared to red bloodcellsofcontrols with trol

Jk (a-b + ) Familial azotemia iscausedby aduplicationoftheUT-B transporter 1000 600 800 200

Con 400

trol 0 s Jk(a+

b -) p=0.4896 a expression levels onred bloodcells II.2 Jk(a+b-)

Controls Jk( Samples a+b+)

III.1 Jk(a+b+) p=0.0512

Cont rol Jk (a-b + ) 167

Chapter 8 Chapter 8

ccc

ddd

Figure 3. UT-B expression in the kidney. a) + b) Renal biopsy of the affected child including outer medulla section showing UT-B staining (brown) of vasa recta, a) magn. 10x, b) magn. 40x. UT-B expression was detected in the descending vasa recta, but not in cortical sections indicating orthotopic expression of UT-B. c) Western blot analysis of kidney biopsy samples of the affected child and an aged matched control probed for UT-B. d) A graph of the densitometries in arbritrary units of the UT-B western blot band normalized for loading differences. The UT-B expression was doubled in the renal cells of III.1 compared to the healthy control.

Discussion In this family with familial azotemia, high serum urea concentrations due to increased reabsorption of urea segregate with a duplication of SLC14A1, the gene coding for urea transporter UT-B. No mutations were detected in the SLC14A1 or the SLC14A2 genes that encode the two known urea transporters in humans. Furthermore, the five additional genes (SIGLEC15, EPG5, PSTPIP2, ATP5A1 and HAUS1) present in the duplicated area do not play a role in urea transport or urea homeostasis. We therefore assume that the phenotype in this family is due to the duplication of SLC14A1. Moreover, UT-B expression levels were significantly increased on red blood cells and in the descending vasa recta of the affected family members. In accordance with previous reports of cases with familial azotemia, renal function is normal in the affected adults.4,5 The affected child (III.1) has stage 2 CKD which appears to be related to his history of glomerulonephritis with acute renal failure and hypercalcemia/nephrocalcinosis from coinciding atypical Williams syndrome. The concerted action of urea transporters in dedicated parts of the renal vasculature and along the renal tubule is essential for the trapping of urea in the renal pyramids and for the concentration of urine. Following filtration across the glomerular basal membrane, intratubular

168 Layton’s intherat mathematicalmodelofurea transport rapidly lowers intracellularurea concentrationsintheascendingvasarecta. prevents back-leak ofurea into thesystemic circulation. To thesameend, UT-B in erythrocytes ascending vasarectadonotexpress butare to urea urea, highlypermeable transporters which ofureathe bloodstream towards leadingto nettransport thetipofrenal pyramid. The alongthedescendingvasarectaby UT-BUrea transport isdirected from theinterstitium to regulated undertheinfluenceofADHleading to increased urea inanti-diuresis. reabsorption in thepresence oflarge urea. concentrationsofosmoticallyactive are up- These transporters are abundantlypresent intheIMCD, where urea isreabsorbed thereby limitingwater losses UT-A1 Urea duct. transporters collecting andUT-A3osmole inthetubularfluidcortical of theinner medulla. UT-A2loop viaurea transporter astubularurea concentrationislower thanintheinterstitium urea asaresult concentrationinitiallyrises ofsecretion into thethindescendinglimbofHenle’s urea excretion by 50%inUT-B mice. knock-out urea inthepathogenesisofcomplicationsassociated withrenal failure. by theseauthors. Patients withisolated azotemia insight into may therole provide of further ortoto changesintheprostaglandin mutationsinurea metabolismassuggested transporters remains to beelucidated whethernon-familialazotemia by Conte described etal. It inerythrocytes. andthedescendingvasarectaincreased activity both inerythrocytes duplication of the conclusion,autosomal dominantfamilial azotemiaIn inthisfamilyisassociated witha affected individualscannotbeexcluded basedonthesefindings. is inlinewithdatafrom otherspecies. However, expression ectopic ofUT-B intheIMCD intheIMCD.transporters UT-B wasnotexpressed intheIMCDhealthy control, which toin III.3wasrestricted theouter medullaanddidnotallow for studyingexpression ofurea in thefamilypresented doesnotinvolve theUT-A gene. Unfortunately, biopsy thekidney ofUT-A.up-regulation Althoughthegenesfor UT-A andUT-B lieadjacent,theduplication i.e.azotemia duct, ismostcompatiblewithanincrease inthecollecting inurea permeability impair urea inourprobants. excretion asobserved ofpatientswithfamilial The phenotype predicts that overexpression of UT-B in the descending vasa recta should duct. concentrations, inthecollecting whichimpairsureaTherefore, re-absorption thismodel ureain thedescendingvasarectaurea inthismodelleadsto increased intramedullary SLC14A1 gene leading to overexpression UT-B of the urea transporter 10 Following inthethin ascending limb, NaCl absorption urea isthe major Familial azotemia iscausedby aduplicationoftheUT-B transporter 12 Simulatingincreased urea permeability 11 correctly predictsadecrease correctly in increase rather than 14

13 isrelated 169

Chapter 8 Chapter 8

References 1. Klein JD, Blount MA, Sands JM. Urea transport in the kidney. Compr.Physiol 2011;1(2):699-729. 2. Frohlich O, Macey RI, Edwards-Moulds J, Gargus JJ, Gunn RB. Urea transport deficiency in Jk(a-b-) erythrocytes. Am.J.Physiol 1991;260(4 Pt 1):C778-C783. 3. Sands JM, Gargus JJ, Frohlich O, Gunn RB, Kokko JP. Urinary concentrating ability in patients with Jk(a-b-) blood type who lack carrier-mediated urea transport. J.Am.Soc.Nephrol. 1992;2(12):1689-96. 4. Hsu CH, Kurtz TW, Massari PU, Ponze SA, Chang BS. Familial azotemia. Impaired urea excretion despite normal renal function. N.Engl.J.Med. 1978;298(3):117-21. 5. Armsen T, Glossmann V, Weinzierl M, Edel HH. [Familial proximal tubular azotemia. Elevated urea plasma levels in normal kidney function]. Dtsch.Med.Wochenschr. 1986;111(18):702-6. 6. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann.Intern.Med. 1999;130(6):461-70. 7. Schwartz GJ, Munoz A, Schneider MF, Mak RH, Kaskel F, Warady BA, Furth SL. New equations to estimate GFR in children with CKD. J.Am.Soc.Nephrol. 2009;20(3):629-37. 8. Barge-Schaapveld DQ, Maas SM, Polstra A, Knegt LC, Hennekam RC. The atypical 16p11.2 deletion: a not so atypical microdeletion syndrome? Am.J.Med.Genet.A 2011;155A(5):1066-72. 9. Doran JJ, Klein JD, Kim YH, Smith TD, Kozlowski SD, Gunn RB, Sands JM. Tissue distribution of UT-A and UT-B mRNA and protein in rat. Am.J.Physiol Regul.Integr.Comp Physiol 2006;290(5):R1446-R1459. 10. Wade JB, Lee AJ, Liu J, Ecelbarger CA, Mitchell C, Bradford AD, Terris J, Kim GH, Knepper MA. UT-A2: a 55-kDa urea transporter in thin descending limb whose abundance is regulated by vasopressin. Am.J.Physiol Renal Physiol 2000;278(1):F52-F62. 11. Layton AT. A mathematical model of the urine concentrating mechanism in the rat renal medulla. I. Formulation and base-case results. Am.J.Physiol Renal Physiol 2011;300(2):F356-F371. 12. Yang B, Bankir L, Gillespie A, Epstein CJ, Verkman AS. Urea-selective concentrating defect in transgenic mice lacking urea transporter UT-B. J.Biol.Chem. 2002;277(12):10633-7. 13. Conte G, Dal CA, Terribile M, Cianciaruso B, Di MG, Pannain M, Russo D, Andreucci VE. Renal handling of urea in subjects with persistent azotemia and normal renal function. Kidney Int. 1987;32(5):721-7. 14. Linthorst GE, Avis HJ, Levi M. Uremic thrombocytopathy is not about urea. J.Am.Soc.Nephrol. 2010;21(5):753-5.

170 Chapter 8 information Supporting Familial azotemia iscausedby aduplicationoftheUT-B transporter 171

Chapter 8 Chapter 8

Supplementary Figure S1. Oligomicroarray results. Values along the y-axis represent log2 ratios of patient : control signal intensities. Genes (with OMIM reference) in the interval are shown as blue horizontal bars. Results showed a single copy gain of 30 probes from 18q12.3-21.1, ~448kb in size (chr 18:41570396-42018874 Based on UCSC 2006 hg18 assembly) in the index patient (III:1). Probes are ordered in the x-axis according to physical mapping position, with the most proximal 18q12.3 probes to the left and the most distal 18q21.1 probes to the right. The genomic position (UCSC v280) of the SLC14A2 gene is chr18:42,792,947-43,263,060, of the SLC14A1 gene chr18:43,304,092-43,332,485, and of the last oligo with normal log2 ratio is 18:41547618-41547677.

172 Chapter 9

General discussion

General Discussion

For a safe transfusion practice it is important to transfuse compatible donor red blood cells to an immunized patient. Blood group antigen status in blood donors and blood recipients is most often determined via serological typing. Nevertheless, serological typing is not always possible, for instance, in patients who recently received a red blood cell transfusion or in patients with auto-antibodies. Furthermore, serological typing can be cumbersome if commercial reagents are not available or when an antigen can have low expression levels. Prediction of blood group antigen expression via genotyping can overcome situations in which serology is impossible or impractical.1 At the moment some blood centers have implemented genotyping assays to accurately predict the blood group status of blood donors and recipients and blood group genotyping has also been used to screen blood donors for rare blood types in a high- throughput fashion. Nevertheless, most blood group genotyping assays are still used as an add-on tool to supplement serological typing, instead as a stand alone assay used to predict antigen expression. In this thesis we focused on the identification of the genetic basis of rare blood group variants, characterization of the effect of variant alleles on antigen expression and used this knowledge for the development of genetic assays to predict blood group antigen status. The results of our studies and the feasibility of blood group genotyping in general are discussed below.

For most blood group systems prediction of antigens via genotyping is relatively simple, because only a single nucleotide variation is responsible for most antigenic differences. However, the reality of blood group genotyping is not as straightforward as just described. Almost all blood group systems have next to the alleles coding for the antithetic antigens so called variant alleles. These variant alleles have next to the antigenic variation (an) additional mutation(s) that can cause weakened expression of the antigen, can alter antigen expression or can cause the complete absence of the antigen. It is clinically relevant to detect variant alleles, because recipients who have altered antigen expression or lack the complete antigen are at risk of immunization when they come in contact with antigen positive red blood cells. Furthermore, red blood cells from individuals with weakened or altered antigen expression, can immunize individuals who are negative for the antigen. Hence, blood group genotyping assays should also target variant alleles, next to the wild-type alleles that code for the antithetic antigens. A general advantage of genotyping tests above standard serology is that in one single test multiple targets (antigens) can be detected. However, most available blood group genotyping assays only target a small number of blood group systems. 2-6 These methods can not cope with the extensive multiplexing that is needed to target all clinically relevant blood group systems. Another disadvantage is that these methods have high start-up and reagents costs.

175 Chapter 9 Chapter Chapter 9

In chapter 2 we developed and validated a novel genotyping assay based on the Multiplex Ligation-dependent probe amplification [MLPA] technique, the blood-MLPA assay can predict the presence or absence of 48 blood group alleles and 112 variant alleles of the MNS, Rh, Lutheran, Kell, Lewis, Duffy, Kidd, Diego, Yt, Scianna, Dombrock, Colton, Landsteiner-Wiener, Gerbich, Cromer, Knops, Indian and Ok blood group systems.7 We determined that the blood- MLPA can reliably predict the presence and absence of virtually all clinically relevant blood group antigens, except the ABO antigens. The phenotype prediction of the blood-MLPA is at least as accurate as the determination of blood group antigens by serological typing. Moreover, the blood-MLPA assay has not yet reached its maximum typing capacity and it is still possible to add 29 new “targets”, for instance mutations that cause the Jr(a-), Lan- or Vel- phenotypes, of which the genetic basis was detected after we validated the blood-MLPA assay. Because only standard laboratory equipment is needed to perform the assay, the blood-MLPA makes genotyping of blood donors and recipients more feasible. The main advantage of the blood-MLPA and of blood group genotyping in general over serological typing is the direct recognition of variant alleles that cause weakened or alter antigen expression. These variant alleles have been described for the ABO, MNS, Rh, Lutheran, Kell, Duffy, Kidd, Dombrock, Colton, Landsteiner-Wiener, I, RHAG, JR, LAN and VEL blood group systems.7 Standard serology has insufficient sensitivity to detect blood donors with very weak antigen expression. Consequently, donors with very weak antigen expression are incorrectly determined antigen negative, while their red blood cells are able to induce immunization and/or a hemolytic transfusion reaction in a recipient that is negative for the antigen.8,9 The D antigen is notorious for the frequent occurrence of variant expression and more than 200 RHD variant alleles have been described. Serology can easily determine normal D+ and D- antigen expression, however, only by extensive serology, sometimes requiring rare patient sera, specific D variants can be recognized. Nowadays blood centers have implemented genetic assays as an additional tool to specify the RHD variant alleles. In chapter 3 we determine that a subset of the blood-MLPA assay, the RH-MLPA that entirely focuses on the detection of RHD and RHCE variant alleles, can reliably determine the majority of RHD variant alleles, while serology could not determined the specific variant. Correct determination of RHD variant alleles is important, because the presence of an RHD variant allele in an individual influences blood transfusion practices. The best way to define RHD variant alleles to efficiently facilitate blood transfusion practices is still being debated. Classically, variant RhD expression is divided into weak D expression and partial D expression, in which weak D individuals are assumed to be incapable of D-immunization, because they express all D epitopes. While partial D individuals are prone to be immunized to the D antigen, because they lack one or more D-epitopes. The partial D and weak D terms are however misleading for transfusion practices. Carriers of some weak D variants can become immunized to the D antigen.10 On the other hand for some partial D an immunization event to D antigen has never been described, for instance

176 General Discussion for the RHD*05.05, RHD*16 and RHD*18 variant alleles11. Recently, Daniels proposed to replace the terms weak D and partial D with a single collective term: D variant.12 In donor typing the collective term D variant can be implemented, because it is irrelevant whether a blood donor caries a weak D or partial D variant; both are able to immunize a D- recipient. However, I believe that in blood recipient typing the collective term D variant is impractical, because it is of clinical relevance (carriers of many RHD variant alleles can become immunized to the D antigen) and of practical relevance (to minimize the unnecessary use of relatively scarce D- red blood cells) to distinguish between D variants. To keep the amount of D immunization in blood recipients with an RhD variant as low as possible and the use of D- red blood cells acceptable, blood recipients with the frequently occurring weak D type 1, 2 and 3 variants (as proposed by Daniels) and also blood recipients with the weak D type 5 variant that all have no to a very low risk of immunization to the D antigen, should receive D+ red blood cells. While individuals with all other D variants should receive D- blood. To be able to follow these criteria RhD typing should be supplemented with a genotyping assay, because serology can not recognize the weak D type 1, 2, 3 or 5 variants. Before we can discuss whether D antigen prediction via genotyping can even completely replace serological D typing, the main drawback of genotyping assays must be discussed. The main drawback of genotyping assays is the presence of null alleles. Null alleles are variant alleles that have mutations which cause the complete lack of antigen expression. If not tested for these mutations, blood recipients carrying a null allele might be incorrectly typed as antigen positive. In most blood group systems the frequency of null alleles is very low, for instance for the Lutheran, Kidd, Diego, Dombrock, Colton, Cromer and Vel blood group systems. Furthermore, for most blood group systems only a small number of null alleles have been described that can easily be targeted in a genotyping assay, for instance in the MNS, Duffy, Diego, Dombrock, Cromer, RHAG and Vel blood group systems. The D antigen is one of the few antigens in which null alleles frequently occur. Especially in African Blacks the frequency of D- null alleles is high, in 81% of D-negative individuals the RHD*Ψ or RHD*03N.01 null allele is present.13 Moreover, because these null alleles are most often next to a “deleted” allele, carriers of D- null alleles are at high risk of immunization. In several studies among D- Caucasian blood donors over 50 different D- null alleles have been described14 The blood- MLPA targets twelve of the frequently occurring D- null alleles, including the RHD*Ψ and RHD*03N.01 alleles. As was demonstrated in chapter 3, we failed to detect only one D- null variant allele, the novel RHD*443G allele, in a selected set of DNA samples. To determine the true false positive rate of the blood-MLPA in Dutch D- individuals, an assessment of a cohort of extensively typed D- individuals needs to be performed. Moreover, the analysis of the missed D- null alleles might help us to improve the blood-MLPA. Yearly ~25.000 D- Dutch pregnant women are tested using a quantitative fetal RHD genotyping assay. This gave us the opportunity to investigate the presence of RHD variant alleles in a

177 Chapter 9 Chapter Chapter 9 cohort of 37,764 D- pregnant woman. In this quantitative fetal RHD genotyping assay all variants carrying RHD exon 5 and/or exon 7 will be detected, which is the far majority of the variant alleles and only RHD*14.01, RHD*14.02 and RHD*03N.01 are missed. In chapter 4 we show that in 0.96% (95% CI 0.86% - 1.05%) of the Dutch D- pregnant women an RHD variant allele is present, which disturbs the quantitative fetal RHD genotyping assay. In 53% of the women with a variant allele a D- null allele is present. In 84% of the women with a D- null allele the RHD*Ψ allele was present, in 4% of the women one of five different D- null alleles, that were detected by the blood-MLPA, was present. In the remaining 12%, whom carried five known and eight novel D- null alleles, the blood-MLPA gave normal wild type RHD results. Hence, to increase the specificity of the blood-MLPA for the Dutch population, the assay should be extended with probes targeting the known RHD*922C>T and novel RHD*1074- 1G>A D- null alleles, that both were detected in three cases of this study. It is impractical to extend the blood-MLPA to target the other eleven D- null alleles that were present in this study, but for which the MLPA has no detecting probes. These alleles are very rare, they were only detected in one or two individuals of this study, respectively. The variant alleles that were not detected by the blood-MLPA assay were all, except one allele, linked with the RHCE*02 (RhCe; n = 9) or RHCE*03 allele (RhcE; n = 3). A total of 45 different RHD variant alleles, including 14 novel variant alleles, were detected and the variant alleles caused in 53% of the woman the D- phenotype, in 32% a partial D phenotype and in 15% a weak D phenotype. Taken the complete cohort into account, we conclude that in 0.26% (CI 0.21% - 0.31%) of the pregnant women the blood-MLPA demonstrated that serology was incorrect, these cases were actually positive for a variant allele that causes weak D expression, while 0.05% (95% CI 0.03% - 0.08%) of the D- women would be incorrectly determined D+ when the phenotype was determined solely on the blood-MLPA assay. Indicating that the blood-MLPA assay can more accurately determine the D- phenotype compared to serological typing. Three large studies, performed in Germany15, Poland16 and Austria17, on the presence of RHD variant alleles in D- donors have been performed. These studies detected a lower percentage of RHD variant alleles, 0.6%, 0.2% and 0.4%, respectively. However, these studies were performed on serologically typed D- donors. Serological donor typing has been developed to determine donors who carry the RHD*06 allele as D+, in contrast to serological typing of blood recipients and pregnant women, who are determined D-. Hence, these studies did not detect individuals carrying the RHD*06 allele. Furthermore, all three studies deliberately did not detect persons positive for the RHD*Ψ variant allele. If the frequency of variant alleles in our study is recalculated and the individuals positive for the RHD*Ψ or the RHD*06 allele are removed, the variant allele frequency of 0.33% (95% CI 0.28% - 0.39%) fits exactly in the frequency determined in the three other studies. In the German, Polish and Austrian studies, the blood- MLPA would not have recognized the presence of a D- null allele in 0.02%, 0.02%, and 0.004% of the donors, respectively. This lower percentage of undetected D- null alleles might be due

178 General Discussion to the more variable ethnic background of our Dutch pregnant women, compared to the Caucasian background of the tested blood donors. In agreement with the results of chapter 3, we again demonstrated the superiority of the blood-MLPA over serological typing of blood donors with weak D expression; in 0.07%, 0.09% or 0.21% of the individuals, respectively, the false negative serological results would have been corrected by the blood-MLPA. We can therefore conclude that, at least in West and Central Europe, genotyping of the D- phenotype of blood donors via the blood-MLPA assay is more accurate compared to serological typing. Furthermore, due to the very low false-positive rate of the blood-MLPA, this assay can also be used to type blood recipients or pregnant women for the D phenotype.

Genotyping of the Vel blood group has been a long desire, because serological typing of the Vel antigen is very cumbersome, due the variable antigen expression and due to the lack of commercially available reagents.18 Genotyping was, however, not possible, since the genetic basis of the Vel antigen was not yet determined. The discovery of the genetic basis of the Vel blood group antigen is described in chapter 5. The Vel blood group antigen is encoded by the SMIM1 gene and the Vel- phenotype is caused by the homozygous presence of a 17-nucleotide frame shift deletion that most likely causes the complete absence of the SMIM1 protein. This is supported by the fact that in an individual heterozygous for the SMIM1*64_80del allele wild- type SMIM1 mRNA is detected, while mRNA with the 17-nucleotide deletion was absent.19 In chapter 6 we determine that weakened Vel expression is caused by the major allele of rs1175550 small nuclear polymorphism in intron 2 of SMIM1 in combination with a SMIM1*64- 80del, SMIM1*152T>A or SMIM1*152T>G allele. Nevertheless, not in all cases with weak Vel expression levels genetic variation in SMIM1 could be hold responsible for the weakened expression and we conclude that genetic factors outside the SMIM1 gene or environmental factors can also be responsible for weak Vel expression. To overcome the difficult serological typing of the Vel antigen we developed and validated a high-throughput genotyping assay to detect blood donors with the Vel- phenotype. In 3,366 screened Caucasian donors we were able to detect two new Vel- blood donors.

In the previous paragraphs we have showed that for most blood group systems genotyping can accurately predict antigen expression. The most clinical relevant blood group system, the ABO system, has, however, not yet been mentioned. The ABO blood group system is actually the most difficult blood group system to genotype. The ABO gene codes for an enzyme that determines the antigen expression, therefore mutations in ABO have an indirect effect on antigen expression. Furthermore, to correctly predict antigen expression of the ABO blood group system it must be decided if mutations that are almost 2 kilo bases apart are present in cis or in trans. The currently available genotyping assays target specific mutations and none can determine whether mutations are in cis or in trans. The correct ABO (variant) allele can

179 Chapter 9 Chapter Chapter 9 only be determined when genetic data is supplemented with serological information. Also for other blood group system antigen prediction via genotyping is still impossible or impractical. In chapter 7 we show that the phenotypic prediction of the JR and LAN blood group systems by standard genotyping assays is impractical. Due to the large heterogeneity of mutations that cause the Jr(a-) or Lan phenotype the developed genotyping assays were not efficient in a Caucasian population. Until suitable reagents are broadly available, serological screening for the Jra and Lan antigen needs to be combined with genetic confirmation of the Jr(a-) and Lan- phenotype to detect Jr(a-) and Lan- individuals.

In chapter 8 of this thesis we describe the functional consequences of a rare variation in a gene encoding the Kidd blood group antigen. The genotyping assays we developed to predict the blood group phenotype, were shown to be equally useful to explain a functional phenotype. A duplication of the SLC14A1 gene encoding the Kidd blood group antigen was detected in a family with the rare dominant familial azotemia, of which affected individuals have increased urea concentrations in the blood and decreased urea clearance, while other renal tubular functions are normal. The protein, the UT-B urea transporter, on which the antigens of the Kidd blood group are present, is expressed on red blood cells and also on kidney cells. No serological difference for the Kidd blood group antigens were detected in the affected family members, however, a significantly increased Kidd antigen expression on red blood cells and also a significantly increased UT-B expression in the kidney were detected.

In conclusion, the results presented in this thesis indicate that the prediction of most blood group antigens, including the D antigen, via genotyping is accurate and reliable. Presently available genotyping assay, such as the blood-MLPA, can be used to replace serology for blood group typing of blood recipients, blood donors and pregnant women, when serology is not possible. Genotyping has proven its value in the prediction of antigen expression in blood donors, because unlike genotyping, serology is unable to detect and therefore correctly type very weak expression of an antigen. Furthermore, current genotyping techniques have enough capacity to target the most frequently occurring null alleles in a single assay and can achieve a very low rate of false-positive antigen detection in blood recipients, as we show for D antigen prediction using the blood-MLPA assay. The continuous improvement of genotyping techniques will ensure that the prediction of blood group antigen expression will become even more accurate and more complete than at present time. For instance, blood group genotyping via next generation sequencing can be used to predict the JR and LAN blood group system, with a large heterogeneity of causative mutations. Because next generation sequencing does not target specific mutations, but can theoretically target entire blood group genes. The first next generation blood group genotyping assay has been developed, however, first test showed that improvement of the

180 General Discussion assay is needed.20 Also for the ABO blood group system next generation might be able to correctly predict antigen expression, when the read lengths become long enough to cover the almost 2 kilo bases to determine whether mutations are in cis or in trans. A fourth generation sequencer in which a nanopore is used to “read” a DNA strand21, could easily cover the 2 kilo bases. But despite the very positive views from the companies that develop this technique, nanopore next generation sequencing is still in a developmental stage. When genetic typing of any antigen will become the golden standard over serological typing, the implementation of genotyping for other blood group antigens will gain momentum, because genotyping assays become relatively cheaper in comparison with serology when more antigens are tested. However, before genotyping assays can replace serology, large population studies need to confirm that the frequency of incorrect antigen prediction, due to variants that are not detected by a genotyping assay, is at a minimum and most importantly lower than the error rate seen with serological typing. When blood group genotyping assays are implemented in blood centers, the greatest benefit for transfusion practices will arise from comprehensive matching between blood donors and recipients. In the Netherlands, as in most countries, only ABO and D matched red blood cells are transfused in recipients. With the exception of women under 45 and patients with thalassaemia and sickle cell disease, who receive more extensively matched blood, as well as all patients after development of a clinically relevant alloantibody. When genotyping is implemented, all blood donors and recipients will be typed for a large set of blood group antigens in a high-throughput fashion and more comprehensive preventive matching will be possible. Routine preventive matching of for instance the clinically relevant Rh, K, Fy, Jk and Ss antigens, will prevent the occurrence of most immunization reactions and as a result the occurrence of severe immediate or delayed hemolytic transfusion reaction.22,23

181 Chapter 9 Chapter Chapter 9

References 1. Reid ME, Denomme GA. DNA-based methods in the immunohematology reference laboratory. Transfus.Apher. Sci. 2011;44(1):65-72. 2. Di CJ, Silvy M, Chiaroni J, Bailly P. Single PCR multiplex SNaPshot reaction for detection of eleven blood group nucleotide polymorphisms: optimization, validation, and one year of routine clinical use. J.Mol.Diagn. 2010;12(4):453-60. 3. Hashmi G, Shariff T, Zhang Y, Cristobal J, Chau C, Seul M, Vissavajjhala P, Baldwin C, Hue-Roye K, Charles-Pierre D, et al. Determination of 24 minor red blood cell antigens for more than 2000 blood donors by high-throughput DNA analysis. Transfusion 2007;47(4):736-47. 4. Avent ND, Martinez A, Flegel WA, Olsson ML, Scott M, Nogues N, Pisacka M, Daniels G, Muniz-Diaz E, Madgett TE, et al. The Bloodgen Project of the European Union, 2003-2009. Transfus.Med.Hemother. 2009;36(3):162-7. 5. Montpetit A, Phillips MS, Mongrain I, Lemieux R, St-Louis M. High-throughput molecular profiling of blood donors for minor red blood cell and platelet antigens. Transfusion 2006;46(5):841-8. 6. Hopp K, Weber K, Bellissimo D, Johnson ST, Pietz B. High-throughput red blood cell antigen genotyping using a nanofluidic real-time polymerase chain reaction platform. Transfusion 2010;50(1):40-6. 7. Reid ME, Lomas-Francis C, Olsson ML. The Blood Group Antigen Facts Book. 3 ed. San Diego: Academic Press; 2012. 8. Wagner T, Kormoczi GF, Buchta C, Vadon M, Lanzer G, Mayr WR, Legler TJ. Anti-D immunization by DEL red blood cells. Transfusion 2005;45(4):520-6. 9. Beattie KM, Sigmund KE, McGraw J, Shurafa M. U-variant blood in sickle cell patients. Transfusion 1982;22(3):257. 10. Wagner FF, Frohmajer A, Ladewig B, Eicher NI, Lonicer CB, Muller TH, Siegel MH, Flegel WA. Weak D alleles express distinct phenotypes. Blood 2000;95(8):2699-708. 11. Daniels G. Human Blood Groups. 3 ed. Oxford: John Wiley & Sons; 2013. 12. Daniels G. Variants of RhD--current testing and clinical consequences. Br.J.Haematol. 2013;161(4):461-70. 13. Chou ST, Westhoff CM. The Rh and RhAG blood group systems. Immunohematology. 2010;26(4):178-86. 14. Wagner FF, Flegel WA. Rhesus Base 2 2013 Feb 3 Available from http://www.uni-ulm.de/~fwagner/RH/RB2/. 15. Flegel WA, von Z, I, Wagner FF. Six years’ experience performing RHD genotyping to confirm D- red blood cell units in Germany for preventing anti-D immunizations. Transfusion 2009;49(3):465-71. 16. Orzinska A, Guz K, Polin H, Pelc-Klopotowska M, Bednarz J, Gielezynska A, Sliwa B, Kowalewska M, Pawlowska E, Wlodarczyk B, et al. RHD variants in Polish blood donors routinely typed as D-. Transfusion 2013. 17. Polin H, Danzer M, Gaszner W, Broda D, St-Louis M, Proll J, Hofer K, Gabriel C. Identification of RHD alleles with the potential of anti-D immunization among seemingly D- blood donors in Upper Austria. Transfusion 2009;49(4):676-81. 18. Issitt PD, Oyen R, Reihart JK, Adebahr ME, Allen FH, Jr., Kuhns WJ. Anti-Vel 2, a new antibody showing heterogeneity of Vel system antibodies. Vox Sang. 1968;15(2):125-32. 19. Storry JR, Joud M, Christophersen MK, Thuresson B, Akerstrom B, Sojka BN, Nilsson B, Olsson ML. Homozygosity for a null allele of SMIM1 defines the Vel-negative blood group phenotype. Nat.Genet. 2013;45(5):537-41. 20. Fichou Y, Audrezet MP, Gueguen P, Dupont I, Jamet D, Le Marechal C, Ferec C. Next-Generation Sequencing (NGS) for blood group genotyping. Vox Sang. 105[Suppl 1], 59. 2013. 21. Thompson JF, Oliver JS. Mapping and sequencing DNA using nanopores and nanodetectors. Electrophoresis 2012;33(23):3429-36. 22. Heddle NM, Soutar RL, O’Hoski PL, Singer J, McBride JA, Ali MA, Kelton JG. A prospective study to determine the frequency and clinical significance of alloimmunization post-transfusion. Br.J.Haematol. 1995;91(4):1000-5. 23. Redman M, Regan F, Contreras M. A prospective study of the incidence of red cell allo-immunisation following transfusion. Vox Sang. 1996;71(4):216-20.

182 Appendix

Nederlandse Samenvatting

Dankwoord

Curriculum Vitae

Lijst met publicaties

Portfolio

Nederlandse Samenvatting

Rode bloedcellen vervoeren zuurstof naar alle cellen van het lichaam en zijn daarom essentieel voor het menselijk leven. Wanneer er tijdens een trauma omvangrijk bloedverlies optreedt, kan er niet meer genoeg zuurstof naar de vitale organen (zoals de hersenen en het hart) worden getransporteerd, met een mogelijke fatale afloop. Daarnaast hebben sommige mensen een defect in de aanmaak van rode bloedcellen (de erytropoëse) wat chronische bloedarmoede kan veroorzaken. Om massaal bloedverlies na een trauma te compenseren of om bloedarmoede in patiënten met een defect in de erytropoëse te compenseren, kunnen donor rode bloedcellen worden getransfundeerd. Maar donorbloed kan niet simpelweg aan iedere patiënt, die een bloedtransfusie moet ondergaan, worden toegediend. Op het oppervlak van rode bloedcellen bevinden zich bepaalde structuren, de bloedgroepantigenen. Wanneer een ontvanger van een bloedtransfusie antistoffen heeft tegen een bloedgroepantigeen dat aanwezig is op de rode bloedcellen van de donor, zal het immuunsysteem van de ontvanger alle donor rode bloedcellen vernietigen, wat kan leiden tot een hemolytische transfusiereactie met soms een fatale afloop. Daarom is het van belang om compatibele rode bloedcellen te transfunderen in een geïmmuniseerde ontvanger. In hoofdstuk 1 worden de verschillende bloedgroepantigenen en de genetische basis van de bloedgroepantigenen geïntroduceerd, daarnaast wordt de klinische relevantie van antistoffen tegen bloedgroepantigenen beschreven.

Op dit moment wordt de bloedgroepantigeenstatus van donoren meestal bepaald via serologische bepalingen, waarbij op de rode bloedcellen zelf wordt gekeken of een bloedgroep antigeen aan- of afwezig is. Maar serologie is niet altijd mogelijk of onhandig, bijvoorbeeld in mensen met autoantistoffen of wanneer een groot aantal bloedgroepantigenen bepaald moet worden. Het voorspellen van de aan- of afwezigheid van een bloedgroepantigeen, door middel van het genotyperen van de allelen die coderen voor het bloedgroepantigeen, kan de serologische typering vervangen. Het voorspellen van bloedgroepantigenen via een genotyperingstest wordt op dit moment alleen nog op kleine schaal uitgevoerd door bloedcentra, omdat voor de huidige bloedgroep genotyperingstesten geïnvesteerd moet worden in gespecialiseerde apparatuur en omdat ze maar een klein aantal bloedgroepantigenen kunnen voorspellen. In hoofdstuk 2 en hoofdstuk 3 hebben wij gevalideerd of een bloedgroepgenotyperingstest, gebaseerd op de Multiplex Ligation-dependent Probe Amplification [MLPA] techniek, alle klinisch belangrijke bloedgroepantigenen, behalve de bloedgroepantigenen van het ABO bloedgroepsysteem, correct kan detecteren. De bloed-MLPA test kan de aan- of afwezigheid en het copy number van 48 bloedgroepallelen en 112 variant allelen van achttien bloedgroepsystemen bepalen. De bloed-MLPA omvat verreweg de grootste set van bloedgroepantigenen ten opzichte van de huidige bloedgroep genotyperingstesten. In hoofdstuk 2 laten wij zien dat de MLPA techniek in staat is om de aan- en afwezigheid en copy number van bloedgroepallelen

185 Appendix correct the detecteren. Om correct de bloedgroepantigeenstatus te voorspellen kan de bloed-MLPA ook een groot aantal variant allelen herkenen, vooral allelen van het complexe Rh bloedgroepsysteem. In hoofdstuk 3 wordt een onderdeel van de bloed-MLPA, de RH-MLPA, die volledig gespecialiseerd is in de detectie van RHD en RHCE allelen, gevalideerd. De RH- MLPA kan de frequent voorkomende RHD en RHCE variant allelen correct identificeren, terwijl de serologische typering in de meeste gevallen alleen de aanwezigheid van een variant kon bepalen. In hoofdstuk 2 en hoofdstuk 3 laten we zien dat de bloed-MLPA op betrouwbare wijze de aan- en afwezigheid van bijna alle klinische relevante bloedgroepantigenen correct kan voorspellen. De voorspelling door de bloed-MLPA is minstens net zo nauwkeurig als de serologische bloedgroepantigeen bepaling. Omdat alleen standaard laboratoriumapparatuur nodig is om de bloed-MLPA uit te voeren, maakt deze test het genotyperen van bloeddonoren en bloedontvangers op grote schaal eindelijk meer toegankelijk. Het gebruik van deze test zal leiden tot een betere afstemming van bloedtransfusies, waardoor er minder immunisaties zullen plaatsvinden en de transfusiegeneeskunde nog veiliger wordt.

Alle Nederlandse D-negatieve [D-] zwangere vrouwen krijgen de keus om een kwantitatieve foetale RHD genotyperingstest uit te voeren om de D status van hun kind te bepalen en zo de onnodige toediening van anti-D profylaxe te voorkomen. Jaarlijks worden ~25,000 D- zwangere vrouwen onderzocht met behulp van deze test. Dit gaf ons de gelegenheid om de aanwezigheid van RHD variant allelen in een cohort van 37.764 D- zwangere vrouwen te onderzoeken. In hoofdstuk 4 bepalen wij dat ~0,96% van de D- zwangeren een RHD variant allel draagt en ~47% van deze vrouwen waren positief voor het RHD*Ψ allel. Toch werden er in 275 D- vrouwen met een variant allel 45 verschillende RHD variant allelen ontdekt, waaronder veertien nieuwe allelen. Extra serologische testen bevestigden dat 55% van de vrouwen met een variant allel D- waren en 26% van de vrouwen had partiële D expressie, 16% van de vrouwen had Del expressie en 3% van de vrouwen had zwakke D expressie. Opmerkelijk was dat bij twee vrouwen met een Del fenotype en normale RhCE expressie, geen mutaties werden gevonden in alle exonen, introngrenzen of de promotorregio van het RHD gen. Wij veronderstellen dat in deze twee gevallen een gen, dat nodig is voor de membraanexpressie van het RhD eiwit gemuteerd is.

De ontdekking van de genetische basis van het Vel bloedgroepantigeen is beschreven in hoofdstuk 5. Het Vel bloedgroepantigeen wordt gecodeerd door het SMIM1 gen. Het Vel- negatieve [Vel-] fenotype wordt veroorzaakt door de homozygote aanwezigheid van een 17 nucleotide deletie, die meest waarschijnlijk de afwezigheid van het SMIM1 eiwit op het membraan van de rode bloedcel veroorzaakt. Met de huidige kennis van de genetische basis van het Vel bloedgroepantigeen kunnen donoren onmiskenbaar als Vel- worden bepaald. In het verleden zijn meerdere donoren met zeer zwakke Vel expressie serologisch onjuist als

186 Nederlandse Samenvatting

Vel- bepaald, wat tot een hemolytische transfusie reactie had kunnen leiden als deze rode bloedcellen waren getransfundeerd bij een ontvanger met anti-Vel. Verder denken wij dat het Vel/SMIM1 eiwit een rol speelt in het ijzermetabolisme en in de ontwikkeling van rode bloedcellen. Aangezien een knock-down van smim1 in zebravissen een lichte maar consequente daling van het aantal rode bloedcellen gaf en Vel- individuen een zwakke, maar niet significante, verlaging van de ijzerreserves hadden. Daarnaast is het minor allel van de rs1175550 variatie in intron 2 van SMIM1 gekoppeld aan significant verhoogde SMIM1 transcriptie niveaus en een significant verhoogde hemoglobine concentratie in rode bloedcellen. In hoofdstuk 6 laten wij zien dat het minor allel van de rs1175550 variatie ook gekoppeld is met significant hogere SMIM1/Vel expressie niveaus op het rode bloedcel membraan. Zwakke Vel expressie wordt veroorzaakt door de aanwezigheid van de major allel van rs1175550 in combinatie met een SMIM1 allel met de c.64_80del, c.152T>A of c.152T>G mutatie. In ~6% van de gevallen kon een genetische variatie in de exonen en intronen van SMIM1 echter niet verantwoordelijk worden gehouden voor de verzwakte expressie. In deze gevallen zijn genetische veranderingen buiten SMIM1 of omgevingsfactoren verantwoordelijk voor de verzwakte Vel expressie. Vanwege de variabele expressieniveaus van het Vel antigeen is het serologische screenen van donoren met het Vel- fenotype erg lastig. De ontwikkelde hoge-capaciteit genotyperingstest detecteerde twee Vel- donoren in 3.366 Kaukasische donoren. Deze test kan worden gebruikt om het serologische screenen van donoren met het Vel- fenotype te vervangen.

Ook voor de Jra en Lan bloedgroepantigenen is serologisch screenen erg lastig. In hoofdstuk 7 hebben wij een hoge-capaciteit genotyperingstest ontwikkeld en gevalideerd om donoren te screenen voor het Jra-negatieve [Jr(a-)] en Lan-negatieve [Lan-] fenotype. Sinds meerdere mutaties in ABCG2 of ABCB6, respectievelijk, het Jr(a-) en Lan- fenotype kunnen veroorzaken, werd eerst de genetische basis van het Jr(a-) en Lan- fenotype onderzocht in de Nederlandse bevolking. In de Nederlandse Jr(a-) individuen werden drie eerder beschreven nonsense mutaties gevonden en in de Nederlandse Lan- individuen werden twaalf mutaties in ABCB6, waaronder acht nieuwe mutaties gevonden. Daarnaast werd bepaald dat het Jr(a-) of Lan- fenotype niet veroorzaakt wordt door een gehele of gedeeltelijke deletie van, respectievelijk, het ABCG2 of ABCB6 gen. De ontwikkelde hoge capaciteit genotyperingstest detecteerde de meeste voorkomende mutaties die respectievelijk het Jr(a-) of Lan- fenotype veroorzaakten. In totaal werden 3.366 Kaukasische, 621 Afrikaanse en 333 Chinese donoren gescreend. Zeventien Kaukasische donoren werden heterozygoot positief gevonden voor een mutatie die het Jr(a-) (n = 5) of Lan- (n = 12) fenotype veroorzaakt. Geen van de Afrikaanse of Chinese donoren was positief voor een van de mutaties, die werden gedetecteerd in onze test. Wij concluderen daarom dat

187 Appendix genotypisch screenen van donoren met het Jr(a-) of Lan- fenotype mogelijk is. Maar door het hoge aantal mutaties dat respectievelijk het Jr(a-) of Lan- fenotype kan veroorzaken, is de test niet efficiënt.

In hoofdstuk 8 beschrijven we de functionele consequenties van een zeldzame variatie in het gen dat codeert voor de Kidd bloedgroep. Een duplicatie van het SLC14A1 gen dat codeert voor het Kidd bloedgroepsysteem werd gedetecteerd in een familie met zeldzame dominantie familiale azotemia. De aangedane familieleden hadden verhoogde ureumconcentraties in het bloed en een verminderde ureumklaring in de nieren, terwijl andere nierfuncties normaal waren. Het SLC14A1 gen codeert voor het UT-B ureum transporter eiwit, dat tot expressie komt in de nier en op rode bloedcellen In rode bloedcellen staat de UT-B transporter beter bekent als de Kidd bloedgroep. De duplicatie werd gevonden in combinatie met een significant verhoogde UT-B/Kidd expressie op de rode bloedcellen en in de nier. Opmerkelijk is dat muizen met een knock-out van het SLC14A1 gen precies hetzelfde fenotype vertonen als de aangedane familieleden met de SLC14A1 duplicatie. Het moet nog worden uitgezocht hoe de afwezigheid en verhoogde expressie van de UT-B ureum transporter kan resulteren in hetzelfde fenotype.

In hoofdstuk 9 wordt er geconcludeerd dat de voorspelling van bijna alle klinisch belangrijke bloedgroepantigenen (waaronder het D-antigeen) via genotypering nauwkeurig en betrouwbaar is. De momenteel beschikbare genotyperingstesten, zoals de bloed- MLPA kunnen daarom de serologie vervangen. Daarnaast zal de continue verbetering van genotyperingstechnieken ervoor zorgen dat de bloedgroepsystemen die op dit moment nog niet met genotypering kunnen worden voorspeld (zoals de ABO, JR en LAN bloedgroepsystemen) in de toekomst wel correct kunnen worden voorspeld. Wanneer genetische typering van bloedgroepantigenen de gouden standaard boven serologische typering zal worden, kunnen rode bloed cellen van donoren beter op een patiënt afgestemd worden waardoor immunisaties tegen bloedgroepantigenen worden voorkomen. Als gevolg daarvan zullen er minder hemolytische transfusie reacties plaats vinden.

188

Dankwoord

Dankwoord In dik vier jaar kan je heel veel nuttige en iets minder nuttige dingen (AH spaarfiguurtjes, dopjes van epjes, etc.) verzamelen. Toen ik op mijn laatste officiële werkdag alle nuttige en iets mindere nuttige dingen aan het opruimen was, kwamen er dan ook vele mooie herinneringen boven van de tijd die ik op Sanquin heb rondgelopen (en ook vaak gerend). Ik sta nog steeds versteld van de onvoorwaardelijke hulp die ik van collega’s van de vele verschillende afdelingen heb gekregen en de ontzettend gezellige werksfeer die op Sanquin heerst. Ik heb dan ook op mijn laatste werkdag Sanquin met een stiekeme traan verlaten en wil graag iedereen bedanken die op de een of andere manier een bijdrage heeft geleverd aan dit prachtige proefschrift.

Ellen, al tijdens de eerste minuten van mijn sollicitatie gesprek voelde ik me direct op mijn gemak en had ik meer het idee dat ik een gezellig en interessant gesprek aan het voeren was, dan dat ik aan het solliciteren was. Wat was ik dan ook blij dat je mij een dag later al belde dat ik de promotieplek had. Door jouw eindeloze inzet (ook tijdens je vakanties) bevat dit proefschrift negen mooie hoofdstukken. Als ik eindelijk een smartphone koop zullen we zeker met elkaar gaan WhatsAppen!

Masja, wat had jij altijd goed door wat mij dwars zat (zelfs zonder dat ik maar één woord had gezegd). Bedankt voor al je advies en hulp en bedankt voor alle fleurige bloemetjes die je correcties van mijn manuscripten altijd sierden. Ik heb veel waardering voor jou en hoop je dan ook nog vaak te spreken!

Peter, Piet, Alwetende Bloedgroep Grootheid en natuurlijk liefste paranimpf. Mijn eerste twee weken bij Sanquin heb jij mij alle ins en outs van de bloedgroepwereld geleerd. Naast jouw oneindige kennis over bloedgroepen heb je ook veel humor en we hebben die twee weken dan ook samen veel afgelachen. Gelukkig bleef de samenwerking niet beperkt tot mijn eerste twee weken en kon ik de afgelopen vier jaar altijd bij jou terecht voor een vraag of voor een lach. Ook in de toekomst schakel ik zeker nog mijn Piet-hulplijn in!

Florentine, wat was ik blij dat jij terug kwam bij Sanquin en de gezelligheid opvulde, die ik na Peter zijn vertrek tekort kwam. Bedankt voor alle nootjes en gesprekken die je met mij hebt gedeeld. Al voordat je terug was bij Sanquin was het voor mij duidelijk dat je als paranimf naast mij zou staan!

Prof. Dr. Willem Ouwehand, bedankt dat je zitting hebt genomen in mijn promotiecommissie en bedankt voor de geweldige samenwerking die heeft geleid tot hoofdstuk 5. Ik heb veel geleerd van jouw manier van aanpak en ik heb genoten van mijn spoedbezoek aan Cambridge.

191 Appendix

Prof. dr. F. Baas, Prof. dr. E. Bakker, Prof. dr. M.H.J. van Oers en Prof. dr. J.J. Zwaginga wil ik bedanken voor het zitting nemen in mijn promotiecommissie.

Suilan, jij was mijn eerste analist helemaal voor mezelf en wat was ik blij met je. Sinds jij was begonnen hoefde ik alleen nog maar aan te geven wat er gedaan moest worden en werd het allemaal netjes en accuraat uitgevoerd. Veel succes met je eigen wetenschappelijke carrière.

Aicha, ook jij hebt geweldig werk voor mij geleverd. Aan een half woord had je genoeg. Daarnaast waren het bufferlab en pre-PCR lab perfecte ruimtes om gezellig bij te kletsen over alle dagelijkse beslommeringen.

Peter, mijn superkamergenootje, ik ben erg blij dat ik mijn promotietijd met jou heb kunnen delen en heb dan ook veel (ook wetenschappelijk) van je geleerd. Bedankt voor alles en ik kijk alweer uit naar het volgende dinertje met rode wijn.

Yanli, the year that you were in Holland was amazing. I really enjoyed working with you and hope to see you soon.

Bernadette, ook jij hebt een speciaal plekje in mijn proefschrift veroverd. Ik zal de gezelligheid vooral toen we samen de MLPA’s uitvoerden nooit meer vergeten. Door jou zijn mijn jaren bij Sanquin er fleuriger op geworden.

Anita en Mo, zonder jullie zou de afdeling in een grote chaos veranderen. Bedankt voor alle hulp en al het uitzoekwerk zodat ik nu kan promoveren.

Remco, Ridge, Renate en Shabnam, wat heb ik een geluk gehad met vier hardwerkende en gemotiveerde studenten. Het resultaat mag er dan ook zijn en pronken jullie namen in vier hoofdstukken van dit proefschrift. Ik weet zeker dat ik jullie in de toekomst nog tegen kom!

Helga, Lusy, Rick, Sabrina en Sofieke, voor jullie is het einde bijna in zicht. Heel veel succes met het afronden van jullie proefschriften, ik kijk uit naar het resultaat!

Tamara, Esther, Jalenka en Gillian, jullie hebben nog even te gaan. Het is mij niet gelukt om me eraan te houden, maar hier is mijn goede raad: stress niet te veel en ga naar alle borrels van Sanquin!

Annemieke, Arthur, Barbera, Magdalena en Onno, bedankt voor alle gezelligheid als kamergenootjes.

192 Dankwoord

Gestur en Emile, de deur stond altijd open tussen onze kamers en dat heeft voor hele mooie spontane discussies gezorgd. Daphne en Carlijn, jullie kamer is dan wat verder weg, maar tijdens de vrijdagochtend besprekingen kwamen en komen jullie altijd met goede ideeën en adviezen.

Erytrocytenserologie en natuurlijk erytrocytengenetica (ook al bestaat deze afdeling dan nog niet), af en toe was ik vaker op jullie lab dan op mijn eigen lab te vinden, maar jullie zorgden er altijd voor dat ik me bij jullie thuis voelde.

Aniska, Carlijn, Franca, Marion, Suzanne en Remko, bedankt voor jullie tomeloze inzet op de labs.

Martin Lodén, zonder jou zou dit proefschrift twee hoofdstukken korter zijn. Ik waardeer je inzet voor de bloed-MLPA enorm en ik was altijd erg blij als je weer eens op de fiets langs kwam om wat epjes uit je zak te toveren.

Gabor Linthorst en Arend Bökenkamp, bedankt voor de fijne samenwerking en het vertrouwen van jullie in mij als wetenschapper.

Robin van Bruggen en Sacha Zeerleder, bedankt voor het zitting nemen in mijn OIO- begeleidingscommissie. Ondanks dat niet al mijn bijeenkomsten zijn door gegaan, heb ik er veel aan gehad.

Ada, Christa, Denise, Naomi, Margreet, Ine, Ingrid, Lily, Mark, Erik, Rob en met name Mariëtte, ik kon altijd bij jullie terecht als de hoeveelheid werk me weer eens boven het hoofd groeide. Bedankt!

Martin, Mariëtte en Karin, ik weet niet hoeveel sequentiereacties ik heb ingezet, maar het waren er in ieder geval heeeeeeel veeeeeeeeeeeeel. Bedankt dat jullie altijd zonder problemen al mijn samples op de sequencer hebben gezet.

Collega’s van Hematopoiese, door de oprichting van deze nieuwe afdeling werd de kantoorruimte twee keer zo krap, maar de gezelligheid minstens twee keer zo groot.

Josephine, Max, Mohamed, Vincent en Werner van Cryobiologie, bedankt voor het snelle leveren van de 450 rietjes en alle andere “aparte” aanvragen van Piet (dus stiekem voor mij). Ik zie de rietjes nog steeds voorbij huppelen in mijn dromen!

193 Appendix

Bert Tomson, Fikreta Danovic en Rianne Koopman van de Klinisch Consultative Dienst, bedankt voor de hulp bij het aanvragen van de Vel- donoren, mede door jullie snelle handelen is hoofdstuk 5 een prachtstuk geworden.

Simon, Sander, Carmen en Jacqueline, jullie behulpzaamheid was overweldigend, maar de Mass Spec zal nooit mijn beste vriend worden.

Collega’s van de afdelingen Laboratorium voor Celtherapie, Rodecel Diagnostiek, HLA Diagnostiek, Vaderschapsonderzoek en het Nationaal Screeninglaboratorium, door jullie bijdrages kon ik mijn hoofdstukken nog mooier afronden.

Marloes, als enige van het biologengroepje ben jij ook bezig met een promotie en ik heb er alle vertrouwen dat jij een glansrijke carrière in de wetenschap zal hebben. Bioloogjes, ondanks dat bijna iedereen Nijmegen heeft verlaten is het altijd super als we weer met zijn allen samen zijn.

Jan en Joke, bij jullie kan ik altijd ontspannen met een crackertjes met castello of tapenade, een glas rode wijn en een potje boerenbridge.

Marloes, Jordan, Louis en Arthur, wat geniet ik (en Job) ervan om langs te komen in Putte- Kapelle. Jullie deur staat altijd open en daar maken wij graag gebruik van.

Wouter, bedankt voor het maken van de prachtige omslag voor dit proefschrift. Ik waardeer het enorm dat je altijd bereid bent om je creatieve inbreng toe te voegen. Mirjam, binnenkort zijn wij ook officieel familie, ik kan niet wachten op jullie grote dag.

Lieve Papa en Mama, zonder jullie had dit alles niet plaats kunnen vinden. Er zijn niet genoeg woorden of kaartjes om jullie te bedanken voor alles wat jullie voor mij hebben gedaan.

Job, bedankt dat je er altijd voor mij bent!!!

194 Curriculum Vitea

Curriculum Vitea Lonneke Haer-Wigman was born on 13th of June in Philippine, Zeeuws-Vlaanderen, The Netherlands. From 1998 to 2003 she attended the Stedelijke Scholengemeenschap De Rede high school in Terneuzen. In 2003 she started her study Medical Biology at the Radboud University Nijmegen. As a student she worked as a tutor/practical assistant at courses of the departments of Statistics, Molecular Biology and Biochemistry. As part of her study she performed three internships at the departments of Pathology, Biomolecular Chemistry and Farmacology and Toxicology and Human Genetics. All her internships had a common main theme; the role of genetic variation in a specific disease, Glioma brain tumors, Cartilage Hair Hypoplasia and Rheumatoid Arthritis, respectively. After obtaining her Master in Medical Biology in 2008, she started her PhD project at the department of Experimental Immunohematology under the supervision of Prof. Dr. Ellen van der Schoot and Dr. Masja de Haas. The results of the research performed are described in this thesis.

195

List of publications

List of publications 1. Wagener FA, Toonen EJ, Wigman L, Fransen J, Creemers MC, Radstake TR, Coenen MJ, Barrera P, van Riel PL, Russel FG. HMOX1 promoter polymorphism modulates the relationship between disease activity and joint damage in rheumatoid arthritis. Arthritis Rheum. 2008;58(11):3388-93. 2. Cvejic A, Haer-Wigman L, Stephens JC, Kostadima M, Smethurst PA, Frontini M, van den Akker E, Bertone P, Bielczyk-Maczynska E, Farrow S, et al. SMIM1 underlies the Vel blood group and influences red blood cell traits. Nat.Genet. 2013;45(5):542-5. 3. Haer-Wigman L, Veldhuisen B, Jonkers R, Loden M, Madgett TE, Avent ND, de Haas M, van der Schoot CE. RHD and RHCE variant and zygosity genotyping via multiplex ligation-dependent probe amplification. Transfusion 2013;53(7):1559-74. 4. Lejon Crottet S, Haer-Wigman L, Gowland P, Fontana S, Niederhauser C, Hustinx H. Serologic and molecular investigations of DAR1 (weak D Type 4.2), DAR1.2, DAR1.3, DAR2 (DARE), and DARA. Transfusion 2013;doi: 10.1111/trf.12363. [Epub ahead of print]. 5. Haer-Wigman L, Ji Y, Loden M, de Haas M, van der Schoot CE, Veldhuisen B. Comprehensive genotyping for 18 blood group systems using a multiplex ligation-dependent probe amplification assay shows a high degree of accuracy. Transfusion 2013;doi: 10.1111/trf.12410. [Epub ahead of print].

197

PhD portfolio

PHD Portfolio Name PhD Student: L. Haer-Wigman PhD Period: April 2009 - July 2013 Promotor: Prof. dr. C.E. van der Schoot Copromotor: Dr. M. de Haas

1. PhD Training Year ECTS Courses - Immuunhematologie 2010 2.0 - DNA Technology 2010 2.1 - Mass Spectrometry, Proteomics and Protein Research 2010 2.1 - Advanced Immunology 2011 2.9 - Scientific Writing in English for Publication 2011 1.5 - Browsing Genomes with Ensembl 2012 0.3

Seminars - Department seminars 2009-2013 4.0 - Landsteiner Lectures and guest speaker seminars 2009-2013 2.0 - Journal Club 2009-2013 3.0 - Sanquin Science Day 2010 2010 0.5 - Sanquin Science Day 2012 2012 0.5

(Inter)national conferences - Nederlandse Vereniging van Bloedtransfusie (NVB) 2009 0.5 symposium 2009 tranfusiegeneeskunde - NVB symposium tranfusiegeneeskunde 2010 2010 1.0 Oral Presentation: Ontwikkeling van een MLPA assay om varianten in het RHD en RHCE gen te bepalen - FBI [First Amsterdam Board of Immunology Students] PhD retraite 2010 1.0 Oral presentation: The genetic basis of the red blood cell antigen Vel - 31st International Congress of the ISBT [International 2010 2.0 Society of Blood Transfusion] Poster presentation: Development of a Multiplex Ligation-dependent Probe Amplification [MLPA] assay to determine variants in the RHD and RHCE gene - XIV° Congrès de la Societe Francaise de Transfusion Sanguine 2010 0.8 - NVB symposium tranfusiegeneeskunde 2011, The Netherlands 2011 1.0

199 Appendix

Oral presentation: Familiaire uremie wordt veroorzaakt door verhoogde UT-B1 (Kidd) expressie - Triple I PhD retraite, The Netherlands 2012 1.0 Oral presentation: RhD antigen presentation in HLA class II of mature DCs - BBTS [British Blood Transfusion Society] Annual Conference 2012 0.8 2012 - 23rd Regional Congress of the ISBT 2013 3.3 Oral presentation: Comprehensive genotyping for 18 blood group systems using Multiplex Ligation-dependent Probe Amplification [MLPA] Oral presentation: Duplication of the urea transporter B gene (Kidd blood group) in a kindred with familial azotemia Oral presentation: The immune response to the Vel antigen is HLA-class II DRB1*11 restricted Poster presentation: Molecular analysis of immunized Lan- or Jra-negative patients and validation of genotyping assays to screen blood donors for Lan- or Jra-negativity Poster presentation: Identification of 10 novel RHD variants resulting in partial, weak or absent RhD expression

2. Teaching Year ECTS Supervising - Bachelor thesis supervision 2010-2013 7.0 R. Jonkers R. Droste R. Bijman - Master thesis supervision 2013 2.0 S. Solati

3. Parameters of Esteem Year Awards and Prizes - ISBT & DGTI Poster Award 2010 2010 - Nominee Sanquin In-House Seminar Award 2013

Total 41.3

200