Six years’ experience performing RHD genotyping to confirm D− red blood cell units in Germany for preventing anti-D immunizations

Abstract

BACKGROUND:

Red blood cell (RBC) units of D+ donors are falsely labeled D− if regular serologic typing fails to detect low D antigen expression or chimerism. The limitations of serology can be overcome by molecular typing.

STUDY DESIGN AND METHODS:

In January 2002, we introduced a polymerase chain reaction (PCR)-based assay for RHD as a routine test for first-time donors who typed D− by serologic methods including the indirect antiglobulin test. Samples were tested in pools of 20 for the RHD-specific polymorphism in Intron 4. RHD alleles were identified by PCR and nucleotide sequencing.

RESULTS:

Within 6 years, 46,133 serologically D− first-time donors were screened for the RHD gene. The prevalence of RHD gene carriers detected by this method was 0.21 percent. Twenty-three RHD alleles were found of which 15 were new. Approximately one-half of the RHD gene carriers harbored alleles expressing a DEL phenotype resulting in a prevalence of 0.1 percent.

CONCLUSION:

The integration of RHD genotyping into the routine screening program was practical. We report 6 years’ experience of this donor testing policy, which is not performed in most transfusion services worldwide. RBC units of donors with DEL phenotype have been reported to anti-D immunize D− recipients. We transferred those donors to the D+ donor pool with the rationale of preventing anti-D immunizations, especially dreaded in pregnancies. For each population, it will be necessary to adapt the RHD genotyping strategy to the spectrum of prevalent alleles.

The D antigen is by far the most immunogenic, diverse, and clinically important protein-based blood group. Anti-D may induce hemolytic transfusion reactions and remains the major cause of hemolytic disease of the fetus and newborn.^1,2 Anti-D prophylaxes become ineffective, once an anti-D immunization has occurred, even if it is subclinical. Therefore, D antigen–compatible transfusion is the established standard therapy. Approximately 0.4 percent of the Central European D+ population carries RHD alleles associated with reduced D antigen expression.³ Evidence has accumulated that some red blood cell (RBC) units with weak D expression may escape detection by standard serologic methods including the indirect antiglobulin test (IAT) and may cause anti-D immunizations when transfused to recipients.^4–6 For example, RBC units with weak D Types 1, 2, and 26 that were incorrectly labeled D− have induced anti-D formation after transfusion.^7–9 The more sensitive technique of anti-D adsorption/elution is not suitable for routine diagnostic use. D variants detectable by adsorption/elution only are known as DEL¹⁰ and have also caused the generation of anti-D.^11,12 In apparently D− individuals small numbers of D+ RBCs may remain undetected due to the limitations of serology; transfusions of such chimeric blood gave rise to multiple anti-D immunizations.¹³

Molecular typing overcomes the limitations of serologic methods.^14,15 Therefore, we screened routinely all serologically D− first-time donors for the presence of the RHD gene since January 2002.^16–18 The D− phenotype is usually caused by the RHD gene deletion. The DEL^10,13 and some rare D−¹⁹ phenotypes are caused by RHD gene mutations. Our results demonstrate the prevalence of RHD gene carriers and their alleles in the D− population, the practicability of the method, and a potential relevance in preventing anti-D immunizations.

MATERIALS AND METHODS

D antigen typing method

D antigen typing of donors was performed with two different immunoglobulin M monoclonal antibodies that do not bind to RBC with a DVI phenotype using an auto-analyzer (PK7300; Olympus, Hamburg, Germany).⁷ Non-reactive samples were tested with oligoclonal anti-D (Seraclone anti-D blend, Clones BS221, H4111B7, and BS232; Biotest, Dreieich, Germany) using the IAT in a gel matrix technique (ID microtyping system, DiaMed, Cressier sur Morat, Switzerland), which detects DVI. This method conformed to the German guidelines for donor testing.

Determination of DEL by anti-D adsorption/elution

For anti-D adsorption to RBCs, 200 μL of RBCs was incubated for 1 hour at 37°C with 200 μL of polyclonal anti-D (anti-D incomplete, Lot 1121.54.20, DiaMed). Elution was effected by a chloroform⁷ or an acid elution technique (DiaCidel, DiaMed). The DEL phenotype was established once for novel RHD alleles. If the RHD allele is known, the decision to label RBC units D+ or D− was based on the molecular data.

Molecular methods

Samples of serologically D− first-time donors were subjected to RHD PCR in pools of 20. Equal volumes of 250 μL of ethylenediaminetetraacetate-anticoagulated blood were pooled, and DNA was extracted. The pools were tested for the RHD-specific polymorphism in Intron 4 by PCR with sequence-specific primers rb12 and re41.¹³ The readout was manual. RHD Intron 4–positive pools were resolved by individual PCR testing of samples with serologically defined C or E and of the residual samples in a pool. Alternatively, all samples of the pool were subjected to single PCR. Samples identified as RHD gene–positive were further characterized by PCR and nucleotide sequencing to determine the involved RHD alleles.^13,20

Haplotype frequencies

The frequency of an RHD allele in the population and in its haplotype was calculated. The numbers of the corresponding haplotypes under observation were 89,122 cde, 2067 Cde, 1076 cdE, and 1 CdE, including RHD gene–positive haplotypes that are D− in standard serology. We assumed that all 27 samples of CcddEe phenotype represented a Cde/cdE genotype. Ninety-five percent confidence intervals (CIs) were calculated according to the Poisson distribution. Donors were not generally checked for kinship; however, two pairs of donors with RHD-CE(8–9)-D, three donors with RHCE(1–3)-D(4–10), and two with RHD(147delA) were related.

Nomenclature

The designation DBU indicates a “DBT-1–like allele from Ulm” because DBU (EMBL Accession Number AM945964) is similar to the partial D DBT-1.

RESULTS

RHD gene carriers

Between 2002 and 2007, almost 3 million whole blood donations were collected in Baden-Württemberg (southwestern Germany) at our DRK Blood Donor Service, of which 621,685 were D−. A total of 46,133 D− first-time donors were tested for the presence of the RHD gene by Intron 4 PCR. RHD genotyping revealed 96 D− donors who carried the RHD gene (Table 1). Almost half of these harbored RHD alleles expressing DEL phenotypes.

TABLE 1.

Genotyping for RHD Intron 4 in 46,133 serologically D− first-time donors

	Serologically D− donors*
Rhesus phenotype	RHD Intron 4–negative	RHD Intron 4–positive
		DEL†	D−†
ccddee	43,038	1	14
Ccddee	1,916	39	28
ccddEe	1,023	3	7
CCddee	24	4	0
CcddEe	27	0	0
ccddEE	8	0	0
CCddEe	1	0	0
Total	46,037	47	49

^*Serologically D− by routine donor screening methods not involving adsorption/elution. RHD gene–negative or –positive by PCR with sequence-specific primers for RHD Intron 4.

^†DEL or D− according to anti-D adsorption/elution.

Summary of RHD alleles encountered

We found 11 RHD alleles encoding a DEL phenotype and 12 that did not encode demonstrable D antigen (Table 2). The estimated population frequency of the most frequent DEL phenotype was 1:2883 in D− individuals (1:1816 to 1:4807; 95% CI, Poisson distribution) and 1:124 in individuals with a Ccddee phenotype (1:78 to 1:207). The 4 most frequent alleles made up almost two-thirds of all RHD gene–positive samples. The remaining 19 alleles occurred sporadically. Among these, RHD(K409K) was the most frequent allele in unrelated donors; the estimated population frequency was 1:11,533 in D− and 1:496 in Ccddee individuals. Fifteen alleles were new. Our RHD(V56M, W90X) allele differed from the known RHD(W90X)²¹ by an additional amino acid substitution in Exon 1 and by haplotype.

TABLE 2.

Detection of 23 RHD alleles in 96 serologically D− RHD gene carriers

Allele	Phenotype	Haplotype	Nucleotide change	Number of donors observed	Reference
RHD-CE(8–9)-D*	D−	CDe	RHD-CE-D hybrid	17	Wagner et al.13
RHD(IVS3+1G>A)	DEL	CDe	IVS3+1G>A	16	Wagner et al.13
RHDΨ	D−	cDe	Multiple	14	Singleton et al.19
weak D type 11	DEL	CDe	885G>T	14	Wagner et al.13
RHCE(1–3)-D(4–10)*	D−	cDE	RHCE-D hybrid	4	This study
RHD(147delA)†	DEL	CDe	Deletion of A at 147	4	AM998539
RHD(K409K)	DEL	CDe	1227G>A	4	Wagner et al.13
RHD(IVS3+2T>A)	D−	cDE	IVS3+2T>A	3	AM998540
RHD(W16X)	D−	CDe	48G>A	2	Wagner et al.13
RHD(93_94insT)	DEL	CDe	Insertion of T at 93_94	2	AM998541, Nogues et al.30
RHD(343delC)	D−	CDe	Deletion of C at 343	2	AM998542
RHD(R318X)	D−	CDe	952C>T	2	AM998543
RHD(X418L)	DEL	CDe	Insertion of T at 1252_53	2	Gassner et al.9
RHD(V56M, W90X)	D−	CDe	166G>A, 270G>A	1	AM998544
RHD(L153P)	DEL	cDE	458T>C	1	AM998545
DBU*	DEL	cDE	RHD-CE(5–7)-D hybrid	1	AM945964
RHD(G212R)	DEL	cDe	634G>C	1	AM998546
RHD(660delG)	D−	CDe	Deletion of G at 660	1	AM998547
RHD(712delG)	D−	CDe	Deletion of G at 712	1	AM998548
RHD(786delA)	D−	CDe	Deletion of A at 786	1	AM998549
RHD(Y269X)	D−	CDe	807T>A	1	AM998550
RHD(Y401X)‡	DEL	cDE	1203T>A	1	Gassner et al.9
RHD §	DEL	CDe	Normal RHD	1	This study

^*These possible RHD-CE-D hybrid alleles could also be caused by partial deletions and, in case of RHCE(1–3)-D(4–10) and DBU, by combinations of hybrid alleles.

^†Intron deletion IVS1+6delA present in all four samples.

^‡While this mutant was previously described as D−,⁹ our sample was DEL.

^§Corresponding to normal RHD with respect to the nucleotide sequence of the 10 exons and flanking regions, including 115 nucleotides in 5′UTR, 38 in 3′UTR, and at least 10 nucleotides of the intron sequences flanking the exons.

Incidence of donors and alleles

The number of RHD gene carriers detected among serologically D− first-time donors remained constant each year during the study period (Fig. 1). An exception was 2002 (not shown), because the study was initially restricted to a part of the donor recruitment area. As expected the number of donors with newly detected alleles declined over the years; in 2007 there was no novel allele observed. The number of different alleles in the donor pool was similar each year (Fig. 1).

Fig. 1.

RHD+ donors and RHD alleles during the years 2003 until 2007. The serologically D− RHD gene–positive donors are shown per year (□) along with the different RHD alleles encountered (▧) and the newly discovered RHD alleles (■). The total number of D− donors screened in 2003 was 6313 (2004 = 9690; 2005 = 10202; 2006 = 8287; and 2007 = 8023).

Partial D DBU

DBU was identical to the known DBT-1 with few exceptions: It carried the 676G>C nucleotide substitution in RHD Exon 5 encoding the amino acid substitution A226P, which is typical for the E antigen, was associated with a cDE haplotype, and the antigen expression was greatly reduced. In congruence with its ccDEe phenotype, the heterozygous E/e typical nucleotide pattern was present in RHCE Exon 5.

DISCUSSION

We routinely screened serologically D− donors for the presence of the RHD gene. This method was established at our blood service because current serologic practice fails to discriminate the DEL from the true D− phenotype. To avoid immunization of D− recipients, we transfer RBC units carrying a DEL phenotype to the D+ pool.

Our PCR screening strategy detected the RHD-specific polymorphism in Intron 4. This method was deliberately devised to miss distinct RHD-CE-D hybrid alleles in which large segments derive from RHCE, like in RHD-CE(2–9)-D₂. These hybrid RHD alleles, where at least Exons 4 to 7 are replaced by RHCE, are not uncommon among D− donors;¹³ nevertheless, such RBCs do not pose any risk to D− recipients due to the known lack of D antigen expression. Clinically relevant D+ hybrids, like DVI, were invariably detected by the serologic screening including the IAT, which is presently mandatory according to the German guidelines.

The study revealed a large number of different alleles among serologically D− RHD gene carriers. The 23 RHD alleles comprised 12 true D− and 11 DEL (Table 2). Most alleles conformed to the expected phenotype (Table 3). However, the DEL phenotypes of RHD(147delA) and RHD(93_94insT) were not easily explained by the usual translation process, because frameshift mutations are expected to result in nonfunctional proteins (Table 3). The underlying molecular mechanism remained obscure, but RNA transcription or translation errors, known as transcription slippage and ribosomal frameshift, may be implicated in both cases. For instance, in RHD(147delA) a slippage in RHD Exon 1 at the AAAAGGGG sequence 16 nucleotides 5′ prime of the deletion at Position 147 could restore the reading frame, permitting the production of trace amounts of functional protein. Partial correction of similar frameshift mutations by compensatory ribosomal frameshift during expression has been reported for several genes, like apoB,²² CA-II,²³ and FVIII.^24–26

TABLE 3.

RHD alleles observed and D antigen phenotype

Molecular variation	Phenotype
	D−	DEL	Possible cause for phenotype
Amino acid substitution		Weak D Type 11	Transmembraneous mutation with C in trans
		RHD(L153P)	Transmembraneous mutation
		RHD(X418L)	Prolongation of amino acid chain
Stop codon	RHD(W16X)		Truncated nonfunctional protein
	RHD(V56M, W90X)		Truncated nonfunctional protein
	RHD(Y269X)		Truncated nonfunctional protein
	RHD(R318X)		Truncated nonfunctional protein
Splice site mutation		RHD(Y401X)	Amino acid chain shortened by 17 amino acids
	RHD(IVS3+2T>A)		Missplicing
		RHD(IVS3+1G>A)	Missplicing
		RHD(K409K)	Missplicing
		RHD(G212R)	Missplicing/transmembraneous amino acid substitution
Frameshift mutation	RHD(343delC)		Premature stop with truncated nonfunctional protein
	RHD(660delG)		Premature stop with truncated nonfunctional protein
	RHD(712delG)		Premature stop with truncated nonfunctional protein
	RHD(786delA)		Premature stop with truncated nonfunctional protein
		RHD(93_94insT)	Replication slippage may permit traces of functional protein
		RHD(147delA)	Replication slippage may permit traces of functional protein
Pseudogene	RHDΨ		No protein19
RHD fragments	RHD-CE(8–9)-D		Possible incomplete RHD gene or hybrid gene
	RHCE(1–3)-D(4–10)		Possible incomplete RHD gene or hybrid gene
		DBU	Similar to partial D DBT-1

The presently described RHD(G212R) is the fourth allele with a mutation at Codon 212, which spans the Exon 4/Exon 5 boundary in the cDNA. Previously, RHD(G212V) representing D−,¹³ RHD(G212C) representing weak D type 23,²⁷ and RHD(G212G) known as DUC-1²⁸ have been described. All of these four alleles expressed different phenotypes (Fig. 2), which may be explained by differing effects on the splicing process or the protein or both. The two possible hybrid alleles RHD-CE(8–9)-D and RHCE(1–3)-D(4–10) did not encode D antigen expression, although they would appear to encode all D-specific exofacial loops. Analysis of cDNA may resolve this apparent contradiction.

Fig. 2.

RHD alleles with nucleotide substitutions at Codon 212. The Exon 4/Exon 5 boundary (arrow) in the cDNA is shown with the nucleotide and amino acid substitutions in correlation to phenotype.

The distribution of alleles in our study was in accordance with the distributions in three limited surveys from Central Europe^9,29 and Spain,³⁰ particularly with respect to the prevalent nonhybrid alleles RHD(IVS3+1G>A), weak D Type 11, and RHD(K409K), considering that we disregarded clinically irrelevant large hybrid alleles.

The integration of RHD genotyping into the routine donor screening program was feasible and required less than 400 PCR procedures per year for the pool testing. We estimated the cost of the RHD genotyping program at our blood service to less than 40,000 US$ per year, i.e., approximately 5 US$ per D− first-time donor or 8 US cents per whole blood donation. The cost-effectiveness may be increased by genotyping D− donors with C or E antigens only, which may be considered an acceptable strategy in European donor cohorts. An inexpensive and less technically demanding approach would be to refrain from resolving the underlying RHD alleles in the RHD gene–positive donors and to transfer all D− RHD gene–positive donors to the D+ donor pool in European or to a new, separate DEL donor pool in East Asian³¹ donor cohorts.

Our technical approach was adjusted to the local variety of alleles and to our serologic screening strategy. The prevalence of RHD gene carriers in the D− population is approximately 0.6 percent in Caucasians,¹³ 10 percent in Africans,¹⁹ and 30 percent in Asians.³¹ These differences are due to predominantly three RHD alleles: the D antigen negative RHDΨ and Ccde^s in Africans¹⁹ and the DEL variant RHD(K409K) in Asians.²¹ Hence, for each population it will be necessary to adapt the RHD genotyping strategy to the spectrum of prevalent alleles. For example, in populations with a relevant frequency of RHDΨ, an approach using pooled testing will require an RHD-specific PCR that fails to amplify the RHDΨ allele.³² Such a screening program would allow efficient pool testing despite the presence of RHDΨ, while single-donor testing remains always an option. Screening strategies have also been proposed for the East Asian population.^31,33 Once the alleles that are frequent in non-European populations have been accounted for, the remainder may represent sporadic RHD alleles with low frequencies. The frequency of diverse, sporadic nonfunctional alleles has been reported to be 1 in 300 or less.^13,34

The sensitivity of current serologic D antigen testing needs tight quality assurance. However, any weak D+ or DEL+ donor missed by serology can be detected by genotyping.³⁵ Hence, genotyping obviates the need for a very sensitive serology. This less stringent demand on serology may save costs. Defining the lower limit of the clinically required serologic sensitivity in donor testing that has been a matter of debate for decades becomes obsolete. The decision to label RBC units D+ or D− could be based on the clinical relevance of distinct alleles.

We propose avoiding transfusion of DEL RBC units to D− recipients, especially to D− girls and women of child-bearing age. Based on a prevalence of the DEL phenotype of 1 in 982 among serologically D− donors (Table 1), we estimated that more than 100 potentially immunogenic RBC units would be dispensed to D− recipients each year in southwestern Germany. Therefore, we permanently transferred donors with alleles encoding a DEL phenotype to the D+ donor pool. The DEL phenotype was almost exclusively associated with CDe or cDE haplotypes; only one DEL+ was found among ccddee donors (Table 2). Generally, D− donors with C or E antigens could be transferred to the D+ donor pool; this practice has been abandoned in some health care systems because of the shortage of D− supplies, which is well founded in Caucasian donors because 98 percent of these donors are truly D− (Table 1). During the 6-year study period we established sufficient knowledge on the prevalence of alleles in our population (Fig. 1). Therefore, it could now be a practical approach, at least for many Central European populations, to restrict genotyping to D− donors with C or E antigens, although few chimeric^13,36 and DEL RBC units will remain undetected. However, since the presented donor testing policy, which was established in 2002 as a routine procedure, is continuing, we anticipate including eventually all repeat donors at our blood service.

RHD genotyping of repeat donors followed by look-back of the transfused RBC units can be instrumental to define the immunization potential of weak D and DEL phenotypes. We also recommend a trace back in any D− RBC recipient with an unexplained anti-D; the presence of weak D or DEL in the involved donors may be recognized.^7,9

The results of the 6 years’ experience reported here validate our rationale to introduce RHD genotyping in blood donors for the quality control of D− RBC units. Blood group genotyping for RHD in donors has the potential to become a routine procedure in blood centers.