Received 19 August 1998; accepted for publication 27 November 1998

Transcription

1 British Journal of Haematology, 1999, 104, Detection of weak D and D VI red cells in D-negative mixtures by flow cytometry: implications for feto-maternal haemorrhage quantification and D typing policies for newborns P. L LOYD-EVA N S, 1 A. R. GUEST, 1 D. VOAK 2 AND M. L. SCOTT 1 1 International Blood Group Reference Laboratory, Bristol, and 2 East Anglian Blood Centre, Cambridge Received 19 August 1998; accepted for publication 27 November 1998 Summary. Quantitation of feto-maternal haemorrhage (FMH) by flow cytometry (FC) has been shown to be more accurate than the Kleihauer-Bekte test. Fetal cells will be predominately of R 1 r or R 2 r phenotype, with antigen site numbers per cell (SPC) of between 9900 and If the fetus is of weak D or partial D VI phenotype, fewer SPC will be present. Red cells from 20 adult weak D samples were mixed with rr red cells to give 1% mixes. Mixtures were stained and analysed by FC, using two different monoclonal reagents. The SPC of each sample was measured using SOL-ELISA with Scatchard plot analysis. 18 samples could not be distinguished and had <1000 SPC. Two samples that could be distinguished had 1350 and 3000 SPC. Red cells from seven samples of D VI were also analysed. None of these samples could be distinguished; SPC were all <1000. Although one of the reagents used reacts with D VI cells, quantitation of a D VI FMH would not be possible due to low SPC. The ability of fetal red cells with low Rh D SPC to cause immunization is questionable; failure to measure FMH in these cases is unlikely to cause clinical problems, as long as suitably sensitive serological reagents and techniques are used to type all weak D and D variant babies as Rh D positive, and thus ensure that the mother is given the appropriate dose of anti-d. Keywords: feto-maternal haemorrhage, FITC-conjugated anti-d, partial D, weak D, D typing. The quantification of feto-maternal haemorrhage (FMH) by flow cytometry (FC) has been shown to be more accurate for larger bleeds than the Kleihauer-Betke test (Lloyd-Evans et al, 1996; Austin et al, 1997; Lubenko et al, 1997). The quantification of the volume of fetal Rh D þ cells in the circulation of D ¹ women is essential to determine the dosage of prophylactic anti-d required. In the U.K. 500 iu (100 mg) of Rh D immune globulin is routinely given to all Rh D negative mothers of Rh D positive (or equivocal/unknown D type) babies post-partum, and is sufficient to remove up to 4 ml of packed fetal D þ red blood cells (RBC) from the maternal circulation. For bleeds >4 ml an additional dosage of anti-d is required, calculated at 125 iu/ml of bleed. If inadequate doses are given, fetal cells may remain in the circulation and an alloimmune response could be mounted by the mother, which may have serious consequences for Correspondence: Dr Paul Lloyd-Evans, International Blood Group Reference Laboratory, Southmead Road, Bristol BS10 5ND. paul.lloyd-evans@nbs.nhs.uk Blackwell Science Ltd future D-positive pregnancies and result in Rh D haemolytic disease of the newborn. Falsely high estimates of bleeds lead to unnecessary over-dosing with anti-d; although this is not of any danger to the mother, it is a waste of a blood product that is often in short supply, and may compromise the availability of the product for true large bleeds, or for routine ante-natal prophylaxis. The first full exercise of the UKNEQAS for feto-maternal haemorrhage indicated that, for bleeds of <4 ml, 40% of laboratories are recommending over-dosing and 4% under-dosing with anti-d (personal communication). In most cases the fetal cells will be of R 1 rorr 2 r phenotype, with antigen site numbers per cell (SPC) of between 9900 and (Rochna & Hughes-Jones, 1965). If the fetus has a weak D or partial D VI phenotype much fewer sites may be present due to decreased Rh D antigen expression (Gorick et al, 1993; Jones et al, 1996). It has been estimated that the SPC in weak D vary between 100 and 1870 sites per cell, and in D VI between 300 and 2900, although a recently characterized D VI category (type III; Wagner et al, 1998) 621

2 622 P. Lloyd-Evans et al has been shown to be able to have near-normal Rh D antigen expression. Whereas the performance of the Kleihauer test is not dependent on the phenotype and Rh D sites of the fetal cells in the FMH sample (since fetal cells are identified by acid elution of maternal cells), this is not the case for flow cytometry when using anti-d antibodies. The sensitization levels with antibody, mean channel fluorescence (MCF) response and consequent discrimination from maternal D ¹ cells by flow cytometry is directly related to the phenotype and Rh D sites per cell of the fetal cells (Jones et al, 1996). Recently, the use of anti-fetal haemoglobin antibodies for the detection of fetal cells by flow cytometry has been described (Davis et al, 1998). This technique is not reliant on Rh D antigen levels, and could be useful for the detection of weak D or partial D fetal bleeds. However, this method is more technically demanding, further complicated by the possible labelling of false positives, and crucially would not identify the Rh D status of the fetal bleed for treatment with additional anti-d. In this study we have investigated the theoretical ability of flow cytometry to detect and measure FMH of weak D and partial D VI phenotypes. A comparison of two fluorescently labelled monoclonal IgG anti-d conjugates was made; FITC- BRAD-3 (IBGRL Research Products, Bristol) and a commercially available reagent consisting of a different monoclonal (BTSN10) directly labelled with R-phycoerythrin: Quanti-D (Quest Biomedical, Knowle). Both antibodies react with most D-positive cells, except for BRAD-3, which fails to recognize the D VI phenotype (Scott, 1996). MATERIALS AND METHODS All the chemicals and reagents used in this study were obtained from Sigma, Poole, unless otherwise stated. Serological typing of weak D and D VI cells. Weak D individuals were identified as those which failed to react in a saline tube test using MAD-2 monoclonal IgM anti-d (IBGRL, Bristol; 30 mg/ml), but reacted in the same test using RUM-1 monoclonal IgM anti-d (IBGRL, Bristol; 10 mg/ml). One volume of 3% RBC suspended in PBS and 2 volumes of antibody were mixed in a mm glass tube and incubated for 5 min at room temperature. Tubes were centrifuged at 110 g for 1 min, the red cell button carefully resuspended, and the degree of agglutination assessed macroscopically and graded on a scale from 6 for complete agglutination to 0. The serological strength of representative weak D samples was compared by observing reactions with a twofold doubling dilution series of RUM-1, prepared in PBS containg 2% BSA. Tests were carried out using the saline tube technique above, and also using a microplate technique. For the microplate technique, tests were set up in polystyrene U-well microplates (Sterilin, U.K.) and were incubated for 15 min at room temperature prior to centrifugation and macroscopic reading by resuspension after agitation on an automated plate shaker. Tube and microplate techniques were also carried out using papain-treated red cells (treated according to the ISCH/ISBT reference technique; Scott et al, 1994). RUM-1-negative samples were screened using BRAD-2 IgG monoclonal anti-d (IBGRL, Bristol; 10 mg/ml). This antibody reacts with D VI cells. Equal volumes of 3% RBC suspended in PBS and antibody were mixed in a mm glass tube and incubated for 15 min at 37 C. Red cells were washed three times with PBS, and then 2 volumes of antiglobulin reagent (NBS, M&SW Zone, U.K.) were added. After mixing, tubes were immediately centrifuged at 110 g for 1 min, the red cell button carefully resuspended, and the degree of agglutination observed macroscopically. Staining of red cells for flow cytometry. FITC-BRAD-3 labelling: FITC-BRAD-3 conjugate was prepared as previously described (Lloyd-Evans et al, 1996). 5% RBC mixes of samples and controls (rr, R 1 r, R 2 r RBC) were made in PBS and 10 ml were added to 90 ml ofa50mg/ml concentration of FITC-BRAD-3 in PBS, incubated for 60 min at 37 C, then washed twice in PBS to remove the unbound conjugate. The sample was finally resuspended in 1 ml PBS containing 0 5% BSA (PBS-A) for analysis. Quanti-D labelling: RBC were prepared as before and the cells were labelled according to the manufacturer s instructions (Quest Biomedical Knowle) at room temperature (RT) for 15 min. Cells were washed twice in PBS and resuspended as above for analysis. Flow cytometric analysis. Samples were analysed on a FACS Calibur (Becton Dickinson, Oxford) flow cytometer. The samples and controls were gated and isolated according to forward and 90 side scatter and a total of events were analysed per sample along with the appropriate controls. For FITC-BRAD-3, histograms were generated using the parameters of log-integrated green fluorescence (FL1) versus the number of events and the MCF noted, whereas for Quanti-D histograms were generated using the parameters of log-integrated red fluorescence (FL2) versus the number of events. For weak D and D VI RBC mixtures in rr cells a marker was put around the D þ (high FL1/FL2) events and the percentage of these of the total events obtained. Quantitation of molecules of FITC-BRAD-3 bound per cell on weak D cells. The method used was adapted from that described by Jones et al (1996). Briefly, a standard curve for FITC-BRAD-3 was prepared by adding 100 ml of a 5% R 2 R 2 RBC suspension (pool of three donors, and a cell count made) in PBS to 900 ml of a serial 2-fold dilution of conjugate (starting concentration of 50 mg/ml), and incubating for 1 h at 37 C. The sensitized RBC were washed three times in PBS- A (PBS ph 7 4 containing 1% albumin), resuspended to 1 ml in PBS-A and 100 ml aliquot taken for FC analysis to obtain the MCF. The amount of RBC-bound FITC-BRAD-3 in the remaining 900 ml was determined by ELISA (Kumpel, 1990). A standard curve was constructed by plotting the number of RBC-bound IgG molecules per cell against the corresponding MCF. The test samples were prepared as described earlier, the MCF noted and the molecules of FITC-BRAD-3 were extrapolated from the standard curve. Since sensitization of each of the test weak D samples was done under antibody excess (saturating levels of bound FITC-BRAD-3), the number of molecules of conjugate bound to the test RBC approximated closely the available Rh D SPC (Jones et al, 1996).

3 Quantitation of Rh D sites per cell. Absolute Rh D SPC were quantitated by using SOL-ELISA with Scatchard plot analysis. A 100 ml of 5% RBC suspension of test sample in PBS was incubated for 1 h at 37 C with 900 ml of a serial 2-fold dilution of FITC-BRAD-3 (starting concentration of 10 mg/ml). The sensitized RBC were washed three times in PBS-A, pelleted and lysed for 15 min with a 100 ml of1% Triton X-100 and resuspended back to a total of 1 ml in PBS- A. The cell-bound IgG was determined by an ELISA as before. Scatchard plots for each test sample were constructed and the total Rh D SPC determined by extrapolation (Rochna & Hughes-Jones, 1965). Analysis of partial D VI phenotype by flow cytometry. Since BRAD-3 does not recognize the D VI phenotype, all staining was performed by the Quanti-D reagent. Three samples of each of category type I and II, and one of category III D VI were prepared for FC as before, and 1% mixtures of each D VI in rr cells were also analysed. The methods for FC analysis were identical to those described earlier. Detection of Weak D and D VI Cells in Artificial FMH Mixtures 623 RESULTS Analysis of weak D samples by FITC-BRAD-3 and Quanti-D Both conjugates were tested against a panel of 20 weak D RBC cells and in both cases 18 samples had very low MCF values and were indistinguishable as a 1% population in D- negative rr cells (equivalent to an 18 ml FMH bleed). For FITC-BRAD-3 sample values ranged from 3 4 to 11 MCF, as compared to rr control of 2 8, whereas for Quanti-D the values ranged from 4 to 20 MCF, as compared to the rr control of 2 1. Two samples had higher MCF values which could be discriminated as a 1% population in rr cells by both conjugates. The first sample gave a MCF value of with FITC-BRAD-3 and 46 8 with Quanti-D (Fig 1), whereas the second sample was greater with a MCF of Fig 2. A 1% weak D (sample 2) in rr cells as analysed by FITC- BRAD-3 (A) and Quanti-D (B) with FITC-BRAD-3 and 74 3 with Quanti-D (Fig 2). The Quanti-D reagent gave a slightly higher MCF response than FITC-BRAD-3 in both cases. Quantitation of weak D Rh D site numbers The molecules of FITC-BRAD-3 bound per cell on each of the 20 weak samples was calculated by extrapolation from a standard curve of MCF response against sensitization levels of FITC-BRAD-3 on R 2 R 2 cells (Fig 3). The 18 samples which gave low MCF showed sensitization levels of less than 1000 molecules of FITC-BRAD-3 per cell. In the two samples which gave a MCF value of and 46 8 the sensitization levels of FITC-BRAD-3 per cell were 1200 and 2600 molecules and absolute Rh D SPC was determined by Scatchard plot analysis and quantified as 1350 and 3000 SPC respectively. Fig 1. A 1% weak D (sample 1) in rr cells as analysed by FITC- BRAD-3 (A) and Quanti-D (B). Fig 3. Standard curve of mean channel fluorescence against sensitization level of FITC-BRAD-3.

4 624 P. Lloyd-Evans et al Serology Seven representative samples (the two analysed above and five of those with <1000 molecules FITC-BRAD3 per cell) were tested serologically against a titration series of RUM-1, in parallel with an R 1 r sample (Table IA). The results showed that the serology titres corresponded well with the MCF response by flow cytometry. Reactivity of one of the weaker samples was compared with a titration series of RUM-1 in tube and microplate techniques, using untreated and papain treated red cells. The results (Table IB) show that with untreated red cells, the tube technique was more sensitive than the microplate technique. Use of papain-treated red cells improved sensitivity in both tube and microplate methods. If the methods are ranked in order of sensitivity, the order is enzyme tube, enzyme microplate, saline tube, saline microplate. D VI cells were readily detected using Table I. (A) Comparison of RUM-1 titre end-point and scores in a saline tube technique versus flow cytometry data on seven weak D samples. MCF Cell FITC BRAD-3 D SPC Titre Score R1r Weak D Weak D Weak D 10 0 < Weak D 7 4 < Weak D* 5 3 < Weak D 3 1 < Weak D 3 5 < rr Nt Nt (B) The weak D marked with the asterisk above was also assessed with a titration series of RUM-1 in enzyme and microplate techniques; titre end-point and score is given. Technique Titre Score Saline tube 4 6 Enzyme tube Saline microplate N 3 Enzyme microplate (C) Detection of D VI by antiglobulin and enzyme techniques using a titration series of BRAD-2; titre end-point and score is given. Technique Titre Score Antiglobulin Enzyme BRAD-2 IgG anti-d in an antiglobulin or enzyme technique. There was little difference in sensitivity between these two techniques (Table IC). Analysis of D VI variant cells by Quanti-D All the samples analysed from each D VI category showed very low MCF responses (values between 2 and 3) as compared to the rr control value of 2 1. No discrimination of a1%d VI population in a rr mix was possible. DISCUSSION Our results showed that when FMH was assessed by flow cytometry using a fluorochrome-labelled anti-d, weak D or D VI cells with <1000 sites per cell were not detected. Such bleeds would be detected by the Kleihauer-Betke test and typed as Rh D positive by serology if the appropriate reagents and methods were employed. At present, policies on reagents and techniques for typing newborn samples vary between hospitals and no published guidelines are available for this purpose. As long as the newborn is typed as D positive, the mothers would automatically receive the standard 500 iu dose of anti-d. Controversy has existed for many years as to whether weak D and D VI phenotype donor blood is capable of causing immunization, and continues to be a matter for discussion (FDA/CBER, 1997). In the Netherlands one incident was identified where a unit of weak D donor blood (1470 Rh D SPC), mistakenly typed as D negative, caused primary immunization in a female transfusion recipient with no previous history of pregnancy or transfusion (Gorick et al, 1993). For D VI blood there has been no reported incidences of immunization after transfusion in D-negative recipients. The immunogenicity of Rh D variant red cells will depend upon a number of factors: D antigen site density, which epitopes are present, volume and number of immunizing events, and the responsiveness of the recipient. Therefore it is unclear what volume of feto-maternal haemorrhage of weak D cells would be required to cause immunization in the mother. There is only one report of Rh D immunization during pregnancy caused by a 4 ml bleed from a weak D Va fetus (Mayne et al, 1990), but unfortunately no antigen site quantitation was undertaken. It is not clear from this case whether the anti-d given failed to protect the mother from primary immunization at birth, or whether primary and secondary immunization occurred during the second pregnancy. We have shown that weak D samples with 1350 and 3000 sites per cell could be readily quantified by flow cytometry which would allow the administration of the appropriate dosage of anti-d. The protective effect of therapeutic anti-d is known to be dose-dependent and dosage has been standardized against normal Rh D positive blood (Pollack et al, 1971). Thus the routine administration of 500 iu dose of anti-d post-partum probably provides adequate protection against much larger Rh D þ bleeds bearing lower antigen site numbers than 4 ml of normal fetal cells. As long as suitable serological reagents and techniques are available to type all weak D babies as D positive, then the adequate

5 administration of therapeutic anti-d in these cases should be ensured. This means selection and use of high-avidity IgM typing reagents in suitably sensitive techniques, such as tube or gel tests, or enzyme microplate tests. In our study it was not possible to detect D VI cells by flow cytometry with either conjugate. To date there have been no reported incidences of D VI blood alloimmunization during pregnancy, and it is known that these red cells lack the most immunogenic part of the Rh D antigen (Jones et al, 1995). Failure to quantify large FMH when the fetus is D VI is unlikely to cause immunization. Policies vary between hospital laboratories as to whether babies are typed with reagents that would type D VI as D positive, and thus some D-negative mothers are given anti-d and some are not. Given the incidence of the Rh D negative phenotype (17%) and the D VI phenotype (0 02%) in the British population, the predicted incidence of a D-negative mother bearing a D VI baby is 1 in pregnancies. The number of pregnancies in which a >3 ml bleed occurs has been reported to be about 1% (Mollison et al, 1997); thus a large bleed from a D VI fetus into a D-negative mother, not detected by flow cytometry, will occur in a predicted 1 in pregnancies. In conclusion, the failure of flow cytometry to detect FMH where the fetus has <1000 antigen sites per cells is thus extremely unlikely to cause a failure of Rh D prophylaxis. Recent UK NEQAS exercises have shown that the flow cytometry method is more accurate for quantifying large FMH than the Kleihauer test. It is being proposed that hospitals should routinely screen by Kleihauer, and refer positive samples for quantification by flow cytometry. Laboratories should be aware that even if the baby types as D positive, it is possible that the Kleihauer test may be positive and the flow cytometry negative if the baby is of a weak D phenotype with <1000 sites per cell. As long as the baby is serologically typed as D positive using suitably sensitive reagents and techniques, the mother should have adequate protection against immunization by administration of the standard dose of anti-d. ACKNOWLEDGMENTS The authors thank both staff at the North London Blood Centre Colindale and the South Thames Blood Centre for collecting the panel of weak D samples which were critical for this study. This work was funded by the National Blood Authority, England. Detection of Weak D and D VI Cells in Artificial FMH Mixtures REFERENCES 625 Austin, E., Powell, H., McIntosh, Y., Hulston, M. & Shwe, K. (1997) Quantification of high feto-maternal bleeds using BRAD-3 FITC and flow cytometry. Transfusion Medicine, 7, (Suppl. 1), 28. Davis, B.H., Olsen, S., Bigelow, N.C. & Chen, J.C. (1998) Detection of fetal red cells in fetomaternal hemorrhage using a fetal hemoglobin monoclonal antibody by flow cytometry. Transfusion, 38, FDA/CBER (1997) Workshop on Current Topics in Immunohematological Testing, 10 December Transcript available from CBER, Bethesda, Maryland, U.S.A. Gorick, B., McDougall, D.C.J., Ouwehand, W.H., Overbeeke, M.A.M., Tippett, P., Hughes-Jones, N.C. & van Rhenen, D.J. (1993) Quantitation of D sites on selected weak D and partial D red cells. Vox Sanguinis, 65, Jones, J., Scott, M.L. & Voak, D. (1995) Monoclonal anti-d specificity and RhD structure: criteria for selection of monoclonal anti-d reagents for routine typing of patients and donors. Transfusion Medicine, 5, Jones, J.W., Lloyd-Evans, P. & Kumpel, B.M. (1996) Quantitation of Rh D antigen sites on weak D and D variant red cells by flow cytometry. Vox Sanguinis, 71, Kumpel, B.M. (1990) A simple non-isotopic method for the quantitation of red cell-bound immunoglobulin. Vox Sanguinis, 59, Lloyd-Evans, P., Kumpel, B.M., Bromelow, I., Austin, E. & Taylor, E. (1996) Use of a directly conjugated monoclonal anti-d (BRAD-3) for quantification of fetomaternal hemorrhage by flow cytometry. Transfusion, 36, Lubenko, A., Collier, R., Williams, M., Hindmarch, D., Wilson, S. & Pluck, J. (1997) Quantitating fetomaternal hemorrhages of D þ red cells using an FITC-conjugated IgG monoclonal anti-d by flow cytometry: a case report. Immunohaematology, 13, Mayne, K., Bowell, T., Woodward, T., Sibley, C., Lomas, C. & Tippett, P. (1990) Rh immunisation by the partial D antigen of category D Va. British Journal of Haematology, 76, Mollison, P.L., Engelfriet, C.P. & Contreras, M. (1997) Blood Transfusion in Clinical Medicine, 10th edn, p Blackwell Science, Oxford. Pollack, W., Ascari, W.Q. & Crispin, J.F. (1971) Studies on Rh D prophylaxis. I. Relationship between doses of anti-rh and size of antigenic stimulus. Transfusion, 11, Rochna, E. & Hughes-Jones, N.C. (1965) The use of 125 I labelled antiglobulin in the determination of the number of D antigen sites on red cells of different phenotypes. Vox Sanguinis, 10, Scott, M.L. (1996) Rh serology: co-ordinator s report. Transfusion Clinique et Biologique, 6, Scott, M.L., Voak, D., Phillips, P.K., Hoppe, P.A. & Kochman, S.A. (1994) Review of the problems involved in using enzymes in blood group serology: provision of freeze-dried ICSH/ISBT protease enzyme and anti-d reference standards. Vox Sanguinis, 67, Wagner, F.W., Gassner, C., Muller, T.H., Schonitzer, D., Schunter, F. & Flegel, W.A. (1998) Three molecular structures cause Rhesus D category VI phenotypes with distinct immunohematologic features. Blood, 91,