J Clin Pathol 1995;48:861-864 861 Comparison of haemoglobin H inclusion bodies with embryonic 4 globin in screening for ct J Clin Pathol: first published as 10.1136/jcp.48.9.861 on 1 September 1995. Downloaded from thalassaemia

L C Chan, J C C So, D H K Chui

Abstract cluster is intact. We have compared the HbH Aims-To compare the haemoglobin (Hb) inclusion test with immunocytochemical H inclusion test with immunocytochem- detection of 4 globin chains in screening for a ical detection of embryonic 4 chains in thalassaemia. screening for a thalassaemia. Methods-Blood samples from 115 patients with relevant clinical history and hypo- Methods chromic microcytic indexes were screened Blood samples from 115 patients with relevant using the HbH inclusion test and the Vari- clinical history and microcytic (mean cor- ant Hemoglobin Testing System (BioRad, puscular volume (MCV) < 80 fl) and hypo- Hercules, CA, USA). chromic indexes were subjected to the following Results-The HbH inclusion test was studies: (1) HbH inclusion test; (2) estimation positive in 61 of 115 cases, three of whom of HbA2 and HbF by the Variant Hemoglobin had HbH disease confirmed by electro- Testing System3 (BioRad, Hercules, California, phoresis. The remaining 58 had a thal- USA); and (3) electrophoresis of the haem- assaemia 1. All three HbH cases and 56 olysate in cellulose acetate at pH 8-55 using of 58 cases of a thalassaemia 1 expressed the Super Z kit Hb system as appropriate. embryonic 4 chains, giving a specificity of Using the Variant Hemoglobin Testing System, 96-7%. Fifty four of115 cases had a negative the normal adult reference ranges for HbA2 HbH inclusion test, of whom 50 had P and HbF in our laboratory are, respectively, thalassaemia trait and three had iron de- 2 3-3O0% and 0-0 9% (unpublished data). The ficiency. No diagnosis was reached for the normal adult reference range for MCV in our remaining patient. local population, based on determination of Conclusion-The immunocytochemical erythrocyte indexes using the Technicon H*1 test is as sensitive as the HbH inclusion blood analyzer, is 83-96 fl.4 Serum iron and test in screening for a thalassaemia. The serum ferritin concentrations were determined http://jcp.bmj.com/ presence of4 chains is highly specific for a in 40 cases. thalassaemia 1 incorporating the (--/SEA) deletion. The specificity and simplicity of the immunocytochemical test make it the HBH INCLUSION TEST test of choice in screening for a thal- Air dried peripheral blood smears were made assaemia. from a mixture of two parts 1% brilliant cresyl (JT Clin Pathol 1995;48:861-864) blue to one part blood following incubation at on September 24, 2021 by guest. Protected copyright. 37°C for exactly 30 minutes. For detection of Keywords: HbH inclusion test, 4 chains, a thalassaemia. inclusions, 1000 to 5000 red blood cells were examined under an oil immersion lens which requires up to 15 minutes per case. In South East Asia, where a and thalassaemia are prevalent, the haemoglobin (Hb) H in- clusion test, based on incubation of eryth- DETECTION OF EMBRYONIC 4 CHAINS rocytes in brilliant cresyl blue, is used Air dried peripheral blood smears were fixed Haematology Section, extensively in many laboratories as a screening in a solution containing 90% absolute acetone Department of Pathology, test for subjects with a thalassaemia. The pres- and 10% absolute methanol at room tem- Queen Mary Hospital ence of HbH inclusions is correlated with the perature for five to 10 minutes. They were then Compound, aocl-- genotype (including those with deletions washed in phosphate buffered saline (PBS) University of Hong Kong, Hong Kong involving the complete 4-a gene cluster) as containing 3% bovine serum albumin (BSA); L C Chan present in oa thalassaemia 1 carriers and the 10 ml of a 1 in 20 dilution of murine mono- J CC So a-/-- genotype associated with HbH disease.' clonal antihuman 4 globin chain conjugated Based on positive couples at risk of with fluorescein isothiocyanate was then added. Department of findings, Pathology, producing fetuses with hydrops fetalis due to The slides were incubated in a humidified McMaster University, Hb Barts are identified for genetic counselling chamber at room temperature for 30 minutes Hamilton, Ontario, and prenatal diagnosis. Recently, an im- and were then washed in BSA free PBS, covered Canada D H K Chui munocytochemical test which detects em- with glycerol and examined under a fluor- bryonic 4 chain in erythrocytes was shown to be escence microscope (Zeiss). Cord blood and Correspondence to: Dr L C Chan. highly specific and sensitive for a thalassaemia 1 blood from an individual with a normal full Accepted for publication carriers of the (--SEA/) deletion'-that is, the blood count were used as positive and negative 30 March 1995 oca/-- genotype in which the 4 globin gene controls, respectively, for 4 globin expression. 862 Chan, So, Chui 0 kb 10kb 20 kb 30kb J Clin Pathol: first published as 10.1136/jcp.48.9.861 on 1 September 1995. Downloaded from 52 inter- CHVR 5V1 Na2 va1 a2 al 01 3 HVR

- C30,, 12

-SEA deletion 3

1=5'-GCGATCTGGGCTCTGTGTTCT-3' 2=5'-GTTCCCTGAGCCCCGACACG -3' 3=5'-ACTGCAGCCTTGAACTCCTG -3' Figure 1 Alpha globin gene loci and primers for PCR analysis.

DNA EXTRACTION AND POLYMERASE CHAIN an automated themocycler under the following REACTION conditions: denaturation at 94°C for 30 sec- DNA was extracted from peripheral blood onds, annealing at 58°C for 45 seconds and using a commercial kit (Intragene Purification extension at 72°C for one minute, except for Matrix; BioRad). To detect the (--/SEA) de- the first cycle when denaturation took place at letion, three primers were synthesised using 95°C for 90 seconds and at 94°C for 30 sec- details of the sequences provided by Chang et onds. After amplification, DNA was elec- al5 (fig 1). Primers 1 and 2 amplify a 314 trophoresed on a 2% agarose gel at 100 V nucleotide band if normal or if a thalassaemia for one hour and visualised using ethidium 2 is present, whereas a thalassaemia 1 with the bromide. (--/SEA) genotype will yield a 195 nucleotide band on amplification with primers 1 and 3. DNA (250 ng) was amplified using the poly- STATISTICAL ANALYSIS merase chain reaction (PCR) with primers at The sensitivity and specificity of the im- a final concentration of 2 mtM for 35 cycles on munocytochemical test for the diagnosis of ax thalassaemia were calculated as described by Simel.6

Results The HbH inclusion test was positive in 61 of the 115 cases studied. Three of these 61 http://jcp.bmj.com/ patients had HbH disease based on the pres- ence of multiple inclusions and a HbH band on electrophoresis. In the remaining 58 the pattern ofHbH inclusions was compatible with a diagnosis of a thalassaemia 1. All three cases of HbH disease expressed embryonic 4 chains as did 56 of the 58 cases of a thalassaemia 1. Therefore, the sensitivity of the immuno- on September 24, 2021 by guest. Protected copyright. cytochemical test compared with conventional methods for the diagnosis of a thalassaemia was 96-7% (59/59+2). Figure 2 shows the fluorescence pattern of erythrocytes, revealing varying staining intensities in a patient with1a thalassaemia 1. Of the two patients (cases 47 and 71) with HbH inclusions, but without embryonic 4 chains, one had iron deficiency anaemia whilst the other had an unexplained hypochromic microcytic anaemia with no family history of thalassaemia. As expected, the (-/SEA) de- letion was not detected on DNA analysis of the a globin locus by PCR in both of these cases. By contrast, an amplified band of 195 base pairs, confirming the presence of the (--/ SEA) deletion, was present in all seven cases of a thalassaemia 1-that is, both HbH and 4 chain positive. The results ofthe DNA analysis are presented in fig 3. Fifty four cases had a negative HbH inclusion Figure 2 Immunofluorescent erythrocytes stained with anti-C chain antibody in a case of test. Fifty had thalassaemia trait based on ax thalassaemia 1. (Magnification x 2000.) raised A2 and F concentrations and three were Screening methods for a thalassaemia 863 Case 1 Case 47 Case 71 M A B A B A B J Clin Pathol: first published as 10.1136/jcp.48.9.861 on 1 September 1995. Downloaded from

404- 367- 314- 242- 195- 190-

Figure 3 An ethidium bromide stained agarose gel of DNA amplifiedfrom cases 1 (HbH positive, 4 chain positive), 47 and 71. Lane A, primers 1 and 2; lane B, primers 1 and 3; M, DNA markers.

1 15 cases (MCV < 80 fl)

HbH positive HbH negative (n = 61) (n = 54)

HbH disease a Thalassaemia P Thalassaemia Iron deficiency (n = 3)b http://jcp.bmj.com/ pattern (n = 3) (n = 58) (n = 50)a Unknown (case 74)c (confirmed by electrophoresis) -all 4 chain positive 4 positive 4 negative = = (n 56)d (n 2)e on September 24, 2021 by guest. Protected copyright. -iron deficiency (case 71) -unexplained (case 47) Figure 4 Flow chart of results of the HbH inclusion test and C globin chain expression. aFive offive cases 4 chain negative; btwo of two cases C chain negative; 'C chain positive; seven of seven cases (CISEA) positive; '(-ISEA) negative.

iron deficient. The 4 chain was expressed in Embryonic 4 chains were present in 56 (96 5%) the patient (case 74) whose iron status was of 58 cases of oa thalassaemia 1 based on a unknown and whose HbH inclusion test was positive HbH inclusion test. Of the two cases negative, but not in the five cases of ,B thal- (47 and 74) which did not express the 4 chain, assaemia and two cases of iron deficiency ex- DNA analysis confirmed that the (--/SEA) de- amined. As no DNA was available to resolve letion was absent as expected. It is possible the oa globin status of case 74, the specificity that, in these two cases, the 4 genes were deleted ofthe immunocytochemical test compared with together with the oa gene loci, as is the case in conventional methods (after excluding case 74) adult carriers of deletions such as (--Fill), for the diagnosis of oa thalassaemia was 100% (-- Thai/), and (--HW/) in South East Asia.78 (7/0 + 7). A summary of the data is shown in Case 47 had a negative discriminant func- fig 4. tion as calculated using red blood cell indexes9 which is in keeping with the diagnosis of a thal- assaemia. Other possibilities which may Discussion account for negative 4 chain expression The immunocytochemical detection of em- include the -al- oa, -a/aCsot and ----/C4aax' bryonic 4 chains is as sensitive as the HbH genotypes, or a "false positive" HbH inclusion inclusion test in screening for a thalassaemia. test. 864 Chan, So, Chui We have also shown that the immuno- cytochemical test will be negative but the HbH cytochemical test is highly specific for the diag- inclusion test will be positive. Therefore, we

nosis ofa thalassaemia 1. However, the 4 globin recommend that screening for oc thalassaemia J Clin Pathol: first published as 10.1136/jcp.48.9.861 on 1 September 1995. Downloaded from chain was also detected in a single case (case be undertaken using 4 chain detection; negative 74) for whom the HbH inclusion test was cases must be confirmed using appropriate mo- negative. As the HbF and A2 concentrations lecular analysis. were normal and no variant haemoglobins were detected, the absence ofHbH inclusions cannot We thank F Lee, Amy Chan and J Y Chan for excellent technical be because of the presence of f thalassaemia assistance and J Kwok for preparing the manuscript. Serum iron and ferritin concentrations were kindly provided by Mr D trait or HbE, both of which are known to Robinson, Clinical Biochemistry Unit, Queen Mary Hospital, decrease significantly the number of HbH in- Hong Kong. clusions.' As no cells were stored for DNA analysis, we cannot confirm whether the (- 1 Skogerboe KJ, West SF, Smith CS, Terashita ST, LeCrone CN, Detter JC, et al. Screening for a-thalassaemia. Cor- SEA/) deletion was present using PCR. relation of hemoglobin H inclusion bodies with DNA- We have shown that the immunocyto- determined genotype. Arch PatholLab Med 1992;116:1012- 18. chemical detection of embryonic 4 chains is a 2 Tang W, Luo H-Y, Eng B, Waye S, Chui HK. Im- practical alternative to the HbH inclusion test munocytochemical test to detect adult carriers of the (-- SEA/) deletional oa-thalassaemia. Lancet 1993;342:1145-7. in the diagnosis of oc thalassaemia. As staining 3 Tan GB, Aw TC, Dunstan TA, Lee SH. Evaluation of high is positive in almost all erythrocytes, only small performance liquid chromatography for routine estimation of haemoglobins A2 and F. J Clin Pathol 1993;46:852-6. numbers of red cells need to be evaluated 4 Robertson EP, Pollock A, Yau KS, Chan LC. Use of Tech- compared with the laborious screening of sev- nicon H* 1 technology in routine thalassaemia screening. Med Lab Sci 1992;49:259-64. eral fields for the detection of HbH inclusions. 5 Chang J-G, Lee L-S, Lin C-P, Chen P-H. Rapid diagnosis Detection of 4 chains using immunofluor- of a-thalassaemia 1 of Southeast Asia type and hydrops fetalis by polymerase chain reaction. Blood 199 1;78:853-4. escence can be carried out in batches and the 6 Simel DL. Playing the odds. Lancet 1985;i:329-30. results can be rapidly scored and assigned as 7 Tang W, Luo H-Y, Albitar M. Human embryonic 4 globin chain expression in deletional a-thalassaemias. Blood 1992; definitively positive or definitively negative. 80:517-22. It must be emphasised that screening for 8 Chui DHK, Wong SC, Chung S-W, Patterson M, Bhargaua S, Poon M-C. Embryonic 1-globin chains in adults: a embryonic 4 chains, like all laboratory tests, marker for a-thalassaemia due to a 17.5kb deletion. N Engl will need to be interpreted together with other J Med 1986;314:76-9. 9 England JM, Fraser PM. Discrimination between iron de- relevant clinical and laboratory data. In a few ficiency and heterozygous-thalassaemia syndromes in patients with 4-oc thalassaemia the immuno- differential diagnosis ofmicrocytosis. Lancet 1979;i: 145-8. http://jcp.bmj.com/ on September 24, 2021 by guest. Protected copyright.