J Clin Pathol: first published as 10.1136/jcp.28.11.920 on 1 November 1975. Downloaded from

920 Technical methods Technical methods

Dye Supplier and Batch No. Sulphated Ash yielded % A standardized or Purchase Date By Dried After desalting Commercial prepared from purified dyes Product BASF 6698403 19-87 0-27 P. N. MARSHALL, S. A. BENTLEY, AND S. M. LEWIS BDH 145810 16 50 <0*05 Eastman PE4 21-59 <0.05 Department of Haematology, Royal Postgraduate Edward Gurr Feb 1973 16-63 0-10 Medical School, Hammersmith Hospital, Du Cane R.A. Lamb0715 19-53 0-48 Road, London W12 OHS R. A. Lamb 2621 18-18 <0 05 Table Sulphated ash analyses ofeosin samples Stains consisting of combinations of various thiazine dyes with , for example, those of Giemsa, Leishman, and Wright, are used routinely to study the cellular morphology ofblood and bone marrow. The major problem in the use ofthese Romanowsky- (d) Finally, the precipitate was dried at 135-140°C type stains is batch-to-batch variations in their for approximately 4 hours. properties as a result of variations in their The purity of this free acid was confirmed by dye composition and the presence of contaminating sulphated ash analyses. metal salts (Lillie, 1943 and 1944a; Lillie and Roe, FORMULATION OF STAIN 1942; Marshall et al, 1975a; Saal, 1964; Scott and copyright. French, 1924). Purified , azure B, and eosin were The extent of these variations in a large number mixed in varying proportions. In some mixtures of commercially available Romanowsky-type stains either methylene blue or azure B was omitted. was demonstrated in a previous study (Marshall The mixtures were dissolved in 1:1 v/v glycerol- et al, 1975a). In this study it was found that the methanol, in which the dyes are readily soluble and simplest successful stain examined contained methy- highly stable (Lillie, 1944b). These solutions were lene blue, azure B, and eosin and was only mini- diluted with varying proportions of Sorensen's mally contaminated with metal salts. Accordingly, phosphate buffer, pH 6-8 or 7-2, 0-00132 M (available http://jcp.bmj.com/ we have produced a stain from purified samples of from Mercia, Watford, Herts). Films prepared on these dyes. good quality microscope slides (Chance Propper Ltd, Smethwick, Warley) were fixed in methanol, Materials and Methods stained for varying times in these solutions, and differentiated in buffer pH 6-8 or 7-2. They were then air-dried and mounted in a synthetic resin (Diatex, PURIFICATION OF DYES R. A. Lamb Ltd).

Methylene blue and azure B were purified as des- on September 26, 2021 by guest. Protected cribed previously (Marshall and Lewis, 1975). The stained films were examined by an observer Commercial samples of eosin (table) were freed of who was unaware of the composition of the stains contaminating metal salts and volatile material used and they were compared with duplicate films and converted to the free acid by the following stained with a traditional May-Grunwald-Giemsa scheme: sequence as routinely used in this laboratory (Dacie (a) The commercial dye was dissolved at 10 g/l and Lewis, 1968). in 0-1 M aqueous ammonia. Any insoluble material was removed by filtration. OPTIMAL STAINING PROCEDURE (b) 0-2 M aqueous hydrochloric acid was added After testing a number of staining solutions and to this solution until precipitation ceased. staining times, the procedure presented below was (c) The precipitate was removed by filtration found to give excellent results: and washed several times with dilute aqueous (a) The following quantities of the purified dyes hydrochloric acid (ca 0'0005 M). were mixed: methylene blue, 9 0 mg; azure B, 17-0 mg; eosin, 13-0 mg. The mixture was dissolved in 20 ml of 1:1 v/v glycerol-methanol by gentle Received for publication 19 May 1975, stirring for a few minutes. (This stock solution 920 J Clin Pathol: first published as 10.1136/jcp.28.11.920 on 1 November 1975. Downloaded from

Technical methods 921

4 copyright.

.5

Fig. 3 .'t. -

U http://jcp.bmj.com/

F. on September 26, 2021 by guest. Protected

... ." I

I. Fig. S Fig. 6

Figs 1 to 6 Material stained using the procedure described in the text. Figs 1 and 2 Photomicrographs ofa normal bloodfilm. Fig 3 Photomicrographs ofa bloodfilmfrom a case of fi-thalassaemia major showing anisochromasia, polychromasia, andpunctate basophil stippling. Fig 4 Photomicrograph ofa bloodfilmfrom a case ofacute myeloblastic leukaemia. The blast cellhasprominentnucleoli and contains an Auer rod. Figs 5 and 6 Photomicrographs of a bone marrow film from a case of megaloblastic anaemia showing erythroid and granular precursors. J Clin Pathol: first published as 10.1136/jcp.28.11.920 on 1 November 1975. Downloaded from

922 Technical methods may be stored for many months without alteration ci in its composition.) ) (b) The solution was diluted with 100 ml of (CH,),23(C,2 C.I. 52015 Sorensen's phosphate buffer, pH 6-8, 0-00132 M. (c) Films of blood (collected into EDTA-K2 anticoagulant, 15 mg/ml of blood) and aspirated Br bone marrow were fixed for 5-10 minutes in meth- () B(CHB NH-CH anol. They were stained by immersion in the C.I. 52010 diluted solution for 10 minutes. (d) Stained films were differentiated for 1 minute in the pH 6-8 buffer, air-dried, and mounted. Results and Discussion C.1.45380 Brr (free acid) Typical examples of the appearances of films stained H- OOH by the recommended technique are shown in figs 1 to 6. These show the typical scheme of colouration produced by successful Romanowsky-type stains. It should be noted, however, that the colour balance Fig 8 Structuralformulae ofthe dyes used to prepare and some of the more subtle features seen in the the staining solution (a) methylene blue; (b) azure B; films, especially fine granulation, have been lost and (c) eosin (free acid). in thereproduction. Although the number ofcoloured components in this stain is small (see figs 7 and 8) compared with that typically seen in commercial subtle colour variation. The presence of methylene stains (Marshall et al, 1975a) the characteristic subtle blue enhanced nucleolar staining and red variation in colour was present. polychromasia. In addition, it resulted in monocytecopyright. Stains not containing methylene blue gave results cytoplasm being stained grey in contrast to the blue- approaching a Romanowsky effect but lacking this grey colouration seen in its absence. Those stains not containing azure B stained chromatin blue instead of purple. Replacement of the pH 6-8 buffer by one of pH Zq SOLVENT FRONT 7*2 resulted in an orange-grey colouration of red cells. The staining of white cells was unaffected. Stains prepared from different batches of purified http://jcp.bmj.com/ dyes gave identical results. To date, about 200 films of normal and pathological blood and bone marrow I Tribromofluores(ein have been stained. On no occasion was any staining variation seen. This reproducibility obtained from lEosin a stain prepared from purified dyes represents a distinct advantage over that of currently available

commercial stains (Marshall et al, 1975a). Another on September 26, 2021 by guest. Protected advantage which makes the stain particularly suitable for routine use is that the stock solution in glycerol-methanol is rapidly prepared and may be used immediately. It may be noted that the prep- aration of stock solutions of Giemsa, Jenner, and Wright stains (Baker et al, 1966; Lillie, 1969) Azure B involves a lengthy solution and ageing process. The Methylene blue stock solution of our new stain is extremely stable. ORIGIN Once diluted with buffer, the stain must be used within a period of about 3 hours; diluted solutions show Fig 7 Thin-layer chromatogram (prepared according to of the conventional Romanowsky-type stains MarshallandLewis (1974)) ofthe Romanowsky stain similar stability. described in the text. The major components are methy- The effectiveness of the procedure for desalting lene blue (Rr 0059), azure B (Rf 0-11), and eosin (Rf 0 60). eosin is illustrated in the table. Commercial samples A trace oftribromofluorescein (Rf 0-72) ispresent. of eosin contain a small proportion of tribromo- J Clin Pathol: first published as 10.1136/jcp.28.11.920 on 1 November 1975. Downloaded from

Technical methods 923 fluorescein (Marshall et al, 1975b). This is not eosin and methylene blue. Publ. Hlth Rep. (Wash.), 58, removed by the desalting procedure, although it 449-452. Lillie, R. D. (1944a). Factors influencing the Romanowsky may be removed chromatographically (Graichen staining of blood films and the role of methylene violet. and Molitor, 1959). However, we have found that J. Lab. clin. Med., 29, 1181-1197. this is unnecessary for the present staining procedure. Lillie, R. D. (1944b). The deterioration of Romanowsky Tribromofluorescein-free eosin and eosin deliber- stain solutions in various organic solvents. Publ. Hlth Rep. (Wash.), Suppl. 178, 1-16. ately contaminated with a much larger proportion Lillie, R. D. (1969). H.J. Conn's BiologicalStains, 8th edition, of tribromofluorescein than is present in commercial pp. 438-439. Williams and Wilkins, Baltimore. samples gave indistinguishable results. Lillie, R. D. and Roe, M. A. (1942). Studies on polychrome methylene blue. I. Eosinates, their spectra and staining capacity. Stain Technol., 17, 57-63. This work was supported by a grant to S. M. Lewis Marshall, P. N., Bentley, S. A., and Lewis, S. M. (1975a) from the Department of Health and Social Security. An evaluation of some commercial Romanowsky stains. J. clin. Path., 28, 680-685. Marshall, P. N., Bentley, S. A., and Lewis, S. M. (1975b). A procedure for assaying commercial samples of eosin. References Stain Technol., 50, 107-113. Marshall, P. N. and Lewis, S. M. (1974). A rapid thin-layer Baker, F. J., Silverton, R. E., and Luckcock, E. D. (1966). chromatographic system for Romanowsky blood stains. An Introduction to Medical Laboratory Technology, 4th Stain Technol., 49, 235-240. edition, p. 524. Butterworths, London. Marshall, P. N. and Lewis, S. M. (1975). The purification of Dacie, J. V. and Lewis, S. M. (1968). Practical Haematology, methylene blue and azure B by solvent extraction and 4th edition. Churchill, London. crystallization. Stain Technol., 50, in press. Graichen, C. and Molitor, J. C. (1959). Studies on coal-tar Saal, J. R. (1964). for the diagnosis of bovine colors. XXII. 4,5-dibromofluorescein and related bromo- babesiosis. I. Staining properties of commercial samples fluoresceins. J. Ass. off. agric. Chem., 42, 149-160. and their component dyes. J. Protozool., 11, 573-582. Lillie, R. D. (1943). A Giemsa stain of quite constant Scott, R. E. and French, R. W. (1924). Standardization of composition and performance, made in the laboratory from biological stains. Milit. Surg., 55, 229-243 and 337-352. copyright.

Modifications to improve the The Rinse and Drain Mechanisms reliability of the Coulter 'S' The most troublesome limitation exposed when the

instrument was run continuously was the failure http://jcp.bmj.com/ P. F. J. NEWMAN, E. ROBERTSON, AND M. J. MUIR of the aperture baths to drain completely. (This Haematology Department, Royal Infirmary, happened several times within a period of three Edinburgh, Scotland months.) In the original version of the instrument, during each cycle waste fluid was drawn under The Coulter Counter Model 'S' System (Coulter vacuum from the aperture baths into the waste Electronics Ltd) measures automatically the haema- chamber and then discharged to waste under 0 35 kgf/cm2 positive pressure. Owing to the low surface

tological parameters haemoglobin concentration, on September 26, 2021 by guest. Protected white cell count, red cell count, and mean cell tension of the effluent, caused by the lysing agent, volume and also computes from them the haemato- and the turbulence created by the vacuum, there was crit, mean cell haemoglobin, and mean cell haemo- a tendency for excess foam to accumulate in the globin concentration. The instrument can handle waste chamber which then overflowed into the trap one sample every 20 seconds and can be both precise bottle and from there into the switching valve (SV) and reliable (Barnard et al, 1969). However, in our No. 6, which controls the vacuum in the pneumatic laboratory the model 'S' is used to process up to 800 card (L2). This caused the switching valve to become samples per day, and we have found that these progressively less efficient with the result that the conditions of virtually continuous operation expose aperture baths were sometimes only partially limitations in the design of the instrument with a drained, and carry-over between successive samples concomitant fall off in its reliability. We now de- was observed. The valve eventually failed com- scribe some of these limitations and suggest ways of pletely. The manufacturers then modified the instru- modifying the machine to overcome them. ment by fitting an anti-foam device (Coulter Altera- tion Number ECO-0367) which gave only partial Received for publication 10 July 1975. improvement; however the excess foam still over-