Guideline for the flow cytometric enumeration of CD34 + haematopoietic stem cells

Clin. Lab. Haem. 1999, 21, 301 308 Guideline for the flow cytometric enumeration of CD34 + haematopoietic stem cells PREPARED BY THE CD34 + HAEMATOPOIETIC STEM CELL WORKING PARTY* *D. Barnett, G. Janossy, A. Lubenko, E. Matutes, A. Newland & J.T. Reilly Members of the General Haematology Task Force of the British Committee for Standards in Haematology: J.T. Reilly (Chairman), B.J. Bain (Secretary), R. Amos, I. Cavill, C. Chapman, K. Hyde, E. Matutes, J. Parker-Williams, I.D. Walker Introduction It is well established that the 3% of cells in the bone marrow which express the CD34 antigen, a heavily glycosylated mucin-like structure, are capable of reconstituting longterm, multilineage haematopoiesis after myeloablative therapy (Berenson et al. 1988; Andrews et al. 1992). CD34 + cells are also found in the peripheral blood of normal individuals, but are extremely rare (approximately 0.01 0.05% of total nucleated cells). However, current treatment and mobilization regimes, including chemotherapy and/or haematopoietic growth factors, can significantly increase circulating CD34 + stem cell numbers in patients and healthy donors. Peripheral blood stem cells (PBSC) have now virtually replaced bone marrow as the primary source of stem cells for autologous transplantation (Gratwohl, Hermans & Baldomero 1996) and the procedure is being increasingly used for allogeneic transplantation between HLA-identical siblings (Russell, Gratwohl & Schmitz 1996). Potential advantages of PBSC transplantation include a more rapid haematopoietic reconstitution, lower risk of contamination by tumour cells, reduced hospitalization costs, increased numbers of T-lymphocytes and NK cells that may reduce post-transplant relapse, and the elimination of a need for general anaesthesia (Dreger et al. 1994; Ager et al. 1995). In addition, the PBSC product is also more suitable for ex vivo manipulation, such as CD34 + cell selection (Brugger, Henschler & Heimfeld 1994), tumour purging (Ross et al. 1995) and gene transfer (Bregni et al. 1992). Transplant centres routinely rely on CD34 + cell quantification by flow cytometry to determine the optimal timing and to confirm the adequacy of PBSC harvests (Haas et al. 1994). In contrast, assessment of haematopoietic pro- Correspondence: BCSH Secretary, British Society for Haematology, 2 Carleton House Terrace, London SW1Y 5AF, UK. Whilst the advice and information in these guidelines is believed to be true and accurate at the time of going to press, neither the authors nor the publishers can accept any legal responsibility or liability for any errors or omissions that may be made. 1999 Blackwell Science Limited genitor cells (HPC) in colony-forming assays, although correlating with CD34 levels (Siena et al. 1991), is handicapped by a lack of reproducibility and the prolonged assay time. A minimum threshold dose, in adults, of between 2 and 5 10 6 CD34 + cells/kg body weight has been observed in multiple clinical settings to result in adequate engraftment (Krause et al. 1996), although the lack of standardization of the assay prevents a more exact definition of threshold level (Bender et al. 1992). This lack of standardization with respect to reagents and gating strategies, as well as to the use of different haematology instruments to derive absolute total nucleated cell counts has contributed to the widespread inter-laboratory variation (Barnett et al. 1998). As a result, there is an urgent need for a nationally agreed protocol to enable standardization of the assay and to enable comparison of clinical and laboratory data. This guideline provides recommendations for: (1) specimen collection and frequency of CD34 testing; (2) monoclonal antibody (mab) selection; (3) cell separation, lysing and counting techniques; (4) gating strategies; (5) isotype controls; (6) determination of absolute counts; (7) instrument quality control; (8) data analysis; (9) data reporting; (10) data storage; and (11) quality assurance. Finally, this guideline will also give a brief overview of several new technologies that still require evaluation. Specimen collection and frequency of CD34 testing All peripheral blood specimens should be collected by venepuncture into either 0.34 M di- or tri-potassium ethylenediaminetetraacetic acid (K 2 EDTA or K 3 EDTA) anticoagulant. The total WBC must be obtained within 6 h of venepuncture. If a specimen is referred to a central laboratory for analysis resulting in a delay of over 6 h, then the result of a total WBC must accompany the sample. All samples must be labelled with the patient s surname and forename(s), a unique patient identifier such as the hospital reference number and date of birth as well as 301

302 CD34 + Haematopoietic Stem Cell Working Party the patient location and the date and time of collection. Limited data are available on the stability of the CD34 antigen at different temperatures. Thus, to minimize the influence of temperature and storage it is recommended that the specimen (peripheral blood or apheresis specimens) is stored at a constant temperature of 4 C and all processing completed within 12 h. Packaging and transportation of peripheral blood, bone marrow or apheresis samples should be in accordance with the regulations of the postal and/or courier service used. Specimen integrity must be checked upon receipt and a repeat specimen requested if there is evidence of clot formation, gross haemolysis and/or if the specimen is received more than 12 h from the time of collection. Timing and adequacy of stem cell harvest should be decided by the CD34 + cell count. Harvesting can commence when the peripheral blood CD34 count is 10 cells/ml with the aim of obtaining a yield of 2 10 6 CD34 + cells/kg (Dreger & Schmitz 1998). Monoclonal antibody selection The CD34 antigen has epitopes that can be classified according to their sensitivity to neuraminidase and glycoprotease enzymes obtained from Pasteurella haemolytica (Sutherland, Marsh & Davidson 1992). Epitopes recognized by class I antibodies (e.g. ICH3, MY10, 12.8, Immuno133, Immuno409) are sensitive to both enzymes. Those detected by class II antibodies (e.g. QBEND10, 43.A1) are sensitive to glycoprotease only, while those detected by class III antibodies (e.g. Tük3, HPCA-2, BIRMA-K3) are resistant to both enzymes. Class I antibodies fail to detect all the glycoforms of the CD34 antigen and may fail therefore to detect CD34 expression on some leukaemic cells (Sutherland & Keating 1992). In addition, the lower avidity and reduced reactivity of conjugated class I antibodies when compared to those of class II and III, further limits their applicability. Overall, the above observations not only underline the importance of selecting an appropriate CD34 antibody clone, but also of selecting one that retains the high specificity and avidity of its binding after conjugation to the designated fluorochrome. Therefore, only class II (phycoerythrin (PE)-conjugated) or class III (any conjugate) antibodies should be used for CD34 + stem cell enumeration. The use of the most discriminating fluorochrome with an argon laser, PE, is recommended for detection of CD34 + cells (Johnsen 1995; Sutherland et al. 1996). Advantages of PE-conjugates include better demarcation between negative and positive cells, reduced binding to Fc receptors and lower non-specific binding to dead cells (Gratama et al. 1997a; Barnett et al. 1998; Gratama et al. 1998). Fluorescein isothiocyanate (FITC) conjugation, because of its induced negative charge, interferes with the binding properties of some class II antibodies, such as QBEND10, to their epitopes. Therefore, FITC-conjugated class II antibodies must not be used. Antibodies conjugated with tandem fluorochromes (i.e. PE Cy5) or allophycocyanin have only recently become readily available and data on their performance is limited. If these latter fluorochromes are used then the performance should be established in comparison to the equivalent PE-conjugate. An FITC-conjugated mab to the CD45 antigen should be used which not only detects all isoforms but also all glycoforms, e.g. anti-hle-1 (Becton- Dickinson Immunocytometry Systems, San Jose, CA, USA) or J33 (Beckman-Coulter, Miami, FL, USA). Used in combination with a PE-conjugated anti-cd34, the anti-cd45 FITC mab will thus provide an additional parameter for the identification of CD34 + cells. This antibody combination enables the identification of true CD34 + cells and simultaneously provides a useful indicator for the effectiveness of red cell lysis. It also facilitates the discrimination of cells from contaminating events, such as platelets, platelet aggregates and other debris that can bind low levels of PE-conjugated CD34 antibodies (Sutherland et al. 1996). Nucleic acid dyes (NAD) as gating reagents have recently been introduced and enable the exclusion of unlysed red cells, platelets and debris from the analysis (e.g. Laser Dye Solution (LDS)-751 and the proprietary NAD reagent provided with the ProCOUNT TM (kit)). However, NADs have a number of limitations. For example, the NAD used in ProCOUNT TM has a significant spectral overlap within the PE channel, which can reduce the resolution between CD34 + and CD34 populations. Furthermore, the detection of LDS-751 on some Coulter flow cytometers can only be achieved after the replacement of the 610 620 nm band pass filter with a 650-nm long pass filter (Gratama et al. 1998). However, LDS-751 is of low cost and has negligible spectral overlap with PE. Finally, if a NAD is used in conjunction with anti-cd45 mab then an instrument capable of four colour analysis will be required for CD34 + subset analysis. All initial data must be acquired ungated because in some circumstances, such as BM regeneration and cytoreductive therapy, some CD34 + cells may express very low levels of CD45 (Gutensohn et al. 1996). Cell separation, lysis and counting techniques It should be noted that the CD34 antigen may be affected by fixatives used with specific cell labelling methods (Macey

Flow cytometric enumeration of CD34 + haematopoietic stem cells 303 et al. 1997). Furthermore, a number of studies have reported improved recovery using whole blood lysis when compared to cell separation techniques (Siena et al. 1991; Owens & Loken 1995). Although the optimal method for sample preparation has yet to be established, data are available suggesting that CD34 + stem cell analysis using a whole blood, lyse-no-wash, technique, with fixative-free NH 4 Cl-based lysing reagents, is the preferred approach (Menéndez et al. 1998). However, it should be stressed that certain proprietary kits, i.e. ProCOUNT TM, use a fixative-based reagent and must therefore be used under manufacturer s recommendations only. Furthermore, to ensure optimum sample condition, samples should be analysed immediately post-lysis, or kept on melting ice until analysis (Serke et al. 1998). However, all analyses must be completed within 1 h of addition of the lysing reagent. It is important when counting rare events that sufficient numbers are collected to produce a statistically valid result. Reducing the number of events analysed per test (tube) reduces the reliability of the estimation. Given the fact that the standard error of the number of positive cells per analysis is given by the square root of the number of positive cells, the larger the acquisition the lower the coefficient of variation (CV) (Wunder et al. 1992). UK NEQAS data has highlighted a marked variation in the number of CD45 + cells counted (ranging from 10,000 to 1.2 million), with a number of laboratories counting significantly less than 50,000 (Barnett et al. 1998). To achieve an intra-assay CV of 10% a minimum of 100 CD34 + events should be collected (Sutherland et al. 1996). Gating strategies A number of assay protocols for the flow cytometric enumeration of CD34 + cells have been published. The first of these was the Milan protocol (Siena et al. 1991) which was later improved upon by the Mulhouse group to incorporate direct immunofluorescence, whole blood staining and a lyse-wash method (Bender, Unverzagt & Walker 1994) and anti-cd45, whilst the SIHON protocol (Gratama et al. 1997b) evolved further to incorporate the laser dye solution (LDS) 751, anti-cd14 and anti-cd66e in order to achieve greater gating accuracy and eliminate non-specific mab binding. However, the ISHAGE (International Society for Hematotherapy and Graft Engineering) gating strategy developed by Sutherland et al. (1996) has gained increased popularity and is the most frequently used analysis method in the UK NEQAS CD34 + stem cell quantitation scheme (Barnett et al. 1998). A study by Chang & Ma (1996) reported that gating strategies are a major contributory factor in result variability and that the gating strategy described in the ISHAGE protocol gave the most reproducible results between centres. Furthermore, the gating strategy described in the ISHAGE protocol overcomes the limitations of the Milan and Bender protocols. Both these latter protocols rely upon the use of isotype controls, and may also be unsatisfactory for samples that are not in optimal condition, or contain high numbers of lysis-resistant nucleated red cells (i.e. cord blood) (Barnett et al. 1998). For these reasons, staining of CD34 + stem cells should employ the use of CD45 (FITC) vs. CD34 (PE) together with the sequential gating strategy as described within the ISHAGE protocol (Sutherland et al. 1996). The ISHAGE sequential gating strategy (Sutherland et al. 1996) takes advantage of the fact that blast cells can be identified by their dim CD45 expression and low sidescatter (SSC) (Borowitz et al. 1993). Two tubes are required, containing anti-cd45-fitc and anti-cd34-pe. Isotype controls are not required because the gating strategy excludes cells that non-specifically bind anti-cd34. An example of a specimen analysed using the ISHAGE gating strategy is shown in Figure 1. An initial gate (R1) (Figure 1: plot 1) is set on a CD45 vs. SSC dot plot, so as to contain all CD45 + events including CD45 dim and CD45 bright. This will exclude CD45 events (i.e. red blood cells, platelets and other debris). The events in gate R1 are then displayed on a CD34 vs. SSC dot plot (Figure 1: plot 2) and a second gate (R2) produced to include the cluster of CD34 + events (Figure 1: plot 3). The third plot is obtained by plotting the events that fulfil the criteria of gates R1 and R2 (i.e. sequential gating) (Figure 1: plot 3). Cells forming a cluster with characteristic low SSC and low to intermediate CD45 fluorescence are then gated on this third plot to produce a third region (R3) (Figure 1: plot 3). Ungated data are displayed on a CD34 vs. CD45 histogram to establish the lower limit of CD45 expression by the CD34 + events (Figure 1: plot 5). Finally, the events fulfilling the criteria of all three gates (R1, R2 and R3) are then displayed on a forward light scatter (FSC) vs. SSC dot plot to confirm that the selected events fall into a generic lymph-blast region (R4) (Figure 1: plot 4). This region (R4) is precisely set to include events no smaller than lymphocytes by back scattering a small number of lymphocytes from plot 1 (Figure 1: plot 6) (high CD45 staining low side-scatter). Any events falling outside region R4 are excluded from the final calculation. CD34 + cell determination must be performed in duplicate and the mean CD34 + value used. Each laboratory should establish its own acceptability criteria for the differences between the two tubes. If the two values obtained fall outside the level of acceptability the analysis

304 CD34 + Haematopoietic Stem Cell Working Party Figure 1. A peripheral blood sample analysed for CD34 + stem cells using the ISHAGE logical gating strategy described by Sutherland et al. 1996. See text for precise logical gating strategy sequence.

Flow cytometric enumeration of CD34 + haematopoietic stem cells 305 should be repeated or a fresh specimen requested. Intraassay variation should also be established with an appropriate target CV of ³ 5%. The percentage of CD34 + cells is determined by using the number of CD34 + events expressed as a percentage of the CD45 + events (both values derived from the mean of the replicate tubes see above). The absolute CD34 + cell count in the sample is then calculated by multiplying this percentage value by the absolute total WBC as provided by the haematology analyser after correction to exclude nucleated red cells. To reduce inter-laboratory variation, a single platform approach should be employed (Barnett et al. 1999). Keeney et al. (1998) have recently adapted the ISHAGE strategy with the addition of a known number of Flow- Count TM fluorospheres (Beckman-Coulter) and implemented NH 4 Cl-lyse, no-wash, no-fix sample processing. These modifications convert the basic protocol into a single platform method to determine the absolute CD34 + cell count directly from a flow cytometer and form the basis of the Stem-Kit TM from Beckman-Coulter. Thus, the absolute count can be calculated from the observed ratio between the number of flow cytometrically counted beads and CD34 + cells. This guideline supports the use of such a system. Recently, Becton-Dickinson (Verwer & Ward 1997) have developed ProCOUNT TM, a single platform approach that uses a purpose-written computer software package. The system is based on bead counting using TruCOUNT TM tubes, and thresholding on nucleated cells, using a proprietary NAD. The TruCOUNT TM tubes contain a known number of lyophilized polyfluorescent microbeads retained in the bottom of the tube by a wire mesh grid. Sample and reagents are mixed in fixed amounts in the tubes, thus reconstituting the beads and enabling the absolute count to be determined. The system uses a lyse-no-wash procedure and sufficient events are collected to obtain a precision of 10%. This guideline supports the use of the ProCOUNT TM and Stem-Kit TM approaches. However, there must be operator inspection when suboptimal samples are being analysed using semi or fully automatic computer-assisted software programs. A limitation of the ProCOUNT TM approach is that all three fluorescence channels are utilized, thus preventing CD34 + subset analysis or viability assessment. Isotype controls Isotype controls are not required for the ISHAGE strategy as sequential gating excludes cells and debris responsible for non-specific staining with anti-cd34. Thus, very few events will meet the conditions of the combined gates and be falsely classified as CD34 +. Furthermore, isotype and fluorochrome-matched antibodies may not be representative of the concentration and fluorochrome protein (F:P) ratio of the anti-cd34 mab in use (Gratama et al. 1998). At present, isotype controls form an integral part of the ProCOUNT TM kit and should therefore be used in accordance with manufacturer s recommendations. Determination of absolute counts Absolute counts should be calculated on all specimens tested and expressed as cells/ml. A recent study has shown that the use of single platform analysis results in lower inter-laboratory variation (Barnett et al. 1999). Therefore, wherever possible, absolute counts should be derived from single platform analysis. The single platform approach also removes the need for correcting for nucleated red cells. However, care should be taken when using a dual-platform approach, particularly if dealing with apheresis, bone marrow or cord blood samples, since many haematology analysers over-estimate the leucocyte count in the presence of nucleated red cells. Correction of the total WBC, using validated methods should be employed in such circumstances. Reverse pipetting should always be used when using single platform systems (see below). In addition, apheresis specimens may require dilution with phosphate-buffered saline before staining. The WBC should be in the range for which the CD34 + HPC assay is linear, 10 10 9 /l being preferable (Gratama et al. 1998). It is important for ProCOUNT TM users to refer to the manufacturer s current instructions with respect to the working WBC concentration. Extreme care is required when diluting the specimen, since errors introduced at this stage will adversely affect the final CD34 + haematopoietic cell count. The final CD34 + count is obtained following correction by the relevant dilution factor. Reverse pipetting The accuracy and reproducibility of single platform techniques are influenced by several factors with reverse pipetting being one of the most important. With conventional pipetting there is an increased risk of inclusion of air bubbles in the dispensed volume. Thus, to improve the accuracy and reproducibility of sample dispensing, reverse pipetting should be employed. With such an approach, a slight excess of sample is aspirated into the pipette and the precise volume ejected. A small amount of the sample remains in the tip, so that deviation from the required volume as a result

306 CD34 + Haematopoietic Stem Cell Working Party of the formation of air bubbles is reduced. It is recommended that a calibrated, electronically adjustable automated pipette be employed for the procedure. Reverse pipetting should be carried out as follows: (1) press the plunger of the pipette to the second stop; (2) gently aspirate fluid; and (3) pipette the fluid against the lower end of the wall of the tube press the plunger to the first stop. Instrument quality control Daily calibration of the instrument must be performed to ensure optimal performance and must be carried out in all laboratories performing immunophenotyping. Commercial beads and process controls are used to achieve this. Commercial beads are used in order to: (i) Monitor the light scatter and fluorescence peak channel coefficients of variation. (ii) Monitor light and fluorescence peak channel drift. (iii) Monitor instrument sensitivity. (iv) Facilitate compensation set-up to adjust for spectral overlap. All values should be logged daily, together with instrument settings. In addition, all settings should be re-established following a change in bead batch or after an instrument service. Beads only provide guidance for the final flow cytometer set-up and the optimization of settings requires the analysis of a fresh, normal specimen. The use of a process control is also recommended to enable the monitoring of reagent performance, staining, lysis and analysis. Such controls should be run at least once a week (prior to any of the week s work being processed) and preferably on a daily basis. They must also be performed if: (i) there has been a change in reagent and/or laboratory personnel; (ii) there has been an instrument service and/or instrument calibration; and/or (iii) the validity of the technique is suspect (i.e. lysis problems). Process controls should be used to test the labelling procedure and if possible to test the lysing step. They should also be stable over time (minimum of 30 days). The manufacturer should have assigned target values to these specimens. Results obtained from these reagents must be plotted on a Levy Jennings type plot, thus providing a visual indication of drift or bias over time. A fresh, previously analysed or cryopreserved specimen must not be used as a process control. Each laboratory must establish its own intra-assay variation and define acceptability criteria for the test as mentioned earlier. Data analysis Operator inspection is recommended if a laboratory uses computer-assisted analysis that automatically defines the analysis regions and calculates the percentage and absolute numbers of CD34 + cells (e.g. ProCOUNT TM ). If there is any doubt as to the validity of the analysis or if replicate CD34 + values exceed the criteria stated earlier then a repeat test must be performed. Furthermore, regardless of gating strategy, no result should be issued until all calculations have been independently checked and validated. Data reporting All results must be validated before issue (see above). To avoid clerical error (e.g. erroneous placement of decimal point, insertion of additional zero, etc.) results should be reported as cells/ml. Wherever possible, cumulative reports should be collated to allow a day-to-day check. If results are telephoned, faxed or e-mailed then the laboratory should have in place a standard operating procedure that will facilitate an audit trail in case of errors or misunderstandings. All results should also be reported in a hard copy format. Data storage All primary data files, worksheets and copy report forms should be kept for a minimum of 6 months and ideally stored electronically. The flow cytometry data should be stored as list-mode files. Deletion of the data or destruction of paper records should be in accordance with the locally agreed guidelines and the Department of Health s document HC(89)20 Preservation, Retention and Destruction: Responsibilities of Health Authorities under the Public Records Act. Quality assurance It is expected that laboratories performing CD34 + haematopoietic cell counting are fully conversant with all procedures employed. Such laboratories should meet Clinical Pathology Accredited (CPA) (UK) Ltd. standards and be accredited as appropriate. In addition, both internal and external quality assurance should be undertaken. Satisfactory performance should be demonstrable in external proficiency testing schemes such as UK NEQAS Leucocyte Immunophenotyping Schemes. All quality assurance activities should be documented.

Flow cytometric enumeration of CD34 + haematopoietic stem cells 307 New technologies for haematopoietic cell analysis A number of new technologies have the potential to simplify haematopoietic analysis including the STELLer assay (Dietz et al. 1996) and haematopoietic cell measurement by haematology analyser, i.e. Sysmex SE-9000 (Takekawa et al. 1997). The STELLer system employs the IMAGN 2000 micro volume fluorimeter and measures the number of cells which are CD34 + using Cy5-conjugated anti-cd34. In contrast, the Sysmex SE-9000 system identifies haematopoietic cells based upon their phospholipid content. However, comparative data is awaited before firm recommendations can be issued. This guideline provides information for the enumeration of CD34 + stem cells using current techniques. However, several issues remain unresolved and will require further investigation before recommendations can be made, including: CD34 + subset analysis and their clinical relevance; the use of fixative-based lysing reagents; the determination of viability and finally the evaluation of the newer fluorochrome conjugates for CD34 mabs. Key points (1) The CD34 + cell count and not the change in total WBC should determine timing and adequacy of stem cell harvest. (2) Total nucleated cell count must be completed within 6 h and the CD34 + haematopoietic enumeration performed within 12 h of venepuncture. (3) All samples must be labelled with the patient s surname and forename(s), a unique patient identifier (such as the hospital reference number), date of birth, patient location, date and time of collection. (4) All samples must be fully processed within 12 h. (5) To minimize the influence of temperature, specimens should be stored at a constant temperature of 4 C. (6) Harvesting should commence when the peripheral blood CD34 + count is 10 cells/ml with the aim of having a final yield of 2 10 6 CD34 + cells/kg. (7) The use of the brightest excitable fluorochromes with an argon laser, i.e. phycoerythrin (PE), is recommended for detection of CD34 + cells. (8) To ensure there is no loss of cells during centrifugation, a lyse-no-wash technique is recommended using a fixative-free NH 4 Cl-based erythrocyte lysis reagent. (9) Class II (PE-conjugated) or class III conjugated anti-cd34 antibodies must be used. (10) Anti-pan-FITC CD45 should be used as an additional parameter for identifying the CD34 + cell population. Pan-CD45 reagents should be used that not only detect all isoforms but also all glycoforms, e.g. anti-hle-1 (Becton-Dickinson Immunocytometry Systems) or J33 (Beckman-Coulter). (11) A minimum of 100 CD34 + events should be collected. (12) The number of CD34 + cells transplanted per kg body weight should be calculated. (13) The sequential gating strategy as described in the ISHAGE protocol is recommended. (14) CD34 + determination must be performed in duplicate and the mean CD34 + value used. (15) All initial data should be collected ungated. (16) Wherever possible absolute counts should be derived using single platform technology employing reverse pipetting. (17) Control tubes for non-specific antibody binding are not required unless using ProCOUNT TM. (18) Process controls to monitor reagent, staining and, ideally, lysis procedures should be used at least once a week, but preferably daily. (19) Flow cytometer performance should be monitored daily using unstained and fluorescent latex beads. (20) Laboratories providing a routine CD34 + haematopoietic cell enumeration service should meet Clinical Pathology Accredited (CPA) (UK) Ltd. standards and be accredited where appropriate. (21) Participation (together with satisfactory performance) in accredited external proficiency testing schemes should be demonstrable. (22) Deletion of the data or destruction of paper records should be in accordance with the locally agreed guidelines and the Department of Health s document HC(89)20 Preservation, Retention and Destruction: Responsibilities of Health Authorities under the Public Records Act. References Ager S., Scott M.A., Mahendra P. et al. (1995) Peripheral blood stem cell transplants after high-dose therapy in patients with malignant lymphoma: a retrospective comparison with autologous bone marrow transplantation. Bone Marrow Transplantation 16, 79 83. Andrews R.G., Bryant E.M., Bartelmez S.H. et al. (1992). CD34 + marrow cells, devoid of T and B lymphocytes, reconstitute stable lymphopoiesis and myelopoiesis in lethally irradiated baboons. Blood 80, 1693 1701. Barnett D., Granger V., Storie I. et al. (1998) Quality assessment of CD34 + stem cell enumeration. Experience of the united kingdom national external quality assessment scheme (UK NEQAS) using

308 CD34 + Haematopoietic Stem Cell Working Party a unique stable whole blood preparation. British Journal of Haematology 102, 553 565. Barnett D., Granger V., Whitby L., Storie I. & Reilly J.T. (1999) Absolute CD4 + T-lymphocyte and CD34 + stem cell counts by single platform flow cytometry. The way forward. British Journal of Haematology 106, 1059 1062. Bender J., To L.B., Williams S. & Schwartzberg L. (1992) Defining a therapeutic dose of peripheral blood stem cells. Journal of Hematotherapy 1, 329 342. Bender J.G., Unverzagt K. & Walker D. (1994) Guidelines for determination of CD34 + cells by flow cytometry: application to the harvesting and transplantation of peripheral blood stem cells. In: Wunder E., Sovalat, H., Henon, P., Serke, S, eds. Hematopoietic Stem Cells The Mulhouse Manual. Alpha Med Press, Dayton, OH, 31 43. Berenson R.J., Andrews R.G., Bensinger W.J. et al. (1988). Antigen CD34-positive marrow cells engraft lethally irradiated baboons. Journal of Clinical Investigation 81, 951 955. Borowitz M.J., Guenther K.L., Schultz K.E. & Stelzer G.T. (1993) Immunophenotyping of acute leukaemia by flow cytometry: use of CD45 and right angle light scatter to gate on leukaemic blasts in three colour analysis. American Journal of Clinical Pathology 100, 534 540. Bregni M., Magni M., Siena S., Di Nicola M., Bonadonna G. & Gianni A.M. (1992) Human peripheral blood haematopoietic progenitors are optimal targets of retroviral-mediated gene transfer. Blood 80, 1418 1422. Brugger W., Henschler R. & Heimfeld S. (1994) Positively selected autologous blood CD34 + cells and unseparated peripheral blood progenitor cells mediate identical haematopoietic engraftment after high dose VP16, ifosamide, carboplatin and epirubicin. Blood 84, 1421 1426. Chang A. & Ma D. (1996) The influence of flow cytometric gating strategy on the standardization of CD34 + cell quantitation: an Australian multi-centre study. Journal of Hematotherapy 5, 605 616. Dietz L.J., Dubrow R.S., Manian B.S. & Sizto N.L. (1996) Volumetric capillary cytometry: a new method for absolute cell enumeration. Cytometry 23, 177 186. Dreger P., Haferlach T., Eckstein V. et al. (1994) G-CSF mobilised peripheral blood progenitor cells for allogeneic transplantation: safety, kinetics of mobilisation, and composition of the graft. British Journal of Haematology 87, 609 613. Dreger P. & Schmitz N. (1998) Sources of stem cells: autografts. In: Apperley J. Gluckman E. & Gratwohlthe A. eds EBMT Handbook: Blood and Marrow Transplantation. Gratama J.W., Van der Linden R., Van der Holt B., Bolhuis R.L.H. & Van de Winkel J.G.J. (1997a) Analysis of factors contributing to the formation of mononuclear cell aggregates (escapees) in flow cytometric immunophenotyping. Cytometry 29, 250 260. Gratama J.W., Kraan J., Levering W., van Bockstaele D.R., Rijkers G.T. & van der Schoot C.E. (1997b) Analysis of variation in results of CD34 + haematopoietic progenitor cell enumeration in a multi-centre study. Cytometry 30, 109 117. Gratama J.W., Orfao A., Barnett D. et al. (1998) Flow cytometric enumeration of CD34 + hematopoietic stem and progenitor cells. Cytometry (Communications in Clinical Cytometry) 34, 128 142. Gratwohl A., Hermans J. & Baldomero H. (1996) Haemopoietic precursor cell transplant activity in Europe 1994. Bone Marrow Transplantation 17, 137 148. Gutensohn K., Serke S., Cassens U. et al. (1996) Durchflu)zytometrische Analyse CD34 exprimierender häma- topoetischer Zellen in Blut und Zytafereseprodukten. Infusionsther Transfusionsmed 23 (Suppl. 2), 1 23. Haas R., Mohle R., Fruehauf S. et al. (1994) Patient characteristic associated with successful mobilising and autografting of peripheral blood progenitor cells in malignant melanoma. Blood 83, 3787 3794. Johnsen H.E. (1995) Report from a Nordic Workshop on CD34 + cell analysis: technical recommendations for progenitor cell enumeration in leukapheresis from multiple myeloma patients. Journal of Hematotherapy 4, 21 28. Keeney M., Chin-Yee I., Weir K., Popma J., Nayar R. & Sutherland D.R. (1998) Single platform flow cytometric absolute CD34+ cell counts based on the ISHAGE guidelines. Cytometry (Communications in Clinical Cytometry) 34, 61 70. Krause D., Fackler M.J., Civin C.I. & May W.S. (1996) CD34: structure, biology, and clinical utility. Blood 87, 1 13. Macey M.G., McCarthy D.A., van Agthoven A.J. & Newland A.C. (1997) How should CD34 + cells be analysed? A study of three classes of antibody and five leucocyte preparation procedures. Journal of Immunological Methods 204, 175 188. Menéndez P., Redondo O., Rodriguez A. et al. (1998) Comparison between a lyse-and-then-wash method and a lyse-non-wash technique for the enumeration of CD34 + hematopoietic progenitor cells. Cytometry 34, 264 271. Owens M.A. & Loken M.P. (1995) Peripheral blood stem cell quantitation. In: Flow Cytometry Principles for Clinical Laboratory Practice. Wiley-Liss Inc, New York, 111. Ross A.A., Miller G.W., Moss T.J. et al. (1995) Immunocytochemical detection of tumour cells in bone marrow and peripheral blood stem cell collections from patients with ovarian cancer. Bone Marrow Transplant 15, 929 933. Russell N.H., Gratwohl A. & Schmitz N. (1996) The place of blood stem cells in allogeneic transplantation. British Journal of Haematology 93, 747 753. Serke S., van Lessen A., Pardo I. & Huhn D. (1998) Selective susceptibility of CD34-expressing cells to acquire flow cytometric features of apoptosis/necrosis on exposure to an ammonium chloride-based red blood cell lysing reagent. Journal of Hematotherapy 7, 315 318. Siena S., Bregni M., Brando B. et al. (1991) Flow cytometry for the clinical estimation of circulating hematopoietic progenitors for autologous transplantation in cancer patients. Blood 77, 400 409. Sutherland D.R., Anderson L., Keeney M., Nayar R. & Chin-Yee I. (1996) The ISHAGE guidelines for CD34 + cell determination by flow cytometry. Journal of Hematotherapy 5, 213 226. Sutherland D.R. & Keating A. (1992) The CD34 antigen: structure, biology and potential clinical applications. Journal of Hematotherapy 1, 115 129. Sutherland D.R., Marsh J.C.W. & Davidson J. (1992) Differential sensitivity of CD34 epitopes to cleavage by Pasteurella haemolytica glycoprotease: implications for purification of CD34- positive progenitor cells. Experimental Hematology 20, 590 599. Takekawa K., Yamane T., Suzuki K., Hino M. & Tatsumi N. (1997) Identification of hematopoietic stem cells by the SE-9000 automated hematology analyzer in peripheral blood stem cell harvest samples. Acta Haematologica 98, 54 55. Verwer B.J.H. & Ward D.M. (1997) An automated classification algorithm for ProCount flow cytometric acquisition and analysis. Journal of Hematotherapy 6, 169. Wunder E., Sovalat H., Fritsch G., Silvestri F., Henon P.R. & Serke S. (1992) Report on the European Workshop on Peripheral Blood Stem Cell Determination and Standardization Mulhouse. France. Journal of Hematotherapy 1, 131 142.