Quality assurance of CFU-GM assays: inter-laboratory variation despite standard reagents.

To investigate the hypothesis that commercial kits for CFU-GM (colony forming unit granulocyte-macrophage) assay will reduce the interlaboratory variation noted by many workers, we carried out a quality assurance exercise in 2 parts. There were 8 participants in the first study and each performed CFU-GM assays using their in-house method and a commercial kit (Stem Cell CFU Kit, Gibco) in parallel. In the second exercise there were 10 participants and each performed CFU-GM with in-house methods and with a different commercial medium (Methocult GF H4534, Stem Cell Technologies). Twelve samples of cryopreserved peripheral blood progenitor cells (PBPC) were analysed by each participant in each part of the study. A very wide range of results was found for the different in-house methods, but standardizing the clonogenic assay with the commercial kits did not reduce the variation seen. To improve the reproducibility of CFU-GM assays between laboratories, scrupulous attention should be paid to all the steps involved in the assays, as little progress will be made by using commercial medium in isolation from efforts to reduce other sources of variation.

Colony counting is a major source of variation in CFU-GM results between centres.

The results for colony forming unit granulocyte-macrophage (CFU-GM) assays vary substantially between centres. It is possible that colony counting is largely responsible for this discrepancy. In order to examine this exclusively from the many factors that make up the CFU-GM assay, we performed a colony counting exercise involving 11 laboratories. Two-way analysis of variance showed a highly significant difference (P = 0.0001) in the counts obtained from the centres. One centre was found to score consistently high and two others scored consistently low numbers of colonies. This suggests that identification of colonies is a major source of variation between centres.

Quality assurance of CD34+ cell estimation in leucapheresis products.

In order to examine the feasibility of an external quality assurance (QA) scheme for CD34+ cell enumeration and to identify causes of the differences between laboratories for CD34 counts, we carried out a pilot QA exercise in two parts. There were eight participating laboratories in the initial study and each performed CD34 counts using their in-house method. A series of 12 samples of cryopreserved peripheral blood progenitor cells (PBPC) were analysed by each of the eight laboratories. A very wide range of values for all the samples was found for the different in-house methods. For the second part of the study, 12 laboratories analysed a different set of 12 PBPC samples and each used the same anti-CD34 antibody (HPCA-2 PE), anti-CD45 antibodies to identify leucocytes, and counted a minimum of 50,000 events. These measures have reduced the interlaboratory variation in results but this variation is still too high to allow us to realistically compare values between centres. Overall, most centres performed comparably, but there was one centre in part one of the study which gave results that were significantly different from the other centres.

BIRMA-K3, a new monoclonal antibody for CD34 immunophenotyping and stem and progenitor cell assay.

A murine monoclonal antibody with specificity for the CD34 antigen has been produced and designated BIRMA-K3. The antibody characterized as IgG1(kappa) has been shown to react with KG-1a cells following treatment of the cells with glycoprotease enzyme, indicating reactivity with the class III epitope of CD34. It was possible to show that class I and class II anti-CD34 antibodies were not able to inhibit binding of BIRMA-K3. Investigation of FITC-labeled as well as PE-labeled BIRMA-K3 resulted in a clear cut-off staining of acute leukemias and CD34+ cell counts in patients submitted to high-dose chemotherapy and stem cell transplantation. The results obtained correlate strongly with those from HPCA-2, the Becton-Dickinson class III antibody.

High lactate dehydrogenase level is associated with an adverse outlook in autografting for Hodgkin's disease.

Forty-two patients with relapsed or refractory Hodgkin's disease (HD) were treated with high-dose chemotherapy (BEAM regimen) followed by autologous bone marrow and/or peripheral blood progenitor cell (PBPC) rescue. There was one procedure-related death and the overall response rate at 6 months was 88% (95% confidence interval 78-98%). The 2 year overall and event-free survival was 81% (95% confidence interval 65-96%) and 74% (95% confidence interval 58-89%) respectively. Median follow-up was 33 months. The use of PBPC instead of marrow resulted in a significant shortening of the time to engraftment (P < 0.01). Multivariate analysis identified the pre-transplant LDH level as a highly significant factor in predicting overall survival (P = 0.007). The BEAM regimen is an effective conditioning schedule that is well tolerated but patients with a raised LDH at the time of transplant remain at high risk of early relapse and death due to disease.

Influence of Gm allotype on the IgG subclass response to streptococcal M protein and outer membrane proteins of Moraxella catarrhalis.

The IgG antibody response to streptococcal M protein is distributed between the IgG1 and IgG3 subclasses, however individual sera vary with respect to the relative amounts of these two subclasses. The basis of this variation was investigated. Sera were also analysed for IgG subclass antibodies to the outer membrane proteins (OMP) of Moraxella catarrhalis, as these have also been reported to have a major IgG3 component. The mean percentage of IgG3 was higher in the antibody response to OMP and there was less variability between sera for this antigen than was seen for M protein. Non-specific binding of IgG3 in ELISA, which has been reported for some bacterial proteins (including M protein of some serotypes) was excluded as an explanation for the apparent IgG3 bias of these antibodies. The relative amount of IgG3 antibody to the two antigens showed a positive correlation, suggesting that some individuals tended to make a greater IgG3 response to unrelated antigens. Serial bleeds from two individuals maintained a relatively constant subclass profile over several months, suggesting that time since infection did not play a major role in determining the proportion of IgG1 and IgG3. Gm allotypes for the sera were determined, and found to correlate with both total serum IgG3 concentrations and with IgG subclass composition of specific antibodies. Mean serum IgG3 concentrations were highest in sera typed as Gm(fb/fb) homozygous and lowest in sera typed as Gm(ag/ag) homozygous. Similarly, in the M protein-specific antibodies, the mean percentage of IgG3 was much lower in the Gm(ag/ag) sera than in the Gm(fb/fb) homozygous sera. Sera which typed as Gm(fb/ag) heterozygous were not significantly different from the Gm(fb/fb) homozygous sera for either total serum IgG3 or for M protein-specific IgG3. Moreover, both Gm(fb/fb) homozygous and Gm(fb/ag) heterozygous sera included samples in which IgG1 was the predominant antibody subclass and the percentage of IgG3 was very low. In contrast to the M protein-specific antibodies, for the OMP-specific antibodies there was no correlation between Gm phenotype and the proportion of IgG3. The data suggest that Gm allotype may influence the IgG subclass composition of antibody responses to bacterial surface protein, but that other factors are also likely to be involved.

Human monoclonal anti-D secreting heterohybridomas from peripheral B lymphocytes expanded in the CD40 system.

Peripheral human B lymphocytes isolated from four immunised anti-D donors have been cultured in the CD40 system (Banchereau et al., 1991), prior to fusion to murine X63Ag8.653 plasmacytoma cells. High fusion efficiency was noted in all cases and immunoglobulin (IgG and IgM) was detected in nearly all heterohybrid containing wells. Only fusions using boosted donor lymphocytes generated specific anti-D secreting hybrids. A short period of culture of committed B cell precursors in the CD40 system before fusion appears to be both adequate and efficient in permitting generation of specific antibody-secreting hybrids.

A murine monoclonal anti-N, BIRMA-N, suitable for blood grouping in an automated system.

BIRMA-N, a murine monoclonal anti-N antibody of the IgG1 subclass, was assessed for suitability as a blood grouping reagent on the Olympus PK7100 automated blood grouping machine. At a selected dilution and over the pH range 5.0-8.0, the antibody performed accurately in this system as confirmed by parallel manual testing of donor blood samples with human anti-N and commercial monoclonal anti-N reagents. These findings show BIRMA-N to be extremely suitable for N typing blood samples in an automated system providing a convenient, objective and cost-effective method for large scale typing of the blood donor population.

A new monoclonal anti-A antibody BIRMA-1. A potent culture supernatant which agglutinates Ax cells, but does not give undesirable reactions with B cells.

A monoclonal anti-A antibody, BIRMA-1, has been evaluated and found to be eminently suitable as a blood-grouping reagent. The culture supernatant is potent, avid and specific as demonstrated by its reactivity with a large sample of donor bloods tested manually and on the Olympus PK 7100 automated blood grouping machine. Over 100 cord blood samples were tested manually, and all were correctly grouped using this antibody. No false-positive reactions were obtained with papainized group B cells. Reagent prepared from BIRMA-1 detected most examples of Ax and all other A subgroups and A variants tested. Negative reactions with B(A) cells indicate that it is possible to have an antibody capable of detecting Ax which will not react with B cells from individuals with high levels of galactosyltransferase.

A histologic comparison of aseptic loosening of cemented, press-fit, and biologic ingrowth prostheses.

The histology of interface membranes from aseptic loosened prostheses of various types including cemented, press-fit, and biologic ingrowth varieties was compared. Pseudosynovial implant-facing surfaces were present in specimens from all types. The remaining portions of these membranes showed distinct characteristics as well. Cemented implant membranes contained many macrophages and giant cells and evidenced frequent granuloma formation. Press-fit membranes consisted of poorly vascularized, dense fibrous tissue within the loosened press-fit membrane. Macrophages and giant cells were rare, except in one specimen containing ceramic debris particles. Biologic ingrowth membranes were the most vascular and contained loosely organized connective tissue and islands of woven bone. Macrophages were common. One out of six specimens from patients with rheumatoid arthritis contained massive numbers of lymphocytes and plasma cells but not mast cells. The greatest numbers of mast cells were present in membranes from patients with osteoarthritis and in all cases were associated with the presence of stainless steel and/or chrome cobalt particles.

Histochemical localization of cathepsin B, dipeptidyl peptidase I, and dipeptidyl peptidase II in rat bone.

The histochemical distribution of the thiol proteases cathepsin B and dipeptidyl peptidase I and the serine protease dipeptidyl peptidase II was examined in rat bone and joint using amino acid derivatives of 4-methoxy-2-naphthylamine (MNA). The liberated MNA was then visualized by simultaneous coupling with fast blue B. Cathepsin B was examined with CBZ-Arg-Arg-MNA, dipeptidyl peptidase I (DPP I) with Gly-Arg- or Pro-Arg-MNA, and dipeptidyl peptidase II (DPP II) with Lys-ALA- or Lys-Pro-MNA. Bright red reaction product indicative of proteolytic activity was observed in most cell types associated with bone and its surrounding connective tissues, including osteocytes, osteoblasts, chondrocytes, chondroblasts, fibroblasts, and macrophages. Surprisingly, protease activity in osteoclasts could not be established with certainty, and it was concluded that these enzymes are either absent, present in very low amounts, or secreted as soon as they are synthesized rather than stored within the cell. The cells of the resting zone of the growth plate were intensely reactive for DPP II but were only moderately reactive for cathepsin B and DPP I. The reverse was true of the proliferating and hypertrophic layers. The protease activity observed in bone, cartilage, tendon, ligament, and synovium would be expected to contribute significantly to normal protein metabolism as well as to pathological destruction in these tissues.

Establishment of a large collection of extracted teeth for research.

A collection of over 14,000 teeth extracted at the Prince Philip Dental Hospital since 1982 has been organized and catalogued on a computerized database management system. The computer catalogue provides, for each tooth in the collection, information on the age and sex of the patient, and the date of extraction and condition of the tooth. The catalogue can be searched according to any combination of the descriptive variables in the database record. Researchers, including visiting scientists, can borrow teeth from the central "tooth library" on a temporary or permanent basis. Further information on particular teeth (e.g. patient's medical and dental histories, dental radiographs) can be obtained from patients' charts. Establishment of this collection has greatly facilitated the work of researchers in clinical dentistry, dental anatomy, and dental anthropology.

Direct bone resorbing activity of murine myeloma cells.

The cellular mechanism(s) responsible for tumor-associated bone resorption in multiple myeloma remain uncertain. Both in vivo and in vitro evidence is presented for the direct resorption of bone by mouse plasmacytomas. Morphological examination of autopsy specimens from tumor-bearing mice revealed in vivo erosion of bony surfaces at sites of tumor cell-bone matrix apposition. No osteoclastic bone resorptive activity was evident. Using a 45Ca-labelled, devitalized bone explant assay system, mouse myeloma cells caused the release of isotope at levels from 200-300% above control values. Control cells such as normal spleen lymphocytes and liver cells did not resorb bone. Demonstration of the ability of myeloma cells to independently destroy bone is important to the understanding of the causes of and development of chemotherapeutic approaches to myelomatous bone resorption.

A comparative study of new substrates for the histochemical demonstration of acid phosphomonoesterase activity in tissues which secrete acid phosphatase.

The histochemical demonstration of acid phosphatase activities against phosphoethanolamine (PEA), phosphorylcholine (PC), and D-ephedrine phosphate (DEP) are reported for a variety of rat tissues and are compared to acid beta-glycerophosphatase (beta GPase) activity. Intense acid beta GPase activity was demonstrated in all tissues examined. However, liver, kidney, intestine, spleen and bone marrow cells failed to exhibit any enzyme activity against PEA, PC, or DEP. In addition, significant differences in the hydrolysis of these substrates were noted among the tissues that did demonstrate activity (bone, tooth, oral mucosa, sebaceous gland, and prostate gland). These observations suggest that PEA, PC, and DEP are more specific substrates for acid phosphatase than beta GP and permit the differential localization of several distinct acid phosphatase isoenzymes.

Selective localization of a Golgi apparatus acid phosphatase isoenzyme in bone using pyridoxal-5'-phosphate.

Substrates commonly used for localizing bone Golgi apparatus (GA) acid phosphatase (AcPase), e.g., beta-glycerophosphate, p-nitrophenylphosphate, cytidine-5'-monophosphate, and di(dicyclohexylammonium)-2-napthylthiolphosphate, give strong staining not only of GA but also of lysosomes. Thiamine pyrophosphate and inosine-5'-monophosphate--substrates that give good GA staining in some soft tissues--give only lysosomal staining in bone. No previously used substrate or substrate-effector combination has selectively localized the GA acid phosphatase in bone. This article describes results using a new AcPase medium having pyridoxal-5'-phosphate (PLP) as substrate. In bone this medium produced strong staining of the osteoblast GA, but relatively little staining of lysosomes, including lysosomes in osteoclasts. The weak lysosomal staining was almost totally eliminated, without affecting the GA reaction, by pretreating the tissue in 0.3% NH3 solution. Conversely, elevated ionic strength of the substrate medium eliminated the GA reaction, while having little effect on lysosomal staining. The GA enzyme was very sensitive to 1 mM tartrate whereas the lysosomal enzyme was not. These differences suggest the presence of distinct isoenzymes in the two locations. The distribution of osteoblasts with stained GA coincided with the distribution of strongest alkaline phosphatase activity and rapid bone mineralization, supporting previous suggestions that osteoblast GA AcPase is involved in the processing of one or more newly synthesized bone matrix components.