| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
OPEN ACCESS ARTICLE
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
letter |
aUniversity Health Network, Toronto General Hospital, Toronto, Ontario, Canada;
bLawson Health Research Institute, London Health Sciences Centre, London, Ontario, Canada
Key Words. Bone marrow transplantation • CD34 selection • Flow cytometry • Monoclonal antibodies
Correspondence: D. Robert Sutherland, M.Sc., University Health Network, Toronto General Hospital, 200 Elizabeth Street, Toronto, Ontario M5G 2C4, Canada. Telephone: 416-340-5443; Fax: 416-340-5543; e-mail: rob.sutherland{at}utoronto.ca
Received on April 2, 2007;
accepted for publication on June 5, 2007.
First published online in STEM CELLS EXPRESS June 14, 2007.
The recent publication by Otto et al. [1] raises some important issues for the analysis of rare (CD34+) cell populations by flow cytometric methods (FCM). Identification and enumeration of CD34+ cells by FCM provide a major challenge for both clinical and research flow laboratories. Standardized, validated, multiparameter protocols have been developed over the last 10 years to accurately detect and enumerate viable CD34+ cells in all sources of hematopoietic stem cells [2–5] and are in widespread clinical use. However, many research flow laboratories still rely on simple single-parameter CD34 staining to identify CD34+ cells and misguidedly use isotype controls [6] that are not robust to delineate unstained cells from "CD34+ events," although specific information on how bona fide CD34+ cells were separated from other events in this study was not provided.
By way of background to some of these issues, the CD34 molecule itself is a mucin-like structure in which the carbohydrate moieties probably play a fundamental role in determining the fine functional characteristics of the various glycoforms [7, 8]. As the authors note, there are three broad classes of epitopes recognized by the various CD34 monoclonal antibodies, classified by their sensitivities to neuraminidase and the O-sialoglycoprotease from Pasteurella hemolytica [9]. Chymopapain was not used in our initial definition of the 3-class epitope classification scheme that utilized biochemical as well as serologic/FCM techniques, and is of little if any value in epitope mapping studies of the CD34 molecule [8]. Regardless, the simple epitope classification scheme has been very useful for selecting the best CD34 antibody for a specific application, and is still in use today. For example, because of their dependence on carbohydrate moieties (sialic acids), their consequent inability to detect all glycoforms of the CD34 antigen, their generally lower avidity, and their general inability to retain reactivity after conjugation with certain charged fluorochromes, class I antibodies are not recommended for use in immunodiagnosis or the enumeration of CD34+ cells in the transplant setting. However, the epitope detected by the class I CD34 antibody 12.8 is known to be located in the far N terminus of the molecule (reviewed in [8]) and ironically was the first antibody to demonstrate that infusing positively-selected CD34+ cells using the CellPro device (Baxter Healthcare, Deerfield, IL, http://www.baxter.com) was both clinically safe and potentially therapeutically useful [9, 10]. Unfortunately, in the paper of Otto et al., it is difficult to discern any specific staining of the thymic CD34+ cells with the 12.8-based cocktail shown in Figure 1A. How was this reagent combination validated prior to its use in this study?
With regard to QBEnd10, the class II epitope it detects is insensitive to denaturation and, consequently, this high affinity antibody is the reagent of choice for certain research applications and clinical immunohistochemical protocols using paraffin-embedded tissues. Furthermore, the class II epitope has been determined to be a linear stretch of seven amino acids at positions 10 through 16 at the far N terminus of this "flag-pole"-like structure. This short sequence is flanked on each side by threonine residues that are almost certainly conjugated with sialylated O-linked glycans and which are probably involved in the formation of the sialic acid-dependent class I epitopes detected by 12.8 and others. It is perhaps because the class II epitope is invariant and is located at the far N terminus of the CD34 molecule that antibodies to this epitope have been successfully used in a variety of commercial reagent kits for CD34+ cell purification schemes [10, 11], the best known of which are the QBEnd10-based CliniMACS and the 9C5-based Isolex clinical cell separators (Miltenyi Biotech Inc., Auburn, CA, http://www.miltenyibiotech.com).
As part of our validation of suitable CD34 antibody clones/conjugates for deployment in multiparameter-with-Boolean-gating International Society for Hematotherapy and Graft Engineering (ISHAGE)-type methodologies to enumerate CD34+ cells in the transplant setting [2–5], we found that phycoerythrin-conjugates of class III antibodies such as 8G12 and 581 and the class II antibody QBEnd10 detected virtually identical numbers of CD34+ cells in the same samples in the same day. Fluorescein-conjugated versions of the class III antibodies similarly generated nearly identical numbers on the same samples. However, fluorescein isothiocyanate (FITC)-conjugated versions of QBEnd10 (we have personally tested approximately 15 of them from a variety of vendors) detected markedly fewer CD34+ cells than either the phycoerythrin (PE) conjugate of the same antibody clone or the PE- or FITC-conjugated versions of the class III reagents [2, 8]. The explanation for this observation is that FITC confers negative charge on the QBEnd10 antibody that reduces its ability to bind to its epitope, surrounded as it is by negatively charged, sialylated O-linked glycans. Prior treatment of the cells with neuraminidase corrected this anomaly. A significant number of studies have been published that unfortunately failed to take such basic science of CD34 epitopes into account to conclude that class III epitopes are expressed more widely across the CD34+ cell compartment than class II epitopes. Invariably, such studies, and several were cited by Otto et al., combined single parameter staining methods with isotype controls and unsophisticated gating strategies and, in most cases, inappropriately utilized FITC conjugates of QBEnd10.
These observations underscore the importance not only of selecting CD34 clones that detect all glycoforms of CD34 (i.e., class II or class III antibodies) but also of selecting only validated and properly titrated conjugates of such clones. The only way to do this for the accurate analysis of rare events (like thymic CD34+ cells) is to employ a state-of-the-art, multiparameter assay that can specifically identify bona fide CD34+ cells. In the Otto study [1], similarly to the situation with the 12.8 detection cocktail (Fig. 1A), there is no evidence of any specific staining with the QBEnd10Cy5 conjugate employed (Fig. 1B). Again, how was this conjugate tested and validated for the assay? Furthermore, the two class III conjugates are different again, with allophycocyanin conjugates of 8G12 and ECD conjugates of 581 used in Figure 1C and 1D, respectively. Thus, the authors assessed the expression of three different CD34 epitopes with four different antibody clones, each of which was conjugated to a different fluorochrome (or biotin)! Of note, the "CD34+" fraction on Figure 1C appears to be present on the "bright" CD7+ cells, whereas they are present on CD7 "dim" cells in Figure 1D. Ironically, both CD34 clones (Fig. 1C, 8G12; Fig. 1D, 581) detect class III epitopes and have never shown a significant discrepancy in our hands. To properly evaluate the impact of CD34 epitope class on CD34 selection, the authors need to stain all samples with appropriately validated and titrated antibodies conjugated with the same fluorochrome (preferably PE) in a multiparameter gating strategy that includes CD45 and takes into account the latter's staining intensity [2–5].
Another point we find difficult to understand concerns the data in Table 1. The 8G12 antibody stains more CD34+ cells than QBEnd10, as claimed by the authors. However, when looking at cells dual-stained by both 8G12 and QBEnd10, levels appear very low, suggesting that the cells stained by QBEnd10 are not stained by 8G12 (particularly for patients 9, 10, and 12). It has been our experience that CD34+ cells from samples selected by QBEnd10 are all stained with 8G12 and/or 581. Based on the data presented, the authors' suggestion that there may be clinical relevance to their findings is premature at best and flies in the face of the hundreds of successful transplants performed on CD34+ cell infusates purified by class I-based (CellPro/Baxter) and class II-based CliniMACS (Miltenyi Biotec) and Isolex (Baxter) devices.
 |
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST |
|---|
 Top Disclosure of Potential... References |
|---|
|
REFERENCES |
|---|
|
TopDisclosure of Potential...References |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Barfield, X. Chen, M. Otto, and R. Handgretinger Beyond the FACS Stem Cells, November 1, 2007; 25(11): 2972 - 2972. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| STEM CELLS | THE ONCOLOGIST | CME | ALPHAMED PRESS JOURNALS |
