Expression of cell-associated prion protein on normal human platelets - reply

Abstract

In reply to the letter by Starke et al that comments on our recent paper ( Barclay et al, 1999 ), there is general agreement (using flow cytometry) between ourselves, Starke et al and Holada et al (1998) that by far the major proportion of normal prion protein (PrP) in healthy adult human blood is found on platelets. This agrees with our findings using a DELFIA immunoassay system to study prion protein distribution in adult blood ( MacGregor et al, 1999 ). Starke et al comment on their lower values for estimation of platelet PrP expression compared with the higher, but similar, values found by us and by Holada et al (1998) for resting and activated platelet prion protein expression. They suggest that their lower values relate to their having used gentler methods without any washing steps, as recommended by the calibrant manufacturer (Biocytex), instead of the more aggressive washing methods used by ourselves and by Holada et al (1998) . We have all used similar methods for platelet studies based on the mouse IgG-coated bead standard from the Biocytex Platelet PrP kit. In our studies, we used this standard for platelets, but used the larger Quifikit (Dako) mouse IgG-coated bead as a standard for leucocyte studies. These beads are also manufactured by Biocytex. The fluorescent intensities that we found and the resultant calibration curves from the stated mouse IgG-coating densities for both bead standards were shown ( Barclay et al, 1999 ). We had intended to use only the Quifikit beads, running them through parallel stain and wash procedures as were used in our whole blood method for leucocyte and platelet prion staining. However, we had questions (relating to the stated method) on how to perform calculations to estimate cellular expression of prion protein from the comparative binding of fluorescence-labelled anti-mouse IgG to mouse IgG-coated beads and to mouse IgG monoclonal anti-PrP bound to different blood cell populations. Communications (personal) with Biocytex persuaded us to use their smaller Biocytex Platelet PrP kit bead standard instead of the Quifikit beads for platelet studies, but it was not emphasized to us that these could not be used in parallel to the stain, wash and lyse system used in our blood studies. The calculated values for the same studies of platelets using the two different standards are shown in Table I. In fact, the net PrP per cell values determined from the Quifikit bead standard, which we did not report, are closer to the values found by Starke et al from the Platelet PrP kit standard using a ‘no-wash’ technique. It appears that these quantitative measurements are very dependent on the behaviour of the bead standards under different assay conditions. If we accept that the Quifikit beads may offer more robust retention of mouse IgG label and/or bound fluorescence-conjugated anti-mouse antibody during a stain and wash technique, then the lower (Quifikit) estimated values may be a better reflection of true levels of PrP expression on human platelets. Starke et al make further comment on the possible reasons for the differences in findings and, although they do not themselves report any studies of PrP expression on leucocytes, they speculate that the high concentrations of primary anti-PrP antibody (monoclonal) that we used in our studies might result in high non-specific (background) labelling of leucocytes. High concentrations were used because a whole blood technique was utilized and we were aware of several sources of PrP in whole blood that might saturate anti-PrP binding. We have clearly stated in our publication (see Fig. 4, Barclay et al, 1999 ) that we have used different anti-PrP monoclonal antibodies in our study. The anti-PrP mAb DF7, that has specificity for a C-terminal proximal epitope of denatured PrP, was clearly shown by us to be indistinguishable from controls with no prestaining with anti-PrP and does not suggest any non-specific background staining. Quantitative calculations were based on 3F4 mAb staining with the appropriate background subtracted. Our initial studies suffered from non-specific or inappropriate binding of secondary (anti-mouse IgG) fluorescence-conjugated antibody. Reagents that we tried either bound apparently through Fc receptors on the target cells or through cross-reactivity with human immunoglobulins. Careful selection of the human-absorbed F(ab′)2 anti-mouse IgG fluorescence conjugate used in the reported study abolished these problems. We retained the use of this antibody in quantitative studies on beads and leucocytes as we were reluctant to use the anti-mouse IgG supplied with the bead standards. It is unlikely that this would have affected quantitative studies.