Comment on “Aberrant Regulation of Synovial T Cell Activation by Soluble Costimulatory Molecules in Rheumatoid Arthritis”

We have read with interest the article by Wan et al. (1) concerning the detection of a soluble form of PD-1. In relation to our recent identification of alternatively spliced PD-1 mRNA transcripts (2), we have ourselves been working on developing an ELISA to measure putative soluble forms of PD-1. Using the exact same combination of primary Abs, in our hands, the results were not valid because the primary Abs reacted with purified IgG. We were therefore wondering whether Wan et al. had tested reactivity with human IgG in their ELISA.

The first preliminary ELISA results were promising with detectable OD signals in serum, and we obtained a linear standard curve using diluted series of the PD-1.Fc protein (R&D Systems)—similar to Wan et al. In the particular setup, a polyclonal goat anti-human PD-1 (AF1086; R&D Systems) was used as the capture Ab (like Wan et al.) and a mouse monoclonal anti-human PD-1 (clone MIH4; BD Pharmingen). Polyclonal goat anti-mouse IgG/HRP (DakoCytomation) was added as secondary detection Ab.

As the polyclonal goat anti-human PD-1 (AF1086; R&D Systems) was raised against recombinant PD-1 fused with an Fc region, we wanted to exclude the possibility that IgG was wrongfully detected in the assay. To eliminate this possibility, we tested a dilution series of purified human IgG1. Unfortunately, this was actually the case. Using diluted series of purified human IgG1 (500 ng/ml to 1.95 ng/ml) we obtained a linear standard curve in this range with comparable OD levels. Our interpretation is that the polyclonal goat anti-human PD-1 Ab from R&D Systems was produced in goats immunized with purified the PD-1.Fc protein, so the epitopes recognized by this Ab could be located in the IgG part (human IgG1 (Pro¹⁰⁰–Lys³³⁰)) of the PD-1 fusion protein as well.

To overcome this problem, we conjugated the monoclonal secondary Ab to biotin and used this in the ELISA together with strepavidin-peroxidase (DakoCytomation) as detection substrate.

Again, the OD signals were measurable in serum, and we obtained a linear standard curve using diluted series of the PD-1.Fc using this new sandwich ELISA assay. Surprisingly and very unfortunately, we obtained comparable OD levels using purified human IgG1. We have at present no explanation for this observation, but in our hands, the current use of these particular Abs cannot be used for measuring soluble PD-1.

As Wan et al. (1) have not included human IgG as control in their sPD-1 ELISA; there is a possibility that the results could be biased and correlate with the concentration of IgG in their samples.

• Copyright © 2007 by The American Association of Immunologists