We have read with great interest the article by Edwards and colleagues, recently published in Clinical Cancer Research (1). However, we have concerns about the conclusions on GITR expression in tumor-infiltrating cells.
Figure 1D of the article depicts the expression of T-cell markers in the melanoma microenvironment. The images reveal that ICOS, PD-1, and GITR are mainly expressed by T cells in all or the majority of the patients, and OX40 and Tim-3 are mainly expressed by T cells in less than half of the patients. On the basis of the information in Fig. 1C, the markers can be divided into two groups: (i) GITR, OX40, and Tim-3 (expressed by a low number of T cells, e.g., regulatory T cells) and (ii) ICOS and PD-1 (expressed by a high number of T cells). According to the data in Figure 3B, this is true for all markers, with the exception of GITR, which shows the highest expression levels in monocytes and B cells (in disagreement with the data in Figure 1D) and is not expressed at all by regulatory T cells (in disagreement with the literature; ref. 2).
We have evidence that the anti-GITR antibody clone 621 (621 mAb), which was used in the experiments in Figure 3B, but not those in Figure 1 of the Edwards and colleagues' study, does not bind GITR, or, at least, its principal isoform. Indeed, a cell line transfected with human GITR (NP_004186.1) was confirmed to overexpress GITR by the anti-GITR clone DT5D3 mAb (Miltenyi Biotec) and not by 621 mAb (Figure 1A shown here). On the contrary, comparison of GITR staining by the two mAbs in purified CD4+ T cells (Figure 1B shown here) may only indicate that 621 mAb has a relatively low affinity to GITR.
In conclusion, Figure 1A shown here and Figure 1D of the Edwards and colleagues manuscript contradict the data published by the Edwards and colleagues study in Figure 3B. Thus, we propose that the anti-GITR 621 mAb is not reliable for assessing GITR expression, as suggested by another group (3). It is unclear whether this is a result of the aspecific binding of 621 mAb or its specific binding to another protein or to shedded GITR bound to its transmembrane ligand. Nonetheless, we are of the opinion that the conclusions drawn by Edwards and colleagues and other similar studies using this antibody (such as refs. 3–5) must be reconsidered.
See the Response, p. 5424
Disclosure of Potential Conflicts of Interest
G. Nocentini is a consultant for Pieris Pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.
C. Riccardi was supported by the Ministero dell'Istruzione, dell'Università e della Ricerca (MIUR) under grant number 2015ZT9MXY_001