Lack of Foxp3⁺ macrophages in both untreated and B16 melanoma-bearing mice

To the editor:

Foxp3⁺ Tregs are essential for maintaining immune tolerance in mice and men. Except for expression in a minor population of CD8⁺ T cells,¹  Foxp3 is currently believed to be restricted to CD4⁺ Tregs in mice, because widespread nonhematopoietic Foxp3 expression and its putative implication in tolerance have been refuted.^2,3  Thus, the employment of Foxp3 reporter mice, including Foxp3^DTR-eGFP (DEREG) mice,⁴  has been valuable for the investigation of Foxp3⁺ Treg biology. Recently, we demonstrated that the depletion of Foxp3⁺ Tregs unleashes potent therapeutic tumor-specific immunity,^5,6  unlike less selective agents, eg, targeting CD25. The recent postulation of immunoregulatory Foxp3⁺ macrophages⁷  implies that macrophage deletion in DEREG mice could contribute to the striking effects of Foxp3⁺ cell ablation as cancer immunotherapy. We thus specifically investigated this issue. Anti–Foxp3-PE stainings of WT spleens revealed a CD11b⁻Foxp3⁺ Treg population as expected, whereas the CD11b⁺ fraction contained a clearly weaker PE signal (Figure 1A). The latter was identified as autofluorescence since it persisted in isotype stainings (Figure 1A) or unstained samples (not shown), in contrast to CD11b⁻Foxp3⁺ Tregs. To further assess if the CD11b⁺PE^low population actively transcribes the foxp3 locus, we analyzed spleens of DT-treated DEREG mice. Strikingly, CD11b⁺PE^low cells persisted on DT administration, whereas CD11b⁻Foxp3⁺ Tregs were depleted (Figure 1A), suggesting that the autofluorescent CD11b⁺ population does not express Foxp3^DTR-eGFP. Similar results were obtained with BM (not shown). In line with these findings, CD11b⁺PE^low cells were present at rather increased percentages in the BM of scurfy mice harbouring a mutated nondetectable Foxp3 protein,⁸  compared with Foxp3-sufficient litters (Figure 1B). In contrast, CD11b⁻Foxp3⁺ Tregs were absent from scurfy mice as expected, demonstrating a lack of Foxp3 staining in CD11b⁺ cells (Figure 1B). Because Manrique et al used the FITC/GFP channel for revealing Foxp3 in their study,⁷  which is more prone to autofluorescence than the PE channel, we next assessed the deletion of Foxp3^DTR-eGFP+ cells in DT-treated DEREG mice. Approximately 4% of T cell–excluded BM cells in WT mice displayed strong autofluorescence (Figure 1C), comparable with the reported frequency⁷ . DEREG mice did not harbor increased frequencies of GFP/autofluorescence⁺ cells within the T cell–excluded CD11b⁺ BM fraction (Figure 1C), which would be expected from Foxp3-reporting cells. Similarly, DT administration to DEREG mice did not result in depletion of the GFP/autofluorescence⁺ population (Figure 1C). DT efficiently accessed the BM because GFP⁺ Tregs were eliminated (not shown). Of note, autofluorescent CD11b⁺ cells are FSC^hiSSC^hi compared with nonautofluorescent CD11b⁺ cells and T cells (not shown). Consistent with Manrique et al,⁷  we readily detected CD11b⁺ autofluorescent cells in spleens and BM of Rag KO mice, but those similarly lacked Foxp3 expression (not shown). We next generated BM-derived macrophages (BMM) from WT or DEREG mice in presence or absence of DT during the culture period. Again, we could not detect Foxp3 staining in CD11b⁺ BMM, and no increased rate of cell death was observed after culture in presence of DT (Figure 1D and not shown). LPS stimulation—which was implicated in de novo Foxp3 expression by Foxp3⁻ macrophages⁷ —did not induce Foxp3 expression in BMM (not shown). The authors demonstrated a regulatory role for putative Foxp3⁺ macrophages in B16 melanoma.⁷  To exclude that Foxp3 is induced in macrophages by melanoma-dependent mechanisms, we analyzed B16-OVA melanoma-bearing mice. Spleens and BM lacked CD11b⁺Foxp3⁺ cells (not shown), as did tumor-free mice (Figure 1A-C). Furthermore, single cell suspensions of tumors raised in WT mice revealed no specific Foxp3 expression in CD11b⁺ cells, whereas CD11b⁻Foxp3⁺ Tregs were present (Figure 1E), in line with previous results.⁵  Similarly, immunofluorescence analyses of tumor sections revealed that Foxp3⁺ cells were exclusively CD3⁺, but did not coexpress F4/80 or CD11b, whereas Foxp3⁺ cells were completely absent from tumors elicited in Rag KO mice (Figure 1F, not shown).

Concluding, we found no evidence for Foxp3 expression in macrophages by combining flow cytometry, immunofluorescence microscopy and genetic approaches. This is fully in line with the previously noted lack of Foxp3 protein expression in Rag KO spleens.²  Thus, macrophage subpopulations are not targeted in genetically engineered DEREG mice and do not contribute to the onset of self/tumor-specific immunity upon Foxp3⁺ Treg ablation. These results are of immediate relevance for the use of Foxp3 reporter mice and for the translational design of novel cancer immunotherapies.