Making sense of Foxp3 as a prognostic value in tumors

A transcription factor Foxp3 controls functionality of T cell subset called regulatory T cells (T_regs). Seminal studies in mice clearly showed that Foxp3⁺ T cells play a dominant role in suppressing excessive immune response.

But human studies introduced an additional confusion in T_regs field. Unlike, mouse Foxp3, human Foxp3 is up-regulated in human naive CD4 T cells upon activation. In fact, human Foxp3⁺ T cell population is divided into Foxp3^low and Foxp3^high population. Foxp3^low population lack regulatory properties, whereas Foxp3^high population represents a bona fide T_regs cells.

However, such separation of Foxp3 marker within individual tumor clinical samples are not always possible and ordinarily cancer biopsies are grouped into Foxp3^hi or Foxp3^lo for Foxp3 mRNA total expression levels that are above or below the [arbitrary] median. So, it is not surprising that [total] expression level Foxp3 biomarker in tumor tissue does not correlate with overall survival.

In this regard, it is interesting to read a new paper in Nature Medicine showing that inclusion of two additional biomarkers, IL-12 and TGFβ1, could provide that differentiating power to transform Foxp3 expression into a prognostic factor in cancer diagnostics.

For this study the authors analyzed colorectal cancer samples. They found that cancer samples could be grouped into A and B types based on presence of two separate Foxp3⁺ T cell population: Foxp3^low and Foxp3^high population. Type A samples mostly have Foxp3^high population whereas type B samples have both Foxp3^low and Foxp3^high populations.

As expected Foxp3^low T cell population lacked suppressive [cell proliferation inhibition] quality and most likely represent conventional activated T cells.

Gene expression analysis of type A and B cancer tissue showed that type B were enriched with several cytokines such as IL-12, TNF-α and also TGFβ1. In vitro Foxp3 conversion assay showed that when combined with TGFβ1, interleukin-12 [but not TNF-α] could specifically increase naive T cell differentiation into Foxp3^low population (most likely by preventing their differentiation into Foxp3^high population).

Interestingly, when cancer samples were classified based on mRNA expression levels of Foxp3, IL-12 and TGFβ1, there were clear differences in patients survival between groups:

(1) as expected Foxp3 expression by itself had no prognostic value

(2) in samples with IL-12^loTGFβ1^lo expression, mRNA Foxp3^hi was associated with poor prognosis

and (3) unexpectedly, in samples with IL-12^hiTGFβ1^hi expression, mRNA Foxp3^hi was associated with a better prognosis.

Finally, the authors find that cancer samples with IL-12^hiTGFβ1^hi expression harbored more of gut microbe Fusobacterium nucleatum.

In summary, this study indicates following scenario:

(a) in colorectal cancers, detection of mRNA Foxp3 by itself has no prognostic value.
(b) however, cancer biopsies with Foxp3^hiIL-12^loTGFβ1^lo mRNA expression could mean poor prognosis, because such samples contain mostly bona fide T_regs cells.
(c) but, cancer biopsies with Foxp3^lo/hiIL-12^hiTGFβ1^hi mRNA expression could mean good prognosis, because such samples contain mostly Foxp3^low population or it has sufficient inflammatory milieu to override any suppression by bona fide T_regs cells.

David Usharauli