STAT1 signaling shields T cells from NK cell-mediated cytotoxicity

The JAK-STAT pathway critically regulates T-cell differentiation, and STAT1 is postulated to regulate several immune-mediated diseases by inducing proinflammatory subsets. Here we show that STAT1 enables CD4+ T-cell-mediated intestinal inflammation by protecting them from natural killer (NK) cell-mediated elimination. Stat1−/− T cells fail to expand and establish colitis in lymphopenic mice. This defect is not fully recapitulated by the combinatorial loss of type I and II IFN signaling. Mechanistically, Stat1−/− T cells have reduced expression of Nlrc5 and multiple MHC class I molecules that serve to protect cells from NK cell-mediated killing. Consequently, the depletion of NK cells significantly rescues the survival and spontaneous proliferation of Stat1−/− T cells, and restores their ability to induce colitis in adoptive transfer mouse models. Stat1−/− mice however have normal CD4+ T cell numbers as innate STAT1 signaling is required for their elimination. Overall, our findings reveal a critical perspective on JAK-STAT1 signaling that might apply to multiple inflammatory diseases.

1. The data shown in the manuscript are obtained through intraperitoneal transfer of 1x106 cells of unfractionated CD4+ T cells into RAG-deficient mice, in which homeostatic proliferation of the transferred cells is thought to take place. Cytokines such as IL-7 and IL-15 are indispensable for such homeostatic expansion of naïve and memory T cells, and Stat1 is one of the signaling molecules downstream of IL-15 receptor. In addition, a recent report suggested that T cells undergoing lymphopenia-induced proliferation would upregulate Stat1 protein, which would trigger an alternate IL-7-dependent program. Therefore, it is possible that failed expansion of Stat1-deficient T cells may be due to lack of IL-7 or IL-15 signaling, rather than failure in NK cell-mediated elimination of the T cells. The authors should conduct more experiments to challenge this kind of criticism. 2. How long does it take to get a full reconstitution by transferred T cells? Is it possible to obtain similar results by using tamoxifen-induced deletion of Stat1 gene after full reconstitution of T cells? Again, it is not clear to me whether the authors really observe what they are arguing in the manuscript. 3. If a marked reduction of Stat1-deficient T cells was observed even in the absence of colonic inflammation, the title of the manuscript is misleading. Colonic inflammation is just a result of manipulating animals under the specific experimental setting. 4. What are the physiological or pathological relevance of the findings? An experiment such as transfer of irrelevant of T cells into RAG-deficient mice should be considered.

Point by Point Critique:
Essentials from the editor: "We ask that you include additional empirical data to test the effects of STAT1 deficiency on CD4 T homeostatic expansion in the context of adoptive cell transfer colitis, as requested by referee #2. We also ask the other raised points by both referees to be addressed." We agree with the editor and reviewer 2's comments and have tested the role of STAT1 deficiency in homeostatic expansion. The data is provided in the manuscript and is explained in reviewer 2's comment below.  2b. The effect of IL-10 receptor-deficient NK cells on STAT1-deficient CD4 T cells was confirmed, so the role of IL-10 in NK cells should be discussed in the discussion section. We have updated the discussion section accordingly.

We agree with the reviewer that macrophages can interact with Stat1 -/-T cells, as it has been shown previously that they can phagocytose β 2m -/tumors (Barkal et. al. Nat Immunol 2017), and have
preliminary data addressing this concern in the 3 rd comment below.

Does the STAT1 signal in the T cell affect the action of other cells besides the NK cell?
The value recovered through anti-NK 1.1 treatment is not so large, and other roles through STAT1 in T cell may play a role.

We agree with the reviewer that other cell types might play a role in the elimination of Stat1 -/-T cells as we do not observe a complete rescue of these cells compared to WT in NK depleted mice. We have done preliminary experiments depleting several cell types in addition to NK cells and transferring CellTrace Violet (CTV) labelled Stat1 -/-T cells to see if we can get further enhancement of its expansion 2 weeks after T cell transfer.
We explored the role of macrophages (explained in the 2 nd comment above) by using an anti-CSF1R blocking antibody, which effectively reduced splenic red pulp macrophages (Fig. R2a). Compared to NK-depleted Rag1 -/mice that received the control antibody, the depletion of macrophages did not lead to an increase in T cell levels (Fig. R2b).

Il2rg
-/mice were cohoused for at least 2 weeks with Rag1 -/mice to control for the microbiota. Macrophage depletion was achieved by i.p. injection of 400μg Isotype or anti-CSF1R antibody (BioXCell) twice weekly beginning 1 week before T cell transfer. NK depletion was achieved by i.p. injection of 400μg anti-NK1.1 antibody into Rag1 -/mice 1 day before T cell transfer, as well as 200μg 1 week after to maintain depletion.
We also explored the role of ILCs using Rag2 -/-Il2rg -/mice, which naturally also do not have NK cells. We also do not see an enhancement of Stat1 -/-T cell expansion in this setting (Fig. R2b). However, the CTV profile suggested that the T cells display less proliferation compared to the NK-deplete Rag1 -/mice. While this suggests that ILCs might promote the expansion of Stat1 -/-T cells, further work is needed to address other factors such as γ c signaling on innate immune cells and stromal cells, as well as slight differences in the genetic background (Rag2 -/-Il2rg -/mice are B6;10 vs Rag1 -/which are B6).
All in all, these data suggest that CSF1R-dependent macrophages and ILCs are not the reason behind the incomplete rescue of Stat1 -/-T cells. It is possible that other cell types that we've not looked at mediate this difference. It is also possible that in addition to protecting T cells from NK cell killing, STAT1 signaling itself might promote T cell proliferation in vivo downstream of cytokines other than Type I+II IFN, such as IL-6, which is discussed in the manuscript.

The authors have demonstrated that Stat1 and associated expression of MHC class 1 molecules would protect CD4+ T cells from elimination by natural killer cells, leading to intestinal inflammation.
Although the manuscript is potentially interesting, it remains unclear whether the deletion of Stat1 gene actually affects the survival of CD4+ T cells, since an effect of Stat1 may be directed toward both NK cell-mediated T cell deletion and homeostatic proliferation.

The data shown in the manuscript are obtained through intraperitoneal transfer of 1x106 cells of unfractionated CD4+ T cells into RAG-deficient mice, in which homeostatic proliferation of the transferred cells is thought to take place. Cytokines such as IL-7 and IL-15 are indispensable for such homeostatic expansion of naïve and memory T cells, and Stat1 is one of the signaling molecules downstream of IL-15 receptor. In addition, a recent report suggested that T cells undergoing lymphopenia-induced proliferation would upregulate Stat1 protein, which would trigger an alternate IL-7-dependent program. Therefore, it is possible that failed expansion of Stat1-deficient T cells may be due to lack of IL-7 or IL-15 signaling, rather than failure in NK cellmediated elimination of the T cells. The authors should conduct more experiments to challenge this kind of criticism.
We thank the reviewer for this important comment, and provide data addressing this issue that is in the updated manuscript (Figure 7). The data is briefly explained below.

We note that the degree of rescue of SP in Stat1 -/-T cells is incomplete, which correlates with the incomplete rescue of T cell levels and the colitis in Il10rb -/-Rag1 -/mice. We are unclear as to the reason behind this. It is possible that there might be other cell types involved in clearing Stat1 -/-T cells, although NK cell depletion is sufficient to reverse the cell death phenotype in Stat1 -/-T cells.
Alternatively, IL-6, which drives SP, also activates STAT1, but this might not be the sole factor as a recent report showed that IL-6 regulates T cell expansion only during inflammation. We refer the reviewer to the manuscript for a more detailed discussion.

How long does it take to get a full reconstitution by transferred T cells? Is it possible to obtain similar results by using tamoxifen-induced deletion of Stat1 gene after full reconstitution of T cells? Again, it is not clear to me whether the authors really observe what they are arguing in the manuscript.
We are unfortunately unable to obtain commercially available Cre-ERT2 Stat1 fl/fl mice or labs which have published with these mice which will directly address the reviewer's question within a reasonable time frame. We do however understand the reviewer's underlying concern of whether T cell specific STAT1 deletion in lymphoreplete mice also leads to its elimination.
To address the concern, we transferred WT or Stat1 -/-T cells (CD45.2 + ) into WT CD45.1 mice to see if Stat1 -/-T cells are also reduced when there's no lymphopenia. Consistent with earlier reports, there is reduced overall proliferation when T cells are transferred into lymphoreplete mice compared to lymphopenic mice. We do not observe a reduction in overall T cell percentages when Stat1 -/-T cells are transferred into lymphoreplete mice ( Figure R3a). However, in the SP population, there is a mild but consistent decrease when T cells lack STAT1 ( Figure R3b). This supports the earlier data showing that NK cells specifically target Stat1 -/-T cells undergoing SP. We thus believe that should STAT1 be inducibly deleted after reconstitution, they would not be eliminated unless there's a trigger for SP.

If a marked reduction of Stat1-deficient T cells was observed even in the absence of colonic inflammation, the title of the manuscript is misleading. Colonic inflammation is just a result of manipulating animals under the specific experimental setting.
We agree with the reviewer that the primary effect of STAT1 is to promote T cell survival with colonic inflammation being a secondary effect and will alter the title to reflect this emphasis. other groups to be driven by the microbiota. We think that the reason why Rag1 -/mice don't develop colitis with unfractionated CD4 + T cells is likely due to the presence of Tregs.

What are the physiological or pathological relevance of the findings? An experiment such as transfer of irrelevant of T cells into RAG-deficient mice should be considered.
The T cell transfer model of colitis that we've utilized is a relevant and robust model that has been utilized by several groups to understand IBD pathogenesis. As an example, we've used this model to uncover a role for IL-1 in T cell driven colitis in Il10rb -/-Rag1 -/mice, which led us to utilize anakinra as a therapeutic approach in patients with Il-10R deficiency. (Shouval et. al. Gastroenterology 2016) With regards to the findings of STAT1 deficiency, we believe this is important as several groups have utilized the IBD model, as well as other similar models (GvHD, Autoimmunity) to draw conclusions about the role of STAT1 in T cell driven disease without examining the role of NK cells, as described in the discussion section of the manuscript. Moreover, our finding that STAT1 specifically protects T cells that undergo the microbiota driven SP provides a mechanism by which NK cells regulate commensal driven T cell responses. We note that earlier reports showed that Ifnar1 -/antiviral T cells are eliminated by NK cells only in the setting of viral infection, and our work adds to this paradigm by showing that this mechanism is relevant in regulating microbiota driven T cell responses, which will be helpful in studying intestinal disorders such as IBD. It also suggests that, therapeutically, on can consider specifically target STAT1 in T cells as a way of reducing commensal driven T cell inflammation.
Finally, STAT1's intrinsic antiproliferative activity has led scientists to consider targeting STAT1 for improving T cell reconstitution in clinical lymphopenia e.g. HIV infection, transplantation. Our findings suggest that this approach will have to be balanced with ensuring sufficient MHC-I levels to prevent NK cell mediated T cell depletion. . The use of these irrelevant T cells was reported to lead to reduced overall SP, which is analogous to our data in lymphoreplete mice where there is also reduced overall SP. Thus, we believe that performing this experiment would not yield significant insight over the previous data ( Figure R3).