CD4 T cells need to have their cake and eat it too to become Th17

     I want to review two new and complementary papers addressing natural specificity of gut Th17 cells. One paper was published in Immunity and came from Ivaylo I. Ivanov's Lab (1) and second one was published in Nature and came from Dan Littman's Lab (2). This is not surprising since Ivaylo I. Ivanov is Dan Littman's former postdoc who was the first author in original paper discovering relationship between segmented filamentous bacteria (SFB) and IL-17 producing T helper cells.

     To understand significance of these papers one needs to remember that proper development of immune system requires presence of gut microbiota. Some of this microbes induces development regulatory T cells and some of them can induce development Th17 cells.

    

     These papers address the role of SFB antigen in driving Th17 induction. In Immunity paper, the authors first showed that induction of new Th17 cells requires class II (MHCII) expression on host cells. However, since CD4 T cells will need to interact anyway with MHCII just to become properly activated, result is not surprising or even relevant. Second, they used adoptive transfer of TCR transgenic CD4 T cells on RAG KO background to reveal that antigen alone or SFB alone is not sufficient to promote Th17 conversation from random T cells. Third, to understand the specificity of Th17 cells, the authors utilized IL-17-GFP reporter mice and prepared hybridomas from GFP+ and GFP- T cells and stimulated them with different gut microbial lysate or alternatively with purified SFB. The vast majority of GFP+ cells (consisting of natural Th17 cells) responded to SFB antigen. Lastly, the authors showed that MHCII expression on dendritic cells was necessary and sufficient to induce Th17 cells.

     

     Nature paper went further to analyze SFB antigen recognized by Th17 cells. They utilized IL-23R-GFP reporter mice that labels natural occurring IL-17 producing T cells. The authors generated hybridomas that specifically recognized two SFB antigen nominally named SFBNYU 3340 and SFBNYU 4990. Based on hybridoma specificities, the authors generated several transgenic mice expressing SFB-specific TCRs. Adoptive transfer experiment revealed that upon transfer almost all donor SFB-3340 specific T cells started to express Th17 specific transcription factor, RORgt. Interestingly, SFB antigen itself did not drive the polarization into Th17 cells since its expression in Listeria Monocytogenes (Listeria-3340) induced typical Th1 response. Moreover, dual colonization of mice with SFB and Listeria did not affect differentiation of Th17 driven by SBF.

   

    In summary, these studies showed that SFB colonization induce Th17 cells in antigen and context dependent manner. The most likely scenario involves gut dendritic cells internalizing SFB and expressing its antigen to CD4 T cells to drive their polarization towards Th17.

     These studies reveal that even though gut microbiota is made of diverse species, few species have extraordinary effect on development of immune system. This implies that there is some evolutionary pressure for development of such cohabitation. The knowledge that particular bacterial specie can drive particular polarization of T helper subset even in presence of competing polarization programs (SFB vs. Listeria) could lead to development of natural gut flora based medicine (transgenic microbial flora) to treat many autoimmune and allergic diseases.

David

microbiota with retro taste

This is a really good paper to understand how we can be mistaken to rely on knock-out mice models without taking into account indirect effects of the mutations (1).

Published in journal Nature in November 2012, this study from Kassiotis Lab looked at the expression of retroviruses in immune-deficient B6 mice. Under  normal situation, B6 have replication-defective endogenous murine leukaemia viruses that can't produce infectious particles. However, as shown in paper, under certain immune-deficient conditions, like RAG-KO, Ighm-KO, MyD88-KO, TLR7-KO, but not in T cell-deficient mice, mouse starts to produce infectious retroviral particles by recombining eMLV with other retroviruses. In this conditions many immune cells started to express MLV surface glycoprotein. This, in turn, can promote formation of viral-induced lymphomas in aged mice. Anyone who has worked  with RAG-KO mice can attest to this observation. The author showed that MLV was transmitted from mother to offspring.

Next, the authors thought to understand the mechanism of this retroviral re-activation. It has been known that inflammatory stimuli can re-activate MLV expression in mouse cells. Indeed LPS and Poly(I:C), but not Pam3CSK4, could increase MLV expression in mouse bone marrow-derived dendritic cells.

Lastly, based on this observations, the authors reasoned that changes in commensal microbiota in Antibody-deficient mice contributed to the reactivation of retroviruses. This was confirmed by the absence of MLV in germ-free RAG-KO mice. In addition, giving to mice acidified drinking water  to reduce bacterial colonization and number in the gut dramatically cut MLV expression.

The conclusion of the paper is following: natural antibody produced in normal mice in a MyD88 and TLR7-dependent manners prevents commensal bacterial translocation and re-activation of MLV. Very similar observations regarding microbiota and viruses infectivity were made by two other research groups that was reported in journal Science in 2011 (2, 3).

Why these results are relevant: first, it can explain how retroviral-induced tumors can develop. Secondly, many studies in immune-deficient mice may require re-evaluation to avoid erroneous conclusions because of the unintended presence of MLV.

David

CD4 T cells with killer instincts

CD4 T cells with killer instincts

These two, back to back papers from Nature Immunology caught my attention this week (1, 2). Both papers deal with CD4 T cells that acquire "abnormal" CD8 T cell functions. It has been known for some time that occasionally CD4 T helper cells can show cytotoxic potential or that some tissues, for example, intestinal tissue harbors hybrid population positive for both CD4 and CD8.

To understand the origin of this double positive T cell population, the authors developed reporter mice expressing Thpok-GFP and Runx3-YFP. Transciption factors Thpok and Runx3 control CD4 T helper and CD8 T cytotoxic cell developmental program s respectively. First finding was that unlike spleen CD4 T cells, more than 50% of intestinal CD4 T cells are Thpok-low and express CD8-alpha, granzyme B and CD107a, all markers of cytotoxic T cells. Runx3-YFP reporter mouse confirmed that this CD4 killer cells express Runx3. Importantly, the authors showed that development of double positive T cells was dependent of TGF-beta and vitamin A signaling in T cells. Next, using Rosa26-YFP / Thpok-cre reporter mouse model, they showed that Thpok-low CD4 T cell population developed from Thpok-high population. Next, the authors showed that sorted Thpok-high CD4 T cell upon adoptive transfer into RAG KO mice lost Thpok expression specifically in the gut, but not in spleen or lymph nodes. At the genomic level, the authors showed that conditional deletion of Thpok silencer abrogated acquisition of cytotoxic phenotype by CD4 T cells by stabilizing Thpok. Interestingly, CD4 killer cells do not develop in germ-free mice and deficiency of MyD88 had no effect. Oral feeding with ovalbumin antigen led to the accumulation of OT-II CD4 killer T cells in the gut that express low level of Thpok. Interestingly, regulatory CD4 T cells expressing Foxp3 also developed in similar conditions (after oral feeding). At this stage it is not known what determines this difference. Interestingly, absence of TGF-beta signaling in OT-II cells drove development of severe diarrhea in these mice after oral feeding with ovalbumin, compared to wild-type control. However this mouse also showed deficiency in Foxp3 T cell development, so interpretation of these data is complicated by this fact. These studies showed that Thpok expression was necessary for intestinal damage by Th17 cells in adoptive transfer model and that Runx3 expression counteracted this effect with the help from TGF-beta and retinoic acid.

While these authors suggest that they identify unique T cell population, I would rather call it adaption to local tissue environment. At this stage it is not clear what is the advantage of having CD4 killer cells in the gut or what are their physiological function. They may be needed to maintain local tolerance to commensal microbiota. One interesting observation is that they still maintain class II restriction implying that function of CD8-alpha may be just a by-product of the transcriptional switch.

David