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

An invisible friends

If you appreciate unusual scientific results then these two recent papers are for you to read. One paper by David Hill et al. (1), was published in Nature immunology, and another one by Olszak and An et al. (2), was published in Science. Both papers are trying to explain how hygiene hypothesis might work by showing how early exposure to commensial bacteria-derived signals educates immune system so it can control allergic responses later on. What is interesting, however, is that fact that both papers reached similar conclusions by examining different arms of immune system.

In paper by David Hill et al., the authors observed an increase in basophils and IgE in germ-free mice or in mice treated with broad-spectrum antibiotics. Using house dust mite allergen or papain allergy model, the authors showed that microbial flora-deficient mice display exaggerated inflammatory response and produce more IgE promoting cytokine IL-4. This increase in basophil numbers however did not occur in either T and B cell-deficient mice or IgE-deficient mice, indicating that IgE promoted basophil accumulation. Direct injection of IgE supported this conclusion. Mechanistically, IgE driven basophil accumulation was shown to be mediated through increase of basophil differentiation from its precursors via augmentation of sensitivity to IL-3. In addition, B-cell specific MyD88 signaling was shown to play a role in limiting IgE production. In short, this paper explained hygiene hypothesis based on B cells-MyD88/IgE/ basophil axis.

In paper by Olszak and An et al., however, authors reached similar conclusion regarding protective effect of microbial flora by examining biology of NKT cells in germ-free mice. The authors showed that germ-free mice had more tissue NKT cells (I wonder how authors of other papers could miss this fact). Using chemically-induced colitis model or allergen-induced asthma model, the authors showed that blocking CD1d molecule by specific antibody inhibited allergic-prone response in germ-free mice (reduced tissue eosiphilia and IgE accumulation), indicating that NKT cells are involved. This conclusion was supported by NKT-specific deficient mice. Crucially, the authors showed that early, but not late life exposure to microbes could protect germ-free mice from allergy. Mechanistically, accumulation of NKT cell in germ-free mice was linked to CXCL16 over-production as a result of microbial-deficiency. It was MyD88 independent effect.

While both paper shed some new lights into hygiene hypothesis, it also highlight how narrow scientific research has become. Each group looked at germ-free mice and observed only what made more sense to them. We know that NKT cells can drive IgE response and I can even argue that NKT cell activation in germ-free mice helps B cells to switch to IgE production. So in a way these two paper complement each other. However, the point is that we still don't understand why absence of microbial signals should lead to IgE response in general. Does it mean that IgE response is somehow helpful in protecting tissue against insults of non-microbial origin? Why? How?

David

1. http://www.nature.com/nm/journal/v18/n4/full/nm.2657.html

2. http://www.sciencemag.org/content/early/2012/03/21/science.1219328.abstract