Norovirus dirty job

Last week Science magazine published a new article that literally took my breath away. This study is a classical example why is the scientific research so exciting.

In this new study, the research group led by Stephanie Karst at the University of Florida, studied Noroviruses target cell population. 

Based on earlier studies on RAG KO and B cell KO mice (both lack B cells and show low NoVs titers), the authors speculated that B cell represent NoVs target population.  

Initially, in vitro experiment confirmed that mouse NoVs could infect B cell lines (and macrophage cell line, RAW264.7 too) but not mouse epithelial cell line.

Of note, while mouse B cell line, WEHI, showed a high level of productive infection with NoVs (>90%), mouse B cell line, M12, showed a minimal level of productive infection with NoVs (~10%). However, both cell lines could amplify NoVs to the same level by day 3 post infection in vitro. 

For some reason, the authors did not include primary mouse B cells for comparison in these series of experiments.

Next, the authors examined in vivo infectivity of mouse NoVs in wild-type or uMT mice. uMT mice lack B cells (but may have small number of IgA producing B cells). These experiments showed that uMT mice showed reduced viral load in tissues rich with B cells, like MLN. Interestingly, colon tissue showed comparable viral load in both control and B cell deficient host. 

To further validate the hypothesis that NoVs infects B cells, the authors purified B cells from payer's patches from infected mice and run virus specific RT-PCR. The authors found that 1 out of 1000 B cells contained NoVs. However, in bulk cells from payer's patches, 1 out 100 cells contained NoVs (10-fold more). It seems other cell type(s), besides B cells, can be infected by NoVs more efficiently in vivo. 

Next, the authors moved to human system. Unlike mouse B cell lines, human B cell line did not show productive infection with human NoVs. Surprisingly, when unfiltered sample from NoVs infected human stool was added to the human B cell line, the authors observed significant increase in viral load. 

The transwell experiment with human epithelial cell line at the top and B cell line at the bottom, confirmed that B cell were required for NoVs amplification from unfiltered stool sample.

The authors speculate that human stool contained co-factor that facilitated B cell line infection by NoVs. Since it was known that human NoVs interact with histo-blood group antigen H, the authors tested the role of purified H antigen or H+ Enterobacter cloacae in in vitro NoVs infection model of human B cell line. As predicted, both purified H antigen and H+ Enterobacter cloacae, but not LPS or H- E. coli, could amplify NoVs infectivity.

It appears that H antigen improved human NoVs attachment to the target B cell line.

Finally, to verify the role of endogenous gut flora in mouse NoVs infectivity, the authors examined NoVs load in mice treated with broad spectrum antibiotics. Indeed, antibiotic treatment significantly reduced mouse NoVs infection. Interestingly, unlike uMT mice results, antibiotic treatment almost abolished NoVs infection of colon tissue. 

In summary, this study expands our understanding as to how viruses interact with the host, in this case with human host. The observation that endogenous bacteria "helps" NoVs to infect the human host is an important step for better treatment of NoVs infection in humans.

As mentioned earlier, there are few results in this article that require further discussion. For example, experiments with primary B cells would have been more informative. In addition, uMT mice may still have small number of IgA secreting B cells, and may not be the ideal model to study the role of B cells in NoVs infection. Use of another B cell KO mice, JH-/-, or AID/μS double KO mice, would be more informative. And finally, it seems that colon tissue infectivity by NoVs was not influenced by absence of B cells. Why? Does density of H+ Enterobacter cloacae determines the role of B cells?            

David Usharauli

          

         

   

    

Regulatory T cells need more than Foxp3

Absence of Foxp3+ Tregs leads to uncontrollable activation of self-specific T cells. 

Natural Foxp3+ Tregs are selected in the thymus. This selection is TCR mediated. It is safe to assume that at least in the thymus, TCR transmits signal(s) essential for the induction of Foxp3 protein. 

Does TCR signal have the same function in the periphery as well?

New research published in journal Immunity has examined an effect of TCR ablation on Tregs. To do this, research group led by Marc Schmidt-Supprian have used Mx1-cre TCRalpha fl/fl mice. Mx1 is a IFN-alpha responsive element. IFN-alpha can be induced by Poly(I:C) injection. 

When the authors injected Poly(I:C) into Mx1-cre TCRalpha fl/fl mice they found that ~20% of Foxp3 Tregs lost TCR expression. 

Unexpectedly, examination of TCR-deficient Tregs (tracked by eGFP expression) 6 weeks post treatment revealed that both TCR+ and TCR- Tregs expressed similar level of Foxp3 (both at protein and mRNA level). 

Actually, TCR ablation had effect mostly on molecules which define Tregs' activation phenotype, such as OX40, ICOS, CD44.      

Separate experiments showed that TCR ablation reduced half-life of Tregs.

To examine a consequence of TCR loss on Foxp3+ Tregs' functionality, the authors transferred naive T cells together with Tregs recovered from tamoxifen treated CD4-creER TCR alpha fl/fl (or from control mice) into T cell deficient recipient mice. In this setting, Tregs from tamoxifen treated CD4-creER TCR alpha fl/fl mice were less potent in controlling naive T cell-induced inflammatory response.

Similarly, Tregs from tamoxifen treated CD4-creER TCR alpha fl/fl mice were less potent in controlling EAE induction in the recipient DEREG mice after MOG immunization.

In summary, this study and another one from Rudensky's lab published in Nature Immunology this month, indicate that while Foxp3 expression is essential for Treg identity, Tregs' functionality is a more complex trait and depends on continues stimulation through TCR signaling. 

What is puzzling for me is the fact that according to this study CD4-creER TCR alpha fl/fl mice retain ~30% TCR+ Tregs after tamoxifen injection and still show loss of overall Treg potency upon adoptive transfer. There are data to suggest that adoptive transfer of a single CD8 T cell can protect recipient mice from infection. Why are millions of supposedly normally functioning Tregs so inefficient?  

David Usharauli

          

           

Firmicutes for a slimmer waist?

Obesity is a complex metabolic disorder. Both genetic and environmental factors play a role. More recently, third component, gut microbiota was implicated as well.

Several seminal studies from Jeffrey Gordon's lab suggested that abundance of bacterial family of Firmicutes in the human gut flora contributed to obesity due to their efficacy in harvesting energy from the food source.   

However, new study in journal Cell, surprisingly points to opposite direction. This study led by Ruth Ley (US), in collaboration with Timothy Spector (UK), has examined a stability of human gut microbial community across genetically related (twins) or unrelated human population. 

Twins are ideal to study Nature (Darwinian genetics) versus Nurture (Lamarckian epigenetics) effects. Initial studies on twins indicated that environment (food, habits, etc), rather than genes played more fundamental role in altering gut flora composition. However, it seems those prior studies have used smaller sample size. In this study, however, the authors have used larger pool size of twins (416 twins) for comparison and for validation of their observation.

First, the authors found that gut flora was less variable between monozygotic twins, as compared to dizygotic twins or unrelated individuals. This suggests that genes play a more dominant role in influencing gut microbiota.

Interestingly, gene effect was significant for bacterial family of Firmicutes. 

However, surprisingly, no significant difference were found for bacterial family of bacteroidaceae between twins or unrelated individuals.

Second, the authors found that between monozygotic twins, the most heritable bacterial family was Christensenellaceae. This is a newly described bacterial family.

Christensenellaceae belong to Firmicutes family.

Interestingly, the authors made observation that abundance of Christensenellaceae in the gut flora correlated with low BMI (<25). Indeed, fecal transfer experiment in germ-free mice showed that weight gain was inversely correlated with the level of Christensenellaceae in donor stool. The authors also made an observation, that for some unknown reason, Christensenellaceae effect on BMI was most prominent when Christensenellaceae was associated with metanogenic bacteria.   

Finally, to validate their observation, the authors added live Christensenella minuta to fecal transplant from obese individual with no detectable level of Christensenellaceae and transferred into germ-free mice. At day 21 post transfer, mice with live Christensenella minuta gained less weight and had less adipose tissue.

In summary, the results from study indicate that genetic factors may play a dominant factor in determining the presence of some but not all bacterial communities in the gut. In addition, the data indicate that few bacteria species play disproportionately large role in influencing low BMI. The most intriguing observation of this study is the fact that presence of bacteria from Firmicutes family correlated with low BMI.

David Usharauli