Friday, May 23, 2014

Understanding self-nonself discrimination by adaptive immune system

modified from 

Usharauli, David (2010, October 30). Understanding self-nonself discrimination by adaptive immune system. 


In 1969, Niels Jerne, later a Nobel laureate, predicted that "immunology will be completely solved within fifty years from now”, i.e. by 2019 (Jerne, 1969). Even earlier, in 1964, Frank Burnet, a Nobel laureate, complained that “the infant science [immunology] would soon run out of problems to solve” (Anderson, 1994). What was the basis for such a confidence? Maybe they knew better? After all, both were science demigods, acclaimed and worshipped and the modern immunology was of their creation. So what was this all about? 

The answer is the principle of self-nonself discrimination by adaptive immune system.
The adaptive immune system consists of T cells and B cells (clones). Each clone expresses a unique membrane receptor that (as the product of random gene recombination) is specific for an antigen, either self or nonself. Since an immune response can be initiated from a single clone (Stemberger et al, 2007), it is mandatory for the adaptive immune system to keep self-specific clones in check (tolerance) while at the same time allowing nonself-specific clones to respond (immunity). So how is this achieved? 

The value of any conceptual model rests on its predictive power. Initially, based on Peter Medawar’s (Nobel Lecture) experiments and Niels Jerne’s theoretical concept, the favored model was the one proposed by Frank Burnet. Burnet’s model of self-nonself discrimination required that all clones being generated during the embryonic stage when presumably only self-antigens were present (Burnet, 1957). At this stage, any clone that expressed a self-specific receptor would be deleted. Only nonself-specific clones would accumulate and subsequently respond to nonself-antigens when introduced by pathogen-invader. 

Soon, however, this model required major modification when it was shown experimentally that new clones of T and B cells are continuously generated long after birth. The problem was that according to the Burnet model, immediately after birth, each clone becomes spontaneously fully capable of responding to its cognate antigen, and given that newly generated clones can be self-specific, this scenario would have probably led to an unacceptably high frequency of self-specific responses, i.e. autoimmunity. A solution to this problem was introduced by Joshua Lederberg. According to the Lederberg model, each new clone (irrespective of whether it was generated during the embryonic stage or after birth) transits through a deletion-only stage and then it becomes spontaneously fully capable of responding to its cognate antigen (Lederberg, 1959). This deletion-only stage ensured that new clones with self-specificity would encounter their cognate self-antigens at this stage and become deleted. 

The Lederberg model, however, also failed for two reasons: first, deletion-only stage for newly generated clones was difficult to document experimentally, and second and more importantly, the Lederberg model did not account for the scenario when new self-antigens are introduced in the body, for example during puberty, pregnancy, lactation, etc. According to the Lederberg model these new self-antigens would have been considered as nonself (foreign) by the adaptive immune system and attacked, leading to unacceptably high frequency of autoimmune responses. In other words, the Lederberg model only works if self-antigens are constant (static) and it fails if self-antigens are changing (dynamic). 

All classical self-nonself discrimination models fail if self-antigens are changing. More recently, Melvin Cohn tried to save the classical self-nonself discrimination model by suggesting that the primary function for AIRE gene is to drive expression of “future” self-antigens during embryonic stage to create static view of self-antigen dynamics (Cohn, 2009).
The challenge posed by self-antigen dynamics was the major conceptual driving force behind the development of dendritic cell based innate “self-nonself” discrimination models of the past 20 years. Two models in particular completely revolutionized the modern immunology. The first one was introduced by late Charles Janeway and the second one by legendary Polly Matzinger (Janeway, 1992; Matzinger, 1994). 

In these models, “self-nonself” discrimination actually refers to immunity/tolerance fate of individual clones. Many weaknesses of Janeway’s model as it was originally understood (for example, how viruses stimulate innate immunity, how nonbacterial adjuvants, such as alum, work) is nowadays mainly resolved. While both these models can explain the immunological tolerance to changing “self”, both dendritic cell based “self-nonself” discrimination models so far failed (one in theory and another in practice) to explain the persistence of adaptive immune response to nonself (foreign) transplanted tissues. 

This maybe the result of downplaying the role that effector/memory T cells can play in perpetuating the immune response. Especially important is the relationship between effector/memory T cells and dendritic cells. If an effector/memory T cell with the self-specificity were able to deliver maturation signals to resting dendritic cells (and be equivalent in this regard to signals introduced by pathogen-invader or generated during tissue damage), then both Janeway’s and Matzinger’s models would fail because this scenario would lead to unacceptably high frequency of autoimmune responses. Interestingly, one paper published in Nature Medicine found that memory CD4 T cells were capable of inducing signals similar in nature to that induced by pathogen-invader (Strutt et al, 2010). If confirmed, this observation will necessitate yet another major modification of self-nonself discrimination model. In addition, to best of my knowledge, no one yet able to propose a working hypothesis how to incorporate Foxp3+ T regs in a predicable model of immune system function. After all, there are still 5 years left until 2019.

David Usharauli

Tuesday, May 20, 2014

Autophagy, metabolic switch and long-term antibody response

A cellular proliferation, differentiation or survival require fine-tuned metabolic processes (oxidation, glycolysis, etc). Autophagy is a metabolic process that supplies energy trough recycling of cellular byproducts, kind of cellular green technology. Some type of cells, like neurons which are exceptionally long-lived, are critically dependent on autophagy.

Immune system also have few cell types that are exceptionally long-lived, like memory B cells or plasma cells. They are responsible for protection against re-infection, the basis for immunity. However, how memory B cells or plasma cells are able to survive so long (years, or even decades) is unknown.

Here, new paper published in Nature Medicine may provide some clues to this questions. The authors, Min Chen et al., examined the role of autophagy in memory B cell maintenance (1). I find their research to be of quite good quality with minimal deficiencies discussed below.

First, the authors showed that unlike germinal center B cells, sorted memory B cells are resistant to spontaneous cell death in an in vitro culture and do not display pro-death caspase activity. In addition, memory B cells showed high degree of expression of several autophagy markers, like LC3 and Atg7.

To undress the role of authophagy in memory B cells, the authors created mice with B cell specific deletion of autophagy gene Atg7 (CD19-cre Atg7 fl/fl). Indeed, deletion of Atg7 in B cells resulted in a loss of resistance of memory B cells to spontaneous cell death in an in vitro culture. However, this increase of death in CD19-cre Atg7 fl/fl memory B cell was not caspase-dependent but was due to increase in sensitivity to the oxidative stress, since it was significantly reversed by the use of anti-oxidative agents like N-acethyl-L-cysteine (NAC) or α-tocopherol.

Interestingly, CD19-cre Atg7 fl/fl mice displayed normal primary antibody response to NP-KLH immunization (day 14), however, secondary memory response (~ day 60 + day 5) was dramatically diminished compared to control, wild-type mice. There was diminished secondary response from bone marrow samples as well.

Kinetic analysis indicated that memory B cell formation was normal until day 14 and then declined in CD19-cre Atg7 fl/fl mice.

Significantly, use of NAC or α-tocopherol in vivo could rescue memory B cell decline and secondary antibody response after immunization.

Finally, the authors showed that CD19-cre Atg7 fl/fl mice were highly sensitive to influenza infection even after prior immunization.

In summary, the authors proposed that deletion of Atg7 in B cells leads to severe loss of memory B cells and failure to protect against re-infection. Mechanistically, Atg7 deficiency impairs mitochondrial function and leads to premature death of memory B cells due to excess of ROS generation and lipid peroxidation.

One difficulty interpreting these data has to do with the fact that both memory B cell and long-lived plasma cells (LLPC) contribute to secondary antibody response. Earlier study by Pengo N et al. (2), clearly showed that autophagy deficiency impairs LLPC formation. It is very difficult to differentiate between memory B cells role in secondary response from that of LLPC. In Figures 4 and 5, there is significant staining for IgG1-negative but antigen-specific population in CD19-cre Atg7 fl/fl mice. Are those cells LLPCs? Is autophagy required for memory B cell survival or their secondary differentiation into plasma cells? Why is autophagy essential for survival of memory B cells which supposedly are quiet population in absence of antigen?


Saturday, May 17, 2014

How antibiotics can make us vulnerable

   Since birth human [infant] immune system is constantly exposed to environmental antigens. These antigens are either of harmless or harmful nature. In general, there are two ways for immune system to learn the difference between harmless and harmful antigens.

   First detection system is a genetically fixed trait programmed to detect evolutionary stable signature associated with presence of harmful microorganisms, like endotoxin (LPS).

   Second detection system is a de novo acquired phenotypic trait involving collaboration between commensal microbiota, dendritic cells and T cells. It complexity, especially its dependency on commensal microbiota renders it more prone to errors.

  Until very recent times, humans used to spend their entire lives in a farm-like environment, being exposed to diverse set of environmental antigens and microorganisms. However, urbanization and non-discriminate use of antibiotics changed the playing field, especially undermining the proper development of second detection system. This led to the increase in frequency of allergic or autoimmune inflammatory disorders and several types of cancer.

    To specifically show how non-discriminatory use of antibiotics could lead to immune deficiencies, lets review recent paper published in Nature Medicine.

    This study by Deshmukh H, et al. (1), examined the effect of broad spectrum antibiotics treatment on mice neonates.

    As expected continued treatment of pregnant female mice and neonates (- day 5 / + day 14)  with combination of 3 or 5 broad-spectrum antibiotics resulted in significant (20X-50X) reduction of gut microbiota. This in turn led to reduction of circulating and bone-marrow residing neutrophils. Plasma level of G-CSF, a cytokine responsible for neutrophil mobilization were also reduced. Analysis of germ-free mice confirmed that these effects were due to reduction of microbiota.

   To understand how this reduction of neutrophils affected neonatal mice, the authors injected antibiotic-treated neonatal mice with pathogenic strains of E. Coli or K. Pneumoniae. Compared to control mice, antibiotic-treated neonatal mice became highly susceptible to infections. Similar effect was seen with neonatal mice treated with neutrophil-depleting antibody. However, neonatal mice simultaneously treated with either G-CSF or transplanted with gut microbiota showed improved survival and resistance to infection.

    To further elucidate the mechanism of susceptibility, the authors examined the role of IL-17. It has been recognized that IL-17 plays a role in G-CSF induction and in neutrophil biology. It turned out that antibiotic-treated neonatal mice or germ-free mice have reduced level of IL-17. Alternatively, treatment of mice with anti-IL-17 antibody or use IL-17R alpha KO mice confirmed that IL-17 was critical factor in neutrophil mobilization.

   Finally, the authors showed that TLR4 and MyD88 pathways were involved in microbiota-driven neutrophil mobilization. Importantly, injection of low dose of LPS could restore IL-17 and G-CSF level in antibiotic-treated mice and improve neutrophil mobilization in blood and bone-marrow. This effect was again IL-17 dependent. However, the authors did not examine if injection of low dose of LPS could improve survival of antibiotic-treated mice.

   In summary, the authors' proposed model suggests that early post-natal colonization of neonatal mice with microbiota activates innate immune system leading to IL-17 production that in turn induces G-CSF secretion and neutrophil mobilization. This process prepares neonatal mice to resist harmful effect of pathogenic microorganisms.

   One relevant finding is that injection of G-CSF or low dose of pure LPS could improve neutrophil mobilization in neonatal mice treated with antibiotics. This could have clinical application for human neonates.

   One things that is really confusing about this study is the fact that antibiotic-treated neonatal mice were highly susceptible to pathogenic microorganisms which are it turn supposedly susceptible to the effects of broad-spectrum antibiotics used to treat mice. I have no idea how to interpret these particular experiments.


Saturday, May 10, 2014

reverse sensing of IL-10 by innate system prevent IBD

           IL-10 is a multipurpose cytokine. The research suggest IL-10 dampens immune response. Many cell types including T cells, B cells or macrophages produce IL-10. IL-10 receptor is made of two subunits, alpha and beta. Alpha subunit is specific for IL-10 and beta subunit is shared with other cytokines, for example IL-22. Since many cell types produce and sense IL-10, it is not immediately clear how redundant is IL-10 system. The animal model mostly used to test the role of IL-10 is an equivalent of human inflammatory bowel diseases (IBD). These studies showed that IL-10 played critical role in “educating” immune system to tolerate the presence of gut microbiota. IL-10KO mice develop spontaneous IBD. However, germ-free IL-10KO mice has no IBD.

        Recently, two back to back studies published in Immunity try to explain mechanistically how exactly IL-10 prevents IBD.

      First study came from Steffen Jung's lab (1). In my opinion, this particular paper is very simple and it is unlikely it would have been accepted for publication in Immunity by itself. I will highlight only the most relevant data from this study.

       First, the authors showed that in the absence of IL-10 gut resident macrophages, identified as CX3CR1-GFP positive cells, adopt pro-inflammatory phenotype. This is expected.

     Next, the authors generated gut macrophage-specific IL-10 or IL-10Ralpha deficient mice by crossing CX3CR1-cre mice to IL-10 fl/fl or IL-10Ralpha fl/fl mice.

     Interestingly, CX3CR1-cre IL-10 fl/fl mice develop normally and did not show any sign of IBD. This suggested that gut macrophage-derived IL-10 is dispensable for protection against IBD. This is a new finding.

      However, CX3CR1-cre IL-10Ralpha fl/fl mice developed IBD, similar to total IL-10KO mice. This suggested that sensing of IL-10 by gut macrophage was essential for protection against IBD. This is a new and unexpected finding.

      The most obvious weakness of this paper is the direct use of KO mice without adoptive transfer experiments. Since CX3CR1 can be expressed by other cell types (Dendritic cells, T cells), CX3CR1-cre IL-10Ralpha fl/fl mice phenotype could have been affected by function of non-macrophage population.

       Luckily, second paper (2) from Scott Snapper's Lab provided necessary data to fill the holes in the story. It contains enormous amount of data. I will highlight the most important findings.

       The authors showed that IL-10Rbeta KO mice develop IBD. However, RAG2 KO / IL-10Rbeta KO mice lacking T and B cells are healthy. This suggests that presence of adaptive immune system is required for IBD development in IL-10Rbeta KO mice.

         Next, the authors showed that adoptive transfer of wild-type (WT) total T cells into RAG2 KO / IL-10Rbeta KO mice drives IBD. This suggests that IL-10 signaling defect in non-T cells (innate system, other tissues) system drives IBD.

        Next, to further narrow down IL-10 signaling defect that drives IBD, the authors generated bone-marrow (BM) chimeric mice by reconstituting lethally irradiated RAG2 KO or RAG2 KO / IL-10Rbeta KO mice with RAG2 KO or RAG2 KO / IL-10Rbeta KO BM. Upon transfer of total WT T cells, only mice with hematopoietic cells derived from RAG2 KO / IL-10Rbeta KO BM developed IBD. This definitely showed IL-10 signaling defect in innate immune system drives IBD.

     The authors also conducted several in vitro experiments and showed that transfer of WT BM-derived macrophages generated in M2 polarizing condition could prevent IBD development in RAG2 KO / IL-10Rbeta KO recipient mice following T cells transfer, whereas transfer of IL-10Rbeta KO BM-derived macrophages failed to do the same.

       Next, the authors showed that adoptive transfer of total WT CD4 T cells into RAG2 KO / IL-10Rbeta mice leads to fewer Foxp3-positive T cell generation in lamina propria (LP) compared to RAG2 KO recipient mice.

       One unusual result was the evidence that exogenous supplementation with IL-10 following transfer of WT CD4 T cells failed to prevent IBD development in RAG2 KO / IL-10Rbeta recipient mice.

      In summary, we can imagine the following scenario how IL-10 prevents IBD: first, gut microbiota is internalized by gut dendritic cells or gut macrophages and presented to T cells (either to naive T cells or thymus derived Foxp3+ T cells). Second, this process generates local pool of regulatory T cells that secrete IL-10 and conditions local innate system or specifically macrophages to adopt anti-inflammatory phenotype. In absence of this feedback, however, macrophages acquire pro-inflammatory phenotype, overrides regulatory T cell effect and drives IBD through generation of colitogenic effector T cells. Of course, the question how initial T cells “know” to secrete IL-10 to condition gut macrophages is unclear.