Nurture Immunity: Immune system influenced more by environment than by genes

Differences in immune protection presumably explain why some people exposed to infection resist disease or recover while others succumb.  These authors sought to distinguish the influences of genes and environment on immunity. They compared the cellular and molecular components of the immune system among 210 twins: 78 monozygotic (MZ, “identical”) and 27 dizygotic (DZ, fraternal) pairs.  They measured 43 serum proteins and 72 immune cell populations repeatedly and longitudinally (over time) to assess actual variations and account for technical variations.  MZ twins, who have practically identical genomes, and DZ twins, who share half their genes, are especially valuable for assessing the relative contributions of “nature or nurture” (genes or environment) to phenotype.  Their analysis allowed them to detect as little as 20% heritability.

The levels of few proteins and cell populations are under strong genetic control, such as interleukin-6 and CD4+ “central memory” T cells, but most are only weakly heritable or not at all (Fig 1). They found that a common, chronic infection, by cytomegalovirus (CMV), influences the levels of most (58%) cell populations and proteins (Fig 5).  Variation between twins increased as they age, probably reflecting different environmental stimuli and epigenetic changes (Fig 4).  Most intriguing, they correlate the heritability of response to vaccines to the age of immunization, whereby early childhood vaccines are highly heritable while vaccines after early adolescence have no detectable heritability (Table 1, shown below).  

Brodin et al. Cell. 2015 Jan 15;160(1-2):37-47. Variation inthe human immune system is largely driven by non-heritable influences. 

Levels of a common, chronic virus (TTV) reflects the immune competence of transplant recipients

Successful organ transplantation requires careful immune suppression: enough to block the rejection of transplant while permitting host defense against infectious microbes.   Viruses that are not cleared by our immune systems, are common in healthy people, and can complicate transplantation include cytomegalovirus (CMV) and Torque teno virus (TTV), which was first described in 1997 [review].  TTV is a small (3.8 kb), single-stranded, transfusion transmitted DNA virus, representative of a highly diverse family of anelloviruses.   

The authors examined the influence of immune-suppressive drugs (e.g., tacrolimus, mycophenolate mofitil, cyclosporine) and the anti-CMV drug valgancyclovir on chronic, endogenous microbes.  From 96 heart or lung transplant recipients they collected 656 blood samples over time, some up to a year post-transplant, removed the cells, and identified remaining DNA by sequencing.  They found that 0.12% matched viral or bacterial or fungal sequences.  They validated some ‘hits’ with quantitative PCR.  Control preparations using water or bacteriophage demonstrated no relevant artifacts or contamination.  

They found that treatment with valgancyclovir reduced herpesviruses, including CMV, but dramatically increased the relative and absolute levels of anelloviruses, including TTVs (fig. 2, 3, 4).   Moreover, those patients who did not reject their transplants tended to have a greater increase in anelloviruses (Fig. 5A, shown; rejecting patients plotted in red).  The authors conclude that anellovirus levels might be used to monitor immune competence.   Focosi et al. made a related observation following autologous stem cell transplantation. 

 Cell. 2013 Nov 21;155(5):1178-87. Temporal response of the human virome to immunosuppression and antiviral therapy.  De Vlaminck I, Khush KK, Strehl C, Kohli B, Luikart H, Neff NF, Okamoto J, Snyder TM, Cornfield DN, Nicolls MR, Weill D, BernsteinD, Valantine HA, Quake SR. 

Salt develops a taste for Th17 lymphocytes

Helper T lymphocytes that make the hormone interleukin-17 (IL-17), called Th17 cells, contribute to inflammation and autoimmune diseases (review). The development of Th17 cells was known to require IL-23 but it was not known exactly why.  To gain some perspective, the authors measured gene transcripts found in Th17 cells as they develop over time from naïve mouse T lymphocytes treated with transforming growth factor-beta (TGFb) and IL-6.  They found that SGK1, an enzyme that phosphorylates proteins and has been shown to regulate sodium (Na+) transport and salt (NaCl) balance in other cells, was induced nearly 200-fold.  They emphasize that IL-23 is “critical” to the induction and maintenance of SGK1 but much of that evidence is relegated to the supplement data.  “Network analysis” with a computer program strengthened their suspicion that SGK1 is a “node” in the IL-23 signaling pathway.  

Mice without SGK1 (SGK1-knockouts, KO) have fewer Th17 cells that make less IL-17 when treated with IL-23; notably SGK1 deficiency also alters genes regulating other T cell subsets, including interferon-gamma (Ifng), Tbx21, and Gata3 (see also the previous gloss on transcription factors regulating Th17).  To test the role of SGK1, they immunized “floxed” SGK1 (conditional KO) mice with a myelin protein (MOG), which induces in some mouse strains a multiple sclerosis (MS) like disease called experimental autoimmune encephalitis (EAE).  EAE severity was significantly reduced in mice without SGK1 in Th17 cells or CD4+ helper T cells,  (fig 2a, KO score <1 normal="" vs.="">3), which corresponded with a greatly reduced number of Th17 cells in the organ targeted by this autoimmune disease, the central nervous system (CNS).  They also saw that CNS-infiltrating cells in EAE had expressed IL-17 at one time (eYFP+, fig 2e) but that expression of IL-17 by CD4 cells was lost in SGK1-KO animals (eYFP+ IL17-), suggesting that SGK1 was required to maintain expression.

That was nice but now the spice – could dietary salt modulate immunity through SGK1?    Indeed, they found that a high salt diet (HSD) accelerates the development of EAE in normal mice (fig 4e, top 2 lines, trend line offset to the left is HSD) while it does nothing to the GSK1-KO animals (lower, lines).  And connecting at least one of the dots between diet and Th17, they found that HSD also increased more than 2-fold Th17 cells and to a lesser degree interferon-gamma expressing cells in the CNS,  and the induction depended on SGK1 (fig 5f, shown, Th17 left, IFNg right; open bars SGK1-CD4-KO).  A companion paper pursued the role of dietary salt in EAE   

Nature. 2013 Apr 25;496(7446):513-7.  Induction of pathogenic TH17 cells by inducible salt-sensing kinase SGK1.  Wu C, Yosef N, Thalhamer T, Zhu C, Xiao S, Kishi Y, Regev A, Kuchroo VK.

Recipe for Developing Th17 cells

Thymus-dependent “T” lymphocytes develop into several effector and regulatory lineages, including the well-characterized regulatory “helper” T (Th) cells that express the cellular differentiation marker 4 (CD4+) and CD8+ cytotoxic T lymphocytes (CTL, or Tc) that kill virus-infected cells. The CD4+ Th lineages further differentiate into Th1, Th2, and Treg cells that help protect against intracellular microbes, or helminthes, or specifically regulate immune responses, as well as Th17 cells, so-called because they make the interleukin-17 (IL17) that is required for protecting the mucosa against infection by bacteria and fungi.  

Development of cell lineages is controlled externally by cytokines and internally by “master” transcription factors.  RORgt (retinoic-acid-receptor-related orphan receptors gamma t) is expressed by Th17 cells and forced expression of RORgt gene in naïve CD4+ T cells (Th0) makes them express some genes characteristic of Th17 cells such as the IL-23 receptor and the chemokine receptor CCR5 but not the full range of Th17 products, which requires other TFs including STAT3, IRF4, BATF, and IkappaBzeta.  Other TFs may replace these for inducing some Th17 genes.  The myriad of TFs required for more or less full Th17 function led these investigators to try to sort out how they work together. 

The authors first looked where these implicated TFs bind on the genome of Th0 cells treated with Th17-inducing cytokines using chromosome immune-precipitation (ChIP).  They then compared genes that are transcribed, measured using RNA seq, in the absence of specific TFs, reduced using siRNA, to “build a network model for Th17 cells”.  

They propose that TFs BATF and IRF4 bind cooperatively and open chromosomes to STAT3, which drives transcription of many genes including the lineage-specifying TF RORgt.  They also identify several putative new Th17 regulators, including the AP-1 family member Fosl2, which they belive is a key TF for Th17 development.  They derive many complicated, colorful figures.  A largely understandable, intriguing single figure is 5D, copied here, which shows that a block of Th17-related genes is increased (red) or decreased (blue) “log2 fold” (NB the genes, named on the right, are NOT the same) in Th17 cells but not other T cell subsets (Th1, Th2) treated with siRNA suppressing Satb1 (a “chromatin organizer"), Bcl11b (a zinc finger TF), Jmjd3 (a histone demethylase), and the old familiar RorC (encoding RORg).   But why no siRNA for Fosl2? 

A validated regulatory network for th17 cell specification.  Ciofani M, Madar A, Galan C, Sellars M, Mace K, Pauli F, Agarwal A, Huang W, Parkurst CN, Muratet M, Newberry KM, Meadows S, Greenfield A, Yang Y, Jain P, Kirigin FK, Birchmeier C, Wagner EF, Murphy KM, Myers RM, Bonneau R, Littman DR.   Cell. 2012 Oct 12;151(2):289-303.

The peptide binding site accounts for the MHC link to Rheumatoid Arthritis

That genes within the major histocompatibility complex (MHC, human HLA) influence susceptibility to rheumatoid arthritis (RA) has been known for over 40 years, even before HLA nomenclature was well established (e.g., Dick et al. 1975). However, the few “classical” HLA genes constitute only a small fraction of the hundreds genes within the MHC, which include the inflammatory cytokine tumor necrosis factor (TNF), a key player in RA. Which genes are responsible for the association?

The authors investigated 5,018 “cases” of RA, all with antibodies against cyclic citrullinated peptides, CCP (i.e, seropositive, accounting for 70% of RA patients [review]) and 14,974 health, ethnically matched controls.

First they tested their ability to “impute” HLA alleles from their SNP data using a reference panel of 2,767 individuals. Conclusion, not bad: 98% accuracy for “two digit” mapping and >80% for 4 digit (allele). Then they found the most significant nucleotide (p<10^-526!) is part of a codon for amino acid 11 of the HLA-DR beta 1, and thus not part of the “shared epitope” (a 5-amino acid sequence and antibody epitope linked to RA [review]). A valine at this position confers a 3.8-fold higher risk whereas a (polar) serine is protective (the converse of risk). Comparing cases and controls shows a clear difference (shown here, from fig 3). Adding amino acids at positions 71 and 74 improved significance slightly, and alleles with these amino acids were independently shown to confer risk.

The authors conclude “These results are consistent with a disease model in which classical HLA genes and proteins are the dominant factors in rheumatoid arthritis pathogenesis, with only a minor contribution from non-HLA loci in the MHC”. It seems that someone might also explore whether these variants bind CCP better!

Raychaudhuri S, Sandor C, Stahl EA, Freudenberg J, Lee HS, Jia X, Alfredsson L, Padyukov L, Klareskog L, Worthington J, Siminovitch KA, Bae SC, Plenge RM, Gregersen PK, de Bakker PI. “Five amino acids in three HLA proteins explain most of the association between MHC and seropositive rheumatoid arthritis.” Nat Genet. 2012 Jan 29;44(3):291-6

ENCODE salvages “junk” DNA

The “ENCyclopedia Of DNA Elements”, ENCODE, founded in 2003 with grants from the NIH Genome Institute, seeks to identify all the functional parts of the human genome, assessed by DNA and histone modifications, chromatin looping, transcription factor binding, chromatin compaction (DNAse accessibility), and transcripts.  The collaboration of ~37 groups, first developed technology. Recently they published their first salvo of 30 research papers, several published in Nature along with a News & Views.

The paper by Djebali and scores of colleagues offers “a genome-wide catalogue of human transcripts”, together with their location (nucleus or cytoplasm), and whether they have a 7mG cap 5’ or a poly-A tail 3’.  They prepared RNA from 15 human cells lines after fractionation (whole cell, nucleus and cytosol) and separation of RNA into short and long (>200 nucleotides).  Long RNAs were further separated into +/- polyA tails.  They sequenced these RNAs and determined their initiation sites and their 5’ and 3’ termini (using technologies felicitously named CAGE and PET).  Then they did bioinformatics: compared to annotated genome (GENCODE) statistics, etc.,  All these data are available for your perusal using the RNA Dashboard. 

They made many interesting observations; e.g., they conclude there is very little “junk” DNA.  Nearly 75% of the genome is transcribed in at least one of the cell lines, though only a little over 50% in any given line.  (This is similar to previous findings, albeit not as “encyclopedic”).  Only 28% of the 7,053 small RNAs (including snRNAs, snoRNAs, miRNAs, and tRNAs) annotated by GENCODE are found in any of these cell lines, suggesting the expression of many annotated small RNAs is cell type specific.   

They also find that protein-coding transcripts are more abundant than long non-coding RNAs (lncRNAs) and that the same genes are transcribed in different cells.  Figure 3, shown here, plots the number of transcripts (r.p.k.m., reads per kilobase per million reads) on the x axis vs. the ratio of nuclear/cytoplasmic for protein-coding (orange), which are abundant (right) in the cytoplasm (down),  non-coding (blue), and novel intergenic (green), which tend to be expressed at lower levels (left) and mostly nuclear (up).  A few individual transcripts are also identified, giving appreciation for the range of expression.

Also not for the first time, they suggest that shrinking “intergenic” regions  “prompts the reconsideration of the definition of a gene”.   They “propose that the transcript be considered as the basic atomic unit of inheritance” and that “gene … denote … all those transcripts …. that contribute to a given phenotypic trait".  Mendel would approve. 

PubMed. Djebali et al. "Landscape of transcription in human cells." 2012 Nature Sep 6;489(7414):101-8.

Monkey’s Uncle? Your HLA might be Neanderthal

Humans’ genomes are extremely similar to those of other primates because the species diverged relatively recently, approximately 6 million years ago in the case of our nearest cousins, Chimpanzees. Modern humans and Denisovans separated 250,000 years ago (10,000 generations). With the recent sequencing of extinct, ancient hominids, such as Neanderthals and their Denisovan relatives, it was realized that up to 6% of the genomes of humans now in Europe and Asia derive from these older lineages.

HLA genes are by far the most polymorphic within the human genome, with thousands of variants (alleles). Here, investigators first identified one particular HLA allele, HLA-B*73:01, as being more similar to homologous Chimpanzee alleles than other human HLA-B alleles. This allele diverged from other HLA-B alleles 16 million years ago, before the separation of humans and Chimps, and was lost from the majority of modern humans. Its reappearance in the human genome was most likely, they reckoned, a result of “introgression”, introduction from ancient humans such as Neanderthal. An alternative model, which computer simulations indicate is 100 times less probable, is that this allele came out of Africa late

They also simply “typed” (sequenced and matched) the most important HLA loci, HLA-A, -B and –C from 1 Denisovan and 2 Neanderthal subjects. Surprisingly, most of these archaic HLA alleles were identical to common HLA types of modern humans. HLA-A2, the most widespread allele at the HLA-A locus, was shared with and might have been acquired from Denisovans. Putative archaic HLA-A alleles are now more common in China and Europe than in Africa (Figure, from fig. 4d). The authors conclude that although a small minority of our genomes overall derived from archaic humans, about half of our HLA was acquired through interbreeding between modern humans migrating out of Africa and locally established archaic humans. These archaic alleles conferred fitness in the new environment, e.g., pathogen and allergen resistance, and so outcompeted and displaced previous human HLA alleles.


The shaping of modern human immune systems by multiregional admixture with archaic humans. Abi-Rached et al. Science. 2011 Oct 7;334(6052):89-94.

Gut Feeling – intestinal microbes influence immune system tolerance of central nervous system

Multiple Sclerosis (MS: Wikipedia, PubMedHealth) is an autoimmune disease wherein lymphocytes attack the central nervous system (CNS), including the brain and spinal cord, leading to relapsing, progressive loss of neurons. Lesions containing B and T lymphocytes develop in the CNS. The cause of MS is unknown.

A mouse model of MS, called experimental autoimmune encephalomyelitis (EAE), can be induced when mice of certain strains are immunized with spinal cord proteins, or it can occur spontaneously in genetically engineered strains in which many CD4+ “helper” T cells express a transgenic T cell receptor specific for myelin oligodendrocyte glycoprotein (MOG), a protein abundant on the surface of key non-neuronal cells of the CNS.

These authors observed that depending on the animal housing facility, between 35-90% of MOG-specific-TCR-transgenic mice spontaneously develop EAE at between 3-8 months of age. The wide range in the disease incidence reminded the authors of a 1993 report by Goverman that mice with T cells expressing transgenic antigen receptors specific for another nerve protein, MBP, developed EAE ‘spontaneously’ in non-sterile housing but not in sterile housing.

They compared EAE incidence in mice that possess normal gut microbes but harbor no known pathogens, termed Specific Pathogen Free (SPF), and mice that possess no microbes at all, termed “germ free” (GF), and found that GF mice were protected (Fig 1a, shown, left panel).

Gut microbes are known to contribute to lymphocyte maturation, stimulated by , e.g., segmented filamentous bacteria) or polysaccharides of Bacteroides fragilis. However, the authors argue this does not explain protection because GF mice colonized with “conventional commensal” microbes developed EAE “promptly”, starting about a month later (Fig 1b, shown, right panel). They add that colonization with segmented filamentous bacteria – shown to trigger autoimmunity in another model – conferred EAE susceptibility only inefficiently. They also argue that GF mice immunized with MOG in complete adjuvant develop EAE (though again with a delay of about a month) and produce specific antibodies (though measured crudely, not titered), demonstrating that their lymphocytes are mature.

Instead, the authors argue that some lymphocyte activities are reduced in GF mice, particularly T cell production of the pro-inflammatory interleukin-17 and spontaneous B cell production of MOG-specific antibodies (which is also “promptly” albeit only partially corrected by colonization, Fig 3a). Moreover, MOG-specific B cells – but not polyclonal normal B cells – transferred into MOG-specific-TCR-transgenic mice – but not MOG-deficient mice – home to germinal centers where they mature and make antibodies that are IgG2a class-switched, and therefore implicitly effective in cooperating with specific T cells to induce EAE. They conclude that commensal gut microbes activate autoreactive T cells that recruit autoreactive B cells, which together mediate disease.

Berer K, Mues M, Koutrolos M, Rasbi ZA, Boziki M, Johner C, Wekerle H, Krishnamoorthy G. Commensal microbiota and myelin autoantigen cooperate to trigger autoimmune demyelination. Nature. 2011 Oct 26;479(7374):538-41. doi: 10.1038/nature10554. PubMed PMID: 22031325

Would you like some E. coli with that?

An epidemic of bloody stools and failing kidneys, some with hemolytic uremic syndrome (HUS) appeared in Germany in May 2011 and subsequently 15 other countries. By late July when the epidemic had subsided, a total of 3,816 cases - including 54 deaths - were reported in Germany, 845 of which included HUS. Rasko and colleagues cultured E. coli bacteria isolated from a 64 year old woman from Hamburg, Germany, who did not develop HUS. They characterized this bacterium, designated C227-11, as enteroaggregative, which means a gut pathogen that aggregates and forms “biofilms” that are resistant to treatment.

They sequenced the bacterium’s genome and found it was a unique strain of the O104:H4 serotype of E. coli bacteria, distinguished by possession of a prophage (http://en.wikipedia.org/wiki/Prophage ) producing the Shiga toxin. Shiga toxin binds to cells, inhibits protein synthesis, and kills by inducing apoptosis [review]. The O104 serotype is rare; the most frequent cause of HUS worldwide is the shiga-toxin–producing E. coli O157 (Tarr 2005).
Although they isolate only one strain themselves, they analyzed also 3 additional sequences from the current outbreak that had been made public (that’s data mine-ing!) together with 7 other O104:H4 serotype isolates, all from Africa, and 4 other reference strains. The authors conclude that the outbreak was caused by a difficult (enteroaggregative) strain made more virulent by its acquisition of the Shiga toxin gene in addition to antibiotic-resistance and “additional virulence and antibiotic-resistance factors”. Rohde and colleagues reached the same conclusion using "rapid, bench-top DNA sequencing technology, open-source data release, and prompt crowd-sourced analyses".

Where did the E. coli O104:H4 come from? A subsequent publication reported the results of trace-back and –forward investigations by Buchholz and colleagues who analyzed 26 HUS patients and 81 healthy controls. They concluded that despite only about a quarter of the patients recalling in exploratory interviews having eaten bean sprouts during the 14 days before the onset of illness, 100% of these illnesses were attributable to the consumption of sprouts – and not other raw foods such as tomatoes or cucumbers or lettuce – at a particular restaurant, and for other patients, sprouts obtained from a single, common supplier (figure).

N Engl J Med. 2011 Aug 25;365(8):709-17. Origins of the E. coli strain causing an outbreak of hemolytic-uremic syndrome in Germany. Rasko DA, Webster DR, Sahl JW, Bashir A, Boisen N, Scheutz F, Paxinos EE, Sebra R, Chin CS, Iliopoulos D, Klammer A, Peluso P, Lee L, Kislyuk AO, Bullard J, Kasarskis A, Wang S, Eid J, Rank D, Redman JC, Steyert SR, Frimodt-Møller J, Struve C, Petersen AM, Krogfelt KA, Nataro JP, Schadt EE, Waldor MK.

Helpful mutants usually won’t cooperate with each other

How genes interact – the phenomenon termed “epistasis” – is complex, yet must be understood to accurately interpret the contribution of individual genes to the development and behavior of the organism. Two groups -- Chou et al. and Khan et al. -- recently reported in Science the results of their remarkably similar experiments; they isolated individual, beneficial mutations in bacteria then tested how the mutations interact when possessed by the same bacterium.

Previous studies (cited by Chou) suggested that two deleterious mutations in the same pathway are generally less-than-additive but were greater-than-additive when they were in parallel pathways, which seems intuitive. Interactions between mutations in single genes were shown to depend on the background (other genes). However, epistasis among beneficial mutations in different genes was “unexplored”, and the focus of these new experiments. Evolution in laboratory conditions is initially rapid but quickly slows, which fits a model of mutually antagonistic beneficial mutations. Khan states the deceleration is due to either (1) negative epistasis or (2) because the most beneficial would “tend to be incorporated earlier owing to their faster spread and greater success in the face of competing beneficial mutations” (which sounds suspiciously convenient – the best arrive early – but they give a reference: Gerrish 1998 so you can look it up).

Khan grew E. coli with a glucose supplement for 20,000 generations, when they sequenced a clone and identified 45 mutations. They say other beneficial mutations arose but were lost due to “interference” with more-beneficial mutations (24, 26) (which seems to answer their question), or because they were “less able to evolve than the eventual winners” (33). They took the first 5 “that fixed … and whose spread coincided with the period of fastest adaptation” (arbitrary? How many generations was that, 200? 2,000?), which were, in order of appearance: rbs operon, topA, spoT, glmUS promoter, and pykF. Together, these 5 mutations increased fitness ~30%, accounting for ~80% of the fitness increase over the full 20,000 generations.

Then they produced 32 populations of E. coli, one for each possible combination of the 5 genes (Fig. 1 from Khan, copied here, shows the ancestral genotype at the top and the 32 combinations of mutations, with increasing fitness downward). They found that although each combination improved fitness, improvement was less than expected from a “multiplicative null model” (in which the individual fitness effects are multiplied).

Similarly, Chou selected a bacterium to grow with methanol as its sole carbon source and then replaced a key metabolic pathway with a less efficient pathway from a different bacterium. The resulting bacterium grew only one-third as fast as the original. Eight separate populations improved the efficiency over 600 generations. In the fittest strain, 9 mutations were identified; 3 were clearly related to the pathway while 6 others were deemed “unlikely” to contribute to fitness. To determine the interactions among these 3 genes, they constructed 16 strains, one with each combinations of these 3 mutations plus the WT allele. They also observed that combination strains were much less fit than expected if they were acting independently (our old favorite, the multiplicative null model).

These findings should influence our expectations in genetic studies in humans; e.g., two disease-risk alleles may have little additive or even negative effects (ergo protective?), contradicting the simple expectations of most biologists. [Chou distinguished beneficial from detrimental effects.] I wonder if these experiments might be misleading because they first fixed the individual mutants, thereby eliminating mutants that interact positively during selection.

Khan AI, Dinh DM, Schneider D, Lenski RE, Cooper TF."Negative epistasis between beneficial mutations in an evolving bacterial population. " Science. 2011 Jun 3;332(6034):1193-6.

The same issue has a commentary.