Content uploaded by Geraint Rogers
Author content
All content in this area was uploaded by Geraint Rogers on Dec 21, 2015
Content may be subject to copyright.
news and views
nature medicine volume 21 | number 8
|
AuGuST 2015 839
represented MLGs revealed substantial micro-
biome divergence between RA patients and
healthy controls, not only in feces but also in
salivary and dental samples. Further, analysis
revealed common elements to the dysbiosis
seen at each site, including the depletion in
Haemophilus species and an overrepresenta-
tion of Lactobacillus salivarius, which were
most marked in severe cases
3
.
Fecal samples from RA patients were found
to be enriched in Gram-positive bacteria and
depleted in Gram-negative when compared
with control individuals, including some
Proteobacteria and Gram-negative Firmicutes.
Notably, P. c o p r i in RA-affected individuals
showed a trend of increasing relative abundance
in the first year, consistent with its reported
expansion in NORA patients
2
. Owing to its
ability to generate citrullinated neoantigens, the
presence of oral Porphyromonas gingivalis has
previously been proposed to be a factor under-
pinning the association of periodontal disease
with RA
6
. However, the reported enrichment of
P. gingivalis in control oral samples is in keep-
ing with a number of recent studies refuting
this hypothetical link
7
. Across sites, a number
of MLGs that were enriched in the healthy
samples correlated negatively with markers of
acute inflammation (C-reactive protein) and
RA-specific autoantibodies (anti-cyclic cit-
rullinated peptide (CCP) and/or rheumatoid
factor), whereas some MLGs enriched in RA
individuals showed positive correlations with
anti-CCP, rheumatoid factor, IgG and IgA.
These findings suggest potential utility for
MLGs as markers for RA pathophysiology.
The dysbiosis that was observed at each
of the three sample sites was accompanied
by shifts in the relative abundance of genes
encoding specific functional traits
3
. In fecal
samples from control individuals, modules
RA is a common debilitating autoimmune
disorder associated with progressive disabil-
ity, systemic complications, and early death.
Although its etiology remains elusive, recent
studies have highlighted the potential influence
of the mucosal microbiome on RA onset and
progression
1,2
. In this issue of Nature Medicine,
Zhang et al.
3
report a case-control metage-
nome-wide association study (MGWAS) of
the fecal, dental and salivary microbiome in a
large cohort of treatment-naïve and treated RA
patients. They reveal the RA-associated micro-
biome to deviate substantially from healthy con-
trols in all sites; they show that these imbalances
can be partially redressed by disease-modifying
antirheumatic drugs (DMARDs).
The development of RA is associated with
the dysregulation of normal immune func-
tion, involving an increased production of both
self-reactive antibodies and pro-inflammatory
T lymphocytes. Genome-wide association
studies (GWASs) have identified a number
of susceptibility alleles, including those pre-
disposing T cell repertoire selection, antigen
presentation, or alteration in peptide affinity
4
.
However, genetic predisposition is not sufficient
for RA development, as the genetic heritability
is estimated to be approximately 60%, based on
concordance in monozygotic twins
5
. It is likely
that an environmental trigger is necessary
for disease onset in genetically predisposed
individuals. An intriguing possibility is that
such a trigger might be provided by abnormali-
ties in the bidirectional cross-talk between the
host and the resident microbiome (Fig. 1).
The concept that bacterial populations,
particularly those in the gut, could contribute
to the development of autoimmune arthritis
is more than a century old and is supported
by a large body of research in both humans
and animal models. Notably, multiple studies
using arthritis-prone animals have shown
that bacterial colonization is a requirement
for the emergence of disease
1
. Only recently,
however, sequencing and computational capa-
bilities have allowed the complex relationships
between the microbiome, immune modula-
tion, and pathogenesis to be investigated.
The potential of a MGWAS approach was high-
lighted by a recent study comparing the intestinal
microbiome in RA-affected individuals and
controls that showed an overrepresentation
of Prevotella species (particularly P. c o p r i ) in
the feces of new-onset rheumatoid arthritis
(NORA) patients, and a concurrent reduction
in Bacteroides species
2
.
Whereas the mucosal surface of the gut offers
arguably the greatest potential for host-microbe
interactions, almost all mucosae have resident
microbiota and respond dynamically to their
presence through immune modulation. In this
study Zhang et al.
3
therefore aimed to assess
whether the dysbiosis reported in the gut of
RA patients was also evident in the mouth—a
highly plausible hypothesis given the frequent
co-occurrence of RA and periodontal disease,
and the shared pathogenic mechanisms and
immunologic pathways of these conditions
1
.
By comparing the differential representation
of metagenomic linkage groups (MLGs are
groups of genetic material in a metagenome
that are probably physically linked; MLG is a
term used in lieu of ‘species’) constructed from
shotgun sequence data, the authors were able to
identify bacterial markers that were enriched
in each sample site
3
. Analysis of differentially
Germs and joints: the contribution of the human
microbiome to rheumatoid arthritis
Geraint B Rogers
Rheumatoid arthritis (RA) is a debilitating autoimmune disorder, the etiology of which is poorly understood. A new
study reveals dysbiosis in gut and oral microbiomes of affected individuals, potentially providing a basis for patient
stratification and clues to pathophysiological mechanisms of RA onset and progression.
Geraint B. Rogers is in the Infection & Immunity
Theme of the South Australian Health and Medical
Research Institute, and is based at the School of
Medicine, Flinders University, Adelaide, Australia.
e-mail: geraint.rogers@sahmri.com
npg
© 2015 Nature America, Inc. All rights reserved.
news and views
840 volume 21 | number 8
|
AuGuST 2015 nature medicine
zinc-dependent matrix metalloproteinases
(MMPs), although the paths by which these could
contribute to cartilage destruction at remote
sites of inflammation remains unclear. Finally,
the relative distribution of MLGs is also infor-
mative with respect to the molecular mimicry
hypothesis, in which the microbial production
of cross-reactive epitopes results in an immune
response against self-antigens
8
. The potential for
this phenomenon to contribute to RA is sup-
ported by the enrichment of both gut and oral
RA microbiomes with sequences that mimic
key motifs in the RA-related human antigens.
Interestingly, Zhang et al.
3
found that
MLG distribution changed significantly with
DMARD treatment (methotrexate and/or
Tripterygium wilfordii glycosides), with an
increase in the abundance of control-associated
microbiome markers. These changes were
responsible for lipopolysaccharide biosynthesis
and transport, and type II, III, and IV secretion
systems were enriched, in accordance with the
relative depletion of Gram-negative bacteria in
RA. To some extent, such differences in gene
relative abundance represent indirect mark-
ers of bacterial taxon distribution. However,
they also provide insight into potential drivers
of dysbiosis. L. salivarius, for example, is toler-
ant of the high reactive oxygen species (ROS)
concentrations found at sites of inflammation.
The high relative abundance of this species in
RA-affected individuals may therefore reflect an
elevated immune response at the sites sampled.
In addition, the distribution of functional traits
provides clues to possible bacterial contribu-
tions to pathogenesis. The enrichment of zinc
transport systems genes in RA oral samples
could, for example, influence the activity of
most marked in patients who showed the
greatest clinical improvement. Such findings
highlight the likely bidirectional relationship
between the microbiome and RA, and they
underline the need for caution when assign-
ing causality based on microbial association.
Perhaps most importantly, models based on
gut, dental or salivary MLGs were predictive
of treatment response, with disease classifiers
constructed from sequence data comparable or
superior to existing RA serum markers
3
.
The potential for members of the gut or
oral microbiota to contribute to pathogenesis
or to beneficially modulate immune response
in RA raises the possibility of microbiome
manipulation for therapeutic benefit. Probiotic
administration of immunomodulatory spe-
cies has shown some promise in autoimmune
diseases
9
. However, the complexity of the
Gut dysbiosis
Clostridium asparagiforme,
Gordonibacter pamelaeae,
Eggerthella lenta, Lachnospiraceae
bacterium, Bifidobacterium dentium,
Lactobacillus, Ruminococcus lactaris
Veillonella, Haemophilus, Klebsiella
pneumoniae, Bifidobacterium bifidum,
Sutterella wadsworthensis,
Megamonas hypermegale
Environmental
triggers
Genetic
predisposition
e.g., HLA-DR4/1,
PADI4, TNF-α
Immune
dysregulation
e.g., autoimmune antibodies,
dysregulation of B and
T cell populations
Inflammation
e.g., TNF-α, IL-1, IL-8,
MMPs, CRP
Comorbidities
Cardiovascular disease,
liver and metabolic
function, periodontitis
DMARD
RA
Microbiome of
RA-affected individual
Oral dysbiosis
Veillonella, Atopium,
Lactobacillus salivarius,
Cryptobacterium curtum,
Selenomonas flueggei
Haemophilus, Aggregibacter,
Cardiobacterium, Eikenella,
Kingella, Neisseria, Leptotrichia,
Porphyromonas gingivalis,
Rothia aeria, Capnocytophaga
ochraea
Molecular mimcry
Collagen XI, HLA-DR4/1
Inflammation
Reduction of tolerogenic species
Increases in pro-inflammatory
species
Altered mucosal environment
e.g., high levels of ROS, altered redox
environment
Production of
immunomodulatory molecules
e.g., hydrogen sulfide, homocysteine,
altered SCFA production
Figure 1 The potential influence of the human microbiome on rheumatoid arthritis. Zhang et al.
3
find that in rheumatoid arthritis there is disruption
to normal microbiome composition (dysbiosis) in the gut and mouth. This can contribute to immune dysregulation, both through molecular mimicry
(the microbial production of epitopes cross-reactive with self antigens, resulting in an immune response against self-antigens) and through changes
in the relative abundance of tolerogenic and pro-inflammatory bacterial species. Dysbiosis may also contribute to disease onset through differences in
the functional characteristics of the constituent species, including the ability to produce metabolites that are pro-inflammatory (for example, hydrogen
sulfide, homocysteine) or immunomodulatory (for example, short-chain fatty acids (SCFAs)). The ability of inflammation to shape the microbiome is
reflected in compositional shifts identified by Zhang et al.
3
in patients receiving DMARDs. Proposed environmental risk factors may also act, in part,
by contributing to dysbiosis, as might genetic predisposition. These may also lead to comorbidities. CRP, C-reactive protein; HLA-DR, human leukocyte
antigen death receptor; IL, interleukin; PADI4, peptidyl arginine deiminase, type IV; TNF-α, tumor necrosis factor.
npg
© 2015 Nature America, Inc. All rights reserved.
news and views
nature medicine volume 21
|
number 8
|
AuGuST 2015 841
similar environmental drivers, translated
according to genetic predisposition. Such a
model could explain common bidirectional
RA comorbidities with underlying inflam-
matory components, most notably, heart
disease and depression
13
. The development
of the case-control MGWAS now offers an
exciting opportunity to identify common-
alities in microbiome characteristics across
these conditions, and to extend analysis of
relationships between dysbiosis and inflam-
matory dysregulation to other body sites. The
extension of shotgun sequencing approaches
to metatranscriptomic analysis, allowing
characterization of not only bacterial func-
tional potential but also differential gene
expression (both microbial and human), may
provide yet-greater insight and represent a
logical next step.
relationship between the microbiome and
immune regulation suggests that a more sophis-
ticated approach may be apposite. The capacity
to engineer a complete artificial microbiome for
introduction to a specific body site, or to use
synthetic biology to incorporate inducible ther-
apeutic traits in the microbiome, are intriguing
possibilities
10
.
The work by Zhang et al.
3
has importance
beyond RA. Aberrant immune regulation
resulting from disruption of the healthy
microbiome may be a common underlying
feature of many of the idiopathic inflam-
matory disorders, including metabolic syn-
drome, type 2 diabetes, atherosclerosis
11
and
inflammatory bowel disease
12
. It is interest-
ing to consider whether these conditions
might be manifestations of a common dys-
biotic phenomenon, perhaps resulting from
Competing finanCial interests
The author declares no competing financial interests.
1. Brusca, S.B., Abramson, S.B. & Scher, J.U. Curr. Opin.
Rheumatol. 26, 101–107 (2014).
2. Scher, J.U. et al. eLife 2, e01202 (2013).
3. Zhang, X. et al. Nat. Med. 21, 895–905 (2015).
4. Wellcome Trust Case Control Consortium. Nature 447,
661–678 (2007).
5. MacGregor, A.J. et al. Arthritis Rheum. 43, 30–37
(2000).
6. Rosenstein, E.D., Greenwald, R.A., Kushner, L.J. &
Weissmann, G. Inflamm. 28, 311–318 (2004).
7. Konig, M.F., Paracha, A.S., Moni, M., Bingham,
C.O. III & Andrade, F. Ann. Rheum. Dis. doi:10.1136/
annrheumdis-2014–205385 (26 May 2014).
8. Kamphuis, S. et al. Lancet 366, 50–56 (2005).
9. Kverka, M. & Tlaskalova-Hogenova, H. APMIS 121,
403–421 (2013).
10. Sonnenburg, J.L. Nature 518, S10 (2015).
11. Devaraj, S., Hemarajata, P. & Versalovic, J. Clin. Chem.
59, 617–628 (2013).
12. Hedin, C. et al. Gut doi:10.1136/gutjnl-2014–308896
(8 April 2015).
13. Halaris, A. Curr. Psychiatry Rep. 15, 400 (2013).
Supporting itch: a new role for astrocytes in chronic itch
Dustin Green & Xinzhong Dong
A new study shows that astrocytes are involved in the development of chronic itch in a mouse model. This is
dependent on upregulation of lipocalin 2 (LCN2) by the transcription factor STAT-3 and astrogliosis.
Dustin Green is in the Solomon H. Snyder
Department of Neuroscience, Department
of Neurosurgery, Department of Dermatology
and Center for Sensory Biology, Johns Hopkins
University, School of Medicine, Baltimore,
Maryland, USA. Xinzhong Dong is in the
Solomon H. Snyder Department of Neuroscience,
Department of Neurosurgery, Department of
Dermatology, Center for Sensory Biology and
the Howard Hughes Medical Institute,
Johns Hopkins University, School of Medicine,
Johns Hopkins University School of Medicine,
Baltimore, Maryland, USA.
e-mail: dgreen89@jhmi.edu or xdong2@jhmi.edu
Acute itch, or pruritus, serves as a biological
warning that protects against parasites,
disease-carrying insects and poisonous plants.
The past decade has provided new insights
into the molecular mechanisms that under-
lie the transmission of itch, identifying new
receptors and neural circuits
1–3
. Although
acute itch is protective, chronic itch can be a
devastating illness. Pruritus that lasts longer
than 6 weeks is deemed chronic and is accom-
panied by a host of co-morbidities includ-
ing depression, sleep and anxiety disorders
4
.
Millions of people suffer from chronic itch, and
it is associated with a wide variety of diseases,
ranging from diabetes and cancer to kidney
disorders
5
. Up until now, our understanding
of how acute itch transitions into a state of
chronic itch has not been well understood.
Here Shiratori-Hayashi and colleagues pres-
ent exciting work that links astrocytes to the
maintenance of chronic itch
6
.
Previous studies of itch have focused on the
peripheral afferent and spinal neural circuits
and molecular pathways. Characterizing pruri-
tus has been no easy task, with studies showing
multiple receptors responsible for mediating
differing itch-inducing substances such as
histamine or the malarial drug chloroquine
7,8
.
Moreover, itch neurons seem to be a hetero-
geneous population with subtypes that share
only certain commonalities. However, the con-
sensus at present is that these peripheral itch
neurons project signals to the dorsal horn, an
area of the spinal cord that receives and pro-
cesses sensory information. Shiratori-Hayashi
and colleagues
6
took a different approach to
study chronic itch by investigating the role of
glia—specifically astrocytes—a type of support
cell in the central nervous system.
To explore the role of non-neuronal cells in
maintaining itch, the authors used a mouse
model of atopic dermatitis. These mice, when
removed from a pathogen-free environment,
developed the hallmarks of human chronic
itch, including repetitive spontaneous scratch-
ing and skin lesions. Using this animal model,
Shiratori-Hayashi et al.
6
then examined how
the morphology of astrocytes in the spinal cord
had changed. When the spinal cord was fluores-
cently labeled and examined using a confocal
microscope, the astrocytes displayed enlarged
cell bodies and overly arborized processes char-
acteristic of astrogliosis, a type of change in glia
that occurs only after injury or infection (Fig. 1).
Compared to what they observed in healthy
mice, the authors also saw increased levels of
the astrocyte marker GFAP in segments of the
spinal cord that signal to areas of the body that
the animal scratched
6
. This phenotype could
also be recapitulated by injecting wild-type
mice with diphenylcyclopropenone (DCP), an
immunotherapy drug known to produce pruri-
tus as a side effect. However, glia from segments
of the spinal cord that did not show itch behav-
ior were normal. Up until now, astrogliosis has
primarily been thought of as a feature found
in nerve-injury models of chronic pain; how-
ever, the findings here also point to it being an
important marker for chronic itch.
To study the possible role that scratching
has in producing astrogliosis, the investiga-
tors ablated a subset of itch-sensing nerve
fibers expressing the ion channel TRPV1,
using high concentrations of the potent plant
toxin resiniferatoxin. Mice whose TRPV1
+
fibers were abolished showed both reduced
GFAP expression and scratching behavior.
These results provide a possible link between
npg
© 2015 Nature America, Inc. All rights reserved.
