REVIEW
Airway remodeling in asthma: what really matters
Heinz Fehrenbach
1,2
& Christina Wagner
3,4
& Michael Wegmann
2,5
Received: 20 October 2016 /Accepted: 21 December 2016 / Published online: 11 February 2017
#
The Author(s) 2017. This article is published with open access at Springerlink.com
Abstract Airway remodeling is generally quite broadly de-
fined as any change in composition, distribution, thickness,
mass or volume and/or number of structural components ob-
served in the airway wall of patients relative to healthy indi-
viduals. However, two types of airway remodeling should be
distinguished more clearly: (1) physiological airway remodel-
ing, which encompasses structural changes that occur regular-
ly during normal lung development and growth leading to a
normal mature airway wall or as an acute and transient re-
sponse to injury and/or inflammation, which ultimately results
in restoration of a normal airway structures; and (2) patholog-
ical airway remodeling, which comprises those structural al-
terations that occur as a result of either disturbed lung devel-
opment or as a response to chronic injury and/or inflammation
leading to persistently altered airway wall structures and func-
tion. This review will address a few major aspects: (1) what
are reliable quantitative approaches to assess airway remodel-
ing? (2) Are there any indications supporting the notion that
airway remodeling can occur as a primary event, i.e., before
any inflammatory process was initiated? (3) What is known
about airway remodeling being a secondary event to inflam-
mation? And (4), what can we learn from the different animal
models ranging from invertebrate to primate models in the
study of airway remodeling? Future studies are required ad-
dressing particularly pheno-/endotype-specific aspects of air-
way remodeling using both endotype-specific animal models
and B endotyped^ human asthmatics. Hopefully, novel in vivo
imaging techniques will be further advanced to allow moni-
toring development, growth and inflammation of the airways
already at a very early stage in life.
Keywords Asthma
.
Airway remodeling
.
Airway pathology
Introduction
With more than 241 million prevalent cases in 2013 (Global
Burden of Disease Study 2013 Collaborators 2015), asthma is
one of the most common lung diseases worldwide and it is
expected that the number of people suffering from this chronic
lung disease increases to about 300 million in 2025 (Croisant
2014). In 2013, asthma ranked 15th among all diseases world-
wide with regard to years lived with the disability (YLD) over
all age groups (Global Burden of Disease Study 2013
Collaborators 2015). Notably, asthma is most prevalent in
children with highest YLDs observed in the age group 5–14
years (ranked 6th worldwide, even 2nd in the group of devel-
oped cou ntries) (see http://vizhub.healthdata.org/gbd-
compare/). On the basis of disability-adjusted life years, the
impact of asthma is estimated being similar to other major
chronic diseases such as diabetes or Alzheimer disease
(Croisant 2014). Even though asthma is generally not seen
as a major ca use of death, the Global Burden of Disease
* Heinz Fehrenbach
hfehrenbach@fz-borstel.de
1
Division of Experimental Pneumology, Priority Area Asthma &
Allergy, Research Center Borstel, Leibniz Center for Medicine and
Biosciences, Parkallee 1-40, 23845 Borstel, Germany
2
Airway Research Center North (ARCN), German Center for Lung
Research (DZL), Borstel, Großhansdorf, Kiel, Lübeck, Germany
3
Junior Research Group of Invertebrate Models, Priority Area Asthma
& Allergy, Research Center Borstel, Leibniz Center for Medicine and
Biosciences, Parkallee 1-40, 23845 Borstel, Germany
4
Leibniz-ScienceCampus Evolutionary Medicine of the Lung
(EvoLUNG), Kiel, Germany
5
Junior Research Group of Asthma Mouse Models, Priority Area
Asthma & Allergy, Research Center Borstel, Leibniz Center for
Medicine and Biosciences, Parkallee 1-40, 23845 Borstel, Germany
Cell Tissue Res (2017) 367:551–569
DOI 10.1007/s00441-016-2566-8
Study 2013 reported a global age-standardized death rate of
8.0 per 100,000 in 2013, which is equivalent to breast cancer
(7.4) or pedestrian road injuries (8.0). Hence, asthma remains
among the top 50 causes of global years of life lost (GBD
2013 Mortality and Causes of Death Collaborators 2015).
A brief search for the available literature in PubMed using
the strategy Bairway AND asthma AND (remodeling OR
remodelling)^ revealed that, in 1993, the first two papers fall-
ing into this category were published and that numbers slowly
increased to 203 in 2007. Since then, the number of papers per
year are on a relatively constant level of between 200 and 240,
which represent only approx. 15% of all publications on
Basthma AND airway^. Notably, among these publications,
about 50–60 per year are review papers, i.e., 20–40% of all
publications. To avoid carrying coals to Newcastle, in the
present review we will not try to review everything already
well reviewed by others (Elias 2000; Holgate et al. 2000;
Saetta and Turato 2001; Jeffery 2004;Hogg2004; Bai and
Knight 2005; Boulet and Sterk 2007; Al-Muhsen et al. 2011;
Martinez and Vercelli 2013;SaglaniandLloyd2015). Instead,
we will focus on some aspects related to a methodologically
sound quantification as a prerequisite for the reliable assess-
ment of airway remodeling and on what we can learn from
invertebrate to primate animal models to better understand the
mechanisms underlying physiological versus pathological
remodeling.
Today’s perception of bronchial asthma:
the phenotype/endotype concept
The perception of bronchial asthma has fundamentally
changed during the last d ecade. Conseque ntly, the Global
Initiative for Asthma (GINA) suggested a new definition of
asthma: BAsthma is a heterogeneous disease, usually charac-
terized by chronic airway inflammation. It is defined by the
history of respiratory symptoms such as wheeze, shortness of
breath, chest tightness and cough that vary over time and in
intensity, together with variable expiratory airflow limitation^
(Reddel et al. 2015). The definition takes into account that
during the last decade asthma turned out to be quite heteroge-
neous encompassing patients with different phenotypes (i.e.,
the entity of observable characteristics), which, however, ex-
hibit overlap to variable degrees. Adult asthma was suggested
to be not a single disease but a syndrome with one common
feature being a variable expiratory airflow limitation (Lötvall
et al. 2011; Wenzel 2006). The concept of asthma as a syn-
drome is also expected to be very helpful in pediatric asthma
(Spycher et al. 2010; Lødrup Carlsen and Carlsen 2 012).
However, the characteristics used for the definition of the var-
ious phenotypes are not necessarily directly related to the un-
derlying pathogenetic process(es). Aiming at the development
of personalized therapies, i.e., therapeutic approaches targeting
key elements of the causative pathomechanism(s), the concept
of asthma endotypes was proposed with each endotype being
the result of a specific molecular pathomechanism that is dis-
tinctly different from the other endotypes (Lötvall et al. 2011;
Wenzel 2012; Agache et al. 2012). Until today, probably the
(one and only) most clearly defined asthma endotype is the
one termed the Th2 (T helper type 2)-high endotype, which is
characterized by a T helper lymphocyte type 2-driven inflam-
mation (Fahy 2015). Anti-inflammatory therapies targeting
Th2 cytokines such as interleukin (IL)-4, IL-5 and IL-13 have
consistently exhibited beneficial effects in adult patients diag-
nosed as Th2-high asthmatics on the basis of some emerging
biomarkers related to the IL-13 response, such as periostin or
FeNO (Bhakta and Woodruff 2011;IngramandKraft2012;
Fahy 2015; Fajt and Wenzel 2016). Efforts are being under-
taken to reveal additional novel biomarkers that may help in
identifying and distinguishing further endotypes (Zissler et al.
2016).
The current phenotype/endotype concept of asthma has a
strong focus on clinical and inflammatory characteristics and
omitted aspects of airway remodeling, one important feature
of asthma, as was recently emphasized (Saglani and Lloyd
2015). Today, it is unclear whether the differences in airway
remodeling parameter values observed between individual pa-
tients define specific remodeling phenotypes and how these
may relate to clinical or inflammatory phenotypes or are even
linked to a specific endotype.
Defining airway remodeling in asthma: physiological
versus pathological processes
Most review papers define airway remodeling quite broadly as
any change in composition, distribution, thickness, mass or
volume and/or number of structural components observed
in the airway wall of patients relative to the airway wall of
normal healthy individuals (Bergeron et al. 2009;Bai2010;
Hirota and Martin 2013). Changes have been described for
various tissues in asthma patients, such as airway epithelium
(e.g., epithelial shedding, goblet cell hyperplasia, basal mem-
brane thick ening), perib ronchial int erstitial tissue (e.g.,
subepithelial fibrosis), airway smooth muscle cells (e.g., hy-
perplasia and/or hypertrophy), nerve tissue (e.g., increased
neurite sprouting) and bronchial vasculature (e.g., barrier dys-
function, angiogenesis) (Undem et al. 1999; Beckett et al.
2003; Jeffery 2004; Al-Muhsen et al. 2011). The observation
of an inflammatory infiltrate characterized by eosinophilic
granulocytes and CD4
+
Th cells in the airways of (probably
Th2-high) asthmatics (Saetta and Turato 2001), sometimes
being the result of a longstanding inflammatory process, was
suggested as a conditio sine qua non of the definition of air-
way remodeling in asthma (Hirota and Martin 2013).
Although airway remodeling has been reported for other
552 Cell Tissue Res (2017) 367:551–569
chronic lung diseases such as chronic obstructive pulmonary
disease (COPD), some structural changes of the airways ap-
pear to be distinctly different when comparing asthma and
COPD as reviewed recently (Jones et al. 2016). The evidence
accumulated until now suggests that airway remodeling is
associated with a progressive loss of lung function, a view
which still has to be considered a hypothesis because therapies
targeting airway remodeling are still missing (Pascual and
Peters 2005).
In this review, we adopt the suggestions made by
Jeffery (2001, 2004) and distinguish two types of airway
remodeling, i.e., physiological remodeling on the one hand
and pathological remodeling on the other. Physiological
airway remodeling comprises those structural changes,
which occur regularly during normal lung development
and growth leading to a normal mature airway wall or
that occur as an acute and transient response to injury
and/or inflammation ultimately resulting in restoration of
a normal airway structure. Structural alterations that occur
as a result of either disturbed lung development or as a
response to chronic injury and/or inflammation leading to
persistently altered airway wall structures and function are
considered as pathological airway remodeling.
The most relevant implications of these definitions are:
& Unless quantitative analyses of airway structural charac-
teristics are used, objective evidence of structural devia-
tions from the normal healthy condition cannot be provid-
ed without doubt.
& Although airway remodeling is frequently associated with
airway inflammation, remodeling cannot be considered
being a secondary phenomenon to inflammation in every
single case.
& Airway remodeling may be a primary event in asthma
pathogenesis if it is the result of disturbed lung
development.
& Unless the kinetics of these processes can be revealed,
which is very difficult in humans, it will be very difficult
to distinguish an acute and t ransient response from a
chronic reaction and, thereby, unequivocally differentiate
between physiological and pathological processes. In the
absence of data on the kinetics, additional criteria/
biomarkers are badly needed and animal models can be
very helpful in that they allow for kinetic studies.
Consequently, the following major questions will be
addressed:
& what are reliable quantitative approaches to assess airway
remodeling,
& are there any indications supporting the notion that airway
remodeling can occur as a primary event, i.e., before any
inflammatory process was initiated,
& what do we know about airway remodeling being a sec-
ondary event to inflammation and
& what can we learn from animal models in the study of
airway remodeling for distinguishing physiological and
pathological airway remodeling and the underlying, po-
tentially differing pathomechanisms?
Quantitative approaches to assess airway remodeling
The initial approach to assess airway remodeling, both in
humans and in animal models, has been the histologic
analysis of two-dimensional sections by means of light,
fluorescence or electron microscopy. Recent technologi-
cal advances allowed the implementation of high-
resolution imaging into radio logic a nalyses of airwa y
morphology (Hartley e t al. 2016). These approaches,
however, are beyond the expertise of the authors and
therefore this review will focus on microscopy-based ap-
proaches only.
In 2010, a joint task force of the American Tho racic
Society (ATS) and the European Respiratory Society (ERS)
published an Official Research Policy Statement paper that
critically reviewed the state-of-the-art stereological methods
in lung morphometry and defined standards to promote com-
parability of morphometric studies in pulmonary research
(Hsia et al. 2010). This landmark paper is suggested as the
starting point for everyone designing new studies of airway
remodeling and is a benchmark paper when evaluating pub-
lished data. As was emphasized by this task force, the quanti-
fication of structures is based upon the three-dimensional (3D)
physical attributes of its components. When two-dimensional
(2D) sections are used for quantitative analysis, only incom-
plete information about the 3D structure are obtained, which
bears a high risk of misinterpretations and false conclusions
(Hsia et al. 2010). This risk is particularly prominent for the
airway tree, which is highly anisotropic and exhibits marked
qualitative and quantitative changes in airway wall structure
from proximal to distal airway generations (Crystal et al.
1997; Mauroy et al. 2004). Consequently, quantitative ap-
proaches to assess remodeling of airways have to take into
account that both the orientation of a 2D section relative to
the airway’s longitudinal axis and the location along the air-
way tract (i.e., airway hierarchy) have marked effects on the
quantitative parameters analyzed (Hsia et al. 2010). Therefore,
obtaining a collection of an unbiased set of representative
tissue samples requires a few but important additional steps
during sampling, as described previously for whole lungs ob-
tained from animal models (Hyde et al. 2007;Mühlfeldand
Ochs 2013) but also for studying human biopsies (Ferrando
et al. 2003; Woodruff and Innes 2006; Bratu et al. 2014).
Cell Tissue Res (2017) 367:551–569 553
A few ex amples could illustrate that using quantitative
stereological approaches may challenge some generally ac-
cepted views.
Epithelial cell shedding Today, it is widely accepted that the
airway epithelium, which is in almost
1
direct contact to the
inhaled air, is far more than just a passive physical barrier to
what is inhaled during breathing (Tam et al. 2011). The epi-
thelium exerts v arious functions that help maintain a
healthy lung such as particle clearance, fluid balance, innate
immune responses. Direct injury to the airway epithelium in-
duced by various triggers has long been widely accepted as a
very early, may be the initial, step in asthma pathogenesis (Al-
Muhsen et al. 2011; Hirota and Martin 2013;Holtetal.2014).
This notion is in part based on the qualitative observation of a
denuded ep ithelial basal lamina in biopsies of asthmatics,
which was interpreted as the result of epithelial cell shedding
and a histological reflection of the loss of airway epithelial
function (f or references of original studies, see, e.g.,
Bergeron et al. 2009; Fajt and Wenzel 2016). A computer-
based quantitative study of bronchial biopsies obtained from
14 mild and moderate human asthmatics, however, revealed
that there were no differences in the degree of epithelial des-
quamation in comparison to biopsies from 12 healthy subjects
(Ordoñez et al. 2000). Using glycol methacrylate as embed-
ding medium for the biopsies, a section thickness of 2 μm
could be achieved, which allowed an excellent presentation
of all structures in the microscopic images. The authors sug-
gested that epithelial desquamation in endobronchial biopsies
in asthmatics is an artifact of tissue sampling and not a true
pathologic feature of asthma. Although this was controversial-
ly discussed (Holgate et al. 2001), our own data demonstrate
that the degree of epithelial desquamation in human
endobronchial biopsies increases with decreasing biopsy size
(Fig. 1). The smaller the biopsy, the more mechanical forces
may affect the tissues during collection and embedding, which
supports the n otion that epithelial desquamation is highly
prone to artefactual damage. Therefore, epithelial shedding
is a que stionable phe notypic characteristic of asthma.
Without any doubt, however, it is very well supported that
dysregulated airway epithelial cell functions are a central ele-
ment in the pathogenesis of asthma (Holgate 2007, 2011a, b;
Fahy and Locksley 2011).
Epithelial basement membrane thickening The basement
membrane or basement membrane zone of the airways is an
extracellular structure specialized for the attachment of the
epithelium to the underlying extracellular matrix (Evans
et al. 2010). In the electron microscope, three layers can be
distinguished: the lamina lucida, the lamina densa and the
lamina reticularis, which have been shown to also differ in
their chemical composition. The lamina reticularis, the basal
portion of the basement membrane, can also be seen with the
light microscope and is also referred to as reticular basement
membrane (RBM) or subepithelial basement membrane.
RBM thickening was suggested to be pathognomonic of asth-
ma (Jeffery 2004). However, it has been reported to also occur
in children with cystic fibrosis (Hilliard et al. 2007).
Moreover, RBM thickening was also observed in adult
COPD patients with RBM thickness being not significantly
different from adult subjects with asthma (Liesker et al. 2009).
When analyzing RBM thickening in the airways, again one
has to take into account that there is marked variation along
the airway tree, with the RBM becoming thinner as it extends
from the trachea into the small airways (Evans et al. 2010). In
addition, RBM thickness as represented in microscopic sec-
tions strongly depends on how much the sectioning angle
deviates from the ideal situation of a section normal to the
RBM surface. Therefore, exclusion of obliquely to tangential-
ly cut tissue is a step regularly implemented into the measure-
ment of RBM thickness, although using largely subjective
criteria. Implementation of design-based stereological proto-
cols is rare, although this is the only way to guarantee random
tissue orientation (i.e., isotropic, uniformly random orienta-
tion) ensuring that tangential cuts occur with a known distri-
bution and can be handle d systematically (Ferrando et al.
2003;Hsiaetal.2010
). Ferrando et al. (20
03)madecompar-
isons using two classical procedures in parallel. Although
RBM was significantly thicker in asthmatics than in healthy
subjects, the measurements made by using design-based ste-
reology were approximately 30% smaller than measurements
made with the two classical procedures. Notably, the mean
coefficient of error for repeat measurements (i.e., the repro-
ducibility of the measurements) was 0.06 for the stereological
1
Epithelial secretory products such as mucus and/or surfactant are, however,
components that separate the air from the epithelial tissue layer.
Fig. 1 Fraction of epithelial basal membrane (BM) o f human
endobronchial biopsies exhibiting complete denudation is inversely
correlated with biopsy volume (=size), which was estimated according
to the Cavalieri Principle. Figure by courtesy of Dr. V.A. Bratu, modified
from Bratu (2008); Fig. 3.4c
554 Cell Tissue Res (2017) 367:551–569
approach, which is by far preferable compared with the 0.19–
0.30 in the other approaches (Ferrando et al. 2003).
Implementation of such an unbiased procedure as an objective
correction for tangential cuts could be an important step for
standardization of protocols and thus would help ensure better
comparisons of measurement data across studies and labora-
tories. Accepting that reliable methods for the quantitative
assessment of RBM thickening are crucial for obtaining sound
data, reports on therapies successfully reducing RBM thick-
ness in asthma patients, reviewed recently by Durrani et al.
(2011), should be critically re-evaluated.
Smooth muscle hyperplasia and hypertrophy The potential
role(s) of airway smooth muscle cells in the pathogenesis of
asthma symptoms, in particular with regard to airway
hyperresponsiveness, has been comprehensively reviewed
by others (An et al. 2007; Gosens and Grainge 2015). It is
widely accepted that the total amount of smooth muscle is
increased in asthma (Jeffery 2004;Fixmanetal.2007;
Durrani et al. 2011). The increase seems to involve both small
and large airways (James et al. 2012) and may be related to the
clinical severity and duration of asthma (Bai et al. 2000). In
principle, the increase in muscle mass can be achieved by an
increase in cell number (hyperplasia), by increase in cell vol-
ume (hypertrophy), or a combination of both and by addition-
al immigration of myofibroblasts (Bara et al. 2010). Distinctly
different molecular mechanisms may be considered leading to
hyperplasia or hypertrophy and both may be associated with
distinct functional consequences. Conclusive demonstration
of hyperplasia, i.e., increase in cell number, is only feasible
if an unbiased design-based stereology approach such as the
physical or optical disector is applied (Sterio 1984;Hsiaetal.
2010;Gruberetal.2012). Using an assumption-based quan-
titative approach and the mean cell diameter across the nucle-
us as surrogate for cell size, hypertrophy was reported to al-
ready be present in mild-to-moderate asthmatics and being
even more pronounced in severe asthmatics (Benayoun et al.
2003). However, using design-based stereology, a biopsy
study in adult patients with mild-to-moderate asthma demon-
strated hyperplasia of smooth muscle cells, i.e., an increase in
cell number per volume of tissue, whereas no significant in-
crease in cell size, i.e., mean cell volume in μm
3
, was seen
(Woodruff et al. 2004). Notably, Woodruff et al. (2004) did not
find any differences in gene expression of smooth muscle cells
isolated by laser-capture microdissection from bronchial biop-
sies of asthmatics and healthy controls with regard to the
cellular phenotype. Regamey et al. (2008)demonstratedby
design-based stereology that both hyperplasia and hypertro-
phy occur in children with mild-to-severe asthma but that
hyperplasia was the most prominent contributor to the in-
crease in smooth muscle mass. In this study, smooth muscle
cell hyperplasia revealed to be not limited to children with
asthma but to also be present in cystic fibrosis and n on-
cystic fibrosis bronchiectasis. Differentiation of hyperplasia
versus hypertrophy is important when considering airway
smooth muscle as a potential therapeutic target. Excit ing
new data suggest that bronchoconstriction as a result of airway
smooth muscle contraction appears sufficient to induce airway
remodeling via processes triggered by mechanical forces and
independent of the inflammatory response (for review see
Grainge et al. 2011; Gosens and Grainge 2015). These studies
indicate that airway smooth muscle cells may not simply be
secondary effector cells responding to an already ongoing
pathogenic process but in contrast may also be at the forefront
of disease initiation.
Airway remodeling as a secondary event
to inflammation
In fact, airway remodeling in allergic bronchial asthma is
discussed to be the result of a chronic inflammatory response
entailing on the one hand permanent airway tissue destruction
and on the other hand chronic tissue repair. Thus, chronic
airway inflammation can be described as the major force driv-
ing the processes leading to most aspects of airway remodel-
ing. This Binflammation theory^ is mainly supported by the
finding that steroid treatment in asthmatic patients does not
only reduce airway inflammation but also has beneficial ef-
fectsonairwayremodeling(Triggetal.1994;Olivierietal.
1997;Laitinenetal.1997; Hoshino et al. 1998, 1999, 2001;
Sont et al. 199 9;Wardetal.2002; Chetta et al. 2003).
Infiltrating cells like T helper (Th) cells, eosinophils, neutro-
phils and mast cells interact with resident cells of the airways
such as fibroblasts, smooth muscle cells, neuronal cells, epi-
thelial cells and endothelial cells by the release of a plethora of
cytokines, enzymes, metabolites and growth factors creating a
signaling environment that—under chronic conditions—re-
sults in airway remodeling.
Thelpercells—in allergic bronchial especially Th2 cells—
orchestrate the allergic inflammatory response by releasing a
characteristic array of cytokines including IL-4, IL-5, IL-9 and
IL-13. Each of these cytokines has prominent functions in
directing the production of allergen-specific IgE, recruitment
of eosinophils or development of AHR; however, whether
they directly have an impact on airway remodeling is still a
matter of debate. As already mentioned, gene-targeted mouse
strains overexpressing the cytokines IL-4, IL-5, IL-9, IL-11, or
IL-13 spontaneously develop airway inflammation, AHR,
mucus hyperproduct ion and airway remodeling (Rankin
et al. 1996;Tangetal.1996; Lee et al. 1997; Temann et al.
1998; Zhu et al. 1999). After these initial studies, further ex-
periments provided deeper insight into the contribution of
each of these Th2-type cytokines to airway remodeling. At
least the effects of IL-4, IL-5 and IL-9 are either dependent
on IL-13 or promote airway remodeling by supporting the
Cell Tissue Res (2017) 367:551–569 555