Physiol Rev
82: 19 – 46, 2002; 10.1152/physrev.00020.2001.
Immunity, Inflammation, and Remodeling in the Airway
Epithelial Barrier: Epithelial-Viral-Allergic Paradigm
MICHAEL J. HOLTZMAN, JEFFREY D. MORTON, LAURIE P. SHORNICK, JEFFREY W. TYNER,
MARY P. O’SULLIVAN, AURITA ANTAO, MINDY LO, MARIO CASTRO, AND MICHAEL J. WALTER
Departments of Medicine and Cell Biology/Physiology, Washington University School of Medicine, St. Louis, Missouri
I. Introduction: Traditional and Alternative Views of Airway Inflammatory Disease
II. Molecular Basis of Airway Epithelial Immune Function From In Vitro Studies
A. Gene products for mediating transepithelial traffic of immune cells
B. Cytokine-dependent gene network: transcriptional regulation
C. Virus-dependent gene network: posttranscriptional regulation
D. The final common pathway: epithelial cell death
III. Airway Immunity and Inflammation in a Mouse Model of Viral Infection
A. Epithelial gene expression: interaction between virus and host cell
B. Toward epithelial gene knockouts: defining an active role in innate immunity
IV. Airway Immunity and Inflammation in Human Subjects With Asthma
A. Constitutive abnormalities in epithelial gene expression (Stat1 and IL-12)
B. Glucocorticoid-sensitive abnormalities in epithelial gene expression (RANTES)
C. A revised model for Th1/Th2 contributions to asthma
V. Long-Term Airway Hyperreactivity and Remodeling in Mice and Humans
A. Segregating acute from chronic phenotypes
B. Genetic and viral determinants for persistence
VI. Smart Strategies for Correcting Epithelial Inflammation and Remodeling
A. Reversing viral mimicry using mutant E1A oncoprotein
B. Modifying epithelial signaling with mutant Stat1
C. Future considerations
VII. Summary
19
21
21
24
27
28
30
31
31
34
34
35
36
37
37
37
38
39
41
41
42
Holtzman, Michael J., Jeffrey D. Morton, Laurie P. Shornick, Jeffrey W. Tyner, Mary P. O’Sullivan, Aurita
Antao, Mindy Lo, Mario Castro, and Michael J. Walter. Immunity, Inflammation, and Remodeling in the Airway
Epithelial Barrier: Epithelial-Viral-Allergic Paradigm. Physiol Rev 82: 19 – 46, 2002; 10.1152/physrev.00020.2001.—The
concept that airway inflammation leads to airway disease has led to a widening search for the types of cellular and
molecular interactions responsible for linking the initial stimulus to the final abnormality in airway function. It has not yet
been possible to integrate all of this information into a single model for the development of airway inflammation and
remodeling, but a useful framework has been based on the behavior of the adaptive immune system. In that paradigm,
an exaggeration of T-helper type 2 (Th2) over Th1 responses to allergic and nonallergic stimuli leads to airway
inflammatory disease, especially asthma. In this review, we summarize alternative evidence that the innate immune
system, typified by actions of airway epithelial cells and macrophages, may also be specially programmed for antiviral
defense and abnormally programmed in inflammatory disease. Furthermore, this abnormality may be inducible by
paramyxoviral infection and, in the proper genetic background, may persist indefinitely. Taken together, we propose a
new model that highlights specific interactions between epithelial, viral, and allergic components and so better explains
the basis for airway immunity, inflammation, and remodeling in response to viral infection and the development of
long-term disease phenotypes typical of asthma and other hypersecretory airway diseases.
I. INTRODUCTION: TRADITIONAL AND
ALTERNATIVE VIEWS OF AIRWAY
INFLAMMATORY DISEASE
A critical step toward defining molecular mechanisms of airway disease came with formal recognition of
www.prv.org
the role of immunity and inflammation. In the case of
asthma, evidence of immune abnormalities and excessive
airway inflammation (induced by allergic and nonallergic
stimuli) has led to a widening search for the types of
inflammatory cells and mediators that might be responsible for abnormal airway function. Cell types implicated in
0031-9333/02 $15.00 Copyright © 2002 the American Physiological Society
19
20
HOLTZMAN ET AL.
the development of airway inflammation include immune
cells as well as parenchymal cells. Cell-cell interactions
are attributed to classes of mediators that include lipids,
proteases, peptides, glycoproteins, and cytokines. The
leading scheme for integrating this information has been
based on the classification of the adaptive immune system, and especially the responses of T helper (Th) cells. In
this scheme, CD4⫹ T cell-dependent responses are classified into T helper type 1 (Th1) or type 2 (Th2). Th1 cells
characteristically mediate delayed-type hypersensitivity
reactions and selectively produce interleukin (IL)-2 and
interferon (IFN)-␥, whereas Th2 cells promote B celldependent humoral immunity and selectively produce
IL-4, IL-5, and IL-13. Thus Th2 reactions may underlie the
airway hyperreactivity and inflammation characteristic of
the late response to allergen inhalation and so account for
the overproduction of Th2-derived cytokines that is characteristic of asthma (109, 151). How then can a Th2polarized response account for asthma that is also triggered by exposure to nonallergic stimuli (especially
respiratory viruses) that would ordinarily trigger a Th1
response? On the basis of the possibility that Th2-skewed
responses may develop in this setting as well (21, 62, 121),
it is possible that a Th2-dominant response may mediate
inflammation and hyperreactivity in response to nonallergic stimuli as well (as modeled in Fig. 1). However, this
dichotomous view of the immune response may not be
completely accurate, since there is often ambiguity in the
type of Th response triggered by most stimuli as well as
significant cross-regulation between the two types of responses (57). In fact, recent studies in mice using adoptive transfer with Th1 and Th2 cells have indicated that
Th1 cells may even be involved in initiating the allergic
response (16, 110).
Because of these uncertainties, we have questioned
whether other aspects of immunity and inflammation
might also be critical for the pathogenesis of airway disease. In particular, we aimed to develop a model that
better accounted for the dissociation between the development of allergy and asthma in many subjects and was
based on a more precise appraisal of the Th1 system in
the airway. Furthermore, the “Th2 hypothesis” does not
account for the possibility that the airway epithelial cells
may act as active sentinels of innate immunity, and so,
like other components of the innate immune system, may
provide critical signals to the adaptive immune system.
This review summarizes how our view of the barrier
epithelial cell has evolved based on the identification of
its specialized programming for host defense and abnormal programming in disease.
The review is divided into six major sections. Section
II summarizes studies of isolated airway epithelial cells.
The section proceeds from a relatively simple scheme for
leukocyte recruitment (based on cell adhesion and chemoattraction) to one that depends on the coordinated
Physiol Rev • VOL
FIG. 1. Scheme for the role of the airway immune response in the
development of asthmatic airway inflammation. A: illustration of how
decreases in virus-driven production of Th1 cytokines [e.g., interferon-␥
(IFN-␥)] and increases in allergen-dependent production of Th2 cytokines [e.g., interleukin (IL)-4, IL-5, and IL-13] are characteristic of
asthma (designated as an “A”). This Th2-skewed setting thereby provides for flares of the disease driven either by virus infection (with a
functional blockade in the normal Th1 response and a concomitant shift
to an increase in the Th2 response) or by allergen (with an increase in
the Th2 response). B: illustration of a modified scheme based on an
increased activity of both Th1 and Th2 responses. In this case, pathogenactivated Th1 cells in the airway tissue generate factors [e.g., tumor
necrosis factor-␣ (TNF-␣)] that mediate subsequent recruitment of Th2
cells [e.g., via vascular cell adhesion molecule-1 (VCAM-1)] that provide
a setting for enhanced allergen responsiveness. [Modified from Castro et
al. (16).]
expression of a network of epithelial immune-response
genes under distinct transcriptional and posttranscriptional controls. The transcriptional program depends primarily on interferon-Jak/Stat signaling while posttranscriptional regulation uses a distinct RNA-protein
interaction that is responsive directly to viral replication.
Section III summarizes studies that extend these same
molecular pathways to studies done in vivo using a mouse
model of viral bronchiolitis that takes advantage of a
genetically defined host. Section IV summarizes studies of
human subjects with airway disease, focusing on how
these same epithelial gene networks behave in subjects
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
with stable or flared asthma in the presence or absence of
anti-inflammatory treatment with glucocorticoids. This
section presents a model for how the epithelial gene
network might interact with the Th1/Th2 balance to cause
airway inflammation in asthma and introduces the possibility that this same gene network may also underlie
persistent abnormalities in airway epithelial mucosal behavior found in chronic airway disease. Section V summarizes studies of this possibility done in vitro and in vivo in
mice and humans, highlighting the role of genetic susceptibility to virus-induced modification of epithelial phenotype and consequent remodeling. Section VI, the final section, summarizes the review.
II. MOLECULAR BASIS OF AIRWAY
EPITHELIAL IMMUNE FUNCTION
FROM IN VITRO STUDIES
This section summarizes studies of isolated airway
epithelial cells, placing special emphasis on the use of
primary-culture human cells that exhibit high-level fidelity
to in vivo behavior. Section IIA begins with a model of
epithelial barrier function and so presents a working
scheme for epithelial-dependent traffic of immune cells
into the airway. This model depends on at least one major
ligand for cell adhesion, i.e., intercellular adhesion molecule-1 (ICAM-1) and another for chemotaxis, i.e., regulated upon activation, normal T cell expressed and secreted (RANTES) chemokine, each of which depends on
the level of gene expression for function. Accordingly,
section II, C and D, proceeds to summarize the basis for
epithelial expression of these two corresponding genes.
In doing so, these studies define two larger gene networks. One is typified by ICAM-1 and is under the control
of IFN-driven Jak/Stat signal transduction and uses Stat1
as a key intermediate. The other is typified by RANTES
and relies on posttranscriptional regulation that is more
directly inducible by viral replication. In this section, we
also introduce recent findings related to the IL-12 p40
gene product that also appears chemotactic, but the regulation of this pathway is not yet fully defined. Later
sections will summarize evidence for the role of these
same systems in mediating immunity and inflammation in
vivo in animal models and human subjects.
A. Gene Products for Mediating Transepithelial
Traffic of Immune Cells
The initial approach to defining epithelial control of
immune cells and specific epithelial-leukocyte interactions relied on studies of immune cell interactions with
each other and with the endothelium. In these systems,
the development of monoclonal antibodies against immune cell determinants indicated that a critical step in
Physiol Rev • VOL
21
leukocyte influx from the circulation is the coordinated
interaction of specific cellular receptors on the migrating
leukocytes with corresponding ligands on adjacent endothelial cells. A scheme for leukocyte recruitment was
developed that revolved around the regulated expression
of distinct families of cell communication molecules, notably the selectins, the integrins, and the cell adhesion
molecule (CAM) members of the immunoglobulin (Ig)
supergene family (12). In addition to these cell adhesion
systems for direct cell-cell contact, endothelial cells and
immune cells also appear capable of generating a series of
chemoattractants [e.g., platelet activating factor (PAF)]
and chemoattractant cytokines (or chemokines) that may
act over a greater distance to direct immune cell movement and activation in tissue (85). These three systems, 1)
selectin binding to carbohydrate mucinlike molecules, 2)
cell adhesion molecules from the Ig and the extracellular
matrix protein families binding to integrins, and 3) chemoattractants and chemokines binding to G protein-coupled receptors, may act in a specific combination to dictate the type of immune cells that enter and get retained
in the tissue. Especially in the case of endothelial-leukocyte interaction, this information has enabled the development of a stepwise molecular scheme for leukocyte
influx into tissue from the circulation (120).
1. Cell adhesion molecules and the role of ICAM-1
in the epithelium
Our initial studies of epithelial-immune cell adhesion
were undertaken with airway epithelial cell monolayers
established under conventional cell culture conditions,
thereby exposing only the apical epithelial surface to the
immune cell (an approach that is more logically taken for
studies of endothelial cells). Improving on earlier methods, however, we studied the basis for immune cell adhesion to airway epithelial and venular endothelial cells
using a quantitative flow cytometry-based assay (98). This
technique avoids extensive leukocyte purification, culture, or labeling steps that may alter leukocyte function
and by their nature may eliminate heterotypic cell-cell
interactions that may be important in vivo. The quantitative aspect of the flow cytometry methodology was also
critical for examining the low levels of constitutive adherence that might be expected for resting T cells. Compared
with standard cell-labeling methods, the flow cytometrybased assay yielded a lower level of constitutive T cell
adhesion despite a similar level of stimulated adhesion
(after T cell activation with phorbol dibutyrate) using
endothelial or epithelial cell monolayers. Endothelial-T
cell adhesion was further increased by monolayer treatment with tumor necrosis factor-␣ (TNF-␣) (less so with
IL-1␤ and least with IFN-␥), whereas epithelial-T cell adhesion was most sensitive to IFN-␥. Cytokine stimulation
of adhesion was invariably concentration dependent and
82 • JANUARY 2002 •
www.prv.org
22
HOLTZMAN ET AL.
closely matched to the cellular levels of ICAM-1 (79).
Accordingly, T cell adhesion was markedly inhibited by
anti-ICAM-1 or anti-␤2-integrin antibody (95–97% inhibition for epithelial cells and 57– 67% inhibition for endothelial cells) directed against ICAM-1 interaction with
leukocyte function-associated antigen 1 (LFA-1). Residual
endothelial-T cell adhesion that correlated with endothelial vascular cell adhesion molecule-1 (VCAM-1) levels
was blocked by an anti-␣4-integrin antibody directed
against VCAM-1 interaction with very late-activation antigen 4 (VLA-4). The results suggest that 1) peripheral
blood T cells without exogenous activation exhibit little
LFA-1- or VLA-4-dependent adherence except to endothelial or epithelial cells expressing high levels of ICAM-1
and/or VCAM-1; and 2) differences in endothelial versus
epithelial cell mechanisms to bind activated and unactivated T cells (e.g., dependence on a mixed- versus a
single-ligand system and distinct cytokine-responsiveness
of ligand levels) may help to coordinate T cell traffic to
epithelial barriers. These findings also support the view
that T cell trafficking into inflamed tissues (especially in
mucosal barriers) depends critically on local activation
events.
These initial studies highlight two distinct features of
airway epithelial cell behavior: relative hypersensitivity to
IFN-␥ and a critical role for ICAM-1 in epithelial-immune
cell adhesion. This pattern of cell adhesion molecule expression and function on airway epithelial cells appears
distinct from the one for venular endothelial cells. In the
case of endothelial cells, at least three interactions of cell
adhesion receptors have been implicated in the development of inflammation in epithelial barriers: 1) E-selectin
binding to its leukocyte sialyl-Lewis X carbohydrate ligand (38, 49, 145); 2) VCAM-1 binding to the leukocyte
␤1-integrin VLA-4 (␣4␤1) (7, 97); and 3) ICAM-1 binding to
the leukocyte ␤2-integrins LFA-1 (␣L␤2) or Mac-1 (␣M␤2)
(101, 146). In each case, there is evidence that the ligand
may be overexpressed and/or activated under basal or
allergen-challenge conditions in asthmatic subjects and
that pretreatment with a blocking monoclonal antibodies
can inhibit immune cell influx in a model of asthmatic
airway inflammation. The findings provide for two systems (E-selectin and VCAM-1) capable of initiating leukocyte rolling and tethering to the vascular endothelium as
well as two that may mediate diapedesis through the
vascular wall (VCAM-1 and ICAM-1). However, neither
isolated cells in culture nor endobronchial tissue from
normal and asthmatic subjects indicated that airway epithelial cells (or fibroblasts or smooth muscle cells) expressed significant levels of E-selectin or VCAM-1. These
findings are therefore consistent with the view that airway
epithelial cells present fewer adhesion ligands (than endothelial cells) for directing leukocyte movement. The
biologic basis for this difference may reflect the capacity
of endothelial cells to be armed with ligands that slow
Physiol Rev • VOL
down passing leukocytes (to allow tethering and triggering) and that select a specific leukocyte subset (to allow
adhesion and transmigration) from the diverse circulating
pool. This requires multiple molecular interactions with
varying specificities. In contrast, other parenchymal cells
may come into contact with immune cells after some
degree of selection and activation so that they are required only to facilitate further leukocyte migration and
retention. ICAM-1 is well suited to mediate this process
because nearly all immune cells constitutively express its
receptors, i.e., the ␤2-integrins LFA-1 and Mac-1. In the
case of T cell traffic, the endothelial cells (and likely
epithelial cells) also may express a series of receptors
(designated homing receptors or addressins) that serve to
direct distinct subsets of lymphocytes to appropriate locations of lymphoid tissue. This type of cell adhesion
(exemplified by some of the ␤7-integrin interactions) is
probably most important in maintaining a resident population of immune cells, but whether this system is also
regulated during airway inflammation remains uncertain.
Interestingly, a specific addressin for T cell localization to the lung has not yet been identified. We reasoned
that patterned expression of more common adhesion molecules might also facilitate traffic of specific T cell subsets
to the airway mucosa. To define the relationship between
T cell phenotype and adhesiveness, we examined T cell
adhesion to endothelial cell, fibroblast, and epithelial cell
monolayers as well as extracellular matrix proteins (collagen and fibronectin) using a three-color flow-cytometrybased adherence assay that minimizes basal adhesion
levels and facilitates quantitative lymphocyte subtyping
(99). Regardless of monolayer type, monolayer stimulation conditions, or T cell activation status, we found that
the ␥␦-TCR bearing T cells adhered more efficiently than
␣␤ T cells. The difference was based predominantly on
increased levels of activatable LFA-1 (and to a lesser
degree VLA-4) because 1) it correlated precisely with
inhibitability by anti-LFA-1 (and VLA-4) monoclonal antibodies and the levels of LFA-1 (and VLA-4) on the cell
surface; and 2) it persisted after maximal LFA-1 (and
VLA-4) activation with phorbol dibutyrate. In contrast to
most cases of ␣␤ T cell behavior, ␥␦ T cell adhesion to cell
monolayers was not linked to memory status, i.e., there
was no difference between naive V␦1⫹ and memory V␦2⫹
populations in levels of LFA-1 (or VLA-4) expression or
LFA-1- (or VLA-4-)dependent adhesion to cell monolayers.
However, V␦1⫹ cells exhibited higher levels of VLA-5 that
correlated with an increased adhesiveness to fibronectin
and to a 120-kDa fibronectin fragment (FN-120) that contains only the VLA-5-binding domain but not to type I
collagen or to a fibronectin fragment (FN-40) that binds
only VLA-4. Taken together, the results define a hierarchy
for integrin (LFA-1, VLA-4, and VLA-5) expression and
consequent adhesion among T cell subsets that is linked
to TCR gene usage (but not necessarily linked to memory
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
status) and may thereby help to explain the accumulation
and retention of V␦1⫹ ␥␦ T cells in epithelial and connective tissues. The findings underscore the general concept
that immune cells (like parenchymal cells) may regulate
their expression of cell adhesion molecules to achieve
preferential localization in tissues. The relationship of
regulation of more general cell adhesion receptors (such
as LFA-1) to more specific systems directed by addressins
or by specific antigens needs to be further defined. At
present, however, it appears that these more general receptors may work in concert with more specific ones, a
concept that is illustrated during antigen presentation
where cell adhesion molecules facilitate engagement and
activation of the T cell receptor for antigen in the appropriate major histocompatibility complex (MHC) context.
Recruitment of specific subsets of leukocytes may also be
mediated by a diverse capacity for epithelial production
of chemokines as noted in section IIA2.
2. Chemokines and the role of RANTES
in the epithelium
As noted above, schemes for transendothelial movement
of immune cells (extravasation) depend on the coordinated
expression of cell adhesion molecules and chemoattractants
that interact with corresponding receptors on the immune cell
surface (12, 85, 120). Thus it appeared reasonable to propose
that similar molecular mechanisms may regulate transepithelial
movement of immune cells. However, there are critical differences between epithelial and endothelial cell adhesion and
transmigration, the most obvious of which may be that immune
cell recruitment through endothelium and epithelium generally
occur in opposite directions with respect to the luminal surface
of cells. For endothelium, immune cells leave the circulation in
an abluminal direction moving from apical-to-basal endothelial
surfaces, whereas for epithelium, cells move toward the lumen
traversing a basal-to-apical direction with respect to epithelial
cell surfaces. In the case of the endothelium, this directional
process may be coordinated by the actions of selectins, cell
adhesion molecules, and chemokines all acting as haptotaxins
(90), but the driving force for immune cell movement across
the epithelial barrier (if it exists at all) was uncertain. Thus
expression of epithelial ICAM-1 and consequent interaction
with T cell LFA-1 appeared to be the major determinant of T
cell adhesion to the apical surface of the airway epithelial cell
monolayers (98, 99), but the extent to which ICAM-1/LFA-1
interaction or other molecular interactions might regulate T
cell traffic through the epithelium and the directionality of
movement was uncertain for airway and other epithelia.
Accordingly, we next developed a system for monitoring immune cell adhesion and transmigration through
an epithelial model in apical-to-basal and basal-to-apical
directions (129). Immune cell (in this case, T cell) behavior was again monitored by quantitative flow cytometry to
avoid a need for extensive leukocyte purification, culture,
Physiol Rev • VOL
23
and labeling. Epithelial monolayers were established with
primary-culture human tracheobronchial epithelial (hTBE)
cells that exhibit differentiated structural and functional
features of polarized epithelial barriers found in situ (149). In
particular, monolayers of hTBE cells emulate in vivo behavior with low basal levels of ICAM-1 and cytokine-dependent
increases in ICAM-1 expression (66, 79, 114). In this ex vivo
system, T cell adhesion and subsequent transmigration were
blocked in both directions by monoclonal antibodies against
LFA-1 or ICAM-1 (induced by IFN-␥ treatment of epithelial
cells). The total number of adherent plus transmigrated T
cells was also similar in both directions, and this pattern fit
with uniform presentation of ICAM-1 along the apical and
basolateral cell surfaces. However, the relative number of
transmigrated to adherent T cells (i.e., the efficiency of
transmigration) was increased in the basal-to-apical relative
to the apical-to-basal direction, so an additional mechanism
was needed to mediate directional movement toward the
apical surface. Screening for epithelial-derived T cell chemokines indicated that IFN-␥ treatment caused predominant
expression of RANTES (68). The functional significance of
RANTES production was then demonstrated by inhibition of
epithelial-T cell adhesion and transepithelial migration by
anti-RANTES monoclonal antibody. In addition, we found
that epithelial (but not endothelial) cells preferentially secreted RANTES through the apical cell surface, thereby
establishing a chemical gradient for chemotaxis across the
epithelium to a site where they may be retained by high
levels of RANTES and apical ICAM-1. In this system,
RANTES did not account for all chemotactic activity, and
subsequent studies have indicated that other chemokines
(and other chemotaxins such as IL-12 p80) provide for functional redundancy (see sect. III). Nonetheless, the results
define how the patterns for cell-specific apical sorting of
RANTES serve to mediate the level and direction of T cell
traffic and provide a basis for how this process is precisely
coordinated to route immune cells to the mucosal surface
and maintain them there (Fig. 2). Potent effects of RANTES
on T cells (4), eosinophils (2), and macrophages (as described below), distinct from effects on cell movement, indicate an additional role of epithelial cells in regulating the
activation status of local immune cells.
In sum, it appears that airway epithelial-T cell adhesion
and transmigration depend on the cytokine-dependent expression of ICAM-1 on the epithelial cell surface (98), concomitant
levels of expression and activation of LFA-1 on the T cell
surface (99), and polarized secretion of RANTES through the
apical cell surface (129). For ICAM-1, distribution on both cell
surfaces mediates efficient cell adhesion at the basal cell surface (to aid in transmigration) and at the apical cell surface (for
retention and movement along the airway). For RANTES, the
pattern of preferential apical secretion provides for a soluble
chemical gradient for T cell movement from the subepithelium
(where levels are low) to the mucosal surface (where levels
are higher). These patterns fit well with available data for
82 • JANUARY 2002 •
www.prv.org
24
HOLTZMAN ET AL.
FIG. 2. Molecular interactions that mediate immune cell adhesion to airway epithelial cells, transepithelial cell
migration, and retention at the airway lumen. Critical steps include the following: 1) initial integrin [e.g., leukocyte
function-associated antigen 1 (LFA-1)] binding to cell adhesion molecule [e.g., intercellular adhesion molecule-1
(ICAM-1)] on the basolateral epithelial cell surface; 2) subsequent ␤-chemokine receptor (e.g., CCR5) binding to
chemokine (e.g., RANTES) and migration along the chemical gradient for RANTES, which is secreted preferentially to
the apical epithelial cell surface; and 3) renewed LFA-1 binding to ICAM-1 for retention and/or migration along the apical
cell surface. Examples for cell adhesion molecules and chemokine and their receptors are most applicable to mononuclear cell traffic. [Modified from Holtzman et al. (53).]
ICAM-1 and RANTES expression in airway epithelium in
vivo (see below), but we emphasize that this pattern of cell
adhesion molecule expression and chemokine secretions is
distinct. Other types of epithelial cells (notably colonocytes
and pneumocytes) express ICAM-1 only along the apical cell
surface (11, 59, 106). Perhaps in this setting, ICAM-1 functions mostly in cell movement along that surface or acts in
concert with other cell adhesion receptors to aid in host
defense (e.g., as opsonins) (102). Similarly, epithelial capacity for polarized secretion of RANTES may be distinct from
other cell types. In the case of endothelial cells, it appears
that chemokines with immobilization domains (including
RANTES and IL-8) are anchored to the cell surface to act as
haptotaxins rather than freely secreted to act as chemotaxins (41, 113, 131). Apparently in this haptotactic mode, there
is no mechanism for polarization because both luminal and
abluminal surfaces of endothelial cells are coated with chemokine (90). It is also possible that the cellular source of
endothelial-bound chemokine is the epithelial cell, since
Physiol Rev • VOL
transcytosis and subsequent presentation of exogenous chemokine may occur in either direction. Taken together with
comparative data for endothelial cells, our studies of airway
epithelial cell-T cell interaction offer a means for progressive
movement of T cells from endothelium to the airway lumen
through distinct cell-specific mechanisms for cell adhesion
molecule expression and chemokine secretion. As noted
below in studies of human subjects, these same molecular
systems appear activated in asthma because there are increased numbers of activated T cells as well as increased
levels of ICAM-1 and RANTES expression in the epithelium
of asthmatic subjects depending on the severity of disease
and treatment conditions.
B. Cytokine-Dependent Gene Network:
Transcriptional Regulation
Section IIA concentrated on the topological basis for
how epithelial cells coordinate their expression of cell
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
adhesion molecules and chemoattractants with that in the
underlying vascular plexus so that leukocytes migrate
efficiently through the tissue. Another challenging question is how this process is coordinated to be stimulus
specific. The first clue to this process was the finding that
vascular endothelial and airway epithelial cells have distinct but complimentary profiles of cytokine responsiveness for induction of ICAM-1 and ICAM-1-dependent cell
adhesion and transmigration. In particular, it appears that
airway epithelial cells exhibit a selective sensitivity to
IFN-␥ (79), and this selectivity in cytokine responsiveness
of ICAM-1 expression offered an opportunity to decipher
the genetic code of airway inflammation. In general, inducible gene expression may be mediated by signal transduction leading to regulation at transcriptional or posttranscriptional levels (controlled by DNA- or RNA-protein
interactions, respectively). In the case of the ICAM-1
gene, nuclear run-off assays aimed at monitoring transcriptional initiation rate indicated that IFN-␥ regulation
of epithelial ICAM-1 levels occurs at least in part at a
transcriptional level (80). Because gene transcription is
often regulated by DNA-protein interactions in the promoter region, the basis for selective cytokine control of
ICAM-1 expression could be investigated by structurefunction analysis of the ICAM-1 gene promoter region.
This section therefore summarizes findings on one aspect
of this genetic code: the DNA-protein interactions in the
ICAM-1 gene promoter region that control IFN-␥-inducible transcription of the ICAM-1 gene. As will be developed, the molecular building blocks used to regulate the
ICAM-1 gene are also used to control an entire immuneresponse gene network in the epithelium.
Initial identification of DNA-protein interactions is
often undertaken with a functional analysis of promoter/
reporter gene constructs and a concomitant analysis of
proteins that bind to any identified regulatory sites. When
this approach was used to analyze IFN-␥-inducible expression of the ICAM-1 gene in primary-culture airway
epithelial cells, a 37-bp region of the gene from nucleotides ⫺130 to ⫺93 was found to be responsible for selective IFN-␥ responsiveness of airway epithelial cell ICAM-1
expression (80). This region (designated the IFN-␥-response element or IRE) contained an 11-bp inverted repeat (the gamma-activation site) that was necessary and
sufficient for IFN-␥-responsiveness of the ICAM-1 gene.
This function was closely linked to the capacity of this
site to bind to Stat1 (the first member of the signal transduction and activation of transcription factor family). Because vascular endothelial cells also appear capable of
Stat1 activation and binding in response to IFN-␥ treatment, their failure to fully respond to IFN-␥ may be due to
the presence of a concomitant interaction of a repressor
protein that prevents the response. This possibility may
be reflected by restoration of IFN-␥ responsiveness in
endothelial cells by protein synthesis inhibition with cyPhysiol Rev • VOL
25
cloheximide treatment (D. C. Look and M. J. Holtzman,
unpublished observations). The relatively small ICAM-1
response to IFN-␥ in endothelial cells may be reconciled
with the fact that IFN-␥-targeted viruses are not often
encountered at the endothelial barrier. However, as discussed next, the shared capacity of both endothelial and
epithelial cells to utilize an alternative DNA element for
IFN-␥-activation of the MHC class I genes may provide a
common pathway for antigen recognition and processing
that is essential for immune defense against inhaled as
well as circulating agents.
Additional studies of airway epithelial cells in the
context of work on other cell types by us and others
indicate that transcriptional regulation of the epithelial
ICAM-1 gene is typical of genes controlled by an IFN-␥responsive Janus family kinase (Jak)-Stat signaling pathway (summarized in Fig. 3). This pathway consists of the
IFN-␥-receptor, receptor-associated Jak1 and Jak2 tyrosine kinases, the Stat1 transcription factor, and specific
Stat1/DNA and Stat1/protein interactions in the ICAM-1
gene promoter region (79, 80, 141). Signaling depends on
the capacity of Stat1 to relay the signal from the cytoplasm to the nucleus through its capacity to interact both
with the IFN-␥-receptor ␣-chain (46) as well as the GAS
site in the gene promoter region (80). In addition to
Stat1/DNA interaction, it appears that Stat1 interaction
with the constitutively active transcription factor Sp1 is
also critical for full activation of the ICAM-1 gene (81). At
the promoter, direct and indirect interactions with the
coactivators p300/CBP and p/CIP may further facilitate
enhanceosome formation and more efficient gene transcription, and these components appear active in airway
epithelial cells as well (58, 82, 134, 153). In other cell
systems, an additional action of a serine/threonine mitogen-activated protein kinase is needed to convey full activity, but this step appears less influential in airway epithelial cells (141).
In further work, it appears that these same molecular
building blocks serve to regulate IFN-␥ responsiveness of
other epithelial immune-response genes. Thus Stat1 binding also confers IFN-␥ responsiveness of genes for transporter and antigen processor-1 (TAP-1), interferon regulatory factor-1 (IRF-1), and Stat1 itself (80 – 82, 115, 141).
As discussed further below, induction of Stat1 and consequent autoamplification of this pathway may serve to
exaggerate the inflammatory response in experimental
models and in disease. Similarly, the IRF-1 gene product is
also a member of a transcription factor family and so may
amplify the pathway (in concert with other factors) by
activating the genes for inducible nitric oxide synthase
(iNOS), MHC class I molecules, and IFN-␣/␤ (35, 42, 65).
In turn, IFN-␣/␤ (generated as a result of these events or
more directly in response to replicating virus) may activate the IFN-␣ receptor complex and overlapping signaling components to activate additional antiviral genes, in-
82 • JANUARY 2002 •
www.prv.org
26
HOLTZMAN ET AL.
FIG. 3. Signal transduction pathway from cell surface (left) to nucleus (right) for cytokine (IFN) activation of an
airway epithelial immune-response gene network. IFN-␥ signaling begins when IFN-␥ binds to its heterodimeric receptor
(IFN-␥R) and triggers activation of ␣-chain-associated Jak1 and ␤-chain-associated Jak2 tyrosine kinases and consequent
␣-chain phosphorylation. This step enables ␣-chain recruitment of Stat1 via its SH2 domain and subsequent Stat1
phosphorylation and release of the phosphorylated Stat1 homodimer. Activated Stat1 homodimer translocates to the
nucleus where it binds to an inverted-repeated motif designated the gamma-activation site (GAS) and activates (in
concert with Sp1, p300/CBP, and p/CIP) transcription of a subset that includes the ICAM-1, TAP-1, IRF-1, and Stat1 genes.
Increased levels of Stat1 may then act to autoamplify Stat1-dependent gene activation, while newly generated IRF-1 may
bind to an IFN-response stimulation (IRS) site and activate (in concert with other factors) a second wave of transcription, including the HLA-B, inducible nitric oxide synthase (iNOS), and IFN-␤ genes. IFN-␤ driven gene expression is
initiated by activation of the IFN-␣/␤ receptor (IFN-␣R) and subsequent activation of IFN-␣R1-associated Tyk2 and
IFN-␣R2-associated Jak1 with consequent IFN-␣R1 phosphorylation and recruitment of Stat2. Phosphorylation of Stat2
enables recruitment of Stat1 and release of the phosphorylated Stat1/Stat2-heterodimer. This heterodimer in concert
with IRF-9 (p48) forms the interferon-stimulated gene factor 3 (ISGF3) complex that binds to the interferon-stimulated
response element (ISRE) and increases transcription of a subset that includes the MxA, IRF-1, and Stat1 genes. For IFN-␤
signaling, increased levels of Stat1 may autoamplify Stat1-dependent gene activation, but this pathway may also be
effectively downregulated by paramyxoviral (e.g., RSV) infection. However, as discussed in Fig. 4, paramyxoviruses may
activate other epithelial immune-response pathways for host defense.
cluding IRF-1 and Stat1 (24; M. Lo and M. J. Holtzman,
unpublished observations). Thus a cascade of common signaling components enables IFN-␥ to efficiently activate a
subset of immune-response genes that are oriented toward
several levels of antiviral defense. These include 1) antigen
recognition and T cell costimulation as well as immune cell
recruitment (ICAM-1); 2) antigen processing (TAP-1); and 3)
amplification of the immune response (Stat1 and IRF-1) with
additional capacity for antigen recognition (MHC class I)
and antiviral toxicity (iNOS and IFN-␣/␤).
Based on the apparent efficiency of epithelial immune-response genes for antiviral defense, one might exPhysiol Rev • VOL
pect little problem in coping with respiratory viruses.
However, this is clearly not the case, and these failures of
the immune response provide another informative strategy for unraveling the genetic code of epithelial-dependent immunity. Thus viruses have established an array of
strategies for immune subversion (108). In the case of
riboviruses with relatively small genomes, molecular adaptation may take advantage of host cell machinery to
subvert the immune response. In fact, ICAM-1 itself was
discovered in part by its capacity to also serve as the
receptor for the major group of human rhinoviruses (47,
123). Paramyxoviruses appear to block type I IFN signal-
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
ing (Lo and Holtzman, unpublished observations). Other
viruses may encode proteins with the capacity for molecular mimicry that specially target IFN-␥-dependent immunity by modifying host gene expression. In particular,
adenoviral E1A oncoprotein may interfere with p300/CBP
action and may also directly target Stat1 (82). The molecular basis for this action and for how it can be used to
develop anti-inflammatory strategies is discussed further
in the section on strategies for correcting abnormalities in
epithelial signaling and remodeling (sect. VI).
C. Virus-Dependent Gene Network:
Posttranscriptional Regulation
Similar to the case for ICAM-1, only little was previously known for mechanisms that regulate RANTES gene
expression (or ␤-chemokine gene expression in general) (6).
On the basis of the standard approach noted above for
defining cis- and trans-acting controls for gene transcription, initial studies by others suggested that RANTES gene
expression depended on NF-␬B sites in the RANTES promoter region (93, 100, 111). However, no attempt was made
to determine the functional importance of these sites by
direct determination of transcriptional initiation rate, and no
information was available on any possible role of posttranscriptional regulation of the RANTES gene. This section
summarizes our recent findings related to transcriptional
and posttranscriptional regulation of the RANTES gene in
airway epithelial cells using systems for IFN-␥- and virusinducible expression in isolated cells. Later sections define
how this system behaves in vivo.
In initial experiments (as noted above), we determined
the capacity of primary culture airway epithelial cells to
produce chemotaxins that could mediate T cell transmigration (68, 129). In this setting, we found that IFN-␥ caused
production of RANTES consistent with the IFN responsiveness of this cell type. However, we were interested in determining whether airway epithelial cells could respond directly to viral infection without a requirement for signals
from immune cells. We were especially interested in
whether the types of viruses associated with asthma might
produce distinct effects on epithelial immune-response gene
expression. The effects of paramyxoviruses appeared particularly relevant, based on the epidemiological evidence
that members of the Paramyxoviradae family, especially
respiratory syncytial virus (RSV) and parainfluenza viruses,
were closely associated with recurrent wheezing and
asthma in young children (19, 26, 107, 119, 122, 124). In that
context, we were interested to find that RSV infection of
human airway epithelial cells (using an ex vivo system lacking IFN-␥ or other immune cell contribution) caused induction of RANTES gene expression in marked excess of other
␤-chemokine genes (69).
We next aimed to define the mechanism responsible
Physiol Rev • VOL
27
for virus-inducible RANTES production. Previous analysis
of the host response to viral infection has generally focused on the capacity of viruses to activate or repress
transcription of cellular genes (52, 87), and this approach
is also characteristic of work on riboviruses. In related
examples from the Paramyxoviruses, the effect on host
genes is mediated by DNA regulatory elements that bind
NF-␬B, IRF-1, ATF-2/c-Jun, and high mobility group protein HMG-I(Y) in the IFN-␤ gene or HMG-I(C) in the
RANTES gene (78, 130, 132). Accordingly, we assumed
that NF-␬B sites in the RANTES gene promoter (93, 100,
111) might be responsible for virus induction of RANTES
gene expression. Indeed, it initially appeared that RSVdriven expression of epithelial RANTES also depended on
inducible gene transcription because expression was accompanied by coordinate increases in transcriptional initiation rate and gene promoter activity. However, RSV-driven
increases in RANTES gene transcription and promoter activity were small and transient relative to RANTES expression, and they were no different in size and duration than for
inactivated RSV that was incapable of inducing RANTES
expression. These findings suggested that the increase in
RANTES gene transcription was required but not sufficient
for inducible expression and that critical regulatory effects
occurred at a posttranscriptional level. This type of mechanism for virus-inducible expression of RANTES was established when we found that replicating RSV markedly increased RANTES mRNA half-life (69).
In contrast to what appear to be likely DNA-protein
interactions for mediating transcriptional activation of
host genes by viruses and other stimuli, little is known
about viral mechanisms for controlling mRNA stability. In
concomitant work on cytokine responsiveness of epithelial immune-response genes (including ICAM-1, IRF-1, and
RANTES), we have noted that IFN (like RSV) stimulates
RANTES production via mRNA stabilization (68). It is
therefore possible that RSV-driven signals for altering
stability of RANTES mRNA may overlap with those for
IFN-␥ signal transduction. However, the available examples for IFN-␥-dependent increases in mRNA stability appear distinct from the characteristics of RANTES expression. For example, IFN-␥ stabilization of ICAM-1 mRNA is
mediated by a region of the translated sequence encoding
the ICAM-1 cytoplasmic domain (103), but the RANTES
gene (encoding a secreted protein) does not contain this
sequence. Furthermore, this instability mechanism (in
contrast to the one for RANTES) is uninfluenced by actinomycin D treatment. Similarly, IFN-␥ upregulates expression of the complement components C3 and C4 by
stabilization of mRNA, but this system is also uninfluenced by transcriptional inhibition (91). In addition, the
RANTES mRNA does not contain consensus sites for
previously identified mRNA turnover elements, including
AU-rich elements in other cytokine (as well as ICAM-1)
mRNAs or UC-rich cleavage sites in gro ␣ and 9E3 mRNAs
82 • JANUARY 2002 •
www.prv.org
28
HOLTZMAN ET AL.
(13, 103, 126, 127). Thus the mechanism underlying basal
instability as well as RSV-dependent stabilization of
RANTES mRNA may be biochemically distinct from other
genes. Indeed, RNase protection assays of heterologous promoter/reporter plasmids indicate that basal instability of
RANTES mRNA is mediated at least in part by nucleotides
11–389 of the RANTES gene, and this region is also the
target for induction by virus (69). This region contains a
distinct RNA turnover element in the 3⬘-untranslated region
(UTR) that forms a complex with a putative RANTES-binding factor (RBF) under basal conditions but not during RSV
infection (A. Antao, W. Roswit, M. Pelletier, and M. J. Holtzman, unpublished observations) (Fig. 4). The precise nature
of this RNA-protein interaction and how it is regulated by
viral replication still needs to be determined.
Even at this point, however, the findings provide the
basis for an alternative model for virus-dependent induction
of epithelial immune-response genes (depicted in Fig. 5). In
this model, cytokine- (especially IFN-␥)-dependent induction of immune-response genes depends on transcriptional
activation of gene expression, whereas direct viral induction
of expression may be regulated by transcriptional or posttranscriptional events. Transcription may depend on viral
interaction with Toll-like receptors (TLR) and activation of
NF-␬B dependent pathways (71, 132, 133). This type of
regulation depends on viral surface proteins (e.g., RSV F
protein) and may therefore be triggered by dead or live
virus. In addition, viruses may also act downstream at the
posttranscriptional level. This action requires viral replication and alters gene expression by stabilizing mRNA (which
is the case for RANTES) or improving protein translation
and processing (which appears to be the case for other
epithelial immune-response genes). These two actions, activation of gene transcription and stabilization of mRNA,
would be highly synergistic for gene expression. In fact, our
experience has been that viral induction of epithelial gene
expression is minimal in the absence of viral replication and
consequent action at a posttranscriptional level (69; A. Antao, M. Pelletier, and M. J. Holtzman, unpublished observations).
D. The Final Common Pathway:
Epithelial Cell Death
Each of the systems described above aims to explain
how epithelial cells communicate with the immune sys-
FIG. 4. Regulation of RANTES mRNA turnover under basal conditions and during viral infection. Under basal
conditions, a trans-acting mRNA turnover factor (i.e., RBF) may be generated and activated and so be available to bind
to a cis-acting mRNA turnover element in the 3⬘-untranslated region (UTR). RBF-mRNA interaction alters the conformation of the mRNA to render it susceptible to degradation by active endoribonuclease (RNase) at the degradation target
site. During viral infection (e.g., by RSV), the RBF is prevented from binding (by downregulating the level or activation
status) and from mediating RNase action. This mRNA stabilization may synergize with transcription to mediate marked
increases in gene expression. [Modified from Holtzman et al. (53).]
Physiol Rev • VOL
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
29
FIG. 5. Regulation of epithelial immune-response gene expression by IFN-␥ or paramyxovirus (e.g., RSV). For IFN-␥
signaling (1), sequential steps include IFN-␥R activation/phosphorylation, Stat1 translocation and interaction at the
interferon regulatory element (IRE) in the gene promoter region, generation of pre-mRNA transcripts, and mRNA
maturation and release to ribosomal machinery for translation and then processing of protein for expression at the
appropriate cellular site. The mRNA transcripts may alternatively interact with mRNA turnover factors such as
RANTES-binding factor (RBF) that mediate rapid mRNA degradation by endoribonuclease (RNase). For paramyxovirus
signaling (2), RSV may interact with Toll-like receptors (TLR) and so activate NF-␬B and consequent NF-␬B-dependent
gene transcription. This type of transcriptional regulation depends on viral surface proteins and may therefore be
triggered by dead or live virus. In addition, RSV may act downstream at the posttranscriptional level and alter gene
expression by stablizing mRNA or improving protein translation and processing. These posttranscriptional events require
live, replicating virus.
tem. In the setting of viral infection, this communication
is generally designed to recruit immune cells that destroy
the infected host (epithelial) cells and then clear the
cellular debris from the site. In this setting, cytotoxic T
cells can mediate epithelial cell death via Fas-dependent
and perforin/granzyme-dependent pathways (96), and
macrophages may clear the debris from the site. We will
return to the specifics of this process in section III, which
defines the host response to viral infection in vivo, but at
this point, we introduce the role of intrinsic pathways to
regulate airway epithelial cell death and survival during
the course of viral infection.
The approach to defining epithelial life and death
pathways that are targeted during acute respiratory viral
infection begins with the premise that the virus and the
host have opposing motives. The virus aims to first maintain host viability (to allow for viral replication) and then
trigger cell death (to allow for lysis and viral spread to
adjacent cells). In contrast, the epithelial host aims to
destroy itself as rapidly as possible, presumably via apoptosis, to protect its neighbors from infection and further
Physiol Rev • VOL
inflammation. Thus self-programmed epithelial cell death
is an effective innate host defense mechanism. For example, activation of caspase-dependent epithelial death pathways by the Fas death receptor may be critical for host
defense against Pseudomonas aeruginosa infection in
mice (45). However, we have found that primary culture
human airway epithelial cells exhibit minimal Fas-dependent cell death unless the response is augmented by concomitant treatment with cycloheximide or actinomycin D,
and paramyxovirus-inducible cell death is not influenced
by Fas blockade (105). Thus the relevance of this system
remains uncertain, especially in response to intracellular
stimuli like respiratory viruses.
Recognizing that the pathways governing cell death
versus survival are complex, we have analyzed the behavior of airway epithelial cells during viral infection using an
oligonucleotide microarray in conjunction with assays of
cell death parameters, e.g., caspase activation and mitochondrial dysfunction. Our preliminary results indicate
that RSV infection causes mitochondrial dysfunction, but
lethal effects appear to be delayed until viral replication is
82 • JANUARY 2002 •
www.prv.org
30
HOLTZMAN ET AL.
FIG. 6. Schematic diagram for representative pathways that mediate epithelial cell death (left box) or survival (right
box) during acute viral infection. The death box features pathway initiation by TNFR superfamily members (e.g., Fas)
that recruit death domain-containing adapter proteins (e.g., FADD and TRADD) and then activate caspases that mediate
degradation of structural proteins and apoptosis. Other signals trigger pro-apoptotic BCL-2 family members (e.g., BAX)
that lead to mitochondrial damage and then to apopotsome formation and apoptosis or to necrosis. This pathway
appears to be the primary target for cell death induced by TNF-␣ plus IFN-␥. The survival box features inhibition of this
same pathway by anti-apoptotic members of the BCL-2 protein family (e.g., BCL-2). Other survival pathways include
activation of Rho family members, growth factor receptor tyrosine kinases (RTK), or TNFRII receptors, each leading to
activation of MAP kinases (MAPK) that may influence BCL-2 proteins and/or gene transcription that supports cell
survival. In addition, inhibition of cell cycle proteins may be necessary to prevent inactivation retinoblastoma protein
(pRb) and so allow for cell cycle progression to S phase. Preliminary data indicate that paramyxoviruses (PMV) may
influence each of these steps: 1) downregulation of BAX; 2) upregulation of BCL-2; 3) downregulation of cyclins; 4)
upregulation of Rho; and 5) upregulation of TNFRII pathways to facilitate viral replication in the epithelial host cell.
well developed (M. O’Sullivan, K. Takami, and M. J. Holtzman, unpublished observations). The initial phase may
rely on alterations in Bax-, TNF-, growth factor-, and cell
cycle-dependent pathways that are each pushed toward
gene expression that would favor cell survival during the
acute phase of RSV infection (Fig. 6). Presumably,
paramyxoviruses have developed a diverse program for
preserving host cell viability during this early phase of
infection, because their survival depends on it. Concomitant with these events, we also observe changes in Bcl2-dependent pathways that would lead to mitochondrial
dysfunction and cell death. Whether this change reflects
host defense or the initiation of cell lysis by the virus will
require additional work. In either case, an alteration in
this step may provide a target for improving control of
paramyxoviral infection, even before immune cells arrive
at the scene. We still know little about how these life and
Physiol Rev • VOL
death pathways may be altered in airway disease, but the
indispensable role for them to govern cell growth and
differentiation argues that they must be involved in pathogenesis. For example, preliminary evidence suggests that
the Fas death pathway may be abnormally regulated in
immune cells in asthma (61). In section V, we present
experimental evidence for virus-inducible alterations in
the epithelial repair process, and these alterations are
also linked to the controls for epithelial cell death versus
growth and differentiation.
III. AIRWAY IMMUNITY AND INFLAMMATION
IN A MOUSE MODEL OF VIRAL INFECTION
Results from murine models of airway inflammation
have been reported on numerous occasions (21, 22, 36, 43,
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
83, 97, 116, 121) and reviewed by us and others (28,
55–57). In the present context, we aimed to develop a
mouse model to better determine how the airway epithelial system operates in vivo and to subject the system to
genetic modification. Because the epithelial system is programmed for antiviral defense, we also aimed for a model
with high fidelity to viral bronchiolitis in humans. Although RSV is often used for studies of human airway
epithelial cells and human subjects, based on its link to
childhood asthma, this virus does not cause a similar type
of bronchiolitis in the mouse. For that reason, we chose
another Paramyxoviridae family member, mouse parainfluenza type I or Sendai virus based on its capacity, as a
natural pathogen in rodents, to cause top-down infection
leading from the nose to the bronchi to the bronchioles to
the alveoli. By limiting the inoculum, infection is limited
to the airways and so resembles the pathology and pathophysiology of the human condition (136, 137, 139, 140).
This pathogenesis is the same as observed in human
subjects with paramyxoviral infection and so offers an
appropriate model for analysis. Sendai virus also infects
isolated human airway epithelial cells, presumably because the allantoic fluid medium provides the Clara cell
trytpase that is ordinarily provided by the rodent airway.
Thus heterologous cell systems can be used to validate
fidelity to RSV behavior when possible. This section summarizes our initial experience with this model for first
identifying gene expression that is prominently inducible
in airway epithelial cells during paramyxoviral infection
and then determining the response in same-strain mice
with targeted mutagenesis of these genes.
A. Epithelial Gene Expression: Interaction
Between Virus and Host Cell
The initial step in defining epithelial cell-dependent
immunity came with viral induction of epithelial immuneresponse genes in vitro. The second step came with induction of the same genes in vivo. In particular, we have
found that paramyxoviral bronchiolitis causes induction
of epithelial ICAM-1, Stat1, and RANTES gene expression
in a pattern that is remarkably similar to the one observed
in isolated airway epithelial cells. Accordingly, it appears
that the mechanism for viral induction of epithelial immune-response genes in vivo may be similar to the posttranscriptional mechanism observed in vitro for RANTES.
Thus, like RANTES, Stat1 gene expression is markedly
upregulated in the epithelial host cell by Sendai viral
infection in vivo (138), similar to the earlier experience
with RSV infection of isolated human airway epithelial
cells (T. Koga, D. Sampath, M. Lo, and M. J. Holtzman,
unpublished observations). The sequence of the 5⬘-regulatory region of the mouse Stat1 gene shows a GAS consensus site for binding Stat1 homodimer in a location
Physiol Rev • VOL
31
comparable to the human gene promoter region. However, this type of activation would require activation of
Stat1 by IFN-␥, and we have found that IFN-␥-deficient
mice exhibit the same virus-inducible expression of epithelial Stat1 in vivo. In contrast to human, there are no
putative regulatory sites for binding ISGF3, NF-␬B, or
IRF-1 in the mouse Stat1 promoter. These findings suggest
that paramyxovirus may directly alter Stat1 gene expression without a direct requirement for IFN- or NF-␬Bdependent signaling. Consistent with these findings, experiments using isolated airway epithelial cells indicate
that RSV causes little change in transcription rate of the
Stat1 gene despite marked increases in gene expression
(T. Koga, M. Pelletier, A. Antao, and M. J. Holtzman,
unpublished observations). Additional definition of Stat1
(and other epithelial) gene expression is needed, but each
of the available findings points to the possibility that
viruses trigger the network of epithelial immune-response
genes at a posttranscriptional level using a mechanism
that is sensitive to viral replication.
Despite the uncertainty over the precise mechanism
for viral regulation of epithelial gene expression, it is
evident that the site and the timing of viral replication
correlate closely with induction of expression. Thus both
events are localized predominantly and coordinately to
epithelial cells lining the airway. The entire profile of
epithelial gene expression in this setting is still being
defined by oligonucleotide microarray, but initial results
continue to suggest close correlation between in vitro and
in vivo changes in expression. As was the case in vitro,
induction of epithelial gene expression in vivo also requires viral replication and is uninfluenced by preparations of inactivated virus. Furthermore, epithelial gene
expression is followed closely by immune cell accumulation and activation at the site of viral infection. Each of
these findings reinforces the proposal that the intrinsic
response of the epithelial host cell to replicating virus
may be critical for innate immunity, and so begs the
question to more precisely define the role of the barrier
epithelial cell in this setting. The next section describes
our current understanding of epithelial function for host
defense in the short term, and subsequent sections will
define the implications of epithelial-viral interaction for a
more chronic process.
B. Toward Epithelial Gene Knockouts: Defining
an Active Role in Innate Immunity
This section summarizes the behavior of mice with
targeted null mutations in selected immune-response
genes in the setting of paramyxoviral infection. Because
the tools are still being developed to achieve specific
targeting, in airway epithelial cells and epithelial cell subsets (i.e., ciliated, Clara, goblet, or basal cells), the results
82 • JANUARY 2002 •
www.prv.org
32
HOLTZMAN ET AL.
TABLE 1. Influence of immune-response genes in paramyxoviral bronchitis/bronchiolitis and bronchopneumonia
based on behavior of inbred mice with targeted null mutations and SdV infection
Gene Target
Viral Clearance
Infiltrate
Mortality
Airway Reactivity
IFN-␥*
IL-12 p40*
IL-12 p35
ICAM-1*
Stat1*
RANTES*
No change
No change
No change
Decreased
Decreased
Decreased
No change
No change
Increased (Mac)
Decreased (PMN/L)
Increased (PMN)
Increased (Mac)
No change
No change
Increased
Decreased
Increased
Increased
No change
No change
No change
Decreased
Mac, macrophage; PMN, polymorphonuclear leukocyte/neutrophil; L, lymphocyte; IFN-␥, interferon-␥; IL, interleukin; ICAM-1, intracellular
adhesion molecule-1.
* Genes that exhibit virus-inducible expression in airway epithelial cells during infection.
do not yet fully determine whether and which type of
epithelial cell may be critical for innate immunity to respiratory viruses. Nonetheless, even the current evidence
from the mouse model strongly favors the hypothesis that
epithelial cells contain a specialized genetic program that
is critical for antiviral host defense as catalogued below
and summarized in Table 1.
1. IFN-␥ and IL-12
The possibility that epithelial barrier cells might contribute to host defense first came from experiments that
examined the role of immune cells. Thus the Th1-dependent response has been traditionally designated as critical
for defense against respiratory viruses, and essential regulators of Th1 cell responses are IFN-␥ and IL-12 (94).
Accordingly, these cytokines might be expected to be
derived from immune cells and be essential for host defense in the setting of viral infection. However, IFN-␥
deficiency caused no significant change in the response to
paramyxoviral infection (139). In particular, induction of
epithelial immune-response genes proceeded at the usual
level, indicating that viral effects on the epithelial cell did
not depend on IFN-␥ and instead might reflect direct
actions of replicating virus within the host cell (as noted
above for viral effects on epithelial cells in vitro). In
addition, IFN-␥-deficient mice exhibited no defect in recovery from paramyxoviral infection, indicating that the
epithelial program may be capable of directly helping
with protection or for arming other aspects of the mucosal immune response.
The response of IL-12-null mice to paramyxoviral
infection proceeded with a similar lack of immunocompromise but was even more informative for the role of the
epithelium. Thus IL-12 is a heterodimeric protein consisting of p35 and p40 protein subunits. Initial results with
null mutations targeted to the IL-12 p35, p40, or both IL-12
subunit genes produced mice that exhibited no defect in
clearance of Sendai viral infection, indicating that IL-12
like IFN-␥ production was also not essential for antiviral
defense (137). Somewhat unexpectedly, however, mice
with IL-12p35 deficiency exhibited increased airway inPhysiol Rev • VOL
flammation (characterized by excessive macrophage accumulation) and increased mortality during infection. Because IL-12 is generally produced by antigen-presenting
cells (i.e., macrophages, dendritic cells, B cells) (23, 84,
125, 135), we next defined the site of IL-12 induction and
surprisingly found that expression was inducible by viral
infection and was predominantly localized to airway epithelial cells. Initial IL-12 induction was followed by excessive expression of IL-12 p40 (often as homodimer IL12p80) that could be further enhanced in IL-12 p35deficient mice. Others have provided evidence that IL-12
p40 may function as an antagonist of IL-12 action (50, 67,
76, 152), but in the present case, its production was
associated with increased mortality and epithelial macrophage accumulation. Although toxicity has been observed
for overproduction of IL-12 (37, 112), inflammation due to
IL-12 p40 had not been observed. Thus the results placed
epithelial cell overgeneration of IL-12 p40 as a key intermediate for virus-inducible inflammation and as a candidate for epithelial immune-response genes that are abnormally programmed in inflammatory disease. As noted
below, this possibility was further supported when we
observed increased expression of IL-12 p40 selectively in
airway epithelial cells in subjects with asthma and concomitant increases in airway levels of IL-12 p40 (as homodimer) and airway macrophages.
Taken together, these experiments with IL-12 suggest
a novel role for epithelial-derived IL-12 p40 in modifying
the level of airway inflammation during mucosal defense
and disease. The results also serve to introduce a theme
that will develop for each of the epithelial immune-response genes, i.e., the usual level of epithelial gene expression may aid in host defense, whereas excessive expression may lead to inflammatory disease. In this case,
epithelial IL-12 p80 production may help to mediate macrophage recruitment and/or activation, whereas higher or
inappropriate levels of production may lead to macrophage-dependent inflammation. This view favors cellular
recruitment as a feature of host defense, but from the
viral perspective, recruitment may facilitate viral spread
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
to additional cell populations, including neighboring epithelial cells that may also respond to chemotaxins.
2. ICAM-1
ICAM-1 is a predominant determinant of epithelial-immune cell adhesion, and transmigration in vitro and epithelial expression is prominently inducible by paramyxoviral
infection in vitro and in vivo. Thus mutagenesis of the
ICAM-1 gene was a natural target to influence airway inflammation in paramyxoviral bronchiolitis. As might be predicted, ICAM-1-null mice are indeed protected from airway
inflammation (as assessed by accumulation of immune cells
in airway tissue) induced by paramyxoviral infection compared with same-strain controls (136). The blunted inflammatory response in ICAM-1-null mice was accompanied by
less efficient viral clearance but was still beneficial to the
host, since this cohort experienced less weight loss after
bronchiolitis and lower mortality rates after a larger inoculum that causes bronchopneumonia.
We have long proposed that airway inflammation
may lead to airway hyperreactivity (53, 54), but definitive
proof has been difficult to obtain through pharmacological and physiological approaches. Accordingly, we next
determined if ICAM-1-null mice that are relatively protected from virus-induced airway inflammation are also
protected from postviral hyperreactivity. Indeed, using
whole body barometric plethysmography to measure enhanced pause (Penh) as an index of airway obstruction at
baseline and after methacholine challenge in wild-type
and ICAM-1-null mice, we found 1) slightly increased
baseline and reactivity by 1 wk postinoculation that was
partially blocked in ICAM-1-null mice, and 2) normal
baseline but markedly increased reactivity by 3 wk after
inoculation that was completely blocked in ICAM-1-null
mice. As noted above, the predominant site of ICAM-1
expression during paramyxoviral bronchiolitis is the airway epithelial cells that are home to the virus, but
whether this cellular site is the one responsible for downregulating inflammation and hyperreactivity will require
more cell-specific targeting. As discussed further in section V, genetic susceptibility to virus-inducible hyperreactivity can also be defined in this system, since only certain
strains of inbred mice develop the acute inflammation/
hyperreactivity phenotype as well as the subsequent and
persistent remodeling response.
3. RANTES
Experimental paramyxoviral bronchiolitis in mice
causes marked induction of RANTES gene expression in
the lung, and expression is localized predominantly to
airway epithelial cells and to adjacent tissue macrophages
(N. Kajiwara, M. J. Walter, M. O’Sullivan, J. Tyner, O.
Uchida, D. N. Cook, Y. Makino, T. M. Danoff, and M. J.
Holtzman, unpublished observations). This pattern correPhysiol Rev • VOL
33
lates precisely with the location of viral replication and so
fits closely with findings in humans with paramyxoviral
infection and in isolated cells, indicating that viral replication is a potent and direct inducer of RANTES gene
expression (69). Additional experiments indicate that
mice with targeted disruption of the SCYa5/RANTES gene
are immunocompromised to the point of overwhelming
viral infection and death (Kajiwara et al., unpublished
observations). This defect appears to be manifest because
RANTES is required to block apoptosis of virus-infected
macrophages. The physiological source of RANTES in
vivo could be either the epithelial cell or the macrophage,
so these experiments did not yet add to the primary role
of the airway epithelial cell in this setting. Nonetheless,
this chemokine function is distinct from ones that have
been previously identified in the setting of infection, such
as recruiting and activating immune cells, interfering with
viral entry receptors, or triggering cell death receptors,
and so establish a novel mechanism for host defense
based on preserving viability of the infected macrophage
via a distinct combination of antiapoptotic and antiviral
actions of a chemokine. Thus defense against intracellular
pathogens, particularly viruses, depends on programmed
death of infected host cells and then clearance of these
cells by phagocytic macrophages. For effective clearance
to take place, the viability of macrophages must be maintained in the face of infection, and these results suggest a
molecular basis for preservation of virus-infected macrophages. The precise molecular mechanism for RANTES to
influence cell death pathways still needs to be determined, but initial observations indicate that signaling to
the death pathway proceeds via specific chemokine receptors that are also susceptible to regulation by viral
replication (J. Tyner, O. Uchida, and M. J. Holtzman,
unpublished observations). However, the results with
RANTES- and IL-12-deficient mice reinforce another
theme of this system, i.e., both epithelial and macrophage
components of the innate immune response are critical
targets for the virus and for effective antiviral defense.
4. Stat1
Stat1 mediates the expression of a subset of interferon-inducible genes (typified by ICAM-1) in airway epithelial cells, so Stat1 deficiency should have a major impact on
epithelial function during viral infection. Indeed, Stat1-null
mice exhibit markedly increased weight loss and decreased
survival after Sendai viral infection (118). Thus intranasal
inoculation with low levels of Sendai virus causes little
effect in same-strain control mice but 100% mortality in
Stat1-null mice. These results are similar to reports of Stat1null mice in other viral models, but the basis for the defect
in host defense has not been determined (29). In the present
setting, Stat1 deficiency is associated with marked increases
in viral replication rates and severe airway inflammation
82 • JANUARY 2002 •
www.prv.org
34
HOLTZMAN ET AL.
with cellular infiltrate and debris in the lumen. This host
response may not simply reflect increased viral tissue load,
since the response is not observed in wild-type mice even at
a high inoculum that results in comparable viral load and
levels of mortality. The luminal infiltrate is comprised mainly
of neutrophils that contain virus and exhibit apoptosis and is
fully manifest even at 8 days after infection. Because neutrophils are ordinarily cleared by this time in wild-type mice
(ordinarily the responsibility of activated macrophages), the
results suggest a possible delay in neutrophil clearance in
Stat1-null compared with wild-type control mice.
These findings suggested that Stat1 in epithelial cells
(leading to enhanced viral replication) or macrophages
(leading to decreased activation) might underlie the defect in antiviral defense. Additional experiments with
bone marrow radiation chimeras were used to dissect the
role of Stat1 in the radiation-resistant compartment (especially the airway epithelium) versus radiation-sensitive
hematopoietic cells (including macrophages). In this
case, we found that lethally irradiated Stat1-deficient mice
reconstituted with wild-type bone marrow were still susceptible to infection with Sendai virus, whereas wild-type
mice that received Stat1-deficient bone marrow retained
resistance to virus (L. Shornick, D. Briner, M. Lo, and M. J.
Holtzman, unpublished observations). The viral susceptibility of chimeras with Stat1-deficient epithelium exhibited the same pattern of persistent luminal inflammation.
Taken together, the results suggest that epithelial Stat1
may be critical for antiviral defense, perhaps by limiting
viral replication and so aiding the removal of virus-infected, apoptotic neutrophils.
Because Stat1-deficient mice may have a selective
and profound defect in interferon signaling (30, 88) and
IFN-␥-deficient mice exhibit no significant immunocompromise with Sendai viral infection, the findings implicate
Stat1-dependent effects of IFN-␣/␤ signaling as critical for
antiviral defense against respiratory paramyxoviral infection. However, related experiments in isolated airway
epithelial cells indicate that paramyxoviruses downregulate IFN-␣/␤ signaling as part of a strategy to establish
infection (M. Lo and M. J. Holtzman, unpublished observations). Whether further decreases in this pathway or in
other pathways lead to Stat1-deficient immunocompromise in vivo still needs to be determined. Defining the
mechanism for Stat1-dependent protection in paramyxoviral bronchiolitis and the particular role of epithelial
(versus immune cell) function of Stat1 may finally define
a primary and distinct role of the epithelial barrier cell in
innate immunity.
IV. AIRWAY IMMUNITY AND INFLAMMATION
IN HUMAN SUBJECTS WITH ASTHMA
On the basis of the rationale that an abnormal immune response is part of inflammatory disease, it apPhysiol Rev • VOL
peared reasonable to next determine the behavior of epithelial immune-response genes in asthma. Initial studies
indicate that airway epithelial cell expression of Stat1 and
its target genes as well as IL-12 p40 and RANTES are all
altered in asthma, and this section summarizes these
findings. Because each of these epithelial immune-response genes are highly responsive to paramyxoviral infection (with or without concomitant production of IFN␥), their activation in asthma brings into question the
hypothesis that asthma develops due solely to decreased
Th1- and increased Th2-type T cell responses. This issue
is also discussed in this section.
A. Constitutive Abnormalities in Epithelial Gene
Expression (Stat1 And IL-12)
Cytokine effects on immunity and inflammation often
depend on STAT signaling pathways, so these are ideal
candidates for influencing inflammatory disease. We reasoned that selective interferon responsiveness of the first
STAT family member (Stat1) and Stat1-dependent immune-response genes such as ICAM-1, IRF-1, and Stat1
itself in airway epithelial cells provided a basis for detecting cytokine signaling abnormalities in inflammatory
airway disease. Based on nuclear localization and phosphorylation, epithelial Stat1 (but not other control transcription factors such as Stat3, AP-1, and NF-␬B) was
invariably activated in asthmatic compared with normal
control or chronic bronchitis subjects (114). Stat1 activation was relatively selective for epithelial cells, since activation was not detectable in neighboring macrophages.
Furthermore, epithelial levels of activated Stat1 correlated with levels of expression for epithelial ICAM-1,
IRF-1, and Stat1, and in turn, ICAM-1 level correlated with
T cell accumulation in tissue. However, only low levels of
IFN-␥ or IFN-␥-producing cells were detected in airway
tissue in all subjects. Since tissue levels of IFN-␥ are not
increased, asthma appears to be characterized by constitutive activity of Th1-like cytokine signaling even without
evidence of stimuli (such as IFN-␥ or viral infection) (69,
138) that normally drive this type of immune reaction. The
results therefore provide initial evidence linking abnormal behavior of STAT pathways for cytokine signaling (as
opposed to cytokine production) to the development of
an inflammatory disease. In that context, the results also
change the current scheme for asthma pathogenesis to
one that must include a localized gain in transcriptional
signal ordinarily used for a Th1 cytokine (IFN-␥) and so is
distinct from mechanisms that depend on allergy-driven
overproduction of Th2 cytokines.
Localizing the abnormality in Stat1 activity to airway
epithelial cells is critical to the impact of these findings.
Constitutive STAT activities in lymphocytes are associated with malignant transformation, whereas excessive
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
Stat1 activity in chondrocytes driven by a mutant fibroblast growth factor receptor may be associated with
growth arrest (128). Thus depending on the programmed
cytokine response of specific cell types in specific tissues,
abnormal Jak-STAT activity may lead to diseases as diverse as lymphoma or dwarfism. Consequently, analysis
of cytokine-dependent responses in asthma using sites
distant from the airways may not reflect behavior in airway cells. For example, others have indicated that decreased Th1 responses may predispose to asthma based
on an association with diminished delayed-type hypersensitivity reactions to tuberculin (117). However, this association was based on immune responses in the skin and so
does not reflect the specific activity of Th1-driven signaling pathways in airway epithelial cells.
To next determine whether the pattern of IL-12 upregulation in murine viral bronchitis also develops in
asthma, we measured levels of IL-12 expression in airway
tissue and bronchoalveolar lavage fluid obtained from
asthmatic subjects. In fact, subjects with asthma exhibited higher levels of IL-12 as well, with expression predominantly in airway epithelial cells and in the form of
IL-12 p40 (and p80 homodimer). Upregulation of IL-12 p40
expression also correlated with increased number of macrophages in the airway, so each of the features of IL-12
behavior was similar to ones found in the mouse during
paramyxoviral infection. This finding offers two new possibilities for a role of IL-12 p40 in asthma: 1) antagonism
of endogenous IL-12, and so skewing the local cytokine
environment toward a Th2 immune response, and/or 2)
function as an agonist, e.g., as a macrophage chemotactic
and activating factor, and so causing airway inflammation. In fact, some (but not all) previous studies find
significant macrophage accumulation in the submucosal
and intraepithelial airway tissue of asthmatic compared
with normal subjects (104, 109), whereas others provide
evidence of macrophage activation in asthma (18) as well
as an increase in number and activation state of airway
macrophages during allergen challenge in asthma (89). In
either case, these results provide initial evidence that
asthma, often characterized as a condition that depends
on overexpression of Th2 and underexpression of Th1
cytokines by immune cells, does in fact also exhibit overexpression of IL-12 p40 that appears to be chiefly derived
from airway epithelial cells. In conjunction with previous
observations of constitutive activation of Stat1 and Stat1dependent genes, the findings further support the possibility that pathways normally responsive to Th1 cytokines
are also dysregulated in airway inflammatory disease. As
discussed further below, upregulation of epithelial Stat1
and IL-12 pathways in asthma is found in both allergic and
nonallergic subjects and with or without treatment with
glucocorticoids.
Physiol Rev • VOL
35
B. Glucocorticoid-Sensitive Abnormalities
in Epithelial Gene Expression (RANTES)
Recognizing that RANTES and Stat1-dependent
genes exhibit distinct but complementary function (chemotaxis versus cell adhesion) and regulation (posttranscriptional versus transcriptional), we also determined
the level of epithelial RANTES expression in asthma.
Most previous reports indicated that RANTES is expressed at similar basal levels in the airway epithelium of
normal versus asthmatic subjects with mild disease (8, 33,
60, 144). Similarly, we did not expect much change in
epithelial RANTES levels in response to allergen, based
on our experience with segmental allergen challenge (M.
Castro and M. J. Holtzman, unpublished observations).
Accordingly, we sought an experimental protocol for endogenous exacerbation of asthma that might better reflect
natural flares of the disease and might better define the
mechanism of action of anti-inflammatory treatment.
In that context, we developed a protocol for controlled glucocorticoid withdrawal in stable asthmatic subjects that includes measurements of airway function and
endobronchial biopsy and lavage during treatment and
then after withdrawal of inhaled glucocorticoids (15). In
some subjects, glucocorticoid withdrawal results in
asthma exacerbation characterized by increases in airway
hyperreactivity and immune cell (predominantly T cell)
infiltration. Interestingly, these disease flares are consistently associated with increased expression of RANTES
(by in situ hybridization for RANTES mRNA) that appears
localized predominantly to airway epithelial cells (129; M.
Castro, S. Block, D. L. Hamilos, M. V. Jenkersen, Q. Li, T.
Horiuchi, K. Schechtman, and M. J. Holtzman, unpublished observations). Concomitant work has indicated
that there is no increase in IFN-␥ production or NF-␬B
activation in this setting, similar to the findings in stable
asthma. Thus epithelial RANTES production is inducible
during asthma exacerbation even in the apparent absence
of viral infection or IFN-␥ production. Furthermore, induction is sensitive to glucocorticoid treatment but appears to proceed via an NF-␬B-independent pathway.
These findings are analogous to those in paramyxoviral
infection, so it is possible that posttranscriptional events
underlying RANTES expression during paramyxoviral infection may also be relevant to expression in asthma. As
discussed in section IVC, we have proposed that the increased expression of each of these pathways, i.e., Stat1,
IL-12 p40, and RANTES, may reflect a response to virus that
no longer causes signs of infection, but in the case of
RANTES, expression is unmasked by glucocorticoid withdrawal. Defining the basis for induction of RANTES gene
expression and its functional implications will require
additional work on the molecular basis for RANTES expression in isolated cell and animal model systems as well
as correlative protocols in human subjects. For example,
82 • JANUARY 2002 •
www.prv.org
36
HOLTZMAN ET AL.
glucocorticoid effects on posttranscriptional regulation of
the RANTES genes still need to be determined. Nonetheless, the findings for Stat1, IL-12, and RANTES already
suggest that a propensity for epithelial gene expression
may be conditioned by viral exposure and so underlie the
relationship of childhood viral infection to ongoing
asthma even later in life.
C. A Revised Model for Th1/Th2 Contributions
to Asthma
Asthmatic inflammation has been attributed to an
abnormal sensitivity to inhaled allergens, and by extension, to a skewed T helper cell response with increased
Th2 and decreased Th1 components compared with normal. In some settings, Th1 cells (perhaps activated by
viral infection) may assist Th2 cells in initiating an allergic
response, but this scheme also relies on Th2-dependent
production of cytokines, notably IL-4, IL-5, and IL-13 to
drive asthmatic inflammation and consequent pathophysiology (as summarized in Fig. 1). However, this pathogenesis does not account for the discrepancy between allergic and asthmatic phenotypes and, as indicated in the
previous section, does not provide for the heightened
antiviral state of the airway epithelium in asthma. In
particular, increased activity of innate antiviral programming in the epithelium appears to be a fundamental feature of asthma. This alternative pathway is normally useful for airway barrier cells to defend against respiratory
viruses, but if overactive, could lead to inflammation, as
appears to be the case in asthma. Some epithelial programs appear insensitive to glucocorticoids and so are
present constitutively even in stable disease (i.e., Stat1,
IL-12 p40 networks), whereas others are sensitive to glucocorticoids and so inducible by glucocorticoid withdrawal and flares of the disease (i.e., RANTES). As perhaps expected for an antiviral system, this epithelial
network is generally oriented toward a Th1 response.
To better integrate these concepts and so better explain the pathogenesis of airway disease, we have revised
the model for the role of the airway immune response in
asthma. The revised model accounts for how an alternative Th1-oriented epithelial network may act in combination with an enhanced Th2 response, and the combination
of epithelial, viral, and allergic components led to its
designation as an Epi-Vir-All paradigm (Fig. 7). In this
scheme, increases in epithelial antiviral signals (e.g., Stat1
activation and IL-12 p40 expression) and allergen-driven
FIG. 7. Revised model for the role of the airway immune response in the development of airway inflammation and
remodeling. A: illustration of how increases in epithelial antiviral signals (e.g., Stat1 activation and IL-12 p40 expression)
and allergen-driven production of Th2 cytokines (e.g., IL-4, IL-5, and IL-13) are characteristic of subjects with asthma
(designated as an “A”) studied under stable conditions during treatment with inhaled glucocorticoids. B: illustration of
how further increases in epithelial signaling (driven by viral infection) or Th2 cytokine production (driven by further
allergen exposure) may develop in subjects with asthma studied during a flare without glucocorticoid treatment. The
abnormality in epithelial cell behavior may depend on persistence of virus at low copy number and proper genetic
background. The combination of epithelial, viral, and allergic components led to designation of this pathogenesis model
as an Epi-Vir-All paradigm.
Physiol Rev • VOL
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
production of Th2 cytokines (e.g., IL-4, IL-5, and IL-13) are
characteristic of subjects with asthma under stable conditions during treatment with inhaled glucocorticoids.
Further increases in epithelial signaling (driven by viral
infection) or Th2 cytokine production (driven by further
allergen exposure) would develop in subjects with
asthma during a flare of the disease. In addition, increased
levels of Stat1 may mediate a hypersensitive Th1-type
response in the airway. At least in vitro, high levels of
Stat1 are capable of priming the Stat1-dependent pathway
and so lead to exaggerated gene expression in response to
normal levels of stimulation (D. Sampath, Y. Zhang, and
M. J. Holtzman, unpublished observations). In addition,
flares of the disease are closely associated with induction
of RANTES gene expression. This system, which appears
to be regulated at the posttranscriptional level during viral
infection, may rely on a similar regulatory mechanism in
asthmatic flares. Each of these possibilities begs the question as to how the abnormality in epithelial cell behavior
originally develops. In section V, we summarize initial
evidence that viral infection and the genetic background
of the host may interact to permanently reprogram the
behavior of airway epithelial cells.
V. LONG-TERM AIRWAY HYPERREACTIVITY
AND REMODELING IN MICE AND HUMANS
A relationship between viral infection and the development of chronic inflammatory disease has been proposed for diverse clinical syndromes, but the basis for this
relationship is still uncertain. In the particular context of
asthma, paramyxoviral infections are the leading cause of
lower respiratory tract illness in infants and young children (19, 26), and children with clinically significant viral
bronchiolitis appear to be marked for the subsequent
development of a chronic wheezing illness that is independent of allergy (107, 119, 124). Presumably paramyxoviral infection triggers a switch to an abnormal host response, since paramyxoviruses (or other respiratory RNA
viruses) are not thought to persist in airway tissue as a
cause of chronic respiratory disease (1). With or without
viral persistence, however, the role of specific host factors in the development of chronic wheezing or life-long
asthma has not been determined. This section describes
our initial results from studies of mice and human subjects in defining the linkage between respiratory viral
infection and persistent asthma and in so doing extends
the concepts of the previous sections that were focused
predominantly on short-term immunity and inflammation.
A. Segregating Acute From Chronic Phenotypes
To better define viral and host factors in the development of the asthma phenotype, we took further advanPhysiol Rev • VOL
37
tage of the mouse model of paramyxoviral bronchiolitis.
As noted above, inhibition of the acute inflammatory response could be achieved by disruption of epithelial immune-response genes. On the basis of the work in vitro,
this epithelial gene network is directly inducible by viral
replication and is dominated by an array of interferonresponsive genes, but among candidates that might mediate immune cell traffic, ICAM-1 appears to be a predominant determinant for immune cell transmigration (69, 82,
129). In fact, as noted above, we found that ICAM-1 is
inducible primarily on host airway epithelial cells by viral
infection and is necessary for full development of acute
inflammation and concomitant postviral airway hyperreactivity that peaks at ⬃3 wk after infection in the mouse
model. Unexpectedly, however, as we followed wild-type
and ICAM-1-null mice in a homogeneous genetic background, we also found that primary viral infection caused
essentially permanent airway hyperreactivity and concomitant epithelial remodeling that were manifest despite
ICAM-1 deficiency and were not accompanied by lowlevel persistence of viral transcripts (136, 140; M. J.
Walter, J. D. Morton, N. Kajiwara, E. Agapov, T. Horiuchi,
M. Castro, and M. J. Holtzman, unpublished observations). The predominant features of this remodeling phenotype, i.e., goblet cell hyperplasia and mucin production,
are also inducible in human subjects with asthma. Taken
together, the results indicate that paramyxoviral infection
may cause both acute airway inflammation/hyperreactivity and chronic airway remodeling/hyperreactivity phenotypes that can be segregated by their dependence on
ICAM-1 and time and so may depend on distinct genetic
controls (as modeled in Fig. 8). In a more general context,
the findings establish the capacity of a single paramyxoviral infection to cause both acute and chronic manifestations of the phenotype for hypersecretory disease and
the relevance of specific host defense genes in moderating
the acute but not necessarily the chronic phenotype.
B. Genetic and Viral Determinants for Persistence
The striking possibility that viral infection at an early
age may lead to permanent alterations in airway epithelial
behavior and airway function requires significant further
investigation. Initial results indicate that both viral and
host factors may determine the outcome of infection.
Thus only specific strains of mice are susceptible to developing the chronic phenotype, analogous but distinct
from the diversity in resistance to acute infection (10;
J. D. Morton and M. J. Holtzman, unpublished observations). Mouse age also affects the severity of the phenotype, indicating that developmental factors also influence
the outcome. Improvements in the technology for genetic
mapping (using single-nucleotide polymorphisms as informative markers) and gene expression (using oligonucle-
82 • JANUARY 2002 •
www.prv.org
38
HOLTZMAN ET AL.
FIG. 8. Segregation of virus-inducible phenotypes. Model for genetic- and time-dependent events in the development
of the asthma phenotype induced by respiratory paramyxoviral infection. On the basis of studies of genetically defined
mice, it appears that viral replication peaks on postinfection days 3–5, epithelial immune-response gene expression
peaks on day 5, immune cell infiltration peaks on day 8, acute airway hyperreactivity (that depends on ICAM-1 gene
expression) peaks on approximately day 21, and chronic remodeling (e.g., goblet cell hyperplasia) and chronic
hyperreactivity are present thereafter.
otide microarray) will allow more precise definition of
epithelial programming that is responsible for asthma and
bronchitis phenotypes in the mouse model (39). Extension of these results to analysis of candidate genes in
human subjects using precise case-control studies may
then provide further insight into the genetic susceptibility
for virus-inducible disease in humans.
Although genetic susceptibility for the host response
has been the target of studies by our lab and others, it
appears likely that viral genetics are also critical to the
development of the chronic remodeling/hyperreactivity
phenotype. Thus previous studies of Sendai virus and
related paramyxoviruses indicated that viral particles are
normally eliminated from tissue after 10 –12 days (25).
However, these studies relied on relatively insensitive
methods that are responsive to replicating (nonmutant)
forms of virus, and recent studies of other riboviruses
indicate that they may persist in the tissue as mutant
quasi-species. In these cases, viral persistence may be
facilitated by high mutation rates at critical epitopes and
consequent escape from immune surveillance (27), but
whether persistence is necessary or sufficient for chronic
alterations in host cell behavior still needs to be determined. Initial analysis of Sendai viral titers using kinetic
PCR suggests that viral persistence in host tissues at low
copy number is not a common feature of paramyxoviral
infection (J. D. Morton, E. Agapov, D. Palamand, and M. J.
Physiol Rev • VOL
Holtzman, unpublished observations). Moreover, initial
results with the Sendai viral mouse model indicate that
viral persistence is not necessary or sufficient for the
chronic phenotype, since sensitive and resistant strains of
mice exhibit similar effectiveness in viral clearance. The
results suggest a hit-and-run hypothesis for the viral effect, i.e., transient infection causes permanent alteration
in host cell behavior. Even so, however, the viral gene
products responsible for triggering the change in host
phenotype still need to be defined. New systems for recovery of negative-strand RNA viruses entirely from
cDNA may allow better definition of these issues as well
as general understanding of replication and pathogenesis
for this order of viruses (20, 95).
VI. SMART STRATEGIES FOR CORRECTING
EPITHELIAL INFLAMMATION AND
REMODELING
The airway epithelium (by virtue of location) is directly accessible to inhaled agents, so modifying epithelial
function is a natural target for therapy of asthma and for
combating respiratory viral infection. In the case of
asthma, even current agents (such as glucocorticoids and
sodium cromoglycate) may owe part of their efficacy to
the attenuation of cell adhesion molecule and chemoat-
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
39
FIG. 9. Strategies for modifying epithelial gene expression to preserve immunity but inhibit inflammation and
remodeling. Anti-asthma strategies (top box) are aimed at downregulating the Stat1 pathway using mutant viral proteins
(e.g., adenoviral E1A) or mutant Stat1 that each interrupts Stat1-dependent gene transcription. Anti-viral strategies
(bottom box) include expression of mutant RBF or native RANTES to increase RANTES levels and so enhance innate
antiviral defense.
tractant function (as noted above for RANTES) and so
could target the epithelium to achieve therapeutic effects.
Similarly, delivery of blocking antibodies, soluble ligands,
or synthetic peptide analogs via the airway may indicate
the importance of local epithelial events if the reagents
cannot effectively reach subepithelial sites (34, 49, 75, 86,
101). Using the epithelial antiviral network as a model, we
have begun to develop more specific and perhaps smarter
strategies for modifying epithelial gene expression and so
influencing epithelial function. The approach derives directly from the molecular pathways defined in this review
and is summarized in Figure 9. The overall aim is to
improve airway disease, with special reference to asthma
and related phenotypes in other forms of bronchitis/bronchiolitis. In view of the proposed nature of epithelial
programming against respiratory viruses, another natural
therapeutic target is to improve antiviral host defense.
A. Reversing Viral Mimicry Using Mutant
E1A Oncoprotein
Immunity to viral pathogens depends on coordinated
control of host cell genes, so viruses (including adenoviruses) have developed strategies to redirect the normal
genetic program (108). As one of the most informative
examples, the adenoviral early region 1A (E1A) gene enPhysiol Rev • VOL
codes for an oncoprotein that activates host cell cycle and
so enables DNA synthesis needed for viral replication. To
accomplish this, E1A oncoprotein presents two distinct
binding sites for competing with cell cycle regulators: one
site uses conserved regions 1 and 2 (CR1 and CR2) to bind
a family of antioncoproteins typified by retinoblastoma
(Rb) protein; the other site uses NH2-terminal residues
and CR1 to bind a family of transcriptional coactivators
typified by p300 (92, 148) (Fig. 10). The Rb protein and
Rb-related p107 and p130 proteins normally bind to the
E2F transcription factor and silence genes needed for
progression through the cell cycle (147). The p300 protein
(17) and the related cAMP response element binding protein (CREB)-binding protein (CBP) (31) serve as adaptor
proteins that link transcription factors such as CREB (3,
72) as well as p53 antioncoprotein (48, 74) to the basal
transcription complex. In addition, p300/CBP proteins influence gene transcription by intrinsic histone acetyltransferase activity and by linking additional coactivators
(such as p300/CBP/cointegrator-associated protein or
p/CIP) to core histones (134). Thus, by sequestering (and
perhaps linking) Rb- and p300/CBP-type proteins in competition with endogenous transcription factors and coactivators, three E1A domains (NH2 terminus, CR1, and
CR2) are sufficient for mediating cellular proliferation
(143).
82 • JANUARY 2002 •
www.prv.org
40
HOLTZMAN ET AL.
FIG. 10. Reverse viral mimicry using adenoviral E1A oncoprotein to downregulate Stat1 function. A: diagram for E1A
(13S and 12S splice variants) that includes the NH2-terminal (N-term), conserved region 1 (CR1), conserved region 2
(CR2), conserved region 3 (CR3), and COOH-terminal (C-term) domains. B: proposed structure and the critical sites
(including Arg2) for binding and sequestering p300/CBP and pRb. C: illustration of how E1A wild-type (E1A-WT) and E1A
with Arg to Gly mutation (E1A-RG2) may act to disrupt p300/CBP and Stat1 interaction and so inhibit transcription of
the ICAM-1 gene. DNA/protein and protein/protein interactions that control transcription rate include the following: 1)
Stat1/DNA interaction mediated by the GAS site at ⫺106 to 106; 2) Sp1/DNA interaction mediated by the GC box at ⫺99
to ⫺94; 3) Stat1/Sp1 interaction; 4) Sp1 interaction with the Pol II-specific basal transcription complex that binds at the
TATA box and initiates transcription at the transcription initiation site (TIS); 5) Stat1 interaction with p300; and 6) p300
interaction with the basal transcription complex via p/CIP. [Modified from Holtzman et al. (53) and Moran (92).]
In addition to altering control of the cell cycle, adenoviruses also aim to subvert host defense by evading
immune detection. Early studies in cell lines suggested
that E1A might inhibit activation of immune-response
genes by disrupting events (such as protein phosphorylation) that lead to protein-DNA interactions at interferonresponsive enhancers (64). However, subsequent studies
demonstrated that E1A capacity for immune suppression
might also depend on targeting p300/CBP. Thus p300/CBP
coactivator function was extended to c-fos and c-Jun
components of activator protein-1 (3, 5), the p65 component of NF-␬B (40), steroid and nuclear-hormone receptors (70, 150), and to the first two members of the STAT
family (9, 58, 153) that mediate interferon-driven gene
activation. In all cases, including Stat1 and Stat2, E1A
influenced host gene expression by competing with these
endogenous factors for binding to p300/CBP (5, 9, 40, 70,
150, 153). Because Stat1 is critical for IFN-␥-driven gene
transcription, competition for p300/CBP by E1A may underlie its capacity to subvert IFN-␥-stimulated immunity.
Physiol Rev • VOL
In reviewing this previous work, we questioned
whether E1A might also act directly on a specific DNAbinding transcription factor and thereby provide determinants for more precisely influencing gene expression. Precedent for this possibility may be found in interactions of
the transactivating domain of E1A (conserved region 3)
with transcription factors needed for viral gene expression (77), but at least to date, E1A NH2-terminal interactions appear more highly restricted (92). Nonetheless,
additional interactions might not be detected if E1A determinants for binding p300/CBP overlapped with those
for binding more specific cellular activators. In addition,
previous studies of adenoviral infection and host gene
behavior often used transformed cell lines rather than the
natural adenoviral host cell and so could not exclude
confounding effects of other oncogenes (such as SV40
large T antigen) on E1A action (32).
In that context, we have used primary culture hTBE
cells (along with endobronchial tissue and mouse models
of viral bronchitis) to define epithelial cell-dependent im-
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
munity to respiratory pathogens (129, 138). Our work on
isolated cells indicates that a subset of epithelial immuneresponse genes typified by ICAM-1 are controlled by an
IFN-␥-responsive Jak-STAT signaling pathway that relies
on Stat1 to first be phosphorylated at the IFN-␥ receptor
complex and then to bind to the ICAM-1 promoter region
(80, 81, 141). At the promoter, Stat1 interacts with adjacent transcription factors such as Sp1 and with p300/CBP
(58, 81, 153) to facilitate enhanceosome formation and
induce gene transcription. As noted above, E1A-p300/CBP
interaction may therefore underlie its capacity to subvert
IFN-␥-driven gene activation, or alternatively, may explain the antiviral effect of IFN-␥ (153).
Recently, however, we demonstrated that E1A can
also modify Stat1-dependent gene activation by direct
action on Stat1 itself (82). Using mutant forms of E1A, we
defined an NH2-terminal determinant (Arg2) for binding
p300/CBP that is not needed for binding Stat1. Furthermore, we showed that mutant E1A with no capacity for
binding p300/CBP or inhibiting p300/CBP-dependent
events (73, 143, 148) still binds Stat1 and inhibits IFN-␥induced, Stat1-dependent transactivation of exogenous
and endogenous target genes early in the course of adenoviral infection. This effect is distinct from downregulation of Stat1 phosphorylation that occurs later in the
course of infection. Taken together, these results provide
for an overall scheme for adenoviral disruption of host
gene expression during early and late-phase infection, a
novel action of the E1A NH2 terminus in mediating immune suppression, and a revised model of Stat1-dependent gene expression. The results also provide an appropriate E1A-mediated strategy for more selectively
downregulating Stat1-dependent events in the airway.
B. Modifying Epithelial Signaling
With Mutant Stat1
Transcription factors generally contain at least two
independent domains for DNA binding versus activation
of transcription (44). Removal of the transactivation domain has been shown in many cases to result in an
inactive factor that can bind to the DNA element and
displace wild-type protein, thereby creating a dominantnegative action (51). Targeting such a dominant-negative
construct so that it is expressed in a specific tissue has
been useful in understanding the function of specific transcription factors in different tissues. Accordingly, this
strategy offers an advantage over complete deletion of the
factor by homologous recombination with the endogenous gene if there is a goal of defining function in a
specific tissue or cell type. The creation of a dominantnegative mutation for Stat1 is more challenging than for
transcription factors with less complex function, but also
offers an opportunity for dissecting the relative importance of Stat1 modular function.
Physiol Rev • VOL
41
Accordingly, we aimed to selectively downregulate
the pathway using a dominant-negative strategy for inhibition of epithelial Stat1 in the primary culture airway
epithelial cell model (141). In initial experiments using a
Stat1-deficient cell line, we demonstrated that transfection of wild-type Stat1 expression plasmid restored appropriate Stat1 expression and IFN-␥-dependent phosphorylation as well as consequent IFN-␥ activation of
cotransfected ICAM-1 promoter constructs and endogenous ICAM-1 gene expression. However, mutations of
Stat1 at Tyr-701 (the Jak kinase phosphorylation site),
Glu-428/429 (the putative DNA-binding site), His-713 (the
splice-site resulting in Stat1␤ formation), or Ser-727 (the
mitogen-activated protein kinase phosphorylation site) all
decreased Stat1 capacity to activate the ICAM-1 promoter. The Tyr-701 mutant (followed by the His-713 mutant) was most effective in disabling Stat1 function and in
overcoming the activating effect of cotransfected wildtype Stat1 in this cell system, thereby highlighting the
effectiveness of blocking Stat1 homo- and heterodimerization. In experiments using primary culture airway epithelial cells and each of the four Stat1-mutant plasmids,
transfection with the Tyr-701 and His-713 mutants again
most effectively inhibited IFN-␥ activation of an ICAM-1
gene promoter construct. By then transfecting airway
epithelial cells with wild-type or mutant Stat1 tagged with
a Flag-reporter sequence, we used dual immunofluorescence to show that cells expressing the Tyr-701 or His-713
mutants but not the two other Stat1 mutants or wild-type
Stat1 were prevented from expressing endogenous
ICAM-1 in response to IFN-␥ treatment. The results provide the initial indication that loss of function may correlate with dominant-negative activity for Stat1 in a biologically relevant human cell model. The capacity of specific
Stat1 mutations to exert a potent dominant-negative effect on IFN-␥ signal transduction provides for further
definition of Stat1 structure/function. The strategy also
provides a means for natural or engineered expression of
mutant or truncated Stat1 to selectively downregulate
activity of this pathway in a cell type- or tissue-specific
manner during immune and/or inflammatory responses in
vivo.
C. Future Considerations
The studies summarized above indicate that molecular strategies can be devised that selectively target epithelial signaling pathways. Thus downregulation of epithelial
Stat1 signaling (to correct the abnormality in asthma) and
possibly upregulation of RANTES signaling (to aid in
antiviral defense) may be devised from the regulatory
features of these pathways, and these strategies appear
effective in vitro. Studies in animal models are poised to
determine whether modifying epithelial immune function
82 • JANUARY 2002 •
www.prv.org
42
HOLTZMAN ET AL.
may be effective as a sole treatment for airway inflammation or may render other anti-inflammatory drugs more
effective in vivo. However, the efficacy of this approach
will depend on developing methods to achieve selective
and efficient expression in the airway epithelium. Thus
proof of concept may be feasible using transgenic animals
(limited only by specificity of available gene promoter
systems for epithelial expression), but vectors with reproducible, sustainable, and high-level expression in the airway epithelium still need to be developed.
VII. SUMMARY
This review summarizes increasing evidence that airway epithelial barrier cells actively mediate airway immunity and inflammation. In fact, the role of the epithelium
has evolved from a relatively simple scheme for mediating
leukocyte recruitment to one that depends on the coordinated expression of a network of epithelial immuneresponse genes coordinated for host defense under distinct transcriptional and posttranscriptional controls. The
transcriptional program is typified by an interferon-driven
Jak/Stat signaling pathway, while posttranscriptional regulation uses RNA-protein interactions that are responsive
directly to viral replication. Respiratory viruses ordinarily
interact with this Th1-style gene network in a battle of
host defense versus immune subversion, but the same
network is also activated in asthma even in the absence of
overt viral infection. Thus the barrier epithelial cell population appears specially programmed for normal host
defense but abnormally programmed in inflammatory airway disease. Recent results indicate that airway epithelial
cells may be reprogrammed for permanent proliferation
and skewed mucus cell differentiation by asthmagenic
RNA viruses that persist at low copy number. This alteration in the epithelial repair process exhibits genetic susceptibility and so appears analogous to abnormal mucosal
phenotypes in asthma and other hypersecretory diseases.
These concepts can be integrated into a scheme that
incorporates epithelial, viral, and allergic components
(designated an Epi-Vir-All paradigm) for a more complete
explanation of the pathogenesis of airway disease. By
extension, the same epithelial network is a target for
therapy in airway disease and for improving host defense
against respiratory viruses. Smart strategies have already
been defined for reversing viral mimicry and engineering
dominant-negative mutations that alter epithelial behavior, but full utilization of these approaches will depend on
also achieving selective and high-level gene expression in
the epithelium. This goal depends in turn on determining
the critical biology of the airway epithelium in antiviral
defense and airway disease and in particular on defining
the contributions of specific viral and host genetic components.
Physiol Rev • VOL
We acknowledge the contributions of our colleagues Eugene Agapov, Yael Alevy, Jan Bragdon, Steve Brody, David
Briner, Naohiro Kajiwara, Edy Kim, Takeharu Koga, Dwight
Look, Shin Nakajima, Divya Palamand, Mark Pelletier, William
Roswit, Deepak Sampath, Massaguchi Taguchi, Kazutaka
Takami, and Yong Zhang.
The research underlying this review was supported by
grants from the National Heart, Lung, and Blood Institute; Martin Schaefer Fund; and Alan A. and Edith L. Wolff Charitable
Trust.
Address for reprint requests and other correspondence:
M. J. Holtzman, Washington University School of Medicine,
Campus Box 8052, 660 South Euclid Ave., St. Louis, MO 63110
(E-mail: holtzmanm@msnotes.wustl.edu).
REFERENCES
1. AHMED R, MORRISON LA, AND KNIPE DM. Persistence of viruses. In:
Fields Virology, edited by Fields BN, Knipe DM, and Howley PM.
Philadelphia, PA: Lippincott-Raven, 1996, p. 219 –249.
2. ALAM R, STAFFORD S, FORSYTHE P, HARRISON R, FAUBION D, LETTBROWN MA, AND GRANT JA. RANTES is a chemotactic and activating
factor for human eosinophils. J Immunol 150: 3442–3447, 1993.
3. ARIAS J. Activation of cAMP and mitogen responsive gene relies on
a common nuclear factor. Nature 370: 226 –229, 1997.
4. BACON KB, PREMACK BA, GARDNER P, AND SCHALL TJ. Activation of
dual T cell signaling pathways by the chemokine RANTES. Science
269: 1727–1730, 1995.
5. BANNISTER AJ AND KOUZARIDES T. CBP-induced stimulation of c-Fos
activity is abrogated by E1A. EMBO J 14: 4758 – 4762, 1995.
6. BEN-BARUCH A, MICHIEL DF, AND OPPENHEIM JJ. Signals and receptors
involved in recruitment of inflammatory cells. J Biol Chem 270:
11703–11706, 1995.
7. BENTLEY AM, DURHAM SR, ROBINSON DS, MENZ G, STORZ C, CROMWELL
O, KAY AB, AND WARDLAW AJ. Expression of endothelial and leukocyte adhesion molecules intercellular adhesion molecule-1, E-selectin, and vascular cell adhesion molecule-1 in the bronchial mucosa in steady-state and allergen-induced asthma. J Allergy Clin
Immunol 92: 857– 868, 1993.
8. BERKMAN N, KRISHNAN VL, GILBEY T, NEWTON R, O’CONNOR B, BARNES
PJ, AND CHUNG KF. Expression of RANTES mRNA and protein in
airways of patients with asthma. Am J Respir Crit Care Med 154:
1804 –1811, 1996.
9. BHATTACHARYA S, ECKNER R, GROSSMAN S, OLDREAD E, ARANY Z,
D’ANDREA A, AND LIVINGSTON DM. Cooperation of Stat2 and p300/
CBP in signalling induced by interferon-␣. Nature 383: 344 –347,
1996.
10. BROWNSTEIN DG. Genetics of natural resistance to Sendai virus
infection in mice. Infect Immun 41: 308 –312, 1983.
11. BURNS AR, TAKEI F, AND DOERSCHUK CM. Quantitation of ICAM-1
expression in mouse lung during pneumonia. J Immunol 153:
3189 –3198, 1994.
12. BUTCHER EC AND PICKER LJ. Lymphocyte homing and homeostasis.
Science 272: 60 – 66, 1996.
13. CAPUT D, BEUTLER B, HARTOG K, THAYER R, BROWN-SHIMER S, AND
CERAMI A. Identification of a common nucleotide sequence in the 3⬘
untranslated region of mRNA molecules specifying inflammatory
mediators. Proc Natl Acad Sci USA 83: 1670 –1674, 1986.
15. CASTRO M, HOLTZMAN MJ, LANGBECKER C, SAMPATH D, AND HAMILOS
DL. Inflammatory changes in asthma following inhaled glucocorticoid withdrawal (Abstract). Am J Respir Crit Care Med 155: A287,
1997.
16. CASTRO M, WALTER MJ, CHAPLIN DD, AND HOLTZMAN MJ. Could
asthma worsen by stimulating the T helper type 1 (Th1) response?
Am J Respir Cell Mol Biol 22: 143–146, 2000.
17. CHRIVIA JC, KWOK RPS, LAMB N, HAGIWARA M, MONTMINY MR, AND
GOODMAN RH. Phosphorylated CREB binds specifically to the nuclear protein CBP. Nature 365: 855– 859, 1993.
18. CLUZEL M, DAMON M, CHANEZ P, BOUSQUET J, CRASTES DE PAULET A,
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
MICHEL FB, AND GODARD P. Enhanced alveolar cell luminol-dependent chemiluminescence in asthma. J Allergy Clin Immunol 80:
195–201, 1987.
COLLINS PL, CHANOCK RM, AND MCINTOSH K. Parainfluenza viruses.
In: Fields Virology, edited by Fields BN, Knipe DM, and Howley
PM. Philadelphia, PA: Lippincott-Raven, 1996, p. 1205–1241.
CONZELMANN K-K. Nonsegmented negative-strand RNA viruses: genetics and manipulation of viral genomes. Annu Rev Genet 32:
123–162, 1998.
COYLE AJ, ERARD F, BERTRAND C, WALTI S, PIRCHER H, AND LE GROS G.
Virus-specific CD8⫹ cells can switch to interleukin 5 production
and induce airway eosinophilia. J Exp Med 181: 1229 –1233, 1995.
COYLE AJ, TSUYUKI S, BERTRAND C, HUANG S, AGUET M, ALKAN SS, AND
ANDERSON GP. Mice lacking the IFN-␥ receptor have an impaired
ability to resolve a lung eosinophilic inflammatory response associated with a prolonged capacity of T cells to exhibit a Th2 cytokine
profile. J Immunol 156: 2680 –2685, 1996.
D’ANDREA A, RENGARAJU M, VALIANTE NM, CHEHIMI J, KUBIN M, ASTE
M, CHAN SH, KOBAYASHI M, YOUNG D, NICKBARG E, CHIZZONITE R, WOLF
SF, AND TRINCHIERI G. Production of natural killer cell stimulatory
factor (interleukin 12) by peripheral blood mononuclear cells. J
Exp Med 176: 1387–1398, 1992.
DARNELL JE JR. STATs and gene regulation. Science 277: 1630 –1635,
1997.
DOHERTY PC AND CHRISTENSEN JP. Accessing complexity: the dynamics of virus-specific T cell responses. Annu Rev Immunol 18:
561–592, 2000.
DOMACHOWSKE JB AND ROSENBERG HF. Respiratory syncytial virus
infection: immune response, immunopathogenesis, and treatment.
Clin Microbiol Rev 12: 298 –309, 1999.
DRAKE JW AND HOLLAND JJ. Mutation rates among RNA viruses. Proc
Natl Acad Sci USA 96: 13910 –13913, 1999.
DRAZEN JM, ARM JP, AND AUSTEN KF. Sorting out the cytokines of
asthma. J Exp Med 183: 1–5, 1996.
DURBIN JE, FERNANDEZ-SESMA A, LEE C-K, DHARMA RAO T, FREY AB,
MORAN TM, VUKMANOVIC S, GARCIA-SASTRE A, AND LEVY DE. Type I
IFN modulates innate and specific antiviral immunity. J Immunol
164: 4220 – 4228, 2000.
DURBIN JE, HACKENMILLER R, SIMON MC, AND LEVY DE. Targeted
disruption of the mouse Stat1 gene results in compromised innate
immunity to viral disease. Cell 84: 443– 450, 1996.
ECKNER R, EWEN ME, NEWSOME D, GERDES M, DECAPRIO JA, LAWRENCE JB, AND LIVINGSTON DM. Molecular cloning and functional
analysis of the adenovirus E1A-associated 300-kD protein (p300)
reveals a protein with properties of a transcriptional adaptor.
Genes Dev 8: 869 – 884, 1994.
ECKNER R, LUDLOW JW, LILL NL, OLDREAD E, ARANY Z, MODJTAHEDI N,
DECAPRIO JA, LIVINGSTON DM, AND MORGAN JA. Association of p300
and CBP with simian virus 40 large T antigen. Mol Cell Biol 16:
3454 –3464, 1996.
FAHY JV, FIGUEROA DJ, WONG HH, LIU JT, AND ABRAMS JS. Similar
RANTES levels in healthy and asthmatic airways by immunoassay
and in situ hybridization. Am J Respir Crit Care Med 155: 1095–
1100, 1997.
FECONDO JV, KENT SBH, AND BOYD AW. Inhibition of intercellular
adhesion molecule 1-dependent biological activities by a synthetic
peptide analog. Proc Natl Acad Sci USA 88: 2879 –2882, 1991.
FUJITA T, SAKAKIBARA J, SUDO Y, MIYAMOTO M, KIMURA Y, AND TANIGUCHI T. Evidence for a nuclear factor(s), IRF-1, mediating induction
and silencing properties to human IFN-␤ gene regulatory elements.
EMBO J 7: 3397–3405, 1988.
GAVETT SH, O’HEARN DJ, LI X, HUANG S-K, FINKELMAN FD, AND
WILLS-KARP M. Interleukin 12 inhibits antigen-induced airway hyperresponsiveness, inflammation, and Th2 cytokine expression in
mice. J Exp Med 182: 1527–1536, 1995.
GAZZINELLI RT, WYSOCKA M, HIENY S, SCHARTON-KERSTEN T, CHEEVER
A, KUHN R, MULLER W, TRINCHIERI G, AND SHER A. In the absence of
endogenous IL-10, mice acutely infected with Toxoplasma gondii
succumb to a lethal immune response dependent on CD4⫹ T cells
and accompanied by overproduction of IL-12, IFN-␥, and TNF-␣.
J Immunol 157: 798 – 805, 1996.
GEORAS SN, LIU MC, NEWMAN W, BEALL WD, STEALEY BA, AND BOCHNER BS. Altered adhesion molecule expression and endothelial cell
Physiol Rev • VOL
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
43
activation accompany the recruitment of human granulocytes to
the lung after segmental antigen challenge. Am J Respir Cell Mol
Biol 7: 261–269, 1992.
GERMER S, HOLLAND MJ, AND HIGUCHI R. High-throughput SNP allelefrequency determination in pooled DNA samples by kinetic PCR.
Genome Res 10: 258 –266, 2000.
GERRITSEN ME, WILLIAMS AJ, NEISH AS, MOORE S, SHI Y, AND COLLINS
T. CREB-binding protein/p300 are transcriptional coactivators of
p65. Proc Natl Acad Sci USA 94: 2927–2932, 1997.
GILAT D, HERSHOVIZ R, MEKORI YA, VLODAVSKY I, AND LIDER O. Regulation of adhesion of CD4⫹ T lymphocytes to intact or heparinasetreated subendothelial extracellular matrix by diffusible or anchored RANTES and MIP-1␤. J Immunol 153: 4899 – 4906, 1994.
GIRDLESTONE J, ISSAMAT M, GEWERT D, AND MILSTEIN C. Transcriptional regulation of HLA-A and -B: differential binding of members
of the Rel and IRF families of transcription factors. Proc Natl Acad
Sci USA 90: 11568 –11572, 1993.
GONZALO J-A, LLOYD CM, KREMER L, FINGER E, MARTINEZ-A C,
SIEGELMAN MH, CYBULSKY M, AND GUTIERREZ-RAMOS J-C. Eosinophil
recruitment to the lung in a murine model of allergic inflammation.
J Clin Invest 98: 2332–2345, 1996.
GOODRICH JA, CUTLER G, AND TJIAN R. Contacts in context: promoter
specificity and macromolecular interactions in transcription. Cell
84: 825– 830, 1996.
GRASSME H, KIRSCHNEK S, RIETHMUELLER J, RIEHLE A, VON KURTHY G,
LANG F, WELLER M, AND GULBINS E. CD95/CD95 ligand interactions
on epithelial cells in host defense to Pseudomonas aeruginosa.
Science 290: 527–530, 2000.
GREENLUND AC, MORALES MO, VIVIANO BL, YAN H, KROLEWSKI J, AND
SCHREIBER RD. Stat recruitment by tyrosine-phosphorylated cytokine receptors: an ordered reversible affinity-driven process. Immunity 2: 677– 687, 1995.
GREVE JM, DAVIS G, MEYER AM, FORTE CP, YOST SC, MARLOR CW,
KAMARCK ME, AND MCCLELLAND A. The major human rhinovirus
receptor is ICAM-1. Cell 56: 839 – 847, 1989.
GU W, SHI X-L, AND ROEDER RG. Synergistic activation of transcription by CBP and p53. Nature 387: 819 – 822, 1997.
GUNDEL RH, WEGNER CD, TORCELLINI CA, CLARKE CC, HAYNES N,
ROTHLEIN R, SMITH CW, AND LETTS LG. Endothelial leukocyte adhesion molecule-1 mediate antigen-induced acute airway inflammation and late-phase airway obstruction in monkeys. J Clin Invest
88: 1407–1411, 1991.
HEINZEL FP, HUJER AM, AHMED FN, AND RERKO RM. In vivo production and function of IL-12 p40 homodimers. J Immunol 158: 4381–
4388, 1997.
HERSKOWITZ I. Functional inactivation of genes by dominant negative mutations. Nature 329: 219 –222, 1987.
HISCOTT J, BEAUPARLANT P, CREPIEUX P, DELUCA C, KWON H, LIN R,
AND PETROPOULOS L. Cellular and viral protein interactions regulating I␬B␣ activity during human retrovirus infection. J Leukoc Biol
62: 82–92, 1997.
HOLTZMAN MJ, CASTRO M, LOOK DC, O’SULLIVAN M, AND WALTER MJ.
Regulation of epithelial-leukocyte interaction and epithelial immune-response genes. In: Asthma and Rhinitis, edited by Busse
WW and Holgate ST. Cambridge, MA: Blackwell Scientific, 2000, p.
784 – 800.
HOLTZMAN MJ, FABBRI LM, O’BYRNE PM, GOLD BD, AIZAWA H,
WALTERS EH, ALPERT SE, AND NADEL JA. Importance of airway
inflammation for hyperresponsiveness induced by ozone in dogs.
Am Rev Respir Dis 127: 686 – 690, 1983.
HOLTZMAN MJ, LOOK DC, IADEMARCO MF, DEAN DC, SAMPATH D, AND
CASTRO M. Asthma. In: Principles of Molecular Medicine, edited by
Jameson JL. Totawa, NJ: Humana, 1998, p. 319 –327.
HOLTZMAN MJ, LOOK DC, SAMPATH D, CASTRO M, KOGA T, AND WALTER
MJ. Control of epithelial immune-response genes and implications
for airway immunity and inflammation. Proc Assoc Am Phys 110:
1–11, 1998.
HOLTZMAN MJ, SAMPATH D, CASTRO M, LOOK DC, AND JAYARAMAN S.
The one-two of T helper cells: does interferon-␥ knockout the Th2
hypothesis for asthma? Am J Respir Cell Mol Biol 14: 316 –318,
1996.
HORVAI AE, XU L, KORZUS E, BRARD G, KALAFUS D, MULLEN T-M, ROSE
DW, ROSENFELD MG, AND GLASS CK. Nuclear integration of JAK/
82 • JANUARY 2002 •
www.prv.org
44
59.
60.
61.
62.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
HOLTZMAN ET AL.
STAT and Ras/AP-1 signaling by CBP and p300. Proc Natl Acad Sci
USA 94: 1074 –1079, 1997.
HUANG GT-J, ECKMANN L, SAVIDGE TC, AND KAGNOFF MF. Infection of
human intestinal epithelial cells with invasive bacteria upregulates
apical intercellular adhesion molecule-1 (ICAM-1) expression and
neutrophil adhesion. J Clin Invest 98: 572–583, 1996.
HUMBERT M, YING S, CORRIGAN C, MENZ G, BARKANS J, PFISTER R,
MENG Q, VAN DAMME J, OPDENAKKER G, DURHAM SR, AND KAY AB.
Bronchial mucosal expression of the genes encoding chemokines
RANTES and MCP-3 in symptomatic atopic and nonatopic asthmatics: relationship to the eosinophil-active cytokines interleukin
(IL)-5, granulocyte macrophage-colony-stimulating factor, and IL-3.
Am J Respir Cell Mol Biol 16: 1– 8, 1997.
JAYARAMAN S, CASTRO M, O’SULLIVAN M, BRAGDON MJ, AND HOLTZMAN
MJ. Resistance to Fas-mediated T cell apoptosis in asthma. J Immunol 162: 1717–1722, 1999.
JOHNSON TR, JOHNSON JE, ROBERTS SR, WERTZ GW, PARKER RA, AND
GRAHAM BS. Priming with secreted glycoprotein G of respiratory
syncytial virus (RSV) augments interleukin-5 production and tissue
eosinophilia after RSV challenge. J Virol 72: 2871–2880, 1998.
KALVAKOLANU DVR, BANDYOPADHYAY SK, HARTER ML, AND SEN GC.
Inhibition of interferon-inducible gene expression by adenovirus
E1A proteins: block in transcriptional complex formation. Proc
Natl Acad Sci USA 88: 7459 –7463, 1991.
KAMIJO R, HARADA H, MATSUYAMA T, BOSLAND M, GERECITANO J,
SHAPIRO D, LE J, KOH SI, KIMURA T, GREEN SJ, MAK TW, TANIGUCHI T,
AND VILCEK J. Requirement for transcription factor IRF-1 in NO
synthase induction in macrophages. Science 263: 1612–1615, 1994.
KANG B-H, MANDERSCHIED BD, HUANG Y-CT, CRAPO JD, AND CHANG
L-Y. Contrasting response of lung parenchymal cells to instilled
TNF␣ and IFN␥: the inducibility of specific cell ICAM-1 in vivo.
Am J Respir Cell Mol Biol 15: 540 –550, 1996.
KATO K, SHIMOZATO O, HOSHI K, WAKIMOTO H, HAMADA H, YAGITA H,
AND OKUMURA K. Local production of the p40 subunit of interleukin-12 suppresses T-helper 1-mediated immune responses and prevents allogeneic myoblast rejection. Proc Natl Acad Sci USA 93:
9085–9089, 1996.
KOGA T, LOOK DC, TIDWELL RM, AND HOLTZMAN MJ. Mechanisms for
cytokine regulation of ␤-chemokine expression in airway epithelial
cells (Abstract). Am J Respir Crit Care Med 157: A744, 1998.
KOGA T, SARDINA E, TIDWELL RM, PELLETIER MR, LOOK DC, AND
HOLTZMAN MJ. Virus-inducible expression of a host chemokine gene
relies on replication-linked mRNA stabilization. Proc Natl Acad Sci
USA 96: 5680 –5685, 1999.
KUROKAWA R, KALAFUS D, OGLIASTRO M-H, KIOUSSI C, XU L, TORCHIA J,
ROSENFELD MG, AND GLASS CK. Differential use of CREB binding
protein-coactivator complexes. Science 279: 700 –703, 1998.
KURT-JONES EA, POPOVA L, KWINN L, HAYNES LM, JONES LP, TRIPP RA,
WALSH EE, FREEMAN MW, GOLENBOCK DT, ANDERSON LJ, AND FINBERG
RW. Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus. Nat Immunol 1: 398 – 401,
2000.
KWOK RP, LUNDBLAD JR, CHRIVIA JC, RICHARDS JP, BACHINGER HP,
BRENNAN RG, ROBERTS SG, GREEN MR, AND GOODMAN RH. Nuclear
protein CBP is a coactivator for the transcription factor CREB.
Nature 370: 223–226, 1994.
LEE J-S, SEE RH, DENG T, AND SHI Y. Adenovirus E1A downregulates
cJun- and JunB-mediated transcription by targeting their coactivator p300. Mol Cell Biol 16: 4312– 4326, 1996.
LILL NL, GROSSMAN SR, GINSBERG D, DECAPRIO J, AND LIVINGSTON DM.
Binding and modulation of p53 by p300/CBP coactivators. Nature
387: 823– 827, 1997.
LINEBERGER DW, GRAHAM DJ, TOMASSINI JE, AND COLONNO RJ. Antibodies that block rhinovirus attachment map to domain 1 of the
major group receptor. J Virol 64: 2582–2587, 1990.
LING P, GATELY MK, GUBLER U, STERN AS, LIN P, HOLLFELDER K, SU C,
PAN Y-CE, AND HAKIMI J. Human IL-12 p40 homodimer binds to the
IL-12 receptor but does not mediate biologic activity. J Immunol
154: 116 –127, 1995.
LIU F AND GREEN MR. Promoter targeting by adenovirus E1a
through interaction with different cellular DNA-binding domains.
Nature 368: 520 –525, 1994.
LOKUTA MA, MAHER J, NOE KH, PITHA PM, SHIN ML, AND SHIN HS.
Physiol Rev • VOL
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
Mechanisms of murine RANTES chemokine gene induction by
Newcastle disease virus. J Biol Chem 271: 13731–13738, 1996.
LOOK DC, KELLER BT, RAPP SR, AND HOLTZMAN MJ. Selective induction of intercellular adhesion molecule-1 by interferon-␥ in human
airway epithelial cells. Am J Physiol Lung Cell Mol Physiol 263:
L79 –L87, 1992.
LOOK DC, PELLETIER MR, AND HOLTZMAN MJ. Selective interaction of
a subset of interferon-␥ response element binding proteins with the
intercellular adhesion molecule-1 (ICAM-1) gene promoter controls
the pattern of expression on epithelial cells. J Biol Chem 269:
8952– 8958, 1994.
LOOK DC, PELLETIER MR, TIDWELL RM, ROSWIT WT, AND HOLTZMAN MJ.
Stat1 depends on transcriptional synergy with Sp1. J Biol Chem
270: 30264 –30267, 1995.
LOOK DC, ROSWIT WT, FRICK AG, GRIS-ALEVY Y, DICKHAUS DM,
WALTER MJ, AND HOLTZMAN MJ. Direct suppression of Stat1 function
during adenoviral infection. Immunity 9: 871– 880, 1998.
LUKACS NW, STRIETER RM, CHENSUE SW, AND KUNKEL SL. Interleukin4-dependent pulmonary eosinophil infiltration in a murine model of
asthma. Am J Respir Cell Mol Biol 10: 526 –532, 1994.
MACATONIA SE, HOSKEN NA, LITTON M, VIEIRA P, HSIEH CS, CULPEPPER
JA, WYSOCKA M, TRINCHIERI G, MURPHY KM, AND O’GARRA A. Dendritic
cells produce IL-12 and direct the development of Th1 cells from
naive CD4⫹ T cells. J Immunol 154: 5071–5079, 1995.
MACKAY CR. Chemokines: immunology’s high impact factors. Nat
Immunol 2: 95–101, 2001.
MARLIN SD, STAUNTON DE, SPRINGER TA, STRATOWA C, SOMMERGRUBER
W, AND MERLUZZI VJ. A soluble form of intercellular adhesion molecule-1 inhibits rhinovirus infection. Nature 344: 70 –72, 1990.
MARTIN KJ AND GREEN MR. Transcriptional activation by viral immediate-early proteins: variations on a common theme. In: Transcriptional Regulation, edited by McKnight SL and Yamamoto KR.
Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1992, p.
695–725.
MERAZ MA, WHITE JM, SHEEHAN KCF, BACH EA, RODIG SJ, DIGHE AS,
KAPLAN DH, RILEY JK, GREENLUND AC, CAMPBELL D, CARVER-MOORE K,
DUBOIS RN, CLARK R, AGUET M, AND SCHREIBER RD. Targeted disruption of the Stat1 gene in mice reveals unexpected physiologic
specificity in the JAK-STAT signaling pathway. Cell 84: 431– 442,
1996.
METZGER WJ, ZAVALA D, RICHERSON HB, MOSELEY P, IWAMOTA P,
MONICK M, SJOERDSMA K, AND HUNNINGHAKE GW. Local allergen challenge and bronchoalveolar lavage of allergic asthmatic lungs. Am
Rev Respir Dis 135: 433– 440, 1987.
MIDDLETON J, NEIL S, WINTLE J, CLARK-LEWIS I, MOORE H, LAM C, AUER
M, HUB E, AND ROT A. Transcytosis and surface presentation of IL-8
by venular endothelial cells. Cell 91: 385–395, 1997.
MITCHELL TJ, NAUGHTON M, NORSWORTHY P, DAVIES KA, WALPORT MJ,
AND MORLEY BJ. IFN-␥ up-regulates expression of the complement
components C3 and C4 by stabilization of mRNA. J Immunol 156:
4429 – 4434, 1996.
MORAN E. Mammalian cell growth controls reflected through protein interactions with the adenovirus E1A gene products. Semin
Virol 5: 327–340, 1994.
MORIUCHI H, MORIUCHI M, AND FAUCI AS. Nuclear factor-kB potently
up-regulates the promoter activity of RANTES, a chemokine that
blocks HIV infection. J Immunol 158: 3483–3491, 1997.
MURPHY KM, OUYANG W, FARRAR JD, YANG J, RANGANATH S, ASNAGLI H,
AFKARIAN M, AND MURPHY TL. Signaling and transcription in T helper
development. Annu Rev Immunol 18: 451– 494, 2000.
NAGAI Y. Paramyxovirus replication and pathogenesis. Reverse genetics transforms understanding. Rev Med Virol 9: 83–99, 1999.
NAGATA S. Apoptosis by death factor. Cell 88: 355–365, 1997.
NAKAJIMA H, SANO H, NISHIMURA T, YOSHIDA S, AND IWAMOTO I. Role of
vascular cell adhesion molecule 1/very late activation antigen 4 and
intercellular adhesion molecule 1/lymphocyte function-associated
antigen 1 interactions in antigen-induced eosinophil and T cell
recruitment into tissue. J Exp Med 179: 1145–1154, 1994.
NAKAJIMA S, LOOK DC, ROSWIT WT, BRAGDON MJ, AND HOLTZMAN MJ.
Selective differences in vascular endothelial- vs. airway epithelial-T
cell adhesion mechanisms. Am J Physiol Lung Cell Mol Physiol
267: L422–L432, 1994.
NAKAJIMA S, ROSWIT WT, LOOK DC, AND HOLTZMAN MJ. A hierarchy for
82 • JANUARY 2002 •
www.prv.org
EPITHELIAL-VIRAL-ALLERGIC PARADIGM
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
integrin expression and adhesiveness among T cell subsets that is
linked to TCR gene usage and emphasizes V␦1⫹ ␥␦ T cell adherence
and tissue retention. J Immunol 155: 1117–1131, 1995.
NELSON PJ, ORTIZ BD, PATTISON JM, AND KRENSKY AM. Identification
of a novel regulatory region critical for expression of the RANTES
chemokine in activated T lymphocytes. J Immunol 157: 1139 –1148,
1996.
NOONAN TC, GUNDEL RH, DESAI SN, STEARNS C, BARTPM RW, ROTHLEIN
R, LETTS LG, AND PIPER PJ. The effects of an anti-CD18 antibody
(R15.7) in antigen-induced airway hyperresponsiveness (AH) and
cell influx in guinea pigs. Agents Actions 34: 211–213, 1991.
O’BRIEN AD, STANDIFORD TJ, BUCKNELL KA, WILCOXEN SE, AND PAINE
RI. Role of alveolar epithelial cell intercellular adhesion molecule-1
in host defense against Klebsiella pneumoniae. Am J Physiol Lung
Cell Mol Physiol 276: L961–L970, 1999.
OHH M AND TAKEI F. IFN-␥ and phorbol myristate acetate-responsive
elements involved in the intercellular adhesion molecule-1 stabilization. J Biol Chem 269: 30117–30120, 1994.
OLLERENSHAW S AND WOOLCOCK AJ. Characteristics of the inflammation in biopsies from large airways of subjects with asthma and
subjects with chronic airflow limitaiton. Am Rev Respir Dis 145:
922–927, 1992.
O’SULLIVAN MP, KULCZYSKI A, MARSHALL DL, BRAGDON MJ, JAYARAMAN
S, AND HOLTZMAN MJ. Distinct Fas- versus cytokine-dependent pathways for death in airway epithelial cells (Abstract). Am J Respir
Crit Care Med 157: A28, 1998.
PARKOS CA, COLGAN SP, DIAMOND MS, NUSRAT A, LIANG TW, SPRINGER
TA, AND MADARA JL. Expression and polarization of intercelluar
adhesion molecule-1 on human intestinal epithelia: consequences
for CD11b/CD18-mediated interactions with neutrophils. Mol Med
2: 489 –505, 1996.
PATTEMORE PK, JOHNSTON SL, AND BARDIN PG. Viruses as precipitants
of asthma symptoms. I. Epidemiology. Clin Exp Allergy 22: 325–
336, 1992.
PLOEGH HL. Viral strategies of immune evasion. Science 280: 248 –
253, 1998.
POSTON RN, CHANEZ P, LACOSTE JY, LITCHFIELD T, LEE TH, AND
BOUSQUET J. Immunohistochemical characterization of the cellular
infiltration in asthmatic bronchi. Am Rev Respir Dis 145: 918 –921,
1992.
RANDOLPH DA, STEPHENS R, CARRUTHERS CJL, AND CHAPLIN DD. Synergy between Th1 and Th2 cells promotes eosinophilic airway
inflammation in a mouse model of asthma. J Clin Invest. In press.
RAY P, YANG L, ZHANG D-H, GHOSH SK, AND RAY A. Selective upregulation of cytokine-induced RANTES gene expression in lung
epithelial cells by overexpression of I␬BR. J Biol Chem 272: 20191–
20197, 1997.
REIS E SOUSA C, YAP G, SCHULZ O, ROGERS N, SCHITO M, ALIBERTI J,
HIENY S, AND SHER A. Paralysis of dendritic cell IL-12 production by
microbial products prevents infection-induced immunopathology.
Immunity 11: 637– 647, 1999.
ROT A. Endothelial cell binding of NAP-1/IL-8: role of neutrophil
emigration. Immunol Today 13: 291–294, 1992.
SAMPATH D, CASTRO M, LOOK DC, AND HOLTZMAN MJ. Constitutive
activation of an epithelial signal transducer and activator of transcription (Stat1) pathway in asthma. J Clin Invest 103: 1353–1361,
1999.
SAMPATH D, ROSWIT WT, KOGA T, WALTER MJ, LOOK DC, CASTRO M,
AND HOLTZMAN MJ. Autoamplification of Stat1-dependent gene expression (Abstract). FASEB J 12: A1390, 1998.
SCHWARZE J, HAMELMANN E, BRADLEY KL, TAKEDA K, AND GELFAND EW.
Respiratory syncytial virus infection results in airway hyperresponsiveness and enhanced airway sensitization to allergen. J Clin
Invest 100: 226 –233, 1997.
SHIRAKAWA T, ENOMOTO T, SHIMAZU S, AND HOPKIN JM. The inverse
association between tuberculin responses and atopic disorders.
Science 275: 77–79, 1997.
SHORNICK L, BRINER D, LO M, AND HOLTZMAN MJ. Role of Stat1 in
defense against respiratory viruses. In: Proceedings of Keystone
Symposium: Molecular Aspects of Viral Immunity. 2001, p. 56.
SIGURS N, BJARNASON R, SIGURBERGSSON F, KJELLMAN B, AND BJORKSTEN B. Asthma and immunoglobulin E antibodies after respiratory
Physiol Rev • VOL
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
45
syncytial virus bronchiolitis: a prospective cohort study with
matched controls. Pediatrics 95: 500 –505, 1995.
SPRINGER TA. Traffic signals on endothelium for lymphocyte recirculation and leukocyte emigration. Annu Rev Physiol 57: 827– 872,
1995.
SRIKIATKHACHORN A AND BRACIALE TJ. Virus-specific CD8⫹ T lymphocytes downregulate T helper cell type 2 cytokine secretion and
pulmonary eosinophila during experimental murine respiratory
syncytial virus infection. J Exp Med 186: 421– 432, 1997.
STARK JM AND GRAZIANO FM. Lower airway responses to viruses. In:
Asthma and Rhinitis, edited by Busse WW and Holgate ST. Boston, MA: Blackwell Scientific, 1995, p. 1229 –1243.
STAUNTON DE, MERLUZZI VJ, ROTHLEIN R, BARTON R, MARLIN SD, AND
SPRINGER TA. A cell adhesion molecule, ICAM-1, is the major surface receptor for rhinoviruses. Cell 56: 849 – 853, 1989.
STEIN RT, SHERRILL D, MORGAN WJ, HOLBERG CJ, HALONEN M, TAUSSIG
LM, WRIGHT AL, AND MARTINEZ FD. Respiratory syncytial virus in
early life and risk of wheeze and allergy by age 13 years. Lancet
345: 541–545, 1999.
STERN AS, PODLASKI FJ, HULMES JD, PAN YC, QUINN PM, WOLITZKY AG,
FAMILLETTI PC, STREMLO DL, TRUITT T, CHIZZONITE R, AND GATELY MK.
Purification to homogeneity and partial characterization of cytotoxic lymphocyte maturation factor from human B-lymphoblastoid
cells. Proc Natl Acad Sci USA 87: 6808 – 6812, 1990.
STOECKLE MY. Removal of a 3⬘ non-coding sequence is an initial step
in degradation of groa mRNA and is regulated by interleukin-1.
Nucleic Acids Res 20: 1123–1127, 1992.
STOECKLE MY AND HANAFUSA H. Processing of 9E3 mRNA and regulation of its stability in normal and Rous sarcoma virus-transformed cells. Mol Cell Biol 9: 4738 – 4745, 1989.
SU W-C, KITAGAWA M, XUE N, XIE B, GAROFALO S, CHO J, DENG C,
HORTON WA, AND FU X-Y. Activation of Stat1 by mutant fibroblast
growth-factor receptor in thanatophoric dysplasia type II dwarfism.
Nature 386: 288 –292, 1997.
TAGUCHI M, SAMPATH D, KOGA T, CASTRO M, LOOK DC, NAKAJIMA S, AND
HOLTZMAN MJ. Patterns for RANTES secretion and intercellular
adhesion molecule-1 expression mediate transepithelial T cell traffic based on analyses in vitro and in vivo. J Exp Med 187: 1927–
1940, 1998.
TANAKA N AND TANIGUCHI T. Cytokine gene regulation: regulatory
cis-elements and DNA binding factors involved in the interferon
system. Adv Immunol 52: 263–281, 1992.
TANAKA Y, ADAMS DH, HUBSCHER S, HIRANO H, SIEBENLIST U, AND SHAW
S. T-cell adhesion induced by proteoglycan-immobilized cytokine
MIP-1␤. Nature 361: 79 – 82, 1993.
THANOS D AND MANIATIS T. Virus induction of human IFN␤ gene
expression requires the assembly of an enhanceosome. Cell 83:
1091–1100, 1995.
THOMAS LH, FRIEDLAND JS, SHARLAND M, AND BECKER S. Respiratory
syncytial virus-induced RANTES production from human bronchial
epithelial cells is dependent on nuclear factor-␬B nuclear binding
and is inhibited by adenovirus-mediated expression of inhibitor of
␬B␣. J Immunol 161: 1007–1016, 1998.
TORCHIA J, ROSE DW, INOSTROZA J, KAMEI Y, WESTIN S, GLASS CK, AND
ROSENFELD MG. The transcriptional co-activator p/CIP binds CBP
and mediates nuclear-receptor function. Nature 387: 677– 684,
1997.
TRINCHIERI G AND SCOTT P. Interleukin-12: basic principles and clinical applications. Curr Top Micro Immunol 238: 57–78, 1999.
WALTER MJ, KAJIWARA N, AND HOLTZMAN MJ. Diminished airway
inflammation and hyperreactivity in ICAM-1-deficient mice during
primary viral bronchitis (Abstract). Am J Respir Crit Care Med
161: A778, 2000.
WALTER MJ, KAJIWARA N, KARANJA P, CASTRO M, AND HOLTZMAN MJ.
IL-12 p40 production by barrier epithelial cells during airway inflammation. J Exp Med 193: 339 –352, 2001.
WALTER MJ, KAJIWARA N, SAMPATH D, RUCKER J, AND HOLTZMAN MJ.
Induction of epithelial immune-response genes in a mouse model
of viral bronchitis and hyperreactivity (Abstract). FASEB J 12:
1453, 1998.
WALTER MJ, KAJIWARA N, SAMPATH D, RUCKER J, XIA D, AND HOLTZMAN
MJ. Epithelial immune-response gene expression during viral bron-
82 • JANUARY 2002 •
www.prv.org
46
140.
141.
143.
144.
145.
146.
HOLTZMAN ET AL.
chitis and hyperreactivity in wild-type and IFN-␥-deficient mice
(Abstract). Am J Respir Crit Care Med 159: A437, 1999.
WALTER MJ, KAJIWARA N, XIA D, AND HOLTZMAN MJ. Early-phase
innate immunity and late-phase remodeling of the epithelium during primary viral bronchitis and hyprreactivity (Abstract). J Invest
Med 47: 256A, 1999.
WALTER MJ, LOOK DC, TIDWELL RM, ROSWIT WT, AND HOLTZMAN MJ.
Targeted inhibition of interferon-␥-dependent ICAM-1 expression
using dominant-negative Stat1. J Biol Chem 272: 28582–28589,
1997.
WANG HH, RIKITAKE Y, CARTER MC, YACIUK P, ABRAHAM SE, ZERLER B,
AND MORAN E. Identification of specific adenovirus E1A N-terminal
residues critical to the binding of cellular proteins and to the
control of cell growth. J Virol 67: 476 – 488, 1993.
WANG JH, DEVALIA JL, XIA C, SAPSFORD RJ, AND DAVIES RJ. Expression of RANTES by human bronchial epithelial cells in vitro and in
vivo and the effect of corticosteroids. Am J Respir Cell Mol Biol 14:
27–35, 1996.
WARDLAW AJ, BENTLEY AM, MENZ G, STORZ C, DURHAM SR, AND KAY
AB. Expression of the adhesion molecules ICAM-1 and ELAM-1 in
the bronchial mucosa in asthma. J Allergy Clin Immunol 89: 164,
1992.
WEGNER CD, GUNDEL RH, REILLY P, HAYNES N, LETTS LG, AND ROTHLEIN R. Intercellular adhesion molecule-1 (ICAM-1) in the pathogenesis of asthma. Science 247: 456 – 459, 1990.
Physiol Rev • VOL
147. WEINTRAUB SJ, CHOW KNB, LUO RX, ZHANG SH, HE S, AND DEAN DC.
Mechanism of active transcriptional repression by the retinoblastoma protein. Nature 375: 812– 815, 1995.
148. WHYTE P, WILLIAMSON NM, AND HARLOW E. Cellular targets for transformation by the adenovirus E1A proteins. Cell 56: 67–75, 1989.
149. YAMAYA M, FINKBEINER WE, CHUN SY, AND WIDDICOMBE JH. Differentiated structure and function of cultures of human tracheal epithelium. Am J Physiol Lung Cell Mol Physiol 262: L713–L724, 1992.
150. YANG X-J, OGRYZKO VV, NISHIKAWA J, HOWARD BH, AND NAKATANI Y. A
p300/CBP-associated factor that competes with the adenoviral oncoprotein E1A. Nature 382: 319 –324, 1996.
151. YING S, DURHAM SR, CORRIGAN CJ, HAMID Q, AND KAY AB. Phenotype
of cells expressing mRNA for TH2-type (interleukin 4 and interleukin 5) and TH1-type (interleukin 2 and interferon ␥) cytokines in
bronchoalveolar lavage and bronchial biopsies from atopic asthmatic and normal control subjects. Am J Respir Cell Mol Biol 12:
477– 487, 1995.
152. YOSHIMOTO T, WANG C-R, YONETO T, WAKI S, SUNAGA S, KOMAGATA Y,
MITSUYAMA M, MIYAZAKI J, AND NARIUCHI H. Reduced T helper 1
responses in IL-12 p40 transgenic mice. J Immunol 160: 588 –594,
1998.
153. ZHANG JJ, VINKEMEIER U, GU W, CHATRAVARTI D, HORVATH CM, AND
DARNELL JEJ. Two contact regions between Stat1 and CBP/p300 in
interferon-␥ signalling. Proc Natl Acad Sci USA 93: 15092–15096,
1996.
82 • JANUARY 2002 •
www.prv.org