Inactivation of the Integrin [36 Subunit Gene Reveals a Role of
Epithelial Integrins in Regulating Inflammation in the
Lungs and Skin
Xiao-Zhu Huang, *§ll Jian Feng Wu,*ll Darrell Cass, ~ David J. Erle,*ll David Corry,*ll Stephen G. Young, *§ll
Robert V. Farese, Jr., *~l[ and Dean Sheppard *~FI
*Lung Biology Center, *J. David Gladstone Institute for Cardiovascular Research, §Cardiovascular Research Institute; and the
Departments of I/Medicine and tSurgery, University of California, San Francisco, California 94143
Abstract. The integrin av[36 is only expressed in epi-
thelial cells. In healthy adult epithelia, this receptor is
barely detectable, but expression is rapidly induced fol-
lowing epithelial injury. Mice homozygous for a null
mutation in the gene encoding the 1~6 subunit had juve-
nile baldness associated with infiltration of macro-
phages into the skin, and accumulated activated lym-
phocytes around conducting airways in the lungs. 136 -/-
mice also demonstrated airway hyperresponsiveness to
acetylcholine, a hallmark feature of asthma. These re-
suits suggest that the epithelial integrin av136 partici-
pates in the modulation of epithelial inflammation. Ge-
netic or acquired alterations in this integrin could thus
contribute to the development of inflammatory dis-
eases of epithelial organs, such as the lungs and skin.
flammation, wound healing, and tumorigenesis (Hynes,
1987, 1992; Ruoslahti and Pierschbacher, 1987). In vitro,
integrins have been shown to contribute to cell adhesion,
spreading and migration, and to more complex processes
including cell proliferation, apoptosis, and the regulation
of gene expression (Hynes, 1992; Juliano and Haskill,
1993). In vivo, the biological importance of integrins has
been most clearly demonstrated for control of leukocyte
migration and platelet aggregation. At least some integrins
play critical roles in development, as demonstrated by the
findings that the inactivation of the genes encoding two
different integrin ~ subunits, a4 (Yang et al., 1995) and ~5
(Yang et al., 1993), are embryonic lethal mutations. How-
ever, although epithelial cells express several different in-
tegrins, the roles epithelial integrins play in health and dis-
ease remain largely unknown.
The integrin [36 subunit is expressed exclusively in epi-
thelial cells, and only in a single integrin heterodimer,
av[36, a receptor for the extracellular matrix proteins fi-
bronectin (Busk et al., 1992; Weinacker et al., 1994) and
tenascin (Prieto et al., 1993). av[36 is highly expressed in
the lung, skin, and kidney during organogenesis (Breuss et
al., 1995). In epithelia of healthy adults, this receptor is ex-
pressed at very low levels, except in the endometrium,
where av[36 is highly expressed during the secretory phase
NTEGRINS are heterodimeric receptors for extracetlular
matrix and cell surface ligands that have been sug-
gested to play important roles in development, in-
Please address all correspondence to D. Sheppard, Lung Biology Center,
University of California, Box 0854, San Francisco, CA 94143-0854. Tel.:
(415) 206-5901. Fax: (415) 206-4123. e-mail: firstname.lastname@example.org
of the menstrual cycle, and in the lung and kidney, where
cw136 is often expressed in a patchy distribution associated
with subclinical inflammation (Breuss et al., 1995). av[36 is
highly expressed in response to injury or inflammation,
and in malignant epithelial neoplasms (Breuss et al., 1995).
For example, in experimental skin wounds, av136 is highly
expressed in the keratinocytes at the wound edge within a
few days of wounding, where it remains expressed until
soon after wound closure is complete. In the respiratory
epithelium, av136 mRNA expression is induced within 5 h
of acute injury, and c~v136 protein can be detected in the
epithelium of patients with a variety of inflammatory lung
diseases (Breuss et al., 1995; Weinacker et al., 1995).
These findings suggest a role for this receptor in the re-
sponse of epithelia to injury.
To examine the role(s) that cxv[36 plays in vivo, we have
generated mice lacking 136 expression using homologous
recombination in embryonic stem cells. These mice de-
velop and reproduce normally, but develop functionally
significant infiltration of their skin and lungs with inflam-
matory cells. In the conducting airways of the lung, these
morphologic changes are associated with enhanced bron-
choconstrictor sensitivity to acetylcholine, the central
physiologic abnormality in human asthma.
Materials and Methods
Inactivation of the f16 Subunit Gene in Mouse
Embryonic Stem Cells
We initially amplified a 240-bp fragment of mouse [36 cDNA by poly-
merase chain reaction (PCR) with degenerate mixtures of oligonucle-
otides based on the human and guinea pig 136 sequences (Sheppard et al.,
© The Rockefeller University Press. 0021-9525/96/05/921/8 $2.00
The Journal of Celt Biology, Volume 133, Number 4, May 1996 921-928 921
1990). We then used the resultant fragment as a probe to screen a genomic
mouse 129 strain library, and obtained a 15-kb clone containing two exons
and portions of three large introns. The most 5' exon extended to within
50 amino acids of the predicted 5' end of the mature protein, predicted on
the basis of the human sequence. We used a 7-kb EcoRI fragment of this
clone to construct a replacement vector that contained a neomycin resis-
tance gene inserted into the second of the two exons in our clone, and a
thymidine kinase gene at the 5' end (see Fig. 1). We introduced this vector
into 129 strain mouse embryonic stem (ES) I cells and identified targeted
clones by Southern blotting with two different probes and by PCR. The
ES cell line used, RF8, was derived from agouti 129/terSV mice (a gift
from Dennis Huszar, GenPharm International, Mountain View, CA) and
cultured on SNL76/7 mitotically inactive feeder cells (a gift from Allan
Bradley, Baylor College of Medicine, Houston, TX). Only targeted clones
resulting from a single integration event (as judged by Southern blotting
using a neo probe) were used for blastocyst injection. Targeted clones were
injected into C57B1/6 blastocysts, and one of these clones produced two
90% chimeric male offspring that transmitted the inactivated gene
through the germline, as determined by both Southern blotting and PCR.
Heterozygous offspring of crosses between these high percentage chimeras
and pure C57B1/6 females were crossed to produce mice that were ho-
mozygous for the null mutation, and homozygous wild-type litter mates
that served as controls in subsequent experiments. The 240-bp fragment
used for library screening was originally amplified by PCR with 13 subunit
(T)TIATGGA -3') and B2AR (5'-GGICTT(C)CCACCIA(G)AICTA-
(G)CGG(T)TAITACG-3'). The resultant fragment was reamplified with
degenerate 136 primers 1363F (5'-GA(TC)GA(TC)CTIAA(CT)ACIAT-
(CAT)AA(GA)GA-3') and 1364R (5'-TC(GA)TI'(GA)AA(CT)CT(CT)-
TCIGC(GA)TC(GA)Tr-3') designed based on human and guinea pig 136
sequences (Sheppard et al., 1990).
Detection of Recombinant Clones by PCR and
RF8 embryonic stem cells were grown in embryonic stem cell complete
media. The targeting vector was linearized at a unique SacII site and trans-
fected into RF8 ES cells by electroporation. Selection medium containing
the neomycin analogue G418 (0.15 mg/ml) and FIAU (0.2 p.m) was used to
obtain resistant clones. Individual colonies were screened by PCR and South-
ern blotting. PCR was performed with a forward primer from within the
introduced neomycin resistance gene (neoF-5'CAGTAAATCGTTGTC-
AACAG) and a reverse primer from the mouse 136 gene 3' of the target-
ing vector (Km136R-5'GTGGATCTGCTAAGTI'AACC). For Southern
blotting, genomic DNA digested with BamHI was blotted with two differ-
ent probes, one specific for mouse 136 and the other for the inserted neo-
mycin resistance gene, and only clones with a single integration were used
for blastocyst injection.
Generation of Germline Chimeras
Chimeras were generated as described by Bradley (1987). ES cells were
injected into C57B1/6 blastocysts and the injected embryos were trans-
ferred into the uteri of pseudopregnant recipients. Chimeras identified by
the presence of an agouti coat color were test-mated with C57Bl/6J fe-
males. Offspring were tested for the targeted 136 gene by PCR and South-
Keratinocyte Culture and Immunoprecipitation
Mouse skin was removed and placed in 0.1% protease (Sigma Chem. Co.,
St. Louis, MO) at 4°C overnight. The following day the skin was trans-
ferred to 0.05% trypsin and the epithelial layer was scraped off with a sur-
gical blade. After 30 min incubation in trypsin, cells were disaggregated by
pipetting, and passed through a 150-~m nylon mesh to remove residual
hair. The cells were suspended in keratinocyte growth medium (Clonetics,
San Diego, CA), plated onto dishes coated with collagen, and grown to
confluence. After overnight labeling with [35S]methionine, cells were lysed
in immunoprecipitation buffer (100 mM Tris-HC1, pH 7.5, 0.1% SDS, 1%
Triton X-100, 0.1% NP-40, 300 mM NaC1) and immunoprecipitated with
Ab 206 which was raised against the cytoplasmic domain of human 136.
1. Abbreviations used in this paper: ES, embryonic stem; IL-4, interleukin-4;
RL, pulmonary resistance.
Samples were analyzed by SDS-PAGE on 7.5% acrylamide gels and ex-
posed to film at -80°C.
Total RNA was harvested from lung and kidney tissue using Trizol re-
agent (BRL, Grand Island, NY) and cDNA was synthesized using Su-
perscript reverse transcriptase (BRL). A 360-bp fragment of murine 136
flanking the neomycin resistance gene insertion site using primers m1366F
(5' CAGTrCTGACATTGTTCAGA 3') and m1365R (5' TGTTAATGG-
Collection of Lung Cells and Flow Cytometry
Mouse lungs were perfused with PBS via the main pulmonary artery to re-
move intravascular cells. The lungs were removed and minced into fine
fragments that were gently dispersed into RPMI medium (GIBCO BRL)
using a syringe plunger and passed through a 0.75-tLm nylon mesh filter.
Cells from minced mouse lungs were stained with phycoerythrin-, fluores-
cein isothiocyanate-, or biotin-conjugated antibodies against CD4, CD8,
B220, and CD25 (Caltag, South San Francisco, CA) and analyzed for surface
expression using a FACSCAN flow cytometer (Becton Dickinson, Moun-
tain View, CA).
Histology and Immunohistochemistry
Freshly isolated organs were embedded in OCT and quick frozen in liquid
nitrogen. Serial 5-~m sections were prepared and fixed in Histochoice
(Fisher Scientific, Pittsburgh, PA) for hematoxylin and eosin staining. For
immunohistochemistry, frozen sections were fixed in cold acetone for 5-10
rain and air-dried. Sections were blocked for endogenous peroxidase and
biotin activities with Peroxoblock solution (Zymed Labs, S. San Francisco,
CA) and Avidin/Biotin Blocking Kit (Vector Labs, Inc., Burlingame, CA)
at room temperature. After rinsing, sections were blocked with 0.25%
casein/0.025% thimerosal in PBS for 15 min and then incubated overnight
at 4°C in biotin-labeled primary antibodies against CD3 (T cells), B220 (B
cells), and F4/80 (macrophages) (all from Caltag). After washing, sections
were incubated in ABC avidin/peroxidase reagent (Vector) for 1 h at
room temperature. Chromagen was developed using the DAB Plus Kit
(Zymed). Finally, sections were dehydrated and mounted with permount
onto clean slides.
Measurement of Airway Resistance
Mice were anesthetized with pentobarbital (50 mg/kg, i.p.), the chest was
opened, and a tracheotomy tube was inserted. The mice were paralyzed
with pancuronium bromide (0.1 mg/kg), and then ventilated with 100%
oxygen by a Harvard small animal ventilator at a rate of 150 breaths/
minute and a tidal volume of 9 I~l/gm. In pilot experiments, we determined
that these settings result in near normal values of arterial (left ventricular)
pH and pCO 2. A heparinized, indwelling catheter was placed in the tail
vein, and the animal was placed in an airtight plexiglass plethysmograph,
with the venous catheter threaded through a small hole in the plethysmo-
graph. Airway pressure and plethysmograph pressure were continuously
measured by differential pressure transducers and recorded on a Hewlett-
Packard chart recorder. Pulmonary resistance (RL), tidal volume, flow,
and dynamic lung compliance were continuously calculated by a Buxco
Pulmonary Mechanics analyzer. RL was calculated on each inspired
breath as the ratio of driving pressure to airflow at 70% of inspired tidal
volume. Increasing concentrations of acetylcholine were administered
through the tail vein catheter at 2-min intervals, and peak pulmonary re-
sistance in response to each concentration was determined.
Elispot assay were performed as described (Corry et ak, 1996). This assay
is a modification of a sandwich ELISA that allows the identification of in-
dividual cells that are secreting specific antigens. Briefly, wells of 96-well
microtiter plates (Dynatech, Chantilly, VA) were coated with either mAb
against IFN-7 (R46A2) or IL-4 (BVD4-1Dll.2) at 4°C overnight and
blocked with 10% FBS. Pooled cells from minced lungs were plated and
incubated at 37°C overnight. Plates were incubated for I h with biotiny-
luted secondary antibodies XMG-1.2 to IFN-7 and BVD6-24G.2 to IL-4,
and then with streptavidin-conjugated alkaline phosphatase (Jackson Im-
munoResearch, West Grove, PA) for another hour. Between incubations,
plates were washed with phosphate buffered saline supplemented with
The Journal of Cell Biology, Volume 133, 1996 922
0.05% Tween-20. Final color development was obtained by incubating
plates with 5-bromo-4-chloro-3-indolyl phosphate (Sigma) in 2-amino-2-
methyl-l-propanol buffer (Sigma) suspended in 0.6% agarose low-melt
gel. After solidification of the agarose, individual blue spots were counted
under an inverted microscope.
Mice Homozygous for a Null Mutation in the ~6
Subunit Develop Juvenile Baldness
To assess whether /36 -/- mice produced [36 mRNA, we
amplified cDNA obtained from the lung and kidney of
/36 -/- and/36 +/+ mice by PCR using oligonucleotide prim-
ers flanking the exon disrupted by our targeting vector
(Fig. 1). The expected 360-bp fragment could be amplified
from the cDNA of/36 +/+ mice, but no amplification prod-
uct was detectable from the cDNA of/36 -/- mice (Fig. 2 A).
To confirm that/36 -/- mice were not capable of making [36
protein, we attempted to immunoprecipitate 136 from met-
abolically labeled lysates of cultured keratinocytes with an
antibody raised against a peptide based on the cytoplasmic
domain sequence of human 136 (Ab 206 [Busk et al.,
1992]). The anti-J36 antibody immunoprecipitated two
bands of the appropriate apparent molecular masses to be
av and 136, from/36 +/+ keratinocytes, but no bands were
immunoprecipitated from/36 -/- keratinocytes (Fig. 2 B).
Mice homozygous for the null mutation were born at ap-
proximately the expected Mendelian frequency from het-
erozygous intercrosses (30% +/+, 48% +/-, and 22% -/-
of 234 offspring analyzed), demonstrating that this integrin
subunit, in contrast to a4 (Yang et al., 1995) and a5 (Yang
et al., 1993), is not absolutely required for embryonic de-
velopment. However, /36 -/- mice were not completely
normal. All /36 -/- mice failed to develop hair normally
over the tops of their heads, the backs of their necks, and
the inner surface of their thighs (Fig. 3 A). These abnor-
malities were visually apparent by postnatal day 5, and
persisted through day 20-30. After this time, hair growth
resumed over the head and neck but remained sparse over
the inner surface of the thighs. No gross abnormalities in
Figure 2. (A) RT-PCR analysis of mRNA from/36 +/+ and/36 -/-
mice. Total RNA was extracted from mouse lung and kidney and
transcribed to complementary DNA (cDNA). A 360-bp amplifi-
cation fragment was detectable in tissues of/36 +/+ mice but not of
/36 -/- mice with primers specific for wild-type 136 cDNA. (B) Im-
munoprecipitation of av136. Primary cultures of keratinocytes
from/36 +/+ and/36 -/- were labeled overnight with [3SS]methio-
nine and cell lysates were immunoprecipitated with Ab 206,
raised against the human [36 cytoplasmic domain. Immunoprecip-
itated proteins were analyzed by 7.5% SDS-PAGE under nonre-
ducing conditions. The positions of molecular size markers (in
kD) are shown to the left.
any other organs have been apparent in mice followed for
up to 6 mo. Furthermore,/36 -/- mice gained weight nor-
mally and were fertile.
The Dermis in Bald Areas of f16 -/- Mice Is Infiltrated
The hair loss seen in/36 -/- mice was associated with mor-
phologic abnormalities in the dermis of affected areas. In
comparison to/36 +/+ mice, 136 -/- mice had fewer hair folli-
cles, and numerous degenerating hair follicles surrounded
by foci of mononuclear cells (Fig. 3 C). The mononuclear
cells resembled tissue macrophages and stained with the
monocyte/macrophage marker F4/80. Furthermore, stain-
ing with F4/80 demonstrated increased numbers of mono-
cytes/macrophages throughout the dermis of the bald ar-
eas in /36 -/- mice (Fig. 3 E). Staining with antibodies
specific for mouse T cells and B cells did not demonstrate
Figure 1. (A) Replacement
vector for inactivation of the
136 gene in mouse embryonic
stem (ES) cells. Top panel
shows the structure of a 15-kb
fragment of the mouse 136 gene
isolated from a 129 strain ge-
nomic library. The two exons
in this fragment are shown as
shows the targeting plasmid
(Stratagene, La Jolla, CA) and
linearized at a unique SacII
site. A neomycin-resistance gene under the control of the RNA polymerase II promoter (neo) and herpes simplex virus thymidine ki-
nase (HSV-tk) gene (gifts from Kirk Thomas, University of Utah, Salt Lake City, UT) are shown as shaded boxes. Bottom panel shows
the expected structure of the [36 gene after homologous recombination. The arrowheads represent PCR primers used to identify homol-
ogous recombination events in ES cell colonies. The forward primer is located in the neo gene and the reverse primer is 3' of the target-
ing vector. These primers generated a 1.2-kb fragment in targeted clones. The locations of the two probes used for Southern blotting are
shown as lines. Hybridization of either probe with genomic DNA digested with BamHI yielded a new 6.8-kb band in targeted clones.
(B) Southern blot analysis of wild-type (first two lanes) and targeted (last two lanes) ES clones digested with BamHI and hybridized with
the external probe (probe 2).
Huang et al. Integrin ~6 Subunit Knockout
Figure 3. (A) Photograph of a f16 -/- mouse (left) and a t36 +/+ littermate (right) at 10 d of age. In contrast to the/36 +/+ mouse, the 136 -j-
mouse demonstrates the typical pattern of juvenile baldness over the head and neck. 136 -/- mice also lack hair over the inside of their
thighs. B and C are low power photomicrographs of sections of frozen inner thigh skin from a/36 +/+ (B) and a/36 -/- mouse (C) stained
with hematoxylin and eosin (H and E), and demonstrate loss of hair follicles and infiltration of the dermis in the/36 -/- animal. D and E
are higher power photomicrographs of frozen sections of skin from the same areas of the same 136 +/+ (D) or 136 -/- mouse (E) stained
with the monocyte/macrophage marker F4180, and lightly counterstained with H and E. Hair follicles are marked with arrows. Infiltrat-
ing macrophages are marked with arrowheads. Note the marked increase in F4/80 positive cells in the/36 -/- mouse.
any increase in lymphocytes in the skin lesions (data not
shown). No morphologic abnormalities were seen in skin
taken from unaffected areas.
Expression of av136 is induced within a few days of cuta-
neous wounding, is expressed only in the keratinocytes at
the wound edge, and remains expressed until wounds are
completely closed (Breuss et al., 1995). The two known
ligands for etv136, fibronectin (Busk et al., 1992; Weinacker
et al., 1994), and tenascin (Prieto et al., 1993), are compo-
nents of the provisional matrix across which keratinocytes
must migrate during wound repair. Therefore, we hypoth-
esized that this integrin might be critical for normal cuta-
neous wound healing. However, when either incisional or
excisional wounds (2 or 4 mm in diameter punch biopsy
wounds) were made in the backs or necks of/36 -t- or/36 +I+
mice, the rate of healing and the morphology of healing
wounds was similar in both groups for mice examined 2, 4, 6,
or 12 d after wounding. In both groups, all incisional
wounds were completely healed by day 6, and all exci-
sional wounds were healed by day 12. These data indicate
that etv136 is not required for normal wound healing. This
result could be explained by the expression of other kera-
tinocyte fibronectin and tenascin receptors. For example,
the fibronectin receptor, a5131, was found to be increased
at the wound edge in both 136 -/- and ~6 ÷/+ mice, and the
tenascin receptor, et9131 (Palmer et al., 1993; Yokosaki et
al., 1994), was constitutively expressed on keratinocytes
from both types of mice (data not shown).
f16 -/- Mice Demonstrate Infiltration of the Conducting
Airways of the Lung by Activated Lymphocytes
The other significant pathologic finding observed in 136 -/-
mice was in the lungs. Lung morphology was normal in
mice examined up to 13 d after birth. However, beginning
at ~21 d,/36 -t- mice developed infiltration of the walls of
the conducting airways with mononuclear cells (Fig. 4).
These cells morphologically resembled lymphocytes. In
contrast to the skin lesions, the lung lesions did not contain
increased numbers of F4/80 + cells, but did contain a mix-
ture of cells that stained with the B cell marker, B220, or
the T cell marker CD3 (data not shown). This pathology
The Journal of Cell Biology, Volume 133, 1996 924
Figure 4. Lung histology. Low
power photomicrograph of H
and E stained frozen section of
a conducting airway from a
/36 +/+ (A) and a /36 -/- mouse
(B) demonstrate accumula-
tion of mononuclear cells (ar-
rowheads) diffusely around a
large conducting airway in the
/36 -/- mouse. This pathology
was seen in all 11 /36 -/- mice
examined, but in none of 16
/36 +/- or /36 +/+ mice. (C)
Higher power photomicro-
graph of the same airway
shown in B.
was seen in all 11/36 -/- mice analyzed after 21 d of age. No
similar lesions were observed in 16/36 +/+ or/36 +/- mice.
Flow cytometry of cells obtained from minced lungs dem-
onstrated approximately threefold increases in the per-
centages of B cells and CD4+ and CD8+ T cells in/36 -/-
mice. These mice also demonstrated a marked increase in
the number of CD4+ T cells that expressed the activation
marker CD25 (Fig. 5). These abnormalities in lung mor-
phology and cellularity were not likely to be due to lung
infection, since all animals were housed together in a bar-
rier facility. Sentinel mice were examined weekly and
found to be free of specific pathogenic viruses.
86 -/- Mice Demonstrate Increased
Despite the progressive abnormalities in lung morphology
noted in the /36 -/- mice, these mice appeared healthy,
without noticeable respiratory distress. However, because
activated T cells have been implicated in the development
of asthma (Gavett et al., 1994), we sought to determine
whether these animals would demonstrate one of the hall-
mark features of asthma, airway hyperresponsiveness to
bronchoconstrictor agents such as acetylcholine (Boushey
et al., 1980). Groups of/36 -/- and/36 +/+ mice were anes-
thetized, tracheostomized, and ventilated in a whole body
plethysmograph, and pulmonary resistance (RL) was mea-
sured at baseline and after intravenous administration of
increasing concentrations of acetylcholine. Baseline RL
was the same in both groups of mice (Fig. 6). The/36 +/+
mice developed only small increases in RL, even after ad-
ministration of the highest concentrations of acetylcholine
used, whereas the /36 -/- mice demonstrated markedly
larger responses to each of the two highest concentrations
(P < 0.001).
Cells Obtained from the Lungs of 86 -I- Mice
Express the Th2-associated Cytokine IL-4, but Not
the Th rassociated Cytokine lnterferon-~
In the most widely studied experimental model of asthma,
antigen-induced airway hyperresponsiveness, increased
Huang et al. lntegrin [36 Subunit Knockout
B Cells CD8+ CD4+ CD25+!
Figure 5. Lymphocyte subsets in mouse lungs. Lung cells were
obtained by mincing lungs pooled from groups of/36 +/+ (open
bars) and/36 -/ mice (shaded bars). Cells were stained with anti-
bodies specific for mouse B cells (B220), or T cell markers CD4
or CD8, and analyzed by flow cytometry. To identify activated
CD4+ T cells, cells were simultaneously stained with antibodies
to CD4 and to the lymphocyte activation antigen CD25.
airway responses to acetylcholine are associated with the
presence of T cells in the airway wall that express the so-
called Th2 (T helper 2) phenotype, characterized by ex-
pression of the cytokines interleukin-4 (IL-4), IL-5, and
IL-10. In contrast, Thl ceils, thought to be important for
cytotoxic T cell responses, express the cytokine Inter-
feron--/(IFN--/), but not IL-4, IL-5, or IL-10. Recent data
using antibodies against IL-4 and IL-4 knockout mice, sug-
gest that IL-4, in particular, is crucial to the development
= + T
---o-- +/+ *_~T
of airway hyperresponsiveness in this model (Brusselle et
al., 1995; Corry et al., 1996). To determine whether the
lymphocytes in the lungs of/76 -/- mice were expressing ei-
ther Thl or Th2 cytokines, we performed Elispot assays to
detect IL-4 or IFN-~/ secretion from cells obtained from
minced lungs of 136 -/- or/36 +/+ mice. These assays, which
allow us to count individual lung cells secreting each cyto-
kine, demonstrated a fourfold increase in the percentage
of lung cells that secrete IL-4 with no increase in the per-
centage of cells that secrete IFN-~/ in /36 -/- mice (Fig. 7).
Multiple members of the integrin family are expressed in
vivo in epithelial cells, including a2~31, et3131, a9131, et6134,
etv~35, and etv136. Of these, a6134 is a critical component of
the hemidesmosomes that basal epithelial cells use for at-
tachment to the underlying basement membrane, but the
functions each of the other integrins play in epithelial tis-
sues remain largely unknown, o~vl36 is the only known
member of the integrin family that is restricted in its distri-
bution to epithelial cells. However, unlike most epithelial
integrins, av136 is not constitutively expressed in healthy
epithelia, but is rapidly and transiently induced in re-
sponse to local injury or inflammation. The principal
ligands identified for oLv136, fibronectin and tenascin, are
also absent from healthy epithelia, but present in most in-
jured and inflamed epithelia. Based on these observations,
we and others hypothesized that otv~6 might play some
critical role in the repair of injured epithelial tissues. Be-
cause of the known in vitro roles played by integrins in cell
spreading, migration and proliferation, we presumed that
any in vivo role of av136 would probably involve one or
more of these functions. Such effects of etv136 would also
be consistent with the induction of av136 during organo-
genesis, its high level expression in secretory phase en-
dometrium, and its expression in tumors derived from epi-
thelia; since cell migration, spreading, and proliferation
are important in organ development, endometrial regener-
ation, and tumorigenesis.
The results of the present study are therefore somewhat
surprising. /36 -/- mice develop and reproduce normally,
and are fully capable of healing cutaneous wounds. Al-
though these observations by no means prove that ave6
1.0 3.0 10 30
Ach dose (~g/g)
Figure 6. Airway responsiveness to acetylcholine in /36 +/+ and
/36 -/- mice. Total pulmonary resistance (RL) was measured in a
whole body plethysmograph in seven /36 +/+ mice (open circles)
and in six /36 -/- mice (closed squares) from matched litters, at
baseline (B.L.) and after the administration of successively in-
creasing doses of acetylcholine into an indwelling tail vein cathe-
ter. Data are plotted as the mean (-SD). Pulmonary resistance
was calculated using an analogue computer (model 6, Buxco,
Sharon, CT) (22) in mice that were anesthetized, paralyzed, and
ventilated breathing 100% oxygen with a rodent ventilator (Har-
vard-Ealing, Millis, MA) at a rate of 150 breaths/min and a tidal
volume of 9 Ixl/g, settings that resulted in normal arterial blood
gases in pilot experiments. *P < 0.0001 as determined by Stu-
dent's t test for unpaired data, adjusted for multiple comparisons.
Figure 7. Cytokine production by lung cells. Purified lung cells
were assessed for secretion of IL-4 and IFN-',/using an Elispot as-
say to detect cytokine secretion from individual cells. Data are
expressed as the mean (_+ SD) number of positive spots/106 cells.
The Journal of Cell Biology, Volume 133, 1996 926
does not normally contribute to development reproduc-
tion, and/or wound healing, they do suggest that other re-
ceptors can serve the same functions. However, our obser-
vation that 136 -/- mice all develop inflammatory cell
infiltrates in the skin and lungs suggest a previously unex-
pected role for this integrin in modulating inflammation in
these epithelial organs.
It is unclear why the pathology seen in [36 -/- mice has
thus far been confined to the lungs and skin. In addition to
these two organs, av[~6 can be expressed on epithelial cells
in a variety of other organs, including the kidney, uterus,
testes, ovary, salivary glands, and gall bladder (Breuss et
al., 1993). At least in the kidney, as in the lung and skin,
av136 expression is dramatically upregulated by injury or
inflammation. In the present study, the mice we describe
were housed continuously in a specific pathogenic virus
free barrier facility, so they were largely protected from
environmental insults. However, the conducting airways
of healthy mammals are repeatedly exposed to aspirated
gastric contents, and we hypothesize that focal areas of in-
jury or inflammation might be routine events, even in ani-
mals maintained in an environmentally controlled barrier
facility. In addition, even in a barrier facility, mice are po-
tentially exposed to inhaled antigens from their food and
bedding. Similarly, the pattern of abnormality seen in the
skin of/36 -/- mice might be explained by repeated low-
level injury. Baldness and macrophage infiltration of the
skin was most prominent in the head and neck skin of mice
that had not yet been weaned, and resolved by ~30 d of
age. This pattern corresponds to the area used by mothers
to lift and move mouse pups. The inside of the thighs is
also an area that could be subjected to continuous low-
level irritation. In this region, hair loss persists even in
adult/36 -/- mice.
One explanation for our findings is that c~vl36 partici-
pates in the regulation of local factors that are responsible
for activation, recruitment and/or proliferation of lympho-
cytes and/or monocytes. For example, expression of av136
following epithelial injury could modulate the synthesis of
inflammatory cytokines by epithelial cells, contributing to
the termination of the local inflammatory response. Inacti-
vation of c~vl36, especially in sites such as the airways and
skin that are repeatedly exposed to injurious stimuli,
would thus lead to persistent inflammation. A role for in-
tegrins in regulation of cytokine gene expression has been
demonstrated in mononuclear cells, where the regulation
of expression of several cytokine genes can be dramati-
cally altered by plating cells on different extracellular ma-
trix substrates, an effect that is mediated by integrins
(Haskill et al., 1988; Juliano and Haskill, 1993; Miyake et
al., 1993). It is now well-recognized that epithelial cell-
derived chemokines and other cytokines play important
roles in initiating and modulating inflammatory responses
in epithelial organs, including the lungs and skin (Becker
et al., 1994; Bellini et al., 1993; DiCosmo et al., 1994; Elias et
al., 1994; Sousa et al., 1994). For example, one of the cyto-
kines synthesized by airway epithelial cells, IL-6, induces
the proliferation of both B lymphocytes and T lympho-
cytes, and also induces lymphocyte activation, as detected
by expression of the interleukin-2 receptor, CD-25 (Lotz
et al., 1988; Tosato et al., 1988; Tosato and Pike, 1988).
Overexpression of IL-6 in the airway epithelium of trans-
genic mice induced a very similar pathology to that seen in
the/36 -/- mice we describe, including large accumulations of
B cells and T cells adjacent to conducting airways (DiCosmo
et al., 1994).
One surprising feature of our results is that the charac-
ter of the inflammatory ceils in the skin and lungs of/36 -/-
mice is different. The cells in the skin are principally macro-
phages, whereas the cells in the lung are principally B and T
lymphocytes. This difference suggests that either the ef-
fects of ~v~6 ligation on keratinocytes and airway epithe-
lial cells are different, or that the subsequent cellular re-
sponses to any effects of avt36 ligation are different in the
microenvironments of the skin and lungs. Differential ef-
fects of cytokines in different tissues are well described.
For example, transgenic mice overexpressing interleukin-6
in airway epithelial cells develop loci of lymphocytes sur-
rounding conducting airways in a pattern quite similar to
that seen in the ~6 -/- mice we describe (DiCosmo et al.,
1994). However, mice overexpressing IL-6 in keratinocytes
do not develop lymphocytic infiltrates in the skin (Turksen
et al., 1992).
The /36 -/- mice we describe demonstrate dramatic in-
creases in bronchomotor responsiveness to acetylcholine,
the hallmark feature of asthma in humans (Boushey et al.,
1980). Lymphocytes are thought to play an important role
in induction of airway hyperresponsiveness in human
asthma (Robinson et al., 1992; Walker et al., 1991) and are
the principal cell type recruited to the airways in these
mice. Furthermore, as in human asthma, the CD4+ T cells
in the airways of 136 -/- mice appear to be activated. Fi-
nally, the lungs of ~6 -/- mice contain many more IL-4 se-
creting cells than do the lungs of 136 +/+ mice. Two recent
studies in which airway hyperresponsiveness was tran-
siently induced in ovalbumin-sensitized mice by inhalation
of ovalbumin suggested that IL-4 played an important role
in the induction of airway hyperresponsiveness. In one of
these studies, IL-4 deficient mice were shown to be resis-
tant to ovalbumin-induced airway hyperresponsiveness
(Brusselle et al., 1995), and in the other study, ovalbumin-
induced airway hyperresponsiveness was prevented by ad-
ministration of an anti-IL-4 antibody (Corry et al., 1996).
One feature that distinguishes the animals we describe
from most people with asthma, and from ovalbumin-sensi-
tized and challenged animals is the absence of eosinophils
in the airways. Despite the absence of eosinophils,/36 -/-
mice demonstrated marked airway hyperresponsiveness,
suggesting that eosinophils are not required for induction
of this physiologic abnormality. These data are consistent
with the recent report that treatment of ovalbumin-sensi-
tized and challenged mice with anti-IL-5 antibody pre-
vents airway eosinophilia but has no effect on the induc-
tion of airway hyperresponsiveness.
As discussed above, in the most widely studied experi-
mental models of asthma, lymphocytes are recruited to the
airways and activated in response to inhalation of specific
allergens (Gavett et al., 1994). These models suffer from a
high degree of variability and from the transient nature of
the airway hyperresponsiveness induced. Moreover, many
patients with asthma are not atopic, and most clinical exac-
erbations of asthma are provoked by nonimmunologic
stimuli (Boushey et al., 1980). Many of these stimuli, in-
cluding viral infection and inhalation of irritant gases, are
Huang et al. Integrin 86 Subunit Knockout
likely to directly affect the airway epithelium. Further-
more, one nearly universal feature of human asthma is an
alteration in the amount and composition of extracellular
matrix in the airway wall (Djukankovic et al., 1990; James
et al., 1989; Laitinen and Laitinen, 1994). The results of the
present study suggest that these changes in the matrix, de-
tected by integrins, could modulate recruitment and acti-
vation of airway lymphocytes, and that genetic or environ-
mental inactivation of the airway epithelial integrin, c~vl36,
could induce or potentiate the airway lymphocytosis and
airway hyperresponsiveness that characterize human asthma.
Similarly, alterations in signals initiated through epithelial
integrins could contribute to inflammatory diseases affect-
ing other epithelial organs.
The authors gratefully acknowledge excellent technical assistance from
Eric Somde, Heather Myers, and John Chen.
This work was supported in part by National Institutes of Health grants
HL/A133259, HL47412, HL53949, and HL47660.
Received for publication 4 January 1996 and in revised form 14 February
Becker, S., J. Quay, H.S. Koren, and J.S. HaskiU. 1994. Constitutive and stimu-
lated MCP-1, GRO c~, 13, and 3, expression in human airway epithelium and
bronchoalveolar macrophages. Am. J. Physiol. 266:L278-286.
Bellini, A., H. Yoshimura, E. Vittori, M. Marini, and S. Mattoli. 1993. Bronchial
epithelial cells of patients with asthma release chemoattractant factors for T
lymphocytes. J. Allergy Clin. lmmunol. 92:412-424.
Boushey, H.A., MJ. Holtzman, J.A. Sheller, and J.A. Nadel. 1980. State of the
art: bronchial hyperreaetivity. Am. Rev. Respir. Dis. 121:389-413.
Bradley, A. 1987. Production and analysis of chimaeric mice. In Teratocarcino-
mas and Embryonic Stem Ceils: A Practice Approach. E.J. Robertson, edi-
tor. IRL Press, Oxford. pp. 113-151.
Breuss, J.M., J. Gallo, H.M. DeLisser, I.V. Klimanskaya, H.G. Folkesson, J.F.
Pittet, S.L. Nishimura, K. Aldape, D.V. Landers, W. Carpenter, et al. 1995.
Expression of the 136 integrin in development, neoplasia, and tissue repair
suggests a role in epithelial remodeling. Z Cell Sci. 108:2241-2251.
Breuss, J.M., N. Ginett, L. Lu, D. Sheppard, and R. Pytela. 1993. Restricted dis-
tribution of integrin 136 mRNA in primate epithelial tissues. 3". Histochem.
Brusselle, G., J. Kips, G. Joos, H. Bluethmann, and R. Pauwels. 1995. Allergen-
induced airway inflammation and bronchial responsiveness in wild-type and
interleukin-4 deficient mice. Am. J. Respir. Cell Mol. Biol. 12:254-262.
Busk, M., R. Pytela, and D. Sheppard. 1992. Characterization of the integrin
otv136 as a flbronectin-binding protein. J. BioL Chem. 267:5790-5796.
Corry, D., H.G. Folkesson, M.L. Warnock, D.E. Erie, M.A. Matthay, J.P.
Wiener-Kronisch, and R.M. Locksley. 1996. Interleukin 4, but not interleu-
kin 5 or eosinophils, is required in a murine model of acute airway hyperre-
activity. J. Exp. Med. 183:109-118.
DiCosmo, B.F., G.P. Geba, D. Picarella, J.A. Elias, J.A. Rankin, B.R. Stripp,
J.A. Whitsett, and R.A. FlaveU. 1994. Airway epithelial cell expression of in-
terleukin-6 in transgenic mice. Uncoupling of airway inflammation and
bronchial hyperreactivity. J. Clin. Invest. 94:2028-2035.
Djukankovic, R., W.R. Roche, J.W. Wilson, O.P. Twentyman, P.H. Howarth,
and S.T. Holgate. 1990. State of the art: mucosal inflammation in asthma.
Am. Rev. Respir. Dis. 142:434--457.
Elias, J.A., T. Zheng, O. Einarsson, M. Landry, T. Trow, N. Rebert, and J. Pa-
nuska. 1994. Epithelial interleukin-ll. Regulation by cytokines, respiratory
syncytial virus, and retinoic acid. J. BioL Chem. 269:22261-22268.
Gavett, S.H., X. Chen, F. Finkelman, and M. Wills-Karp. 1994. Depletion of
murine CD4+ T lymphocytes prevents antigen-induced airway hyperreac-
tivity and pulmonary eosinophilia. Am. J. Respir. Cell Mol. Biol. 10:587-594.
Haskill, S., C. Johnson, D. Eirman, S. Becker, and K. Warren. 1988. Adherence
induces selective mRNA expression of monocyte mediators and proto-onco-
genes. J. Immunol. 140:1690-1694.
Hynes, R.O. 1987. Integrins: a family of cell surface receptors. Cell. 48:549-554.
Hynes, R.O. 1992. Integrins: versatility, modulation, and signaling in cell adhe-
sion. Cell. 69:11-25.
James, A.L., P.D. Pare, and J.C. Hogg. 1989. The mechanics of airway narrow-
ing in asthma. Am. Rev. Respir. Dis. 139:242-246.
Juliano, R.L., and S. Haskill. 1993. Signal transduction from the extracellular
matrix. J. Cell Biol. 120:577-585.
Laitinen, A., and L.A. Laitinen. 1994. Airway morphology: epithelium/base-
ment membrane. Am. ,L Respir. Crit. Care Met. 150:$14-$17.
Lotz, M., F. Jirik, P. Kabouridis, C. Tsoukas, T. Hirano, T. Kishimoto, and D.A.
Carson. 1988. B celt stimulating factor 2/interleukin 6 is a costimulant for hu-
man thymocytes and T lymphocytes. 3". Exp. Met. 167:1253-1258.
Miyake, S., H. Yagita, T. Maruyama, H. Hashimoto, N. Miyasaka, and K. Oku-
mura. 1993. 131 integrin-mediated interaction with extracellular matrix pro-
teins regulates cytokine gene expression in synovial fluid cells of rheumatoid
arthritis patients. J. Exp. Med. 177:863-868.
Palmer, E.L., C. Ruegg, R. Ferrando, R. Pytela, and D. Sheppard. 1993. Se-
quence and tissue distribution of the integrin a9 subunit, a novel partner of
131 that is widely distributed in epithelia and muscle. ,L Cell Biol. 123:1289-
Prieto, A.L., G.M. Edelman, and K.L. Crossin. 1993. Multiple integrins mediate
cell attachment to cytotactin/tenascin. Proc. Natl. Acad. Sci. USA. 90:10154-
Robinson, D.S., Q. Harold, S. Ying, A. Tsicopoulos, J. Barkans, A.M. Bentley,
C. Corrigan, S.R. Durham, and A.B. Kay. 1992. Predominant TH2-1ike
bronchoalveolar T-lymphocyte population in atopic asthma. N. Engl. J. Med.
Ruoslahti, E., and M.D. Pierschbacher. 1987. New perspectives in cell adhesion:
RGD and integrins. Science (Wash. DC). 238:491-497.
Sheppard, D., C. Rozzo, L. Starr, V. Quaranta, D.J. Erie, and R. Pytela. 1990.
Complete amino acid sequence of a novel integrin 13 subunit (136) identified
in epithelial cells using the polymerase chain reaction. J. Biol. Chem. 265:
Sousa, A.R., S.J. Lane, J.A. Nakhosteen, T. Yoshimura, T.H. Lee, and R.N.
Poston. 1994. Increased expression of the monocyte chemoattractant pro-
rein-1 in bronchial tissue from asthmatic subjects. Am. J. Respir. Cell Mol.
Tosato, G., and S.E. Pike. 1988. Interferon-13 2/interleukin 6 is a co-stimulant
for human T lymphocytes. Z lmmunoL 141:1556-1562.
Tosato, G., T.L. Gerrard, N.G. Goldman, and S.E. Pike. 1988. Stimulation of
EBV-activated human B cells by monocytes and monocyte products. Role of
IFN-13 2/B cell stimulatory factor 2/IL-6. J. lmmunoL t40:4329-4336.
Turksen, K., T. Kupper, L. Degenstein, I. Williams, and E. Fuchs. 1992. Inter-
leukin 6: insights to its function in skin by overexpression in transgenic mice.
Proc. Natl. Acad. Sci. USA. 89:5068-5072.
Walker, C., J.-C. Virchow, P.L.B. Bruijnzeel, and K. Blaser. 1991. T cell subsets
and their soluble products regulate eosinophilia in allergic and non-allergic
asthma. Z lmmunol. 146:1477-1484.
Weinacker, A., A. Chen, M. Agrez, R. Cone, S. Nishimura, E. Wayner, R.
Pytela, and D. Sheppard. 1994. Role of the integrin av136 in cell attachment
to fibronectin: heterologous expression of intact and secreted forms of the
receptor. J. Biol. Chem. 269:6940q5948.
Weinacker, A., R. Ferrando, M. Elliot, J. Hogg, J. Balmes, and D. Sheppard.
1995. Distribution of integrins av136 and a9131 and their known ligands, fi-
bronectin and tenascin, in human airways. Am. J. Respir. Cell Mol. Biol. 12:
Yang, J.T., H. Rayburn, and R.O. Hynes. 1993. Embryonic mesodermal defects
in a5 integrin-deficient mice. Development. 119:1093-1105.
Yang, J.T., H. Rayburn, and R.O. Hymes. 1995. Cell adhesion events mediated
by a4 integrins are essential in placental and cardiac development. Develop-
Yokosaki, Y., E.L. Palmer, A.L. Prieto, K.L. Crossin, M.A. Bourdon, R. Pytela,
and D. Sheppard. 1994. The integrin ~x9131 mediates cell attachment to a
non-RGD site in the third fibronectin type III repeat of tenascin. J. BioL
The Journal of Cell Biology, Volume 133, 1996 928