|
|
||||||||


* Division of Pulmonary and Critical Care Medicine, University of Miami School of Medicine, Miami, Florida 33136, USA;
Division of Neonatology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, USA; and
Department of Cell Biology and Anatomy, University of Miami School of Medicine, Miami, Florida 33136, USA
1Correspondence: Division of Pulmonary and Critical Care Medicine (R-47), University of Miami School of Medicine, 1600 NW 10th Ave., RMSB 7063, Miami, FL 33136, USA. E-mail: msalathe{at}miami.edu
| ABSTRACT |
|---|
|
|
|---|
Key Words: kallikrein lactoperoxidase cilia epithelium airway
| INTRODUCTION |
|---|
|
|
|---|
We have been studying two key enzymes of airway mucosal defense, both
of which are secreted onto the airway surface: tissue kallikrein (TK),
which is involved in allergic bronchoconstriction to limit the
inhalation of noxious substances (2)
, and airway
lactoperoxidase (LPO), which increases bacterial clearance from ovine
airways (3)
. TK is produced in submucosal gland cells
(4
5
6)
. LPO is produced in both surface epithelial goblet
cells and submucosal gland cells (7
8
9)
. We have
previously shown that TK from broncho-alveolar lavage is bound to
secreted hyaluronan that regulates its activity (4)
. We
have also shown that ciliated cells in culture bind endogenously
produced hyaluronan and that exogenously applied hyaluronan can
stimulate ciliary beat frequency (CBF) (10)
.
Hyaluronan is a glycosaminoglycan made of repeating disaccharides
formed from glucuronic acid and N-acetylglucosamine, linked through
alternating beta-1,4 and beta-1,3 glycosidic bonds (11)
.
In recent years, evidence has accumulated that hyaluronan serves
important biological roles beyond its generally accepted function as a
structural component of interstitial and connective tissues. For
instance, hyaluronan has been shown to promote wound healing
(12)
, tumor transformation, and metastasis (13
, 14)
, as well as cell motility and migration
(15
16
17
18)
. Many of these biological activities are mediated
by either CD44 or the receptor for hyaluronic acid-mediated motility
(RHAMM or CD168) (15)
. RHAMM is expressed in a variety of
mammalian cell types including fibroblasts, smooth muscle cells,
endothelial cells, macrophages, sperm, nerve cells, and malignant
tumors (19)
, but no data on RHAMM in the airways are
available. Since the ultrastructure of sperm tails is similar to the
one of cilia in the tracheobronchial tree, data showing that hyaluronan
addition to media was critical to preserve sperm motility and able to
increase sperm velocity in vitro (20
, 21)
through its
interaction with RHAMM (22)
hinted that RHAMM could have
functional relevance in the airways. We therefore conducted a series of
experiments to examine the role of hyaluronan in airway host defense,
specifically as it relates to TK, LPO, and ciliary beating.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Preparation of tracheal epithelial cultures
Primary cultures of tracheal epithelial cells were prepared as
described previously (25)
.
Immunohisto- and cytochemistry
Sheep trachea and cell cultures were fixed with acid formalin
(26)
and processed according to standard procedures for
immunohisto- and immunocytochemistry. Primary antibodies were used at
the following dilutions: anti-RHAMM antibody (10 µg/ml); anti-TK
(1:500); and anti-LPO (1:500). Preimmune serum was diluted 1:500.
Secondary antibodies were alkaline phosphatase or horseradish
peroxidase-labeled, affinity-purified goat anti-rabbit IgG used at 5
µg/ml in 50 mM Tris-buffered saline (Kirkegaard and Perry,
Gaithersburg, MD). Color was developed with nitro blue tetrazolium and
5-bromo-4-chloro-3-indolyl-phosphate or diaminobenzidine, respectively.
Hyaluronan was visualized according to Bray et al. (27)
using a biotinylated hyaluronan binding protein (Seikagaku, Tokyo,
Japan) and avidin coupled with alkaline phosphatase. Hyaluronidase
digestion (50 IU/ml from Streptomyces hyalurolyticus;
Seikagaku) was performed in 10 mM potassium-acetate, pH 5.6, containing
pepstatin 0.1 µg/ml, aprotinin and leupeptin 0.1 µM, at 37°C
overnight. Chondroitinase ABC (ICN, Costa Mesa, CA) was used as a
control at 0.5 IU/ml in 50 mM Tris-buffered saline containing pepstatin
(0.1 µg/ml), aprotinin and leupeptin (0.1 µM), at pH 7.5 and 37°C
overnight. The chondroitinase ABC units used represent a 10-fold excess
over those used for experiments digesting chondroitin sulfate produced
by submucosal gland cells (28)
. Chondroitinase ABC was
also used in 10 mM potassium-acetate, pH 5.6, at 0.5 IU/ml.
PCR
PCR was done with our ovine mucosa cDNA library (3)
as template and a specific 5' oligonucleotide (TGTTGAATGAACATGGTGCA) as
well as a sixfold degenerate 3' primer (CCTTDGAAGGRTCAAAGTG), both
designed from consensus RHAMM sequences. The
FailSafeTM PCR system (Epicentre Technologies,
Madison, WI) was used successfully with all buffers provided, annealing
at 52°C for a total of 30 cycles.
Measurement of CBF
CBF was measured and analyzed as described previously using
continuous perfusion at room temperature (29)
. To reveal
the CBF response to exogenously applied hyaluronan, endogenous
hyaluronan was removed from cells with 5 IU/ml hyaluronidase (from
bovis testis; Worthington, Freehold, NJ) in buffered Hanks balanced
salt solution, 10 mM HEPES, pH 7.4 (referred to as HBSS) for 2 h
at room temperature (10)
. For blocking experiments, cells
were incubated for 30 min with anti-RHAMM or control rabbit
anti-chicken IgG antibodies (both at 25 µg/ml in HBSS). This
concentration of anti-RHAMM antibody inhibited locomotion of smooth
muscle cells (24)
. The cells were then exposed within 60
min to hyaluronan (Seikagaku or Worthington; see Results; both with an
average molecular mass of 200 kDa) at 50 µg/ml, previously
demonstrated to increase CBF (10)
. As a control, the cells
were exposed to 20 µM ATP to ascertain their response to a known
stimulator of CBF. Exchanges of the chamber volume with HBSS alone had
no influence on CBF at the perfusion rates used, ruling out that
changes in CBF were due to mechanical disturbance of cilia.
Tracheal transport measurements
Recombinant TK (gift kindly provided by Amgen Pharmaceuticals,
Thousand Oaks, CA), purified bovine milk LPO, and bovine serum albumin
(Sigma, St. Louis, MO) were labeled with fluorescein or rhodamine
isothiocyanate (Sigma) according to published methods
(30)
. The products were purified on Sephadex G50 and
concentrated to 1 mg/ml in PBS. The coupling efficiency was assessed to
be optimal with less than 2 mol fluorochrome per mole of protein by the
ratio of absorbance at 495/280 nm for fluorescein (measured to be 0.9)
and 575/280 nm for rhodamine (measured to be 0.5). Labeled enzyme (LPO
or TK) and albumin were applied in equimolar amounts to the mucosal
surface of a trachea in a total volume of 10 µl. Tracheas were
obtained from freshly killed sheep, opened by cutting through the
membranous portion, and kept in a humidified chamber at 37°C.
Movement of the applied fluorescent substances was monitored using a
broad-spectrum UV illuminator and a digital camera every 10 min for a
total of 30 min.
| RESULTS |
|---|
|
|
|---|
|
TK and LPO interact with hyaluronan
Using a nondenaturing gel system and affinity chromatography, we
previously showed that TK binds to hyaluronan (4)
. Since
TK has an acid pI, nonspecific electrostatic interactions are not
expected to account for this interaction. We therefore searched for and
found the putative hyaluronan binding motif
B(X7)B (32)
in the amino acid
sequence of TK, where B is a basic residue and X is any amino acid not
negatively charged. Since this binding motif contains a portion of the
active TK site (Fig. 2
), it may provide a basis for specific interactions between hyaluronan
and TK. Using nondenaturing agarose gel electrophoresis, bovine milk
LPO was also found to bind to hyaluronan (200 kDa average molecular
mass, Seikagaku) after incubation times of 60 min (not shown). The
binding could be reversed by hyaluronan digestion with hyaluronidase
(20 IU/ml for 2 h at 37°C). Analysis of the LPO amino acid
sequence, however, did not reveal the presence of the
B(X7)B hyaluronan binding motif. Since LPO has an
alkaline pI, nonspecific electrostatic interactions could account for
its association with hyaluronan.
|
Although binding to hyaluronan inhibits TK activity (4)
,
measurements of LPO activity using the 3,3',5,5'-tetramethylbenzidine
assay (9)
revealed that LPO was not inhibited by addition
of hyaluronan (200 kDa average) in molar ratios of 5:1, 10:1, 50:1, and
100:1 (hyaluronan:LPO). Relative LPO activity after 60 min of
incubation with hyaluronan compared to LPO alone (control) was 95 ± 3%, 92 ± 5%, 99 ± 0.5%, and 92 ± 6%
(mean±SE) of control, respectively, not reaching
statistical significance at any ratio using ANOVA (n=4 at
each ratio measured in triplicates). Hyaluronan alone did not affect
the 3,3',5,5'-tetramethylbenzidine assay of LPOs activity.
RHAMM is expressed on the apical airway epithelial surface
Since hyaluronan appeared to retain LPO and TK at the surface, we
examined whether hyaluronan binding receptors are expressed at the
apical surface of the epithelium. Previous reports indicated that CD44,
a common extracellular hyaluronan receptor, is found on basal cells but
not on the apex of normal, ciliated airway epithelial cells
(33)
. Although this does not rule out its expression on
the apical surface of ovine airway epithelial cells, it makes it less
likely, and we concentrated on the expression of RHAMM in ovine
trachea. Immunohistochemistry revealed specific staining for RHAMM in
the apical portion of ciliated cells, but no staining in goblet cells
(Fig. 3
). To confirm expression of RHAMM in tracheal epithelial cells, we used
an ovine tracheal mucosa cDNA library (3)
and primers for
RHAMM (Fig. 3)
designed according to consensus regions of the published
sequences. PCR reactions yielded a band of expected size (249 bp) and
the fragment was sequenced (Fig. 3
; GenBank accession no. BankIt354819
AF310973). The deduced amino acid sequence was 89% and 81% identical
to the published human and mouse sequences, respectively. Together,
these data show that RHAMM is expressed in and localized to the apical
portion of polarized ciliated airway epithelial cells.
|
CBF is regulated by hyaluronan-mediated signaling through RHAMM
To examine whether our previously reported hyaluronan-mediated
increase in CBF (10)
was mediated through RHAMM, we used
primary cultures of ovine airway epithelial cells (25)
.
Using anti-RHAMM antibody and fixed, nonpermeabilized cells, expression
of RHAMM could be shown to occur on the ciliated surface of cultured
cells (Fig. 3)
. These results were confirmed by adding anti-RHAMM
antibody to live cells before fixation. The expression of RHAMM
increased during the time in culture (18% of all ciliated cells
stained positive on day 3 after plating, 57% on day 5, 65% on day 8,
and 76% on day 11). This increase in expression correlated with our
reported increase in the percentage of ciliated cells in culture
staining positive for surface hyaluronan (stemming from submucosal
gland cells usually found in primary airway epithelial cell cultures;
10
) as well as the increase in the percentage of ciliated
cells in culture responding to exogenous hyaluronan with an increase in
CBF (Fig. 4
; ref 10
). This correlation is significant, since other
responses of ovine ciliated airway epithelial cells, such as CBF
increases in response to ATP or acetylcholine, can be elicited
immediately after plating, are maintained over a 1420 day period, and
do not depend on cell growth or adaptation to culture (25
, 29
, 34)
.
|
We next made sure that the CBF stimulating effect of hyaluronan did not
depend on the commercial source used. Hyaluronan from two suppliers
(Seikagaku and Worthington) yielded the same CBF stimulation when
measured in six cells each over 10 days in culture (
CBF of
1.8±0.6 Hz for Seikagaku and
CBF of 1.2±0.4 Hz for Worthington,
P>0.5). Hyaluronan digested with hyaluronidase, however,
had no effect on CBF.
Then we tested whether the hyaluronan-induced CBF increase could be
inhibited with our functionally blocking anti-RHAMM antibody. At room
temperature, none of 10 cells (>10 days in culture) preincubated with
anti-RHAMM antibody (24)
responded with a CBF change
(baseline 7.4±0.6 Hz; Fig. 5
). However, six of eight cells exposed to a nonspecific, control rabbit
anti-chicken IgG responded to 50 µg/ml hyaluronan with an increase in
CBF from 7.2 ± 0.6 to 9.1 ± 0.4 Hz (P<0.05;
Fig. 5
). This CBF increase in the presence of control antibody was
similar to responses measured without any antibody (10)
.
The percentage of responding cells also corresponded to the percentage
of RHAMM-expressing ciliated cells. Finally, control responses to 20
µM ATP, a well-known stimulator of CBF (35
36
37)
, were
statistically indistinguishable between both groups (
CBF in the
anti-RHAMM group was 2.5±0.5 Hz and in the anti-IgG control group
2.7±0.5 Hz, P=0.45) and indistinguishable from cells not
exposed to antibodies.
|
Tracheal transport of TK, LPO, and albumin
To examine whether hyaluronan actually protected these enzymes
from removal by mucociliary clearance, we labeled recombinant TK with
fluorescein and both LPO and albumin with rhodamine. Labeled TK and
albumin were applied together (as a mixture) onto the same region of
the surface tracheal epithelium and the migration of the fluorescence
was measured over a 30 min period. TK was not transported after
application whereas albumin moved forward over the entire 30 min
period. Thus, the two substances moved apart, as indicated by a
separation of the original orange fluorescence (mixture) into clearly
defined green (TK) and red (albumin) bands (Fig. 6
). Movement of albumin was measured to range between 1 and 5 cm (dye
front) at the end of 30 min, whereas TK did not move (n=3).
To show that LPO behaved similarly to TK and that fluorescein
modification was not responsible for TKs immobilization, we used
rhodamine-labeled LPO with the same result (not shown). The
immobilization of the enzymes was in fact due to hyaluronan, since TK
and albumin did not separate on tracheas pretreated with hyaluronidase,
moving at the same rate over the 30 min observation period (Fig. 6)
. These data show that airway LPO and TK are both bound to the airway
epithelial surface by hyaluronan and are not transported away by
mucociliary clearance as labeled albumin is.
|
| DISCUSSION |
|---|
|
|
|---|
Proteins can bind to hyaluronan by several mechanisms (32
, 39
, 40)
. We have shown here that the deduced amino acid sequence of
TK contains the putative hyaluronan binding motif
B(X7)B (32)
inthe active site of
TK (Fig. 2)
. We have also shown that LPO binds to hyaluronan, but that
the LPO amino acid sequence does not contain the
B(X7)B binding motif. LPO has an alkaline pI
(
10) that by itself allows nonspecific electrostatic interactions
with hyaluronan. In fact, hyaluronan may act in general as a cation
exchanger (39)
and therefore may be able to bind several
other cationic proteins present in the airway to the epithelial
surface. Among those could be a variety of cationic antimicrobial
substancesfor example, those studied by Cole et al.
(41)
.
In addition to binding TK to the epithelial surface, hyaluronan also
inhibits its activity (4)
. This is important because TK
can lead to bronchoconstriction, which is useful only during exposure
to certain stimuli. Airway LPO, on the other hand, should be active at
all times, because it contributes to host defense against inhaled
bacteria (3)
. In fact, measurements of airway LPO activity
in vitro presented here revealed that the enzyme did not change its
activity when bound to hyaluronan over a wide range of molar ratios.
Thus, hyaluronan perhaps serves as a broad anchor for many enzymes at
the airway surface, either through electrostatic interactions or
through specific binding. Note that the specific interaction occurs
with TK, an enzyme that needs to be inhibited while present at the
airway surface. On the other hand, electrostatic interactions seem to
be important for enzymes immobilized in their active form. In this
sense, it is remarkable that many antimicrobial substances in nasal
secretions and in airway secretions are cationic in nature
(41)
.
The data showing RHAMM expression at the apical surface suggest a
mechanism for hyaluronan binding to the airway epithelial surface.
Whereas the presence of RHAMM is certainly important, we cannot rule
out that other hyaluronan anchoring proteins are present at the apical
surface including CD44, although it has been reported not to be
expressed there in the normal, undamaged epithelium (33)
.
Whether or not RHAMM is the sole anchor for hyaluronan, the experiments
using anti-RHAMM antibody reveal a clear functional role of RHAMM in
the airway epithelium: upon hyaluronan binding, RHAMM signals to
increase CBF. Although it would be interesting to know how RHAMM
increases CBF without having a transmembrane signaling component, such
experiments go beyond the scope of this paper, especially since the few
signaling pathways known to be activated by RHAMM (e.g., ERK
stimulation) have not been reported to regulate CBF and are usually
slower than the observed reactions of CBF to hyaluronan.
RHAMM expression on the surface was initially lost and slowly
reappeared on the cell surface in culture. This phenomenon could be due
to initial removal of RHAMM from these cells by proteolytic action
during cell isolation. The time course of RHAMM reappearance correlated
with hyaluronan binding to these cells as well as CBF responses to
exogenously applied hyaluronan. RHAMM expression in fact predicted CBF
responses to hyaluronan. The correlation between RHAMM expression,
hyaluronan binding, and CBF increases over time in culture is
significant since CBF responses to other stimuli such as ATP and
acetylcholine do not show such a time course. These responses can be
elicited right after plating and are not lost during a 1420 day
culture period. These data therefore show that hyaluronan signals
through RHAMM to increase cilia beating on airway epithelial cells
similar to the effect of hyaluronan on sperm flagellar bending
(22)
.
Hyaluronan is most often thought to be a component of the extracellular
matrix and until now, the airways have not been an exception.
Therefore, an increase in the amount of soluble hyaluronan in bronchial
lavage fluid has been interpreted as a marker of interstitial damage to
the lung (42
43
44)
. In the context of chronic airway
diseases, where increased amounts of soluble hyaluronan could be
measured in bronchial lavage fluid (e.g., ref 45
),
hyaluronan has been interpreted as a sign of tissue remodeling.
These interpretations are surprising since hyaluronan has been known to
be secreted from submucosal gland cells, specifically from serous cells
(46)
. Therefore, and as shown here, hyaluronan is
secreted into the airways under normal conditions and must have
specific functions in the normal airway. At least part of this function
is to immobilize TK and LPO on the epithelial surface. In diseased
airways, altered amounts of secreted or surface-released hyaluronan
could be responsible for changes seen in the airway in asthma and other
inflammatory conditions.
In summary, we propose a model in which hyaluronan serves a previously
unrecognized pivotal role in mucosal host defense (Fig. 7
). Hyaluronan stimulates ciliary beating (through its interaction with
RHAMM) and hence the clearance of foreign material from mucosal
surfaces, but simultaneously retains and regulates enzymes important
for homeostasis at the apical mucosal surface. Therefore, the common
belief that constitutive and stimulated secretion onto the mucosal
surface is the major determinant of enzyme availability has to be
revisited. The data shown here suggest a new paradigm that involves an
apical enzyme pool ready for use and protected from ciliary
clearance. This pool will have to be considered in enzymatic reactions
at the mucosal surface, be it in health or disease.
|
| ACKNOWLEDGMENTS |
|---|
Received for publication March 21, 2001.
Revision received June 12, 2001.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. E. Monzon, D. Manzanares, N. Schmid, S. M. Casalino-Matsuda, and R. M. Forteza Hyaluronidase Expression and Activity Is Regulated by Pro-Inflammatory Cytokines in Human Airway Epithelial Cells Am. J. Respir. Cell Mol. Biol., September 1, 2008; 39(3): 289 - 295. [Abstract] [Full Text] [PDF] |
||||
![]() |
F. Gao, J. R. Koenitzer, J. M. Tobolewski, D. Jiang, J. Liang, P. W. Noble, and T. D. Oury Extracellular Superoxide Dismutase Inhibits Inflammation by Preventing Oxidative Fragmentation of Hyaluronan J. Biol. Chem., March 7, 2008; 283(10): 6058 - 6066. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Manzanares, M.-E. Monzon, R. C. Savani, and M. Salathe Apical Oxidative Hyaluronan Degradation Stimulates Airway Ciliary Beating via RHAMM and RON Am. J. Respir. Cell Mol. Biol., August 1, 2007; 37(2): 160 - 168. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Scuri, J. R. Sabater, and W. M. Abraham Hyaluronan blocks porcine pancreatic elastase-induced mucociliary dysfunction in allergic sheep J Appl Physiol, June 1, 2007; 102(6): 2324 - 2331. [Abstract] [Full Text] [PDF] |
||||
![]() |
V. Vigdorovich, A. D. Miller, and R. K. Strong Ability of Hyaluronidase 2 To Degrade Extracellular Hyaluronan Is Not Required for Its Function as a Receptor for Jaagsiekte Sheep Retrovirus J. Virol., April 1, 2007; 81(7): 3124 - 3129. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Forteza, S. M. Casalino-Matsuda, M. E. Monzon, E. Fries, M. S. Rugg, C. M. Milner, and A. J. Day TSG-6 Potentiates the Antitissue Kallikrein Activity of Inter-{alpha}-inhibitor through Bikunin Release Am. J. Respir. Cell Mol. Biol., January 1, 2007; 36(1): 20 - 31. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. M. Casalino-Matsuda, M. E. Monzon, and R. M. Forteza Epidermal Growth Factor Receptor Activation by Epidermal Growth Factor Mediates Oxidant-Induced Goblet Cell Metaplasia in Human Airway Epithelium Am. J. Respir. Cell Mol. Biol., May 1, 2006; 34(5): 581 - 591. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. E. Monzon, S. M. Casalino-Matsuda, and R. M. Forteza Identification of Glycosaminoglycans in Human Airway Secretions Am. J. Respir. Cell Mol. Biol., February 1, 2006; 34(2): 135 - 141. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. S. Wilkinson, S. Potter-Perigo, C. Tsoi, L. C. Altman, and T. N. Wight Pro- and Anti-Inflammatory Factors Cooperate to Control Hyaluronan Synthesis in Lung Fibroblasts Am. J. Respir. Cell Mol. Biol., July 1, 2004; 31(1): 92 - 99. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. M. Casalino-Matsuda, M. E. Monzon, G. E. Conner, M. Salathe, and R. M. Forteza Role of Hyaluronan and Reactive Oxygen Species in Tissue Kallikrein-mediated Epidermal Growth Factor Receptor Activation in Human Airways J. Biol. Chem., May 14, 2004; 279(20): 21606 - 21616. [Abstract] [Full Text] [PDF] |
||||
![]() |
I. T. Lauredo, R. M. Forteza, Y. Botvinnikova, and W. M. Abraham Leukocytic cell sources of airway tissue kallikrein Am J Physiol Lung Cell Mol Physiol, April 1, 2004; 286(4): L734 - L740. [Abstract] [Full Text] [PDF] |
||||
![]() |
S.-H. Leir, S. T. Holgate, and P. M. Lackie Inflammatory cytokines can enhance CD44-mediated airway epithelial cell adhesion independently of CD44 expression Am J Physiol Lung Cell Mol Physiol, December 1, 2003; 285(6): L1305 - L1311. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Schoenfelder and R. Einspanier Expression of Hyaluronan Synthases and Corresponding Hyaluronan Receptors Is Differentially Regulated During Oocyte Maturation in Cattle Biol Reprod, July 1, 2003; 69(1): 269 - 277. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. M. Turino and J. O. Cantor Hyaluronan in Respiratory Injury and Repair Am. J. Respir. Crit. Care Med., May 1, 2003; 167(9): 1169 - 1175. [Full Text] [PDF] |
||||
![]() |
M. Stojkovic, O. Krebs, S. Kolle, K. Prelle, V. Assmann, V. Zakhartchenko, F. Sinowatz, and E. Wolf Developmental Regulation of Hyaluronan-Binding Protein (RHAMM/IHABP) Expression in Early Bovine Embryos Biol Reprod, January 1, 2003; 68(1): 60 - 66. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||