(Received for publication, August 22, 1996, and in revised form, January 7, 1997)
From the Pulmonary Center and the Department of Biochemistry, Boston University School of Medicine and the Boston Veterans Affairs Medical Center, Boston, Massachusetts 02118
The steady state levels of 1(I) collagen
mRNA are decreased by retinoic acid and prostaglandin
E2. These effector substances decrease the uptake of
A system amino acids. We examined the effect of amino acid deprivation
on the steady state levels of
1(I) collagen in human lung
fibroblasts. Maintenance of fibroblasts in amino acid-free medium
decreased
1(I) collagen mRNA levels by 29% at 24 h and
78% at 72 h. Frequent refeeding of cultures with amino acid-free
medium resulted in more rapid decreases in intracellular amino acids
and in
1(I) collagen mRNA levels. The decrease in
1(I)
collagen mRNA levels was mediated by decreases in mRNA
stability as assessed by a half-life determination using actinomycin D
and by decreases in the rate of transcription as assessed by nuclear
run-on assay. Treatment of fibroblasts with medium containing amino
acids resulted in rapid restoration of
1(I) collagen mRNA
levels. This increase in
1(I) collagen mRNA expression required
protein synthesis as determined by cycloheximide sensitivity and was
inhibited by prostaglandin E2. These data indicate that
1(I) collagen mRNA levels are sensitive to alterations in the
amount of intracellular amino acids and suggest a potential mechanism
whereby
1(I) collagen accumulation may be regulated independent of
inflammatory mediators following lung injury.
The biosynthesis of type I collagen is complex and involves both
intracellular and extracellular sites of regulation. The accumulation
of type I collagen by lung fibroblasts is increased by transforming
growth factor- and insulin and is decreased by prostaglandin
E2 (PGE2),1
retinoic acid, and interferon-
(1-7). We and others previously reported that both PGE2 and retinoic acid induce large
decreases in the expression of
1(I) collagen mRNA by these cells
(2-4). These effects require protein synthesis and are mediated by
decreases in the rate of transcription of the
1(I) collagen gene as
well as by decreases in the stability of the
1(I) collagen
mRNA.
PGE2 and retinoic acid decrease the uptake of neutral amino
acids transported by A system amino acids in lung fibroblasts (7, 8).
The activity of the A system is Na+-dependent
and is responsive to hormones, extracellular amino acids, and other
effector substances (9, 10). The decreases in amino acid transport were
rapid and occurred within 1 h of exposure to the effector
substance. In contrast, the decreases in the steady state levels of
1(I) collagen mRNA occurred between 8 and 12 h following
exposure. This kinetic relation whereby decreases in amino acid uptake
preceded decreases in collagen mRNA levels suggests that depletion
of amino acid levels are associated with decreases in
1(I) mRNA
levels. Certain other mRNAs are regulated by amino acid
availability. For example, asparagine synthetase mRNA is
up-regulated by amino acid starvation, whereas
-actin and
glyceraldehyde 3-phosphate are decreased in abundance (11-13).
In these studies, we examined the relation between amino acid depletion
and 1(I) collagen mRNA levels in human lung fibroblasts. We find
that amino acid depletion caused marked decreases in
1(I) collagen
mRNA levels that rapidly reaccumulate following the addition of
amino acids. This decrease in
1(I) collagen mRNA was mediated by
decreasing the rate of transcription of the
1(I) collagen gene and
by decreasing the stability of the mRNA.
Human embryonic lung fibroblasts (IMR-90, Institute for Medical Research, Camden, NJ) were grown in Dulbecco's modified Eagle's medium with 0.37 g of sodium bicarbonate/100 ml, 10% fetal bovine serum, 100 units/ml penicillin, 10 µg/ml streptomycin, and 0.1 mM nonessential amino acids. After confluence was reached, the serum content of the medium was reduced to 0.4%. The numbers of cells were determined by triplicate cell counts with an electronic particle counter (Coulter Counter ZM).
RNA Isolation and Northern AnalysisTotal cellular RNA was
isolated by the single-step method employing guanidinium
thiocyanate/phenol/chloroform extraction as described by Chomczynski
and Sacchi (14). RNA was quantitated by absorbance at 260 nm. Purity
was determined by absorbance at 280 and 310 nm. RNA (10 µg) was
electrophoresed on a 1% agarose, 6% formaldehyde gel and transferred
to a nitrocellulose filter. RNA loading was assessed by ethidium
bromide staining of ribosomal bands fractionated on
agarose-formaldehyde gels and by co-hybridization with Gs,
which is a constitutively expressed mRNA that codes for a
GTP-binding protein (15). Hybridization was performed using 0.5-1.0 × 106 cpm/lane labeled probe (specific
activity, 4-10 × 108 cpm/µg), and the filter was
washed according to methods described by Thomas (16). The filter was
exposed to x-ray film for autoradiography at several different times to
ensure that the bands could be quantified by densitometry within the
linear range. The loading was verified by probing with an
oligonucleotide that specifically identifies the 18 S ribosomal subunit
(17). The probes utilized in these experiments were a rat cDNA
1(1) collagen clone (18) that specifically identifies corresponding
human
1(1) mRNA, Gs, a GTP-binding protein (kindly
provided by Dr. R. Reed, John Hopkins University School of Medicine)
(15), and cyclooxygenase-1 (19).
Quiescent fibroblast cultures were incubated with fresh medium containing no amino acids at 37 °C for the indicated time periods. Following the incubation, the medium or the cellular material was extracted with 1 ml of 10% trichloroacetic acid. The amount of free amino acids in the supernatant was determined using an automatic amino acid analyzer as described previously (7, 8).
Nuclear Run-on AssayMedium was removed from 150-mm dishes, and the cells were washed twice with Puck's saline and scraped into a Nonidet P-40 lysis buffer. Following two low speed spins, the pellet was reconstituted in a glycerol buffer. In vitro labeling of nascent RNA and hybridization with cDNAs immobilized on nitrocellulose filters were performed according to the methods outlined by Greenberg and Ziff (20) and Groudine et al. (21). No hybridization occurred to filters containing plasmids without inserts.
StatisticsA Student's t test was used for means of unequal size (22). Probability values <0.05 were considered significant.
Quiescent confluent lung fibroblasts were maintained in medium
without amino acids to deplete intracellular amino acids. After varying
periods of time, we determined the steady state levels of 1(I)
collagen mRNA. Gel loading was assessed by ethidium bromide staining and by probing the filter with an oligonucleotide that recognizes the 18 S ribosomal subunit. The levels of
1(I) collagen mRNA were slightly decreased (29 ± 7%, mean ± S.E.,
n = 3) in fibroblasts maintained in amino
acid-deficient medium for 24 h. Maintenance in amino acid-free
medium for 72 h resulted in a large decrease in
1(I) collagen
mRNA levels (Fig. 1). The results of five such
experiments revealed that amino acid deprivation for 72 h
decreased
1(I) collagen mRNA levels by 78% ± 5 (mean ± S.E.).
The addition of the usual complement of 15 amino acids present in
complete medium restored 1(I) collagen mRNA levels, indicating that the absence of added amino acids was not toxic to the cells (Fig.
1). The availability of amino acids markedly affected
1(I) collagen
mRNA levels, but only minimally affected the steady state levels of
Gs mRNA (which encodes a GTP-binding protein) and
cyclooxygenase-1 mRNA (Fig. 2). The amino
acid-mediated induction of
1(I) collagen mRNA levels required
protein synthesis. Cycloheximide alone at 5 µM inhibits
protein synthesis by greater than 90% (23) and does not affect
1(I)
collagen mRNA levels in these cells ((24) and data not shown).
However, the presence of cycloheximide blocked the amino acid-induced
increase in
1(I) collagen mRNA levels (Fig.
3).
We examined the kinetic relation between the addition of amino acids
and the reexpression of 1(I) mRNA. Quiescent fibroblasts were
maintained in medium with or without amino acids for 48 h. Northern analysis was performed on amino acid-deficient cells that were
refed with serum-free medium containing amino acids. The steady state
level of
1(I) collagen mRNA was increased at 8 h
(approximately 4-fold as assessed by densitometry of the two major
1(I) collagen mRNA signals) and further increased at 24 h
following refeeding with medium containing amino acids (Fig. 4).
To determine whether frequent refeeding with amino acid-free medium
increased the depletion of intracellular amino acids by increasing
efflux and preventing reuptake, we determined the amount of free amino
acids in the medium and in the intracellular compartment. The
fibroblasts were placed in amino acid-deficient medium, and the amount
of A system amino acids in the medium was determined after 6 h. We
found that glycine, alanine, and lesser amounts of proline accumulated
in the medium by efflux (Fig. 5). Small amounts of
leucine transported by the L system were also detected in the medium.
The intracellular concentration of the A system amino acids glycine,
proline, and alanine and the L system amino acid leucine was determined
in cells incubated in medium with or without amino acids. The
intracellular levels of the amino acids were variably decreased.
Glycine decreased by 17%, proline by 64%, alanine by 71%, and
leucine by 88% after 12 h (Fig. 6). When the
medium was changed at 4 and 8 h to remove amino acids accumulating
in the extracellular space from efflux (and therefore potentially
available for reuptake), the intracellular levels of glycine, proline,
and alanine decreased further, whereas leucine levels were
unchanged.
We examined the effect of rapidly altering the levels of amino acids by
frequent refeeding with amino acid-deficient medium on 1(I) collagen
mRNA levels. Rapid depletion of intracellular amino acids was
accomplished by refeeding the cells every 5 h (four changes)
during the 24-h period with amino acid-deficient medium. We assessed
the steady state levels for
1(I) collagen and Gs. After
24 h, we found that
1(I) mRNA levels fell by 71% in cells
frequently refed with amino acid-free medium (Fig. 7). In contrast, the level of
1(I) mRNA decreased only 39% in cells cultured in amino acid-deficient medium but not frequently refed.
To examine the mechanism whereby amino acid depletion decreased
collagen mRNA levels, we determined the half-life for 1(I) collagen mRNA and the rate of transcription of the
1(I) gene. We
found that amino acid depletion decreased the stability of the
1(I)
mRNA (Fig. 8). The half-life of the mRNA was
assessed by measuring the decay of the mRNA after the addition of
actinomycin D. Linear regression analysis using the results of three
such experiments employing a variety of time points (2.5, 5, 7, and 9 h) revealed that amino acid depletion decreased the half-life of
the
1(I) mRNA from 9.7 to 4.3 h. Because this decrease in half-life cannot account entirely for the decreases in
1(I) mRNA following amino acid deprivation, we assessed the rate of transcription by a nuclear run-on assay. Nuclei were isolated from fibroblasts maintained in medium with or without amino acids for 48 h. The rate of transcription of the
1(I) gene was reduced by 73% (mean of
two experiments) in nuclei obtained from fibroblasts maintained without
amino acids. In contrast, the rate of transcription for Gs
was unchanged. The addition of amino acids restored the transcription rate for
1(I) collagen (Fig. 9).
We examined the effect of PGE2 on levels of intracellular
amino acids and on 1(I) collagen levels in cells deprived of amino acids. The addition of PGE2 at 10
7
M to cultures maintained in amino acid-free medium for
8 h further decreased the intracellular levels of glycine by 35%
to 75.4 ± 3.5 nmol/106 cells and proline by 20% to
29.5 ± 1.1 nmol/106 cells as compared with cells
maintained in amino acid-free medium. PGE2 at
10
7 M also induced a small additional
decrease in
1(I) mRNA levels in cells maintained in amino
acid-deficient medium (Fig. 10). When fibroblasts were
reexposed to amino acids, PGE2 inhibited the restoration of
1(I) mRNA levels by 76 ± 7% (mean ± S.E.,
n = 3).
We found that the steady state level of collagen 1(I) mRNA
was decreased by treatment of fibroblasts with amino acid-deficient medium. In contrast, changes in amino acid availability minimally affected the levels of the constitutively expressed cyclooxygenase-1 and Gs mRNAs. The decrease in
1(I) mRNA levels
was dependent on the time of exposure to amino acid-deficient medium
and the rate of decrease in intracellular amino acids. Short exposures to amino acid-deficient medium resulted in a 29% decrease in mRNA levels, and prolonged exposure resulted in a 78% decrease. When the
intracellular levels of amino acids were rapidly lowered by frequent
refeeding with amino acid-deficient medium, the
1(I) mRNA levels
decreased more markedly. Frequent refeeding decreased intracellular
amino acids by removing amino acids appearing in the medium by efflux
and preventing reuptake. The decrease in
1(I) collagen mRNA
levels was not the result of toxicity because restoring amino acids to
the medium resulted in a rapid restoration of mRNA levels.
Alteration of intracellular amino acid levels affects the steady state
levels of several other mRNAs (9-11, 25). The steady state
mRNA level for asparagine synthetase is up-regulated by amino acid
derivation in baby hamster kidney cells (11, 12). A novel cis-acting
element (5-CATGATG-3
) located in the proximal asparagine promoter
mediates this effect and binds nuclear proteins (26). However, analysis
of the promoter region of the
1(I) collagen gene did not reveal a
similar binding site. Other examples of amino acid-regulated mRNAs
include Cu-Zn superoxide dismutase, glyceraldehyde 3-phosphate, and
histone H2. These mRNAs are decreased by amino acid starvation in
rat Fao hepatoma cells (13). In addition, insulin-like growth factor I
mRNA levels in hepatocytes is decreased by amino acid deprivation
and increased by excess (27).
Certain effector substances regulate gene expression by altering
intracellular amino acid levels. Interferon- decreases expression of
collagenase and stromelysin by decreasing the intracellular concentration of tryptophan (25, 28). Interferon-
also decreases production of
1(I) collagen mRNA (29). However, we found that the addition of tryptophan to amino acid-deficient medium did not
restore
1(I) mRNA levels2
(unpublished results) suggesting that one or more other amino acids are
necessary to induce this affect.
Depletion of amino acids decreased the rate of transcription of the
1(I) collagen gene and the stability of the mRNA. This process
was relatively selective because the rate of transcription of
Gs was unchanged by decreased amino acid availability. The molecular mechanism whereby amino acid depletion causes alterations in
gene expression and mRNA stability is unknown. The sensing mechanism may involve changes in the level of aminoacylation of tRNA.
For example, the rate of protein breakdown in histidine starvation of
the Chinese hamster ovary cell is regulated by the levels of
aminoacylation of tRNA (30). Studies employing
temperature-dependent mutant cell lines suggest that
regulation of asparagine synthetase activity may involve tRNA acylation
(31, 32). This regulatory mechanism may also affect mRNA levels of
selected genes, possibly by altering the production of a
transcriptional activator under circumstances of amino acid
deprivation. Our results employing cycloheximide also suggest that
synthesis of one or more specific proteins (perhaps a transcription
factor) is initially required for activation of
1(I) collagen
mRNA levels by amino acid exposure. We find that cycloheximide
inhibited the induction of
1(I) collagen mRNA following the
addition of amino acids to amino acid-deficient cells.
Our previous results showed that PGE2 and retinoic acid
decreased the transport of A system amino acids by increasing the Km of the A system transporter (7, 8). Both of these
effectors decrease the rate of transcription of the 1(I) collagen
gene. Our results employing amino acid-deficient medium suggest that
the PGE2-induced decreases in
1(I) collagen mRNA levels may be mediated through decreases in intracellular amino acids.
PGE2 further decreased
1(I) collagen levels when
fibroblasts were cultured in amino acid-deficient medium, perhaps by
inhibited reuptake of amino acids accumulating in the medium by efflux. In addition, PGE2 blocked the restoration of
1(I)
mRNA levels following reexposure to amino acids. We previously
found that PGE2 was more effective in inhibiting
1(I)
collagen accumulation induced by transforming growth factor-
than
that induced by insulin (33). Because insulin is a strong inducer of
amino acid transport, it may inhibit the effect of PGE2 by
directly stimulating uptake.
Alterations in amino acid availability may affect 1(I) collagen
mRNA expression by fibroblasts residing within a biomatrix and
in vivo. The steady state levels of
1(I) collagen
mRNA decreases in fibroblasts that are placed within collagen gels,
perhaps as a result of matrix-fibroblast interactions that decrease the
availability of nutrients by alterations in transporter function
(34-36). It is noteworthy in this regard that suspension of 3T3
fibroblasts in methylcellulose decreases the rate of transcription of
the
1(I) collagen gene and decreases the stability of the
1(I)
collagen mRNA (37). In the lung, fibroblasts that reside in the
pulmonary interstitium do not express detectable levels of
1(I)
mRNA as assessed by in situ hybridization (38).
Following lung injury,
1(I) collagen levels rapidly increase,
presumably from exposure to fibrogenic cytokines such as transforming
growth factor-
. However, alteration in vascular permeability and
proteolytic disruption of the extracellular matrix may provide greater
access for amino acids into the interstitium, resulting in fibroblast
activation.