| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
-Induced MHC Class II Expression in Fibroblasts1
,
* Ontario Cancer Institute, Toronto, Canada;
Departments of Immunology, Medicine, and Medical Biophysics, University of Toronto, Toronto, Canada; and
St. Michael’s Hospital, Toronto, Canada
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
signaling. We used mouse embryo
fibroblasts derived from SOCS1-/-
mice to investigate the role of SOCS1 in IFN- signaling pathways.
SOCS1-/- fibroblasts were
exquisitely sensitive to the IFN--mediated growth arrest and showed
sustained STAT1 phosphorylation. However,
SOCS1-/- fibroblasts were
inefficient in MHC class II surface expression following IFN-
stimulation, despite a marked induction of the MHC class II
transactivator and MHC class II gene expression. Retroviral
transduction of wild-type SOCS1 relieved the growth-inhibitory effects
of IFN- in SOCS1-/- fibroblasts
by inhibiting STAT1 activation. SOCS1R105K, carrying a mutation within
the phosphotyrosine-binding pocket of the Src homology 2 domain, did
not inhibit STAT1 phosphorylation, yet considerably inhibited
IFN--mediated growth arrest. Strikingly, expression of SOCS1R105K
restored the IFN--induced MHC class II expression in
SOCS1-/- cells, indicating that
expression of SOCS1 facilitates MHC class II expression in fibroblasts.
Our results show that SOCS1, in addition to its negative regulatory
role of inhibiting Janus kinases, has an unanticipated positive
regulatory function in retarding the degradation of IFN--induced MHC
class II proteins in fibroblasts. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MHC class II molecules expressed on restricted cell types present
peptide Ags to CD4+ Th cells. Professional APCs
constitutively express MHC class II, while in other cell types its
expression is induced by certain cytokines. IFN- not only elevates
MHC class II expression in professional APCs, but also induces de novo
expression in monocyte/macrophage lineage, fibroblasts, endothelial
cells, epithelial cells, neuronal cells, and muscle cells (see Ref.
6 for a review). Induction of MHC class II expression by
IFN- is regulated primarily at the transcriptional level and is
dependent on STAT1 (7, 8). IFN- stimulates expression
of the transcription factors MHC class II transactivator (CIITA) and
IFN response factors, IRF1-IRF4 (9). CIITA is obligatory
for MHC class II expression (10), and IRF-1 is required
for MHC class I up-regulation (11). CIITA does not bind to
DNA directly, but functions to orchestrate the assembly of multiple
transcription factors such as the RFX complex, X2BP/CREB, and NF-Y to
the S, X, X2, and Y boxes at the promoter region of MHC class II genes
(7, 12). Additional promoter-binding factors such as Oct-2
modulate cell- and stage-specific expression of MHC class II
(13).
Binding of IFN- to IFN-R1 and IFN-R2 stimulates the
receptor-associated Janus kinases 2 (JAK2) and 1 that phosphorylate
STAT1. Binding of phospho-STAT1 homodimers to IFN- activation site
elements in the promoter regions of IFN--responsive genes induces
their transcription and subsequent pleiotropic cellular responses,
including growth arrest, antiviral activity, differentiation, and MHC
class II and class I expression (see Ref. 14 for a
review). IFN- strongly induces SOCS1 (15, 16), and
SOCS1-/- mice die within 2–3 wk after
birth from fatty degeneration of liver and monocytic infiltration
of visceral organs (17). This phenotype is suppressed in
the IFN--deficient background, suggesting that SOCS1 is the major
intracellular regulator of IFN- signaling (2, 18).
All members of the SOCS family have a central Src homology 2 (SH2) domain, a conserved carboxyl-terminal motif called the SOCS box, and a variable amino-terminal segment. SOCS1 inhibits cytokine signaling primarily by inhibiting the catalytic activity of all four members of the JAK family tyrosine kinases JAK1–3 and Tyk-2, thereby blocking the critical JAK-STAT pathway in cytokine receptor signaling (19, 20, 21, 22, 23). Structure-function analyses of the SOCS1-JAK interaction have revealed that the SH2 domain is critical for JAK binding, as mutation of the Arg105 residue within the phosphotyrosine-binding pocket abrogates JAK binding (23, 24). Recent studies have suggested that the SOCS box could also participate in the suppression of cytokine signaling. Similar to the von Hippel-Lindau tumor suppressor protein, SOCS1, via the SOCS box, binds to ElonginBC within a multiprotein E3 ubiquitin ligase complex containing Cullin2, Roc1/Rbx, and an E2 ubiquitin-conjugating enzyme (25, 26, 27, 28). This interaction targets the SOCS1-interacting proteins such as Vav and Tel-JAK2 to ubiquitination and subsequent proteasomal degradation (29, 30, 31, 32).
Overexpression of SOCS1 completely blocks IFN--mediated growth
arrest, antiviral activity, differentiation, and MHC class II
expression through inhibition of JAK1 and JAK2 activation, and STAT1
phosphorylation (15, 17, 33, 34). However, endogenous
SOCS1, expressed within 1 h after IFN- stimulation
(15), does not seem to interfere with the multitude of
IFN--induced cellular responses that span several hours following
IFN- treatment. This raises an interesting possibility that SOCS1
may actually be mediating some of the temporally distant cellular
effects of IFN-, in addition to attenuating the early signaling
event. To address this issue in a defined genetic background, we
established fibroblast lines (mouse embryonic fibroblasts; MEFs)
from SOCS1-/- embryos, reconstituted
them with SOCS1 expression constructs, and evaluated the effects of
IFN- in terms of growth arrest and MHC class II expression. Our
results show that SOCS1 plays an agonistic function in IFN--induced
MHC class II expression in fibroblasts.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Breeding stocks of SOCS1+/- and
SOCS1+/- IFN--/- mice were
kind gifts from J. Ihle (St. Jude’s Children’s Research
Hospital), and the SOCS1-/- mice
were bred in our animal facility. Murine rIFN- was a gift from E.
Fish (University of Toronto, Toronto, Canada) and was also
purchased from R&D Systems (Minneapolis, MN). NIH3T3 cell line
expressing murine CSF-1 was a gift from P. Greer (Queen’s University).
pBMN- internal ribosomal entry site (IRES)-Lyt-2 originated in G.
Nolan’s laboratory (Stanford University, Stanford, CA). pBMN-green
fluorescent protein (GFP) retroviral vector expressing human c-Myc was
a gift from L. Penn (University of Toronto). PE-conjugated M5/114.15.2
(M5) mAb that recognizes the murine MHC class II molecules I-A and I-E
of H-2b and H-2d
haplotypes, biotinylated anti-MHC class I
(H-2Db), and APC-conjugated Mac1 were purchased
from BD PharMingen (San Diego, CA). Phosphoantibodies to STAT1, p38,
extracellular signal-regulated kinase, and c-Jun N-terminal kinase were
from New England Biolabs (Beverly, MA). Anti-upstream stimulatory
factor-1 (USF-1) was from Santa Cruz Biotechnology (Santa Cruz, CA).
Streptavidin-spectral red was from Southern Biotechnology Associates
(Birmingham, AL).
Cell lines
Mouse embryos at 14–15 days of gestation were obtained after euthanasia, eviscerated, minced, and digested with trypsin. Single cell suspensions were plated in DMEM containing 10% FBS and 50 µM 2-ME to culture fibroblasts. The monolayer cultures were continually passaged every 3–4 days, and the cells were used for experiments after 10 passages. Some of the SOCS1-/-, SOCS1+/-, and SOCS1+/+ fibroblast lines were immortalized by continuous passage. To culture macrophages, single cell suspensions prepared from the fetal liver were stimulated with CSF-1 (added as culture supernatant) every 3–4 days. The cultures were maintained in OPTI-MEM (Life Technologies, Burlington, Ontario, Canada) containing 10% FBS and 50 µM 2-ME. When the cultures became homogeneous and cleared of debris, the cells were infected with c-Myc-expressing retrovirus to establish permanent lines. The human c-Myc oncoprotein was expressed from a bicistronic vector with GFP expressed from an IRES. After 4 wk, the cultures contained mostly GFP+Mac1+ adherent macrophages.
Plasmids and retroviral infections
N-terminal hemagglutinin (HA)-tagged murine SOCS1 DNA fragment
was amplified from pMT3-SOCS1 and subcloned in the XhoI and
SalI sites of pMIEV retroviral vector expressing GFP as a
selection marker from a bicistronic promoter (24) (see
Fig. 4
A). SOCS1R105K carrying a point mutation (CGC changed
to AAG) within the phosphotyrosine-binding pocket of the SH2 domain was
generated by overlapping PCR (24) and subcloned into
pMIEV. SOCS1
CT lacking the C-terminal SOCS box spanning the last 38
aa was generated by digesting pMIEV/SOCS1 with NotI and
ligation. The retroviral SOCS1 constructs were stably transfected into
GP + E retrovirus packaging line, and fresh culture supernatants from
the virus-producing GP + E cells were used to infect embryonic
fibroblasts in the presence of 2 µg/ml polybrene (Sigma-Aldrich, St.
Louis, MO) for 2 h. After two to three consecutive infections, the
cells were expanded and sorted for GFP-positive cells using FACStar (BD
Biosciences, Mississauga, Ontario, Canada). Supernatants from
the sorted GP + E cells were used subsequently to infect
SOCS1-/- fibroblasts.
|
- and 
-chains of I-Ab were
cloned by RT-PCR from C57BL/6 splenocytes into pCDNA3.1 in frame with
V5 and 6xHis epitope tags. I-Ab was
additionally tagged with HA epitope. Protein expression was confirmed
in COS-7 cells by transient transfection and Western blot detection of
HA and His tags. The constructs were subcloned into pBMN-IRES-Lyt2
retroviral vector (Fig. 8C). Stable GP + E virus producer
cell lines were established by sorting CD8+
cells. Culture supernatants from GP + E cells expressing
I-Ab were used to infect fibroblast lines;
cells expressing CD8 were sorted, expanded, and subsequently infected
with I-Ab-expressing retrovirus. Cells
constitutively expressing the MHC class II molecule
I-Ab on surface were sorted after staining
with M5 Ab.
|
stimulation, growth kinetics, and flow cytometry
Fibroblasts and macrophages at 
80% confluency were
stimulated with IFN- in the culture medium for indicated periods.
For growth curve and MHC class II expression, 0.5–1.0 x
105 cells were plated in 24- or 12-well tissue
culture plates in 2 or 3 ml vol and cultured for 48–96 h, in the
presence or absence of IFN-. At the end of culture period, cells
were trypsinized and counted by a Corixa (Seattle, WA) cell
counter for growth analysis. Alternatively, cell growth was monitored
by the Alamar Blue dye reduction assay (Biosource International,
Camillo, CA). A total of 5 x 103 cells were
plated per well in 96-well plates in the presence of various
concentrations of IFN-. At the end of a 48-h incubation period, 20
µl of Alamar Blue was added to each well, and the plates were
incubated further for 6–9 h. The plates were read in a microplate
reader at 570 and 600 nm, and the amount of reduced dye (proportional
to the number of cells and their metabolic activity) was calculated by
following manufacturer’s instructions. To evaluate the surface density
of MHC class II molecules following IFN- stimulation, 1–2
x 105 cells were stained with PE-conjugated M5
mAb that recognizes I-A and I-E of b and d haplotypes. MHC class I
molecules were labeled with biotinylated
anti-H-2Db, followed by streptavidin-spectral
red. Stained cells were analyzed by FACScan (BD Biosciences). To stain
intracellular MHC class II, surface molecules were blocked by 100
µg/ml of unlabeled M5/114.15.2 for 15 min on ice, followed by
PE-labeled M5/114.15.2 in PBS/5% FBS containing 0.3% saponin. After
20-min incubation at room temperature, cells were washed in PBS/5% FBS
containing 0.1% saponin and resuspended in PBS/5% FBS for FACS
analysis.
Western blot and RT-PCR
For biochemical analysis of signaling proteins, 5 x
105 cells were plated in 60-mm culture dishes and
stimulated with IFN- at 80% confluent stage. At indicated time
points, the cells were lysed in 100 µl of SDS lysis buffer (1% SDS,
50 mM Tris, pH 7.0, 1 mM EDTA, 1 mM DTT), scraped into Eppendorf tubes,
and boiled. Lysates containing equivalent amounts of total proteins
were electrophoresed in SDS-PAGE gels, transferred to polyvinylidene
difluoride membranes, and probed with phospho-STAT1 Ab following the
manufacturer’s instructions. The blots were developed by Renaissance
ECL reagent from New England Nuclear (Boston, MA) life science
products. The blots were stripped by incubating in stripping solution
(2% SDS, 62.5 mM Tris, pH 6.8, 100 mM 2-ME) for 30 min at 55°C,
blocked, and reprobed for STAT1. A similar protocol was followed for
other phosphospecific Abs.
For RT-PCR analysis of IFN--induced gene expression, 5 x
105 cells were plated in 60-mm culture dishes and
stimulated with IFN- at 80% confluent stage for 48 h. Total
RNA was isolated using 1 ml of TRIzol (Life Technologies) following the
manufacturer’s instructions. Equivalent amounts of RNA were primed
with oligo(dT), and first strand cDNA was synthesized using
Thermoscript reverse transcriptase (Life Technologies), according to
the manufacturer’s protocol. PCR was performed on 10-fold serially
diluted cDNA using Platinum Taq DNA polymerase (Life
Technologies) in a GeneAmp PCR System 9700 thermocycler (PE Applied
Biosystems, Foster City, CA). The PCR primers and product sizes are as
follows: I-Ab, 5'-GGCATTTCGTGTACCAGTTC-3'
(sense) and 5'-AGGTCCTTTCTGACTCCTGT-3' (antisense), 664 bp;
I-Eb, 5'-GGAGAGTTGAGCCTACGGTGA-3' (sense)
and 5'-AGACTGTCCTTTCTGGTTCCT-3' (antisense), 406 bp; I-A,
5'-TCCTGGAGACATTGGCCAGT-3' (sense) and 5'-TCTCAGGTTCCCAGTGTTTC-3'
(antisense), 499 bp; mouse CIITA, 5'-GCAGCTACCTGGAACTCCTTA-3' (sense)
and 5'-CTCATTTACACGGGAGGTCAG-3' (antisense), 685 bp; mouse
invariant chain (Ii), 5'-GTGTCTGTTTCATCGTCCCAG-3' (sense) and
5'-AAGGCAGCAAATGTGTCCAGC-3' (antisense), 205 bp; mouse -actin,
5'-GGCATTGTTACCAACTGGGAC-3' (sense) and 5'-ACCAGAGGCATACAGGGACAG-3'
(antisense), 219 bp. For actin, CIITA, I-A,
I-Ab, and I-Eb, the
following PCR conditions were used: predenaturation at 94°C for
60 s, followed by 30 cycles of denaturation at 94°C for 30
s, annealing at 67°C for 30 s, and elongation at 72°C for
60 s. Ii fragment was amplified by predenaturation at 94°C for
60 s, followed by 30 cycles of denaturation at 94°C for 30
s, annealing at 65°C for 30 s, and elongation at 72°C for
30 s.
Metabolic labeling of MHC class II molecules
Subconfluent cultures of fibroblast lines grown in 60-mm dishes
were stimulated with 1000 U/ml IFN-. After 36 h, cells were
washed and incubated in Cys- and Met-free DMEM (ICN, Montreal,
Quebec, Canada) for 30 min. The medium was replaced with Cys, Met-free
DMEM containing 2.5% dialyzed calf serum and 200 µCi/ml Trans
35S-label (ICN; sp. act. 1175 Ci/mmol) to label
newly synthesized proteins. After 60-min pulse, the cells were washed
in PBS and lysed in buffer containing 50 mM Tris, pH 8, 150 mM NaCl,
1% Nonidet P-40, 6 mM
3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, 50 µM
PMSF, 50 µg/ml
N-tosyl-lysine-chloromethylketone, and 10 µg/ml
aprotinin (35). The lysates were precleared sequentially
with rabbit Ig + pansorbin (Calbiochem, La Jolla, CA), pansorbin, and
protein A-Sepharose (Amersham Biosciences, Uppsala, Sweden).
Purified M5 mAb (10 µg/sample) was added to the precleared lysate,
and MHC class II moecules were immunoprecipitated with protein-A
Sepharose precoated with rabbit anti-rat Ig at 4°C overnight.
Immunoprecipitated proteins were separated by SDS-PAGE and detected by
autoradiography.
Measurement of protein turnover by cycloheximide treatment
Subconfluent cultures of fibroblast lines grown in 60-mm dishes
were stimulated with 1000 U/ml IFN- or left untreated. After 36
h, the cells were treated with 10 µM cycloheximide (CHX; Calbiochem).
At indicated time points, the cells were rinsed in PBS and lysed in SDS
lysis buffer. Amounts of I-Ab and
I-Ab proteins were estimated by Western blot
using anti-His Ab. The relative amounts of the MHC class II -
and -chains remaining at different time points after blockade of
protein synthesis were estimated using a laser-scanning densitometer
(Molecular Dynamics, Sunnyvale, CA). The amount of proteins present in
cells that did not receive CHX was used as reference to compare the
decline after CHX treatment.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-mediated growth arrest
IFN- induces cell growth arrest in a number of cell types,
including macrophages, fibroblasts, and T cells (reviewed in Refs.
14 and 36). The growth-inhibitory property of
IFN- is mediated through the induction of CDK inhibitors
p27KIP1 and p21CIP1/WAF1 as
well as by general inhibition of transcription and translation
(36, 37, 38, 39). Because IFN--induced growth arrest requires
STAT1 activation by JAKs (37, 40), we investigated the
effect of IFN- on cells lacking SOCS1, the negative regulator of JAK
family kinases. MEFs established from
SOCS1-/- embryos were highly susceptible
to growth arrest by IFN- (Fig. 1A). IFN- inhibited the
growth of SOCS1-/- MEFs by >80%
compared with 30% inhibition in wild-type cells after 96 h in
culture. The SOCS1+/- MEFs showed an
intermediate phenotype with 50% growth inhibition. To assess whether
SOCS1 deficiency increases the sensitivity to IFN-, we used the
Alamar Blue dye reduction assay that reflects both the number and
metabolic activity of the cells. Increased susceptibility of
SOCS1-/- fibroblasts to IFN--mediated
growth arrest was evident over a wide concentration range of IFN-
after a 48-h culture period (Fig. 1B). These results showed
that SOCS1-/- fibroblasts are highly
susceptible to IFN--mediated growth arrest compared with wild-type
cells.
|
-mediated growth arrest (Fig. 2A, top).
Strikingly, phospho-STAT1 level decreased within 3 h following
IFN- stimulation in wild-type fibroblasts, while it did not decrease
in SOCS1-/- MEFs even at 24 h after
stimulation. The STAT1 protein level increased 6–24 h after IFN-
stimulation, as reported earlier (15). This increase in
STAT1 protein levels was comparable between wild-type and
SOCS1-/- fibroblasts. Densitometric
quantitation of the phospho-STAT1 and STAT1 levels showed that the
ratio of phospho-STAT1 to STAT1 decreased more slowly in
SOCS1-/- fibroblasts compared with
wild-type cells (Fig. 2A, bottom). The elevated
phospho-STAT1 level persisted longer in
SOCS1-/- fibroblasts even after IFN-
withdrawal, up to 4 h following 30 min of IFN- stimulation,
while it was substantially decreased in wild-type cells (see Fig. 5B). However, absence of SOCS1 did not lower the threshold
level of IFN- required to stimulate STAT1phosphorylation, because
both SOCS1-/- and wild-type fibroblasts
showed identical IFN- dose kinetics (Fig. 2B). This
sustained intracellular signaling in
SOCS1-/- fibroblasts is specific to
IFN- stimulation, as TNF--stimulated activation of
mitogen-activated protein kinase family proteins was comparable between
SOCS1-/- and
SOCS1+/+ fibroblasts (Fig. 2C).
|
|
stimulation is not sufficient to induce MHC class II
expression
Because IFN--induced MHC class II expression is dependent on
STAT1 activation (8, 41), we investigated whether the
sustained STAT1 phosphorylation in
SOCS1-/- fibroblasts resulted in
elevated MHC class II levels. Surprisingly, we observed that in
SOCS1-/- fibroblasts, MHC class II
induction by IFN- was substantially decreased compared with the
SOCS1+/+ or
SOCS1+/- cells (Fig. 3A). Although 45% of the
wild-type or SOCS1+/- cells expressed MHC class
II molecules after 48–60 h of IFN- stimulation, only 14% of the
cells were positive in two different
SOCS1-/- fibroblast lines. The staining
for MHC class II molecules is specific, and was not due to the binding
of anti-MHC class II mAb (M5) to FcR because preincubation
of cells with either a nonspecific mAb of the same isotype, or Abs
blocking the FcR did not diminish the M5 mAb binding. The
responsiveness of SOCS1-/- fibroblasts
to IFN- was defective only in MHC class II expression, as the
induction of MHC class I molecules was identical with that of wild-type
cells (Fig. 3B). The defective MHC class II expression in
SOCS1-/- fibroblasts following IFN-
stimulation was cell type specific because constitutive expression of
MHC class II molecules in splenic B lymphocytes was comparable between
SOCS1-/- and
SOCS1+/- mice in
IFN--/- background (Fig. 3C).
Similarly, c-Myc-transformed Mac1+
macrophage lines established from
SOCS1-/- mice responded equally well as
the SOCS1+/- macrophages (Fig. 3D).
These results showed that SOCS1-/-
fibroblasts have a cell type-specific defect in inducible MHC class II
expression, despite a sustained and elevated STAT1 activation.
|
-induced growth
arrest
The requirement of SOCS1 for IFN--induced MHC class II
expression on fibroblasts suggested that SOCS1 might have an agonistic
function in IFN- signaling independent of its role as a JAK
inhibitor. To verify this notion, we generated stable cell lines
expressing wild-type SOCS1, SOCS1 with R105K point mutation within the
SH2 domain (SOCS1R105K), or SOCS1 lacking the C-terminal SOCS box
(SOCS1CT) by retroviral transduction and GFP selection (Fig. 4
, A and B). The
steady state levels of the ectopically expressed SOCS1 proteins were
not affected by IFN- (Fig. 4C). However, SOCS1R105K was
expressed at a much lower level compared with wild-type SOCS1 or
SOCS1CT. This decrease is unlikely due to reduced transcription or
mRNA stability because GFP expressed from the same transcript is
comparable between the SOCS1R105K and SOCS1CT constructs (Fig. 4B). Because SOCS1 protein is rapidly turned over
(42), we tested whether the low steady state level of
SOCS1R105K resulted from increased susceptibility to ubiquitin-mediated
proteasomal degradation. Proteasomal protease inhibitors LLnL and
MG-132 caused an accumulation of wild-type SOCS1, whereas SOCS1CT
was unaffected by either of the inhibitors (Fig. 4D). LLnL
had no effect on SOCS1R105K expression, whereas MG-132 caused a partial
stabilization. These results suggest that: 1) the C-terminal SOCS box
regulates the turnover of SOCS1, and 2) the decreased expression of the
R105K mutant protein cannot be fully attributed to proteasomal
degradation.
Analysis of the IFN- responses of
SOCS1-/- MEFs reconstituted with the
SOCS1 constructs showed that constitutive expression of the wild-type
SOCS1 completely blocked IFN--mediated growth arrest (Fig. 5A). These results are
consistent with the inhibitory effects of SOCS1 on JAK2
(23), and are in agreement with earlier reports on the
effect of SOCS1 in IFN--stimulated wild-type fibroblasts (15, 34). Although the parental and vector-transduced
SOCS1-/- fibroblasts showed >80%
inhibition of their growth in the presence of IFN-,
SOCS1-/- cells expressing wild-type
SOCS1 were not inhibited at all, as a consequence of the total
abrogation of IFN--stimulated STAT1 activation in these cells (Fig. 5B). As anticipated, forced expression of SOCS1R105K did not
inhibit STAT1 phosphorylation, but surprisingly rendered the
SOCS1-/- cells less sensitive to
IFN--induced growth arrest (Fig. 5A). These results
suggest that in addition to the JAK-STAT pathway, other unknown
IFN--stimulated pathway(s) is involved in mediating the growth
arrest. Furthermore, SOCS1 blocks the IFN--mediated growth arrest
not only by blocking the JAK-STAT pathway, but also by blocking the
additional pathway(s). Although the SH2 domain of SOCS1 is required for
blocking the JAK-STAT pathway, it does not seem to be necessary for
blocking this additional pathway(s).
SOCS1 is required for inducible MHC class II expression in fibroblasts
Failure of SOCS1-/- fibroblasts to
up-regulate MHC class II molecules after IFN- stimulation was
specific to MHC class II molecules, as the induction of MHC class I
molecules by IFN- was comparable between wild-type and
SOCS1-/- cells (see Fig. 3B).
In SOCS1-/- cells, constitutive
expression of wild-type SOCS1 completely abolished IFN--stimulated
expression of both MHC class I and class II molecules (Fig. 6). This could be attributed to the
inhibition of JAK activity and subsequent STAT1 phosphorylation by the
constitutively expressed SOCS1 (Fig. 5B), because STAT1
phosphorylation is required for both MHC class II and class I
expression. Surprisingly, SOCS1R105K restored MHC class II expression
in SOCS1-/- fibroblasts and did not
affect MHC class I induction (Fig. 6). Failure of
SOCS1-/- cells to up-regulate MHC class
II molecules despite sustained STAT1 activation, and its reversal by
SOCS1R105K strongly suggest an essential role for SOCS1 in MHC class II
expression that does not require the phosphotyrosine-binding capacity
of the SOCS1 SH2 domain.
|
Up-regulation of MHC class II surface expression by IFN- is
regulated primarily at the transcriptional level (43) and
requires, in addition to STAT1 activation, up-regulation of CIITA and
presence of the constitutively expressed USF-1 (41). To
understand the defect in SOCS1-/-
fibroblasts that underlies the inefficient MHC class II expression, we
examined the levels of IFN--induced gene expression by
semiquantitative RT-PCR (Fig. 7A). Both
I-Ab and I-Eb MHC
class II genes were induced in SOCS1-/-
fibroblasts to levels identical with the wild-type cells. Similarly,
induction of CIITA gene was comparable between
SOCS1-/- and wild-type fibroblasts. In
addition, there was no discernible alteration in the levels of a USF-1
in SOCS1-/- fibroblasts (Fig. 7B). Up-regulation of PA28a, a component of the 19S
regulatory subunit of the proteasome, served as a control for IFN-
activity (Fig. 7B). These results showed that assembly of
the multicomponent transcriptional machinery is normal at the MHC class
II promoter in SOCS1-/- cells.
Constitutive expression of SOCS1 in the
SOCS1-/- cells completely blocked the
expression of CIITA, with concomitant lack of transcripts for
I-Ab and I-Eb (Fig. 7A), emphasizing the importance of the JAK-STAT pathway in
MHC class II gene induction. In contrast, in
SOCS1-/- fibroblasts reconstituted with
SOCS1R105K, which permitted MHC class II up-regulation, IFN--induced
CIITA, I-Ab, and I-Eb
expression were not significantly higher than the parental cells (Fig. 7A). These results show that the defect in MHC class II
induction in SOCS1-/- cells occurs at
the posttranscriptional level.
|
Cell surface expression of the newly synthesized MHC class II -
and -chains requires their association with Ii for transport to the
cell surface (44). In the absence of Ii, the MHC class II
molecules accumulate within the cell (45). Because the
expression of Ii is also up-regulated by IFN-, we examined the
induction of Ii following IFN- stimulation. RT-PCR analysis showed
that the IFN--stimulated increase in Ii transcript was comparable
between SOCS1-/- and
SOCS1+/+ cells (Fig. 7A). Surprisingly, IFN--induced Ii up-regulation was not
inhibited by constitutively expressed SOCS1, suggesting that the
induction of Ii following IFN- stimulation occurs independently of
the JAK-STAT pathway. To evaluate whether the MHC class II proteins
accumulate intracellularly due to a defect in transport to the cell
surface, we stained the IFN--stimulated cells for intracellular MHC
class II molecules after blocking the surface molecules with unlabeled
Ab. Unlabeled blocking Ab completely inhibited surface staining of MHC
class II (Fig. 8), whereas a nonspecific
Ab of the same isotype did not (see Fig. 3A). Intracellular
staining following surface Ag blocking showed intense staining in
IFN--stimulated wild-type cells and
SOCS1-/- fibroblasts reconstituted with
SOCS1R105K mutant, whereas parental
SOCS1-/- fibroblasts showed no evidence
for intracellular accumulation of MHC class II proteins. These results
show that the defective MHC class II expression in
SOCS1-/- fibroblasts is not due to a
defect in transport to the cell surface.
IFN- stimulates degradation of MHC class II proteins in
SOCS1-deficient fibroblasts
Our results suggested that the markedly reduced MHC class II
up-regulation in IFN--stimulated
SOCS1-/- cells might result either from
inefficient translation of MHC class II transcripts or from rapid
turnover of these proteins in the absence of SOCS1, or both. Metabolic
labeling of IFN--treated SOCS1-/- and
control fibroblasts showed very small amounts of MHC class II proteins
in SOCS1-deficient cells (Fig. 8B), too little to carry out
a meaningful pulse-chase analysis for protein stability. To overcome
this obstacle, we generated SOCS1-/- and
SOCS1+/+ fibroblasts that constitutively
expressed both - and -chains of I-Ab and
transported the I-Ab heterodimer to the
cell surface (Fig. 8, C and D). The steady state
levels of - and -chains in
SOCS1-/- fibroblasts were slightly lower
than in wild-type cells (Fig. 8E). However, examination of
protein stability using CHX, a general inhibitor of protein
translation, did not reveal any appreciable difference in protein
turnover (Fig. 8F, upper panels), indicating that
the difference in steady state levels could be due to variability in
the copy number of the transgene. Strikingly, following 36-h IFN-
stimulation, SOCS1-/- fibroblasts showed
a faster turnover of both - and -chains of
I-Ab than wild-type cells (Fig. 8E,
bottom and right panels). These results show that
MHC class II proteins are less stable in IFN--stimulated
SOCS1-/- fibroblasts.
SOCS box is not required to block IFN--mediated growth arrest
and MHC class II expression
SOCS1 interferes with IFN- signaling by inhibiting the JAK
kinases through the binding of SOCS1 SH2 domain to the activation loop
phosphotyrosine residue of the JAKs (23). The C-terminal
end of the SOCS1 molecule, the SOCS box, is implicated in targeting the
bound Tel-JAK2 to the VHL-Elongin C-Elongin B complex ubiquitin
ligase (30, 31, 32). Ensuing ubiquitination targets the bound
JAKs to the proteasomal degradation machinery. To determine whether the
SH2 domain and the SOCS box should function coordinately to attenuate
IFN- signaling, we introduced SOCS1 with a truncation deleting the
entire SOCS box (SOCS1CT) into
SOCS1-/- cells (Fig. 4A).
Analysis of IFN- responses of the
SOCS1-/- fibroblasts reconstituted with
SOCS1CT showed that SOCS box-deleted SOCS1 was equally potent in
blocking the IFN--mediated growth arrest as the wild-type SOCS1, and
efficiently blocked IFN--stimulated STAT1 phosphorylation (Fig. 5).
These results showed that the SOCS box is dispensable for suppressing
the growth-inhibitory activity of IFN-. Moreover, SOCS1CT did not
block the up-regulation of CIITA, I-Ab, and
I-Eb gene expression (Fig. 7A),
suggesting that inhibition of IFN- signaling by SOCS1CT may not
be complete without the SOCS box-mediated capture of activated JAKs and
probably other downstream signaling molecules for ubiquitination and
proteasomal degradation.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
stimulation (15). Overexpression of SOCS1 inhibits both
JAK1 and JAK2, and completely blocks IFN- responses (15, 17, 33, 34, 46). Therefore, in the absence of SOCS1, we expected an
exacerbation of the IFN- responses in
SOCS1-/- fibroblasts. Indeed, we
observed an increased sensitivity of the
SOCS1-/- fibroblasts to IFN--mediated
growth arrest (Fig. 1). However, contrary to our expectation,
IFN--stimulated MHC class II surface expression was markedly
diminished in SOCS1-/- fibroblasts. Both
growth inhibition and induction of CIITA transcription factor required
for MHC class II expression by IFN- require activation of STAT1
(40, 41). The sustained STAT1 phosphorylation in
SOCS1-/- fibroblasts correlated well
with their increased susceptibility to IFN--induced growth arrest,
but not with MHC class II surface expression, indicating that: 1) STAT1
activation alone is not sufficient, and 2) SOCS1 is required for
inducible MHC class II surface expression in fibroblasts.
Reconstitution of the SOCS1-/-
fibroblasts with mutant SOCS1 constructs confirmed these predictions,
and revealed that SOCS1 does not require its phosphotyrosine-binding
capacity to facilitate inducible MHC class II expression in
fibroblasts.
MHC class II expression is regulated at multiple levels. Both
constitutive and inducible MHC class II expression require the critical
transcription factors CIITA, RFX5, and USF-1 for the transcription of
MHC class II genes (7). Inherited mutations within these
genes are characterized by the lack of MHC class II expression and
defective T cell development leading to the Bare lymphocyte syndrome
(6). A number of reports have documented other
pathological and physiological conditions that result in defective MHC
class II gene transcription. Suppression of CIITA expression by human
CMV underlies its inhibition of IFN--induced MHC class II expression
(47). Chlamydiae interferes with MHC class II expression
by targeting USF-1 and RFX5 to proteasomal degradation (48, 49). However, we did not find any change in the induction of
CIITA message or in the level of USF-1 protein in
SOCS1-/- fibroblasts compared with
wild-type cells (Fig. 7). Fibroblasts deficient in retinoblastoma
protein pRB show defective I-A gene expression despite normal CIITA
induction (50). MHC class II gene expression has been also
reported to be actively repressed by PU.1 (51), and by an
IFN--induced transcriptional repressor p202 that interacts with pRB
(52, 53). Recently, the inducible MHC class II expression
in murine macrophages has been shown to be transcriptionally regulated
in a cell cycle-dependent manner (54). The fact that
SOCS1-/- fibroblasts showed normal
induction of I-Ab and
I-Eb genes indicates that assembly of the
complex transcriptional machinery at the promoter region of MHC class
II genes is not impaired in SOCS1-/-
fibroblasts. Therefore, the decreased surface expression of MHC class
II in SOCS1-/- fibroblasts must arise
from a posttranscriptional defect. Lack of intracellular accumulation
of MHC class II within the SOCS1-/-
fibroblasts favors a hypothesis that the defect might arise from
alterations in mRNA stability, translation initiation, and progression,
or the t1/2 of the protein.
Recent studies show that MHC class II protein expression is
additionally regulated at the translational level (55, 56). For example, up-regulation of the - and -chains of
I-A in a B cell line (A20) following IFN- stimulation is not
accompanied by an increase in mRNA levels or mRNA stability, even
though there is a 3-fold increase in surface expression
(56). This increase in surface expression is accompanied
by a marked increase in the amount of mRNA bound to polyribosomes.
Neither the factors nor the mechanism that contribute to more efficient
translation of I-A and I-A mRNA following IFN- stimulation
have been defined yet, but these observations clearly indicate that
translational regulation is another potential control point in MHC
class II expression and SOCS1 may be involved in translational
activation or translational derepression. One of the well-defined
translational activation pathways involves the phosphatidylinositol-3
kinase (PI3K)-Akt/PKB-mTOR pathway, leading to phosphorylation of 4E-BP
and release of eIF-4E to form an active translation initiation complex
at the 5' cap of the mRNA (57). Nguyen et al.
(58) have recently shown activation of the PI3K pathway
and phosphorylation of Akt in murine fibroblasts following IFN-
stimulation. We examined whether MHC class II expression was inhibited
by Wortmannin added at different time points after IFN- stimulation,
and we could not find evidence for a role of PI3K-dependent
translational activation in MHC class II surface expression (data not
shown).
We have used constitutively expressed - and -chains of
I-Ab to address whether MHC class II proteins are
degraded in SOCS1-/- fibroblasts. We
show that MHC class II proteins are degraded in IFN--stimulated
SOCS1-/- fibroblasts with a faster
kinetic than in control cells. These results strongly suggest that
IFN- stimulation not only induces MHC class II expression, but also
initiates a protein degradation program. In normal cells, this
degradation process appears to be delayed by the concomitantly induced
SOCS1, so that the MHC class II proteins could carry out their Ag
presentation function for longer duration. IFN- induces significant
changes to the proteasomal protein degradation machinery to generate
immunogenic peptides for MHC class I Ag presentation (59).
We are investigating whether MHC class II proteins are degraded by the
proteasomal pathway. A recent report showing transcriptional activation
of several genes by IFN- independently of STAT1 (60)
raises the possibility that additional IFN--induced pathways, which
are probably regulated by SOCS1, may be involved in the coordinated
regulation of MHC class II surface expression.
Constitutive expression of wild-type SOCS1 in
SOCS1-/- fibroblasts completely blocked
IFN--induced growth arrest and MHC class II and class I gene
expression by inhibiting the JAKs required for STAT1 phosphorylation.
The JAK proteins bound to SOCS1 have been shown to be targeted for
ubiquitin-dependent proteasomal degradation, a process that requires an
intact SOCS box (30, 31). However, similar to the
observations of Kamizono et al. (31), the degradation
function of SOCS1 is not essential for blocking IFN- responses, as
the SOCS1CT was as efficient as the wild-type SOCS1 in inhibiting
IFN--mediated growth arrest (Fig. 5A). Our results also
suggest that the SOCS box is involved in the regulation of SOCS1
protein levels, because wild-type SOCS1, but not SOCS1CT, was
stabilized by proteasomal inhibitors (Fig. 4D). The SH2
domain may also be involved in regulating SOCS1 protein levels, as
accumulation of SOCS1R105K in the presence of proteasome inhibitors was
not comparable with that of wild-type SOCS1. Because SOCS1 contains
only one lysine residue, and a conserved point mutation of this lysine
does not alter the t1/2 of SOCS1 (our
unpublished results), molecular dissection of the SOCS1 protein is
required to understand the mechanisms underlying its rapid
turnover.
Consistent with earlier reports (23), SOCS1R105K failed to
inhibit STAT1 phosphorylation in
SOCS1-/- cells. Despite a low level of
expression, SOCS1R105K in SOCS1-/- cells
has revealed two unexpected roles for SOCS1 in IFN- signaling.
First, despite its inability to block STAT1 phosphorylation, SOCS1R105K
inhibited IFN--induced growth arrest (Fig. 5A),
suggesting that SOCS1R105K mediates this effect downstream of STAT1
phosphorylation. Second, SOCS1R105K restored the MHC class II
expression in SOCS1-/- fibroblasts,
which failed to up-regulate MHC class II despite prolonged STAT1
phosphorylation (Fig. 6), suggesting that a certain critical level of
SOCS1 expression is required for MHC class II expression in
IFN--stimulated fibroblasts. In both instances, SOCS1 did not
require the phosphotyrosine-binding capacity of its SH2 domain to
mediate the novel functions. Several SH2 domains have recently been
shown to bind to their physiological protein partners in a
phosphotyrosine-independent manner (61, 62). In such a
scenario, proteins bound to the mutant SH2 domain of SOCS1 may still be
targeted for ubiquitination and proteasomal degradation via the
E3-ligase substrate adaptor function conferred by the SOCS
box.
In conclusion, we have observed that IFN--induced growth arrest and
MHC class II induction occur through overlapping, but distinct
signaling pathways. Although persistent JAK activation and STAT1
phosphorylation in SOCS1-deficient fibroblasts are sufficient to cause
growth arrest, they are insufficient to up-regulate MHC class II
expression. The requirement for SOCS1 for MHC class II expression
suggests that SOCS1 functions not only as a negative feedback regulator
of the JAK kinases, but also as a positive regulator in some aspects of
cytokine signaling. One such positive regulatory function appears to be
prolongation of the t1/2 of
IFN--induced MHC class II expression.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: Institut National de la Santé de la Recherche Médicale Unité 19, 13009 Marseille, France.
3 Address correspondence and reprint requests to Dr. Robert Rottapel Experimental Therapeutics, Ontario Cancer Institute, Princess Margaret Hospital, Toronto M5G 2A1, Canada. E-mail address: rottapel{at}uhnres.utoronto.ca
4 Abbreviations used in this paper: SOCS1, suppressor of cytokine signaling 1; CHX, cycloheximide; CIITA, class II transactivator; GFP, green fluorescent protein; HA, hemagglutinin; Ii, invariant chain; IRES, internal ribosomal entry site; IRF, IFN response factor; JAK, Janus kinase; MEF, mouse embryonic fibroblast; PI3K, phosphatidylinositol-3 kinase; SH2, Src homology 2; USF, upstream stimulatory factor.
Received for publication December 12, 2001. Accepted for publication September 4, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Annu. Rev. Immunol. 15:749.[Medline]
-inducible gene and confers resistance to interferons. Blood 92:1668. signaling and prevents the potentially fatal neonatal actions of this cytokine. Cell 98:597.[Medline]
