migration of T cells by eicosanoids. DAVID LEPPERT,* STEPHEN L. HAUSER,* JEFFREY L. KISHIYAMA, SONGZHU AN, U ZENG, AND. EDWARD J. GOETZL'.
I
RESEARCH
Stimulation migration
COMMUNICATIONSI
of matrix metalloproteinase-dependent of T cells by eicosanoids
DAVID LEPPERT,* STEPHEN L. HAUSER,* JEFFREY L. KISHIYAMA, EDWARD J. GOETZL’ Departments Center,
ABSTRACT
otriene
B4
of Neurology,*
Medicine
and
Microbiology-Immunology,
San Francisco, California 94143-0711,
concentrations, elicited migration of human blood T cells and cultured T lymphoblastoma cells of the Tsup- 1 line through a layer of Matrigel basement membrane matrix. The density of Tsup-1 cell highaffinity receptors was low for PGE2 and high for LTB4, resulting in respectively predominant chemokinetic and chemotactic stimulation of migration. Migration-enhancing concentrations of PGE2 and LTB4 also increased Tsup-1 cell content and secretion of matrix metalloproteinases (MMPs) 2, 3, and 9, which were quantified by Western blots and zymography, and augmented Tsup- 1 cell-surface expression of the MMPs, as shown by flow cytometry. That a specific MMP inhibitor suppressed migration of blood T cells and Tsup-1 cells through Matrigel, but did not affect PGE2- and LTB4-initiated T cell migration through micropore filters without Matrigel, suggests dual requirements for MMP expression and enhanced motility in T cell passage through basement membranes.-Leppert, D., Hauser, S. L., Kishlyama, J. L., An, S., Zeng, L., Goetzl, E. J. Stimulation of matrix metalloproteinase-dependent migration of T cells by eicosanoids. FASEB J. 9, 1473-1481 (1995) acid
protease
‘ ‘ connective
tissue
im-
Prostaglandin E2 (PGE2),2 leukotriene B4 (LTB4), and some other eicosanoid products of the oxygenation of arachidonic acid are generated in immune tissues, principally by macrophages, mast cells, and B cells, and have potent effects on T cell functions (1-5). These eicosanoid mediators affect T cell proliferation, expression of membrane receptors (Rs), and secretion of diverse cytokines (5). PGE2 preferentially inhibits proliferative and cytokine secretory responses of CD4+ T cells, which exhibit helper functions. The subset selectivity of PGE2 effects on cytokine generation by CD44 T cells is exemplified by reductions in the secretion of interleukin-2 (IL-2) and interferon-y(IFN-y) by Thi cellsand of IL-3 and GM-CSF by Th 1 and Th2 cells,without effectson IL-4 or IL-5 secretionby Th2 cells(6).The immunosuppressive
0892-6638/95/0009-1
473/$01
.50. © FASEB
University
of California
AN, U ZENG, AND Medical
USA
Prostaglandin E2 (PGE2) and leuk(LTB4), at nanomolar to micromolar
Key Words: arachidonic munity
The
SONGZHU
activities
ofLTB4 appear tobe attributable toconcurrentinhibition of proliferation of CD4+ cells and enhancement of proliferation of CD8cells of the suppressor phenotype (7).The most prominent effects of LTB4 on I cell cytokines are increased secretion of IL-2 and IFN-’y by helper T cells, with lesser inhibitionofgenerationoflFN-ybysuppressorTcells(8,9).T cells express distinct sets of C-protein-associated receptors for eicosanoid mediators that evoke functional responses, of which the EP2R for PGE2 and the LTB4R are especially prominent ( 10-13). PGE2 suppresses lymph node entry and egress of lymphocytes by unknown mechanisms (14, 15). Studies in in vitro models of nonlymphoid tissues have documented diverse effects of PGE2 and LTB4 on human T cell migration, which have included chemokinesis
PGE2 suppression evoked by casein
of chemotaxis (16), elicitation
to C5a and of of chemoki-
nesis by products of the cyclo-oxygenation of arachidonic
acid (17),LTB4 stimulationof chemotaxis (18),and PGE2 inhibition of migration through monolayers of human umbilical cord venous endothelial cells on nitrocellulose filters by actions on both T cells and endothelial cells (19). The interpretation of some previous results has been complicated by the heterogeneity of responding T cells and a frequent requirement for high concentrations of PGE2 that act predominantly by elevating the cellular concentration of cyclic AMP and have complex effects on cells other than lymphocytesinthemodel systems. The spontaneous and IL-2-enhanced migration of human blood T cells across a layer of Matrigel, which consists of basement membrane matrix constituents, has recently been shown to depend on the activity of endogenous matrix metalloproteinase-9 (MMP-9), as migration was suppressed by the GM6001 inhibitor of MMPs (20). LTB4 and PGE2 now are shown to elicit chemotaxis and chemokinesis, respectively, of human blood I cells and T lymphoblastoma cells of the
1To whom correspondence and reprint requests should be addressed. at: Immunology and Allergy, UB8B, Box 0711, University of California MedicalCenter,533 Parnassus, San Francisco, CA 94143-07 11, USA. 2Abbreviations: PCE2, prostaglandin E2; LTB4, leukotriene B4; LTC4, leukotriene C4; LTD4, leukotriene D4; MMP, matrix metalloproteinase; R, receptor; Th, helper T cell; IL, interleukin; IFN, interferon; P, penicillin; 5, streptomycin; FBS, fetal bovine serum; CM6001, HOHNCOCH2CH(i-Bu)CO-L-Trp-methylamide; CM2454, L-Leu-L-Trp-methylatnide; PMSF, phenylmethylsulfonylfluoride; BM, basement membrane; DFP. diisopropylfluorophosphate; HRP, horseradish peroxidase.
1473
COMMUNICATION
RESEARCH
CD4+8+3t Tsup-1 line through the Matrigel model basement membrane by enhancing both intrinsic motility and expression of multiple MMPs, which are requiredforT cell penetration of the basement membrane.
in the
lower
lower
MATERIALS
AND METHODS
Quantification
of migration
Tsup-1
cells
(American
Type
Culture
compartment
for chemnotactic
experiments,
and
was
compartment,
cell
suspension,
and
Mairigel
layer
at
concentrationof 108 M to 10 M. Chambers with uncoated incubated forthe previouslyestablishedoptimal time of 4 with Matrigel-coated filtersfor 18-54 h at 37 C in 5% Insertsthen were shaken for 10 mm at room temperature to
Collection.
Rockville,
Md.)
were
in RPMI-modified Dulbecco’s medium (UCSF Cell Culture Facility, Sun Francisco,Calif.)with 10% (v:v)fetalbovine serum (FBS) (Hyclone, Inc.,Logan, Utah), 100 U/nsl of penicillinC (F),and 100 .tg/mlof streptomnycin(5) at a density of 0.5-i x 106/ml, as described (21), and washed twice in protein-fieeIscove’s medium (UCSF Cell Culture Facility).Replicate suspensions of 0.8 x 106 washed Tsup-1 cellsin 0.2 ml of Eseove’smedium with F, S. and 2% (v:v)Nutridoma (Boehringer-Mannheimn. Indianapolis, Ind.) were adided to the 6.5 mm diameter insertsof Transwell chambers (Costar. Cambridge, Mass.) over 5 JIm pore polycarbotsatefilters (Nucleopore Corp., Pleasanton,Calif.) without or with a continuous even coating of 15 .tlof standard Matrigel or Matrigeldepleted of growth factors(CollaborativeResearch, Bedford, Mass.), which separated the cells from 500 jsl of buffer in the lower compartment. A stimulus of 10b0 M to 106 M PCE2, LTB4, or a structuralanalog (Biomo!. Inc.. Plymouth Meeting, Pa.) was present only cultured
added
at the same concentration to the lower compartment and cell suspension forchemokinetic studies.The hydroxamnicacid dipeptide analog inhibitor of MMPs, CMoOO1, and the inactivepeptide control L-Leu-L-TrpNHMe-HCI (CM2454)(Clycomed, Alameda, Calif.)were added to the the
same
were Isand those Co2 in air. detach cells filters
adherent to the bottom surface of the filters and the number of cells in the lower compartment was determined. The migration responses were expressed as a percentage of the totalnumber of Tsup-1 cellsadded to the chamnber or, for some studies,as a percentage of the controlvalue (100%) for migration in the absence of an inhibitor.En studies of the time and
course
of migration,
chemokinesis
respectively,
the
were
nsagnitudes
a mean
of 35%
and
at 54 h of the corresponding
131% other
studies
were
performed.
of stimulation
lsy i0
Matrigel
through
52%
at
responses
Mixed
of chemotaxis
M LTB4 and 108 M PCE2, 18
at 36
T cells
were
(n2),
isolated
from
venous blood of normal human subjects, not taken any medications for at least 96 Ii, by sequential sedimentation of erythrocytes, removal of granulocytes and cia,
by centrifugation
Piscataway,
Petri
dishes,
Sephadex
N.J.), and
on cushions
elimination
filtration
of the
ClO, as described
of Ficoll-Hypaque
of monocytes (22).
lymphocytes The
T cells
through
sodium
who had dextran residual (Pharma-
by adherence recovered
all
when
citrate-anticoagulated
erythrocytes
and
123%
h and h
to plastic a column had
of
a CD3
DChe,notaas
I
CtnoIones,s
ChernoIa.,.
#{149}wr6M6001
ocsemokress #{149}W95GOOl
I
2s
j..
I I
0
9
0
10
o(Iog
N)
CO,,..,0’.,,
E2
o
7
6
N).
Figure 1. Effects of eicosanoids PCE2 on Tsup-1 cell migration
on Tsup-1 cell tnigration. A) Effects of through Matrigel. Each bar and bracket
represents the mean ± so of the results of three separate studies performed in duplicate. The concentration of PCE2 or PCF20 indicated was present
B::
DCI,e,,,otaxa
in the buffer compartment alone for chemotaxis (open bars) and in both the buffer and cell suspension compartments for chemokinesis (stippled bars). The effects of 1 (tM CMoOO1 on chemokinesis stimulated by io M and 10’ M PCE2 are depicted by the solid black bars. The statistical significance of differences between stimulated values and that of the concurrent buffer control was determined by a paired test and is depicted by + = P < 0.05 and * = P < 0.01. B) Effects of LTB4 on Tsup-1 cell migration through Matrigel. Each bar and bracket represents
I
Chemoks,eas
#{149}wn, GM6I
the mean
± so of the results of three separate studies performed in The effects of I .iM GM600I on chemotaxis stimulated by io M and 10 M LTB4 are depicted by the solid black bars. The experimental procedures, depiction of results, and levels of statistical significance are as described in panel A. C) Effects of PGE2 and LTB4 on Tsup-1 cell migration without Mutrigel. Each bar and bracket represents the mean ± st of the results of three separate studies performed in duplicate. The effects of! JIM CM#{212}OO1 on chemokinesis evoked by 10
duplicate.
20
I
L
‘I
11:
0
10
S
9
C(-g0
E1
7
4
LTC4
M to !0
M PCE2 and chemotaxis
depicted
by solid
results,
1474
Vol.9
November
1995
The FASEB Journal
and
levels
black
bars.
of statistical
elicited
The
experimental
significance
by 10
M to 10 procedures,
are as described
M LTB4 are depiction of in panelA.
LEPPERT Er AL
RESEARCH COMMUNICATION Chemotaxis
mChemolsnesls #{149}WithGM600E
1
I I
Figure 2. Stimulation of human bloodlcell migration through Matrigel by PGE2 and LTB4. Each bar and bracket represents the mean ± so of the results of three separate studies performed in duplicate. The effects of 1 .tM CM6001 on chemokinesis evoked by !0 M and i0 M PGE2 and chemotaxis elicited by 108 M and i0 M LTB4 are depicted by solid black bars. The
I 0
a
9
1
S
10
9
experimental
Conc.qpb’,fico, of (.109 U)
procedures,
els of statistical for Fig. IA.
significance
depiction of results, and levare as described in legend
4’C.The
purity of greater than 95% with a range of 61-78% CD4 and 19-37% CD8+. as assessed by flow cytometry with FITC-labeled mouse monoclonal antibodies (Zymned, Inc., So. San Francisco, Calif.). Migration of freshly harvested human blood-derived T cells, which were not pre-
thickness
treated, was quantified as described for Tsup-1 cells, with migration times of 4 and 24 h at 37 C for filters without and with Matrigel, respectively. CD4+ Tcells were isolated by fluorescence-activated cell sorting, as described (7).
3 (stromelysin)antibodies(Biodesign International, Kennebunk, Maine) and mouse monoclonal IgG! anti-MMP 2 (Ab 75-7F7; 24) and antiMMP 9 (Ab 7-11C; 25) (Oncogene Science, Cambridge, Mass.), fol-
of binding
Characterization
of [3H]LTB4
by Tsup-
1
SDS-10%
polyacrylamide
gel
at
150
V for 3 h at
resolved were transferred by electroblotting to a 0.45 (tm pore nitrocellulose membrane (Hybond, Amershamn, Arlington Heights. Ill.) at 100 V for 1 h. The membranes were developed with sheep anti-MMPproteins
by horseradish peroxidase (HRP) -conjugated rabbit anti-sheep lgC and HRP-conjugated goat anti-mouse lgC and luminescence analysis (ECL system. Amersham). lowed
cell membranes To prepare membrane-enriched subcellular particles, suspensions of 2 x i0 Tsup-I cells/mI of l-IBSS were incubated with 1 mM diisopropylfluorophosphate (DFP) for 1 h at 20 C, washed three times with HBSS, resuspended at 5 x !07/ml in 0.1 M sucrose containing 2 mM EDTA. 2 mM phenylmethylsulfonylfluoride (PMSF), 0.1 mM leupeptin. 0.1 mM DL-thiorphan, and 5 mM Tris-HCI (pH 7.2) (all from Sigma Chemical Corp., St. Louis, Mo.), and disrupted by sonication twice for 30 s with 200 watts at 4 C. The sonicate was centrifuged at 200 x g for 10 main at 4 C, the 200 x g supernatant was centrifuged at 200.000 X g for 20 mm at 4 C, and the pellet was washed twice and resuspended in the original volumne of HBSS with 0.2 g/100 ml of ovalbumin and 15 mM HEPES (pH 7.4). Replicate 100 l ahiquots of resuspended 200,000 x g pellet of sonicated Tsup-1 cells were incubated for 60 mm at 4 C with 1.1 pmol of [3H]LTB4 (Dupont-NEN. Boston, Mass.; 204 Ci/mmol), without and with picomolar-nanomolar nonradioactive LTB4. Bound and unbound [3H]LTB4 were resolved by filtration, binding data were analyzed, and the Kd and number of sites per cell were calculated as described (23).
Western
blot
analysis
of MMPs secreted
by Tsup-1
cells
suspensions of 4.-5 x 10 Tsup-1 cells/ml of Iscove’s mnedium 1% Nutridoma were incubated for 4 and 24 h with i0 M to 10-6 M PGE2, LTB4, or a structural analog at 37 C in 5% C02, and centrifuged for 5 mm at 400 x g. The supernatant medium was transferred to clean plastic test tubes and the Tsup-1 cells were washed twice in PBS. suspended in 0.25 ml of 250 mM sucrose containing 1 mM PMSF, 0.1 mM leupeptin, and 15 mM HEPES (pH 7.4), and lysed by sonication for 30 s with 200 watts at 4 C. The sonicates were centrifuged at 600 X g for 10 mm at 4 C, and the 600 X g supernatants were centrifuged at 100,000 X g for 30 mm at 4 C in a TL-100 centrifuge (Beckman, Fullerton, Calif.). Replicate aliquots of the media supernatants and 100,000 x g supernatants were electrophoresed in 1 mm Replicate
with P. S, and
I CELL METALLOPROTEINASE-DEPENDENCE
OF MIGRATION
Zymographic quantification MMPs of Tsup-1 cells
of proteolytic
activities
of
and 100,000 x g supernatants of sonicates of Tsupas for Western blots and electrophoresed in nonreducing SDS-10% PAGs. which were copolymerized with 1 mg/mI of type A gelatin from Porcine skin or casein from bovine milk (Sigma), as described (20). The gels were then incubated in 2% Triton X-100 to remove SDS and for 24 h at 37 C in 50 mM Tris-HCI-5 mM CaCI2 (pH7.6). After staining with Coomassie blue,the decrease in staining Media supernatants
I cells were prepared
of each band of protease activity was quantified results of stimulation of untreated controls. Densitometric
or inhibition
by densitometry
were expressed
as a relative
and the percent
analysis
The intensity of bands in Western blots and zymography gels quantified with a Scan-jet IIC (Hewlett-Packard. Boise, Idaho). were analyzed with NIH Image 1.41 software.
Flow-cytometric MMPs
of Tsup-1
semiquantitative
evaluation
were Data
of surface
cells
Replicate suspensions of Tsup-1 cellswere washed twice in PBS and incubated for 90 mm at 4 C with 1/500 FITC-sheep IgG anti-MMP 2, 1/500 FITC-sheep IgG anti-MMP 3, 20 ng/ml of mouse monoclonal IgG! anti-MMP 9, and the nonimmune sheep lgG controls and mouse MOPC 21 control IgGi (Sigma). The mouse monoclonal anti-MMP 9 and control lgG!-treated suspensions were washed three times with PBS containing 1% FBS and incubated for 30 mm at 4 C with 1/350 FITC goat anti-mouse IgGi Abs (Zymed, Inc.). Fluorescence flow cytometry was performed with a FACScan (Becton-Dickinson, Inc., San Jose, Calif.). using the LYSIS II software program. Analysis gates included the viable populations of cells determined by low-intensity fluorescence of
1475
RESEARCH COMMUNICATION DNA
staining with propidium
intensity
values
population
were
in each
iodide.
determined
for
The the
log
channel
propidium
fluorescence iodide-negative PGE2
sample.
RESULTS Tsup-1
cell receptors
for PGE2
and LTB4
Cultured Tsup-1 cells were known to express a mean of 350 PGE2 Rs percell, which bind [3H}PGE2 with a mean K(1of2.5 nM and a specificity characteristic of EP2-/EP4-type Rs (11, 21). The EP-subset identity of the Tsup-1 cell Rs had been confirmed by immunofluorescence and Western blot analyses with monospecific antibodies and by the relative binding affinities of subset-selective synthetic analogs of PGE2 (21). LTB4Rs now are demonstrated by the binding of [3H] LTB4 to a mean of 27,500 sites per Tsup-1 cell, which is rapid, saturable, rapidly reversible by a 1,000-fold excess of nonradioactive LTB4, and characterized by a Kd of 204 ± 36 nM (mean ± SEM, n 3) similar to that of the low-affinity subset of human neutrophil LTB4Rs (23). The specificity of binding of LTB4 by Tsup-1 cells, as determined by the rank order of competitive potencies of 20-OH-LTB4, 6-trans-LTB4, LTC4 and LTD4 (Biomol, Inc.) is also the same as that of low-affinityLTB4Rs of human neutrophils. Stimulation Matrigel
of Tsup- 1 cell migration and LTB4
Vol.9
N
0 a
0
N
I0
2
4
3
Log of Relative Intensity of Fluorescence
N
E z
LTB4
N N
through
10
by PGE2
PGE2 significantly enhanced the migration of Tsup-1 cells through the Matrigel model basement menibrane (BM) at concentrations of iO M to 10-6 M when presented in both compartments of the chamber at equal concentrations to elicit chemokinesis and when presented as a gradient from the lower compartment to elicit apparent chemotaxis (Fig. 1A). The principal mechanism by which PGE2 stimulated the migration ofTsup-1 cells through the Matrigel BM appears to be chemokinesis, as maximal mean enhancement of sevenfold was attained at 10 M PGE2 in the absence of a concentration gradient and the magnitude of chemokinesis exceeded that of apparent chernotaxis at all concentrations except 106 M PGE2. At 10-6 M PGE2, an optimally chemokinetic concentration of 10-8 M could be reached in the cell compartment by diffusion from the stimulus compartment (Fig. 1A). PGE1, which binds to the EP2-type of PGE2R on Tsup-1 cells with an affinity equal to that of PGE2 (11,21,26), had an identical stimulatory effect on chemokinesis through the BM in three additional studies, where mean chemokinetic enhancement of Tsup-1 cell migration by 10-8 M PGE2 was 6.1-fold (P< 0.01), that by 10 M, 10-8 M, and 10 M PGE1 were 2.4-, 6.5-, and 3.0-fold (P< 0.05, 0.01 and 0.01), respectively. In contrast, 10-8 M to 10-6 M PGF2c, which has less than 0.001 the affinity of PGE2 for the EP2 R (26), had no chemokinetic effect on Tsup- 1 migration through the MatrigelBM (Fig. 1A). LTB4 enhanced the migration of Tsup-1 cells through the BM almost solely by stimulating chemotaxis significantly at concentrations of 10.10 M to 10-6 M, with a maximal mean
1476
N
November
1995
1
2
Log ci RciatIv
Figure
3. Enhancement
of Tsup-l
3
tnt.nelty ci Fluorssc.r.c.
cell-surface
expression
of MMP-2
by
PCE2 and LTB4. The results of one representative analysis of the intensity of fluorescent labeling of Tsup-1 cells with FETC-conjugated sheep anti-MMP-2
antibodies
are
depicted
to illustrate
the
stimnulatory
effects
of24 h incubationswith 10 nM and 1 RM PCE2 (upper)and LTB4 (lower). A summary of the resultsof similar studies of MMPs-2, -3, and -9 is presented
in Table
1.
increase of 14-fold at 10 M and significant chemokinetic stimulation only at 1 .tM (Fig. 1B). In three additional studies where 10 M LTB4 stimulated Tsup-1 cell chemotaxis through the BM by a mean of 11-fold (P< 0.01), 10.8 M and 10 M 20-OH LTB4 stimulated chemotaxis by 9.2- and 11-fold (P< 0.01 for both), respectively, which is consistent with the equal affinity of binding of LTB4 and 20-OH LTB4 to the LTB4 R (23,27). In contrast, 6-trans-LTB4, which binds to Tsup-1 cells with 0.01 the affinity of LTB4, stimulated chemotaxis by only 1.1-fold at 10.8 M (P> 0.1) and by 5.0-fold
The FASEB Journal
LEPPERT El AL
RESEARCH COMMUNICATION TABLE 1. Flotv-cytometric
Concentration
of stimulus,
assessment
-
of mncrea.mes in Tsup-1 Cell-surface expression of MMPs by PCE2 and LTB4U
log M
MMP2
0
7.3 ± 2.1
M
15
PCE2, i06
M
93 ± 5l
LTB4,
M
17 ± 3.5
LTB4, 10’ M
85 ± 4l
PGE2,
MMP3
1o 10
± ‘LOt
MMP9
12 ± 1.5
6.3 ± 2.5
22
8.7
± 3.2*
111 ± 6O 20,
sI
156 ± 109
± 2.1*
15 ± 3.1* 8.7
± 4.0
16 ±
#{176}Each value is the mean ± so of the results iodine-negative) Isup-1 cells. The range of viability
from three studies expressed in units of log mean relative fluorescence for the population of viable (propidium for the samples in the three studies was 27-55%, which is consistent with the susceptibility of Tsup-1 cells to apoptosis. The mean values for control samples in the three studies were: 262 for FI’IC-labeled mouse monoclonal anti-CD4 antibody, 5 for FITC-laheled mouse monoclonal anti-CD3 antibody (Zymed Laboratories), as expected for CD4CD8CD3#{176}” Tsup-i cells; 4 in the absence of any FITC-labeled antibody; and 3 for FITC-labeled goat antimouse IgC second antibody alone. al = mean of only two values, as samples from one study were lost. The significance of increases in fluorescence, compared to control values, was calculated by a paired 1-test: P=0.05, tP
