Int. J. Cancer: 66, 98-103 (1996) Publication of the International Union Against Cancer Publication de I'Union lnternationale Contre le Cancel 1

0 1996 Wiley-Liss, Inc.

TOLERANCE TO THE ANTI-METASTATIC EFFECT OF LIPOPOLYSACCHARIDE AGAINST LIVER METASTASIS IN MICE Katsuaki sATO'.3, Yung Choon Yoof, Kaori MATSUZAWA', Ryosuke WATANABE~, Ikuo SAIKI~ , Shuichi TONO-OKA' and Ichiro AZUMA'

2Research Institute for WAKAN-YAKU (Traditional Sino-Japanese Medicines), Toyama Medical and Pharmaceutical Univenity, 2630 Sugitani, Toyama 930-01, Japan.

Institute of Immunological Science, Hokkaido University, Kita-15, Nishi- 7, Kita-ku, Sapporo 060; and

We describe the involvement of endotoxin tolerance in the refractoriness of i ts anti-metastatic effect against murine synge- neic tumors. Three i.v. administrations of LPS at intervals of 4 days after tumor inoculation inhibited liver metastasis of L5 I78Y- ML25 cells, whereas 3 consecutive i.v. administrations of LPS showed only a slight suppressive effect. Multiple i.v. administra- tions of LPS, synthetic lipid A, i ts synthetic derivative DT-546 I, Staphylococcus aureus (S. oureus) BioParticles or Staphylococ- cal enterotoxin B (SEE) on days I, 5 and 9 after tumor inoculation inhibited liver metastasis of T-lymphoma cells in normal mice. The anti-metastatic effects of LPS, synthetic lipid A or DT-5461 but not S. aureus BioParticles or SEB were diminished in mice injected with LPS at daily intervals for 7 days before tumor inoculation. Mice receiving 3 consecutive i.v. administrations of LPS at daily intervals exhibited suppression of LPS-induced production of endogenous tumor necrosis fac- tor-cy (TNF-a), tumoricidal activity of macrophages, and natural- killer (NK) activity of splenocytes when compared with those of normal mice. Macrophages from mice receiving consecutive daily i.v. administrations of LPS for 3 days showed reduction of LPS-induced tyrosine phosphorylation of several intracellular proteins, including p42maPk/ERKZ when compared with that of the cells obtained from normal mice. These data suggest that the LPS-induced anergic state of rnonocyteslmacrophages plays a crucial role in endotoxin tolerance with respect to the metastasis of T lymphoma in the liver. o 1996 Wiley-Liss, Inc.

Bacterial endotoxin (LPS), a major constituent of the outer membrane of gram-negative bacteria, has many biological effects, and a lipid-A component represents the active center responsible for most of the biological functions of LPS (Billiau and Vandckcrckhove, 1991). The interaction of LPS with cells, especially monocytesirnacrophages, causes cellular activation with the subsequent synthesis and secretion of numerous inflammatory cytokines, including interleukin (1L)-1, IL-6 and TNF-a and these cytokines are involved in several pathophysi- ological effects of LPS (Billiau and Vandekerckhove, 1991). Several serum and cell-surface proteins that bind LPS have been identified in various mammalian tissues (Morrison et al., 1993). LPS appears to bind via its lipid-A moiety to an LPS-binding protein (LBP), a 60-kDa glycoprotein found in normal and acute phase, greatly enhancing the sensitivity of the cells to LPS (Wright etaZ., 1990). CD14, a glycosylphospha- tidylinositol-anchorcd protein expressed on the surface of monocytesimacrophages and activated neutrophils, functions primarily as a receptor for a complex of LPS and LBP (Wright et al., 1990).

The early intracellular events regulating the LPS-induced activation of monocytesimacrophages are being extensively studied. LPS binding to CD14 causes the rapid tyrosine phosphorylation of several cellular proteins and the src kinase pS3/S6iYn is associated with CD14 in monocytesimacrophages (Stefanova et aL, 1993). LPS is reported to induce the transient activation of several protein kinases, such as protein kinase C (PKC) (Shapira et al., 1994), protein tyrosine kinase (PTK) (Shapira et al., 1994), and mitogen-activated protein (MAP) kinase (Weinstein et al., 1992; Dong et al., 1993). The phosphor- ylation of intracellular proteins is essential to LPS-induced

cytokine production and the development of tumoricidal activity in monocytesimacrophages (Shapira et al., 1994).

The development of refractoriness by the repeated adminis- tration of LPS is recognized in humans and experimental animals, and tolerance develops to many endotoxic effects, including lethality and pyrogenicity (Fraker et al., 1988). This phenomenon is most likely a result of hyporesponsiveness of cells of the monocyteimacrophage lineage to LPS (Virca et al., 1989; LaRue and McCall, 1994). Relatively little is known about the molecular basis of this endotoxin tolerance.

LPS exhibits suppressive effects on the progressive growth and metastasis of syngeneic tumors in experimental animal models (North and Havell, 1988; Nakatsuka et al., 1991). Interest in the potential use of LPS, lipid A and its relative compounds in cancer therapy has increased, and human clinical trials with these agents have been conducted (Kumazawa et al., 1992). We have reported that the administra- tion of LPS, synthetic lipid A or its derivative, DT-5461, inhibited the metastasis of LS178Y-ML25 T lymphoma in mouse liver (Sato et al., 1992). However, tolerance to anti- tumor effect of LPS against syngeneic tumors has not been clarified.

Here, we examine tolerance to the anti-metastatic effect of LPS against murine T-lymphoma cells and the mechanism involved in endotoxin tolerance.

MATERIAL AND METHODS Reagents

LPS from E. coli 0127:BS was purchased from Difco (De- troit, MI). Synthetic lipid A (LA-15-PP) and its derivative DT-5461 were provided by Daiichi Pharmaceutical (Tokyo, Japan). S. aureus BioParticles were purchased from Molecular Probes (Eugene, OR). SEB was purchased from Toxin Tech- nology (Sarasota, FL). Recombinant mouse TNF-LY was pur- chased from Genzyme (Cambridge, MA). Horseradish peroxi- dase(HRP)-coupled recombinant anti-phosphotyrosine MAb (clone RC-20) was purchased from Transduction Laboratories (Lexington, KY). Anti-mouse MAP kinase (ERK2) was pur- chased from UBI (Lake Placid, NY). Actinomycin D was purchased from Wako Pure Chemical Industries (Osaka, Japan). BSA (fraction V) and crystal violet were purchased from Sigma (St. Louis, MO).

Animals Specific pathogen-free female 7- to 10-week-old CDFl

(BALBlc x DBAi2) purchased from the Shizuoka Labora- tory Animal Center, Hamamatsu, Japan, were maintained in the Laboratory of Animal Experiment, the Institute of Immu- nological Science, Hokkaido University, in laminar airflow.

3To whom correspondence and reprint requests should be ad- dressed. Fax: (81)11709-6482.

Received: July 29,1995 and in revised form November 28,1995.

LPS TOLERANCE IN TUMOR METASTASIS 99

Cells and cell cultures Liver metastatic L5178Y-ML25 T-lymphoma cells were

provided by Dr. A. Okura, Banyu Pharmaceutical (Tokyo, Japan). The NK-activity-sensitive leukemia cell line Yac-1 was provided by Dr. K. Onoe, Institute of Immunological Science, Hokkaido University. L5178Y-ML25 and Yac-1 cells were maintained in RPMI-1640 supplemented with 7.5% FCS and L-glutamine. The TNF-a-sensitive fibroblast cell line originally derived from mice of the C3H strain, L929, was provided by Chugai Pharmaceutical (Tokyo, Japan) and maintained as monolayer cultures in DMEM supplemented with 7.5% FCS, vitamins, sodium pyruvate, non-essential amino acids and L-glutamine.

Assay for liver metastasis of T-lymphoma cells Liver metastasis was assessed by means of tumor-cell injcc-

tion into the lateral tail vein of mice (Sato et al., 1992). Five CDF, mice per group were given an i.v. injection of 4 x lo4 L5178Y-ML25 T-lymphoma cells. Treatment with LPS (1, 10 or 100 pg/mousc), synthetic lipid A (10 pg/mouse), DT-5461 (10 pg/mouse), S. aureus BioParticles (10 pgimouse) or SEB (1 pg/mouse) began the day after tumor inoculation, and was given every 4 days for a total of 3 treatments. The mice were killed 14 days after tumor inoculation and the livers were weighed to evaluate tumor metastasis. In another experiment, mice were injected i.v. with LPS (10 pg/mouse) at daily intervals 1 to 9 injections before tumor inoculation, to induce refractoriness to subsequent injections of LPS in terms of an anti-metastastic effect.

Preparation of serum f iom mice Three CDFl mice per group were given 1 or 3 i.v. injections

of LPS (10 pg/mouse) at daily or 4-day intervals, and bled 2 hr after the last injection. Sera obtained following centrifugation at 800 g for 5 min and passage through a 0.22-pm filter (Millipore, Bedford, MA) were stored at -20°C until use.

Preparation ofperitoneal exudate macrophages (PEM) Three CDFl mice pcr group were given a single i.p. injection

or 3 i.p. injections at daily or 4-day intervals of 10 pg LPS, and macrophages were collected by peritoneal lavage 4 days after the last injection. The macrophages were centrifuged at 200 g for 5 min, re-suspended in serum-free medium, and 0.1 ml of cell suspension (1 x 10h/well) was seeded into 96-well tissue- culture plates (Costar, Cambridge, MA). After 2-hr incuba- tion, the non-adherent cells were removed by vigorous washing and medium containing 7.5% FCS was added.

Preparation of splenocytes Three CDF, mice per group were given a single i.p. injection

or 3 injections of 10 pg LPS at daily or 4-day intervals, and the spleens were collected the day after the last injection. The spleens were homogenized in serum-free RPMI-1640 and whole spleen cells (1.25 x lo5 to 1 X 10°/well) were plated into 96-well U-bottomed plates (Corning Glass Works, Corn- ing, NY).

TNF-a assay TNF-a activity in sera was measured by means of a cytotoxic-

ity assay with actinomycin D-treated, L929 fibroblasts. Briefly, L929 cells (5 x 103/well) suspended in DMEM supplemented with 2.5% FCS were seeded into 96-well plates. Several hours later, serial 2-fold dilutions of sera or standard TNF-a were added to the wells, followed by the addition of actinomycin D at a final concentration of 0.05 pgiml. The cultures were incubated at 37°C for 24 hr, stained with 0.5% crystal violet in 20% methanol for 15 min, and washed with tapwater. Residual cells were solubilized with 0.1 ml 30% acetic acid. The absorbance at 590 nm was monitored photometrically by an MTP-22 2-wavelength microplate photometer (Corona, To- kyo, Japan), The value for TNF-a activity was expressed as

pgiml calculated from a standard curve generated from triplicate cultures.

Assay of macrophagemediated tumor cytotoxicity in vitro Macrophages (1 x 105/100 pl/well) were incubated with

L5178Y-ML25 T-lymphoma cells (1 x 104/100 pl/well) for 24 hr at 37°C. The cultures were pulsed with 0.5 pCi/well of [?H]thymidine (specific activity, 26 Ci/mmol; Amersham, Ayles- bury, UK) for the final 4 h. The cells were then harvested using a Filtermate 196 (Packard, Meriden, CT) and the radioactivity was measured in a Matrix 96 direct beta counter (Packard). The radioactivity was expressed as cpm, mean & S.D. of triplicatc cultures. The percentage of macrophage cytotoxicity generated was calculated from the radioactivity (in cpm) as follows: 100 - (radioactivity of target cells with macrophages)/ (radioactivity of target cells) x 100.

Assay of NK-cell-mediated tumor cytotoxicity in vitro Cell-mediated cytotoxicity was assessed by a radioactive-

release assay using the spleen cells and 51Cr-labelled target cells. Yac-1(3 x lo6) target cells were incubated with 0.1 ml of 51Cr (1 mCi/ml; specific activity, 561.39 mCi/mg; NEN, Bos- ton, MA) for 1 hr at 37°C in a humidified atmosphere containing 5% COz. The cells were washed 3 times with PBS to remove unbound radiolabel and re-suspended in RPMI-1640 containing 7.5% FCS. The 51Cr-labelled target cells (1 x lo4/ well) were added to the effector cells (splenocytes) at effector- to-target-cell ratios of 12.5:1, 25:1, 50:l and 100:1, and incu- bated for 4 hr at 37°C in a 5% COz/air atmosphere. Thereafter, the cells were centrifuged for 10 rnin at 8OOg. The supernatant from each well was recovered and the radioactivity count was determined in a gamma counter (Packard). The spontaneous release was determined by measuring the radiolabel released from an equivalent number of 51Cr-labelled target cells during a 4-hr incubation period at 37°C. The maximum release was measured by determining the radioactivity released by an equivalent number of cells treated with 0.1 M NaOH. The percentage of specific cytotoxicity was calculated as: (experi- mental release - spontaneous release)/(maximum release - spontaneous release) x 100.

Immunoprecipita~ion and Western blotting The macrophages (5 x 106/well) in 6-well plates (Costar,

Cambridge, MA) were stimulated with 0.1 pg/ml LPS for 20 rnin at 37°C. Thereafter, the stimulus was aspirated and rapidly replaced with 250 p1 of ice-cold lysis buffer (PBS, 1% Nonidet P-40,0.5% sodium deoxycholate, 0.1% SDS) containing phos- phatase and protease inhibitors (1 mM NalVO 4, 20 mM NaF, 1 mM EDTA, 10 pg/ml aprotinin, 0.5 mM benzamidine, 5 pg/ml leupeptin, 0.1 mM PMSF). Culture plates were shaken on ice for 10 min, the cells were then scraped off, and the lysates were transferred to microcentrifuge tubes. The nuclei and the insoluble cell debris were pelleted by brief centrifuga- tion, and supernatants were obtained. These were mixed with 2x loading dye (313 mM Tris-HCI, pH 6.8, 10% SDS, 99% mercaptoethanol, 50% glycerol, 0.01% bromphenol blue). For immunoprecipitation of MAP kinase, cell lysates (0.1 ml) were cleared with 1 pg of affinity-purified rabbit anti-mouse IgG and 20 pl of packed protein A-Sepharose beads for 1 hr. The lysates were then incubated with 5 pg of anti-MAP-kinase polyclonal antibodies for 2 hr, then with 1 pg of affinity- purified rabbit anti-mouse IgG for 1 hr. All incubations proceeded at 4°C. The beads were washed 4 times with 1 ml of lysis buffer and re-suspended in 2x loading dye. Cell lysates or the beads were boiled for 3 rnin and the supernatant fraction was obtained. Protein samples were resolved by SDS-PAGE on 10% acrylamide gels and transferred onto PVDF mem- branes (Bio-Rad, Hercules, CA). The non-specific antibody- binding sites on the membrane were blocked with 1 % BSA, 0.1% Tween 20 in saline (10 mM Tris-HC1, pH 7.4; 100 mM NaCI) for 20 rnin at 37°C. The membrane was then incubated

100 SAT0 ETAL.

for 1 hr at room temperature with 0.1 pg/ml of HRP-coupled anti-phosphotyrosine mAb (clone RC-20), and washed for 15 min with 0.5% Tween 20 in saline. The immunoreactive proteins were visualized using an ECL kit (Amersham).

Statistical analysis

was determined by applying Student's 2-tailed t-test. The statistical significance of differences between groups

RESULTS The effect of administration schedule of LPS on liver metastasis of T-lymphoma cells

Preliminary experiments showcd that repeated injections of LPS had no effect on liver weight in normal mice (data not shown). We investigated the effect of multiple systemic admin- istrations of LPS on liver metastasis of L5178Y-ML25 T- lymphoma cells. Mice were given 3 i.v. injections of LPS (10 pg) after an i.v. injection of 4 x lo4 L5178Y-ML25 cells, and the liver was weighed 14 days after tumor inoculation (Fig. 1). When the tumor cells were injected i.v. into CDF, mice, liver weight increased to more than 4-fold that of normal mice. Three intermittent i.v. injections of LPS at intervals of 2 , 4 or 6 days significantly inhibited liver metastasis, whilc 3 Consecutive injections of LPS only slightly inhibited liver metastasis. However, the effect was not significant. Furthermore, we observed that multiple i.v. administration of LPS on days 9, 10 and 11 or 10,11 and 12, as well as on days 1 , 2 and 3 after tumor inoculation, did not caused significant inhibition of liver metastasis of L5178Y-ML25 cells (data not shown).

The induction of hyporesponsiveness to anti-metastatic effects of LPS against liver metastasis of T-lymphoma cells

We investigated whether consecutive multiple injections of LPS would induce refractoriness to its own anti-metastatic effect (Fig. 2). A single i.v. injection of LPS one day before tumor inoculation effectively inhibited the liver metastasis of T-lymphoma cells. However, mice receiving sequential daily administrations of LPS before the tumor inoculation showed a low anti-metastatic effect on liver metastasis of L5178Y-ML25 cells, and this hyporesponsiveness increased in parallel with the increasing number of injectiom. We also examined the influence of thc LPS-induced refractory state on its anti- metastatic effect on liver metastasis of L5178Y-ML25 cells. Mice were given consecutive i.v. LPS injections for 7 days before tumor implantation, and the multiple LPS injections were given 1, 5 and 9 days after tumor inoculation (Fig. 3). Multiple i.v. administrations of LPS after tumor inoculation

0 1 2 3 4 S

Mean Weight cp) f S.D.

FIGURE 1 - Effect of the administration schedule of LPS on liver metastasis of L5178Y-MLZS T-lymphoma cells. Five CDFl mice per group were inoculated i.v. with 4 x lo4 L5178Y-ML25 T-lymphoma cells and injected i.v. with 10 &mouse of LPS on the indicated days after tumor inoculation. Mice were killed 14 days after tumor inoculation. "p < 0.01; **p < 0.001.

0 1 2 3 4

Mean Welght cs) f S.D.

FIGURE 2 - Induction of LPS tolerance in liver metastasis of L5178Y-ML25 T-lymphoma cells. Five CDFl mice per group were given i.v. injections of 10 pg/mouse of LPS for various times at daily intervals before an i.v. inoculation of L5178Y-ML25 T- lymphoma cells (4 x lo4), they were then killed 14 days after tumor inoculation. *p < 0.05; **p < 0.01; ***p < 0.001.

significantly inhibited liver metastasis of T-lymphoma cells in a dose-dependent manner, in the range from 1 to 100 pgimouse. On the other hand, consecutive exposures to LPS before tumor inoculation reduced the anti-metastatic effect of subsequent LPS treatment on liver metastasis of L5178Y-ML25 cells. The lipid-A domain is responsible for most of the biological functions of LPS (Billiau and Vandekerckhove, 1991). We tested the development of endotoxin tolerance involving the lipid-A portion (Fig. 4). Multiple i.v. administrations of syn- thetic lipid A or its synthetic derivative DT-5461 as well as LPS on days 1, 5 and 9 after tumor inoculation significantly inhibited liver metastasis of T lymphoma. Furthermore, re- peated i.v. injections of LPS daily for 7 days before tumor implantation markedly diminished the anti-metastatic effect of synthetic lipid A and DT-5461. W e also examined effect of LPS-induced hyporesponsiveness on the anti-metastatic effect of S. aureus BioParticles or SEB against T lymphoma. Figure 4 shows that multiple i.v. administrations of S. aureus BioPar- ticles or SEB on days 1, 5 and 9 after tumor inoculation markedly reduced the increase in liver weight caused by T-lymphoma metastasis in normal mice, and that the suppres- sive effects of S. aureus BioParticles or SEB on liver metastasis of L5178Y-ML25 cells were not affected in mice receiving consecutive i.v. injections of LPS for 7 days before tumor implantation.

The effect of systemic multiple administrations of LPS on endogenous TNF-a level in mice

The importance of endogenous TNF-a derived from macro- phages as a mediator of LPS-induced suppression of the progressive growth of tumors is well documented (North and Havell, 1988). We therefore examined the kinetics of LPS- induced TNF-a production in mice (Table I). A single injec- tion of LPS induced the marked production of serum TNF-a. Furthermore, 3 intermittent injection of LPS at intervals of 4 days induced even higher levels of circulating TNF-a, while 3 sequential injcctions had no effect.

The influence of multiple administrations of LPS on tumoricidal activities of macrophages and NK cells in mice

We tested the influence of multiple administrations on LPS-induced tumoricidal activities of macrophages (Fig. 5). A single injection of LPS enhanced the cytotoxic activities of macrophages against L5178Y-ML25 T-lymphoma cells. Three intermittent injections of LPS at intervals of 4 days induced greater cytotoxic activity than a single injection, whereas the daily consecutive administrations of LPS failed to induce

LPS TOLERANCE IN TUMOR METASTASIS 101

Normal

-

0 1 2 3 4 5

Mean Weight (s) i S.D.

FIGURE 3 - Effect of LPS tolerance on its anti-metastatic effect on liver metastasis of L.5178Y-ML25 T-lymphoma cells. Five CDFl mice per group were injected i.v. with LPS (10 p,g/mouse) (hatched bar) or saline (closed bar) a t daily intervals for 7 days before an i.v. inoculation of L5178Y-ML2S T-lymphoma cells (4 x lo4). Mice were injected i.v. with various doses of LPS or with saline alone (-) on days 1 , s and 9, then killed 14 days after tumor inoculation. *p < 0.05; **p < 0.01; ***p < 0.001.

Normal I - LPS

synthetic lipid A

DT-5461 S. uureus BioPartlclea

SEB

4

0 1 2 3 4 5 Mean Weight (s) i S.D.

FIGURE 4 - Involvement of the lipid-A portion in the induction of LPS tolerance on liver metastasis of L5178Y-ML25 T- lymphoma cells. Five CDFl mice per group were injected i.v. with LPS (10 pgimouse) (hatched bar) or saline (closed bar) at daily intervals for 7 days before an i.v. inoculation of L5178Y-ML25 T-lymphoma cells (4 x lo4). Mice were injected i.v. with 10 pg/mouse of LPS, synthetic lipid A, DT-5461, S. aureus BioPar- ticles or 1 pgimouse of SEB or with saline alone (-) on days 1, 5 and 9, then killed 14 days after tumor inoculation. *p < 0.01; **p < 0.001.

TABLE I - THE EFFECT OF MULTIPLE INJECTIONS OF LPS ON ENDOGENOUS TNF-a PRODUCTION

Treatment Injections (number) Intervals (days) TNF-a (pgiml)

- < 100 Normal - LPS 1 - 1200 2 so

3 4 2000 2 80* 3 1 250 2 30*

~~

Three CDF, mice per group were given i.v. injections of 10 pgimouse of LPS on the indicated days, and sera were collected 2 hr after the last injection. TNF-a were measured in serum samples by means of the cell proliferation assay. 'p < 0.001 vs. the value of the group of the mice receiving a single injection of LPS.

tumoricidal activity. We also examined the NK activity of splenocytes in mice that received multiple injections of LPS. As shown in Figure 6, splenocytes from mice that received a single injection of LPS had markedly enhanced NK activity as compared with that of normal mice. In contrast, the cytolytic activity of the splenocytes from mice that received 3 injections of LPS at daily or 4-day intervals was much lower than that of the cells obtained from mice given a single injection.

60 $ * CA I 1 *

i i 40-1 -r '

IqjectiOhs I 1 3 3 (numbers) Intervals - - 4 1 (days)

FIGURE 5 - Down-regulation of LPS-induced tumoricidal activ- ity of macrophages. Three CDFl mice per group were given i.p. injections of 10 pg/mouse of LPS on the indicated days, and the PEM were obtained 3 days after the last injection. Macrophages (lo5) were cultured with L5178Y-ML2S cells (lo4) for 24 hr, and the cytotoxic activity was measured. * p < 0.001.

I

100 50 25 12.5

E/T ratio

~ Suppression of NK activity of splenocytes by mul- tiple injections of LPS. Three CDFl mice per group were given i.v. injections of 10 pgimouse of LPS on the indicated days, and splenocytes were collected the day after the last injection. Cyto- toxic activity of splenocytes was measured by means of the 4-hr 5Tr-release assay. (0), normal; (O), single injection of LPS; (A), 3 injections of LPS at intervals of 4 days; (m), 3 consecutive injections of LPS.

LPS-induced tyvosine phosphorylation in LPS-tolerant macrophages

LPS rapidly increases the tyrosine phosphorylation of sev- eral proteins in monocytes/macrophages, and this intracellular event appears to be crucial for some functional consequences in these cells (Shapka et al., 1994; Dong et al., 1993). We examined the potential involvement of LPS-induced protein tyrosine phosphorylation in anergic macrophages (Fig. 7a). Some proteins with apparent molecular masses of 15 to 80 kDa were constitutively phosphorylated on tyrosine in macro- phages from normal mice (Fig. 7a, lane I) and similar results were obtained from the cells of mice given 3 consecutive injections of LPS (Fig. 7a, lane 5 ) . The increased tyrosine phosphorylation appeared in proteins with apparent molecular masses of 74 and 80 kDa (p74 and p82) in macrophages

102 SAT0 ETAL

FIGURE 7 - Suppression of LPS-induced protein tyrosine phos- phorylation in macrophages from LPS-tolerant mice. Three CDFl mice per group were given i.v. injections of 10 pg/mouse of LPS on the indicated days, and the macrophages were obtained 3 days after the last injection. The cells (5 x lob ) were either unstimu- lated (lane 1,3,5) or incubated (lane 2,4,6) with LPS (0.1 pgiml) for 20 min at 37°C. The cell lysates (a) or the immunoprecipita- tions with anti-MAP kinase anti-sera (b) were fractionated by SDS-PAGE and blotted onto PVDF membranes, as described in "Material and Methods". Tyrosine-phosphorylated proteins were detected by ECL using HRP-conjugated anti-phosphotyrosine MAb. Lanes 1 and 2, normal mice; lanes 3 and 4, mice given 3 injections of LPS at intervals of 4 days; lanes 5 and 6, mice given 3 consecutive injections of LPS.

derived from mice given 3 intermittent injections of LPS (Fig. 70, lane 3). Incubating macrophages from normal mice with LPS (0.1 pg/ml) induced protein tyrosine phosphorylation at 5 bands of 24,32,39,42 and 74 kDa (p24, p32, p39, p42 and p74) (Fig. 7a, lane 2). Increased tyrosine phosphorylation of p24, p32, p39 and p42, but not p74, was induced by LPS in macrophages from mice given 3 injections of LPS at 4-day intervals (Fig. 7a, lane 4). In contrast, with cells from mice given 3 continuous injections, LPS stimulation induced signifi- cantly Iowcr tyrosine phosphorylation of p24, p32, p39, p42 and p74 (Fig. 7a, lane 6) when compared to normal LPS-stimulated macrophages (Fig. 7a, lane 2). p42 correspond to p42mapk/ ERK2 in monocytesimacrophages (Weinstein et al., 1992). To determine whether the repression of LPS-induced tyrosine phosphorylation of ~ 4 2 " ~ p ~ / E R K 2 is associated with the aner- gic state of macrophages, the anti-MAP-kinase-immunoprecipi- tated fractions were analyzed by immunoblotting with anti- phosphotyrosine antibody (Fig. 7b). We found that p42"'"pk/ ERK2 were phosphorylated on tyrosine residues in LPS- stimulated macrophages from normal mice and mice receiving 3 intermittent injections of LPS at intervals of 4 days (Fig. 76, lane 2 and 4). On the other hand, LPS-induced tyrosine phosphorylation of p42""pk/ERK2 were suppressed in the cells from mice given 3 continuous injections of LPS (Fig. 7b, lane 6).

DISCUSSION

A series of studies has indicated the potential use of LPS, lipid A and related compounds in the immunotherapy of cancer. However, the suitability of these agents remains unclear. In this study, we demonstrated that LPS induced refractoriness to its anti-metastatic effect against murine T lymphoma. We observed that the development of anti- metastatic activity of LPS depends on the application sched- ule, and that shorter intervals between repeated LPS injections promote the development of refractoriness (Fig. 1). Adequate intervals of LPS application may therefore be necessary to prevent establishing endotoxin tolerance and to enhance its suppressive effect on tumor metastasis. We showed that a single administration of LPS before tumor inoculation exhib- ited an anti-metastatic effect on liver metastasis of T- lymphoma cells, whereas increasing the challenge doses of LPS elicited a diminished response (Fig. 2). These results indicated that continued daily administration of LPS could generate a state of tolerance to its anti-metastatic activity. Whereas animals given sequential doses of LPS displayed a markedly reduced ability to inhibit tumor metastasis in response to LPS, synthetic lipid A or DT-5461, the anti-metastatic effects of S. aureus BioParticles or SEE were unaffected when compared with those in control animals treated with these agents alone (Fig. 4). We also observed that the anti-metastatic effects of LPS were significantly suppressed in mice receiving sequential injections of synthetic lipid A or DT-5461 before tumor inoculation (data not shown). These results indicated that LPS-induced hyporesponsiveness to its own anti-metastatic activity is specific for the action of the lipid-A moiety.

A variety of mechanisms have been proposed to explain the LPS-induced hyporesponsive state. LPS tolerance does not apparently involve the formation of anti-LPS antibodies in the serum or negative feedback by LPS-induced mediators (La- beta et d., 1993). Endotoxin tolerance may be associated with selective down-regulation of LPS-induced production of sev- eral cytokines (Erroi et al., 1993). LPS-tolerant mice exhibited low production of circulating TNF-a (Table I). We found that TNF-a inhibited the in vitro proliferation of L5178Y-ML25 cells (data not shown). These results implied that suppression of endogenous TNF-a secretion is correlated with refractori- ness to the anti-metastatic effect of LPS in LPS-tolerant mice. LPS tolerance to tumoricidal activities of NK cells or macro- phages has still not been clarified. We found that cytotoxic activity of macrophages and NK activity of splenocytes were significantly diminished in LPS-tolerant mice (Figs. 5, 6). TGF-P is a multifunctional cytokine with predominantly sup- pressive effects on several functions of NK cells (Gray et al., 1994). We observed that LPS-tolerant mice exhibited normal levels of circulating TGF-P and exogenous TGF-P suppressed LPS-induced enhancement of NK activity of splenocytes (data not shown). Taken together, our findings suggest that the ability of LPS to induce functional down-regulation of mono- cytes/macrophages and NK cells contributes to endotoxin desensitization in liver metastasis of T lymphoma in mice.

Several molecular mechanisms account for the development of the LPS-induced anergic state of monocytes/macrophages. Endotoxin tolerance may be associated with reduced numbers of LPS receptors and/or decreased receptor-binding affinity for LPS on the surface of the cells (Labeta et aL, 1993). We observed that LPS application schedules could provide no significant effect on the binding capacity to FITC-LPS in macrophages (data not shown). These findings indicate that the occupancy or modulation of cellular LPS receptors is not or not entirely responsible for tolerance. Relatively little is known about the intracellular signaling mechanisms control-

ling the development of the endotoxin tolerant state. We found that LPS-induced tyrosine phosphorylation of p24, p32, p39, p42 and p74 was suppressed in LPS-tolerant mice (Fig. 7a) . These results indicate that the suppression of protein tyrosine phosphorylation contributes to the down-regulation of LPS- elicited activation of macrophages. LPS-induced phosphoryla- tion of MAP kinases is correlated with the functional activa- tion of murine macrophages (Dong et al., 1993). LPS-induced tyrosine phosphorylation of p42""pk/ERK2 was repressed in LPS-tolerant macrophages (Fig. 7b) . We suggest that inactiva- tion the MAP kinase cascade via suppression of tyrosine phosphorylation is involved in the anergic state of macro- phages. On the other hand, the cellular expression of IKB-OI as a labile transcriptional repressor regulates LPS-induced gene expression in a state of endotoxin tolerance in human periph- eral-blood monocytes (LaRue and McCall, 1994). These phe- nomena imply that repression of early intracellular signaling pathways leading to subsequent gene expression is responsible for the LPS-induced anergic state in macrophages.

In conclusion, we demonstrate that tolerance of LPS abro- gated its suppressive effect on liver metastasis of T lymphoma in mice. The LPS-induced anergic state of macrophages may be crucial for developing this phenomenon. Our findings regarding LPS-mediated tolerance to its own anti-metastatic effects extend the mounting evidence that the clinical potential of LPS, lipid A and related compounds is limited for immuno-

therapy of cancer diseases in patients with gram-negative bacterial infection or septicemia.

ACKNOWLEDGEMENTS

We thank Dr. A. Tohgo and the late Dr. Y. Osada (Daiichi Pharmaceutical Co. Ltd., Tokyo, Japan), for providing syn- thetic lipid A and DT-5461, and Dr. K. On06 (Hokkaido University, Sapporo, Japan) for providing the Yac-1 cell line. We thank Dr. M. Hosokawa (Hokkaido University, Sapporo, Japan) for his invaluable suggestion concerning the measure- ment of TGF-@, and Drs. K. Kikuchi and s. Matsuzawa (Hokkaido University, Sapporo, Japan) for helpful suggestions regarding immunoprecipitation and Western blotting. We also thank Dr. M. Murakami (Hokkaido University, Sapporo, Japan) for helpful discussion and critical comments on this manuscript. We are grateful to Ms. S. Kuriyama and Ms. M. Sat0 for their assistance. This work was supported in part by the Special Coordination Funds for Science and Technology Agency of the Japanese Government; by a grant-in-aid for Scientific Research from the Japanese Ministry of Education, Science and Culture; by the Osaka Foundation for Promotion of Clinical Immunology; by grants-in-aid for Special Project Research from Hokkaido University provided by the Japanese Ministry of Education, Science and Culture; and by the Uehara Memorial Foundation.

LPS TOLERANCE IN TUMOR METASTASIS 103

REFERENCES

BILLIAU, A. and VANDEKERCKHOVE, F., Cytokines and their interac- tions with other inflammatory mediators in the pathogenesis of sepsis and septic shock. Europ. J. clin. Invest., 21, 559-573 (1991). DONG. Z.. 01. X. and FIDLER. I.J.. Tvrosine DhosDhorvlation of mitogen-activated protein kinases is necessary for hv&on'of murine macrophages by natural and synthetic bacterial products. J. exp. Med., 177,1071-1077 (1993). ERROI, A,, FANTUZZI, G., MENGOZZI, M., SIRONI, M., ORENCOLE, F.S., CLARK, B.D., DINARELLO, C.A., ISETTA, A., GNOCCHI, P., GIOVARELLI, M. and GHEZZI, P., Differential regulation of cytokine production in lipopolysaccharide tolerance in mice. Infect. Immun., 61, 4356-4359 (1993). FRAKER, D.L., STOVROFF, M.C., MERINO, M.J. and NORTON, J.A., Tolerance to tumor necrosis factor in rats and the relationship to endotoxin tolerance and t0xicity.J. exp. Med., 168,95-105 (1988). GRAY, D.J., HIROKAWA, M. and HORWITZ, D.A., The role of transform- ing growth factor p in the generation of suppression: an interaction between CD8+ T and NK ce1ls.J. exp. Med., 180,1937-1942 (1994). KUMAZAWA, E., TOHGO, A., SOGA, T., KUSAMA, T. and OSADA, Y., Significant anti-tumor effect of a synthetic lipid-A analogue, DT-5461, on murine syn eneic tumor models. Cancer Immunol. Immunother., 35, 307-317 (19927. LABETA, M.O., DURIEUX, J.J., SPAGNOLI, G., FERNANDEZ, N., WI- JDENES, J. and HERRMANN, R., CD14 and tolerance to lipopolysaccha- ride: biochemical and functional analysis. Immunology, 80, 415-423 (1993). LARUE, K.E.A. and MCCALL, C.E., A labile transcriptional repressor modulates endotoxin tolerance. J. exp. Med., 180,2269-2275 (1994). MORRISON, D.C., LEI, C.M., KIRIKAE, T. and CHEN, Y.T., Endotoxin receptors on mammalian cells. Immunobiobgy, 187,212-226 (1993). NAKATSUKA, M., KUMAZAWA, Y., MATSUURA, M., HOMMA, J.Y., KISO,

M. and HASEGAWA, A,, Enhancement of non-specific resistance to bacterial infections and tumor regressions by treatment with synthetic lipid-A-subunit analogs. Critical role of N- and 0-linked acyl groups in 4-0-phosphono-D-glucosamine derivatives. Int. J. Immunophannacol.,

NORTH, R.J. and HAVELL, E.A., The anti-tumor function of tumor necrosis factor (TNF) 11. Analysis of the role of endogenous TNF in endotoxin-induced hemorrhagic necrosis and regression of an estab- lished sarc0ma.J. exp. Med., 167, 1086-1099 (1988). SATO, K., SAIKI, I., Yoo, Y.C., IGARASHI, Y., KISO, M., HASEGAWA, A. and AZUMA, I., DT-5461, a new synthetic lipid-A analogue, inhibits lung and liver metastasis of tumor in mice. Jap. J. Cancer Res., 83,

SHAPIRA, L., TAKASHIBA, S., CHAMPAGNE, C. , AMAR, S. and DYKE, E.T., Involvement of protein kinase C and protein tyrosine kinase in lipopolysaccharide-induced TNF-a and IL-1p production by human monocytes. J. Immunol., 153,1818-1824 (1994). STEFANOVA, I., CORCORAN, M.L., HORAK, E.M., WAHL, L.M., BOLEN, J.B. and HORAK, I.D., Lipopolysaccharide induces activation of CD14-associated protein tyrosine kinase p53156'Y". J. b i d Chem., 268,

VIRCA, D.G., KIM, S.Y., GLASER, K.B. and ULEVITCH, R.J., Lipopoly- saccharide induces hyporesponsiveness to its own action in RAW 264.7 cells. J. bzol. Chem., 264,21951-21956 (1989). WEINSTEIN, S.L., SANGHERA, J.S., LEMKE, K., DEFRANCO, A.L. and PELECH, S.L., Bacterial lipopolysaccharide induces tyrosine phosphor- ylation and activation of mitogen-activated protein kinases in macro- phages.J. biol. Chem., 267,14955-14962 (1992). WRIGHT, S.D., RAMOS, R.A., TOBIAS, P.S., ULEVITCH, R.J. and MATHISON, J.C., CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS-binding protein. Science, 249,1431-1433 (1990).

11,349-358 (1991).

1081-1087 (1992).

20725-20728 (1993).