(CANCER RESEARCH 50. 1834-1840. March 15. 1990]

Role of Asialo-GjviiPositive Liver Cells from Athymic Nude or Polyinosinic-Polycytidylic Acid-treated Mice in Suppressing Colon-derived ExperimentalHepatic Metastasis1

Stefan A. Cohen,2 Shie-pon Tzung, Ralph J. Doerr, and Martin H. Goldrosen

Departments of Medicine and Surgery, Division of Gastroenterology and Hepatology, VA Medical Center and Buffalo General Hospital, Buffalo, New York 14215(S. A. C., S. T., R. J. D.], and Division of Clinical Immunology, Roswell Park Memorial Institute, Buffalo, Neu- York 14263 [M. H. G.J

ABSTRACT

Liver-derived (LD) murine colon adenocarcinoma MCA-38 cells injected into the ileocolic vein (ICV) of C57BL/6 mice developed distincthepatic foci within 14-21 days and survived for an average of 19-35days. In contrast, C57BL/6-nu/ni/ mice given injections of LD-MCA cellsby the same route did not develop hepatic lesions. Furthermore, '"In-

labeled LD-MCA-38 tumor cells were rapidly taken up by the liver ofconventional mice within l h and 73% of the radioactivity remained after24 h. However, about 60% of the '"In-labeled LD-MCA-38 tumor cells

were cleared from the liver of nude mice after 24 h. Nonparenchymalliver cells isolated from untreated conventional mice displayed littlecytotoxicity against freshly excised "Cr-labeled LD-MCA-38 cells butdid lyse the standard natural killer target, YAC-1 tumor cells, in 4 hchromium release assays. On the other hand, nonparenchymal liver cellsbut not spleen cells from nude mice were cytotoxic to "Cr-labeled LD-MCA-38 in vitro. The nonparenchymal liver cell population responsiblefor tumor killing was phenotypically nonadherent and asialo-GMi (AsGMi)positive. C57BL/6 mice treated with polyinosinic-polycytidylic acid|poly(IC)| also displayed cytotoxic activity against LD-MCA-38 tumorcells in vitro. Furthermore, poly(IC) treatment of mice 1-8 days aftertumor inoculation suppressed the number of hepatic foci and also significantly increased the life span of tumor-bearing mice. Treatment ofathymic nude mice or poly(IC)-treated conventional mice with anti-AsGM,induced significant numbers of foci and significantly decreased the lifespan of MCA-38-bearing mice suggesting that AsG.Mi-positive cells inthe liver of these mice may inhibit tumor growth in vivo. In conclusion,the host defense system of the liver from athymic nude or poly(IC)-treated mice possess AsGMi-positive cells that can suppress tumor implantation or tumor growth in the early stages of metastasis in liver.

INTRODUCTION

Hepatic metastasis is the most common cause of death fromcolon cancer (1,2). About 70% of patients with advanced colontumors develop hepatic metastasis (3, 4). After the invasion ofthe tumor through the portal blood to the liver, there is aprogression of colon-derived metastasis from hepatic to extra-hepatic sites (5). Despite the concomitant development of ex-trahepatic cancer, the extent of tumor within the liver determines the length of survival of most patients. Unfortunately,the prognosis for this disease has not changed over the last fewdecades. Little is known about the role of the host defensesystem of the liver in controlling tumor invasion. To study thisproblem with the goal of creating new therapies, more meaningful experimental models need to be generated that moreclosely simulate the human condition. Several syngeneic (6-8)and xenogeneic (9, 10) systems of colon-derived hepatic metastasis have been developed in rodents. The metastatic capacity

Received 8/7/89; revised 11/20/89.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported by Medical Research Funds from the Veterans Administration(S. A. C.) and from NIH R01-28835 (S. A. C.) and the Troup Fund from BuffaloGeneral Hospital (S. A. C, R. J. D.).

2To whom requests for reprints should be addressed, at Buffalo VeteransHospital, 3495 Bailey Ave., Buffalo, NY 14215.

of these tumors depends on both the origin of the tumor andthe route of tumor injection. This laboratory (11) has developeda murine model of experimental hepatic metastasis in which amurine colon adenocarcinoma (MCA-38') passaged in vivo is

injected into the ICV and subsequently colonizes the liver ofC57BL/6 mice in a reproducible fashion. In this unique syngeneic model, the number of colonies or foci formed within theliver correlates with survival of the mouse. A syngeneic murinemodel is particularly suited to study the interaction betweenhost organ microenvironment and tumor.

Few studies have focused on the immunological ability ofNPC which may also function to protect the parenchyma fromviruses, bacteria, and toxic insult ( 12) to inhibit the colonizationof colon tumors in vivo and their growth in vitro. Recently ourlaboratory (13-15) as well as others (16, 17) have shown thatAsGM,-positive cells from the NPC population of untreatedrodents have the ability to lyse tumor cells in vitro. However,their in vivo function in controlling colon tumor growth isunknown.

In the present study, using our experimental hepatic metastasis model, we have investigated liver metastasis of LD-MCA-38 cells in immunocompetent, immunostimulated, and congen-itally immunodeficient mice. In vivo, we have demonstratedthat AsGM,-positive liver cells from nude mice or poly(IC)-treated conventional mice play a role in suppressing the numberof metastatic liver foci as well as prolonging the life span ofmice. In vitro, these cells also display cytostatic and cytotoxicactivity against LD-MCA-38 cells.

MATERIALS AND METHODS

Mice. Adult 6- to 10-week-old male C57BL/6J mice were obtainedfrom The Jackson Laboratory (Bar Harbor, ME). Athymic nu/nu micewere obtained from Taconic Farms, Germantown, NY. C57BL/6-/!«/+ and C57BL/6-/1U/H«were bred in animal facilities at Roswell ParkMemorial Institute (Buffalo, NY). All mice were maintained on astandard mouse diet of pellets and water ad libitum.

Reagents. Poly(IC) was obtained from Sigma Chemical Co. (St.Louis, MO). Mice were given i.p. injections of 200 Mgof poly(IC) asindicated.

Preparation of LD-MCA-38 Tumor Cells (11). Mice bearing livertumors between the 3rd and 5th weeks post-ICV injection were sacrificed and pinned to a dissecting board. The liver was exteriorized by amidline incision, and a lobe-by-lobe hepatectomy was performed into aglass Petri dish containing 10 ml 0.25% trypsin-EDTA maintained at4°C.

Tumor nodules were carefully dissected from normal liver tissue, andnonnecrotic portions were cut into 2-3-mm fragments by a scalpelblade and fine forceps. These were collected into a 50-ml plastic tubecontaining 10 ml fresh 0.25% trypsin-EDTA and rotated for 15 min atroom temperature. The resulting tumor enzyme suspension was diluted

3The abbreviations used are: MCA-38, murine colon adenocarcinoma 38; LD,liver-derived; CD, culture-derived; NK. natural killer cell; NPC, nonparenchymalliver cells; ICV, ileocolic vein; AsGMi. asialo-GMi; IFN, interferon; poly(IC),polyinosinic-polycytidylic acid; HBSS, Hanks' balanced salt solution: Ad-NPC,

adherent nonparenchymal liver cells; RS. rabbit serum.

1834

on June 19, 2020. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SUPPRESSION OF HEPATIC METASTASIS BY AsGMI* LIVER CELLS

up to 50 ml with RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum plus antibiotics, filtered through a four-layer No. 200 nylon mesh, and spun at 200 x g. The pellet was washedonce in tissue culture medium and resuspended to a final concentrationof 1.5 x 10" cells/ml. The viability of injected cells as determined by

trypan blue dye exclusion was between 90 and 95%.ICV Injection (11). Mice 8-10 weeks old were anesthetized by i.p.

injection of 55 mg pentobarbital/kg. A midline incision was made andthe ICV was exposed; 1.5 x 10' tumor cells in 0.1 ml were injectedinto the vein with a 30-gauge needle attached to a 1-ml tuberculinsyringe, and a cotton-tipped applicator was placed over the injectionsite for 1 min to prevent excessive bleeding. After the cecum wasreturned to its proper position in the abdomen, the peritoneal membrane was closed with the use of a 5-0 chromic suture (Ethicon), andthe skin was closed with stainless steel clips.

Laparotomy. Two to 3 weeks after ICV injection, mice were reanes-thetized by i.p. injection of pentobarbital (50 mg/kg), and a 3-cmmidline incision was made from the tip of the sternum to the pubis. Allsix lobes of the liver were carefully examined anteriorly and posteriorlyfor tumors by a small stainless steel retractor and saline-moistenedcotton-tipped applicator. The numbers of visible liver foci were recordedfor each lobe, and the abdomen was closed by a 5-0 Ethilene suture(Ethicon) and stainless steel clips. Hepatic foci greater than 400 couldnot be accurately counted and were recorded as 400.

Survival. All experimental animals were palpated once per weekbeginning at the 2nd week until death to assess hepatic tumor growth.The day of death was recorded.

Preparation of NPC. NPC were isolated by modifying a previouslydescribed method (14, 15, 18). Briefly, mice were sacrificed by cervicaldislocation, the skin over the abdomen and chest was reflected, and theinferior vena cava was ligated above the diaphragm. The inferior venacava below the liver was cannulatcd with a heparinized 20-gauge needle,through which the liver was perfused at 37°Cwith 40 ml of HBSSwithout Ca2* and Mg:* (pH 7.2), delivered over 2 min with the outflow

through an incision in the portal vein. The liver was then perfused with100 ml of sterile HBSS (pH 7.6) supplemented with 0.05% collagenase.delivered over 5 min. The distended and blanched liver was removed,rinsed with HBSS, teased apart with a rubber policeman, and passedthrough a 100 ^m nylon mesh screen. Five ml of a single cell suspensionwere mixed with 7 ml of 30% w/v Metrizamide (Accurate and ScientificCo., Hicksville, NY), dissolved in Guy's balanced salt solution (withoutNaCI), then gently overlaid with I ml of Guy's balanced salt solution(with NaCI), and centrifuged at 550 x g for 15 min at 23°C.NPC were

collected from the interface, washed twice in HBSS, and resuspendedin media. NPC obtained by this method were >95% viable (as assessedby trypan blue exclusion) and contained less than 3% hepatocytes.Approximately 1 x IO7NPC were recovered per liver.

Isolation of NPC Adherent to Plastic (Kupffer Cells). Various numbers of NPC were seeded into microwells of 96-well flat bottomed tissueculture plates and allowed to adhere for 3 h at 37"C in tissue culture

medium (containing 20% fetal bovine serum. After incubation, thenonadherent cells were removed and the plates were washed 4 timeswith warm medium. Greater than 90% of adherent cells were livermacrophages or Kupffer cells as shown by uptake of latex beads andnonspecific esterase staining.

In Vivo Treatment of Mice with Anti-AsGM,. Rabbit anli-AsGMi waspurchased from Wako Chemical Co. (Dallas, TX). Mice were given 25-H\ injections of anti-AsGMi i.p.. 1 day prior to and 4 days after tumorinoculation. Rabbit anti-AsGMi reacts with murine NK cells, fetal livercells, and mouse monocytes (19-22).

Assay of Cytotoxic Activity. Target cells were labeled with Na/'CrO.,for l h at 37°C(100¿iCi/106cells). The "Cr release assay was performed

in 96-well fiat bottomed tissue culture plates (Costar, Cambridge, MA).NPC were mixed with 2 x IO4tumor target cells in triplicate in a total

volume of 200 ^1 and incubated for 4 h in a humidified atmosphere of5% CC>2in air. Supernatants were harvested and counted in a Packardgamma counter. The results were expressed as the percentage of specificrelease calculated as

100A - BC- B

in which A is cpm of experimental, B is cpm from spontaneous release,and C is cpm from target cells. The target cells used in this study wereCD-MCA-38 and YAC-1 maintained in tissue culture medium and freeof Mycoplasma and LD-MCA-38.

Assay of Cytostatic Activity. The cytostasis assay was performed in96-well fiat bottomed tissue culture plates as described previously (23).From 4 to 8 x IO5 NPC were mixed with 2 x IO4 cold tumor target

cells in triplicate in a total volume of 200 //I and incubated for 72 h ina humidified atmosphere of 5% CO2 in air. Wells were pulsed with 0.5fjCiofthymidine for 18 h. the cells were harvested with a semiautomaticcell harvester (Skatron Model 1701), and the cpm were determined byliquid scintillation counting. The percentage of inhibition was determined as

%>inhibition = 1 —¿� cpm (NPC + target cells)cpm (NPC) + cpm (target cells)

x 100

In Vivo Clearance of Tumor Cells. LD-MCA-38 tumor cells werelabeled with '"In oxide as described previously (24). Radiolabeledtumor cells were injected into the ICV (IO5 cell/0.1 ml) of mice and the

level of radioactivity in blood, lung, liver, spleen, heart, and kidney wasdetermined in a gamma counter (Beckman Model 5500B) 1 and 24 hlater.

RESULTS

Colonization of LD-MCA-38 Tumor Cells after ICV Injectioninto C57BL/6, C57BL/6-/iw/+, and C57BL/6-/i«//n/. LD-MCA-38 tumor cells after injection into the ICV of C57BL/6 orC57BL/6-/IM/+ mice produced distinct hepatic foci that couldbe easily quantitated. The hepatic foci formed by this procedureare microscopically apparent 11 days after injection and arevisible to the eye 14-21 days after injection. In a typical experiment, up to a maximum of 400 foci/liver were enumerated 21days after inoculation of LD-MCA-38 cells (Table 1). C57BL/6-/IM/+ mice with 400 hepatic foci on day 21 survived an averageof 27 days (Fig. 1). Furthermore, 21 days after tumor inoculation into untreated C57BL/6 mice, the liver was the only organ

Table I Induction of experimental hepatic metastasis in C57BL/6, C57BL/6-nu/+, and C57BL/6-nu/nu mice after ICI ' injection of LD-MCA-38 cells'

No. of

No. of liverfoci/mouse (day 2 1)

StrainC57BL/6

C57BL/6-nu/+C57BL/6-n«/nwmice5

10toMean3004000Median400

4000Range105-400

4000P"0.001colonization100

1000

" LD-MCA-38 tumor cells (2.5 x 10') were injected into C57BL/6. C57BL/6-nu/+, and C57BL/6-nu/nu mice by the ICV route.

* Wilcoxon rank sum test; compared to nu/+ mice.

XSURVIVAL755025nÕV.

0 —¿�OC57BI/6.nu/+\Q

•¿�—¿�•¿�C57BI/6.nu/nu*

V A AC57BI/6'

\\\°R

10 20 30 40 50 60 70

DAYS POST LO-MCA-3B INJECTION

80

Fig. 1. C57BL/6. C57BL/6-Õ1U/+.or C57BL/6-nu/n« mice were given injections of 2.5 x 10' LD-MCA-38 tumor cells into the ICV and survival was

determined.

1835

on June 19, 2020. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SUPPRESSION OF HEPATIC METASTASIS BY AsGMi*LIVER CELLS

Table 2 Clearance of'"In-LD-MCA-3S tumor cells by various organs ofC57BL/6-nu/+ and CS7BL/6-nu/nu mice'

StrainC57BL/6-nu/+

C57BL/6-nu/nuC57BL/6-/IU/+C57BL/6-nu//iuTime

after'"In-LD-MCA-38

injection<h)1

12424Organ

(10~J x cpm/g ±SD)*Blood6.1

±0.046.7 ±0.010.7 ±0.18.8 ±0.05Heart1.4

±0.11.9 ±0.30.9 ±0.061.3 + 0.02Lung33.2

±0.34.1 ±0.62.2 ±0.71.9 + 0.3Liver293.2

±25270.8 ±0.8'

213.6 ±22112.3 ±7''Spleen8.4

±22.8 ±0.7

12.6 ±0.23.5 ±0.4Kidney6.9

±0.37.8 ±0.1

20.1 ±317.1 ±1

" '"In-LD-MCA-38 tumor cells (2.5 x 10') were injected into C57BL/6-nu/+ or C57BL/6-mi/nu mice by ICV route and various organs were harvested after

and 24 h.* Mean ±SD of 5 mice/group.c Not significant, Wilcoxon rank sum: compared to nu/+ mice.* P < 0.003. Wilcoxon rank sum; compared to nu/+ mice.

Table 3 Cytotoxic effect ofNPC and spleen cells from various strains of untreated or poly(IC)-treated mice against "Cr-labeled tumor cells

ExperimentStrain"A

C57BL/6-OU/+

C57BL/6-/iu/nuB

C57BL/6-/IU/+C57BL/6-nu/nuC57BL/6-/IU/+

C57BL/6-nu/nuC

Athymic nu/nuAthymicnu/nuD

C57BL/6CellsNPCNPCAd-NPCNPCSpleenNPCSpleenNPCTreatment

invivoNone

LD-MCA-38 (day +14)Anti-AsGwi''

NoneNoneNone

RS"Anti-AsGMi*'None

NoneNone

NoneNone

Poly(IC)'%

ofspecificLD-MCA-38

(5h)2±2

6+ 1NDC35

±73±20±016±

20±01

±14±210±

20±00±

015±2release*YAC-1(4h)12±

17± 10±0

35 + 44±12+

132 + 20±0I0±3

49±4I8±

217+111

±266 ±4

" Three mice/group.* Mean + SD of triplicate samples. NPC or spleen cells (8 x 10s) were incubated with 2 x IO4"Cr-YAC-1 or LD-MCA-38 tumor cells.' ND, not determined.* RS or anti-AsGMi (25 n\) was injected i.p. into 3 mice 1 day prior to sacrifice.' Three C57BL/6 mice were treated i.p. with 200 »jgpoly(IC) 1 day prior to sacrifice.

where the tumor was detected. When these tumor cells wereinjected into C57BL/6-nu/nu mice, hepatic foci were not observed 21 days after tumor inoculation. These mice also survived for 70 days (Fig. 1) at which time they were sacrificedand the number of foci were determined. No foci were found inthe liver at 70 days (data not shown).

Biodistribution of '"In-LD-MCA-38 after ICV injection intoC57BL/6-IIII/+ and C57BL/6-/i«//i«mice. To determine thebiodistribution and clearance of LD-MCA-38 cells, 10' '"In-

LD-MCA-38 tumor cells were injected into the ICV of C57BL/6-WW/+and C57BL/6-/IM/««mice and the amount of radioactivity in various organs was assessed 1 and 24 h later. As shownin Table 2, within l h after tumor administration, >85% of the'"In was detected in the liver compared with other organs

(blood, heart, lung, spleen, and kidney) of either strain. After24 h, the liver of C57BL/6-WM/+ mice retained 73% of the '"In

incorporated in 1 h (213,000 cpm versus 293,000 cpm) andmost of the radioactivity was cleared from the other organsexcept for the kidney and spleen. The increase in the radioactivity in the kidney likely represents the release of free "'In.

Thus, the liver of conventional mice has only a marginal abilityto clear the tumor cells during the first 24 h. In contrast, after24 h the liver from nude mice cleared about 60% of the '"In

incorporated into the liver in 1 h (112,300 cpm versus 270,800cpm). Consistent with this observation was the 12-fold increaseof radioactivity (8800 cpm versus 700 cpm) in the blood fromnude mice compared with that in the blood from conventionalmice at 24 h.

Cytotoxicity and Cytostasis of NPC and Spleen Cells fromC57BL/6, PolydQ-treated C57BL/6, C57BL/6-/iu/+, andC57BL/6-/IH//IH Mice against Freshly Excised LD-MCA-38 andYAC-1 Tumor Cells. Previously, this laboratory and others (13-17) have demonstrated that NPC from untreated mice and ratshave NK activity against "Cr-YAC-1 tumor cells in a 4-h assay.Therefore, the cytotoxicity of NPC from C57BL/6, C57BL/6-HH/+, and C57BL/6-WM//IMmice against freshly excised "Cr-LD-MCA-38 was compared with that against YAC-1 cells. Asshown in Table 3, Experiment A, NPC from C57BL/6 micewere cytotoxic against YAC-1 tumor cells but not LD-MCA-

38 tumor cells. As shown by this laboratory previously (14),treatment of mice with anti-AsGMi, an antibody that inhibitsNK activity (20), abolished the ability of NPC to lyse YAC-1tumor cells. NPC from nude mice demonstrated significantactivity against 51Cr-LD-MCA-38 and augmented cytotoxicactivity against "Cr-YAC-1 compared with conventional mice

(Table 3, Experiment A). The cytotoxic activity against YAC-1 and LD-MCA-38 exhibited by NPC from nude mice was alsoabrogated by prior treatment of mice with anti-AsGMi (Table3, Experiment B). Ad-NPC (>90% Kupffer cells) did not exertany cytotoxicity against YAC-1 or LD-MCA-38 tumor cells.

Interestingly, spleen cells from nude mice did not display cytotoxic activity against LD-MCA-38 but did display significantlysis against YAC-1 tumor cells. NPC from athymic nude nu/nu mice also lysed LD-MCA-38 and YAC-1 tumor cells (Table

3, Experiment C). Poly(IC) which augments NK activity in theliver and spleen (25) and induces serum interferon activity (26)

1836

on June 19, 2020. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SUPPRESSION OF HEPATIC METASTASIS BY AsGM1* LIVER TELLS

was injected into C57BL/6 mice and 1 day later the cytotoxicityagainst both tumors was evaluated. NPC from poly(IC)-treatedmice induced significant activity against "Cr-LD-MCA-38 tu

mor cells and displayed augmented cytotoxic activity against"Cr-YAC-1 compared with untreated mice (Table 3, Experi

ment D).Because no cytotoxic activity by NPC from conventional mice

against LD-MCA-38 cells was detected using the short-term 4-h "Cr release assay, a longer term 72-h cytostatic ['Hjthymidine

uptake assay was used. As illustrated in Table 4, NPC fromuntreated conventional mice from 4 separate experiments exerted 30-50% cytostatic activity against LD-MCA-38 cells.This activity was not detected with Ad-NPC (>90% Kupffercells). In contrast, NPC from nude mice exerted augmentedcytostatic activity (95-97%) against LD-MCA-38 tumor cells.Furthermore, treatment of nude mice with anti-AsGMi totallyreversed the cytostatic activity of NPC in one experiment (Table4, Experiment B) and partially reversed in another experiment(Table 4, Experiment C). Poly(IC) given 1 day prior to sacrificealso boosted the cytostatic activity of NPC against LD-MCA-

38 tumor cells (Table 4, Experiments D and E).In Vivo Effect of Anti-AsGM, on the Number of Hepatic Foci

and Survival of LD-MCA-38 and CD-MCA-38 Tumor Cells inAthymic nu/nu Mice and Poly(IC)-treated C57BL/6 Mice. Previously, we have shown (27) that treatment of C57BL/6 micewith anti-AsGMi 1 day prior to and 3 days after LD-MCA-38tumor inoculation reduced survival from 35 to 17 days (P <0.005). Since anti-AsGM1 treatment of nude mice caused adecrease in the cytotoxic activity of NPC against YAC-1 cells

and LD-MCA-38 cells and cytostatic activity of NPC againstLD-MCA-38 tumor cells (Tables 3 and 4), the effect of thistreatment on the number of hepatic foci and survival of nudemice given LD-MCA-38 tumor cells was determined. As shownin Table 5, athymic nu/nu mice developed only a median of 2foci/liver and as illustrated in Fig. 2 survived an average of 60days. In parallel experiments, when these mice were given anti-AsGM, followed by ICV injection of LD-MCA-38 cells, they

developed dramatically more hepatic foci (400) and their survival time was significantly decreased to an average of only 19day (P< 0.0005). C57BL/6 mice inoculated with CD-MCA-38

cells by the ICV route did not develop any hepatic foci (day 21)and prior treatment of mice with anti-AsGMi did not inducecolonization or death (data not shown). BALB/c-nu/nu mice

were also able to resist hepatic metastasis and 80% did not havehepatic foci on day 21 (Table 4) and survived for at least 80days.

The effect of poly(IC) on the hepatic metastasis and survivalof tumor bearing mice was also determined. As illustrated inTable 5, poly(IC) given to C57BL/6 mice 1 day after LD-MCA-

38 tumor cell administration dramatically inhibited the numberof hepatic foci on day 14 (median, 400 versus 1). This treatmentalso increased their mean life span from 20 to 70 days (P <0.005) (Fig. 3). The administration of anti-AsGM) before andafter tumor inoculation and poly(IC) treatment resulted in asignificant increase in the number of hepatic foci (1 versus 400)as well as a decrease in mean survival from >70 to 24 days(Fig. 3). Poly(IC) also suppressed tumor progression and in-

Table 4 Cytostatic activity of NPC from C57BL/6-nu/+, C57BL/6-nu/nu, and poly(IC)-treated mice against LD-MCA-38 cells

TreatmentExperiment Strain* invivoAC57BL/6-/IU/+

NoneNoneC57BL/6-nu/nu

NoneNoneBC57BL/6-nu/nu

RS'RSAnti-AsGM,'Anti-AsGM,CC57BL/6-/1U/+

NoneNoneC57BL/6-nu/nu

RS'Anti-AsGmi'Anti-AsGwiDC57BL/6

NoneNoneNoneNonePolydcVPoly(IC)Poly(IC)Poly(IC)EC57BL/6

NonePoly(1C)7Poly(IC)*Cells*LD-MCA-38NPCNPC

+LD-MCA-38NPCNPC

+LD-MCA-38LD-MCA-38NPCNPC

+LD-MCA-38NPCNPC

+LD-MCA-38LD-MCA-38NPC

+LD-MCA-38Ad-NPC+LD-MCA-38NPC

+LD-MCA-38NPC+LD-MCA-38Ad-NPC

+LD-MCA-38LD-MCA-38NPCNPC

+LD-MCA-38AdNPCAd-NPC

+LD-MCA-38NPCNPC

+LD-MCA-38Ad-NPCAd-NPC

-1-LD-MCA-38LD-MCA-38NPC

+LD-MCA-38NPC+LD-MCA-38NPC+ LD-MCA-38['HIThymidine

(10-'xcpm)c29

±31±119±

21±13±032

±51±11±15±336

±522

±311±121±11±114±518±

154

±52±134

±21±164±11±111±11±150±119±314±

25±118±

2%

inhibition''409097050s953519390780307510

°Three mice/group.* NPC. poly(IC) NPC. or Ad-NPC (4 x 10'), LD-MCA-38 tumor cells (2 x IO4),or NPC. poly(IC) NPC, or Ad-NPC (4 x IO5)plus 2 x 10" LD-MCA-38 tumor

cells »erecultured in medium containing I0rr fetal bovine serum for 72 h and pulsed with 0.5 »¿Ciof ['H]thymidine for 18 h.' Mean ±SD of triplicate samples.d See "Materials and Methods."' RS or anti-AsGMi (25 ^1) was injected i.p. into 3 mice 1 day prior to sacrifice.

Three C57BL/6 mice were treated i.p. with 200 ^g poly(IC) 1 day prior to sacrifice.'Three C57BL/6 mice »eretreated i.p. with 200 »igpoly(lC) 9 days prior to sacrifice.

1837

on June 19, 2020. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SUPPRESSION OF HEPATIC METASTASIS BY AsGMi* LIVER CELLS

Table 5 Induction of experimental hepatic metastasis in athymic nude or poly(IC)-treated C57BL/6 mice after ¡CVinjection ofLD-MCA-38; effect ofRS,and anti-AsGmi

StrainAthymic

nu/nu

BALB/c-nu/nuTreatment

invivoRSCAnti-AsGM,cNoneNo.of

mice7

79Mean12 40015No.

of liverfoci/mouse"Median2

4000Range0-59 4000-60%

P* colonization57

0.0005 10022

C57BL/6C57BL/6NonePoly(IC)rfRS'Polype/

+Poly(lC)"+RSCanti-AsGMir55666150340023561380400140014-3700-134000-8138-4000.00150.0001NSr"1002010050100

" Number of liver foci was determined 14 days after tumor injection for all groups except BALB/c-nu/nu mice which was determined 21 days after tumor injection.* Wilcoxon rank sum test; compared to control.' RS or anti-AsGMI (25 ¡i\)was injected i.p. into mice I day before and 3 days after injection of 2.5 X 10' LD-MCA-38 tumor cells into the ICV.d C57BL/6 mice were given i.p. injections of 200 ng poly(IC) 1 day after injection of 2.5 x 10' LD-MCA-38 tumor cells into the ICV.' NS, not significant.

1001908070601

5°§40-viM

30

2010n.\^\

O —¿�ONu/nu•O\ •¿�•¿�Nu/nu+anti-AsGI•

1\•0i

V^S***SN^•coo0

Table 6 Effects ofpoly(IC) treatment on survival ofC57BL/6 mice giren LD-MCA-38 tumor cells by ICV route

10 20 30 40 50 60 70 80DAYSPOSTLD-MCA-38 INJECTION

90 100

Fig. 2. RS or anti-AsGMi (25 ¿il)was injected i.p. into athymic nu/nu mice 1day before and 3 days after injection of 2.5 x 10! LD-MCA-38 tumor cells into

the ICV and survival was determined.

O—OC57BI/6

•¿�•¿�Polyl-C C57BI/6\ A A Polyl-C C57BI/6+anti-AsGM1

-t- -I-

20 30 40 50 60 70 80 90 100110120130DAYSPOSTLD-MCA-38 INJECTION

Fig. 3. C57BL/6 mice were given i.p. injections of 200 ^g poly(IC) 1 day afterinjection of 2.5 x 10* LD-MCA-38 tumor cells into the ICV. RS or anti-AsGMi

(25 M')was injected i.p. into C57BL/6 mice 1 day before and 3 days after injectionof 2.5 x 10* LD-MCA-38 tumor cells into the ICV. Survival was determined in

all three groups.

creased life span if given 8 days after tumor inoculation but lostits effect if given 14 days after tumor inoculation (Table 6).

DISCUSSION

In the present study, we correlated the ability of the liver andits subpopulations from immunocompetent, immunostimu-lated, and congenitally immunodeficient mice to resist colon-derived tumor colonization after ICV administration of LD-MCA-38 tumor cells and lyse or inhibit colon tumor growth invitro. We demonstrated that (a) NPC from immunocompetent

Treatmentinvivo"None

Poly(IC)NonePoly(IC)NonePoly(IC)Day

oftreatment814No.

ofmice9

68

87

7Mean

survival(days)34.5

72.534.778.440.238.0Median

survival(days)38.077.0

34.572.538.035.0P"0.0001

0.0001NS

" C57BL/6 mice were given injections of 2.5 X 10' LD-MCA-38 tumor cellsinto the ICV and treated i.p. with either phosphate-buffered saline or 200 ngpoly(IC) 1, 8. and 14 days later.

Wilcoxon rank sum test; compared to control.

C57BL/6 mice although containing AsGM)-positive liver cellshad little or no cytotoxic activity but did exert cytostatic activityagainst freshly excised LD-MCA-38 in vitro; (b) AsGMi-posi-tive liver cells from athymic nude mice were cytotoxic andexhibited augmented cytostatic activity against LD-MCA-38 invitro and suppressed the colonization of the liver by LD-MCA-38 cells; (e) AsGMi-positive liver cells from poly(IC)-treatedC57BL/6 mice also were cytotoxic and exhibited augmentedcytostatic activity against LD-MCA-38 in vitro and inhibitedthe invasion of LD-MCA-38 tumor cells into the liver.

Spontaneous metastasis initially involves local invasion fromthe primary neoplasm and penetration of the tumor cells intothe lumen of blood vessels (28, 29). However, there are fewmurine tumors that spontaneously metastasize to the liver.Those that do are reticuloendothelial sarcomas such as M 5076and RCS2 which are histologically different from colon tumors(8, 30). Transplantable mouse colon tumors such as MCA-38and MCA-26 cells do not spontaneously metastasize to the liverbut when injected into the ICV can form hepatic foci (11, 31).Colonization or experimental metastasis represents only a portion of the metastatic cascade, the entrapment and growth inthe hepatic microenvironment. Therefore, this model is suitableto assess the role of effector cells in the liver cells to controlcolon-derived hepatic metastasis.

In our previous studies ( 11), we demonstrated that LD-MCA-38 tumor cells injected into the ICV of conventional micecolonized the liver in a stepwise and reproducible fashion. Thismodel offered several advantages over other previously described systems of experimental metastasis. It more closelyfollowed the normal pathway of human colon tumors than theintrasplenic model and required no splenectomy after tumorinoculation. MCA-26 administered by the ICV route also resulted in colonization of the liver (31). However, this tumor

1838

on June 19, 2020. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SUPPRESSION OF HEPATIC METASTASIS BY AsGM1* LIVER CELLS

formed a diffuse pattern of metastasis in the liver which couldnot be easily quantitated.

Our previous results (27) have shown that untreated C57BL/6 mice do mount an antimetastatic response in vivo against LD-MCA-38 tumor since anti-AsGMi treatment significantly reduced their life span from 35 to 17 days. In addition, NPCfrom C57BL/6 mice exerted cytostatic but not cytotoxic activityagainst LD-MCA-38 cells in vitro. However, NPC from thesemice can lyse the standard NK target YAC-1 and lysis could beabrogated by prior treatment with anti-AsGM, (14, 20). Itoh etal. (17) demonstrated using immunofluorescence that non-adherent NPC from C57BL/6 mice contained 18% AsGM1-positive lymphocytes. Thus, resident AsGMi-positive cells present in the livers of conventional C57BL/6 mice do not exert adirect cytotoxic effect on freshly excised tumor cells in shortterm 4-h assays. Other studies have also shown that freshlyexcised tumor cells are resistant to natural cytotoxic activity(32, 33). The resident cytostatic activity exhibited by AsGM,NPC from conventional mice in vitro may not prevent tumorimplantation or colonization of the liver but may instead contribute to suppression of tumor growth. However, the cytostaticactivity displayed by NPC does not appear to be due to conventional NK cells since NPC from NK-deficient beige mice wereas cytostatic against LD-MCA-38 as those from their conventional littermates. Furthermore, no difference was seen in thesurvival of beige mice compared to their matched littermates inresponse to ICV-administered LD-MCA-38 cells (data notshown). It is known that AsGM, cells exist in the livers of micedevoid of NK activity (21, 22).

Wiltrout et al. (34) have shown that anti-AsGNn treatment of

C57BL/6 mice exhibited increased formation of experimentalmetastasis in the lung and induced metastasis in the liver afteri.v. challenge with B16 melanoma. This increased metastasiscoincided with decreased blood, spleen, and liver NK (YAC-1)activity. In contrast to LD-MCA-38 tumor cells, B16 tumorcells are from a cultured cell line which is usually sensitive toNK cytotoxicity. Since cultured tumor cells tend to drift fromtheir original phenotype (35), LD-MCA-38 may be a morerelevant tumor to study. Furthermore, LD-MCA-38 tumor cellsgrown in culture (CD-MCA-38 cells) do not produce hepaticcolonies following ICV injection. This inability of CD-MCA-38 cells to colonize the liver is not the result of susceptibilityto AsGMi-positive cells since the abrogation of in vivo AsGM¡activity did not result in colonization of the liver (data notshown). The reason for the lack of colonization by CD-MCA-38 cells may be more dependent on other properties of thetumor such as their decreased motility rather than an increasein susceptibility to immune attack (36).

Nude mice have a significantly greater capacity to resisttumor colonization than untreated conventional mice. The liverfrom nude mice may more efficiently suppress or prevent thetumor from initially seeding the liver. In this regard, the liverof C57BL/6-/ÕH//1Hmice can clear '"In-LD-MCA-38 cells at afaster rate than the liver of C57BL/6-HH/+. Furthermore, whennude mice were pretreated with anti-AsGMi antibody, then thesemice survived about the same length of time as anti-AsGMi-treated conventional mice (19.5 versus 17 days, respectively).Since the cytotoxic and cytostatic activity in the liver of nudemice reappeared 7 days after cessation of antibody treatment(data not shown), this suggests that AsGM,-positive cells fromnude mice inhibit tumor implantation or early tumor growthbut not established tumors.

There are many reports that demonstrate that N K cells fromnude mice have greater cytotoxicity than from conventional

mice (37, 38). However, more efficient cytotoxicity may not beentirely responsible for the antimetastatic effect since splenicNK cells from nude mice which have greater cytotoxicity againstYAC-1 cells than spleen from conventional mice still do notlyse freshly excised LD-MCA-38 cells. It is likely the microen-vironment of liver from nude mice may be different from thespleen. In this regard, NPC from nude mice produced detectableIFN-«//i(~20-40 units/ml) in culture while spleen cells do not(data not shown). We speculate that hepatic NK cells fromnude mice are activated in situ in part by local production ofIFN-«//3which broadens their specificity and directly lyses LD-MCA-38 cells. This is consistent with the fact that freshlyexcised tumors have been shown to be susceptible to IFN-activated NK cells (32, 33).

The interferon inducer, poly(IC), injected into C57BL/6 miceinduced cytotoxicity and augmented cytostasis against LD-MCA-38 tumor cells in vitro. Furthermore, the number ofhepatic foci in the liver of poly(IC)-treated mice was significantly reduced when measured 14 days after tumor inoculation.When mice were given anti-AsGM, prior to poly(IC), then the

number of hepatic foci increased to those of untreated mice.Poly(IC) treatment of mice bearing covert or micrometastasis(day 8) was also effective (Table 6). In contrast, poly(IC) hadno effect if administered to mice bearing visible hepatic tumors(day 14). AsGMi-positive cells have been shown to contributeto the therapeutic efficacy of IL-2 against MCA sarcoma soonafter implantation. However, once tumors were established,AsGMi-positive cells had no detectable role in mediating tumorregression (39). As in the case with nude mice, AsGMI-positivecells from poly(IC)-treated mice may influence tumor implantation but have little or no influence on established tumors.

Kupffer cells did not exhibit antitumor activity in vitro. Ad-NPC from nude mice which consist of greater than 90% Kupffercells did not lyse or prevent the proliferation of LD-MCA-38cells (Tables 4 and 5). Furthermore, the administration of anti-AsGivn into mice or treatment of NPC with anti-AsGM, pluscomplement does not abolish the number of Kupffer cells ortheir phagocytic activity (14, 17). Anti-AsGMi also does notinhibit the tumoricidal activity of macrophages (40). AsGMitreatment of nude mice led to increased numbers of hepatic focibut no extrahepatic métastasessuggesting that other mechanisms may operate to prevent the release of tumor from theliver.

In conclusion, the experiments in this paper indicate thatAsGM,-positive cells in the liver from conventional mice do notprevent tumor implantation but have the capacity to slow focaltumor growth. In contrast, AsGM|-positive cells in the liverfrom nude or poly(IC)-treated mice which differ from conventional AsGMi-positive liver cells can suppress tumor implantation or early tumor growth.

ACKNOWLEDGMENTS

The authors wish to thank Paul Evans and Norman Paolini for theirexcellent technical help.

REFERENCES

1. Bengtsson, G.. Carlsson, G., Hafstrom. L., and Jonsson. P. Natural historyof patients with untreated liver métastasesfrom colorectal cancer. Am. J.Surg., 141: 586-589. 1981.

2. Carter. S. K. Adjuvant chemotherapy of large bowel cancer. In: S. K. Carter(ed.). Principles of Cancer Treatment, pp. 39-56. New York: McGraw-HillBook Co.. 1982.

3. Stehlin, J., Ipolyi, P., Greeff. P.. McGaff, C., Davis, B., and McNary, L.Treatment of cancer of the liver. Ann. Surg., 20«:23-35, 1988.

1839

on June 19, 2020. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SUPPRESSION OF HEPATIC METASTASIS BY AsGMi* LIVER CELLS

4. Weiss, L., and Mayhew, E. An approach to the therapy of métastasesfromcancer of the upper rectum: A working hypothesis. Cancer Drug Delivery. 2:19-54, 1985.

5. Weiss, L., Haydock, K., Pickering. J. P.. and Lane. W. Metastatic patternsin patients with carcinoma of the lower esophagus and upper rectum. InvasionMetastasis, 4: 47-60, 1984.

6. Tsuruo. T.. Tamori, T.. Naganuma. K.. Tsukagoshi, S., and Sakurai. Y.Characterization of metastatic clones derived from a metastatic variant ofmouse colon adenocarcinoma-26. Cancer Res., 43: 5437-5442, 1983.

7. Dunnington, D., Buscarino, C., Gennaro, D., Greig, R., and Poste, G.Characterization of an animal model of metastatic colon carcinoma. Int. J.Cancer. 39: 248-254, 1987.

8. Williams, D., Sherwood. E.. McNamee. R., Jones. E., Browder, I., andDiLuzio, N. Chemotherapy of experimental hepatic métastases.Hepatology,7: 1296-1304, 1987.

9. Giavazzi, R., Jessup. J., Campbell. D.. Walker, S.. and Fidler. I. J. Experimental nude mouse model of human coloréela!cancer liver métastases.J.Nail. Cancer Inst., 77: 1303-1308, 1986.

10. Giavazzi, R., Campbell, D., Jessup. J., Cleary, K., and Fidler. I. J. Metastaticbehavior of tumor cells isolated from primary and metastatic human colorec-tal carcinomas implanted into different sites in nude mice. Cancer Res., 46:1928-1933. 1986.

11. Goldrosen, H., Paolini. N., and Holyoke, D. Description of a murine modelof experimental hepatic métastases.J. Nati. Cancer Inst., 77:823-828. 1986.

12. Wisse. E., and Knook, D. The investigation of sinusoidal cells. A newapproach to the study of liver function. In: H. Popper and F. Schaffner (eds.).Progress of Liver Disease, Vol. 6, pp. 153-175. New York: GruñeandStratton, 1979.

13. Cohen, S. A., Salazar, D.. and Nolan. J. P. Natural cytotoxicity of isolatedrat liver cells. J. Immunol., 129: 495-508. 1982.

14. Cohen, S. A., Salazar, D., von Muenchhausen. W., Werner-Wasik, M., andNolan, J. P. Natural antitumor defense system of the murine liver. J.Leukocyte Biol.. 37: 559-569. 1985.

15. Leung, K.. Salazar. D.. Ip. M., and Cohen. S. A. Characterization of naturalcytotoxic effector cells isolated from rat liver. Nat. Immun. Cell GrowthRegul.. 6: 150-159, 1987.

16. Kaneda. K.. Dan. C., and Wake, K. Pit cells as natural killer cells. Biomed.Res., 4: 567-576. 1983.

17. Itoh, H., Abo, T., Sugawara. S.. Kanno. A., and Kumagai, K. Age-relatedvariation in the proportion and activity of murine liver natural killer cellsand their cytotoxicity against regenerating hepatocytes. J. Immunol., 141:315-323, 1988.

18. Richman, L., Klingenstein, R.. Richman. J., Slrober. W., and Berzofsky. J.The murine Kupffer cell: 1. Characterization of the cell serving accessoryfunction in antigen specific T cell proliferation. J. Immunol., 123: 2602-2609, 1979.

19. Kasai, M., Iwamori. M., Nagai, Okumura. K.. and Tada. T. A glycolipid onthe surface of mouse natural killer cells. Eur. J. Immunol.. 10: 175-180.1980.

20. Kasai, M., Yoneka, T., Habu, S., Maruyama, Y., Okumura, K., and Toku-naga, T. In rivo effect of anti-asialo <i-,, antibody on natural killer activity.Nature (Lond.). 291: 334-336, 1981.

21. Shimamura, Kasai, M., Okamura. K., Fukiu, A., Akasuka, A., Okumura, K.,and Taaoki, N. Morphological and functional study of ganglio-A'-tetrasylgly-ceramide positive lymphocytes which mediate natural killer activity. J. Nati.Cancer Inst., 68:449-455. 1982.

22. Schwarting, G., and Gajewski. A. Glycolipids of murine lymphocyte subpop-ulations: a defect in the levels of sialidase sensitive sialosylated asialoGM> in

beige mouse lymphocytes. J. Immunol.. 126: 2403-2409. 1981.23. Kaplan, A. Cytostasis of tumor and non tumor cells. In: D. Adams. P.

Edelson, and H. Koren (eds.). Methods for Studying Mononuclear Phagocytes, pp. 775-783. New York: Academic Press. 1981.

24. Wiltrout, R.. Gorelik, E.. Brunda. M., Holden. H., and Herberman, R.Assessment of in vivonatural antitumor resistance and lymphocyte migrationin mice: comparison of '"-iododeoxyuridine with '"indium oxide and"chromium as cell labels. Cancer Immunol. Immunother.. 14: 172-179,

1983.25. Werner-Wasik, M.. von Muenchhausen, W., Nolan. J., and Cohen, S. A.

Endogenous interferon a/ßproduced by murine Kupffer cells augments liver-associated natural killer activity. Cancer Immunol. Immunother., 35: 103-112, 1989.

26. Buckler, C.. DuBuy, H., Johnson. H., and Baron, S. Kinetics of seruminterferon response in mice after single and multiple injections ofpolyl:polyC. Proc. Soc. Exp. Biol. Med.. 136: 394-401. 1971.

27. Cohen, S. A., and Goldrosen. M. Modulation of colon-derived hepaticmetastasis by murine nonparenchymal liver cells. Immunol. Invest., IS: 351-363, 1989.

28. Fidler, I. J.. Gersten, D., and Hart, I. The biology of cancer invasion andmétastases.Adv. Cancer Res., 28: 150-251, 1978.

29. Nicolson, G. L. Organ specificity of tumor metastasis: role of preferentialadhesion, invasion and growth of malignant cells at specific secondary sites.Cancer Metastasis Rev., 7: 143-188, 1988.

30. Jones, C., and Goldrosen, M. H. Characterization of a murine tumor ofspontaneous origin with selective hepatic metastasis. Invasion Immun., 7:144-157, 1987.

31. Tan, M., Holyoke, D., and Goldrosen. M. H. Murine colon adenocarcinomas:syngeneic orthotopic transplantation and subsequent hepatic metastasis. J.Nail. Cancer Inst.. 59: 1537-1544, 1977.

32. Pattengale. P.. Gidlund. M., Nilsson, K.. Sundstrom. C.. Sallstrom, J.,Simmonsson. B., and Wigzell, H. Lysis of fresh human B-lymphocyte-derivedleukemia cells by interferon activated natural killer cells. Int. J. Cancer. 29:1-7. 1982.

33. Moore, M., Taylor, G., and White, W. Susceptibility of human leukemias tocell-mediated cytotoxicity by interferon-treated allogeneic lymphocytes. Cancer Immunol. Immunother.. 13: 56-61. 1982.

34. Wiltroul. R., Herberman, R.. Zhang. S., Chirigos. M., Ortaldo. J., Green.K., and Talmadge, J. Role of organ-associated NK cells in decreased formation of experimental métastasesin lung and liver. J. Immunol., /.(-/: 4267-4275, 1985.

35. Hewitt, H. Animal tumor models and their relevance to human tumorimmunology. J. Biol. Response Modif., /: 107-119. 1982.

36. Maslow, D., and Goldrosen, M. H. Relationship of in vitro cell motility andcolonization potential in a mouse colon adenocarcinoma (MCA-38) cell line.Invasion Metastasis, 8: 133-142, 1988.

37. Herberman, R.. and Holden, H. Natural cell-mediated immunity. Adv. Cancer Res., 27: 305-354. 1978.

38. Habu, S.. Shimaru. K., Tamaoki, N.. Kumagai. Y., and Okumura, K. Twonatural killer cell populations which belong to T cell lineage in nude mice.Eur. J. Immunol., 16: 1453-1456. 1986.

39. Mule, J., Yang. C., Lafreniere. R., Shu, S., and Rosenberg, S. Identificationof the cellular mechanisms operational in vivo during the regression ofestablished pulmonary métastasesby the systemic administration of highdose recombinant interleukin 2. J. Immunol., 139: 285-292, 1987.

40. Riser, B., Laybourne. K.. and Varani, J. Treatment of mice with anti-AsGMIantibody or poly-I:C: effects on metastasis dissociable from modulation ofmacrophage antitumor activity. Nat. Immun. Cell Growth Regul.. 7: 305-315. 1988.

1840

on June 19, 2020. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1990;50:1834-1840. Cancer Res   Stefan A. Cohen, Shie-pon Tzung, Ralph J. Doerr, et al.   Colon-derived Experimental Hepatic MetastasisPolyinosinic-Polycytidylic Acid-treated Mice in Suppressing

Positive Liver Cells from Athymic Nude orM1Role of Asialo-G

  Updated version

  http://cancerres.aacrjournals.org/content/50/6/1834

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/50/6/1834To request permission to re-use all or part of this article, use this link

on June 19, 2020. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from