Experimental colorectal liver metastases

53

Experimental

colorectal liver metastases

Influence of sex, immunological status and liver regeneration Y. Panis’,B. Nordlinger‘,*,R. Delelo’, J.P. Her&, J. Infante’, M. Kuhnle* and F. Ballet’ ‘INSERM U.18,. *Cm,,c dr Chin‘,@ Di~c~riveand 31NSERM 0.535.“~~i,~,soin,.anroi”r

Pm& Froncp

(Received 2.5May 1989, (AEecptcd 25hwaly w-0)

The liver is the most frequent site of metastares from colon cancer. To improve our knowledge of liver metastases and to develop tte.w adjuvant therapies, a good animal model is necessary. The aims of this study were to obtain a model of liver metmaser with intraponal injection of colon adrenocarcinoma cell aggregates (DHDKl2 cell line) and to study the effect of various !actots, i.e., sex, liver regeneration and immunosuppression. 011the development of liver metastasis. Cell aggregates were injected into the portal vein of 59 syngenic male and female BD IX rats following randomization into !hree groups. Group 1. (control 12 males and 10 females) received only cell aggregate& group 2 (12 males and 10 females) undetwent a 70% bepatectomy before cell injection; gmup 3 (IS males and 10 females) received cychx4x1rin A injections at a doje of 10 tug/kg per day for 28 days following cell injection. Autopsy was perfomted at 10 weeks. Livermetastaseswere morefrequent in the male rats ia group 3 than in those in group 1 @OS vs. 30%. p < 0.04). The rate of liver metastases in females was not increasea by immunosuppression

(22.2% vs. lZ.JW, N.S.). Liver resection (group 2) did not significantly

modify the incideta of liver metastasis. No female had liver metastases in this group. This relatively simple model rapidly produces liver mnasrasis with a high yield, but only in male rats. Besides sexual factors, immunosuppression also increased the rate of expetimental liver metastasis. while liver regeneration failed to do so.

The pmgttc& for c&rectal

cancer has not improved in

recent years and the survival rate at 5 years remains close

potlal

infruionaf S-fluorouracik

afler resection of a colon

to 40% after surgical excision (1). The liver is the mast frequent site 0: meta.stases from colon cancer. Metastases are present in 20% of the cases at the time of excision of ptimatyoolcm cancer, and appear later in a futtber2O% of patients. Once liver mewstases have occutted, surgical excision is the only method to prolong survival in some pa-

adenocarcinatta (5). These results remain to be confirmed. Unfommately, there is no gocd animal model to help improve OUTknowledge of the behavior of such tumors and to develop new treatment modalities. In a previous study (6). azoxymethane, a compound derived ftum Cimetbylbydraltine, was used to induce colon adenocati-

tients (2). With few exceptions, results with adjuvam chemotherapy for mlorectal cancer have been dissppobtting (3.4). Taylor, however, has observed a statistically significant decrease in tbe incidence of liver metastasis and a prolonged survival in cenain patients who had rxxived intra-

nomas. After surgical excision of the primary tumor, liver metastases occurred spontaeouly in 22% of the animals. Although the model mimicked the natural history of the development of liver metastases in humans quite well, the yield was ttw low for use in all apedmental studies on c&rectal liver metastases. While the azoxymethane

54 model is appropriate for testing new adjuvsnt therapies to prevent liver metastais after surgical excision of primary cancer, a model with a higher yield is necessary to study the development and treatment of patent liver metas-

denvent a 70% hepatectomy 5 min before i.p. injection of the cell aggregates. Group 3 rats (15 males end 10 females) were given cyclosporin A (Sandoz, France) in olive oil at a dose of 10 mg/kg body weight daily by subcu-

tases. Patent liver metastasa can be achieved by direct injection ofcancercells into the portal system (7). The aim of this studv was two-fold. First, to obtain e mproducible high-yield model of liver metastases from colon cancer and second, to study the effect of sex, liver regeneration and immunosuppresrion on the development Of such metestases.

teneous injections for 4 weeks. This treatment cornmenced 48 h after i.p. injection of the cell aggregates. A laparotomy was performed for complete abdominal exploration at 4,6 and 8 weeks after i.p. injection of cells. At 10 weeks. all the animals were killed. The oeritoneal cavity end the thorax were carefully examined; the liver

A transplantable colon carcinoma cell line, using cell aggregates, was administered in the portal vein a~cordii toVan derElst (i), inordertoobtein ahigherlivermetes-

and any suspected visceral or peritoneal metastases subjected to systematic histologic examination.

were

with the injection of

A Fisher exact test with Yates correction was used to compare the number of rats with liver metestares in the different groups.

Syngenic male and female BD IX rats (CNRS, France) werre used. At the beginning of the study. body weight war 350 g for the males and u)o g for the females. The animals were given free access to standard laboratory chow

Fifty-nine rats were evaluable: three early postaperative deaths due to bleeding at the puncture site of the partel vein occurred (one male end one female in group 1 and

tasis yield than is currently isolated cells.

obtained

Msderinls and Metbeds

and Water throughout

the experiment.

one female in group 3). Two rats died during the experiment and could not be evaluated acettrately (One male and one female in group 1). In group 2, five females died following surgery.

The cells were provided by F. Martin (Dijon, France) and obtained from an established transplantable carcinama line originating from a 1,2-dimethylhydrazine-induced colon adenocarcinoma in syngeneic ED IX rats (8.9). They were maintained in monolayer culture with a culture medium consisting of Ham’s FlO (O.S.I.), 10% fetal calf serum (Gibco), 5% t_-glutamine, 100 IU penicillin/ml and 100 #g streptomycin/ml. Single cells were obtained by trypsinization of a confluent monolayer.

According to the technique described by Van der Elst (7), approx. 500 aggregates (150-200 pm in diameter) were obtained from the incubation of 1.106 cells on a giratory shaker for 48 h. Using diethyl ether anesthesia and a clean surgical technique, the abdomen was opened through a 4 cm midline incision. Approx. 2.50 aggregates (SW.101 cells) were injected into the portal vein through a minicatheter (buttefiy No. 21, Abbott). The site of injection was covered with a hemostatic sponge. ‘Ike animals were then randomly divided into three groups: group 1 (the control group) (12 males and IO females) received only an intrapatal (i.p.) injection of cell aggregates; group 2 rats (12 males and 10 females), un-

Group I No macroscopic

liver metestases

were observed

at 4,6

or 8 weeks. On death (10 we&s after the cell injection), metestases from colonic adenocabwma, confirmed by bistopmhoiogic examination, were &served in four ratx three males had macroscopic liver metestases. Two of these had two livermetastnaes (2 and3 mm) and one had a

.:

single large liver metastasis (20 mm). In one these rfd8, associated lung metastasis was present. One female had two liver metestases (1 mm). No liver metestaSes were discovered during macroscopic and histologic exambaation of the livers of the remaining I4 animals. Group 2 No macroscopic liver metastascs were observed at 4,6 or8 weeks. On death of the animals, metastawswete conlirmcd by pathologic examination In only one male rat which had a single macroscopic liver metastasis (5 mm). No liver metestases were discovered in the temainii 16 animals in this group. The number of rats with liver metastases in group 2 was not statistically different from that in group 1 (Table 1). Autopsy post-operativedeaths.

was negative in the five early

55

EXPERIMENTAL COLORECTAL LIVER METASTASES

trapottai injection of cell aggregates derived from a colon carcinoma cell line into syngeneic rats. After immuttosuppression with cyclosporin A, the number of liver metastases was significantly higher, with 80% of the male rats affected. This clearly shows the role of immunity and sex in the development of liver metestases. Cell aggregates wete ured instead of singlecell suspensions (7) because the injection of single cells into the pmtal vein pmduw associated extmhepatic metastxes. particularly in the lungs, due to celk escaping from the liver sinusoids. In out study only one ret had lutg metastases,

3 mm); notte bad

which is similar to the situation in humans, where liver metastases from colon cancer are mote frequent than lung metastexs. Injecting aggregates tether t!tan single cells may also mimic what-in humam. For example, Ra bzts et al. repott a series of 100 cancer patients with cancer cells in the vecau blood. Although some single cancer cells were observed, the majority were aggregated in clumps, particularly in the seven patients with c&m cancer (10). Furthemtore, when single cells are wed in animal models (1 to 1C miltion cells/rat), a *eatn number must be injected (11.12). In our study, the results were obtained with 0.5 million ce,Wrat. After partial liver tmection, the regetteraing liver releases gnvth factam which promote hepatic mitotic ectivity. These facton have been shown to increase the growth

earlier in this group than in

rate of sukutatteonsly implanted tumors and hepatic metastases of some temors (13,14). Fisher 0lnerve.d that a

Group 3 No macroscopic liver tttetastaeee were obsetved at 4 weeks. Six weeks after the injection of cells, 7115 males (47%) and 219 females (22%) had macroscopic liver metesteses. At 8 weeks, ll/lS males (73%) end the same two females had mescroscqic liver metestases. On death of the animals, mctastaserwere observed in WI5 males (80%) and 2/9 females (22%). In the males, the metas&es were nutnetow end &ge (mean dieter 8 mm); none had exttabepatic metastases. In the females, themetastavs wete smeller (mean diameter exwahepatic m&stases. Liver metastases

appeared

gmup 1 (Table 2). The mtmher of male rats with liver metastpges was significently

higher

than in group

0.04). (Table 1). In the females, the difference grcups 1 and 3 was not statistically sig&icant.

1@ c between

Out data show that Liver m&stases can be obtained within 10 weeks in our experimental model involving itt-

70% hepatetiomy

performed

01 eftet the intteportal

in the tat within 24 h before

injectfott

from a Walker ‘256cmci imsarcotna,

of tumor cells derived promoted

the devel-

opment of liver metastases, with the greatest effect occurring when the resection was performed 5 min aftet the inoculation of the cancer cells (15). In the present study, a 70% hepatectomy was performed 5 min before i.p. injection of the cells rather tkett efter, in order to avoid losing e large proportion

of the injected cells in the resected sped-

men. The number of liver tttetastases observed wes in fact reduced after partial liver resect&n. This may be explained by the liver simwidal obstruction and acute partal hypertension resulting from the migration of all the injected cell aggregates to the 70% reduced liver. The same phenomenon may have been responsible for the early post-operative deaths observed in this group (particularly in females with smaller livers) and for the reduced liver regeneration and tumor promotion observed in the 31iviting rats. The immtmosuppressioe induced by cyclosporin A promote* metastatic dissemination in several experimental tumors in the rat (16). Subattaneow admioistration of cyclosputin A was used es it is well tolerated, easy to administet and yields repmduaile cyckqatin A plasma levels

Y. PANS

56

et

a,.

with an adequate peak and few variations with time (17). In the present study, cyclosporin A administration was associated with a significant increase in the number of liver metastases in the male rats (Table 1). Van der Elst abserved that the host cell infiltration which surrounded experimental liver metastases was reduced in previously cyclosporin A-immtmos~pp~es~ed animals (7). The level of host cell infiltration around the liver metastases, due to the high immunogenicity of the DHDKlZ celi line, seems to be related to the metastatic potential of the tumor. In our study, the effect of cyclosporin A which inhibits T-

strayed by the host. I, can thus be hypothesized that ‘dormant’ cells begin to proliferate when immunosuppression is induced by cyclosporin A, by repeated surgical pmce-

lymphocytes confirms the role ofcellular immunity in metastatic spread and growth. In the animals receiving cyclosporin A, the meta~tases were larger and more “umerous than in the control group. Other factors have been shown to modulate the growth of liver tnetastases. The stimulation of the Kupffer cells can reduce the number of liver metastases after i.p. cell injection (18). La Frdnibre, injecting colon cancer cells into the splenic vein of rats prior to splenectomy observed a liver metastasis frequency comparable to ours. This ten be explained by the greater number of cells injected Cl@‘) and the immunosupprev sive effect of splenectomy (11). It is well known that the operative period is critical for

in females, but the ~eascms are unclear (21). The incidence of experimental colonic carcinomas in male rats CRD be decreased by castration and increased by hormonal substitution (22,23). For other organs, certain estrogens, e.g., tamoxifen (24), can inhibit the growth and the mtdig“ant trattsfommtion of liver tumors produced by acetylaminofluorene in Fischer 344 rats (25). Steroid hormone receptors, present in normal colonic ntucosa, exist in greater numbers in colonie cawxs, both in humans and in experimental models (2.5). However, the biologica significence of andmgen receptors in wltmic cxcbmgenesis is unknown. In our model, the difference betweeo males and females appeared only after immunosuppression with cyclosporin. The possibility remains that the metabolism

the development muncdepression.

of metastases (19), due to profound imFisher has shown that animals under-

going imraportal

injection of tumor cells develop more ttt-

dttres or by liver resections (15). This would also explain why some supposedly curative liver resections performed for liver metastases in humans are followed by early recurrence in the remaining liver (2). In the present study, cyclosporin A increased the rate of liver metastasis in males only, suggesting B role of sexual factors. Chemically induced primmy colon cancas in rats are known to develop more frequently in males than

of cyclospmin A is different in males and females, leadbtg to a higher promoting effect in males.

mars when the injection is followed by several explorato. ty laparotomies (13). Eccles found that cyclosporin A had no effect on the growth of primary tumors in the rat, but promoted distant tnetasrases even after excision of the primary tutnor (16). These findings, as well as our own results, support the existence of ‘domtant cells’ in the liver (20) which remain in a dormant state and are not de-

1 Faivre J, Milan C, Munsch P, Hillon P, Boutran MC, Klepping c. Le cawercolorectal dans ,e d$Iartemo”t de ,a cLite&Or. Bull Cancer 1984:7,: SO-56

The DHDKl2 cell line was a generous pin from Dr. Martin (Dijon, France). We are grateful to Sattdoz Ltd. France). We are grateful to Sandoz Ltd. (France) for providing cyclospain A.

EXPERfMENTAL

COLORECTAL

LNER

METASTASES

ercnse to liver injury. Ann Surg 1959; 193: 731-44. 14 Rwga J, Tsmaka N. Jeppwm B, Hargenrraod I, Bengmark S. Tumor gwti in liver atrophy and growth. An expm+meatal smdyiaratr. Eurf CancsrCliaOwol1985:21: 135-40. 15 FisherB, Fisher E. Experimental studies of facton influencing hcpatic mctastascs. II. Effect of partial hepatectomy. Cancer 1959; 12:929-32. 16 Ealer S.4, Hcrckfwd SE, Alexander P. Effect of cyclmparinc A on the gmwtk and rpontaneolu metastasis of ryllgeaeic timal tumors. BrJ Caxer1980;42: 252-M. 17 Wassef R, Cohen 2, Lager 8. Pbamtacoktnnic pmfikx of cy~Iqwrfne in rats. Transplant&m 1985: 40: 489-93. 18 Pearson HJ. Andersw J. chamberlain J, Bell PRF. The effect of Kupffcr cell atimulatin or depression on tbe development of liver qelplitass in the rat. Cancer Ymmuno! Jmmunother 19g6; u: 21616. 19 Were JL, GilbcrtwmEM, Sytjala SE, Starling JR. Provendon of rat c&n cancer met85taz.s by periopcrative immunostimulatioo. slugcry 1984; 96: 420-m.

51 20 W&r B, Fisher E. Expetiemaf evidence in supfmrtof the dormant tumor cc”. Scic@a 195% 130: 918-19. 21 Lament IT. o%o,,,,aa IA. Experimenta! mkm caocer. Gastraenterology 197B; 75: twl-69. 22 Izbicki JR, Schmitz R, Kantran D, fzbfcki W. Andmgcns asprcnwters of colon carcinogencsis. Cancer Detect Rev I%% 6: x5-62. 23 Mow RC. Frickr CM. Influence of gonadal h-ncs nod age on I.2-dimethyl-hydmztne&iwccd carcimgmnis. Cacxer 1pTI: 140: 2%-s 24 Misbkin SY. Farkr E, Kor R. Evidence for the bonnone depndewy of hcpatk hypqlatic ncduks. fnhibftian of malignard transformation aher exogemrus 178_crtradiol and tamoxifen. Hepatotogy 19e3: Y 3&i-16. 25 lzbicki JR, Wamhach G, HardIton SR, Hsmiscb E, Hogenschlnz R. trbicki w, Kwche I. Aodrcgen reapton in apedmentally induced c&m rarcbmgenedr. J Cancer Res Clin Owl 1% 112: 39-46.