(CANCER RESEARCH 51, 880-885, February 1, 1991]

Effects of Indomcthacin, Cytokines, and Cyclosporin A on Tumor Growth and theSubsequent Development of Cancer Cachexia1

Johan Gelin, Christer Andersson, and Kent Lundholm2

Department of Surgery, Institution I, University of Gothenburg, Sahlgrenska Hospital, S-413 45 (iothenburg, Sweden

ABSTRACT

C57BL/6J mice bearing a low or undifferentiated rapidly growingtumor were treated daily with either i.p. injections of the recombinantcytokines interleukin(IL)-la, II.-1J (21 ng/g), II -2 (21 ng/g), and tumor

necrosis factor a (21 ng/g) or s.c. injections of cyclosporin A (60 ¿ig/g)oriiulomci lud n (l Mg/g)-In some experiments, indomethacin was administered in the drinking water corresponding to the amount of s.c. injections. Survival and the time course of tumor growth and food and waterintake were measured. The nutritional state (body composition) of theanimals was registered at spontaneous death in the course of tumordisease.

Indomethacin prolonged survival from 14 ±1 to 22 ±1 days in tumor-bearing mice when administered either in the drinking water or as s.c.injections. This effect, which was due to tumor growth inhibition, wasequally effective irrespective of whether indomethacin was instituted onDay 1, 5, 7, or 9 following tumor implantation. Indomethacin did notinhibit tumor cell growth in vitro. Indomethacin-treated tumor-bearingmice were also less anorectic than untreated tumor-bearing mice, andtheir nutritional state, particularly lean body mass, was significantlyimproved by indomethacin at doses (1 ng/g) that did not influence thefood intake or body composition in non-tumor-bearing mice. At spontaneous death, indomethacin-treated tumor-bearing mice had a significantlylarger tumor burden when accounting for their degree of malnutrition ascompared with untreated tumor bearers. Indomethacin did not decreasethe elevation in hepatic concentrations of RNA seen in response to tumorprogression. Adherent peritoneal macrophages from tumor-bearing micehad a lower prostaglandin 1 ..synthesis compared with macrophages fromnon-tumor-bearing controls in the basal state (1100 ±ISO pg/10' cellsversus 3700 ±922 pg/106 cells). Lipopolysaccharide stimulated macro

phages from tumor-bearing mice to produce significantly more prostaglandin 1 _.in vitro compared with control macrophages (39,500 ±4208pg/10' cells versus 12,500 ±4330 pg/10' cells). IL-la and IL-1/3 suppressed tumor growth in vivo, which tumor necrosis factor a, IL-2, orcyclosporin A did not, but this effect of II -I was probably secondary topotentiating anorexia in tumor-bearing mice. II -I had no effect onsurvival in tumor disease.

We conclude that many previously reported effects on tumor growthby cytokines can be ascribed to alterations in food intake and hostsubstrate metabolism if not otherwise confirmed, which was done forindomethacin in this study. Positive host effects and tumor inhibition byindomethacin in this model represent so far an outstanding monodrugexperimental therapy that has no equal alternative that we know of. Wealso conclude that immune control of tumor growth is insignificant in thismodel, although the tumor is highly "immunogenic."

INTRODUCTION

The integrated response in vivo to factors restricting or promoting tumor growth is essentially unknown in experimental

Received 5/7/90; accepied 11/9/90.The cos)s of publication of ¡hisarlicle were defrayed in par) by ¡hepayment

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 in part by grants from the Swedish Cancer Society (93-B89-22XA, 20I4-B88-01XA. 2147-B89-04XA), the Medical Research Council (B89-17X-00536-25A. B89-17K-08712-01A), Tore Nilson Foundation. Assar Ga-brielsson Foundation (AB Volvo. Sweden), Jubileumskliniken Foundation, In-gabritt and Arne Lundbergs Research Foundation. Swedish and GothenburgMedical Societies, and the Medical Faculiy. University of Gothenburg. Sweden.

2To whom requests for reprints should be addressed, at Department ofSurgery, Sahlgrenska Hospital, S-413 45 Gothenburg, Sweden.

and clinical cancer. It is believed that the immune system has asignificant role to restrict tumor progression in most circumstances. In this context, cytokines such as IL-1,3 IL-2, and TNF

have been suggested as critical factors for the up regulation ofthe immune system against certain tumors (1). The proposedmechanisms behind positive treatment effects by cytokinesagainst tumors vary, but one effect may be that down regulationin the expression of the IL-1 and IL-2 gene that occurs incancer-bearing hosts is compensated for by exogeneous supplementation of cytokines (2, 3). Thus, exogeneous cytokineswould simply support the host in its deficiency of endogenousproduction. In recent studies we have, however, demonstratedthat in vivo administration of neutralizing antibodies to tumor-bearing animals, either against the high affinity receptor for IL-1 or against the TNF molecule, in fact caused tumor growthinhibition (4, 5). These results suggested to us that both IL-1and TNF, which are produced endogenously in many tumorsincluding our model, may also act as tumor growth factors inaddition to being immune stimulatory. However, such a theoryis conceptually in disagreement with several experiments demonstrating that in vivo administration of cytokines has tumor-inhibitory effects (6).

In addition, we and others have speculated that increasedendogenous cytokine production explains the development ofcancer cachexia during the progression of a tumor (7, 8). Arecent study where the TNF gene was inserted close to an activepromoter in tumor DNA, leading to endogenous TNF synthesiswith significant increase in circulating plasma concentrationsof TNF, was in support of this theory (9). Studies demonstratingthat injections of recombinant cytokines to normal animalscause tissue wasting, which appears to be comparable to spontaneous cancer cachexia, are also in favor of such an idea (10).It has also been demonstrated that indomethacin is almostregularly a potent inhibitor of experimental tumor growth (11),although it is also an inducer of increased susceptibility toinfections in non-tumor-bearing animals and patients probablyby attenuating the immune system (12). Indomethacin alsoimproves appetite, a mechanism that may in theory be secondary to attenuation of the cytokine cascade and decreased inflammation (13). The above-mentioned but contradictory impressions make it difficult to sort out what are primary or secondaryeffects of cytokines either on the host or on the tumor. Theoverall effect of the immune system on tumor growth and cancercachexia remains unknown, although a formidable amount ofinformation exists.

The aim of the present study was to evaluate the net effectsof IL-1, IL-2, TNF, indomethacin, and cyclosporin A on tumorgrowth, survival, and cachexia (body composition) in a well-defined and -characterized tumor model. Such results mayhopefully improve our understanding of what are primary or

3The abbreviations used are: IL-1, interleukin 1; IL-2, interleukin 2; 11-6,interleukin 6: lili 1... human interleukin 1»;hIL-2, human interleukin 2; IL-1/3,interleukin 1/j; TNF, tumor necrosis factor; hTNF-o, human tumor necrosisfactor «;PGE2, prostaglandin E2; LPS, lipopolysaccharide; ANOVA, analysis ofvariance.

880

on July 20, 2021. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF INDOMETHACIN, CYTOKINES, AND CYCLOSPORIN ON TUMORS

secondary events behind tumor progression and death in experimental cancer cachexia.

MATERIALS AND METHODSAnimals. Weight-stable female C57BL/6J mice (ALAB, Stockholm,

Sweden) were used for all the studies. The animals were housed in atemperature-controlled room with a 12-h light/dark schedule. Micewere placed in groups of 5 and housed in plastic cages which wereprovided with screen floors to allow collection and quantitation ofspilled food. The animals were allowed an initial adaptation period tobe in stable body weight and have normal food intake before experiments started. They received a known amount of food and waterbetween 4 and 6 p.m. daily. The following day, the quantity of foodconsumed and spilled was measured. Water intake was measured onthe same occasion. Body weights were recorded daily. The food suppliedwas ordinary rodent chow (ALAB, Stockholm, Sweden) and drinkingwater.

Under pentobarbital anesthesia (60 mg/kg of body weight ¡.p.),themice were implanted s.c. with 3 to 5 mm3 of a transplantable methyl-cholanthrene-induced sarcoma (MCG 101) bilaterally on the dorsalregion. This tumor has been continuously transplanted in vivo for morethan 15 yr in our laboratory. It has a reproducible growth pattern, leadsto 100% tumor take, and does not give rise to visible métastaseswithinthe time period the tumor kills its host. Although this tumor wasoriginally induced chemically as a sarcoma, recent histological evaluation has revealed that it has few, if any, characteristics of a sarcoma. Itshould, therefore, rather be classified as a low or undifferentiatedrapidly growing solid tumor with a predominantly monoclonal appearance determined by flow cytometry. Animals have generally been studied 10 to 11 days following implantation when the tumor comprises 10to 15% of body weight. In this study the major part of the investigationswas survival experiments with spontaneous death in cancer cachexia.Previous studies from this laboratory have shown that untreated tumor-bearing mice uniformly die 13 to 17 days following tumor implantation,depending on the amount of tumor that is implanted and with somevariation in survival among various tumor batches. Therefore, survivalexperiments were always done on a strictly comparative basis withstudy and control tumor-bearing mice. Anorexia, host wasting, andtumor growth are progressive (14). The tumor growth rate correspondsto a doubling time around 57 h (15). Non-tumor-bearing animals were

sham implanted.Treatment Regimens. All animals were randomly assigned to different

treatment groups and received the following per day: recombinant hlL-la at 20.5 ng/g of animal weight (8.6 x IO6 units/mg of proteinprovided by P. Lomedico, Hoffman-La Roche, Inc., Nutley, NJ) or IL-\ßat 20.5 ng/g (16.7 x IO6 units/mg of protein, provided by C.Dinarello, Tufts University Medical School); hTNF-a, 22.5 ng/g (3.1x IO7units/mg of protein provided by Genentech, Inc., San Francisco,CA); hIL-2, 20.5 ng/g (3.2 x 10' units/mg of protein, provided byBiogen, Inc. (Boston, MA); combination of IL-1 and TNF, all together21 ng/g; cyclosporin A 60 Mg/g (Sandimmun, 50 mg/ml; Sandoz);indomethacin (Confortid, Dumex), 1 ng/g. As control, 0.9% NaClsolution was used. All of the recombinant products contained less thanlOpgofendotoxin per ¡igof protein. All the water-soluble preparationswere administered in 0.9% sodium chloride solution. Cyclosporin Awas dissolved in ethanol and ricine oil as used for i.v. infusion tohumans.

The amount injected each time was 0.2 ml, and the cytokines (IL-1,IL-2, TNF) were administered as twice daily i.p. injections, whereascyclosporin A and indomethacin were given once daily as s.c. injections(0.2 ml). As indicated separately, indomethacin was also given as p.o.treatment supplied in the drinking water in the same amount as above(1 Mg/g). The appropriate dilution in drinking water was calculated,based on the previous day's water consumption. The treatments started

with the first dose given on Day 1 after tumor implantation. Theamounts given of IL-1 and TNF were chosen based on our previousexperiments on the metabolic effects of "physiological" doses of recom

binant IL-1 and TNF on healthy non-tumor-bearing mice (10, 16). Thedose of cyclosporin A was chosen to correspond to the amount used inour transplantation practice, a dose which is in the range of parenteral

doses previously used in experiments on rats and mice (20 to 80 Mg/g).Cyclosporin A at 20 Mg/g given daily as s.c. injections has clearlydemonstrated immunosuppressive effects (17,18), which was confirmedin our laboratory by cyclosporin to mice infected by sublethal doses ofbacteria (Listeria monocytogenes). The dose of indomethacin to betested was tolerable. Initial experiments were conducted where the doseof indomethacin (3 Mg/g) was found to be toxic when given on acontinuous basis to both tumor-bearing and non-tumor-bearing controls, and the animals uniformally died on Day 7. A nontoxic tolerabledose was titrated out on reduced dosage of indomethacin (1.5 to 0.37Mg/g), and 1 Mg/g was eventually found to be nontoxic. In additionalexperiments, indomethacin treatment was instituted at different timeintervals when the tumor just became visible (Day 5), when it was wellestablished (Day 7), and when cachexia was obvious (Day 9) as compared with the continuous treatment from Day 1. The effect of indomethacin (10 Mgto 500 Mg/ml) on tumor cell proliferation was investigated in experiments on tumor cells cultured at standard conditions inMcCoy's Medium 5A (Flow Laboratories) in the presence of 10% fetal

calf serum during logarithmic growth. In some experiments, the tumor"immunogenicity" was tested essentially as described elsewhere (19).

The tumor-bearing mice were inoculated by our standard technique.Control mice were sham inoculated. The growing tumors were thenextirpated when weighing around 0.5 g. Five days later varying amountsof tumor cells (10* to IO8) were inoculated s.c. The subsequent tumor

take and tumor growth rate were monitored in order to confirm whetheror not previous "tumor-vaccination" had caused any resistance to tumor

growth.Daily recordings of food and water intake as well as of body weight

were performed. Physical activity, condition of the fur, and other signsof the general well-being of the animals were registered. Net tumorgrowth was determined according to our previous experience by excisingthe entire tumor and determining wet and dry tumor weight (20). Thetime course of tumor growth was measured by sacrificing various tumor-bearing mice on Days 7, 9, and 11 for determination of tumor dry'weight. When tumor-bearing mice died spontaneously, the tumors weredissected free, and the wet weight of carcass and tumor were recorded.Carcass and tumor were dried to constant weight in an oven at 80°C,

and dry weight was recorded. Carcass composition was determinedgravimetrically as previously described (14, 20). In some experimentsadherent peritoneal macrophages were harvested from tumor-bearingand non-tumor-bearing mice as described elsewhere (2). The cells wereincubated in RPMI 1640 medium (Flow Laboratories, Sollentuna,Sweden) with 10% fetal calf serum. The PGE2 synthesis was measuredin the basal state and after in vitro stimulation of macrophages by LPS(Staph. enter.; 10 Mg/ml). PGE2 was determined byaradioimmunoassaykit from New England Nuclear Laboratories (Germany). This studyprotocol was approved by the ethical committee for animal experiments,Medical Faculty, University of Gothenburg, Sweden.

Statistics. Results are presented as the mean ±SE. One-factorANOVA was used when more than two animal groups were compared.Longitudinal measurements as food and water intake and body weightalterations were tested by one-factor ANOVA for repeated measures.Survival was tested by a log-rank test (life table analysis). Two-tailedtests or 95% confidence intervals were used. A P < 0.05 was consideredstatistically significant (21).

RESULTS

In the first experiment tumor-bearing mice had a meansurvival time of 17.2 ±1 days following tumor implantation(Table 1). Injections of IL-1, TNF, or cyclosporin A to additional tumor-bearing mice in this experiment did not changesurvival time, although tumor dry weight, carcass dry weight,and body fat were significantly decreased in animals receivingeither IL-1 or the combination of IL-1 and TNF. CyclosporinA or IL-2 did not alter survival, tumor growth, or body composition in tumor-bearing animals. Results for IL-2 are notgiven, since IL-2 in the dose used here did not influence any ofthe variables measured in either tumor-bearing or control mice.

881

on July 20, 2021. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF INDOMETHACIN. CYTOKINES, AND CYCLOSPORIN ON TUMORS

Table 1 Body composition and survival time in tumor-bearing animals given daily injections with IL-1, TNF, and cyclosporin A in doses given in"Materials and Methods, "

Tumor-bearing controlTumor-bearing + 1L-ITumor-bearing + TNFTumor-bearing + TNF + IL-1Tumor-bearing -f cyclosporin ANon-tumor-bearing controlInitial

bodywt.(g)22.0

±0.3°

20.9 ±0.320.9 ±0.321.0 ±0.222.0 ±0.420.8 ±0.3Final

bodywt.(g)24.2:

18.326.418.625.2

22.0t

1.30.3*

0.70.3*

1.10.4Carcass

drywt.(g)4.72

±0.104.29 ±0.07*

4.77 ±0.094.18 ±0.04*

4.91 ±0.148.35 ±0.69Tumor

drywt.(g)1.62

±0.230.58 ±0.04*

1.43 ±0.040.69 ±0.04*

1.50 ±0.17Body

fat(g)0.86

±0.130.67 ±0.05e'1'

1.08 ±0.100.74 ±0.05C'</

1.01 ±0.113.35 ±0.17Survival

(days)17.2

±1.018.0 ±0.318.6 ±1.017.6 ±0.619.5 ±1.2

ANOVA NS' NS°Mean ±SE (10 animals/group).* P < 0.05 versus tumor-bearing control, tumor-bearing TNF, and tumor-bearing cyclosporin A.' P < 0.05 versus tumor-bearing cyclosporin A.d P < 0.05 versus tumor-bearing TNF.' NS, not significant.

Tumor-bearing mice receiving IL-1 and the combination of IL-1 and TNF had a slower body weight gain (carcass + tumor)compared with TNF-, cyclosporin A- and saline-injected tumor-bearing animals (Fig. 1). This was due to a significantly lowerfood intake in IL-1-treated animals, while the food intake inTNF-treated tumor-bearing mice did not differ from that ofsaline-injected tumor-bearing animals. The water intake wasseveralfold higher in IL-1-treated tumor-bearing animals compared with tumor-bearing mice exposed to TNF or cyclosporinA or the combination of TNF and IL-1 (Fig. 2).

Provision of indomethacin either in the drinking water or asdaily s.c. injections to tumor-bearing mice prolonged survivalfrom 14.7 ±0.7 to 22.0 ±1.0 (Fig. 3; Table 2). Tumor-bearingmice treated with indomethacin from Day 1 died with a morepreserved body composition (carcass dry weight) despite numerically larger tumors. In these animals, it was lean body mass(carcass fat-free dry weight) that was clearly improved by indomethacin (P < 0.0004), but the difference in body fat did notreach the level of statistical significance. The ratio betweentumor dry weight and carcass dry weight was significantlyhigher in indomethacin-treated mice compared with sham-treated controls (0.34 ±0.02 versus 0.17 ±0.02, P < 0.0001).This means that such animals had larger tumors in relation totheir degree of undernutrition compared with untreated tumor-bearing mice, although the tumor grew at slower rates inindomethacin-treated mice. Tumor dry weight on Days 7, 9,and II was in untreated mice 0.12 ±0.01, 0.23 ±0.02, and0.53 ±0.04 g, and in indomethacin-treated mice, 0.08 ±0.01,0.17 ±0.02, and 0.36 ±0.03 g, respectively (P < 0.05).

26-

,I »;

18 -

16•4-2 16180 2 4 6 8101214

Days after tumor implantationFig. 1. Animal weight (including tumor) in tumor-bearing animals treated with

daily i.p. injections of cytokines (IL-1, TNF). cyclosporin A (CyA), and shaminjections (NaCI) as described in "Materials and Methods." The injections and

the tumor implantation started on Day zero. Fifteen to 20 animals were in eachgroup. IL-1-and IL-1 + TNF-treated animals had significantly lower body weight(P < 0.0001) compared with the other groups tested by one-factor ANOVA forrepeated measures.

0.4-

•JS2 0,3 -

Ifw"0,1-

IL-1

TNF

TNFtlL-1

CyA

IMI

•5-3-1 1 3 5 7 9 1113151719

Days

Fig. 2. Daily water intake in tumor-bearing animals treated with cytokines (/£-/, 7W), cyclosporin A (CyA), and sham injections (A/aC7) following tumorimplantation, which was done on Day zero; 15 to 20 animals were in each group.The water intake in IL-1-treated animals was statistically significantly highercompared with the other groups (P < 0.001) tested by one-factor ANOVA forrepeated measures, bw, body weight.

100

Days after tumor implantationFig. 3. Survival of tumor-bearing animals (T.B.) treated with either indometh

acin (Indo) in the drinking water or as daily s.c. injections as compared withsaline injections following tumor implantation as described in "Materials andMethods." The difference in survival was statistically significant (P < 0.01) astested by a log-rank test.

However, indomethacin at concentrations of 10 to 500 Me/mldid not reduce tumor cell proliferation in vitro in the presenceof 10% fetal calf serum. Indomethacin-treated mice also had asignificantly larger food intake compared with untreated tumor-bearing mice (Fig. 4). Most interestingly, indomethacin couldalso prolong survival in tumor-bearing animals to the sameextent irrespective of whether treatment started on Day 1, 5, 7,or 9 following tumor implantation (Fig. 5).

Indomethacin did not decrease the hepatic acute-phase re-

882

on July 20, 2021. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF 1NDOMETHACIN, CYTOKINES, AND CYCLOSPORIN ON TUMORS

Table 2 Survival lime and body composition at death in tumor-bearing animals treated daily with indomethacin in the drinking water as described in"Materials and Methods. "

Tumor bearing + indomethacinTumor-bearing controlInitial

bodywt.(g)20.5

±0.2°

19.9 ±0.2Final

bodywt.(g)25.6

±0.918.4 ±0.7Carcass

drywt.(g)4.48

+ 0.133.90 ±0.05Carcass

fat-freedry wt.(g)3.76

±0.063.27 ±0.12Body

fat(g)0.84

±0.080.68 ±0.04Tumor

drywt.(g)1.50

±0.111.06 ±0.40Survival

(days)22.0

±1.014.7 ±0.7

Student's t test NS P< 0.001 P < 0.005 P < 0.0004 NS NS />< 0.0001

°Mean ±SE (15 animals/group).* NS, not significant.

16

14-

12-

10-

2-

XTB Indo

TB control

•101 23456789101112

days after tumor implantation.

Fig. 4. Time course of food intake in tumor-bearing mice (TB) treated withindomethacin (indo) in the drinking water compared with untreated tumor-bearing control animals as described in "Materials and Methods." The differencein food intake was tested by one-factor ANOVA for repeated measures. Points,mean; bars, SE; *,P< 0.05; 20 animals in each group. There was 100% survivalin both groups within the experimental period (1 to 11 days).

TB control

TBDay9

10 15 20

Survival time in days

2 s

Fig. 5. Survival time in tumor-bearing mice treated with indomethacin by dailys.c. injections as described in "Materials and Methods." The treatment was

instituted on different days (Days 1 to 9) following implantation of the tumorcompared with sham-injected tumor-bearing controls (TB control); *,P< 0.0001versus TB controls. The tumor grew exponentially in TB control animals with anet doubling time around 57 h. Columns, mean; linn, SE.

sponse measured as the increase in liver RNA content frommice with progressive tumors (11.2 ±0.4 versus 9.5 ±0.4 mg/g in liver tissue from non-tumor-bearing controls, P< 0.01).

Peritoneal macrophages from tumor-bearing mice synthesized spontaneously PGE2 levels corresponding to 1100 ±150pg/106 cells determined 10 to 11 days following tumor implantation, while macrophages from freely fed non-tumor-bearingmice synthesized 3700 ±922 pg/106 cells. These lower levelsin tumor-bearing mice were not due to an impaired cellularprostaglandin synthesis, since these cells could be stimulatedby LPS (10 Mg/ml) in vitro to produce steady-state levels ofPGE2 corresponding to 39,500 ±4208 pg compared with12,400 ±4330 pg/106 cells (P < 0.01) in non-tumor-bearingcontrols. Previous "vaccination" of tumor-bearing mice against

the tumor caused a significantly lower tumor take and a reducedtumor-growth rate.

DISCUSSION

The present study confirms previous reports demonstratingthat prostaglandin synthesis inhibitors, particularly indomethacin, inhibit tumor growth (22-25). The main effect of indomethacin is probably to block the synthesis of PGE2, which isa cellular effector to many cytokines in the process of immuneregulation and inflammation. It can, however, not be excludedthat other nonclassical prostaglandin effect(s) are involved,since the effect is not clearly related to the prostaglandinsynthesis in the tumors (26-28). It is likely that the inhibitionwas confined to host cells, since indomethacin caused no growthinhibition of cultured tumor cells. New information is also thatinitiation of the indomethacin treatment, late in the diseasecourse, caused the same inhibitory effects as did immediate orearly treatment. This is a very unusual effect, considering various therapy modalities in experimental tumor models.

Indomethacin also protected the host from deterioration inbody composition, particularly lean body mass, despite the factthat the tumor could grow 40% larger when indomethacin-treated tumor-bearing animals died spontaneously with a 50%prolonged survival. This is a remarkable finding, since untreatedanimals with solid tumors usually die with similar relationsbetween tumor burden and carcass composition. Present resultssuggest that a "metabolic barrier" between the host and the

tumor was enhanced or preserved by indomethacin, so that thetumor could grow to large volumes but without depletion of themetabolic reserves (fat, proteins). The fact that the animals diebefore depletion of metabolic stores demonstrates that something other than metabolic reserves limits survival in indometh-acin-treated tumor-bearing mice. In untreated tumor-bearingmice we have found that spontaneous death occurs due tohypoglycemia when the total glucose consumption exceedsglucose production (intake + gluconeogenesis) (29).

Appetite was improved in the indomethacin-treated animals.This is compatible with the suggestion that cytokines (IL-1, IL-6, and TNF) induce anorexia in tumor-bearing and infectedanimals (30-32). However, our recent experiments on tumor-bearing rats have shown that indomethacin had no effect onanorexia unless tumor growth itself was inhibited (13). Therefore, it is possible that the improved appetite by indomethacinwas in part secondary to tumor growth inhibition. Our previousstudies have also demonstrated that indomethacin does notdecrease the increased level of circulating IL-6 in our tumor-bearing mice (33); in fact it rather seems to increase plasma IL-6. Therefore, it is unlikely that a decrease in IL-6 productionshould be the major explanation for improved appetite intumor-bearing animals treated with indomethacin. In addition,it has been suggested that some brain cytokines may not besensitive to classic prostaglandin synthesis inhibitors (34).

The inhibitory effect of indomethacin in a variety of tumormodels has in general been assumed to be mediated by immuneeffects (35, 36). At a superficial look, this may be a reasonableconclusion. However, tumor inhibition by indomethacin has

883

on July 20, 2021. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF INDOMETHACIN, CYTOKINES. AND CYCLOSPORIN ON TUMORS

also been described to occur to the same extent in both naturalkiller cell- and T-cell-deficient mice, suggesting that classicalimmune effects are not a prerequisite (37). To elucidate theimportance of net immune effects for the tumor growth in ourmodel, we treated tumor-bearing animals with cyclosporin A ina dose which is immune suppressive (17, 18). Studies withrodent tumors in syngeneic mice have shown that cyclosporinA has no antitumor effect in itself, but it prevents immunolog-ical regression of transplanted human tumors in allogeneic mice(38). We have also confirmed in unpublished experiments thatthe immune system in our mice is severely depressed by cyclosporin A at a dose of 30 ^g/day, although cyclosporin A at 60jig/day had no effect on either tumor growth or body composition. This finding is particularly pertinent, since our tumorbelongs to a group classified as highly immunogenic tumors(38), which was confirmed in the present study by "tumorvaccination" experiments. The biological meaning and the im-

munological significance of the traditional classification of immunogenic and nonimmunogenic tumors remain an unexplained phenomenon, although it is related to some chemicalstructures in the tumor (39).

Our results agree with previous observations demonstratingthat tumor growth is not influenced by cyclosporin A, but theappearance of secondary tumors may be in some models (40,41), although not consistently so (42). In our opinion, theimportance of both known and unknown virus contaminationsof experimental tumors for described immune responses in vivoand in vitro has not been accounted for in the interpretation ofexperimental results. By means of electron microscopy, unclassified virus particles have been observed within tumor cells inbiopsies from our tumor-bearing mice.4 Whether this may have

anything to do with the observed effects in the present study isdifficult to evaluate, but if so, indomethacin should ratheraggravate tumor progression, if expression of virus antigens oninfected tumor cells were of major significance for immunecontrol of tumor growth in our model. It has recently beenreported that both indomethacin and neutralizing antibodies toTNF-«made sublethal infections lethal in C57BL/6 mice (12,43, 44). These kinds of experiments represent a simple andpowerful demonstration of the role that the immune systemplays in the control of infections. If significant, it should be aseasy to demonstrate the significance of immune control oftumor progression as for bacterial infection and viremia. Ourarguments about the weak or nonexisting immune control inour model are further strengthened if experiments with potentimmunosuppressive steroids are considered, since the immune-depressant effects of glucocorticoids are both well understoodand undisputable (45, 46). In fact, glucocorticosteroids have asimilar tumor-growth inhibition as does indomethacin4 in our

model as reported for some other experimental tumors (47). Inaddition, measurements of prostaglandin production in peritoneal macrophages from our tumor-bearing animals did notconfirm elevated basal synthesis of PGE2, which may be expected in an immune-compromized tumor-bearing host withdecreased IL-1 production in macrophages (2). This negativefinding was not due to a cellular impairment, since LPS stimulation in vitro increased prostaglandin synthesis in tumor-hostmacrophages more than in control macrophages. This resultalso argues against that indomethacin exerted its effects viaimmune mechanisms (47) as upregulation of cytokines andcytokine receptors in immune cells (2, 48). Therefore, we propose that growth of our tumor is not under immune control,

' Unpublished results.

although belonging to a group of highly "immunogenic tumors." Classic immune effects toward some experimental tu

mors, both in vivo and in vitro, may theoretically be explainedby the presence of virus/virus antigens in laboratory-propagatedtumor cells. Expression of virus antigen seems to be a potentmechanism to extinguish tumor cells in vivo (49).

In the context of immune modulation, it is also important toconsider the effects of IL-1, TNF, and IL-2 administration atreasonable physiological concentrations. IL-2, the most well-recognized T-cell activator, had no effect on either tumorgrowth, survival, hepatic acute phase response, or body composition when given in equimolar concentrations to IL-1 andTNF. However, IL-1 but not TNF had a potent inhibitory effecton tumor growth. This effect was not at all translated intoprolonged survival, which it should if this effect was primarilyon tumor growth. IL-1-treated, tumor-bearing animals had aworse nutritional state including the lowest reserves of body fatand lean body mass compared with any other group of tumor-bearing mice. Thus, IL-1 supplementation to tumor-bearinganimals potentiated the tumor-induced anorexia leading to aslower growth of the tumor-bearing animal (carcass + tumor),an effect which was not seen with TNF alone. Interestingly, IL-1 had also a polydipsic and polyuric effect on tumor-bearinganimals. It is our opinion that the tumor growth inhibition byIL-1 was to potentiate the anorexia, which immediately leadsto decreased tumor growth. Starvation, leading to substratedeficiency, is one of the most potent regulators of experimentaltumor growth.5 This effect is in part mediated by alterations in

the energy charge of tumor cells leading to a prolonged cellcycle transit time.6-7 This effect is mediated by or associatedwith the carbohydrate intake.5 Thus, effects of cytokines on

tumor growth must always take appetite into account, sincetumor-inhibitory effects may be secondary to a decreased foodintake and tumor substrate supply. This has not been done inprevious reports on positive effects by cytokines on tumorgrowth in vivo. Most likely, the seemingly positive effect of IL-1 on tumor growth was primarily due to increase anorexia andperhaps whole-body energy expenditure, leading to a substrate-induced tumor-growth inhibition. The effect of indomethacinon tumor growth was not mediated by this mechanism, sincefood intake was increased in these experiments.

In conclusion, this study demonstrates that indomethacininduces or promotes tumor growth inhibition, which was lessdependent on the stage of the tumor disease than described forany other monodrug treatment that we know of. The mechanism is unknown, but may be related to neoangiogenesis andtumor growth factor(s) regulation within or close to the tumor(50). The improved appetite and nutritional state in tumor-bearing animals on indomethacin are probably secondary totumor inhibition, although some primary effects on food intakecan not be excluded. We have recently reported that both IL-1and TNF may represent growth factors in our tumor model (5).Systemic administration of recombinant cytokines, particularlyat high doses, may therefore simply result in both positive andnegative secondary effects on host metabolism, which some-

5T. Westin, S. Edström,and K. G. Lundholm, Ornithine decarboxylase activity

in tumor tissue in response to refeeding and its dependancy on diet components,submitted for publication.

' T. Westin, J-P. Idstrom, B. Soussi, S. Edström, E. Lydén,B. Gustavsson,and K. Lundholm. Energy state and nutrition-induced tumor growth retardationin experimental cancer evaluated by 31P-NMR spectroscopy, submitted forpublication.

'T. Westin, B. Gustavsson, S. Edström, K. Hellander, L. Reinholdtsen, L.Tibell, and K. Lundholm. Tumor cytokinetic effects of acute starvation versuspolyamine depletion in tumor-bearing mice, submitted for publication.

884

on July 20, 2021. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF INDOMETHACIN, CYTOKINES, AND CYCLOSPORIN ON TUMORS

times influence tumor growth. Nonimmune mechanisms arecertainly of significance (51).

REFERENCES

1. Durant, J. R., Immunotherapy of cancer. N. Engl. J. Med., 316: 939-941,1987.

2. Moldawer, L. L., Lonnroth, C., Mizel, S., and Lundholm. K. Down-regulation of ¡nterleukin1 production by macrophages of sarcoma-bearing mice. J.Immunol., 138: 4270-4274, 1987.

3. Hakim, A. A. Peripheral blood lymphocytes from patients with cancer lackinterleukin-2 receptors. Cancer (Phila.), 61: 689-701. 1988.

4. Sherry, B., Gelin, J., Fong, Y., Maraño,M., Wei, H., Cerami, A., Lowry, S.F., Lundholm, K. G., and Moldawer, L. L. Anticachectin-tumor necrosisfactor-alpha antibodies attenuate development of cachexia in tumor models.FASEB (Fed. Am. Soc. Exp. Biol.) J., 3: 1956-1962, 1989.

5. Gelin, J., Moldawer, L. L., Lonnroth, C., Sherry, B., Chizzonite, R., andLundholm, K. The role of endogenous TNF-a and interleukin-1 for experimental tumor growth and the development of cancer cachexia. Cancer Res.,SO:415-421, 1990.

6. Mclntosh, J. K., Mulé,J. J., Krosnick, J. A., and Rosenberg, S. A. Combination cytokine immunotherapy with tumor necrosis factor alpha, interleukin-2 and alpha-interferon and its synergistic antitumor effects in mice.Cancer Res., 49: 1408-1414, 1989.

7. Moldawer, L. L., Georgieff, M., and Lundholm, K. Interleukin 1, tumournecrosis factor-alpha (cachectin), and the pathogenesis of cancer cachexia.Clin. Physiol., 7: 263-274, 1987.

8. Oliff, A. The role of tumor necrosis factor (cachectin) in cachexia. Cell, 54:141-142, 1988.

9. Oliff, A., Defeo-Jones, D., Boyer, M., Martinez, D., Kiefer, D., Vuocolo, G.,Wolfe, A., and Socher, S. H. Tumors secreting human TNF-cachectin inducecachexia in mice. Cell, 50: 555-563, 1987.

10. Moldawer, L. L., Andersson, C., Gelin, J., and Lundholm, K. G. Regulationof food intake and hepatic protein synthesis by recombinant-derived cyto-kines. Am. J. Physiol., 254: 450-456, 1988.

11. Lynch, N. R., Castes, M., Astoin, M., and Salomon, J. C. Mechanism ofinhibition of tumour growth by aspirin and indomethacin. Br. J. Cancer, 38:503-512, 1978.

12. Tripp, C. S., Needleman, P., and Unanue, E. R. Indomethacin in viroincreases the sensitivity to Listeria infection in mice. A possible role formacrophage thromboxane A2 synthesis. J. Clin. Invest., 79: 399-403, 1987.

13. Sandström, R., Gelin, J., and Lundholm, K. The effect of indomethacin onfood and water intake, motor activity, and survival in tumor-bearing rats.Eur. J. Cancer, 26: 811-814, 1990.

14. Lundholm, K., Edström, S., Karlberg, I., Ekman, L., and Scherstén,T.Relationship of food intake, body composition, and tumor growth to hostmetabolism in non-growing mice with sarcoma. Cancer Res., 40: 2516-2522,1980.

15. Karlberg, I., Edström, S., Ekman, L., Johansson, S., Scherstén,T., andLundholm, K. The metabolic host-reaction in response to the proliferationof non-malignant cells versus that of malignant cells in vivo. Cancer Res., 41:4154-4161,1981.

16. Moldawer, L. L., Svaninger, G., Gelin, J., and Lundholm, K. Interleukin-1(alpha or beta) and tumor necrosis factor alpha do not regulate proteinbalance in skeletal muscle. Am. J. Physiol., 253: C766-C773, 1987.

17. Wassef, R., Cohen, Z., and Langer, B. Pharmacokinetic profiles of cyclo-sporine in rats. Influence of route of administration and dosage. Transplantation (Baltimore), 40: 489-493, 1985.

18. Fukuzawa, M., Sharrow, S. O., and Shearer, G. M. Effect of cyclosporin Aon T cell immunity. II. Defective thymic education of CD4 T helper cellfunction in cyclosporin A-treated mice. Eur. J. Immunol., 19: 1147-1152,1989.

19. Suurkula, M., and Boeryd, B. Tumor metastasis in mice with reduced immunereactivity. II. Studies with highly antigenic MCA-induced sarcoma in thy-mectomized and/or sub-lethally irradiated C57BL/6J mice. Int. J. Cancer,14: 633-641, 1974.

20. Edén,E., Lindmark, L., Karlberg, I., and Lundholm, K. Role of whole bodylipids and nitrogen as limiting factors for survival in tumor-bearing micewith anorexia and cachexia. Cancer Res., 43: 3707-3711, 1983.

21. Armitage, P., and Berry, G. Statistical Methods in Medical Research, Ed. 2.Oxford, United Kingdom: Blackwell Scientific Publications, 1987.

22. Furnia. Y., Hunter, N., Barkley, T., Jr., Hall, E.. and Milas. L. Increase inradioresponse of murine tumors by treatment with indomethacin. CancerRes., 48: 3008-3013, 1988.

23. Panje, W. R. Regression of head and neck carcinoma with a prostaglandin-synthesis inhibitor. Arch. Otolaryngol., 107: 658-663, 1981.

24. Kim, B., and Warnaka, P. Indomethacin-enhanced immunotherapy of pulmonary métastasesusing IL-2 and IFN-a. Surgery (St. Louis), 106: 248-256, 1989.

25. Rao, A. R.. and Hussain, S. P. Modulation of methylcholanthrene-inducedcarcinogenesis in the uterine cervix of mouse by indomethacin. Cancer Lett.,«.•15-19,1988.

26. Furuta, Y., Hall, E. R., Sanduja, S., Barkely, T., Jr., and Milas, L. Prosta-glandin production by murine tumors as a predictor for therapeutic responseto indomethacin. Cancer Res., 48:3002-3007, 1988.

27. Rodan, S. B., Wesolowski, G., and Rodan. G. A. Clonal differences inprostaglandin synthesis among osteosarcoma cell lines. J. Bone Mineral Res.,7:213-220. 1986.

28. Fulton. A. M., Hanchin, C. M., and Butler, W. B. Heterogeneity of prostaglandin E2 binding in murine mammary tumor cells differing in metastaticpotential. Invasion Metastasis, 8: 17-30, 1988.

29. Svaninger, G., Gelin, J., and Lundholm. K. G. The cause of death innonmetastasizing sarcoma-bearing mice. A study with relevance for tumortreatment experiments in mice. Eur. J. Cancer Clin. Oncol., 25: 1295-1302,1989.

30. Stovroff, M. C., Fraker, L. D., Swedenborg, J. A., and Norton, J. A.Cachectin tumor necrosis factor: a possible mediator of cancer anorexia inthe rat. Cancer Res., 48:4567-4572, 1988.

31. McCartney, D. O., Kluger, M. J., and Vander, A. J. Suppression of foodintake during infection: is interleukin 1 involved? Am. J. Clin. Nutr., 42:1179-1182.1985.

32. Plata-Salamán, C. R., Oomura, Y., and Kai, Y. Tumor necrosis factor andinterleukin-IB: suppression of food intake by direct action in the centralnervous system. Brain Res., 448: 106-114, 1988.

33. Gelin, J., Moldawer, L. L., Lonnroth, C., deMan, P., Svanborg-Edén,C.,Lowry. S.. and Lundholm, K. G. Appearance of hybridoma growth factor-interleukin-6 in the serum of mice bearing a methylcholanthrene-inducedsarcoma. Biochem. Biophys. Res. Commun., 157: 575-579, 1988.

34. Hellerstein, M. K.. Meydani, S. N., Meydani, J., Wu, K., and Dinarello, C.A. Interleukin-1 induced anorexia in the rat. Influence of prostaglandins. J.Clin. Invest., 84: 228-235, 1989.

35. Fulton, A. M. Inhibition of experimental metastasis with indomethacin: roleof macrophages and natural killer cells. Prostaglandins, 35: 413-425, 1988.

36. Fulton, A. M. The role of eicosanoids in tumor metastasis. ProstaglandinsLeukotrienes and Essential Fatty Acids, 34: 229-237, 1988.

37. Maca, R. D. Inhibition of the growth of Lewis lung carcinoma by indomethacin in conventional, nude, and beige mice. J. Biol. Response Modifiers, 7:568-580, 1988.

38. Fingen, H. J., Treiman, A., and Pardee, A. B. Transplantation of human orrodent tumors into cyclosporine-treated mice: a feasible model for studies oftumor biology and chemotherapy. Proc. Nati. Acad. Sci. USA, 81: 7927-7931, 1984.

39. Srivastava, P. K., DeLeo, A. B., and Old, L. J. Tumor rejection antigens ofchemically induced sarcomas of inbred mice. Proc. Nati. Acad. Sci. USA,«5:3407-3411. 1986.

40. Eccles, S. A., Heckford, S. E., and Alexander. P. Effect of cyclosporin A onthe growth and spontaneous metastasis of syngeneic animal tumours. Br. J.Cancer, 42: 252-259, 1980.

41. Shimizu. T., Martin, M. S., Pelletier, H., Lagadec, P.. and Martin, F. Effectsof cyclosporin A on progressive and regressive tumors induced by two cancerlines derived from a single colon carcinoma chemically induced in the rat.Immunobiology, 178:401-415, 1989.

42. Zylstra, S., Saji, S.. and Takita, H. Effects of cyclosporin A on mice tumorsand the host. J. Exp. Clin. Cancer Res., 4:47-53, 1985.

43. Havel!, E. A. Production of tumor necrosis factor during murine listeriosis.J. Immunol., 139:4225-4231, 1987.

44. Nakane, A., Minagawa, T., and Kato, K. Endogenous tumor necrosis factor(cachectin) is essential to host resistance against Listeria monocytogenesinfection. Infect. Immun., 56: 2563-2569, 1988.

45. Guyre, P. M., Girard, M. T., Morganelli, P. M., and Manganiello, P. D.Glucocorticoid effects on the production and actions of immune cytokines.J. Steroid Biochem., 30: 89-93, 1988.

46. Tsujimoto, M., and Adachi, H. Effects of steroid hormones on the cytotoxicactivity of tumor necrosis factor. J. Biochem., 103: 393-395, 1988.

47. Crowley, P., Lai, N. Y., De Young. N., Pearce, P., Funder, J. W., and Gill,P. G. Inhibition of growth of B16 melanoma by glucocorticoids does notresult directly from receptor-mediated inhibition of tumour cells. Oncology(Basel), 45:331-335, 1988.

48. Rincón, M., Tugores, A., Lópes-Rivas, A., Silva, A., Alonso, M., De-Landázuri, M., and Lópes-Botet, M. Prostaglandin E2 and the increase ofcellular cAMP inhibit the expression of interleukin 2 receptors in human Tcells. Eur. J. Immunol., 18:1791-1796, 1988.

49. Plata. F., Langlade-Demoyen. P., Abastado, J. P.. Berbar. T., and Kourilsky,P. Retrovirus antigens recognized by cytolytic T lymphocytes activate tumorrejection in vivo. Cell, 48: 231-240, 1987.

50. Peterson, H-I. Effects of prostaglandin synthesis inhibitors on tumor growthand vascularization. Invasion Metastasis, 3: 151-159, 1983.

51. Manda, T., Nishigaki, F., Mori, J., and Shimomura, K. Important role ofserotonin in the antitumor effects of recombinant human tumor necrosisfactor «in mice. Cancer Res., 48: 4250-4255, 1988.

885

on July 20, 2021. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1991;51:880-885. Cancer Res   Johan Gelin, Christer Andersson and Kent Lundholm  CachexiaTumor Growth and the Subsequent Development of Cancer Effects of Indomethacin, Cytokines, and Cyclosporin A on

  Updated version

  http://cancerres.aacrjournals.org/content/51/3/880

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/51/3/880To request permission to re-use all or part of this article, use this link

on July 20, 2021. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from