Metal Carcinogenesis

II. A Study on the Carcinogenic Activity of Cobalt,

Copper, Iron, and Nickel Compounds*

J. P. W. GILMAN

(Division of Histo1o@y,Oniarlo Veterinary College,Gustph, Ontario, Canada)

SUMMARY

Nickel sulfide, nickel oxide, cobalt sulfide, and cobalt oxide were shown to be carcinogenic on single intramuscular injection in rats. The sulfide of each metal induceda significantly higher tumor incidence than did its oxide. Comparisons between averagelatent periods and between measures of tumor progression also indicated that thepresence of the sulfide enhances the carcinogenic activity of these compounds. Nickelsulfate, iron and copper sulfides, and oxides failed to induce tumors in rats under theseconditions.

Both CSH and Swiss mice developed tumors from single intramuscular exposure toNi@S2and to NiO; however, no tumors were induced with CoO in this species. Tumorresponse in mice was consistently lower than in rats, and no apparent enhancing effectof the sulfide was evidenced.

In rats, almost all tumors examined histologically were rhabdomyosarcomas, witha few fibrosarcomas, particularly among those tumors induced with NiO. Difficulty wasencountered in definitely classifying the mouse tumors; whereas most appeared to befibrosarcomas, many of these were not typical, and a few were almost certainlymyomas.

It was concluded that nickel sulfide was probably the compound responsible for thecarcinogenic activity of the sample of metallurgical dust originally investigated (collected from the dust flue of a nickel refinery).

The apparent specificity of nickel and cobalt compounds for striated muscle tumorigenesis is discussed.

A metallurgical powder' collected from the dustflue of a nickel refinery and milled to particlesizes of less than S z has been shown to be a potentlocal carcinogen when introduced intramuscularly

S Tkn work was supported in part by grants-in-aid of re

,earch from the National Cancer Institute of Canada.

1 Estimated composition:

as a single injection in either rats or mice (5).In the same study, cobalt oxide (accounting forapproximately 1 per cent of the dust) was alsoshown to be carcinogenic to rats but not to mice.A majority of the sarcomas examined in theseinitial experiments appeared to be of muscle-cellorigin.

The present report deals with a determinationof the components of this refinery dust responsiblefor its tumorigenicity, and our observations on thecomparative tumorigenic activities of these andrelated compounds. The following six compounds,present in the original sample of refinery dust,were tested against rats : CuO, Fe2O3, NiO, Ni3S@,NiSO4 . 6H2O, CoO, as were also several 8uffidesof copper, iron, and cobalt. Subsequently, several

Per cent3.420.057.06.81.01.81.22.07.8

Cupric oxide (CuO)Nickel sulfate (NiSO4.6H2O)Nickel sulfide (Ni,S,)Nickel oxide (NiO)Cobalt oxide (CoO)Ferricoxide(Fe,O,)Silicon dioxide (SiO2)MiscellaneousMoisture

Received for publication July 10, 1961.

158

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Group no.and compound,

No. onexp.Effect.

@o.rats5Total

. .inject.

.sitesTotal

tumors.

at siteNo.tumorrat..Days

to 1sttumorAv.

latent.

period(days)Days

onexp.Sur

.vivoraat end

exp..

Miscellaneous

tumors1.CuO

2.FeOS. NiSO4•6H@O

4. NiO

5. Ni@S26.CoO7.CoS8. CuS

9. Cu,S10. FeS (47% 5)11. FeS(88.5% S)$2

3252

32

3232so80

3030SO28

5227

32

28242930

30305056f

64M@

64

45@29#5860

6060600

00

2611

36111335

0

00'@021

251228180

919697S02@j

15O@178196595

001608

@595

365842365603

60562762710

1418

S

051

19

2021181

lymphoma1 uterine fibroma4 mam. fibroadenomas1 mam. fibroadenoma1 lymphoma

1 mam. fibroadenoma1 reticulocytoma1 mam. fibroadenoma2 mam. fibroadenomas2 lymphomas

GIL@&N—ifetal Carcinogenesis. II 159

groups of mice were exposed to single intramuscular injection with those components of the refinerydust that had proved tumorigenic to the rat.

MATERIALS AND METHODS

The experimental animals used were a cornmercial strain of Wistar rats@ and CSH and Swissmice bred in our own colony. All animals werefrom@ to S months of age when placed on experinient. Housing and care were the same as reportedpreviously, as also was the method of preparationand administration of suspensions of the metallurgical powders (3).

showed no carcinogenic activity through observation periods of approximately @0months followingexposure.

Both the sulfide and the oxide of nickel inducedtumors at the site of injection in a majority of therats treated. However, marked differences in tumorigenic activity were observed between thesetwo nickel compounds both in proportion of injection sites at which tumors developed and in average latent period. Thus, the difference betweenthe 80 per cent tumor incidence at sites injectedwith Ni3S2 and the 41 per cent incidence withNiO is significant at the 1 per cent level (x2 =

TABLE 1

LocALTUMORRESPONSEOPRATSTOSINGLEINTRAMUSCULARINJECTIONS OF METALLURGICAL CoMPoUNDs

S Number surviving; 90 days after treatment.

t 20 mg. right leg;6 mg. left leg.@ 5-mg dose/thigh.

§Seventeenratstreatedinboththighs,eleveninonethighonly.#Fiveratstreatedinboththighs,nineteeninonethighonly.IIDifferencessignificant at 1 per cent level.*5 One rat (died 139 days) with large hematoma at injection site with indications of early sarcoma formation in thickened

fibrous wall.

As a general rule, a single @0-mg.dose of thepowder under test was injected into both theleft and right thigh muscles of each rat, whereasthe mice received injections of 5 mg. per thigh.Exceptions to this were made for reasons of toxicity and are noted in the text where they occurred.

RESULTSTable 1 summarizes the findings in the initial

series of tests for the carcinogenicity of the cornpounds listed, with the rat used as the test animal.

Apparently, cupric oxide and ferric oxide, at a20-mg. site dosage, and nickel sulfate at 5 mg/site

16.6@, df = 1, P < .01). Similarly, the S0@-daylatent period required on the average for NiOto induce a palpable tumor is significantly longerthan the 150 days required by Ni3S2 (“t―= 7.855,df = 44, P < .01).

Cobalt oxide had previously been shown to becarcinogenic to the rat under these conditions ofexperiment (3). A further test of this compoundagain resulted in approximately a 50 per centtumor response (Group 6).

The fact that injection with nickel sulfide resuited in tumors developing in 89 per cent of allthe exposed animals, whereas only 66 per centof the NiO-treated rats and 50 per cent of those2 Woodlyn Farms, Guelph, Ontario, Canada.

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Cs.scniooisNo.a.&rs

AUTOP

sianMETAITASESDISTRIBTJTION

-_______No.PercentLung

andlymphnodesLung

onlyLymphnode

onlyNickel

sulfideCobalt sulfideNickeloxideCobalt oxide21

29201220

167895

65852514

9322

1214

62

CarcinogenNo. ratsAv.

pro

time(days)Differ

ems(days)Nickel

sulfide 20

Nickel oxide 16

Cobalt oxide 12

Cobalt sulfide J 2455

88

71

4228*

12

29t

Cancer Research Vol.@ February 196@160

receiving CoO responded to the same or heavierdosages, suggested that possibly the sulfide itselfmight be directly concerned in this enhancementof responsiveness. To test this hypothesis, groupsof 30 rats each were exposed to single intramuscular injections of the following metal sulfides:

Cuprous sulfideCupric sulfideIron suffide (47 per cent S)Iron sulfide (38.3 per cent S)Cobalt suffide

Results obtained in these tests are included inTable 1 (Groups 7—11). Copper and iron cornpounds again failed to induce tumors. However,the suffide of cobalt proved markedly more activethan the oxide in numbers of tumors induced,although no difference developed in average latentperiod.

The great majority, possible all, of tumors induced by nickel sulfide and cobalt oxide werepleomorphic, highly cellular sarcomas, obviouslyof striated muscle origin (Fig. 1), showing frequent

metastasis to the lung and lymph nodes (Fig.@).Although a majority of the nickel oxide-induced

tumors could also be shown to be rhabdomyosarcomas (Fig. 3), about @0per cent were classedas fibrosarcomas (Fig. 4) . Several of the tumorsexamined from the CoS group were highly anaplastic and difficult to categorize definitely; however, here again the predominant histological typewas a striated muscle tumor (Fig. 5).

Metastases were very frequent in animals ofthe nickel suffide group, occurring in almost allanimals autopsied (Table @).Fifty-five per centof the cobalt suffide group also showed metastases,although these frequencies were progressively reduced when the primary tumor was induced byeither nickel oxide or cobalt oxide.

In all cases the primary tumors, once established, tended to grow rather rapidly and oftento a very large size around the site of injection.The rapidity with which the induced tumors resuited in the death of their hosts (tumor progression time) is compared in Table 3. Tumor-bearinganimals whose death was attributable to causes

TABLE 2

FREQUENCY AND DISTRIBUTION OF METASTASES IN INDUCED RAT TUMORS(MOSTLY RIIABDOMYOSARcOMA)

other than tumor growth and metastasis havebeen omitted from these calculations. The difference of 28 days in progression time that existedbetween animals supporting Ni3S2- and NiO-induced tumors was significant (“t―= 2.@19, df =34, P < 0.05), as also was the difference of 29days between average progression times of CoOand CoS-induced tumors (“t―= 3.41, 4f = 34,P < 0.01). No apparent difference existed in therapidity with which the cobalt oxide and nickeloxide-induced tumors caused the death of theirhosts.

Tumor response of mice.—Table 4 comparesthe tumor response of mice of two different strainsto nickel suffide and nickel oxide. The response ofSwiss mice exposed to CoO at a heavier dosehas been included for comparison.

TABLE 3

CoMPARisoNOF THE AVERAGETIME PROMFIRST APPEARANCE OF TUMOR TILL

DEATH IN RATS ExPoSED TO FouR CARCINOGENIC METALS

C Significant at the 5 per cent level.

t Significantat the1percentlevel.

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

TREATMENTTUMORRESPONSEMetalStrain

and.no. miceTotal

sitesinjectedDose!

site(mg.)No.

tumoranimalsPer

centaffected

sitesAv.

latentperiod(days)Av.

progr'ntime

(days) tDurationof exp.(days)Survivors

at endofexp..

Ni,S@

NiO

CoOSwiss-45

CSH.18Swiss-SOCSH-52Swiss-4651@

30100104925

555

1027

93358

333523250

225287131

13978478

4484768

2428

025671476 7518 12

GIu@N—Metal Carcinogenesis. II 161

The fact that both nickel compounds provedcarcinogenic to mice is in sharp contrast to theinactivity of CoO, even when administered indouble the dose and kept under observation for2 years. No differences can be noted in the tumorigenic activity of the two nickel compounds inmice either in average latent period or numbers oftumor-bearing animals. However, in this speciesit would appear that the progression time of theNiO-induced tumors was appreciably shorter thanthat of Ni3S2 tumors (Table 4), an observationin direct contrast to the condition observed inrats (Table 3). There also appeared to be a straindifference in susceptibility to the nickel compounds, with the C3H being more refractory thanSwiss on the basis of per cent tumor response.

DISCUSSIONAlthough the findings reported here demon

strate that several compounds of both nickel and

cobalt are carcinogenic, they also indicate thatin all probability nickel suffide was the effectivecarcinogenic agent in the original crude refinerydust investigated. This view is supported by therapidity with which Ni3S2 induced tumors in therat, the failure of CoO to induce tumors in mice,and the fact that 57 per cent of the refinery dustwas made up of the suffide of nickel.

Needless to say, direct extrapolation betweenthe induction of rhabdomyosarcomas in the ratand the occurrence of tumors of the sinus and lungin refinery workers is not justifiable—a fact thatis underlined by our observation that parenterally

TABLE 4

CARCINOGENICITYOF NIO, Ni,S,, AND CoO ON INTRAMUSCULARiNJECTIONINTO C3H AND SwISs MICE

a Øflly a few Swiss mice were injected in both thighs with Ni,SI, owing to its toxicity.t Averagenumber days from appearanceof first tumor to death.

administered cobalt oxide, while highly carcinogenic to the rat, is apparently inactive in mice.Nevertheless, the demonstrable tumorigenic capabilities of finely powdered compounds of nickeland cobalt in laboratory animals lends furthersupport to the view that certain such dusts in therefinery industry may constitute an industrialcancer hazard (9, 11). Sinus and lung canceroccurring in plants where nickel is produced bythe decomposition of gaseous nickel compoundshas been included in the list of Proscribed Diseasesin Britain for many years (4). During the lastdecade, however, evidence has been accumulatingthat this restriction of the industrial cancer hazardto the extremely finely divided nickel liberatedfrom carbonyl, Ni(CO)4, is probably not justifiable(1,4).

The reports of Heath (7) on the carcinogenicityof metallic cobalt and particularly of the apparentspecific affinity of this metal for muscle in therat are corroborated by our own findings with both

Both sexes were included in about equal numbersin the composition of these groups, with no sexdifferences being observed.

Exact histological classification of these mousetumors proved to be difficult. All were sarcomas,many of which show the characteristics of a fibrosarcoma. Frequent examples, however, were encountered of both NiO- and Ni3S@,-induced tumorswith the superficial features of a fibrosarcoma(Fig. 6), which on closer examinationshoweda lack of collagen, a considerable cellularity, andnumerous oval nuclei suggestive of a myoma (Fig.7). A few of the mesenchymal tumors of thistype also showed areas of cellular pleomorphism,with giant cells that may have been of muscle-cellorigin (Fig. 8). Metastases to the lung occurredonly occasionally, while the lymph nodes havenot been seen to be involved in mice. This absenceof dissemination through the lymphatics is inmarked contrast to the situation commonly observed in the rat (Table 2).

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

162 Cancer Research Vol. 22, February 1962

the oxide and suffide of cobalt. Of even moreinterest is the fact that this peculiar selectivityfor muscle has, in our experience, been shownto be particularly pronounced when the sulfidesof nickel or cobalt were used as the carcinogenicagent. Hueper (10) has reported a wide variationin histogenetic types of tumors resulting fromintrafemoral injection of nickel and has concluded“. . . that no tissue specificity exists in relation

to the carcinogenic action of nickel.― This is incontrast to our observations on the rat which arestrongly suggestive of a definite tissue preferenceby this carcinogen for striated muscle. A possiblereason (other than the different injection siteused) for this apparent divergence may have beenthe high incidence of a rather wide variety of“spontaneous―tumors in Hueper's strain of rats,as indicated from his report of ten neoplasms ofat least five different histogenic types in 23 controlanimals. In the light of both Heath's and ourown observations and the relative rarity of occurrence of striated muscle tumors, Hueper's findingof several rhabdomyosarcomas developing alongthe site of intrafemoral injection is of considerableinterest.

It is tempting to speculate what role the sulfidesthemselves may play in metal carcinogenesis, particularly in view of the significant enhancementto tumorigenic activity shown by suffides of bothnickel and cobalt over that of the oxides of thesemetals. As neither iron nor copper suffides inducedtumors, it seems safe to conclude that the suffidesper se have no carcinogenic activity. Thus, their

function, if any, must be assumed to be one oftrue co-carcinogenesis, possibly involving alterations in the solubility and/or binding of the metalcompounds. It has been suggested speculatively(5) that the introduction of excess metal into asystem of delicately balanced metals and metalenzymes might readily interfere with normal function in such a way as to lead to a specific respiratory change and mutation. Excesses of eithernickel or cobalt compounds might conceivablyaffect the same muscle enzyme system in the rat.

Heath (8) has reported that, in an experimentstill in progress, cobalt has failed to induce tumordevelopment in mice. This supports a similar observation by us, on the basis of which we hadconcluded that cobalt oxide could not be theeffective carcinogen in the original refinery dustinvestigated (3). Furthermore, in mice, tumorsclassifiable as definitely of muscle cell origin rarely

occurred in response to nickel compounds, and thefibrosarcomas that predominated appeared to beof a rather low order of malignancy, developingrelatively slowly and rarely metastasizing. Nomarked enhancing effect of the sulfide of nickeloccurred in this species. Such differences in response to the same carcinogens underscore boththe desirability of utilizing more than one speciesin screening and the risk of error inherent ininterpolation of findings from one species to another.

The consistently lower tumor response of miceof the CSH strain to nickel compounds, whencompared with the Swiss mice tested, is probablya reflection of a general or systemic tumor resistance rather than of a specific refractoriness oflocal tissues to the carcinogenic agent used (2).

ACKNOWLEDGMENTS

The assistance rendered by Doctors J. D. Schroder andK. V. Jubb of the Pathology Division,Ontario Veterinary College, Guelph, in histopathology and in the taking of photomicrographs, respectively, is gratefully acknowledged. Thehelp and cooperation received from Mrs. Agnes Anderson andMrs. N. Gohlich in both the care and routine inspection of theexperimental animals are gratefully acknowledged.

REFERENCES

1. ECKARDT, R. E. Industrial Carcinogens, pp. 98-99. NewYork: Grune & Stratton, 1959.

2. GILMAN,J. P. W. Heredity and the Response of MouseSkin to Carcinogens. Can. Cancer Conf., 3: 135-48. NewYork: AcademicPress, 1959.

3. Giuwe, J. P. W., and RUCICERBAUER,G. Metal Carcinogenesis I. Observations on the Carcinogemcity of a Refinery Dust, Cobalt Oxide and Thorium Dioxide. Cancer Research, 22: 152—67,1962.

4. GOLDBLATr,M. W., and GOLDBLATr,J. Industrial Carcinogenesis and Toxicology. In: E. R. A. Maaaws'raait(ad.), Industrial Medicine and Hygiene, 3:296. London:Butterworth Ltd., 1956.

5. HADDOW,A., and HORNING, E. S. On the Careinogenicityof an Iron-Dextran Complex. J. Nat. Cancer Inst., 24:109—47,1960.

6. Hna@rii, J. C. The Production of Malignant Tumours byCobalt in the Rat. Brit. J. Cancer, 10:668-78, i956.

7. . The Carcinogenic Action of Cobalt Ann. ReportBrit. Emp. Cancer Campaign, pp. 804-5, 1958.

8. —. Attempts To Induce Tumors with Powdered Metallic Cobalt in Other Sites and Species. Ibid., pp. 322-23,1959.

9. Husn'na, W. C. Experimental Studies in Metal Carcinogenesis. I. Nickel Cancers in Rats. Texas Rep, Biol. &Med., 10:167—86,1952.

10. - . Experimental Studies in Metal Carcinogenesis. IV.Cancer Produced by Parenterally Introduced MetallicNickeL J. Nat. Cancer Inst., 16:55—78,1955.

11. MORGAN,J. G. Some Observations on the Incidence ofRespiratory Cancer in Nickel Workers. Brit. J. led. Med.,15:224-34, 1958.

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

FIG. 1.—Nickelsulfide-induced primary rhabdomyosarcomain rat. X200.

Fia. 2.—Iliac lymph node, metastasis of rhabdomyosarcoma shown in Fig. 1. X200.

FIG. 3.—Nickel oxide-induced primary rhabdomyosarcomain rat. X100.

FIG. 4.—Nickel oxide-induced primary fibrosarcoma in rat.x100.

FIG. 5.—Cobalt sulfide-induced rhabdomyosarcoma in rat.X200.

FIG. 6.—Nickel oxide-induced primary sarcoma in mouse.Note the whorled pattern suggestive of leiomyoma. X 100.

FiG. 7.—High-power view of mouse sarcoma shown inFig. 6. Note the elongation and tendency to chromatin bandingin some of the nuclei. X600.

FIG. 8.—Nickel sulfide-induced primary sarcoma in mouseshowing giant cells and syncytial arrangements suggestiveof muscle cell origin. X150.

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

I

@1

‘0

,.1

br'

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1962;22:158-162. Cancer Res   J. P. W. Gilman  of Cobalt, Copper, Iron, and Nickel CompoundsMetal Carcinogenesis: II. A Study on the Carcinogenic Activity

  Updated version

  http://cancerres.aacrjournals.org/content/22/2/158

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/22/2/158To request permission to re-use all or part of this article, use this link

on May 21, 2018. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from