Prdtn f lnn n Vtr. V. Th Fbrblt ltr nd hn n n th Lvn ll · z f th lp. ltr tht hd rvd th nt ln 4 d...

Production of Malignancy in Vitro.

IV. The Mouse Fibroblast Cultures andChanges Seen in the Living Cells

By WILTON R. EARLE, senior cytologist, with the technical assistance of EDWARD L. SCHILL-ING, THOMAS H. STARK, NANCY P. STRAUS, MARY F. BROWN, and EMMA SHELTON,National Cancer Institute, National Institute of Health, United States Public Health Service

INTRODUCTION

Earle and Voegtlin (1) showed that20-methylcholanthrene added to culturefluid as a fine suspension in concentrationsranging from 1,000y to 0.2y per cubiccentimeter, has a definitely injuriouseffect on cultures of subcutaneous rat andmouse fibroblasts grown in chicken fibrinclot with a supernatant-fluid culturemedium of horse serum and chick-embryoextract diluted with saline solution. Thecarcinogen caused partial or completenecrosis, particularly at higher concen-trations, and always caused a retardationin the rate of growth. The study wascarried to 54 days after the first additionof the carcinogen.

Using the same culture medium andfibroblasts from the strain C3H mouse,Earle and Voegtlin (2) continued the studywith a concentration of 10y of carcinogenper cubic centimeter of culture fluid forperiods up to 250 days and found lensdegeneration than was shown at the higherconcentrations (1), but there was the samesevere initial retardation of the increase insize of the clump. Cultures that hadreceived the agent as long as 114 daysshowed a change in the architecture of theclump to a distinctly more epithelialliketype, which seemed to result from markedchanges in the cells, at least some of whichwas in the cell surfaces. The cells becamelaterally coherent; this change was gradual,progressing with prolonged exposure to

the carcinogen and affecting the cultureas a whole rather than a few isolated cells.

Unfortunately, the type cultures did notallow an accurate analysis of the cellchanges in the early stages. Changes after265 days in methylcholanthrene were alsonot investigated, owing to loss of the seriesfrom a bacterial contamination. This losscut short studies on the behavior of thecarcinogen-treated cells on reinjection.Only 12 C3H mice were injected with 1culture each of cells that had been subjectedto the carcinogen about 114 days and thencarried 83 days longer without additionof the carcinogen. The results of theseinjections were negative.

Just before the series was lost, it wasnoted that the control cultures were show-ing a trace of the same morphologic changeas that in cultures of the treated cells.This was ascribed to trace contaminationwith methylcholanthrene, which couldpossibly have occurred with the techniqueused.

The studies were later extended, withfour general aims:

(1) To confirm the previously observedcell changes with a more accurate controlof the concentration of methylcholan-threne.

(2) To follow with greater accuracythe progression of changes in the cells,both in the carcinogen and after removalfrom it, with particular attention to thesequence of changes during the first 60days after exposure to the carcinogen

165

at University of Illinois at U

rbana-Cham

paign on Novem

ber 14, 2012http://jnci.oxfordjournals.org/

Dow

nloaded from

http://jnci.oxfordjournals.org/

166 JOURNAL OF THE NATIONAL CANCER INSTITUTE

(this interval corresponding approximatelyto the latent interval for production oftumors by injections of the carcinogen invivo) and the cell changes after exposurefor more than 260 days.

(3) To attempt to demonstrate conclu-sively a change from the normai to themalignant cell by the action . of the car-cinogen in vitro. This of necessity in-volved the production of tumors byinjection of cells treated with the carcino-gen in vitro.

(4) To correlate the relative tumor-producing ability of the cells with morpho-logic changes observed in them.

Although these studies are as yet incom-plete, the present paper reports thetechnical methods used and the data ac-cumulated on a primary strain of mousefibroblasts, cultures of which were carriedin methylcholanthrene for various inter-vals up to 406 days. While the generalsequence of morphologic changes as ob-served in thé living culture is given, thispaper does not include any work withstained preparations. A description of theproduction of the tumors obtained by in-oculation is given in the next paper (3),while the pathology of the tumors will bepresented in the paper following (3a).Other studies in the series will appear inlater issues of this Journal. These willcover (1) observations on the mitochondriaand Golgi apparatus, and (2) metabolismof the tumors produced. Data from otherstained preparations, from motion picturereels, and analysis of the growth rates ofthe tumors are as yet incomplete and willbe presented at a later date. The equip-ment and method used in cleaning theglassware and the photomicrographic andthe microcinematographic equipment de-signed for use in these experiments aredescribed in the three preceding papers(4,5,6).

MATERIALS AND METHODS

In addition to the acid cleaning equip-ment and method already described (4),this investigation has necessitated the con-struction of considerable special apparatusand the elaboration of specialized pro-cedures. Sufficient details are given tomake clear the course of the experiment, aswell as the various precautions taken.

CLEANINO OF EQUIPMENT

All glassware on hand was first cleaned by usingthe acid-cleaning method (4). Then the glass-ware, used with normal-tissue-culture solutionsonly and not contaminated with heavy metallicsalts or with carcinogenic substances, was washedwith soap and water as usual, rinsed with distilledwater, and drained.

Witli the single exception of the cultures thatwere fixed in situ in Carrel flasks with Orth'sformalin-dichromate solution, the glassware usedwith tissue cultures was never employed with anysolution for fixation or staining. After fixation ofcultures in the Carrel flasks, the flasks were given apreliminary cleaning in a mixture of equal parts ofconcentrated nitric and sulfuric acids at 80° C. for24 hours and then, after being drained at a sinkreserved for waste acids, were rerun through theregular acid cleaning bath. In this way, therewas no contamination of the large acid cleaningbaths with chromium, and yet the carcinogen wasdestroyed. In flasks thus cleaned, cultures havegrown satisfactorily and have shown no sign of theirregular and poor growth so common in thosegrown in glassware cleaned with dichromate-sulfuric acid.

All waste carcinogenic solutions were destroyedby sulfuric acid in the acid bath if exceedinglyconcentrated; but if the concentration was notmore than 1y per cubic centimeter, the solutionswere flushed down a small covered sink used forthat purpose only.

After being cleaned, all glassware and othermaterials were wrapped in paper and autoclavedat 20 to 22 pounds steam pressure for 1 % hours,or dry sterilized at about 170° to 175° C. for 3 to4 hours, after the desired temperature was reached.

Rubber stoppers and rubber equipment thatmight come in contact with the culture fluid wereboiled repeatedly with an approximately 5-per-cent solution of sodium hydroxide, then washed


rbana-Cham

paign on Novem


Dow

nloaded from


CELL CHANGES OBSERVED

167

free from alkali, and rinsed repeatedly with dis-tilled water. An attempt was made to have aslittle rubber as possible come in contact withculture solutions. All rubber equipment wassterilized by autoclaving.

Until about June 8, 1942, all rubber items, in-cluding stoppers, tubing, etc., employed with solu-tions containing carcinogen or with carcinogen-treated cultures, were used once and then de-stroyed. Since then because of the critical short-age of rubber, stoppers used with cultures removedfrom the carcinogen for at least 60 days, arecleaned with sodium hydroxide solution as de-scribed and used over for such cultures. They arenever used with untreated cultures.

The dissecting knives and needles used in trans-ferring carcinogen-treated cultures to fresh flaskswere cleaned by soaking them in five to six con-secutive changes of acetone and benzene. Theywere then sterilized with dry heat at 170° C.

DISTILLED WATER

All distilled water used in rinsing glassware wasfrom the general distilled water system of the lab-oratory. All solutions used with the cultures wereprepared from this distilled water, which had beencarried through two additional distillations in adouble all-glass still and collected and stored in5-gallon pyrex bottles. The rubber stoppers usedwith the bottles were cleaned by repeated boilingin sodium hydroxide solution. Care was takenthat the distilled water did not touch the rubber.

ISOTONIC SALINE SOLUTION

The saline solution used for washing cultures andfor diluting culture media consisted of the follow-ing ingredients:

GramsSodium chloride ...................... 6.80Potassium chloride .................... . 40Calcium chloride ..................... . 20Magnesium sulfate .................... . 10Sodium dihydrogen phosphate.......... .125Sodium bicarbonate ................... 2.20Dextrose ............................. 1.00Glass-distilled water to 1,000 cc.

All concentrations were calculated in terms ofanhydrous salts, which were of reagent qualityand were purchased in relatively large lots so thatthere was no change in lot number for most ofthe salts used. With the exception of the calciumchloride, the salts were dissolved, in the ordergiven, in about 600 cc. of water for Bach liter ofsolution. The calcium chloride was kept in theform of an analyzed stock solution containing 0.1

gm. anhydrous salt per cubic centimeter. Thecorrect amount of this solution was added to 100to 200 ce. of water for Bach liter of final solution,and this in turn poured into the rest of the solu-tion with agitation. The pH of the solution wasset at 7.4; if necessary, a little carbon dioxide wasrun through it from a cylinder of compressedgas. The solution was then filtered. Inasmuchas no such physiologic saline relying on bicar-bonate for its buffer action can be filtered undervacuum without a severe shift in pH to the alka-line side, all filtrations were carried out by usingair pressure at 3 to 5 pounds. The filter candlesemployed were never used for any other purpose.During the early part of the experiment the solu-tion was made up and filtered in lots of 41. througha 1-by-5-inch Mandler candle of normal porositydesigned to handle 7 to 9 pounds of air pressure;later the solution was made up in lots of 32 1. andwas simultaneously filtered through two 2-by-10-inch Mandler candles to give two final lotsof 16 1. each. The filtering equipment is shownin figure 1, A and B.

In autoclaving such large units as the 32-1. salineapparatus, the stopcock permitting access of airto the apparatus was wrapped but left open, whilein Bach of the two large containers about 100 cc.of glass-distilled water was placed, thus insuringthe formation of enough steam within Bach con-tainer to act as a sterilizing agent. These unitswere autoclaved at 20 pounds for 2 hours andthen allowed to dry out. After filtration, thesaline solution was run into flasks which weresealed with rubber stoppers and stored at 3° C.

During the whole interval of preparation andstorage the pH of the saline solution would riseabout 0.1 unit. Since the solution was desiredat pH 7.5, it was originally set at about pH 7.4.The pH of the other solutions was similarly con-trolled so that the pH of the final culture mediawas about 7.5.

HORSE SERUM

While no particular selection was made as tothe age and sex of the horses from which the serumwas obtained, records show that all the horseswere adult geldings ranging up to 33 years of age.Most of the serum was from horses estimated tobe about 16 years old. All serums were clear,and no trouble was ever experienced fromhemolysis.

The jugular vein was punctured with a 4-mm.(outside diameter) Graefe trocar, with stopcock,connected with a rubber tube, and the blood col-lected in 3-1, lots in sterile glass cylinders about


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 1. A, Equipment used to filter saline solution and horse serum in lots of 41. or less; B, Equipmentused to filter saline solution in lots of 32 1. The small bottles are of 5-gallon capacity, the larger one

is of 12-gallon capacity< All are made of pyrex.

100 mm. in diameter and 450 mm. high, whichwere kept cool in crushed ice to reduce changes inthe blood. Usually 9 1. was obtadned from onehorse. When horses were bied on Saturday andallowed to rest over the week end, the bleedingdid not interfere with their regular work. Theblood was allowed to clot, and the clots wereloosened from the cylinders which were held at 3°C. overnight. During this handling sterility wasnot maintained although an attempt was made tokeep the blood reasonably clean. Next day, theserum was drawn off with a 200-cc. bulb pipette.The clots were minced with scissors, the fragmentswere centrifuged, and the expreseed serum wasremoved. The total lot of serum was recentri-fuged to eliminate any remaining red cells whichwould be disrupted by freezing, then pooled, andstored in 1,000-ec., flat-bottomed, boiling flaskssealed with rubber stoppers. The flasks were fillednot more than half full to prevent their breakingon freezing, then stored at —10 0 , and held frozenuntil needed. When needed, the serum wasthawed out, recentrifuged to remove any fine pre-cipitate that might (very rapidly) clog the filter,and filtered under pressure through a 1-by-5-inchMandler candle. The smaller equipment andthe method for filtering saline solution were used.The filtered serum was stored at 3° in small flaskssealed with rubber stoppers. Since the filtercandles were used repeatedly, they were carefullywashed with a dilute sodium chloride-sodiumbicarbonate solution, then with glass-distilledwater. Filters used for horse serum were reservedfor that solution only. The serum when used was

usually less than 60 days old. In other work,serums up to 9 months old have been used withoutany observed deleterious action.

PLASMA

Plasma was obtained from hens weighing 5 to 8pounds and not more than 14 years old. It wasnecessary to watch carefully the age of the hens,as the use of plasma from older hens has beenfound to result in poor growth of the cultures. Asimple glass cannula sterilized in heavy minraloil at about 140° C. was inserted in the carotid,and the blood collected in a narrow-mouthed,heavy-walled, iced, paraffined, 100-cc. centrifugeflask, which was stoppered with cotton. The sur-plus melted wax from the tube had been drainedinto the cotton. Great care was taken not to over-heat the minral oil and wax. The blood wascentrifuged in ice for 5 minutes at 3,000 r. p. m.;the plasma was drawn off with a 5-ec. pipette andstored at 3° in a paraffined, cotton-stoppered,50-cc. centrifuge flask. This plasma was usedwithin 30 days. No anticoagulant was used.While some tubes were lost because of spontane-ous clotting, the Toss was never so severe as tonecessitate the use of heparin.

EMBRYO EXTRACT

Embryo extract was prepared from 9-day chickembryos. The whole unwashed embryo wasminced by running it through a 30-cc. Luersyringe in the lower end of the barrel of whichwas a 28-mesh monel-metal wir'e gauze. The


rbana-Cham

paign on Novem


Dow

nloaded from



169

minced tissue was received in a 100-cc. narrow-throated centrifuge flask wrapped integrally withthe barrel of the syringe. Approximately one vol-ume of saline was added and the whole agitatedbriefly by sucking it in and out of a 25-cc. wide-mouthed bulb pipette. The flask was rubberstoppered and centrifuged at 3,000 r. p. m. for20 minutes. The supernatant fluid was pipettedinto 50-cc. centrifuge flasks, rubber stoppered,and frozen at once with carbon dioxide snow tokilt any living cells. The solution was keptfrozen until needed, at which time it was thawedout and recentrifuged at 3,000 r. p. m. for 20minutes to clarify it directly before use.

Embryo extract was prepared two, or moreusually three times a week. An attempt was madeto use it when less than 5 days old, but in a fewinstances as a result of some accident older extractwas used for a single change of fluid on thecultures.

Of the various culture media used, the embryoextract was probably subject to the greatest vari-ation, both from lot to lot of eggs and possiblyfrom season to season. In some earlier work inwhich refractometer readings were made ondifferent lots of embryo extract, a wide differencewas found in lots prepared on different days.Because of these variations, particular care wastaken to use the same lot of embryo extract foreach experimental culture and its control.

PREPARATION OF METHYLCHOLANTHRENE

The 20-methylcholanthrene used was all of thesame lot, which was different from the lots used inBarlier work (1, 2). It was purified by Dr. J. L.Hartwell, of this Institute, and had a melting pointof 179.3°-180.0 ° C., corrected.

Whereas in the previous work methylcholan-threne had been used in the form of a fine sus-pension, in this experiment it was dissolved in theserum. A lot of 2 1. of serum was placed in aglass-stoppered 4-1. pyrex bottle; about 40 mg. ofthe crystalline methylcholanthrene was weighedout into a small agate mortar, one or two drops ofserum were added, and the mixture was rubbedto an extremely fine, creamy paste; more serumwas added and the suspension transferred to theserum in the bottle quantitatively. The serumwas held at 3° C. for 24 hours and shaken occasion-ally, and then shaken in a mechanical shaker for24 hours. It was next run through filter paperto get out any particles that might clog the filter,and then filtered, through a Mandler candle intoflasks that were sealed with rubber stoppers. TheMandler candle was used once, then destroyed.

This method of solution of the carcinogen wassuggested by Dr. Egon Lorenz, of this Institute.Dr. Lorenz also analyzed spectroscopically thefirst lot of solution. The determination showeda concentration of between 2.07 and 2.57 ofmethylcholanthrene per cubic centimeter. Laterlots were not analyzed but were prepared with thesame technique and were considered to haveapproximately the same concentration.

Although an attempt was always made to havethe relative size of the control and carcinogen-containing lots such that they would last for anequal period, it was rare that the two lots ran outat exactly the same time; usually one was ex-hausted a few days ahead of the other. Since nodifferences have been detected in the action ofdifferent lots of serum prepared by the methodused, no hesitancy was felt in using up the last fewflasks of any lot. To avoid any hazard of acci-dentally using a carcinogen-containing solution innormal cultures, the carcinogen-containing solu-tion was always segregated. As a further pre-caution, the container of the normal solutioncarried a blue or black label, that of the carcinogen-containing solution a red label.

DUPLICATION OF CULTURES

To prevent the loss of a whole strain of culturesthrough bacterial contamination of a stock solu-tion, duplicate independent lots of solution wereused, and all culture series were so divided thatcontamination of any single solution would resultin the loss of only half of any one series or strainof cultures. No such contamination of a stocksolution has occurred.

CARREL FLASKS

All cultures were prepared and carried in modi-fied Carrel type D flasks made of pyrex and havingan outside diameter of 33 to 34 mm., a heightthrough the flat area of approximately 10 mm.,and floor and roof thickness of about 0.4 to 0.6mm. This thickness gave good optical definitionwith a 16-mm. objective, yet the flask was not sofragile as to be hazardous when used with carci-nogenic solutions. The throat length was approx-imately 33 mm., measured from the juncture ofthe throat with the top of the flask; the throat hada wall thickness of 1.0 to 1.5 mm. and an insidediameter of 10.0 to 10.5 mm. These changes inthe throat were necessary to permit the sealing ofthe flasks with standard size 00 rubber stoppers. '-

1 Flasks of this type may be obtained from E. Machlett& Son, 220 East 23d St., New York, and from AopfGlass Apparatus Co., 192 3d Ave., New York.


rbana-Cham

paign on Novem


Dow

nloaded from



INCUBATION OF CULTURES

All cultures were maintained at a temperature ofapproximately 38.2° to 38.8° C. in an incubatorhaving an ample water jacket. The heaters in theincubator warmed the water jacket rather than theinternal air space, while the two thermoregulators,which were wired in series, were immersed in thewater bath. In this way sudden overheating of theincubator for a short period after the doors hadbeen opened and closed again was eliminated.One thermoregulator was adjusted for a tempera-ture 0.3° higher than the other so that it did notfunction at all but merely served as a reserve regu-lator. An alarm bell was wired into the circuit ofthis res rve regulator. The air in the incubatorwas subjected to forced circulation to insure evenheating. In the incubator, the culture flasks werearranged in racks of 10 (fig. 2, A) and placed onshelves hinged at the left end, while a motormechanism mounted on the roof of the incubatoralternately raised and lowered the right end of theshelves through a cycle of about ±3 ° from thehorizontal at a rate of six cycles per hour. Thiscaused the culture fluid in the flasks to be washedover the surface of the clots and helped to cum-mate local changes in the culture medium withinthe flasks. The racks held the flasks level. If theracks were ever soiled with carcinogen, they weredestroyed.

A temperature of about 38.5° C. was maintainedduring the visual examination of the cultures byhaving the microscopes surrounded with a thermo-regulated air bath su(ïiciently large to take severalracks of cultures. The photographic equipment

^ Z MM DIAMETER PLATINUM IRIDIUM

EDGE 90 MMGLA55

HONEDLENGTN 5OMM

220 MM OVER ALL

40 MM^DIZJIMI'R

.III, 1I1I III. 1111 I11t

N>WALL NOT LE55 TN ^EXTERNAL DIAMETERIA MM TMICN AFMOXIMCTELY 9-I0MM

E 09 MM.

FiGURE 2.—A, Flask racks, made of wood withmetal ends, and springs of sheet phosphorbronze 0.25 mm. thick. The pins are brassescutcheon pins; B, Platinum-iridium spatulaused to transfer cultures. The curved portionof the spatula is flattened by hammering andthen honed to a sharp edge; C, Pipette forhandling small lots of solution like plasma.

was similarly held to constant temperature.With these precautions, the only serious intervalsof lowering the temperature were when the cul-tures were reduced to room temperature duringthe thiee weekly changes of culture fluid. Theseintervals of cooling averaged about 4 hours eachtime and were constant for the whole series. Ithas as yet been technically impracticable to elimi-nate them.

LIGHT

To eliminate complications that might resultfrom the action of short wavelengths of light oncells photosensitized by methylcholanthrene, allcultures were handled, examined, and photo-graphed by orange light. All light sources ofgeneral illumination in the rooms were shieldedwith five layers of No. 300 Tango shade of cello-phane. The microscope lights were 6-volt, 108-watt, ribbon-filament, incandescent bulbs shieldedwith water cells and deep orange filters (CorningNo. 351), cutting off at about 500 mµ. Forphotomicrography where the light was concen-trated on the culture by a condenser, the light wasfurther shielded with a Corning No. 397 heat-absorbing glass filter. While it cannot be saidthat no white light ever reached the cultures, thetotal amount reaching them during the wholecourse of the experiment was extremely low.

PREPARATION OF THE ORIGINAL EXPLANT

The original explant tissue was dissected out,cut up rapidly into strips about 2.2 x 6.0 mm.,and placed in saline in depression spot plates.The cultures were then planted as described inthe following section.

TRANSFER OF CULTURES

Cultures were transferred to fresh flasks aboutevery 27 to 35 days. The series of flasks containingthe cultures were arranged on the table in orderof transfer; Bach flask was flamed very hot aroundthe throat, the stopper removed, a sterile corkinserted lightly, and the flask set aside to allow thethroat to cool. By the time the last of the groupof flasks had been flamed, the throat of the firstone had cooled sufficiently to allow handling.The clot was loosened from the Hoor of the flaskby means of the spatula (fig. 2, B); and after avery light flaming of the throat which did notheat the glass through, the clot was slid onto theplatform of the dissecting dish assembly, whichconsisted of a covered pyrex petri dish, 15 mm.high and 100 mm. in diameter, with half of a


rbana-Cham

paign on Novem


Dow

nloaded from



171

75-by-10-mm. dish (either top or bottom) placedin it to form a platform just below the surface ofthe lid of the pyrex dish. Dissections were madeon this platform, and unnecessary fragments ofclot and waste fluid were pushed over the edgeinto the surrounding moat. Each individual cul-ture was dissected separately, a heavy dissectingneedle and a No. 3 or 5 Graefe cataract knifebeing used. Separate dissecting dishes and in-struments were used with each culture. Theheavy ridge at the edge of the clot and all parts ofthe clot containing no cells or too few cells forsatisfactory transfer were cut away. From theremainder, as many strips as desired were cut.

The size of the explants varied somewhat. Inthe first few generations an attempt was made toget strips about 3 mm. wide and 15 mm. long.However, better results were obtained with stripshaving a somewhat greater width, and in the laterwork an attempt was made to get them 4 to 5 mm.wide and approximately 20 mm. long. Usuallythree or four strips were cut at right angles acrossthe original explant. In some instances only twostrips, lying parallel, and just lateral to the oldexplant, were cut, and in others diagonally cutstrips were used. The exact type of explant wasso chosen as to give what seemed the best possiblechance of growth from the old sheet of cells.

The explants were handled rapidly to avoid theirdrying out; as soon as they were cut, they wereplaced in approximately 0.7 cc. of saline solutionin one of the depressions of a spot plate (ComingNo. 7220) set in a 150-by-20-mm. flat pyrex prep-aration dish. The adhesive label of the flask wastransferred to the top of the dish to mark the cul-ture, while the order of explants was identical withthat on the werk sheet. Each strain of cultureswas kept in a separate depression plate so thatthere was no chance of confusing individual cul-tures or different strains. Each culture was labeledand records were kept on the origin of each sub-culture.

PREPARATION OF THE CLOT

From 0.4 to 0.5 cc. of chicken plasma was in-serted in each Carrel flask from a 5-cc. pipette(figs. 2, C, and 3). In the early part of the ex-periment 0.4 cc. was used, but this gave a clotwhich in some instances was so soft that holesdeveloped near the end of the transfer interval.When the amount was increased to 0.5 cc., nofurther trouble was experienced. The plasma wasshaken over the entire surface of the Hoor of eachflask, and 0.7 to 0.8 cc. of a mixture of 40 percenthorse serum, 20 percent chick-embryo extract, and

548963-43--4

FIGURE 3.-Pipette and burette unit for plantingcultures. The rubber bulb on this pipette is of5-ee. capacity. For washing cultures and addingfresh culture fluid, a second burette unit wassubstituted for the pipette.

40 percent saline was then added. The explantwas inserted in the flask at once by means of thespatula, flattened out, and centered. Each flaskwas closed with a rubber stopper, marked with apreviously prepared adhesive label, then leveleduntil the mixture had formed a firm clot, andracked. Later all flasks were transferred to therocking shelves in the incubator. Twenty-fourhours later, 1 cc. of a mixture of the horse serum,embryo extract, and saline in the foregoing pro.portions was added. For cultures that were toreceive carcinogen, it was contained in the horseserum in such concentration that each cubiccentimeter of the final supernatant culture fluidcontained approximately 17 of carcinogen.

WASRING AND RENEWAL OF FLUID

CULTURE MEDIUM

The fluid culture medium was changed threetimes weekly, usually on Monday, Wednesday,and Friday. The only serious deviation from thiswas near the end of December 1941 and in earlyJanuary 1942, when the change was made onlyevery 3 or 4 days. The resultant injury to the


rbana-Cham

paign on Novem


Dow

nloaded from



normal cultures was severe and was reflected inthe noticeably poorer growth for nearly threegenerations.

In changing the fluid medium, the flasks werewiped off with a towel moistened with absolutealcohol and the cultures arranged on the table innumerical sequence within each set; the order ofthe sets was as follows: (1) Cultures which hadnever been subjected to carcinogen (all flaskscarried white labels); (2) cultures which had atsome time been subjected to carcinogen but wereno longer (flasks carried green labels), the varioussets being arranged in order so that those mostrecently removed from the carcinogen werealways handled last; and (3) cultures which werebeing carried in carcinogen (flasks carried redlabel) and which were always handled last of all.

Aseptic precautions were taken and all normalsolutions made up before any carcinogen-contain-ing solutions were opened, and the culture fluidswere correctly mixed in the separatory funnels ofthe burette units (fig. 3). Each funnel was thenmounted on its burette which was locked in theburette rack. The other half of the rack carrieda similar unit filled with saline solution.

The throat of each flask was flamed, the rubberstopper slipped out, and the end of the throatagain flamed very hot. The old fluid was suckedout with a sterile glass needle attached to a fluidreservoir and a source of vacuum (fig. 4). Withthe burette arrangement, 3 cc. of saline solutionwas run into the flask, which was lightly stoppered

with a sterile cork, and set aside, until each flaskin the set had received saline. The suction needlewas sterilized with heat after each use. After thelast flask in the set had received saline, the firstone was uncorked, the neck lightly flamed, thcsaline sucked out, 1 cc. of culture fluid added,and the flask at once sealed.

The burettes (fig. 3) were of 25-cc. capacity,graduated to 0.1 cc., and sturdily built to takerelatively rough handling. The tip of each wasprotected, while in use, with a short glass sleeveheld on by a cork collar. The open end of thesleeve was flared slightly for easy insertion of thethroat of the flask which could be slipped in adistance of 7 to 8 mm., while the burette tipprojected through the cork collar about 3 or 4mm. The sleeve fit rather snugly around thethroat of the flask. The freshly flamed throatinserted into the sleeve was free from bacteria,shielded so that bacteria could not fall into it, andso centered around the burette tip that the latternever touched the inside of the throat. In thisway the likelihood of spreading bacterial con-tamination through consecutive flasks was ob-viated.

A wad of cotton in the upper end of the sleevepermitted air to flow into the burette as fluid wasrun in and withdrawn. All stopcocks were lubri-cated with sterile (autoclaved) heavy petroleumjelly. At the end of each run the equipment wasdisassembled, the cork collars wcre destroyed, andthe glass parts of the apparatus were cleaned (with

DETAIL OF VALVE ONSUCTIOPJ

DETAIL OF SUCTION NEEDLE

99

FIGURE 4.—Suction needle unit with heater. The heater diagram is drawn disproportionately large toshow detail. Details of suction needle and valve assemblage are also enlarged.


rbana-Cham

paign on Novem


Dow

nloaded from



173

acid if the unit had been used with culture flasksthat contained or had ever contained carcinogen).

The suction needle (fig. 4), a piece of standardwall pyrex tubing 6 mm. in outside diameter and150 mm. long, was attached to a length of gumrubber tubing 60 mm. long, with 6-mm. bore(1.5-mm. wall thickness). A glass bead of specialdesign was inserted midway the length of thetubing to serve as a valve. When the outside ofthe rubber tube was squeezed gently with thefingers, the resulting deformation caused a leakageof air around the bead. The tubing was suffi-ciently soft and pliable that a relatively gentlepressure near the bead gave a vigorous suction atthe tip of the glass needle. Figure 4 shows indetail the setup. The bottles containing the glassbeads Bach had a 1 -inch layer of sulfuric acid inthe bottom, which served to trap any possiblespray containing traces of carcinogen that mightbe carried over into the house vacuum line. Thesulfuric acid was replaced about every 6 to 8weeks. The first of the three bottles connectingwith the vacuum outlet contained lime-soda toneutralize any acid spray carried to it.

A small electric heater to sterilize the suctionneedle between uses is shown in figure 4. Theheating element was wound around a 15-mm.(inner diameter) metal tube about 6 inches longand insulated from it with several layers of asbes-tos paper. The tube was sloped so that the needlecould be rapidly inserted without danger of itssliding out. A small glass test tube was used asa lining to protect the metal tube from corrosionand to prevent the carrying over of metal oxidesinto the culture flasks. It was changed at the endof Bach series of cultures.

The winding was so designed that, within a fewminutes after being switched en, the heater reachedand maintained a temperature of about 340° C.within the lumen of the glass tube, as measuredby a thermocouple, white a suction needle con-taining several drops of water boiled dry within 7seconds after insertion into the heater tube. Inas-much as the interval between using the suctionneedle on two consecutive flasks was about 20 to 30seconds, the hazard of carrying a bacterial infec-tion from one flask to the next was avoided.

Figure 4 shows the metal casing projecting overthe open end of the heater, and allo the hose con-nection with the acid shield on the vacuum line.Through this line a slight current of air was keptflowing over the front outlet of the heater in orderthat any traces of fumes liberated from the heatedneedle might be carried into the sulfuric acid trapand not released in the room.

When this unit was originally designed, the

question arose whether it could be used safely onthe cultures lest the hot needle injure them by itsown heat or by fluid heated to boiling and ex-pelled from it onto the sheet of cells. The systemfinally worked out has now been used for more than2 years and has given satisfactory results. Toremove fluid from a culture flask, the needle wasremoved from the heater and at once inserted intothe open mouth of the flask. Before it entered,however, the operator had already exerted pres-sure on the rubber tubing around the glass-beadvalve and so established a rapid flow of air throughthe glass tube. The tube was lowered along theinner bottom surface of the flask throat; mean-while an uninterrupted flow of air was maintainedthrough the needle. As the needle tip reached thefloor of the flask, the Jatter was tilted very slightlybut not so much that fluid ran into the open end ofthe throat. The fluid was sucked out cleanly andin a fraction of a second. The needle was removed,the pressure on the valve released, and the needlereturned to the heater. In this way a needle couldbe used for evacuating a series of 150 flasks with-out clogging. If the needle clogged after a shorttime or if it did not suck out the fluid cleanly andalmost instantaneously, it was not being correctlymanipulated. Usually in such cases the valve wasnot open sufficiently. Once an operator wastrained to use this device, the average time neces-sary to handle each flask was approximately 60seconds. Cultures used with this unit have nevershown local injury to cells near the throat of theflask, and only a very few have had to be discardedbecause of some incorrect behavior of the suctiondevice.

After use, the suction-needle equipment and itsfluid reservoir were disassembied. Rubber unitswere destroyed, and glass units given the usual acidtreatment. The tip of the glass inlet tube of thefirst acid bottle on the suction unit was flamed veryhot to eliminate any traces of carcinogen thatmight have adhered from spray passing throughit, and it was capped until the next change of fluid.The aluminum front of the electric heater washeated with an open flame as hot as possible with-out fusing to destroy any traces of methylcholan-threne that might have been left on it by theneedle.

No culture that was receiving carcinogen or thathad ever received it was ever opened until allnormal cultures had been washed, fresh fluidadded, and the cultures sealed. Similarly, no cul-ture that was receiving carcinogen was openeduntil those that had once had carcinogen but wereno longer receiving it were resealed. There wasno exception to this rule. Careful checking and


rbana-Cham

paign on Novem


Dow

nloaded from



the different colored labels on the three seriesprecluded any chance disarrangement in the orderof handling.

In a few instances when one or two drops of acarcinogen-containing solution were accidentallyspilled on the table or the Hoor during the hand-ling of the cultures, the solution was cleaned up atonce by wiping the area repeatedly with a succes-sion of cotton swabs soaked in acetone or benzene.

PROCEDURE FOR RECORDING CHANGES IN

SIZE OF CULTURES

One of the usual techniques, and the simplest,for estimating the enlargement of a culture is todetermine and plot the area at different times afterplanting. Parker (7) and Cunningham and Kirk(8), among others, have pointed out that this is nota true index of growth but that it is complicatedby ether factors. When cells are treated withmethylcholanthrene, the number of cells in Bachunit area of the treated cultures is different fromthat in the controls; the curves obtained are,therefore, certainly not a reliable index of growth.

Even though this be true, the need was feit forsome simple and easily applied index which couldbe used to keep some record, other than that ofcell description, of the deviation of the carcinogen-treated cultures from the normai. Since the ear-lier work had shown that the action of the carci-nogen could be detected within a very short timeby the lagging of the rate of increase in area or

FIGURE 5.—Ruling of ocular micrometer disk forrecording changes in the width of cultures.The ruling is displaced slightly from the centerto facilitate the measuring of small cultures.

width of the treated cultures and since this retar-dation seemed to be associated with the action ofthe carcinogen, it was feit that this index was worthusing as a working guide. The index, as in the pre-vious work with strip-shaped explants (2), con-sisted of a record of the changes in the averagewidth of a culture, the length of the strip beingleft out of consideration. In all instances thewidth was determined by measuring the cultureunder a compound binocular dissecting micro-scope fitted with paired 1.0 X objectives and9 X oculars, ene ocular carrying a special microm-eter disk divided, as shown in figure 5, to readnumbered intervals of 1.0 mm. actual size of theobject, with ruled subdivisions of 0.2 mm. Thisgave a simple and rapid means of estimating thewidth of the cultures to 0.1 mm. without subject-ing them to intense light. The data obtained bythis method of measuring were collected on all cul-tures and plotted to give curves for each genera-tion of Bach culture strain. These curves arenot presented in this paper on account of the dif-ficulties in reproducing them for publication.

The interruption of these curves because ofperiodic transplantation of the cultures intofresh flasks offered a serieus difficulty to obtaininga satisfactorily connected picture in the culturestrains under experiment. To show this progres-sive change, the width of the zone of new growthwas taken at 5, 10, and 15 days after explantation.These three readings were averaged and plottedas an average width at a time 10 days after ex-plantation (fig. 6). In a few instances where thewidth curves ran only 13 or 14 days, the valuefor 15 days was obtained by extrapolation. Thesevarious average points on any ene strain of cul-tures were then joined by a continuous line. Inthis way, fór any ene carcinogen-treated strainene continuous line was obtained, and for itsnormal control cultures another line was simi-larly obtained. The number of cultures averagedto give each point on the curve is shown by thenumber opposite the point, while the letteradjacent indicates the cell strain used to givethat point. This notation is necessary since oftenene set of controls was used for more than eneexperimental series. To show the relationship ofcontrol and experimental culture when this oc-curred, the curve from the control was repeatedfor each series for which it was used. Similarly,where strain N, for instance, originated fromstrain 0, the early part of the curve from strain 0has been drawn in for the strain N curve to com-plete the picture of the curve. The time Bachexperimental strain was left in the carcinogen is


rbana-Cham

paign on Novem


Dow

nloaded from



175

1941 1942 1943AUG. SEPT OCT NOV. DEC. JAN FEB. NAR. APR. MAY JUNE JULY AUG. SEPT OCT. NOV. DEC. JAN. FEB. MAR. APR MAY5 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 I 15 1 15 1 15 1 . 15 115115 1 15 1 15 1 15 1 15 1 15 30

DAYS--+O 25 50 75 100 125 150 ' 175 200 225 250 275 300 325 350 375 400 425 450 475 500 525 550 575 600 625 650

D e NPS MFB 15

IQ.(I D-6 `^y 4 - N-g N 7 H--8__yd0_- H a^N.9 IQH HB --^^ H•10 H•7 M-7 H-9

H-8 H-10 M-8 N-B H-10 H-B H-7

5.Q H-6 H•7 H-8 H-8 SIH-8

O -6 DAYS IN METHYLCHOLANTHRENE o

150

100

50

0

150

100

J w

0

150

K lal --

0-8

D-6 1 '2 i-8 I-8 H-8 H-10

1-3 1-4 1-8 1-8 I-91-8

m13DAYS IN METHYLCHOLANTHRENE

D-7J-9

J-4^• v__ ~^ _- •s- J•-8 -- ^✓

32 DAYS IN METHYLCHOLANTHRENE

K-10

D-9 0- 6_7 K-8 ' K.9 K-9


J 1 J77., o -1 5,0

J-10 J-8 J-90

(l) 150

W 100

w G 500

JJ 150

Z L 100

54

0W

15.0

D D 1 ^0

0= 150

o N 1003 50

Z 0

N I50W 0 IOAFY 50

Z 0

G•8 G.g

G-e~ p10

^ . 76 DAYS IN METHYLCHOLANTHRENE

o-7 alo o; - 8^ D9__D•10

a9

D 10 D-8 L-8 L-40 L-9 L-10 L-9 L-9 L-1O L.9 L 9 L 8 L-g L 8

5.0

III DAYS IN METHYLCHOLANTHRENE 0

D-8 D-9

p •`^\^ -̂p^ ^^^ .-IO- D-10__0- 'p^D^1 •~1 D̀-10 D-7 p 0-^DrB IOA

D-9 al0 ^•

170 DAMS IN METHYLCHOLANTHRENE O

0.5

rj, Si-3 0.7 8 ^ Y

N•9 `` -̂9 ^ X10- '1O X10

N 5 ^--9N.8 N-8 N_7 N•8 N_

Q

150

I0d04 0.7 0.8

o-lo o-s y^ o e o-7 5.0

406 DAYS IN METHYLCHOLANTHRENE C


0-9 _ 10 o-e

DAMS

250 275 300 325 350 375 400 425 450 475 500 525 550 575 600

515 1 15 1. 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 1 15 115 1 15 1 Kf 115 515 1 15 1 15 1 15 70AUG. SEPT. OCT NOV. DEC. JAN. FEB. MAR. APR. MAY JUNE JULY AUG. BEPT. OCT. NOV. DEC. JAK. FEB MAR, APR MAY

1941 1942 1943TIME IN DAYS

FICURE 6.—Continuous width curves for all strains. The carcinogen-treated cultures are shown by thesolid line; the control is shown by the broken line. The time the treated cultures were left in car-cinogen is shown by the heavy black line. Each curve is identified according to its strain of cellsby the large letters opposite it.


rbana-Cham

paign on Novem


Dow

nloaded from



shown by the length of the heavy black line belowthe respective curves. These continuous widthcurves are presented through May 18, 1943.

This type of curve was plotted for each of thestrains studied. The curves (fig. 6) are designatedcontinuous width curves.

RESULTS

ORIGIN AND DESCRIPTION OF THE PRIMARYNORMAL-CELL STRAIN

The strain of fibroblastlike cells used wasoriginally taken from a 100-day-old malemouse of the C3H strain, Andervont sub-strain. On October 18, 1940, the mousewas decapitated. Without including mus-cle tissue and the regional lymph node, theconnective and fatty tissue pad along theside and just in front of one of the hind legswas removed and cut into six strips eachabout 2.2 mm. wide and 6.0 mm. long.Each strip was planted in a separate flask.When examined just after planting, muchof the tissue was seen to consist of fat cells.Five days later, the only sign of migrationwas an occasional projecting cell. At 10days, the width of the strips had increasedfrom 2.2 to 10.2 mm., and the growth wasrecorded as luxuriant and even but some-what loose; there were numerous cells inthe growth zone which contained large fatdroplets, and it appeared certain that thefat cells had contributed materially to thiszone. From the reduction in size of the fatdroplets, it also appeared that many ofthese cells were losing their fat droplets.Growth continued until 42 days afterexplantation. At this time, one of thecultures had been lost through accident.The remaining cultures showed an averagewidth of 18.0 mm. The fringe of cellsshowed even less fat than earlier, because ofa reduction both in the number of cellscontaining fat droplets and in the size ofthe contained droplets.

The cultures were transplanted intofrom 5 to 8 flasks each, the explants being

selected so as to include none of theoriginal explant. All grew luxuriantly,and all except 1 were carried on for anumber of generations and then closed out.This 1 was transferred to 3 flasks 63 daysafter planting. In this third generationgrowth was excellent and even. Oneselected culture was transferred to 6 flasks43 days later. One of these was againtransferred to 10 flasks 39 days later, and ofthese 1 was transferred to 5 flasks 30 dayslater (May 22, 1941). Of these 5 cultures,2 were selected and built up into a serieswhich was used for the experimentsreported herein. All cultures used in thiswork originated from 1 of these 2 cultures.

The amount of fat droplets seen in thecultures had rapidly lessened until by thethird generation often whole cultures wereentirely free from all except the few verysmall fat droplets normal to growing fibro-blasts in this culture medium. When thecultures were carried in the same flask solong that the density of the cell populationcaused overcrowding, particularly amongcelas deep within the clot, some culturesshowed a rapid temporary formation oflarge fat droplets within the cells, but thisformation subsided at once when thecultures were transferred to fresh flasks. Inthe later history of the strain, large fatdroplets showed up sporadically underthese conditions, but the number of celasshowing such droplets was small in relationto the number of cells in the culture. Thedroplets disappeared as soon as the culturewas shifted into a fresh flask with freshculture medium.

The structure of these cultures wasnormal for cultures of fibroblasts previ-ously described (2), the architecture wasloose, and the cells were of characteristicshape, adhering to each other usually byterminal processes. Figures 8, B, 9, B,and 11, B, show typical culture structure.The refractive index of the cells was low,


rbana-Cham

paign on Novem


Dow

nloaded from



177

and particularly in very dense culturesthe ceils were often hard to see. Therewas no clear relationship between relativegrowth on the surface of the clot closestto the fluid and that nearest the glass inthe early life of the cultures. At bothsurfaces the cells seemed to be rathersimilar, although at each they were usuallymore flattened than were those within thebody of the fibrin clot.

The rate of growth of the cells wascomparable with that of earlier strains ofnormal fibroblasts carried in this culturefluid. There was variation from flask toflask, and growth during some generationswas better than during others. Frequentlythese variations could be traced to minordifferences in handling; too frequentshifting of the cultures to a new generationresulted in smaller explants and in lowcell density within the new explants, andthese poorer explants in turn gave rise toeven poorer growth. Similarly the hold-ing of cultures too long before transferresulted in diffuse central necrosis andpoorer growth when the cultures weretransferred. The best interval for transferwith the range of explants used wasprobably about 35 to 40 days.

SEQUENCE OF CHANGES IN FRESHLY TRANS-

PLANTED NORMAL CULTURES

During any one generation the sequenceof changes in these normal cultures wasapproximately as follows:

For about 8 to 12 hours after explanta-tion there was no appreciable sign ofmigration of the cells. At 24 hours therewere often a few cells or cell processes out,which lay at all levels within the thicknessof the clot and often resembled spikes.This fringe got wider and denser; at severaldays there was a luxuriant growth, thecells of which usually showed a definitetendency to concentrate along the inter-faces of the clot, that is, in the layer of clot

closely adherent to the glass and the onenext to the overlying fluid. These layerswill be referred to later as glass and fluidinterface layers, respectively. The cells atthe edge of the migrated zone were typi-cally fibroblastlike, with characteristicspindle, flattened-spindle, or less frequentlytriangular shapes, and with frequent, long,terminal, threadlike processes. At thisstage the culture was often troublesome tophotograph with higher power lenses forseveral reasons: (1) There were irregulari-ties in the thickness of the clot very closeto the explant; (2) in reaching the glassand fluid interfaces, the cells within theclot sometimes lay along a slope so thatwhen examined with higher lenses thewhole length of the cell could not befocused at once; and (3) the layer of cellsat this stage was frequently very thick,often with a lack of areas of cell densitysuitable for photographic records. Norwas the comparison of the increase in sizeof the clump with that of carcinogen-treated cultures as satisfactory during thefirst few days as it was a little later, sincethe experimental and control cultures hadnot grown sufficiently long to indicateclearly divergent rates of growth.

As the growth of the culture continued,there was a stage at about 8 to 10 dayswhich was probably the most useful forcomparing different cultures. The cellswere not latera]ly coherent as the growthof the cultures was typically loose, mitoseswere frequent, and there was no sign ofdegeneration. At this time no regularitywas observed, whether the glass or fluidlayer was the larger, although as a rulethe edges of the fluid layer were somewhatlooser. The edge of the normal culturewas usual]y not too loose to permit theestimating of the culture diameter. At thisstage also, experimental cultures had hadtime for their width curves to develop anydivergences.


rbana-Cham

paign on Novem


Dow

nloaded from



In the later growth of the culture, a num-ber of complications interfered with thedetermination of the rate of increase of thesize of the clump. Measurements of thewidth of the normal culture were oftencomplicated by the extreme looseness ofthe sheet of cells. In some instances, therewas no real edge to the culture, loose cellsreaching the edge of the flask at 18 days ofculture growth. This condition was worseat the fluid interface where sometimes thelack of a definite culture edge was causedby the fact that the cells had been washedloose from the culture and had reanchoredthemselves in the more peripheral parts ofthe clot. Often, too, little clumps of cellsthat had broken loose from the culture atits planting proliferated and at this stagereached such a size as to interfere with de-terminations of the size of the central clump.Consequently, width curves were irregularin this later period. Frequently the cellsheet at the fluid interface had grown solarge and dense that study of the cells atthe glass interface was obscured. With avery dense and rapidly growing clump, the

study was further confused by traces ofdiffuse necrosis in the central area of theglass layer.

Once the loose cells reached the edge ofthe flask, growth could obviously not in-crease the diameter of the culture. In-stead, from about 21 to 30 days on, thegrowth of the culture was manifested asincreased cell density in the more pe-ripheral areas. This increase tended to fillin the body of the clot. The culture wasmore difficult to examine and to photo-graph. In its last stages diffuse necrosisbegan to appear, first along the centralarea of the glass interface. When a largestrip explant was used, the culture in thisinterval usually thickened to a heavy sheetof cells that reached the edge of the flask.With smaller explants, this migration tothe edge of the flask was slower.

The foregoing general description heldapproximately for normal stock culturesand for controls up to the time lateral co-hesion of the cells was observed. Owingto the various complications that occurduring the very early and the very late life

1940s 1941 '^ : n 1942 1943

SERIES 202 G Gu Dure o x5T11- waaa

F '—l-'I FcOirRôks FOR au srnaxs

DAYS IN METHVLCHOLANTHRENE E E

STR IN D IV4 D vF D'COxTP LS M iED Ox—YG

f 36TBa L L

32 K 1 I I F— K55

N IB4 1 I ITI I .lO 90B

t I

H `II II 1H

0 O

Nt II 1 II IN

FIGURE 7.—General relationship of the culture strains. The interval during which any strain of cellsreceived carcinogen is shown by the very heavy lines, while the time at which the strain was removedfrom the carcinogen is shown by the ending of the heavy line or by the vertical dotted lines whichindicate when the culture was lifted out as a separate strain. Approximate times of changes of thecultures into new flasks, with corresponding changes of the fibrin clot, are shown by the short uprightlines arranged along the horizontal lines. The letters denoting the different cell strains correspondto the strain labels in our records and are used to identify them.


rbana-Cham

paign on Novem


Dow

nloaded from



179

of the cultures, in these experiments anattempt was made to correlate visual, pho-tographic, cinematographic, and stained-slide records of the cells within the intervalfrom 8 to 15 days after the culture wasplanted, preferably 8 to 12 days after. Atthis time, there were usually satisfactoryareas of cells clearly defined along the glassand fluid interfaces, the sheets were nottoo dense for examination, and there wasno necrosis of the normai cultures. Fur-thermore, as will be seen later, this periodwas probably the most suitable one forstudy of carcinogen-treated cultures.

RELATIONSHIP OF THE EXPERIMENTAL

CELL STRAINS

The relationship of the different strainsof the carcinogen-treated cultures may beseen from figure 7. Until August 1942each strain of cells was carried with its owncontrol. All cultures were in the sameculture fluid medium; but since not all thecultures could be transferred to fresh flaskssimultaneously, different strains with theirindividual controls were transferred ondifferent days. In order to compare thedifferent strains more accurately, in Augustand early September 1942 the strains wereregrouped. Since all control cultures hada common origin and since recognizedmorphologic differences between themwere extremely slight, the controls of strainD were carried on as the control on allcultures and all other control strains wereclosed out. The carcinogen-treated cul-tures were then regrouped into several sets,so that the cultures of each set were trans-ferred on the same day with the same solu-tions and so that each set contained cul-tures from strain D control, and H, J, L,N, and 0 carcinogen-treated strains. Thisreorganization allowed close comparisonof the different strains under identicalconditions.

In the following description of the

changes in the cells of the different strains,the progress of the continuous widthcurves is given to the end of May 1943 andthe cell changes observed ii1 the living cul-tures to September 1942, whereas the cellchanges from October 1942 through May1943 are given in another section, in whichthe cells of the different strains are moreclosely compared.

COMPARISON OF CONTINUOUS WIDTH

CURVES FOR THE DIFFERENT STRAINS

THROUGH APRIL 1943, AND DESCRIPTION

OF CELL CHANGES OBSERVED IN THE

CULTURES THROUGH SEPTEMBER 1942

Strains N and 0 (Series 208)

Strain 0, the first of the experimentalstrains, was started August 5, 1941, afterthe cells had been carried 291 days invitro. Four cultures were selected, andfrom each, two strips were cut lengthwisealong the explants. Each was in turnhalved transversely. Two explants fromeach original culture were maintained ascontrols; 24 hours later the other twoexplants from each culture were started inthe methylcholanthrene solution. Two ofthese carcinogen-treated explants weredamaged with a hot spatula and were dis-carded shortly after planting. The strainhas been carried on by consecutive trans-fers up to the present time. The continu-ous width curve of this series is shown infigure 6, section O.

During the early days of this strain, therewas no evidence that the carcinogencaused any accelerated rate of increase ofwidth of the cultures. Six to eleven daysafter the addition of the carcinogen, thewidth curve of the treated cultures showeda definite depression relative to that of thecontrols. At 14 days, the depression wasmore severe; in the second generation itwas even more so.

A visual examination of the cells during


rbana-Cham

paign on Novem


Dow

nloaded from



the first few days after addition of thecarcinogen showed no clearly definedchange. There was no sign of lateral co-hesion or of ribbon or sheet formation. Ifthere was any degeneration in the cultures,it was so slight that its existence was un-certain. The only suggestion of an inju-rious influence was some slight lateralshrinkage of some of the cells.

At 45 to 48 days after addition of thecarcinogen and 11 to 14 days after transfer,the treated cells were less spindle-shaped,more sheetlike, with definite shortening oftheir slender terminal processes. Thecultures were more compact than nor-mally, and their edges appeared less reticularand loose and showed fewer loose cells.There was a definitely increased irregu-larity of cell size. Numerous mitoses wereseen in both experimental and controlcultures.

At 52 days, 17 days after transfer, thechanges described had progressed. Thecells were very granular; some cells wereup to eight times their normal size; numer-ous cells appeared in mitosis in bothexperimental and control cultures.

At 73 to 76 days, 9 to 12 days after trans-fer, the terminal processes of the cells wereshorter, more blunted, and showed definitelateral irregularities suggestive of amoe-boid rippling or frothing. The cells werestill spindle-shaped and adherent laterallyso that the edge of the culture appearedeven more compact and less reticular orlacelike in architecture (compare fig. 8,A and B). There were few, if any, loosecells. The glass interface layer was domi-nant, and there was little or no fluid layerat this stage.

At 97 days, 7 days after planting, theglass interface layer was extremely com-pact, with few or no loose cells at its edges.The fluid interface layer was less pro-nounced and less characteristically co-herent than was the glass layer. The

culture edge showed numerous short cellprocesses, and there were only occasionalepitheliallike lobes at the edges. The ter-minal processes of the cells were usuallyvery short, and the amoeboid structure ofthe lateral edge of these processes had ex-tended farther toward the mid region ofthe cell. The whole lateral edge of theprocesses often had an amoeboid appear-ance (cf. fig. 9, A and B), which was so pro-nounced that in some instances it was hardto delimit the exact edge of the cell. Thecells seemed more granular than the con-trols, and there was more local celldisintegration.

At 106 days after addition of the methyl-cholanthrene and 16 days after planting,the cultures were very dense, and there wasa large area of central necrosis. The cellswere very closely coherent into sheets overextensive areas, and the culture edge wasabrupt (figs. 10, A and B, and 11, A).The cells shown in figure 11, B, are thecontrol for those in figure 11, A.

In some instances the edge was made upof short projecting spikes of cells, in othersthe cells were flattened, often epitheliallike.The cell processes were extremely short,while the amoeboid rippling or bubblingalong the edges of the cells often extendedthe whole length of the cells. There wasextensive cytoplasmic granulation.

During the subsequent interval up to184 days after the first addition of carcino-gen, there was no feature of note except agradual progression of the processes al-ready described. Often in this period thecultures showed lobed edges typical ofepithelium in culture.

On February 6, 1942, after 184 days inthe carcinogen and 31 days after planting,the cultures had gone through six genera-tions of transplantation. At this time agroup of four cultures was removed fromthe carcinogen and set up as strain N(fig. 7). From this time on the cultures of


rbana-Cham

paign on Novem


Dow

nloaded from


CELL CHANGES OBSERVED 181

Fioui 8.—A, Culture from strain 0, 76 days after first adding carcinogen and 12 days after planting.The black circle is an optical imperfection; B, Normal control cells. Both A and B are cells at theglass interface. X 20.


rbana-Cham

paign on Novem


Dow

nloaded from



FicuRE 9.—A, Edge of culture of strain 0, 97 days after addition of carcinogen and 7 days after planting.Note shortening of cell processes and occurrence of lateral frothing or rippling; B, Normal controlcells. Both A and B are cells at the glass interface. X 200.


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 10, A and B.—Edge of culture of strain 0, 106 days after addition of carcinogen and 16 daysafter planting. Note extremely short cell processes and definitely amoeboid appearing edges. Figure9, B, can be taken as representative of the normai culture at this time. The dark areas of shadowin the print arise from droplets of fluid on the roof of the flask. X 200.


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 11.—A, Edge of culture of strain 0, 132 days after addition of carcinogen and 12 days afterplanting. Note epitheliallike cell shape, the suggestion of granularity, the less clearly defined cellaxis, and the shortened processes; B, Control cells. Both A and B are cells at the glass interface.Oblique lighting. X 200.


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 12.—An extremely small cell clump which arose from a fragment in a flask with a larger sizeclump of strain N, 57 days after removal from carcinogen and 18 days after planting. The clumpis situated near and at the glass interface. One end of the clump is at a lower level and is out of focus.Note the close cell cohesion and the irregularity of the cell surface, also the complete absence ofnecrosis at a time when the central area of the larger clump showed extensive necrosis. X 200.

strain N were carried without further addi-tion of carcinogen. Through the end ofMay 1943, the strain had been carried atotal of 663 days since the original additionof carcinogen and 479 days since removalfrom it.

During the four transfers following theremoval of this strain from the carcinogen,the continuous width curve (fig. 6) held tothe same even, depressed level it had shownwhile the cells were in the carcinogen.With the omission of the first reading in thecurve following the original addition ofcarcinogen, at which time the curve hadnot reached this low level, the averagevalue for the curve through the fourth gen-eration following removal from carcinogenwas 4.6 mm. Beginning with the fifthgeneration after removal from carcinogen

and including the 11 transfer generationsthrough May 1943, the curve showed muchgreater fluctuation, while the average levelof the curve for these 11 generations roseto 6.0 mm.

No great change in the appearance ofthe cells was noted in the period imme-diately following removal from the carcino-gen. There was no evidence suggesting afurther progression of the cell changesnoted. The close coherente of the cells andthe altered character of the cell surface atthis stage are particularly well shown infigure 12, 57 days after removal frommethylcholanthrene and 18 days afterplanting.

The strain of cells continued in the car-cinogen after 184 days (strain 0) was car-ried on uninterruptedly until September


rbana-Cham

paign on Novem


Dow

nloaded from



16, 1942, or 406 days after the first addi-tion of carcinogen. Following the initialdepression in the first generation afteraddition of carcinogen, the continuouswidth curve of this strain held to a relative-ly constant depressed level through theeleventh transfer generation (fig. 6, sect. 0).The average level of the curve during thistime was 5.0 mm. Beginning with thetwelfth transfer generation, there was a fur-ther depression of the curve which held toa relatively constant lower level for twogenerations.

At the end of that time so much difficultywas being experienced in transferring thestrain (the average fringe of migrated cellswas only 2.7 mm. wide at the end of 16days, and even the more tentral cells ofthis fringe were already necrotic) thatthere seemed to be no doubt but that con-tinuation of the cultures in the carcinogenwould result in loss of the strain within thenext two or three transfers. The carcino-gen was therefore discontinued September16, 1942, 406 days after it was first addedto the cultures.

For the next three generations followingremoval of the culture from the carcinogen,the continuous width curve held to thesame low level, the average level for thesethree generations being 3.0 mm. Begin-ning with the fourth generation alterremcval of the culture from the carcinogen,the curve showed increased fluctuation,while the average height of the curveduring the next five generations (ending inMay) was 5.0 mm.

The progression of cell changes in strain0 during the period 184 through 406days was a continuation of those shownBarlier (fig. 13, A and B, and fig. 14).The adhesion became increasingly severe,with a further increase in production ofdense sheets and masses of cells, the sheetsat their edges often showing lobe-shapedprojections similar to these seen in cultures

of epithelial ceils. The celas were short,often more or less rounded, and adheredlaterally to those around them. Cellstructure in such massive sheets was veryhard to study; but where more isolatedcells were found on the glans interface, thecells no longer showed the characteristicspindle shape of the fibroblast. The indi-vidual cells were usually irregular in shape,and there were often ahort lateral proc-esses. The edges were often frothy andrufed along the whole length of the cell(fig. 14). During the same period therewas a further increase in the granulationof the cytoplasm so that the cells wereextremely grandular. No well-defined al-terations in the nuclei were observed in theliving cells, although this point has not asyet been critically studied.

Series 202

Series 202 was planned to obtain moredetailed data on the earlier stages of theaction of methylcholanthrene, particularlythe influence of graded doses. StrainsD, E, F, and G were, therefore, startedNovember 20 and 27 and December 4 and11, 1941, respectively, while the carcinogenwas added December 5, 19, 12, and 12.The addition of carcinogen to the serieswas so planned that in each of the strainsit was made at a different time afterplanting, at 15, 22, 8, and 1 day, respec-tively.

Because of circumstances that had noconnection with the behavior of the cul-tures, it was necessary to discontinuestrains E, F, and G at 29, 36, and 75 daysafter the original addition of the carcino-gen. For that reason, these strains con-tribute little to the data. The continuouswidth curve of strain G is shown in figure6, section G. The early cellular descrip-tion of these strains is similar to that ofstrain 0.

Strain D, however, was continued and


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 1 3.—Edge of culture of strain 0, 403 days after addition of carcinogen and 11 days after planting.A, Cells of glass interface. Note the massive sheet formation and the very limited width of the zoneof growth; B, Cells of the Huid interface. Note the relatively loose structure of the culture and thegreater width of the zone of migrated cells Both A and B. X 40.


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 14.—Edge of culture strain 0, 403 days after addition of carcinogen and 11 days after planting.Cells of glass interface, just before removal from carcinogen and at a time when growth was so slowthat it seemed the series would be lost. X 200.

was studied in greater detail. The cul-tures of this strain were removed from thecarcinogen at different intervals and werecarried along with their respective con-trols. Strain H was removed at 6 days,strain I at 13, strain J at 32, strain K at 55,and strain L at 111 days after the firstaddition of carcinogen. With the re-moval of strain L from carcinogen, strainD experimental cultures were closed out,while the controls of this strain were rununder their original label (D) as the con-trol cultures for strain L.

Strain D.—The continuous width curves of thisstrain are shown in figure 6, section D. To getan uninterrupted record, these curves have alsobeen included in the early part of the curves ofstrains H, I, J, K, and L.

The continuous width curve of strain D wasdepressed sharply in the second generation at 19days after addition of the carcinogen. It is mostlikely that this depression could have beendetected even Barlier if the time of adding thecarcinogen to the cultures had been different in

relation to their last planting. In the four subse-quent generations the strain was carried, thedepression did not increase but growth was heldto a level uniformly much lower than that of thecontrol cultures. The average level of the curvefor the second through the sixth generation was6.3 mm.

In a group of four photographs each of controlsand treated cells at 25 days after the first additionof carcinogen, one treated culture showed notice-able cell cohesion. There were very few loosecells. The cells were more sheetlike, with shorterand more amoeboid appearing terminal processesthan the controls. In instances the cells adheredterminally to form strands and short ribbons ofrelatively uniform diameter. These formed loopsand twisted into arches, occasionally joining otherloops laterally. At 47 days, five of nine culturesphotographed showed this cohesion. The cells atthis time showed definite lateral amoeboid outlinesof their terminal processes. The cohesion pro-gressed at what appeared to be a rate approxi-mately comparable with that in series 208, untilthe strain was closed at 194 days.

Strain H.—Strain H was removed from strain Dafter 6 days in the carcinogen, and its control wasremoved from the control of strain D at the same


rbana-Cham

paign on Novem


Dow

nloaded from



189

time. When examined at this time and shortlythereafter, the carcinogen-treated cells lookedentirely normal. There was certainly no sign ofchange of culture architecture, cell shape, or anyincrease in cell cohesion. In spite of this, the con-tinuous width curves (fig. 6, sect. H) showed asharp drop in the generation directly following theremoval of the cells from the carcinogen. In thenext generation, however, growth was practicallyequal to that of the controls; and in the generationfollowing, the average width and continuous widthcurves were slightly higher than those of thecontrols. During the last part of the secondgeneration after removal from the carcinogen,that is, 64 days after removal, only one culture ofeight examined showed recognizable cohesion ofthe cells.

In the fourth generation following removal fromthe carcinogen, the continuous width curve of thisstrain showed a sharp drop, and at this same timenearly all the cultures of the strain showed definitecohesion of the cells. This general cohesion was at131 days after the carcinogen was originally addedand 125 days after it was discontinued. The edgesof the cultures were abrupt, and there were fewloose cells. The cells were spindle-shaped,shorter, more sheetlike than normal cells, and hadshort amoeboid-appearing terminal processes.Short cell strands arborescing and re-fusing wereobserved. There was an increased amoeboidappearance of the lateral edges of the cell, particu-larly along the terminal edges of the cell processes.

In the interval from the drop of the continuouswidth curve in the fourth generation to October 1,1942, the series continued to show a consistent butlimited cohesion of the cells with consequentalteration of the architecture of the clump. Thiscohesion never reached as severe a degree as thatshown in either strain N or strain O.

While the continuous width curve of this strainhas fluctuated probably more than any otherexperimental curve, subsequent to the depressionin the fourth generation at the time of cellularcohesion and through May 1943, it showed nodefinite, recognizable trend of change. Theaverage level of the curve after the removal of thecells from carcinogen and through May 1943 was7.6 mm.

Strain I.—Strain I was removed from strain Dafter 13 days in the carcinogen, at which time acontrol was started from the control of strain D.As may be seen from the continuous width curve(fig. 6, sect. I), in the first generation after theremoval from the carcinogen there was no de-pression noted relative to the controls. Since,

however, this determination was made on onlythree flasks for the control and two for the car-cinogen-treated cultures, it cannot be reliedupon. Only in the third generation were thereenough cultures to give a reliable average. Atthat time the continuous width curve was some-what lower than that of its controls and continuedso until the series was closed out July 28, 1942,235 days after the original addition of carcinogen.The average level of the curve from the secondgeneration on was 7.3 mm.

The typical cohesion noted for the other seriesappeared in strain I sometime during the secondgeneration after removal from the carcinogen.An exact record of its first recognized occurrencewas not obtained. The degree of cell cohesionseen in strain I was in general comparable withthat observed in strain H and was less than thatobserved in strain J. A detailed description,however, was not recorded.

Strain ,7.—Strain J was removed from strain Dafter 32 days in the carcinogen. The controls ofthis series were removed from the strain D con-trols at the same time. Even before these cultureswere removed from strain D, the continuouswidth curve of the Jatter had shown a noticeabledepression relative to its controls (fig. 6, sect. J).This depression continued uninterruptedly andat relatively constant level for 12 transfer gener-ations following removal from the carcinogen.The average level during this time was 5.3 mm.Beginning about February 1943, however, thelevel of the curve became more erratic. Theaverage level during this last period, throughMay 1943, was 7.4 mm.

The first record of a recognizable alteration inthe cells of this strain was at 45 days after the firstaddition of carcinogen. At this time, of 10cultures examined 4 showed definite cell cohesion;shortly thereafter all showed cohesion. Cohesionappeared earlier and was more severe than instrains H or I but was less severe than that instrains N or 0, and continued to increase forpossibly 60 days after removal of the celas fromcarcinogen, after which further change was lessnoticeable.

Strain K.—Strain K was removed from strain Dafter 55 days in carcinogen and after strain D hadalready shown a severe depression of the con-tinuous width curve (fig. 6, sect. K). This depres-sion continued at a relatively constant level aslong as the strain was carried, that is, for fivegenerations following removal from carcinogen.The average level of the curve from the secondgeneration on was 6.6 mm. At 64 days after first


rbana-Cham

paign on Novem


Dow

nloaded from



addition of the carcinogen, one flask of eightexamined showed the cells of the culture to betypically coherent. This cohesion became in-creasingly severe, until at 139 days nearly allflasks showed typical cohesion and the formationof short strands, ribbons, and loose, fenestratedsheets. Later changes in the cells were progres-sive for a short interval after their removal fromthe carcinogen, but after this they showed nofurther recognizable progressive changes, nor didthe cells show any recognizable loss of the changesalready induced. The strain was closed at 237days after the first addition of carcinogen. Thecontrols of this strain appeared entirely normalas long as they were carried.

Strain L.—Strain L was removed from strain D,together with its controls, 111 days after the firstaddition of carcinogen. At this time its continu-ous width curve was showing the usual depression(fig.6, sect. L). The depression was much moresevere than that for either strain H or strain I andwas comparable with that shown by strains J andN. The curve continued at this low level throughthe ninth generation after removal from the car-cinogen, the average level being 6.2 mm. Duringthe next five generations (through May 1943) thecurve showed erratic fluctuations, while the aver-age level during this period was 9.5 mm.

The progression of cell changes observed in thecultures was similar to that described for the cellsof strain N at comparable ages, up to the time ofremoval from the carcinogen, while after that timethe cell changes were similar to but less severethan those of strain N.

Control Cultures

During the early part of the experiment,that is, until May 1942, the control cul-tures had been watched very closely andhad appeared entirely normal. During theinterval from May to July tl-ey wereexamined regularly in determining thewidth curves, but the cellular structurewas watched somewhat less closely, moreattention being given to increasing thenumber of cultures in the experimentalseries in order to study the behavior ofthe cells on reinjection into mice.

The first evidente of any abnormalityin the morphology of the control culturesdated from about August 5, 1942, atwhich time three cultures in the strain H

controls had definitely coherent architec-ture, typical of the cultures treated withmethylcholanthrene. These three cultureswere discarded. The other controls werenorma] at this time. On June 18 a seriesof normal cultures was reinjected into C3Hmice, and some of them showed tumorsafter a latent period of about 60 days, thatis, about August 18. It was only when thetumors appeared that a re-examinationrevealed a general slight cohesive altera-tion in the architecture of the control cul-tures. These alterations were of the samenature as those observed earlier in methyl-cholanthrene-treated cultures.

From August 18 to September 9, 1942,the normal continuous width curve forthe control cultures remained at a rela-tively high level. About October 9, 1942,however, it showed a sharp depressionwhich became increasingly severe untilthe middle of December and reached alevel as low as that of strain H carcinogen-treated c el Is. The carcinogen-treatedstrains may also have shown some depres-sion, but this was not conclusive. A criti-cal viswal examination of all control cul-tures was made in the Jatter part of August1942. The findings may be summarizedas follows:

(1) Of the various strains of controlcultures, not one showed the normalloose structure that the controls hadshown up to May 1942. In each straina large fraction of the cultures showedrecognizable but slight cel] cohesion ofthe type which had been considered char-acteristic of the action of methylcholan-threne on these cells.

(2) The degree of alteration in some cul-tures was so slight that it could be con-sidered only as probable rather than asclearly recognizable. In other instancesthere was no question but that the cellswere cohering in the usual carcinogen-type architecture. In no instante was this


rbana-Cham

paign on Novem


Dow

nloaded from



191

cohesion nearly so severe as that seen inthe carcinogen-treated flasks of the Hstrain, which was the least altered of anyof the experimental strains.

(3) No unaltered strain or line of cul-tures could be segregated from the rest ofthe controls.

(4) Not all strains were equally altered.The controls on strain H were probablythe least altered of any, and they veere soslightly altered that many of them seemedpractically normal. In numerous in-stances, cultures which had arisen from thesame explant at the last transfer showeddistinctly different degrees of alteration.

CULTURES FROM SEPTEMBER 15, 1942, TOMAY 1, 1943

A repeated detailed examination wasmade of all living cultures in late Sep-tember and early October 1942, the cul-

tures being studieel at 2, 4, 11, and 16 daysafter planting. The conditions prevailingin cultures of the different strains are sum-marized herein, whereas the series of pho-tographs are from a similar study madeabout December 17, 1942.

The descriptions of these various strainsmay be considered as characteristic. Inall strains there was some overlap of thedifferent cultures of different strains. Forinstance, in strain D an occasional culturewas seen which was changed to a degreecharacteristic of strain H; strain Hshowed some cultures typical of those ofstrain D and some typical of strain J;similarly strain J had some characteristicsof strain H and of strain L. Probably theclosest overlap was with strains J and L,but even in these the average change forcultures of strain L was somewhat greaterthan that for strain J.

FIGURE 15.—Culture from strain D, photographed December 17, 1942, 9 days after planting. Noteslight shortening of the cell processes and slight increase in lateral cohesion, with loss of the loosereticular structure of the normal fibroblast cultures and with the formation of loose cell strands.Compare with figures 9, B, and 11, B. X 200.


rbana-Cham

paign on Novem


Dow

nloaded from



Strain D controls.—The architecture of the strainD controls was loose, although more coherent thanthat of the normal cultures at the beginning of thisseries of experiments in 1941. There were rela-tively few loose cells at the edges of the cultures,and there was a luxuriant growth at the glassinterface. Growth at the Huid interface was'irreg-ular but in some instances was even greater thanthat at the glass interface. The characteristicgrowth at the Jatter is shown in figure 15. Thecells were loosely coherent laterally and formed nodefinite ribbons but gave an architecture thatlacked the characteristic reticular structure of thenormal fibroblast cultures. The terminal proc-esses of the cells were very slightly shortened andwere amoeboid-appearing. There was usually noappreciable rippling or frothing of the lateraledges of the body of the cell, although the extremetips and lateral edges of the terminal processessometimes showed a slightly amoeboid pattern.The long axis of the cell was well-defined; onlyoccasionally was a cell seen which departed from amore or less definite spindle or related shape; cyto-plasmic granulation was not recognizably in-creased; nuclei were grossly normal; many

mitoses were seen. Cell size seemed grossly normalalthough this estimate was extremely rough.

Strain H.—The growth of the cultures wasluxuriant, practically indistinguishable from thatof strain D controls; the structure was somewhatless loose (fig. 16), and the cells had a greatertendency to adhere laterally, forming more of asheetlike structure in the culture. There were noloose cells at the edge of the culture, and the edgewas sharp. The cells in general showed a spindleshape, although their terminal processes wereshorter than those in strain D. The lateral edgesof the terminal processes were often amoeboidalong their whole- length, down to where theyjoined with the body of the cell. The cytoplasmicgranulation was not recognizably increased.Nuclei appeared grossly normal, many mitoseswere seen, and the cell size seemed roughlynormal.

Strain,7.—There was a noticeable average reduc-tion in the width of the growing zone of the culturesof strain J. The structure of the clump was alteredfrom that of either of the preceding strains; thecell layer at the glass interface was often made upof characteristically slender interconnected loop

FIGURE 16.—Culture from strain H, 371 days after removal from the carcinogen and 9 days after plant-ing. Note the progressive cohesion, shortening of cell processes, and amoeboid pattern of cell edgesas compared with figures 11, B, and 15. X 200.


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 17.—Culture from strain J, 345 days after removal from the carcinogen and 9 days after planting.Note the progressive alteration from strain H (fig. 15). The clump of cells at the upper left illustratesthe tendency of the cells at this stage to send off lateral processes, with the result that the cell axisis far less clear. X 200.

of cells. In other areas the cells were even moreclosely coherent laterally than in D or H. Suchcells were no Jonger spindle-shaped but weremore sheetlike, often with a number of lateralprocesses which gave them the appearance ofhaving a far less definite axis (figs. 17 and 18).Cell processes were shorter, and their lateral anddistal margins were frothy or irregular. Theypresented a pattern suggestive of amoeboid activ-ity. The free edges of the body of the cell werealso rippled or amoeboid in appearance. Thegrowth at the edge of the culture was very muchless clearly radial, the radial structure being dis-torted by the irregular loops and strands or ribbonsof cells, which often were arched, interlaced, andfused, enclosing islands of clot. The granulationof the cells was only slightly greater than normai.Nuclei looked normal, and a few mitoses wereobserved, possibly fewer than in strains D and H.

Strain L.—In strain L the cells were even moresheetlike and more closely laterally coherent,tending to form sheeted ribbons and solid sheets,with fewer fenestrations and irregularities (fig. 19).

At this stage they often resembled epithelial cells.The cell processes were even shorter than in thepreceding stage, and the design of all free edgesof the cell was frequently or usually suggestive ofamoeboid rippling. Cell size, granulation, andnuclei did not seem greatly altered.

Strain N.—This strain was progressively alteredfrom the preceding strains, and there was an evencloser cohesion of the cells. Many cultures showedribboning or combination of ribboning and sheetformation as the chief form of growth. Therewas a much greater tendency to sheet formation,and the sheets appeared thicker and denser (fig.20). The edge was more regular than in strainsJ and L, and there was a tendency to form epithe-Iiallike projecting lobes along the edges of thecultures. Cell processes were even shorter thanin the preceding strain, but the amoeboid orfrothy structure of the whole free edge of the cellwas more accentuated, often so much so that theexact edge of the cell could not be clearly defined.There was a recognizable increase in the cytoplas-mie granulation. Mitoses were fewer in number,


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 18.—Another culture from strain J, 345 days after removal from the carcinogen and 13 daysafter planting. This culture shows the tendency of the cells at this stage to send off lateral processes,with resultant obscuring of the cell axis. The photographic contrast is less as this was taken on asofter film. Approximately X 206.

and cells in mitosis were less rounded up and lessseparated from those around them.

Strain 0.—The growing fringe of the culture instrain 0 was even narrower than in strain N, andthe cells were more altered. Individual ribbonsof cells or ribbons closely united into sheetlikestructures were seen. The culture was often amass of cells so closely united that it appearedalmost as a syncytium (fig. 21), which terminatedin an amoeboid-appearing edge which in someinstances was lobose and epitheliallike, and inothers broken only by very short amoeboid-appearing processes. (In stained preparationsthere was no evidence of a true syncytium.)Mitoses were often most difficult to recognize, asthe cells showed little or no rounding up or separa-tion from the surrounding cells of the sheet. Onlyin areas of less cell density could they be recog-nized. A careful examination of such areas re-vealed almost no cells in mitosis. In areas wherethe individual cells could be observed, nuclei weregrossly normal. One, two, and three nucleoliwere visible in each nucleus, and they appeared

grossly normal. Nuclei were otherwise free fromgross granulation. The cytoplasm was extremelygranular, the granules being irregular in shapeand size and easily visible with a 16-mm. lens.They often gave the cytoplasm a peculiarly moth-eaten and characteristic appearance. These gran-ules were not to be confused witn the occasional,small, nighly refractile oil droplets seén in thecells, or with the slightly increased granularityseen in the early stages.

Examination of the cultures in Decembershowed no recognizable progression of thelater stages (L, N, 0), but at that timethere was thought to be possibly a slightprogressive alteration of cultures in theearlier stages. For instance, numerousstrain D cultures showed structure typicalof strain H and in one instance of strain J,while strain H culture often showed struc-ture typical of strain J.


rbana-Cham

paign on Novem


Dow

nloaded from


CELL CHANGES OBSERVED 1,95

FIGURE 19: Edge of culture from strain L, 266 days after removal from the carcinogen and 9 days afterplanting. Note progressive alteration from strain J, with formation of sheets and further shorteningof cell processes. X 200.

Examination of the series May 15, 1943,showed the following results for the glassinterface layers:

Strain D cultures were characterized by figure15, a photograph of cells of strain D taken De-cember 17, 1942. A few cultures were character-ized by figure 16 (strain H).

Strain H cultures were characterized by figure16, and a few by figure 15 (strain D control).None were observed showing the structure offigure 17 (strain J).

Strain J cultures showed a cell structure likethose in figures 17 and 18 (strain J), and a fewcultures were like figure 16 (strain H), while someshowed coherent sheets of cells although the cellsheets were not so closely coherent as in figure 19(strain L).

Most cultures of strain L presented a structurelike that in figure 19 (strain L), and some likethose in figures 17 and 18 (strain J). No culturesobserved resembled figure 18 (strain N).

The degree of cohesion in strain N was definitelygreater than in strain L. The cells were coherent,

548963-43-5

as in figure 20 (strain N), and often showed thesame type of terminal edge. Cells were definitelymore granular than the cells of strain L.

Strain 0 cultures were plainly more altered onan average than those of strain N. Culture edgessimilar to those in figures 20 and 21 were observed;the cells showed greater granulation than instrain N.

In all these cultures the fluid interfacecells were looser and showed less coherencethan did those at the glass interface. Instrain N, two cultures were seen in whichthe retraction of the terminal and lateralcell processes had gone so far that the fluidlayer cells had often lost all connectionwith each other and were separated androunded up.

DISCUSSION

The cultures described by Earle andVoegtlin (1) were transferred at various


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 20.—Culture from strain N, 314 days after removal from the carcinogen and 9 days after planting.Note the extreme lateral frothing and rippling of the cell edges and the close cohesion of the cellsinto a sheet with definitely lobar edges. X 200.

intervals up to about 18 days. It waslater feit that the frequency of transferhad interfered with the life of the cultures,and in the later series (2) length of thetransfer interval was increased to as muchas 112 days in some instances. Thisincrease allowed the cultures to survive;but later work with these series and withother tumor strains indicated that theinterval had been extended too long forthe best life of the carcinogen-treatedcells, while detail of the culture was ob-scured by the clouding of the fibrin clot.Accordingly, in the present studies theinterval between transfers was reducedto 28 to 36 days. This period gave muchpetter results than did either of the othertwo. About 30 days is probably theoptimal common period for the strainsof cells studied. Under the conditions

used, even this period must be con-sidered a compromise, since the normalfibroblast culture would probably givebetter results with transfer intervals of 36days; while, because of tentral necrosis,carcinogen-treated cultures of verg co-herent architecture probably give betterresults with somewhat more frequenttransfers.

In considering the results, the mast out-standing question ic: that concerned withthe behavior of the control cultures. Inthe earlier work (1, 2), great care was takenin the handling of the dibenzanthraceneand methy]cholanthrene used. In spite ofthe precautions taken, just at the last ofthe experiments the control culturesshowed a slight, though unquestionable,alteration in architecture. This altera-tion was similar to that observed in cultures


rbana-Cham

paign on Novem


Dow

nloaded from



FIGURE 21.—Edge of culture from strain 0, 92 days after removal from the carcinogen and 9 days afterplanting. Note the formation of a dense sheetlike edge often entirely devoid of cell. processes for

extensive lengths. X 200.

deliberately treated with methylcholan-threne, and the conclusion was reachedthat in spite of the care taken lome trace ofcontamination of the controls with methyl-cholanthrene had occurred. In planningthe present experiments an attempt wasmade to eliminate any possibility of con-tamination of the cultures with the car-cinogen. The alteration of the controlcultures must, therefore, be interpretedin the light of the care taken to avoid suchcontamination.

The possibility that the cells of the orig-inal parent strain used in this work werenot normal cells of the fibroblast type butwere malignant cells is most improbablebecause of the extreme rarity of subcutane-ous tumors in this strain of mouse at theage at which it was used (100 days). Italso seems improbable, since the appear-

ance of the control cultures, certainly upto May 1942 (559 days after original plant-ing), was entirely normal for fibroblastsfrom the subcutaneous tissue of mice andrats; and, when development of tumors inmice injected with control cultures led tocareful examination of the cultures inAugust, there was no difficulty in recog-nizing that mant' of these control cultureshad undergone a definite though slightchange in morphology since their lastdetailed examination. The remeinder ofthe control cultures showed suggestive butless conclusive morphologic changes, whichin al] instances became conclusive in thosecultures that were carried on.

This alteration of the cells was of muchless extent but was of the same morphologictype as that induced Barlier in the strains ofcultures treated with methylcholanthrene


rbana-Cham

paign on Novem


Dow

nloaded from



and of the same type as the alterations wehad formerly observed in tissues cultures oftumors arising in rats from the injection ofmethylcholanthrene. 2 It was also thesame as that which Jacoby (9) had de-scribed for cells of tumors induced in vivoby the action of 1, 2, 5, 6-dibenzanthra-cene.

From these data it appears probablethat the assumption of malignant proper-ties by the control cells was closely asso-ciated with the alteration in the morphologyof the cells which occurred about June1942.

In accepting these original parent straincultures as normal fibroblasts whichunderwent a morphologic and probablya malignant transformation about June1942, several possibilities must be con-sidered as to what induced this change.The possibility that the control cultureswere contaminated with altered fibro-blasts from the experimental cultures iseliminated - by the unvarying order inwhich the cultures were handled, andby the fact that when the alteration inthe controls was first observed the dis-tribution of the altered cultures throughthe series was such that it could havebeen produced only by general con-tamination of all sets of the controls.This hazard seems entirely eliminatedin view of the technical care employed.When the change in the controls wasfirst observed, there was no sign of foci ofgreatly altered cells, distributed in rela-tively unaltered cultures. The appear-ance was rather that of a very slight,relatively uniform alteration of the cells,most easily recognizable at the glass inter-face of the culture. The possibility ofcontamination from any tumor culturesmay similarly be dismissed since duringthe whole course of this experiment up to

2 Unpublished work.

October 1942 no tumor cultures werebeing carried.

The possibility must be consideredwhether apparently normal fibroblasts ofthis strain of mice, when grown in aculture medium of chicken plasma,horse serum, and chick-embryo extract,can spontaneously go over into the alteredform finally seen in the control culturesand give rise to sarcomas on injection . 3

3 Personal communication from Dr. G. 0. Gey, out-lining some results previously presented bui never pub-lished.

Gey, using a culture technique and culture medium ofwhich the details are not at present available, describesalterations in several strains of cells, all of which aroseoriginally from mesenchyme, of subcutaneous areolartissue origin, from a young adult rat of pure strainPhiladelphia albino stock. This stock of rat, accordingto Gey, rarely develops spontaneous tumors.

Three cell strains described showed spontaneous cellalterations; and when cultures of these altered cells wereinjected, they gave rise to tumors. In the first instantethe alteration in the cells was first detected 4 monthsafter primary explantation and C° consisted of a change tocells varying greatly in site and with a great number ofcells showing a typical and unequal mitosis. inoculationof these greatly altered cells produced a variety of tumorsdifferent in cytological and cultural characteristics fromtheir cells of origin and from each other. Transplanta-tion from tissue culture to host and back again to tissueculture showed that each of the transplantable tumorswas of a fixed type. Prolonged cultivation for longperiods and further animal transfer revealed that theystayedfixed in type."

The t4ird conversion occurred in a strain oj cells relatedto the other two in that it came from the same small piecesof normal tissue. This strain had been onder cultivationfor several years without showing any changes of fixedtype. When the change occurred, it showed up asmicroscopic areas of transformed cells within the thinlyspread out periphery of only a few of the cultures. Thetransformation then spread through all the cells of thecultures during a period of several transfers and sub-cultures. There was thus produced a strain of malignantcells which on inoculation produced tumors of similarcell type. Other normal-appearing cultures of the samestrain did not show any changes nor did they producetumors on inoculation.

Gey also records that other normai cell strains of ratmesenchyme derived from another rat of the same stockwere maintained for over 4 years without showing anychange in type, and that to date he has not been able toreproduce these teil conversions by any deliberate experi-mental alteration of the cell culture environment. Hestates:

"The period of cultivation, whether months oryears, isapparently unimportant. We have so far been unableto isolate any virus from these new malignant cell strains,Cell-free filtrates do not produce tumors on inoculationsnor does the filtrate alter normal strains maintained incontinuous culture. We were at one time inclined tobelieve that stray gamma radiation had some effect.


rbana-Cham

paign on Novem


Dow

nloaded from



199

We know that the conditions in the tissuecultures used were different from thosein vivo. The culture media and otherculture conditions were such that thefibroblast continued to proliferate at arate far greater than that in the adultanimal. That this alone induced such achange in the cells seems unlikely in viewof the extended work of Carrel and hisassociates (10, 11, 7), who carried fibro-blasts from the embryo chick heart incultures, in which they showed a rapidrate of growth for more than 25 years,apparently without any change in mor-phology. If these cells had undergoneany such drastic morphologic changesas that seen in strains J, L, N, or 0, theycould hardly have been overlooked. Itis emphasized, however, that whereas theCarrel strain of fibroblasts was from theheart of an embryo chick and was carriedin solutions obtained from chicken bloodand chicken eggs, in the present series ofstudies the celas were from subcutaneoustissue of a mouse and were carried duringtheir whole period of culture in an entirelyheterologous culture fluid. Obviously, ina number of respects the two sets of cultureconditions were dissimilar, and the be-havior of cells under the one might nothold for the other.

With the limited data available, no con-clusion can be reached whether or not nor-mal mouse subcutaneous fibroblasts bysimple culturing in the heterologous horseserum-chick embryo extract can be in-

Our observations to date on continuous weak irradiationdo not confirm such a conclusion. Otherfactors are beinginvestigated which may have some bearing on theseunusual normal to tumor cell conversions which oc-curred in our cultures."

While Gey described a malignant alteration in severalstrains of mesenchymal cells, all from one rat, no con-clusions are presented as to what induced such changes,nor do the data available indicate whether these trans-formations could have arisen from some trace contamina-tion of the cultures with a recognized carcinogen. Atten-tion is called to this work; but until more data are avail-able, no attempt will be made to correlate it in detailwith the resuits reported in the present paper.

duced to change over into sarcoma cells.This question is fundamental, and inanswering it much important informationconcerning the mechanism of carcino-genesis may be obtained. Obviously,experiments designed to settle the issueshould be carried out as soon as practi-cable.

The second possibility is that in spite ofthe precautions taken, some active amountof methylcholanthrene got into the controlcultures and effected the transformation.The change noted in the control cells wasmore limited in degree but was of the sametype as that induced through the action ofmethylcholanthrene. This fact, however,cannot be considered as evidence that thetumors arose from the action of methylchol-anthrene, since Jacoby (9) noted a similarcoherent architecture in cultures of malig-nant cells from a tumor induced in vivoby means of an entirely different carcino-gen, 1, 2, 5, 6-dibenzanthracene, and thepresumption is that a similar change couldbe effected with other carcinogens.

The evaluation of the possibility that thecontrols were contaminated is complicatedby the fact that as yet no data are availableon how little methylcholanthrene is neces-sary to change a strain of normal fibro-blasts in tissue culture into one with suchaltered morphology and physiology. Bry-an and Shimkin (12) have estimated thatinjection in vivo of about 2.4-y of methyl-cholanthrene is necessary to produce asarcoma in strain C3H mice, in only about1 percent of the injections, whereas 4.5yproduces tumors in about 5 percent of theinjections. But conditions in vivo are dif-ferent from those in vitro. It seems likelythat a clump of fibroblasts in vitro is farmore reactive or is under conditions farmore apt to cause reaction with the car-cinogen than are fibroblasts within themouse. Hollaender, Cole, and Brackett(13) showed a definite photosensitizing ac-


rbana-Cham

paign on Novem


Dow

nloaded from



tion of methylcholanthrene on yeast in theconcentration of 10 -9 . This reaction oc-curred within a few hours after additionof the carcinogen to the cells. In the ab-sence of any data on the amount of methyl-cholanthrene necessary to alter a clumpof mouse fibroblasts under the culture con-ditions in the present experiments, the pos-sibility exists that when allowed to act onthe cells for such an extended interval as ayear or more, a concentration comparablewith or even lower than that which theseauthors found active for yeast may be ade-quate. It is obvious that if such minutetraces are active, the problem of treatingstrains of cells with the carcinogen for longintervals and of keeping other controlstrains entirely normal becomes most difpi-cult.

While careful reconsideration of the tech-nique used has Biven no definite reason tosuspect any step in the handling of the cul-tures, the solutions, or the soiled glassware,less positive assurance is feit with referenceto the re-use of rubber stoppers whichwere possibly contaminated with traces ofmethylcholanthrene. Although stoppersfrom carcinogen-treated cultures werenever used on control cultures, therecleaning, sorting, wrapping, sterilizing,and other handling of these rubber stoppersmay have introduced a trace-contamina-tion hazard.

The laboratories in which this work wascarried out are in the same building withother laboratories and animal rooms wherelarge amounts of methylcholanthrene andsimilar carcinogens were handled. Thedesign of the air conditioning and heatingsystem of the building is such that therehas been some trace recirculation of usedair from many rooms, in some of which thecarcinogen was certainly handled. Be-cause of insufficient knowledge of theactivity of low concentrations of methyl-cholanthrene and other carcinogens in

tissue culture, the hazard of such tracecontamination from this source also can-not be evaluated at present, although it isfeit that its possible significance must notbe overlooked.

The evaluation of another possibility,that the changes were induced by someunrecognized agent of unknown source,must await further work.

As the most probable explanation of thealterations in the control cultures and as aworking basis for further study, the hypoth-esis is advanced that a trace contamina-tion of methylcholanthrene occurred inspite of all precautions taken. If this isthe correct explanation, the amount ofcarcinogen needed to effect such a changein the cultures is probably extremelysmall. It is suggested that, until moredata are available, the most scrupulouscare should be taken in all laboratoriesdoing work with long-term tissue culturesso that such trace contamination of thecultures with methylcholanthrene or simi-lar carcinogens does not lead to complica-tions difficult to eliminate, and to erro-neous conclusions.

With reference to the other observa-tions reported in this paper, it seems clearthat the changes in the control culturesarose at a substantially later date thanthe changes shown in cultures of evenstrains H and I, and more than 8 monthsafter recognizable morphologic changeswere observed in strains N and 0. Fur-thermore, while of the same type, thechanges in the controls have at all timesbeen less accentuated than those seeneven in strain H cultures. Regardless ofwhether the changes finally noted in thecontrols were spontaneous or arose fromsome contamination of the controls by thecarcinogen, there seems no doubt that thecell changes observed in the experimentalcultures are directly correlated with theaction of the carcinogen that was delib-


rbana-Cham

paign on Novem


Dow

nloaded from



201

erately added. This correlation is alsoshown by the orderly nature of the changesreported for different intentional expo-sures of the cultures to the carcinogen.

The retardation of growth and thedegeneration of rat and mouse fibroblastsunder the action of methylcholanthrene invarious concentrations from (roughly)0.2y to 1007 have already been describedby Earle and Voegtlin (1, 2). Thesepapers should be consulted for Barlierreferences to the literature. Lebensonand Magat (14) carried out a furtherextensive series of studies on this point.They cultivated the skin of mouse embryosin chicken plasma and mouse and chick-embryo extract, in hanging-drop cultures,later shifting the skin to Carrel flasks. Tothese cultures was added 1, 2, 5, 6-dibenz-anthracene, which was dissolved byletting the fluid plasma stand 2 days, withoccasional shaking, in contact with anexcess of the carcinogen. The cultureswere left in contact with the carcinogenfor intervals of 3 days and were thenretransferred to normal medium and car-ried on. Some of the cultures were passedthrough dibenzanthracene once (from thefourth to the sixth day), whereas somereceived an additional exposure from theeighteenth to the twenty-first day. Cul-tures were apparently carried as long as54 days.

During the first 3 weeks of growth nomanifestation of the action of the carcino-gen was noted. Most of the cells of thecultures were fibroblasts with a littleepithelium. About the fourth week, cul-ture growth improved considerably, andthere was an intense liquefaction of theplasma in most of the cup, xres. Explan-tation of fragments of muscle onto the cul-tures gave rapid invasion of the muscle bythe culture cells, destruction of the muscle,and a great acceleration of culture growth,

the growth zone reaching 20 to 40 mm. 2

in 3 days. There was also noted a greatlyincreased ability of very small clumps ofcells to live and to grow rapidly.

Cultures were inoculated into micefrom the twenty-second or the thirty-eighth to the fiftieth day after addition ofthe carcinogen. No mice showed tumors.

Lebenson and Magat concluded thatwhile there was no clear demonstrationof production of malignancy in their exper-iments, it was evident that the treated cellshad acquired certain new properties underthe influence of the carcinogen, and thatthe properties had appeared after a con-siderable latent period following treatmentwith the carcinogen. Furthermore, theyconsidered that the induced changes werein a direction corresponding to an ap-proach to the blastoma cell.

Larinov, Chertkova, and Samokhvalova(15), in an extensive series of studies, grewfibroblasts from the skeletal (femoral) mus-cles of new-bomn mice, in Carrel flasks, ina medium of chicken and rabbit plasmaand dilute chick-embryo extract. Tothis medium benzpyrene or dibenzanthra-cene was added in the form of a fine sus-pension. Benzpyrene in a concentrationof 107 to 507 per cubic centimeter gave aconsiderable inhibition of the growth of thefibroblasts, while dibenzanthracene dis-played no such toxic action. For Jongerterm tissue cultures, this carcinogen wasused in a concentration of several gammaper cubic centimeter of culture fluid.

Cultures were carried up to 6% months.There was nothing particularly distinc-tive in those exposed to the dibenzanthra-cene 17 days. When they were carried onwithout further addition of carcinogeu,some of them began to _grow far morerapidly at 40 days. Benzpyrene had notoxic action on them. In the following 3%months the rapid growth continued, and


rbana-Cham

paign on Novem


Dow

nloaded from



during this time many cultures were im-planted in young mice. Of 55 mice inocu-lated, none developed a tumor.

In these treated cultures, frequently atthe periphery of the culture zone, a culturewould arise from a very small, isolated,clump of daughter cells. These daughterclumps grew very rapidly, and their cellsshowed a greater mobility, sometimes alarger size, and more disordered celiarrangement. In all cases the cells ofthese daughter cultures showed muchmore fat than did the parent cultures.

From our own experience with small cellclumps in culture, it is difficult to evaluatethe significante of some of these findings ofLarinov, Chertkova, and Samokhvalova.The increased storage of fat droplets withinthe cells is often seen in small clumps ofnormal fibroblasts (16), and in the presentauthor's experience the cells of theseclumps are often markedly less orderly inarrangement than are those in large clumps.The acceleration of growth of small clumpsof cells treated with carcinogen agrees withthe observations of Lebenson and Magat;and in the series of studies reported herein,the same phenomenon was frequentlyobserved, at a stage of carcinogen treat-ment after definite cohesion of the cellshad appeared. For instante, the cellclump shown in figure 12 continued toproliferate rapidly, while substantiallylarger clumps of control cells died.

The observations of Lebenson and Magat(14) and of Larinov, Chertkova, and Samo-khvalova (15) and the observations re-ported in the present paper are in entireagreement that the cel] transformationsappeared after a definite latent interval ofseveral weeks following initial exposure tothe carcinogen.

Benevolenskaya (17) used cultures ofembryonic chick heart and mesenchymegrown in hanging-drop cultures and inCarre] D3.5 flasks in a mixture of chicken

plasma and chick-embryo extract, towhich methylcholanthrene or 3, 4-benz-p yrene was added, usually as a suspension.Concentrations of carcinogen ranged from50-y to 17 per cubic centimeter of culturemedium, and cultures were carried in thecarcinogen to 88 days. Frequently thecultures were transferred to normal culturemedia and carried for a time after thisexposure to carcinogen.

Both carcinogene were toxic and inducedstoppage of growth after 3 to 7 days in aconcentration of 507 per cubic centimeter.When treated cells were transplanted intoa normal medium after such treatment,degeneration and death still ensued. Aslower concentrations were tried the toxicaction was less marked, until at 2y percubic centimeter the cultures were keptliving in methylcholanthrene for 85 days,after which they were carried in normalmedia 65 days. Such concentrationsproved slowly toxic. No changes in thecells were observed to suggest that theyhad undergone a malignant alteration; thechanges were those of toxicity and degen-eration.

Eight fowls were each injected with from6 to 12 cultures that had been carried in aconcentration of 2y of methylcholanthreneper cubic centimeter from 37 to 62 days,and then carried without carcinogen 11 to42 days, while 6 were each injected with 8to 9 cultures that had been subjected to aconcentration of 17 of methylcholanthreneper cubic centimeter from 22 to 30 daysand then carried 17 to 18 days in normalmedia. Nine chickens were injected withcultures that had been subjected to 2y or37 of benzpyrene per cubic centimeter from12 to 19 days, then carried from 6 to 37days in normal culture medium, and in-jected. All injections were uniformlynegative.

Cooper and Reller (18) treated the earsof mice twice weekly with a 0.6-percent


rbana-Cham

paign on Novem


Dow

nloaded from



203

solution of methylcholanthrene in benzene.They found an increase in the frequency ofmitosis of the epithelial cells from a normalaverage of 0.11 percent to 1.0 percent at23 to 37 days after first treatment. Thiscount subsided to 0.49 at 65 days and roseto 1.5 percent at 93 days, at which timethe experiment was terminated. From 16days on the ears became definitely hyper-plastic and showed a fairly uniform thick-ening of the entire epidermis, and from thesixty-fifth day the ears of some of the micebegan to show "precancerous hyper-plasia." No definite carcinomas wereobserved within the duration of the experi-ment.

Creech (19), working with fibroblastsfrom connective tissue surrounding the ribof embryonic mice of pure strain, grewthe cells for short periods as cover-glasspreparations in a medium of fowl plasmaand chick-embryo extract containing 20-methylcholanthrene-choleic acid or ace-naphthene-choleic acid in concentrationsof 1007, 107, and 17 per cubic centimeter.1,2,5,6-Dibenzanthracene-choleic acid andphenanthrene-choleic acid were used inconcentrations of 1007 and 107 per cubiccentimeter. The outgrowth of the cultureswas measured at 45 and 70 hours, andrelative areas of outgrowth weredetermined by camera lucida drawingsand planimetric measurements. The cul-tures were finally fixed and stained andthe cells studied. Nearly 1,700 cultureswere studied. Methylcholanthrene-cho-leic acid in a concentration of 17 per cubiccentimeter (equivalent to 0.157 of methyl-cholanthrene per cubic centimeter) anddibenzanthracene-choleic acid in a con-centration of 107 per cubic centimeter(equivalent. to 1.57 of dibenzanthraceneper cubic centimeter) caused a significantincrease in cell proliferation as indicatedby measurements of outgrowth and bycounts of mitoses, while the same carcino-

548963-43-6

gens in a tenfold or greater concentrationcaused a retardation. Desoxycholic acid(107 per cubic centimeter), phenanthrene-choleic acid (107 and 1007 per cubic centi-meter), and acenaphthene-choleic acid(17, 107, and 1007 per cubic centimeter)all caused a decrease in cell proliferation.With the methylcholanthrene and dibenz-anthracene, a premature separation ofthe chromosomes in the prophase andmetaphase was observed. This did notoccur with the other substances tried.

The observation of Creech that methyl-cholanthrene-choleic acid retarded cellgrowth in a concentration equivalent to1.57 of the carcinogen, per cubic centi-meter accords closely with the data ofEarle and Voegtlin for methylcholan-threne. The observation that a concen-tration equal to 0.157 causes a significantincrease in the rate of cell cleavage em-phasizes further the necessity of investi-gating, over Jonger intervals, lower con-centrations of the carcinogen than anyyet used.

In the present work a concentration of17 of methylcholanthrene per cubic centi-meter of fluid culture medium was used.With this concentration, after the firstaddition of the carcinogen to the culturesthere was a sharp depression of the rateof increase in width of the cultures in everystrain examined. The continuous widthcurves of the different strains showed thata great part of the depression of the curvewas manifest by the end of the first 20 to35 days after addition of the carcinogen.In strain N, which was continued in thecarcinogen for as long as 184 days, thefurther depression of the continuous widthcurve was slight, while even in strain 0there was no further depression recogniza-ble until after 350 days in carcinogen, atwhich time there was a further sharpdepression of the curve. This broughtthe rate of increase of width of the cul-


rbana-Cham

paign on Novem


Dow

nloaded from



tures to such low levels that furthermaintenance of the strain was most diffi-cult. The strain would almost certainlyhave been lost during the next severalgenerations if it had not been removedfrom the carcinogen.

Examination of the cells just after theonset of the initial sharp depression follow-ing the first addition of carcinogen showedno evidence of an increase in cohesiveness,nor was there any increase in cell densityof the cultures. If there was any change indensity, the cultures seemed somewhat lessdensely cellular than normally. Mor-phologic evidence of injury to the cells ofyoung cultures, if it existed at all, wasextremely slight. Certainly, there was nosign of widespread degeneration or disinte-gration of the cells. The depression of thecurves, then, was not caused by any ofthese factors. While no determination hasas yet been made of the actual rate ofmitosis in the cultures, the data indicate adefinite retardation in the frequency ofmitosis in the celas. This would agree

with Creech's data (19). Whether thisretardation continued after cohesion ofthe cells set in, and up to 325 days afterthe first addition of carcinogen, was notdetermined and cannot be inferred fromthe data to date since the retarded rate ofincrease of the diameter of the culturewas at least partially offset by an increasein cell density.

The data suggest, however, that such aretardation of mitosis did occur in culturessubjected to the carcinogen about 400 days(strain 0) since the rate of increase ofwidth of these cultures was greatly re-tarded, and few cells were observed inmitosis in the living cultures. The studyof the action of the carcinogen on the rateof mitosis of the fibroblast for various inter-vals in long-term cultures under the condi-tions used must be left for future work.The conclusion is reached that, under theconditions of culture and of concentrationof carcinogen used, the changes observedin the cells in the early stages (D, H, J, L)of action of the carcinogen are not recog-

0 50 100 150 200 250 300 350 400 450

NUMBER OF DAYS IN METHYLCHOLANTHRENEFIGURE 22.—Continued average level of the continuous width curves after removal of culture from

carcinogen and through May 1943. The number of days the cultures were in the carcinogen isshown along the abscissa, the average culture width along the ordinate.


rbana-Cham

paign on Novem


Dow

nloaded from



205

nizably such as to lead to the death of thecell. However, the cumulative, progressiveaction of the carcinogen for longer intervalsand in the concentration used is not onlyclearly injurious but is almost certainlylethal.

In order to demonstrate more clearlythe continued depression of the continuouswidth curves after removal of the differentstrains from carcinogen, the average levelof the respective curves after removal ofthe strains from methylcholanthrene andthrough May 1943 is plotted against theinterval of time each strain was subjectedto the carcinogen (fig. 22). As a normalcontrol point on these curves, the averageof points on the continuous width curves,of the control cultures of strains D, H, I,J, K, L, N, and 0 through September 11,1942, have been used. Values later thanthat have been discarded because of thedepression in the curve after that date andthe possibility that this depression resultedfrom trace contamination with the car-cinogen.

The general trend of the curve is regularwith the exceptions of point J (which islow) and point L (which is high). Thelonger the cells were left in the carcinogen,the less was their continued rate of increaseof culture width after removal therefrom.The general morphologic appearance ofthe cells of strain J shows that they weremore altered than those of strain H, butless altered than those of strain L, whilethe cells of strain L were less altered thanthose of strain N. This intermediatemorphology confirms the correctness of thelateral position of points J and L along thecurve. Other data (3) also indicate thatthis dip in the curve at point J and therise at point L are real, although no expla-nation can he offered for this irregularity atthe present time.

Leaving out of consideration the depres-sion at point J and the rise at point L,

the conclusion is obvious that, as far as isindicated by the continuous width curves,the degree of change induced in the cellswith methylcholanthrene under the ex-perimental conditions was a direct func-tion of the duration of exposure (dose) tothe carcinogen, and that even after thecell was removed therefrom the inducedchanges persisted. In the instance ofthe cells of strains H, J, L, and N, thisalteration has now persisted more than 1year since their removal from carcinogen.

The question comes up whether or notthere has been any rise at all in thecontinuous width curves since the cellswere removed from carcinogen. Thecurve for strain H fluctuated over awide range and showed no clearly definedrise. The curves of all the other strainsmaintained their low level for some timeafter remova] from the carcinogen. Thistime included, for instance, 3 generationsin strain 0 and 12 generations in strain J.From these times on, the curves lost theirlevel, which had as a rule characterizedthem during the interval of carcinogentreatment and directly thereafter, andbegan to show erratic variations, and insome strains of cells, a rise. The signifi-cance of these variations is not yet clear,and the interpretation will be left to alater date.

From the various data presented, itseems evident that the cel] changes previ-ously described by Earle and Voegtlinhave been confirmed, while the accumu-lation of additional data makes possible aclearer description of the sequence of thecel] changes and of the later stages of thecells. As shown herein, in those strainsleft in the carcinogen for an extendedperiod, at about 25 to 50 days after firstaddition of the carcinogen, there was aprogressive shortening of the slenderterminal processes of the cell, and anincreased amoeboid appearance of the


rbana-Cham

paign on Novem


Dow

nloaded from



lateral edges of these processes, an in-crease that extended farther and fartheralong the surface of the cell toward a zonemidway of its length. Meanwhile the cellsbecame more coherent, particularly later-ally, and tended to form loose sheets andribbons, then sheets, and finally massive,thick sheets, particularly prominent at theglass interface of the culture. The individ-ual cells showed a progressively less clearlydefined major axis, and there was a definitetendency (at about stage J, as shown in fig-ures 17 and 18 (to send off amoeboid-ap-pearing lateral processes. In later stagesthe cells were so closely coherent that theextremely short, amoeboid-appearing proc-esses at the edge of the massive cel] sheetwere the only ones observed. In one or twoinstances in strain 0, it seemed that, theprocess of ceil rounding had gone so farthat the massive cell sheet or the heavyribbons of cells had apparently disinte-grated into living, isolated, rounded cells.Unfortunately, no photographs were madeof this phenomenon. It is interesting tospeculate whether such rounding and cellseparation might increase the tendencyof a tumor to metastasize in vivo.

This progression of changes, as describedpreviously (2), was gradual and appearedto affect the whole culture rather than anisolated cell or a small clump of cells.There was no suggestion whatsoever of asudden drastic change. The gradualchange may be seen from a comparisonof figuren 15 to 21, inclusive. Further-more, the changes, particularly those ofcell cohesiveness and rippling of thesurface, and changes in cell processen,seemed to be such as would arise fromchanges in the cell cytoplasm and moreparticularly in the cell surface.

The cells of strains H, J, L, and N, over1 year after their removal from the 1 yconcentration of methylcholanthrene, seemto have lost none of the typical cohesion

and other morphologic changes inducedwith the carcinogen. While strain 0 hasbeen carried for a shorter time sinceremoval from the carcinogen, it, too, haspreserved its structure. At eight genera-tions after the removal from carcinogen, thecells of strain 0 were more altered thanthose of strain N, but the architecture ofthe cultures appeared possibly a triflelooser than it had directly after removal.This change was so slight, however, thatat present it must be considered uncertain.

The conclusion is reached that over theextended period these cultures have beenstudied, once these morphologic changeshave been induced in the cells by the actionof the carcinogen, the cells to a very greatdegree seem to stabilize at that level ordegree of alteration induced in their mor-phology.

While occasional instances of very largenuclei, or of very large cells, or of cellscontaining as many as eight very smallnuclei were seen, usually these aberrationsoccurred in cultures more than 10 daysold and were far ruol e prevalent in thoseat about 25 to 30 days after planting. Theimpression was obtained from the livingcultures that in the greater number ofinstances such aberrations occurred underculture conditions of cell overcrowding anddegeneration. A more detailed study ofnuclear changes will be made when thestained slides are studied.

Some carcinogen certainly persisted inthe cultures for some time following theirremoval from it. However, in the morethan 1 year that strains H, J, L, and Nhave grown since they were removed frommethylcholanthrene at the concentrationof 1.Oy per cubic centimeter, each hasundergone more than 160 washings, morethan 160 additions of fresh fluid, and 14 ormore consecutive replantings. In each ofthe replantings it may be conservativelyestimated that not in excess of one-fifth


rbana-Cham

paign on Novem


Dow

nloaded from



207

of the original plasma (and cell) mass wastransferred to the new culture. If any ofthe original carcinogen used on the culturestill remains, it must be an extremely slighttrace.

From the changes undergone by the con-trol cultures, the presence of carcinogen inactive though trace concentrations mustbe assumed as probable in the experimen-tal cultures as well as the controls sometimein June 1942.

While a trace concentration of carcino-gen might conceivably have had an effecton a normal cell, or on one which had hadonly a very low concentration of carcinogenfor an extremely short interval, it appearsmost unlikely that such a trace contamina-tion, if it occurred at all, has substantiallyaltered the course of experimental culturesof strains H, J, L, N, and O. In spite ofthis, these strains and the altered controlshave to a most remarkable degree sta-bilized at levels of alteration of the cellswhich are proportional to thc time of theirexposure to carcinogen. Minor changeswhich have occurred in the last few monthsor which may occur in the future find theirmost probable explanation in cell selectionand in the overgrowth of more rapidlygrowing cell types within the cultures.

Bryan and Shimkin (12) injected methyl-cholanthrene in tricaprylin into the axillaeof male mice of strain C3H and deter-mined the final tumor incidente, as wellas the time-frequency relationship, forvarious doses of the hydrocarbon. Table1 is derived from the equations to theirestimated relationships. The recognitionof a tumor was determined when the cel!mass reached the size of a palpable nodulewhich continued to increase in size.

Obviously these latent periods must eachbe considered as consisting of an initialtransformative latent period during whichthe cells were changed from normal tomalignant, and of a latent period of

TABLE 1. Eslimated dose-response relationships for20-methylcholanthrene

Mice with tu- Esti-coated

Estimated latent periods

1-------mors (percent) dose Extreme Extreme

lower upper Averagelimit 1 limit 2

Gamma Days ! .Days Days182 39 163 9199-----------------

95---------------- 96 40 205 11050---------------- - 21 1 43 306 154

4.5 ! 46 407 1971_________________ 2.4 48 449 217

Derived from the data of Bryan and Shimkin (12).

1 Time at which not more than 1 percent of the mice in thetumor population would be expected to have developedtumors.

2 Time at which 99 percent of the mice in the tumor popu-lation would be expected to have developed tumors.

growth, an interval necessary for the cellmass to attain a palpable size. The datain another paper (3) show that the timerequired for subinoculated tumors, whichwere originally induced by the inoculationof cells treated with methylcholanthrenein vitro, to reach a palpable size is roughly7 days. If this interval is accepted asbeing a rough index of the latent periodof growth of such tumors and is subtractedfrom the latent periods given by Bryanand Shimkin, the time remaining may beconsidered a rough guide to the durationof the latent period of transformation.

The comparison of these results withthose obtained from the treatment ofcells in vitro with methvlcholanthrene isobviously complicated by differences, inconcentration of the carcinogen and bythe solvent used. In the results of Bryanand Shimkin (12), the minimal latentperiod for all doses of carcinogen rangesfrom 39 to 48 days, which would indicatea minimal transformative latent period offrom 32 to 41 days. This period coin-cides with the period necessary to effectgeneral characteristic morphologic alter-ations of the cells treated in vitro. Thecorrelation, however, is incomplete sincewith Bryan and Shimkin's data the periodfrom 32 to 41 days was the minimal, from


rbana-Cham

paign on Novem


Dow

nloaded from



156 to 442 days the maximal, and from 84to 210 days the average transformativelatent period. The closer correlation ofthese results must await further work.

Grady and Stewart (20), working withpulmonary tumors induced with methyl-cholanthrene and dibenzanthracene instrain A mice, showed by section of thelungs that definite tumors appear about 5weeks after injection. This latent intervalfor production of malignant tumors is inclose correlation with the general periodrequired to induce cell cohesion throughthe action of methylcholanthrene in vitrounder the conditions of the presentexperiment.

Furthermore, confirmation that cell co-hesiveness is intimately linked with theassumption of malignancy by the fibro-blast is shown by some unpublished studiesmade on tumors that arose in vivo throughinjection of methylcholanthrene into rats.Several of these tumors were cultured forshort intervals and showed definite cellcohesion. In a recent short study of astrain of tumor cells that originated from asubcutaneous injection of methylcholan-threne into a C3H mouse, cultures of thethirty-third passage of the tumors in miceshowed a degree of cohesion roughlycomparable with or slightly greater thanthat shown in strain D (control) cells (fig.15).

Jacoby (9), using a pure strain of fibro-blastic sarcoma cells derived from aspindle-cell sarcoma originally induced inmice with dibenzanthracene, described asimilar cohesion of the cells. Jacoby'sdescription, a part of which is quoted(9,p. 301), agrees with description of cellcohesion by Earle and Voegtlin and thechanges in vitro reported herein.

When grown in Carrel flasks in a hen plasmacoagulum and fed with either Heparin hen plasmaor a hen serum-chick embryo juice-Tyrode mix-ture, in which the embryo juice concentration iskept low, the cells regularly show a tendency to

grow out in close association with one another, andto form ribbon-like strands or even broad sheetswhich resembie very much the epithelial type ofgrowth in vitro. These ribbons frequently arborizeand their branches often join up with one another,forming loops and bridges enclosing the coagulum,as shown in the accompanying illustration. Thesheet-like growth is found especially at the inter-face coagulum-glass, whereas the cell ribbonsoccur also within the clot. This architectureseems to be very characteristic, and has been main-tained through frequent passages. It permits thediagnosis of such a sarcomatous colony with thenaked eye, being vastly different from the archi-tecture of normai fibroblast colonies; aided by themore highly refractive cytoplasm of the tumourcells it brings the whole colony much more intorelief against the background of the coagulum sothat, even for the naked eye or under low power,the entire edge of the colony is sharply outlinedand clearly defined.

The culture architecture shown inJacoby's illustration is typically similar tosome shown by Earle and Voegtlin (2) andcomparable with a type of ribbon growthseen at about stage J or L of the presentpaper. It should be noted that at the timeJacoby's article was written, he had notseen the second one (2) by Earle andVoegtlin in which this cell cohesion wasdescribed.

These data all confirm the concept thatthe change in culture architecture is closelyassociated with the assumption of malig-nancy by the cells. This concept is furtherstrengthened by the demonstration (3)that fibroblasts that have undergone ascarcely recognizable cohesion are capableof inducing typical sarcomas in a smallpercentage of injections into mice. Thepoint cannot yet be considered as conclu-sively demonstrated, however, because,owing to the transformation of the controls,no study has as yet been possible to deter-mine whether or not fibroblasts possessingthe normai loose radial growth pattern canbe made to produce sarcomas on injection.

Earle and Voegtlin (2) stated that thecohesiveness of cells under the action of


rbana-Cham

paign on Novem


Dow

nloaded from



209

methylcholanthrene regularly occurred atboth glass and fluid interfaces of the cul-ture. Later data from all carcinogen-treated strains necessitated modification ofthis statement. Jacoby's recognition (9)that the cohesion was especially prominenton the glass interface of the culture isentirely in accord with later observations.

No definite relationship has been workedout for the relative growth of the cells atthe glass and fluid interfaces, although theimpression has been that in the carcinogen-treated culture up to about 12 days of age,the cells were particularly prominent at theglass interface, and often there was nofluid interface cell layer. The cells of theglass interface showed conclusively theincrease in the cohesiveness of the cells fromthe action of the carcinogen. In some cul-tures, about 10 days old, the prominentlayer of ce]ls was at the fluid interface.When this occurred the cohesion wasalmost without exception lens marked andparticularly in cultures subjected to carci-nogen for shorter intervals was sometimesunrecognizable. In cultures of an olderage group the layer of cells at the glassinterface of carcinogen-treated cultures wasoften overgrown or had become entirelynecrotic; and in these cultures, too, thecells at the fluid interface were often notclearly cohesive, nor did they show soplainly such extreme shortening of theterminal cell processes in the early stages ofcarcinogenic alteration. They did showthe granulation observed in the later stages.Because of these differences in the cells atthe two interfaces, care must be taken todetermine correctly the positions of thecells within the depth of the culture beforereaching an} conclusion as to the degree ofcarcinogen-induced morphologic alterationthey have undergone.

The difference in the structure of glassand fluid layers is shown in figure 13, Aand B, from the glass and fluid interfaces,

respectively, of companion cultures ofstrain 0 at 403 days after addition ofcarcinogen and 11 days after planting.While these photomicrographs are entirelytypical and illustrate the point, they areunfortunately not nearly so remarkable incontrast with each other as the two layershave often appeared, sometimes in thesame culture in directly overlying regions.

The question arises why this increase incohesiveness of the cells has not been com-mented on by other workers who havegrown in vitro carcinogen-induced sar-comas which presumably arose from fibro-blasts. The possibility is that this cohesive-ness is prominent or occurs only in tumorsthat have arisen from certain strains offibroblasts or from tumors that have arisenfrom a very limited group of carcinogens,or in a very limited group of culturemedia. The fact, however, that the cohe-siveness is so much more prominent atthe glass interface suggests another ex-planation which seems more likely. Intissue-culture preparations of the type used(in Carrel flasks) the thickness of thefibrin clot is substantially greater than thethickness of the clot in either "roller tube"cultures (21) or in many, if not all,hanging- or lying-drop slide preparations.This would give a greater separation ofglass and fluid interface cell layers. Sincethe glass interface layer is the one closerto the microscope objective, in the flaskculture it naturally assumes a prominence,definition, and isolation which are prob-ably not so accentuated in other types ofcultures. With many slide cultures theextremely short interval during which thecells may be carried undisturbed, usuallyjust a few days, would itself interferegreatly with the appearance of this cohe-sion architecture as it is most easily recog-nized at about 8 to 10 days of culture.

A further question is why this change incell cohesiveness should be more promi-


rbana-Cham

paign on Novem


Dow

nloaded from



nent at the glass than at the fluid interface.While with the data available differericesin such factors as oxygen tension or rela-tive nutrition at the two levels in theplasma clot cannot be ruled out, it seemsmore .probable that the differente lies inthe reaction of the surface of the cell tointerfacial forces, which would be differentat the plasma-fluid interface from those atthe plasma-glass interface.

Central necrosis in carcinogen-treatedcultures appeared regularly at an earlierdate than in the controls. It usually ap-peared first at the glass interface. Thiscentral necrosis seemed to result from theextremely compact architecture of theculture with consequent crowding andinterference with the metabolism of thedeepest and most central cells. Youngcultui es, where the cells were not over-crowded, appeared to be free of necrosis.

During the whole course of the experi-ment, there was no detectable increase inthe rate of liquefaction of the clot, nor didany cell strain show an increase. A similarabsence of unusual liquefaction has alsobeen noted in other malignant°cells grownin this chicken plasma-horse serum- chickembryo extract medium.

In considering the nature of the cellchanges observed, it is desired to callattention to certain points. These are:(1) The drastic morphologic changes incells treated with carcinogen resulted inthe production of cell types greatly alteredfrom the original cell type used; (2) themorphologic changes were gradual; (3)they could be considered as changes thatmight arise from alteration in the cellsurface; (4) these changes stabilized afterthe removal of the cells from carcinogen,and different cell strains treated for dif-ferent periods stabilized at different levelsof morphologic alteration; (5) the mor-phologic changes were correlated with theaction of a definite, highly purified, crys-

talline chemical substance in extremely lowconcentration; and (6) the changes wereinduced in cells in an entirely heterologousculture medium quite removed from thesystemic influence of the parent host.

With our present Jack of knowledgethese considerations cannot lead to anydefinite conclusion. Even with the recog-nition of this fact, the suggestion is madethat the cell changes observed may weltbe of the same category as those seen inthe process of cell differentiation of theorganisrn, and it is emphasized that thesechanges are considered by some workersto result from definite chemical organizersacting in extremely low concentrations. Itis felt that this similarity is too close to beoverlooked or to be dismissed withoutmost careful consideration and that thedata call for further evidente either toconfirm the similarity or to reveal dis-similarities.

SUMMARYA primary strain of fibroblasts from the

subcutaneous and adipose tissue of a 100-day-old male mouse of the C3H strain,Andervont substrain ; was started in vitroin horse serum-chick émbry o extract-salinesolution on October 18, 1940. Cultureswere grown in Carrel D3.4 flasks, slightlymodified. When the strain reached anage of 291 days in vitro, the first additionrof purified 20-methylcholanthrene wasmade to selected groupp of the cultures ina solution of such strength that the finalconcentration of carcinogen in the fluidculture medium was 1y per cubic centi-meter.

Various selected groups of cultures werecarried in the carcinogen for differentintervals. Particular attention is directedto cell strains designated H, J, L, N, and0, which were subjected to the carcinogenfor 6, 32, 111, 184, and 406 days, respec-tively, at the end of which time they werecarried on without further addition of


rbana-Cham

paign on Novem


Dow

nloaded from



211

carcinogen. To june 1, 1943, these strainshad been carried 542, 542, 542, 663, and663 days, respectively, since first additionof carcinogen, and 536, 510, 431, 479, and257 days since their removal from it.

The first noticeable effect of the carcino-gen was a definite decline in the rate ofincrease in width of the cultures. Thisinitial reaction appeared after a vert fewdays in the carcinogen and in its prelimi-nary stages at least seemed to arise from aretardation of cell proliferation. This de-pression of the rate of increase of size ofthe cell clump continued even after thecells were removed from methylcholan-threne. It appeared that the depresssionof the rate of cel] proliferation was alsopresent at about 400 days in the carcino-gen, but whether the depression of therate of increase in width of the cell clumparose from this factor alone was not de-termined. The conclusion is that theaction of the carcinogen on the cells wasprogressive and gradual, and in the courseof time almost certainly lethal.

The earliest recognizable morphologicchange in the cultures appeared about 40days after first exposure to the carcinogen.In genera], the cells showed a gradualdiminution in the length of the terminalcell processes. The lateral edges of theprocesses became increasingly amoeboidin appearance, and this change extendedprogressively from the tips of the ce]1 pro-cesses toward the middle part of the cellas time in the carcinogen went on. Thecells gradually tended to send out numer-ous short lateral processes at their freeedges; meanwhile the cells became in-creasingly coherent, particularly later-ally, with the production of cell strands,cell ribbons, and finally of sheets whichshowed many characteristics of epithelialsheets in tissue culture. As time went on,the sheets became increasingly massiveand the cell cytoplasm extremely granular.

The changes induced in different cellstrains with different intervals of exposurecontinued after the carcinogen was dis-continued. To date, four cell strains, sub-jected to the carcinogen in the concen-tration mentioned for 6, 32, 111, and 184days, respectively, have been carried formore than a year since the carcinogen wasdiscontinued, without loss of their inducedcharacteristics. One strain has not beencarried so long without carcinogen, but italso shows no clearly defined diminution.

Over 8 months after the first recogniz-able morphologic cell alterations in theearliest group of experimental cultures(strain N) and approximately 2 monthsafter the last experimental strain (H) hadshown recognizable cell alterations, thecontrol cultures started to show definitealteration of the same type as but to a lesserdegree than the experimental cultures.Of the various control strains examined,strain H control showed least alteration,the change in the architecture of the cul-tures of this strain being barely discern-ible morphologically. The possibilitieswhether this change arose from a sponta-neous tendency of the apparently normalfibroblast to undergo a malignant changeunder cultivation in vitro or whether thechange was induced from trace contamina-tion of the control cultures with carcinogen,are discussed. The data available do notpermit a definite conclusion, but the work-ing hypothesis is advanced that such tracecontamination with carcinogen did occur,although the mechanism bv which itoccurred is unknown. In view of theprecautions taken to guard against con-tamination, it is emphasized that the con-tamination, if it occurred, was probablyextremely slight. Until further data areavailable to define more clearly the activeconcentration of carcinogen necessary toinduce changes in cells in long-term tissuecultures, great care should be taken in the


rbana-Cham

paign on Novem


Dow

nloaded from



handling of such cultures with carcinogento eliminate the possibility of complicationsthat might arise from trace contamination.

After all deliberate treatment of the cellswith carcinogen had been discontinued,and after strains H, J, L, and N had beencarried for more than a year after discon-tinuance of the carcinogen whereas theother strains had been carried for shorterperiods, the characteristics of the controlcultures and cultures subjected to carci-

nogen for 6, 32,111,184, and 406 days werecompared. The progression of morpho-logic changes with Jonger exposure to thecarcinogen was similar to that describedpreviously.

The similarity of the latent period for theearliest in vivo production of tumors withmethylcholanthrene and the time neces-sary for the production of recognizablemorphologic alteration of the cells in vitrois pointed out.

REFERENCES

(1) EARLE, W. R., and VOEGTLIN, C.: Themode of action of methylcholanthrene oncultures of normal tissues. Am. J. Cancer,34: 373-390 (1938).

(2) EARLE, W. R., and VOEGTLIN, C.: A furtherstudy of the mode of action of methylchol-anthrene on normal tissue cultures. Pub.Health Rep., 55: 303-322 (1940).

(3) EARLE, W. R., and NETTLESHIP, A.: Produc-tion of malignancy in vitro. V. Resultsof injections of cultures into mice. J.Nat. Cancer Inst., 4: 213-227 (1943).

(3a) NETTLESHIP, A., and EARLE, W. R.: Pro-duction of malignancy in vitro. VI.Pathology of tumors produced. J. Nat.Cancer Inst., 4: 229-248 (1943).

(4) EARLE, W. R.: Production of malignancy invitro. 1. Method of cleaning glassware.J. Nat. Cancer Inst., 4: 131-133 (1943).

(5) : Production of malignancy in vitro.II. Photomicrographic equipment. J.Nat. Cancer Inst., 4: 135-145 (1943).

(6) EARLE, W. R., and CRISP, L. R.: Productionof malignancy in vitro. III. Microcine-matographic equipment. J. Nat. CancerInst., 4: 147-164 (1943).

(7) PARKER, R. C.: Methods of tissue culture.Paul B. Hoeber, Inc., New York (1936).

(8) CUNNINGHAM, B., and KIRK, P. L.: Measureof "growth" in tissue culture. J. Cell. &Comp. Physiol., 20: 343-358 (1942).

(9) JACOBY, F.: Architecture of colonies of a purestrain of fibroblastic sarcomatous cellsderived from a dibenzanthracene mousetumor. (Letter to the editor.) Nature,London, 146: 301-302 (1940).

(10) EBELING, A. H.: A strain of connectivetissue Beven years old. J. Exper. Med.,30: 531-537 919).

(11) : A ten year old strain of fibroblasts.J. Exper. Med., 35: 755-759 (1922).

(12) BRYAN, W. R., and SHIMKIN, M. B.:

Quantitative analysis of dose-responsedata obtained with three carninogenichydrocarbons in strain C3H male mice.J. Nat. Cancer Inst., 3: 503-531 (1943).

(13) HOLLAENDER, A., COLS, P. A., and BRACKErIF. S.: Absorption and fluorescence spectrain relation to the photolethal action ofmethylcholanthrene on yeast. Am. J.Cancer, 37: 265-272 (1939).

(14) LEBENSON, E., and MAGAT, M.: Changes inthe biologica) properties of cells in skincultures of mouse embryos upon pro-longed action of dibenzanthracene. J.Méd. de l'Acad. d. sc., RSS d' Ukraine,7: 381-391 (1937). [In Russian, Englishsummary, pp. 389-391.]

(15) LARINOV, T., CHERTKOVA, M. A., andSAMOKHVALOVA: Alterations of the bio-logical properties of cells in tissue cul-tures under the action of cancerogenoussubstances. Bull. biol. et med. exper. de1'U. S. S. R., 9: 515-517 (1940).

(16) EARLE, W. R., and TIIOMPSON, J. W.: Theinfluence of the size of the explant uponcultures of chick fibroblasts in vitro.Pub. Health Rep., 45: 2672-2698 (1930).

(17) BENEVOLENSKAYA, S. V.: The effect ofmethylcholanthrene and 3:4-benzpyreneon cultured chick mesenchyme. Arch.d. sc. biol., 57: 94-105 (1940). [Englishsummary, p. 105.]

(18) COOPER, Z. K., and BELLER, H. C.: Mitoticfrequency in methylcholanthrene epider-mal carcinogenesis in mice. J. Nat.Cancer Inst., 2: 335-344 (1942).

(19) CREECH, E. M. H.: Carcinogenic and relatednon-carcinogenic hydrocarbons in tissueculture. Am. J. Cancer, 39: 149-160(1940).

(20) GRADY, H. G., and STEWART, H. L.: His-togenesis of induced pulmonary tumors instrain A mice. Am. J. Path., 16: 417-432(1940).

(21) GEY, G. 0., and GEY, M. K.: The mainte-nance of human normal cells and tumorcells in continuous culture. I. Prelimi-nary report: cultivation of mesoblastietumors and normai tissue and notes onmethod of cultivation. Am. J. Cancer,27: 45-76 (1936).


rbana-Cham

paign on Novem


Dow

nloaded from


Prdtn f lnn n Vtr. V. Th Fbrblt ltr nd hn n n th Lvn ll · z f th lp. ltr tht hd rvd th nt ln 4 d...

Documents

Transcript of Prdtn f lnn n Vtr. V. Th Fbrblt ltr nd hn n n th Lvn ll · z f th lp. ltr tht hd rvd th nt ln 4 d...