Studies on Resistance against 5-Fluorouracil

*

PETER REICHARD, OLA SKöLD, GEORGE KLEIN, LASZLÔ REv@sz,

AND PER-HENRIK MAGNUSSON

(Department of Chemistry I and Department of Tumor Biology, Karolinaka InstiMei, Stockholm, Sweden)

SUMMARY

Four different 5-fluorouracil (FU)-resistant tumor lines were developed by treatinga drug-sensitive Ehrlich ascites tumor, ELD, with FU during 925—SOpassages. Acetonepowder extracts of the tumor lines were assayed for the following enzymes : uridinephosphorylase and kinase, deoxyuridine phosphorylase and kinase.

In all four tumor lines significant decreases in uridine kinase activity occurred duringFU treatment, whereas the other enzymes were not affected. The decrease in uridinekinase activity commenced around the tenth tumor passage and showed either a continuous or stepwise character. The enzyme values after twenty passages were between5 and 15 per cent of the original values. From the available evidence it is proposed

that the enzyme changes were caused by a multistep process, probably of geneticorigin.

After 925—SOpassages the FU-resistant lines were cross-resistant against 5-fluorouridine and 5-fluorodeoxyuridine. Furthermore, two tumor lines, after eight and ninepassages, respectively, were resistant against FU before a decrease in uridine kinaseactivity had occurred. One of these early tumors was also resistant against 5-fluorodeoxyuridine, but not against 5-fluorouridine.

The loss of uridine kinase activity is thus only one biochemical factor which contributes to the resistance of the tumors against fluorouracil, and it is evident that oneor several other unknown factors exist which explain the early resistance against thedrug.

The resistant tumor lines showed no significant changes in permeability for the drug.

The pyrimidine, 5-fluorouracil (FU), originallysynthesized and studied by Heidelberger and Duschinsky and their co-workers (4, 7), greatly inhibits the growth of a variety of tumor cells(7, 11). However, after prolonged treatment withFU a population of cells is obtained which showsa greatly increased resistance toward the growthinhibiting effect of the drug (1, 6, 192).

We have attempted to study some of the changesin enzyme activities which occur in tumor cells

S This investigation was supported by grants from the

Swedish Cancer Society and the Damon Runyon MemorialFund for Cancer Research (DRG-470 B) and from the UnitedStates Public Health Service (Grant No. Cy 4619).

Received for publication September 1, 1961.

developing resistance against FU. In an earlierstudy (192) we compared the capacity of FUsensitive and -resistant cells to transform FUto the nucleotide level. It was found that extractsfrom two resistant lines contained considerablyless of several enzymes involved in this process,as compared with extracts from the original FUsensitive lines.

The present study represents a continuationof our earlier experiments. Four FU-resistant linesof Ehrlich ascites tumor were developed by prolonged FU treatment of tumor cells during serialtransplantations. For each passage assays weremade of the enzymes of the “uracilpathway― ofpyrimidine biosynthesis. These enzymes catalyze

9235

I. Enzymes of the Uracil Pathway during

Development of Resistance

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

9236 Cancer Research Vol. 9292,February 19692

the formation of UMP and deoxyUMP from uracilby the following reactions:

(a) Uracil + ribose 1-phosphate @±uridine +phosphate (uridine phosphorylase).

(b) Uridine + ATP —pUMP + ADP (uridinekinase).

(c) Uracil + deoxyribose 1-phosphate ±deoxyuridine + phosphate (deoxyuridine phosphorylase).

(d) Deoxyuridine + ATP —pdeoxyUMP +ADP (deoxyuridine kinase).

It was earlier shown that FU can substitutefor uracil in reactions (a)—(d) and is thus transformed by these enzymes to FUMP and F-deoxyUMP (15).

ELD

R0 R@

6 77 8, I I

I , II II I

I , I I@ I I I

31• .25 .29 .26

CHART 1.—Derivation of H., Re., B,, and R@lines from theELD tumor. The figures indicate the number of serial intraperitoneal transplantation passages. During each passage thetumor-bearing animals were treated with intraperitoneal FUinjections on alternate days.

MATERIALS AND METHODS

Mice—Heterozygous albino mice of both sexesobtained from a commercial breeder were used.The animals were between 92and 4 months oldand weighed between 16 and 926gm. They werekept on a standard pellet diet. Food and drinkingwater were available ad libitum.

Tumor—An Ehrlich ascites tumor, designatedELD, was used. Several properties of this tumorline have been studied previously and are summarized in a recent publication (13). The mainline of this tumor was maintained by serial, intraperitoneal transfer of about 92X 1O@cells in 0.1ml. ascites, diluted tenfold with Ringer's solution,containing 100 I.U. penicillin and 100 @gstreptomycin/ml. During several periods of varyinglength, the cells were preserved by frozen storageat —79°C. in a tumor bank as described previously (10).

FU-treated lines, denoted R0 and R1, respectively, were derived from the ascites of a mousethat had been given inoculations of ELD. Two

further lines, Rb and R@, were derived from R0after its third and sixth passage, respectively(see Chart 1). Each of these lines was propagated

by serial, intraperitoneal transfer of 0.1 ml. dilutedascitic fluid containing 10@cells. At each transfer,fifteen mice were given inoculations. Starting theday after inoculation, ten mice were treated bythe intraperitoneal injection of 920 mg/kg fluorouracil every 92dday, while five were left as untreated controls. For a subsequent transfer theascites was used which developed in any one ofthe animals of the FU-treated group. During theearly passages ascites did not become availableuntil 3—4weeks after inoculation, and in onlya few of the FU-treated animals; during the laterpassages aLmost all the inoculated animals developed ascites within 10 days. On a few occasionssamples of the ascites were frozen in several tubesand preserved viable in the tumor bank for varioustests at a later date. The survival time of each

of the inoculated animals was registered.Campound@—Nonlabeled FU, FUR, and FUdR

were generous gifts from the Hoffmann LaRocheCompany; FU-92-C'4 and uracil-92-C'4 were obtamed from the California Corporation for Biochemical Research. Uridine-92-C'4, deoxyuridine92-C'4 and orotic acid-92-C'4 were prepared as described earlier (14).

Enzyme assays.—Tumors were collected fromgroups of mice (usually ten mice per group) onthe 9th—l92thday after intraperitoneal transplantation of ascites taken either from material inthe course of an in vivo passage, or from a frozensample. The enzymes catalyzing reactions (a)—(d)were measured in acetone powder extracts, asdescribed earlier (14). In this paper one unit ofenzyme activity is defined as the amount of enzyme which, under the conditions of our assays,gave rise to the formation of 1 mj@mole ofproduct per mg. of acetone powder.

Biochemicat experivwnts with whole cell,.—Thetumors from ten mice were pooled 9—10 daysafter transplantation. For the experiments eachvessel received 3 ml. of the pooled ascites, 3ml. of Tyrode solution, 0.5 ml. of 0.92,5M glucoseand FU-92-C'4 (0.04—0.46 mmoles). Incubationswere carried out at 37°C. for 30 minutes, withshaking, in an 02 :CO2 (94 :6) atmosphere.

The reaction was terminated by cooling in ice.All the following manipulations before chromatography were carried out at temperatures slightlyabove 0°C. The incubation mixture was centrifuged for 5 minutes at 92000 X g, and the sedimented cells were washed twice with 5 ml. ofsaline. The cells were then homogenized with 92ml. of 0.6 M HC1O4, centrifuged, and washed

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

REICHARD et al.—Resistance against 5-Fluorouracil 9237

twice with 92ml. of 0.6 M HC1O4. The combinedacid supernatant solutions were neutralized with4 M KOH (phenol red) and centrifuged after about1 hour. The supernatant solution was then passedthrough a column of Do*ex-92-formate (0.9 X 4cm.). The column was first washed with 920ml.of 0.05 M HCOOH. This fraction contained FU,FUR, and FUdR. Next the column was washedwith 30 ml. of M ammonium formate : 4 M formicacid in order to elute FU-nucleotides.

The total radioactivity present in each fractionwas determined by measuring the radioactivityof aliquots at infinite thinness with a windowlessgas-flow counter.

Growth curves.—A suspension of 1 X 106 cellsof the tumor to be tested and taken either frommaterial in the course of an in vivo passage orfrom a frozen sample was injected intraperitoneally into each of a group of 80 mice of the samesex and similar weight. Starting the day afterinoculation and every 92d day thereafter the animals were treated by intraperitoneal injectionsof either 0.1 ml. of a solution of one of the fluorinated compounds (920 mg/kg of PU and FUdR,

@ @C

Number of possagesCHART 2.—Activity of uridine phosphorylase of tumor lines

B,,, Ri,, R., and R@,transplanted serially during treatment withPU.

3 mg/kg of FUR) in physiological saline, or 0.1ml. of a Ringer solution. On alternate days thesize of the tumor cell population was determinedin four randomly chosen animals. A quantitativerinsing procedure was used in combination withquantitative and differential cell counts. The dotails of this technic have been described previously(9).

RESULTS

Changes in enzyme activities during treatmentwith FU.—The activities of the enzymes catalyzing reactions (a)—(d)-—--uridinephosphorylase andkinase, and deoxyuridine phosphorylase and Idnase—were determined in extracts of four lines(R0, Rb, R0, and R@) of ELD ascites tumor duringthe course of treatment with FU. The analyses

I'o 20 30

Number of passagesRb CHART 3.—Activity of uridine kinase during treatment with

PU.

were carried out for almost every passage, andthe results, recorded in Charts 92—5,were obtainedover a period of about 1 year.

It is evident that the two phosphorylases and also deoxyuridine kinase showed considerable, rather irregular variations during this period. The

R@ changesareprobablyexplainedby physiologicalvariations, at least in part caused by the inhomogeneity of the animals used for the experiments.

—r-- Although the variations of these three enzyme

30 activities are of a random nature the results foruridine kinase, as represented in Chart 3, clearlyshow a general trend downward, starting aroundthe eighth passage in all four tumor lines. Thisdecrease of uridine kinase activity was slow andspread out over many passages. In two of thelines (R0 and R@)the final enzyme values observedafter 92.5—SOpassages were around 4 units, whereaslines Rb and R0, after about the same number ofpassages, had reached levels of around 10 units.The starting values were around 60. The decreasein activity after the eighth passage was apparentlymore or less continuous for lines R, and R0, untilthe lowest values of 4 and 10 units, respectively,

R080

40

80

40.@

@801 Rc

@@4o4@

1@ :@wT@@ R@

80

40

80

40

80

40

80

40

LI)

C

ci@E>@NJC

U

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

@@J\@QRb

Rc

.hI*@\.A.

9238 Cancer Research Vol. 9292,February 19692

were reached. On the other hand, line R@showeda stepwise decrease. In this line the enzyme values,which started around 65 units, reached a levelof around 30 units between the tenth and eighteenth passage and finally fell to 4 units after thetwentieth passage.

From line R0 (after the thirteenth passage)and line Rb (fourteenth passage), respectively,separate lines were developed and transferred during ten more passages without treatment withFU (“noFU― in Charts 92—5).Under such conditions no further changes in uridine kinase activity occurred.

Similarly, line R0 (after 81 passages with FU)was transferred during ten more passages withoutthe drug. The uridine kinase level remained unchanged at 4 units.

As a further general control the activities ofthe enzymes involved in the transformation oforotic acid to UMP were determined occasionallyin all four lines. Considerable variations wereobserved, but no general trend was apparent inany line. The values are not given here.

SLEW@ ‘

‘@O'2b'

Number of passagesCHART4.—Activity of deoxyuridine phosphorylase during

treatment with PU.

Finally, two separate lines of Ehrlich ascitestumor were treated with uracil (same dose asFU) during 928 passages, and enzyme activitieswere determined. No changes in any of the enzymes catalyzing reactions (a)—(d)were observed.

Determination of resistance against fluorouracilcompound&.—The mean survival time of the FUtreated mice decreased considerably (from about60 days to about 12 days) in the course of subse

quent intraperitoneal transfers of the differentsublines of the ELD tumor. This indicated that,during prolonged treatment, a marked increasein resistance of the tumor cells developed againstthe growth-inhibiting effect of FU. A similar trendfor a decreased survival time was apparent ifmice were given inoculations of FU-resistant cellsand subsequently treated with FUR or FUdR,

R02

2(I)

E:,

a.)2E>@lNJC

U

R@20

I0

20

I0

20

to

20@

I0@

0 20 30

Number of passagesCHART 5.—Activity of deoxyuridine kinase during treat

ment with fiT.

as compared with the survival of animals bearingpreviously untreated ELD cells and treated withthe same fluorinated compounds.

The survival time of the tumor-bearing animalsappeared, however, to be a rather unreliable indexof sensitivity (and resistance) of the tumor fortreatment. Inconsistent and often contradictoryresults were obtained. The poor reproducibilityof the survival data may have been due to the

_@T6_ immunizing action of the tumor cells upon their

immunogenetically foreign, homologous hosts. Thestrength of such a homograft reaction is an unknown variable, and may be superimposed uponthe' inherent sensitivity of the graft toward acertain treatment (13). Furthermore, the hostsurvival may also be influenced unspecifically bythe accumulated toxic effects of the drug duringprolonged use.

In an attempt to study the effect of the differentcompounds on tumor cell growth more directlythan by observing the survival of the host, themultiplication of the cells of the different lineswas followed quantitatively after inoculation. During the first 10 days, before an efficient host

C.,)C

:@

ci)E>.‘NJC

LU

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

2 6 tO 14 8 2 6 tO

REICHARD et al.—Resistance against 5-Fluorouracil 9239

response develops, the growth curves can be considered to be influenced to only a small, if any,extent by the homograft reaction or by an accumulated toxic effect of the compound on thehost, and may thus reflect a direct influence of thedrug on tumor cell multiplication.

Chart 6 gives the growth curves of tumor linesR0, Rb, R0, and R@after 925,929,923,and nineteenpassages with FU treatment, respectively. Eachline was tested not only for its sensitivity to FUtreatment, but also for cross-resistance againstFUR and FUdR (Chart 7). For comparison, Chart8 shows similar experiments with previously untreated ELD cells.

0

CHART 6.—Intraperitoneal growth of an inoculum of 1 X10$ cells derived from tumor lines H., Rb, R., and R@, pretreated with fiT in the courseof several serial transplantationpassages. The tumor-bearing animals were given injections ofPU or Ringer solution (control) every 2d day. Each pointrepresents a mean value calculated from the tumor cell numhers determined in four animals.

The results represented by Charts 6—8can besummarized by stating that lines R0, Rb, R0, andR@were not only resistant to treatment with FUbut also to treatment with FUR and FUdR.In most cases there was probably a very sinai!inhibition of tumor growth as compared withthat in the nontreated controls, indicating a slightantitumor effect of the different drugs. These

effects, however, were of doubtful significanceand can in no way be compared with the growthinhibiting effect of the three compounds on thesensitive tumor, as demonstrated by Chart 8.

The growth curves described above were ob

———FUR-.—. FUdR— conirol

IoI@R0

. 31 passages

CHART 7.—Growth curves of B,., Re., B,, and R1 lines duringtreatment with FUR, FUdR, or Ringer solution (control)on alternate days.

EJ.fl ::: FUR—.—.. FUdR

— control

Days after inoculation

C@ur 8.—Growth curves of ELD cells during treatmentwith PU, FUR, PUdR, and Ringer solution (control) onalternate days.

Rj26 passages

los@

E

tO@

0E

0'

.0E

;@ 08

Days af'er incculoton

Days after inoculation

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

the amounts of radioactivity in (a) the intracellular FU + FU-nucleoside fraction; (b) the intracellular FU-nucleotide fraction; and (c) thelast saline wash. The latter values are included,since it is believed that a large part of the radioactivity of this fraction may correspond to FU,originally located within the cells.

It is seen that the radioactivity in (a) above(= intracellular FU) from two of the resistanttumors (R@ and Rb) showed higher values ascompared with the FU-sensitive tumor (ELD),whereas the two other resistant tumors (R.@andR@) showed lower values. On the other hand,the latter two tumors showed the largest amountsof radioactivity in the last saline wash (= fraction[c]). When the radioactivity of fractions (a) and(c) are added together all five tumor lines showalmost identical values. It therefore appears thatthere is little difference in permeability for FUbetween the normal and the four resistant tumors.

Chart 11 also gives the amounts of radioactivitycorresponding to FU nucleotides (= fraction [b]).These values were considerably higher in thesensitive tumor than in any of the resistant lines.

.@t ———FUR

8passages—@—@—FUdRcontrol

U,a,I.,

0E

0

a,.0

Ez

CHART 10.—Growth curves of R3 line, pretreated withPU under eight serial passages, in the course of FUR, PUdR,or Ringer solution treatment on alternate days (two separateexperiments).

There is, however, no direct correlation betweenthe decrease in the amounts of FU-nuc!eotidesand the decrease in uridine kinase activities, asdescribed in Chart 3. The results thus show that,under the conditions of the whole cell experiments,the formation of FU nucleotides was only in partlimited by the amounts of uridine kinase presentin the cells.

2

Cancer Research Vol. 9292,February 196929240

tamed with tumors which had been treated oxtensively with FU and which showed a largedecrease in their uridine kinase activity. In twoof the tumor lines (R0 and R@) the effect ofFU on tumor growth was investigated alreadyafter nine and eight passages, respectively—i.e.,at a stage when there was very little change inuridine kinase activity (cf. Chart 3). Chart 9

tO@0E

@- t0@0

a,.0Ez

——— FU

6 tO 4 2 6Days after inoculation

CHART 9.—Growth curves of B,, and R4 lines during treatment with PU or Ringer solution every 2d day. The twolines had been pretreated with fiT in the course of eight andnine serial passages, respectively.

demonstrates that already at this point the twotumors were highly resistant to treatment withFU.

With line R2, treated during eight passageswith FU, we also investigated cross-resistanceof the tumor to FUR and FUdR (Chart 10).In two different experiments it was found thatthe tumor was resistant to FUdR, whereas therewas still a considerable growth inhibition by FUR.

Biochemical experiments with whole cells.—Further information about the state of the four FUresistant tumor lines, as compared with the original FU-sensitive ascites tumor, was obtained inexperiments in which tumor cells from the differentlines were incubated in vitro with radioactiveFU. After a brief incubation the amounts of radioactive acid-soluble compounds inside the cells weredetermined. These experiments were primarily dosigned to test whether there existed a differencein permeability to FU between the resistant linesand the original tumor.

The tumor cells were incubated for 30 minuteswith different amounts of radioactive FU, andthe cells obtained after centrifugation were washedtwice with saline. The washings were considerednecessary to remove all extracellular FU.

Chart 11 shows the results from such experiments. The curves for each tumor correspond to

— control

R09 passages

R@8 passages

to'

to 4 8

2 6 tO 2Days after inoculation

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

REICHARD et al.—Re@n@tance against 5-Fluorouracil 9241

DISCUSSIONIn connection with a search for agents useful

for the chemotherapy of malignant tumors a largevariety of different pyrimidines and purines inhibiting cell growth have been synthesized (5).Most of these substances are structural analogsof the purines and pyrimidines which form partof the nucleic acids, and the biological effectof the synthetic bases depends on their interferencewith the function of the normal bases. In mostcases the analogs must first be transformed to thenucleotide stage (mononucleotide or polynucleotide) before they exert their growth-inhibitinginfluence (5). These transformations occur throughpyrophosphory!ase, phosphorylase and kinase reactions, during which the analogs substitute forthe “normal―purines and pyrimidines.

Once one accepts the view that the action ofan analog depends on its transformation to anuc!eotide, then it follows that a cell which losesthis capacity will be resistant to the action of theanalog. By now many cases have been establishedin which enzyme activities necessary for the formation of nucleotides disappear on prolonged exposure of bacterial or neoplastic cell populationsto different types of antipurines or antipyrimidines(5). Such losses of enzyme activities actually seemto be a general rule, albeit with some exceptions,and it is therefore tempting to see a causal relationship between the enzyme changes and thedevelopment of resistance.

In our earlier experiments with one FU-resistantline of Ehrlich ascites tumor and one line oflymphoma L19210 we found in both cases a docreased capacity of the resistant cells to transformuracil (and therefore FU) to the nucleotide level

(192). In the Ehrlich tumor this depended on analmost complete loss of uridine and deoxyuridinephosphorylase activities, whereas the FU-resistantline of L19210 showed a considerable decrease ofuridine kinase activity and a smaller loss of uridinephosphorylase. In the present experiments wefound that uridine kinase activity decreased greatly in four different lines of Ehrlich ascites tumorduring prolonged treatment with FU. None of theother enzymes of the “uracilpathway― were affected.

In the following we shall first discuss this lossof uridine kinase activity during FU treatmentas a phenomenon per se and subsequently considerthe significance of the enzyme loss for the development of resistance.

It is evident that the decrease in enzyme activi

ty required prolonged treatment with FU. The firstdefinite changes were observed after the eighthpassage in all four tumor lines. This would correspond to about 100 cell generations. On continued treatment with the drug a further decreaseof uridine kinase activity occurred, until, aftera total of 925—30passages, the enzyme activitymay be as low as 5 per cent of the original (linesR0 and Ri). This further decrease in enzymeactivity was always slow and occurred either asan apparently continuous decrease (e.g., line Re),or in a stepwise fashion (most pronounced in R,).

When the tumor cells were transplanted in theabsence of the drug no further changes in enzymeactivity occurred. This was true when treatmentwas stopped either at intermediate or at thelowest levels of enzyme activities and demonstrates that in both instances a permanent changehad occurred. The results speak against enzyme

ELD //@@

R0

FU @FU-nucleosidesLost saline washFU - nucleotides

I00

50

r@)0

E

0

0

/

Rc ,@R@

.@.-, G ,0p@

2424

(C'4-Fu)

. 0@4

Rb

Mx104CHART 11.—Permeability for PU-C'4 of tumor cells from passages),@ (29 passages), R. (23 passages), and R4 (26

PU-sensitive line ELD, and from PU-resistant lines B,, (81 passages).

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

‘Cancer&8earch92492 Vol. 9292,February 19692

repression as the cause of the decreased uridinekinase activity. Against enzyme repression is alsothe observation that treatment with uracil inplace of FU did not cause any changes in enzymeactivities.

For the interpretation of our results it wasimportant to demonstrate the extent of selectiveadvantage which the resistant cells possess in thepresence of FU. In experiments to be reportedlater a mixture consisting of one resistant cellper 10@sensitive cells was transplanted in micewhich were treated with FU. Already after one ortwo passages the resulting cell population showeda uridine kinase activity characteristic of theresistant celLs, demonstrating that the resistantcells very rapidly overgrew the sensitive cellsin the presence of FU.

These results rule out one type of mechanismwhich might be considered as an explanation forthe slow decrease in uridine kinase activity duringtreatment with FU. This would be that a singleevent around the eighth passage had resulted ina tumor cell with very low uridine kinase activity,and that the changes in enzyme activity observedduring subsequent passages depended on a gradualincrease of the relative proportion of this typeof cell in the total cell population. Against sucha concept also speak the results obtained withline R,, where one can clearly observe a stepwisedecrease of uridine kinase activity.

Instead, we would like to explain our resultsby a mechanism which involves a series of events,probably of genetic origin, which, each in a stepwise fashion, decreased the amount of uridinekinase activity present in the tumor cells. Duringeach event one cell with a decreased amountof enzyme activity arose. The presence of FUon subsequent passages conferred a selective advantage to such a cell, and during a relativelyshort time period it gave rise to a new cell population with a decreased content of uridine kinaseactivity.

It is not yet clear whether the changes inenzyme activity reflect an actual loss of enzymeprotein or a change in the structure of the protein,resulting in a less active enzyme. A reproducibleand fairly simple method is now available for theextensive purification of uridine Idnase from Ehrlich ascites tumor, and experiments are at presentunder way to differentiate between these twopossibilities.

The second question to be discussed is whetherthe loss in uridine kinase activity may be considered as the cause of the resistance against thedrug. At the bottom of this page are summarizedthe enzyme reactions which lead to the formationof nucleotides from FU.

There are two pathways leading to the synthesisof F-deoxyllMP, the most important antimetabolite formed from FU (92, 8, 11). Only one of thepathways involves the participation of uridinekinase, whereas the other involves the sequentialaction of deoxyuridine phosphoryl&se and kinase.From the work of deVerdier and Potter (3) itseems probable that this latter pathway is oflittle importance for deoxynucleotide synthesisin vivo, since the substrate (= deoxyribose 1-phosphate) for the anabolic action of deoxyuridinephosphorylase does not seem to be available in thecell. It may therefore be assumed that a loss ofuridine kinase activity would lead to a substantialdecrease in the formation of F-deoxyUMP fromFU and therefore would result in cells with a higherresistance against FU.

Even though it seems clear that a decreaseof uridine kinase activity would result in an increased resistance against FU, we do not wishto infer that the enzyme loss is the cause of resistance. Significant for this question are our resultson cross-resistance against FUR and FUdR. Allfour FU-resistant lines were also completely resistant against FUR and FUdR. From the abovescheme it can be seen that the transformationof FUR to the nucleotide requires phosphorylationby uridine kinase. However, this does not applyto FUdR. This nucleotide is phosphorylated bydeoxyuridine kinase, which did not change duringtreatment with FU. Therefore, the resistance ofthe tumor cells against the deoxynucleoside couldnot have been due to insufficient phosphorylationof FUd.R, and at least one more change musthave occurred which rendered the cells resistantagainst FUIR.'

Further information on this point was obtainedfrom experiments on the development of resistance

1 A large part of the injected F-deoxyuridine is no doubt

cleaved to PU, which then may be utilized for nucleotide for'.mation. This reutilization would involve the action @furidinekinase. Thus, if F-deoxyuridine would exert all its carcinostaticaction only after cleavage to fiT, a large decrease in uridinekinase might well be considered to give the observed cross-resistance against F-deoxyuridine. However, we do not considerthis to be a likely possibility.

F-uridine uridme@F@P @FUDPFU7 kmase

\I .. ,- -.

F-deoxyuridme---.+LF-deoxyUMP*-—F-deoxyUDP

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

REICHARD et al.—Re.nstance against 5-Fluorouracil 9243

against FU, FUR, and FUdR in lines R@and R0,respectively, during their early passages. Aftereight and nine passages, respectively, these linesdid not yet show any significant decrease in uridine kinase activity. Nevertheless, both lines wereclearly resistant against treatment with FU (Chart9). Furthermore, after eight passages R@was alsocross-resistant against FUdR, but not againstFUR (Chart 10).

Since these two tumors were resistant againstFU (and FUd.R) without showing a decreaseduridine kinase activity, it is evident that a differentbiochemical change must be responsible for theresistance observed. However, the lack of resistance against FUR might indicate that the resistance against the ribonucleoside observed onfurther treatment with FU (Chart 7) was causedby the loss of uridine kinase.

Heidelberger et a!. (8) have shown that theaffinity of TMP-synthetase for F-deoxyUMP wasgreatly decreased in one FU-resistant line, andit was proposed that this was the biochemicalchange responsible for resistance. We have investigated the levels of TMP-synthetase and theinhibition of this enzyme by F-deoxytMP inour lines and could not find any significant changesfrom the original PU-susceptible cells.2 Heidelberger also recently found that his FU-resistant linehad lost the decreased affinity of TMP-synthetasefor F-deoxyUMP, without losing the resistanceagainst FU, and that therefore the altered TMPsynthetase could not have been responsible forthe observed resistance against FU.'

No gross changes in permeability for FU couldbe observed in any of the resistant lines, and suchchanges can therefore not be regarded as thecause of resistance.

Thus, we arrived at the conclusion that severalchanges did occur during prolonged treatmentwith FU and that all of them together expressthemselves in the final FU-resistant cell population. Relatively early during FU treatment(around the eighth to ninth passage) one or more

unknown biochemical changes have taken placewhich render the cell populations resistant toFU and FUdR. Later on we observe a stepwise

2 A. Hllggmark, to be published.

decrease in uridine kinase activity. We considerthis to be a factor which further contributesto the resistance against FU and which also rendersthe tumor cells resistant to FUR.

REFERENCES

1. BROCKMAN,H. W.; DAvis, J. M.; and S@rurrs,P. Metabo11amof Uracil and 5-Fluorouracil by Drug-Sensitive and byDrug-Resistant Bacteria. Biochim. et Biophys. Acta, 40:22—82,1960.

2. COHEN, S. S.; FLAx8, J. G.; B@iuraa, H. D.; Loss, M. R.;and LICHTENSTEIN,J. The Mode of Action of 5-Fluorouradl and Its Derivatives. Proc. Nat. Acad. Sd., 44:1004-12,1958.

S. DnVsnwzms, C.-H., and Porrsst, V. It. Alternative Pathways of Thymine and Uracil Metabolism in Liver andHepatoma. J. Nat. Cancer Inst., 24:18-29,1960.

4. DusclnNsxr, R., and Pi@vnsr, E. The Synthesis of 5-Fluoro-pyrimidines. .1.Am. Chem. Soc., 79:4559-00, 1957.

5. HANDSCHUMACHER,H. E., and Wnr.ca, A. D. AgentsWhich InfluenceNudeic Acid Metabolism. In: E. CatRGAFF and J. N. DAVIDSON (eds.), The Nucleic Adds, 8:458-526. New York & London: Academic Press, Inc. 1960.

6. HumEi.BEnoEn, C.; GROBAR,A.; B@xsm,IL K.; and MuXHEBJEE, K. L. Studies on Fluorinated Pyrimidines. X. InVivo Studies on Tumor Resistance. Cancer Research, 20:897—902,1960.

7. HEIDELBERGER, C. ; GRIESBACH, L.; MONTAG. B. J.;MOOREN, D.; and Cauz, 0. Studies on Fluorinated Pyrimidines. II. Effects on Transplanted Tumor@ Cancer Research,18:805—17,1958.

8. HEIDELBERGER, C.; K@woa, G.; Muxnxiaaxa, K. L.; andDANNEBERG, P. B. Studies on Fluorinated Pyrimidines.XI. ln Vitro Studies on Tumor Resistance. Cancer Research,20:908-9,1960.

9. KLEIN,G., and RgV@Z,L. Quantitative Studies on theMultiplication of Neoplastic Cells in Vise. I. GrowthCurves of the Ehrlich and MC1M Ascites Tumors. J. Nat.Cancer Inst., 14:229—77, 1958.

10. KLEIN, G.; R@v@@sz,L.; and KLEIN, E. Experiences witha Frozen Tumor Bank. Transplantation Bull., 4:81-88,1957.

11. LINDNER, A. Cytochemical Effects of 5-Fluorouracil onSensitive and Resistant Ehrlich Ascites Tumor Cells. Cancer Research, 19:189—94, 1959.

12. REICHARD, P.; SKöW, 0.; and Ki.xzN, G. Possible Enzymic Mechanism for the Development of Resistanceagainst Fluorouracil in Ascites Tumors. Nature, 183:939—41, 1959.

13. Rs@v@sz,L., and Noisse@.x, U. Relationship between Chromosome Ploidy and Radiation Sensitivity in Selected Tumor Sublines of Common Origin. J. Nat. Cancer Inst., 25:1041—64,1960.

14. Sxöw, 0. Enzymes of Uracil Metabolism in Tissues withDifferent Growth Characteristics. Biochim. et Biophys.Acta,44:1—12,1960.

15. . Nucleoside and Nucleotide Derivatives of 5-Fluorouracil. Arkiv Kemi, 17:59—71, 1960.3PersonalcommunicationfromDr.C.Heidelberger.

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

1962;22:235-243. Cancer Res Peter Reichard, Ola Sköld, George Klein, et al. Uracil Pathway during Development of ResistanceStudies on Resistance against 5-Fluorouracil: I. Enzymes of the

Updated version

http://cancerres.aacrjournals.org/content/22/2/235

Access the most recent version of this article at:

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

Subscriptions

Reprints and

.pubs@aacr.orgDepartment at

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

Permissions

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

.http://cancerres.aacrjournals.org/content/22/2/235To request permission to re-use all or part of this article, use this link

on July 5, 2021. © 1962 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/content/22/2/235http://cancerres.aacrjournals.org/cgi/alertsmailto:pubs@aacr.orghttp://cancerres.aacrjournals.org/content/22/2/235http://cancerres.aacrjournals.org/