1963 irving johnson vinca alkaloids (paper)

The Vinca Alkaloids : A New Class of Oncolytic Agents
IRVING S. JOHNSON, JAMES G. ARMSTRONG, MARVIN GORMAN,
AND J. PAUL BURNETT, JR.
(Lilly Laboratoriesfor Research and the Lilly Laboratotiesfor Clinical Research, Indianapolis, Indiana)
SUMMARY
A phytochemical investigation of the plant Vincaro@eaLinn. has demonstrated that
a number of alkaloidal substances can be obtained with antitumon activity. Over 30
alkaloids have been obtained, of which four—vinbiastine, vinleunosine, vincnistine,
and vinrosidine—are known definitely to be active. Chemically these compounds are
closely related to one another and to two monomeric alkaloids, vindoline and catharan
thine. The structure of these latter two compounds has been determined, and partial
structures for the biologically active alkaloids have been proposed. They represent a
new class of large complex dimenic alkaloids containing both indole and dihydnoindole
moieties.
Experimentally, a strain-specific, transplantable, acute, lymphocytic leukemia
(P-1534) carried in DBA/@mice served as a bioassay for obtaining these compounds
and for predicting their clinical activity. Vinblastine, vincnistine, and vinrosidine are
capable of prolonging and/on “curing―mice of the P-1534 leukemia even when therapy
is delayed until a near-terminal state of generalized disease. Resistance to an additional
challenge of leukemic cells has been observed in these “cured―animals. Parenteral ad
ministration of vincnistine has been demonstrated to “cure―mice given intracranial
implants of the P-1534. The experimental tumor spectrum and toxicological studies
are presented and discussed.
Biochemical studies performed to date do not reveal any effect on cellular nespira
tion, glycolysis, protein or nucleic acid synthesis. The mechanisms of action of these
compounds, which may differ within the group as well as from those of other known
agents, remain to be determined.
Only two of these compounds, vinbiastine and vincnistine, have received extensive
clinical evaluation. In spite of their close similarity, chemically, a somewhat different
group of human neoplasms responds to these compounds, and there has been a singular
lack of cross-resistance between these two drugs and any other oncolytic drug now in

wide use. Vinblastine has proved effective in chonioepithelioma, Hodgkin's disease,
and other lymphomas, and a number of beneficial results have been obtained in car
cinoma of the breast and bronchus. In addition, there have been smaller numbers of a
variety of other neoplasias reported as responding to this compound.
Vincnistine has been striking in its ability to induce complete hematological remission
of the acute leukemias of childhood, both lymphocytic and myelogenous in type. Re
sponses have also been reported in a number of other malignancies.
The problems and obstacles encountered in obtaining a full realization of the clinical
efficacy of these new types of oncolytic compounds are discussed in addition to areas
of clinical application other than those previously reported.
The plant Vinca rosea Linn. (periwinkle) of the the natural state pink and white color varieties are
family Apocynaceae is an ever-blooming pubescent found, and hybrids such as blush pink with red
herb on sub-shrub which is widely cultivated as an eye, crimson, and white with red eye are commen
ornamental in gardens throughout the world. In cially available. In a recent review (77) of current
1390
JOHNSON et al.—Vinca Alkaloids . 1391
phytochemical research on this plant, Svoboda has
pointed out that the probable correct botanical
name for this plant is Catharanthus roseus G. Don,
but, owing to the frequent and prevalent use of
Vinca ro8ea, the latter name will be used synony
mously.
This plant has enjoyed a popular reputation in
indigenous medicine in various parts of the world.
Peckholt (63) in 1910 described the use in Brazil
of an infusion of the leaves to control hemorrhage
and scurvy, as a mouthwash for toothaches, and
for the healing and cleaning of chronic wounds. In
the British West Indies it has been used to treat
diabetic ulcer (86) and in the Philippines has been

reported as being an effective oral hypoglycemic
agent (fl). The plant is also used in South Africa
as a hypoglycemic, and, in fact, a preparation un
der the name of “Covinca―is marketed for this
purpose. In England the plant has been sold for
diabetic treatment under the name “Vin-q-lin.―
The folklore reputation which this plant enjoyed
as an oral hypoglycemic agent independently stim
ulated its phytochemical investigation in two dif
ferent laboratories, unknown to each other. One of
these groups included Noble, Beer, and Cutts, then
at the Collip Laboratories, University of Western
Ontario, and the other a group in the Lilly Re
search Laboratories including Svoboda, Johnson,
Neuss, and Gonman. Although neither group could
substantiate the hypoglycemic activity, the Ca
nadian group observed a peripheral granulocyto
penia and bone marrow depression in rats asso
ciated with certain fractions (15, 18). Continued
investigation led to their preparation of vincaleu
koblastine (VLB) sulfate, an alkoloid capable of
producing severe leukopenia in rats (17, 60, 61).
During this period the Lilly group had demon
strated that certain alkaloidal fractions gave a no
producible prolongation oflife of DBA/@ mice given
implants of the acute lymphocytic neoplasm, the
P-1534 leukemia (41).
The detection of activity against the P-1534 leu
kemia was considered particularly significant, ow
ing to the fact that this tumor system has detected

other clinically useful antitumor agents in our lab
oratory (40) and had been sensitive enough to
study structure-activity relationships of active
compounds which correlated with the clinical ac
tivity (@3,46, 53). An intensive phytochemical in
vestigation resulted in the obtaining by Svoboda
(7@) of leunosine, a new dimeric alkaloid closely
related chemically to VLB, as well as VLB sulfate.
The effectiveness of both of these alkaloids against
the P-1534 leukemia was first demonstrated in the
Lilly Laboratories.
Early in this investigation it became apparent
that “indefinite―survival of animals implanted
with the P-1534 leukemia was being obtained by
tneatment with crude fractions of the plant which
were chemically free of both leurosine and VLB
(38). Further investigation of those fractions ne
sulted in Svoboda's obtaining still two other new
active alkaloids, leurocnistine and leurosidine (74).
The A.M.A. Council on Drugs has approved yin
blastine' (VLB), vinleurosine (VLR), vincnistine2
(VCR), and vinnosidine(VRD) as generic names
for these alkaloids; and they will be referred to by
these names in the balance of this report.
The purpose of this report is to summarize our
current knowledge of this new class of antineoplas
tic agents, including the chemistry, pharmacology,
possible mechanisms of action, and experimental
and clinical activity.
ISOLATION AND CHEMISTRY

Following the initial observation that extracts
of Vinca rosea Linn. produced prolongation of life
in mice with P-1534 leukemia, a detailed fractiona
tion of the plant was undertaken. It was shown
(41) that the activity was found entirely in the
alkaloidal constituents of this material, and that
the alkaloids of the leaves were far more active
than those contained in either the stems or the
roots (73). The leaf material was therefore used for
the preparation of the compounds described hero
in, and a procedure of differential extractions was
developed (78) which separated the alkaloids into
several groups, as shown in Chart 1, according to
their varying basicities This procedure, which in
volves the conversion of the bases to their tartrates
followed by extraction into organic solvents, was
first used by Svoboda (71) to prepare alkaloids from
another plant of the same family, ALitonia con
stricta F. Muell. It was found that the antitu
mor activity was primarily located in fraction A
(Chart 1).
Fraction A was partially purified by chromatog
raphy on alumina, deactivated by treatment with
10 per cent acetic acid. This yielded, in addition
to a number of inactive alkaloids (the results of
purification of fraction A are shown in Chart @),
two pure compounds possessing antitumor activity
(7@), vinleurosine and VLB (isolated as sulfate).
When the noncnystalline residues obtained from
this chromatography were tested for antitumor ac

tivity, certain of the post-VLB fractions appeared
to be more active than either vinleurosine or VLB
(74). These materials were subjected to a gradient
pH technic involving partition of the mixtures
between benzene and buffers from pH @.8to 7.5
1 Velban® (Vinblastine, Lilly).
2 OncovinTM (Vincristine, Lilly).
“AlkalineBenzeneExtract
1) 2% Tartaric Acid
2) EtCI2
GroundPlant
I SkellyB
Extract
1) HCI(2N)
2) NH4OH.CHCI3
Drug
1) NH3
2) Benzene
1) NH4OH
Alkaline
EtOH
Extract
Phenolic
Alkaloids
(C,D)
Benzene Benz-CHCI,(1:1) CHCI,
I I@ 1
I@ @.9.4.4pH4.@.6.4 [CATHARICINEJ
IVINDOLININE(.2HCl)@ISOLEUROSINBenzBenz(deactivated)FVIROSINE11CHCI,residues
I CATHARANTHINE LEUROSINE Moth Llq. Al,03 LOCHNERIDINE Combined 1

IA.JMALICINE@j VLB(-H,S04) chromatography@ chromatography
I Ai,O,
. CHCI3 Benz CHCI3 CHCI3
I I
________________________________________ Benz.CHCI, CHCI3
OSINEJ (1:3)@ Gradient
@ GradIent@ pH
____________I2.1.3.4
Benz@CHCI3
IVINDOUDINEI
(1:3)
ICAROSID1N.!JICombined
I1:3ResIdues1
‘Gradient
IpH(4.4.5.4) Ichromatog.
[VINCARODINEI
CHCI3. CH3OH
(99:1)
, [NE0LEuR0sIDINE]
CHART 2.—Isolation of alkaloids from fraction A
Phase
1) NaOH (pH 11)
2) EtCI,
2)
chromatography
Gradient
pHofmoth.
Iiq.(3.4)
[itEUROCRI@@j
DetailedDrug

1) 2% Tartarlc Acid
2) Benzene
CHART 1.—Extraction scheme for Vinca rosea L. leaves
[Frac@on@
Ai,03
Chromatography
@ii@iy-SolubleI
@lkalolds(E)]
.L@ I I
Benz-CHCI3Benz@CHCl,CHCI,
(3:1) (1:1)
I I
LAjmaucinel[LOIHNERI1@iE]___________
r—1-PERIVINEI
@ Moth.Liq.
Benzene
I (3:1)
IS1TSIRIKINE( Y2H2S04)@ Catharanthinetlne@1@ndolinej
H,S04
NameFormulaAjmalicine
Tetrahydroalstonine
Serpentine
Lochnerine
Akuammine*
Reserpine*C,1H,4N@O@
C,1H34N,O,
C21H,2N20,
C20H24N,02
C,@N@.N@O4
C@H@N2O,

JOHNSON et al.—Vinca Alkaloids 1393
in 0.5 pH unit increments. The resulting materials
were then crystallized as the bases or sulfate salts
as shown in Chart @.By this procedure two addi
tional biologically active alkaloids, vincnistine and
vinrosidine, were obtained pure (74) . These com
pounds differed in their activity from the previous
two in that they routinely gave survivors (38),
rather than just limited prolongation of life in
P-1534 leukemia. The remaining fractions (Chart
1) were purified as described above for fraction
A, and the results obtained are outlined in Chart
3 (75, 76). In addition to the alkaloids listed in
Chant 3 from fraction E, Moza and Trojanek (54)
At the time of our entry into this area a number
of well-known alkaloids had been reported from
this plant (78) ; these are listed in Table 1. They
are all found in other genera and species of the
family Apocynaceae, and their structures are well
documented. No direct chemical relationship could
be seen between the alkaloids in Table 1 and the
four active alkaloids. The remaining alkaloids
which are listed in Table@ with their tentative
empincal formulae, melting points, rotations, and
pK'@ values were all of unknown structure but
were, by and large, indole and dihydroindole in
nature (77). A comparison of physical properties
[FractionB1I
Al,O,
Chromatography

Benz@CHCl3
(3:1)
(Ajmalicine]
I Fraction
Al,03
Chromatography
Benz.CHCI3
(3:1)
IVindolineI
ITETRAHYDROALSTONINEI
Benzene
[FractionB
IA120,Chromatography
CHART 3.—Isolation of alkaloids from fractions A1, B1, B (A + B), E and F
have recently reported two additional compounds
tentatively named vindolidine3 and locherinine.
The biological comparison of the four active
alkaloids from this plant (38) indicated consider
able differences in the nature and scope of their
activities, and it was felt that an understanding of
the chemistry of this apparently new class of on
colytic agents was important to the eventual un
derstanding of their mode of action as well as to
enable one to study chemical modifications of these
compounds which might show enhanced activities.
3 Since the name vindolidine had previously been given to
another alkaloid (Table 2) Prof. Trojanek has suggested the
alternate name vindorosine for his compound.
TABLE 1
ALKALOIDS FROM Vinca roses LINN.

a Akuammine (vincamajoridine) and reser
pine have not been encountered in our investiga
tion, whereas the presence of the other four has
been confirmed.
IFractionE
IAl,03
IChromatography
F I
Benzene Benz-CHCI,
I (3:1)
I LOCHNERICINEI
[Tetrahydroalstonine@ [VlndOJifld
IFractionF]
IA120,
@J@hromatography
CHC1,
ISERPENTINE( HJ1O,I
Benz@CHCl3
(2:1)
I-CATHARINE
IVINDOLICINE
NameFormulaM.P., °C.*(a)@
CHCIspK'@ in DMFEtOH
U.v.)@max.,m@Vinleurosine
Vinblastine
Virosine
Perivine
Catharanthine
Lochnericine
Vindolinine-2HC1

Vindoline
Isoleurosine
Lochnenidine
Sitsirikine' 1/2 H@SO4
Vincamicine
Catharine
Vindolicine
Vinrosidine
VincristineC4H@N4O,t
C4H@N4O,
C@H,.N@O4f
CHH@N,O$
C@H24N,O,
C,iHUN@O@
C@H@N@O,•2HCl
C,@HI,N,O
C4@HoN4O,t
C,OHUN,OS@
@ 1/2 H@SO4
(Dimeric)
C@.H@N4O,.CU,OH
()
(Dimeric)
C@.H@N4O@0202—205
(decompn.
211—216(decompn.
258—264(decompn.
180—181
126—128
190—193(decompn.)

210-212 (decompn.)
154—155
202—206(decompn.
211—214(decompn.
289-241 (decompn.
224—228(decompn.
271—275(decompn.
248-251 (melts, re
cryst.)
265—267(decompn.
208—211(decompn.
218—220(decompn.+
72
+ 42@
—160 .5
—121.4
+ 29 .8
—482
— 9 (H,O)
—18
+ 61 .2
—607.5
+ 23 (Base)
+418
—54 .2
—48 .4
+ 55.8
+ 17 .0 (EtCh)5
.5, 7 .5 (H20)
S .4, 7 .4 (H,O)

5 .85 (66%)
7.5(66%)
6 .8 (66%)
4 .2 (66%)
3 .8, 7 . 1 (66%)
5 .5 (66%)
4 .8, 7.3 (66%)
5 .5 (66%)
7 .6 (66%)
4 .80, 5 .85 (66%)
5 .34 (66%)
5 .4 (66%)
5 .0, 8 .8 (33%)
5 .0, 7 .4 (38%)214,
259
214, 259
226, 270
226,314
226, 284, 292
226, 297, 827
245, 300
212, 250, 304
214,261,287
230, 298, 328
224,282,288
214, 264, 815, 341
222, 265, 292
212,257,308
214, 265
220, 255,296Carosidine

Carosine
Pleurosine
Neoleurosidine
Vincarodine
Catharicine
Vindolidine
Neoleurocristine
Vindolidi@ie#(Dimeric)
C@H@N4O@0f
C4H,sN4Oiot
C@4@H@N4Oi@t
@ C441L,N4O@ot
C@H,,N4O,ot
C4,H.@N4O@ot
C4,H,N4O@,t
C,,H,0N@O5263-278,
288(de
compn.)
214-218
191—194 (decompn.
219—225(decompn.
258—256(decompn.
281—284(decompn.)
244—250(decompn.)
188—196(decompn.)
167—
89 .8
+ 6 .0
+ 61 .0
+ 41 .6

—197.4
+ 34 .8
—113.2
—57 .87
—814
.4, 5 .5 (88%)
4 .4, 5 .55 (33%)
5 .1 (88%)
5 .8 (66%)
5 .8, 6 .3 (88%)
5 .3 (83%)
4 .68 (38%)212,254,308
255, 294
267,308
214, 268
280, 272, 298
214, 268, 293, 315
261, 311
220, 257, 298
250,302Lochneridinine#C@H,sN@O4163-169424247,
326
Cancer Research
Vol. @3,September 19631394
such as titration, infrared and ultraviolet spectra,
and nuclear magnetic resonance spectra of the four
active alkaloids indicated first of all that they
are closely related chemically, and secondly that
they are nonsymmetrical “dimeric―alkaloids (30).
Thus, for vinleurosine and VLB titration showed
the presence of two basic nitnogens, pK'a 5.5 and

7.5, one of which would quatennize when treated
with methyl iodide or similar reagents. Analyses of
Comparison of these spectra with those of sev
eral other Vinca rosea alkaloids, notably catharan
thine and vindoline (31) (Table @),clearly mdi
cated a close interrelationship between these com
pounds. When the infrared spectrum of a solution
containing equimolar quantities of catharanthine
and vindoline was compared with those of vinleu
rosine and VLB, it was virtually supenimposable
from@ to 8@ and quite similar up to 16@ (Chant
TABLE 2
NEW ALKALOIDS FROM Vinca rosea LINN.
a The melting points were determined on a Kofler microstage. The ultraviolet absorption spectra were obtained with a Cary
model 14 spectrophotometer.
t Whilethesemolecularformulasagreewellwiththeanalyticalresultsforeachparticularalkaloid,it shouldbenotedthatthey
are to be considered as proximate at this time, in light of our experience with the other dimeric alkaloids (57).
@ Determined on VLB etherate.
§OnthebasisofmassspectrometricevidenceobtainedbyH.BudzildewiczandI. M.Wilsonof Stanford,wenowpreferthe
C,@,formulationratherthanthe C1,firstreported(54).
#SeeRef.58.
the free bases, sulfates, dihydrochlonides, and qua
ternary salts indicated that they are C@H@_58N4O9
compounds (57). A study of the infrared spectra
(Chart 4, A—D)of the bases further supported this
contention (thus the intensity of the indole N-H
was approximately one-half that expected for a
typical indolic C21 alkaloid) and indicated that
VLB contains a hydroxyl group (X CHC13 @.8,
1O.0@)that is not present in vinleunosine. Aside

from this difference, the infrared spectra are very
similar (Chart 4, A—D)(30).
5). It was assumed that the double molecules yin
leurosine and VLB were composed of catharan
thine and vindoline moieties, with minor molecular
modification, bonded together in some unique
manner (30) . It was therefore possible to investi
gate the structures of the more plentiful smaller
molecules and then relate this informatioa to the
dimenic compounds.
Cathananthine (31, 56) (I)@ was found to be a
C21H2@N@O2compound and to form a methiodide
4 Catharanthine and derivatives are shown in Chart 8.
FREQUENCY(CM@')
WAVELENGTH (MICRONS)
I
WAVELENGTH (MICRONS)
CHART 4.—Infrared spectra of biologically active Vines alkaloids
A. Vinleurosine C.
B. Vinblastine D.
1396
Cancer Research
Vol. p23, September 1963
readily. The physical characteristics of the quater
nary salt showed that the basic nitrogen (pK'a 7.0)
was analogous to the one forming this salt in the
dimeric compounds (77). Spectral data indicated
that the alkaloid was a simple @,3-substituted pen
tacyclic indole containing an isolated double bond,
ester group, and closely related to the isoquinucli

dine alkaloid coronanidine5 (II), C21H26O2N2(@7).
The N.M.R. spectrum (Chart 6B) confirmed the
above assignments and indicated that the double
bond contained one vinyl proton at 5.Sö,and that
a C-ethyl group characterized by a methyl triplet
a 1.1öand an allylic methylene quartet at 1.9öwas
present (56). These data clearly suggested structure
(I) for cathananthine (Chant 8). Supporting cvi
dence for the position of the double bond was ob
double bond at C3-C4—viz., voacangine (III), di
hydrocatharanthine (IV) (vide infra)—the Ci-C18
bond is broken to afford 3-methyl-S-ethyl pynidine
(83).
Mild hydrogenation of I afforded only one iso
men, dihydrocatharanthine (IV). Its infrared spec
trum was strikingly similar to that of coronar
idine. The differences in these compounds could
be explained by examining Dreiding models which
showed that the hydrogens could only have en
tered the molecule from the side nearest to Nb to
give the axial ethyl group in dihydnocatharanthine.
The ethyl group in the Iboga alkaloids has been
shown to be equatorial (@).7
The reduction of catharanthine and dihydro
catharanthine with LiA1H4 afforded corresponding
FREQUENCY (CM1)
CHART 5.—Infrared spectral comparison of vinleurosine and an equimolar solution of catharanthine and vindoline
z
0
WAVELENGTH (MICRONS)

alcohols. The formation of the tetrahydro-1,3-
oxazine derivative (70) from cathananthinol gave
additional evidence for the position of canbometh
oxyl at C18.
The Iboga alkaloids possessing a carbomethoxy
function at C13are known to decarboxylate smooth
7 The following evidence can also be correlated with this
assignment:
a) The pK'a values of the axial series are higher than those
in the equatorial series; dihydrocatharanthine pK'@ 6.4,
coronaridine pK'. 6.1 (38% DMF).
b) The rate of methiodide formation has been found to be
a sensitive indicator of the configuration of the C-ethyl group,
the rate being much faster in the axial series (private commu
nication, Prof. M. Shamma, Penn. State Univ., State College,
Pennsylvania).
c) The Re values of these compounds vary when they are
run by thin-layer chromatography on silica gel plates. (RT of
dihydrocatharanthine 0.2, coronaridine 0.8, solvent ethyl ace
tate-chloroform, 1: 1.) This difference may be ascribed to the
acidity of the silica.
tamed recently from decoupling experiments per
formed on the vinyl proton of the alkaloid.6
Saturation of this proton led to changes in the
shape of the C-@proton at 2.6@6,and, in addition,
small decoupling effects were seen on the protons
at 4.18@(C-5 proton) and 1.96 (methylene of ethyl
group). As expected, dehydrogenation of catharan
thine with Pd on carbon yielded 3-ethyl pyridine
by the cleavage of the allylic C1-C2 bond. In iso

quinuclidine Iboga alkaloids which do not have a
6 The correlation of isoquinucidine alkaloids (with no aro
matic enethexyl groups) is readily found by the observance of
a triplet centered at 6.8@ in the infrared spectrum. N. Neuss,
“LillyCollection of Physical Data on Indole and Dihydroin
dole Alkaloids,― Eli Lilly and Company, Indianapolis, Indiana
(1961).
6 The decoupling experiments were carried out by Mr. Paul
Landis, Eli Lilly and Company, on a proton-proton decoupler
patterned after that described by Mr. L. F. Johnson, Varian
Associates, Palo Alto, California (private communication).
JOHNSON et al.—Vinca Alkaloids
1397
ly on saponification (@9), followed by mild acid
treatment, or alternately after prolonged refiuxing
of the ester with hydrazine (65) in ethanol. Anal
ogously, dihydrocatharanthine afforded the con
responding descarbomethoxy base epi-ibogamine
(V), differing from the known ibogamine (VI) only
in the orientation of the C-4 ethyl group as pre
viously mentioned. The final confirmation of the
nature of the ring system was given by the isola
tion of 4-ethyl-@,6-dimethyl-11-H-indolo (@, 3-C)
quinoline from Se-dehydrogenation of epi-iboga
mine. The identical compound has been obtained
B
@@@ @u
‘ ‘ ‘ I I
400
300 260 100 0

5@O
C.Ps.
CHART 6.—N.M.R. spectra in CDC13 (60 mc.) A, vindoline; B, catharanthine
1398
Cancer Research
Vol. @3,September 1963
earlier by the same treatment of ibogamine (57).
Catharanthine (I) must therefore be &-dehydno
coronanidine.
The nature of the second half of the dimeric
compounds could best be understood by a study
of the dihydroindole alkaloid vindoline (VII) (@6,
31),a C@ basicnitrogen did
not form a methiodide in analogy to the nitrogen
with PK'a 5.4 in the dimenic alkaloids. The nature
of the six oxygens was determined by the forma
tion of suitable derivatives (31). Thus, brief treat
ment of vindoline with acid afforded desacetylvin
doline, C2@H30N2O5,corresponding to the loss of an
a Chart 9 shows vindoline and related compounds.
acetyl group. This was further corroborated in the
infrared spectrum by the appearance of a hydnoxyl
band (@.7@i)and corresponding reduction of inten
sity of the canbonyl band.
The presence of the methyl ester was shown by
lithium aluminum hydride reduction of the base
with simultaneous removal of the acetyl to yield
vindolinol, C@H30N2O4. The infrared spectrum of
vindoline, in which a broad band at 3.5—4.0@twas
conspicuous, showed that the fifth oxygen was a

hydrogen-bonded hydroxyl. Formation of the di
acetate C27H3@N2O7(acetic anhydnide in pynidine)
was accompanied by the disappearance of this
broad band in the infrared spectrum. From the
4.
VL.B
CHART 7.—N.M.R. spectrum vinbiastine in CDC1, (60 mc.)
I R :COOCH3
II R1 : COOCH3; R2: C2H5;
R3 H;R4 : H; Coronaridine
ill R1 : COOCH3; R2 : C2 H5;
R3 H; R4 : OCH3; Voacangine
IV R1@ COOCH3; R2: H;
R3@ C2H5; R4@ H; Dihydro
cat haranthine
V R1:H;R2:H;R3:C2H5;
R4 H; Epi-ibogamine
VI R1 : H; R2 : C2H5; R3 : H;
R4 H; lbogamine
CHART 8.—Structures of catharanthine and related compounds
JOHNSON et al.—Vinca Alkaloids
1399
study of the ultraviolet and infrared spectra, the
sixth oxygen was shown to be present as an ano
matic methoxyl on a dihydroindole chromophore.
The confirmation of the assignment of the oxygen
functional groups was obtained from the N.M.R.
spectrum (Chart 6A).
Vindoline consumed 1 mole of hydrogen at at
inosphenic pressure, showing it to be a pentacyclic

compound and afforded dihydrovindoline (VIII)
which could be converted to an amorphous hy
groscopic hydrochloride (Chart 9). Pyrolysis of
this salt at 195°—@00°C. in vacuum gave a distil
late from which a C21H2@N@O2compound (IX) was
obtained in a 15 per cent over-all yield by direct
crystallization from hexane. The methoxydihydno
(8).Thecleavagecanberationalizedasoccurring
in the manner shown in Chant 10.@
The position of the carbonyl in the ketone (IX)
was assigned on the basis of the presence of the
peak at m/e @98(M-4@ is equivalent with loss of
ketene in IX or ethylene in XI) which suggested
that this group involved either C-3 or C-4. Position
4 was selected for two reasons; the typical ABX
pattern at low field (3—3.58)in the N.M.R. spec
trum which was consistent with structure VIII in
dicated three protons between the canbonyl group
and nitrogen; and equilibration with CH3OD/
methoxide resulted in the introduction of only two
deuterium atoms per molecule (M = 34@); where
as if the carbonyl was at C-3 there would be three
I DELTA VALUES
2.65
3.4
5.u
2.1—3.0
6.91
5.23
6.30

5.43
3.75@ __
900
3.80
2.68
VII
CH3O R p
x R:CH3C@-
XI R=CH3
H
CH)O@@@J1@H
CH3
XII
VIII
H COOCH3
XIV
CH3d''211@
CH3
XIII
CHART 9.—Structures of vindoline and related compounds
indole portion of this compound was the same as
in vindoline, and the second oxygen was found to
be present as a ketonic canbonyl (X@fi'c',5.85j@)(@8).
Vindoline (VII) was found to possess the same
ring system as the alkaloid aspidospermine (X)
(7), obtained from plants of the Aspidosperma
genus which is in the same family as Vinca. This
observation was made by comparing the mass
spectral fragmentation patterns of the pyrolysis
ketone and dihydnovindoline with that of the aspi

dospermine (X) derivative desacetyl-N-methyl as
pidospermine (XI). In each compound the identical
intense peaks were found at rn/c 1@4,174, 188, and
@98(@8). A study by Biemann and co-workers has
shown that these four peaks are uniquely formed
by cleavage of this type of a pentacyclic system
B
@N@YD
CH3Q IC
CH3
Ix
XI
CLEAVAGE AT A,C rn/c •174
B.C rn/c =188
C.D rn/c =298
(HEAVY LINES) A,D rn/c :124
CHART 10.—Mass spectral fragmentation of the vindoline
pyrolysis ketone (IX, Y =C= 0) and desacetyl-Nmethyl
aspidospermine (Y
1400
Cancer Research
Vol. ‘23,September 1963
exchangeable hydrogens. The other fragmentation
peaks occurring in these vindoline derivatives (p28)
are consistent with the proposed skeletal structure
and are due to more complex fragmentation which
will be discussed elsewhere (7) . The similarities in
the mass spectra of dihydrovindoline and the C21-
ketone indicate that no rearrangements of the ring
system occurred during the pyrolysis. The correla

tion of the mass spectrum of vindoline with dihy
drovindoline is rendered more complex because of
the double bond in the former which alters the
typical fragmentation patterns somewhat; how
ever, all other physical measurements are con
sistent with the simple hydrogenation of an iso
lated double bond in the conversion of VII to VIII.
There remained then the problem of the place
ment of functional groups and the stereochemistry
still be consistent with the remainder of the spec
trum). Since in the N.M.R. spectrum of vindoline
(Chart 6A) there is a single unsplit proton peak at
S.43ô which shifts on removal of the 0-acetyl
moiety to 4.07ôthe carbon bearing the acetoxyl
must contain a single proton and those carbons on
either side of it must contain no protons. Likewise,
acetylation of the 3-hydroxyl in vindoline does not
result in an analogous shift of any protons to a
lower field, and it must therefore be tertiary. The
only arrangement consistent with these data is
that shown for vindoline (VII) with the hydroxyl
and carbomethoxyl groups attached to carbon 3
and the acetoxyl on carbon 4.
The double bond was placed as shown also on
the basis of the N.M.R. spectrum which shows two
cis vinyl protons (J = 10 c.p.s.), one of which
(5.885) is further split by two nonequivalent adja
cent protons with coupling constants J = 5 and
@2c.p.s. Since vindoline is stable to both borohy
dnide reduction and zinc and acid reduction, the

double bond is not present as an enamine (7, S
position) and therefore can only be in the 6, 7
position as shown. It is interesting to note that a.
related dihydroindole alkaloid, vindolinine (19)
(Tablep2),foundinthesameplant,hasbeenshown
to have structure XIV and possess an ester and
double bond in the same position as those of yin
doline.
The stereochemistry of the ring system of vindo
line (VII) is assumed to be as found in aspido
spermine (X), and this conclusion is supported by
the observation that the methyl (of the ethyl
group) in vindoline yields a 3 proton triplet in the
N.M.R. at abnormally high field (0.48ö).One can
explain this shift on the basis of increased shielding
due to ring currents from the aromatic ring. Ob
servation of models suggests that, with the aspido
spermine stereochemistry, the methyl is in the
proper position to exhibit such enhanced shielding.
The stereochemistry of the functional groups at
positions S and 4 is, as yet, not certain, and a.
discussion of this problem will be published else
where.
As mentioned above, spectral evidence strongly
suggested that the dimeric alkaloids were com
posed of a molecular combination of catharanthine
(I) and vindoline (VII) (59). This assumption
could be further substantiated by the products
of acid cleavage (concentrated hydrochloric acid,
stannous chloride, tin, refiux) carried out on the

four active dimenic alkaloids. In each case there
was obtained upon chromatography of the reac
tion mixture an indole compound (vide infra) , fol
lowed by vindoline derivatives (Chart 11). Vin
XVII XVIII
COOCH3 CH3O@ •OOCH3
XV R=CH3
XVI R=CHO
CHART 11.—Structures of dimeric alkaloids and derivatives
of the molecule. The position of the aromatic
methoxyl was shown to be at either C-is or C-16
by examination of the typical 1,@,4 aromatic pro
ton pattern (ortho splitting J = 8 c.p.s., meta split
ting J = @.Sc.p.s.). The final selection of C-16 is
based on the comparison of the infrared and ultra
violet spectra of vindoline with 6-methyoxy-N-
methyldihydroindole (XII) and by the isolation of
ind-N-methylnorhanmine (XIII), from a soda lime
distillation of vindoline at 3@50•This derivative
also indicates the position of the N-CH3 as being
on the anilino nitrogen. A second product (IX)
which was identical to the C21-ketone obtained
above was also isolated from the soda lime reaction.
The formation of the four major fragmentation
peaks in the mass spectrum of dihydrovindoline
(vide supra) not only indicated the ring system of
the alkaloid but showed that all the oxygen func
tions other than the aromatic methoxyl must be
located at carbons 3 and 4 (i.e., the peak m/e @98
could not contain the ester, acetyl, or hydroxyl and

JOHNSON et al.—Vinca Alkaloids
1401
blastine (XV),9 vinleurosine, and vinnosidine af
forded desacetylvindoline, thus proving the iden
tity of the dihydroindole portion of these alkaloids.
The corresponding fraction from the cleavage of
vincristine (XVI) yielded des-N(a@-methyldesace
tylvindoline, M = 400, C@ mass
spectrum, as well as ultraviolet and infrared spec
tra, demonstrates the relationship of this com
pound to desacetylvindoline. Both VLB and yin
cnistine yield the same tetracyclic indole derivative
velbanamine (XVII), C19H2@N2O, the infrared
spectrum of which clearly shows the oxygen to be
present as a hydnoxyl (57).
A related tetracyclic indole derivative, cleava
mine (XVIII), C19H24N2, was obtained from the
cleavage of vinleurosine (the correct formula of
vinleurosine is still in doubt owing to difficulty of
preparing a solvent-free sample) . Isolation of cleav
amine (XVIII) demonstrates the relationship of
the indole portion of C46alkaloids to catharanthine
(I), since it was also obtained by a similar acid
treatment of the latter (@6, @7,@9).Mass spectral
evidence indicates that cleavamine has structure
XVIII and that velbanamine is its dihydrohy
droxy analog XVIII. Recently, the structure of
cleavamine has been unambiguously substantiated
by an x-ray structure determination (47). The
structure of the tetracyclic indole derived from

vinrosidine has not, as yet, been determined.
Thus, vinleurosine, vinrosidine, and VLB differ
from one another only with respect to the catha
ranthine portion of the molecule, whereas VLB
(XV) and vincnistine (XVI) differ only with re
spect to the vindoline portion. The nature of this
latter difference was demonstrated in the following
way:
A comparison of the empirical formulae of VLB
(XV) and vincnistine (XVI) (Table @)shows that
the former has one less oxygen and two more hy
drogens than the latter. The U.V. spectrum of yin
cnistine is quite different from those of the other
three dimeric compounds (Table @),and this dif
ference is indicative of a different substitution on
N(a) of the dihydroindole moiety. The infrared
spectra of VLB and vincristine (Chart 4) are simi
lar, with the exception of the presence of a strong
additional band,@ S.94j@in the spectrum of
the latter. The N.M.R. spectra differ in that the
N-CH3 proton resonance at @.73ôpresent in VLB
(Chart7)ismissinginvincnistine;and,conversely,
in place of only one low field proton 9.85 in VLB,
there are two in vincnistine at 9.5 and 8.9ö.
Lithium aluminum hydride reduction of the two
alkaloids afforded good yields of the same penta
a Chart 11 contains structures of dimeric compounds and
derivatives.
hydnoxy derivative, C42H5,N406 (XIX). This for
mulation is consistent with the reduction of two

methyl esters, one acetate and an N-fonmyl group
in the case of vincnistine. The N.M.R. spectrum
of XIX shows, accordingly, only two methyl sig
nals, aromatic OCH3 at 3.88 and N-CH3 at @.73ö.
The N-CHO of vincristine is, therefore, at the
anilino nitrogen in place of the N-CH3 in VLB, and
therefore vincnistine is des-N(a)-methyl N(a)formyl
VLB (57).
The mode of attachment of the indole portion
of the dimenic alkaloids to the dihydroindole frag
ment must be through the C-is aromatic carbon,
since the N.M.R. spectrum shows only two protons
in a 1 :4 relationship in the aromatic portion of the
molecule rather than the 1,@,4 three-proton pat
tern present in vindoline (Charts 6 and 7).
From the data presented thus fan one can draw
a fairly complete picture of the structures of the
dimenic alkaloids. Vincnistine and vinblastine are
best represented by structures XV and XVI, ne
spectively.'° The point of attachment of the vindo
line portion as well as the location of the hydroxyl
in the catharanthine moiety is, as yet, obscure;
however, on the basis of the biogenetic considera
tions discussed below, we prefer position 4' for the
hydroxyl group. An N.M.R. study of the fully
acetylated VLB shows that this hydroxyl is ten
tiary. The point of attachment of vindoline in the
other two biologically active alkaloids is probably
the same as for VLB, but the nature of their oxy
genation in this region of the molecule is, as yet,

unknown and under investigation.
The remaining alkaloids in Table@ can be classi
fled into three general groups on the basis of the
chemical data available to date. Eleven of the
alkaloids (Table @)are definitely of the dimenic
indole-indoline class and closely related to the four
already discussed. These are listed in Table 3 The
resemblance is seen in their infrared, ultraviolet
and N.M.R. spectra as well as the results of acid
cleavage of several of them.
The second group (Table 3) which includes ca
thananthine and vindoline consists of ten mono
menic alkaloids ranging in size from C18to C25corn
pounds. The known alkaloids of Table 1 are also
all in this category. In addition to vindoline (@8),
catharanthine (56), and vindolinine (19) the struc
tune of lochnenidine (47) and vindonosine (54)
(desmethoxyl vindoline) have been described, and
partial structures of penivine (78), lochnenicine
10 The possibility of the indole portion being in the open
tetracycic form as in velbanamine (XVII) cannot be entirely
excluded on the basis of analytical (57) or infrared data (58).
Note Added in Proof: Recent mass spectral studies have
shown that the formula of VLB should be CuH58O9N4, and
thus the indole moiety must be tetracycic as in velbanamine.
ActivityAlkyl
halides(%
prolonga
tion)Leurosine:Methyl
iodide

Methyl chloride67-144
64Methyl
bromide41Methyl
tosylate
Dimethyl sulfate
Ethyl iodide45
44
128N-Propyl
iodide33Ally!
iodide71Bencyl
bromide
Carbethoxymethyl bromide
@9-Hydroxyethyl bromideInactive
Inactive
45VLB:Methyl
iodide
Ethyl iodide26
25Propy!
iodide
Dimethyl sulfate73
60
Cancer Research
Vol. @3,September 1963
140@
(31), sitsirikine (76), and lochneninine(54) have
been proposed.
The third group (Table 3) consists of dimenic
alkaloids which are not of the indole-indoline type.
Vindolicine (76) and vindolidine (75) appear to be
dimers of vindoline, and vincanodine (75) seems to

be related to the alkaloid vincine from V. minor
Linn.
With this background of knowledge concerning
the chemical nature of the dimenic alkaloids our
TABLE 3
CLASSIFICATION OF THE ALKALOIDS
FROM Vines rosea LINN.
@1Band Leurosine-like Alkaloids:
1. Vincaleukoblastine (Vinblastine)
2. Leurosine (Vinleurosine)
3. Leurocristine (Vincristine)
4. Leurosidine (Vinrosidine)
5. Isoleurosine
6.
7. Carosine
8.
9. Catharicine
10. Neoleurocristine
11. Neoleurosidine
Monomeric Alkaloids:
1.
2. Virosine
8. Catharanthine
4. Vindoline
5. Lochnericine
6. Vindolinine
7. Lochneridine
8. Sitsirildne
9.
10.

Miscellaneous Dimeric Alkaloids:
1. Vindolicine
2. Vindolidine
8. Vincamicine
4. Vincarodine
5. Carosidine
attention was turned toward the effect of struc
tuna! changes on biological activity. One can con
sider the four active alkaloids as minor modifica
tions of one another, and, as will be shown below,
each of these compounds has a somewhat different
antitumor spectrum; as far as is known, they also
exert a varying clinical spectrum and possess dif
ferent toxic manifestations (vide infra). A number
of the dimenic alkaloids of Table 3 such as neoleu
rosidine, neoleurocnistine, and cathanine possess
only one pK'. value (Table @),indicating that the
basic nitrogen in the catharanthine portion of the
molecule has been changed, most probably by
amide formation. None of these derivatives has
any activity.
It is apparent that all the alkaloids possessan
acetyl group at C4 (vindoline portion). Removal of
this group destroyed antileukemic activity in VLB.
VLB contains two hydroxyls, whereas vinleurosine
has only one (58). Complete acetylation of these
hydroxyls yields the respective acetates devoid of
antileukemic properties. LiA1H4 reduction of yin
leurosine and VLB, yielding the corresponding
carbinols (58), likewise destroyed activity. The hy

drogenation of the double bond in these compounds
also reduces activity.―
An interesting change in the spectrum of bio
logical activity of the active alkaloids was observed
upon quatennization of the basic nitrogen in the
catharanthine moiety of these compounds with
various alkyl halides (Table 4). Catharanthine
(pK'a 7.5) readily forms a series of quaternary salts
from the appropriate alkyl halide in benzene at
room temperature, whereas the basicity of the
basic N (pK'a 5.5) in vindoline is such that it does
not react under these conditions. In the case of
vinleurosine the methiodide and ethiodides were
TABLE 4
PER CENT PROLONGATION OF DIFFERENT
ALKYL HALIDES OF LEUROSINE AND
VBLIN P.1534LEUKEMIA
much more active than the parent base. With
VLB, vinrosidine, and vincristine, the quaternary
salts were considerably less active than the parent
compounds.'2 These results will be described in
more detail below. It therefore appears that the
requirements for activity in the vindoline portion
of the molecule are much more rigid than in the
catharanthine portion. A number of other deniva
tives are currently under investigation.
The nature of the biogenesis of these interesting
new dimeric compounds is of great interest. Wen
kert and others have outlined detailed schemes for
the biosynthesis of many monomeric alkaloids of

1@Private communication, Dr. N. Neuss, Lilly Research
Laboratories.
12 Personal communication, Dr. G. H. Svoboda, Lilly Or
ganic Development Laboratories.
MaterialDosage
(mg/kg/day)Av.
wt. change
(gm) T/CAv.
survival
time (days) TICPer
cent
increase in
survival
timePer
cent
indefinite
.
survivorsDefatted
whole plant
FractionA
FractionB
Totalalkaloids120
.0
0.5
30.0
6.0
7.5
15.0
15.0
75(Oral)+0

1/+2 7
—0.6/+1.2
—1.2/+1.2
+0.8/+2.9
+0.2/+0.6
—2.3/+0.6
+0.3/+0.5
—1.5/+0.626/15
27/19
24/19
28/17
30/13
20/13
30/13
21/1373
38
25
61
122
51
130
530
20
40
20
0
60
0
0
JOHNSON el al.—Vinca Alkaloids

1403
the Iboga (catharanthine) on Aspidosperirta (yin
doline) types. These are formed by the interaction
of an appropriately oriented unit with tryptophan,
and it has been shown that tryptophan is an excel
lent precursor for the Vinca alkaloids.'3 The non
tryptophan portion is thought to be derived from
either shikimic and pyruvic acids or, alternately,
from acetate via mevalonate to cyclopentane mon
oterpenoid intermediates. The details of these
schemes can be found in the paper by Wenkert
(87), and we shall discuss only the method of at
tachment. This is most probably photochemical
since, as mentioned above, the major portion of
the dimenic alkaloids are found in the leaves of the
plant. The first step can be visualized as an oxida
tion of vindoline to a free radical by a peroxidase
type of enzyme followed by attack at the double
bond of catharanthine by this radical. Alternately,
A possibly analogous chemical reaction to the
biochemical addition of an aromatic radical to a
double bond is the Meerwein anylation reaction
(67), in which the aromatic radical formed by de
composition of a diazonium salt adds to the least
substituted position of a double bond according to
the reaction in Chart 1@. Further investigations
into the nature of this biosynthetic mechanism are
planned.
EXPERIMENTAL BIOLOGICAL
PROPERTIES

A@rITtmzoR STUDIES
The most striking experimental biological effect
of the four Vinca alkaloids under discussion are
their effectiveness in prolonging life or, in some
cases, “curing―DBA/@ mice given implants of the
P-1534 leukemia. The procedures for antitumor
TABLE 5
AcTIvITY OF ORIGINAL EXTRACT OF WHOLE PLANT AND CRUDE
FRACTIONS AGAINST P.1534 LEUKEMIA
the vindoline may react in the phenolic state be
fore methylation, and the cathananthine may be
already oxygenated at the double bond (hydnoxy!
or epoxide).'4 The resulting intermediate could
then be reduced to a neutral molecule in various
ways (such as reaction with water) to yield the
wide variety of dimeric products found in the
plant. The large number of dimeric products speaks
for the addition step of the free radical not being
enzymatical!y controlled. The formation of the
N-CHO derivative of VLB (i.e., vincristine) is
analogous to the oxidation of various dimethylani
lino compounds by horseradish penoxidase enzyme
and hydrogen peroxide where oxidative demethyl
ation occurs (p24). The peroxidase-catalyzed di
menization of aniline derivatives is known to yield
compounds similar to vindolicine (a vindoline
dimer).
13 Unpublished data, Drs. R. E. McMahon and M. Gorman,
Lilly Research Laboratories.
‘4This scheme resulted from a discussion with Prof. E.

Wenkert, Indiana University.
AR-N2C(@ [AR.] H2-C=CR@
AR@CH2@CR2@X [AR-CHf CR2]
x
CHART 12.—Meerwein arylation reaction
testing in this laboratory have been previously de
scribed (4@) but essentially consist of intrapeni
toneal inoculation of leukemic on ascites cells and
subcutaneous trocar implantation of solid tumors,
with therapy being initiated the day following im
plantation and continuing for ten daily injections.
Effectiveness is described in terms of per cent pro
longation over saline-treated control groups in the
case of leukemic and ascites-beaning animals and
per cent inhibition in terms of comparison of mean
tumor diameters of treated groups to those of
saline-treated controls in the case of solid tumors.
The results obtained with the first extracts of
Vinca rosea in December of 1957 against the P
1534 leukemia are seen in Table S. From extracts
MaterialDosage
(mg/kg/day)Toxic
deathsPer
cent
prolonga
tionJndef
i
nite
survivorsVinleurosine7
.5

10.0
1125
20.0
150.0(Oral)0
0
0
1
00
32
41
76
460
0
0
1
0Vinblastine0.05
0.1
0.3
0.45
0.45
0.6
1.5(Oral)0
0
0
0
0
1
041
53
70

98
131
150
700
0
0
0
0
0
0Vincristine0
35
0.30
0.25
0.20
0.15
0.12
0.09
0.002
2
0
1
0
0
0
030
?
226
110
55
49

32
241
S
S
2
0
0
0Vinrosidine10
.0
7.5
5.0
4.0
3.0
2.01
0
0
0
0
0?
?
127
75
30
214
5
0
0
0
0
Dosage

(mg/kg/
day)Toxic
deathsPer
cent
prolongationIndefinitesurvivors3.0
3.0
3.0
0.6
18.0
4.5
9.0
9.0
6.0
20.0
30.00
0
2
2
2
0
1
0
0
0
0?
?
?
181
147
151

147
174
123
148
2385
5
3
0
0
1
2
1
4
1
4
TUMORHOSTALKALOIDS*ABCDLilly
mammary
Sarcoma 180
Adenocarcinoma 755
C-1498 leukemia
P-l534leukemia
L1210 leukemia
Ridgeway osteogenic sarcoma
Mecca lymphosarcoma
AKR leukemia
Ehrlich ascites
Freund ascites
S-180 ascites
B82A leukemia
Walker carcinosarcoma 256

Lilly mammary
Gardner lymphosarcoma
S-91melanoma
X-5568 myeloma
High malignancy clone
Lilly rhabdomyosarcomaDBA/1
CAP'
C57BL/6
C57BL/6
DBA/2
DBA/2
AKR
AKR
AKR
Cox std.
CAl?'
CAF@
C58
Rat
C3H
CSH
DBA/1
C3H
CSH
Rat+
—
—
—
+++
—

—
—
—
+++
+++
+++
++
+
—
—
N.D.
N.D.
+—
—
—
—
++
—
—
—
+
+++
+++
+
+
—
—
N.D.
N.D.
——

—
—
—
+++
—
+++
+
-
—
++
+
++
+
+
—
+
—
—+
—
+
—
+++
—
+++
+
++
—
++
+++
+++

—
—
—
+
—
—
Cancer Research
1404
with this type of potency and with the P-1534 as
the bioassay, Svoboda obtained vinleurosine and
vinblastine. Results obtained with crude extracts
chemically free of vinleurosine and vinblastine are
seen in Table 6. From extracts of this type of
potency Svoboda obtained vincnistine and vinro
sidine. A comparison of the anti-P-1534 activity of
the four compounds under discussion is seen in
Table 7. In this connection, it must be stated,
however, that the response of the P-1534 leukemia
TABLE 6
ANTI-P-1534 ACTIVITY OF AMORPHOUSFRAC
TIONS FREE OF VINBLASTINE
AND VINLEUROSINE
(5 mice/group)
TABLE 7
ANTI-P-1534 ACTIVITY OF VINCA ALKALOIDS
(5 mice/group)
TABLE 8
COMPARISON OF ANTITUMOR AcTivITY OF FOUR “ACTIvE―ALKALoIDs
S A, vinbiastine; B, vinleurosine; C, vinrosidine; D, vincristine. + = 30-50% inhibition of solid tumors
or prolongation of survival time in leukemias; ++ = 50—100%inhibition of solid tumors or prolongation

of leukemias; +++ = 100% inhibition of solid tumors of > 100% prolongation of leukemias; N.D. = not
done.
Vol. @23,September 1963
CompoundDosage
(mg/kgX 1 X 10)Av.
wt.
change
(gm)Av.
surv.
(days)Per
cent
prolonga
tionIndefinite
.
survivorsVinrosidine
(VRS)
Vinleurosine (VIAl)
Vinblastine (VLB)
Vincristine (VCR)
VRS+VLR
VR@S+VLB
VRS+VCR
Saline3
750
15 .000
0.200
0 125
Asabove
‘4 a
U ii+0

9
+1 .2
+1 .5
+0 .4
—0.8
—0.6
—0.219
18
22
24
31
23
29
1437
29
61
74
128
69
1090
0
0
1
0
3
3
0
Treatment
started
(days after

implantation)Mg/kg
Xfreq.Av.
surv.
(days) T/CToxic
deathsPer
cent pro
longationIndefinite
survivors1
3
6
1
1
3
6
100.2SX1X1O
025X1X10
0.25X1X10
0.25X1X10
2.5X1
2.5X1
2.5X1
2.5X1?/14
28/14
33/14
30/15
26/15
23/15
31/15
33/153/5
3/5

0/5
0/10
3/10
3/10
0/10
0/10?
108
146
93
70
81
98
1122/5
0/5
4/5
2/10
4/10
0/10
0/10
1/10
JOHNSON et al.—Vinca Alkaloids
1405
as maintained in our laboratories has changed in
response to vinleurosine and vinblastine since 1958—
196@2,but the response to vincristine and vinrosi
dine during the period 1960 to the present has
remained unchanged. Vinb!astine currently gives
a lower order of activity (50 per cent prolongation)
at a maximum tolerated dose, whereas vinleurosine
gives a somewhat greater and certainly more uni

form response than previously. Unfortunately a
“tumorbank―was not available to us during this
greater therapeutic effect in the combination then
apy. Combinations of the various active Vinca
alkaloids themselves gave no evidence of any
greater activity in this tumor system either, with
the possible exception of vinrosidine (Table 9),
which seems to be additive with both vinleurosine
and vincristine but not vinblastine.
In a different tumor system, the B8@A leukemia,
there was a suggestion of additive effect with yin
leurosine and vincristine under certain regimens of
TABLE 9
COMBINATION THERAPY OF THE P-1534 LEUKEMIA WITH VINIt0SIDINE
AND OTHER VINCA ALKALOIDS
(5 mice/group)
TABLE 10
EFFECTS OF DELAYED TREATMENT WITH VINCRISTINE SULFATE ON P-1534
period, and we are unable to check sensitivity ret
roactively.
The four “active―alkaloids have similar expeni
mental tumor spectra but with some significant
differences (Table 8). The most striking difference
in spectrum is the activity of vinnosidine and yin
cristine against several AKR tumors.
A number of studies have been made expeni
mentally of combining vinbiastinc or vincnistine
therapy with therapy with other “standard,―use
ful antitumor compounds against the P-i534 icu
kemia. These studies will be described in some

detail elsewhere, but in all cases there was a no
ticeable lack of any evidence for potentiation or
therapy. The two regimens in which this sugges
tion might be made was by giving one compound
daily for a week followed by a week's therapy with
the other compound, and secondly by giving one
compound in the A.M. and the other in the P.M.
In the same system vinblastine and vincristine
were not additive.
Vinbiastine has previously been shown to be
effective against the P-1534 leukemia when thera
py was delayed for several days after implantation
(4@). Vincnistine is also effective in this manner
(Table 10), More remarkable, perhaps, was the
fact that “cures―or indefinite survivors could be
obtained when therapy was initiated late in the
GroupNo.
in groupNo.
surviving
at 40 daysAv.
surv.
time(days)A
B
C
D4)Challenged
with
4@, spleensfrom
@i “cured―
4) P-lS34animals@)ChaIIe@@@ed
with

4@, P-l534leuke
4
4J miacellS.2
11.5
i@.o
12.5
.
Lines of L1210Transfer
genera
.
tion
numberSubtelo
.
centric
chromo
some.
Relative
effect
ofVLB5L1210
V, parent line
L1210/MP ifi
L1210/A IV
L1210/FU Xlii
L1210/AMPFUXYI2
L1210/AMPFU XYI2, frozen
L1210/AMPFU XVI
L1210/AMPFU XYI4, frozen49
79
38
73

64
97
41
60Present
Present
Absent
Present
Present
Absent
Present
Absent
Present
Present
Present
Present1
1
0
3
4
1
4
4
4
4
4
4
1406
Cancer Research
Vol. @3,September 1963
disease. These animals were proved to be free of

leukemia by passage of sp!eens of such “cured―
animals to normal DBA/@ mice. These animals
remained free of disease but were not resistant to
a challenge with P-1534 cells. A typical experi
ment is seen in Table 11.
Resistance to a new challenge to P-1534 cells
after “cures―has been seen, however, and de
serves some comment in reference to the Vinca
alkaloids in general. This resistance, in our hands,
has not been a regularly predictable phenomenon.
It is our impression that this occurs most fre
quently in “survivors― obtained by delayed treat
TABLE 11
PASSAGE EVIDENCE FOR “CURES―OF YIN
CRISTINE-S04-TREATED ANIMALS
B-8174 leukemias and to a lessen extent in the
B8@. He has also observed a peculiar relationship
in terms of cross-resistance between vincnistine
and certain compounds in sensitive and resistant
lines of these leukemias (10) . Vincnistine was mac
tive in P-815 and P-388 leukemias resistant to
NSC 38@80 but is not additive to this compound
in a sensitive line. NSC 38@80 is potentiated in
these systems by stilbamidine, tniparanol, and
amphotericin, whereas vincnistine is not. Cross
resistance of these compounds is difficult to inter
pret in view of the obvious differences in chemo
therapeutic behavior. Burchenal has also observed
cross-resistance of vincnistine and vinblastine in a
line of P-815 made resistant to vinblastine.

Dr. Dornis J. Hutchinson'6 has observed a
marked effect of vinb!astine on several lines of
L1@10 leukemia resistant to amethopterin, 6-MP,
5-FU, and all three compounds (Table 1@).These
TABLE 12
EFFECT OF VINBLASTINE ON RESISTANT
LINES OF L1210
Four groups of normal DBA/2 mice were given intraperito
neat injections of suspensions of spleen cells from “cured―ani
mals. A!! survived until 40 days, when they were challenged
with P-1534 leukemia cells indicating their susceptibility, if
these cells had been present in the spleens of VCR.S04
“cured―animals.
ment rather than by therapy initiated soon after
implantation. It should be pointed out that Noble
and his associates have also reported resistance in
“cured―animals (17). They reported resistance in
BDF, hybrid mice “cured― of P-i534 leukemia
and subsequently re-challenged with P1534 cells.
We previously reported (39) our inability to repeat
this phenomenon but suggested the probability of
tissue incompatibility. Goldin (@5) has reported
a similar phenomenon with halogenated deriva
tives of amethoptenin in the case of L1@i0 leu
kemia, but again this was in an F1 hybrid and did
not occur in the inbred DBA/@ animal. Noble also
reported “cures―and subsequent resistance in a
transplantable AKR leukemia in AKR mice. The
curing and subsequent resistance to re-challenge
of a strain-specific tumor in the inbred host is

unique in experimental chemotherapy though not
unusual in nonstnain-specific tumors in random
bred animals.
Studies in other laboratories have revealed in
teresting data on resistant lines of mouse neo
plasms. Burchenal16 has found vincnistine to be
effective by several routes in the P-388, P-815, and
S New code designations: 0 = inactive; 1 = slight activi.
ty; 2 = active; 3 = <6 “cures―;4 —>5 “cures.―For corn
parison, the old code designations were: —, <24% increase
in survival time; ±, 25—59% increase in survival time; +,
<60% increase in survival time.
lines are characterized by the presence or absence
of a subte!ocentric chromosome. The presence or
absence of the chromosome appears to have little
to do with sensitivity to vinblastine but in general
it does appear that the more compounds the van
ous lines are resistant to, the greater the sensitivity
to vinbiastine. Similar studies with vincnistine are
in progress in Dr. Hutchinson's laboratory.
‘SPersonal communication, Dr. Joseph Burchenal, Sloan
Kettering Institute.
10 Personal communication, Dr. Dorris Hutchinson, Sloan
Kettering Institute.
Intraperitoneal
Intravenous
Oral3.2
±0.02 mg/kg
17 .47 ±1 .57 mg/kg
33.12±6.03mg/kg

JOHNSON et al.—Vinca Alkaloids
1407
ToxIcoLoGY
Pharmacology and toxicological studies of these
alkaloids in laboratory animals have not been pub
lished in detail. Vinblastine and vincristine have
been the most extensively studied by Anderson
and his associates.'7 Acute intravenous LD50's for
yinblastine and vincristine in mice are approxi
mately 17.0 and @.0mg/kg, respectively. Multiple
doses of vinblastine and vincnistine caused marked
leukopenia in rats and dogs at 0.50 and 0.1 mg/kg,
respectively, causing death in some instances.
Dogs have tolerated daily doses of @.0mg/kg of
vinleurosine with slight reversible leukopenia but
few other toxic effects. A daily dose of 5.0 mg/kg,
however, was eventually lethal. Neither vincnistine
nor vinleurosine is as potent a producer of leuko
penia in animals as vinblastine.
An interesting phenomenon with vinblastine is
the markedly different LD50 by intravenous and
intrapenitoneal administration in mice (Table 13).
The mouse tolerated larger doses of vinblastine
than did the rat or the guinea pig. Subacute toxici
ty studies in rats given various doses of vinblastine
revealed loss of weight, with reduction in food
intake. The degree of leukopenia produced bone a
direct relationship to the size of the dosage, and,
with sublethal doses, the leukocyte count returned
to normal after dosage was discontinued. Thirty

rats that survived repeated !eukopenia-pnoducing
doses of vinbiastine were found to be normal at
necropsy.
Given intravenously in leukopenia-producing
but sublethal doses to dogs, vinblastine was found
to cause a reversible reduction of the myeloid
components of the bone marrow.
The main pathological finding in dogs given
purposely lethal doses was hypoplasia of the bone
marrow. Of lesser prominence and consistency
were scattered petechial hemorrhages in other or
gans, with slight hepatitis and entenitis. Hypen
plasia of the marrow and extnamedullany hema
topoiesis were found at necropsy in those dogs that
were given courses of vinbiastine in sublethal doses
and that were not sacrificed until after recovery
from the !eukopenic effects of the drug.
The cause of death in animals following lethal
doses of vinb!astine was bacterial disease second
any to leukopenia. Death did not occur immedi
ately after the dosage but was delayed until infec
tion had occurred. Potentiation of phenobarbital
anesthesia in dogs and prolongation of hexobarbi
tal “sleepingtime―in mice with the use of vinblas
tine have been observed.
17 Personal communication, Dr. R. C. Anderson, Lilly Re
search Laboratories.
Toxicity studies with vincnistine have been con
ducted in rats, rabbits, cats, dogs, and monkeys
with the following results:

Rats.—Intravenous doses administered 5 times
weekly caused leukopenia, weight loss, and re
duced food intake in direct relationship to the
doses, which were 0.5 and 0.@5 mg/kg. Rats re
ceiving 0.@ mg/kg daily by the intraperitoneal
route survived 10—@1doses. They developed
leukopenia, weight loss, diarrhea, alopecia, and
blanching of tails, ears, and feet. Clonic extensor
convulsions and catalepsy occurred before death
in one animal.
Rabbits.—Intnavenous doses given S times week
ly caused leukopenia and death within three to
five doses at levels of 0.1 and 0.@ mg/kg. Before
death, signs of muscular weakness (head drop,
kneeling and sprawling positions) appeared. Ani
mals receiving 0.0@5 and 0.05 mg/kg appeared
normal after @8doses.
TABLE 13
LDHOFVINBLASTINEINMICE
BY DIFFERENT ROUTES
Cats.—The animals received the drug by the
intravenous route 3 times weekly. Those given 0.@
mg/kg were dead after two doses. A cat adminis
tered 0.1 mg/kg was killed after two doses because
it was moribund. When 0.05 mg/kg was given to
one animal, it died after eight doses. A cat receiv
ing 0.0@5 mg/kg was killed after thirteen doses
because of its extremely poor condition. Weight
loss, general debilitation, clonic convulsions, and
leukopenia were observed.

Dogs.—Vincnistine was administered by the in
travenous route 5 times weekly. The dogs given
0.05 mg/kg were dead or had been killed after the
fifth dose. One animal receiving 0.O@5mg/kg was
killed after seven doses because it was in poor con
dition. Three animals survived twenty doses at
this level. Leukopenia, tremors, vomiting, weak
ness, and bloody stools were observed in these ani
mals.
Monkeys.—Vincnistine has been administered
orally in solution (1 mg/mi) to two male monkeys
at dosage levels of@ and 4 mg/kg, respectively.
Hematological and blood chemistry studies on
each monkey were essentially unchanged through
out a @9-dayobservation period. Vomiting was
observed in the animal that received 4 mg/kg.
A single intravenous dose of 4 mg/kg killed one
VincristineFolic
acidFolinic acidAv.
survival
time and
range
(days)Mortality
at SOdays2.5X1
I.Y.
a52X1X10
I_P.14OX1
X10
I.P.11
(2—26)

7 (2—19)
7 (4—10)17/19
16/18
4/16
VincristineFolic
acidFolinic
acidAv.
su@
vival
time
(days)Per
cent
prolonga
tioiiIndefinite
.
survivors0.2
0.2
0.2
0.2
0.250 140
50140
140
5085
?
2.3
13*
21
6
?
22

?
15*172
?
77
38
0
?
41
?0
5
4
0
4
1
5
3
5
0
Cancer Research
Vol. p23, September 1963
1408
male monkey in 30 hours. Constricted pupils, ano
nexia, and diarrhea were observed prior to death.
Reduced hematocrit reading, leukocytosis, and
elevated blood sugar, blood urea nitrogen, and
serum transaminase were also observed. The major
pathological finding was a severe acute cytotoxic
effect on the granulocytic and erthnocytic series in
the bone marrow.
Two male monkeys survived two doses of yin

cristine,@ mg/kg each, administered 36 days apart.
Vomiting, weakness, tremors, weight loss, leuko
penia, and thrombocytopenia were observed in
both animals. Decrease in blood sugar was noted
TABLE 14
EFFECT OFFOLIC ANDFOLINIC ACIDS ON YCR ToxICITY
Dosaoz (un/au)
These effects, as well as mild to moderate degen
erative changes in the liver and kidney, were re
lated to dosage. An unusual reaction noted was
proliferation of the epithelium of the renal collect
ing system.
A recent observation has made been in this
laboratory in reference to host protection of mice
to lethal doses of vincnistine (Table 14). In normal
mice folinic acid has a protective effect, whereas
folic does not. Neither folic nor folinic appears to
be involved in the anti-P-1534 activity of vincnis
tine. Indeed, rather than reversing the antitumor
effect both compounds given with vincristine ap
pear to give somewhat better antitumor effects
than vincnistine alone (Table 15).
In view of this host protection phenomenon, we
have investigated the effect of vincristine and yin
blastine on both folic acid reductase and fonmyl
tetrahydrofolate synthetase in vitro. The enzymes
were assayed by either a crude homogenate or par
TABLE 15
EFFECT OF F0LIc AND FOLINIC ACIDS ON
VCRTHERAPYOFP.1534

(5 mice/group)
Dosaux(uo/aoX 1X 10)
in one of the animals. Pathological studies after
sacrifice revealed hypoplastic marrow infiltrated
by lymphocytes. The acinan cells of the pancreas
were unevenly depleted of secretion granules.
Widespread atrophy of the mucosa of the large
intestine occurred in one monkey. There were mild
toxic manifestations in the liven and kidney and
interstitial edema and adipose infiltration in the
myocardium of the left ventricle of one monkey.
Eight monkeys (one male and one female per
dosage level) were given repeated doses of vincris
tine intravenously at levels of 0.1 mg/kg/day, 0.@
mg/kg/day, 1 mg/kg twice weekly, and 1 mg/kg
weekly. Genera! body weakness, anorexia, tremors,
diarrhea, and ataxia were noted in all groups. No
deaths were observed at the 0.1 mg/kg/day dosage
level, but death did occur after seven on twelve
doses of 0.@ mg. vincnistine per kg. daily, two or
three doses of 1 mg. vincnistine pen kg. twice week
ly, and five doses of 1 mg. vincnistine per kg.
weekly.
Leukopenia was generally observed. No major
blood biochemical changes were noted except ten
minal blood values from moribund animals and an
occasional elevation of serum transaminase.
Autopsy of these monkeys revealed no charac
tenistic gross findings. Histological studies showed
reduced celluanity of myeloid and lymphoid tissue.

S Saline controls.
tially purified fraction from mouse liver. Folic acid
reductase activity was measured by the procedure
of Werkheimer et a!. (88) and formyl tetrahydro
folate by the method of Rabinowitz and Pricer
(64). Neither alkaloid has any influence on either
of these enzymatic activities in vitro at concentra
tions of drug up to 100 @g/ml.
BIOCHEMICAL AND MECHANISM STUDIES
Since the alkaloids of Vinca rosea comprise a
new class of antitumon agents, it was of consider
able interest to us to study the effect of these com
pounds on various biochemical reaction sequences
CaLL TYPETiua
(MR.)LACTATE
PRODUCTION
@ CONTROL)Aerobic-
AnaerobicS-180
Freund0—2
2—4
4—8
0—2
2-4
4—8100
98
96
99
98
102100
101

98
97
99
96
Cell typeAddition
of VCR
(pg/mI)0,
uptake
(%control)S.180
Freund0
10
50
200
0
10
50
200100
100
98
85
100
101
99
89
Addition
of VLB
(jig/mi)Per
cent
control15
150103

94
CellAddition
of VLB
(jig/mI)@‘
uptake
(%control)Normal
spleen
Leukemic spleen5
(P.1534)
S-iso Ascites cellst0
20.5
41.0
0
20.5
41.0
0
100100
98
102
100
99
95
100
99
JOHNSON et al.—Vinca Alkaloids
1409
with the hope of uncovering information concern
ing their mode of action.
Our initial experiments were devoted to a study
of the effects of vinb!astine and vincristine on the

rate of respiration and glycolysis of tumor cells
under various conditions. S-180 ascites cells were
chiefly employed ; however, some studies were per
formed with Freund ascites cells and also spleen
tissue from normal mice and mice carrying P-1534
leukemia. The results of experiments in this area
are summarized in Tables 16 to 19. Each table
represents the results of at least two separate ex
peniments.
TABLE 16
RESPIRATION OF S-18O AND
FREUND ASCITES CELLs
In intact ascites cells (Table 16), either 8-180
or Freund, concentrations of less than 50 ,@g/ml
vincnistine appear to have no effect on respiration.
However, when the concentration of drug is raised
to very high levels some inhibition appears. Al
though data are not presented for vinbiastine, es
sentially the same results were obtained (10—iS
per cent inhibition with 100 @&g/mi).In Table 17
TABLE 18
GLYCOLYSIS OF S-180 AND FREUND ASCITES
Cr@LLsIN PRESENCE OF YIN
CRISTINE (50 @iG/ML)
Glycolysis measured as lactate production.
Krebs-Ringer bicarbonate buffer containing
0.2% glucose was employed. Incubation was at
370 C. with 5% CO@-95% 02 or 5% CO@-95% N3
as gas phase.
TABLE 19

ANAEROBIC GLYCOLYSISH0M0G-
ENATE OF S-iSO ASCITES CELLs
Respiration measured by standard
Warburg technic over 4-hour period.
Krebs-Ringer phosphate buffer contain
ing 0.1% glucose used with air as atmos
phere.
TABLE 17
RESPIRATION OF S-iSO HOMOGENATE
AND SPLEEN MINCE
Anaerobic glycolysis measured
as lactate production. Medium em
ployed was essentially that of La
Page (48).
it can be seen that vinblastine was without effect
on the respiration of spleen mince from normal or
leukemic animals. In addition, the complete lack
of influence of vinblastine, at a relatively high con
centration, on homogenates of S-iSO ascites cells
is noted. This latter result is in contrast to the re
suits obtainedwith intactcells, since corresponding
ly high levels of vinblastine do give rise to a iO—i5
per cent inhibition of respiration. In Tables 18 and
i9, the inability of vinblastine and vincnistine to
affect the rate of glycolysis as measured by lactate
production is shown.
To summarize our results, it can be said that we
S Spleen from DBA/2 mice. Respiration
measured with mince in Krebs-Ringer phosphate
buffer by standard technic over period of 30

minutes.
t S-180 homogenate prepared in 0.25 M su
crose. Medium employed was essentially that of
Aisenberg (1). Results are similar whether suc
cinate or pyruvate and fumarate are used as sub
strate. 02 uptake measured at intervals over a
1-hour period.
ADDITIONTIME
(uz@.)Spac.
ACTIVITY(DPM/uu)RNA
@DNA0
10@gVCR/ml
20 @gYLB/ml30
90
270
30
90
270
30
90
27016,500
39,600
40,000
15,900
37,200
38,700
14,600
36,300
38,60027,400
68,400

72,400
24,600
69,100
73,300
26,000
66,100
61,600
ADDITIONSpac.
ACTIVITY(DPM/MG)RNADNA0
1 zg YCR/ml
1OOj@gVCR/ml
20 j@gVLB/ml14,800±740
14,600±780
18,300±670
18,800±6904,520±280
4,610 ±280
4,560±230
4,470 ±280
ADDITIONINCUBATION TIME
(MIN.)SPECIFIC
ACTIVITY
(DPM/uo)RNADNA50@gYLB/ml5
25 @gYCR/mlt
10@sgMTX/ml@
Saline30
90
30
90
30
90

30
901,270
1,470
1,380
1,500
1,010
1,280
1,270
1,5902,660
3,290
2,570
3,020
2.090
2,640
2,570
3,120
Cancer Research
Vol. 23, September 1963
1410
observe neither inhibition nor stimulation of nes
pination or glycolysis in the presence of these alka
bids at levels approaching physiological concen
trations. At very high concentration values, we
begin to obtain inhibition. However, we would not
consider this particularly meaningful. Our results
agree with those of Beer (6), who measured the
respiratory quotient of liver from animals treated
with yinblastine, but differ from those of Hunter,'8
who reported that vinbiastine stimulates aerobic
acid production and inhibits nespiration. However,

Hunter employed different cell lines so that a di
rect comparison with our data is not possible.
The pronounced effect of vinblastine on mitosis
led us to studies on the biosynthesis of nucleic
acids. We have measured the incorporation of
fonmate-C'4 and glycine-@-C'4 into the nucleic acids
of 5-180 ascites cells in vitro. In Table @0the results
of a typical experiment of this type are shown. In
general, washed 6-day ascites cells were employed
TABLE 20
FORMATE-C―INCORPORATION
IN S-180 AScITES
The results of an experiment in which animals
were pretreated in vivo with drug is given in Table
@-Twenty-four hours after treatment the ascites
cells were removed and incubated in vitro with
TABLE 21
GLYCINE-2-C'4 INCORPORATION
IN S-180 ASCITES
Incubation conditions: 1 ml packed S.180 ascit.es/15 ml
Krebs-Ringer phosphate buffer containing 0.2% glucose;
0.2 @sc(12.5 ,.@g)glycine-2-C―/ml.
TABLE22
GLYCINE-2-C'4 INCORPORATION IN 8-180 AsCITES
Final concentration: 0.2 @il.packed in 2.5 ml. total volume
Krebs-Ringer bicarbonate buffer containing 3.5 gig. (0.3 iic.)
C'4-formate/ml; 5% CO,-95% N3 gas phase.
with Krebs-Ringer bicarbonate on Krebs-Ringen
phosphate buffer containing Od—0.4pen cent glu
cose.

After 60 minutes incubation at 37°C., proteins
and nucleic acids were precipitated with perchlonic
acid and determined after separation by the meth
od of Scott et a]. (68). The results shown in Table
@0indicate that neither vincristine nor vinblastine
has any measurable effect on the incorporation of
formate into nucleic acid in 5-180 ascites.
Table @ipresents the results of a similar expeni
ment in which incorporation of glycine-@-C'4 was
used as a measure of nucleic acid formation. Again
vinblastine and vincristine demonstrate no appre
ciable effects. In this particular experiment, aeno
bic conditions were employed. However, anaerobic
experiments with glycine-@-C'4 as substrate gave
similar results.
18 Personal communication, Dr. J. C. Hunter, National
Cancer Institute.
a Cells from group of six mice treated 24 hours earlier with
2.5 mg/kg VCR.
t Cellsfromgroupofsixmicetreated24hoursearlierwith
5.0 mg/kg VLB.
@ Cells from group of six mice treated 24 hours earlier with
1.0 mg/kg methotrexate.
Incubation: 0.3 ml. cells (packed) in 2.5 ml. Krebs-Ringer
phosphate buffer containing 0.1 mmole amino acids (all except
glycine), 1 rnmole glucose, and 0.9 @c.(22.5 gig.) glycine-2-C'4.
Incubated at 37°C. in 02 atmosphere.
glycine-@-C'4 under aerobic conditions. As in the
previous experiment, vinblastine and vincnistine
have no effect, whereas methotrexate treatment

depresses the incorporation to a significant extent.
Very recently, however, some preliminary ne
Exper.pg
VLB/miliiliterCounts/mm/mg
proteinA*
Bt10
50
150
0
10
100
0160±7
150±6
124±5
168±8
176±6
166±5
172±6
JOHNSON et al.—Vinca Alkaloids 1411
sults have been obtained with HeLa cells in tissue
culture experiments, lasting over periods of 1 week,
which suggest that vinblastine and possibly also
vincristine may act as inhibitors of de novo syn
thesis of nucleic acids. These studies, if confirmed,
will be reported elsewhere in detail.
The capacity of 8-180 cells to incorporate gly
cine-1-C'4 into protein is presented in Table @3.At
relatively low concentrations of vinblastine, no ap
preciable effect on protein synthesis is noted. At
a higher concentration, however, some inhibition

becomes apparent. In the second half of Table @3,
the influence of vinblastine on the capacity of a
cell-free microsomal system to synthesize protein
was tested. Again no large differences are to be
found between the treated and untreated samples.
In agreement with the results reported by Beer
(6)fornucleicacidsynthesisinregeneratingliver
and Ehnlich ascites, we conclude from our data
that neither vinblastine nor vincnistine influences
the rate of nucleic acid formation of purine syn
thesis in S-180 ascites. Furthermore, we can find
no effect by these agents on protein biosynthesis
at relatively low concentrations of drug. The ne
sults of all of our experiments tend to follow the
pattern on little of no activity in vitro at concen
tnations approaching physiological levels; whereas
we begin to observe inhibition at very high con
centrations. It should be pointed out that our cx
periments reported in this paper were carried out
with chiefly a single line of ascites tumor cells in
vitro. One must consider the possibility that in
vitro activity cannot be equated with in vivoactivi
ty. That is, the injected alkaloid may be modified
in some way by the host to provide an “active
component.―
CLINICAL STUDIES WITH THE
VINCA ALKALOIDS
This clinical discussion of the Vinca alkaloids is
confined to vinblastine sulfate and vincnistine sul
fate, which are the only members of this family of

drugs which at this time have undergone extensive
clinical trial. Before a detailed discussion of yin
blastine and vincnistine individually, we will brief
ly compare and contrast the more salient clinical
features of both.
Both vinblastine and vincnistine have been
found to be useful in the treatment of some types
of human cancer. There is no clinical evidence of
cross-resistance with radiation or with other pres
ently used oncolytic agents (4, 9, ii, 13, 14, @0,
3@,44, 45, 80, 81, 85). Despite only a minor differ
ence in their chemical structures, the clinical ef
fects of vinblastine and vincristine differ consid
erably ; and, surprisingly, there is no clinical cvi
dence of cross-resistance between them (3, 11, 14).
The dosage requirements of vinblastine and yin
cnistine differ markedly (3). Whereas the weekly
intravenous dose for the average solid-tumor pa
tient lies between 0.10 and 0.@0mg/kg of vinbias
tine, the dose of vincnistine is only one-tenth to
one-quarter of this (i.e., 0.01—0.03mg/kg).
Both vinblastine and vincristine are alike in
sparing hemoglobin and platelet production (3, 4,
ii, is, 33, 44, 69, 81, 84, 89). The limiting factor
to dosage with vinblastine is leukopenia (31, SO).
With vincnistine, on the other hand, leukopenia is
uncommon; and when it has occurred its degree
and duration have generally not constituted a seni
ous problem (3, ii, 13, 44, 45, 66, 69).
TABLE 28

INCORPORATION OF GLYCINE-1-C―
INTO PROTEIN in Vitro
S Intact S-18O ascites cells incubated in
Krebs-Ringer bicarbonate buffer at 37°C. for
30 minutes with varying concentrations of VLB
and glycine-1-C'4.
t Microsomal system derived from S-iSO
ascites cells. Glycine-l-C'4 incorporation meas
ured essentially according to the system of
Littlefield and Keller (49).
With very high doses of vinblastine, manifesta
tions interpreted as central nervous system toxici
ty has been encountered (@0,31, 43) ; but this
occurs only rarely, and then only if dosage is in
creased above the leukopenia-producing magni
tude. In contrast, at least some neuromuscular
abnormality has been encountered in most patients
who have been treated with vincnistine, and it is
this problem rather than leukopenia which limits
the magnitude of vincnistine dosage (3). Neuno
muscular toxicity apparently may also involve the
autonomic system. Indeed, severe and even ob
structive constipation may develop with chronic
vincnistine dosage if stool-softening agents and
lubricants are not given to prevent this. Constipa
tion so rarely follows vinbiastine dosage that there
is no need to medicate prophylactically on its ac
count.
In our experience, epilation occurs less frequent
ly with vinblastine than it does with vincnistine

Cancer Research Vol. @23,September 1963
141@
approximately 10 per cent of patients being af
fected by the former and perhaps twice this num
ber by the latter drug.
Both vinblastine and vincnistine have proved
clinically useful in the treatment of Hodgkin's
disease and the other lymphomas including lym
phosarcoma (3, 4, 9, ii, 14, @0).Effectiveness has
been shown not only when given as de novo therapy
but also following development of refractoriness to
alkylating agents and/or radiation. Although it
remains to be determined which of the two drugs
will prove more effective in these situations, it is
the early impression of the Lilly Clinic group that
vinblastine may be better than vincnistine for
Hodgkin's disease, whereas vincristine may be
more useful in lymphosarcoma (3, 4). However,
clinical studies with comparable groups of patients
in the same stages of their illness will have to be
carried out to determine the correctness of these
impressions.
The leukemias constitute another area in which
it is of interest to note apparent differences@ in ne
sponses to the two drugs. Although published pa
pens report remissions with vinblastine in the leu
kemias as relatively few in number and often
incomplete, in some patients the drug has been
reported of definite benefit following development
of resistance to other agents. Warwick et al. (84)

reported four of nine patients with acute stem-cell
leukemia responding to vinblastine. Whitelaw and
Teasdale (89) report eighteen cases of acute blast
cell leukemias treated with vinblastine. Of these,
three of eight patients without previous therapy,
and four of ten resistant to previous therapy, mani
fested favorable response for varying periods of
time. (One of the latter group of four patients re
sponding had simultaneous 6-mencaptopunine then
apy.) Hill and Loch (33) included reports on the
use of vinblastine (often combined with other
agents) in the leukemias and concluded there was
sufficient benefit to warrant further clinical trial.
The Acute Leukemia Cooperative Group B (44,
45) obtained remissions in only one of 2.9 children
with acute leukemia. Details of this study have not
as yet been published.
On the other hand, vincnistine has frequently
been found useful in the leukemias, especially the
acute leukemias of children. Karonet al. (45), using
vincnistine, obtained a complete remission rate in
children with acute leukemia at least as good
as with any of the presently accepted standard
agents (6-mencaptopurine, steroids, on methotnex
ate). Other investigators have verified the useful
ness of this agent in acute leukemia in children.'9
Indeed, some centers now use vincnistine first
in the series of drugs they employ for the treatment
of acute childhood leukemia because vincnistine
is unlikely to exacerbate the neutropenia, anemia,

and thrombocytopenia which may be present in
acute leukemia even before treatment is begun.2°
Overlap in the spectra of activity of vinblastine
and vincnistine has been reported to involve a van
ety of neoplasms including carcinomas of the
breast2' (4, 43, 84), ovary (4, @0,69, 79, 80), kidney
(ii, 33, 79), and cervix (4, 13, 14), chonionepithe
lioma (13, 14, 3@), nhabdomyosarcoma (3, 79),
reticulum-cell sarcoma (14, 33, 84, 89), and neuro
blastoma (ii, 69, 79). As is the case with Hodg
kin's disease, comparison between the effectiveness
of vinblastine and vincnistine against these tumors
must await further trials as well as a better under
standing of administration and dosage.
Vinblastine has been reported to produce tumor
regression in patients with metastatic adenocanci
noma of stomach (4, @0,84), colon (34), and rec
tum (35). Vincristine has thus far not produced
significant benefit in any patient with adenocar
cinoma of the gastrointestinal tract. Vincristine
has been reported to have some useful effect in a
single patient with carcinoma of the prostate (13,
14). Similarly, benefit has been reported with yin
blastine in isolated instances of carcinoma of the
prostate.'9 Obviously more patients with carci
noma of the prostate should be included in trials
of these drugs.
VINBLASTINE SULFATE
Vinblastine is the official generic name for the
alkaloid formerly known as vincaleukoblastine.

This contraction was necessitated by the U.S. Phan
macopeia regulation that a generic name should
not imply a therapeutic use. The earlier name was
felt to suggest usefulness in leukemia.
Vinblastine was made generally available in
1961,with recommendations for its use in Hodg
kin's disease (4, 14, @0,@i,84, 85, 89) and chonio
carcinoma (3@,85). Since then it has been reported
to be active against many other malignancies in
cluding carcinomas of the lung (4, 33, 50, 81, 90),
breast (4, 43, 84), cervix (4), uterus (35), ovary
(4, 80, @0,81, 90), stomach (4, @0,81, 84), colon
(33, 34, 81, 90), rectum (35), and kidney (33, 8i,
90), squamous-cell carcinomas@ (33, 81), nhab
domyosarcoma (89), melanoma (37, 90), seminoma
20 Personal communication, Dr. Sidney Farber, Children's
Hospital and Dr. Emil J. Freireich, National Cancer Institute.
31 Personal communication, Dr. James F. Holland, Roswell
Park Memorial Institute.
13 Personal communication, Dr. Donald Rochlin, University
of California, Med. Ctr., Los Angeles.
19 Data received at the Lilly Laboratories for clinical re
search.
TumorObjective improve
mentSubjective
improve
mentNo
responseBronchus:
Squamous-cell
Adenocarcinoma

Undifferentiated
Oat-cell
Not specified2
1
1
11
1
17
1
5
JOHNSON et al.—Vinca Alkaloids
1413
(84), tenatoma (84), and embryonal-cell carcinoma
of the testis (37), reticulum-cell sarcoma (33, 35,
36, 51, 84, 89) monocytic leukemia (34, 36, 37,
81), erythroleukemia (36), astrocytoma (4, 5@,91),
mycosis fungoides (33, 51).
Delay in recognition of this broaden spectrum is
in part due to the short duration of vinblastine
treatment which was used in earlier trials. In
many of the early trials, vinblastine therapy was
discontinued if no tumor shrinkage had occurred
after@ weeks of dosage. It is now evident that in
the case of solid tumors a response to vinblastine
may occur only after 10—if weeks of therapy (4,
34, 43). This is in contrast to the rapid response
which occurs in cases of Hodgkin's disease and
choniocancinoma, where benefit usually follows the
first few doses of vinblastine.
Not only were some early trials too brief, but

large and toxic doses were often employed. Pro
found leukopenia and other severe side-effects re
sulted from “pressingdosage to toxicity.― We pre
fer a careful incremental approach to dosage. We
believe a patient's dosage with vinblastine may be
established as follows: 1st dose, 0.10 mg/kg; @d
dose, 0.15 mg/kg; 3d dose, 0.@0 mg/kg; 4th dose,
0.@5 mg/kg; and 5th dose, 0.30 mg/kg, etc.
One intravenous dose every 7 days is given ac
cording to this schedule until the leukocyte count
is reduced to approximately 3,000 cells per cu.
mm. The amount of vinbiastine which will produce
this degree of leukopenia varies from patient to
patient. Once the leukopenia-producing dose is es
tablished, doses 1 increment smaller than those
may be administered at weekly intervals for main
tenance. It should be emphasized that, even though
7 days have elapsed, the next dose of vinblastine
should not be given until the white cell count has
returned to at least 4,000 cells per cubic millimeter.
Should benefit occur with a dose below that which
causes leukopenia, there is no need to increase the
size of subsequent doses. In this way many pa
tients may be treated with doses well below the
leukopenia-producing level (4, 84). Our experience
has shown that the maintenance dose for most pa
tients is 0.15 on 0.@0 mg/kg of body weight.
We believe it best to continue with vinblastine
dosage at regular 7-day or 14-day intervals for as
long as a patient will respond (4, 43, 50). Where

successful therapy has been discontinued, patients
have usually relapsed about 3 weeks after the last
dose. Since long-continued dosage at regular 7- to
14-day intervals was only rarely used in the early
trials, it now seems probable that the lack of main
tenance dosage may explain the short duration of
many of the early responses reported.
It is apparent that the spectrum of activity of
a new oncolytic drug cannot be properly evaluated
until the optimal magnitude and duration of dos
age are ascertained. Needless to say this requires
study over a considerable period of time.
An interesting example of what might be accom
plished without significant side effects, with use of
an incremental approach to dosage, is suggested in
the preliminary report of J. C. W. MacFanlane et
al. (50). This group reported that vinblastine had
prolonged the survival of a series of patients with
carcinoma of the bronchus as compared with that
of untreated patients or patients treated by radia
tion and/or nitrogen mustard. Although some of
these patients whose survival was significantly
prolonged exhibited tumor shrinkage, others mere
ly showed an unchanged radiological picture
throughout their several months of treatment.
This suggests that to require prompt demonstra
TABLE 24
RESPONSE OF CARCINOMA OF BRONCHUS TO YINBLASTINE
tion of gross tumor shrinkage in order to decide
that a drug is effective may result in our missing

more subtle long-term tumor-inhibiting and host
supporting effects. We may not always render
a service to the cancer patient when we try as
quickly as possible to produce tumor shrinkage
by pressing the dose of an oncolytic drug to
toxicity. This is not meant to imply that it is
unnecessary to demonstrate objective evidence
of tumor shrinkage with a new drug, but rather
to suggest that a new drug might be found to
prolong survival and inhibit tumor growth if
administered in subtoxic doeses for long periods.
Obviously such trials must be conducted in full nec
ognition of the natural variation in tumor growth
which can occur spontaneously from time to time.
To avoid drawing falselypositive conclusions, such
trials must be statistically controlled and should
include adequate numbers of patients for a suffi
cient length of time.
Carcinoma of the bronchus has also been re
ported as responsive to vinbiastine by the Midwest
Cooperative Chemotherapy Group23 (Table @4).
23 Paper presented at the Nov., 1961, meeting of the Nation
al Cancer Chemotherapy Service Center in Washington, D.C.
Cancer Research
Vol. @23,September 1963
1414
To quote from their papen on this topic : “Thirty
one patients were studied, of whom @1were evalu
ated. All but 4 patients had received prior therapy
with x-ray, alkylating agents, or 5-fluonouracil.

The priming dose of vinblastine ranged from 0.15
to 0.6 (mean 0.33) mg/kg. Twenty patients re
ceived maintenance therapy varying from 0.15 to
@.37(mean .7@) mg/kg over 7 to @6(mean 33)
days. Five patients showed objective improvement
of short duration (@i to 48 days), but one patient
with undifferentiated carcinoma showed marked
regression of a right hilan mass which, on main
tenance therapy, was sustained for @i5days. It is
of interest that four histologic types of carcinoma,
squamous, oat cell, adenocarcinoma, and undif
ferentiated, were included in those with objective
improvement. Three patients obtained some de
gree of pain relief lasting 7, 8, and 67 days.―
In this series the duration of the trial of vinblas
tine therapy varied markedly from patient to
patient and i4 days' trial was considered adequate
for evaluation. In the light of other experiences, it
would appear that this duration of trial was too
brief for the demonstration of a maximal response
rate.
In the first clinical trial of vinblastine by the
Eastern Cooperative Group in Solid Tumor Chem
otherapy (@O),nine patients with carcinoma of the
bronchus were considered evaluable. None of these
was reported to have shown tumor regression. In
this series an inadequate duration of trial was de
fined : “ . when no toxicity occurred and when
the cumulated duration of drug administration and
toxicity was less than @1days.―Again the duration

of trial, though it produced a high percentage of
responses in Hodgkin's disease, was brief for solid
tumors.
Warwick et al. (84) have stated: “Therewas
no definite evidence of a correlation between the
degree of leukopenia and the beneficial effect.
Remissions in Hodgkin's disease have been ob
served in the absence of leukopenia following
therapy.― The toxicity which can follow the use
of excessive initial dosage may only deter an
investigator from continuing the use of a new drug
for an adequate period of trial.
The earliest reference to the use of vinblastine
in the treatment of breast carcinoma was by War
wick et a!. (85) in 1960. They wrote: “ . on the
i@th day the supraclaviculan nodes were no longer
palpable, and induration and thickening of the
skin had disappeared. There was less swelling of
the arm. . . The patient was discharged feeling
well, apart from stiffness in the left shoulder.
The ulcer on the chest wall had completely epi
thelialized. This remission was maintained for 6
weeks.― The dosage used for this patient was a
single course of 0.15 mg/kg of vinblastine, intra
venously, daily for 4 consecutive days. No further
doses were given. In the light of subsequent expe
niences this dosage was excessive and lacked the
maintenance treatment necessary for prolonged
benefit.
In 1961 Barbara Johnston et a!. (43) of St

Luke's Hospital, New York, published on eight
patients with carcinoma of the breast treated with
vinblastine. Six patients were evaluable; and of
these, three were improved both objectively and
subjectively. Dr. Johnston24 has stated that some
of these remissions have now been maintained for
more than 1 year and that her series, now grown
larger, continues to show the same order of respon
siveness.
The dosage schedule employed by Johnston et
al. (43) was : “ . 0.15 mg/kg/day for@ days fol
lowed by 3 days of rest. If no leukopenia was
noted, one or two additional doses were given until
leukopenia occurred. Weekly doses were given
thereafter depending on the w.b.c.―Johnston oh
served that vinblastine treatment for at least 6@
weeks to 3 months was necessary to demonstrate
objective evidence of tumor regression. This is in
agreement with the experiences of Armstrong et al.
(4) (i96@2)in which three of seven patients with
carcinoma of the breast showed both objective and
subjective improvement, in one case only after 3
months of vinblastine treatment. Armstrong et al.
reported less toxicity than did Johnston et al.,
probably because they avoided initial heavy load
ing doses. They followed the incremental approach
to dosage and so avoided leukopenia as much as
possible. The patients of both Johnston et al. and
Armstrong et at. were treated for at least 1@ weeks
and, if they responded, were thereafter given

maintenance dosage indefinitely once every 7 days.
For various reasons not related to drug toxicity,
vinblastine dosage was discontinued in two of the
patients of Armstrong et a!. while the drug was
still effective. Both patients relapsed during the
3d week after their last dose. This again illustrates
the need for continued maintenance dosage for as
long as a patient will respond.
In the Midwest Cooperative Chemotherapy
Group study (1961) of vinblastine, seven patients
with carcinoma of the breast were included, and
none showed any response. Likewise none of eight
evaluable patients with carcinoma of the breast
showed any response in the study (1961) of the
24 Personal communication, Dr. Barbara Johnston, St.
Luke's Hospital, N.Y.C.