A STUDY OF SPONTANEOUS TUMORS O F T H E
MOUSE BY THE TISSUE CULTURE METHOD
MARGARET REED LEWIS
AND
LEONELL C. STRONa
(From the Mount Desert Island Biological Laboratory and the Department
Embryology of the Carnegie Zmtitution of Tashington, und from the
Roscoe B. Jackson Memorial Laboratory, Bar Harbor, Maine)
6f
Descriptions of the cytology and behavior of a number of transplantable tumors of the mouse in tissue cultures have been published, but in only a few instances have similar studies been undertaken on the spontaneous tumors arising in this animal. The wide
range of tumor-susceptible as well as tumor-refractory inbred
strains of mice available at the Roscoe B. Jackson Memorial Laboratory offered an unusual opportunity f o r such observations upon
spontaneous tumors arising in these mice.l
MATERIALA N D METHOD
The investigations extended over a period of two summers
spent at the Maine laboratories. Forty-eight tumor-bearing mice
were used, 7 of which had more than one tumor. Two of the mice
died, and one tumor regressed. I n all, cultures of 52 different tumors were prepared. The tumor-bearing mice were selected from
representatives of several distinct genetic stocks. The majority
of them belonged to inbred stocks ; the others were in hybrids of the
first or second filial generation produced by crossing mice of the
inbred stocks. The tumor mice may be grouped, therefore, into
eight distinct classes as follows :
5. A
6. AF, ( A X D )
7. DF, ( D X A)
4. D
8. F, ( D X B )
These classes will be briefly described (Strong).
(1) Stock C,H: These mice belonged to a relatively inbred
stock produced by brother-to-sister matings for twenty generations. The mice are all wild type, black-agouti. I n this stock
practically all females used for breeding purposes develop spon1. C,H
2. EI
3. L
1 The writers wish to express their thanks to Dr. Clarence C. Little, director of the
Roscoe B. Jackson Memorial Laboratory, and to the members of the laboratory, for their
generous help in this investigation.
72
A STUDY
OB SPONTANEOUS TUMORS OP THE MOUSE
73
taneous tumors of the bread. The tumors are usually medullary
mrcinomata or adenocarcinomata of moderate malignancy. The
mice live an average of fifty-three days after the appearance of
the tumor. Metastases into the lungs occur, but infrequently.
Three females of this strain with spontaneous tumors were used
in the culture work (S58815, 558104, 558814).
(2) Stock E l : This stock is less inbred than any of the others.
Tbe mice are in the eighth generation of brother-to-sister matings.
!he tumors are relatively benign, never infiltrating adjacent tissue,
and never metastasizing. The mawes gradually increase in size
until death occurs. In many instances the weight of the tumor exceeds the weight of the host. One representative of this stock was
used in cultures (555779).
(3) Stock L: These mice are in the fourteenth generation of
inbreeding. The individuals never have carcinoma of the breast,
but both males and females have enlargements of the lymph nodes.
The diagnosis of the pathological condition (by Warthin) was
lymphoblastoma. One individual ( 55843%’) showing enlargements of the superficial lymph nodes was used in tissue culture.
(4) Stock D : This is the highly inbred dilute brown stock
which has been used extensively in studies on cancer. Adenocarcinoma of the breast occurs in more than 80 per cent of all
breeding females. Eleventeen affected mice were used in this
study. Of these, 11 were obtained from Dr. W. S. Murray, 4 from
Dr. J. B. Bittner, and 2 from Dr. L. C. Strong. The mice from
Murray’s laboratory were W76781, W75664, W81138, W80568,
W79777, W80453, W80088, W79778, W79354, W74053, and W81126 ;
those from Bittner were B5495, B5770, B3378 and B4427; while
those from Strong’s colony were 558090 and S54386.
( 5 ) Stock A: These mice are descendents of the Bagg albinos
which have been in Dr. Strong’s laboratory for eleven years.
They are a t present in the fortieth generation of brother-to-sister
matings. Approximately 85 to 90 per cent of all breeding females
develop carcinoma of the breast. Clinically, these A mice give rise
to the most malignant of all the mouse tumors that have been investigated by Strong. Mice with these spontaneous tumors die of
advanced cancer on an average of forty-five days after the tumor
appears. Multiple nodules, as well as metastases into the lungs,
occur. Twelve affected individuals of this stock were used in this
study. Of these, 8 were obtained from Bittner, B4558, 85117,
B5798, B10207, B11119, B11436, B6166 and B5786, and 4 from
Strong, 553378,554012, S60 and 862.
(6 and 7) Stocks BPI am’ DP,: These mice belonged to the first
filial generation from a cross between individuals of the A and the
D stocks. If an A female was mated to a D male, the resulting
74
MARGARET REED LEWIS A N D LEONELL 0. STRONG
15
-
13
-
12
-
14
11
10
I
-
,
,
I
9 -
e 7 -
k
+
5 6 -
g53
@ 4 -
,
8 2 -
s
z 1 -
I
4
progeny were called A F , ; whereas if a D female wm mated to an
A male, the progeny were then called DF,. Spontaneous tumors
occur quite frequently in individuals of each of these F, generations. The tumors are clinically malignant, with frequent metastases into the lungs.
A STUDY OF SPONTANEOUS TUM0R.S O F THE MOUSE
75
Tumors from these two hybrid generations are of importance
in genetic investigations on the nature of cancer. If it could be
demonstrated that peculiarities of tumors were genetic in nature,
then it should be possible to trace the transmission of these genetic
determiners into the hybrid-mice that eventually give rise to spontaneous tumors. The 8 AF, mice used in the culture work were
B5465, B5916, B5254, B5690, B4052, €33746, I315013 and R14465.
The DF, were B820.5, B3493, and R7921. These mice were all received from Bittner.
(8) Stock P, (DX B ) : The three mice from this strain that
were used, F,4649, F,4086, and F,4634, belonged to the second
filial generation of a cross between mice of the dilute brown and
the black C,, stocks. In the dilute browns the incidence of spontaneous breast carcinoma is high, while the C,, blacks are practically immune to all neoplasia. These hybrids were obtained from
Murray’s laboratory.
There was a wide variation in the age at which the tumor appeared and also in the age of the different tumors used (Chart J).
The youngest mouse used was a dilute brown, 193 days old; the
oldest was also a dilute brown, 588 days old. The majority of mice
were less than a year old, the average length of life being 350 days.
They were examined for tumors about once a week. Where the
record shows the age of the tumor as one day, it means that the tumor was used one day after its appearance had been observed, although it may have appeared several days before it was recorded
(Table I ) . The oldest tumor used had been present in an albino
mouse for 66 days; the youngest ones were recorded as present in
a dilute brown mouse for one day (Table I). Sixteen tumors were
used in the first weck of their growth, 14 in the second week, 8 in
the third week, 5 in the fourth week, 1 in the fifth week, 6 in the
sixth week, 1in the seventh week, and 1in the tenth week.
As may be seen in Table 1, there was no marked difference in
the strains used as regards age of the animal or age of the tumors.
The mice of the albino strain were, on the whole, a little older than
those of the other strains.
The tumors selected for this investigation were not open to the
surface. Although they contained some cystic and some necrotic
areas, none of them was infected, as is frequently the case with tumors in the rat that originate from Cysticercus infestation or that
are transplanted. With the exception of L58436, they were all
found in female mice and were located in one of the mammary
glands. The tumors were characteristic for the mouse, and, as a
group, differed quite markedly from the spontaneous tumors of human beings or of other animals. They were all circumscribed, radiating growths, composed largely of epithelial cells, with a delicate
76
MARQARET REED LEWIB AND LEONELL C. STRONG
TABLE
I
Strain of Mouse
Genetic No.
C3H
S 58815
A
S 58104
S 58814
186
205
232
224
213
238
38
8
6
s 54012
286
305
'I
11
s 53378
326
447
447
335
454
455
19
19
9
7
8
4558
6166
5786
5798
10207
272
359
287
309
305
501
338
371
301
330
326
537
11
'I
I1
B 11436
B 11119
476
500
519
522
W 76781
361
366
1'
16
I'
8
S
62
60
I3 5117
B
B
B
B
B
D
Age of Mouse Age of Mouse
Tumor Noticed Cultures Made Age of Tumor
I1
W 75664
W 81138
W 80568
66
12
14
21
21
36
36
43
22
5
5
7
1
1
4
13
12
2
26
370
192
216
257
263
273
308
294 Died at
402
586
377
193
217
26 1
276
285
310
320
409
588
8 54386
210
294
213
299
3
5
B
B
B
B
307
286
481 Died at
367
332
309
493
394
25
23
12
27
w 79777
W 80453
w 80088
w 79354
w 79778
W 74053
W 81126
S 58080
5495
5770
3378
4427
7
2
capsule made up of several layers of cells (Fig. 2). The tumors
were profusely supplied with blood vessels but had little stroma
and almost no fibrous tissue. Most of them were hematophorous
and contained large lakes and sinuses of circulating blood (Fig. 3).
Some were hemorrhagic and contained regions of freshly extravasated blood and of old blood clots. I n all of the tumors there were
regions of closely packed white epithelial cells, sometimes growing
as a spongy mass within the bloody part of the tumor, again as
firmer white areas occupying a considerable portion of the growth.
A few of the tumors mere composed almost entirely of the compact
77
A STUDY OF SPONTANEOUS TUMORS OF THE MOZTSE
TABLEI (Continued)
Strain of Mousr
Cknrtic No.
A It‘,
B 5916
I3 5254
11
Age of Mouse Age of Mouse
Tumor Noticed Cultures Made Age of Tumor
250
314
29 1
325
41
11
11
16
21
8
10
21
11
11
II
1I
282
404
27 1
426
403
439
298
425
279
436
424
450
1‘
I1
446
486
I I
It
B 8205
B 7921
233
530
Itegressed
565
35
EI
s 55779
284
290
6
L
8 58438
257
262
5
W 4086
391
365
394
371
3
6
6
17
D 5690
I3 4052
B 5465
B 3746
B 15013
B 14465
II
D 141‘
I3 3493
‘I
DBF2
w 4649
I1
11’ 4634
11
I1
365
382
40
40
RevumE of Table: Forty-eight tumor bearing mice were received. Seven had two tumors,
making a total of 55 tumors. One tumor regressed and the mouse wm returned alive. Two
mice died and their tumors were not used. Forty-five mice bearing 52 tumors were used for
cultures. One tumor proved to be an enlarged lymph node. One tumor (D X BF24649)
was a carcinoma containing some cartilage and bone and the other 50 tumors were of the
adenocarcinoma type.
cancerous tissue. The masses of epithelial cells were soft and, in
the few cases where there was considerable stroma, this also was
soft-and pliable, so that none of these growths were fibrous or hard,
as is often characteristic of spontaneous tumors of human beings
or of other animals.
FIXED
AND STAINED
SECTIONS
A small piece of the more solid part of the tumor was removed
for tissue cultures, after whirh tlic remainder was fixed in Helly’s
solution o r 10 per cent formalin. Sections were cut and stained
with hematoxylin a i d eosin. Stained preparations were made of
42 of the tumors used.
All but two of these tumors, T) X B F,4649 and LS58438, exhibited what may be termed an adeiiocarcinoma type of growth,
with some regions made up of ramifying masses of cuboidal cells
and others of atypical tnbular epithelial cells. I n some regions of
certain of the tumors the tubular cpithelial growth resembled the
78
MARQARET -ED
LEWIS AND LEONELL C. STRONG
papillary form, in others the growth approached a type of only
slightly irregular glandular tissue (Fig. 4), while in other tumors
the ramifying masses of cells were more solid and approached the
type of growth that has been termed medullary cancer (Fig. 5).
All of these tumors had relatively little stroma; in places the
irregular tubular growth seemed to be separated only by the endothelium of the intermediate blood vessels.
I n many instances the epithelial cells mere rather small, and
closely packed together, but not greatly changed from the normal
mammary gland cells. They were, however, slightly larger, with
larger nuclei containing a little more nucleolar material, a slightly
crumpled nuclear wall, and occupying a greater proportion of tho
cell than normally. On the other hand in some tumors, as for instance A62 and A53378, the cells were quite large and contained
large, somewhat granular nuclei with heavily marked, crumpled
walls and enlarged nucleoli. The stained preparations show that
many dividing cells were present in the epithelium, but only a few
were found among the stroma or in the capsule. Many of the dividing epithelial cells had nn increased number of chromosomes
with correspondingly larger mitotic figures. None of the tumors
was characterized by the presence of a definite abnormal number
of chromosomes, such as tetraploid described by Lewis and Lockwood (21), although most of them contained many cells with
double the normal number of chromosomes. From the sections it
was difficult to classify the tumors as more or less malignant
growths according to the strain of mice in which they were found.
While all of the tumors were encapsulated, some of them showed
infiltrations of the epithelial cells into the tissue of the capsule,
with the formation of a new capsule around them. The majority
of the tumors were used within the Arst or second week of their
growth, which may account for the fact that in only one of the mice
did a metastasis develop. This was found as a minute nodule in
the lung.
The two exceptions to this general description of the cancerous
growths occurred in LS58436 and I3 X B F,4649. The tissue mass
removed from LS58436 for culture purposes proved to be an enlarged lymph node. D X B F,4649 (Fig. 6) was a large, soft tumor composed of large round and somewhat spindle-shaped cells,
and resembled certain so-called medullary carcinomata, particularly Walker No. 256. The tumor contained little fibrous tissue,
but there were some fragments of bone and cartilage in it. It occurred as a large white growth just beneath the skin in the vaginal
region, and was separated from the abdominal wall and from the
skin by a thin capsule. Adjoining it, in the region of the lymph
node of the groin, was a small, soft, bloody tumor, composed of
A STUDY OF SPONTANEOUS TUMORS OF THE MOUSE
79
FEWTVMORS STVDlED BY THE TISSUECULTURE METHOD
( X 200)
Fig. 1, stock D, shows a region with papillary type of growth; Fig. 2, stock EI, the
capsule; Fig. 3, stock C,H, cystic areas and soine blood spaces; Pig. 4, stock AF,, tubular
growth; Fig. 5, stock A, a more malignant type of growth with some stroma; Fig. 6,
the hybrid D X B F., a type of carcinoma different from the others.
FIG$.
1-6.
SECTIOSS OF A
cells of the same type, which may have been an infiltration from
the larger tumor, although the two were no longer connected.
The more or less cancerous nature of the different tumors
studied could not always be attributed to the period of growth of
the tumor. One of the most malignant of the tumors had been
noticed for only eight days, another for eleven days ; while a tumor
80
MAROARET REED LEWIS AND LEONELL 0. STRONG
present in the same strain of mice ( A s , ) for forty-one days was
not as malignant in appearance as either of these. The least malignant tumor used had been present in a mouse of the dilute
brown strain for twenty-seven days.
TAB^ I1
-1- 7
8-14
16-21
Age of mice from which tumors of given number of days growth were taken
193 200-250 251-300 301-350 351-400 401450 451-500 601-650 651-800
1
3
1
4
1
5
4
3
3
1
22-28
0
1
1
1
1
3
2
1
1
1
1
1
29-35
36-42
43-49
86
1
1
-
2
2
1
4
4
1
I1
~~~
9
14
9
0
2
16 tumors were in let week of
14 tumors were in 2nd week of
8 tumors were in 3rd week of
5 tumors were in 4th week of
1 tumor waa in 5th week of
6 tumors were in 6th week of
1 tumor was in 7th week of
1 tumor waa in 10th week of
growth.
growth.
growth.
growth.
growth.
growth.
growth.
growth.
A number of these tumors were comparatively benign, aa is indicated by the length of
time preeent aa compared with the span of life of the host.
TISSUECULTURES
Cultures were prepared by the usual hanging drop method.
Pieces of tumor, free from capsule and from bloody areas, were
explanted into various media, some into chicken plasma, some into
mouse plasmrs, and borne into a mixture of the two. They were
incubated at 38” C.
About 800 of the cultures, showing typical growths, were selected at various times and fixed and stained for further study.
Various fixatives were used, including methyl alcohol, absolute
alcohol, alcohol sublimate, Zenker ’s fluid without acid, Zenker ’s
with one per cent up to 5 per cent glacial acetic acid, Hells’s solution, silver nitrate and Tellyesniczky’s fluid. The stains used
were Harris ’a hematoxylin, hematoxylin with various counterstains such as eosin, acid fuchsin, orange G, light green and safranine, gentian violet, Gram’s stain, methyl green and acid fuchsin,
methyl blue and eosin, Van Gieson’s stain, Wright’s blood stain,
and Oiemsa’s stain.
A STUDY OF SF'ONTANECOUS TUMORS OF THE MOUSE
81
Fischer (11) studied a large series of primary spontaneous
carcinomata of the mouse in tissue cultures but found that these
tumors grew poorly and that the inoculation into normal mice, not
only of the tissue cultures which grew but also of pieces o€ the
spontaneous tumors, failed to produce tumors in the inoculated animals. I n our investigation all of the 51 tumors studied grew well
in tissue cultures; in fact, most of these tumors grew quite as well
as some of the mouse transplantable tumors, as for instance M63
and CRF18O.
A n abundant growth of stroma and sometimes a fairly good
growth of epithelial cells took place in the cultures in mouse plasma
and in the mixtures of mouse and chicken plasma. However, as
these cultures rapidly became liquefied and frequently failed to
exhibit a growth of cancer cells, only the cultures made in chicken
plasma have been considered in this publication.
I n the chicken plasma the tumors almost always exhibited an
extensive growth of epithelial cells free from any growth of stroma
cells. However, in a few of the tumors that contained more stroma
than usual it was not always possible to avoid including these
stimulated cells in selecting the pieces for explantation. I n such
cultures an extensive growth of stroma cells occurred within fortyeight hours. This growth began quite as soon as that of the malignant cells, increased as rapidly, and remained in a healthy condition after the malignant cells had degenerated. The growth of
these cells was quite different from that exhibited by the stroma or
connective tissue taken from a normal region of the same mouse.
The normal stroma, as, for instance, that from a mammary gland
or from a lymph node, did not begin to grow until after one or two
days ; the cells multiplied slowly, attained an extensive growth
only after seven to ten days, and continued to grow for several
weeks. The normal cells of the mouse grew well in mouse plasma
but only moderately in chicken plasma, while the stimulated stroma
grew as abundantly in one medium as in the other, and in cultures
of a few of the tumors it grew even more extensively in the chicken
plasma than did the malignant epithelium in the same culture. It
could not be said that the epithelial cells outgrew the stroma cells
nor that the stroma cells outgrew the cancer cells, but rather that
the type of cell that predominated in a given culture depended
upon the piece of tumor selected for explantation and upon the tumor from which it was taken.
I n most of the cultures the growth took the form of membranes
of epithelial cells without any growth of stroma cells (Figs. 8, 9,
10, and 11). Except in the case of D X B F,4649, the type of
growth exhibited by the malignant cells of these mammary gland
tumors in tissue cultures was the same. The epithelial cells grew
82
MARGARET REED LEWIS AND LEONELL 0. STRONG
in the form of thin, wiclespread membranes, usually only one or
two cells thick. There were often two membranes, one growing
on the cover glass from the upper surface of the esplant and the
other proliferating from the lower part of the piece out along the
surface of the hanging drop. Occasionally there were finger-like
and narrow membranes stretching out a little way from the explant between the two larger membranes. These membranous
growths remained in good condition for only three or four days,
after which proliferation of the cells ceased and the growth degenerated within two or three daya.
The epithelial membranes grew out with the cells closely joined
together (Fig. 13). The intercellular boundaries were difficult to
determine, but when stained with silver nitrate they became sharp
brown lines, showing that the cells were clearly separated from
one another (Fig. 12). The edges of the membranes were quite
irregular unless the membrane had contracted back as it frequently
did. The cells forming the edge had broad, irregular processes
ending in wavy filamentous sheets, somewhat like those of the
macrophages.
In a few instances the growth of epithelium resembled that lrorn
the normal mammary gland, and in such growths the cells contained large fat globules (Fig. 13). Many of the membranes, however, were free from these globules and resembled, in general, the
usual growth of epithelium i n vitro except that the cells degenerated within a few days, while the normal cells lived ten days to two
weeks. As may be seen in Figures 14 to 39 in the epithelium proliferated from the tumors the nuclei were somewhat larger and the
cytoplasm less in proportion to the size of the cell. The nuclear
membranes were slightly more marked and the nuclei a little folded
and irregular ; the nuclei were more granular and contained more
nucleolar material than those of normal cells. The chromatin of
the resting cells was slightly increased in amount, but not as much
as one would expect considering that many of these cells had more
than the normal number of chromosomes. I n a few of the growths
the nucleoli were enlarged, but in none did they assume the large
size and brilliant non-chromatin staining that have been observed
in some of the rat carcinomata, particularly in the Walker adenocarcinoma 76 and medullary carcinoma 256 (Lewis, 17). On the
whole, the cells forming the growths from the spontaneous mouse
tumors did not exhibit as great cytological abnormalities as did
those from many of the spontaneous human and rat tumors previously studied.
The most outstanding characteristic of cancerous growth, both
in vivo and i 9 ~vitro, is not alone the uncontrolled proliferation of
the cells, but also the rapidity with which many of these vigorously
growing cells die here and ther? throughout the membrane, while
83
A STUDY OF SPONTANEOUS TUMORS OF THE MOUSE
TOTALCULTURES OF TUMORSFRON DIFFERENT
STIlAINS OF MICE, SHOWING
ABOUT THE SIZE AND MANNER
OF GROWTH OF TITE EPITHELIAL
MENBUANES
24)
FIGS.7-11.
(x
Figs. 7, 8, 9, 10, and 11 show cultures of tumors from micc of the D X B F,, AF,,
D, EI, and C,H, respectively.
FIGS.12-13. IIIGHER nfAGNIFICATION OF THE MEMBRANE FROM CULTURES OF T U MORS OF STOCKS DF, AND A RESPECTIVELY,
WITH TIIE SILVER NITRATE STAIN AND WITH
THlP USUAL SPAIN
( X 480)
in cultures of normal tissue the cells grow f o r periods of ten days
to two weeks and remain in good condition until death of the culture takes place. The degenerated cells were rapidly ingested by
the surrounding tumor cells or by macrophages.
The results from the study of the many tissue cultures prepared
84
MARGARET REED LEWIS AND LEONELL C. STRONG
from these tumors showed that the type of growth exhibited was
similar except in two of the tumors. One of these was the LS58436,
which grew as a stimulated lymph node with extensive outgrowth
of stroma and abundant proliferation of mononuclear cells. These
cultures, prepared in chicken plasma, exhibited greatly stimulated
growth, which in some instances was more extensive than that generally shown by the normal lymph node growing in autoplasma.
The other was D X B F24649. The cultures of this tumor had
large outgrowths of stroma accompanying the growth of epithelial
cells (Fig. 7 ) . These malignant epithelial cells did not grow in
the form of the membrane characteristic of the growth in vitro of
adenocarcinomata. They were larger and migrated out more o r
less separately, although in places they were loosely joined together into sheets. Some of the cells contained an increased number of chromosomes and some exhibited abnormal mitotic figures.
All of the other 50 tumors exhibited growth composed of membranes of epithelial cells, usually free from any growth of stroma
cells but in a few instances accompanied by a more or less luxuriant
outgrowth of stroma cells.
Some of the D X B F, tumors exhibited in cultures the type of
growth that is usually shown by glandular tissue. It was composed of an abundant proliferation of large spindle cells, of macrophages, and of mononuclear cells, with a rather meager growth of
epithelium. Cultures of other F, tumors, however, exhibited the
usual growth of epithelial membranes, without an accompanying
growth of stroma (Fig. 10).
Cultures of the EI tumor contained more growth of stroma
than did those of the CRH,the A, the D, or the F, strains, but many
of them had rather extensive membranes of epithelial cells, and in
some instances the growth was composed of only epithelial membranes without any spindle-shaped cells (Fig. 8). The epithelial
cells of these cultures were only slightly less atypical than those
that grew from some of the tumors of the other strains of mice.
The growth from the tumors of the A, D, C,H, and AF, strains
of mice took the form of extensive membranes of rapidly dividing
epithelial cells (Figs. 9 and l l ) , usually free from stroma cells.
Many of the epithelial cells had an increased number of chromosomes and, except in the cultures of the tumors from the D strain,
there was present an occasioiial abnormal mitotic figure.
The three tumors of the DF, strain each behaved differently.
DF,3493 grew in the form of membranes of epithelial cells with
large fa t globules and resembled the growth from mammary gland
tissue. These cells grew slowly and did not show abnormal mitoses. The cultures of DB’,7921 exhibited growth quite like that
from the more malignant tumors of the A and the AF , strains.
A
STUDY OF SPONTANEOUS TUMORS OF THE MOUSE
85
They had extensive membranes of somewhat atypical epithelial
cells with many mitotic figures, some of which were abnormal.
One cell had five spindles massed together and many times the
normal number of chromosomes. The tumor in one mouse
(DF,8205) regressed, so that cultures could not be made from this
tissue.
MITOSIS
While all but 58 of the 2550 cultures grew, the extent of the
growth and the number of days (3-5) that the cells remained in a
good condition varied greatly with the different tumors. Also, although the membranes that grew around the explant were large,
the growth from some of the tumors exhibited few mitotic figures.
The growths from tumors of all the strains varied greatly in the
number of dividing cells exhibited, some of each strain having
many and others few mitotic figures.
The cells growing in tissue cultures made as hanging drop
preparations become spread out as a thin layer along the cover
glass. In consequence the mitotic figures assume a position parallel to the cover glass and during metaphase the spindles form horizontally along the plane of the cover glass with the equatorial
plate flattened into an ellipsoid. In the thin epithelial cells forming the membranous growths of the cancerous tissue, all the
phases of the mitotic division were clearly depicted (Figs. 15, 16,
26, 27, and 28). I n these cells there were no so-called polar views
of the equatorial plate, although the late prophase and beginning
metaphase often assumed a ring form. All the chromosomes were
present, not scattered through one or more sections, and although
they were often overlapping or massed in the metaphase, they
could usually be counted in the late prophase (Figs. 14, 25, and
26). Also, it was possible to determine the approximate number
of chromosomes in the other stages of division by the general size
of the mitotic figure.
The cells of these spontaneous tumors did not have enlarged
centrospheres (Plimmer’s body; see Lewis, 22) nor aster rays in
the living condition, nor was it possible to produce centrospheres
or aster rays by gelation of the cells with acid or other coagulative
agents. Neither was it possible to demonstrate spindles or to show
the position of the spindle in the prophase. The spindle area became clearly evident in the living cells at metaphase. It did not
contain spindle fibers and could be seen only as a slightly clearer
zone within the granular cytoplasm. I n the fixed preparations the
spindle could usually be seen in the metaphase. It did not contain
fibrils unless care had been taken to gelate it. The spindle fibrils
could be rather clearly observed when an acid fixative was selected;
86
MARGARET REED LEWIS AND LEONELL 0. STRONQ
however, all of the coagulative fixing solutions used brought about
some distortion of the chromosomes and of the mitotic figures,
The most beautiful preparations which exhibited the spindle area
and the chromosomes most nearly as they were in the living state,
were obtained by fixing the cultures in Helly’s solution for fifteen
minutes and then passing them directly into weak alcohol without
washing them in water.
Although theoretically the chromosomes are supposed to become split longitudinally as they form in the early prophase, so
that the individual chromosomes of the prophase and of the metaphase are in reality pairs of chromosomes rather than single ones,
this phenomenon is seldom evident during the division of somatic
cells. Lewis (18), however, found that the malignant cells of
the tumors studied in tissue cultures showed a split condition of
the chromosomes previous to the metaphase, and this was also
found to be true in the growths of these spontaneous mouse tumors
(Figs. 14,27,and 36).
The growths from the 51 tumors taken from seven strains of
mice exhibited chromosomes that were so widely split during prophase and metaphase that they often appeared as pairs of chromosomes of the same shape and size. The longitudinal division of
the chromosomes was sometimes evident in the early prophase,
usually in the late prophase, and practically always in the metaphase. At the beginning of anaphase the halves of the chromosomes became separated, .one half going to each anaphase plate.
The split condition of the chromosomes was evident in mitotic figures in which the number of chromosomes was increased, as well as
those that had the normal number (Figs. 27 and 34). The chromosomes that lagged behind during the migration to form the metaphase were split (Fig. 36), but those that lagged in the separation
to form the anaphase, telophase, and daughter nuclei were usually
separated and appeared singly rather than split (Fig. 39).
Although many of these tumors, both in sections and in cultures, exhibited more dividing cells with an increased than with a
normal number of chromosomes, there was no given number of
chromosomes characteristic for any one tumor or for the tumors
of any one strain of mice. Most of the membranes growing in
cultures contained some cells of both types (Figs. 30-39 and 2530). In no instance were mitotic fignres found in which the
chromosomes were increased in size rather than in number.
Where more than the normal number were present, the whole
nuclear picture was usually correspondingly larger. These cancerous cells growing in tissue cultures showed rather clearly that
the size of the resting nucleus of a given kind of cell under given
conditions depended upon the number of chromosomes that took
A STUDY OF hlPONTANEOUS TUMORS OF THE MOUSE
87
ABNORMAL
TYPESOF MITOB~S
FOUND IN CULTURES OF TUMORS
DIFFEXENT
STRAINS OF MICE( X 600)
Fig. 14, stock C,H, shows a double ring of chromosomes; Fig. 15, stock A, a double
ring in which each ring had an increased nuniber of chromosomes; Fig. 16, stock AF,, a
prophase with a n increased number of chromosomes. Figs. 17, 18, and 19, each from a
different tumor from stock A, show dividing cells, each with three spindles arranged in
a tripolar figure. Figs. 20 and 21, from two tumors of stock A, show t h e telophase of
two cells, in each of which the daughter cells rereive fewer chromosomes than were present in the mother cell. R g . 22, stock A, shows those cliromosomes t h a t do not become
arranged on the spindle and a r e known as aberr:int cliroinosomes. Fig. 23, stock A,
shows two dividing cells, each contaiuing four spindles with four times the normal number of chromosomes. Fig. 24, from a different tumor of stock A, shows the telophase of
a dividing cell similar t o the ones i n Fig. 23. The four daughter cells each receive
chromosomes from two spindles, and, therefore, have twice the normal number of
chromosomes or one-half the number that the mother cell had.
6
FIQB.14-24.
SOME
FROM
88
MARQARET REED LEWIS AND LEONELL 0. STRONQ
part in its formation, and that the size of its later mitotic structures, prophase, metaphase, spindle, anaphase, and telophase, corresponded to some extent to the number of chromosomes involved
(Figs. 15, 30,32, and 38).
During the early prophase the chromosomes began to appear
within the nuclear areas, where they became concentrated into definite chromatin bodies. They remained in this region for some
time after the nuclear wall had disappeared. During the prophase, as the chromosomes began to take on the appearance of
long and short rods and threads, they frequently became radially
arranged, more or lees into a ring. In some instances the ring was
thicker at one side, where it was composed of more-than one layer
of chromosomes, but thinned out into a single layer on the opposite
side (Figs. 25 and 26). There was often a space free from chromosomes just opposite the thickest part of the ring. The chromosomes at this stage extended out beyond the original nuclear area.
However, they never became scattered out into the cytoplasm except under the abnormal condition known as aberrant chromosomes
(Fig. 22).
Aberrant chromosomes are difficult to explain. They are not
present in many tumors but were abundant in the rat sarcoma 338,
described by Lewis and Lewis (20). They are usually found in
tumors and even normal tissue that has been subjected to radium
(Whitman, 42). Aberrant chromosomes were seldom found in
growth from spontaneous mouse caroinomats unless they had been
exposed to radium (Lewis and Hunt, 19). However, a few abnormal mitoses of this type were observed in the present study.
These chromosomes were sometimes split, sometimes single ; they
were gathered into large or small groups or remained as single
rods, but they were always located away from the zone of the
spindle, as though they had been propelled away from the center
of mitotic activity rather than drawn toward it. I n some instances
there were only a few aberrant chromosomes in a cell, while in
others there were about as many chromosomes scattered in the
cytoplasm as there were on the spindles taking part in the division
(Fig. 22). The aberrant chromosomes did not take part in the
division, but later, when the daughter groups of chromosomes
developed into daughter nuclei, the aberrant chromosomes also
swelled up and formed nuclear bodies, some large, some small, and
some single chromosomal vesicles (Lewis and Lewis, 20). Some
of the nuclear bodies that arose in this way were quite large and
contained nucleoli. It is probably such swollen chromosomal
vesicles that Levine (14)interpreted as the dissolution of chromosomes.
Soon after the broken ring arrangement of the chromosomeR
A STUDY OF SPONTANEOUS TUMORS OF THE NOUSE
89
FIQS.25-39.
DfVIDINQ CANCER CELLS, SOME WITH TEE N O R M A L NUMBER OF CHROMOMIMES AND SOME WITH AN INCREASED NUMBER
OF CHRQMOBOMEf4 ( X 800)
Except for Fig. 27, which is from stock DF,, Figs. 31 and 35 from AF,, and Fig. 37
from stock C,H, all the mitoses shown are from cultures of tumors of the A stock. Figs.
25-29 show mitotic figures with about the normal number of chromosomes, while Figs.
30-39 show mitotic figures with an increased number of chromosomes. Several of those
figures show the split condition of the chromosomes, but Fig. 27 shows this clearly. Fig.
86 shows the chromosomes lagging in metnphase and Fig. 39 shows two chromosomes delayed in telophase.
took place, the beginning indication of the formation of spindle
material with its apex and centrioles located in the portion of
the ring free from chromosomes could be vaguely determined.
Fischer noted in cultures of Ehrlich’s mouse carcinoma that the
90
MARGARET REED LEWIS AND LEONELL C. STRONG
chromosomes assumed a ring-shaped figure with an empty place in
the ring. He attributed this to the absence of certain chromosomes in the malignant cell. In our culture of the mouse spontaneous tumors it was possible to demonstrate bymeans of acidfixing solutions that the beginning of the formation of the spindle
was taking place in the region of the ring free from the chromosomes. Also such empty spaces among the chromosomes occurred
in prophases that had more than the normal number and the normal number, as well as in those that had less than the normal number of chromosomes.
The beginning formation of the spindle became clearer when
the chromosome began to move toward a more equatorial region.
From this time on the spindle gradually could be more dehitelp
determined, usually with one pole sharper and clearer than the
other, until a distinct spindle became present in the metaphase.
M. R. Lewis (15) demonstrated the reversible gelation of living
cells by means of acid solutions (pH 4.4 to 4.6) and showed that
during the gelation of cells fibrils appeared in the spindle. I n
these tumor cells the spindle could be gelated readily in all the
stages of division where migration of the chromosomes toward the
equatorial region had begun. As the chromosomes migrated, they
became arranged in a position at a slight angle to or parallel with
the surface of the spindle. During metaphase the individual
chromosomes sometimes shifted hack and forth across the equatorial plate (as is shown in the Lewis (W. H.) cinema of dividing
tumor cells), a little toward one pole, then a little toward the other
pole, until all of them became arranged in a more or less straight
band along the equatorial region, after which the two portions of
each split chromosome moved apart toward opposite poles. The
size of the spindle formed depended largely upon the size of the
nucleus from which it arose and the number of chromosomes present. When the cell had twice the normal number of chromosomes,
the spindle was not only broader but also considerably longer than
usual (Figs. 15 and 36). Some cells had two (Figs. 14 and 31),
and one cell had three (Fig. 15) rings of chromosomes with the
thicker region of the rings overlapping. These cells had more
than the normal number of chromosomes and probably developed
more than one spindle.
Most of the tumors from A, D, C,H, AF,, and DB’, strains of
mice exhibited, both ia vivo and in vitro, many cells with more than
the normal number of chromosomes. Such abnormal mitosis usually had a bipolar spindle arrangement (Fig. 28). Certain of the
tumors, especially those from the A, C,H, and the AF, strains, exhibited a number of abnormal spindle figures. On the other hand,
only one abnormal spindle, a tripolar spindle, was found in the cultures of tumors from the D strain.
A STUDY OF SPONTANEOUS TUMOBS
OF THB MOUSE
91
The abnormal arrangements of the spindles were of many kinds
(Yigs. 17 to 24). Usually the spindle itself was difficult to see in
the stained preparations, and its position had to be determined
from the arrangement of the chromosomes. Sometimes two spindles lying parallel were present in a cell, so that the chromosomes
were continuous across the two spindles, but giving rise to two binucleate cells, each nucleus with the normal number of chromosomes. Sometimes the spindles Bormed a V, so that the chromosomal equatorial plate was continuous but bent and the anaphase
showed one double group of chromosomes passing to the junction
of the two limbs of the V and one single group passing to each end
of the limbs of the V. This resulted in three nuclei, one with twice
the norm91 number of chromosomes and two each with half that
number.
I n some instances one spindle was double in size and the other
one was of normal size (Fig. 37), but they had one pole in common.
This resulted in one nucleus with three times the normal number
of chromosomes and two others, one with a n increased and one with
a normal number of chromosomes. I n some cells three spindles
were present. These might be equal in size, or again one was large
and two were small or two were large and one was small. The
three spindles usually formed a tripolar figure with the chromosomes arranged in the form of a Y, either with limbs of equal size
forming three nuclei each with twice the number of chromosomes
(Fig. 20), or with one long limb and two short ones (Fig. 18) resulting in one nucleus with double the number of chromosomes and
two with three times the number, or with two long limbs .and one
short one (Fig. 17), giving rise to one nucleus with four times and
two with three times the normal number of chromosomes. Sometimes there were four spindles (Fig. 23), resulting in four nuclei
each with twice the normal number of chromosomes (Fig. 24).
Usually the cells with three spindles and those with four spindles
divided into three and into four daughter cells, but in one cell
which probably had four spindles the four nuclei became clumped
into a cross forming a single nucleus. I n some cells several spindles, arranged irregularly, were present.
One of the well known characteristics of malignant growths is
the appearance of abnormal figures of mitosis and abnormal numbers of chromosomes during the division of the cells (Levine, 14).
Most of the abnormal mitotic figures that arise in tumor growth
seem to indicate that at some previous time there occurred a duplication of some or of all of the chromosomes of the cell. Boveri (7)
advanced. the idea that more than the normal number of spindles
and of chromosomes had been introduced into a cell possibly by the
entrance of more than one spermatozoon into the egg, and that later
92
MARGARET REED LEWIS AND LEONELL C. STRONG
divisions of such a cell led to the duplication, in a given cell, of the
proper combination of those chromosomes necessary to produce a
particular type of malignant cell. The further growth of such a
cell brought about the given type of tumor. Our observations in
tissue cultures, however, show that it is possible to derive most of
the abnormal mitotic figures met with in the tumors from mitosis
of binucleate cells. Binucleate cells occur both in normal and in
malignant tissue, and the investigations of Kozhukhova (13) and
Rosenfeld (29) show that binucleate and potentially double nucleate cells can be produced at will in tissue cultures of normal cells
by simple chemical procedure.
However, any theory that attempts to explain the cancer cell
on the basis of a particular combination of chromosomes ,must take
into account the fact that in many types of tumors various sorts of
abnormal mitotic divisions (Figs. 30 to 39) giving rise to cells with
various numbers of chromosomes, often different even from that of
the mother cell (Figs. 20, 21, and 24), continue to take place
throughout the growth of the tumor, while in other tumors most of
the mitotic divisions may be on the whole quite normal.
Geneticists have hoped that the cancer factor may be found to
be attached to some particular chromosome. Unfortunately investigations have so far failed to demonstrate any one chromosomal or even any one cytological picture common to all or to one
type of malignant growth, In addition to this, there is the tissue
and the species susceptibility to be explained ( Apolant, 1 ; Bullock
and Curtis, 8; Slye, 31; Little, 24).
However, the results of the study, by the tissue culture method,
of the spontaneous tumors of the carefully inbred strains of mice
of the colonies of the members of the Roscoe B. Jackson Memorial
Laboratory, indicate that with one exception (Figs. 6 and 7),
which occurred in a hybrid mouse, the mutant cell that forms the
cancerous growth had become changed in much the same manner
in all these tumors (Figs. 1 to 11). These results are different
from what we have found in tissue cultures of human tumors and
of the spontaneous tumors of the inbred rats of the Philadelphia
albino strain carried on for a number of years by Dr. George
Walker. In Dr. Walker’s colony there arose many different types
of tumors, some sarcoma, some carcinoma, and some Cysticercus
tumors. Instead of the majority of the tumors being adenocarcinoma of the mammary gland, as they were in the mouse tumors
studied, the prevailing type was a benign tumor of the mammary
gland grading from adenoma to adenofibroma to fibroma. The
spontaneous tumors of man and of the rat behaved differently one
from another not only irt v i m hut also irc uitro.
A STUDY OF SPONTANEOUS TUMORS OF THE MOUSE
93
DISCUSSION
(Strong)
These studies on tissue cultures of mouse carcinoma are suggestive for an analysis of the factors that undoubtedly underlie the
cancer problem. Biologists recognize that both extrinsic and intrinsic factors are involved in all living phenomena. The cancer
cell, as a unit of structure, must obey these same biological laws.
The extensive genetic work on transplantation, initiated by Tyzzer
(41) and Little (24-28), elaborated by Strong (32-39), and continued by Bittner (2-6) and Cloudman (%lo), suggests rather
convincingly that certain genetic or intrinsic determiners control
the physiological behavior of the tumor cell. There must be a
definite intrinsic mechanism within the structure of the cell, else
it could not perpetuate its own specificity during years of transplantation both in viva and in vifro. The meaning of the irregular mitoses is not clear. This process would lead to an irregular
distribution of chromosomes and may be involved in the cancer
process, as Boveri (7) maintained. It is significant, however, that
the-cancer cell can grow and reproduce itself without undergoing
irregular mitosis. I n some tumors the irregularity of the mitotic
process is an exception and not the rule. Thus from the study of
tissue cultures it is evident that peculiarities of chromosome behavior may be not fundamental but only coincident or the result of
the original production of certain types of cancerous tissue. Thus
the intrinsic factors, such as must be assumed to explain the findings of transplantation, may also be called upon to explain the retention of tissue specificity found to persist in tissue cultures.
It is evident that if genetic determiners control the constitution
of individuals that give rise to nmplasia, then by selective inbreeding from different sources one ought to be able to produce different
“constitutional” types, each one of which eventually develops
peculiar tumors. We used tumors from various genetic stocks of
mice and could not detect genetic dissimilarities. This was rather
disappointing, but nevertheless, something of a significant nature
may have been detected. In the first place, all the tumors were
carcinomata of the breast. This condition is a rather variable
quantity-ranging all the way from adenocarcinomn to medullary
carcinoma, depending upon the relative frequency and arrangement of the cancer cells. I n spite of this diversity of histological
structure, we are probably dealing with one condition-a progresgive alteration from gland-like arranged cells to more rapidly
dividing packed sheets of cells. Consequently, even if genetic determiners were a t the basis of the neoplastic condition, the end
product-adenocarcinoma or medullary carcinoma-may not have
94
MARGARET REED LEWIR A N D LEONELL C. STRONG
enough variability to be fixed by the selective mating of the individuals that give rise to cancer. We did not encounter the variability that was found in the rat and in the human material, either
in sections or in tissue cultures. This may be involved in the peculiarities of mouse physiology. If not, one would have to conclude that certain agencies other than genetic determiners were instrumental in the production of the neoplastic state.
It is an interesting observation that many abnormal mitoses
were found in tumors from the Strong albino strain. On the other
hand, abnormal mitoses were exceedingly rare in tumors from the
Little dilute brown strains. These are well established genetic
stocks and it may well be that this difference is genetic. At the
same time, however, it must be kept in mind that the albinos develop the most malignant of all the tumors encountered, the dilute
browns show somewhst less malignancy (as determined by the rate
of growth of the tumor and the survival time of the mouse growing
the spontaneous tumor). One would have to conclude, therefore,
that degrees of malignancy also must be genetic-a point of view
that is not inconsistent with the evidence obtained from several
fields. At the same time, one must be cautious not to overplay .the
r81e of the intrinsic factors in the cancer problem. Cancer is still
an unknown quantity; the solution may lie in any direction.
In closing, we would like to emphasize the presence of dying
cells in vitro. That process seems to be peculiar to cancer both
in vivo and in vitro. Should this finding be further emphasized o r
should it be ignored? It may be that the process by which the
organism rids itself of this great mass of dead material is involved
in the mcchanism that controls the cancer process.
REFEREN~ES
1. APOLANT,
H.: Arb. a. d. Konigl. Inst. f. exp. Therap. Heft I, 19OG.
2 BIITNER,J. J.: Mich. Acad. Sci. Pub. 11: 349, 1929.
J. J.: J. Cancer Research 14: 466, 1930.
3. BITTNER,
4. B I ~ E RJ., J.: Am. J. Cancer 15: 2202, 1931.
5. B I ~ E RJ., J.: Am. J. Cancer 16: 322, 1932.
6. B I ~ E R J.
, J.: Am. J. Cancer 16: 1144, 1932.
7. BOVERI, T.: Zur F r a p der Enstehung maligner Tumoren, O. Fischer, Jena,
1907. Trans. The Origin of Malignant Cells, by Marcella Boveri, Williams
and Wilkins, Baltimore, 1914.
8. BULLOCK,
F. D., AND CURTIS, M. R. : J. Cancer Research 14: 1, 1930.
A. M.:Am. J. Cancer 16: 568, 1932.
9. CLOUDMAN,
A. M.:Genetics 17: 468, 1932.
10. CLOUDMAN,
11. FISCHER,A.: Oewebeziichtung, Rudolph Miiller and Steinicke, Miinchen, 1930.
12. HANCE,R. T. : Biol. Bull. 50: 155, 1926.
Z. A.: Mem. Soc. Nat. Kiew 27: 86, 1928.
13. KOZHUKHOVA,
14. LEXINE, M.:Am. J. Cancer 15: 144, 1931.
15. LEWIB, M. R.: Bull. Johns Hopkins Hosp. 34: 373, 1923.
16. LEWIS,M.R. : Anat. Record 45 : 269,1930.
A STUDY OF SPONTANEOUS TUMORS OF THE MOUSE
95
17. LEWIS, M. R.: Anat. Record 45: 268, 1930.
18. LEWIS,M. R.: Anat. Record 52: 65, 1932.
19. LEWIS,M. R., AND HUNT,B.: Mt. Desert Island Biol. Lab. Annual Report 31,
1932.
20. LEWIS,M. R., A N D LEWIS,W. H.: Am. J. Cancer 16: 1153, 1932.
21. LEWIS,M. R., AND LOCKWOOD,
J.: Bull. Johns Hopkins Hosp. 44: 187, 1929.
22. LEWIS,W. H.: J. Exper. M. 31: 275, 1920.
23. LEWIS,W. H., AND GEY, (3. 0.: Bull. Johns Hopkins Hosp. 34: 369, 1923.
24. LITTLE,
C. C. : Science 40: 904, 1914.
25. L I ~ EC., C.: J. Exper. Zool. 31: 307, 1920.
26. LITTLE,
C. C. : J. Cancer Research 8: 75, 1924.
27. LITI'LE, C. C., AND STRONG,
L. C.: J. Exper. Zool. 41 : 93, 1924.
28. LITTLE,C. C., AND TYZZER,E. E.: J. Med. Research 33 : 393, 1916.
29. ROSENFELD,
M.: Arch. f. exper. Zellforsch. 12: 570, 1932.
30. ROSENFELD,
M.: Anat. Record 52: 33, 1932.
31. SLYE,MAUD: J. Cancer Research 8: 240, 1924.
32. STRONG,
L. C.: J. Exper. Zool. 36: 67, 1922.
33. STRONG,
L. C.: J. Exper. Med. 39: 447, 1924.
34. STRONG,
L. C.: J. Exper. Med. 43: 713, 1926.
35. STRONG,
L. C.: Genetics 11: 294, 1926.
36. STRONG,
L. C.: Amer. Nat. 60: 201, 1926.
37. STRONG,
L. C.: J. Exper. Zool. 45: 231, 1926.
38. STRONG,L. C.: J. Cancer Research 10: 219, 1926.
39. STRONG,
L. C. : J. Cancer Research 13 : 103, 1929.
40. STRONG,
L. C., AND CLOUDNAN,
A. M.: Mich. Aoad. Sci. Pub. 11: 1929.
41. TYZZER,E. E.: J. Med. Research 16: 519, 1909.
42. WHITMAN,
W. G.: Am. J. Cancer 17: 932, 1933.