CANCE R RESEARCH Experimental Brain Tumors

CANCE R RESEARCH A MONTHLY JOURNAL OF ARTICLES AND ABSTRACTS REPORTING CANCER RESEARCH

VOLUME 1 DECEMBER, 1941 NUMBER 12

Experimental Brain Tumors I. Tumors Produced with Methylcholanthrene* H. M. Zimmerman, M.D., and Hildegarde Arnold, M.D.

(Frorn the Laboratory o/Pathology, Yale University School o/ Medicine, New Haven, Conn.) (Received for publication October 9, I941)

Efforts to produce intracranial neoplasia by various of 4 in pyrex glass jars having wire mesh covers. The chemical carcinogens have been attended with scant jars were sterilized weekly. Each group of animals was success prior to the work of Seligman and Shear (2). inspected at least twice daily for evidences of tumor By intracerebral implantation of pellets of 2o-methyl- development. cholanthrene, these workers produced ii gliomas and The diet consisted of Purina Fox Chow and oats. 2 fibrosarcomas in a series of 2o male mice of the C3H This and tap water were available to the animals at all strain. Seligman and Shear reported also successful times ad libitum. subcutaneous transplantation of several of these tumors, Carcinogen.--The carcinogen employed for intra- one of which was stated to be a glioma. cranial implantation was 2o-methylcholanthrene (Hoff- Utilizing the same technic, the present writers re- man-LaRoche, Inc., Nutley, N. J.) which was purified ported a preliminary experiment (3) in which they by chromatographic adsorption on A120:~.1 The speci- found u6 intracranial tumors in 51 C3H mice. These men used had a corrected melting point of i79.8- tumors, occurring during the first io months of the I8o.4 ~ C. Cylindrical pellets of this hydrocarbon were experiment, consisted of oligodendroglioma, glioblas- prepared with a diameter of about I mm. and a length toma multiforme, medulloblastoma, unclassified gli- of about i. 5 mm., the average weight of each pellet oma, and meningeal sarcoma. being i. 5 mgm. More recently Peers (1) implanted cholesterol pellets Operation.--Anesthesia was accomplished by the containing lO per cent methylcholanthrene in the subcutaneous injection of 0.25 cc. of a solution con- brains of 99 mice, of which 87 survived into the tumor- taining ioo mgm. nembutal in x5 cc. of o. 9 per cent bearing period. In all, 32 intracranial tumors were sodium chloride solution. The top of the head was produced--i 7 sarcomas and 15 gliomas. shaved and washed with 7 ~ per cent alcohol. For intra- These observations were significant in that they sug- cerebral and subdural implantations a right paramed- gested the future possibility of studying the incidence, Jan incision 5 mm. in length was made in the skin; the histiogenesis, and growth behavior of experimentally periosteum was scraped off the right parietal bone; a induced primary brain tumors. From such a study it hole about 2 ram. in diameter and anterior to the oc- was felt that certain deductions could be drawn regard- cipital suture was made in this bone with a dental burr. ing the human counterparts of these neoplasms. The For intracerebellar implantations the scalp incision was present investigation was thus undertaken for a three- made in the midline over the occipital bone and the fold purpose; namely, the determination of the inci- burr hole in the area between the occipital suture and dence, the histiogenetic origin, and the growth be- the attachments of the occipitalis muscles. Muscle havior of experimental brain tumors. bleeding, which sometimes occurred in this location as the result of trauma, was readily controlled with hot MATERIALS AND METHOD wet sponges. With fine forceps the pellets were pushed through the craniotomy opening and dura for about Care o[ animals.--All the animals employed in these 3 mm. into the right parietal lobe subcortex or into the experiments were male mice of the C3H strain between cerebellum. Those intended for subdural implantation 3 and 4 months of age. They were housed in groups 1 The purified methylcholanthrene was kindly prepared for us * This investigation was aided by a grant from The Jane Coffin in the laboratory of Dr. M. J. Shear of the United States Public Childs Memorial Fund for Medical Research. Health Service. 919

were left in the longitudinal fissure in contact with the chrome stain and with the Wilder silver carbonate meninges between the two cerebral hemispheres. The method for reticulin. Where it seemed indicated, skin margins were then approximated and a drop of preparations were stained with Heidenhain's iron- collodion was applied over the wound. This never re- hematoxylin and with phosphotungstic acid-hematoxy- quired any further attention, healing occurring lin. Frozen sections impregnated with silver and gold promptly and without infection. Until full recovery salts offered such little additional aid in histologic study from the anesthetic, which usually occurred in from that their preparation was abandoned early. 2 to 3 hours, the animals were kept warm on a padded hot plate which was set at about 4 ~ o C. When recovery RESULTS was complete, the mice were returned to their cages in the specially ventilated and heated animal room. For purposes of convenience in presentation, the data Subcutaneous transplantation o/ tumors.--To study of this study will be presented in two parts; I. the certain phases of the growth behavior of the cerebral results of the intracranial implantation of the methyl- neoplasms more effectively, subcutaneous transplanta- cholanthrene pellets and, II. the subcutaneous trans- tion of a number of these tumors was made into male plantation of the intracranial neoplasms thus induced. and female mice of the C3H strain. The animals used were 2 to 3 months of age and the sexes were repre- I. INTRACRANIAL IMPLANTATION OF METHYLCHOLANTHRENE sented in about equal numbers. Mating was prevented by strict segregation. As a rule, 8 mice received trans- The pellets of methylcholanthrene were implanted plants from each tumor, although in some instances 4 in three different locations in groups of mice as fol- mice were used. Subtransplants were made from the lows: the right parietal subcortex, 57 mice; the cere- subcutaneous growths when the latter attained sizes of i cm. or more in diameter, subtransplantation being TABLE I: INCIDENCE OF BRAIN TUMORS INDUCED WITH METHYL- carried out through 9 to 14 passages. These mice re- CHOLANTHRENE ACCORDING TO SITE OF PELLET ceived the same general care and food as those in which I M PLANTATION Group I Group II Group III the pellets of methylcholanthrene were implanted. (Cerebrum) (Cerebellum) (Meninges) The material used for subcutaneous'transplantation Total number of animals... 57 3o consisted of a piece of tissue removed from the main Negative for tumor ...... 3I 17 tumor mass, with careful avoidance of obviously ne- Total number of tumors .... 26 I3 Unclassified tumor ...... t 2 crotic and hemorrhagic areas. This was cut in sterile Gliomas ...... I5 8 saline into fragments about x mm. in diameter. The Astrocytoma ...... I fragment of tumor tissue was deposited in the sub- Glioblastoma multiforme. 4 cutis of the right axillary region or in the right flank Medulloblastoma ...... l Oligodendroglioma ..... 3 by means of a trocar introduced into the right groin Spongioblastoma polare .... through skin washed with alcohol. Unclassified glioma ..... 5 Necropsy technic.--Mice which were moribund were Multiple gliomas ...... i invariably killed by sectioning the cervical spinal cord Sarcomas ...... 5 i Rhabdomyosarcoma ..... 2 and a complete necropsy was performed immediately Meningeal sarcoma ..... 2 thereafter. The brain was removed with sterile pre- Cerebral sarcoma ...... I cautions and, when a tumor was encountered, a piece Mixed sarcoma and glioma. 5 2 was usually excised for subcutaneous transplantation. The brain was then fixed in neutral formalin (U.S.P. bellum, 3 ~ mice; and the subdural space, i6 mice. The formaldehyde I :Io). Animals which were found dead incidence of the various types of neoplasms as they were necropsied immediately on discovery and the occurred at the different sites of pellet implantation in brains fixed in the same manner. It proved expedient the three groups of the io 3 mice of this experiment is in a few instances, because of onsetting post-mortem shown in Table I. softening, to harden the contents of the partially opened Extreme conservativism was employed in the classi- crania in the fixative for :z4 hours before removing the fication of the neoplasms, which accounts for the fact brains. Since tumors were not found in any of the that such noncommittal designations as unclassified other viscera of these animals the organs were not tumor and unclassified glioma appear in Table I. The saved. three tumors under the first designation, although gen- With a few exceptions the brains were embedded in uine neoplasms as indicated by their invasiveness and paraffin and sectioned serially. Hematoxylin-eosin was the presence of cells in mitosis, were nevertheless too the stain employed routinely, but at regularly spaced small to permit more detailed study and classification. intervals sections were prepared with the Masson tri- The ten tumors designated as unclassified glioma were

large enough to permit special study. None of them after the intracerebellar implantation of methylcholanthrene and contained reticulin fibers and, although glial elements after a course which was characterized by disturbances of balance and paralysis of the left hind leg. At necropsy, the brain was were identified in all, they failed to present such char- found adherent to the skull. The cerebral hemispheres were acteristic architectural patterns as are requisite for the uninvolved, but the left half of the cerebellum was bulging and classification of tumors of the glioma group. As an brown in color. On section it was seen that this was due to a unclassified glioma, for example, is listed the tumor of hemorrhagic and semigelatinous tumor mass surrounding the mouse No. 74, which will be described in some detail pellet (MCA) (Figs. 3-A and 3-B). The microscopic picture revealed an infiltrating glioma com- later and which had certain features of glioblastoma posed of a remarkably uniform type of cell. This had a promi- multiforme. To avoid arbitrariness, however, this tu- nent oval or round nucleus of medium size with numerous mor was left unclassified. chromatin granules. The cell body had scant cytoplasm and In one instance, mouse No. ~ I, the brain tumor was was inconspicuous. There were no cellular processes. Cells of this variety were frequently arranged in pseudorosettes (Fig. found to consist of two distinctly different component 3-D). Mitotic division was present in abundance. The tumor parts~one, oligodendroglioma and the other, glio- had almost no stroma; the Wilder preparations were negative blastoma multiforme. This is not to mean that oligo- (Fig. 3-C). Calcium salt deposits were found in several regions dendroglia cells were found scattered in a tumor that (Fig. 3-E). There was widespread invasion of the cerebral otherwise resembled a glioblastoma, but rather that two leptomeninges and of the Virchow-Robin spaces in many parts of the brain. Tumor cells were found in the fourth ventricle, different tumors were found side by side. In Table I, the aqueduct of Sylvius, and the third ventricle (Fig. 3-F). these new growths are listed as In multiple gliomas. MouSe No. 49- This mouse was killed 296 days after methyl- the same sense multiple tumors composed of gliogenous cholanthrene was implanted in the cerebellum, when the animal and sarcomatous portions were also discovered. These, developed paralysis of both hind legs and a disturbance in 7 in number, are classified as mixed sarcoma and balance which was characterized by a tendency to fall to the left in walking. The crania[ sutures were found separated at glioma in Table I. necropsy, which accounted for a slight increase in the size of the Space does not permit a detailed description of each head of this animal. The hydrocarbon was found embedded in of the neoplasms produced, but representative tumors a somewhat gelatinous tumor mass replacing the left half of the will be described below. cerebellum (Fig. 4-A). Part of this tumor was removed for transplantation in other mice, the results of which will be Glioblastoma multilorme.--Mouse No. I4. This animal died reported below. 2o 7 days after the intracranial implantation of the hydrocarbon. The tumor was found microscopically to be poorly demarcated At necropsy the methylcholanthrene pellet (MCA) was found from the surrounding cerebe[lar parenchyma and was composed buried deep in the right parietal lobe, which was replaced in of a remarkably uniform" type of cell. The latter resembled in large part by a hemorrhagic neoplasm measuring x cm. in all important features the cell described in mouse No. 52 diameter (Figs. x-A and l-B). This tumor eroded the overlying (Figs. 4-B and 4-C). For the most part the neoplastic elements calvarium and lay as a flattened, partially necrotic, and partially failed to form any definite architectural pattern, lying helter- calcified mass beneath the scalp. It was composed of pleo- skelter, but a faint suggestion of pseudorosette formation was morphic cells, many of which had bipolar processes. These seen in some regions. The nearby meninges and choroid plexus cells were frequently arranged in pseudopalisades around foci were infiltrated with tumor cells (Fig. 4-D). The stroma was of necrosis (Fig. x-C). Many were in mitotic division and many scant and indifferently stained; Wilder preparations were nega- were multinucleated, some nuclei containing spheroid, pink- tive (Fig. 4-E). staining structures resembling inclusion bodies (Fig. I-D). The Oligodendroglioma.--Mouse No. I6. The animal died 366 choroid plexus of the right lateral ventricle was infiltrated with days after pellet implantation in the right cerebral hemisphere tumor cells and the leptomeninges likewise contained clusters of where, at necropsy, a gray, opaque, and partly hemorrhagic these cells. There was no stroma of reticulin in this neoplasm tumor was found (Figs. 5-A and 5-B). The midline of the as demonstrated by the Wilder silver impregnation method brain had shifted to the left. An infiltrating glioma was found, (Fig. I-E). microscopically composed of cells having an exceptionally uni- Mouse No. 4 ~ . This animal survived 314 days the pellet form appearance. They were of small size with scant, pink- implantation in the cerebellum. At necropsy the calvarium was staining cytoplasm and small, dark, round nuclei (Fig. 5-C). found intact, but much of the cerebellum was replaced by a Frequently the nuclei seemed to lie naked in clear, unstained gray, semigelatinous, infiltrating glioma (Figs. 2-A and z-B). halos, but occasionally they were surrounded by narrow rings Many of the cellular elements of this tumor were identified as of cytoplasm which lay within the halos. Some of the cells were unipolar and bipolar spongioblasts, but other glial elements such in mitotic division. The tumor was hemorrhagic in spots but as astrocytes were also present. There were moderate numbers was devoid of reticulin. of cells in mitotic division and of multinucleated giant cells. Spongioblastoma polare.--Mouse No. 55. Four days before Numerous zones of necrosis were seen around which spongio- this animal was found dead on the 3I 4th day of the experiment, blasts were arranged in pseudopalisades (Fig. e-C). Small loci its head was noted to be peculiarly deformed. Necropsy disclosed of hemorrhage and of calcium salt deposition were found this to be due to a bulge of the cranium in the right parietal scattered in the tumor. The neoplastic cells had infiltrated the region due to a large tumor mass which replaced most of the leptomeninges and extended along the Virchow-Robin spaces right cerebral hemisphere (Fig. 6-A). The pellet of methyl- into the nervous parenchyma (Fig. e-D). There was no evidence cholanthrene presumably had been placed in the subdural space of a vascular proliferative reaction. What little stroma was in contact with the meninges, but it could not be located there visible in the neoplasm was not formed by reticulin, as the at necropsy, nor could it be found in the cerebral tumor mass. Wilder preparations revealed (Fig. 2-E). The histologic picture of the intracerebral neoplasm (Fig. 6-B) Medttlloblastoma.--Mouse No. 52. This mouse died 295 days revealed invasiveness and a number of deep-seated hemorrhages.

Large zones were composed of parallel rows, bands, or whorls neoplasm which arose at the site of the pellet implantation of spindle-shaped cells with elongated nuclei and bipolar proc- (Figs. io-A and io-B). It was apparent that the bulk of the esses. These cells resembled spongioblasts. Intermingled with tumor within the cerebellum itself was composed of small round them were medium-sized cells with uniform dark nuclei (Fig. cells forming one distinctive pattern and that the outer, cap-like 6-C). A few calcium salt deposits were encountered in this portion of the tumor was composed of densely packed spindle- portion of the tumor. In other parts there were present closely shaped cells forming another pattern (Figs. Io-B and Io-C). packed bizarre-shaped giant cells, some multinucleated (Fig. In the larger portion the cells were easily identified as oligo- 6-D). The cytoplasm of these cells was abundant and formed dendroglia (Fig. I0-D). They were round and small with in- prominent processes. There were many glia cells in mitotic tensely stained, round nuclei surrounded by halos of pale or division. The tumor had very little stroma, none of it of meso- unstained cytoplasm. Occasional astrocytes were present as well dermal origin as the negative Wilder stain proved (Fig. 6-E). as groups of multinucleated cells and cells in atypical mitotic Multiple gliomas.--Mouse No. 11. The animal was found dead ctivision. This part of the neoplasm contained no reticulin fibers on the ~92nd day of the experiment. On reflecting the scalp, a (Fig. I o-E). tumor mass of about 4 ram. diameter was found protruding The smaller, dorsally lying portion of tumor was definitely through the skull at the site of the trephine wound. It was of mesodermal origin. Its cells had oval, chromatin-rich nuclei found originating in the right cerebral hemisphere, destroying and bipolar processes (Fig. to-F). The stroma of this portion the basal ganglia and shifting the midline to the left (Figs. 7-A was readily impregnated with silver by the Wilder method for and 7-B). There was some hemorrhage in this tumor and somc reticulin (Fig. 10-G). necrosis. Microscopically the neoplasm was composed of several different II. SUBCUTANEOUS TRANSPLANTATION OF EXPERI- zones. In one, the cells were nearly all spongioblasts, the MENTALLY PRODUCED BRAIN TUMORS processes of which pointed to centers of necrosis. These cells were arranged in pseudopalisades (Figs. 7-C and 7-D). The The following section includes descriptions of repre- endothelial cells of many blood vessels were proliferated. Giant sentative cases of successful transplantation of various cells and mitotic figures were present in small numbers. The induced brain tumors. cell constituents and architecture were typical of gliobIastoma multi[orme. In other large areas the tumor cells contained round, Medulloblastoma.--Mouse No. 49- The tumor of this animal remarkably uniform nuclei surrounded by clear halos (Fig. 7-E). was described above as an example of an invasive malignant Here mitotic figures were rare. These portions of tumor repre- cerebellar neoplasm which was classified as a medulloblastoma. sented oligodendroglioma. Reference to the description of this tumor and to Fig. 4 in which Meningeal sarcoma.~Mouse No. 83. A tumor appeared it is illustrated will, perhaps, raise some doubt as to the justifica- beneath the scalp on the 3izth day of the experiment. During tion for this classification. The results of subcutaneous trans- the next 3 days it increased so rapidly in size that it was plantation, however, leave no doubt on this point and demon- deemcd essential to kill the mouse in order to save material for stratc the value of this mcthod of study, especially in difficult subcutaneous transplantation. The vertex of the skull was eroded cases. by a neoplasm which arose in the longitudinal fissure region In Fig. I I-A is shown the tumor which developed in 45 clays near the pellet in the right parietal lobe (Fig. 8-A). The carci- after implantation in the subcutaneous tissues of the right flank. nogen was embedded in the base of the tumor which had a Subtransplantation was carried out in i0 generations, involving a loose structure and seemed to arise from the leptomeningcs. Its total of 44 animals. In general, the microscopic appearance of cells were bipolar and had long processes. The nuclei were the transplants resembled rather closely the primary neoplasm, oval or elongated and contained many large chromatin granules. but the characteristic architecture was more pronounced. Thus, There were many mitotic figures. The neoplastic cells were pseudorosette formations were more numerous and better formed arranged in long strands or in whorls, frequently around capil- (Fig. I I-B). It is of interest to note that in spite of the fact laries (Fig. 8-B). A delicate reticulin network permeated the that the gliogenous tumor grew in the subcutaneous tissues, it whole neoplasm including the whorls (Fig. 8-C), but there were grew as a "pure" medulloblastoma uncontaminated by meso- no collagenous fibers. dermal elements (Fig. I I-C). Cerebral sarcoma.--Mousc No. I0. Two clays before this Unclassified glioma.--Mouse No. 74. Reference has already animal was killed on the 372nd clay of the experiment, a small been made above to the problem presented by this tumor. The tumor appeared on the top of the head beneath the scalp. At primary neoplasm appeared in the right parietal lobe 24o days necropsy this tumor nodule was found to be an extension after the implantation of the carcinogen. The microscopic struc- through the craniotomy wound of a much larger tumor mass ture of the tumor (Fig. I2-A) was densely cellular with poor in the right parietal lobe (Figs. 9-A and 9-B). The neoplasm demarcation from the surrounding brain tissue. There was great was gray in color, firm in texture, and lay above the pellet of cellular pleomorphism with many cells identifiable as astrocytes, methylcholanthrene which was partially submerged in the right spongioblasts, and even medulloblasts. Cells in mitotic division lateral ventricle. The midline of the brain was shifted to the were seen frequently; however, no characteristic architecture was left; the right cerebral hemisphere was compressed by the present to aid classification. Tumor cells were found diffusely neoplasm, a portion of which was utilized for subcutaneous infiltrating the leptomeninges (Fig. 12-B). transplantation. The first subcutaneous transplant revealed tumor cells of two The cells forming this tumor were of two types (Fig. 9-C). varieties. One was a rather small cell with scant cytoplasm and One had large, pate vesicular nuclei, one or two nucleoli, and a round, chromatin-rich nucleus. Groups of such cells had a ten- moderate amount of pale cytoplasm. The other had small, dark, dency to form pseudorosettes (Fig. I2-C). The other was a round nuclei and inconspicuous cytoplasmic bodies. There was larger cell with a vesicular nucleus and a conspicuous cytoplasmic no characteristic architecture. The stroma was scant, but was body from which originated many processes. In hematoxylin- readily impregnated with silver in the Wilder preparations eosin preparations the latter cell type resembled epithelioid cells (Fig. 9-D). There were no collagen fibers. which were frequently arranged in wide bands around blood Mixed sarcoma and glioma.--Mouse No. 44. On the 222nd vessels in the manner of astroblasts. This is illustrated in (lay of the experiment this animal succumbed to a cerebellar Fig. I2-D, which is derived from the I I th subcutaneous sub-

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 1941 American Association for Cancer Research. Zimmerman and Arnold--Experimental Brain Tumors 923 transplant. Multinucleated giant cells, like those seen so fre- No. 1o6. At 177 days of the experiment the head of this animal quently in glioblastoma multiforme, were present in small num- was malformed by a bulging tumor mass which had its origin bers in all of the I2 generations of transplants from this in the right parietal lobe and had eroded the overlying calvarium tumor (Fig. I2-E). to present itself beneath the scalp (Figs. I4-A and I4-B). The A total of 96 mice received subcutaneous transplants from this tumor was firm, fleshy, and gray. It had produced a shift of the neoplasm, only 5 mice in all revealing a completely similar midline of the brain to the left. tumor architecture. The fluctuation in the histologic appearance The microscopic appearance of the neoplasm was difficult to from one generation of transplants to another made a rigid evaluate since it presented a confused mixture of several different diagnosis of the tumor unwarranted, but certainly the resem- types of cells. Many small round elements were present which blance to glioblastoma multiforme was the most frequently had deeply stained nuclei and scant cytoplasmic bodies (Fig. observed. ~4-C). Large numbers of these cells were in different stages Astrocytoma.--Mouse No. 69. After 27o days this animal suc- of mitotic division. Here and there were noted giant cells with cumbed to the neoplasm which arose in the right parietal lobe more than one nucleus. In the background of these cells were at the site of methylcholanthrene implantation (Fig. I3-A). The found interlacing strands and whorls of spindle-shaped elements eyes were constantly closed during the last days of life and the with elongated nuclei and cytoplasmic bodies that seemed to give head was misshaped from the tumor growth and as a result origin to a fibrillary stroma. Some of these cells were also in of the separation of the cranial sutures. mitotic division. The absence of a characteristic structural pat- At necropsy the tumor was found involving both hemispheres, tern microscopically precluded a diagnosis of this tumor. and the pellet was located in the subcortex at the junction of Portions of the primary neoplasm were transplanted subcu- frontal antl parietal lobes. This tumor had a gelatinous con- taneously into 8 mice, 7 of which developed local tumor growths sistency and was poorly demarcated from the surrounding brain only after more than 4 weeks. In all 7 mice the transplanted tissue. It consisted of cells whose cytoplasm was scant and gave tumors differed considerably from the primary tumor. In 5 of rise to multipolar processes. The nuclei were intensely stained the mice the microscopic appearance of the neoplasm was that and often had one or more nucleoli. Frequently the)' were of spindle-shaped cells forming strands and whorls (Fig. I4-D) arranged around spaces containing homogeneously pink-staining in an abundant stroma of reticulin fibers (Fig. I4-E ). The (in hematoxylin and eosin preparations), colloid-like material tumors of the remaining 2 mice, although similar to each other, (Fig. 13-B). Cells in division were rather numerous and occa- differed entirely from those just described. They consisted of sional large bizarre elements were also seen. There was no uniformly round, small cells with deeply stained nuclei and a invasion of the meninges except for one small focus. The vast minimum amount of stroma (Fig. I4-F). Many of these cells majority of the cells were thus readily identifiable as astrocytes, were dividing and there was a definite tendency for them to be although the cellular division was a discordant feature. Meso- arranged around blood vessels or tissue spaces in the form of dermal constituents were absent (Fig. 13-C). pseudorosettes. Reticulin could not be demonstrated in either A series of 1o subtransplants inw)lving 48 mice was made of of these tumors (Fig. 14-G). this neoplasm. In general, the microscopic features of the trans- Thus it was shown conclusively that the primary neoplasm plants were similar to those in the original tumor. There were was composed in part of mesodermal and in part of gliogenous seen the same cystic spaces filled with coagulated material and elements. Unfortunately, subtransplants were made only of the the same multipolar astrocytcs. In addition, however, the cells mesodermal tumor, which went through 9 generations in a total began to assume a pseudorosette formation in the second sub- of 37 mice. In each animal the tumor remained a "pure" transplant (Figs. I3-D and I3-E) which became progressively sarcoma without a gliogenous component. From the previous more distinct in subsequent subtransplants and reachcd its full experiences with subcutaneous transplants of gliomas there is no developmcnt in the 7th gcneration (Fig. :3-F). Here the reason to suspect that the gliomatous portion of the primary pseudoglandular structure was quite conspicuous and was neoplasm could not have been perpetuated had subtransplants strongly suggestive of the human "piloid" or "malignant" been made. astrocytoma. Meningeal sarcoma.--Mousc No. 83. The primary tumor DISCUSSION which this animal developed at the site of the implanted carcinogen received attention above. Its histologic appearance was The incidence of brain tumors induced by the intra- that of a reticulin-forming sarcoma whose cells had a tendency cranial implantation of methylcholanthrene in C3H to form whorls. Transplantation of this tumor was carried out male mice was 46.6 per cent (48 out of IO3 mice). through 8 generations, involving a total of 44 mice. The in- The pellets of carcinogen were found within the variable microscopic picture seen in the transplants was that of strands of elongated spindle-shaped cells sometimes caught in cranial cavities in association with the neoplasms in a longitudinal and sometimes in a transverse plane (Fig. 8-D). the animals which developed tumors and were em- At no time were the whorls of the original tumor encountered. bedded in normal tissue in the mice which failed to Numerous cells were in division and connective tissue reticulin develop them. Yet the pellets removed from both the was produced abundantly (Fig. 8-E). Cerebral sarcorna.--Mouse No. I0. The primary tumor which positive and negative tumor groups were equally ef- this mouse developed was described above. It consisted of two fective in inducing brain tumors later when implanted cell types, a large and a small, in a scant argcntophile stroma. in other C3H mice. This experience suggests that other There was no characteristic architectural pattern. A total of factors in addition to the carcinogen are important in 32 mice received transplants from this neoplasm, its perpetuation the induction of tumors. Also the fact that only 46 per being voluntarily stopped at the end of the 6th generation. Prac- tically all the transplanted tumors showed a predominance of cent of the mice developed brain tumors points to fac- the large cells, many of which were in mitotic division (Fig. tors, in addition to the carcinogen, which influence 9-F). A characteristic structure failed to develop although reticu- tumor production. The present experiment, however, lin continued to be formed in small amounts (Fig. 9-E). was not devised to throw light on this question. Mixed sarcoma and glioma.--Perhaps the greatest value of the technic of subcutaneous transplantation was demonstrated for It was originally felt that the site of origin of the the group of mixed tumors, an example of which is mouse brain tumor might influence its type and it was for this

reason that the pellets of carcinogen were implanted in environment. This is perhaps all the more surprising 3 different locations; namely, in contact with the since it is a well known fact that the human tumors of meninges, the cerebrum, and the cerebellum. Of the this variety are never found as extracranial metastases. 9 tumors which developed in the first named location, Apparently there is no local tissue resistance against the 7 were sarcomas, as was to be expected, and 2 were ectodermal gliomas, but rather an absence of an avail- gliomas. In each of the latter 2 instances, however, the able pathway for metastasis. Even the most malignant pellet actually came in contact with the nervous paren- of the gliomas fail to invade blood vessels. chyma. The converse of this also occurred; namely, the Among the more interesting results of these experi- development of sarcomas following the implantation of ments were those obtained with the transplantation of pellets in the cerebrum. Thus, even 2 rhabdomyosarco- the tumors designated as mixed sarcoma and glioma. mas were produced, but this undoubtedly could be In transplants it proved possible to grow the constituent ascribed to the fact that the carcinogen had worked its parts of these neoplasms in pure form; i.e., as glioma way out through the burr hole and came to lie in con- or sarcoma. The subtransplants of these "purified" tact w~th the temporalis muscle. tumors remained true in many subsequent generations. Of the gliomas that were produced in the cerebrum, Such results are comparable to the every day experience only 2 examples need special mention. One was the of the bacteriologist who subcultures colonies from a medulloblastoma, an unusual site for this tumor, and mixed bacterial growth to obtain several pure strains of the other was the instance of multiple gliomas. In the organisms. cerebellum, the glioblastomas that were found there could not have been anticipated, but the 3 medullo- SUMMARY AND CONCLUSIONS blastomas were entirely in keeping with clinical ex- Pellets of purified 2o-methylcholanthrene were im- perience. With certain definite reservations, therefore, planted in the cerebral meninges, the right cerebral it could be stated that the site of origin of the neoplasm hemisphere, and the cerebellum of Io 3 C3H mice of had an influence in determining its type. This, how- the male sex. ever, is not to be interpreted as implying that any con- In all, 48 tumors were produced in this manner: 25 siderable light has been shed on the histiogenesis of gliomas, 13 sarcomas, 7 mixed gliomas and sarcomas, the gliomas. That still remains a moot question. It and 3 unclassified. Among the gliomas were present was at first hoped that a study of the beginnings of a examples of astrocytoma, glioblastoma multiforme, me- glioma in a young tumor would explain its histio- dulloblastoma, oligodendroglioma, and spongioblas- genesis, in reality, however, the malignant cells in early toma polare. Within certain limits the site of pellet stages were unidentifiable with respect to the subse- implantation was a determinant of the type of intra- quent type of tumor and remained so until they pro- cranial neoplasm which developed. liferated sufficiently to produce a recognizable archi- The rate of growth of the sarcomas was much greater tectural pattern. The subcutaneous transplants aided than of the gliomas. The average time when the sar- almost as much in the study of the histiogenesis as of comas appeared was x95 days as against 279 for the the growth behavior of these tumors, but it is still not gliomas. possible to state what the factors are which determine The method of subcutaneous transplantation was em- whether a given glial cell will form an astrocytoma, an ployed for the study of the growth behavior of these oligodendroglioma, or any other type of glioma. intracranial neoplasms. From 9 to 14 subtransplants A striking difference was noted in the rate of devel- were made of many of these tumors with results that opment between the primary sarcomas and gliomas. indicated a much more rapid growth of the sarcomas Of the 13 sarcomas, the first appeared on the I25th than the gliomas. Frequently, unclassifiable primary day and the last on the 372nd day. The average, how- gliomas developed characteristic structural patterns in ever, was i95 days and, more striking still, was the the transplants which made identification possible. mean of 165 days. Of the 25 gliomas, the first was This method of study also permitted the separation of noted on the I27th day and the last on the 378th day. so-called "mixed" tumors into their component parts. The average day of appearance was the 279th and the mean, the 33oth day. Accurate determinations of the REFERENCES first appearance and the rates of growth of the tumor I. PEERS, I. H. The Response of the Central Nervous System transplants were not made, but the generalization is to the Application of Carcinogenic Hydrocarbons. II. justified that the gliomas grew much more slowly than Methylcholanthrene. Am. J. Path., I6:799-816. I94 o. 2. SELmMAN, A. M., and M. J. SHEAR. Studies in Carcino- the sarcomas. Signs of growth of the latter tumors genesis. VIII. Experimental Production of Brain Tumors were often noted within 3 weeks, whereas the gliomas in Mice with Methylcholanthrene. Am. J. Cancer, 37: frequently took at least twice this time. 364-395. 1939. 3. ZIMMERMAN, H. M., and H. ARNOLD. Experimentally In- It is of some interest to note with what great facility duced Primary Intracranial Neoplasms. Tr. Am. Neurol. the glioma transplants grew in their new mesodermal A., 66:191-193. 194o.

F1o. 1.--Mouse No. ~4. Glioblastoma multiforme. A. Pellet mag. X 6. C. Pseudopalisade around focus of necrosis. H & E of methylcholanthrene (MCA) deep in right parietal lobe and stain; mag. X 2oo. D. Multinucleated giant cells. Note pale surrounded by tumor. Note extracranial position of neoplasm spheroid structure resembling inclusion body in nucleus of upper in upper figure. B. Transverse section of brain showing intra- giant cell. H & E stain; mag. X 270. E. Absence of reticulin cerebral and extracranial portions of tumor. H & E stain *; fibers. Wilder stain; mag. X 4oo.

*Hematoxylin and eosin stain abbreviated in legends as H & E.

Fro. 2.--Mouse No. 4o. Glioblastoma multiforme. A. Draw- Calcification in lower left corner. H & E stain; mag. 5( ~6o. ing of tumor in cerebellum. B. Photomicrograph of cerebellar D. Tumor cells invading leptomeninges and molecular layer of neoplasm. Clear space represents site of methylcholanthrene cerebellar cortex. H & E stain; mag. X I6O. E. Absence of pellet. H & E stain; mag. X 8. C. Pseudopalisading of cells. reticulin in tumor. Wilder stain; mag. X 5oo.

#" .... "*a~ ...... ir : # ..... MCA~i

' ~,~.,~. ~"

B ...... ,~ ~ ~ ...... ,~ ......

Fro. 3.--Mouse No. 52. Medulloblastoma. A. Drawing of D. Pseudorosette formed by tumor cells. H & E stain; mag. tumor in cerebellum at site of carcinogen (MCA). B. Photo- X 35o. E. Calcium salt deposit in tumor. H & E stain; mag. micrograph of neoplasm in cerebellum. H & E stain; mag. X 6. X 35o. F. Tumor invasion of aqueduct of Sylvius. H & E C. No reticulin fibers in tumor. Wilder stain; mag. X 45o. stain; mag. X ~75.

F1c. 4.--Mouse No. 49. Medulloblastoma. A. Tumor in left differently stained stroma. H & E stain; mag. X 450. D. Infil- half of cerebellum containing empty pellet space. H & E stain; tration of choroid plexus by tumor cells. H & E stain; mag. mag. X 9. B. Photomicrograph of tumor cells. H & E stain; 5< 3oo. E. Note absence of mesodermal fibers in tumor stroma. mag. X 25o. C. Higher magnification of same cells in in- Wilder stain; mag. X 45o.

FxG. 5.--Mouse No. I6. Oligodendroglioma. A. Drawing of stain; mag. X 6. C. Note characteristic perinuclear halos in hemorrhagic tumor in situ. B. Note pellet space at top of infil- tumor cells. H & E stain; mag. X 4oo. trating tumor which has shifted the midline to the left. H & E

Fro. 6.--Mouse No. 55. Spongioblastoma polare. A. Note spongioblasts in interlacing strands. H & E stain; mag. )< 200. tumor mass replacing most of right cerebral hemisphere. B. The D. Multinucleated giant cells. H & E stain; mag. X 40o. E. invasive neoplasm is seen in right hemisphere with hemorrhages Note absence of reticulin fibers. Wilder stain; mag. X 2oo. on its margins. H & E stain; mag. X 5- C. Arrangement of

.... {,~ ..... , ...... a,s,

FIG. 7.--Mouse No. I i. Multiple gliomas. A. Drawing of 5( Ioo. D. Spongioblasts undcr higher magnification. H & E tumor in right cerebral hemisphere. B. Note destruction of basal stain; mag. X 4oo. E. Photomicrograph of oligodendroglioma ganglia by neoplasm and shift of midline to left. H & E stain; portion of neoplasm. Note typical cells with perinuclear halos. mag. X 5. C. Pseudopalisade formation. H & E stain; mag. H & E stain; mag. X 400.

s~w

o ' ' i

Fro. 8.--Mouse No. 83. Meningeal sarcoma. A. Loosely con- mag. X 30o. D. Interlacing bands of cells in subcutaneous structcd neoplasm invading right parietal cortex near longitudinal transplant. H & E stain; mag. )< 3o0. E. Abundant reticulin in fissure. H & E stain; mag. X 6. B. Whorl formation in tumor. transplant. Wilder stain; mag. X 3o0. H & E stain; mag. X ~oo. C. Reticulin in whorl. Wilder stain;

W~CA.

FIG. 9,--Mouse No, 1o. Cerebral sarcoma. A. Drawing of Wilder stain; rnag. X 5oo. E. Wilder preparation positive for sharply circumscribed tumor mass lying above carcinogen reticulin in 1st transplant. Mag. X 6oo. F. Appearance of (MCA). B. Photomicrograph of neoplasm. Note empty pellet tumor in Ist transplant. Note cells in mitotic division. H & E space. H & E stain; mag. X 5. C. Cellular detail of sarcoma. stain; mag. )< 3oo. H & E stain; mag. M 3oo. D. Well impregnated reticulin fibers.

FJ(;. lo.--Mouse No. 44- Mixed sarcoma and glioma. A. Ap- tumor. H & E stain; mag. X I5o. E. Wihlcr preparation of pearance of cerebellar neoplasm. Drawing. B. Bulk of tumor-- same part of tumor to show absence of reticu[in. Mag. X 15o. oligodendroglioma. Note cap of sarcoma in right upper corner. F. Cellular detail of sarcomatous portion of tumor. H & E stain; H & E stain; mag. N{ 7. C. Sarcomamus portion in upper half mag. X I5O. G. Wilder preparation of latter part of tumor to ar, d gliomatous in lower half of photomicrograph. H & E stain; show abundant argyrophile fibers. Mag. M 15o. mag. X 3 ~ D. Cells of oligodendroglioma in larger portion of

Fro. 11.--Mouse No. 49, Medulloblastoma. A, Tumor trans- H & E stain; mag, X 400. C. Negative Wilder stain of same plant in subcutaneous tissues. B. Pseudorosette in transplant, transplant. Mag_ )< 350.

e ,,

.... ;i:::~: ili~ : i~ ....,~ :~.._

L N ii~iiiii " , ~:~,~'~ ~

~a~'~ ~'~ * *" "-:r ~ ,~i

*%.li it*"

FIG. 12.--Mouse No. 74. Unclassified glioma. A. Microscopic cells arranged around blood vessel in IIth transplant. Note appearance of original tumor in right parietal lobe. H & E epithelioid nature of cells. H & E stain; mag. X 3oo. E. Mul- stain; mag. X 300. B. Tumor cells infiltrating leptomeninges, tinucleated giant cells in second transplant. H & E stain; mag. H & E stain; mag. X 300. C. Pseudorosette formation of tumor X 300. cells in Ist transplant. H & E stain; mag. X 300. D. Tumor

FIG. 13.--Mouse No. 69. Astrocytoma. A. Drawing of neo- tumor in 2nd transplant. H & E stain; mag. X 30o. E. Note plasm in brain at site of carcinogen. B. Cystic spaces filled with absence of reticulin in this transplant. Wilder stain; mag. X 300. colloid-like material. H & E stain; mag. X 200. C. Negative F. Appearance of same tumor in 7th transplant. Note pseudo- Wilder stain of same tumor. Mag. X 200. D. Appearance of rosette formations. H & E stain; mag. X 40o.

MCA B

...... j

F1o. 14.--Mouse No. 1o6. Mixed sarcoma anti glioma. A. l)e- No. 5 of Ist transplant. H & E stain; mag. X 300. E. Same formed head due to prcsentation of tumor beneath scalp through tumor in Wilder preparation. Note abundant reticulin. Mag. eroded calvarium. B. Drawing of tumor at site of carcinogen X 500. F. Glioma in mouse No. 3 of 2nd transplant. H & E (MCA) in right cerebral hemisphere. C. Original tumor. Note stain; mag. X 300. G. Same tumor in Wilder preparation. mixture of spindle-shaped mesodermal elements and round gliog- Negative for reticulin. Mag. X 3oo. enous cells. H & E stain; mag. X 3oo. D. Sarcoma in mouse

H. M. Zimmerman and Hildegarde Arnold

Cancer Res 1941;1:919-938.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/1/12/919.citation

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/1/12/919.citation. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.