Juvenile Nasopharyngeal Angiofibroma Tumor Models Failure of Androgens to Stimulate Growth in Nude Mice and In Vitro Alan H.

Shikani, MD, William J. Richtsmeier, MD, PhD

nasopharyngeal angiofibroma is a tumor with a predilection for adolescent boys. It has been shown to contain cytosolic androgen receptors and to regress with estrogen therapy; however, the results have not been consistent. Extensive investigation has been unable to settle this issue in patients owing, in part, to the rarity of these tumors. We have attempted to establish a tumor model for juvenile nasopharyngeal angiofibroma by transplanting the tumor into the subdermal space of athymic mice and also by culturing it in vitro, to study the effect of hormonal manipulation. The tumor did survive in male and female athymic mice but has failed to grow. Androgen treatment of the mice of either sex did not alter its survival or growth behavior. The in vitro

\s=b\ Juvenile

tissue culture grew fibroblastoid cells that were not stimulated by androgen supplementation. This study suggests that factors other than androgens are at least complementary, if not essential, in promoting the growth of juvenile nasopharyngeal angiofibroma in tumor models, and that androgens are not, in and of themselves, sufficient growth stimuli. (Arch Otolaryngol Head Neck Surg. 1992;118:256-259) uvenile

nasopharyngeal angiofibromas (JNAs) are rare

lesions, comprising only 0.05% of all head and neck

striking characteristic of these tumors is their propensity for occurring in the peripubescent male pa¬ tient. Although a few cases have been reported in female patients, many pathologists believe that such female pa¬ tients were incorrectly diagnosed.2-3 In a large review of 120 cases of JNA, not a single female was found.4 Based

hydroxysteroid, ketosteroid, and gonadotropin levels, and correlation of the degree of sexual maturation with chronological age, were all normal.1-14,1517 Studies of the

hormonal receptors demonstrated the presence of cyto¬ solic androgen receptors but the absence of estrogen or progesterone receptors in JNA,18"21 suggesting that the action of the estrogen on these tumors is indirect, possi¬ bly through hypothalamic suppression and reduction in the secretion of luteinizing hormone and subsequently in the secretion of testosterone.18 All the observations point to the dependence of these tumors on androgens and emphasize the rationale of antiandrogen therapy as a most appropriate treatment. The rarity of these lesions, however, makes it difficult significantly to evaluate the effect of hormonal manipulation. Establishing a tumor model was, in our minds, the first step in approaching this problem. If successful, the nude mouse-human tumor model would allow studying the effect of the sex of the animal and the addition of exogenous testosterone on transplantability and growth of the tumor. The in vitro tissue culture techniques would allow more extensive hormonal manipulation and investigation of androgen

receptor regulation.

tumors.1 A



We used (RNU) athymic nude NuBr mice (Harlan SpragueDawley Ine, Annapolis, Md). They were housed in the Nude Mice Facility of the Oncology Center at The Johns Hopkins Hos¬ pital, Baltimore, Md, in a controlled environment.

patients, including

Three different JNAs were transplanted into 20 mice each. The donor material was obtained from three adolescent male patients undergoing a surgical resection at The Johns Hopkins Hospital. AH tumors were clinically observed to be in active growth just before treatment. One patient had previously failed estrogen therapy and radiotherapy. Diagnosis of JNA was confirmed by histological examination of the donor material adjacent to that used for this study (Fig 1).

these observations, exogenous estrogens in varying doses have been administered either solely as or adjuvant therapy before surgery or irradiation to diminish the mass and vascularity. Although many authors have reported a reduction in tumor size and intraoperative blood loss with estrogen therapy, the results have not always been consistent.515 Some have demonstrated an increase in tu¬ mor size in response to testosterone treatment.16 Extensive studies of the pituitary-gonadal axis in these




levels, urinary

publication June 27, 1991. From the Department of Otolaryngology\p=n-\Headand Neck Surgery, The Johns Hopkins Medical Institutions, Baltimore, Md. Reprint requests to Department of Otolaryngology\p=n-\Headand Neck Surgery, The Johns Hopkins Medical Institutions, 600 N Wolfe St, Carnegie 400, Baltimore, MD 21205 (Dr Shikani). for


Transplantation and Observation Samples from the freshly resected specimens were cut into 12-mm3 pieces and transplanted subcutaneously into the flank region of the nude mouse recipients, which had been anesthe¬ tized with inhaled methoxyflurane (Metofane). The transplan¬ tation procedure followed the principles of Povlsen and Ryto

Downloaded From: http://archotol.jamanetwork.com/ by a University of Manitoba User on 06/10/2015

Fig 1. Original juvenile nasopharyngeal angiofibroma from one of the patients. Typical angiofibroma with prominent slitlike vessels surrounded by a thin fibromuscular zone, composed mainly of collagenous eosinophilic stroma with a prominent amount of stromal myofibroblasts (hematoxylin-eosin, x 0).

Fig 3. Nine-day specimens from mouse. Well-demarcated tumor composed of abundant eosinophilic collagenous stroma contain¬ ing moderate amounts of spindle-shaped hyperchromatic nuclei suggesting atrophied fibrocytes. A few scattered slitlike vessels are seen (hematoxylin-eosin, 770,).

Fig 2. —Nine-day specimens from mouse (hematoxylin-eosin,



4.— Two-month


specimens from



gaard22-23 and Sharkey and Fogh.24 For each tumor, a total of 20

were transplanted. These were divided into four groups of five mice each and consisted of two groups of females (groups 1 and 3) and two groups of males (groups 2 and 4). Groups 1 and 2 were transplanted without hormonal manipulation; groups 3 and 4 were transplanted 24 hours after receiving an intramuscu¬ lar depot injection (10 mg) of testosterone enanthrate in oil (Delatestryl) designed to provide a sustained drug level for 3 weeks. The testosterone injection was repeated in groups 3 and 4 every 2 weeks for a period of 2 months, at which time the mice were killed.


In Vitro Tissue Culture Fresh tumor tissue was taken from operative specimens immediately after removal and washed three times in RPMI1640 solution containing gentamicin sulfate (50 mg/mL) and penicil¬ lin (50 IU/mL) to control potential bacterial contamination. The tissue was then finely minced using a scalpel into 1-mm3 or smaller pieces and placed into plastic flasks containing RPMI 1640 with phenol red supplemented with 10% vol/vol fetal bo¬ vine serum, glutamine, gentamicin sulfate (50 mg/mL), and penicillin (50 IU/mL). Within 4 weeks, a sheet of fibroblastoid

cells covered the entire flasks. No other cellular outgrowth was observed. The cells were trypsinized and recultured in 25-cm2 plastic flasks containing media enriched with testosterone con¬ centrations (0, 102, 103, 104, and 105 ng/dL). For reference, the normal serum testosterone level in the male is 575±150 ng/dL (19.9±5.2 nmol/L) and in the female is 50±13 ng/dL (1.7± 0.5 nmol/L), so our concentrations corresponded to 0, IN, ION, 100N, and 1000N. Cellular growth was assessed by direct cell count using a hemocytometer after trypsinization. The flasks were sequentially trypsinized every 2 or 3 days and growth was followed up for 16 days.


Macroscopic Examination All 60 mice that received transplants lived without any unexpected wasting or death. In each group of five, two were killed 9 days after transplantation and the remain¬ ing three were killed after 2 months. All the treated groups showed the anabolic effects of the testosterone consisting of a significant increase in the size and weight

Downloaded From: http://archotol.jamanetwork.com/ by a University of Manitoba User on 06/10/2015




Fig 5. Two-month specimens from mouse. Well-demarcated and shrunken tumors surrounded by the mouse's subcutaneous tissue, composed of eosinophilic collagenous stroma with focal mineral¬ ized areas and some areas of central necrosis. In one periphery, there is infiltration by chronic inflammatory cells (hematoxylin—





Fig 7. —Growth curve of juvenile nasopharyngeal angiofibroma fibroblasts in vitro, with and without androgen stimulation. Open squares with black dots indicate test with 0 ng/dL of testosterone; closed diamonds, TO2 ng/dL; closed squares with white dots, Iff ng/dL; open diamonds, 70* ng/dL; and closed squares, 105 ng/dL.

'S ò


Fig 6. —Fibroblasts from juvenile nasopharyngeal angiofibroma in tissue culture (unstained Hoffman modulation optics). of the mouse. None of the mice showed obvious growth of transplanted tumor. In all, the transplanted material appeared as a white, fibrous mass that appeared viable. In some of the mice, one or two grossly visible blood ves¬ sels were seen going from the subcutaneous tissue into the graft. No difference in size or appearance of the tumor was noted between groups 1 through 4.

Microscopic Examination Light microscopy was performed on each of the origi¬ nal JNAs (from the patients) and on specimens from the mice at 9-day and then at 2-month intervals from the time of transplantation. No significant difference in micro¬ scopic findings was noted between groups 1 through 4. Nature of Original JNA's (Fig 1).—The rîistological consisted of focal areas of a benign, slightly specimens keratinizing, stratified squamous epithelial surface mu¬ cosa

beneath which is


demarcated vascular-fibrotic tu-

Days 8. Growth curve of juvenile nasopharyngeal angiofibroma fibroblasts in vitro, with and without androgen stimulation. Open squares with black dots indicate test with 0 ng/dL of testosterone; closed diamonds, IO2 ng/dL; closed squares with white dots, 10s ng/dL; open diamonds, IO4 ng/dL; and closed squares, 10s ng/dL.


process. The vascular element revealed small benign vessels varying from a slitlike to round structure. Occa¬ sional large vessels were surrounded by a thin hyalinelike zone containing a spindle cell element. Most of the tumor was made up of a stroma consisting of a collagen tissue matrix with focal hemorrhagic infarcted areas. The stroma contained prominent distribution of spindle- to starshaped cells considered to be of a myofibroblastic nature. These findings are those of a classic JNA. Nine-Day Specimens From Mice (Figs 2 and 3).— Microscopy revealed well-demonstrated tumors commor

Downloaded From: http://archotol.jamanetwork.com/ by a University of Manitoba User on 06/10/2015

posed essentially of an eosinophilic collagenlike stroma containing moderate amounts of spindle-shaped hyperchromatic nuclei suggesting atrophied fibrocytes. A few scattered benign slitlike vessels were noted. Two-Month Specimens From Mice (Figs 4 and 5).— Microscopy revealed small, shrunken tumors, well de¬ marcated by the surrounding subcutaneous tissue, com¬ posed of an eosinophilic collagenlike stroma with focal concentrations of the spindle-shaped hyperchromatic nucleated cells similar to those described in the 9-day specimens. Few areas of focal central necrosis were seen. In addition, there were prominent focal mineralized areas throughout the stroma with some suggestion of early cal¬ cification. In peripheral areas of the tumors, there was in¬ filtration by chronic inflammatory cells, indicating some immune response by the mouse (with time, the nude mouse recovers part of its immunological function). Tissue Culture The cultures obtained from the angiofibroma were composed of elongated fibroblastoid cells (Fig 6). The growth patterns with the different concentrations of tes¬ tosterone are shown in Figs 7 and 8 for two of the JNAs. The third JNA had a similar growth pattern. There was no significant enhancement of growth for the testosteronesupplemented cells. COMMENT This study confirms that JNA transplants in athymic nude mice survive but do not grow. This is not unex¬ pected in view of the fact that, with rare exceptions (such

pleomorphic adenoma), only malignant tumors are re¬ ported to grow in these animals. To our knowledge, there are no reports in the literature on JNA transplantation in nude mice. The sex of the mice and the large doses of in¬ jected testosterone did not affect growth or survivability as

of the tumors.

Although we have succeeded in growing fibroblastoid tissue from the JNA in vitro, no vascular or endothelial cell proliferation was noted in our flasks. This and previous electron-microscopic evidence of nuclear hyperactivity found exclusively in the fibroblasts25 (dense intranuclear granules, large nucleoli and nuclear bodies) suggests that these cells are the reactive component of the angiofi¬ broma. Whether the fibrous component is the site of or¬ igin or a desmoplastic tissue reaction remains elusive. The lack of growth of the JNA with hormonal stimula¬ tion alone suggests that other factors play a complemen¬ tary, if not essential, role in promoting growth of this tu¬ mor. In the absence of a tumor model, we are restricted to study the patients themselves. Because of the rarity of these tumors, a multi-institutional collaborative study may be needed to clarify the biological characteristics of

JNAs. We

gratefully acknowledge

the advice and assistance of Vincent

J. Hyams of The Johns Hopkins Medical Institutions for the histological interpretation of the JNAs in the patients and in the mice. References 1. Schiff M.

Juvenile nasopharyngeal angiofibromas. Laryngoscope. 1959;69:981-1016. 2. Coppa FCW, Irvine G, Timmis S. Four recent cases of juvenile fibroangioma of the last nasal space. J Larnygol Otol. 1961;75:924-931. 3. MacComb WS. Juvenile nasopharyngeal fibroma. Am J Surg. 1973;106:754-763.

4. Neel HB. Juvenile angiofibroma: review of 120 cases. Am J Surg. 1973;126:547-556. 5. Hulthen UL, Toremalm NG. Juvenile nasopharyngeal angiofibroma. J Otolaryngol. 1976;5:55-63. 6. Walike JW, Mackay B. Nasopharyngeal angiofibroma: light and


microscopic changes

after stilbestrol

therapy. Laryngoscope.

1970;80:1109-1121. 7. Duckert LG, Carley RB, Hilger JA. Computerized axial tomography in the

preoperative evaluation of


angiofibroma. Laryngoscope.

1978;88:613-618. 8. Biller HF, Sessions DG, Ogura JM. Angiofibroma: a treatment approach. Laryngoscope. 1974;84:666-706. 9. Butler RM, Nahumn AM, Harifee W. New surgical approach to na-

sopharyngeal angiofibroma.

Trans Am Acad

1967;71:92-104. 10. Conley J, Healty WV,


in the

11. Henderson


Ophthamol Otolaryngol.

Blangrund SM, Erzin KH. Nasopharyngeal juvenile. Surg Gynecol Obstet. 1968;126:824-837.


Patternson CN. Further


in treatment

juvenile nasopharyngeal angiofibroma. Laryngoscope. 1959;79:561\x=req-\

580. 12.


Jakok BW, Nahumn AM, Butler RM, Ward PH. Surgical treatment juvenile nasopharyngeal angiofibroma. Laryngoscope. 1973;83:707\x=req-\

720. 13. Kuttner K,

Katenkamp D, Stiller D. Hormonal treatment of the juangiofibroma. Arch Otorhinolaryngol. 1977;214:331-338. 14. Ward RH, Thompson RW. Juvenile angiofibroma: a more rational therapeutic approach based on clinical and experimental studies. Laryngoscope. 1974;84:2181-2194. 15. Fitzpatrick PJ, Rider WD. The radiotherapy of nasopharyngeal angiofibroma. Radiology. 1973;109:171-178. 16. Johnsen S, Kloster JH, Schiff M. The action of hormones on juvenile nasopharyngeal angiofibroma. Acta Otolaryngol. 1966;61:153-160. 17. Briart TDR, Fitzpatrick PJ, Berman J. Nasopharyngeal angiofibroma: a twenty year study. Laryngoscope. 1978;88:1242-1251. 18. Johns ME, MacLeod RM, Cantrell RW. Estrogen receptors in nasopharyngeal angiofibromas. Laryngoscope. 1980;90:628-634. venile

19. Lee DA, Ramanath R, Meyer JS, Prioleau PG, Bauer WC. Hormonal receptor determination in juvenile nasopharyngeal angiofibromas. Cancer. 1980;46:547-551. 20. Antonelli AR, Cappiello J, DiLorenzo D, Donajo CA, Nicolai S.

Diagnosis, staging and treatment of juvenile nasopharyngeal angiofiLaryngoscope. 1987;97:1319-1325. 21. Faraj MM, Ghanimah SE, Rajai A, Saleem TH. Hormonal receptors in juvenile nasopharyngeal angiofibromas. Laryngoscope. 1987;97:208\x=req-\

broma (JNA).

211. 22. Povlsen CO.

Heterotransplantation of humane maligne tumorer til nude mus. Copenhagen, Denmark: FADL; 1977. 23. Povlsen CO, Rygaard J. Heterotransplantation of human epidermal carcinoma to the mouse mutant nude. Acta Pathol Microbiol.

1971;80:713-717. 24. Sharkey FE, Fogh J. Considerations in the use of nude mice for cancer research. Cancer Metastasis Rev. 1984;3:341-360. 25. Stiller D, Katenkamp D, Kuttner K. Cellular differentiation and structural characteristics in nasopharyngeal angiofibromas; electron-microscopic study. Virchows Arch A. 1976;371:273-282.


Editorial Footnote Our reviewers noted the fact that it appears through the examinations that the transplantation process itself affected the viability of the tumors, or at least that it may have done so. The complex biochemical phenomena as¬ sociated with a host's response to foreign tissue, even in nude mice, must be considered along with androgenic stimulation in this type of study. While failure to grow juvenile nasopharyngeal angiofibroma tumor in the mouse model should not have been unexpected (it is unusual to be able to grow benign transplants of any kind), the apparent absence of any androgenic effect is still useful infor¬ mation. One of the most difficult aspects of interpreting these observations is the lack of clarity as to whether the "fibroblas¬ toid" cells are actually the angiofibroma cells. Although the information in this article may ultimately serve to increase our understanding of juvenile nasopharyngeal an¬ giofibroma, it appears at present that the main conclusion to be drawn is that we still lack an acceptable experimental model for evaluating the role of androgenic stimulation in angiofibromas.

sequential histological


J. Bailey, MD Galveston, Tex

Downloaded From: http://archotol.jamanetwork.com/ by a University of Manitoba User on 06/10/2015

Juvenile nasopharyngeal angiofibroma tumor models. Failure of androgens to stimulate growth in nude mice and in vitro.

Juvenile nasopharyngeal angiofibroma is a tumor with a predilection for adolescent boys. It has been shown to contain cytosolic androgen receptors and...
6MB Sizes 0 Downloads 0 Views