investigated the expressions of BMP subtypes in human salivary adenocarcinoma cell line. (HSG-S8), tongue squamous cell (HSC-4) and gingival squamous ...
Vol. 42, No. 3, July 1997
BIOCHEMISTRY and MOLECULAR BIOLOGY INTERNATIONAL Pages497-505
EXPRESSION OF BONE MORPHOGENETIC PROTEINS OF HUMAN NEOPLASTIC
EPITHELIAL CELLS Setsuko Hatakeyama*, Yu-Hao Gaow Yuko Ohara-Nemoto ~, Hiroaki Kataokaa', and Masanobu Satoh Departments of Oral Pathology, and ~Oral Microbiology, School of Dentistry, Iwate Medical University, Uchimaru 19-1, Morioka, 020; ~The Fourth Military Medical University, 1 Kang Fu Road, Xi'an Shaanxi Province, 710032 People's Republic of China; eThe second Department of Pathology, School of Medicine, Miyazaki Medical College, 5200 Kihara, Kiyotake, Miyazaki 889-16, Japan. Received April 3, 1997
SUMMARY: Bone morphogenetic proteins (BMPs) are crucial factors of osteogenesis. We
investigated the expressions of BMP subtypes in human salivary adenocarcinoma cell line (HSG-S8), tongue squamous cell (HSC-4) and gingival squamous cell (Ca9-22) carcinoma cell lines, gastric poorly differentiated adenocacinoma cell (MNK45) and signet ring cell (KATOIII) carcinoma cell lines, rectal adenocarcinoma (RCM-1, RCM-2, and RCM-3), and thyroid (8505C) and bladder (T24) carcinoma cell lines by reverse transcription-polymerase chain reaction (RT-PCR). RT-PCR disclosed that BMP-1 was expressed in all cell lines examined, and BMP-2 was amplified in almost all cells except MKN45. Two squamous cell carcinomas, HSC-4 and Ca9-22, and KATOIII expressed only BMP-1 and BMP-2. MKN45 did not express BMP-2, but expressed BMP-7 and weakly BMP-4 and BMP-5. In addition to the expression of BMP-1 and BMP-2, three rectal adenocarcinoma cell lines commonly expressed BMP-7, and HSG-S8 expressed BMP-6. These findings indicated that the neoplastic epithelial cells possessed a rather great potency to express BMP mRNAs. On the other hand, among these carcinoma cells, HSG-S8 solely induced bone in nude mouse tumors, and HSC-4 and KATOIII contained many calcified masses in tumors while the rest did not induce either. Key words: bone morphogenetic protein, ectopic ossification, reverse transcription-polymerase
chain reaction, neoplastic epithelial cells. INTRODUCTION The BMP proteins were originally identified by their ability to induce bone when implanted subcuianeously in rodents (1). Wozney et aL (2) cloned the four novel proteins with bone-inductive activity from bovine bone extract. They called these new proteins bone morphogenetic protein (BMP)-I, -2A, -2B (renamed as BMP-4), and -3. By successive studies, more than ten subtypes of BMPs have been totally cloned by screening cDNA and genomic libraries of human
This work was supported in part by a Grant from the Japan China Medical Association in 1995. The abbreviations used are: BMP, bone morphogenetic protein; RT-PCR, reverse transcriptionpolymerase chain reaction *To whom correspondence should be addressed. Fax: 81-19-652-4131,
1039-9712/97/030497-09505.00/0
497
Copyright 9 1997 by Academic Press Australia, All rights of reproduction in any,~rm reserved.
Vol. 42, No. 3, 1997
BIOCHEMISTRYGnd MOLECULAR BIOLOGY INTERNATIONAL
tissues (3). Molecular structure and localization of BMP-2 through BMP-7 have been previously reviewed (4). These subtypes are members of the transforming growth factor (TGF)- fl superfamily which is a highly conservative one of extracellular polypeptide signaling molecules. BMPs appear to be of considerable importance in bone and cartilage development, because they have been localized in the sites of intramembranous and endochondral bone formation (5, 6). In addition, expressions of BMPs have been shown not only in bone matrix, but also in nonosseous tissues such as the kidney and brain (7), and embryonal carcinoma cells (8). A substantial amount of evidence suggests that the BMPs are involved in multiple developmental processes during embryogenesis involved in the specification of ventral mesoderm (9). On the other hand, the functions of BMPs in epithelial tissues remain to be completely specified. In previous histopathological studies, ectopic bone formation has been observed in the rare cases of varied tumor tissues and metastasis of extraskeletal origin (10). It has been an interesting question by what mechanism bone formation is induced in nonosseous tumor tissues originating from gastric, colorectal, and bladder epithelia (11, 12).
It was speculated that unknown
factors released from tumor cells stimulate osteogenesis in the stroma (13).
Recently, BMPs
have been selected as the candidate for bone-inductive activities produced by epithelial tumor cells, which induce ectopic bone formation (14, 15). Human salivary adenocarcinoma cell line, HSG-S8, synthesized and secreted BMP-2, which induced bone formation in tumor tissues (14). The expression of B M P - 2 in HSG-S8 cells was increased by the treatment of retinoic acid, accompanied by the growth inhibition (16). In this paper we examined the expressions of BMPs mRNA of neoplastic epithelial cells originating from the stomach, rectum, bladder, and thyroid gland, and discussed their roles in ectopic bone formation. MATERIALS AND METHODS Human neoplastic epithelial cell lines, HSC-4, Ca9-22, MKN45, KATOIII, 8505C and T24 were provided by the Japanese Cancer Research Resources Bank (JCRB). HSC-4 (17) is a lingual, and Ca9-22 is a gingival squamous cell carcinoma cell line. MKN45 (18) is a gastric undifferentiated adenocarcinorna cell line. KATOIII (19) is a signet ring cell carcinoma cell line. RCM-1 (20), RCM-2, and RCM-3 (21) are rectal adenocarcinoma cell lines. 8505C is a thyroidal adenocarcinoma cell line, and T24 (22) is a urinary bladder carcinoma cell line. HSG-S8 is a subclone (14) of salivary adenocarcinoma cell line (23). The culture media of Eagle's minimum essential medium supplemented with 10% FBS was prepared for HSC-4, Ca9-22, 8505C and T24, RPMI1640 supplemented with 10% FBS for MKN45, RPMI: Ham's F12 (1:1) supplemented with 10% FBS for KATOIII and RCM-1, RCM-2 and RCM-3. Chemically defined serum-free medium (SFM101, Nissui, Japan) was used for HSG-S8. Cell culture:
498
Vol. 42, No. 3, 1997
BIOCHEMISTRYond MOLECULAR BIOLOGY INTERNATIONAL
Reverse transcription-polymerase chain reaction (RT-PCR): PoIy(A)+RNA (0.1/zg) was converted to cDNA by reverse transcription using 100 units Molony murine leukemia virus reverse transcriptase (Pharmacia, Sweden) in the presence of 12.5 nmol random hexamer primers and the four dNTPs (0.126mM) in a buffer containing 50 mM Tris-HC1, pH 8.3, 75 mM KCL, 3 mM MgCL, 10 mM dithiothreitol, and 12.5 units ribonuclease inhibitor in a total volume of 20/A The reaction was incubated at 37 C for 1 h. The primers used for amphficatlon of BMP-2 were AATI'CCCCGTGACCAGACTIT(877-897) and CTAGCAATGGCCTTATCFGTG (11161096), and those of B M P - 4 were T T T G A T A C C T G A G A C G G G G A A ( 4 6 9 - 4 8 9 ) and TGTCCAGTAGTCGTGTGATGA (998-978). Each primer for BMP-1, BMP-3, BMP-5 or BMP-6 was synthesized according to the report of Bentley et aL (24), that of BMP-7 was made according to Ogose et al. (15), and primers for G3PDH were according to Ercolani et al. (25). The PCR reaction was carried out in a total volume of 12.5 /~1 comprising cDNA preparation (0.1 /~g), lx reaction buffer (Pharmacia, as 10x with Taq DNA polymerase), dNTPs (0.128 mM each), 0.025 units Taq DNA polymerase (Phamaasia) and 5 pmol of each primer. PCR conditions of BMP-1,-3,-5, -6, and - 7 were denaturing at 94~ for 1 min followed by 23-28 cycles at 94~ for 1 min, 60~ for 1.5 min, and at 72~ for 30 s, and extension at 72~ for 6 rain. The amplification conditions for BMP-2 and G3PDH were denaturation at 94~ for 1 min followed by 20 cycles of denaturation at 94~ for 1 min, annealing at 55~ for 30 s, and extension at 72~ for 30 s, and those of BMP-4 denaturation at 94~ for 2 min and followed by 40 cycles of denaturation at 94~ for 1 min, annealing at 60~ for 2 rain, and extension at 72~ for 2 min. These cycles were further followed by one cycle of 72~ for 6 min. The reaction products were analyzed by electrophoresis on 1.8% agarose gels containing 0.1 #g ethidium bromide/ml. Negative control reactions without cDNA were carried out on each primer set and no product was found (data not shown). Heterotransplantation into athymic nude mice and immunohistochemistry: Cells (10 7) were subcutaneously injected into the back of athymic nude mice. The same number of RCM-3 ceils were also injected under the capsule of the spleen. At one month after injection, tumors were removed, and fixed with phosphate-buffered formalin. Then, they were processed rountinely, and embedded in paraffin. Specific antibodies against BMP-2 (W12, Wll), which were given by Genetics Institute (Boston, USA) and type IV collagen (CHEMICON, USA) were used. Each of the molecules was detected by the strcptavidin-biotin complex method using Pathostain kit (WAKO, Japan). z~
.
.
o
.
.
.
.
"
RESULTS PCR products were considered to be validated when a clean product of predicted cDNA size was observed after amplification. The cycle numbers which were selected had the most unsaturated bands. RT-PCR revealed that the neoplastic epithelial cells expressed several subtypes of BMPs with a relatively wide spectrum (Fig. 1, Table 1). All cell lines expressed B M P - 1 to various degrees. BMP-2 was also expressed in all of the cell lines except MKN45. Among these cell lines, two squamous cell carcinomas (HSC-4, Ca9-22) and gastric signet ring cell carcinoma (KATOIII) expressed only BMP-1 and BMP-2. Three rectal adenocarcinoma cell lines commonly expressed BMP-1, BMP-2 and BMP-7. In addition to these BMP subtypes, RCM-1 expressed BMP-4 and BMP-6 weakly, and RCM-3 expressed BMP-5, while RCM-2 expressed all of the other subtypes except BMP-3.
Salivary adenocarcinoma cell line, HSG-
$8, was shown to express BMP-1, BMP-2, BMP-6, and BMP-7 weakly. Also thyroid (8505C)
499
Vol. 42, No. 3, 1997
BIOCHEMISTRYand MOLECULAR BIOLOGY INTERNATIONAL
-~'-'~" , ~~J-~~,~q~ ,~" ,-6 ~ ~,.v,t.~" c.'~" &" ~ . ~ ~ ~- " x,,~ c..'J ~.'~q,.'~
BMP-1 BMP-2 BMP-3
BMP-4
BMP-5
BMP-6
BMP-7
G3PDH
F i g . 1. D e t e c t i o n o f B M P s ase chain reaction.
Table
expression
in epithelial tumor cells by reverse
1. E x p r e s s i o n s o f B M P m R N A BMP-I
BMP-2
BMP-3
HSC-4
+
+
.
Ca9-22
+
+
.
MKN45
+
-
-
KATOIII
+
+
.
transcription-polymer-
in C u l t u r e d C e l l s b y R T - P C R BMP-4 .
BMP-5 .
.
.
. +
.
BMP-7
. .
• .
BMP-6
.
-
+
.
RCM-I
+
+
-
+
-
-+
+
RCM-2
+
+
--
+
+
+
+
RCM-3
+
+
-
-
§
-
+
8505C
+
+
-
-
+
+
-
T24
+
+
+
--
-+
-
-
HSG-S8
+
+
-
-
+
•
(23)*
(19)
*: N u m b e r s
in p a r e n t h e s e s
-+: w e a k l y p o s i t i v e . RT-PCR was performed
(26)
(40)
(29)
are the cycle conditions.
at l e a s t t h r e e t i m e s w i t h e a c h s a m p l e .
500
(25)
(28)
Vol, 42, No. 3, 1997
BIOCHEMISTRY and MOLECULAR BIOLOGY INTERNATIONAL
and b l a d d e r (T24) carcinoma cell lines expressed several subtypes of BMPs. These resuhs suggested that the neoplastic epithelial cells possessed a rather great potency to express B M P mRNAs. Table 2 shows the take rates of tumor ceils subcutaneously transp|anted into the nude mice and tumor sizes. All the epithelial tumor cells except KATOIII proliferated with doubling times from 30 to 36 hours. K A T O I I I cells did not elongate on the plastic dishes and were r o u n d - s h a p e d . They proliferated most slowly with a doubling time of about 41 hours.
Though R C M - 3 e l o n -
gated on the plastic dishes, it also proliferated slowly in the monolayer culture and did not form a t u m o r in subcutaneous sites. Therefore, R C M - 3 cells were injected under the capsule o f the spleen, where the cells grew and formed a small tumor. After one month of transplantation, both H S C - 4 and KATOIII cells were difficult to grow in nude mice, and barely m a d e small tumors. However, many calcificated masses were observed in these tumors, though bone tissues were not induced. Calcificated masses were defined by yon Kossa's staining (Fig. 2A). I m m u n o h i s t o c h e m i c a l study with antibody against B M P - 2 confirmed the existence of the B M P - 2 protein in the cytoplasm of H S C - 4 and KATOIII (Fig. 2B) cells around the masses.
On the
other hand, the tumor tissues formed with H S G - S 8 included apparent bone tissues o f various sizes. The c y t o p l a s m of H S G - S 8 was weakly positive to a n t i - B M P - 2 antibody, while bone
Table 2. Doubling Times and Take Rates of Tumor Cells and Histology of Nude Mouse Tumors
Cells
Doubling time* No . of tumors/ (h) No. of transplantation ~
HSC-4 Ca9-22 MKN45 KATOIII RCM-1 RCM-2 RCM-3
8505C T24 HSG-S8
35.4 35.6 30.9 41.3 34.4 NE NE
34.2 30.9 32.3
3/4 3/3 4/4 1/4 5/5 t/1
(75%) (100%) (100%) (25.0%) (100%) (100%)
2/2
(I00%)'
Weight of tumors (g) 0.18 0.33 0.38 0.11 0.74 0.54
4/4 (100%) 1/6 (11.6%) 7/8 (87.5%)
+ 0.21 -++ 0.47 -+ 0.18 -+ 0.22
0.55 +_ 0.35 1.29 0.40 _+ 0.30
* Means of values in two independent experiments. NE : not examined 8 subcutaneous transplantation '- intrasplenic transplantation
501
Calcification
hmnunostain for BMP-
+ + -
+ + -
--
--
-
-
-+ (bone)
+
VOI. 42, NO. 3, 1997
BIOCHEMISTRYond MOLECULAR BIOLOGY INTERNATIONAL
Fig. 2. Histology of nude mouse tumors. (A) Calcified masses in KATOIII nude mouse tumor are stained black by von Kossa's stain. (x300) (B) Immunohistochemical stain for BMP-2 at the same site. A small number of cells around calcified masses are positive to B M P - 2 antibody. (x300) (C) Immunohistochemical detection of BMP-2 at the site of bone in HSG-S8 tumor. (x300) (D) Immunohistochemical stain for type IV collagen of the bor, e tissue in the tumor of
HSG-S8. (x300)
tissue was clearly positive (Fig. 2C). Immunostaining with the antibody against type IV collagen showed the localization of type IV collagen in the marginal site of bone tissues (Fig. 2D). The nude mouse tumor cells formed by other cell lines induced neither bone nor calcification. In spite of the BMP-2 mRNA expression in vitro, BMP-2 protein was not detected in these tunlor cells by immunohistochemistry (Table 2). DISCUSSION This study demonstrated that the neoplastic epithelial cell lines originating from gastrocolic organs expressed various BMPs covering BMP-1 through BMP-7, and that the epithelial tumor ceils from the same organ did not always show the same pattern of the expression of BMP subtypes. In previous studies, B M P - 2 and BMP-4 appeared to have a similar activity to induce cartilage and bone in the rat ectopic assay system (25, 26). Also, BMP-6 (27), BMP-7 (28), and BMP-5, although this showed a lower activity than BMP-2 (29), were ascertained to be osteoinductive without the coexistence of other BMPs. From this evidence, it was expected that all of the epithelial tumor cells had the capacity to induce bone formation in nude mouse
502
Vol. 42, No. 3, 1997
BIOCHEMISTRY and MOLECULAR BIOLOGY INTERNATIONAL
tumor. However, in spite of the expression of BMP-2 and/or other subtypes mRNA in all cells examined, the subcutaneous transplantation system showed only the osteoinduction in HSG-S8 tumor, and calcification in HSC-4 and KATOIII tumors. To explain the failure of bone formation in tumor tissues except the one formed with HSG-S8, three possible reasons have been generally offered: (a) an insufficient amount of BMP proteins to stimulate the targeting cells was released from the cells, (b) responding cells to BMP did not reach the BMP periphery, (c) the extracellular matrix trapping BMP molecules and causing their gradual diffusion to the periphery was not present. As shown in Table 2, both calcification and bone induction in the tumors were closely related with the immunoreaction to B M P - 2 in vivo. Since the cytoplasms of the tumor cells except H S G - S 8 , H S C - 4 and KATOIII were scarcely immunostained with the antibody to BMP-2, these cells may not secrete BMP-2 protein or secrete a very small amount. Althenatively, the expression of BMP genes may be in some carcinoma cells modulated under the subcutaneous transplantation condition. It was also taken into account that the subcutaneous sites contained a relatively small number of responding ceils to BMPs. In fact, we have obtained the data indicating the bone formation by MKN45 cells, when they were transplanted into the thigh muscle of a nude mouse (S. Hatakeyama, unpublished data). The muscle tissue is thought to contain more responding cells than the subcutaneous sites (14). In addition, the existence of the extracelluar matrix trapping BMPs is noticeable. That is, if cells are not covered with extracellular matrix, BMPs secreted from the cells are easily hydrolyzed with extracellular protease and/or dispersed. As one of the extracellular matrix molecules, type IV collagen may be important for the action of BMPs, because BMP-7 was found to bind strongly to it (31). In our results, the positive sites of BMP-2 and type IV collagen were similarly the margin of bone. This finding suggested that type IV collagen was related to the trapping of BMP proteins. We demonstrated that BMP-1 were expressed in all cell lines examined. Recently, BMP-1 has been shown to be identical to procollagen C-proteinase which cleaves the COOH-terminal of procollagen I, II, and III (32). This activity of BMP-1 is expected to work effectively for osteoinduction to provide mature collagen. Coexistence of BMP-1 and the other BMPs may be beneficial for BMPs to act effectively.
503
Vol. 42, No. 3, 1997
BIOCHEMISTRYand MOLECULAR BIOLOGYINTERNATIONAL
In past reports, neoplasms originating from epithelial tissues of such organs as the stomach (10, 11), rectum (12), colon (13), and thyroid (33) infrequently involved bone tissues and calcispherites. Our results indicated that neoplastic epithelial cells originating from these organs express multiple subtypes of BMPs, which indicates that these carcinoma cells possess the latent potency for ectopic bone formation and calcification found in tumor tissues. ACNOWLEDGMENT: We thank the Genetics Institute (Boston, USA) for polyclonal antibodies
to human BMP-2. REFERENCES 1. Urist, M. R. (1965) Science 150, 893-899. 2. Wozney, J. M., Rosen, V., Celeste, A. J, Mitsock, L. M., Whitters, M. J., Kriz, R. W., Hewick, R. M., and Wang, E. A. (1988) Science 242, 1528-1534. 3. Dube, J. L., and Celeste, A. J. (1996) J. Bone Miner. Res. 11, Suppl. $382. 4. Wozney, J. M. (1993) Cellular and Molecular Biology of Bone (Masaki Noda Ed.) pp. 131167, Academic Press. 5. Vukicevic, S., Paralkar, V. M., Cunningham, N. S., Gutkind, J. S., and Reddi, A. H. (1990) Dev. Biol. 140, 209-214. 6. Yang, L. and Yan, J. (1990) Clin. Orthop. 257, 249-256. 7. Ozkaynak, E., Schnegelsberg, P. N. J., and Oppermann, H. (1991) Biochem. Biophys. Res. Communn. 179, 116-123. 8. Rogers, M. B., Rosen, V., Wozney, J. M., and Gudas, L. J. (1992) Mol. Biol. Cell 3, 189196. 9. Dale, L., Howes, G., Price, B. M. J., and Smith, J. C. (1992) Development 115, 573-585. 10. Yasuma, T., Hashimoto, K., Miyazawa, R., and Hiyama, Y. (1973) Acta Path. Jap. 23, 155172. 11. Rhone, D. P., and Horowitz, R. N. (1976) Cancer 38, 1773-1780. 12. Ansari, M. Q., Sachs, I.L., Max, E., and Alpert, L. C. (1992) Dig. Dis. Sci. 37, 1624-1629. 13. Randall, J. C., Morris, D. C., Tomiya, T., Anderson, H. C. (1989) Human Pathol. 20, 8688. 14. Hatakeyama, S., Ohara-Nemoto, Y., Kyakumoto, S., Satoh, M. (1993) Biochem. Biophys. Res. Commun. 190, 695-701. 15. Ogose, A., Motoyama, T., Hotta, T., and Watanabe, H. (1996) Pathol. Int. 45, 9-14. 16. Hatakeyama, S., Ohara-Nemoto, Y., Kyakumoto, S., Satoh, M. (1996) Biochem. Mol. Biol. Int. 38, 1235-1243. 17. Yamamoto, T., Kamata, N., Kawano, H., Shimizu, S., Kuroki, T., Toyoshima, K., Rikimaru, K., Nomura, N., Ishizaki, R., Pastan, I., Gamou, S., Shimizu, N. (1986) Cancer Res., 46, 414416. 18. Motoyama, T., Hojo, H., and Watanabe, H. (1986) Acta Pathol. Jpn. 35, 65-83. 19. Sekiguchi, M., Sakakibara, K., and Fujii, G. (1978) Japan J. Exp. Med. 48. 61-68. 20. Kataoka, H., Nabeshima, K., Komada, N., and Koono, M. (1989) Virchows Arch. [B] 57, 157-165. 21. Kataoka, H., Seguchi, K., Kuroki, T., and Koono, M. (1993) Human Cell 6, 131-132. 22. Bubenik, J., Baresova, M., Viklicky, V., Jakoubkova, J., Sainerova, H., and Donner, J. (1973) Int. J. Cancer 11, 765-773. 23. Shirasuna, K., Sato, M., and Miyazaki, T. (1981) Cancer 48, 745-752. 24. Bentley, H., Handy, F. C., Hart, K. A., (1992) Br. J. Cancer (1992) 66, 1159-1163. 25. Ercolani, L., Florence, B., Denaro, M., and Alexander, M. (1988) J. Biol. Chem. 263, 1533515341. 26. Wang, E. A., Rosen, V., D'Alessandro, J. S., Bauduy, M., Cordes, P., Harada, T.. Israel, D., Howick, R. M., Kems, K., LaPan, P., Luzxenberg, D. P., McQuaid, D., Moutsatsos, I., Nove, J., and Wozney, J. M. (1990) Proc. Natl. Acad. Sci. USA 87, 2220-2224.
504
Vol. 42, No. 3, 1997
BIOCHEMISTRYand MOLECULAR BIOLOGY INTERNATIONAL
27. LaPan, P., Baudy, M., Cox, K. A., D'Alessandro, J. S., Israel, D. I., Nove, J., Rosen, V., Wozney, J. M., Moutsatsos, I. K., and Wang, E. A. (1991) J. Bone Miner. Res., Suppl. $153. 28. Gitelman, S. E., Kobrin, M. S., Ye, J-Q., Lopez, A. R., Lee, A., and Derynck, R. (1994) J. Cell Biol. 126, 1595-1609. 29. Sampath, T. K., Ozkaynak, E., Jones, W. K., Sasaki, H., Tucker, R., Tucker, M., Kusmik, W., Lightholder, J., Pang, R., Corbett, C., Pooermann, H., and Rueger, D. C. (1992) J. Bone Miner. Res. 6, $155. 30. Cox, K., Holtrop, M., D'Alessandro, J. S. D., Wang, E., Wozney, J. M., and Rosen, V. (1991) J. Bone Miner. Res., Suppl. $155. 31. Paralkar, V. M., Nandedkar, A. K. N., Pointer, R. H., Kleinman, H. K., and Reddi, A. H. (1990) J. Biol. Chem. 28, 17281-17284. 32. Kessler, E., Takahara, K., Biniaminov, L., Brusel, M., and Greenspan, D. S. (1996) Science 271, 360-362. 33. McNicol, A. M~ (1992) in Endocrine Pathology in Muir's Text Book of Pathology, 13th ed. (MacSween RNM and Whaley K, Eds.) pp. 1075-1106, Edward Arnold, London.
505
