lsoantigens A, B, H(0) in Nasopharyngeal Carcinoma

RAJ K. GUPTA, MD, FACP, FIAC*

Fifty cases of nasopharyngeal carcinoma (NPC) and its metastasis in ten lymph nodes were studied for specific red cell adherence (SRCA) to determine if any loss of blood group isoantigens was present to explain some of the differences such as their preponderance in local Chinese population and variable clinicopathologic behavior. While a patchy positive SRCA was noted on the surface of tumor cells in all cases of NPCs, it was not so in any of their metastasis. The findings suggest that in NPCs a few tumor cells may be capable of partially retaining isoantigens, and that this is neither influenced by their histologic type nor in any way seems to account for the variable prognosis. While the exact mechanism for negative SRCA in metastatic deposits in lymph nodes is unknown, it is possible that metastatic tumor cells may have lost isoantigens on reaching lymph nodes. The remote possibility of a 'preferential discharge' of isoantigen-free tumor cells from the primary tumor seemed unlikely in view of a mildly positive SRCA in all the NPCs.

Cancer 51:256-259, 1983.

ASOPHARYNGEAL CARCINOMAS (NPCs) are consid- N ered somewhat unique tumors for several reasons. They are known to show a high incidence among Chinese of Southeast Asia, Alaskan natives (Eskimos, Indians, Aleuts) and Canadian Eskimos, while it is a relatively rare tumor in Caucasians and other races.'-' It is also of some interest to indicate that, in the Chinese population of Singapore with these tumors, a signifi- cantly higher frequency of HLA A2, a second locus blank (Sin-2), and a lymphocyte-determined antigen (Sin-2a) have been found in comparison with the normal Chinese population.' Recent studies also suggest that NPCs may be related to Epstein-Barr (EB) v i r ~ s " ' ~ - ' ~ and in patients with NPC-impaired cell-mediated im- mune functions (T-cell functions)."

The histologic pattern of NPC is usually pleomorphic. Such appearances have resulted in some confusion in regard to its histogenesis and nomenclature. The World Health Organization (WHO) regards all these tumors as belonging to a homogenous group arising from squa- mous epithelium, and has proposed three histologic sub- types, i.e., well differentiated squamous cell carcinomas, nonkeratinizing carcinomas, and undifferentiated car-

cinomas." It has also been proposed that the histologic subtypes of NPCs may be associated with certain bio- logical differences, thus accounting for the variations in survival rates and responses to therapy."

Since the clinical and pathologic behavior of these tumors and their preponderance in ethnic Chinese are well known, the current study was designed to ascertain if any variations in tumor cell surface antigenic com- ponents related to blood group isoantigens A, B, H(0) were present in the various histologic subtypes of NPCs and/or their metastasis to explain the differences in their natural progression.

Materials and Methods

A total of 50 selected biopsy specimens from histo- logically proven NPCs was included in this study. The tumor types included were in accordance with the re- cently proposed WHO c1assification," and comprised five cases of the well differentiated squamous cell type, 18 of the nonkeratinizing type, and 27 of the undiffer- entiated type of NPCs. In addition, ten lymph nodes from these cases with their metastasis were also included. The site of excised lymph nodes were either from the cervical chain from the immediate drainage site of the tumor. or further below from areas not constituting a

From the Department of Pathology. Faculty of Medicine, National

* Associate Professor of Patholoav. University of Singapore. Republic of Singapore.

1. - Address for reprints: Raj K. Gupta, MD, Division of Laboratory

Services, Wellington Hospital, Wellington 2, New Zealand. The author gratefully acknowledges the support of Professor K.

Shanmugaratnam during the period ofthis investigation. The technical assistance of c. K. ow, photographic assistance of T. C. Tan, and typing and secretarial assistance of Ms. A. T. Tan and Mrs. J. M. Mooney are also acknowledged.

direct anatomical site from the nasopharynx. N~ other tissue from Other sites in any Of these cases since such biopsy material was not available in view of no clinically demonstrable metastasis.

The presence or loss of isoantigens A, B, H(0) was studied using the original mixed cell agglutination re-

be

Accepted for publication November 16. 198 I .

0008-543>(/83/0 I IS/02S6 $I .OO GI American Cancer Society

256

No. 2 NASOPHARY NGEAL CA - Gupta 257

action (MCAR) technique of Davidsohn.” This tech- nique was preferred because of its applicability on for- malin fixed material apart from the simplicity and re- liable sensitivity. Five micron sections from all tissues were cut from formalin-fixed paraffin-embedded mate- rial and mounted on clean glass slides. Human antisera (Anti A & Anti B) with agglutination titres of 5 12 were used.”

An Ulex Europeus extract (F. W. Shumacher, Sand- wich, MA) with an agglutination titre of above 2000 was used as a source of 0 agglutinin. Details of method of preparation of this extract have been previously de- scribed.lg The indicator system included erythrocytes of Groups A, B, AB and 0. These were washed three times in tris-buffered isotonic saline at pH 7.4 and a 1% sus- pension was made in the same saline. MCAR was per- formed on all of the tissues using the same procedures as have been previously published.lg All of the tissues were studied using a Leitz microscope through the full thickness of the slide with the tissue section remaining on the lower surface. Microphotographs were taken with a Leitz (E. Leitz, D6330, Germany) microscope equipped with an Orthomat Camera using a panatomic-x Kodak black-and-white film.

A variety of tissues, including squamous epithelium of the cervix, adenomas of the breast, and blood vessels in the tissues served as positive controls, since these are known to contain isoantigens A. B, H.’9-26

Results

The intensity of reaction was graded based on specific red cell adherence (SRCA) to the tumor cell surfaces. The reaction was considered as being strong when many erythrocytes were noted all over the individual tumor cell surface, while it was considered as moderate when the specific adherence of erythrocytes was only seen in moderate numbers. A mild reaction was suggested when the erythrocytes were distributed in a patchy manner on only some tumor cell surfaces, leaving other cells devoid of them. A negative reaction was recorded when no red cell adherence was observed on any of the tumor cell surfaces.

Patchy positivity was found in all of the 50 (100%) NPCs, irrespective of the histologic type (Figs. 1 and 2). The intervening stroma between tumor cells consistently showed a negative reaction.

All of the ten lymph nodes from these patients with metastatic deposits of NPCs showed a uniformly nega- tive reaction (Fig. 3). The various tissues, including squa- mous epithelium of the uterine cervix, adenomas of breast (Fig. 4), and blood vessels in all tissues served as positive controls and were always found to show a strongly positive reaction. In all cases studied, it was further noted that erythrocyte agglutination only oc-

curred when patients and indicator cells possessed com- mon A, B. 0 antigens.

Discussion

Several studies in recent years have shown that blood group isoantigens, normally present in epithelial tissues, are lost during malignant tran~formation.’”’~ In an early study, Kovarik et aL19 used a variety of normal tissues and benign and malignant tumors of squamous and transitional cell types and showed that, while isoan- tigens A, B, H(0) were found to be absent or reduced in the majority of cases of carcinomas, these were always present in normal tissues and benign tumors. Studies of Davidsohn ec a1.,20-22 in cases of carcinoma of the cervix, lung, and their metastasis, have also shown significant reduction and loss of isoantigens in a number of their cases, and suggested a possible relationship in the loss of isoantigens in these tumors and their ability to me- tastasize. Gupta et al.23-2’ studied adenocarcinomas of the prostate, breast, and endometrium, and showed an almost complete loss of isoantigens in all their cases, and suggested that the loss of isoantigens was much greater in adenocarcinomas in comparison with carci- nomas of the cervix and lungs. A sudden destruction or masking of antigens, decrease in serologic activity of A, B, 0 blood groups, and a loss of secretory products in the derivative cells of adenocarcinomas, was suggested by these workers as a possible mechanism. In another recent study by Mathew and Gupta,26 a much higher degree of positivity indicating the presence of isoantigens has been shown in 88% of the cases of cholangiocarci- nomas. It has been suggested26 that, in mucus-secreting adenocarcinomas, e.g., cholangiocarcinomas. although no definite correlation can be established between his- tologic grading or metastatic potential, there may be some differences in the mucus-secreting adenocarcino- mas in comparison with other types of adenocarcino-

Such differences may be due to the presence of two biochemically different forms of blood group an- tigens: an alcohol-soluble form (extracted during the processing of tissue) and a water-soluble form, the latter being wholly or partially preserved in cholangiocarci- nomas. Limas rt ~ 1 . ~ ’ studied periodic follow-up biopsy specimens for a five-year period in cases of transitional cell carcinomas of the urinary bladder, and have shown that 8 1 % of the patients with presence of isoantigens in the initial biopsies of tumor did not develop an invasive carcinoma, and 27% of these had no recurrences. In the same study it was also shown that, in patients in whom a negative reaction (loss of isoantigens) in the initial biopsy was noted, recurrences of tumors did occur, and 62% of the cases developed invasive carcinomas of the urinary bladder. Interestingly, 8 1% of these cases with

258 CANCER January 15 1983 Vol. 51

recurrent tumors had also shown a negative reaction in serial biopsies. Limas et uL2’ concluded that the presence of readily detectable isoantigens probably does correlate with a favorable prognosis, while its absence denotes an aggressive potential of transitional cell carcinomas of the urinary bladder.

The results of this study show that, in all of the 50 cases of NPCs, a mild-positive SRCA is invariably found. Such a finding seems to be at variance with that of other investigators who have reported negative SRCA

FIG. 1. Histologic features of NPC from a patient diagnosed as having undifferentiated nasopha- ryngeal carcinoma (H & E, X200).

FIG. 2. Above case showing a patchy positive SRCA on tumor cells (MCAR X200).

in the majority of a d e n o c a r c i n ~ m a s ~ ~ - ~ ~ and in as much as 90% of invasive squamous cell carcinomas of the lung and The finding of a consistently mild-pos- itive SRCA in all of the cases of NPCs in the current study, although suggestive of a partial retention of isoan- tigens, nevertheless seemed obscure, and the exact rea- sons for this are difficult to explain. Whether it is solely attributable to the degree of differentiation or histologic subtype is also very difficult to exclude or accept, al- though it is true that, in the past, Davidsohn22 suggested

FIG. 3 . Histologic features of metastatic deposits in lymph node in above case of nasopharyngeal carcinoma (H & E, X120).

FIG. 4. A case of adenoma of the breast used as one of the control showing positive SRCA (MCAR x 120).

No. 2 NASOPHARYNGEAL CA - Gupta 259

this to be a possible factor in some malignancies. In a similar study in bronchogenic carcinomas, Davidsohn” also stated that loss of isoantigens is not an all-or-none phenomenon as documented by him from his obser- vations in bronchogenic carcinomas showing both pos- itive and negative MCAR in tumor cells. Also, according to Davidsohn, the finding of a positive and negative MCAR in bronchogenic carcinomas may be indicative of a progressive loss of isoantigens in cancerous cells.2’*22 While in cases of NPC no single factor seemed applicable to explain or even speculate the findings of a mild pos- itivity in all cases, of interest was that the partial reten- tion of isoantigens in all cases of NPCs, as observed in this study, seems to be a somewhat unique finding and, as far as we are aware, has not been reported in other tumors. Whether an ethnic predisposition or environ- mental factors can be responsible for such differences in NPCs may seem likely, although cannot be conclu- sively incriminated as a sole reason. Also in these tu- mors, specific HLA patterns, impaired cell mediated im- mune functions and correlation with EB virus have been previously described in Chinese patients.’-I6

The finding of a consistently negative reaction in all lymph nodes involved by metastatic NPCs appeared to be an additional finding of interest in this study. Again, whether such a finding could be solely due to the loss of isoantigens in metastatic tumor cells on reaching the lymph nodes, or due to a ‘preferential discharge’ of isoantigen-free tumor cells from primary NPCs in the metastatic deposits in lymph nodes, although appearing likely, is speculative in view of the constant finding of a mild-positive reaction in all cases of NPCs. Such a suggestion has also been made previously by David- sohn.22 Based on the above findings, only a few conclu- sions can be drawn at this time, and they are as follows: ( 1 ) The cells of NPCs may be somewhat unique in that only occasional tumor cells may have a capacity to par- tially retain isoantigens A, B & H(0). (2) No differences in SRCA are found, irrespective of the histologic sub- types. (3) A correlation based on SRCA pattern and histologic subtype of NPC with ultimate prognosis is not possible. (4) The exact reason for a consistently negative SRCA in metastatic deposits of NPC in lymph nodes is unknown.

REFERENCES

I . Clifford P. On the epidemiology of nasopharyngeal cancer. Int J Cancer 1970; 5:287-309. 2. Ho HC. Current knowledge of the epidemiology of nasopharyn-

geal carcinoma: A review. In: Biggs PM, de The G, Payne LN, eds. Oncogenesis and Herpes Viruses. IRAC Scientific Publications, no. 2. Lyon: International Agency for Research on Cancer, 1972; 357-366. 3. Muir CS. Nasopharyngeal carcinoma in non-Chinese popula-

tions with special reference to Southeast Asia and Africa. Br J Cancer

4. Blot WJ, Lanier AP, Fraumeni JF, Bender TR. Cancer mortality 1971; 8~351-363.

among Alaskan natives 1960-69. J Nut1 Cancer Inst 1975; 55347- 554.

5 . Lanier AP, Bender TR, Blot WJ, Fraumeni JF, Hurlbert WB. Cancer incidence in Alaskan natives. Int J Cancer 1976; I8:409-412. 6. Mallen RW, Shandro WG. Nasopharyngeal carcinoma in Es-

kimos. Can J Otolaryngol 1974; 3: 175-1 79. 7. Schaefer 0, Hildes JA, Medd LM, Cameron DC. The changing

pattern of neoplastic disease in Canadian Eskimos. Can Med Assoc

8. Lanier AP, Bender T, Talbot M et al. Nasopharyngeal carcinoma in Alaskan Eskimos, Indians and Aleuts: A review of cases and study of Epstein Barr virus, HLA, and environmental risk factors. Cancer

9. Simons MJ, Wee GB, Goh EH el al. Immunogenetic aspects of NPC. IV: Increased risk in Chinese of NPC associated with a Chinese- related HLA profile (A2, Singapore 2). J Null Cancer Inst 1976;

10. de Schryver A, Friberg GS Jr, Klein G et al. Epstein-Barr virus- associated antibody patterns in carcinoma of the postnasal space. Clin Exp Immunol 1969; 5:443-459.

11. Anderson-Anvert M, Forsby N, Klein G, Henle W. Relation between the Epstein-Barr virus and undifferentiated nasopharyngeal carcinoma: Correlated nucleic acid hybridization and histopatholog- ical examination. Int J Cancer 1977; 20486-494. 12. Desgranges C, Wolf H, de The G el al. Nasopharyngeal car-

cinoma. X: Presence of Epstein-Barr genomes in separated epithelial cells of tumours in patients from Singapore, Tunisia and Kenya. Int J Cancer 1975; 16:7-15.

13. Pagano JS, Huang CH, Klein G, de The G, Shanmugaratnam K, Yang CS. Homology of Epstein-Barr virus DNA in nasopharyngeal carcinomas from Kenya, Taiwan, Singapore and Tunisia. In: de The G, Epstein MA, zur Hausen H, eds. Oncogenesis and Herpes Viruses II, part 2. IRAC Scientific Publications, no. I I . Lyon: International Agency for Research on Cancer, 1975; 179-190. 14. Wolf H, zur Hausen H, Becker V. E.B. viral genomes in epi-

thelial nasopharyngeal carcinoma cells. Nature 1973; 244:245-247. 15. Henle W, Henle G, Ho HC et al. Antibodies to Epstein-Barr

virus in nasopharyngeal carcinoma, other head and neck neoplasms, and control groups. J Nut1 Cancer Inst 1970; 44:225-23 1. 16. Chan SH, Chew TS, Goh EH, Simons MJ, Shanmugaratnam

K. Impaired general cell mediated immune functions in vivo and in vitro in patients with nasopharyngeal carcinoma. Int J Cancer 1976;

17. Shanmugaratnam K, Sobin L. Histological Typing of Upper Respiratory Tract Turnours. International Histological Typing of Tu- mours, no. 19. Geneva, WHO, 1978; 32-33. 18. Shanmugaratnam K, Chan SH, de The G el al. Histopathology

of nasopharyngeal carcinoma: Correlations with epidemiology, sur- vival rates and other biological characteristics. Cancer 1979; 44 1029- 1044. 19. Kovarik S, Davidsohn I, Stejskal R. ABO antigens in cancer:

Detection with the mixed cell agglutination reaction. Arch Pathol1968;

20. Davidsohn I, Kovarik S, Ni LY. Isoantigens A, B and H in benign and malignant lesions ofthe cervix. Arch Pathol 1969; 87:306- 3 14. 21. Davidsohn I, Ni LY. Loss of isoantigens A, B and H in car-

cinoma of the lung. Am J Pathol 1969; 57:307-334. 22. Davidsohn I, Ni LY. lmmunocytology of cancer. Acta Cytol

1970; 14:276-282. 23. Gupta RK, Schuster R, Christian D. Loss of isoantigens A, B

and H in prostate. Am J Pathol 1973; 70:439-443. 24. Gupta RK, Schuster R. lsoantigens A, B and H in benign and

malignant lesions of the breast. Am J Pathol 1973; 72:253-257. 25. Gupta RK. lmmunocytologic loss of isoantigens A, B and H

in the endometrium and brain. Am J Clin Pathol 1976; 66:390-394. 26. Mathew T, Gupta RK. lsoantigens A, Band H in primary liver

malignancies, Pathology. J Coll Pathol Aust 1981; 1357 1-577. 27. Limas C, Lange P, Fraley EE. A, B, H antigens in transitional

cell tumours of the urinary bladder: Correlations with the clinical course. Cancer 1979; 44:2099-2 107.

J 1975; 112:1399-1404.

1980; 46:2100-2106.

57~977-980.

18: 139- 144.

86: 12-2 1.