9 INTRODUCTION
Preliminary
Communication
hepatic carcinoma (P.H.C.) is the most commalignant neoplasm in several of the world’s major populations (Chinese, Africans, Filipinos). Therefore, it may be the most common cancer in man. On the basis of necropsy studies showing an association between chronic hepatitis and/or macronodular post-necrotic cirrhosis with P.H.C., Payet 1 and Steiner and Davies 2 suggested that viral hepatitis may be the major cause of P.H.C. in Africa. When tests for infection with hepatitis-B virus (H.B.v.) became available it was possible to evaluate the association of H.B.V. with PRIMARY
mon
ANTIBODY TO HEPATITIS-B CORE ANTIGEN IN PATIENTS WITH PRIMARY HEPATIC CARCINOMA PHILIPPE MAUPAS BARBARA WERNER IRVING MILLMAN BERNARD LAROUZÉ ANNA O’CONNELL W. THOMAS LONDON BARUCH S. BLUMBERG Institute for Cancer Research, Fox Chase Cancer Center, Philadelphia, Pennsylvania, U.S.A., and Department of Microbiology, University of Tours, Tours, France
GERARD SAIMOT Institut de Medicine Hôpital Claude
MAURICE PAYET et
d’Épidémiologie Africaine,
Bernard, Paris,
France
Antibody to hepatitis-B core antigen (anti-HBc) was assayed in the serum of patients with primary hepatic carcinoma (P.H.C.) and controls from Hong Kong, West Africa, and the United States. In each region the prevalence of antiHBc was higher in P.H.C. patients than in controls, ranging from 70 to 95 % in the patients and from 20 to 68% in the controls from Asia and Africa; 24% of P.H.C. patients and 4% of controls from the U.S. had anti-HBc. These data support the hypothesis that chronic infection with hepatitis-B virus is ætiologically related to P.H.C., especially in Asia and Africa, although Summary
other factors must also be involved.
Several studies in Africa and Asia showed an increased prevalence (35-80%) of Australia antigen (hepatitis-B surface antigen, HBsAg) among patients with P.H.C. These high prevalences were generally detected only when sensitive methods (radioimmunoassay, immune adherence hæmagglutination) for assaying HBgAg were used.3.4 This observation suggested that either HBsAg was present in decreased quantity in P.H.C. patients or there was some antigenic difference between the HBgAg in P.H.C. and that found in HBsAg carriers.4 No increase in the prevalence of antibody to hepatitis-B surface antigen (anti-HBs) has been reported in P.H.C. patients If patients with P.H.C. are infected with H.B.v. but are producing small quantities of HBgAg, there may be defects in viral assembly or viral replication in the malignant cells. Hoofnagle et al.6 have suggested that antibody to hepatitis-B core antigen (anti-HBc) is a better indicator of intracellular replication of H.B.V. than anti-HBs or HBsAg. In this paper we report P.H.C.
SONICATED LIVER HOMOGENATE
(a) 1/2 HBcB2 L2-65B.
Scheme for isolation and purification of HBcAg from liver of a human HBsAg carrier. concentrated x 10 in Amicon apparatus (SM-300 90 mm. filter). centrifuged 106,000×g×2hr. in Beckman (c) Isopycnic centrifugation in CsClz (1’316 g./c.cm.) 108,000 dialysed against 0.1M P.B.S. pH 7’2x24 hr. ×g×72 hr.
(b) 1/2 HBcB2 ppt. = precipitate.
sup. = supernatant.
C.E.P. = counter-electrophoresis.
E.M.= electron
microscopy.
10
the detection of anti-HB in a very high proportion of patients with P.H.C. These data support the hypothesis that chronic infection with H.B.V. is at least one ætiological factor in P.H.C. and suggest a method of preventing this form of cancer.
tested by C.E.P. against HBcB2 Amicon concentrate and the gradient fractions containing HBcAg. If a precipitin was observed, such sera were scored anti-HBc positive. In general, C.E.P. and LD. using the HBcB2 fraction gave definitive results.
were
PATIENTS AND CONTROLS
MATERIALS
AND
METHODS
HBcAg was extracted by the method of Hoofnagle et al.7 from the liver of a patient who had chronic hepatitis, was positive for HBsAg, and had received a renal transplant. Electron microscopy of a biopsy specimen of this patient’s liver showed numerous 27-29 nm. particles in the nuclei of parenchyma cells. (Attempts by the methods described below to extract HBcAg from the livers of four other HBsAg carriers and one liver from an HBsAg-negative The liver was removed at person were unsuccessful.) necropsy, perfused with 0’25M sucrose, and frozen in dry ice. For preparation of core antigen the liver was thawed and homogenised in a Waring blender as a 25 % suspension in cold phosphate-buffered saline. The homogenate was sonicated twice for 60 seconds each at maximum setting in a Bronson W 185E sonicator (Heat Systems Inc.) with special microtip. The procedures used to concentrate and isolate HBcAg are shown in the figure. HBcAg in the liver homogenate extract (HBcB2) reacted by C.E.P. and immunodiffusion (I.D.) with the rhesus antiHBo and the precipitin in I.D. produced a line of identity with anti-HBc in the serum of HBsAg carriers. All sera from the P.H.C. patients and controls were tested first by C.E.P. and I.D. against HBcB2. C.E.P. was carried out in 1’1 % agarose in barbitone buffer, pH 8’2. I.D. was performed in 0’9% Noble agar in barbitone buffer, pH 8’2. HBcB2 was placed in the centre well and known anti-HBc was placed in the first and fourth wells of a seven-well I.D. pattern. Sera were considered to contain anti-HBc if they showed a line of identity with known anti-HBc by I.D. and Sera which gave equivocal results a precipitin by C.E.P. ANTI-HBc
IN PATIENTS WITH P.H.C.
The P.H.C. patients and controls tested for anti-HBc are listed in the table. The diagnosis of P.H.C. was accepted
only if patients had histological evidence of P.H.C. on biopsy or at necropsy or if their clinical picture was compatible with P.H.C. and they had a-fetoprotein detectable in their
serum
by
C.E.P. or LD.
RESULTS
The table shows that, of 153 patients with P.H.C. from three continents, 114 had anti-HBc detected by the relatively insensitive C.E.P. and I.D. tests. In Senegal over 90% of the patients had anti-HBc, whereas in Mali and Hong Kong about 70% were positive. Of the 5 anti-HBc-positive American P.H.C. patients, 1 was Chinese and another was from Puerto Rico. DISCUSSION
The prevalence of anti-HBo among these P.H.C. patients is higher than the reported prevalence of HBsAg or anti-HBs among P.H.C. patients.3-5 The frequency of these factors in P.H.C. patients will be reported elsewhere in studies describing the variation
of response of West Africans to H.B.V. infection. Here we wish to emphasise the likelihood that virtually all P.H.C. patients in West Africa and South-East Asia are (or have been) infected with H.B.V. In the U.S.A., H.B.V. is associated with a considerable proportion of, but not all, cases of P.H.C. Furthermore, anti-HBc must be present in high titre to be detectable by C.E.P. and I.D., whereas HBSAg is present only at low concentrations in P.H.C. patients.4,5 This suggests that in P.H.C. there is a continuous production of core antigen and limited synthesis of HBsAg. Hadziyannis 8 has reported that, by immunofluorescent techniques, HBBAg is found only in the parts of the liver unaffected by P .H.C. On the basis of these observations and the data reported here, we suggest that both the unaffected and malignant liver cells are
producing HBcAg. The data do
&bul et;
(1) Panents with
P.H.C.
in
Hôpital Le Dantec, Dakar. Controls are 1 Senegal matched for each case by
2 persons without P.H.C. from age, sex, and ethnic group.
or
(2) Patients with P.H.C. at Hôpital Le Dantec, Dakar. Controls are individuals without P.H.C. matched for each case by age, sex, and ethnic group. (3) Patients w-ith P.H.C. at Hôpital Point G, Bamako. Controls are 1 or 2 persons from Mali without P.H.C. matched for each case by age, sex, and ethnic group. (4) Patients with P.H.C. collected by Dr David Todd at the Queen Mary Hospital, Hong Kong. Controls are 58 patients without P.H.C. from the same hospital with a variety of other diseases. 7 of the 58 controls had cirrhosis; 5 of these had anti-HBc. (5) Sera from patienta with P.H.C. were submitted by physicians at several hospitals in Philadelphia. Controls were 25 healthy volunteer blood-donors and 25 prisoner blood-donors. t Calculated from Fisher’s exact 2 x 2 test.
not
in themselves indicate
a
causal
relationship between hepatitis Band P.H.C., but, taken together with previous findings, they support the hypothesis that H.B.V. infection is actiologically related to P.H.C. Since infection with H.B.V., as detected by the presence of HBsAg, anti-HBS, or anti-HBc, is common in the general populations of West Africa and Hong Kong, additional factors must be involved in the causation of P.H.C. Some of these are under investigation and will undoubtedly be identified in due course. No matter what other factors are involved, however, if H.B.V. infection is one necessary factor then prevention of infection with H.B.V. could lead to prevenWith the development of a vaccine tion of P.H.C. against H.B.V., prevention of many cases of P.H.C. may be a realistic expectation. We thank Dr S. N. Huang for supplying the human liver for the source of HBcAg, Dr David Todd for the sera from the Chinese patients and controls, Dr Jay Hoofnagle for the
11 rhesus anti-HBc antibody, and Barbara Peticolas, Ellen Halpern, Theresa Halbherr, and Francis Kern for technical assistance. This work was supported by U.S. Public Health Service grants CA-06551, RR-05539, and CA-06927 from the National Institutes of Health, and by an appropriation from the Commonwealth of Pennsylvania. Requests for reprints should be addressed to W. T. L. Institute for Cancer Research, Fox Chase Cancer Center, Philadelphia, Pennsylvania 19111, U.S.A. REFERENCES
Payet, M., Camain, R., Pene, P. Revue int. Hépat. 1956, 4, 1. Steiner, P. E., Davies, J. N. P. Br. J. Cancer, 1957, 11, 523. Vogel, C. C., Anthony, P. P., Mody, N. J., Barker, L. F. Lancet, 1970, ii, 621. 4. Nishioka, K., Hirayama, T., Sekine, T., Okochi, K., Mayumi, M., Juei-Low, S., Chen-Hui, L., Tong-Min, L. in Alpha-Fetoprotein and Hepatoma (edited by H. Hidematsu and T. Miyaji); p. 167. 1. 2. 3.
Baltimore,
1973.
5. Nishioka, K., Mayumi, M., Okochi, K., Okada, K., Hirayama, T. in Analytic and Experimental Epidemiology of Cancer (edited by W. Nakahara, T. Hirayama, K. Nishioka, and H. Sugano); p. 137.
Baltimore, 1974. Hoofnagle, J., Gerety, R., Hi, L., Barker, L. New Engl. J. Med. 1974, 290, 1336. 7. Hoofnagle, J., Gerety, R., Barker, L. Lancet, 1973, ii, 869. 8. Hadziyannis, S. Proc. natn. Acad. Sci. U.S.A. (in the press). 6.
Hypothesis SCHIZOPHRENIA, EPILEPSY, CANCER, METHIONINE, AND FOLATE METABOLISM PATHOGENESIS OF SCHIZOPHRENIA REFAEL N. LEVI
SAMUEL WAXMAN and Medicine, Mount Sinai
Departments of Psychiatry School of Medicine of City University of New York, New York, N.Y. 10029, U.S.A.
WE base our hypothesis upon the seemingly unrelated observations that folate-responsive schizophreniclike syndromes develop in some epileptics treated with anticonvulsant drugs, the low incidence of carcinoma in schizophrenics, and the absolute need for methionine for the growth of malignant tissues in vitro.
to be at relatively low risk of cancer. al.8 found that the percentage of mental patients who died from cancer was considerably lower
phrenia,
seem
Rassidakis
et
than that of the general population; 15 % of deaths were caused by malignant neoplasms in the general population Other compared with 4’9% among the mentally ill. randomly selected causes of death (cardiovascular, diabetes) showed no appreciable differences. In England and Wales about 20% of deaths are caused by neoplasms compared with 6’9% of deaths in mentally ill populations; for Scotland the figures are 17 % and 5%, respectively. Other investigations suggest that schizophrenic patients may be even more resistant to neoplasms than patients with other forms of mental illness. Of 2145 State hospital deaths over fifteen years, 16 were neoplastic and only 6 of these were among schizophrenics. In another large mental hospital, of a total of 520 deaths during a two-year period, only 4 schizophrenics died of cancer. The Kashenko Hospital, in Moscow, reports the incidence of deaths from malignant neoplasms in schizophrenics to be only 0-1-0-2% from a yearly population of 2500. Katz and colleagues 9 also reported lower cancer mortality in mental patients than that in matched segments of the general population. The low incidence of cancer is particularly evident among patients who have been in The reason is mental hospitals for over ten years. unknown. Lindelius and Kay,10 studying 355 male schizophrenics admitted to the hospital from 1947 to 1958, reported 32 deaths; only 1 was due to cancer. Csatary11 suggests that specific environmental factors and, above all, a peculiarity in the as yet unravelled bio-
chemistry of mental diseases, especially schizophrenia, hold
an
tumour
important clue relationships.
to
the
understanding
may
of host/
FORMULATION OF BIOCHEMICAL
HYPOTHESIS
Methionine is an essential aminoacid which plays an important role in transmethylation processes which provide a major catabolic pathway for biogenic amines.12,13 L-methionine cannot provide labile methyl
FOLATE-RESPONSIVE ACQUIRED SCHIZOPHRENIC-LIKE SYNDROMES
The relation between schizophrenia and epilepsy was first observed by Meduna 1 who said that schizophrenia and epilepsy were biologically antagonistic diseases. Reynolds 2 formulated Meduna’s observation in biochemical terms when he postulated that the biochemical disturbances in untreated epileptics render them less prone to schizophrenia, and, similarly, that the schizophrenic is less liable to develop epilepsy. However, schizophreniclike psychoses in treated epilepsy are seen,3 and Reynolds 4,5 attributed these to the effect of anticonvulsants on folate metabolism. Anticonvulsants, especially diphenylhydantoin, lower serum-folate in significant numbers of patients, and cause frank megaloblastic anæmia in a few.6 Reynolds et al.4 suggested that this antifolate effect is important in controlling the epilepsy but may provoke mental symptoms which resemble schizophrenia. When these patients were treated with folic acid and vitamin B12, the anticonvulsant drug did not control the epilepsy, but it did ameliorate the
schizophrenic-like syndrome.7 SCHIZOPHRENIA AND CANCER
Mentally ill populations, especially patients
with schizo-
Metabolic pathways for production of labile methyl groups. Reaction 1: methionine adenosyltransferase pathway for for-
mation of SAMe. Reaction 2: homocysteine transmethylation pathway through demethylation of 5-M.T.H.F. and resynthesis of L-methionine. Reaction 3: methylene T.H.F. reductase (inhibited by SAMe). Reaction 4: thymidylate synthetase pathway for de-novo D.N.A.
synthesis. Reaction 5: proposed methylation of S-adenosylhomocysteine through demethylation of 5-M.T.H.F. for endogenous resynthesis of SAMe. This reaction should not require the presence of methionine adenosyltransferase. Solid lines represent proven pathways. Broken lines represent proposed pathways for endogenous SAMe formation.
Preliminary
Communication
hepatic carcinoma (P.H.C.) is the most commalignant neoplasm in several of the world’s major populations (Chinese, Africans, Filipinos). Therefore, it may be the most common cancer in man. On the basis of necropsy studies showing an association between chronic hepatitis and/or macronodular post-necrotic cirrhosis with P.H.C., Payet 1 and Steiner and Davies 2 suggested that viral hepatitis may be the major cause of P.H.C. in Africa. When tests for infection with hepatitis-B virus (H.B.v.) became available it was possible to evaluate the association of H.B.V. with PRIMARY
mon
ANTIBODY TO HEPATITIS-B CORE ANTIGEN IN PATIENTS WITH PRIMARY HEPATIC CARCINOMA PHILIPPE MAUPAS BARBARA WERNER IRVING MILLMAN BERNARD LAROUZÉ ANNA O’CONNELL W. THOMAS LONDON BARUCH S. BLUMBERG Institute for Cancer Research, Fox Chase Cancer Center, Philadelphia, Pennsylvania, U.S.A., and Department of Microbiology, University of Tours, Tours, France
GERARD SAIMOT Institut de Medicine Hôpital Claude
MAURICE PAYET et
d’Épidémiologie Africaine,
Bernard, Paris,
France
Antibody to hepatitis-B core antigen (anti-HBc) was assayed in the serum of patients with primary hepatic carcinoma (P.H.C.) and controls from Hong Kong, West Africa, and the United States. In each region the prevalence of antiHBc was higher in P.H.C. patients than in controls, ranging from 70 to 95 % in the patients and from 20 to 68% in the controls from Asia and Africa; 24% of P.H.C. patients and 4% of controls from the U.S. had anti-HBc. These data support the hypothesis that chronic infection with hepatitis-B virus is ætiologically related to P.H.C., especially in Asia and Africa, although Summary
other factors must also be involved.
Several studies in Africa and Asia showed an increased prevalence (35-80%) of Australia antigen (hepatitis-B surface antigen, HBsAg) among patients with P.H.C. These high prevalences were generally detected only when sensitive methods (radioimmunoassay, immune adherence hæmagglutination) for assaying HBgAg were used.3.4 This observation suggested that either HBsAg was present in decreased quantity in P.H.C. patients or there was some antigenic difference between the HBgAg in P.H.C. and that found in HBsAg carriers.4 No increase in the prevalence of antibody to hepatitis-B surface antigen (anti-HBs) has been reported in P.H.C. patients If patients with P.H.C. are infected with H.B.v. but are producing small quantities of HBgAg, there may be defects in viral assembly or viral replication in the malignant cells. Hoofnagle et al.6 have suggested that antibody to hepatitis-B core antigen (anti-HBc) is a better indicator of intracellular replication of H.B.V. than anti-HBs or HBsAg. In this paper we report P.H.C.
SONICATED LIVER HOMOGENATE
(a) 1/2 HBcB2 L2-65B.
Scheme for isolation and purification of HBcAg from liver of a human HBsAg carrier. concentrated x 10 in Amicon apparatus (SM-300 90 mm. filter). centrifuged 106,000×g×2hr. in Beckman (c) Isopycnic centrifugation in CsClz (1’316 g./c.cm.) 108,000 dialysed against 0.1M P.B.S. pH 7’2x24 hr. ×g×72 hr.
(b) 1/2 HBcB2 ppt. = precipitate.
sup. = supernatant.
C.E.P. = counter-electrophoresis.
E.M.= electron
microscopy.
10
the detection of anti-HB in a very high proportion of patients with P.H.C. These data support the hypothesis that chronic infection with H.B.V. is at least one ætiological factor in P.H.C. and suggest a method of preventing this form of cancer.
tested by C.E.P. against HBcB2 Amicon concentrate and the gradient fractions containing HBcAg. If a precipitin was observed, such sera were scored anti-HBc positive. In general, C.E.P. and LD. using the HBcB2 fraction gave definitive results.
were
PATIENTS AND CONTROLS
MATERIALS
AND
METHODS
HBcAg was extracted by the method of Hoofnagle et al.7 from the liver of a patient who had chronic hepatitis, was positive for HBsAg, and had received a renal transplant. Electron microscopy of a biopsy specimen of this patient’s liver showed numerous 27-29 nm. particles in the nuclei of parenchyma cells. (Attempts by the methods described below to extract HBcAg from the livers of four other HBsAg carriers and one liver from an HBsAg-negative The liver was removed at person were unsuccessful.) necropsy, perfused with 0’25M sucrose, and frozen in dry ice. For preparation of core antigen the liver was thawed and homogenised in a Waring blender as a 25 % suspension in cold phosphate-buffered saline. The homogenate was sonicated twice for 60 seconds each at maximum setting in a Bronson W 185E sonicator (Heat Systems Inc.) with special microtip. The procedures used to concentrate and isolate HBcAg are shown in the figure. HBcAg in the liver homogenate extract (HBcB2) reacted by C.E.P. and immunodiffusion (I.D.) with the rhesus antiHBo and the precipitin in I.D. produced a line of identity with anti-HBc in the serum of HBsAg carriers. All sera from the P.H.C. patients and controls were tested first by C.E.P. and I.D. against HBcB2. C.E.P. was carried out in 1’1 % agarose in barbitone buffer, pH 8’2. I.D. was performed in 0’9% Noble agar in barbitone buffer, pH 8’2. HBcB2 was placed in the centre well and known anti-HBc was placed in the first and fourth wells of a seven-well I.D. pattern. Sera were considered to contain anti-HBc if they showed a line of identity with known anti-HBc by I.D. and Sera which gave equivocal results a precipitin by C.E.P. ANTI-HBc
IN PATIENTS WITH P.H.C.
The P.H.C. patients and controls tested for anti-HBc are listed in the table. The diagnosis of P.H.C. was accepted
only if patients had histological evidence of P.H.C. on biopsy or at necropsy or if their clinical picture was compatible with P.H.C. and they had a-fetoprotein detectable in their
serum
by
C.E.P. or LD.
RESULTS
The table shows that, of 153 patients with P.H.C. from three continents, 114 had anti-HBc detected by the relatively insensitive C.E.P. and I.D. tests. In Senegal over 90% of the patients had anti-HBc, whereas in Mali and Hong Kong about 70% were positive. Of the 5 anti-HBc-positive American P.H.C. patients, 1 was Chinese and another was from Puerto Rico. DISCUSSION
The prevalence of anti-HBo among these P.H.C. patients is higher than the reported prevalence of HBsAg or anti-HBs among P.H.C. patients.3-5 The frequency of these factors in P.H.C. patients will be reported elsewhere in studies describing the variation
of response of West Africans to H.B.V. infection. Here we wish to emphasise the likelihood that virtually all P.H.C. patients in West Africa and South-East Asia are (or have been) infected with H.B.V. In the U.S.A., H.B.V. is associated with a considerable proportion of, but not all, cases of P.H.C. Furthermore, anti-HBc must be present in high titre to be detectable by C.E.P. and I.D., whereas HBSAg is present only at low concentrations in P.H.C. patients.4,5 This suggests that in P.H.C. there is a continuous production of core antigen and limited synthesis of HBsAg. Hadziyannis 8 has reported that, by immunofluorescent techniques, HBBAg is found only in the parts of the liver unaffected by P .H.C. On the basis of these observations and the data reported here, we suggest that both the unaffected and malignant liver cells are
producing HBcAg. The data do
&bul et;
(1) Panents with
P.H.C.
in
Hôpital Le Dantec, Dakar. Controls are 1 Senegal matched for each case by
2 persons without P.H.C. from age, sex, and ethnic group.
or
(2) Patients with P.H.C. at Hôpital Le Dantec, Dakar. Controls are individuals without P.H.C. matched for each case by age, sex, and ethnic group. (3) Patients w-ith P.H.C. at Hôpital Point G, Bamako. Controls are 1 or 2 persons from Mali without P.H.C. matched for each case by age, sex, and ethnic group. (4) Patients with P.H.C. collected by Dr David Todd at the Queen Mary Hospital, Hong Kong. Controls are 58 patients without P.H.C. from the same hospital with a variety of other diseases. 7 of the 58 controls had cirrhosis; 5 of these had anti-HBc. (5) Sera from patienta with P.H.C. were submitted by physicians at several hospitals in Philadelphia. Controls were 25 healthy volunteer blood-donors and 25 prisoner blood-donors. t Calculated from Fisher’s exact 2 x 2 test.
not
in themselves indicate
a
causal
relationship between hepatitis Band P.H.C., but, taken together with previous findings, they support the hypothesis that H.B.V. infection is actiologically related to P.H.C. Since infection with H.B.V., as detected by the presence of HBsAg, anti-HBS, or anti-HBc, is common in the general populations of West Africa and Hong Kong, additional factors must be involved in the causation of P.H.C. Some of these are under investigation and will undoubtedly be identified in due course. No matter what other factors are involved, however, if H.B.V. infection is one necessary factor then prevention of infection with H.B.V. could lead to prevenWith the development of a vaccine tion of P.H.C. against H.B.V., prevention of many cases of P.H.C. may be a realistic expectation. We thank Dr S. N. Huang for supplying the human liver for the source of HBcAg, Dr David Todd for the sera from the Chinese patients and controls, Dr Jay Hoofnagle for the
11 rhesus anti-HBc antibody, and Barbara Peticolas, Ellen Halpern, Theresa Halbherr, and Francis Kern for technical assistance. This work was supported by U.S. Public Health Service grants CA-06551, RR-05539, and CA-06927 from the National Institutes of Health, and by an appropriation from the Commonwealth of Pennsylvania. Requests for reprints should be addressed to W. T. L. Institute for Cancer Research, Fox Chase Cancer Center, Philadelphia, Pennsylvania 19111, U.S.A. REFERENCES
Payet, M., Camain, R., Pene, P. Revue int. Hépat. 1956, 4, 1. Steiner, P. E., Davies, J. N. P. Br. J. Cancer, 1957, 11, 523. Vogel, C. C., Anthony, P. P., Mody, N. J., Barker, L. F. Lancet, 1970, ii, 621. 4. Nishioka, K., Hirayama, T., Sekine, T., Okochi, K., Mayumi, M., Juei-Low, S., Chen-Hui, L., Tong-Min, L. in Alpha-Fetoprotein and Hepatoma (edited by H. Hidematsu and T. Miyaji); p. 167. 1. 2. 3.
Baltimore,
1973.
5. Nishioka, K., Mayumi, M., Okochi, K., Okada, K., Hirayama, T. in Analytic and Experimental Epidemiology of Cancer (edited by W. Nakahara, T. Hirayama, K. Nishioka, and H. Sugano); p. 137.
Baltimore, 1974. Hoofnagle, J., Gerety, R., Hi, L., Barker, L. New Engl. J. Med. 1974, 290, 1336. 7. Hoofnagle, J., Gerety, R., Barker, L. Lancet, 1973, ii, 869. 8. Hadziyannis, S. Proc. natn. Acad. Sci. U.S.A. (in the press). 6.
Hypothesis SCHIZOPHRENIA, EPILEPSY, CANCER, METHIONINE, AND FOLATE METABOLISM PATHOGENESIS OF SCHIZOPHRENIA REFAEL N. LEVI
SAMUEL WAXMAN and Medicine, Mount Sinai
Departments of Psychiatry School of Medicine of City University of New York, New York, N.Y. 10029, U.S.A.
WE base our hypothesis upon the seemingly unrelated observations that folate-responsive schizophreniclike syndromes develop in some epileptics treated with anticonvulsant drugs, the low incidence of carcinoma in schizophrenics, and the absolute need for methionine for the growth of malignant tissues in vitro.
to be at relatively low risk of cancer. al.8 found that the percentage of mental patients who died from cancer was considerably lower
phrenia,
seem
Rassidakis
et
than that of the general population; 15 % of deaths were caused by malignant neoplasms in the general population Other compared with 4’9% among the mentally ill. randomly selected causes of death (cardiovascular, diabetes) showed no appreciable differences. In England and Wales about 20% of deaths are caused by neoplasms compared with 6’9% of deaths in mentally ill populations; for Scotland the figures are 17 % and 5%, respectively. Other investigations suggest that schizophrenic patients may be even more resistant to neoplasms than patients with other forms of mental illness. Of 2145 State hospital deaths over fifteen years, 16 were neoplastic and only 6 of these were among schizophrenics. In another large mental hospital, of a total of 520 deaths during a two-year period, only 4 schizophrenics died of cancer. The Kashenko Hospital, in Moscow, reports the incidence of deaths from malignant neoplasms in schizophrenics to be only 0-1-0-2% from a yearly population of 2500. Katz and colleagues 9 also reported lower cancer mortality in mental patients than that in matched segments of the general population. The low incidence of cancer is particularly evident among patients who have been in The reason is mental hospitals for over ten years. unknown. Lindelius and Kay,10 studying 355 male schizophrenics admitted to the hospital from 1947 to 1958, reported 32 deaths; only 1 was due to cancer. Csatary11 suggests that specific environmental factors and, above all, a peculiarity in the as yet unravelled bio-
chemistry of mental diseases, especially schizophrenia, hold
an
tumour
important clue relationships.
to
the
understanding
may
of host/
FORMULATION OF BIOCHEMICAL
HYPOTHESIS
Methionine is an essential aminoacid which plays an important role in transmethylation processes which provide a major catabolic pathway for biogenic amines.12,13 L-methionine cannot provide labile methyl
FOLATE-RESPONSIVE ACQUIRED SCHIZOPHRENIC-LIKE SYNDROMES
The relation between schizophrenia and epilepsy was first observed by Meduna 1 who said that schizophrenia and epilepsy were biologically antagonistic diseases. Reynolds 2 formulated Meduna’s observation in biochemical terms when he postulated that the biochemical disturbances in untreated epileptics render them less prone to schizophrenia, and, similarly, that the schizophrenic is less liable to develop epilepsy. However, schizophreniclike psychoses in treated epilepsy are seen,3 and Reynolds 4,5 attributed these to the effect of anticonvulsants on folate metabolism. Anticonvulsants, especially diphenylhydantoin, lower serum-folate in significant numbers of patients, and cause frank megaloblastic anæmia in a few.6 Reynolds et al.4 suggested that this antifolate effect is important in controlling the epilepsy but may provoke mental symptoms which resemble schizophrenia. When these patients were treated with folic acid and vitamin B12, the anticonvulsant drug did not control the epilepsy, but it did ameliorate the
schizophrenic-like syndrome.7 SCHIZOPHRENIA AND CANCER
Mentally ill populations, especially patients
with schizo-
Metabolic pathways for production of labile methyl groups. Reaction 1: methionine adenosyltransferase pathway for for-
mation of SAMe. Reaction 2: homocysteine transmethylation pathway through demethylation of 5-M.T.H.F. and resynthesis of L-methionine. Reaction 3: methylene T.H.F. reductase (inhibited by SAMe). Reaction 4: thymidylate synthetase pathway for de-novo D.N.A.
synthesis. Reaction 5: proposed methylation of S-adenosylhomocysteine through demethylation of 5-M.T.H.F. for endogenous resynthesis of SAMe. This reaction should not require the presence of methionine adenosyltransferase. Solid lines represent proven pathways. Broken lines represent proposed pathways for endogenous SAMe formation.
