- Email: [email protected]
EXOGENOUS ŒSTROGENS AND OVARIAN CANCER
325 intestinal resection. Fasting blood-samples were taken and lipoprotein concentrations were determined by preparative ultracentrifugation and compared with values for ageadjusted, randomly selected controls.4 Total serum triglycerides and triglycerides in the very-low-density lipoproteins (V.L.D.L.) were not significantly different in patients and controls. In the low-density lipoprotein (L.D.L.) fraction the cholesterol concentration of the patients was about half that of the controls. There was a negative correlation between L.D.L. cholesterol concentration and the length of the resected intestine (see figure). There was no absolute reduction in mean H.D.L. cholesterol concentration but concentrations were negatively correlated with the length of the resected small intestine. No correlation was observed between v.L.D.L. triglycerides and H.D.L. cholesterol. There was no significant correlation between body-weight and H.D.L. cholesterol concentration, indicating that the nutritional state of the patients was not a major determinant of their H.D.r.. cholesterol. The finding of a lower L.D.L. cholesterol concentration in patients with the most extensive resections accord with corresponding reductions found in patients operated on for massive obesity with jejunoileal bypass. The negative correlation between H.D.L. cholesterol and length of small-intestinal resecsuggests the possibility of an intestinal secretion of H.D.L. in man. The finding of unchanged H.D.L. cholesterol concentrations in shunt-operated obese patients, in whom the small intestine is intact although bypassed,s lends indirect support for such a hypothesis.
tion particles
an average period of 6.3 weeks (range 3-13 weeks) all patients received insulin injections, while the oral drugs were omitted. Throughout this period the diet remained the same. Each patient thus served as his own control. Glucose was measured by an automated method. Cholesterol, triglycerides, and lipoprotein concentrations were measured as previously described.4.5 H.D.L.-cholesterol was measured after precipitation of the lowdensity lipoproteins in the fraction with density > 1-006 g/ml. Treatment with insulin did not significantly alter the H.D.L. level as measured by its cholesterol content (see table). In the 9 patients on oral hypoglycaemic agents, the H.D.L.-cholesterol level was also unchanged-1.01+0.32 mmol/1 during oral therapy, 1-04+0-32 during insulin therapy (means ± S.D.). Our results are at variance with those of Bar-On et al. and suggest that the conclusions of the University Group Diabetes Program are not explained by the lowering of H.D.L. levels by
oral therapy.
C. JOHANSSON S. RÖSSNER G. WALLDIUS
S-104 01 Stockholm 60, Sweden
HIGH-DENSITY-LIPOPROTEIN AND MATURITY-ONSET DIABETES
SIR,-A low level of high-density lipoprotein (H.D.L.) in seems to accelerate the development of atheroscleroSiS.l.2 Bar-On et a1.3 found that H.D.L. and cholesterol in serum from 11 elderly patients on oral therapy were lower than in 12 diabetics on insulin and suggested that this might explain the University Group Diabetes Program conclusion that treatment with oral hypoglycaemic agents (tolbutamide and phenformin) increases cardiovascular mortality. However, Bar-On et al. studied two different groups of patients with an unequal sex ratio. This and other discrepancies between the groups could explain the difference in lipoprotein levels. We measured plasma-lipoprotein before and after treatment with insulin in a group of patients (8 males, 7 females, mean age 58 years, range 51-67 years) with maturity-onset diabetes and type iv hyperlipoproteinaemia. They had been treated for several years with diet alone (6 patients) or diet and oral hypoglycaemic agents sulphonylureas 8, phenformin 1. During
plasma
4. Carlson, K. J. clin. Path. 1973, 26, suppl. 5, p. 32. 5. Rössner, S. Hallberg, D. Acta med. scand. (in the press). 1. Miller, G. J., Miller, N. E. Lancet, 1975, i, 16. 2. Berg, K., Børresen, A. L., Dahlen, G. ibid. 1976, i, 499. 3. Bar-On, H., Landau, D., Berry, E. ibid. 1977, i, 761.
University of Nijmegen, Nijmegen, Netherlands
EXOGENOUS ŒSTROGENS AND OVARIAN CANCER
SIR, -Dr Annegers and his colleagues’ suggest that the data in
our
preliminary communication2"do
=
=
not
point
as
strongly
with exogenous oestrogen use" as we suggested. They note that the "expected" values calculated for ovarian cancer did not take into account the proportion of women in the general population who are not at risk-namely, those with surgically removed ovaries. By adjusting for the 10% prevalence-rate’ of bilateral oophorectomy prevailing in the Mayo Clinic population, they estimate that our expected numbers for oestrogen-treated women are 11 % too low, and that two of the three relative risks values are not statistically significant. In case others may have been misled by our paper, we would like to emphasise the following points: (1) We did not claim that ovarian cancer was associated with the use of exogenous oestrogens generally, but rather with specific oestrogen, diethylstilboestrol (D.E.S.). We tried to stress this by entitling the paper Stilboestrol (Diethylstilboestrol) and the Risk of Ovarian Cancer. (2) The relative risk (R.R.) for the association between D.E.s. use and ovarian cancer was 30.0 (confidence interval, 6.2-87-7). These values are not changed by adjusting for the prevalence of oophorectomy in the general population, whether one uses the Mayo Clinic oophorectomy prevalencerate or makes the extreme assumption that every woman in the general population with a hysterectomy has also undergone bilateral oophorectomy, (3) Our most conservative estimate of the R.R. for all
increased risk of ovarian
cancer
4.
Demacker, P.
2.
Hoover, R., Gray, L. A., Sr, Fraumeni, J. F., Jr ibid. p. 533.
N. M., v. Oppenraay, J., Baadenhuysen, H., Jansen, A. P. Clin. chim. Acta, 1975, 64, 45. 5. Demacker, P. N. M., Vos-Janssen, H. E., Jansen, A. P., van ’t Laar, A. Clin. Chem. 1977, 23, 1238. 1. Annegers, J. F., O’Fallon, W., Kurland, L. T. Lancet, 1977, ii, 869 (see also
p. 1188).
PLASMA-LIPOPROTEINS DURING ORAL THERAPY AND INSULIN THERAPY
Results are means < S.D. Student’s t test for paired data. *L.D.L. low-density lipoprotein. V.L.D.L. very-low-density lipoprotein.
P. N. M. DEMACKER J. A. LUTTERMAN A. VAN ’T LAAR
Department of Medicine,
to an
Department of Internal Medicine and King Gustaf V Research Institute, Karolinska Hospital,
A. F. H. STALENHOEF
326 oestrogen users was 2.4 (confidence interval, 1.0-4.8). This declines only slightly to 2.2 (0.9-4.3) when the Mayo Clinic adjustment factor for oophorectomy is used. In a more detailed discussion of the study population,3 we, calculated an estimate of R.R. based on the extreme assumption noted in (2) above. Even this "overadjustment" yields a 2-fold increase in risk for ovarian cancer. We see little difference in all of these estimates. The corrections have shifted the lower bound of the confidence interval below 1.0, so the values are not statistically significant at the 5% level. However, nominal levels of significance should not be overemphasised, since our risk estimate with a P value of 0.067 seems as meaningful as it was with a p value of 0.040. (4) The associations were based on only 8 cases of ovarian cancer (including 3 for the D.E.S. group), and thus may be spurious for a number of reasons (e.g., bias, confounding, chance). Our suspicion of a causal relationship was based not on statistical probabilities but on the magnitude of the risk after D.E.S. (30-fold), plus the capacity of D.E.S. to cause ovarian tumours in dogs, and the possibility that it may explain the rising incidence of this cancer in postmenopausal women. These findings, coupled with the magnitude of the exposure to D.E.S. in the population, led to our preliminary report in the hope that others would investigate the question. Since then, results of a follow-up of women who participated in a randomised clinical trial of D.E.S. for the prevention of miscarriage in the early 1950s has become available.4 4 women in the D.E.S. group have had ovarian cancer, compared with 1 in the placebo group. The numbers are again small, but provide further cause for concern. Annegers et al. report preliminary results from a case-control study of ovarian cancer, showing no relation to the use of exogenous cestrogens, While this is useful information, it is not relevant to the issue of the risk associated with D.E.S. exposure since an earlir Mayo Clinic survey’ indicated very limited long-term use of D.E.S. in that population. ,
Environmental Epidemiology Branch, National Cancer Institute, Bethesda, Maryland 20014, U.S.A.
Department of Obstetrics and Gynecology, University of Louisville, Louisville, Kentucky
ROBERT HOOVER JOSEPH F. FRAUMENI, JR
maternal cellular immunity is impaired or the fetus is not immunogenic. There have been many reports of serum factors than can depress cellular immunity, although this effect is often measured by the non-specific criterion of inhibition of mitogen stimulation. However, most of these factors are not specific for paternal antigens they are not found in all pregnancies, and most of them appear late in pregnancy. They cannot be primarily responsible for the lack of fetal rejection, though they may play a role in later events during pregnancy. We have shown’ that purified plasma membranes from term placentae contain amounts of HLA A and B antigens sufficiently small for their presence to be accounted for by contamination from connective tissue. Others have found,2 by immunofluorescence studies, that these antigens are absent from the human trophoblast. These results strongly suggest that the syncytiotrophoblast expresses HLA A and B antigens not at all or at most in greatly reduced amounts. A similar conclusion has been reached for the mouse.3 Lack of HLA A and B antigens on the tissue in contact with the maternal circulation could abrogate cell-mediated attack in several ways. The work of Doherty and Zinkernagel4 suggests that, in the mouse, cell-mediated killing of cells depends upon recognition of H 2K or D determinants either complexed to a foreign determinant giving rise to an "altered self’ or,’ in an early effector stage of the killing reaction. A similar role has been suggested for the HLA A, B, and C antigens in man.’ The work of Bach et al. and others suggests that the HLA A, B, and C antigens are a necessary part of the antigenic stimulus required to generate cytotoxic T lymphocytes. Thus, lack of HLA A, B, and C antigens on trophoblasts could act either to prevent the generation of cytotoxic T lymphocytes or to prevent the action of such cells when directed against any fetal or polymorphic antigens for which there is fetomaternal incompa-
tibility. It therefore seems most probable that lack of immunogenicity of the trophoblast with respect to cellular immunity, rather than depression of maternal immunocompetence, is the primary mechanism involved in the "immunological inertia of
viviparity". Switching off HLA A, B, and C antigen expression to escape immunological rejection may not be confined to the trophoblast. Such regulation may also sometimes play an important -
LAMAN A. GRAY, SR
role in tumour progression.
8,9
Genetics
IMMUNOLOGY AND THE FETUS p. 27) on the lack of imof the fetus you stress the role of interac-
SIR,-In your editorial (Jan. 7,
munological rejection
tions between fetal lymphocytes, or other fetal factors, and maternal tissues. The development of a lymphoid system is, however, a late event in embryogenesis; it may be important in certain diseases involving feto-maternal relationships such as severe toxaemia and choriocarcinoma, but it can play little more than a seconciary role in the prevention of fetal rejection, which must be ensured from the moment of implantation. Avoidance of rejection is more probably due to a lack of a cell-mediated rather than lack of antibody-based immune response. 20-30% of pregnant women make antibodies against paternally derived HLA A, B, C, and Drw antigens. These antibodies do no obvious harm and their existence indicates that the maternal humoral immune system is not specifically (or non-specifically), suppressed. On the other hand, since not all pregnant women have these antibodies their presence cannot represent a general mechanism of enhancement, such as by masking antigenic sites on the trophoblast that might be the target of maternal rejection mechanisms. The problem of viviparity then becomes one of whether the
Laboratory, Department of Biochemistry, University of Oxford, Oxford OX1 3QU
C. J. BARNSTABLE W. F. BODMER
SIR,-Your editorial draws attention to the possible biologi-
significance of experiments showing that fetal lymphocytes suppress the responses of maternal lymphocytes. Umbilical-cord blood often contains small numbers of maternal erythrocytes1O so the extreme rarity of maternal (XX) lymphocytes in karyotype analyses of cord-blood lymphocyte cultures from males" was surprising. Both Olding and Oldstone" and Lawler cal
can
Goodfellow, P. N., Barnstable, C. J., Bodmer, W. F., Snary, D., Crumpton, M. J. Transplantation, 1976, 22, 595. 2. Faulk, W. P., Temple, A. Nature, 1976, 262, 799. 3. Sellens, M. H. ibid. 1977, 269, 60. 4. Doherty, P. C., Zinkernagel, R. M. Lancet, 1975, i, 1406. 5. Barnstable, C. J., Jones, E. A., Bodmer, W. F., Bodmer, J. G., Arce-Gomez, B., Snary, D., Crumpton, M. J. Cold Spring Harbor Symp. 1976, 41, 443. 6. Goulmy, E., Termijtelen, A., Bradley, B. A., van Rood, J. J. Nature, 1977, 266, 544. 7. Bach, M. L., Bach, F. H., Zarling, J. M. Lancet, 1977, i, 20. 8. Arce-Gomez, B., Jones, E. A., Barnstable, C. J., Solomon, E., Bodmer, W. F. Tissue Antigens, (in the press). 9. Zeuthen, J., Friedrick, V., Rosen, A., Klein, E. Immunogenetics, 1977, 4, 1.
567. 3. Hoover, R., Gray, L. A., Sr J. natn. Cancer Inst. (in the press). 4. Bibbo, M. Progress Report, National Institute of Child Health & Human Development, August, 1977. 5. McDonald, T. W., Annegers, J. F., O’Fallon, W. M., Dockerty, M. B., Malkasian, G. D., Jr., Kurland, L. T. Am. J. Obstet. Gynec. 1977, 127, 572.
Duhring, J. L., Smith, K., Greene, J. W., Rocklin, D. B., Blackmore, W. S. Surg. Forum, 1960, 10, 720. 11. Turner, J. H., Wald, N., Quinlivan, W. L. Am. J. Obstet. Gynec. 1966, 95,
10.
831. 12.
Olding,
L. B., Oldstone, M. B. A. Clin. Immun. Immunopath. 1974,
3, 70.
tion particles
an average period of 6.3 weeks (range 3-13 weeks) all patients received insulin injections, while the oral drugs were omitted. Throughout this period the diet remained the same. Each patient thus served as his own control. Glucose was measured by an automated method. Cholesterol, triglycerides, and lipoprotein concentrations were measured as previously described.4.5 H.D.L.-cholesterol was measured after precipitation of the lowdensity lipoproteins in the fraction with density > 1-006 g/ml. Treatment with insulin did not significantly alter the H.D.L. level as measured by its cholesterol content (see table). In the 9 patients on oral hypoglycaemic agents, the H.D.L.-cholesterol level was also unchanged-1.01+0.32 mmol/1 during oral therapy, 1-04+0-32 during insulin therapy (means ± S.D.). Our results are at variance with those of Bar-On et al. and suggest that the conclusions of the University Group Diabetes Program are not explained by the lowering of H.D.L. levels by
oral therapy.
C. JOHANSSON S. RÖSSNER G. WALLDIUS
S-104 01 Stockholm 60, Sweden
HIGH-DENSITY-LIPOPROTEIN AND MATURITY-ONSET DIABETES
SIR,-A low level of high-density lipoprotein (H.D.L.) in seems to accelerate the development of atheroscleroSiS.l.2 Bar-On et a1.3 found that H.D.L. and cholesterol in serum from 11 elderly patients on oral therapy were lower than in 12 diabetics on insulin and suggested that this might explain the University Group Diabetes Program conclusion that treatment with oral hypoglycaemic agents (tolbutamide and phenformin) increases cardiovascular mortality. However, Bar-On et al. studied two different groups of patients with an unequal sex ratio. This and other discrepancies between the groups could explain the difference in lipoprotein levels. We measured plasma-lipoprotein before and after treatment with insulin in a group of patients (8 males, 7 females, mean age 58 years, range 51-67 years) with maturity-onset diabetes and type iv hyperlipoproteinaemia. They had been treated for several years with diet alone (6 patients) or diet and oral hypoglycaemic agents sulphonylureas 8, phenformin 1. During
plasma
4. Carlson, K. J. clin. Path. 1973, 26, suppl. 5, p. 32. 5. Rössner, S. Hallberg, D. Acta med. scand. (in the press). 1. Miller, G. J., Miller, N. E. Lancet, 1975, i, 16. 2. Berg, K., Børresen, A. L., Dahlen, G. ibid. 1976, i, 499. 3. Bar-On, H., Landau, D., Berry, E. ibid. 1977, i, 761.
University of Nijmegen, Nijmegen, Netherlands
EXOGENOUS ŒSTROGENS AND OVARIAN CANCER
SIR, -Dr Annegers and his colleagues’ suggest that the data in
our
preliminary communication2"do
=
=
not
point
as
strongly
with exogenous oestrogen use" as we suggested. They note that the "expected" values calculated for ovarian cancer did not take into account the proportion of women in the general population who are not at risk-namely, those with surgically removed ovaries. By adjusting for the 10% prevalence-rate’ of bilateral oophorectomy prevailing in the Mayo Clinic population, they estimate that our expected numbers for oestrogen-treated women are 11 % too low, and that two of the three relative risks values are not statistically significant. In case others may have been misled by our paper, we would like to emphasise the following points: (1) We did not claim that ovarian cancer was associated with the use of exogenous oestrogens generally, but rather with specific oestrogen, diethylstilboestrol (D.E.S.). We tried to stress this by entitling the paper Stilboestrol (Diethylstilboestrol) and the Risk of Ovarian Cancer. (2) The relative risk (R.R.) for the association between D.E.s. use and ovarian cancer was 30.0 (confidence interval, 6.2-87-7). These values are not changed by adjusting for the prevalence of oophorectomy in the general population, whether one uses the Mayo Clinic oophorectomy prevalencerate or makes the extreme assumption that every woman in the general population with a hysterectomy has also undergone bilateral oophorectomy, (3) Our most conservative estimate of the R.R. for all
increased risk of ovarian
cancer
4.
Demacker, P.
2.
Hoover, R., Gray, L. A., Sr, Fraumeni, J. F., Jr ibid. p. 533.
N. M., v. Oppenraay, J., Baadenhuysen, H., Jansen, A. P. Clin. chim. Acta, 1975, 64, 45. 5. Demacker, P. N. M., Vos-Janssen, H. E., Jansen, A. P., van ’t Laar, A. Clin. Chem. 1977, 23, 1238. 1. Annegers, J. F., O’Fallon, W., Kurland, L. T. Lancet, 1977, ii, 869 (see also
p. 1188).
PLASMA-LIPOPROTEINS DURING ORAL THERAPY AND INSULIN THERAPY
Results are means < S.D. Student’s t test for paired data. *L.D.L. low-density lipoprotein. V.L.D.L. very-low-density lipoprotein.
P. N. M. DEMACKER J. A. LUTTERMAN A. VAN ’T LAAR
Department of Medicine,
to an
Department of Internal Medicine and King Gustaf V Research Institute, Karolinska Hospital,
A. F. H. STALENHOEF
326 oestrogen users was 2.4 (confidence interval, 1.0-4.8). This declines only slightly to 2.2 (0.9-4.3) when the Mayo Clinic adjustment factor for oophorectomy is used. In a more detailed discussion of the study population,3 we, calculated an estimate of R.R. based on the extreme assumption noted in (2) above. Even this "overadjustment" yields a 2-fold increase in risk for ovarian cancer. We see little difference in all of these estimates. The corrections have shifted the lower bound of the confidence interval below 1.0, so the values are not statistically significant at the 5% level. However, nominal levels of significance should not be overemphasised, since our risk estimate with a P value of 0.067 seems as meaningful as it was with a p value of 0.040. (4) The associations were based on only 8 cases of ovarian cancer (including 3 for the D.E.S. group), and thus may be spurious for a number of reasons (e.g., bias, confounding, chance). Our suspicion of a causal relationship was based not on statistical probabilities but on the magnitude of the risk after D.E.S. (30-fold), plus the capacity of D.E.S. to cause ovarian tumours in dogs, and the possibility that it may explain the rising incidence of this cancer in postmenopausal women. These findings, coupled with the magnitude of the exposure to D.E.S. in the population, led to our preliminary report in the hope that others would investigate the question. Since then, results of a follow-up of women who participated in a randomised clinical trial of D.E.S. for the prevention of miscarriage in the early 1950s has become available.4 4 women in the D.E.S. group have had ovarian cancer, compared with 1 in the placebo group. The numbers are again small, but provide further cause for concern. Annegers et al. report preliminary results from a case-control study of ovarian cancer, showing no relation to the use of exogenous cestrogens, While this is useful information, it is not relevant to the issue of the risk associated with D.E.S. exposure since an earlir Mayo Clinic survey’ indicated very limited long-term use of D.E.S. in that population. ,
Environmental Epidemiology Branch, National Cancer Institute, Bethesda, Maryland 20014, U.S.A.
Department of Obstetrics and Gynecology, University of Louisville, Louisville, Kentucky
ROBERT HOOVER JOSEPH F. FRAUMENI, JR
maternal cellular immunity is impaired or the fetus is not immunogenic. There have been many reports of serum factors than can depress cellular immunity, although this effect is often measured by the non-specific criterion of inhibition of mitogen stimulation. However, most of these factors are not specific for paternal antigens they are not found in all pregnancies, and most of them appear late in pregnancy. They cannot be primarily responsible for the lack of fetal rejection, though they may play a role in later events during pregnancy. We have shown’ that purified plasma membranes from term placentae contain amounts of HLA A and B antigens sufficiently small for their presence to be accounted for by contamination from connective tissue. Others have found,2 by immunofluorescence studies, that these antigens are absent from the human trophoblast. These results strongly suggest that the syncytiotrophoblast expresses HLA A and B antigens not at all or at most in greatly reduced amounts. A similar conclusion has been reached for the mouse.3 Lack of HLA A and B antigens on the tissue in contact with the maternal circulation could abrogate cell-mediated attack in several ways. The work of Doherty and Zinkernagel4 suggests that, in the mouse, cell-mediated killing of cells depends upon recognition of H 2K or D determinants either complexed to a foreign determinant giving rise to an "altered self’ or,’ in an early effector stage of the killing reaction. A similar role has been suggested for the HLA A, B, and C antigens in man.’ The work of Bach et al. and others suggests that the HLA A, B, and C antigens are a necessary part of the antigenic stimulus required to generate cytotoxic T lymphocytes. Thus, lack of HLA A, B, and C antigens on trophoblasts could act either to prevent the generation of cytotoxic T lymphocytes or to prevent the action of such cells when directed against any fetal or polymorphic antigens for which there is fetomaternal incompa-
tibility. It therefore seems most probable that lack of immunogenicity of the trophoblast with respect to cellular immunity, rather than depression of maternal immunocompetence, is the primary mechanism involved in the "immunological inertia of
viviparity". Switching off HLA A, B, and C antigen expression to escape immunological rejection may not be confined to the trophoblast. Such regulation may also sometimes play an important -
LAMAN A. GRAY, SR
role in tumour progression.
8,9
Genetics
IMMUNOLOGY AND THE FETUS p. 27) on the lack of imof the fetus you stress the role of interac-
SIR,-In your editorial (Jan. 7,
munological rejection
tions between fetal lymphocytes, or other fetal factors, and maternal tissues. The development of a lymphoid system is, however, a late event in embryogenesis; it may be important in certain diseases involving feto-maternal relationships such as severe toxaemia and choriocarcinoma, but it can play little more than a seconciary role in the prevention of fetal rejection, which must be ensured from the moment of implantation. Avoidance of rejection is more probably due to a lack of a cell-mediated rather than lack of antibody-based immune response. 20-30% of pregnant women make antibodies against paternally derived HLA A, B, C, and Drw antigens. These antibodies do no obvious harm and their existence indicates that the maternal humoral immune system is not specifically (or non-specifically), suppressed. On the other hand, since not all pregnant women have these antibodies their presence cannot represent a general mechanism of enhancement, such as by masking antigenic sites on the trophoblast that might be the target of maternal rejection mechanisms. The problem of viviparity then becomes one of whether the
Laboratory, Department of Biochemistry, University of Oxford, Oxford OX1 3QU
C. J. BARNSTABLE W. F. BODMER
SIR,-Your editorial draws attention to the possible biologi-
significance of experiments showing that fetal lymphocytes suppress the responses of maternal lymphocytes. Umbilical-cord blood often contains small numbers of maternal erythrocytes1O so the extreme rarity of maternal (XX) lymphocytes in karyotype analyses of cord-blood lymphocyte cultures from males" was surprising. Both Olding and Oldstone" and Lawler cal
can
Goodfellow, P. N., Barnstable, C. J., Bodmer, W. F., Snary, D., Crumpton, M. J. Transplantation, 1976, 22, 595. 2. Faulk, W. P., Temple, A. Nature, 1976, 262, 799. 3. Sellens, M. H. ibid. 1977, 269, 60. 4. Doherty, P. C., Zinkernagel, R. M. Lancet, 1975, i, 1406. 5. Barnstable, C. J., Jones, E. A., Bodmer, W. F., Bodmer, J. G., Arce-Gomez, B., Snary, D., Crumpton, M. J. Cold Spring Harbor Symp. 1976, 41, 443. 6. Goulmy, E., Termijtelen, A., Bradley, B. A., van Rood, J. J. Nature, 1977, 266, 544. 7. Bach, M. L., Bach, F. H., Zarling, J. M. Lancet, 1977, i, 20. 8. Arce-Gomez, B., Jones, E. A., Barnstable, C. J., Solomon, E., Bodmer, W. F. Tissue Antigens, (in the press). 9. Zeuthen, J., Friedrick, V., Rosen, A., Klein, E. Immunogenetics, 1977, 4, 1.
567. 3. Hoover, R., Gray, L. A., Sr J. natn. Cancer Inst. (in the press). 4. Bibbo, M. Progress Report, National Institute of Child Health & Human Development, August, 1977. 5. McDonald, T. W., Annegers, J. F., O’Fallon, W. M., Dockerty, M. B., Malkasian, G. D., Jr., Kurland, L. T. Am. J. Obstet. Gynec. 1977, 127, 572.
Duhring, J. L., Smith, K., Greene, J. W., Rocklin, D. B., Blackmore, W. S. Surg. Forum, 1960, 10, 720. 11. Turner, J. H., Wald, N., Quinlivan, W. L. Am. J. Obstet. Gynec. 1966, 95,
10.
831. 12.
Olding,
L. B., Oldstone, M. B. A. Clin. Immun. Immunopath. 1974,
3, 70.