Marburgvirus is but one example of the evils of nature that are our real enemies in the living world. It is very unlikely to discriminate between Democrat or Communist or Maoist. And as human society is now organized, our encounters with such threats will not for long be just near misses.

[From the Oak Ridger, Oak Ridge, Tenn., Sept. 9, 1968]

Science and Man

HAILING ORNL EXAMPLE OF "GENETIC ENGINEERING"

(By Joshua Lederberg)

Recent advances in fundamental biochemistry have generated much speculation about genetic engineering, the calculated modification or design of an organism based on new knowledge of the genetic code. Now the first factual example to illustrate such an effort has been reported.

(The report in Nature magazine, by Drs. Stanfield Rogers and Peter Pfuderer, of Oak Ridge National Laboratory, also indicates the broadening functions of these laboratories, for the research was sponsored jointly by the Atomic Energy Commission and by the National Cancer Institute.)

In higher organisms, the genetic code is embodied in the DNA of the chromosomes of each cell. This is then translated into multiple copies of RNA "messenger" molecules and these in turn regulate the formation of the various proteins that make up the cell.

Some simple viruses, like the tobacco mosaic virus, have a simplier life cycle that bypasses the DNA. Instead, their RNA functions both as the hereditary material, being able to replicate as such, and as its own messenger. The tobacco mosaic virus, a disease of tobacco plants, has been much used for basic studies of virus biochemistry.

Previous studies pioneered by Dr. Marshall Nirenberg had established that the triplet in messenger RNA that reads "A-A-A-" (A standing for adenine) is translated into the amino acid protein chain. Rogers and Pfuderer used a specific enzyme to add a chain of about 18 A's to the end of the natural tobacco mosaic virus molecule. After purifying the product, they then used the modified complex to infect tobacco plants. If the dangling chain of AAA-triplets was translated according to modern theory, they reasoned, they should find new substances in the infected leaves, namely strings of up to six lysines.

This is precisely what they report.

The most important point that remains to be investigated is how the modified complex functions as a virus-that is, whether it replicates itself according to expectation. This information should be available very soon.

The result has mainly theoretical interest, by corroborating contemporary speculations about ways in which modified viruses-and, by implication, genes of higher organisms-can be artificially designed. We can expect to hear soon of methods of grafting two preexisting sequences together, so that tobacco mosaic virus could be extended with a more meaningful message than one for monotonous lysine strings. This could then lead, for example, to the production of virusinfected plants of precious and rare products like human growth hormone.

Even monotonous sequences may have great economic significance. The AAA . . modification theoretically should exaggerate the amount of lysine manufactured by the infected plant. Lysine is one of the essential amino acids and many cases of malnutrition stem from a deficiency of lysine in the locally available food crops. We can readily visualize the engineering of tempered plant viruses that would augment the value of the protein produced by conventional crops.

The extension of this approach to animal viruses, furthermore, is one of the most promising approaches to a fundamental attack on genetic defects in man: the missing genes might be reintroduced into afflicted children by vaccinating them with specially engineered viruses.

There is, as always, another side to the coin. The lead editorial in the same issue of Nature remarks: "It is hard to decide why discussion of chemical and biological warfare has become so fashionable." But the obvious application of this new science to military progress in anti-human, viral weapons certainly adds new substance to our anxieties about biological warfare.

[From Book World, June 15, 1969]

ONE-THIRD OF EUROPE DEAD OF THE PLAGUE

(By Joshua Lederberg)

The plague has always slumbered here and there around the world. About the year 1340, it flared up in Central Asia near Lake Issyk Kul, to the east of Tashkent. Its ravages in Asia and India are uncounted. By 1346 it had reached the Crimea, a scene of brush-fire conflicts between the Genoese and Tartar empires. With an intuitive sophistication in bacteriological warfare, the Tartars infected the Genoese fort by catapulting their own corpses over its walls. The stricken Genoese fled back to Italy, carrying the plague with them and igniting an epidemic, known to history as the Black Death, that spread rapidly across Europe. During the next two-and-a-half years it is estimated that about one-third of Europe's population was killed by it. But we do not have accurate statistics, and other guesses vary from 20 to 60 per cent, which seems a reasonable range for the plague's impact on different communities. The survivors were beset by repeated but much milder outbreaks for the next sixty years.

In this excellent book, Philip Ziegler suggests that Europe had been weakened by several decades of economic and agricultural decline, but we have very little scientific information to explain the ferocity of the epidemic. Most specialists believe that the present-day plague bacillus would be quite capable of a similar recrudescence except for the availability of antibiotics and pesticides to control rat fleas, if not the rats themselves, both of which maintain the potential contagion.

Fourteenth-century Europe only dimly understood the role of contagion, and was told by its leaders that the plague was above all a visitation of a vengeful God for vaguely specified sins. In Central Europe the Jews were persecuted as scapegoats, according to long-established custom, but the persecutors discovered no immunity to plague by that route.

One historical side effect of the Black Death in Scandinavia may have been a weakening of support for the dying colonies of Norsemen in Greenland. This foothold collapsed before the rediscovery of the New World by a Genoese in 1492— and the fact that we celebrate Columbus Day is at least one historical consequence.

Ziegler tells the story with ample but not stifling documentation and with an engaging balance of detached statement and illustration. The statistical arguments about the mortality percentages are a necessary building stone and, happily, are assembled in one chapter. With equal discretion, Ziegler has written another chapter devoted to a fictional synthesis of the experience of a composite village in England. Though the primary data are sparse and Ziegler takes proper pains to point this out, he builds a most credible, restrained account.

His larger view brings out few conclusions that will surprise anyone who has read shorter accounts of the Black Death (like William L. Langer's in Scientific American, February, 1964) or chapters in history texts and encyclopedias.

The surprising thing about the Black Death is that it was assimilated into the historical life of the period without working any extraordinary change in its larger institutions. Despite the misery of individuals and families, most communities survived; government rarely broke down; inheritance taxes were dutifully collected on the suddenly increased numbers of estates; there were few direct challenges to authority or shifts in centers of power. One can speculate about the role of the Black Death in accelerating the breakdown of feudalism, but it is hard to substantiate an influence commensurate with the mortality figures. The Black Death was an unprecedented source of private bereavement, but it had minimal social impact, perhaps because it visited rich and poor alike, with little discrimination. Ziegler points out that the loss of teachers and clergy may have sped up the vernacularization of education and of the liturgy. We might also guess that it inspired some skepticism about faith in the power and truth of the Church, but this assumes a degree of rational intellectual commitment that is rarely evoked by real distress.

APPENDIX D

STATISTICAL ANALYSIS OF TERATOGENIC EFFECTS OF PESTICIDES

(By D. W. Gaylor, M. D. Hogan)

BIOMETRY BRANCH, NATIONAL INSTITUTE OF ENVIRONMENTAL HEALTH SCIENCES, SEPTEMBER 5, 1969

Introduction

This report contains a statistical analysis of the teratogen data collected by the Bionetics Research Laboratories of Litton Industries during the period 19651968 for the National Cancer Institute. The teratogen screening study contained 53 compounds chosen as representative of several chemical classes of pesticides and industrial compounds.

Four strains of mice were used: C57BL/6, AKR, C3H, and A/Ha. Most of the studies were performed with the C57BL/6 strain. A hybrid fetus resulting from mating a C57BL/6 female with an AKR male was used to study a few compounds.

Most compounds were administered subcutaneously in solutions of dimethylsulfoxide (DMSO). Water soluble compounds were administered in saline solutions. In some cases, the water soluble compounds were studied using both DMSO and saline. Compounds administered orally were suspended in a 50% solution of honey. A group of untreated controls were included as well as controls receiving only DMSO, saline, or honey. Females entered the study at the onset of their first pregnancy as indicated by a vaginal plug. Subcutaneous injections of 100 μl per mouse were made at the nape of the neck. Oral administration was by gastric intubation using a volume of 100 μl per mouse. Compounds were administered daily beginning on the 6th day of pregnancy and continuing through the 14th day (15th day for AKR mice). The mice were sacrificed on the 18th day (19th day for AKR mice) of gestation. Upon sacrifice, fetuses were examined for anomalies. Approximately two-thirds of the fetuses were then stored in Bouin's solution until they were necropsied. The remaining fetuses were treated with alizarin red stain. The numbers of resorption sites and dead fetuses were also noted.

Because of the marked complexity of teratologic investigations in general and, in particular, because of the nature of the experimental design underlying the Bionetics study, several assumptions had to be made in order to statistically analyze the data from this study.

All statistical analyses were performed on a per litter basis rather than a per fetus basis, since initial investigations indicated that the occurrences of anomalies among fetuses within litters were correlated. The large litter-to-litter variation may be a reflection of some maternal effect, an indication of the effective dose level of the compound actually reaching the fetuses, experimental variation, or, as is most likely, some combination of the three factors.

Another important consideration that should be noted is that while controls were run periodically throughout the duration of the study, no effort was made to match compounds and controls with respect to either route or dates of administration since no time trends were anticipated. While there were no statistically significant time trends within the various control groups in terms of the onset of fetal anomalies in the C57BL/6 mice, the incidence of fetal mortality was certainly time-dependent in this strain, with 1965 being characterized by a low incidence of pre-natal deaths. Furthermore, there was a period of approximately six months, extending from the latter part of 1965 into early 1966, during which no control animals were run. During this period a change in the substrain of C57BL/6 mice used in the study took place. Finally, among abnormal litters, i.e., litters containing at least one abnormal fetus, there was some suggestion that the distribution of abnormal fetuses per litter was stochastically larger in the DMSO controls than it was in the untreated controls. Thus, one cannot

ignore the possibility of the existence of a time x strain x solvent interaction that is undetectable in the controls because of the level of background teratologie activity is relatively low. This potential interaction effect could either enhance or dissipate the effect of any given compound, depending on the conditions under which it was administered. Thus, the data were separated by both time period and solvent for the purposes of analysis.

Similarly, an increase in fetal anomalies in the DMSO controls of the AKR mice was noted after November, 1966. Thus, the AKR data were analyzed separately in two time periods.

One should also bear in mind that not all compounds were administered at more than one point in time or in more than one solvent or strain. Thus, in general the compounds in the study cannot be compared for teratogenic potential, since those that were tested extensively were more likely to show some adverse effect and, perhaps, less likely to appear consistent over time, solvent, and/or strain.

As has already been noted, approximately two-thirds of the fetuses were stored in Bouin's solution until they were necropsied, with the remainder being treated with alizarin red stain. However, in many instances the proportion of necropsied fetuses was slightly higher for the compound under investigation than for the corresponding controls. It is doubtful if this discrepancy could have any appreciable effect on the conclusions since the incidence of anomalies detectable only by necropsy among control animals was relatively low. Furthermore, if all of the control and pesticide-treated mice had been necropsied, the significance of the differences observed in this study would be intensified. Thus, no effort was made to correct for inequalities in the necropsy/stain ratio in the present analysis.

Next one should note that no attempt was made to correct for differences in litter sizes or sex-ratios within litters, since both of these factors may be, at least in part, a reflection of the effect of the compound under investigation.

In conclusion it must be re-emphasized that the Bionetics Study was a screening study for detecting teratogenic potential in a group of pesticides and commercial compounds, and that the results apply only to the strains of mice used with specific routes of administration of relatively high doses.

Results

Those test conditions with pesticides and industrial compounds which resulted in statistically increased levels of anomalies in C57BL/6 mice are listed in Table 1 and for AKR mice in Table 2. The proportion of abnormal litters gives the proportion of litters containing one or more abnormal fetuses. This gives a measure of the prevalence of anomalies across litters. The proportion of abnormal fetuses per litter gives a measure of the prevalence of anomalies within litters. The proportion of abnormal fetuses per litter for litters containing abnormal fetuses gives a measure of the prevalence of anomalies within affected litters. A significant increase of dead fetuses and resorptions is listed. Some tests were conducted on only one day or adjacent days as listed. Eye anomalies (mainly microphthalmia and anophthalmia) accounted for approximately 50% of the individual anomalies in C57BL/6 mice. To a large extent the results in Table 1 reflect changes in the incidence of eye anomalies. Thus, statistically significant increases in anomalies other than eye anomalies are listed in the last column of Table 1.

The negative controls consisted of untreated animals and mice administered 100 μl of DMSO, saline, or honey.

The known teretogens, Trypan Blue and Ethylene Imine, Table 1. and 6-aminonicotinamide, Table 2, showed elevated levels of anomalies, although the latter control wasn't consistent over all dose levels.

Only those test conditions which resulted in statistically elevated incidences of anomalies are listed in Tables 1 and 2. Some of these pesticides and industrial compounds gave no increase in anomalies when tested in other solvents, species. or dose levels. These conditions are listed in Table 3. It must be emphasized that failure to detect a statistically significant increase in anomalies may only be a reflection of experimental insensitivity due to experimental and biological variation and insufficient number of litters. Thus, compounds showing no increases cannot be considered non-teratogenic. For example, Trypan Blue in DMSO at the highest dose level tested. 37.5 mg/kg, did not show an increase in anomalies possibly due to higher fetal mortality.

Standard corrected 2 x 2 chi-square tests [ref. 1] were used to compare the proportion of abnormal litters for the compound with the controls in the same solvent. In the cases where tests were conducted in two time periods, the results from the two chi-squares were combined by the method described in ref. [1]. The levels of statistical significance for the combined tests are listed under the total column for proportion of abnormal litters.

The distribution of the proportion of abnormal fetuses per litter for compounds were compared with the appropriate control distribution by use of the nonparametric Mann-Whitney U-test, ref. [2]. This test requires that the proportion of abnormal fetuses per litter is independent from litter to litter, but requires no assumption about the frequency distribution of these proportions. Again, where litters were run in both time periods, the significance level for the combined tests is given under the total column.

The brackets include groups which were combined before statistical tests were conducted.

REFERENCES

1. Snedecor, G. W. and Cochran, W. G., Statistical Methods, 6th ed. Iowa State Univ. Press, Ames, Iowa (1967).

2. Steel, R. G. D. and Torrie, J. H., Principles and Procedures of Statistics. McGraw-Hill Book Co., Inc., New York (1960).

Summary

One compound, 2,4,5-T was tested more extensively than the others. It showed statistically increased proportions of litters affected and increased proportions of abnormal fetuses within litters in both DMSO and Honey for both C57BL/6 and AKR mice. In particular, cleft palate and cystic kidneys were significantly more prevalent. In addition a hybrid strain resulting from a C57BL/6 female and AKR male showed significant increases in anomalies, in particular cystic kidney, when administered at 113 mg/kg of body weight in DMSO.

PCNB produced an increase in renal agenesis across litters and within litters when administered from days 6-14 or days 6-10 of pregnancy. However, renal agenesis was not produced when PCNB was administered only from days 10-14 of pregnancy.

In addition to 2,4,5-T and PCNB, compounds which resulted in statistically significant increases in the proportion of litters containing abnormal fetuses and in the increased incidence of abnormal fetuses within litters were: Captan, Folpet, 2,4-D Isooctyl Ester, 2,4-D Butyl Ester, 2,4-D Isopropyl Ester, Sevin, IPC, 2,4-Dichlorophenol, and 2,4-D. With the exception of 2,4,5-T, the compounds cited above produced elevated incidence of anomalies in one strain and one solvent only.

Phenylisothiocyanate and Captax showed only an increase in the proportion of litters affected. Compounds showing only an increase in the proportion of abnormal fetuses within litters were: a-naphthol, and 2,4-D Methyl Ester.

Sevin plus Piperonyl Butoxide did not show an increase in non-specific anomalies, but resulted in significantly more cystic kidneys for doses above 10 mg/kg Sevin and 100 μl/kg Piperronyl Butoxide in DMSO.

As noted previously, failure to detect statistically significant increases of anomalies may be due to insensitivity resulting from experimental variation and small numbers of litters tested. In addition, higher fetal mortality among some of the negative compounds may be selectively eliminating abnormal fetuses.