Food systems: perspectives on demographics and affluence, food supply and consumption.

Global population may double by 2020 but the Malthusian specter of rapid population growth outracing slower increases in production will continue to be a false alarm. A vast array of agricultural technologies have the capacity to increase output 10-fold, perhaps as much as 100-fold. Discovery of a sweetener 54,000 times sweeter than sucrose (cane or beet sugar) indicates the magnitude of prodigious increases portended by new technologies. Productive agriculture, however, has become capital intense, limiting its availability in poorer nations. Increased production is the key to low prices and affordable supplies. In a world continuing to face starvation, there is no place for government policies purposely limiting supplies and artificially propping prices at high levels that place life-sustaining food beyond means of the poor. Affluence provides financial wherewithal to secure an adequate diet. Unfortunately, an estimated 25% of the world's population go hungry and face starvation. The specter of starvation may afflict as many as 600 million, and malnutrition, another 150 million by the year 2020. Improving self-sufficiency in these nations will remain a top humanitarian concern.

The sulphydryl compound, WR 2721, has been developed as a selective protector of normal tissues against ionising radiation (Yuhas, 1980a). More recently a number of papers have appeared which indicate that this compound is also able to selectively protect normal tissues in the mouse or rat against a number of cytotoxic drugs whilst having little or no effect upon the anti-tumour efficacy of these agents (Yuhas, 1979;Yuhas & Culo, 1980;Yuhas et al., 1980;Wasserman et al., 1981). In our own study of cyclophosphamide (CTX) in combination with WR 2721, however, we found significant protection of two mouse tumours against CTX, whilst seeing less protection of normal tissues than reported by others (Twentyman, 1981). In this paper, we report the results of a much larger series of experiments in which protection by WR 2721 against the effects of a range of cytotoxic drugs has been studied in both tumour and normal tissues of the mouse.

Materials and methods Mice and tumours
The mice used in these studies were inbred C3H/He supplied by OLAC. Females were used in most experiments, but males were used occasionally. Mice entered experiments at age 12-16 weeks and weighed 20-28 g.
Tumours used were the KHT and RIF-1 sarcomas, both of which originated in C3H/Km mice at Stanford University, California, and which have been previously described (Kallman et al., 1967;Twentyman, et al., 1980). The methods used for tumour cell inoculation into the gastrocnemius Received 14 June 1982;accepted 23 September 1982. 0007-0920/83/010057-07 $01.00 muscle of the hind limb and subsequent measurement of tumour growth, including conversion of leg measurement to tumour weight, have also been described (Twentyman et al., 1979). The endpoint of growth delay was calculated from the geometric means of the times taken for individual tumours to reach 4 x the initial groupmean treatment volume. Tumours were treated in the size range 300-600mm3.
Nine to 12 mice were used in each treatment group.
White-cell counts Blood samples were taken from unanaesthetized mice by cutting a few mm from the end of the tail with a scalpel. A capillary pipette was then used to draw up 0.015ml of blood, which was diluted in 20 ml of "Isoton" (Coulter Electronics Ltd). Six drops of "Zapoglobin" were added to lyse the red cells, and counts were made on an electronic particle counter (Coulter Electronics-Model ZBI).

Drugs
WR 2721 (S,2-(3-aminopropylamlno)ethyl-phosphorothioc acid) was kindly supplied by the Drug Development Branch of the U.S. National Cancer Institute. Most of the experiments were carried out with a sample of batch NF LOT AJ 68.2 supplied in December 1979. We have recently (March 1982), however, obtained a sample of batch NH LOT AJ 68.4 and this was used in a number of experiments specified in the Results section. Both specimens arrived packed with dry ice and were subsequently kept at -20°C. The drug was dissolved in Hanks balanced salt solution immediately before use and was injected by the i.p. route at a volume of O.Olmlg-1 30min before cytotoxic drug administration. Cytotoxic drugs were obtained, dissolved and administered as shown in Table I.

Toxicity of WR 2721 alone
In an earlier study we reported the acute LD50 (7 days) of batch AJ 68.2 in female C3H mice to be 550mg kg-. Two recent small experiments with this same batch have produced values of 650 and 550mgkg-. In the second of these experiments batch AJ 68.4 was also tested and gave an identical value of 550mgkg-'.
We also found (Honess & Twentyman, unpublished) that whereas 200mg kg-1 of batch AJ 68.2 produced essentially no change in mouse body temperature, a dose of 400mgkg-' produced a fall of 4-50C at 1 h after administration with recovery by 4-6h. A comparative study of batches AJ 68.2 and AJ 68.4 has now been made and the change in 1 h body temperature with dose of WR 2721 was similar for the 2 batches.
In view of the potential complications due to hypothermia produced by 400mgkg-1 of WR 2721 and our previous finding of marked tumour protection against CTX by this dose (Twentyman, 1981), many of the current experiments have used a WR 2721 dose of 200mgkg-. A number of experiments have, however, also included the higher dose of 400mgkg-. Acute lethality of cytotoxic drugs Results of experiments to determine the median  Table II. These data are based on results of experiments in which 6 groups of 5 mice were treated with graded doses of the cytotoxic drugs. For CHL, most deaths occurred within 48 h of drug administration. For CCNU and cis-P almost all deaths occurred 5-8 days after treatment. For CTX deaths began at Day 7 and continued throughout the 30-day observation period.
White cell count The depression of peripheral white cell count at 3 days after CTX is shown in Figure 1. It may be seen that no significant change in the pattern was brought about by pretreatment with WR 2721 at either of the dose levels used. This conclusion was confirmed in a repeat experiment. Similar experiments in which CCNU and cis-P were combined with either 200 or 400mgkg-1 of WR 2721 produced similar results to that for CTX (i.e. no difference in depression of white cell count at Day 3). For CCNU, white cell counts were followed for a further 7 days and recovery was not modified by WR 2721 treatment.
Tumour response The effect of WR 2721 pretreatment on the anti-   Tables III and IV. It may be seen that in the RIF-1 tumour we confirm our previous finding of marked protection against CTX by 400mgkg-1 of WR 2721 (Twentyman, 1981). In addition we have now also looked at the effect of growing the tumour intradermally in the flank instead of intramuscularly in the leg. There was again a tendency for WR 2721 to reduce the CTX-induced growth delay but the differences in this experiment were not significant at the 95% confidence level. Also in the RIF-1 tumour, protection of the tumour against cis-P is brought about by WR 2721. The data shown in Table III and Figure 2 indicate that the main effect is at lower doses of cis-P, and that for a growth delay of 4 days, the protection factor is around 1.4. The data shown in Table IV indicate that only very small protection factors are seen for WR 2721 in combination with CCNU, MEL or CHL in the KHT tumour. We do not believe that this is a tumour difference, since in our earlier study (Twentyman. 1981) a clear protection by WR 2721 of the KHT tumour against CTX was seen. With such small effects, the differences in Table IV     the 7 experiments in which a direct comparison was possible, the groups receiving WR 2721 showed a smaller growth delay in 6 cases and the delays were identical in the seventh.

Discussion
The original data on the action of WR 2721 as a differential radioprotector of normal tissues (summarised by Yuhas, 1980a) showed two notable features. Firstly, the protection of a variety of different normal tissues in the mouse and in the rat could be obtained, with protection factors frequently in excess of 2.0 (neural tissue, however, being a notable exception). Secondly, with only a rare exception, no protection of tumours was found.
More recently, however, a number of papers have appeared which report significant radioprotection of a number of different murine solid tumours (Rojas et al., 1982;Clement & Johnson, 1982) and of micrometastases in the lungs of mice (Milas et al., 1982).
This same course of events has been followed with regard to data on the ability of WR 2721 to protect normal tissues selectively from the effects of cytotoxic drugs. Reports of protection of normal tissues against nitrogen mustard (Yuhas, 1979), cis-P (Yuhas & Culo, 1980) and CTX  were all accompanied by findings of a lack of protection of tumours. Using the bone marrow CFUs assay in LAF1 mice, Wasserman et al., (1981) examined the ability of WR 2721 to protect against nitrogen mustard, CTX, BCNU, cis-P and 5fluorouracil and obtained protection factors in the range 1.5 to 4.6. In a parallel study of the growth delay in EMT6 tumours little or no protection was seen (although it is not possible to estimate protection factors from the data). More recently, however, our own studies with CTX (Twentyman, 1981) and those of Clement & Johnson (1982) with MEL and CTX have shown significant protection. It must therefore be recognised that differential radioand chemo-protection by WR 2721 are by no means absolute and that detailed studies of therapeutic ratios using a variety of normal tissue endpoints are required.
The dose of WR 2721 used and its time of administration with respect to radiation or cytotoxic drugs are important parameters to be considered in any study of interactions. In most radiation studies a time of 15-30min has been chosen as this allows peak levels of drug to be attained in most normal tissues whilst allowing little time for the slower absorption into tumours (Yuhas, 1980b). A similar rationale has been used in chemoprotection experiments (Yuhas, 1979;Yuhas & Culo, 1980;Yuhas et al., 1980;Twentyman, 1981). A potential artefact of drug interaction in the bloodstream has, however, been pointed out when very short times (5-15min for nitrogen mustard) are used (Yuhas, 1979).
Doses of WR 2721 used have generally been in the range of 200-600mgkg-1. In studies of radiation-induced haemopoietic death in 4 strains of mice, Yuhas (1980a) found that a protection factor of 2.0 was achieved at a WR 2721 dose of 200mgkg-1, compared with a value of 2.7 in the dose range 400-600mgkg-. For radiation damage to the mouse jejunum, a plateau of protection was achieved at a WR 2721 dose of 200mgkg-t, although further protection was seen in the mouse testis by increasing the dose from 300-500mgkg-1 (Milas et al., 1982). Chemoprotection against nitrogen mustard (30-day survival) gave a protection factor of 1.5 for 200mgkg1 of WR 2721 compared with 1.9-2.0 at 400 mgkg 1, subsequently falling back to 1.5 at 500mgkg-' (Yuhas, 1979).
Protection against the renal toxicity of cis-P in the mouse (as measured by day-5 elevation of blood urea nitrogen) was by factors of 1.2 and 1.5 at WR 2721 doses of 100 and 200mgkg-' respectively . In summary, therefore, although increases in protection factors are sometimes seen above a dose of 200 mg kg-1 of WR 2721, the major component of protection is usually seen at this dose level with typical protection factors of around 1.5.
The data presented in this paper are, therefore, somewhat at variance with other reports on differential chemoprotection. In our earlier paper (Twentyman, 1981) we obtained a mean protection factor of 1.25 for 400mgkg-t of WR 2721 with CTX. A single experiment using 200mg kg-I, however, gave a higher protection factor of 1.48. Two further experiments, now reported in this paper give a combined value of 1.17 for 200mg kg-1, thus reducing the weight of the earlier relatively high value. For CCNU, the mean value at 200mg kg-is 1.1, whereas a single experiment at 400mg kggave no protection (factor =0.85). Similarly for 300mgkg-1 of WR 2721 in conjunction with CHL a protection factor of 1.19 was obtained. For WR 2721 together with cis-P, a marked dose dependence was seen, values of 1.06 and 1.15 being obtained at 200mg kgof WR 2721 but four values between 1.5 and 1.9 at 400 mg kg-'.
Using depression of peripheral white cell count at Day 3 after drug treatment, however, we found no protection against CTX, CCNU, or cis-P at either 200 or 400mgkg-' of WR 2721. This result, in particular, is in marked contrast with data of Wasserman et al., (1981) where protection factors for mouse bone marrow CFUs of 2.4 and 3.2 were obtained for CTX and cis-P respectively (together with values of 4.6 and 1.5 for nitrogen mustard and BCNU). Clearly there are considerable differences in the target cell populations for the two assays. The 3-day nadir of white cell count is likely to reflect the effect of the drugs mainly upon the proliferating committed precursors of the white cell series with the subsequent recovery rate being determined by the earlier precursors (including CFUs). Although most progenitors of the granulocyte series are located in the bone marrow, lymphoid precursors are more widely distributed. As around 70% of peripheral white cells in the mouse are lymphocytes, drug effects in sites other than the marrow will be very important in determining the Day 3 peripheral count. Site-dependent differences in drug levels or in the degree of interaction between WR 2721 and the various cytotoxic drugs may, therefore, be involved in these conflicting results. In addition, it should be noted that the CFUs assays in the study of Wasserman et al., were carried out at 2h after drug administration. It is likely that a 5°C drop in mouse body temperature (as caused by 400mgkg-t of WR 2721-see Materials and methods) will cause considerable changes in cytotoxic drug activation and metabolism and the use of such a short time of assay may be misleading if drug availability times are considerably prolonged. This problem does not, of course, arise when considering an in situ endpoint such as white cell count depression.
We have confirmed our earlier finding of tumour protection against CTX by WR 2721, being greater at 400mgkg-1 than at 200mgkg-l. We have also found protection against cis-P at 200 and 400mgkg-' of WR 2721. For the other cytotoxic agents there appeared to be a trend towards tumour protection, but with very small protection factors.
In conclusion, therefore, our data for differential chemoprotection by WR 2721 are distinctly less encouraging than the balance of other data in the literature. In most of our LD50 experiments, the protection factors were very small and in none of our white cell count experiments was significant protection seen. Only for cis-P in combination with 400mgkg-1 of WR 2721 was protection in excess of 1.5 seen for LD50. For this drug (and for CCNU) almost all deaths in the LD50/30 experiments occur 5-8 days after drug administration and are likely to be due to gastrointestinal damage. This LD50 factor may therefore be indicative of considerable variations in protection factors between various critical tissues. As the tumour protection factor for this combination is < 1.2 at cis-P doses in the LD50 region, then differential chemoprotection may occur and thus this combination may be worthy of further study.