Controlling malaria.

The antitumour activities of N-methylformamide, N-ethylformamide and formamide against a number of murine tumours in vivo (Sarcoma 180, M5076 ovarian sarcoma and TLX5 lymphoma) have been estimated. In all cases N-methyl-formamide had significant activity, formamide had marginal or no activity and N-ethylformamide had no significant activity. N-methylformamide and N-ethylformamide were equitoxic to the TLX5 lymphoma in vitro. Formamide was found as a metabolite in the plasma and urine of animals given N-methylformamide and N-ethylformamide, but excretion profiles do not support the hypothesis that formamide is an active antitumour species formed from N-alkylformamides. No appreciable metabolism of N-methylformamide occurred under a variety of conditions with liver preparations in vitro. N-methylformamide, but not N-ethylformamide or formamide, reduced liver soluble non-protein thiols by 59.8% 1 h after administration of an effective antitumour dose.

against murine tumours was first described in the early 1 950s (Clarke et al., 1953;Furst et al., 1955). In structure activity studies of almost 150 formamides and related compounds, NMF was found to be the most potent inhibitor of tumour growth and even small changes in molecular structure, such as substitution of the methyl group with an ethyl group, completely abolished activity. Formamide (HCONH2) and dimethylformamide (HCONMe2), however, showed marginal activity. Early studies on the mechanism of action of NMF suggested that it interferes with nucleic acid base synthesis (Clarke et al., 1 953 Skipper et al., 1955: Wheeler & Grammer, 1960Eidinoff et al., 1961;Sartorelli & Le Page, 1958;Morrison & Higgins, 1955), though it appeared to stimulate the incorporation of formate into nucleic acids in the liver (Barclay & Garfinkel, 1954), an organ to which it is toxic.
The hepatotoxicity of NMF reported in an early clinical trial in man (Myers et al., 1956) was considered to be an unacceptable side effect, and interest in its clinical use and pharmacology was largely lost. Recently, NMF has been found to be active against 3 human tumour xenografts in mice (Wolpert, personal communication) and there is renewe(d interest in its potential for clinical use in man. W!e were particularly interested in the observation that the substitution of an N-ethyl group in the formamide molecule led to a considerable reduction in activity, an observation also made regarding the antitumour activity of three other types of antitumour agent which we have been investigating, namely hexamethylmelamine, * To vlioom reprints requests shlould lie addtressed. aryldimethyltriazenes and procarbazine (Hickman, 1978). Our studies on the mechanism of action of these compounds, and in particular of their fate upon metabolism by the host, have centred upon the role and fate of the N-methyl group (Gescher et al., 1979;Gescher & Raymont, 1981). It was considered that a study of NMF, a simple and stable molecule compared to chemically labile triazenes and hydrazines, may contribute to an understanding of the mechanism of action of agents which require an N-methyl group for activity. Accordingly, we have studied the antitumour activity of NMF and certain analogues against 3 murine tumours, Sarcoma 180, on which the original data was reported (Clarke et al., 1953), the M5076 ovarian sarcoma (previously classified as a carcinoma but recently reclassified (Talmadge et al., 1981)), a tumour particularly sensitive to hexamethylmelamine (Simpson-Herren et al., 1979) and the TLX5 lymphoma, the tumour on which our other studies of the drugs containing N-methyl group have been made . Additionally, we report here on both in vivo and in vitro studies of the metabolism of Nalkylformamides, in particular, attempts to identify and quantitate metabolites resulting from their oxidative dealkylation. These studies are related to both the possible mechanism of cytotoxicity of NMF and to a possible biochemical explanation of the hepatotoxic lesions found in animals and man.

MATERIALS AND METHODS
The derivatives of formamide used in this study were all commercially available (Aldrich Chemicals, Gillingham) except Nethylformamide (NEF) which was synthesized in our laboratories by Dr R. J. Simmonds according to published methods (Saegusa et al., 1969). Drugs were administered i.p. or s.c. dissolved in sterile saline. Nicotinamide adenine diphosphonucleotide and glucose-6-phosphate were purchased from Sigma (U.K.) Ltd. Media and serum were purchased from Gibco (Glasgow) Ltd.
Aitditumour assays M5076 ovarian sarcoma. A suspension prepared by homogenization in saline of 106 cells from a routine passage, grown as a solid s.c. tumour in female BDF1 mice, was injected i.m. into the left lhind leg of groups of 10 female BDF1 mice (18-23 g). Drugs were administered i.p. for up to 17 days, (i.e. approximately half the life-span of the control tumour-bearing animals). Mean tumour volumes were measured by calipers every 4th day from Day 12 until death, and the mean tumour-volume index calculated by the standard method (Geran et al., 1972).
Sarcoma 180.-A suspension of 106 cells, from a routine passage of the tumour as an ascites in female BDF1 mice, was injected i.m. into the left hind leg of 18-23 g female BDF1 mice. Drugs were administered i.p. for up to 9 days. Mean tumour volumes Awere measured by calipers on Days 6, 10, 13 and 16 and the mean tumour-volume index calculated by the standard method (Geran et al., 1972).
TLX5 lymphomas.-Two TLX5 lymphomas were used, one with sensitivity to DTIC (5 -(3,3 -Dimethyl -1 -triazeno)imidazole -4carboxamide) and procarbazine (1 -methyl-2 -p -(isopropylcarbamoyl)benzylhydrazine) known as the TLX5S and one with resistance to these drugs, known as TLX5R (Hickman, 1978;. 105 TLX5 cells from routine passage as ascites in male CBA/CA mice were injected s.c. into the inguinal region of female CBA/CA mice (18-23 g). Drugs were administered i.p. up to Day 7 and the survival time of treated animals compared to untreated controls, a protocol used by others for this tumour (Connors & Hare, 1975). TLX5 lymphoma IN VITRO-IN VIVO assay. 106 cells/ml from a routine passage were incubated in 2 ml of 6 parts of RPMI 1640 medium and 4 parts of horse serum with drugs for 2 h. 105 cells were injected i.p. into 20 g female CBA mice and the survival time of animals receiving treated cells was compared to those receiving untreated cells.
IN VITRO metabolism. Livers of male CBA/CA or BALB/c mice (20-25 g) were homogenized in 0-25M sucrose to give a 200/ homogenate. The 9000g supernatant was prepared by differential centrifugation. Microsomes were obtained after addition of CaCl2 (Schenkman & Cinti, 1972) and resuspended in Earl's buffer. Different con-centrations of NMF, up to 5mM, were incubated with liver preparations equivalent to 0 4 g liver/ml incubation medium (Earl's buffer). The incubation mixtures wiere fortified with 3mM MgCl2 and cofactors which generated 1mM NADPH in a final volume of 2-5 ml. The incubations were carried out at 37°C, shaking for periods of up to 60 min with access of air or 02.
IN VIVO metabolism.-The plasma profiles for formamides were obtained after i.p. administration of either 400 mg/kg NMF or 495 mg/kg NEF to male CBA/CA mice (20-25 g). Blood was collected into heparinized syringes by cardiac puncture, using a mixture of NO, 02 and halothane as anaesthetic. Blood samples were centrifuged to remove red blood cells and analysed. Urine was collected from mice kept in mouse metabowl cages (Jencons, U.K.).

Analytical methods and pharmacokinetic analysis
In order to detect metabolically generated formaldehyde, the incubation mixtures were deproteinized with 0 5 ml of a 10% trichloracetic acid (TCA) solution, centrifuged and the supernatant used for the colorimetric assay according to Nash (1953).
For the gas-chromatographic analysis, samples of incubation mixtures, plasma or urine, were diluted with 5 parts (plasma and metabolic incubations) or 9 parts (urine) acetone containing dimethylacetamide as internal standard. After centrifugation the supernatant was chromatographed in a Pye Unicam 204 gas chromatograph fitted with a selective nitrogen/phosphorus detector and a glass column (2 m long, 4 mm i.d.) packed with 100-120 mesh Chromosorb W17 WDMCS (Phase Separation Ltd., Clwyd) coated with 8% Carbowax 2% KOH. The following temperatures were applied: injector 200°, column 190°, detector 250°C. Gas flow rates were 40 ml/min for N2, 30 ml/min for H2 and 300 ml/min for air. The limits of sensitivity of the assay were 8.5 nmol/ml for NMF and NEF and 45 nmol/ml for formamide (F). The recovery of these agents were 106+10% (NMF), 102+ 6% (NEF) and 109+12% (F) in 6 determinations. The area under the curve for plasma NMF concentrationi vs time was estimated by the trapezoid rule. The plasma concentration values on the declining part of the plasma-concentration vs time curve (Fig. 5) were subjected to a linear regression analysis which gave the apparent eliminationrate constant. This was used for the calculation of the apparent elimination half-life.

Non-protein-thiol assay
Male BALB/c mice (20-25 g) with livers weighing an average of 1 1 g were killed between 8 and 10 a.m. One hour after drug administration tissues were removed, blotted, weighed and immediately homogenized in 5% TCA solution (5 ml/g liver). The preparations were centrifuged at 2000 g for 10 min. Portions of the tissue supernatants (150 ,ul for liver and 300 ,ul for the other tissues) were diluted with 0-4M phosphate buffer (pH 8.0) and 0 3 ml of OO1M 5,5'dithiobis-(2-nitrobenzoic acid) (DTNB) (Ellman, 1955) to give a final solution of 4 ml.
Samples were analysed in duplicate in a Cecil CE 5095 spectrophotometer after 30 min at room temperature by reading the absorbance at 412 nm against a blank (sample without tissue homogenate). Glutathione (GSH) standards in a 5% TCA solution were assayed concurrently with samples. The recovery of GSH was >90% when added to the tissue samples.

Antitumour tests
The results for the in vivo antitumour tests of various formamides against the M5076 ovarian sarcoma, the S180 sarcoma and the TLX5 lymphoma are shown in Figs 1-4. The dosage schedule chosen for the treatment of the M5076 and S180 sarcomas represents a treatment period corresponding to approximately half the life span of the control animals. The data are presented to show tumour volumes both during and after treatment in the cases of the M5076 and S180 tumours.
The results for the in vitro-in vivo assay which compares the cytotoxicities of NMF, NEF and formamide are shown in the Table. There is no significant difference between the cytotoxicities of these alkylformamides measured under these conditions. In a comparison of the antitumour activity of 400 mg/kg NMF to both the  NEF (Table) suggested that their equitoxicity was nonspecific; since only NMF was active against the TLX5 lymphoma in vivo, NMF may be metabolized by the host to a selective species which presumably was not formed by metabolism of NEF. NMF was incubated with mouse liver preparations (whole homogenate, 9000g supernatant and microsomes) with and without 02, at varying substrate concentrations and for varying incubation periods. Oxidative N-demethylation, as measured by the production of formaldehyde, was not detected by any of these methods. Additionally, gas chromatographic analysis showed no significant disappearance of substrate or appearance of formamide (F), the product of Ndemethylation. IN VIVO metabolism and pharmacokinetics.-Determination of plasma concentrations of NMF after administration of the optimal antitumour dose (400 mg/kg) for the TLX5 lymphoma gave an area under the plasma-concentration-time curve of 62 ,umol.h/ml, and an apparent plasma-elimination half-life of 3-6 h, as determined by gas chromatography (Fig.   5). Formamide appeared in the plasma at a peak concentration of 95 nmol/ml after 6 h; at other times the concentration of formamide did not exceed the detection limit.
N-alkylformamides and formamide were detected in the urine of 6 animals given either NMF or NEF (Fig. 6). Within 24 h from the administration of NMF 7-3 + 2.90% of the dose was excreted as unchanged drug and 12 + 0.50/ as formamide. Within 24 h after an equimolar dose of NEF 4*4 + 1.4% appeared unchanged and 2-4 + 0 70/% as formamide. Effect of N-alkylformamides on soluble non-protein thiols in liver.-The optimal antitumour dose of NMF which had minimal toxicity but a good antitumour effect against the TLX5 lymphoma (400 mg/kg) caused a 59-8% depletion of hepatic non-protein thiol (NPT) in CBA mice measured 1 h after administration. This depletion was not brought about by NEF or formamide (Fig. 7). Of the other tissues investigated (heart, lungs, spleen and kidneys) only the kidneys exhibited depletion, though at 21.6% this was less marked than in the liver. It thus appears that NMF depletes the liver of the "labile" pool of glutathione as defined by Higashi et al. (1977) and which comprises 50-60% of total NPT with a half life of 1-7 h. The extent of hepatic NPT depletion was dose-dependent (Fig. 8) and not abolished by pretreatment with N-acetylcysteine. However, pretreatment of the animals with proadifen (SKF 525A, 60 mg/kg), an inhibitor of hepatic mixed-function oxidases (Cook et al., 1954), 1 h before administration of NMF reduced the depletion of hepatic NPT to 25.8% (Fig. 7). ]

DISCUSSION
The results of the antitumour tests on various formamides reported here (Figs 1-4) confirm previous findings regarding structure-activity relationships for this class of agent, namely, that only Nmethylformamide has significant activity, though in some systems, but not the TLX5 lymphoma (Fig. 4), formamide is also active. The activity of NMF against the ovarian sarcoma M5076 (Figs I & 2) is an encouraging result, as this tumour may be a good predictor of activity of new compounds for the treatment of human ovarian malignancies (Simpson-Herren et al., 1979), though this has been recently questioned (Slavik et al., 1981). The equivalent cytotoxicity of both NMF and NEF to TLX5S cells in vitro (Table) contrasts with a clear difference of potency in vivo, suggesting that this in vitro toxicity is nonspecific, and furthermore that NMF may possibly be bioactivated to a product in vivo that is either not formed from NEF or is formed in insufficient quantities.
The interesting phenomenon that the activity of certain antitumour compounds containing an N-methyl group is greatly reduced when substituted by an Nethyl group has been considered by us to 848 1 T I T I L----i L-----i be possibly due to differences in the metabolic fate of the different N-alkyl groupings (Hickman, 1978). Oxidative N-dealkylation is a major pathway of metabolism for N-alkyl-containing xenobiotics. In the case of NMF, such biotransformation should result in the production of formamide and formaldehyde via an intermediate carbinolamine, thus: HCONH . CH 3-HCONHCH20H-* HCONH2+ HCHO In the case of NEF, the products would be expected to be formamide and acetaldehyde. The activity of formamide in the in vivo antitumour tests reported here migh-t suggest that the antitumour activity of NMF is due to it acting as a progenitoio of formamide. However, the results with the TLX5kS lymphoma ( Fig.  4 and Table) do not support this hypothesis, as formamide has no activity, and it is difficult to explain the very greatly reduced potency of NEF, unless the latter is not metabolized as much as NMF.
Both NMF and NEF are metabolized in vivo in the CBA mouse to what was analytically detected as formamide in urine (Fig. 6). Although the rates of appearance of formamide differed, and the amount detected represented a very small proportion of the total dose, it is felt unlikely that the difference in the in vivo activity of NMF and NEF can be attributed to the quantitative aspects of metabolic formamide production. It is possible that there are metabolic pathways other than that suggested above for N-alkylformamides. However, preliminary experiments have shown that a stable precursor of formaldehyde is indeed a urinary metabolite of NMF, which supports the formation of the intermediate carbinolamine metabolite postulated. This will be the subject of a further report.
The disappearance of NMF from plasma and the appearance of its metabolite in urine indicate that host metabolism, presumably via oxidative metabolism, had occurred, but we were unable to show any significant metabolism of this type in vitro using various liver preparations. This contradicts the observations of Barnes & Ranta (1972), who claimed 100% transformation of I-7 mM NMF to formaldehyde after incubation with a rat liver homogenate for 2 h. The level of formaldehyde produced was, however, close to the detection limit of the colorimetric assay (Nash, 1953), and our results show that incubation of liver fractions for such long times leads to significant levels of control absorbances in the Nash assay, levels equivalent to those found after incubation with NMF.
It is interesting that the structural features required for anti-neoplastic activity of the formamides parallel those found to deplete hepatic non-protein thiols (NPT) (Fig. 7). The finding that proadifen partly reversed the depletion brought about by NMF strongly suggests that metabolic oxidation of NMF produces a reactive metabolite capable of reaction with soluble thiols. It has been suggested that the effect of paracetamol poisoning on the liver is due to an electrophilic metabolite which depletes hepatic glutathione stores and subsequently binds covalently to hepatic macromolecules which cause necrotic lesions (Mitchell et al., 1972). By analogy, the hepatic toxicity of NMF reported in the early clinical trial (Myers et al., 1956) may be related to the NPT depletion reported here, and a reactive metabolite may also be responsible for the antitumour effects of NMF. Both of these hypotheses are under active investigation by us.