Search

This revised Clinical Guideline in Paediatric Dentistry replaces the previously published sixth guideline (Fayle SA. Int J Paediatr Dent 1999; 9: 311-314). The process of guideline production began in 1994, resulting in first publication in 1997. Each guideline has been circulated widely for consultation to all UK consultants in paediatric dentistry, council members of the British Society of Paediatric Dentistry (BSPD), and to people of related specialities recognized to have expertise in the subject. The final version of this guideline is produced from a combination of this input and thorough review of the published literature. The intention is to encourage improvement in clinical practice and to stimulate research and clinical audit in areas where scientific evidence is inadequate. Evidence underlying recommendations is scored according to the SIGN classification and guidelines should be read in this context. Further details regarding the process of paediatric dentistry guideline production in the UK is described in the Int J Paediatr Dent 1997; 7: 267-268.

Nociceptive dorsal root ganglion (DRG) cells can be divided into three main populations, namely (1) small diameter non-peptide-expressing cells, (2) small-diameter peptide-expressing (calcitonin gene related peptide (CGRP), substance P) cells, and (3) medium-diameter peptide-expressing (CGRP) cells. The properties of these cell populations will be reviewed, with a special emphasis on the expression of the vanilloid (capsaicin) receptor VR1 and its regulation by growth factors. Cells in populations 1 and 2 express VR1, a nonselective channel that transduces certain nociceptive stimuli and that is crucial to the functioning of polymodal nociceptors. Cells in population 1 can be regulated by glial cell line derived neurotrophic factor (GDNF) and those in populations 2 and 3 by nerve growth factor (NGF). In vivo, DRG cells express a range of levels of VR1 expression and VR1 is downregulated after axotomy. However, treatment with NGF or GDNF can prevent this downregulation. In vitro, DRG cells also show a range of VR1 expression levels that is NGF and (or) GDNF dependent. Functional studies indicate that freshly dissociated cells also show differences in sensitivity to capsaicin. The significance of this is not known but may indicate a difference in the physiological role of cells in populations 1 and 2.

In this study, we describe the activity of CT1746, an orally-active synthetic MMP inhibitor that has a greater specificity for gelatinase A, gelatinase B and stromelysin than for interstitial collagenase and matrilysin, in a nude mouse model that better mimics the clinical development of human colon cancer. The model is constructed by surgical orthotopic implantation (SOI) of histologically-intact tissue of the metastatic human colon tumor cell line Co-3. Animals were gavaged with CT1746 twice a day at 100 mg/kg for 5 days after the SOI of Co-3 for 43 days. In this model CT1746 significantly prolonged the median survival time of the tumor-bearing animals from 51 to 78 days. Significant efficacy of CT1746 was observed on primary tumor growth (32% reduction in mean tumor area at day 36), total spread and metastasis (6/20 treated animals had no detectable spread and metastasis at autopsy compared to 100% incidence of secondaries in control groups). Efficacy of CT1746 could also be seen on reducing tumor spread and metastasis to individual organ sites such as the abdominal wall, cecum and lymph nodes compared to vehicle and untreated controls. We conclude that chronic administration of a peptidomimetic MMP inhibitor via the oral route is feasible and results in inhibition of solid tumor growth, spread and metastasis with increase in survival in this model of human cancer, thus converting aggressive cancer to a more controlled indolent disease.

Four potent, synthetic inhibitors of matrix metalloproteinases (MMPs) were assessed as inhibitors of tumor growth and spontaneous metastasis to the lung. Mat Ly Lu rat prostate tumor, LOX human melanoma and M27 murine Lewis lung tumor were implanted subcutaneously (s.c.) in mice and allowed to grow for 3-12 days. The lungs of the tumor-bearing mice were then removed and implanted s.c. into untreated mice, and the outgrowth of secondary tumors from the implanted lungs measured. The incidence and rate of outgrowth of secondary tumors increased with the length of primary tumor growth, validating these measurements as indices of spontaneous metastasis to the lung. Compounds were tested by s.c. implantation of minipumps which delivered compound throughout the period of primary tumor growth and spontaneous metastasis to the lung at steady-state drug concentrations orders of magnitude greater than the concentrations needed to either inhibit collagenase, gelatinase or stromelysin in vitro. Inhibitor treatment slowed the growth of primary s.c. Mat Ly Lu and LOX tumors by 40-60% but had no significant effect on the growth of primary M27 tumors. Surprisingly, inhibitor treatment had no significant effect on the ability of the lung to generate secondary tumors when reimplanted s.c. in untreated mice. Because of the possible importance of cathepsins B, H and L in tumor growth and metastasis, the irreversible inhibitor E-64 was also infused by s.c. minipump. E-64 had no effect on the growth or spontaneous metastasis of Mat Ly Lu or M27 tumors.

Cyclic adenosine diphosphate ribose (cADPR) is a potent endogenous calcium-mobilizing agent synthesized from beta-NAD+ by ADP-ribosyl cyclases in sea urchin eggs and in several mammalian cells (Galione, A., and White, A. (1994) Trends Cell Biol. 4, 431 436). Pharmacological studies suggest that cADPR is an endogenous modulator of Ca2+-induced Ca2+ release mediated by ryanodine-sensitive Ca2+ release channels. An unresolved question is whether cADPR can act as a Ca2+-mobilizing intracellular messenger. We show that exogenous application of nitric oxide (NO) mobilizes Ca2+ from intracellular stores in intact sea urchin eggs and that it releases Ca2+ and elevates cADPR levels in egg homogenates. 8-Amino-cADPR, a selective competitive antagonist of cADPR-mediated Ca2+ release, and nicotinamide, an inhibitor of ADP-ribosyl cyclase, inhibit the Ca2+-mobilizing actions of NO, while, heparin, a competitive antagonist of the inositol 1,4,5-trisphosphate receptor, did not affect NO-induced Ca2+ release. Since the Ca2+-mobilizing effects of NO can be mimicked by cGMP, are inhibited by the cGMP-dependent-protein kinase inhibitor, Rp-8-pCPT-cGMPS, and in egg homogenates show a requirement for the guanylyl cyclase substrate, GTP, we suggest a novel action of NO in mobilizing intracellular calcium from microsomal stores via a signaling pathway involving cGMP and cADPR. These results suggest that cADPR has the capacity to act as a Ca2+-mobilizing intracellular messenger.

Regional chemotherapy is an attractive but underevaluated method of treating locally advanced breast cancer. We have combined two novel methods of targeting by delivering a single pulse of adriamycin-loaded albumin microspheres down a radiologically placed internal mammary artery catheter. A complete response was observed and prolonged local control achieved.

Many hormones or neurotransmitters act at cell surface receptors to increase the intracellular free calcium concentration, triggering a wide range of cellular responses. As the source of this Ca2+ is often internal stores, additional messengers are required to convey the hormonal message from the plasma membrane. Cyclic ADP-ribose (cADPR) has been proposed as the endogenous activator of Ca(2+)-induced Ca2+ release by the ryanodine receptor in sea urchin eggs and in several mammalian cell types. A second messenger role for cADPR requires that its intracellular levels be under the control of extracellular stimuli. Here we demonstrate a novel action of 3',5'-cyclic guanosine monophosphate (cGMP) in stimulating the synthesis of cADPR from beta-NAD+ by activating its synthetic enzyme ADP-ribosyl cyclase in sea urchin eggs and egg homogenates. We suggest that cADPR may transduce signals generated by cell surface receptors or gaseous transmitters linked to cGMP production.

Propagating Ca2+ waves are a characteristic feature of Ca(2+)-linked signal transduction pathways. Intracellular Ca2+ waves are formed by regenerative stimulation of Ca2+ release from intracellular stores by Ca2+ itself. Mechanisms that rely on either inositol trisphosphate or ryanodine receptor channels have been proposed to account for Ca2+ waves in various cell types. Both channel types contributed to the Ca2+ wave during fertilization of sea urchin eggs. Alternative mechanisms of Ca2+ release imply redundancy but may also allow for modulation and diversity in the generation of Ca2+ waves.

Doxorubicin (DOX) has been incorporated into five different formulations of protein microspheres, each altering tumour drug disposition in a characteristic manner. There was no correlation between the stimulation in anaerobic quinone bioreduction over the levels produced by free DOX and tumour growth delay against the Sp 107 rat mammary carcinoma. A strong correlation (r2 = 0.948, P < 0.01, two-tailed t statistic) was observed between a slower decline in parent drug levels and antitumour activity. These data support the view that free radical processes are not involved in the mechanism of action of DOX and suggest that the optimum way to delivery the drug to the Sp 107 tumour is through sustained release of lower concentrations (approximately 2 microM) from a large extracellular pool.

A high-performance liquid chromatographic method is described for the determination of mitomycin C (MMC) and its metabolites 2,7-diaminomitosene (2,7-DM), 1,2-cis-1-hydroxy-2,7-diaminomitosene (cis-hydro) and 1,2-trans-1-hydroxy-2,7-diaminomitosene (trans-hydro) in tumour tissue. N-la-Methylmitomycin C (porfiromycin, PM) was used as an internal standard. Two factors were critical in resolving the metabolites: pH and buffer ionic strength, where the retention times of the four components were affected in the order 2,7-DM >> cis-hydro >> trans-hydro >> MMC. The optimal isocratic conditions (flow-rate 1 ml/min) were 18 mM sodium phosphate pH 5.8-methanol (74:26) and a column temperature of 40 degrees C on a Spherisorb ODS-2 column (25 cm x 4.6 mm I.D.). Liquid-liquid extraction [twice with chloroform-propan-2-ol-ethyl acetate (2:2:1)] is described for tumour tissue. Recoveries varied depending on the component: MMC, 71.9 +/- 12.4%; PM, 85.5 +/- 27%; 2,7-DM, 51.7 +/- 5.4%; cis-hydro, 52.0 +/- 16.8%; trans-hydro, 62 +/- 8%. When applied to the analysis of a rat mammary carcinoma treated intra-tumourally with 450 micrograms of MMC five drug-related "metabolite" peaks were detected. Three of these co-chromatographed with standards of 2,7-DM, cis- and trans-hydro, and had identical absorption maxima to their respective standards, with the possible exception of trans-hydro.

Preparation of mitomycin C-loaded human serum albumin (HSA) microspheres using a new technique that avoids the use of heat denaturation, which is known chemically to degrade incorporated drug, is described. This method is based on cross-linking of protein by glutaraldehyde (2.2%) during emulsification (W/O) at room temperature. The resultant particles have a mean (s.d.) diameter of 16.9 (0.34) microns (50% weight average), contain mean (s.d.) 1.15 (0.05%) mitomycin C (MMC) (w/w, n = 17) and maintain sustained release of drug over 20 h. High performance liquid chromatography (HPLC) with diode array detection was used to study the chemical integrity of the drug. Two classes of decomposition products were evaluated: chemical degradation products and drug/nucleophile covalent adducts. The HPLC separation was validated by a number of standards of proposed degradation products. To examine incorporated drug, a complete microsphere system was solubilized with 0.4% trypsin for 24 h, while to examine released drug, microspheres were immobilized on a flow-through glass wool column and fractions were collected. No evidence of significant chemical degradation or covalent coupling to protein was detected in microsphere digests. Two candidate decomposition products, representing approximately 10% of drug released from microspheres (assuming similar molar extinction coefficients to MMC), were identified in column fractions. One of these products appeared to be a covalent adduct, the other possibly an isomeric form of intact MMC. Thus, MMC is predominantly incorporated into and released (90%) chemically intact from HSA microspheres prepared by the technique described.

A clear role for quinone reduction in the mechanism of action of doxorubicin has still to be established. There are three possible outcomes of this form of doxorubicin metabolism: (1) drug free radical formation, redox cycling and generation of reactive oxygen species (ROS) resulting in lipid peroxidation and DNA damage; (2) covalent binding of reactive drug intermediates to DNA; and (3) formation of an inactive 7-deoxyaglycone metabolite. In this work, the occurrence of each of these pathways has been studied in vivo in a subcutaneously growing rat mammary carcinoma (Sp 107). Doxorubicin was administered by direct intratumoural injection either as the free drug or incorporated in albumin microspheres (10-40 microns diameter). There was no evidence of an increase in lipid peroxidation over background after either treatment at any time point studied. In fact, doxorubicin administration resulted in a statistically significant reduction in lipid peroxidation at the later time points studied compared to control (no drug treatment), e.g. 24 hr: control, 21.7 +/- 2.8 SD nmol malondialdehyde/g tissue; free doxorubicin (70 micrograms drug), 14.5 +/- 4.0 SD nmol/g (P < 0.01 Student's t-test) and doxorubicin microspheres (70 micrograms drug), 17.4 +/- 1.1 nmol/g (P < 0.05). Covalent binding to DNA was measured by a 32P-post-labelling technique. Low levels of four putative drug-DNA adducts were detected; however, there were no qualitative or quantitative differences in profiles between free drug and microspheres. High 7-deoxyaglycone metabolite concentrations comparable to the parent drug itself were detected after administration of microspheres (3.0 micrograms/g +/- 1.7 SD at 24 hr and 3.1 micrograms/g +/- 1.1 SD at 48 hr). In contrast, these metabolites were present at levels close to the limit of detection of our HPLC assay after free drug (0.04 microgram/g +/- 0.03 SD at 24 hr and 0.02 microgram/g +/- 0.03 SD at 48 hr). Thus, 7-deoxyaglycone metabolite formation can occur in tumour tissue (indicating active drug quinone reduction) without concomitant increases in the level of lipid peroxidation or the levels of drug-DNA adducts. In conclusion, the main biological consequence of doxorubicin quinone reduction in vivo in tumour tissue would appear to be drug inactivation to a 7-deoxyaglycone metabolite rather than drug activation to DNA reactive species or ROS.

We have reported previously that enzymes present in the Sp 107 rat mammary carcinoma catalyse doxorubicin quinone reduction (QR) to 7-deoxyaglycone metabolites in vivo [Willmott and Cummings, Biochem Pharmacol 36: 521-526, 1987]. In order to provide insights into the role of QR in the antitumour mechanism of action of doxorubicin, we have attempted in this work to identify the enzyme(s) responsible. NAD(P)H: (quinone acceptor) oxidoreductase (DT-diaphorase) was the major quinone reductase in the tumour accounting for approximately 70% of all the activity measured in microsomes and cytosols (microsomal activity, 28.4 +/- 4.6 nmol/min/mg; cytosolic activity, 94.3 +/- 11.9 nmol/min/mg). Its presence was confirmed by western blot analysis. Low levels of NADH cytochrome b5 reductase (15.6 +/- 6.3 nmol/min/mg) and NADPH cytochrome P450 reductase (14.5 +/- 4.0 nmol/min/mg) were detectable in microsomes. The presence of the latter was confirmed by western blot analysis. Pretreatment of tumours with doxorubicin (48 hr) at a therapeutic dose decreased the level of activity of all the reductases studied by at least 2-fold (P < 0.01, Student's t-test). Doxorubicin was shown not to be a substrate for purified rat Walker 256 tumour DT-diaphorase with either NADH or NADPH as co-factor and utilizing up to 20,000 units of enzyme/incubation but was confirmed to be a substrate for purified rat liver cytochrome P450 reductase. 7-Deoxyaglycone metabolite formation by purified cytochrome P450 reductase had an absolute requirement for NADPH as co-factor, was inhibited by molecular oxygen and dicoumarol (IC50 approx. 50 microM), and modulated by specific reductase antiserum. Reductive deglycoslation of doxorubicin to 7-deoxyaglycones was localized to the microsomal fraction of the Sp 107 tumour, with negligible activity being found in cytosols (NADH, NADPH and hypoxanthine as co-factors) and mitochondria (NADH and NADPH). The tumour microsomal enzyme had an absolute co-factor requirement for NADPH, was inhibited by oxygen and dicoumarol, and modulated by cytochrome P450 reductase antiserum. These data indicate strongly that NADPH cytochrome P450 reductase is the principal enzyme responsible for catalysing doxorubicin QR in the Sp 107 tumour.

The relationship between hepatic arterial albumin microsphere distribution and hepatic arterial blood flow and the effects of regional angiotensin II were studied in a rat liver metastases model. Hooded-Lister rats were inoculated subcapsularly with 2 x 10(6) HSN sarcoma cells. At 20 days, hepatic arterial blood flow was measured using the reference microsphere technique. Animals then randomly received 50 microliters hepatic arterial saline or albumin microspheres (40 microns, 20 mg ml-1). Hepatic arterial blood flow measurements were then repeated at 5 min. After 5 min, animals were killed and tissues were weighed and counted in a gamma well counter. There were no significant differences between the hepatic blood flow measurements recorded before and after the control hepatic arterial saline infusion. However, regional albumin microspheres produced a significant reduction in tumour and normal liver blood flow and an 80% reduction in mean T/N blood flow ratio. Regional albumin microspheres were delivered to tumour in greater proportions (mean T/N ratio 3.89, SE 0.49) than would be expected from baseline hepatic arterial blood flow (mean T/N ratio 1.28, SE 0.22. P = 0.006). There was no correlation between T/N for baseline blood flow and albumin microsphere distribution.

We have studied incorporation of [14C]doxorubicin within protease-sensitive casein microspheres both by 14C-activity, measuring total drug, and HPLC, measuring free drug only. It was found that total drug content (27.7 micrograms mg-1) exceeded free drug content (3.2 micrograms mg-1) suggesting that the major portion of doxorubicin was incorporated via a covalent linkage to matrix protein. In-vivo drug disposition and activity studies suggested that this fraction of doxorubicin was the major species within tumour tissue (total vs free: 5 min, 14.3 micrograms g-1 vs 0.7 micrograms g-1; 24 h, 11.7 micrograms g-1 vs 1.1 micrograms g-1; 48 h, 11.2 micrograms g-1 vs 1.2 micrograms g-1; 72 h, 10.0 micrograms g-1 vs 0.8 micrograms g-1), did not exhibit a 'burst' effect, was slowly cleared (30% loss over 3 days), and was equiactive (growth delay = 12 days) compared with drug in solution (growth delay = 10 days). This work clearly implicates in-vivo microsphere matrix biodegradation in drug release and subsequent disposition and activity.

There is increasing interest in the use of microspheres, loaded with chemotherapeutic agents, for regional therapy to hepatic metastases. It is necessary to deliver these particles predominately to tumour rather than to normal liver. This study investigates factors influencing the distribution of regionally injected microspheres. Discreet tumour was induced in rats by subcapsular hepatic inoculations of HSN cells. At 20 days, 12.5 microns, 25 microns or 40 microns diameter, radiolabelled albumin microspheres were administered, in various concentrations, via the gastroduodenal artery. Tumour to normal liver microsphere distribution ratios were determined and median values ranged from 0.1 (0.2 mg ml-1 12.5 microns microspheres) to 1.8 (20 mg ml 40 microns microspheres). Concentrated suspensions (20 mg ml-1) of large microspheres (40 microns) produced the most favourable tumour to normal liver distribution ratios. These results not only have implications for the therapeutic administration of microspheres but also for their use in blood-flow studies.

Regionally-administered, drug-loaded microspheres have a potential role in the treatment of renal tumours. Vasoactive agents, for example, angiotensin II, may allow selective delivery of microspheres to tumour. The present study defines the regional advantage that may be obtained from angiotensin II by quantifying tumour and normal kidney blood flow using radiolabelled microsphere renal perfusion studies and per-operative laser-doppler flow measurements. Angiotensin II increased microsphere distribution to tumour, relative to normal kidney, by a factor of four. This enhancement was associated with an absolute increase in tumour blood flow.

Doxorubicin is shown to be present in albumin microspheres (10-40 microns) in two forms: the native drug and a fraction of drug covalently coupled to the protein matrix probably via glutaraldehyde. Upon trypsin digestion the fraction covalently coupled is released and can be resolved from native doxorubicin by high performance liquid chromatography and quantitated either by using 14C-labelled doxorubicin or by measuring the absorption of the doxorubicin chromophore at 480 nm. Albumin microspheres contained 6.9 micrograms/mg protein covalently bound drug versus 11.1 micrograms/mg native drug when 1% glutaraldehyde was used in microsphere preparation. The covalently bound fraction increased significantly with 2% glutaraldehyde. Albumin/polyaspartic acid microspheres lacked a covalently bound fraction when prepared under the same conditions as pure albumin microspheres (35 micrograms/mg native drug, 1% glutaraldehyde) but transferrin microspheres contained similar amounts of bound and native albumin. In vivo, albumin microspheres altered the disposition of doxorubicin in a rat mammary carcinoma (Sp107) compared to albumin/polyaspartic acid microspheres by reducing the rate of parent drug elimination from the tumour and by reducing its biotransformation to 7-deoxyaglycone metabolites. These data indicate that covalent coupling is a key component in the way doxorubicin is handled in tumours after administration of protein microspheres.

The outlook for patients with colorectal liver metastases is poor. Microspheres have been combined with cytotoxics and administered via the hepatic artery in an attempt to improve tumour drug exposure within the liver. However, it has been suggested that arteriovenous connections might occur in association with intrahepatic tumours causing loss of regional advantage, and that the administration of microspheres further exacerbates arteriovenous shunting. In seven patients with colorectal liver metastases, base-line shunting was measured using a tracer quantity of radio-labelled albumin microspheres. The shunted fraction of a 'therapeutic quantity' of microspheres was subsequently measured in the same group of patients using albumin microspheres carrying a different radio-label. Base-line shunt for 0.5 x 10(6) microspheres was found to be 2.2 +/- 1.8% (mean +/- s.d.); the percentage shunt of a therapeutic quantity (40-80 x 10(6)) of microspheres was 3.0 +/- 0.8%. We conclude that arteriovenous shunting in patients with colorectal liver metastases is minimal, and is not significantly increased by the administration of therapeutic quantity of microspheres during regional chemotherapy.

Cytotoxic microspheres have been developed for intra-arterial use in patients with liver metastases. Following injection, the distribution of microspheres reflects the pattern of hepatic arterial blood-flow. Vasoactive agents, such as angiotensin II, by producing vasoconstriction in normal liver, might divert arterial blood toward tumour and thereby enhance the delivery of drug-loaded particles. Using a double isotope technique, the distribution of radiolabelled microspheres to tumour and normal liver tissue was measured before and after angiotensin II infusion in nine patients with multiple liver metastases. The median increase in tumour: normal ratio following angiotensin II infusion was by a factor of 2.8 (range 0.8-11.7, P less than 0.05). This novel approach to regional chemotherapy, using a combination of angiotensin II infusion and cytotoxic microspheres, increases the exposure of tumour to cytotoxic agents and may, therefore, enhance tumour response rates.

Biodegradable albumin microspheres have been prepared incorporating either 131I or 111In/111In. Using cDTPAA/albumin molar ratios of 1, 3 and 10, approx. 0.7, 2.1 and 7 molecules of DTPA could be coupled to an albumin molecule, labelling efficiency being constant over this range. Because 131I microspheres were more stable in plasma than the system labelled with radionuclides of indium it was selected for clinical evaluation: potential uses of this radiopharmaceutical are illustrated.

5-Fluorouracil (5-FU) was extracted from plasma and converted to its di-N-ditrifluoromethylbenzyl (DTFMBz) derivative by treatment with DTFMBzBr. Under negative ion chemical ionization (NICI) conditions the derivative yielded a mass spectrum in which most of the ion current was carried by the M(-)-DTFMBz ion (m/z 355). Selected ion monitoring enabled detection of this derivative in amounts less than 1 pg, and the practical limit of detection for 5-FU extracted from plasma was ca 400 pg ml-1. 5-FU was quantified by adding a fixed amount of [15N2]5-FU to samples of plasma before extraction, and comparing the ratio of the ions of m/z 355 and 357 against a calibration curve constructed over the concentration range under investigation. The method was used to measure the variation in the concentration of 5-FU with time during continuous infusion of the drug via three, different protocols in patients with hepatic tumour.

The pharmacokinetics of 5-fluorouracil (5FU) following its administration via the hepatic artery in conjunction with biodegradable albumin microspheres and angiotensin II have been studied. Peripheral venous concentrations of 5FU are lower and plasma clearance values higher following intrahepatic arterial administration compared with a similar dose administered by intravenous infusion over both 2 h and 24 h. For the 2 h drug infusions, plasma 5FU concentrations following co-treatment with angiotensin II and microspheres via the hepatic artery were intermediate between those of arterial and venous infusions of 5FU alone. There was a trend towards the peak plasma drug concentrations and the area under the plasma concentration-time curve (AUC) being significantly lower following co-treatment with angiotensin II and microspheres compared with intra-arterial and intravenous infusions of 5FU over 24 h. Co-administration of 5FU, angiotensin II and microspheres via the hepatic artery may reduce drug exposure in the systemic compartment and therefore may increase the therapeutic ratio of 5FU administration via the hepatic artery.

Increased activity against a rat solid tumour of doxorubicin incorporated into protein microspheres and administered intratumourally was associated with both increased duration of exposure of tumour tissue to native drug and anaerobic bioreduction of doxorubicin to 7-deoxyaglycones, indicating formation of reactive drug intermediates within tumour tissue. To investigate which of these aspects of drug disposition determined activity we have compared the in vivo fate (clearance from and metabolism by tumour tissue) of doxorubicin in microspherical form with the analogue 4'-deoxydoxorubicin and related this to the tumour growth delay recorded for these drugs. Within the dose range 42 to 55 micrograms, growth delay (14-18 days) of doxorubicin in microspherical form was markedly superior to drug in solution, whereas growth delay of 4'-deoxydoxorubicin in microspherical form (4.3-7.2 days) was not greater than drug in solution. Metabolism to 7-deoxyaglycones by tumour tissue was not a prominent feature of either drug when administered in solution. However, in microspherical form both drugs were extensively metabolized (peak concentrations: 3.6 micrograms/g doxorubicin 7-deoxyaglycone; 2.5 micrograms/g 4'-deoxydoxorubicin 7-deoxyaglycone). Native drug concentrations in tumour tissue were similar after administration in microspherical form at 48 hr (doxorubicin 3.8 micrograms/g; 4'-deoxydoxorubicin 3.7 micrograms/g) and 72 hr (doxorubicin 2.4 micrograms/g; 4'-deoxydoxorubicin 2.7 micrograms/g). At both time points, following administration in microspherical form, tumour tissue concentrations of doxorubicin were significantly greater than when drug was administered in solution, whereas no significant differences were observed for 4'-deoxydoxorubicin. The results are inconsistent with the process of anaerobic bioreduction of doxorubicin to 7-deoxyaglycones being an important component of its anti-tumour activity in microspherical form and point to the importance of increased duration of exposure to native drug.