Timothy P. Wunz

A choline transporter in renal brush-border membrane vesicles: Energetics and structural specificity

SummaryCholine is a quaternary ammonium compound that is normally reabsorbed by the renal proximal tubule, despite its acknowledged role as a substrate for the renal organic cation (OC) secretory pathway. The basis for choline reabsorption was examined in studies of transport in rabbit renal brush-border membrane vesicles (BBMV). Although an outwardly directed H+ gradient (pH 6.0in ∶ 7.5out) stimulated uptake of tetraethylammonium (TEA), a model substrate of the OC/H+ exchanger in renal BBMV, it had no effect on uptake of 1 μm choline. A 5 mmtrans concentration gradient of choline did, however, drive countertransport of both TEA and choline, although trans TEA had no effect on choline accumulation in BBMV. A 20 mm concentration of unlabeled choline blocked uptake of both choline and TEA by >85%, whereas 20 mm TEA blocked only TEA uptake. The kinetics of choline uptake into vesicles preloaded with 1 mm unlabeled choline appeared to involve two, saturable transport processes, one of high affinity for choline (Kt of 97 μm) and a second of low affinity (Ktof ∼10 mm), the latter presumably reflecting a weak interaction of choline with the OC/H+ exchanger. An inside-negative electrical PD stimulated the rate of uptake and supported the transient concentrative accumulation of choline in BBMV. The high affinity transporter showed a marked specificity for choline and closely related analogues. A model of the molecular determinants of substrate-transporter interaction is described. We conclude that the electrogenic high affinity pathway plays a central role in renal reabsorption of choline.

Structure and interaction of inhibitors with the TEA/H+ exchanger of rabbit renal brush border membranes

The renal secretion of organic cations (OCs) involves a carrier-mediated exchange of OC for H+ in the luminal membrane of proximal cells. To assess the influence of chemical structure on the interaction of potential substrates with this process we examined the effect of a series of quaternary ammonium compounds on the transport of the OC tetraethylammonium (TEA) in a preparation of isolated renal brush-border membrane vesicles. Apparent inhibitory potency varied over a factor of 104, as expressed in inhibitor coefficients (KiTEA) whose approximate values ranged from 0.5 μM to 5 mM. The poorest inhibitors of TEA/H+ exchange were those molecules with carboxyl or hydroxyl residues, whereas the addition of methylene groups to a parent molecule tended to increase inhibitory potency. A plot of apparent KiTEA versus calculated octanol:water partition coefficient (expressed in terms of a relative lipophilicity factor) showed a clear correlation between these two parameters, although there was considerable variability between apparent lipophilicity and KiTEA for molecules with very different parent structures. For select groups of molecules with similar parent structures (e.g., the n-tetraalkylammoniums or the 4-phenylpyridinium, 3-phenylpyridinium, and quinolinium compounds) the correlation between calculated lipophilicity and apparent KiTEA was more marked. However, even within these groups of closely related parent structures, there appeared to be subtle, but systematic, variations in inhibitory potency that may have been related to the influence of steric factors on the binding of inhibitors to the TEA/H+ exchanger. We conclude that the lipophilic nature of a quaternary ammonium compound represents the predominant factor in the binding to, and subsequent inhibition of, luminal TEA/H+ exchange. Specific steric factors may influence the binding of substrate to the exchanger, but play a secondary role in this interaction.

Computer simulation of the binding of quinocarcin to DNA. Prediction of mode of action and absolute configuration

SummaryComputer-based models were derived for the covalent and noncovalent binding of the antitumor antibiotic quinocarcin to a representative DNA segment, d(ATGCAT)2. They showed that a mode of action, involving opening of the oxazolidine ring to give an iminium ion, followed by initial noncovalent binding in the minor groove and subsequent alkylation of the 2-amino group of guanine, was rational and attended by favorable interaction energies in each step. The best model had the aryl ring of quinocarcin lying in the 3′ direction from the covalent binding site and anR configuration at the carbon involved in covalent bond formation. It also showed that the preferred absolute configuration for quinocarcin was the reverse of that arbitrarily assigned in the literature.

In vitro cytotoxicity against fresh human tumors and P388 leukemia predicts the differential in vivo activity of a series of anthracene anticancer drugs.

To date, random anticancer drug screening has proven to be relatively inefficient and non-specific with respect to selecting active compounds for most tumor types (except for leukemia/lymphoma). Although large numbers of compounds from diverse sources were evaluated for many years in the P388 mouse leukemia model, only a few clinically useful drugs have been identified by this in vivo screening method. Thus, there is intense interest in the development of more effective in vitro screening models for new anticancer drugs. In the present paper we have compared the discriminating power for fresh human tumors from patients, human tumor cell lines developed from 11 patients and murine P388 leukemia in tumor colony forming assays as indicators of cytotoxicity for a series of anthracene antitumor agents. Two of a series of 21 novel bisantrene analogs, R6 (N, N1-bis[2-(dimethylamino)ethyl]-9,10-anthracenebis(methylamine)) and R26 (N, N1-bis(1-ethyl-3-piperidinyl)-9,10-anthracene-bis(methylamine)) produced significant cytotoxicity against the 11 human tumor cell lines and were therefore selected for additional in vitro and in vivo studies. R26 was specifically selected for further testing since it had Ion or doxorubicin-resistant 8226 myeiomd ten lines, in contrast to the cell line data, only one of the 22 fresh human tumors showed significant in vitro sensitivity (i.e. <50% survival of tumor colony forming units) to either R6 or R26 tested at high concentrations. Both of these bisantrene analogs also proved inactive at 1.2–1.6 μM concentrations against P388 leukemia in vitro, whereas mitoxantrone and bisantrene were highly active in this model at a concentration of 0.2 μM. In order to compare the in vitro data with antitumor activity in vivo, R26, the most active bisantrene analog, and mitoxantrone, the most active of the two anthracene parent compounds, were tested against P388 leukemia and M5076 ovarian sarcoma in mice. In both models mitoxantrone showed significant activity whereas R26 produced minimal or no antitumor effects. We conclude that fresh human tumors, but not defined human tumor cell lines, predict the in vivo cytotoxicity of a series of anthracene anticancer agents. Although this conclusion may not apply to the screening of other classes of antitumor agents, we propose an in vitro screening process which first utilizes numerous human tumor cell lines of many different biologies (to screen a large number of new compounds each year), followed by confirmatory tests in fresh human tumors using colony forming assays to screen up to a smaller number of 1000 compounds. Finally appropriate in vivo tumor model based on histologic pecificity would be used to screen a few consistently active new compounds for advancement to clinical trials. Thus, the first screening stage would be highly sensitive and non-specific, the second in vitro stage more specific and the third in vivo stage relevant by histologic tumor type.