122
BIOCHEMICAL SOCIETY TRANSACTIONS
Ellory, J. C., Jones, S. E. M. & Young, J. D. (19816) J. Physiol.
(London) 320,403422
Vidaver, G. A. & Shepherd, S. L. (1968) J . Biol. Chem. 243,
6 140-6 150
Rennie, C. M., Thompson, S., Parker, A. R. & Maddy, A. (1979) Clin.
Chim. Acta 98,119-125
Young, J. D., Jones, S. E. M.& Ellory, J. C. (1981) Biochim. Blophys.
Acfa 645, 157-160
...
. .
Selective direction of ricin to hepatic parenchymal cells
DAVID N. SKILLETER,* ALAN J. PAINE* and
PHIL E. T H O R P E t
*Toxicology Unit, MRC Laboratories, Woodmansterne Road,
Carshalton, Surrey SM5 4EF, U.K., and tChester Beatty
Research Institute, Fulham Road, London S WJ 658, U.K.
Ricin, a glycoproprotein from castor beans (Ricinus communis),
is a potent inhibitor of protein synthesis in eukaryote cells
(Olsnes et al., 1974) possibly at concentrations as low as one
molecule per cell (Eiklid et al., 1980). In view of the potency of
rich there is much interest in its potential use as a site-specific
toxin, particularly in cancer and immunotherapy. Thus rich has
been conjugated with molecules that 'target' it to specific cell
types (Olsnes, 198 1).
In the treatment of hepatocellular carcinoma it might be
desirable to administer an agent specifically to hepatic
parenchymal cells. Although a high proportion of rich
administered to experimental animals is removed from the blood
by the liver, we have recently shown that the toxin is taken up by
the hepatic sinusoidal (Kupffer) rather than by parenchymal
cells (Skilleter et a!., 1981). In an attempt to direct ricin to
parenchymal cells we have conjugated rich with the glycoprotein asialofetuin, which is known to be taken up by the liver
exclusively into parenchymal cells (Hubbard et al., 1979) and
report here the accumulation of this conjugate and its effect on
protein synthesis in cultures of hepatic parenchymal and
Kupffer cells.
Ricin, isolated by the method of Nicolson et al. (1972), was
coupled to monomeric asialofetuin with chlorambucil. Briefly
the procedure used was to couple asialofetuin at 4OC with the
N-hydroxysuccinamide ester of chlorambucil, followed by
incubation at 25OC with an equimolar amount of ricin. The
asialofetuin-rich conjugate was then isolated by gel filtration
(Sephadex G-200), followed by affinity chromatography on
asialofetuin-Sepharose. The purified material eluted from the
column comprised a major fraction [asialofetuin-ricin, mol.wt.
1 lOO00) and a minor component of an asialofetuin dimer. The
ratio of asialofetuin/ricin in the preparation was 2.2 :1.
In agreement with our previous studies (Skilleter et al., 198 1)
the results in Table 1 show that '251-labelledricin is taken up by
Kupffer cells to a much greater extent than by parenchymal
cells. In contrast, the uptake of '2sI-labelled asialofetuin was
restricted to parenchymal cells. The accumulation of the
asialofetuin-ricin conjugate by parenchymal cells was twice that
measured in Kupffer cells. In view of the ricin content of the
conjugate, this represents no change in the rate of rich
accumulation by parenchymal cells but over a 90% decrease in
the rate of ricin uptake by Kupffer cells. These results suggest
that the asialofetuin moiety may be involved in the direction of
the conjugate to parenchymal cells. This hypothesis is supported
by the finding (Table 1) that the addition of asialofetuin, but not
ricin, to the culture medium inhibits the uptake of the conjugate
by parenchymal cells. Measurements (Table 1) made on the
uptake of ricin in presence of either galactose or mannose
confirmed our previous observation (Skilleter et al., 198 1) that
ricin appears to be recognized and taken up by Kupffer cells via
mannose residues present in the toxin, whereas uptake by
parenchymal cells is mediated by a process involving galactose.
The uptake of asialofetuin by parenchymal cells is known to
involve galactose (Ashwell & Morell, 1977) and predictably was
inhibited by this sugar (Table 1). Accordingly the parenchymalcell uptake, and resultant inhibition of protein synthesis, by rich
and the asialofetuin-ricin conjugate is prevented by galactose
(Table 1). Kupffer-cell accumulation of the asialofetuin-ricin
conjugate was most affected by mannose (75% decrease),
although galactose also had an inhibitory effect (40% decrease),
Table 1. Uptake of asialofetuin, ricin and asialofetuin-ricin conjugate by cultured liver cells and the concentrations that cause a
5096 inhibition (ID5,,)
ofprotein synthesis
Uptake (pghin per lo6cells)* ID,,? (ng/ml of medium)
\
I
Treatment
Cells . . . Parenchymal
Kupffer
Parenchymal
Kupffer
3
0.04
Ricin
13059
947 f 45
36k2
+20 mM-Galactose
805 f 32
60
0.06
+20m~-Mannose
108f 5
464 f 26
3
0.07
223k14
126f12
400
9.0
Conjugate
63f6
+ 1pg of asialofetuin/ml
215k9
+ 1pg of ricidml
+2Om~-Galactose
18k2
75f3
8600
0.5
+20m~-Mannose
228 k 5
33f3
400
9.3
Asialofetuin
292 f 18
(5
None
None
+ 20m~-Galactose
282 1
None
None
+20 mM-Mannose
None
None
309 k 8
*Measurements m mean?^.^.^., n =4) were made in cell culture (Skilleter et al., 1981) by using 0 . 1 of~ '*'I-labelled materialhl of culture
medium.
7 Cells treated as indicated for either 1h followed by 23 h without treatment (Kupffer cells) or continuously for 24 h (parenchymal cells) before
the measurement of protein synthesis (Villa el al., 1980) with ~-14,5-~H1leucine.
1982
597th MEETING. LONDON
123
suggesting that the uptake of the conjugate by Kupffer cells is a
more complex process than in parenchymal cells. Furthermore,
for Kupffer cells neither galactose or mannose had any
significant effect on the inhibition of protein synthesis by ricin or
the asialofetuin-rich conjugate. However, the results show that
linking rich to asialofetuin directs the toxin, in uitro, more
specifically towards hepatic parenchymal than Kupffer cells
with the retention of the potential to inhibit protein synthesis.
Ashwell, G. & Morell, A. G. (1977)Trends Biochem. Sci. 2,76-78
Eiklid, K., Olsnes, S. & Pihl, A. (1980)Exp. Cell Res. 126,321-326
Hubbard, A. L., Wilson, G., Ashwell, G. & Stukenbrook, H. (1979)J .
Cell Biol. 83,41-64
Nicolson, G.L. & Blaustein, J. (1972)Biochem. Biophys. Acfa 266,
543-547
Olsnes, S., Refsnes, K. & Phil, A. (1974) Nature (London) 249,
627-63 1
Olsnes, S.(1981)Nafure (London) 290,84
Skilleter, D. N., Paine, A. J. & Stripe, F. (1981)Biochim. Biophys. Acfa
611,495-500
Villa, P., Hockin, L. J. & Paine, A. J. (1980)Biochem. Pharmacol. 29,
1773-1777
Nature of the polar urinary metabolites of metiamide and cimetidine in man
S. C. MITCHELL, J. C. RITCHIE, J. R. IDLE and
R. L. SMITH
Department of Biochemical and Experimental Pharmacology,
St. Mary's Hospital Medical School, Norfolk Place,
Paddington, London W2 IPG, U.K.
Metiamide
[ N-methyl-N'-(2-( [(5-methyl- 1H-imidazol-4-y1)methyllthio }ethyl)thioureal and cimetidine [ N-cyano-N'methyl-N"-(2- { [ (5-methyl- lH-imidazol-4-yl)rnethyl1thio)ethyl)guanidinel (Fig. 1) both antagonize the effects of histamine on
H,-receptors. The former compound unfortunately produces
adverse reactions, including agranulocytosis, but the cyanoguanidine analogue is now in extensive use in the treatment and
prophylaxis of gastrointestinal ulcer disease and to reduce
gastric acidity states (Greenberger et al., 1978). It has been
shown that, for both compounds, most of the drug is rapidly
excreted unchanged in the urine together with smaller amounts
of the sulphoxide and an unidentified polar metabolite (Taylor et
al., 1978, 1979). The identity of this major polar metabolite has
now been further investigated.
Adult male volunteers were given single oral doses of
[2-'4C]metiamide (50mg, 10pCi) or [2-I4Clcimetidine (500mg,
1OpCi) and the following 24 h urine collected. Investigations
into the nature of the polar metabolites (20-25% of the
administered dose) were carried out at three levels of purity; on
untreated urine, as methanol extracts of the appropriate areas
from t.1.c. plates [silica-gel 60F,,,, 0.2 mm thick, developed in
either ethyl acetate/methanol/aq. NH, (sp.gr. 0.88) (8 :1: 1 by
vol.) or ethanol/water/aq. NH, (sp.gr. 0.88) (16 :3 :1 by vol.)]
or paper chromatograms (Whatman 1 or 3MM developed by
the descending method in butanol/acetic acid/water (12 :3 :5 by
R
II
II
Fig. 1. Nature of the polar urinary metabolites of metiamide
and cimetidine in man
The diagram illustrates the two possible sites of side-chain
N-glucuronidation for metiamide (R = S) and cimetidine (R =
NCN). It is possible that both isomers exist, especially in the
case of metiamide, where two radioactive areas were resoluble.
The asterisk denotes the position of the radioactive carbon label.
Vol. 10
vol.)], and as purified extracts after lead precipitation (Kamil et
al., 195 1) and paper chromatography (as above).
Incubation of urine samples (about lml) at 37OC for 24h
with ,&glucuronidase (Escherichia coli; Sigma), arylsulphatase
( H . pomatia) or nucleoside phosphorylase (bovine spleen) gave
no detectable breakdown, suggesting that the metabolite was not
an 0-glucuronide, sulphate or ribose conjugate. Acid hydrolysis
(conc. HCl, 1h reflux) led to its complete breakdown, with a
corresponding increase in metiamide or cimetidine hydrolysis
products (Durant et al., 1977). A more extensive investigation
into the hydrolysis of the cimetidine metabolite was undertaken
over a range of pH values and it demonstrated that it was more
susceptible to alkali than acid hydrolysis with maximum
stability occurring around pH 6-7.
The use of various chemical spray reagents gave additional
information about the metabolites (Offord et al., 1969; Sherma
1972). A positive reaction was obtained with both the modified
Grote's reagent and Tollen's reagent for the metiamide
derivative demonstrating the presence of the thiourea moiety,
and both metabolites reacted with Pauly's reagent, indicating an
intact imidazole ring. In addition, the compounds gave a positive
reaction with the naphthoresorcinol reagent, suggestive of
glucuronides, but negative responses to both p-dimethylaminobenzaldehyde in acetic anhydride and the rhodizonic acid test,
indicating the absence of glycine and sulphate conjugates
respectively.
Paper chromatography of the purified extracts, after decomposition of the basic lead precipitate with H,S, showed the
presence of two polar radioactive areas for metiamide (7 and
18% of the dose), whereas only one peak was resoluble for
cimetidine. These were isolated by methanol elution and
subjected to Fourier-transform n.m.r. (100MHz). Both the
metiamide (in ,H,O) and cimetidine (in dimethyl sulphoxide)
compounds showed the presence of a sugar moiety and the
parent drug, all signals being identical except for metiamide,
where the N-CH, protons were deshielded relative to those in the
parent compound, providing supportive evidence for an
electron-withdrawing group on either of the two thiourea
nitrogen atoms, and for cimetidine, where the signal from the
proton attached to one of the side-chain nitrogen atoms was
absent, although it was not possible to determine which one was
involved.
The evidence so far accumulated suggests a sugar conjugate,
probably an N-glucuronide, for the major metabolite of both
metiamide and cimetidine (Fig. 1). These compounds are known
to be virtually resistant to P-glucuronidase treatment and more
susceptible to alkali than acid hydrolysis (Bridges et al., 1965).
The immediate positive reaction with Pauly's reagent suggests
that the imidazole ring is not involved in the conjugation and the
tentative n.m.r. data supports this assignment.
We gratefully acknowledge financial support from Smith Kline and
French Laboratories Ltd. (SKF) and the Science Research Council.