- Email: [email protected]
Molecular changes in erythrocyte membranes induced by long-term administration of clofibrate
European Journal of Pharmacology - Molecular Pharmacology Section, 245 (1993) 89-95
89
.~ 1993 Elsevier Science Publishers B.V. All rights reserved 0922-4106/93/$06.00
EJPMOL 90422
Molecular changes in erythrocyte membranes induced by long-term administration of clofibrate Antonio Morales-Aguilera and Adriana Sampayo-Reyes Dit~isi6n de Farmacolog{a, Unidad de Incestigaci6n Biom~;dica del Noreste. Monterr~% N.L., M~;tico
Received 23 September 1992. revised MS received 14 December 1992, accepted 22 December 1992
It has been reported that an enantiomer ((S)-(- )-4) of clofibric acid, and the racemate, can block the chloride conductance of skeletal muscle membrane. It has also been reported that several analogs of clofibric acid inhibit the HCO 3-C1 exchange of erythrocytes. Since the two effects are probably similar biophysical membrane phenomena, the possibility of a common molecular mechanism arises. We exposed Sprague-Dawley male rats to long-term administration of clofibrate and 20,25-diazacholesterol (20,25-D) for comparison, at equipotent doses. Clofibrate (but not 20,25-diazacholesterol) produced a significant increase in density of the 220,000 Da band (/3-spectrin) and a decrease, also significant, in density of bands 2.1, 2.2, 2.3, 2.6 (syndeins or ankyrins) and of bands 4.1 and 6. Thus, clofibrate exhibits a manifold effect on the protein profile of the erythrocyte membrane cytoskeleton which, due to the lack of effect of 20-25-D, does not seem to be produced by the hypolipidemic effect per se, and thus deserves further study. Erythrocyte membrane; Clofibric acid; Clofibrate; Membrane electrophoresis; HCO~-C1
1. Introduction Feller et al. (1987) reported that an enantiomer ((S)-( - )-4) of clofibric acid (a-chlorophenoxyisobutyric acid), as well as the racemate, can block chloride conductance (G a ) of the skeletal muscle membrane. This blockade is an accepted mechanism of production of myotonia (Bryant and Morales-Aguilera, 1971; Kwiecinsky, 1978); it is, therefore, not surprising that clofibrate induces myotonia as a side-effect, since its active metabolite is clofibric acid which satisfies the molecular requirements for chemically induced myotonia, established by Bryant and Morales-Aguilera (1971) and by Palade and Barchi (1977): it is an aromatic monocarboxylic acid; the molecular structure is light (molecular weight less than 300 (242.7)) but extensive (as it is flat), and has several hydrophobic groups. On lhe other hand, Furhmann and Rudolphi (1988) reported that several analogs of clofibrate inhibit the H C O 3 - C 1 - exchange of erythrocytes and also produce a shift to the right of the hemoglobin dissociation curve. Since the blockade of sarcolemmal chloride con-
Correspondence to: Dr. Antonio Morales-Aguilera, Unidad de lnvestigaci6n Biom~dica del Noreste, Apdo. Postal 020-E, C.P. 64720 Monterrey, N.L., Mexico.
exchange inhibition; (Rat)
ductance and the inhibition of the HCO~-C1 exchange of erythrocytes seem to be similar biophysical m e m b r a n e phenomena, and there are analogies between the two types of membrane, the possible existence of a common molecular mechanism arises. We decided, therefore, to investigate whether there are changes in the protein profile of the erythrocyte membrane cytoskeleton that could explain the ionic changes produced by clofibrate (ethylchlorophenoxyisobutyrate), since this salt has a better bioavailability than clofibric acid via the oral route. In vitro exposure of rat erythrocytes to clofibric acid and similar agents did not show effects on the protein profile of the membranes. Therefore, we exposed Sprague-Dawley rats to longterm administration of similarly potent doses of clofibrate and 20,25-diazacholesterol (20,25-D) for comparison, since this latter drug is also hypolipidemic but does not modify chloride conductance significantly (Furman and Barchi, 1981). The initial expectation that band 3, accepted anion transporter in the erythrocyte m e m b r a n e (Rothstein et al., 1976), would be affected by our drugs was not confirmed by unidimensional electrophoresis. However, clofibrate (but not 20,25-diazacholesterol) produced a significant increase in the density of the band of 220,000 Da and a decrease, also significant, of the bands 2.1, 2.2, 2.3, 2.6 (syndeins or ankyrins), and of band 4.1 and band 6. Thus, clofibrate
90 exhibits a manifold effect on the proteins of the erythrocyte membrane cytoskeleton, most probably due to clofribic acid. This does not appear to be due to the hypolipidemic effects per se, and deserves further analysis, both with regard to the possible mechanism of action and the potential significance for erythrocyte membrane stability or other properties.
2. Materials and methods
2.1. Animals and drug admin&tration Twenty 40-day-old male Sprague-Dawley rats (200250 g) were divided into four cohorts of five animals each. The first cohort was used as a control and only received propyleneglycol. The second cohort received clofibrate (300 mg/kg) with propyleneglycol as vehicle, and was used for the protein profile determinations. The third group also received clofibrate at the same dose and during the same period as the second one, and was used for determinations of osmotic fragility and scanning electron microscopy observations. The fourth group received 20,25-diazacholesterol (50 mg/kg) with distilled water as the vehicle. All drugs were administered once a day per os by means of an esophageal cannula in volumes less than 0.32 ml. All animals received water and commercial rat food (Anderson Clayton) ad libitum.
2.2. Withdrawal and processing of blood samples The animals were sacrificed after 40 days of treatment by means of a sodium pentobarbital injection (50 mg/kg, i.p.), after 16 h of fasting. Two types of blood sample were obtained by cardiac puncture: one with anticoagulant (heparin) and one without. The samples with anticoagulant were used for the determination of hemoglobin and hematocrit and for the preparation of erythrocyte membranes ('ghosts') by the method of Atkinson et al. (1980). These samples were centrifuged at 20,000 x g at 4°C; they were then rinsed several times with a sodium chloride solution (0.15 M) at pH 7.4, and the supernatants decanted until only a cream-colored pellet remained with a negligible concentration of hemoglobin. Samples were obtained from this pellet and were observed by phase-contrast microscopy to confirm the presence of erythrocyte 'ghosts'. From a similar volume of the samples with heparin, hemoglobin content (Hainline, 1958) and hematocrit values (Lynch et al., 1972) were determined. Using samples without anticoagulant cholesterol (Zlatkis, 1953), total serum proteins (Reinhold, 1953) and glucose (Dubowsky, 1982) were also determined.
2.3. Electrophoresis in sodium dodecyl sulfate (SDS)polyacrylamide gels This was performed according to the discontinuous system technique described by Laemmli (1970). Pertinent modifications were as follows: the discontinuous system used was 5% and 8%. The protein concentration in each cell was adjusted to 1 /zg/ml (30/zg per cell). The voltage was maintained at 70 V for 8 h. The slabs were fixed overnight in a solution of water/ methanol/acetic acid (50:40: 10, v/v/v). The fixed slabs were stained with Coomassie blue at 0.1% in the fixing solution.
2.4. Densitometry Densitometry was performed in an R-110 Beckman densitometer at 520 nm with the negatives of the slabs previously photographed with high-contrast film (Kodak Technical Pan Film 2415).
2.5. Osmotic fragility and scanning electron microscopy observations Osmotic fragility of erythrocytes was determined from heparinized samples according to Beutler (1990). Scanning electron microscopy was performed from similar samples. Both osmotic fragility determinations and microscopy were performed at the end of the 40-day period of administration of clofibrate.
2.6. Reagents source The following reagents were used (sources are given in parentheses): Coomassie brilliant blue R-250 and bromophenol blue (BioRad), control serum IQ-Pak I (Cooper Frederick); Biuret (I.M.S.S.); sodium chloride 0.9% (Laboratorios Pisa), acetic acid, sulfuric acid, acetic anhydride, sodium carbonate (anhydrous), methanol, glycerol (Merck de M6xico); Acuglobin (Ortho Diagnostic System); urea, trichloroacetic acid (Productos Qulmicos Monterrey); Drabkin solution (Pro Lab de Jalisco); sodium potassium tartrate (Qu~mica Scott); heparin (Riker); acrylamide, bis-acrylamide, glycine, /3-mercaptoethanol, bovine serum albumin, sodium dodecyl sulfate, tris-(hydroxymethyl)aminoethane, N-N-N'-N'-tetramethylenediamine, phenol reagent (Folin and Ciocaiteus') (Sigma Chemical Company); sodium pentobarbital (Anestesal, Norden de M6xico, Div. of Smith, Kline and French); sodium hydroxide, sodium sulfate pentahydrate, sodium phosphate dibasic (Tdcnica Qulmica).
2. 7. Drugs Because of the poor absorption of clofibric acid, we use clofibrate (Atromid-S; Imperial Chemical Indus-
91
clofibrate group were between 14.2 and 22.3 g /d l but the difference in ranges was not significant. However, the values of the group treated with 20,25-diazacholesterol were in the range 11.5-12.8 g/dl, and this difference was significant at the level of P < 0.05. Control rat hematocrit values ranged from 47 to 51; those of animals receiving clofibrate from 44 to 51, and those of rats treated with 20,25-diazacholesterol from 40 to 42. These latter values differ significantly from the other two groups at the level of P < 0.05. Both clofibrate and 20,25-diazacholesterol significantly diminished serum cholesterol. Control values ranged from 63.74 to 75.55 mg/dl; clofibrate-treated values between 39.42 and 55.20 mg/dl, and 20,25-diazacholesterol-treated values between 40.74 and 47.90 mg/dl. The values obtained in treated animals differed significantly from the controls (P < 0.05), but there were no inter-individual differences. The control levels of glycemia were between 156.16 and 180.16 mg/dl; clofibrate-treated animals had levels between 112.29 and 229.36 mg/dl, and, 20,25-diazacholesterol-treated ones between 75.51 and 164.24 mg/dl. The KruskalWallis test did not show significant differences. Total serum proteins of control animals varied between 5.10 and 6.12 g/dl; clofibrate-treated values were between 5.48 and 6.39 g/dl, while 20,25-D-treated animals varied between 5.91 and 6.12 g/dl. There were no significant differences between the three groups. Fig. 1 shows the percentage of hemolysis of erythrocytes (control group and the group treated for 40 days) in the presence of different concentrations of NaC1. In both groups the percentage of hemolysis at concentrations above 0.6 g/1 was between 0.5 and 3.7%; there were no significant differences between the groups. Nor were
tries). The 20,25-diazacholesterol was a gift from Searle Chemical). 2.8. Statistics An analysis of no parametric variance (ANOVA, Kruskal-Wallis) was used to compare values of serum glucose, cholesterol and total protein, hemoglobin, hematocrit and osmotic fragility of the four groups of animals. When borderline significant differences were found, the method of Newman-Keuls for comparison of ranges was also used (Zar, 1974). To compare the integrations of the areas occupied in the densitometric records by the electrophoretic bands, Student's t-test was used as well as a parametric variance test (ANOVA; Zar, 1974).
3. Results The growth curve of animals treated with both drugs (clofibrate and 20,25-diazacholesterol) showed a shift to the right (i.e., their increase in weight was about 10% smaller than that of the controls), but their behavior and mobility were similar to the normal animals, except for the fact that slight myotonic effects were observed after the 2nd week of treatment. The 20,25diazacholesterol-treated animals also exhibited some of the general changes (hair loss, growth of the testis, etc.) already described by other authors (Kwiecinski, 1981), but the clofibrate-treated rats resembled the normal ones in all respects except weight. The hemoglobin concentration in control animals varied between 13.5 and 17.2 g/dl; the values of the
I00
-I
o 60 hi -r I.I.
O40 Z w u
~ zo
o -o
t III
A
~
127
143 m
159
~=
~--~I
175
191
28:
OSMOLARtTY
Fig. 1. Relationship between concentration of NaC1 solution and percentage hemolysis in erythrocytes from control group rats ( • ) and rats exposed for 40 days (©) to 300 mg/kg of clofibrate.
92 BANDS
there significant differences in hemolysis below 0.35 g/1 (88.2-99.3%). However, at 0.5 and 0.45 g/1 there were significant differences between some groups. The 40-day treatment group showed 1.2% hemolysis at 0.5 g/1 and 41.7% at 0.45 g/1. These values are significantly different (P < 0.05) from the values of hemolysis at that NaCI concentration in the control group (25.0% and 88%, respectively). The erythrocytes obtained from both groups were observed by means of scanning electron microscopy and it was found that almost all the cells from the control group were normal in shape and size (fig. 2A). On the other hand, almost 100% of the erythrocytes obtained from the rats treated for 40 days showed forms of echinocytes (fig. 2B illustrates a typical example). Electrophoresis was performed in each one of the samples. The results were reproducible in our conditions. Fig. 3 permits a comparison of the membrane protein profile of human erythrocytes with that obtained from the rat. The extra bands in the regions indicated by the arrows are not a product of a higher resolution in our method; using the same method we obtained the profile of human erythrocytes and those
MOLECULAR WEIGHT
(x~O-3) t
240
2 2.1 2.2
2.6
3 ~,.1
90 78
b,.2
~,.9
35
23 17
HEMOGLOBIN
Fig. 3. Comparison of the m e m b r a n e protein profile of h u m a n erythrocytes with that obtained from the normal rat.
bands are not apparent. Since the main bands coincide with those of human erythrocytes, we decided to use the nomenclature proposed by Steck (1978) and by Goodman and Schiffer (1983). Fig. 4 shows the normal rat protein pattern (A), compared with the pattern obtained from a rat treated with clofibrate ~B) and with another treated with 20,25-diazacholesterol (C). No changes in the protein profile are observed in the case of the animals treated with 20,25-diazacholesterol. By contrast, in the case of clofibrate there are several changes: an increase in density of the band of 220,000 Da (/3-unit of spectrin), and a decrease in the density of the 200,000 Da (2.1, 2.2, 2.3 and 2.6: syndeins or ankyrins), 78,000 Da (band 4.1) and 35,000 Da bands (band 6: glyceraldehyde-3phosphate dehydrogenase or G3PD). Fig. 5 shows the densitometric profiles of the slab gels obtained from a normal rat (A) and from a rat treated with clofibrate (B) and with 20,25-diazacholesterol (C) confirming the observations made in the gels themselves. Table 1 summarizes the integrations of the areas occupied by each band in the densitometric records as TABLE 1 Areas of the main bands as percentages of total area. Values are means_+ S.E.M. Band
Fig. 2. Scanning electron micrographs showing typical examples of erythrocytes obtained from control rats (A), and rats which received 300 m g / k g of clofibrate for 40 days (B).
Control
Clofibrate
20,25-Diazacholesterol
1 8.56 +0.45 8.52 + 0.46 8.52 + 0.50 2 2.72+0.26 5.76+0.26 * 2.8 +0.32 Fractions of 2 5.92_+0.10 2.84_+0.49 * 5.88_+0.17 3 13.0 _+ 1.00 13.0 _+ 1.00 13.0 _+ 1.00 4.1 3.44_+0.40 1.92_+0.10 * 3.48_+0.43 Fractions of 4 22.08_+2.32 22.8 _+2.22 23.56_+1.55 5 5.05_+0.08 5.0 _+0.14 5.08_+0.17 6 7.20_+0.17 5.20_+0.44 * 7.08_+0.17 Minor bands 24.40_+0.37 24.40-+0.40 24.20-+0.34 8 4.08+0.10 4.04_+0.08 4.20_+0.17 * P < 0.05 ('t', n = 5 and A N O V A , n = 5).
93
BANDS
MOLECULAR WEIGHT
(~m-3)
A
1
2~o
2
220
2.1 2.2
2O0
2.6 3
90 78
Zt.5 4.9 5 35
6 7 8
23
i
17
Fig. 4. Normal rat protein pattern (A), compared with the pattern obtained from a rat treated with clofibrate (B) and with another treated with 20,25-diazacholesterol (C). For explanations see text.
a percentage of the total area of the corresponding records, from all the animals from which blood samples were obtained. The increase in the area of band 2 as well as the decrease of bands 2.1, 2.2, 2.3, 2.6, 4.1 and 6 are significant at the level of P < 0.05 (t-test and parametric ANOVA). Table 2 shows the ratio between the area occupied by each of the bands modified by clofibrate and band 3, which was not modified in any of the cases. The modified ratios were significant at the level of P < 0.05 (t-test and parametric ANOVA). No changes in the erythrocyte electrophoretic profile were detected when the erythrocyte packet was exposed in a ratio 1 : 1 (v/v) to clofibric acid (5 x 10 - 4 M in phosphate buffer) at pH = 7.4, at 4°C or at 37°C forlhor16h.
A
B
C
MOLECULAR WEIGHT (x 1°-3)
,,------'------J
240 .220 200 -90 -78
4. Discussion The effect of clofibrate on the growth of the animals is similar to the effect described by Ramos-Ramlrez et al. (1983) in Wistar rats produced by anthracene-9carboxylic acid (9-AC), a more potent chloride conductance blocker than clofibrate, and by Mendoza et ai. (1986) in Sprague-Dawley rats produced by bezafibrate, an analog of clofibrate. In our study, the effect on growth seems less pronounced than the effect of 9-AC. In both cases, males were more affected than females. Therefore, it seems that aromatic monocarboxylic acid of the chloride-blocker type affects the growth of young rats, although the intake or excretion of food and water is not modified according to observations made in our laboratory with rats in metabolic cages (Jim6nez et al., 1981). Furthermore, we did not find changes in either blood glucose or total serum proteins. Therefore, the reasons for the effect on growth or for the higher sensitivity of males to that effect remain to be determined. The myotonic effects produced by 20,25-diazacholesterol and by clofibrate, although not very marked, were similar to the effects
-47
TABLE 2 Band n / b a n d 3 ratio. Values are means_+ S.E.M.
17
Fig. 5. Densitometric profiles of the slab gels obtained from a normal rat (A) and from a rat treated with clofibrate (B) and with 20,25-diazacholesterol (C) confirming the observations made in the gels themselves.
Band n
Control
Clofibrate
20,25-Diazacholesterol
2 Fractions of 2 4.1 6
0.20 0.45 0.26 0.55
0.44 0.21 0.14 0.40
0.21 0.45 0.26 0.54
* * * *
• P < 0.05 ('t', n = 5 and ANOVA, n = 5).
94
previously described by several authors (see review by Kwiecinsky, 1981). The decreases in serum cholesterol produced by 20,25-diazacholesterol and by clofibrate confirm our contention that we administered doses of both substances with similar effects on serum cholesterol level. The normal values of glycemia and total serum proteins indicate that the treated animals did not suffer gross nutritional deficiencies or metabolic side-effects of the drugs that could explain the observed changes in the protein profile of the erythrocytes. The absence of effects of 20,25-diazacholesterol on the electrophoretic pattern of membrane proteins strongly suggests that the hypolipidemic action or the decrease in serum cholesterol per se could not explain the changes produced by clofibrate. Also, the absence of in vitro effects of clofibrate and of clofibric acid and the temporal course of the effects, suggest that the electrophoretic changes are probably due to an action of clofibric acid on protein synthesis hitherto unknown. In view of our results, any possible effects of clofibric acid on HCO3-CI- exchange could hardly be explained by a quantitative modification of band 3, generally accepted as the anion transport protein. The absence of changes in band 3 density in our case does not rule out other effects on band 3, since the unidimensional electrophoretic technique detects only quantitative changes in bands formed by polypeptides of similar molecular weight, but does not detect structural or functional changes in the proteins. Thus, it is conceivable that clofibric acid might be acting in a locus on band 3 with affinity for the acid, similar to the locus described by Knauf et al. (1987) for flufenamic acid. The decrease in density of the band of 200,000 Da corresponds to a decrease in bands 2.1, 2.2, 2.3 and 2.6 (syndeins or ankyrins) which are normally associated with the cytoplasmic domain of band 3 and bind it to spectrin, anchoring the latter to the membrane's most abundant integral protein (band 3), and decreasing its mobility. A decrease of syndeins might affect the functions of band 3 if its cytoplasmic domain has to be fixed to the microenviroment in order to experience the confirmational changes required by the ping-pong model of anion transport. A decrease in the band of 78,000 Da or protein 4.1 (of which two very closely related forms, 4.1a and 4.1b, have been described) would also leave the tetramers of spectrin partially unbound (since 4.1 protein normally binds the tail-ends of one tetramer to another with the mediation of actin), and might contribute to the instability of the membrane and to the observed changes in erythrocyte shape and osmotic resistance. An increase in the band of 220,000 Da (band 2: /3-unit of spectrin) could be due to an increased synthesis of that protein, since the ratio band 2/band 3 is increased 2-fold, despite the fact that the profile of
band 3, the more abundant membrane protein, did not change. To speculate, the increase in band 2 could be due to a homeostatic mechanism at the erythroblast level intended to compensate for the functional deficiencies of the syndeins and 4.1 protein-deficient circulating erythrocytes. Spectrin determines the shape and the viscoelastic properties of the erythrocyte and it is considered essential for the normal stability of the membrane (Chasis et al., 1988). An increase in the /3-unit of spectrin would probably increase the mechanical stability of the membrane (Waugh and Agre, 1988). Here, we also report increases in osmotic resistance and significant morphological changes in erythrocytes of clofibrate-treated animals. The change in the relation of the percentage hemolysis to the osmolarity observed in the group exposed to clofibrate for 40 days suggests an increase in the osmotic resistance. This increase is probably dependent on an increase in the surface/volume ratio, but possible modifications of erythrocyte membrane deformability or stability cannot be inferred from changes in osmotic resistance, and remain to be determined. In summary, the manifold effects produced by longterm administration of clofibrate on the erythrocyte cytoskeleton seem to be due to a direct or combined effect of clofibric acid and not to the decrease in cholesterol synthesis per se. The functional implications of these effects deserve further study.
References Atkinson, J.B., L.L. Swift, P.G. Lankford and V.S. Lequire, 1980 A generalized membrane defect in hereditable myotonia: studies of erythrocytes in animal model and patients, Proc. Soc. Exp. Biol. Med. 163, 69. Beutler, E., 1990, in: Hematology, eds. J. Williams, E. Beutler, J.E. Allan and A.L. Marshall (McGraw Hill Publishing Co., New York) p. 1726. Bryant, S.H. and A. Morales-Aguilera, 1971, Chloride conductance in normal and myotonic muscle fiber and the action of monocarboxilic aromatic acids, J. Physiol. (London) 219, 367. Chasis, J.A., M. Prenat and N. Mohands, 1988, Skeletal and cytoskeletal remodeling during reticulocyte maturation, Blood 72 (Suppl. 1), Abstr. 24A. Dubowsky, K.M., 1962, An ortoluidin method for body fluid glucose determination, Clin. Chem. 8, 125. Feller, D.R., V.S. Komana, H.A.J. Newman, K.J. Romsted, D.T. Witiak, G. Bettoni, S.H. Bryant, D. Conte-Camerino, F. Loiodice and V. Tortorella, 1987, Dissociation of hypolipidemic and antiplatelet actions from adverse myotonic effects of clofibric acid related enantiomers, J. Med. Chem. 30, 1265. Fuhrman, G.G. and K. Rudolphi, 1988, The effects of hypolipidemic drugs on oxygen dissociation curve of hemoglobin and on the HCL--CI exchange in erythrocytes, Naunyn-Schmied. Arch. Pharmacol. 337, 47. Furman, R.E. and R.L. Barchi, 1981, 20,25-Diazacholesterol myotonia, an electrophysiological study, Ann. Neurol. 10, 225. Goodman, S.R. and K. Schiffer, 1983, The spectrin membrane skeleton of normal and abnormal human erythrocytes, Am. J. Physiol. 244, C121.
95 Hainline, A., 1958, in: Haemoglobin: Standard Methods of Clinical Chemistry, vol. 2 (Academic Press, New York) p. 71. Jim~nez, S.Z., N.A. Villegas and A. Morales-Aguilera, 1984, Estudios en Modelos de Miotonla Inducida Cr6nicamente. Aspectos Hematol6gicos y Bioqulmicos, in: Memorias del VII Congreso Nacional de Farmacologla, 20-25 March, Monterrey, N.L. M6xico ed. Amefar (Monterrey) p. 86. Knauf, P.A., L.J. Spinelle and N.A. Mann, 1987, Affinities of flulenamic acid (FA) for different conformations of the human erythrocyte anion transport protein, Band 3, Fed. Proc. 46, 543. Kwiecinski, H., 1978, Myotonia induced with clofibrate in rats, J. Neurol. 219, 107. Kwiecinski, H., 1981, Myotonia induced by chemical agents, Crit. Rev. Toxicol. 8, 279. Laemmli, U.K., 1970, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227, 680. Lynch, M.J., Raphael, S.S. MeUor, P.D. Spare and M.J.H. Inwood, 1972, in: M&odos de Laboratorio, 2nd edn. (Editorial Interamericana (M6xico, D.F.)) p. 754. Mendoza, G.E., A. Sampayo-Reyes and A. Morales-Aguilera, 1988, Efectos sobre el perfil proteico de membranas de eritrocitos de otro agente bloqueador: bezafibrato, in: Memorias del XII Congreso Nacional de Farmacologia, 27-30 November, P~tzcuaro, Mich. M6xico, p. 17.
Palade, P.T. and R.L. Barchi, 1977, The inhibition of muscle membrane chloride conductance by aromatic carboxilic acid, J. Gen. Physiol. 69, 879. Ramos-Ramlrez, E.G., C. Lazgare-Bello and A. Morales-Aguilera, 1983, Effects of long-term administration of anthrancene-9carboxilic acid (9AC) on young and adult rats, Proc. West. Pharmacol. Soc. 26, 21. Reinhold, J.G., 1953, Total protein, albumin and globulin, in: Standard Methods of Clinical Chemistry, Vol. 1. ed. D. Seligson (Academic Press, New York) p. 21. Rothstein, A.Z.I., K.P. Cabantchnik and P.A. Knauf, 1976, Mechanisms of anion transport in red blood cell: role of membrane proteins, Fed. Proc. 35, 3. Steck, T.L., 1974, The organization of proteins in the human red blood cell membrane, J. Cell Biol. 62, 1. Waugh, R.E. and P. Agre, 1988, Reductions of erythrocyte membrane viscoelastic coefficients reflect spectrin deficiencies in hereditary spherocytosis, J. Clin. Invest. 81, 133. Zar, J.H., 1974, in: Biostatistical Analysis (Prentice-Hall, Englewood Cliffs, NJ) p. 171. Zlatkis, A., B. Zak and A.J. Boyle, 1953, A new method for the direct determination of serum cholesterol, J. Lab. Clin. Med. 41, 486.
89
.~ 1993 Elsevier Science Publishers B.V. All rights reserved 0922-4106/93/$06.00
EJPMOL 90422
Molecular changes in erythrocyte membranes induced by long-term administration of clofibrate Antonio Morales-Aguilera and Adriana Sampayo-Reyes Dit~isi6n de Farmacolog{a, Unidad de Incestigaci6n Biom~;dica del Noreste. Monterr~% N.L., M~;tico
Received 23 September 1992. revised MS received 14 December 1992, accepted 22 December 1992
It has been reported that an enantiomer ((S)-(- )-4) of clofibric acid, and the racemate, can block the chloride conductance of skeletal muscle membrane. It has also been reported that several analogs of clofibric acid inhibit the HCO 3-C1 exchange of erythrocytes. Since the two effects are probably similar biophysical membrane phenomena, the possibility of a common molecular mechanism arises. We exposed Sprague-Dawley male rats to long-term administration of clofibrate and 20,25-diazacholesterol (20,25-D) for comparison, at equipotent doses. Clofibrate (but not 20,25-diazacholesterol) produced a significant increase in density of the 220,000 Da band (/3-spectrin) and a decrease, also significant, in density of bands 2.1, 2.2, 2.3, 2.6 (syndeins or ankyrins) and of bands 4.1 and 6. Thus, clofibrate exhibits a manifold effect on the protein profile of the erythrocyte membrane cytoskeleton which, due to the lack of effect of 20-25-D, does not seem to be produced by the hypolipidemic effect per se, and thus deserves further study. Erythrocyte membrane; Clofibric acid; Clofibrate; Membrane electrophoresis; HCO~-C1
1. Introduction Feller et al. (1987) reported that an enantiomer ((S)-( - )-4) of clofibric acid (a-chlorophenoxyisobutyric acid), as well as the racemate, can block chloride conductance (G a ) of the skeletal muscle membrane. This blockade is an accepted mechanism of production of myotonia (Bryant and Morales-Aguilera, 1971; Kwiecinsky, 1978); it is, therefore, not surprising that clofibrate induces myotonia as a side-effect, since its active metabolite is clofibric acid which satisfies the molecular requirements for chemically induced myotonia, established by Bryant and Morales-Aguilera (1971) and by Palade and Barchi (1977): it is an aromatic monocarboxylic acid; the molecular structure is light (molecular weight less than 300 (242.7)) but extensive (as it is flat), and has several hydrophobic groups. On lhe other hand, Furhmann and Rudolphi (1988) reported that several analogs of clofibrate inhibit the H C O 3 - C 1 - exchange of erythrocytes and also produce a shift to the right of the hemoglobin dissociation curve. Since the blockade of sarcolemmal chloride con-
Correspondence to: Dr. Antonio Morales-Aguilera, Unidad de lnvestigaci6n Biom~dica del Noreste, Apdo. Postal 020-E, C.P. 64720 Monterrey, N.L., Mexico.
exchange inhibition; (Rat)
ductance and the inhibition of the HCO~-C1 exchange of erythrocytes seem to be similar biophysical m e m b r a n e phenomena, and there are analogies between the two types of membrane, the possible existence of a common molecular mechanism arises. We decided, therefore, to investigate whether there are changes in the protein profile of the erythrocyte membrane cytoskeleton that could explain the ionic changes produced by clofibrate (ethylchlorophenoxyisobutyrate), since this salt has a better bioavailability than clofibric acid via the oral route. In vitro exposure of rat erythrocytes to clofibric acid and similar agents did not show effects on the protein profile of the membranes. Therefore, we exposed Sprague-Dawley rats to longterm administration of similarly potent doses of clofibrate and 20,25-diazacholesterol (20,25-D) for comparison, since this latter drug is also hypolipidemic but does not modify chloride conductance significantly (Furman and Barchi, 1981). The initial expectation that band 3, accepted anion transporter in the erythrocyte m e m b r a n e (Rothstein et al., 1976), would be affected by our drugs was not confirmed by unidimensional electrophoresis. However, clofibrate (but not 20,25-diazacholesterol) produced a significant increase in the density of the band of 220,000 Da and a decrease, also significant, of the bands 2.1, 2.2, 2.3, 2.6 (syndeins or ankyrins), and of band 4.1 and band 6. Thus, clofibrate
90 exhibits a manifold effect on the proteins of the erythrocyte membrane cytoskeleton, most probably due to clofribic acid. This does not appear to be due to the hypolipidemic effects per se, and deserves further analysis, both with regard to the possible mechanism of action and the potential significance for erythrocyte membrane stability or other properties.
2. Materials and methods
2.1. Animals and drug admin&tration Twenty 40-day-old male Sprague-Dawley rats (200250 g) were divided into four cohorts of five animals each. The first cohort was used as a control and only received propyleneglycol. The second cohort received clofibrate (300 mg/kg) with propyleneglycol as vehicle, and was used for the protein profile determinations. The third group also received clofibrate at the same dose and during the same period as the second one, and was used for determinations of osmotic fragility and scanning electron microscopy observations. The fourth group received 20,25-diazacholesterol (50 mg/kg) with distilled water as the vehicle. All drugs were administered once a day per os by means of an esophageal cannula in volumes less than 0.32 ml. All animals received water and commercial rat food (Anderson Clayton) ad libitum.
2.2. Withdrawal and processing of blood samples The animals were sacrificed after 40 days of treatment by means of a sodium pentobarbital injection (50 mg/kg, i.p.), after 16 h of fasting. Two types of blood sample were obtained by cardiac puncture: one with anticoagulant (heparin) and one without. The samples with anticoagulant were used for the determination of hemoglobin and hematocrit and for the preparation of erythrocyte membranes ('ghosts') by the method of Atkinson et al. (1980). These samples were centrifuged at 20,000 x g at 4°C; they were then rinsed several times with a sodium chloride solution (0.15 M) at pH 7.4, and the supernatants decanted until only a cream-colored pellet remained with a negligible concentration of hemoglobin. Samples were obtained from this pellet and were observed by phase-contrast microscopy to confirm the presence of erythrocyte 'ghosts'. From a similar volume of the samples with heparin, hemoglobin content (Hainline, 1958) and hematocrit values (Lynch et al., 1972) were determined. Using samples without anticoagulant cholesterol (Zlatkis, 1953), total serum proteins (Reinhold, 1953) and glucose (Dubowsky, 1982) were also determined.
2.3. Electrophoresis in sodium dodecyl sulfate (SDS)polyacrylamide gels This was performed according to the discontinuous system technique described by Laemmli (1970). Pertinent modifications were as follows: the discontinuous system used was 5% and 8%. The protein concentration in each cell was adjusted to 1 /zg/ml (30/zg per cell). The voltage was maintained at 70 V for 8 h. The slabs were fixed overnight in a solution of water/ methanol/acetic acid (50:40: 10, v/v/v). The fixed slabs were stained with Coomassie blue at 0.1% in the fixing solution.
2.4. Densitometry Densitometry was performed in an R-110 Beckman densitometer at 520 nm with the negatives of the slabs previously photographed with high-contrast film (Kodak Technical Pan Film 2415).
2.5. Osmotic fragility and scanning electron microscopy observations Osmotic fragility of erythrocytes was determined from heparinized samples according to Beutler (1990). Scanning electron microscopy was performed from similar samples. Both osmotic fragility determinations and microscopy were performed at the end of the 40-day period of administration of clofibrate.
2.6. Reagents source The following reagents were used (sources are given in parentheses): Coomassie brilliant blue R-250 and bromophenol blue (BioRad), control serum IQ-Pak I (Cooper Frederick); Biuret (I.M.S.S.); sodium chloride 0.9% (Laboratorios Pisa), acetic acid, sulfuric acid, acetic anhydride, sodium carbonate (anhydrous), methanol, glycerol (Merck de M6xico); Acuglobin (Ortho Diagnostic System); urea, trichloroacetic acid (Productos Qulmicos Monterrey); Drabkin solution (Pro Lab de Jalisco); sodium potassium tartrate (Qu~mica Scott); heparin (Riker); acrylamide, bis-acrylamide, glycine, /3-mercaptoethanol, bovine serum albumin, sodium dodecyl sulfate, tris-(hydroxymethyl)aminoethane, N-N-N'-N'-tetramethylenediamine, phenol reagent (Folin and Ciocaiteus') (Sigma Chemical Company); sodium pentobarbital (Anestesal, Norden de M6xico, Div. of Smith, Kline and French); sodium hydroxide, sodium sulfate pentahydrate, sodium phosphate dibasic (Tdcnica Qulmica).
2. 7. Drugs Because of the poor absorption of clofibric acid, we use clofibrate (Atromid-S; Imperial Chemical Indus-
91
clofibrate group were between 14.2 and 22.3 g /d l but the difference in ranges was not significant. However, the values of the group treated with 20,25-diazacholesterol were in the range 11.5-12.8 g/dl, and this difference was significant at the level of P < 0.05. Control rat hematocrit values ranged from 47 to 51; those of animals receiving clofibrate from 44 to 51, and those of rats treated with 20,25-diazacholesterol from 40 to 42. These latter values differ significantly from the other two groups at the level of P < 0.05. Both clofibrate and 20,25-diazacholesterol significantly diminished serum cholesterol. Control values ranged from 63.74 to 75.55 mg/dl; clofibrate-treated values between 39.42 and 55.20 mg/dl, and 20,25-diazacholesterol-treated values between 40.74 and 47.90 mg/dl. The values obtained in treated animals differed significantly from the controls (P < 0.05), but there were no inter-individual differences. The control levels of glycemia were between 156.16 and 180.16 mg/dl; clofibrate-treated animals had levels between 112.29 and 229.36 mg/dl, and, 20,25-diazacholesterol-treated ones between 75.51 and 164.24 mg/dl. The KruskalWallis test did not show significant differences. Total serum proteins of control animals varied between 5.10 and 6.12 g/dl; clofibrate-treated values were between 5.48 and 6.39 g/dl, while 20,25-D-treated animals varied between 5.91 and 6.12 g/dl. There were no significant differences between the three groups. Fig. 1 shows the percentage of hemolysis of erythrocytes (control group and the group treated for 40 days) in the presence of different concentrations of NaC1. In both groups the percentage of hemolysis at concentrations above 0.6 g/1 was between 0.5 and 3.7%; there were no significant differences between the groups. Nor were
tries). The 20,25-diazacholesterol was a gift from Searle Chemical). 2.8. Statistics An analysis of no parametric variance (ANOVA, Kruskal-Wallis) was used to compare values of serum glucose, cholesterol and total protein, hemoglobin, hematocrit and osmotic fragility of the four groups of animals. When borderline significant differences were found, the method of Newman-Keuls for comparison of ranges was also used (Zar, 1974). To compare the integrations of the areas occupied in the densitometric records by the electrophoretic bands, Student's t-test was used as well as a parametric variance test (ANOVA; Zar, 1974).
3. Results The growth curve of animals treated with both drugs (clofibrate and 20,25-diazacholesterol) showed a shift to the right (i.e., their increase in weight was about 10% smaller than that of the controls), but their behavior and mobility were similar to the normal animals, except for the fact that slight myotonic effects were observed after the 2nd week of treatment. The 20,25diazacholesterol-treated animals also exhibited some of the general changes (hair loss, growth of the testis, etc.) already described by other authors (Kwiecinski, 1981), but the clofibrate-treated rats resembled the normal ones in all respects except weight. The hemoglobin concentration in control animals varied between 13.5 and 17.2 g/dl; the values of the
I00
-I
o 60 hi -r I.I.
O40 Z w u
~ zo
o -o
t III
A
~
127
143 m
159
~=
~--~I
175
191
28:
OSMOLARtTY
Fig. 1. Relationship between concentration of NaC1 solution and percentage hemolysis in erythrocytes from control group rats ( • ) and rats exposed for 40 days (©) to 300 mg/kg of clofibrate.
92 BANDS
there significant differences in hemolysis below 0.35 g/1 (88.2-99.3%). However, at 0.5 and 0.45 g/1 there were significant differences between some groups. The 40-day treatment group showed 1.2% hemolysis at 0.5 g/1 and 41.7% at 0.45 g/1. These values are significantly different (P < 0.05) from the values of hemolysis at that NaCI concentration in the control group (25.0% and 88%, respectively). The erythrocytes obtained from both groups were observed by means of scanning electron microscopy and it was found that almost all the cells from the control group were normal in shape and size (fig. 2A). On the other hand, almost 100% of the erythrocytes obtained from the rats treated for 40 days showed forms of echinocytes (fig. 2B illustrates a typical example). Electrophoresis was performed in each one of the samples. The results were reproducible in our conditions. Fig. 3 permits a comparison of the membrane protein profile of human erythrocytes with that obtained from the rat. The extra bands in the regions indicated by the arrows are not a product of a higher resolution in our method; using the same method we obtained the profile of human erythrocytes and those
MOLECULAR WEIGHT
(x~O-3) t
240
2 2.1 2.2
2.6
3 ~,.1
90 78
b,.2
~,.9
35
23 17
HEMOGLOBIN
Fig. 3. Comparison of the m e m b r a n e protein profile of h u m a n erythrocytes with that obtained from the normal rat.
bands are not apparent. Since the main bands coincide with those of human erythrocytes, we decided to use the nomenclature proposed by Steck (1978) and by Goodman and Schiffer (1983). Fig. 4 shows the normal rat protein pattern (A), compared with the pattern obtained from a rat treated with clofibrate ~B) and with another treated with 20,25-diazacholesterol (C). No changes in the protein profile are observed in the case of the animals treated with 20,25-diazacholesterol. By contrast, in the case of clofibrate there are several changes: an increase in density of the band of 220,000 Da (/3-unit of spectrin), and a decrease in the density of the 200,000 Da (2.1, 2.2, 2.3 and 2.6: syndeins or ankyrins), 78,000 Da (band 4.1) and 35,000 Da bands (band 6: glyceraldehyde-3phosphate dehydrogenase or G3PD). Fig. 5 shows the densitometric profiles of the slab gels obtained from a normal rat (A) and from a rat treated with clofibrate (B) and with 20,25-diazacholesterol (C) confirming the observations made in the gels themselves. Table 1 summarizes the integrations of the areas occupied by each band in the densitometric records as TABLE 1 Areas of the main bands as percentages of total area. Values are means_+ S.E.M. Band
Fig. 2. Scanning electron micrographs showing typical examples of erythrocytes obtained from control rats (A), and rats which received 300 m g / k g of clofibrate for 40 days (B).
Control
Clofibrate
20,25-Diazacholesterol
1 8.56 +0.45 8.52 + 0.46 8.52 + 0.50 2 2.72+0.26 5.76+0.26 * 2.8 +0.32 Fractions of 2 5.92_+0.10 2.84_+0.49 * 5.88_+0.17 3 13.0 _+ 1.00 13.0 _+ 1.00 13.0 _+ 1.00 4.1 3.44_+0.40 1.92_+0.10 * 3.48_+0.43 Fractions of 4 22.08_+2.32 22.8 _+2.22 23.56_+1.55 5 5.05_+0.08 5.0 _+0.14 5.08_+0.17 6 7.20_+0.17 5.20_+0.44 * 7.08_+0.17 Minor bands 24.40_+0.37 24.40-+0.40 24.20-+0.34 8 4.08+0.10 4.04_+0.08 4.20_+0.17 * P < 0.05 ('t', n = 5 and A N O V A , n = 5).
93
BANDS
MOLECULAR WEIGHT
(~m-3)
A
1
2~o
2
220
2.1 2.2
2O0
2.6 3
90 78
Zt.5 4.9 5 35
6 7 8
23
i
17
Fig. 4. Normal rat protein pattern (A), compared with the pattern obtained from a rat treated with clofibrate (B) and with another treated with 20,25-diazacholesterol (C). For explanations see text.
a percentage of the total area of the corresponding records, from all the animals from which blood samples were obtained. The increase in the area of band 2 as well as the decrease of bands 2.1, 2.2, 2.3, 2.6, 4.1 and 6 are significant at the level of P < 0.05 (t-test and parametric ANOVA). Table 2 shows the ratio between the area occupied by each of the bands modified by clofibrate and band 3, which was not modified in any of the cases. The modified ratios were significant at the level of P < 0.05 (t-test and parametric ANOVA). No changes in the erythrocyte electrophoretic profile were detected when the erythrocyte packet was exposed in a ratio 1 : 1 (v/v) to clofibric acid (5 x 10 - 4 M in phosphate buffer) at pH = 7.4, at 4°C or at 37°C forlhor16h.
A
B
C
MOLECULAR WEIGHT (x 1°-3)
,,------'------J
240 .220 200 -90 -78
4. Discussion The effect of clofibrate on the growth of the animals is similar to the effect described by Ramos-Ramlrez et al. (1983) in Wistar rats produced by anthracene-9carboxylic acid (9-AC), a more potent chloride conductance blocker than clofibrate, and by Mendoza et ai. (1986) in Sprague-Dawley rats produced by bezafibrate, an analog of clofibrate. In our study, the effect on growth seems less pronounced than the effect of 9-AC. In both cases, males were more affected than females. Therefore, it seems that aromatic monocarboxylic acid of the chloride-blocker type affects the growth of young rats, although the intake or excretion of food and water is not modified according to observations made in our laboratory with rats in metabolic cages (Jim6nez et al., 1981). Furthermore, we did not find changes in either blood glucose or total serum proteins. Therefore, the reasons for the effect on growth or for the higher sensitivity of males to that effect remain to be determined. The myotonic effects produced by 20,25-diazacholesterol and by clofibrate, although not very marked, were similar to the effects
-47
TABLE 2 Band n / b a n d 3 ratio. Values are means_+ S.E.M.
17
Fig. 5. Densitometric profiles of the slab gels obtained from a normal rat (A) and from a rat treated with clofibrate (B) and with 20,25-diazacholesterol (C) confirming the observations made in the gels themselves.
Band n
Control
Clofibrate
20,25-Diazacholesterol
2 Fractions of 2 4.1 6
0.20 0.45 0.26 0.55
0.44 0.21 0.14 0.40
0.21 0.45 0.26 0.54
* * * *
• P < 0.05 ('t', n = 5 and ANOVA, n = 5).
94
previously described by several authors (see review by Kwiecinsky, 1981). The decreases in serum cholesterol produced by 20,25-diazacholesterol and by clofibrate confirm our contention that we administered doses of both substances with similar effects on serum cholesterol level. The normal values of glycemia and total serum proteins indicate that the treated animals did not suffer gross nutritional deficiencies or metabolic side-effects of the drugs that could explain the observed changes in the protein profile of the erythrocytes. The absence of effects of 20,25-diazacholesterol on the electrophoretic pattern of membrane proteins strongly suggests that the hypolipidemic action or the decrease in serum cholesterol per se could not explain the changes produced by clofibrate. Also, the absence of in vitro effects of clofibrate and of clofibric acid and the temporal course of the effects, suggest that the electrophoretic changes are probably due to an action of clofibric acid on protein synthesis hitherto unknown. In view of our results, any possible effects of clofibric acid on HCO3-CI- exchange could hardly be explained by a quantitative modification of band 3, generally accepted as the anion transport protein. The absence of changes in band 3 density in our case does not rule out other effects on band 3, since the unidimensional electrophoretic technique detects only quantitative changes in bands formed by polypeptides of similar molecular weight, but does not detect structural or functional changes in the proteins. Thus, it is conceivable that clofibric acid might be acting in a locus on band 3 with affinity for the acid, similar to the locus described by Knauf et al. (1987) for flufenamic acid. The decrease in density of the band of 200,000 Da corresponds to a decrease in bands 2.1, 2.2, 2.3 and 2.6 (syndeins or ankyrins) which are normally associated with the cytoplasmic domain of band 3 and bind it to spectrin, anchoring the latter to the membrane's most abundant integral protein (band 3), and decreasing its mobility. A decrease of syndeins might affect the functions of band 3 if its cytoplasmic domain has to be fixed to the microenviroment in order to experience the confirmational changes required by the ping-pong model of anion transport. A decrease in the band of 78,000 Da or protein 4.1 (of which two very closely related forms, 4.1a and 4.1b, have been described) would also leave the tetramers of spectrin partially unbound (since 4.1 protein normally binds the tail-ends of one tetramer to another with the mediation of actin), and might contribute to the instability of the membrane and to the observed changes in erythrocyte shape and osmotic resistance. An increase in the band of 220,000 Da (band 2: /3-unit of spectrin) could be due to an increased synthesis of that protein, since the ratio band 2/band 3 is increased 2-fold, despite the fact that the profile of
band 3, the more abundant membrane protein, did not change. To speculate, the increase in band 2 could be due to a homeostatic mechanism at the erythroblast level intended to compensate for the functional deficiencies of the syndeins and 4.1 protein-deficient circulating erythrocytes. Spectrin determines the shape and the viscoelastic properties of the erythrocyte and it is considered essential for the normal stability of the membrane (Chasis et al., 1988). An increase in the /3-unit of spectrin would probably increase the mechanical stability of the membrane (Waugh and Agre, 1988). Here, we also report increases in osmotic resistance and significant morphological changes in erythrocytes of clofibrate-treated animals. The change in the relation of the percentage hemolysis to the osmolarity observed in the group exposed to clofibrate for 40 days suggests an increase in the osmotic resistance. This increase is probably dependent on an increase in the surface/volume ratio, but possible modifications of erythrocyte membrane deformability or stability cannot be inferred from changes in osmotic resistance, and remain to be determined. In summary, the manifold effects produced by longterm administration of clofibrate on the erythrocyte cytoskeleton seem to be due to a direct or combined effect of clofibric acid and not to the decrease in cholesterol synthesis per se. The functional implications of these effects deserve further study.
References Atkinson, J.B., L.L. Swift, P.G. Lankford and V.S. Lequire, 1980 A generalized membrane defect in hereditable myotonia: studies of erythrocytes in animal model and patients, Proc. Soc. Exp. Biol. Med. 163, 69. Beutler, E., 1990, in: Hematology, eds. J. Williams, E. Beutler, J.E. Allan and A.L. Marshall (McGraw Hill Publishing Co., New York) p. 1726. Bryant, S.H. and A. Morales-Aguilera, 1971, Chloride conductance in normal and myotonic muscle fiber and the action of monocarboxilic aromatic acids, J. Physiol. (London) 219, 367. Chasis, J.A., M. Prenat and N. Mohands, 1988, Skeletal and cytoskeletal remodeling during reticulocyte maturation, Blood 72 (Suppl. 1), Abstr. 24A. Dubowsky, K.M., 1962, An ortoluidin method for body fluid glucose determination, Clin. Chem. 8, 125. Feller, D.R., V.S. Komana, H.A.J. Newman, K.J. Romsted, D.T. Witiak, G. Bettoni, S.H. Bryant, D. Conte-Camerino, F. Loiodice and V. Tortorella, 1987, Dissociation of hypolipidemic and antiplatelet actions from adverse myotonic effects of clofibric acid related enantiomers, J. Med. Chem. 30, 1265. Fuhrman, G.G. and K. Rudolphi, 1988, The effects of hypolipidemic drugs on oxygen dissociation curve of hemoglobin and on the HCL--CI exchange in erythrocytes, Naunyn-Schmied. Arch. Pharmacol. 337, 47. Furman, R.E. and R.L. Barchi, 1981, 20,25-Diazacholesterol myotonia, an electrophysiological study, Ann. Neurol. 10, 225. Goodman, S.R. and K. Schiffer, 1983, The spectrin membrane skeleton of normal and abnormal human erythrocytes, Am. J. Physiol. 244, C121.
95 Hainline, A., 1958, in: Haemoglobin: Standard Methods of Clinical Chemistry, vol. 2 (Academic Press, New York) p. 71. Jim~nez, S.Z., N.A. Villegas and A. Morales-Aguilera, 1984, Estudios en Modelos de Miotonla Inducida Cr6nicamente. Aspectos Hematol6gicos y Bioqulmicos, in: Memorias del VII Congreso Nacional de Farmacologla, 20-25 March, Monterrey, N.L. M6xico ed. Amefar (Monterrey) p. 86. Knauf, P.A., L.J. Spinelle and N.A. Mann, 1987, Affinities of flulenamic acid (FA) for different conformations of the human erythrocyte anion transport protein, Band 3, Fed. Proc. 46, 543. Kwiecinski, H., 1978, Myotonia induced with clofibrate in rats, J. Neurol. 219, 107. Kwiecinski, H., 1981, Myotonia induced by chemical agents, Crit. Rev. Toxicol. 8, 279. Laemmli, U.K., 1970, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227, 680. Lynch, M.J., Raphael, S.S. MeUor, P.D. Spare and M.J.H. Inwood, 1972, in: M&odos de Laboratorio, 2nd edn. (Editorial Interamericana (M6xico, D.F.)) p. 754. Mendoza, G.E., A. Sampayo-Reyes and A. Morales-Aguilera, 1988, Efectos sobre el perfil proteico de membranas de eritrocitos de otro agente bloqueador: bezafibrato, in: Memorias del XII Congreso Nacional de Farmacologia, 27-30 November, P~tzcuaro, Mich. M6xico, p. 17.
Palade, P.T. and R.L. Barchi, 1977, The inhibition of muscle membrane chloride conductance by aromatic carboxilic acid, J. Gen. Physiol. 69, 879. Ramos-Ramlrez, E.G., C. Lazgare-Bello and A. Morales-Aguilera, 1983, Effects of long-term administration of anthrancene-9carboxilic acid (9AC) on young and adult rats, Proc. West. Pharmacol. Soc. 26, 21. Reinhold, J.G., 1953, Total protein, albumin and globulin, in: Standard Methods of Clinical Chemistry, Vol. 1. ed. D. Seligson (Academic Press, New York) p. 21. Rothstein, A.Z.I., K.P. Cabantchnik and P.A. Knauf, 1976, Mechanisms of anion transport in red blood cell: role of membrane proteins, Fed. Proc. 35, 3. Steck, T.L., 1974, The organization of proteins in the human red blood cell membrane, J. Cell Biol. 62, 1. Waugh, R.E. and P. Agre, 1988, Reductions of erythrocyte membrane viscoelastic coefficients reflect spectrin deficiencies in hereditary spherocytosis, J. Clin. Invest. 81, 133. Zar, J.H., 1974, in: Biostatistical Analysis (Prentice-Hall, Englewood Cliffs, NJ) p. 171. Zlatkis, A., B. Zak and A.J. Boyle, 1953, A new method for the direct determination of serum cholesterol, J. Lab. Clin. Med. 41, 486.