Attenuation of phosphoinositidase activity and phosphatidylinositol bisphosphate level of bovine retinal capillary pericytes in high glucose

Exp.

Eye Rea. (1989)

48, 999106

Attenuation of Phosphoinositidase Activity and Phosphatidylinositol Bisphosphate Level of Bovine Retinal Capillary Pericytes in High Glucose LI,*

WEIYE

Depa,rtment

LEI

TANG,

&I ZHOU,

MEI

&IN

AND TIANSHENG

of Ophthalmology, Eye Research Center, Peking Union Hospital, Beijing, People’s Republic of China

(Received 31 March

1988 and accepted in revised form

Hu

Medical

College

10 June 1988)

Both phosphoinositidase (PIase) and individual species of inositol phospholipid (IPL) of bovine retinal capillary pericytes (BRCP) were quantitatively determined. When glucose in growth medium was increased from 5 to 15 or 30 mM, PIase activity was attenuated to 82 % or 55 %, respectively. In contrast, when glucose (5-, l&,30 mM) was added to an enzyme extract from cells grown in the standard growth medium (5 mM glucose, 0.04 mre myo-inositol) the PIase activity was not changed, indicating that the reduced PIase activity was-not due to the direct effect of slucose. When IPLs from BRCP were analvsed bv HPLC and TLC. we observed reduction of the iota1 and newly formed IPLs including thesubstrate of PIase, Phosphatidylinositol bisphosphate (PIP,). Reduced levels of IPLs were associated with a decrease in myo-inositol and an increase in sorbitol. The changes in IPL metabolism were reversed by adding either free myo-inositol or AL1576, an aldose reductase inhibitor (ARI), to the high-glucose medium. However, the addition of myo-inositol to the growth medium with a standard concentration of glucose only caused a marked increase in phosphatidylinositol, but not in PIP or PIP,, while the supplement of AL1576 in the standard medium did not cause any changes in IPL formation. These findings suggest that the alteration in IPL metabolism in BRCP may be related to insufficient myo-inositol or activated sorbitol pathway under high-glucose conditions. Further explanation of the role of the altered hydrolysis of PIP, triggered by PIase may provide clues to understanding of the mechanism of decreased pericyte viability in the presence of high glucose concentrations. Key words: phosphoinositidase ; inositol phospholipid ; diabetic retinopathy ; pericvte: aldose reductase inhibitor.

1. Introduction

Selective loss of pericytes is one of the early pathological changes in diabetic retinopathy (Cogan, Toussaint and Kuwabara, 1961). The accumulation of sorbitol, triggered by aldose reductase, may be responsible for the degeneration of pericytes (Kador, Akagi, Terubayashi and Kinoshita, 1987). The alteration of inositol phospholipid (IPL) metabolism may relate to the decreasein cell viability of pericytes induced by high glucose (Li, Hu, Stramm, Rockey and Liu, 1987; Li, Zhou, &in and Hu, 1987). In previous studies we demonstrated that high concentrations of glucose suppress the mitotic rate and cell birth rate of cultured bovine retinal capillary pericytes (BRCP) (Li, Shen, Khatami and Rockey, 1984). High concentrations of glucoseinhibit myo-inositol transport and result in a decreasedmyo-inositol content in cultured pericytes (Li, Chan, Khatami and Rockey, 1986; Li, Lou, Chan, Khatami and Rockey, 1988). Myo-inositol is a precursor of IPLs, whose metabolism is responsible for a number of signal transduction processes(Nishizuka, 1984; Berridge and Irvine, 1984). Phosphoinositidase (PIase, a synonym of phosphoinositide-specific phospholipase C) cleaves the phosphodiester bond of phosphatidylinositol 4.5 bisphosphate (PIP,) to produce two second messengers,inositol trisphosphate (IP,) * To whom

correspondence

0014-4835/89/010099+08

should $03.00/O

be addressed 0 1989 Academic

Press Limited +-2

loo

WEIYE

LI

ET

AL

and diacylglycerol (DAG). Since PIase regulates the production of these important second messengers, we examined the effect of increased glucose concentrations on PIase activity. We recently reported the synergistic activation of BRCP proliferation in vitro by IP, and DAG (Li et al., 1987). We also found that high concentrations of glucose alter the formation of both IPLs and inositol phosphate esters in an organoid culture of retinal microvessels (Li, Zhou, &in and Hu, 1987). The purpose of the present study is to determine whether glucose has similar effects on inositol lipid metabolism in pericytes in vitro. Activation of the sorbitol pathway by high glucose in cultured BRCP has also been demonstrated (Li, Khatami and Rockey, 1985). Sorbinil, an aldose reduetase inhibitor (ARI), reduces the elevat,ed sorbitol content of BRCP grown in high-glucose media. However, sorbinil does not reverse the suppressed collagen synthesis in BRCP grown in high-glucose media (Li, Khatami and Rockey, 1985). High concentrations of glucose inhibit myo-inositol uptake, and this inhibition is partially reversed by sorbinil (Li, Chan, Khatami and Rockey, 1986). Whether ARIs influence inositol phospholipid metabolism in retinal pericytes in vitro is investigated herein. In this study, we observed the attenuation of PIase activity, levels of newly formed IPLs and total IPL in pericytes grown in high-glucose media. This reduction could be largely reversed by either adding myo-inositol or an ARI in the growth medium.

2. Materials

and Methods

Primary cultures of bovine retinal capillary pericytes (BRCP) were established as previously reported (Li, Shen, Khatami and Rockey, 1984). Subcultured pericytes (after five passages) were seeded at a density of 2.5 x lo4 cells cmm2 in loo-mm plastic (Corning) Petri dishes and were initially incubated with a standard medium (Dulbecco Modified Minimum Essential medium containing 20% fetal calf serum, 100 U ml-’ penicillin, 100 pug ml-’ streptomycin, and 5 mM glucose). The cells were synchronized with a serum-deficient medium (Li, Shen, Khatami and Rockey, 1984). The synchronized cells were randomly grouped and incubated with different media (as indicated in the tables). Constant molarity of the media was ensured by adding mannitol (Li, Chan, Khatami and Rockey, 1986). Cells in different groups were labelled by myo-[3H]inositol (20 ,&i ml-‘, Amersham, Arlington, Ill. U.S.A.) for 60 hr. The labelled cell layers were washed and harvested. The cell suspension was deproteinized with trichloroacetic acid (final concentration, 5% w/v) and centrifuged for analysis of TPLs. The pellet was extracted with chloroform-methanol (2 : 1) for 1 hr with shaking at 37’C, and then 2 M KC1 was added. The layers were separated by centrifugation. The aqueous layer was aspirated and the remaining phase washed twice with chloroform-methanol-10 mM NaHCO, (3 :48 :47). The organic phase was dried under N, gas. The extracted lipids from an individual dish were redissolved and dried in successively smaller volumes of chloroformmethanol-H,0 (20: 9: 1) until finally dissolved in 100 pl. The lipids in this solution were separated by high-performance liquid chromatography (HPLC). To separate the lipids of the BRCP, a 250 x 4 mm Bio-Sil HP-10 adsorption column (BioRad, Richmond, CA) was utilized. Two reciprocating piston pumps (Bio-Rad Model 1330) controlled by a solvent programmer (Bio-Rad GPS) were used. The solvent system consisted of 97.5% CH,CN (in water) in Reservoir A and 85% CH,CN (in water) in Reservoir B. A linear gradient from A to B was established in 15 min. In this solvent system, a small amount of water was included to help desorb the more polar phospholipids (Nissen and Kreysel, 1983). The flow rate of solvent was 1 ml min-‘. The wavelength was set at 220 nm based on spectral scans for the IPLs. Detector sensitivity was adjusted to 016 absorbance units full scale. Each individual peak was identified by phospholipid standards and quantified by integration. Fractions (05 min) were collected in tubes. A lo+1 aliquot from each tube was

St~PPRESXED

INOSITOL

PHOSPHOLIPII~

.METAl$()LISM

101

transferred to a scintillation vial to detect radioactivity. Fractions containing radioactive TPLs were combined and dried with N,. The IPLs were co-chromatographed with unlabelled phosphatidylinositol (PI), phosphatidylinositol 4-phosphate (PIP) and phosphatidylinositol 4,Sbisphosphate (PIP,) standards by one-dimensional silica gel (type H, Analtech, Neward. DE) thin layer chromatography (TLC) using a modified Abdel-Latif technique (Abdel-Latif. Akhtar and Hawthorne, 1977). The solvent system was butanol-acetic acid-water (6: 1: 1). Lipids were visualized by exposure to I, vapor, and individual IPL spots were scraped from the plates and counted for radioactivity in a liquid scintillation spectrophotometer. In a separate experiment, the synchronized cells grown in different concentrations of glucose were harvested for the determination of PIase activity and cellular myo-inositol content. myo-Inositol determination with BAD-dependent myo-inositol dehydrogenase was carried out by the method of Weissbach (Weissbach, 1984). Briefly, 61 ml of sodium pyrophosphate buffer (0.1 M, pH 90), 0.1 ml of B-&AD (5 mM). 66 ml of doubly distilled H,O and 61 ml of deproteinized sample were mixed in a 1-ml cuvette. and the optical density was read at 339 nm to obtain E,. Four min after the addition of 61 ml of myo-inositol dehydrogenase solution, the optical density at 339 nm was again determined (E,). Based on AE. myo-inositol content was calculated from the calibration curve of standard myo$msitol solutions. PIase preparation was carried out by the method of Hof’mann (Hofmann and Majerus, 1982). Pericytes from individual dishes were harvested and frozen in 20 ITIM potassium phosphate (pH 6.4) containing 100 mM NaCI, 4 mM EGTA, 65 mM phenylmethylsulfonyl fluoride. and 2 mM di-isopropyl fluorophosphate (DFP), allowed to thaw and homogenized with a tissue homogenizer. The homogenate was adjusted to contain 2 mM additional DFP and then centrifuged at 165OOg for I hr at 4Y’. The supernatant pH was adjusted to 6.3 with 1 M potassium phosphate. For the assay of PIase activity. both labelled and unlabelled PIP, were analysed by TLC (type H. see above) to ensure the substrate specificity. Reaction mixtures contained ZiOpM unlabelled PTP,. 2 pu(‘i ml i [3H]PJ P, (Kew England Nuclear), 1 mg ml-’ of sodium deoxycholate. 1 mM (‘a(‘l,, 0.5 mg ml’ of bovine serum albumin. 100 rnM KaCl. 50 rnM Hepes. pH 7.0. and 100,~l of the enzyme preparation in a total volume of 200~1. Blank controls for PIase assay were run with fractions of the boiled enzyme preparation. After 30 min at 37”C, the reaction was stopped with 1 ml of chloroform-methanol-concentratetl HCI (100: 100:0:6) followed by 63 ml of 1 N HCl containing 5 mM EGTA. After extraction. a 400-,uI portion of the aqueous phase was removed for liquid scintillation counting, and the radioactivity was corrected for the cellular DNA content (Li, Chan, Khatami and Rockey, 1986). Standard phosphatidylethanolamine, phosphatidylserine, PI. PIP and PIP, were purchased from Sigma (St Louis. ,liIO. U.S.A.). AL1576. an aldose rrductase inhibitor. was the gift of Drs Lou and York in Alcon Lab. (Fort Worth. TX). The data were analysed by the Student’s two-tailed t test. 3. Results The effect of glucose on PIase activity is shown in Table 1. When the glucose concentration in the medium was increased to 15. or 30 mM, PIase activity was reduced to 82% or 55X, respectively (P < 601). However, when glucose (5, 15, 30 mM) was added directly to the enzyme extract prior to the assay, the PIase activity was essentially unchanged (Table I). Chloroquine, an inhibitor of phospholipase A and phospholipase C, was used as a control group (Matsuzawa and Hostetler, 1980). Under the standard conditions (5 mM glucose, 604 mM myo-inositol) the free cellular myo-inositol content was 28.7f3.2 nmol ,ug-’ DNA (n = 5). When glucose concentration was increased to 30 mM (604 rnM myo-inositol) the cellular myoinositol content was decreased by 43 % (16.4 k 2.5 nmol pg-’ DNA, n = 5, P < @Oi) of the control. Spectral scans for the three IPLs showed all three wavelength maxima (h,,,) are between 219.5 and 220.5 nm. At a wavelength of 220 nm the molar extinction coefficients of these three lipids differed only slightly. Therefore, the peak of mixed

WEIYE

LI

ET

TABLE

ilL.

I

Effect of glucose on Plase

activity

Glucose in medium (mM)

PIase activity (cpm ,ug-’ DNA)*

15 30

1866+ 148 1537t k 168 1025tk265

5 15 30 925 650

1907+ 178 1913k 116 2011+149 1876+ 178 15521 i: 142 1176jk156

5

Assay conditions Control Glucose

(nm)

Chloroquine

(mM)

of BRCP

Results are means fs.D. from six determinat’ions. * The range of ,ag DNA per assay varied from l.O- to 1.5 ,~g. t Differed significantly from PIase with 5 mM glucose, P < 601 t Differed significantly from control, P < 601.

0.16

In

0 1

2

4

6

8

IO Time

0

2

4

6

8

IO

(min)

FIG. 1. HPLC separation of pericyte phospholipids. Separation was achieved at room temperature on a Bio-Sil HP-10 column (Bio-Rad). Solvent A (97.5% CH,N) and Solvent B (85% CHsCN) were used to create a linear gradient in 15 min. The flow rate was 1 ml mini. The left tracing (A) is the profile of lipids (---) and radioactive lipids (---) of BRCP grown in 5 m&r of glucose. The right tracing (B) is the profile of lipids (-) and radioactive lipids (---) of BRCP in 30 rnM of glucose. In, Injection ; 1, neutral lipids; U, unidentified; 2, inositol lipids; 3, phosphatidylserine ; 4, phosphatidylethanolamine.

SVPPRESSED

ISOSITGL

PHOSPHOLTPID TABLE

Ejj’ects of glucose. myo-inositol Medium

condition

(mM)

G

Ml

AL1576

3: -

O-04 604

@(K, 040

30 30 5 5

290 O-04 O-04 200

0.00 610 0.10 @OO

Newly

METAHOLISM

103

II

am? AR1 on IPLs of BRCP formed

PI

IPLs

(cpm x 10-3/10’ PIP

cells) PIP, ~

6577f343 2419+f620 6007t+653 4038*$+233 6250+431 7866*+813

1872+279 814*+ 112 eoaat * 457 2116t+215 1655+ 162 1821*69

.-

1108i101 489’f 121 827*tf59 841*t+ 129 1311k287 1170+ 101

G, Glucose ; MI, myo-inositol ; IPL, inositol phospholipid ; PI, phosphatidylinositol ; PIP. phosphatidylinositol 4-monophosphate; PIP,, phosphatidylinositol 4,5-bisphosphate. Results are means +s.D. from six determinations. * Differed significantly from the effect of G with 5 mM (004 mre MI). P < 601. t Differed significantly from the effect of G with 30 miw (604 mM MI). P < 001.

FIG. 2. One-dimensional TLC separation of inositol phospholipids of pericytes. Separation was achieved by a silica gel H plate (20 x 20 cm). Numbers in the figure refer to the following compounds : 1, PI; 2, PIP; 3, PIP,.

IPLs may be considered as an indicator of total IPLs from individual samples. IPLs were separated from other phospholipids by HPLC (Fig. 1A). However, the three IPLs co-migrated and were eluted into a single peak (Fig. 1 A). BRCP grown in a high concentration of glucose (30 mM) had a 41% loss of endogenous total IPLs (Fig. 1 B). When this high-glucose medium was supplemented with a high concentration of myoinositol (2 mM) the total concentration of IPLs was virtually restored (data not shown). When the high-glucose medium was supplemented with AL1576 (01 mnx) the total amount of IPLs was significantly increased to 81% (P < 601) of that at standard conditions (data not shown). These findings correlated with the results of the TLC studies (Table II), in which the radioactivity of pooled IPLs under high-glucose conditions from TLC was decreased to 39% of the control (Table II). The further separation of IPLs was achieved by one-dimensional TLC (Fig. 2). When BRCP were incubated in the standard medium the percentage distribution of radioactivity in PI,

104

WEIYE

LI

ET

AL.

PIP and PIP, was 69: 19 : 12 (Table II). When BRCP were grown in a medium with 30 mM glucose there was a significant loss of radioactivity from PI, PIP and PIP, (Table II). The addition of myo-inositol (2 mM) to the high-glucose medium led to restoration of the radioactivity levels of PI and PIP to their standard levels, while PIP, was partially but not completely restored (Table II). The addition of AL1576 to the high-glucose medium caused the significant increase in the radioactivity levels from PI, PIP and PIP, (Table II). In contrast, when AL1576 was added to the standard medium no significant changes were observed in the radioactivity from PI, PIP and PIP, (Table II).

4. Discussion The PIase activity was decreased in BRCP grown in high-glucose media. A dosedependent inhibitory effect of increasing glucose concentration was observed. When glucose ranging from 5- to 30 mM was directly added to the enzyme assay solution, however, PIase activity was unchanged. These findings indicate that the inhibitory effect is not due to glucose per se, and may be related to the metabolism of pericytes under high-glucose conditions. PIase is a PIP,-specific phosphodiesterase (Berridge, 1987). Phospholipid accumulation in tissues caused by the inhibition of PIase has been reported (Yamamoto et al., 1976; Matsuzawa and Hostetler, 1980), indicating the key function of this enzyme in catabolism of phosphoinositides. During the cascade of cell signal transduction, the hydrolysis of PIP, generates second messengers. The reduction of PIase activity in high-glucose incubated BRCP may lead to the inhibition of PIP, hydrolysis. It is interesting to note that the expression of the ras oncogene, whose expression is an analog of GTP binding protein, increases the ability of receptors to activate the PIase (Fleisehman, Chahwala and Cantley, 1986). Certain metabolites of pericytes under high-glucose conditions may play an opposing role to inhibit the PIase. Three species of inositol phospholipids (IPLs) were identified by a combination of HPLC and one-dimensional TLC modified from the method of Abdel-Latif (AbdelLatif, Akhtar and Hawthorne, 1977). Since the IPLs were first isolated from other lipids by HPLC, the subsequent TLC requires only a single plate to run multiple onedimensional samples and hence is more economical and efficient. In agreement with our previous results from the study of retinal microvessels, the major component of IPLs is PI (Li, Zhou, &in and Hu, 1987). A high concentration of glucose (30 mM) in growth medium appears to significantly affect both total IPLs and individual IPL formation (including PIP,) in pericytes. Glucose inhibits myoinositol uptake by BRCP in vitro (Li, Chan, Khatami and Rockey, 1986), which may in turn lead to a decrease in cellular myo-inositol content (Li, Lou, Ghan, Khatami & Rockey, 1988). A sufficient pool of free myo-inositol may be essential for resynthesis of IPLs to balance their degradation. This is supported by the fact that all three IPLs suppressed by high glucose were restimulated simply by adding free myo-inositol, although PIP, was not completely restored. However, in BRCP grown in a standard concentration of glucose the addition of high myo-inositol stimulated PI but not PIP or PIP,. This finding suggests that a high concentration of myo-inositol may facilitate its own incorporation into PI by PI synthetase. In addition to a sufficient myoinositol level, the pool of PIP, also depends upon the kinases responsible for the stepwise phosphorylation of PI to PIP, and the phosphomonoesterases which convert PIP, back to PI.

SVPPRESRED

INOSITOL

PHOSPHOLIPID

High

METABOLISM

yucose-1

I

Polyol

Noncompetitive inhibition of Ml transport

I

Decrease

in cellular

Altered inositol metabolism Decrease of inositol

lipid

3. The

possible

relationship

-I Ml ------I

1

I Decrease DG Altered

Reduced

The possible relationship reduced BRCP proliferation Fro. vitro.

bathway

in cellular level phospholipids I

i Decrease in IP3 I Altered Ca” mobilization

A+

between

105

1

in

pH

BRCP between high in vitro

high

glucose

and

glucose the

and the reduced

SRC!P

proliferation

in

AL1576 (@l mM), a potent inhibitor of aldose reductase, also significantly increased IPL formation in BRCP grown in high glucose concentrations. In contrast, in a standard concentration of glucose, AL1576 itself did not show any effect on IPL formation. These findings indicate that the effect, of AL1576 on IPL formation by BRCP is related to the activated sorbitol pathway induced by high glucose. This view is also supported by the finding that under high-glucose conditions sorbinil significantly suppresses sorbitol content in BRCP (Li, Khatami and Rockey, 1985), and partially reverses the inhibited myo-inositol uptake by BRCP (Li, Chan. Khatami and Rockey, 1986). Both short- and long-term changes of inositol lipid metabolism and their subsequent signal cascades may account for the suppressed viability of pericytes in vitro (Kaibuchi et al., 1986; Moore, Todd, Hesketh and Metcalfe, 1986; Li, Shen, Khatami and Rockey, 1984). The reversal of the glucose-induced suppression of IPL metabolism by either myo-inositol or an AR1 indicates that these compounds may be used as in vitro therapy for the treatment of ‘diabetic pericytes’. These findings may add new impetus to studies of the biochemical link bebween an activated sorbitol pathway and altered IPL metabolism (Fig. 3). ACKNOWLEDGMENTS USPHS grant 1 ROl EYO6563-01 and by research grants from the Chinese Natural Science Foundation, Alcon Laboratories Inc., TX, and the Macula Foundation, NY. The authors thank MS Wang Yan for the preparation of the manuscript. This

work

was

supported

by

106

WEIPE

LI ET AL.

REFERENCES Ahdel-Latif. A. A., Akhtar, R. A. and Hawthorne, J. N. (1977). Acetylcholine increases the breakdown of triphosphoinositide of rabbit iris muscle prelabelled with (32P) phosphate. Biochem. J. 162, 61-73. Herridge, M. J. (1987). Inositol trisphosphate and diacylglycerol: two interacting second messengers. Aw. Rev. Biochem. 56, 159-93. Berridge, M. J. and Irvine, R. F. (1984). Inositol trisphosphate, a novel second messenger in cellular signal transduction. Nature (London) 312, 3 15-2 1. Cogan, D. C., Toussaint. D. and Kuwabara, T. (1961). Retinal vascular patterns. IV. Diabetic retinopathy. Arch. Ophthal~mol. 66, 36G-78. Fleischman, I,. F., Chahwala, S. B. and Cantley, L. (1986). Ras-transformed cells: altered levels of phosphatidylinositol 45bisphosphate and catabolites. Science 231, 407-10. Hofmann, S. L. and Majerus, P. W. (1982). Identification of properties of two distinct phosphatidylinositol-specific phospholipase C enzymes from sheep vesicular. J. Biol. Chem 257, 6461-g. Kador, P. F., Akagi, Y., Terubayashi, H. and Kinoshita, J. H. (1987). Prevention of pericyte ghost formation in retinal capillaries of galactose-fed dogs by aldose reductase inhibitors. International Workshop on ARI, Dec. 1987. Kaibuchi. K., Tsuda. T., Kikuchi, A., Tanimoto, T., Yamashita, T. and Takai, Y. (1986). Possible involvement of protein kinase C and calcium ion in growth factor-induced expression of c-myc oncogene in Swiss 3T3 fibroblasts. J. Biol. Chem. 261, 1187-92. Li, W., Chan, L., Khatami, M. and Rockey, J. H. (1986). Non-competitive inhibition of myoinositol transport in cultured bovine retinal capillary pericytes by glucose and reversal by sorbinil. Biochem. Biophys. Acta 857, 198-208. Li, W., Hu, T., Stramm, L., Rockey, J. H. and Liu, S. (1987). Synergistic activation of retinal capillary pericyte prohferation in culture by inositol trisphosphate and diacylglycerol. Exp. Eye Res. 44, 29-35. Li, W., Khatami, M. and Rockey, J. H. (1985). The effect of glucose and aldose reductase inhibitor on the sorbitol content and collagen synthesis of bovine retinal capillary perieytes in culture. Exp. Eye Res. 40, 43944. Li, W., Lou, M., Chan, L. S., Khatami, M. and Rockey, J. H. (1988). Retinal pericyte myoinositol synthesis and myo-inositol reduction by high glucose. Invest. Ophthalmol. Vis Sri. (in press) Li, W., Shen, S., Khatami, M. and Rockey, J. H. (1984). Stimulation of retinal capillary pericyte protein and collagen synthesis in culture by high-glucose concentration. Diabetes 33, 785-9. Li, W.. Zhou, Q., &in, M. and Hu, T. (1987). Reduction of inositol trisphosphate in retinal microvessels by glucose and restimulation by myo-inositol. Ezp. Eye Res. 45, 517-24. Matsuzawa, Y. and Hostetler, K. Y. (1980). Inhibition of lysosomal phospholipase A and phospholipase C by chloroquine and 4,4-bis (diethylaminoethoxy) 2,b-diethyldiphenylethane. J. Biol. Chem. 255, 5190-4. Moore, J. P., Todd. J. A., Hesketh, T. R. and Metcalfe, J. C. (1986). c-Fos and c-myc gene activation, ionic signals and DNA synthesis in thymocytes. J. Biol. Chem. 261, 8158-62. Nshizuka, Y. (1984). The role of protein kinase C in cell surface signal transduction and tumour promotion. Nature (London) 308, 693-8. Xissen, H. P. and Kreysel, H. W. (1983). Analysis of phospholipids in human semen by highperformance liquid chromatography. J. Chromatogr. 276, 29-35. Weissbach, A. (1984). My o -’mositol. In Methods of Enzymatic Analysis. (Ed. Bergmeyer, H. U.). Pp. 36670. Verlag Chemie. Weinheim. Yamamoto, A., Madachi, S., Matsuzawa, Y., Kitani, T., Hiraoka, A. and Seki, K. (1976). Studies on drug-induced lipidosis. VII. Effects of bis-beta-diethyl-aminoethylether of hexestrol, chloroquine, homochlorocyclizine, prenylamine and diazolcholesterol on the lipid composition of rat liver and kidney. Lipids 11, 616-22.