,3". SteroMBIochem. Molec. Biol. Vol. 42, No. 8, pp. 855--861, 1992

0960-0760/92 $5.00 + 0.00 Copyright © 1992 Pergamon I~ess Lid

Printed in Great Britain. All fights reserved

POLYUNSATURATED FATTY ACIDS DECREASE THE APPARENT AFFINITY OF VITAMIN D METABOLITES FOR HUMAN VITAMIN D-BINDING PROTEIN R. BOUILLON,* D. Z. XIANO,R. C o ~ r r s and H. V^~ BAELEN Laboratorium voor Experimentele Geneeskundeen Endocrinologie (LEGENDO), Onderwijs en Navorsing, Gasthuisberg, 3000 Leuven, Belgium

(Received 21 August 1991; receivedfor publication 4 June 1992)

Smmnry--The affinity of purified human vitamin D-binding protein from serum (DBP) for 25-hydroxyvitamin D3 (25-OHD3) and kt,25-dihydroxyvitamin D3 [1,25-(OH)2D3] was measured in the presence of free fatty acids (FFA), cholesterol, prostagiandins and several drugs. Mono- and polyunsaturated fatty acids markedly decreased the affinity of both 25-OHD3 and 1,25-(OH)2D3 for DBP, whereas saturated fatty acids (stearic and arachidic acid), cholesterol, cholesterol esters, retinol, retinoic acid and prostaglandins (A~ and El) did not affect the apparent affinity. Several chemicals known to decrease the binding of thyroxine to its plasma-binding protein did not affect the affinity of DBP. The apparent affinity of DBP for both 25-OHD3 and 1,25-(OH)2D3 decreased 2.4- to 4.6-fold in the presence of 36/tM of linoleic or arachidonic acid, respectively. Only a molar ratio of FFA: DBP higher than 10,000 was able to decrease the binding of 25-OHD3 to DBP by 20%. Much smaller ratio's of FFA: DBP (25 for arachidonic and 45 for oleic acid), however, decreased the binding of 1,25-(OH)2D3 to DBP. These latter ratio's are well within the physiological range. The addition of human albumin in a physiological albumin:DBP molar ratio did not impair the inhibitory effect of finoleic acid on the binding of [3H]25-OHD3 to DBP. The binding and bioavailability of vitamin D metabolites thus might be altered by mono- and polyunsaturated but not by saturated fatty acids.

INTRODUCTION The transport of vitamin D metabolites in the plasma of vertebrates is mediated by a specific protein, vitamin D-binding protein (DBP) or group-specific component (Gc)[1-4]. DBP has also been reported to bind G-actin [5], endotoxin[6] and fatty acids[7]. Ena et a/.[8] reported that human DBP purified from serum, bound fatty acids (0.4 tool per mol of protein), whereas it contained only about 0.04mol of combined vitamin D metabolites [8]. Several drugs interfere with the binding of thyroid hormones to thyroxine-binding globulin (TBG) and free fatty acids (FFA) can also compete with the binding of steroid or thyroid hormones to their plasma-binding prorein [9-11]. Polyunsaturated but not saturated fatty acids were able to decrease the binding of thyroxine to TBG [9] and similar observations were made for the interaction of estradiol with sex hormone-binding globulin [10, 1I]. Cairo and Ena [12] similarly observed that arachidonic *To whom correspondenceshould be addressed. 855

but not palmitic acid decreased the binding of 25-hydroxyvitamin D3 (25-OHD3) to human DBP [12]. We therefore evaluated the effects of FFA and prostaglandins on the binding of 25-OHD3 and l~,25-dihydroxyvitamin D3 [I,25(OH)2D3] to purified human DBP. We found that polyunsaturated but not saturated fatty acids or prostaglandins markedly decreased the affinity of vitamin D metabolites for DBP. EXPERIMENTAL PROCEDURES Materials

26,27-Methyl-[3H]25-OHD3 and 26,27-methyl[3H]l,25-(OH)2D3 (sp. act. 170 and 160-180 Ci/ mmol, respectively) were purchased from the Radiochemical Centre (Amersham, England). 25-OHD 3 and 1,25-(OH)2D 3 were purchased from Duphar (Weesp, The Netherlands). Human DBP was isolated by affinity chromatography as described previously [13]. The fatty acids, cholesterol esters, prostaglandins and retinoic acid were obtained from Sigma (St Louis, MO, U.S.A.) except for stearic, oleic and erucic acid (Merck, Darmstadt, Germany)

856

R.

bJILLON

and linoleic acid (Aldrich, Beerse, Belgium). Retinol was from Fluka (Buchs, Switzerland). Measurement of afinity of DBP for vitamin D metabolites The affinity of DBP for 25-OHD, and 1,25(OH&D, was measured at 0°C in Sodaglass tubes (Vel, Leuven, Belgium). The final volume of the incubation mixture (3.1 ml) consisted of:

(1) either ‘H-labelled

25OHD, or 1,25(OH),D, (24 and 40 pM, respectively) in 40 ~1 of ethanol; (2) increasing amounts of 25-OHD, or 1,25(OH)zD3 added in 15 ~1 of ethanol; 50~1 of ethanol with or without fatty (3) acids or prostaglandins (final concentration 2.2 to 143 PM); (4) human DBP diluted in Tris-HCl buffer (10 mM Tris, 0.154 mM NaCl, pH 7.4). The dilution which enabled a binding of about 50% of the tracer in the absence of stable vitamin D metabolites was selected for each experiment [about 0.4 nM and 0.2 PM for experiments with 25-OHD3 and 1,25-(OH),D, , respectively]. Thereafter, 0.9 ml was taken for measurement of total counts and two aliquots of 0.9 ml were mixed with 0.5 ml of cold dextran-coated (Dextran T70; Pharmacia, Uppsala, Sweden) charcoal (Norit A, acid washed, ICN, Cleveland, OH, U.S.A.) for 30 or 15 min [for experiments with 25-0HD3 and 1,25-(OH)2D3, respectively]. After centrifugation, 0.9 ml of supernatant was counted in a fi-scintillation counter. The affinity of DBP was calculated according to Scatchard [14] using a computer program designed to optimize the regression line by substracting non-specific binding according to Rosenthal [ 151. The effect of increasing amounts of fatty acids, cholesterol esters, vitamin A metabolites and prostaglandins was evaluated by their incubation with human DBP and labelled vitamin D metabolites using the same procedures as in the alBnity measurements. In addition the effect of linoleic acid on 25-OHD3 binding to DBP was studied in the presence of human albumin. E$ct

of fatty acids on 25-0HD3 assay

To evaluate the effect of FFA on the measurement of 25-0HD3 in plasma, linoleic or stearic acid was added to serum either before or after the usual extraction procedure (1 ml of ethyl acetate-cyclohexane, 1: 1, mixed with 0.1 ml of

et

al.

serum and 0.2 ml of water) for 25-0HD3 assay using a competitive protein binding assay [16]. RESULTS

The affinity constant of purified human DBP for 25-OHD, and 1,25-(OH),D, at 0°C and pH 7.4 was 4.9 + 1.4 x lO’(mean + SD; n = 16) and 1.5 f 0.20 x 10’ M-’ (n = 6), respectively. The binding of [3H]25-OHD3 to DBP decreased by increasing amounts of mono- or polyunsaturated fatty acids, whereas unsaturated fatty acids did not influence the binding. Cholesterol, cholesterol oleate, a-phosphatidyl choline and two prostaglandins (A, and E,) were without significant effects (Fig. 1). Retinol and retinoic acid did not significantly inhibit the binding of [31-Ij25-OHD3in concentrations up to 1O-6 M. Among the unsaturated fatty acids, the inhibitory potency was linoleic acid (18 : 2) > oleic acid (18 : 1) > arachidonic acid (20 : 4) > linolenic acid (18 : 3) > erucic acid (22: 1). The degree of unsaturation, therefore, did not markedly influence the inhibitory potency. The inhibitory effect of linoleic acid on the binding of [‘H]25-OHD, to DBP could be blocked by a large molar excess of human albumin but not under circumstances of a physiological albumin:DBP ratio (Fig. 2). The binding of [3H]1,25-(0H)2D3 was also decreased by addition of increasing concentrations of mono- or polyunsaturated fatty acids whereas unsaturated fatty acids, cholesterol or cholesterol ester, phosphatidyl choline, dilinoleoylglycerol and prostaglandins were virtually without effect (Fig. 3). Arachidonic acid at 26 PM final concentration displaced 50% of [‘HI 1,25-(OH&D3 whereas that concentration of oleic, linoleic, linolenic and erucic acid displaced about 38, 34, 27 and 18%, respectively (Fig. 3). The molar ratio of unsaturated FFA: DBP necessary to decrease the binding of 25-OHD3 for DBP by 20% varied between 1.5 and 3.6 x lo4 for linolenic and arachidonic acid, respectively (Fig. 1). However, much smaller molar ratio’s were able to decrease the binding of 1,25-(OH&D, to DBP. Indeed, a molar excess of 25 (arachidonic acid), 45 (oleic acid) or 84 (linolenic acid) significantly decreased the binding (- 20%) of 1,25-(OH)2D3 to human DBP (Fig. 3). The apparent affinity of DBP for 25-0HD3 and 1,25-(OH),D, was studied in more detail by a complete dose-response curve of vitamin D

FFA and DBP affinity for vitamin D

857

O......

too

.

,,,,

"'. ",\\

a"

~h,~

~

a

•

" ,,', ', ,,

a

',,~', .,,,

60

":,

,.,~

\~

~, ',~, ',',

=

•.

I t

",

',

60

"~

i

},, ~40,

,

40

,

', i

~-" 20

~,,'k....--_

2O

\ t

§

'b I

t0 ~ FFA or' p r o s t a g l a n d i n t0 "~

I

I

10-4 concentration

lo-e lipid

(M}

I

1o-S

Io-4

o r drug c o n c e n t r a t i o n

(M)

Fig. 1. Binding of 25-OHD a to human DBP. Left panel: influence of saturated ( ) and unsaturated ( - - - ) fatty acids or prostal0andins (.--) on the binding of [3H]25-OHD3 to al~nity-purified human DBP. ( - - A - - ) arachidic acid (20:0); (--VI--) stearic acid (18:0); ( - - d k - - ) arachidonic acid (20:4); ( - - O - - ) linolenlc acid (18:3); ( - - O - - ) linoleic acid (18:2); (----) oleic acid (18:1); ( - - I I - - ) erucic acid (22:1); (" "O" ") prostaglandin A,; (. "O" ") prostaglandin E n. Right panel: influence of cholesterol and other lipids on the binding of [SH]25-OHD3 to human DBP. The effect of linoleic acid is given for comparison as in the left panel. ( - - A - - ) cholesterol; (--dk--) cholesterol oleate; ("D" ") l-~-phosphatidylcholine; (. "U") d-~,-phosphatidylcholine; ( - - 0 - - ) 1,3-dilineoylglycerol; ( - - O - - ) linoleic acid. For simplicity mean + 1 SD is given only for iinoleic acid ( - - O - - ) . i / .q

.-..~100 I~ ~ E

80

~

60

'-

40

~,~'

20

0

."./

.

'

,

.

, ,,,,I

i

i

,

, ,,,li

10 - 6

i

i

I

I 0-5

llnolelc acld c o n c e n ~ a U o n (M) FiB. 2. Influence of albumin on the interfere~e of [aH]25-OHD3 bindin8 to human DBP (0.4riM or 21 ng/ml) by linoleic acid. Human serum albmain was cither absent (O) or R_d_d,~dm a phygiological albumin:DBP ratio (1/~g human albumin/ml, S ) or in large molar excess (albumin 250pg/ml V;

500~g/ml A; 1000~g/nd II). SBMB 42/8--F

858

R. BOUILLONet 120

al.

120

I t - " ' " "

IO0

'0 C :3

J.

t,

8(3 r-

-

- . . ~ . ,, ,,

5 >

so

,,,

\

-,

¢.

t I

t

40

I I

40

• \ \ &

W

2O ",&

I 10-15

I0"t5

-!

•

I t0-4

I t0-5

10 "e

FFA or p r o s t a g l B n d i N concentration

(M)

lipid

I 10 .-4

or drug concentration

(N)

Fig. 3. Binding of 1,25-(OH)2D 3 to human DBP. Lefipanei: influenceof saturated and unsaturated fatty acids and prostagiandins on the binding of [aHJl,25-(OH)2 Da to affinity-purified human DBP. R/fht panel: influence of cholesterol and lipids on the binding of [JH]I,25-(OH)2D3 to human DBP. For meaning of symbols see Fig. 1. For simplicity mean + 1 SD is only given for aracbidonic acid (--A--, right panel).

metabolites in the presence or absence of the two most active polyunsaturated and one saturated fatty acid [Figs 4(A and B)]. The apparent

affinity of DBP for 25-OHD3 decreased 2.4- to 2

4.6-fold in the presence of linoleic or arachidonic acid (36/~M), respectively, whereas the affinity remained unchanged in the presence of similar concentrations of stearic acid [Fig. 4(A)]. ,2

o

i

o

-o

o o

!15 ' o

o

g

o° I

0.5~.

o .....

1 ......

L .....

2 3 4 5 bound 25-hydroxyvitamln D3 (N x 10 t2) 1

___L .....

t 2 3 4 bound 25-hydroxyvitaBin D3 04 x t0t2)

2 3 4 5 bound 25-hydroxyvttalln D3 (14 X 10121 I

1.5

1.5

B

==o.51-

o

o\

0.5~

"O

O eO . . . . . . .

-4- . . . . . . . . .

0.5 1 1.5 0.5 I 1.5 0.5 I 1.5 bound t. 25-dthydroxyvitaatn Ds bound t. 25-dthydeoxyvttoatn Da bound 1, 25-dthydroxyvito|tn 133

(. x 1o*)

(. x IOe)

(. x t0.)

Fig. 4. Scatchard plots of the influence of fatty acids on the binding of 25-OHDa (A) and 1,25-(OH)2D3 (13) to affinity-purified human DBP in the presence or abunce of ummturated or saturated fatty acids. Left pare/J: with ( 0 ) or without (O) linoleic acid (35.7pM); ~ pane/s: with (O) or without (O) arachidonic acid (35.7 pM) and r//fht/grab: with (O) or without (O) steazic acid (35.7/JM). The n o n - ~ f i c binding Cmdicated by the bottom duhed line) was subtracted from each point according to Rosenthal [15].

FFA and DBP at~nity for vitamin D

Similarly, the apparent affinity of DBP for 1,25-(OH)2D3 decreased 2.5- to 4.3-fold in the presence of linoleic and arachidonic acid but only minimally when stearic acid was added in the same concentration (36/~M) [Fig. 4(B)]. Several other compounds known to interfere with the binding of thyroxine to TBG (diphenylhydantoin, 1,3-dilinoleoylgiycerol cholesterol, heparin, mitotane, sodium salicylate, diazepam, fenelofenac and ipodate) were unable to displace [31-I]vitamin D metabolites from DBP (data not shown). Addition of stearic or linoleic acid to serum samples, before or after extraction, did not interfere with the subsequent assay of 25-OHD3 by competitive binding assay (mean of 4 samples 27.2 + 8 ng/ml without added FFA vs 27.8 + 7.7 or 29.1 + 6.7 ng/ml after addition of 500/~M of stearic or linoleic acid, respectively). DISCUSSION

FFAs can bind to several plasma p r o t e i n s but albumin is the main serum protein responsible for FFA transport. Since DBP belongs, together with ~-fetoprotein, to the albumin gene family[17] DBP might also be able to bind FFA. However, FFAs bind albumin mainly through interaction with its third domain and the homologous domain of DBP is markedly shorter due to partial loss of the corresponding exon[17]. Previous observations by Williams et a/. [7] and Ena et al. [8] indicated that FFA could bind to DBP. The present data clearly confirm and extend these observations. Indeed mono- or polyunsaturated fatty acids were capable of interfering with the binding of 25-OHD3 and 1,25-(OH)2D3 to human DBP. The stereospecificity of this interaction could be demonstrated by the lack of effect of saturated fatty acids (stearic and arachidic), cholesterol and cholesterol esters. The degree of unsaturation was, however, not crucial as the effect on 25-OHD3 binding was greater for linoleic (18:2) than for oleic (18:1) or linolenic acid (I 8: 3). Moreover, arachidonic acid (20: 4) greatly decreased the binding of vitamin D metaboHtes to DBP but its further cyclic unsaturated metabolites (prostaglandin Ai and E~) did not influence the ligand DBP interaction. Unsaturated FFA esterified to glycerol or cholesterol, however, largely lost their effect on interfering vitamin D-DBP binding (Figs 1 and 3). Moreover, retinol and retinoic acid did not

859

markedly affect the binding of 25-OHD3 to DBP. Could the interference of mono- or polyunsaturated fatty acids be of physiological significance? The total concentration of FFA in normal serum is 0.3-0.9 mM compared to a normal DBP concentration of about 6/tM [17, 18]. The normal concentration of unsaturated FFA with the greatest effect on DBP affinity is about 37% (linoleic acid) and 17% (oleic acid) of the total FFA concentration [19]. The normal molar excess of all fatty acids over DBP is, therefore, about 1000 but the molar ratio of individual unsaturated fatty acids: DBP varies between 2 and 400. In vitro, a molar ratio of several unsaturated fatty acids to DBP between 25 and 84 decreased the binding of [3I-I]l,25-(OH)2D3 by more than 20% (Fig. 3) and such molar concentration ratios are thus within the normal range. The amount of fatty acids bound to purified DBP is, indeed, rather small as the molar ratio of total FFA: DBP was