- Email: [email protected]
Protein and carbohydrate selection respond to changes in dietary saturated fatty acids but not to changes in essential fatty acids
Llfe Sciences, Vol. 47, pp. 67-76 Printed in the U.S.A.
Pergamon Press
PROTEIN AND CARBOHYDRATE SELECTION RESPOND TO CHANGES IN DIETARY SATURATED FATTY ACIDS BUT NOT TO CHANGES IN ESSENTIAL FATTY ACIDS
Cad D. McGee and Carol E. Greenwood Department of Nutritional Sciences Faculty of Medicine University of Toronto Toronto, Canada M5S lAB (Received in final
form A p r t l 27, 1990)
We previously reported differences in protein and carbohydrate selection patterns in post-weanling rats fed beef tallow or soybean oll-besed diets. Two experiments were designed to determine the characteristic of the dietary fat which mediates the selection behavior. For each experiment, dietary fat was 20% (w/w) of diets and fatty acid profiles were obtained by blending fat sources. Rats were randomly assigned to diets (24% protein, 40% carbohydrate) which varied only in fatty acid composition. After 2 weeks, rats selected from 2 diets with the fat composition previously fed, but varying in their protein and carbohydrate composition (55% protein, 4% carbohydrate and 5% protein, 61% carbohydrate). Experiment 1 was designed to test the effect of relative (~6:~3 ratios of I and 20) and absolute (15% or 4% =6, 0.7% or 0.2% ~3) differences in essential fatty acids on macronutrient selection patterns. Differences in dietary essential fatty acids had no effect on energy intake or the proportion of energy consumed as protein and carbohydrate. Experiment 2 examined the effect of differences in the level of saturated fat (310% diet (w/w)) on protein and carbohydrate selection. Animals selecting from diets with higher levels of saturated fat consumed more energy as protein and less as carbohydrate than rats selecting from diets with lower levels of saturated fat 0o<0.0001), Regression analysis was used to examine the relationship between percent protein or carbohydrate energy and classes of dietary fat. The strongest relationship existed between percent dietary saturated fat and percent protein or carbohydrate energy (p<0.0001). Polyunsaturated:saturated fat ratio was also weakly associated with percent protein and carbohydrate energy (p <0.05). Polyunsaturated, monounsaturated, =6 and ~3 fatty acids were not significantly related to percent protein or carbohydrate energy. These results indicated that protein and carbohydrate selection patterns are altered in response to qualitatively different dietary fatty acids, and that the amount of saturated fat in the diet is the important characteristic of dietary fat mediating the behavioral alteration. There has long been interest in the effect of variations in nutrient intake on behavior. Although the earliest studies were concerned with the occurrence of behavioral alterations in nutrient deficient states (1,2), current interest lies in the examination of normal variations in dietary intake (i.e. non-deficient states) on behavior. In general, this research has focussed on the effects of protein and carbohydrate intake on behaviors such as food 0024-3205/90 $3.00 +.00 Copyright (c) 1990 Pergamon Press plc
68
Diet
Saturated
Fat Alters
Food S e l e c t l o n
Vol.
4 7 , No.
1, 1990
intake, sleep states and depression (3-7). There is less information regarding the effect of dietary fat on behavior. However, we have shown that cognitive function, thermal regulation, pain sensitivity "rod food intake behavior respond to qualitative differences in dietary fat (8-12). Alterations in food intake behavior were exhibited by differences in the proportion of energy consumed as protein and carbohydrate (11,12). That is, rats fed a lard or beef tallow-based diet selected more protein and less carbohydrate than animals fed a soybean oil-based diet. While these studies provide evidence of an association between qualitative changes in dietary fat and alteration of behavior, the important characteristic of dietary fat which mediates the effect has not been identified. The beef tallow and soybean oil diets used in our food selection studies are dissimilar in many respects. Although differences with respect to the absolute or relative amounts of the essential fatty acids, =6 and ~3 fatty acids, are most striking, amounts of saturated, polyunsaturated and monounsaturated fatty acids also vary greatly. Therefore, because these experiments used only two diets that varied in their fatty acid composition, differences in the dietary intake of any of these fatty acid classes could explain the alterations in macronutriant selection patterns. The objective of these studies, then, was to determine the characteristic of dietary fat which mediates protein and carbohydrate selection patterns. Materials and Methods
Animals and Diets: Male Wistar rats (Charles River, St. Constant, Quebec) were housed individually in wire-mash cages in a temperature-controlled room (24+0.6°C) with a 12-hour light cycle. Initial body weights were 60 - 90 g. All diets were nutritionally adequate purified, granular mixtures containing 5% non-nutritive fiber, 2.5% vitamin mix (13), 5.1% mineral mix and 0.25% L-mathionine (14). Stabilization diets contained 24% protein and 40% carbohydrate. Selection diets were 55% protein, 4% carbohydrate and 5% protein, 61% carbohydrate. A detailed description of diets has been previously published (12). Protein content was verified with the micro-kjeldhal technique (15). Diets were isoenergetic with 20% (w/w) fat. The desired dietary fat profiles were obtained by blending oils as described in Table IA and lB. Fatty acid composition of mixed diets was determined by gas chromatography. Fatty acid profiles are provided in Table IIA and liB. Experiments were conducted using a blinded protocol. For each study, rats were randomly assigned to 20% (w/w) fat diets containing 24% protein and 40% carbohydrate. Following a 2-week stabilization period, rats were challenged with a selection paradigm for an additional 2 weeks. Rats selected from 2 diets with the same fat source fed during the stabilization period, but differing in protein and carbohydrate composition (55% protein, 4% carbohydrate and 5% protein, 61% carbohydrate). Food intake was recorded at the beginning of the light cycle on alternate days throughout the selection period. Protein and carbohydrate consumption were obtained from food intake data. For both macronutrients, the Atwater value of 4 kcal/g was used to calculate protein and carbohydrate energy. Results are expressed as the proportion of total energy consumed as protein or carbohydrate. Energy density of diets was 4.48 kcal/g. Statistical Analysis: Statistical analysis was conducted with SAS 6.03 (SAS Institute, Inc., Cary, NC) for the microcomputer. Effects of diet fat treatment and time on protein and carbohydrate selection were determined by repeated measures analysis of variance followed by Student Newman-Keul's test for multiple comparisons (16). Regression analysis was conducted using linear regression procedures (16).
Vol. 47, No. 1, 1990
Dte¢ S a e u r a e e d Fa¢ A l e e r s Food S e l e c e t o n
69
TABLE IA Type and Amount of Fats Blendedfor Experiment 1 Diets g/kg diet TAL'
Diets
SBO2
(a) 7% ~ , 7% 0)3
CO) 4% u6, 4% 0)3 (c) 15% u6, 0.7% u)3 (d) 4% u6. 0.2% 0)3
.
123 80 162
SFFOs
LO'
48 17 120 36
152 60 2
TABLE IB Type and Amount of Fats Bonded for Experiment2 Diets g/kg diet Diets %SFA
TAL
SBO
SFOs
lard
10 7.2 6.8 6 4 3
180
10 15 40 75 200 10
10 15 40
170
120
Olive
100
25
50
140
' Beef tallow, Canada Packers. Ltd., Toronto, ON 2 Soybean o,, Hain Pure Food, Inc. s Safflower oU, Haln Pure Food, Inc. 4 Linseedoil, Nu-Life, Inc. 5 Sunfloweroil, Haln Pure Food, Inc. Results
Experiment 1: This study was designed to determine if alterations in the relative or absolute amounts of dietary essential fatty acids, ~6 and 0)3 fatty acids, resulted in differences in macronutrient selection. Four dietary fat profiles were developed to test this hypothesis: (a) 7% ~6, 7% ~3, (b) 4% ~6, 4% 0)3, (c) 15% ~6, 0.7% ~3, and (d) 4% ~6, 0.2% ~3 (Table IIA). All diets contained 20% (w/w) dietary fat. There were no differences in body weights prior to beginning the selection paradigm (183+4, 185+5, 187+4, 185+4 g, mean_.+SEM, diets a - d respectively) or at the end of the selection period (270+9, 285+8, 285+7, 280+6 g, diets a - d respectively). Dietary fat treatment had no affect on total energy intake during either the stabilization period (data not shown) or the 2-week selection period (Table III). In addition, no differences in protein or carbohydrate selection were observed as a result of changes in the relative or absolute amounts of ~6 and ~3 fatty acids when data were examined by 2-day interval (not shown) or when collapsed over the 2-week selection period (Table III). Experiment 2: This experimant was designed to determine if there was a relationship between the level of dietary saturated fat and macronutrient selection. Diets were designed for 2 purposes: (a) to assess macronutrient selection over a broad range of saturated fat intake, and (b) to rule out the concomitant alterations in monounsaturated and polyunsaturated fatty acids as mediators of changes in protein and carbohydrate selection. Fatty acid composition for diets is shown in Table liB.
Diet S a t u r a t e d F a t A l t e r s Food S e l e c t l o n
70
Vol. 47, No. i ,
1990
TABLE IIA Fatty Acid Composition of Expedment 1 Diets' % of toad diet (w/w) Fatty Acid
16:0 16:1 18:0 18:1 18:2 18:3 Saturates Monounsaturates Polyunsaturates u6 fatty acids ~3 fatty acids w6:~03
Diets 7% w6 7% u3
4% aJ6 4% ~3
15% w6 0.7% {~3
4% =6 0.2% u3
1.22 ND= 0.65 3.72 6.78 7.62 1.87 3.72 14.40 6.78 7.62
3.62 0.39 2.21 6.37 3.65 3.79 5,63 6.78 7.44 3.65 3.79
1.78 0.52
4.77 0.55
0.49
?-87
7.08
0.89
0.98
2.50 14.50 0.70 2.26 2.52 15.20 14.50 0.70 20.71
3.98
0.20 7.64 7.63 4.08 3.86 0.26 19.30
TABLE lib Fatty Acid Composition of Experiment 2 Diets' Fatty Acid
16:0 16:1 18:0 18:1 18:2 16:3 Saturates Monounsaturates Po4yunsaturates u6 fatty acids u3 fatty acids aJ6:~3
Diets (%SFA) 10%
7.3%
6.8%
6%
4%
3%
5.72 0.62 4.10 7.36 2.02 0.16 9.62 7.98 2.18 2.02 0.16 12.62
4.78 0.46 2.52 7.58 4.44 0.22 7.30 8.04 4.66 4.44 0.22 20.18
4.28 0.40 2.46 6.18 6.26 0.38 6.76 6.58 6.64 6.26 0.38 16.47
3.74 0.28 2.30 7.04 5.96 0.68 6.04 7.32 6.64 5.96 0.68 8.76
2.30 ND 1.62 2.90 11.70 1.46 3.92 2.90 13.16 11.70 1.46 8.01
3.14 0.28 0.10 13.90 2.36 0.20 3.24 14.18 2.56 2,36 0,20 11.80
' Fat was extracted from premixed diets and fatty acid profBedetermined by gas chromatography, NOt detected
:
All groups had similar body weights at the commencement of the selection paradigm. At the conclusion of the selection period, all groups had similar body weights except that rats fed 3% saturated fat weighed significantly less than rats which consumed diets with 10%, 7.3% or 6.8% saturated fat (309+8, 311+6, 314+9, 298+6, 294+8 and 273+11, for 10%, 7.3%, 6.8%, 6%, 4% and 3% saturated fat diets respectively). There were no significant differences in energy intake (i.e. food intake )during the stabilization period (data not shown) or the selection period (Table IV). There was a significant effect of dietary fat treatment on protein and carbohydrate selection when expressed either as absolute intake (g/14 days) or as percent total energy consumed as protein or carbohydrate. In general, rats fed diets with greater amounts of saturated fatty acids selected more protein and less carbohydrate than other groups. However, not all groups were significanlty different from one another based on Student Newman-Keul's multiple comparisons. More specifically, rats fed diets with 10% saturated fat selected significantly more protein and less carbohydrate than other groups (Table IV). On the other hand, rats fed 3% dietary saturated fat selected more protein and less carbohydrate than rats fed diets with 6.8% or
V o l . 47, No. 1, 1990
D i e t S a t u r a t e d Fat A].ters Food S e l e c t i o n
71
10% saturated fat. When selection patterns were examined by 2-day intervals, a similar trend was evident (Rg. 1). Repeated measures analysis revealed 8 significant change in protein and carbohydrate selection patterns dudng the 2-wsek selection period {p < 0.003). However, there was no significant treatment by time interaction (o>0.2) TABLE III In Rats Fed Varying Amotm~ of Essential Fatty Acids'
Total Food Intake and Macronutrlent ~ Diet
Protein % Energy g/14 days
kcal/14 days (e) 7% ~6, 7% ta3
1419_+32
12+1
41+5
(b) 4% c,~6, 4% ~3 (c) 15% (,,,6, 0.7% (,,'3
1411+56 1388+34
15+2 15+1
51+6
(d) 4% ~5, 0.2% U3
1482+48
15+3
Carbohydrate %Energy g114 days 46+1 42+2 42+1 42+3
32+4 83+10
164+6 151+12 147__.+8
160+17
' Fiats were fed mbllization diets (24% protein, 40% cad)ohydrate) for 2 weeks followed by a 2week selection pedod. Rats selected from diets with 5% protein, 61% carbohydrate and 55% protein, 4% carbohydrate conteinlng the same dietary fat as during the stabilization period. All diets were isoenergetlc with 20% (w/w) dietary fat. Values are mean+SEM. N ,, 11 or 12/diet treatment. TABLE IV Total Food Intake and Macronutrlent Selection in Rats Fed Varying Amounts of Saturated Fatty Acids' Diet % SFA
Energy kcal/14 days
10 7.3 6.8 6
Protein % Energy g/14 days 29+2 a 18~'2 b 20~4 ~ 14~'1 ~ 15____.1~=
89+7" 55+5 ~ 64~__10b 41+2 +
4
1248+24 1272+37 1281+43 1234+29 1434+49
3
1307+81
10_+1c
Carbohydrate % Energy g/14 days 126~9 b= 117+14 ~ 136~5 b+
,~____,5m
26+3" 39__~2b 36+4 b 44+1 ~ 43__~1''+
34+5 =
48+1 =
156+8 +
84+9"
I~+Z6 +
1 Rats were fed stabilization diets (24% protein, 40% carbohydrate) for 2 weeks followed by a 2week selection period. Rats selected from diets with 5% protein, 61% carbohydrate and 55% protein, 4% carbohydrate containing the same dietary fat as dudng the stabilization period. AJl diets were Isoenergetlc with 20% (w/w) dietary fat. Values are mean+SEM. N = 10 or 11/treatment group. Values which do not share • common superscdpted letter are significantly different (0<0.0001). SFA = Saturated fatty acids, % total diet (w/w).
35 3O
Protein ,~
, j
x _
Corbohydrote
_z,~"~,
~ 50
----o
25
~,.----'~'--.~.____,___.~...~
20
•~
g- ;o
-o-7"-°~.~t--~
_
,~,----~+~"-
-"
-~.._.-o~.~_~o-----o
5
20
o--o
.~srA
t0
O--O
4~
SPA
6~
SPA
A--A i
2
i
+
4
i
i
6
A
i
i
i
8 10 Selection Doy
.
.
12
.
.
14
.
.
.
2
.
.
4
.
.
°--°E.B~SVA O - - O 7.3~ SF'A 101K S P A
~--A .
.
.
.
6 8 10 Selection Doy
.
.
.
12
14
Fig. 1 Protein and carbohydrate selection by 2-days Intervals. For expedmental protocol see Table IV. Values are mean for 10 or 11 rats/treatment. Effect of time (0<0.003), time by treatment (0>0.2).
72
D i e t S a t u r a t e d Fat A l t e r s Food S e l e c t i o n
+
3O
V o l . 47, No. 1, 1990
+
20
#
~
#
~o
0
i
2
4 6 8 S Dieto~t SFA
10
i
J
#
#
J
4 6 8 10 S Oietc~y MUF4
12
2
14
+
4 6 8 10 I ~ l l i ~ PUfA
12
14
+
++
~ 20 #
#
#
a.
i i i i J 2 4 6 8 10 I Dietary v 6 Forty Acids
0
Oietory PUFA:$FA Rotio
. 12
.
. . . . . , 0.2 0.4 0.6 0.8 1.0 1.2 1.4 ~ Oietory ~ 3 Fotty/~cidl
FIG. 2A >.
~0
°
:o 40 !
+
?
+
_,2o o
2
4
6
8
10
Z D~tory SFA
2
4
6
8
10
12
14
2
Z D~ltory MUFA
4
6
8
10
12
14
Z Oietory PUFA
5o
~ 4o
+
~ 30 ~ 20
+I
T
I
J
+
i÷
.8
_~0
I
2
3
4
Dietory PUrA:SFA Rotio
4
6
D~etory ~
8
10
F ~ t y Acids
12
0.2 0,4 0.6 0.8
1.0 1.2 1.4
I O~ltm~t ~3 Forty Acids
FIg.2B Unear regressions of dietary fat characteristics with percent protein energy (Fig. 2A) and percent carbohydrate energy (Fig. 2B). Dietary fat characteristics are expressed as percent of total diet (w/w). Percent protetn and percent carbohydrate energy were calculated from cumulative intake dudng the 14-day selection pedod. Each point represents the mean+SEM (N ,, 10 or 11/treatment group).
V o l . 4 7 , Ro.
1, 1990
Linear ~
TABLE V Analy~ of Dietary Fat Chamcterist~ with Percent Protein and ~ r a t e
Dietary Fat Characteristic % SFA= % MUFA = % PUFA4
PUFA:SFA % ~8 lelly __wJ~-~__ % c,~ ~
acids
aXB:u3
Dlet Saturated Fat Alters
,S~pe % PRO=
Y-Intercept
r=
73
Energy' p-value
% PRO
% ClIO
-0.59 0.13 0.10
1.72 21.80 19.93
57.97 34.90 37.05
0.42 0.05 0.03
0.0001 0.08 0.15
-0.83
0.95
19.94 19.49
37.04 37.56
0.03 0.05
0.08
0.45
-0.51
t 1.55
45.68
0.05
0.51 44.11 44.09
-2.37 -0.10
% CliOe
Food S e l e c t l o n
2.72 0.11
20.10
36.85
0.07
0.05 0.16 0.05
' Dietaryfat characteristic (% of total diet) was the X-vade~eand percent protein or carbohydrate energy the Y-variable. Percent protein and percent carbohydrate energy were calculated from oumulatlve Intakeduringthe 14-day=election period. Datafrom Individual8nirnalswere L_-~_ _ In the mBlysls (N = 62). The strength of the correfatlon Is ~ as the coefficient of determination l~S~urated fatty acids = Monounsaturatedfatty acids 4 Pofyunsaturatedfatty acids • Percent energy consumed as protein 6 Percent energy consumed as carbohydrate To further examine the relationship between macronutrient selection and dietary fat, the data were analysed by regression analysis. The strongest relationship existed between dietary saturated fat and percent protein or carbohydrate energy (p <0.0001)(Table V, Fig. 2A and 2B). In addition, weaker relationships between the dietary polyunsaturated: saturated fatty acid ratio and percent protein energy and carbohydrate energy were also evident (p<0.05) (Table V, Fig. 2A and 2B). There were no significant relationships between percent protein or carbohydrate energy and dietary polyunsaturated, monounsatureted, =6 or ~3 fatty acids (Table V, Fig. 2A and 2B). Discussion
To our knowledge, these data represent the first report of 8 direct association between dietary saturated fat and behaviour. Our earlier studies reported that rats fed diets which varied in their fatty acid composition (beef tallow or soybean oil) exhibited differences in macronutrient selection when challenged with a serf-selection paradigm (11,12). These experiments, however, were conducted with only 2 diets. Therefore it was not possible to identify the charactedctic of the dietary fat which mediated the differences in macronutrient selection because the diets have notable differences in the relative and absolute amounts of =6 and ~3 fatty acids, as well as in saturated, polyunsaturated and monounsaturated fatty acids. The experiments described here were designed to determine the important characteristic of dietary fat mediating the behavioral response. The initial experiment was designed to test the hypothesis that differences in relative or absolute amounts of ~6 and ~3 fatty acids mediate macronutrient selection. These fatty acids were identified as potential mediators of the behavioural affect because of their role as essential fatty acids, as well as the postulated role of ~3 fatty acids in electrically active tissues (17). Furthermore, differences in these fatty acids were readily apparent when fatty acid profiles for beef tallow and soybean oil diets were compared. Therefore, diets were designed to examine the effects of different absolute amounts of =6 and ~3 fatty acids at each of two =6:=3 fatty acid ratios (Table IIA). Diets were designed to maximize differences in the absolute amounts of each fatty acid class, and in the ratio of ~6:~3 fatty acids. The
74
Die¢ SacuraCed Fat Alters
Food S e l e c t i o n
V o l . 6 7 , No. 1, 1990
results show that differences in the relative or absolute amounts of ~6 and ~3 fatty acids were not associated with changes in the protein and carbohydrate selection by postweanling rats (Table III). The results of the second experiment demonstrated that the amount of dietary saturated fat mediates changes in macronutrient selection (Table IV). Diets were designed to encompass a continuum of dietary saturated fat intake (3.2% to 9.8% (w/w) total diet) so its effect on macronutrient selection could be examined over a wide range. Thus, the range was the maximum attainable given commonly available dietary fat sources. However, saturated, monounsaturated and polyunsaturated fatty acids comprise total fat intake. Therefore a change in the amount of one consttoJent fat results in an alteration in one or both of the remaining fatty acid classes. Consequently, diets which have a range of intake of dietary saturated fat, will have • similar range of intake for both polyunsaturated and monounsaturated fatty acids. To circumvent the effect of the interdependence of dietary fat classes, dietary fat profiles were designed to extricate these factors. For example, 2 diets (3% and 10% saturated diets) with comparable amounts of polyunsaturated fatty acids were designed to have markedly different amounts of saturated fat. If saturated fat were the important variable mediating the selection behavior, these diets should result in different protein and carbohydrate selection, despite comparable amounts of polyunsaturated fat (Table liB). A similar design was used to control for the effects of changes in the amount of monounsaturated fat (6% end 10% saturated fat diets). The results indicate that changes in the amount of monounsaturated or polyunsaturated have little effect on macronutrient selection. Rats consuming diets with 9.8% saturated fat selected significantly more protein and less carbohydrate than rats consuming diets with 3.2% saturated fat despite comparable amounts of polyunsaturated fat (2.2% and 2.6% respectively). Likewise, rats fed diets with 9.8% saturated fat and 8.0% monounsatureted fat exhibited different macronutrient selection profiles than rats fed diets with 6.0 saturated fat and 7.3% monosaturated fat. Furthermore, regression analysis described a significant relationship only with saturated fat or with the ratio of polyunsaturated:saturated fatty acids (Table V, Fig. 2A and 2B). These studies have been unique with respect to both the fat blends used to achieve specific profiles and the amount of dietary fat. Nevertheless, the ranges in protein and carbohydrate selection described in this study are similar to previously reported data despite quantitative and qualitative differences in the dietary fat (3-5). Thus, the type and level of dietary fat do not interfere with the rat's ability to regulate macronutrient intake. Moreover, the effect of dietary saturated fat on macronutrient selection behavior cannot be attributed to extreme deviations in dietary fatty acid profiles or to essential fatty acid deficiency. Diets contained 20% (w/w) dietary fat which accounted for 40% of total energy, an amount similar to the current North Amedcen mixed diet. In addition, dietary fatty acid composition was qualitatively similar to human diets. For example, the fatty acid profiles of the 10% and 4% (w/w) saturated fat diets are comparable to omnivorous and vegetarian diets respectively. The mechanism by which dietary saturated fat alters protein and carbohydrate selection behavior remains to be elucidated. Changes in metabolism, however, would be expected due to differences in protein and carbohydrate selection. Others have described differences in post-absorptive substrete levels between rats fed high carbohydrate and high protein diets (17). Differences in the insulin response would be expected to favor de novo fatty acid synthesis. However, Mullen and Martin (19) failed to observe differences in either plasma glucose or plasma insulin in an experimental paradigm similar to that described here.
Vol.
47, No. 1, 1990
Diet Saturated Fat Alters
Food S e l e c t i o n
75
Qualitative differences in dietary fat also impact upon metabolism. It is now recognized that the physiochemical properties of membranes are dependent upon the characteristics of its constituent fatty acids (20). In addition, research during the past decade has provided evidence that membrane fatty acid composition is sensitive to rnanipula~ons in dietary fatty acids. While this work has primarily examined the effects of differences in dietary fat on membrane fatty acid compos'fdon and activity of membrane-bound proteins in peripheral tissues (21-23), an effect of dietary fat on central nervous tissue has also been documented (24,25). Our previous studies have shown that differences in dietary fatty acids produce changes in synaptosomal phospholipids concurrently with alterations in protein and carbohydrate selection (11). Furthermore, it has recently been shown that, in neuronal tissue, the cardiolipin fraction of mitochondria is particularly sensitive to alterations in the amount of dietary polyunsaturated fat (J.R. Dyer and C.E. Greenwood, submitted for publication). The relationship between alterations in neuronal membrane phospholipid fatty acid composition and behavioral changes requires further investigation to determine whether modification of membrane fatty acid profile mediates the behavioral effect. In summary, these data provide further evidence that qualitative differences in dietary fatty acids are associated with changes in macronutrient selection behavior. The level of dietary saturated fat was found to be strongly associated with protein and carbohydrate selection. Specifically, rats consuming diets with greater amounts of saturated fat selected more protein and less carbohydrate than those fed diets with lesser amounts of saturated fat. Furthermore, macronutrient selection behavior was not related to differences in the relative or absolute amounts of the the dietary essential fatty acids.
Acknowledgements Supported by grants from the Medical Research Council of Canada and the International Life Sciences Institute - Nutrition Foundation. C. McGee was the recipient of a University of Toronto Open Fellowship and an Ontario Graduate Scholarship.
References 1. E. POLLrl-r and C. THOMSON. Nutrition and the Brain, R.J. Wurtman and J.J. Wurtman, ads., 2 261-306, Raven Press, New York (1977). 2. J.R. GALLER. Nutrition and Behavior. J.R. Galler, ed., pp. 63-117, Plenum Press, New York (1984). 3. R.J. WURTMAN. Ann. N.Y. Acad. Sci. 499 179-190 (1987). 4. R.J. WURTMAN, F. Hefti and e. Melamed. Pharmacol. Rev. 32 315-335 (1981). 5. G.H. ANDERSON. Modern Nutrition in H~alth and pisease, 3'~ ed., M.E. Shils and V.R. Young, eds., pp. 557-569, Lea & Febiger, Philadelphia (1988). 6. B. SPRING. Nutrition and th~ Brain, R.J. Wurtman and J.J. Wurtman, eds., 7 1-47, Raven Press, New York (1985). 7. D.A. BENDER. Bibithca Nutr. Dieta 38 82-86 (1986). 8. C.E. GREENWOOD, C.D. MCGEE AND J.R. DYER. Nutrition 5 278-281 (1989). 9. C.E. GREENWOOD and G. WlNOCUR. Behav. Neural Biol. In press. 10. S. YEHUDA, C.E. LEPROHON-GREENWOOD, LM. DIXON and D.V. COSINA. Pharmacol. Biochem. Beh. 24 1775-1777 (1985). 11. S.B. CRANE and C.E. GREENWOOD. Pharrnacol. Biochem. Bah. 27 1-6 (1987). 12. C.D. MCGEE and C.E. GREENWOOD. J. Nutr. 119 1561-1568 (1989). 13. B. MUSTEN, D. PEACE and G.H. ANDERSON. J. Nutr. 104 563-572 (1974). 14. F.W. BERNHART and R.M. TOMARILLI. J. Nutr. 89 495-500 (1969).
76
Diet
Saturated
Fat Alters
Food S e l e c t i o n
V o l . 4 7 , No. I ,
1990
15. ASSOCIATION OF OFFICIAL AGRICULTURAL CHEMISTS. Official Methods of Analvsis, 10th ed., AOAC, Washington, D.C. (1965). 16. R.G.D. STEEL and J.H. TORRIE. Pdncioles and Procedures of Statistics, pp. 67-78, 161-182, McGraw-Hill, New York (1960). 17. E.A. DRA'rZ and A.J. DEESE. Health Effects of Polvunsaturated Fattv Acids, A.P. Simopoulos, R.R. IOfer and R.E. Martin, eds., pp. 319-351, Academic Press, New York (lg~). 18. B. BOIS-JOYEUX, M. CHANF.Z, B. AZZOUT and J. PERET. J. Nutr. 116 446-454 (1986). 19. B. MUU.~N and R.J. MARTIN. The FASEB aourn~ ~ part I) Abstr. #S46 (1989). 2O. K.W.J. WAHLE Proo. Nutr. Soc. 42 273-2S7 (1963). 21. S.M. INNIS and M.T. CLANDININ. Biochem. J. 198 231-234 (1~I). 22. G. HOLMER, C.E. HOY and D. KIRSTEIN. Upids 17 585.-593 (1982). 23. D.M. CONROY, C.D. STIJBBS, J. BEUN, C.L PRYOR and A.D. SMITH. Biochim. Biophys. Acta 861 457-462 (1986). 24. J.R. DYER and C.E. GREENWOOD. Soc Neurosci. 14 Abstr. #406.1 (1988). 25. M. FOOT, T.F. CRUZ and M.T. CLANDININ. Biochem. J. 208 631-640 (1982).
Pergamon Press
PROTEIN AND CARBOHYDRATE SELECTION RESPOND TO CHANGES IN DIETARY SATURATED FATTY ACIDS BUT NOT TO CHANGES IN ESSENTIAL FATTY ACIDS
Cad D. McGee and Carol E. Greenwood Department of Nutritional Sciences Faculty of Medicine University of Toronto Toronto, Canada M5S lAB (Received in final
form A p r t l 27, 1990)
We previously reported differences in protein and carbohydrate selection patterns in post-weanling rats fed beef tallow or soybean oll-besed diets. Two experiments were designed to determine the characteristic of the dietary fat which mediates the selection behavior. For each experiment, dietary fat was 20% (w/w) of diets and fatty acid profiles were obtained by blending fat sources. Rats were randomly assigned to diets (24% protein, 40% carbohydrate) which varied only in fatty acid composition. After 2 weeks, rats selected from 2 diets with the fat composition previously fed, but varying in their protein and carbohydrate composition (55% protein, 4% carbohydrate and 5% protein, 61% carbohydrate). Experiment 1 was designed to test the effect of relative (~6:~3 ratios of I and 20) and absolute (15% or 4% =6, 0.7% or 0.2% ~3) differences in essential fatty acids on macronutrient selection patterns. Differences in dietary essential fatty acids had no effect on energy intake or the proportion of energy consumed as protein and carbohydrate. Experiment 2 examined the effect of differences in the level of saturated fat (310% diet (w/w)) on protein and carbohydrate selection. Animals selecting from diets with higher levels of saturated fat consumed more energy as protein and less as carbohydrate than rats selecting from diets with lower levels of saturated fat 0o<0.0001), Regression analysis was used to examine the relationship between percent protein or carbohydrate energy and classes of dietary fat. The strongest relationship existed between percent dietary saturated fat and percent protein or carbohydrate energy (p<0.0001). Polyunsaturated:saturated fat ratio was also weakly associated with percent protein and carbohydrate energy (p <0.05). Polyunsaturated, monounsaturated, =6 and ~3 fatty acids were not significantly related to percent protein or carbohydrate energy. These results indicated that protein and carbohydrate selection patterns are altered in response to qualitatively different dietary fatty acids, and that the amount of saturated fat in the diet is the important characteristic of dietary fat mediating the behavioral alteration. There has long been interest in the effect of variations in nutrient intake on behavior. Although the earliest studies were concerned with the occurrence of behavioral alterations in nutrient deficient states (1,2), current interest lies in the examination of normal variations in dietary intake (i.e. non-deficient states) on behavior. In general, this research has focussed on the effects of protein and carbohydrate intake on behaviors such as food 0024-3205/90 $3.00 +.00 Copyright (c) 1990 Pergamon Press plc
68
Diet
Saturated
Fat Alters
Food S e l e c t l o n
Vol.
4 7 , No.
1, 1990
intake, sleep states and depression (3-7). There is less information regarding the effect of dietary fat on behavior. However, we have shown that cognitive function, thermal regulation, pain sensitivity "rod food intake behavior respond to qualitative differences in dietary fat (8-12). Alterations in food intake behavior were exhibited by differences in the proportion of energy consumed as protein and carbohydrate (11,12). That is, rats fed a lard or beef tallow-based diet selected more protein and less carbohydrate than animals fed a soybean oil-based diet. While these studies provide evidence of an association between qualitative changes in dietary fat and alteration of behavior, the important characteristic of dietary fat which mediates the effect has not been identified. The beef tallow and soybean oil diets used in our food selection studies are dissimilar in many respects. Although differences with respect to the absolute or relative amounts of the essential fatty acids, =6 and ~3 fatty acids, are most striking, amounts of saturated, polyunsaturated and monounsaturated fatty acids also vary greatly. Therefore, because these experiments used only two diets that varied in their fatty acid composition, differences in the dietary intake of any of these fatty acid classes could explain the alterations in macronutriant selection patterns. The objective of these studies, then, was to determine the characteristic of dietary fat which mediates protein and carbohydrate selection patterns. Materials and Methods
Animals and Diets: Male Wistar rats (Charles River, St. Constant, Quebec) were housed individually in wire-mash cages in a temperature-controlled room (24+0.6°C) with a 12-hour light cycle. Initial body weights were 60 - 90 g. All diets were nutritionally adequate purified, granular mixtures containing 5% non-nutritive fiber, 2.5% vitamin mix (13), 5.1% mineral mix and 0.25% L-mathionine (14). Stabilization diets contained 24% protein and 40% carbohydrate. Selection diets were 55% protein, 4% carbohydrate and 5% protein, 61% carbohydrate. A detailed description of diets has been previously published (12). Protein content was verified with the micro-kjeldhal technique (15). Diets were isoenergetic with 20% (w/w) fat. The desired dietary fat profiles were obtained by blending oils as described in Table IA and lB. Fatty acid composition of mixed diets was determined by gas chromatography. Fatty acid profiles are provided in Table IIA and liB. Experiments were conducted using a blinded protocol. For each study, rats were randomly assigned to 20% (w/w) fat diets containing 24% protein and 40% carbohydrate. Following a 2-week stabilization period, rats were challenged with a selection paradigm for an additional 2 weeks. Rats selected from 2 diets with the same fat source fed during the stabilization period, but differing in protein and carbohydrate composition (55% protein, 4% carbohydrate and 5% protein, 61% carbohydrate). Food intake was recorded at the beginning of the light cycle on alternate days throughout the selection period. Protein and carbohydrate consumption were obtained from food intake data. For both macronutrients, the Atwater value of 4 kcal/g was used to calculate protein and carbohydrate energy. Results are expressed as the proportion of total energy consumed as protein or carbohydrate. Energy density of diets was 4.48 kcal/g. Statistical Analysis: Statistical analysis was conducted with SAS 6.03 (SAS Institute, Inc., Cary, NC) for the microcomputer. Effects of diet fat treatment and time on protein and carbohydrate selection were determined by repeated measures analysis of variance followed by Student Newman-Keul's test for multiple comparisons (16). Regression analysis was conducted using linear regression procedures (16).
Vol. 47, No. 1, 1990
Dte¢ S a e u r a e e d Fa¢ A l e e r s Food S e l e c e t o n
69
TABLE IA Type and Amount of Fats Blendedfor Experiment 1 Diets g/kg diet TAL'
Diets
SBO2
(a) 7% ~ , 7% 0)3
CO) 4% u6, 4% 0)3 (c) 15% u6, 0.7% u)3 (d) 4% u6. 0.2% 0)3
.
123 80 162
SFFOs
LO'
48 17 120 36
152 60 2
TABLE IB Type and Amount of Fats Bonded for Experiment2 Diets g/kg diet Diets %SFA
TAL
SBO
SFOs
lard
10 7.2 6.8 6 4 3
180
10 15 40 75 200 10
10 15 40
170
120
Olive
100
25
50
140
' Beef tallow, Canada Packers. Ltd., Toronto, ON 2 Soybean o,, Hain Pure Food, Inc. s Safflower oU, Haln Pure Food, Inc. 4 Linseedoil, Nu-Life, Inc. 5 Sunfloweroil, Haln Pure Food, Inc. Results
Experiment 1: This study was designed to determine if alterations in the relative or absolute amounts of dietary essential fatty acids, ~6 and 0)3 fatty acids, resulted in differences in macronutrient selection. Four dietary fat profiles were developed to test this hypothesis: (a) 7% ~6, 7% ~3, (b) 4% ~6, 4% 0)3, (c) 15% ~6, 0.7% ~3, and (d) 4% ~6, 0.2% ~3 (Table IIA). All diets contained 20% (w/w) dietary fat. There were no differences in body weights prior to beginning the selection paradigm (183+4, 185+5, 187+4, 185+4 g, mean_.+SEM, diets a - d respectively) or at the end of the selection period (270+9, 285+8, 285+7, 280+6 g, diets a - d respectively). Dietary fat treatment had no affect on total energy intake during either the stabilization period (data not shown) or the 2-week selection period (Table III). In addition, no differences in protein or carbohydrate selection were observed as a result of changes in the relative or absolute amounts of ~6 and ~3 fatty acids when data were examined by 2-day interval (not shown) or when collapsed over the 2-week selection period (Table III). Experiment 2: This experimant was designed to determine if there was a relationship between the level of dietary saturated fat and macronutrient selection. Diets were designed for 2 purposes: (a) to assess macronutrient selection over a broad range of saturated fat intake, and (b) to rule out the concomitant alterations in monounsaturated and polyunsaturated fatty acids as mediators of changes in protein and carbohydrate selection. Fatty acid composition for diets is shown in Table liB.
Diet S a t u r a t e d F a t A l t e r s Food S e l e c t l o n
70
Vol. 47, No. i ,
1990
TABLE IIA Fatty Acid Composition of Expedment 1 Diets' % of toad diet (w/w) Fatty Acid
16:0 16:1 18:0 18:1 18:2 18:3 Saturates Monounsaturates Polyunsaturates u6 fatty acids ~3 fatty acids w6:~03
Diets 7% w6 7% u3
4% aJ6 4% ~3
15% w6 0.7% {~3
4% =6 0.2% u3
1.22 ND= 0.65 3.72 6.78 7.62 1.87 3.72 14.40 6.78 7.62
3.62 0.39 2.21 6.37 3.65 3.79 5,63 6.78 7.44 3.65 3.79
1.78 0.52
4.77 0.55
0.49
?-87
7.08
0.89
0.98
2.50 14.50 0.70 2.26 2.52 15.20 14.50 0.70 20.71
3.98
0.20 7.64 7.63 4.08 3.86 0.26 19.30
TABLE lib Fatty Acid Composition of Experiment 2 Diets' Fatty Acid
16:0 16:1 18:0 18:1 18:2 16:3 Saturates Monounsaturates Po4yunsaturates u6 fatty acids u3 fatty acids aJ6:~3
Diets (%SFA) 10%
7.3%
6.8%
6%
4%
3%
5.72 0.62 4.10 7.36 2.02 0.16 9.62 7.98 2.18 2.02 0.16 12.62
4.78 0.46 2.52 7.58 4.44 0.22 7.30 8.04 4.66 4.44 0.22 20.18
4.28 0.40 2.46 6.18 6.26 0.38 6.76 6.58 6.64 6.26 0.38 16.47
3.74 0.28 2.30 7.04 5.96 0.68 6.04 7.32 6.64 5.96 0.68 8.76
2.30 ND 1.62 2.90 11.70 1.46 3.92 2.90 13.16 11.70 1.46 8.01
3.14 0.28 0.10 13.90 2.36 0.20 3.24 14.18 2.56 2,36 0,20 11.80
' Fat was extracted from premixed diets and fatty acid profBedetermined by gas chromatography, NOt detected
:
All groups had similar body weights at the commencement of the selection paradigm. At the conclusion of the selection period, all groups had similar body weights except that rats fed 3% saturated fat weighed significantly less than rats which consumed diets with 10%, 7.3% or 6.8% saturated fat (309+8, 311+6, 314+9, 298+6, 294+8 and 273+11, for 10%, 7.3%, 6.8%, 6%, 4% and 3% saturated fat diets respectively). There were no significant differences in energy intake (i.e. food intake )during the stabilization period (data not shown) or the selection period (Table IV). There was a significant effect of dietary fat treatment on protein and carbohydrate selection when expressed either as absolute intake (g/14 days) or as percent total energy consumed as protein or carbohydrate. In general, rats fed diets with greater amounts of saturated fatty acids selected more protein and less carbohydrate than other groups. However, not all groups were significanlty different from one another based on Student Newman-Keul's multiple comparisons. More specifically, rats fed diets with 10% saturated fat selected significantly more protein and less carbohydrate than other groups (Table IV). On the other hand, rats fed 3% dietary saturated fat selected more protein and less carbohydrate than rats fed diets with 6.8% or
V o l . 47, No. 1, 1990
D i e t S a t u r a t e d Fat A].ters Food S e l e c t i o n
71
10% saturated fat. When selection patterns were examined by 2-day intervals, a similar trend was evident (Rg. 1). Repeated measures analysis revealed 8 significant change in protein and carbohydrate selection patterns dudng the 2-wsek selection period {p < 0.003). However, there was no significant treatment by time interaction (o>0.2) TABLE III In Rats Fed Varying Amotm~ of Essential Fatty Acids'
Total Food Intake and Macronutrlent ~ Diet
Protein % Energy g/14 days
kcal/14 days (e) 7% ~6, 7% ta3
1419_+32
12+1
41+5
(b) 4% c,~6, 4% ~3 (c) 15% (,,,6, 0.7% (,,'3
1411+56 1388+34
15+2 15+1
51+6
(d) 4% ~5, 0.2% U3
1482+48
15+3
Carbohydrate %Energy g114 days 46+1 42+2 42+1 42+3
32+4 83+10
164+6 151+12 147__.+8
160+17
' Fiats were fed mbllization diets (24% protein, 40% cad)ohydrate) for 2 weeks followed by a 2week selection pedod. Rats selected from diets with 5% protein, 61% carbohydrate and 55% protein, 4% carbohydrate conteinlng the same dietary fat as during the stabilization period. All diets were isoenergetlc with 20% (w/w) dietary fat. Values are mean+SEM. N ,, 11 or 12/diet treatment. TABLE IV Total Food Intake and Macronutrlent Selection in Rats Fed Varying Amounts of Saturated Fatty Acids' Diet % SFA
Energy kcal/14 days
10 7.3 6.8 6
Protein % Energy g/14 days 29+2 a 18~'2 b 20~4 ~ 14~'1 ~ 15____.1~=
89+7" 55+5 ~ 64~__10b 41+2 +
4
1248+24 1272+37 1281+43 1234+29 1434+49
3
1307+81
10_+1c
Carbohydrate % Energy g/14 days 126~9 b= 117+14 ~ 136~5 b+
,~____,5m
26+3" 39__~2b 36+4 b 44+1 ~ 43__~1''+
34+5 =
48+1 =
156+8 +
84+9"
I~+Z6 +
1 Rats were fed stabilization diets (24% protein, 40% carbohydrate) for 2 weeks followed by a 2week selection period. Rats selected from diets with 5% protein, 61% carbohydrate and 55% protein, 4% carbohydrate containing the same dietary fat as dudng the stabilization period. AJl diets were Isoenergetlc with 20% (w/w) dietary fat. Values are mean+SEM. N = 10 or 11/treatment group. Values which do not share • common superscdpted letter are significantly different (0<0.0001). SFA = Saturated fatty acids, % total diet (w/w).
35 3O
Protein ,~
, j
x _
Corbohydrote
_z,~"~,
~ 50
----o
25
~,.----'~'--.~.____,___.~...~
20
•~
g- ;o
-o-7"-°~.~t--~
_
,~,----~+~"-
-"
-~.._.-o~.~_~o-----o
5
20
o--o
.~srA
t0
O--O
4~
SPA
6~
SPA
A--A i
2
i
+
4
i
i
6
A
i
i
i
8 10 Selection Doy
.
.
12
.
.
14
.
.
.
2
.
.
4
.
.
°--°E.B~SVA O - - O 7.3~ SF'A 101K S P A
~--A .
.
.
.
6 8 10 Selection Doy
.
.
.
12
14
Fig. 1 Protein and carbohydrate selection by 2-days Intervals. For expedmental protocol see Table IV. Values are mean for 10 or 11 rats/treatment. Effect of time (0<0.003), time by treatment (0>0.2).
72
D i e t S a t u r a t e d Fat A l t e r s Food S e l e c t i o n
+
3O
V o l . 47, No. 1, 1990
+
20
#
~
#
~o
0
i
2
4 6 8 S Dieto~t SFA
10
i
J
#
#
J
4 6 8 10 S Oietc~y MUF4
12
2
14
+
4 6 8 10 I ~ l l i ~ PUfA
12
14
+
++
~ 20 #
#
#
a.
i i i i J 2 4 6 8 10 I Dietary v 6 Forty Acids
0
Oietory PUFA:$FA Rotio
. 12
.
. . . . . , 0.2 0.4 0.6 0.8 1.0 1.2 1.4 ~ Oietory ~ 3 Fotty/~cidl
FIG. 2A >.
~0
°
:o 40 !
+
?
+
_,2o o
2
4
6
8
10
Z D~tory SFA
2
4
6
8
10
12
14
2
Z D~ltory MUFA
4
6
8
10
12
14
Z Oietory PUFA
5o
~ 4o
+
~ 30 ~ 20
+I
T
I
J
+
i÷
.8
_~0
I
2
3
4
Dietory PUrA:SFA Rotio
4
6
D~etory ~
8
10
F ~ t y Acids
12
0.2 0,4 0.6 0.8
1.0 1.2 1.4
I O~ltm~t ~3 Forty Acids
FIg.2B Unear regressions of dietary fat characteristics with percent protein energy (Fig. 2A) and percent carbohydrate energy (Fig. 2B). Dietary fat characteristics are expressed as percent of total diet (w/w). Percent protetn and percent carbohydrate energy were calculated from cumulative intake dudng the 14-day selection pedod. Each point represents the mean+SEM (N ,, 10 or 11/treatment group).
V o l . 4 7 , Ro.
1, 1990
Linear ~
TABLE V Analy~ of Dietary Fat Chamcterist~ with Percent Protein and ~ r a t e
Dietary Fat Characteristic % SFA= % MUFA = % PUFA4
PUFA:SFA % ~8 lelly __wJ~-~__ % c,~ ~
acids
aXB:u3
Dlet Saturated Fat Alters
,S~pe % PRO=
Y-Intercept
r=
73
Energy' p-value
% PRO
% ClIO
-0.59 0.13 0.10
1.72 21.80 19.93
57.97 34.90 37.05
0.42 0.05 0.03
0.0001 0.08 0.15
-0.83
0.95
19.94 19.49
37.04 37.56
0.03 0.05
0.08
0.45
-0.51
t 1.55
45.68
0.05
0.51 44.11 44.09
-2.37 -0.10
% CliOe
Food S e l e c t l o n
2.72 0.11
20.10
36.85
0.07
0.05 0.16 0.05
' Dietaryfat characteristic (% of total diet) was the X-vade~eand percent protein or carbohydrate energy the Y-variable. Percent protein and percent carbohydrate energy were calculated from oumulatlve Intakeduringthe 14-day=election period. Datafrom Individual8nirnalswere L_-~_ _ In the mBlysls (N = 62). The strength of the correfatlon Is ~ as the coefficient of determination l~S~urated fatty acids = Monounsaturatedfatty acids 4 Pofyunsaturatedfatty acids • Percent energy consumed as protein 6 Percent energy consumed as carbohydrate To further examine the relationship between macronutrient selection and dietary fat, the data were analysed by regression analysis. The strongest relationship existed between dietary saturated fat and percent protein or carbohydrate energy (p <0.0001)(Table V, Fig. 2A and 2B). In addition, weaker relationships between the dietary polyunsaturated: saturated fatty acid ratio and percent protein energy and carbohydrate energy were also evident (p<0.05) (Table V, Fig. 2A and 2B). There were no significant relationships between percent protein or carbohydrate energy and dietary polyunsaturated, monounsatureted, =6 or ~3 fatty acids (Table V, Fig. 2A and 2B). Discussion
To our knowledge, these data represent the first report of 8 direct association between dietary saturated fat and behaviour. Our earlier studies reported that rats fed diets which varied in their fatty acid composition (beef tallow or soybean oil) exhibited differences in macronutrient selection when challenged with a serf-selection paradigm (11,12). These experiments, however, were conducted with only 2 diets. Therefore it was not possible to identify the charactedctic of the dietary fat which mediated the differences in macronutrient selection because the diets have notable differences in the relative and absolute amounts of =6 and ~3 fatty acids, as well as in saturated, polyunsaturated and monounsaturated fatty acids. The experiments described here were designed to determine the important characteristic of dietary fat mediating the behavioral response. The initial experiment was designed to test the hypothesis that differences in relative or absolute amounts of ~6 and ~3 fatty acids mediate macronutrient selection. These fatty acids were identified as potential mediators of the behavioural affect because of their role as essential fatty acids, as well as the postulated role of ~3 fatty acids in electrically active tissues (17). Furthermore, differences in these fatty acids were readily apparent when fatty acid profiles for beef tallow and soybean oil diets were compared. Therefore, diets were designed to examine the effects of different absolute amounts of =6 and ~3 fatty acids at each of two =6:=3 fatty acid ratios (Table IIA). Diets were designed to maximize differences in the absolute amounts of each fatty acid class, and in the ratio of ~6:~3 fatty acids. The
74
Die¢ SacuraCed Fat Alters
Food S e l e c t i o n
V o l . 6 7 , No. 1, 1990
results show that differences in the relative or absolute amounts of ~6 and ~3 fatty acids were not associated with changes in the protein and carbohydrate selection by postweanling rats (Table III). The results of the second experiment demonstrated that the amount of dietary saturated fat mediates changes in macronutrient selection (Table IV). Diets were designed to encompass a continuum of dietary saturated fat intake (3.2% to 9.8% (w/w) total diet) so its effect on macronutrient selection could be examined over a wide range. Thus, the range was the maximum attainable given commonly available dietary fat sources. However, saturated, monounsaturated and polyunsaturated fatty acids comprise total fat intake. Therefore a change in the amount of one consttoJent fat results in an alteration in one or both of the remaining fatty acid classes. Consequently, diets which have a range of intake of dietary saturated fat, will have • similar range of intake for both polyunsaturated and monounsaturated fatty acids. To circumvent the effect of the interdependence of dietary fat classes, dietary fat profiles were designed to extricate these factors. For example, 2 diets (3% and 10% saturated diets) with comparable amounts of polyunsaturated fatty acids were designed to have markedly different amounts of saturated fat. If saturated fat were the important variable mediating the selection behavior, these diets should result in different protein and carbohydrate selection, despite comparable amounts of polyunsaturated fat (Table liB). A similar design was used to control for the effects of changes in the amount of monounsaturated fat (6% end 10% saturated fat diets). The results indicate that changes in the amount of monounsaturated or polyunsaturated have little effect on macronutrient selection. Rats consuming diets with 9.8% saturated fat selected significantly more protein and less carbohydrate than rats consuming diets with 3.2% saturated fat despite comparable amounts of polyunsaturated fat (2.2% and 2.6% respectively). Likewise, rats fed diets with 9.8% saturated fat and 8.0% monounsatureted fat exhibited different macronutrient selection profiles than rats fed diets with 6.0 saturated fat and 7.3% monosaturated fat. Furthermore, regression analysis described a significant relationship only with saturated fat or with the ratio of polyunsaturated:saturated fatty acids (Table V, Fig. 2A and 2B). These studies have been unique with respect to both the fat blends used to achieve specific profiles and the amount of dietary fat. Nevertheless, the ranges in protein and carbohydrate selection described in this study are similar to previously reported data despite quantitative and qualitative differences in the dietary fat (3-5). Thus, the type and level of dietary fat do not interfere with the rat's ability to regulate macronutrient intake. Moreover, the effect of dietary saturated fat on macronutrient selection behavior cannot be attributed to extreme deviations in dietary fatty acid profiles or to essential fatty acid deficiency. Diets contained 20% (w/w) dietary fat which accounted for 40% of total energy, an amount similar to the current North Amedcen mixed diet. In addition, dietary fatty acid composition was qualitatively similar to human diets. For example, the fatty acid profiles of the 10% and 4% (w/w) saturated fat diets are comparable to omnivorous and vegetarian diets respectively. The mechanism by which dietary saturated fat alters protein and carbohydrate selection behavior remains to be elucidated. Changes in metabolism, however, would be expected due to differences in protein and carbohydrate selection. Others have described differences in post-absorptive substrete levels between rats fed high carbohydrate and high protein diets (17). Differences in the insulin response would be expected to favor de novo fatty acid synthesis. However, Mullen and Martin (19) failed to observe differences in either plasma glucose or plasma insulin in an experimental paradigm similar to that described here.
Vol.
47, No. 1, 1990
Diet Saturated Fat Alters
Food S e l e c t i o n
75
Qualitative differences in dietary fat also impact upon metabolism. It is now recognized that the physiochemical properties of membranes are dependent upon the characteristics of its constituent fatty acids (20). In addition, research during the past decade has provided evidence that membrane fatty acid composition is sensitive to rnanipula~ons in dietary fatty acids. While this work has primarily examined the effects of differences in dietary fat on membrane fatty acid compos'fdon and activity of membrane-bound proteins in peripheral tissues (21-23), an effect of dietary fat on central nervous tissue has also been documented (24,25). Our previous studies have shown that differences in dietary fatty acids produce changes in synaptosomal phospholipids concurrently with alterations in protein and carbohydrate selection (11). Furthermore, it has recently been shown that, in neuronal tissue, the cardiolipin fraction of mitochondria is particularly sensitive to alterations in the amount of dietary polyunsaturated fat (J.R. Dyer and C.E. Greenwood, submitted for publication). The relationship between alterations in neuronal membrane phospholipid fatty acid composition and behavioral changes requires further investigation to determine whether modification of membrane fatty acid profile mediates the behavioral effect. In summary, these data provide further evidence that qualitative differences in dietary fatty acids are associated with changes in macronutrient selection behavior. The level of dietary saturated fat was found to be strongly associated with protein and carbohydrate selection. Specifically, rats consuming diets with greater amounts of saturated fat selected more protein and less carbohydrate than those fed diets with lesser amounts of saturated fat. Furthermore, macronutrient selection behavior was not related to differences in the relative or absolute amounts of the the dietary essential fatty acids.
Acknowledgements Supported by grants from the Medical Research Council of Canada and the International Life Sciences Institute - Nutrition Foundation. C. McGee was the recipient of a University of Toronto Open Fellowship and an Ontario Graduate Scholarship.
References 1. E. POLLrl-r and C. THOMSON. Nutrition and the Brain, R.J. Wurtman and J.J. Wurtman, ads., 2 261-306, Raven Press, New York (1977). 2. J.R. GALLER. Nutrition and Behavior. J.R. Galler, ed., pp. 63-117, Plenum Press, New York (1984). 3. R.J. WURTMAN. Ann. N.Y. Acad. Sci. 499 179-190 (1987). 4. R.J. WURTMAN, F. Hefti and e. Melamed. Pharmacol. Rev. 32 315-335 (1981). 5. G.H. ANDERSON. Modern Nutrition in H~alth and pisease, 3'~ ed., M.E. Shils and V.R. Young, eds., pp. 557-569, Lea & Febiger, Philadelphia (1988). 6. B. SPRING. Nutrition and th~ Brain, R.J. Wurtman and J.J. Wurtman, eds., 7 1-47, Raven Press, New York (1985). 7. D.A. BENDER. Bibithca Nutr. Dieta 38 82-86 (1986). 8. C.E. GREENWOOD, C.D. MCGEE AND J.R. DYER. Nutrition 5 278-281 (1989). 9. C.E. GREENWOOD and G. WlNOCUR. Behav. Neural Biol. In press. 10. S. YEHUDA, C.E. LEPROHON-GREENWOOD, LM. DIXON and D.V. COSINA. Pharmacol. Biochem. Beh. 24 1775-1777 (1985). 11. S.B. CRANE and C.E. GREENWOOD. Pharrnacol. Biochem. Bah. 27 1-6 (1987). 12. C.D. MCGEE and C.E. GREENWOOD. J. Nutr. 119 1561-1568 (1989). 13. B. MUSTEN, D. PEACE and G.H. ANDERSON. J. Nutr. 104 563-572 (1974). 14. F.W. BERNHART and R.M. TOMARILLI. J. Nutr. 89 495-500 (1969).
76
Diet
Saturated
Fat Alters
Food S e l e c t i o n
V o l . 4 7 , No. I ,
1990
15. ASSOCIATION OF OFFICIAL AGRICULTURAL CHEMISTS. Official Methods of Analvsis, 10th ed., AOAC, Washington, D.C. (1965). 16. R.G.D. STEEL and J.H. TORRIE. Pdncioles and Procedures of Statistics, pp. 67-78, 161-182, McGraw-Hill, New York (1960). 17. E.A. DRA'rZ and A.J. DEESE. Health Effects of Polvunsaturated Fattv Acids, A.P. Simopoulos, R.R. IOfer and R.E. Martin, eds., pp. 319-351, Academic Press, New York (lg~). 18. B. BOIS-JOYEUX, M. CHANF.Z, B. AZZOUT and J. PERET. J. Nutr. 116 446-454 (1986). 19. B. MUU.~N and R.J. MARTIN. The FASEB aourn~ ~ part I) Abstr. #S46 (1989). 2O. K.W.J. WAHLE Proo. Nutr. Soc. 42 273-2S7 (1963). 21. S.M. INNIS and M.T. CLANDININ. Biochem. J. 198 231-234 (1~I). 22. G. HOLMER, C.E. HOY and D. KIRSTEIN. Upids 17 585.-593 (1982). 23. D.M. CONROY, C.D. STIJBBS, J. BEUN, C.L PRYOR and A.D. SMITH. Biochim. Biophys. Acta 861 457-462 (1986). 24. J.R. DYER and C.E. GREENWOOD. Soc Neurosci. 14 Abstr. #406.1 (1988). 25. M. FOOT, T.F. CRUZ and M.T. CLANDININ. Biochem. J. 208 631-640 (1982).