Download Adverse Effects of Excessive Leucine Intake Depend on Dietary

J Nutr Sci Vitaminol, 59, 45–55, 2013
Adverse Effects of Excessive Leucine Intake Depend on
Dietary Protein Intake: A Transcriptomic Analysis to
Identify Useful Biomarkers
Wataru Imamura1, Ryoji Yoshimura2, Marie Takai2, Junki Yamamura1,
Ryuhei Kanamoto2,3 and Hisanori Kato1,*
1 Corporate Sponsored Research Program “Food for Life,” Organization for Interdisciplinary Research
Projects, The University of Tokyo, Yayoi, Bunkyo-ku, Tokyo 113–8657, Japan
2 Laboratory of Molecular Nutrition, Graduate School of Life and Environmental Sciences,
Kyoto Prefectural University, Shimogamo, Sakyo-ku, Kyoto 606–8522, Japan
3 Division of Food Science and Biotechnology, Graduate School of Agriculture, Kyoto University,
Gokasho Uji, Kyoto 611–0011, Japan
(Received August 27, 2012)
Summary The present study was conducted to identify reliable gene biomarkers for the
adverse effects of excessive leucine (Leu) in Sprague-Dawley rats by DNA microarray. It has
long been known that the adverse effects of excessive amino acid intake depend on dietary
protein levels. Male rats were divided into 12 groups (n56) and fed for 1 wk a diet containing low (6%), moderate (12%) or high (40%) protein. Different levels of Leu (0, 2, 4, and
8%) were added to the diets. Consumption of diets containing more than 4% Leu in 6%
protein resulted in growth retardation and reduced liver weight, whereas the administration
of the same dose of Leu with 12% or 40% protein did not affect them. By a process of systematic data extraction, 6 candidate gene markers were identified. The liver gene expression
data obtained from another experiment with 0, 2, 3, 4, and 8% Leu in a low-protein diet was
used to examine the validity of these biomarker candidates with receiver operating characteristic (ROC) curve analysis. All of AUC values of the biomarker candidates were more than
0.700, suggesting the effectiveness of the marker candidates as the indices of Leu excess.
The cut-off value for the ROC curve of the gene-marker panel, which was obtained by multiple regression analysis of gene markers, indicated that Leu levels higher than 3% have
adverse effects. In conclusion, the gene-marker panel suggested that for male rats dietary
Leu supplementation of 2% is the NOAEL dose in low-protein (6%) diets.
Key Words leucine, transcriptome, biomarker, DNA microarray, dietary protein
underlying the toxic effects of amino acid excess, which
should provide bases for the identification of an upper
limit (UL) intake level of each amino acid.
Studies addressing beneficial or adverse effects of
nutrients and other food components often rely on the
use of biomarkers that reflect biological responses. For
example, the change of Gpx1 (glutathione peroxidase 1)
mRNA expression in liver and blood is an efficient biomarker reflecting selenium deficiency; therefore Gpx1
mRNA is used to determine dietary selenium requirements (5, 6). Since changes of gene expression are usually early biological events manifested by dietary stimulation, mRNA biomarkers (gene biomarkers) are likely
to be effective tools to assess adverse effects caused by
excessive intakes of nutrients including amino acids.
Gene markers can then be used to determine the NOAEL
and UL. Owing to recent advancement of DNA microarray technology, the expression of whole sets of mRNA in
target organs or cells can be simultaneously determined.
Thus, DNA microarray is a very efficient tool for the discovery of marker genes.
Leu is one of branched chain amino acids and it has
been reported to have a role as a signal factor to activate
protein synthesis and inhibit protein degradation in ani-
Amino acids have recently gained popularity as
dietary supplements since the functions of each amino
acid have been revealed, which include the promotion
of health and the enhancement of sport performance
(1). In this trend, the possibility of excessive intakes of
a single amino acid has become an issue to be considered. At least in experimental animals, it is well known
that the excessive intake of a single amino acid leads
to adverse outcomes such as growth inhibition (2–4).
It is therefore important to elucidate the mechanisms
* To whom correspondence should be addressed.
E-mail: [email protected]
Abbreviations: Actb, actin, beta; Ccnb2, cyclin B2; Ccnd1,
cyclin D1; Cdkn1a, cyclin-dependent kinase inhibitor 1A;
Fgf21, fibroblast growth factor 21; Igf1, insulin-like growth
factor 1; Igfals, insulin-like growth factor binding protein,
acid labile subunit; Igfbp1, insulin-like growth factor binding protein 1; Inhbe, inhibin beta E; Leu, leucine; NOAEL, no
observable adverse effect level; qRT-PCR, quantitative reverse
transcription polymerase chain reaction; ROC curve, receiver
operating characteristic curve; SD rat, Sprague-Dawley rat;
Stat5a, signal transducer and activator of transcription 5A;
Stat5b, signal transducer and activator of transcription 5B;
UL, tolerable upper intake level.
45
46
Imamura W et al.
Table 1. Composition of diets.1
LP groups (6% casein)
Leucine (%)
Casein (g)
l-Leucine (g)2
Cornstarch (g)
Cysteine (g)3
Soybean oil (g)
Cellulose (g)
Mineral mix (AIN-93G) (g)
Vitamin mix (AIN-93G) (g)
Total (g)
0 (LP0)
60
0
774.1
0.9
70
50
35
10
2 (LP2)
60
20
754.1
0.9
70
50
35
10
4 (LP4)
60
40
734.1
0.9
70
50
35
10
8 (LP8)
60
80
694.1
0.9
70
50
35
10
1,000
1,000
1,000
1,000
0 (NP0)
2 (NP2)
4 (NP4)
8 (NP8)
NP groups (12% casein)
Leucine (%)
Casein (g)
l-Leucine (g)
Cornstarch (g)
Cysteine (g)
Soybean oil (g)
Cellulose (g)
Mineral mix (AIN-93G) (g)
Vitamin mix (AIN-93G) (g)
Total (g)
120
0
713.2
1.8
70
50
35
10
120
20
693.2
1.8
70
50
35
10
120
40
673.2
1.8
70
50
35
10
120
80
633.2
1.8
70
50
35
10
1,000
1,000
1,000
1,000
0 (HP0)
2 (HP2)
4 (HP4)
8 (HP8)
400
0
429
6
70
50
35
10
1,000
400
20
409
6
70
50
35
10
1,000
400
40
389
6
70
50
35
10
1,000
400
80
349
6
70
50
35
10
1,000
HP groups (40% casein)
Leucine (%)
Casein (g)
l-Leucine (g)
Cornstarch (g)
Cysteine (g)
Soybean oil (g)
Cellulose (g)
Mineral mix (AIN-93G) (g)
Vitamin mix (AIN-93G) (g)
Total (g)
1 The diets are generally based on the AIN-93G formulation.
l-Leucine was generously provided by Ajinomoto Co., Inc., Japan.
3 Cysteine concentrations were adjusted for protein concentration.
2 mal muscles (7, 8). With the expectation of such effects,
Leu, as well as other branched chain amino acids (Ile
and Val), is taken by athletes in relatively high amounts.
Although supplemental intake of amino acids is considered to be safe without any side effects, excessive Leu has
an adverse effect on growth performance or the immune
system in rats, chicks, pigs, and lambs (2, 9–11). However, the mechanism underlying the toxicities remains
to be elucidated. The present study was conducted to
identify gene expression markers reflecting excessive
intake of Leu and to gain insights into the safe and toxic
doses of Leu supplementation.
It has long been known that the occurrence of adverse
effects of excessive single amino acid intake depends on
dietary protein levels (2, 4). Namely, adverse effects of an
excessive amino acid tend to be more prominent under
dietary conditions of lower protein content than those
of sufficient protein content. For these reasons, the estimation of the UL of each single amino acid remains difficult. The object of this study was using DNA microarray analysis to search for effective biomarkers related to
the adverse effects caused by excessive Leu, which would
provide a basis for the estimation of UL.
Methods
Animals and diets.
Experiment 1: Seventy-two 10-wk-old male SpragueDawley rats (Japan SLC, Inc., Japan) were individually
Gene Markers for Leucine Excess
housed in stainless-steel cages and maintained on a
12-h light : dark cycle at 2362˚C, and at 55610%
humidity. Rats were acclimated for 3 or 4 d and provided
ad libitum access to a 20% casein diet based on the AIN93G (12) and water. After this acclimation period, rats
were divided into 12 groups (n56) and fed for 1 wk a
diet containing low (6%: LP), normal (12%: NP) or high
(40%: HP) concentration of casein supplemented with
one of four levels (0, 2, 4, and 8%) of l-leucine (kindly
provided by Ajinomoto, Co., Inc., Japan). The compositions of the experimental diets are shown in Table 1.
After a week of feeding, the rats were anesthetized by
intraperitoneal administration of pentobarbital (50 mg/
mL, 100 mL/100 g BW) after 4 h of fasting, and killed.
Fresh livers were weighed and then portions of the liver
were soaked in RNAlater (Ambion Inc., Austin, TX)
for collecting total RNA. The protocol of experiments
received prior approval by the animal experiment committees of Kyoto Prefectural University and the University of Tokyo.
Experiment 2: Thirty 10-wk-old male Sprague-Dawley rats (Japan SLC, Inc.) were acclimatized to the same
conditions as described above and were assigned to 5
groups (n56). They were then fed the diet containing
6% casein supplemented with five levels of Leu (0, 2,
3, 4, and 8%). Experimental conditions except for the
inclusion of an additional Leu level (3%) were the same
as in Experiment 1.
Isolation of liver RNA and DNA microarray analysis. Liver total RNA was extracted from the homogenized
liver samples using TRIzol Reagent (Invitrogen, USA)
and the amounts and purity were determined with a
NanoDrop ND-1000 spectrophotometer (Thermo Fisher
Scientific, USA). The quality of total RNA was confirmed
by visualization of 18 s and 28 s rRNAs after agarose gel
electrophoresis analysis.
Equal amounts of total RNA were obtained from the
livers of the rats of LP0, LP2, LP4, and LP8 groups. The
same amount of total RNA from five rats (the individual
with largest body weight in each group was omitted in
DNA microarray analysis) of each group was pooled
and subjected to cRNA synthesis. The mRNAs were
reverse-transcribed with T7 oligo (dT) primer and copied to ds cDNAs. Then biotin-labeled cRNAs were synthesized (Affymetrix GeneChip 3′ IVT Expression Kit,
Affymetrix, USA) and fragmented by heating at 94˚C
for 35 min. cRNAs were hybridized to the Affymetrix
GeneChip Rat Genome 230 2.0 array (Affymetrix) at
45˚C for 16 h. Chips were washed and then scanned
using the Affymetrix Scanner 3000 7G. Microarray
Suite 5.0 (MAS5) was used to correct for background,
normalize and generate expression data with Affymetrix
GeneChip Command Console Software (AGCC, Affymetrix). Expression data of the microarray were analyzed
through the use of Ingenuity Pathway Analysis (IPA,
Ingenuity® Systems, www.ingenuity.com). On microarray data analysis, more than 0.5 or less than 20.5 of
the log ratio {log2 (signal value in experimental group/
signal value in control group)} value was considered as
the differentially expressed genes.
47
Transcript filtering for identification of biomarkers. To
filter the genes having the Leu dose-dependency and
high sensitivity for excessive Leu intake, genes exhibiting changes at log ratio of more than 0.5 or less than
20.5 in LP4, and more than 1.0 or less than 21.0 in
LP8, were selected. Then gene sets of commonly detectable molecules in human blood were selected by using a
IPA-Biomarker Filter, which provides ways to select biomarker candidates from the information of the specific
biological characteristics in the IPA-Knowledge Base.
Then gene sets were compared with the probe list of Rat
Tox FX array 1.0 to select the genes related to the generation of biological toxicity and with the liver microarray
data of rats subjected to 1-wk caloric restriction (13) to
exclude the genes changed by reduced food intake. The
relationship among liver gene-marker candidates was
examined by IPA-Knowledge Base.
Quantitative RT-PCR. qRT-PCR was conducted by
using The SYBR Green EX (Takara Bio, Madison, WI,
USA) and a Thermal Cycler DICE Real-Time PCR System
TP800 (Takara, Japan).The primers used for the analysis of respective mRNAs were as follows:
actin, beta (Actb)
Forward: 5′-TTGCTGACAGGATGCAGAAG-3′,
Reverse: 5′-CAGTGAGGCCAGGATAGAGC-3′
insulin-like growth factor binding protein 1 (Igfbp1)
Forward: 5′-GCAGCTTGTCCGGTTCTCA-3′,
Reverse: 5′-CCTCTGTCATCTCTGGGCTCTC-3′
fibroblast growth factor 21 (Fgf21)
Forward: 5′-AGATCAGGGAGGACGGAACA-3′
Reverse: 5′-AGAGGCTTTGACACCCAGGA-3′
inhibin beta E (Inhbe)
Forward: 5′-GCCGATGACATTGTGGCTTA-3′
Reverse: 5′-CCAGCGTCCTCTTCTGACCT-3′
cyclin-dependent kinase inhibitor 1A (Cdkn1a)
Forward: 5′-GGCTGCCCAAGATCTACCTG-3′
Reverse: 5′-AGTGCAAGACAGCGACAAGG-3′
cyclin B2 (Ccnb2)
Forward: 5′-CTGCTGCTTCCTGCCTCTCT-3′
Reverse: 5′-CATGTGCTGCATGACTTCCA-3′
cyclin D1 (Ccnd1)
Forward: 5′-AAGAAGCTGGTCTGGCTTGC-3′
Reverse: 5′-CTTCGCGATGCCACTACTTG-3′
insulin-like growth factor 1 (Igf1)
Forward: 5′-GACGCTCTTCAGTTCGTGTG-3′
Reverse: 5′-GTCTTGGGCATGTCAGTGTG-3′
insulin-like growth factor binding protein 3 (Igfbp3)
Forward: 5′-TCCACATCCCAAACTGTGAC-3′
Reverse: 5′-GCGGTATCTACTGGCTCTGC-3′
insulin-like growth factor binding protein 5 (Igfbp5)
Forward: 5′-TGCACCTGAGATGAGACAGG-3′
Reverse: 5′-TTGTCCACACACCAGCAGAT-3′
insulin-like growth factor binding protein, acid labile
subunit (Igfals)
Forward: 5′-GGCTGGCAACAAACTGACTT-3′
Reverse: 5′-CACGGCTGTAATGAGGTTC-3′
signal transducer and activator of transcription 5A
(Stat5a)
Forward: 5′-AGGCTGTCCATGAATGATCC-3′
Reverse: 5′-CGCCCTAGGTCTCCTAATCC-3′
48
Body Weight Gain (g)
A
Body Weight Gain (g)
B
B
Body Weight Gain (g)
C
Fig. 1. Body weight gain of rats fed the experimental
diets for 1 wk. Body weight gains (final body weight2
initial body weight) are shown. Columns which have
different letters are significantly different (p,0.05).
Data are presented as means6SE (n56). The numbers
following LP, NP, and HP indicate the amount of Leu
supplemented to the diets (LP, 6% casein diet; NP, 12%
casein diet; HP, 40% casein diet).
signal transducer and activator of transcription 5B
(Stat5b)
Forward: 5′-TACATGGACCAGGCTCCTTC-3′
Reverse: 5′-GGGATCCACTGACTGTCCAT-3′
The amount of each mRNA was normalized relative to
that of Actb.
Data analysis and statistics. Multiple regression analyses against body weight gain were performed for the
results of qRT-PCR of the 6 liver biomarkers identified
by DNA microarray and the amounts of Leu intake.
The measurements of liver Igfbp1, Cdkn1a, Ccnb2, Inhbe,
Fgf21, Ccnd1 mRNA as well as the Leu intake index
obtained by the following formula in LP and NP groups
were used as an explanatory variable.
Leu intake index
5Leu content in diet (%)3food intake (g/d)/[casein
content in diet (%)2casein-derived Leu content
in diet (%)]
Leu content of diet
5Sum of added Leu and casein-derived Leu
Multiple regression analysis was performed by ExcelToukei 2010 (Social Survey Research Information Co.,
Ltd., Japan). The Tukey-Kramer method was used for
multiple comparisons to determine significant differences. Multiple comparisons were conducted within
each protein level (LP, NP, and HP). The differences of
means were considered significant at p,0.05. All results
are expressed as the means6SE for each group.
C
Liver Weight / Body Weight (%) Liver Weight / Body Weight (%)
A
Liver Weight / Body Weight (%)
Imamura W et al.
Fig. 2. Liver weight (g)/100 g body weight. Columns
which have different letters are significantly different
(p,0.05). Data are presented as means6SE (n56). The
numbers following LP, NP, and HP indicate the amount
of Leu supplemented to the diets (LP, 6% casein diet; NP,
12% casein diet; HP, 40% casein diet).
Receiver operating characteristic (ROC) curve analysis. ROC curve analysis was used to validate the effectiveness of biomarker candidates using the data obtained
in Experiment 2. More than 0.900 of AUC-ROC value
was considered as high accuracy and 0.900–0.700 as
moderate accuracy for growth retardation (14, 15). The
cut-off point was determined by the Youden index, i.e.
the point with the value of (sensitivity1specificity–1)
being maximal was adopted as the cut-off point of the
ROC curve (16, 17). In other words, the cut-off value
is the point where the distance between the AUC curve
and the putative line with AUC-ROC value of 0.500 was
maximal. ROC curve analysis was performed by ExcelToukei 2010.
Results
Growth performance and liver weight
Mean body weight gain during the experimental
period (final body weight2initial body weight) of each
group is shown in Fig. 1. The liver weight is shown in
Fig. 2. Consumption of a diet containing 6% casein
and more than 4% Leu resulted in significantly lower
body weight gain (vs LP0, p,0.05; Fig. 1A), whereas
the administration of the same dose of Leu with 12%
or 40% protein did not affect body weight gain (Fig. 1B,
C). Significant decreases in total food intake were also
49
Gene Markers for Leucine Excess
Table 2. Changes of liver gene expression of biomarker candidates1 compared with rats fed 6% casein10% Leu diet.
Log ratio2
Gene name
Insulin-like growth factor binding protein 1
Cyclin-dependent kinase inhibitor 1A
Cyclin B2
Inhibin, beta E
Fibroblast growth factor 21
Cyclin D1
Symbol
Igfbp1
Cdkn1a
Ccnb2
Inhbe
Fgf21
Ccnd1
Affimetrix ID
1368160_at
1388674_at
1389566_at
1369238_at
1387643_at
1371643_at
LP2
LP4
LP8
0.1
0.5
20.6
20.5
22.2
20.1
0.5
1.4
21
21.1
21.5
20.9
1.8
2.2
21.5
22.3
24.0
21.3
1 Biomarker candidates were identified by IPA-Biomarker Filter analysis and the comparison of the LP8 dataset with the
liver microarray data of calorie restricted rat (13), and the probe list of Rat ToxFX 1.0 Array. The analytical procedure is
described in detail in the text.
2 Changes of gene expression are shown as log2 ratio, i.e. log2 (signal value of experimental group/signal value of LP0).
Inhbe Fold Change
D
E
F
Ccnd 1 Fold Change
C
Ccnb2 Fold Change
Cdkn1a Fold Change
B
Igfbp1 Fold Change
A
Fgf21 Fold Change
observed in LP4 and LP8 groups as compared with other
groups (data not shown). To examine if the growth
retardation was solely attributable to the reduction in
food intake, a pair-feeding experiment was conducted, in
which the amount of diet consumed by the LP8 group
was given to the LP0 group. As the result, pair-feeding
also resulted in growth retardation, although the body
weight loss of LP8 was significantly greater than that
of the pair-fed group (Supplemental Fig. 1A). In addition, the rate of body weight loss of the group that was
pair-fed a LP8 diet supplemented with high isoleucine
and valine was similar to that of the pair-fed LP0 group
(Supplemental Fig. 1A, LP8IV). The difference between
the pair-fed groups and LP8 group was augmented by
extending the feeding period to 49 d (Supplemental Fig.
1B). A significant decrease of liver weight/body weight
(%) was also observed in the LP8 (p,0.05, vs LP0; Fig.
2A), and the rats of the LP4 group showed a similar
tendency (vs LP0, p50.080). On the other hand, liver
weight was not affected by Leu supplementation in NP
or HP groups (Fig. 2B, C).
The screening of biomarker candidates from liver through
DNA microarray analysis and multiple regression analysis
The total RNA samples obtained from the livers of LP
group rats were subjected to DNA microarray analysis. The expression profile data was used to select possible biomarker genes as described in “Methods.” Briefly,
genes were filtered in light of the relevance to toxicological events, changes of expression with Leu dose dependency and high sensitivity, and expression in human
blood in consideration of future use as biomarkers in
humans. As the result, the transcript filtering identified
6 genes; two were up-regulated (Igfbp1, Cdkn1a) and
four were down-regulated (Ccnb2, Inhbe, Fgf21, Ccnd1)
in Leu dose-dependent manners (Table 2). Relationship
analysis using IPA-Knowledge Base identified close connection between Igfbp1 and Fgf21 and among Cdkn1a,
Ccnb2, and Ccnd1 (Supplemental Fig. 2). The expression
levels of these genes were then examined by qRT-PCR
analyses, the results of which were highly consistent
with that of DNA microarray (Fig. 3) whereas no significant differences were found among NP groups (data are
Fig. 3. Results of qRT-PCR of (A) Igfbp1, (B) Cdkn1a, (C)
Ccnb2, (D) Inhbe, (E) Fgf21, (F) Ccnd1, which were identified as liver gene-marker candidates in this research.
Actb was used as an internal standard for each measurement. Values are the ratio against the average
value of LP0. Columns which have different letters are
significantly different (p,0.05). Data are presented as
means6SE (n55–6). NP0, NP2, NP4, and NP8 are
12% casein diets supplemented with 0, 2, 4, and 8%
Leu.
not shown). As the expression level of liver Igfbp1 and
Fgf21 was drastically affected by excessive Leu intake,
mRNA levels of their related factors were also examined
by qRT-PCR. No significant change was observed in the
expression of Igf1, Igfbp3, Igfbp5, Igfals, Stat5a, or Stat5b
by Leu supplementation (Supplemental Table 1). Functional annotation of altered genes and Fisher’s exact test
indicated that the pathway relating to the metabolism of
Gly, Ser and Thr and the functional group of amino acid
metabolism were most significantly affected in the liver
of LP-fed rats by 8% Leu (Supplemental Table 2).
50
Imamura W et al.
20
LP2
LP4
LP8
NPO
NP2
NP4
NP8
Body Weight Gain (g)
40
LP0
0
-70
-50
-30
-10
-20
10
30
50
70
BP value
B
-40
-60
Fig. 4. Relationship between individual biomarker
panel (BP) and body weight gain. BP was obtained by
multiple regression analysis with use of values of the
expression of 6 genes and Leu intake index described
in the text in LP and NP groups. The variance inflation
factor (VIF) of all explanatory variables was less than
3.00. The contribution ratio is shown in the right bottom of the figure. The numbers following LP, NP, and
HP indicate the amount of Leu supplemented to the
diets (LP, 6% casein diet; NP, 12% casein diet).
Liver Weight / Body Weight (%)
Body Weight Gain (g)
A
R2 = 0.728
60
Fig. 5. Body weight gain (A) and liver weight (g)/100 g
body weight (B) in Experiment 2. Columns with different letters are significantly different (p,0.05). Data are
presented as means6SE (n56). NP0, NP2, NP3, NP4,
and NP8 are 12% casein diets supplemented with 0, 2,
3, 4, and 8% Leu.
A
B
C
D
E
F
Fig. 6. Receiver operating characteristic (ROC) curves against growth retardation for (A) Igfbp1, (B) Cdkn1a, (C) Ccnb2, (D)
Inhbe, (E) Fgf21, and (F) Ccnd1. The expression values of gene-marker candidates identified by liver transcriptome analysis
in Experiment 1 and body weight gain were used for the analysis. Area under the curve (AUC-ROC) values are shown in
the right bottom of each graph. Rats that lost weight during the experimental period were considered to be individuals
with growth retardation when the ROC curves were drawn.
51
Gene Markers for Leucine Excess
A
LP0
60
LP2
Body Weight Gain (g)
40
LP3
LP4
20
LP8
0
-70
-50
-30
-10
10
-20
30
50
70
BP value
-40
R2 = 0.742
-60
B
ROC Curve of Biomarker Panel
Cutoff
Sensitivity (%)
-15.85
Sensitivity (%)
81.8
4
Sensitivity
Specificity (%)
AUC=0.904
1 - Specificity
Ratio of individuals
with BP values
below the cut-off
value in each
group (%)
94.4
LP0
0
LP2
0
LP3
33.3
LP4
50.0
LP8
100.0
Fig. 7. Relationship between the value for the individual biomarker panel (BP) and body weight gain in Experiment 2 (A)
and ROC curve of the biomarker panel against growth retardation (B). BP values were calculated by assigning measurements in Experiment 2 to the multiple regression equation made in Experiment 1. Area under the curve (AUC-ROC) values
are shown in the right bottom of the ROC curve. Rats that lost weight during the experimental period were considered to
be individuals with growth retardation when the ROC curves were drawn. The cut-off point of the ROC curve was determined by the Youden index (16, 17). LP0, LP2, LP3, LP4, and LP8 are 6% casein diets supplemented with 0, 2, 3, 4, and
8% Leu.
To obtain a more effective marker to predict the toxicity of excessive Leu intake, multiple regression analysis using the qPCR measurements of 6 liver genes was
performed. The resulting equation shown below was
named the biomarker panel (BP).
BP521.313Fgf21111.523Ccnb2113.223Ccnd12
17.83Cdkn1a28.843Inhbe210.93Igfbp1–
0.393(Leu intake index)125.1
The contribution ratio (R2) of the values of the equation generated by multiple regression analysis against
body weight gain was 0.728 in Experiment 1 (Fig. 4).
Validation experiment for biomarker candidates
In Experiment 2, growth retardation and decreased
liver weight per body weight were also observed (Fig. 5).
Both were significantly lower in LP3, LP4, and LP8 as
compared with LP0 (p,0.05).
Next, hepatic mRNA levels of the 6 marker candidate genes that were identified by microarray analysis
of Experiment 1 were quantified. Their responses to Leu
supplementation were similar to the result of Experiment 1 and highly dose-dependent (data not shown).
Their validity as biomarkers was examined by receiver
operating characteristic (ROC) curve analysis against
growth retardation caused by excessive Leu intake. As
a result, area under the curve (AUC)-ROC values for all
gene markers were more than 0.700 (Fig. 6).
BP values were calculated by applying the results of
qRT-PCR of 6 gene markers of each rat in Experiment
2 to the equation obtained in Experiment 1. R2 value
between the body weight gain and the BP values was
0.742 (Fig. 7A). The AUC-ROC value of BP was very
high (0.904) (Fig. 7B). The cut-off point based on the
Youden index in the ROC curve of BP was determined to
be 215.854, and all rats in the LP8 group, and 3 and 2
rats out of 6 in LP4 and LP3, respectively, were considered to be positive (BP values were less than the cut-off
value), whereas all BP values of rats fed LP0 and LP2
were negative (Fig. 7B).
Discussion
The results of Experiment 1 and Experiment 2 showed
that excessive Leu intake causes growth retardation and
a decrease in relative liver weight of rats when included
in a relatively low-protein diet, but not a higher-protein
diet. This is consistent with research of others reporting
that 7% Leu in a 4% casein diet causes growth retardation (18), and 5% Leu in a 6% casein diet, but not in
a 34% casein diet, causes growth retardation (2). In
Experiment 2, growth retardation was observed in rats
fed 3% Leu. To our knowledge, this dose is the minimum
52
Imamura W et al.
Leu amount causing growth retardation in rats. So our
results newly suggest that more than 3% Leu can cause
growth retardation in male SD rats fed a 6% casein
diet. Results of the pair-feeding experiment suggest that
growth retardation caused by excessive Leu intake is not
solely attributable to decreased food intake. Moreover,
growth retardation caused by excessive Leu was suppressed by adding comparable amounts of Val and Ile
during pair-feeding. Similarly, other studies reported the
alleviation of growth retardation by adding Val and Ile
to excessive Leu diets in rats (18) and chicks (19). All
together, previous reports and our results suggest that
the relative amount of BCAAs in the diet is the primary
determinant of the adverse effects of excessive Leu. In
addition, decreases in liver weight were observed in rats
fed not less than 3% Leu. Similarly to other amino acids,
those changes caused by excessive Leu intake were only
observed under conditions of relatively low protein
amounts in foods (2). Thus when considering the safe
intake level of Leu, one must also consider the protein
content of the diet.
In the present study, no obvious adverse effects was
seen in NP or HP groups, suggesting that up to 8% Leu
intake is essentially safe under conditions of moderate
to high protein intakes. However, it should be noted that
we have previously observed that higher supplementation of Leu (15%) to a 20% casein diet caused growth
retardation and other adverse effects (20). Thus, Leu
supplementation of around 8% may have a potential risk
of adverse effects under moderate protein intake conditions. Hence, the use of sensitive biomarkers relating to
the adverse effects caused by excessive Leu intake can be
effective for more precise assessment of the safe intake
level of Leu. Aiming at effective and broad screening of
gene biomarkers, the expression changes by Leu under
low protein conditions in which growth response was
highly dependent on Leu dose was applied in this study.
As the result, 6 genes which are known to be regulators
of growth or the cell cycle were selected by a systematic
analysis method.
IGFBP1 (insulin-like growth factor binding protein
1), known as a regulator of growth, binds to IGF-1
and inhibits IGF activity (21). It has been reported that
growth retardation is observed in mice overexpressing
Igfbp1 in liver (22). The result of this research revealed
that the hepatic expression of Igfbp1 was higher in LP
group rats fed excessive Leu, whereas the expression levels of the genes for Igf1, Igfbp3, Igfbp5, and Igfals were
not changed by excess intake of Leu. IPA-core analysis revealed the amino acid metabolism to be the most
significantly affected function in the livers of the LP8
group. Additionally, the aggravation of protein nutrition is known to up-regulate the expression of Igfbp1 in
liver, which is thought to be a cause of growth retardation (23, 24). We have observed in hepatoma cells that
addition of a disproportionate amount of a single amino
acid to the culture media up-regulates Igfbp1 expression (data not shown), suggesting that the amino acid
directly regulates the Igfbp1 gene. On the other hand,
the expression of the IGF-1 gene is also up-regulated by
deficiency of a single amino acid (25). Taken together,
it could be hypothesized that the disturbance of amino
acid metabolism in the liver of Leu excess may up-regulate the expression of Igfbp1, leading to growth retardation. FGF21 is a member of the FGF family, and its
gene is abundantly expressed in liver (26). It has been
reported that Fgf21-transgenic mice show growth retardation with increased expression of liver Igfbp1 (27).
Increased expression of Igfbp1 is also observed in mice
intravenously administrated FGF-21 protein (28). FGF21 is therefore considered to be an important positive
regulator of the expression of Igfbp1 in liver. However,
the present study revealed that the intake of excessive
Leu down-regulated the liver Fgf21 gene. The down-regulation of Fgf21 by excessive Leu intake is contradictory
to previous observations that FGF21 is an activating
factor of Igfbp1 expression and that Fgf21 expression
is up-regulated by fasting. Thus, a specific mechanism
to regulate Fgf21 expression by an amino acid may
exist. Moreover, decreased expression of Fgf21 was also
observed in rats fed the 2% Leu diet during which no
growth retardation was observed. These results suggest
the decreased Fgf21 expression is a response independent of the growth retardation or the increase of Igfbp1
expression. The qRT-PCR results that Igf1, Igfals, Stat5a,
and Stat5b, genes known to be regulated by FGF21 (27),
were not affected by Leu supplementation support this
idea. Regulation of Fgf21 gene expression by carbohydrate metabolism or lipid metabolism in liver (29–31)
has been reported, while information on the regulation
by amino acid metabolism is limited (32).The result of
the gene expression of Fgf21 in this study raises the possibility that protein nutritional status and amino acid
metabolism in liver are novel regulators of Fgf21 activity. Although further study is needed to elucidate the
biological significance of the change of Fgf21 by excessive Leu, the result suggests that Fgf21 is a very sensitive
and specific biomarker for amino acid excess.
Cyclin B and Cyclin D are proteins which have key
functions in the cell cycle (33). It has been reported that
Cyclin B binds to cyclin-dependent kinase 1 (CDK1) and
acts to proceed to M-phase in the cell cycle, and that
Cyclin D binds to CDK4 and activates the cells to transit
from G1-phase to S-phase. The liver gene expression of
Ccnb2 and Ccnd1 was down-regulated in LP4 and LP8,
respectively, suggesting the possibility of inhibition of
the cell division procedure under the situation of excessive Leu intake. Moreover, the gene expression of liver
Cdkn1a (cyclin-dependent kinase inhibitor 1A) showed
significant up-regulation in rats fed the excessive-Leu
diet accompanied by a decrease in the liver weight. It
has been reported that CDKN1A competitively binds
to CDK4 and acts as an inhibitor of Cyclin D (34). This
report is consistent with our hypothesis that excessive
Leu intake can cause negative effects on the stable cell
division procedure. Taken together, the decrease of the
liver weights by high Leu may be at least in part attributable to the regulation of gene expressions of the regulators of the cell division procedure.
Based on the DNA microarray analysis data, several
53
Gene Markers for Leucine Excess
genes related to growth and the cell cycle have been
identified as candidates for biomarkers responding to
the adverse effects caused by excessive Leu intake. Furthermore, multiple regression analysis provided the biomarker panel, which is expected to have the capability
of more accurate assessment for the safe amount of Leu
intake. Using the dataset of Experiment 2, the effectiveness of biomarker candidates including the biomarker
panel were validated with ROC curve analysis. ROC
curve analysis revealed all biomarker candidates have
good accuracy of prediction for growth retardation
caused by excessive Leu. As the value of the AUC-ROC
of the biomarker panel was highest (0.904), the cut-off
point of the curve was deduced by the Youden index.
The cut-off value of the biomarker panel indicated that
Leu levels no more than 2% with 6% protein have no
adverse effects but those higher than 3% have hazardous potential, which is consistent with the results of
body weight gain or liver weight. The result suggests
that dietary Leu supplementation of 2% is the NOAEL
dose, and a higher dose may cause adverse effects in
rats. Further studies including additional doses of Leu
and the different amounts and sources of protein are
needed to obtain a more conclusive value for the safe
level of Leu intake.
The study showed that exhaustive analysis of gene
expression is effective for the discovery of biomarkers. In
addition, using the same animals as in the present study,
we have obtained a gene expression profile of blood cells,
which provided many biomarker candidates (unpublished data). Metabolome analyses have also been carried out and some metabolites in liver and plasma have
been identified as effective biomarkers (unpublished
data). The data also suggested a Leu dose over 2% could
cause adverse effects under 6% protein conditions. The
combination of different omics disciplines, i.e. integrated omics, is likely a promising strategy in addressing
the issue of safety of food.
Conclusion
In conclusion, DNA microarray analysis has identified several genes whose expression in the liver have the
potential as biomarkers for excessive Leu intake. Dietary
Leu supplementation of 2% may be the NOAEL dose in a
6% casein diet for male SD rats.
Acknowledgments
W.I., R.K. and H.K. designed the research; W.I., R.Y.,
M.T. and R.K. conducted the research; W.I., J.Y. and H.K.
analyzed the data; W.I., J.Y. and H.K wrote the paper; and
H.K. had primary responsibility for the final content. All
authors read and approved the final manuscript. The
authors are grateful to Prof. Yuji Nakai and Ms. Kyoko
Nakazawa (The University of Tokyo) for their technical
assistance and Mr. K. Saito for his help in data analysis.
We thank Drs. H. Jia and L. Otani for their advice during the preparation of the manuscript. We are obliged to
Ajinomoto Co., Inc. for providing leucine. This research
was supported by International Council on Amino Acid
Science (ICAAS). Dr Kato received an honorarium and
travel expenses to attend the 8AAAW and present a lecture based on this paper.
References
1) Deldicque L, Francaux M. 2008. Functional food for
exercise performance: fact or foe? Curr Opin Clin Nutr
Metab Care 11: 774–781.
2) Sauberlich HE. 1961. Studies on the toxicity and antagonism of amino acids for weanling rats. J Nutr 75:
61–72.
3) Edmonds MS, Baker DH. 1987. Amino acid excesses for
young pigs: effects of excess methionine, tryptophan,
threonine or leucine. J Anim Sci 64: 1664–1671.
4) Harper AE, Benevenga NJ, Wohlhueter RM. 1970.
Effects of ingestion of disproportionate amounts of
amino acids. Physiol Rev 50: 428–558.
5) Sunde RA, Thompson KM, Evenson JK, Thompson BM.
2009. Blood glutathione peroxidase-1 mRNA levels can
be used as molecular biomarkers to determine dietary
selenium requirements in rats. Exp Biol Med (Maywood)
234: 1271–1279.
6) Lei XG, Evenson JK, Thompson KM, Sunde RA. 1995.
Glutathione peroxidase and phospholipid hydroperoxide glutathione peroxidase are differentially regulated in
rats by dietary selenium. J Nutr 125: 1438–1446.
7) Anthony JC, Yoshizawa F, Anthony TG, Vary TC, Jefferson LS, Kimball SR. 2000. Leucine stimulates translation
initiation in skeletal muscle of postabsorptive rats via a
rapamycin-sensitive pathway. J Nutr 130: 2413–2419.
8) Kadowaki M, Kanazawa T. 2003. Amino acids as regulators of proteolysis. J Nutr. 133: 2052S–2056S.
9) Smith TK, Austic RE. 1978. The branched-chain amino
acid antagonism in chicks. J Nutr 108: 1180–1191.
10) Gatnau R, Zimmerman DR, Nissen SL, Wannemuehler
M, Ewan RC. 1995. Effects of excess dietary leucine and
leucine catabolites on growth and immune responses in
weanling pigs. J Anim Sci 73: 159–165.
11) Kuhlman G, Roth JA, Flakoll PJ, Vandehaar MJ, Nissen
S. 1988. Effects of dietary leucine, alpha-ketoisocaproate and isovalerate on antibody production and lymphocyte blastogenesis in growing lambs. J Nutr 118:
1564–1569.
12) Reeves PG, Rossow KL, Lindlauf J. 1993. Development
and testing of the AIN-93 purified diets for rodents:
results on growth, kidney calcification and bone mineralization in rats and mice. J Nutr 123: 1923–1931.
13) Saito K, Ohta Y, Sami M, Kanda T, Kato H. 2010. Effect
of mild restriction of food intake on gene expression profile in the liver of young rats: reference data for in vivo
nutrigenomics study. Br J Nutr 104: 941–950.
14) Akobeng AK. 2007. Understanding diagnostic tests 3:
Receiver operating characteristic curves. Acta Paediatr
96: 644–647.
15) Fischer JE, Bachmann LM, Jaeschke R. 2003. A readers’ guide to the interpretation of diagnostic test properties: clinical example of sepsis. Intensive Care Med 29:
1043–1051.
16) Youden WJ. 1950. Index for rating diagnostic tests. Cancer 3: 32–35.
17) Perkins NJ, Schisterman EF. 2006. The inconsistency of
“optimal” cutpoints obtained using two criteria based on
the receiver operating characteristic curve. Am J Epidemiol 163: 670–675.
18) Aschkenasy A. 1979. Prevention of the immunodepressive effects of excess dietary leucine by isoleucine and
54
Imamura W et al.
valine in the rat. J Nutr 109: 1214–1222.
19) Calvert CC, Klasing KC, Austic RE. 1982. Involvement of
food intake and amino acid catabolism in the branchedchain amino acid antagonism in chicks. J Nutr 112:
627–635.
20) Matsuzaki K, Kato H, Sakai R, Toue S, Amao M, Kimura
T. 2005. Transcriptomics and metabolomics of dietary
leucine excess. J Nutr 135: 1571S–1575S.
21) Jones JI, Clemmons DR. 1995. Insulin-like growth factors and their binding proteins: biological actions.
Endocr Rev 16: 3–34.
22) Silha JV, Murphy LJ. 2002. Insights from insulin-like
growth factor binding protein transgenic mice. Endocrinology 143: 3711–3714.
23) Takenaka A, Hirosawa M, Mori M, Yamada S, Miura Y,
Kato H, Takahashi S, Noguchi T. 1993. Effect of protein
nutrition on the mRNA content of insulin-like growth
factor-binding protein-1 in liver and kidney of rats. Br J
Nutr 69: 73–82.
24) Filho JC, Hazel SJ, Anderstam B, Bergstrom J, Lewitt
M, Hall K. 1999. Effect of protein intake on plasma
and erythrocyte free amino acids and serum IGF-I and
IGFBP-1 levels in rats. Am J Physiol 277: E693–701.
25) Averous J, Maurin AC, Bruhat A, Jousse C, Arliguie C,
Fafournoux P. 2005. Induction of IGFBP-1 expression
by amino acid deprivation of HepG2 human hepatoma
cells involves both a transcriptional activation and an
mRNA stabilization due to its 3′UTR. FEBS Lett 579:
2609–2614.
26) Kharitonenkov A, Shiyanova TL, Koester A, Ford AM,
Micanovic R, Galbreath EJ, Sandusky GE, Hammond LJ,
Moyers JS, Owens RA, Gromada J, Brozinick JT, Hawkins
ED, Wroblewski VJ, Li DS, Mehrbod F, Jaskunas SR,
Shanafelt AB. 2005. FGF-21 as a novel metabolic regu-
lator. J Clin Invest 115: 1627–1635.
27) Inagaki T, Lin VY, Goetz R, Mohammadi M, Mangelsdorf
DJ, Kliewer SA. 2008. Inhibition of growth hormone
signaling by the fasting-induced hormone FGF21. Cell
Metab 8: 77–83.
28) Fisher FM, Estall JL, Adams AC, Antonellis PJ, Bina HA,
Flier JS, Kharitonenkov A, Spiegelman BM, Maratos-Flier
E. 2011. Integrated regulation of hepatic metabolism by
fibroblast growth factor 21 (FGF21) in vivo. Endocrinology 152: 2996–3004.
29) Cuevas-Ramos D, Almeda-Valdes P, Aguilar-Salinas CA,
Cuevas-Ramos G, Cuevas-Sosa AA, Gomez-Perez FJ.
2009. The role of fibroblast growth factor 21 (FGF21)
on energy balance, glucose and lipid metabolism. Curr
Diabetes Rev 5: 216–220.
30) Iizuka K, Takeda J, Horikawa Y. 2009. Glucose induces
FGF21 mRNA expression through ChREBP activation in
rat hepatocytes. FEBS Lett 583: 2882–2886.
31) Lundasen T, Hunt MC, Nilsson LM, Sanyal S, Angelin B,
Alexson SE, Rudling M. 2007. PPARalpha is a key regulator of hepatic FGF21. Biochem Biophys Res Commun
360: 437–440.
32) De Sousa-Coelho AL, Marrero PF, Haro D. 2012. Activating transcription factor 4-dependent induction of
FGF21 during amino acid deprivation. Biochem J 443:
165–171.
33) Gerard C, Goldbeter A. 2009. Temporal self-organization
of the cyclin/Cdk network driving the mammalian cell
cycle. Proc Natl Acad Sci USA 106: 21643–21648.
34) Harper JW, Elledge SJ, Keyomarsi K, Dynlacht B, Tsai LH,
Zhang P, Dobrowolski S, Bai C, Connell-Crowley L, Swindell E. 1995. Inhibition of cyclin-dependent kinases by
p21. Mol Biol Cell 6: 387–400.
A
IGFBP1
Expression
FGF21
B
CDKN1A
PP (1)
E, I, L, LO, PP, RB (54)
CCND1
Supplemental Fig. 1. Body weight gain of rats in the
pair-feeding experiment during 10 d (A) and 49 d (B).
The average amount of diet consumed by the LP8 rats
was given to pair-fed groups. The composition of the
pair-fed diet was same as the LP0 diet in Experiment
1. The LP8IV diet was LP8 supplemented with 4.42%
Ile and 5.19% Val, the ratio of Leu/Val and Leu/Ile
being equal to the AIN-93G diet. Multiple comparison
testing was done by the Tukey-Kramer test. Columns
which have different letters are significantly different
(p,0.05). Data are shown as means6SE.
CCNB2
Supplemental Fig. 2. Direct connections among the
biomarker candidates identified by transcriptome analysis based on IPA-knowledge. Genes are represented
as nodes, and the biological relationship between two
nodes is represented as a line. Red nodes mean upregulated genes, and green nodes mean down-regulated genes in DNA microarray. Nodes are connected
by canonical information stored in the IPA-knowledge
base. Edges are displayed with labels that describe the
relationship between the nodes. E, expression; I, inhibition; PP, protein-protein interaction; RB, regulation of
binding; LO, localization.
55
Gene Markers for Leucine Excess
Supplemental Table 1. Changes of liver gene expression of Igf-1, Igfbp3, Igfbp5, Igfals by Leu in rats fed the 6% casein diet.1
mRNA
LP0
LP2
LP4
LP8
Igf1
Igfbp3
Igfbp5 A
Igfals
Stat5a
Stat5b
1.0060.075
1.0060.092
1.0060.223
1.0060.072
1.0060.144
1.0060.168
1.2860.166
0.8860.064
0.9960.221
1.0560.142
0.8560.96
1.0360.134
1.2160.095
0.9960.139
0.9260.232
1.0860.153
0.8160.063
1.0260.124
1.1660.141
1.1160.124
1.760.227
0.6360.086
0.9160.137
0.7460.078
1 Results of qPCR of Igf1, Igfbp3, Igfbp5, and Igfals in LP0, LP2, LP4, and LP8 group in Experiment 1 are shown. Actb was
used as an internal standard for each measurement. Values are presented as the ratio to the average value of LP0. Multiple
comparison testing was done by the Tukey-Kramer test. Data are shown as means6SE.
Supplemental Table 2. Groups of genes exhibiting highly significant changes in the liver of rats fed 6% casein18% Leu
diet compared with rats fed the 6% casein diet.1
Top canonical pathways
Pathway name
Glycine, serine and threonine metabolism
Alanine and aspartate metabolism
Type II diabetes mellitus signaling
Aminoacyl-tRNA biosynthesis
Glycerophospholipid metabolism
2Log (p-value)2
2.10E205
6.25E205
1.14E204
1.49E204
1.76E204
Top bio functions
Functional category name
Amino acid metabolism
Small molecule biochemistry
Carbohydrate metabolism
Lipid metabolism
Cellular movement
1 2Log (p-value)
2.51E204 – 4.52E202
2.51E204 – 4.95E202
2.61E204 – 4.95E202
6.52E204 – 4.95E202
2.11E203 – 3.12E202
The result of IPA-core analysis is shown. The top 5 significantly changed canonical pathways or biological functions in LP8
liver are shown in the table in descending order.
2 Fisher’s exact test was used to calculate a p-value determining the probability that the association between the changed
genes (more than log2 ratio 0.5 or less than 20.5) in LP8 microarray data and each biological functions or canonical pathway is explained by chance alone.