Metabolomics for Medical Science

Masaru Yoshida M.D., Ph.D.Division of Metabolomics Research, Gastroenterology,

The Integrated Center for Mass Spectrometry, Kobe University Graduate School of Medicine

Dr. Metabolo by Dr. Megumi KIBI

Metabolomics for Medical Science

Today’s Contents

1. Background and Present State of Metabolomics

2. Methods to Measure Metabolites

3. Study for Biomarker Discovery

4. Study for Drug (metabolites) Discovery

Genomics

Proteomics

Metabolomics

DNA:

Protein:

Metabolite:(4,000)

Omics Studies

（possible by recent progress of mass spec. & analysis software）

(23,000)

(1,000,000)

The large-scale study of genomeGenome wide association study

The large-scale study of proteins

The systematic study of metabolites

✓Smaller numbers compared to genome, RNA, and proteomeHuman genome = about 23,000Human functional RNA = about 100,000Human proteome = about 1,000,000 Human metabolome = about 3,000-4,000(enzyme related gene, less than 1,100)

✓Metabolites have been examined by traditional assaysTraditionally, metabolites have been well investigated in biochemical fields.

✓Close to phenotypeAlterations in genome and proteome do not always change the phenotype

due to homeostasis.

✓No species-specificityAnalytical methods are available to samples from different species.

Why Metabolomics?

Global Movement of Metabolomics2020 visions (nature, 2010)

・Search Engines・Microbiome・Lasers

・Ecology・Metabolomics

Multi-platform system is required.

Metabolites ・・・・a great variety of physicochemical propertieshydrophobic hydrophilicPolarity

MW Fatty acid

SugarAmino acid

Organic acid

AmineSugar alcohole

Lipid Peptide

Sugar phosphate

GC/MS and Ion-paring LC/MSLC/MS

NucleotideCoA

larg

esm

all

Multi-platform System for Widely Targeted Profiling

Method Ionization Derivatization Mobile phase

代謝物カテゴリー

Fatty acid Lipid Organic

acidSugar phosphate, Co A, Nucleotide Amine Amino

acid

Sugar, Sugar

alcohol

GC/MS EI Essential Gas △ × ○ × ○ ○ ◎

LC-MS ESI No need Liquid Reverse phase ◎ ◎ × × × × ×

Ion pair method × × ◎ ◎ × △ △

PFPP column × × × × ◎ ◎ ×

by Nishiumi S, Izumi Y, Matsubara A et al.

Metabolomics Analysis by GC/MS

Observed EI spectrum

Capilary column for metabolites separationin the colum oven

Database spectrumColumn oven temp.100ºC~ ~320ºC

Metabolites are efficiently separated at their own specific boiling points.

Metabolite X

70eV thermoelectron

Mass (m/z)

73.2 155.0 316.8

Fragmentation

Each metabolite is fragmented by 70 eV thermoelectron.

Pancreatic cancer patient Healthy volunteer

TIC chromatograms obtained by GC/MS of serum

Superposition of TIC chromatograms

Some metabolites are changed in patients. (Nishiumi S et al. Metabolomics 2010)

1 Boric acid2 Trichloroacetic acid3 Phenol4 Lactic acid5 2-Hydroxyisobutyric acid6 Caproic acid7 Glycolic acid8 L-Alanine9 L-Glycine10 Glyoxylic acid11 Oxalic acid12 2-Hydroxybutyric acid13 2-Furoic acid14 Sarcosine15 3-Hydroxypropionic acid16 Pyruvic acid17 Valproic acid18 4-Cresol19 3-Hydroxybutyric acid20 3-Hydroxyisobutyric acid21 2-Hydroxyisovaleric acid22 alpha-Aminobutyric acid23 2-Methyl-3-hydroxybutyric acid24 Malonic acid25 beta-Aminoisobutyric acid26 3-Hydroxyisovaleric acid27 2-Keto-isovaleric acid28 Methylmalonic acid29 L-Valine30 Ethylhydracrylic acid31 Urea32 4-Hydroxybutyric acid33 2-Hydroxyisocaproic acid34 3-Hydroxyvaleric acid35 D,L-Norvaline36 Acetoacetic acid37 2-Hydroxy-3-Methylvaleric acid38 Benzoic acid39 Acetoacetic acid40 Octanoic acid41 Cyclohexanediol42 2-Methyl-3-hydroxyvaleric acid43 2-Propyl hydroxyglutaric acid44 L-Leucine45 Glycerol46 Acetylglycine47 Phosphoric acid48 Ethylmalonic acid49 2-Ketoisocaproic acid50 L-Isoluecine

51 allo-Isoleucine52 Phenylacetic acid53 Maleic acid54 L-Proline55 2-Octenoic acid56 Succinic acid57 Methylsuccinic acid58 Glyceric acid59 Fumaric acid60 Uracil61 Citraconic acid62 Propionylglycine63 L-Serine64 Acetylglycine65 Mevalonic lactone66 Isobutyrylglycine67 2-propyl-3-hydroxy-pentanoic acid68 L-Threonine69 Mesaconic acid70 Glutaric acid71 Thymine72 3-Methylglutaconic acid73 3-Methylglutaric acid74 Propionylglycine75 Isobutyrylglycine76 2-Deoxytetronic acid77 3-Methylglutaconic acid(E)78 Glutaconic acid79 Succinylacetone80 Decanoic acid81 3-Methylglutaconic acid(Z)82 2-Propyl-5-hydroxy-pentanoic acid83 Citramalic acid84 Mandelic acid85 Isovalerylglycine86 Malic acid87 Adipic acid88 Phenyllactic acid89 p-Nitrophenol90 Isovalerylglycine91 2-Hexenedioic acid92 Aspartic acid93 L-Methionine94 5-Oxoproline95 Thiodiglycolic acid96 4-Hydroxyproline 97 3-Methyladipic acid98 Acetylsalicylic acid99 7-Hydroxyoctanoic acid100 2-Propyl-glutaric acid

151 Caffeine152 Hydroxylysine (2 isomers)153 Methylcitric acid154 Vanilmandelic acid155 Sebacic acid156 Decadienedioic acid157 4-Hydroxyphenyllactic acid158 Theophylline159 L-Histidine160 3,4-Dihydroxymandelic acid161 L-Tyrosine162 Indole-3-acetic acid163 Palmitoleic acid164 Palmitic acid165 2-Hydroxysebacic acid166 3-Hydroxysebacic acid 167 2-Hydroxyhippuric acid168 Dodecanedioic acid169 Naproxen170 N-Acetyltyrosine171 Uric acid172 Margaric acid173 3,6-Epoxydodecanedioic acid174 Indolelactic acid175 Stearic acid176 L-Tryptophan177 3-hydroxydodecanedioic acid178 Chloramphenicol

Amino acidsOther organic acidsOther organic acids(Fatty Acids)AlcoholsKetonesNucleosidesCarbohydratesHeterocyclic moleculesInorganic compounds

101 Cinnamic acid102 5-Hydroxy-2-furoic acid103 Tiglylglycine104 3-Methylcrotonylglycine105 Tiglylglycine106 3-Hydroxybenzoic acid107 3-Methylcrotonylglycine108 2-Hydroxyphenylacetic acid109 2-Hydroxyglutaric acid110 Pimelic acid111 3-Hydroxy-3-methylglutaric acid112 3-Hydroxyphenylacetic acid113 L-Glutamic acid114 4-Hydroxybenzoic acid115 2-Ketoglutaric acid116 L-Phenylalanine117 4-Hydroxyphenylacetic acid118 Lauric acid119 Tartaric acid120 Hexanoylglycine121 2-Ketoglutaric acid122 N-Acetylaspartic acid123 Glutaconic acid124 N-Acetylaspartic acid125 Asparagine126 2-Hydroxyadipic acid127 Octenedioic acid128 3-Hydroxyadipic acid129 Suberic acid130 Lysine131 2-Keto-adipic132 alpha-Aminoadipic acid133 Tricarballylic acid134 Glutaconic acid135 Aconitic acid136 Orotic acid137 3-Methoxy-4-hydroxybenzoic acid138 Homovanillic acid139 L-Glutamine140 Azelaic acid141 Hippuric acid142 Isocitric acid143 Citric acid144 Glucuronoic lactone145 Hippuric acid146 Homogentisic acid147 Myristic acid148 Glucuronoic lactone149 Methylcitric acid150 3-(3-Hydroxyphenyl)-3-hydroxypropionic acid

Metabolites Database for Identification by GC/MS

Glycolysis

Citrate cycle

Pentose phosphate pathway

Central metabolism

+ Coenzyme etc.

【Anionic metabolites】

Sugar phosphates

Organic acids

Nucleotides

Cofactors (Acetyl-CoA, NAD(P)H, etc.)

Most of intermediates metabolites are water-soluble anionic metabolites.

Anionic Metabolites Profiling by Ion-Paring-LC/MS/MS

<HPLC condition> Column: Unison UK-C18 column, 3 m, 2.0 X 150 mm (Imtakt Corp.) Column Temp.: 35.0oCInjection: 5 LSolvent A: 10 mM TBA/15 mM acetic acid in water

B: MeOHFlow rate: 0.3 mL・min-1

【ODS C18 column + Ion-pair reagent】

G6Phighly polar anionic metabolite

NH+

Ion paring

Tributylamine (TBA)Cationic ion-paring reagent

ODS C18 particle

cannot be retained on the ODS column.

Hydrophobicity of each polar-anionic metabolites is increased!

Anionic Metabolites Profiling by Ion-Paring-LC/MS/MS

UHPLC Nexera + LCMS-8040 (Shimadzu Co.)

Retention and separation

Serum Lipidomics by LC/MS/MS

Large-scale lipids profiling (one of the metabolomics)

Lipidomics

stimulateLipid metabolism

related enzymeCancer

onset ・ malignantinhibit

Lipids may be associated with the each process of diseases.

Candidates of biomarkers

・Lipids

Simple lipids (neutral lipids: C, H, O)

Complex lipids (C, H, O + P, N, S, Sugars)

Basic structure of lipids

Glycerol

H2C O C R1

O

CH

H2C O R3

OCR2

O

Glycerophospholipids

Sphingophospholipids

Phospholipids

Glycolipids Glyceroglycolipids

Sphingoglycolipids

Target lipids

・Diacylglycerol (DG) → R2, R3:acyl chains・Triacylglycerol (TG) → R1, R2, R3:acyl chains

・lyso-Phosphatidylcholine (LPC) ・lyso-Phosphatidylethanolamine (LPE)・Phosphatidylcholine (PC) ・Phosphatidylethanolamine (PE)

・Monogalactosyldiacylglycerol (MGDG)

・Cerebroside (CB)

・Free fatty acid (FFA): approximately 50 metabolites

Lipids variety: Theoretical → over 30,000 species; Actual → over 1,000 species

・Sphingomyelin (SM)

P

O

OH

O XR3:

・Phosphatidic acid (PA) ・Phosphatidylglycerol (PG)・Phosphatidylinositol (PI)・Phosphatidylserine (PS)

Glycerophospholipids metabolic pathway

Dihydroxyacetone-phosphate (DHAP)

GlycerolGlycerol‐3‐phosphate (G3P)

sn-1-acyl-G3P

Phosphatidic acid (PA)

ATPADP

NADHNAD+

Diacylglycerol (DAG)

Choline

O‐Phosphocholine

CDP‐choline

Acyl‐CoA

CoA‐SH

Acyl‐CoA

CoA‐SH

H2O

Pi

ADP

ATP

ATP

ADP

CTP

PPi

CMP

Sphingomyelin (SM)

Ceramide

DAG

Phosphatidylserine (PS)

SerineCholine

Ethanolamine

O‐Phosphoethanolamine

CDP‐ethanolamine

ATP

ADP

CTP

PPi

CMP

SerineEthanolamine

CO2

CDP‐DAGCTP

PPi

CMP

G3P

Phosphatidyl-glycerophosphate

Phosphatidylglycerol(PG)

Cardiolipin (CL),Diphosphatidylglycerol

H2O Pi

Phosphatidylinositol (PI)

PI3P

PI3,4P2

PI3,4,5P3

PI4P

PI4,5P2 PI3,5P2

CDP‐DAG CMP

myo‐inositol CMP

PI5P

ATP

ADP

ATP

ADP

ATP

ADP

ATP

ADP

ATP

ADP

ATP

ADP

ATP

ADP

ADP

ATP

Phosphatidylcholine (PC) Phosphatidylethanolamine (PE)

Lysophosphatidylcholine (LPC) Lysophosphatidylethanolamine (LPE)

H2O

Fatty acidH2O

Fatty acid

+ Free fatty acid (FFA)

• Free fatty acid (FFA) ・・・ 35 MRM transitions (Negative)• Phosphatidylcholine (PC) ・・・・ 59 MRM transitions (Positive)• Lysophosphatidylcholine (LPC) ・・・ 21 MRM transitions (Positive)• Phosphatidylethanolamine (PE) ・・・・ 67 MRM transitions (Positive)• Lysophosphatidylethanolamine (LPE) ・・・18 MRM transitions (Positive)

A total of 200 MRM transitions settings with posi・nega switching

Precursor-ion scanNeutral loss scan

Positive modePhosphoryl cholineCommon fragment of m/z 184.1

Choice of Precursor Ions

Condition of UHPLC Chromatography for Structural Isomers Separation

Determination of Fatty AcidsProduct-ion scan

Fatty acid

Fatty acid

Negative mode

LC/MS/MS(triple-quqdrupole)

MRM settings for multi-targeted lipid profiling

Identification of lipids using various samples by exact m/z (Mouse liver, intestine, brain, and blood plasma, and Human serum)

Each Cancer Mortality Rate

Source: ‘‘vital statistics’’ by Ministry of Health, Labour and Welfare (MHLW) in Japan

The number of colorectal cancer patients has been increased with a Western-style food.

Gastric cancerPancreatic cancerBreast cancerOvarian cancerLeukemia

Liver cancerLung cancerUterine cancerProstate cancerColorectal cancer

People / 100 thousand people

Male Female

Colorectal Cancer (CRC)

• Occult blood test → Resistance toward stool collection→ False negative

• Conventional tumor makers→ Lower sensitivity at the early stage

• Imaging methods (CT etc.)→Not applicable to very early screening

• Colonoscopy→ Invasive procedure

When CRC is first diagnosed,40-60% are advanced.

Early CRC

Complete remission rate:almost 100%

Advanced CRC

Omics Research using Blood for Diagnosis

Genomics(gene)

Proteomics(protein)

Metabolomics(metabolite)

Number of targets

Difficult

≈ 23,000 ≈ 100,000 ≈ 4,000

Analysis Easy

Not reflect Difficult to reflectHealth condition Easy to reflect

Laborious

Serum Metabolomics by GC/MS

Colorectal cancer patients

Healthy volunteers P value

N 60 60Male 39 39

Female 21 21

Age Average 67.7 64.5 N.S.Median 70 68Range 36-88 39-88

BMI 21.9 22.1 N.S.

Stage 0 121 122 123 124 12 (N.S., Not significant)

• The cancer staging was determined base on the International Union Against Center (UICC) TNM classification• Diagnosis of colorectal cancer patients were performed at Kobe University Hospital or Hyogo Cancer Center.• Healthy volunteers were selected based upon the results of consultations at Kobe University Hospital or those

of health examination at another institutions.

Training set

First ScreeningConfirmation of the metabolites• not-derived from serum• stability through the analysis• intra and inter- day variations• Increased or decreased in CRC patients


Training setA total of 131 metabolites was identified in 50 L of serum.

27 candidates

(Nishiumi S et al. PLos One 2012)

GC/MS血清メタボロミクス

Second ScreeningStepwise selection

Construction of Logistic Regression Model

Exclusion of metabolites from foods Selection of Top 10 metabolites


Metabolites selected by first screeningLactitol (an artificial sweetener)meso-Erythritol (an artificial sweetener)Kynurenine2-Hydroxy-butyrateGlutamic acidp-Hydroxybenzoic acidArabinoseAspartic acidCysteine+CystineCysteamine+CystaminePyruvate+Oxalacetic acidIsoleucineXylitolPyroglutamic acid-AlaninePalmitoleate(C16:1)OrnithineInositolPhosphateAsparagineGlucuronate_1CitrullineGlucosamine_2O-PhosphoethanolamineCreatinineRibuloseNonanoic acid(C9)

Stepwise selectionMethods that select metabolites objectively from candidates.

Stepwise-Multivariate Logistic Regression (MLR) Model

Multivariate Logistic Regression (MLR) model

How can we predict “diagnosis” using variables?

Multivariate linear regression model；prediction of Y using variables.Y = aX1+ bX2 + cX3 + dX4…..+ intercept

Set dummy; healthy = 0, diseased = 1

“Output” of the prediction model needs to be converged within 0 and 1.

P 1

1 e(aX1 + bX 2 + cX 3 + dX 4 .... + intercept )

Multivariate Logistic Regression (MLR) Model

Appropriate P value (cut off value) is determined by ROC analysis.

Prediction model

Sensitivity: 85.0%Specificity: 85.0%Accuracy: 85.0%

) ................dx + cx + bx + ax + Intercept( 432111

eP

Serum Metabolomics by GC/MSCoefficient

（a, b…）

2-Hydroxy-butyrate 286.59Aspartic acid 33.87Kynurenine 1634.96Cystamine 78.78

Intercept -8.32

AUC= 0.9097 (95% CI: 0.8438-0.9495)Cut-off value=0.4945

ROC analysis

SpecificityFalse positive

True

pos

itive

Sens

itivi

ty


Validation

Colorectal cancer patients

Healthy volunteers P value

N 59 63Male 30 32

Female 29 31

Age Average 64.8 62.8 N.S.Median 66 63Range 31-84 47-73

BMI 22.5 22.2 N.S.

Stage 0 151 112 33 114 19


(N.S., Not significant)

Training setCEA CA19-9 Predictive model

stage 0-4

stage 0-2

stage 3-4

stage 0-4

stage 0-2

stage 3-4

stage 0-4

stage 0-2

stage 3-4

Sensitivity 35.0% 30.6% 37.5% 16.7% 5.6% 29.2% 85.0% 83.3% 87.5%Specificity 96.7% 100% 85.0%Accuracy 65.8% 58.3% 85.0%

Validation setCEA CA19-9 Predictive model

stage 0-4

stage 0-2

stage 3-4

stage 0-4

stage 0-2

stage 3-4

stage 0-4

stage 0-2

stage 3-4

Sensitivity 33.9% 6.9% 60.0% 13.6% 0% 26.7% 83.1% 82.8% 83.3%Specificity 96.8% 100% 81.0%Accuracy 66.4% 58.2% 82.0%


Comparison with Tumor Markers


Summary• Construction of stepwise MLR model based on the results of training

set between healthy and CRC patients

• The calculated prediction model with training set had good performance(sensitivity, 85.0%: specificity, 85.0% and accuracy, 85.0%).

• When applied to the validation set, the predictive ability was maintained (sensitivity, 83.1%: specificity, 81.0% and specificity, 79.6%).


Metabolites selected in the prediction model 2-Hydroxy-butyrate（2-HB）

Aspartic acid（Asp）Kynurenine（Kyn）Cystamine（Cyst）

p= 1 + e-{-8.32+286.59(2-HB)+33.87(Asp)+1634.96(Kyn)+78.78(Cyst)}1

Kobayashi et al. Cancer Epidemiol Biomarkers Prev. 2013

Serum Metabolomics for Early Detection of Pancreatic Cancer

Metabolites for FormulaXylitol (Xly)1,5-Anhydro-D-glucitol（1,5AD）Histidine（His）Inositol（Ino）

p= 1 + e-{5.48+167.57(Xly)-15.21(1,5AD)-282.34(His)+60.99(Ino)}

1

Development for Clinical Medicine

blood

Meaduament by Conventional Methods

Diagnosis

) ................dx + cx + bx + ax + Intercept( 432111

eP

Diagnosis Kits for Specific Disease

Pretreatment GCMS analysisIdentification and Quantification

Diagnosis of Multiple Diseases

Extraction and Derivatization ) ................dx + cx + bx + ax + Intercept( 432111

eP

automation

Background for Inflammatory Bowel Disease

Inflammatory bowel disease…

is characterized by chronic and relapsing inflammation of the gastrointestinal tract

Genetic Factors

Immune Abnormalities

Environmental Factors

Intestinal Inflammation

HIbi T, et al. J Gastroenterol. 2006

Inflammatory bowel disease

Utilized metabolomics to examine the pathogenesis of IBD

Aim

?

C57BL/6J

DSS-induced Colitis Model

3.0% DSS Water

0 day 5 day 7 day

Sacrifice

10 day

DSS: dextran sulphate sodium

Oral administration of dextran sulphate sodium (DSS) causes similar clinical features to human UC. (Okayasu et al., 1990; Cooper et al., 1993)

Day 7: The degree of colitis was severe

Day 10: The degree of colitis was almost improved

(x200)

(x40)

DSS (day 10)DSS (day 7)Water

Shiomi et al., Inflamm Bowel Dis. 20111

Methods in Metabolomics

Serum (Start volume: 50 l) / Tissue (20 mg)Extraction (CH3OH:CHCl3:H2O=2.5:1:1)

Soluble FractionLyophilization

Lyophilized Product Oximation & Derivatization

Liquid Solution

Metabolite Data

Measurement by GCMS

Gas Chromatograph Mass Spectrometer (GC/MS)

Results

・ In serum, 77 metabolites were detected.23 Amino acids42 Organic acids6 Fatty acids6 Others

・ In colon tissue, 92 metabolites were detected.24 Amino acids56 Organic acids6 Fatty acids6 Others

Look for the decreased metabolite at day 7

Results ~PLS-DA scores plots～Partial Least Square Discriminant Analysis (PLS-DA)

: one of Multiple Classification Analysis

-10

-5

0

5

10

-10 -5 0 5 10

t[2]

t[1]

-5

0

5

-10 -5 0 5 10t[3

]t[2]

-5

0

5

-10 -5 0 5 10

t[3]

t[1]

control

DSS (day10)

control

DSS (day7)

DSS (day10)

control

DSS (day7)

DSS (day10)DSS (day7)

2D-PLS-DA scores plots

control

DSS (day7)

DSS (day10)

PLS-DA scores plots showed distinct clustering and clear separation of the groups according to the degree of colitis.

3D of the first three principal components

-0.30

-0.20

-0.10

-0.00

0.10

0.20

-0.20 -0.10 -0.00 0.10

w*c

[3]

w*c[1]

T1T2T3T4

T5T6

T7T8

T9

T10

T11

T12T13T14

T15T16

T17

T18T19T20

T21

T22T23

T24

T25

T26T27T28 T29

T30

T31

T32 T33T34

T35

T36

T37

T38

T39

T40

T41

T42T43

T44

T45

T46

T47

T48 T49

T50

T51T52T53

T54

T55

T56

T57

T58

T59T60T61

T62

T63

T64

T65

T66T67 T68

T69

T70

T71

T72

T73

T74

T75 T76T77

T78T79T80T81

T82

T83

T84

T85

T86

T87

T88

T89

T90

T91

T92

-0.30

-0.20

-0.10

-0.00

0.10

0.20

-0.20 -0.10 -0.00 0.10 0.20

w*c

[3]

w*c[2]

T1T2T3

T4

T5T6

T7T8

T9

T10

T11

T12T13T14

T15T16

T17

T18 T19T20

T21

T22T23

T24

T25

T26T27T28T29

T30

T31

T32 T33T34

T35

T36

T37

T38

T39

T40

T41

T42T43

T44

T45

T46

T47

T48T49

T50

T51 T52T53

T54

T55

T56

T57

T58

T59T60T61

T62

T63

T64

T65

T66T67T68

T69

T70

T71

T72

T73

T74

T75T76 T77

T78T79T80T81

T82

T83

T84

T85

T86

T87

T88

T89

T90

T91

T92

-0.20

-0.10

-0.00

0.10

0.20

-0.20 -0.10 -0.00 0.10

w*c

[2]

w*c[1]

T1T2

T3T4

T5

T6

T7

T8

T9 T10

T11

T12

T13T14 T15

T16 T17

T18

T19T20

T21

T22

T23

T24T25T26

T27T28

T29

T30T31

T32

T33T34T35

T36

T37T38T39

T40

T41T42T43

T44

T45

T46

T47

T48

T49

T50T51

T52T53T54

T55

T56

T57

T58

T59T60T61

T62

T63

T64

T65

T66

T67

T68

T69

T70

T71

T72

T73

T74T75T76

T77

T78T79T80

T81

T82

T83T84

T85

T86 T87

T88T89 T90

T91 T92

Results

-10

-5

0

5

10

-10 -5 0 5 10

t[2]

t[1]

-5

0

5

-10 -5 0 5 10

t[3]

t[2]

-5

0

5

-10 -5 0 5 10

t[3]

t[1]

The dereased or increased meatbolits will be found easily.

~PLS-DA scores plots and loadings plots～

control

DSS (day10)DSS (day7)control

DSS (day10) DSS (day7)

control

DSS (day7)

DSS (day10)


Decreased Metabolites at day 7 in colon tissue

Results

00.20.40.60.811.21.4

Succinic acid

DSS7day

DSS10dayCont.

00.20.40.60.811.21.4

L-Glutamine

DSS7day

DSS10dayCont.

DSS10day

00.20.40.60.811.21.41.6

L-Glutamic acid

DSS7dayCont. DSS

7dayDSS

10day

00.20.40.60.811.2Indol-3-acetic acid

Cont. (Avg±SE, n=6)


Supplementation of Glutamine in DSS-induced Colitis

C57BL/6J 3.0% DSS Gln or Water

0 day 5 day 7 day

Sacrifice

Administration of glutamine could attenuateDSS-induced colitis in mice.

DSS+ 2.0 g/dl Gln DSS

(x200)

(x40)

DSS+ 4.0 g/dl Gln

Gln: Glutamine

Histological score

DSS DSS+

2.0 g/dl Gln

DSS+

4.0 g/dl Gln

0

5

10

(Mean±SE, n=5)

*

*

D: DSSG: glutamine

(Avg±SE, n=5)

The glutam

inelevel

D+

4G

0

0.2

0.4

0.6

0.8

1

1.2

W+

2G

W+

4G

D+

2G

W DTh

e glutam

ine level

0

0.2

0.4

0.6

0.8

1

1.2

W+

2G

W+

4G

D+

2G

D+

4G

W D

Glutamine:

◆ The primary source of amino acids in the intestinal mucosa ◆ The main respiratory substrate for enterocytes

Serum Colon tissue

Supplementation of Glutamine in DSS-induced Colitis

• The pathogenesis of colitis led to the alterations of some metabolites in the colon tissue.

• Supplementation of the metabolite in the body; i.e., glutamine, recover rapidly.

Inflamm Bowel Dis, 2011

DSS-induced colitis animal model

N (male/female) 22 (12/10)Age (median/range) 43.9/14-85Years with disease (median/range) 8.4/1-30Inflammation (Proctitis/Left Side/Pan Colitis) 3/7/12Rachmilewitz index (CAI) (remission/active) 16/6Sampling location (normal/lesion) 16/22Matt's classification (median/range) 3/1-5Daily medication5-aminosalicylates 21 (2250-4000 mg/day)Prednisolone 2 (5-10 mg/day)6-mercaptopurine 0Azathioprine 0Tacrolimus 2 (4-8 mg/day)

Ulcerative colitis (UC) patients

Patient Information

Liquid-liquid extraction from each tissue siteNon-inflamed

site

Inflamed site

Colon tissue of UC patient

GCMS-QP2010plus

GC/MS measurementTarget: Amino acids and

TCA-cycle related metabolites

colon

rectum

anus

cecum

Tissue Metabolomics

Fold induction P value(lesion/normal)

N-Acetylaspartic acid 0.66 0.0028a

Alanine 0.58 <0.0001a

Aspartic acid 0.94 0.39Asparagine 0.47 <0.0001a

Glutamic acid 0.73 0.044a

Glutamine 0.25 <0.0001a

Glycine 0.73 0.0021a

Isoleucine 0.67 0.00067a

Leucine 0.74 0.0050a

Lysine 0.59 0.031a

Methionine 0.70 0.0016a

5-Oxoproline 0.89 0.30 Phenylalanine 0.70 0.0016a

Proline 0.59 <0.0001a

Serine 0.67 0.00049a

Threonine 0.70 0.0030a

Tryptophan 0.75 0.051Tyrosine 0.70 0.0011a

Valine 0.70 0.0023a

Fold induction P value(lesion/normal)

Citric acid 0.61 0.011a

Fumaric acid 0.56 0.00031a

Isocitric acid 0.58 0.0031a

Malic acid 0.50 0.00060a

Pyruvic acid 1.03 0.41 Succinic acid 0.63 <0.0001a

Amino acids (19) TCA related metabolites (6)

Result: Comparison of Detected Metabolites

The levels of 16 amino acids and 5 TCA-clcle related metaboliteswere significantly decreased in the lesional site compared with the normal tissue.

(Red color: Significantly decreased metabolites)

(Ooi et al., Inflamm Res, 2011)

GCMS-QP2010plus

• UC patients• Healthy volunteers

Blood collection

Serum metabolomics

Serum Metabolomics Method

Liquid-liquid extraction from blood

GC/MS measurementTarget: Amino acids and

TCA-cycle related metabolites

Ulcerative colitis (UC) patients

N (male/female) 13 (7/6)Age (median/range) 39/26-57Years with disease (median/range) 5.8/1.5-12

N (male/female) 17 (12/5)Age (median/range) 38.9/25-67

Healthy volunteers

(All patients were followed up, and their pathology of UC showed clinical remission.)

Patient Information

Fold induction P valueUC/H UC vs H

Alanine 0.99 0.983Aspartic acid 1.46 0.025a

Asparagine 0.80 0.0032a

Glutamic acid 0.73 0.075Glutamine 0.51 <0.0001a

Glycine 1.74 <0.0001a

Histidine 0.38 <0.0001a

4-Hydroxyproline 1.30 0.305 Isoleucine 1.12 0.174 Leucine 0.97 0.983 Lysine 1.14 0.187 Methionine 1.08 0.754 5-Oxoproline 1.01 0.818 Phenylalanine 0.99 0.691 Proline 0.96 0.601 Serine 1.08 0.464 Threonine 1.10 0.950 Tryptophan 0.63 0.00010a

Tyrosine 0.94 0.161 Valine 0.99 0.884

UC, Ulcerative colitis patientsH, Healthy volunteers

(Ooi et al., Inflamm Res, 2011)

Fold induction P valueUC/H UC vs H

Aconitic acid 1.20 0.069 Citric acid 0.98 0.544 Fumaric acid 1.33 0.013a

Isocitric acid 0.96 0.490 Malic acid 1.18 0.117 Pyruvic acid 1.03 0.851 Succinic acid 0.99 0.722

Result: Comparison of Detected MetabolitesAmino acids (20) TCA related metabolites (6)

The levels of 4 metabolites including asparagine, glutamine, histidine, and tryptophan were significantly decreased in both the lesional tissue and the UC patients serum (P < 0.05).

Summary

• The levels of many metabolites were significantly decreased in the inflamed site.

• The serum levels of some of amino acids were also significantly downregulated in the UC patients.

Human inflammatory bowel disease

✓ The potential of nutritional therapy

The potential of Personalized Medicine

Therapy with in vivo targeted metabolite• Supplementation of the insufficient

metabolites in the body• Normalization of the metabolites which

present excessively in vivo

Metabolic profiling～Metabolomics～

Identification of the specific changing metabolites in

individual patient

Personalized medicine

Improvement in pathological conditions

Conclusion

Metabolomics is capable of providing the greatly useful information in the medical field.

✓The discovery of disease biomarkers

✓The finding of novel therapeutic agents

✓Examination of pathogenetic mechanisms behind various diseases

Documents

Metabolomics for Medical Science