QuantitativeStructureActivity

Relationship(QSAR)models
CeydaOksel

Xue Z. Wang

University ofLeeds
27.03.2013
A pan-European Infrastructure for Quality in Nanomaterials Safety Testing
NUID UCD | NHM | IOM | JRC | BFR | KIT | FUNDP | IST | UNIVLEEDS | NILU | HMGU | LMU | CIC | UU | ICN
DLO/RIKILT | WU | DGUV | TAU | VITO | SMU | TCD | FIOH | UOE | HW | CNRS | INERIS | UBRM | RIVM

Purpose ofToday

Theory ofQSAR
NewDataMining Methods
NanomaterialCharacterization
Nanotoxicity
Application ofQSARmodelinnanotoxicity
modeling

QualityNanoTrainingSchool

Part I
Theory ofQSAR
QualityNanoTrainingSchool

QSARHistory
1865

1893

Crum Brownand
Fraser[1]
expressed the
ideathat there
was a relationship
betweenactivity
andchemical
structure.

Richet[2]
correlated
toxicities of
simple organic
molecules
withtheir
solubilityin
water.

1900

Meyer[3] and
Overton[4]found
linearrelationships
betweenthe
toxicityoforganic
compoundsand
their lipophilicity.

1969

Hansch[5] published
afreeenergyrelated
modelto correlate
biologicalactivities
with
physicochemical
Properties.

Thefatheroftheconceptof
quantitativestructureactivity relationship (QSAR),
thequantitativecorrelationofthephysicochemical
propertiesofmoleculeswiththeirbiological
activities[5]
CorwinHermanHansch
(1918 2011)

[1]A.CrumBrown,T.R.Fraser,Trans.R.Soc.Edinb.25(19681969) 257. [2] M.C.Richet,Compt.Rend.Soc.Biol.45(1893)775.

[3]H.Meyer,Arch.Exp.Pathol.Pharmakol.42(1899)109. [4] E.Overton,Studien
uber dirNarkose,GustavFischer,Jena,1901
QualityNanoTrainingSchool
[5] Hansch,C.(1969)AquantitativeapproachtobiochemicalstructureactivityrelationshipsAccountsofChemicalResearch, 2, 232239.

What isQSAR?

QSAR(QuantitativeStructureActivity
Relationship)
TheVarianceIn

Molecular
Structure

Physicochemical
or Biological Property

A QSAR is
that relates
descriptors
compound
activity.

a statistical model
a set of structural
of a chemical
to its biological

MathematicalDependencies
Toxicity/Ecotoxicity Prediction
BulkFormNanoparticles
(widely used)(notadopted)
The presenceofparticular characteristics increase
the propability that the compund istoxic
QualityNanoTrainingSchool
[1]T.Puzyn etal.(eds.),RecentAdvancesinQSARStudies,103125. DOI10.1007/9781402097836_4,CSpringerScience+Business MediaB.V.2010

nanoQSAR

QSAR(QuantitativeStructureActivity
Relationship)

Successful QSAR studies

What about Nanosize particles?

QSAR has been widely used to predict the toxicity of substances in bulk
form (especially druglike compunds) but, up to date, QSAR studies for the
prediction of nanoparticle toxicity have been rarely reported.
The European system for new chemical management (REACH) promotes
QSAR methods as an alternative way of toxicity testing.
QSARmodels are very useful incase ofthe classic chemicals butthe
concept ofnanoQSARisstill under development.
QualityNanoTrainingSchool

Why dowe need QSARmodels?

Sourcesofinformation[1]

All chemical substances need to betested interms oftheir toxicological

and environmental properties before their use
There are several reasons to use QSARModels :very fast,often free,
reducethenumberofanimalsusedinexperiments
predict

predict
validate

INSILICO
(computational)
Alternative,fast
developingapplications

IN VITROINVIVO
(incontainer)
(onlivingorganism)
Themostimportantsources
ofinformation

Reduction inthe time, themost

QSAR
cost and animal testing promising quantitativestructureactivity
[1]
relationship

[1]A..Gajewicz,etal.,Advancingriskassessmentofengineerednanomaterials:Applicationofcomputationalapproaches,
QualityNanoTrainingSchool
Adv.DrugDeliv.Rev.(2012),doi:10.1016/j.addr.2012.05.014[1]

Purpose ofQSAR
To predict biologicalactivityand physicochemical
propertiesbyrationalmeans
To comprehendandrationalizethemechanismsofaction
withinaseriesof chemicals
Savingsinthecostofproductdevelopment
Predictionscouldreducetherequirementforlengthyand
expensiveanimaltests.
Reductionofanimaltests,thusreducing animaluseand
obviouslypainanddiscomforttoanimals
Otherareasofpromotinggreenandgreenerchemistryto
increaseefficiencyand eliminatewaste
QualityNanoTrainingSchool

Requirements for QSAR

1.Setof
molecules
Molecular
Modelling

2. Setofmolecular
descriptors

3.Measured biological
activity or property

DESCRIPTORS

QSARs

Physcochemical,
biological, structural

e.g. Neural Networks

PLS, Expert Systems

How manycompounds arerequired

todevelopaQSAR?
There isnodirect
andsimple answer

asmany as
possible

Toprovidesomeguide,itis widelyacceptedthat
betweenfiveandtencompoundsarerequired
foreverydescriptor inaQSAR[1]

4.Statistical
methods

Toxicity
Endpoints
Daphnia magna
EC50
Cancinogenicity
Mutagenicity
Rat oral LD50
Mouse inhalation
LC50
Skin sensitisation
Eye irritancy

QualityNanoTrainingSchool
[1]Aptula AO,RobertsDW(2006)Mechanisticapplicabilitydomainsfornonanimalbasedprediction
oftoxicologicalendpoints:Generalprinciples
andapplicationtoreactivetoxicity.Chem ResTox.19:10971105

Applications ofQSAR
There are a large number of applications of these models within
industry, academia and governmental (regulatory) agencies.

The estimation ofphysicochemical properties,biological

activities and understanding the physicochemical features
behind abiological response indrug desing.

Therationaldesignofnumerousotherproductssuch
assurfaceactiveagents, perfumes,dyes,andfine
chemicals.

Thepredictionofavarietyofphysicochemical
propertiesofmolecules.

The prediction offate ofmolecules which are released

into the environment.

Theidentificationofhazardouscompoundsatearly
stagesofproductdevelopment,thepredictionof
QualityNanoTrainingSchool
toxicitytohumans
and environment.

10

What isrequired for agood QSARModel?

The QSARmodelshould meet the
requirements ofthe OECDprinciples [1]:

2
3
4
5

adefinedendpoint;
anunambiguousalgorithm;
adefineddomainofapplicability;
appropriatemeasuresofgoodnessoffit,
robustnessandpredictivity;
amechanisticinterpretationifpossible.

Common QSARModelling
Errors
Uninformativedescriptors
Poordescriptorselectionand
chancecorrelations
Modelling complex,nonlinear
structureproperty
relationshipswithlinearmodels
IncorrectlyvalidatingQSPR
models
Notunderstandingthedomain
ofapplicabilityofmodels

[1]OECD,GuidanceDocumentonthe Validationof(Quantitative)Structure ActivityRelationshipsModels, Organisation forEconomicCooperation andDevelopment

QualityNanoTrainingSchool
11
(2007).

Methods
QSARModelling process consists of5main steps.
Main Steps
Molecular Structure
Representation

Molecular Descriptors

Predictive Model

ModelValidation

Applicability Domainof
QSAR

Begins with the selection ofmolecules to be

used
Selection ofdescriptors;numerical
representer ofmolecular features (e.g.number
ofcarbon)
Original descriptor pool must bereduced in
size
Modelbuilding
The reliability ofthe modelshould betested
Types ofMolecular Descriptors
(topological,geometric,electronic and hybrid)
1DDescriptors
2DDescriptors 3DDescriptors

QualityNanoTrainingSchool

12

Molecular Descriptors[1]
Molecular
Descriptors
ca

are numerical values that characterize properties ofmolecules.

calculated by

Mathematical
formulas
Experimental
measurements

Theoretical
Indexes
Computational
algortihms

Different Theories
(e.g.Quantum
chemistry,
information theory,
organic chemistry)

derived
from

More than 5000[2]

Molecular
Descriptors
processed
by

statistics
chemometrics
chemoinformatics

applied
in

Scientific Fields (e.g.QSAR,

medicinalchemistry,drug
design, toxicology,analytical
chemistry,
Environmetrics, virtual
screening)

QualityNanoTrainingSchool
[1]T.Puzyn etal.(eds.),RecentAdvancesinQSARStudies,103125. DOI10.1007/9781402097836_4,CSpringerScience+Business
MediaB.V.2010
[2] Todeschini R,Consonni V(2000)Handbookofmoleculardescriptors.WileyVCH,Weinheim

13

DataAnalysis Methods[1]
Different Statistical Methods have been used inQSARfor the extraction ofuseful information
from the data.
RegressionProblems

DataCollection

Multiplelinearregression
Partialleastsquares
Feedforwardbackpropagationneuralnetwork
Generalregressionneuralnetwork

ClassificationProblems

Molecular
Descriptor
DataAnalysis
ModelValidation

LinearDiscriminant analysis
Logisticregression
Decisiontree
Knearestneighbor
ProbabilisticNeuralNetwork
Supportvectormachine

RecentlydevelopedMethods

Kernelpartialleastsquare
Robustcontinuumregression
Locallazyregression
Fuzzyintervalnumberknearestneighbor
Fastprojectionplaneclassifier

QualityNanoTrainingSchool
[1] Ekins, S. (2007). Computational Toxicology: Risk Assessment for Pharmaceutical and Environmental Chemicals. New Jersey: Wiley-Interscience.

14

DataMiningMethods
arelativelynewfieldwhichsharesometechniqueswithtraditional
dataanalysisbutalsobringsalotofnewvisionsandtechniques.
Value

Decision

Knowledge

Information

Data Data Data Data

Volume
DM is the discovery of useful Information
and Knowledge from a large collection of
Data, in order to help the Decision
making

Data:recordsofnumerical
data,symbols,images,
documents
Knowledge:
Rules:IF..THEN..
Causeeffectrelationships
Decisiontrees
Patterns:abnormal,normal
operation
Predictiveequations

QualityNanoTrainingSchool

15

DataMiningMethods
Data mining shares many techniques and tools with traditional data
analysis (such as the ones listed below). But it emphasises the analysis
of very large databases, and can deal with complicated cases which
cannot be handled by traditional methods.

Clustering

Summarisation

Classification

Regression

Conceptual Clustering

Case-based Learning

Inductive learning
Data mining research has also generated some new techniques that
were not seen in traditional data analysis, such as Link Analysis
QualityNanoTrainingSchool

16

NewDMTechniqueExample1:LinkAnalysis
Example :Tendatacases,threevariables,x1,x2,andx3
How canweautomaticallyuncoverthedependencyrelationshipbetweenthem?
WHICHRELATIONSHIP?

DATASET

x1

x2

x3

1
1
0
1
0
0
1
0
1
0

0
1
0
1
0
1
1
0
1
0

0
1
1
1
0
1
1
0
1
0

x1

x2

x3

x2
x1
x3

Thisrelationship
means,x2 is
dependent onx1,and
x3isdependent onx2,
butx1isindependent
ofx2andx3

Thisrelationship
means,x2is
dependent onx1,x3
isalso dependent on
x1,butx2andx3are
independentofeach
other
17

QualityNanoTrainingSchool

New DM Technique Example 2: Inductive Data Mining for Automatic Generation

of Decision Trees from Data

x1

x2

x3

x4

x1

x2

x3

x4

55.33

1.72

54

1.66219

92.19

71.31

3.44

55

1.60325

90.51

59.13

1.2

53

1.58399

92.74

72.3

4.02

55

1.66783

90.24

57.39

1.42

55

1.61731

91.88

68.81

6.88

55

1.69836

91.01

56.43

1.78

55

1.66228

92.8

66.61

2.31

52

1.77967

91.9

55.98

1.58

54

1.63195

92.56

63.66

2.99

52

1.81271

91.92

56.16

2.12

56

1.68034

92.61

63.85

0.24

50

1.81485

92.16

54.85

1.17

54

1.58206

92.33

67.25

53

1.72526

91.36

52.83

1.5

58

1.54998

92.22

67.19

52

1.86782

92.16

x1,x2,x3aredifferentmeasuresoffeedcomposition
x4isthelogofacombinationofoperatingconditions
y productqualitymeasure
y=good if>92,
=low ifbetween9192,
=verylow if<91

QualityNanoTrainingSchool

Aim:tofindwhichfeedcomposition
oroperationalconditionvariables
havethemostimportantimpacton
productquality
18

Automaticallygenerateddecisiontree
Productqualityismainlydeterminedbyx1andx4,almosthasnothingtodowithx2andx3
Ifx1<55.4,theprobabilitythatproductqualityisgood(y>92)ishigh
Ifx1>55.4butx4>1.8,theprobabilitythatproductqualityisgoodisalsohigh
Othertworegions,product qualityiseitherloworverylow

x1

Verylow

>55.4

<55.4

Good

2.0

>1.8

<1.8

x1
<70

91<y<92

Low

Good

x4

x4

y>92

Low

Y>92
>70

1.5

Good

Y<91
Verylow

50

60

70

Ifx1=[44.6,55.4],thenconditionalprobabilityofy=3is0.9
QualityNanoTrainingSchool
Ifx4=[1.8,2.3],thentheconditionalprobabilityofy
=3is1.0

x1

19

(1)Generationofapopulationofsolutions
<s >s

<s >s

<s >s

<s >s

Automatic
generationof
decisiontreefrom
data

This isanew
.. .... .. .. .... .. .... .. .. .... ..
approoach
(2) Repeatsteps(i) and(ii) untilthestopcriteriaare based onagenetic
programming
method

satisfied:
(i)calculatethefitnessfunctionvaluesforeach
solutioncandidate
(ii)performcrossoverandmutationtogenerate
thenextgeneration
<s >s

<s >s

<s >s

.. .... .. .. .... .... .. .... ..

<s >s

.... ..

(3) thebestsolutioninallgenerations:thefinalsolution
QualityNanoTrainingSchool

20

New DM Technique Example

Two Data Sets
Data Set 1:
Concentration lethal to 50% of the population, LC50,
1/Log(LC50), of vibrio fischeri, a biolumininescent bactorium
75 compounds

1069 molecular descriptors calculated by

molecular modelling software DRAGON

Data Set 2:
Concentration effecting 50% of the population,
EC50 of algae chlorella vulgaris, by causing fluorescein
diacetate to disappear
80 compounds

1150 descriptors
QualityNanoTrainingSchool

21

YAdaptforQSARToxicityPrediction:DataSet1
Atreeforthebacteria
datausingthree
classes.

Amethod to classify toxicity endpoints into classes

SampletreefromYAdapt

Forbacteriadata:
of 1069 descriptors,
only5descriptorsare
important! Usingthe5
descriptors,thetree
modelcanpredict with
the accuracy>96.7%

Polarsurfacearea40.368
No

Yes
Gearyautocorrelationlag4
weightedbyatomicmass
0.007
Yes
Class1
(21/21)

Gearyautocorrelationlag4
weightedbypolarizability
1.181

No
Hautocorrelationlag0weighted
bySandersonelectro
negativities2.111
Yes
Class1
(8/8)

No

Yes
Class3
(9/10)

No
Class2
(9/9)
QualityNanoTrainingSchool

BrotoMoreauautocorrelation
lag2weightedbypolarizability
4.085
Yes
Class2
(5/6)

No
Class3
(6/6)
22

YAdaptforQSARToxicityPrediction :DataSet1
Class1:
Log(1/EC50)4.0829cases
Class2: 4.08<Log(1/EC50)4.5116cases
Class3: 4.51<Log(1/EC50)4.999cases
Class4:
otherwise6cases

A decision tree from generation 46 for the bacteria data, finding four
classes from the continuous endpoint itself during induction.
Training accuracy: 93.3%
Test accuracy: 80.0%
Polarsurfacearea40.368
Yes

No
3DMoRSEsignal15weighted18weightedby
Sandersonelectronegativites0.644

Hautocorrelationlag0weightedby
Sandersonelectronegativities2.111
Yes
Class1
(16/16)

Gearyautocorrelationlag4weightedby
atomicmass0.007
Yes
Class1
(13/13)

No

Yes

No

Class2
(5/7)

No
Class2
(9/9)
QualityNanoTrainingSchool

Hautocorrelationlag0weightedby
Sandersonelectronegativities2.394
Yes
Class3
(7/9)

No
Class4
(6/6)
23

YAdaptforQSARToxicityPrediction :DataSet2
Solvationconnectivityindex 2.949
No

Yes
Selfreturningwalkcount
order83.798
Yes
Class1
(16/16)

No
Class2
(14/15)

2nd dataset algaedata

Ofthe1150descriptors,only5
descriptorsareimportant!

Molecularmultiplepathcount
order392.813
Yes

No

Class3
(6/8)

Hautocorrelationoflag2weighted
bySandersonelectronegativities
0.401
Yes
Class4
(6/7)

No

Training:92.2%

2ndcomponentsymmetry
directionalWHIMindex
weightedbyvanderWaals
volume0.367

Test:81.3%

Yes
Class3
(9/10)
QualityNanoTrainingSchool

No
Class4
(8/8)
24

IntegratedProcessDataMiningEnvironment
Descriptor
calculation

DataPreprocessing
Scaling
Missingvalues
Outlieridentification
Featureextraction

DataMiningToolbox
Regression
PCAandICA
ART2networks
Kohonennetworks
Knearestneighbour
Fuzzycmeans
Decisiontreesandrules
Feedforwardneural
networks(FFNN)
Summarystatistics
Visualisation
MixtureQSARmodels

DiscoveryValidation
Statisticalsignificance
Resultsfortrainingand
testsets

Dataimport
Excel
ASCIIFiles
Database
XML

Results
Presentation
Graphs
Tables
ASCIIfiles
QualityNanoTrainingSchool

Different datamining
techniques canbe
combined intoasingle
25
system

QSARModelValidation
Crossvalidation (Q2)

Model
Validation
Model
Applicability
Model
Interpretation

Kfold Cros Validation

Leaveoneout
Cross Validation

The statistical fitofmodelR2

StandartError of
Predicton (S)

Modelvalidationisrequiredtoensure
modelreliability (quality ofthe model)
Ensures thatthemodelisnotdueto
chancefactors!

The statistical fitofmodelisusually performed

Cross validation:dividing adatasample into subsets,performing the analysis ona
single subset,using other subsets to confirm and validate the initial analysis
There isnoagreed method for the validation ofQSARModels
QualityNanoTrainingSchool

26

QSARWorkflow[1]

[1] Tropsha,A.,Gramatica,P.,Gombar,V. (2003) Theimportanceofbeingearnest:ValidationistheAbsoluteEssentialforSuccessfulApplicationandInterpretationofQSPR

QualityNanoTrainingSchool
27
Models.Quant.Struct.Act.Relat.Comb.Sci.22,6977.

QSARModelApplicability Domain
Descriptor 1

.
...
..
.
. . TRAININGSET
.
.
. . Descriptor 2
......
. .
...
Inside Applicability
Domain

Outside Applicability
Domain

will bepredicted

will notbepredicted

Modelapplicabilityallowsusto
decidewhetherthemodelwill be
usefulfornewdata (correct use of
the model).
Finding the boundaries ofthe
modelto see how well itwill work
for other compounds.
Notevenasignificant and
validatedQSARmodelcanbe
used for the entireuniverse of
chemicals[1]!

[1] Gramatica P. (2007)PrinciplesofQSARmodelsvalidation:internalandexternal.QSARandCombinatorialScience; 26:694701.

QualityNanoTrainingSchool

28

QSARTools for Toxicity Prediction

1.

TOPKAT(byKomputer AssistedTechnology):
ItusesQSARstogivenumericpredictions,
basedonseveral descriptors
2. HazardExpert (by CompuDrug ):
HazardExpert,producedbyCompuDrug,isa
rulebasedsystem.Therulesuse
toxicophores,orstructuralalertsderivedfrom
QSARs
3. CASE: It canderive QSARmodelfrom users
6.DEREK:DeductiveEstimation
data.
ofRiskfromExistingKnowledge
4. ECOSAR:It usesQSARstopredicttheaquatic
(DEREK)ismarketedbyLHASA
toxicityofchemicalsforavarietyoforganisms
Ltd.It isanexpertknowledge
5. OASIS: Ituseshierarchicalrulesbasedon
basesystemthatpredicts
QSARs,whichcanbeuserderived.
whetherachemicalistoxic

Each ofthese QSARtools have their own advantages and weaknesses for toxicity
predictions
There are also several available softwarepackages for descriptor calculations (e.g.
DRAGON)
QualityNanoTrainingSchool
29

Challenges for QSAR[1]

If there isameasurement error inthe experimental data,itisvery likely

that false correlations may arise

Ifthetrainingdatasetisnotlargeenough,thedatacollectedmaynotreflect
thecompletepropertyspace.Consequently,manyQSARresultscannotbe
usedtoconfidentlypredictthemostlikelycompoundsofthebestactivity.

Therearemanysuccessfulapplicationsofthismethodbut onecannot
expecttheQSARapproachtoworkwellallthetime. QSAR Modelsareeasy
tobuildbutalsoveryeasytogetwrong.

QualityNanoTrainingSchool
30
[1]Puzyn,T.,Leszczynski,J.(2012) TowardsEfficientDesigningofSafeNanomaterials:InnovativeMergeofComputational
ApproachesandExperimentalTechniques,The
RoyalSocietyofChemistry.

Part II
Nanotoxicity and
NanomaterialCharacterization
QualityNanoTrainingSchool

31

Macroconcernswithnanoworld
!
Macroconcernswithnanoworld !
Uncertainities atthe early stages ofanew technology bring new concerns!

NewTechnology

NewBenefits
NewConcerns

[1]

Safetytohumanhealthand environment
Suitability ofriskmanagementstrategies

More than 800nanoproducts

Nostandardizedor
validatedmethods
fornanotoxicity
testing

Available toxicity
datafor NPs are
inconsistent and
confusing

Blocks the
developmentof
ENPsriskreduction
strategies[2]

[1] Chaudhry, Q. et al. (2010). Knowns, Unknowns, and Unknown Unknowns . RSC Nanoscience & Nanotechnology (p. 201-218).
Royal Society of Chemistry.
[2]Tran, L., & Chaudhry, Q. (2010). Engineered Nanoparticles and Food: An Assessment of Exposure and Hazard. RSC Nanoscience &
QualityNanoTrainingSchool
32
Nanotechnology (p 120-134). Royal Society of Chemistry.

[1]
Riskassesment
for
NMS
Riskassesment for NMS
[1]

ENMs are already used inanumber ofcommercial applications.

Potential safety issues ofNMs should beaddressed atthe same timeasthe

tech.isdeveloping
Whatquantitative
Whatdangersto
connectionsexist
human healthmay
betweenthedose
basicallyarisefrom
2.Dose
andtheextentofthe
1.Hazard
theharmful agent ?
response
Identification
expected effect?
relationship

4.Risk
Whatisthenature
andmagnitude ofthe assessment
riskandhow
accuratelycanitbe
estimated?

3.Exposure
assessment

Towhatextentis
therelevant
populationgroup
exposed tothe
harmful agent?

QualityNanoTrainingSchool
OECD (2003): Environment Directorate. Description of Selected
Key Generic Terms Used in Chemical / Hazard Assessment. OECD 33
Series on Testing and Assessment Number 44. ENV/JM/MONO(2003)15. Paris, 30-Oct-2003.

Doseresponse relationship
Theevaluationoftoxicityincludes twosteps:

Hazard Identification
Evaluation oftoxicity
Doseresponse evaluation

involves determiningtheconditionsorlevelsatwhichthepresenceof
acontaminant maytriggerabiologicalresponsefromthebody
Different amount of
CNTdust

At level of0.1mg/m3

Noeffects were
observed

Athigher level of
0.5mg/m3

Sideeffects were
detected in the
lung 34

QualityNanoTrainingSchool

Naidu, B. David (2009). Biotechnology and Nanotechnology - Regulation under Environmental, Health, and Safety Laws.. Oxford University Press.

Particle Characterization
Theonlyreasonable waytoobtain toxicity information for thelarge
numberofnanoparticleswithout testing everysingleone istorelate
thephysicochemicalcharacteristicsofNPs withtheirtoxicityinaSAR
(StructureActivityRelationship)orQSARmodel.
Nanomaterialcharacterization isone ofthe first steps intoxicity test.
However,itisavery challenging issue!!!
NanoparticleStructure
Properties
ParticleSize
SurfaceArea
Morphology
StateofDispersion,

Toxicity

QualityNanoTrainingSchool

35

more research is required to reach a clear understanding of the

Possible toxicityrelated Physicochemical Properties

When amaterial isinananosized form,its properties become different from the
same material inabulk form

Fullerenes
CarbonNTs
Graphene

Carbonbased
Nanomaterials

Different and
Exceptional
Properties

Nocertain information!!!
Recent studies showed that toxicity ofNPs canberelated to:

Sizeandshape

Sizedistribution

Agglomeration state

Porosity

Structuredependentelect.
Electronicproperties

Characterisation:

Inwhatmedium?

Surfacearea
Surfacechemistry
Surfacecharge
Crystalstructure
Composition
Configuration
Coating
Aggregationstate
Metalcontent

POSSIBLE factors[1]
inducingtoxicityof
chemicals
moreresearchisrequired
toreachaclearunderstanding

QualityNanoTrainingSchool
[1]

Puzyn, T., Leszczynska, D. & Leszczynski, J. Toward the development of Nano-QSARs: advances and challenges. Small 5, 24942509 (2009).

36

NanomaterialCharacterization
Several different methods and instruments are used to determine the properties onNMs
Shapeandaggregation
state

HighResolutionMicroscopy

Composition,purity
andsurfacechemistry

Spectroscopyand
Chromatography Methods

Surfacecharge

ZetaPotentialAnalysis

Crystalstructure

XRayDiffraction

Particlenumberand
sizedistribution

SeveralTechniques(e.g.
SEM,TEM,DLS)

SurfaceArea

BrunauerEmmettTeller
Adsorption

There isnosingle instrument that isright tool for every test.There are infact more than 400
QualityNanoTrainingSchool
37
different techniques
for particle counting and characterization

Whathaven'tweconsideredyet ?
SomeNPs
Have atendency toformclusters (e.g.aluminium oxide)
inanagglomeration state,someNPsbehavelikelargerparticles
inthemedia
Toxicitymayalsodependsonthesizeofagglomerate

NOTonoriginalparticlesizeitself
!!!
If the NPs aggregate

Howtocharacterizeaggregation?
itisimportant toidentifyhow
Textureanalysis?
thesechangesaffecttoxicity
38

Hussain, S. e. (2008). Can silver nanoparticles be useful as potential biological labels? Nanotechnology 19 .

Challenges
for nanoQSAR?
ChallengesforQSAR
Table shows the results of3different toxicity studies for silver NMs:
NP Size
type (nm)

Doses
Cell
(mg/ml)

Ag

720

0.850

Human
fibrosorcoma and
skincarcinoma

Ag

<15

0.7

Human hepatoma

Ag

15
and
100

10250

Rat liver deliverd cell

line

Differencesin
NMsused
(size)
[1]

Differencesin
experimental
conditions

Method

Result

XTTassay

Nontoxic

MTTAlamarblue Oxidativestres
andLDHassay
mediatedtoxicity

Kim etal.
(2009)[2]

MTT,LDH,GSH
level,ROSand
MMP

Hussain etal.
(2005)[3]

Toxic

Araro etal.
(2008)[1]

strongly affects
the results

Arora S, Jain J, Rajwade JM, Paknikar KM (2008). Cellular responses induced by silver nanoparticles: in vitro studies. Toxicol Lett 179:93100
Kim S, Choi JE, Choi J, Chung KH, Park K, Yi J, Ryu DY (2009). Oxidative stress-dependent toxicity of silver nanoparticles in human
hepatoma cells. Toxicol In Vitro 23:10761084
[3] Hussain SM, Javorina AK, Schrand AM, Duhart HM, Ali SF,
QualityNanoTrainingSchool
39
Schlager JJ (2006). The interaction of manganese nanoparticles with
PC-12 cells induces dopamine depletion. Toxicol Sci 92:456463
[2]

Challenges for nanoQSAR[1]

NMs are problematic for QSARmodellers because ofthe several reasons:

Lackofhighqualityexperimentaldata

Lackofconceptualframeworksforgrouping NPs accordingtomodeof

physicochemicalpropertiesandtoxicaction

Lack ofsufficient moleculardescriptors abletoexpressspecificityof

nanostructure

Lackofrationalmodelingprocedurestoscreenlargenumbersof
structurallydiversifiednanoparticles.

Limited knowledgeontheinteractionsbetweenNPs andbiological

systems
QualityNanoTrainingSchool

[1]T.Puzyn etal.(eds.),RecentAdvancesinQSARStudies,103125. DOI10.1007/9781402097836_4,CSpringerScience+Business MediaB.V.2010

40

Part III
Case Study
despite all limitations
QualityNanoTrainingSchool

41

Nanotxicology 2012,47September,Beijing
QSARToxicity(invitro)AnalysisforaPanelofEighteen
NPs

The aim of this study was to test the hypothesis that nanoparticle
toxicity is a function of some measurable physical characteristics and
that Structure Activity Relationship (SAR) might ultimately be a useful
tool for nanotoxicology. A panel of 18 nanoparticles was chosen to
include carbonbased materials and metal oxides. Comparative toxicity
of the nanoparticles in the panel was assessed, using a number of
different measures of apoptosis and necrosis; haemolytic effects and
the impact, on cell morphology were also assessed. Structural and
composition properties were measured including size distribution,
surface area, morphological parameters, metal concentration,
reactivity, free radical generation and zeta potential. Toxicity end
points were processed using principal component analysis and
clustering algorithms.
QualityNanoTrainingSchool

42

The Set ofEighteen NPs

Nanooxicology 2012,47September,Beijing

Particles

Qualifier

Carbonblack
Dieselexhaustparticles
Nanotubes
Fullerene

Printex 90
EPA
Japanese

Number

N1
N2
N3
N4
Unmodified
N5
Polystyrenelatex
Amine
N6
Carboxylated
N7
7nm
N8
Aluminiumoxide
50nm
N9
300nm
N10
Ceriumoxide
N11
Nickeloxide
N12
Siliconoxide
N13
Zincoxide
N14
Rutile
N15
Titaniumdioxide
Anatase
N16
N17
QualityNanoTrainingSchool
Silver
*AHT:AlliedHighTech,**NAM:NanostructureandAmorphousMaterialInc
N18

Supplier
Degussa
EPA
VickiStone
Sigma
Polysciences
Sigma
Polysciences
KrahnChemie
AHT*
AHT*
NAM**
NAM**
NAM**
NAM**
NAM**
NAM**
NAM**
NAM**

Form
Dry
Dry
Dry
Dry
Sus.
Sus.
Sus.
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
43
Sus.

CharacterisationofPhysicochemicalProperties

Nanotoxiclogy 2012,47September,Beijing

Particle shape was analysed using LEO 1530 Scanning Electron Microscope (SEM) or Philips
CM20 Transmission Electron Microscope (TEM) aspect ratio and mean size
Surface area and porosity were measured using TriStar 3000 BET. Thirteen measurements,
including five surface areas based on different definitions, three pore volumes, three pore
sizes as well as the mean size and particle density.
Particle size and size distribution were analysed using a Malvern MasterSizer 2000. Size
distribution, and seven other size properties (mass diameter, uniformity, specific surface
area, surface area mean diameter and three mass diameters)
ThefreeradicalactivitiesweremeasuredbyEPR(Electronparamagneticresonance)
using the spin traps, DMPO and Tempone H separately. Tempone H has partial selectivity for
trapping superoxide anions, whilst DMPO traps mainly hydroxyl radicals
Particle reactivity in solution, the dithiothreitol (DTT) consumption test as a reducing species
was carried out. Since the DTT consumption test can only be conducted on dry powders, only
fourteen of the panel of nanoparticles were assessed using this assay.
Metal Concentration water soluble concentration of ten heavy metals (Ti, V, Cr, Mn, Fe, Co,
Ni, Cu, Zn and Cd) was measured
Charge:zpotentialwasmeasuredusingMalvernInstrumentsZetasizer Nano instrument
QualityNanoTrainingSchool

44

TotalNumber ofMeasurements
13 BETmeasurements(for14drysamples)

2 SEM/TEMmeasurements (forallsamples)
7 laserdiffractionsizestatisticalmeasurements(MasterSizer 2000)forallsamples
84 laserdiffractionsizedistribution measurements
2 EPRmeasurements(forallsamples)
1 reactivitymeasurement (fordrysamples)
10 metalcontentmeasurements
Total:119Measurements
Particle
name

BET Measurements

M2000
Measurements

SEM Measurements

Zeta measurements

Other
Measurements

N1

A(1,1),A(1,2),,A(1,12)

B(1,1),B(1,2)

C(1,1),C(1,2)

D(1,1),D(1,2)

E(1,1),E(1,2)

N2

A(2,1),A(2,2),,A(2,12

B(2,1),B(2,2)

C(2,1),C(2,2)

D(2,1),D(2,2)

E(2,1),E(2,2)

N18

A(18,1),A(17,2),

QualityNanoTrainingSchool

B(18,1),B(18,2)

C(18,1),C(18,2)

D(18,1),D(18,2)

E(18,1),E(18,2)

45

SEMandTEMimagesofthe18nanoparticlesamples
Nanotoxicolgy 2012,47September,Beijing

QualityNanoTrainingSchool

46

Table2.Thewatersolubleconcentrationsofvariousmetalsforthe18nanoparticlesamples

Theheavy metalconcentration ofeach NP

Nanotoxicology 2012,47September,Beijing

Particlenames

MetalConcentration(g/g)
Ti

Cr

Mn

Fe

Co

Ni

Cu

Zn

Cd

CarbonBlackN1

0.000

0.057

0.000

0.305

0.000

0.002

0.263

1.161

9.687

0.008

DieselExhaustN2

2.320

0.312

16.47

20.81

208.4

0.508

4.940

2.235

3181

0.387

JapaneseNanotubesN3

0.000

0.100

0.000

0.161

13.10

0.000

0.299

0.123

0.460

0.001

FullereneN4

0.000

0.084

0.000

0.000

0.000

0.000

0.534

0.142

4.837

0.000

PolystyreneLatexBeadsN5

0.000

1.172

0.000

0.000

0.000

0.000

0.000

0.625

19.56

0.000

PolystyreneLatexBeadsN6

4.368

0.779

0.343

0.199

0.000

0.006

0.440

18.23

58.71

0.010

PolystyreneLatexBeadsN7

0.000

1.018

0.000

0.000

0.000

0.000

0.000

2.588

22.323

0.000

Aluminuim OxideN8

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.714

0.000

0.000

AluminuimOxideN9

0.000

0.031

0.000

0.000

0.000

0.000

0.056

0.099

4.987

0.000

AluminuimOxideN10

0.000

0.062

0.000

0.000

0.000

0.000

0.023

0.088

0.000

0.000

CeriumOxideN11

0.000

0.004

1.058

0.128

0.000

0.000

0.616

0.191

112.9

0.007

NickelOxideN12

0.000

0.056

0.000

0.000

0.000

0.042

16963

0.286

0.000

0.038

SiliconOxideN13

0.000

0.088

0.000

0.000

0.000

0.000

14.163

0.106

0.000

0.000

ZincOxideN14

0.000

0.096

0.000

0.000

0.000

0.000

1.336

0.047

8185

9.711

TitaniumDioxideRutile N15

0.000

0.083

0.000

0.000

0.000

0.000

0.286

0.276

39.623

0.000

TitaniumDioxideAnatase N16

0.000

0.104

0.000

0.160

0.000

0.006

0.692

0.303

402.3

0.000

SilverN17

0.000

0.111

0.000

0.503

0.000

0.000

0.475

0.597

17.867

0.019

SilverN18

0.000

0.096

0.000

0.256

0.000

0.059

0.561

162.8

194.8

0.037

QualityNanoTrainingSchool

47

CharacterisationofParticleSizeand Shape

Nanotoxicology 2012,47September,Beijing

CarbonBlackN1

LEO1530GeminiFEGSEM
ScanningElectronMicroscopy

TheinformationaboutSize,
Shape canbeobtainedfrom
SEMimages

MeansizeobtainedbymeasuringtheparticlesontheSEM/TEMimage:Mean_Size_SEM
AspectratioobtainedfromSEM/TEMimages:Aspect_Ratio
QualityNanoTrainingSchool

48

BETMeasurements for 14dry powder samples

Nanotoxicology 2012,47September,Beijing

Micromeritics TriStar3000
SurfaceAreaTechnique:GasAdsorption
MinimumSurfaceArea:0.01m2/g
SupportGases:NitrogenandHelium
MeasurementParameters:BETMeasurement
AdsorptionIsotherms
DesorptionIsotherms
tplots
PoreVolumeAnalysis
PoreSizeAnalysis

Surface Area (m2/g)

Pore Volume (cm3/g)

Single point surface area at P/Po =

0.197 (m/g) SPSA

BET Surface Area: BET_SA

Langmuir Surface Area: LM_SA

BJH Adsorption cumulative surface

area of pores between 17.000 and
3000.000

diameter
:

BJH_AC_SAOP

BJH Desorption cumulative surface

area of pores between 17.000 and
3000.000

diameter:
BJH_DC_SAOP

Single point adsorption total pore

volume of pores at P/Po = 0.98:
SP_TP_VOP

Pore Sizes ()

Adsorption average pore width:

BET_PW

BJH Adsorption average pore

diameter : BJH_A_PD

BJH Adsorption cumulative volume

of pores between 17.000 and

3000.000 diameter:
BJH_AC_VOP
BJH Desorption cumulative volume

of pores between 17.000 and

3000.000 diameter :
BJH_DC_VOP

QualityNanoTrainingSchool

BJH Desorption average pore

diameter: BJH_D_PD
Mean size calculated
from BET surface area:
Mean_Size_BET
Particle density:
Density

**14drypowdersamples,noavailableforthreesuspensions(N5,N6,andN7)andthenewSilversample.

49

CharacterisationofNPsMastersizer

Nanotoxicology 2012,47September,Beijing

10
9
P r o b a b ilit y ( % )

8
7
6
5
4
3
2
1
0
0.01

0.1

10

100

1000

10000

Size (micron)

MalvernMastersizer 2000
TypeofAnalyser:LaserDiffraction
SampleVolume:1000ml,150ml(HydroGandS)
MeasurementRange:20nm 2000m
QualityNanoTrainingSchool

Theinformationaboutmean
SizeandSizedistribution,can
beobtainedfromthissizer.

50

CharacterisationofNPsMastersizer

Nanotoxicology 2012,47September,Beijing

Mastersizer 2000Measurements:91**

D[4,3] theVolumeWeightedMeanorMassMomentMeanDiameter,represent
byD[4,3]
Uniformity:ameasureoftheabsolutedeviationfromthemedian,representedby
Uniformity
Specificsurfacearea:thetotalareaoftheparticlesdividedbythetotalweight,
SSA_2000
D[3,2] theSurfaceWeightedMeanorSurfaceAreaMomentMeanDiameter,
representedbyD[3,2]
D(v,0.5) thesizeinmicronsatwhich50%ofthesampleissmallerand50%is
larger.ThisvalueisalsoknownastheMassMedianDiameterorthemedianofthe
volumedistribution,representedbyD(0.5)
D(v,0.1) thesizeofparticlebelowwhich10%ofthesamplelies,representedby
D(0.1)
D(v,0.9) thesizeofparticlebelowwhich90%ofthesamplelies,representedby
D(0.9)
Sizedistribution:thereare100values.Amongthemthereare16valuesequalto
zeroforallparticles.So84 sizedistributionattributesareusedforfurtheranalysis
** Note:availableforallsamples
QualityNanoTrainingSchool

51

CharacterisationofNPsZeta Potential

Nanotoxicology 2012,47September,Beijing

MalvernZetaSizerNano
instrument

Sizeranges:0.6nm 6m
ZetaPotential
Suitable:suspensions
QualityNanoTrainingSchool

52

Toxicity Results
Cytotoxicity byLDHRelease

Viability

25.00
N6

80.00

40.00

100.00

20.00

Apoptosis

60.00

120.00

N12

N14

Viable

100.00
LDHRelease

Apoptosis

80.00

15.00
60.00
10.00

N6
40.00
N14

20.00

20.00

5.00

N6
0.00
31.25

62.50

125

250

Dose(g/ml)

0.00

0.00
31.25

62.50

125

Proinflammation effects

80

62.50

125

250

Dose(g/ml)

Dose(g/ml)

Necrosis

100

Haemolysis

90
Haemolysosatdose500g/ml

70
N6

60
Necrosis

31.25

250

50
40
N14

30
20
N3

10

80
70
60
50
40
30
20
10

0
31.25

62.50

125

Dose (g/ml)

250

QualityNano Training School

53

Principal Component Analysis ofCytotoxicity Data

LDH

Viability

Apoptosis

MTT
TotalToxicity

Necrosis
Haemolysis

QualityNanoTrainingSchool

54

Principal Component Analysis ofCytotoxicity Data

Themajorclustercontainslowtoxicityparticle
samples

Zincoxide

Aminated beads(N6),zincoxide(N14),Japanese
Nanotubes (N3)andnickeloxide(N12)areseparate
fromthemajorcluster.
Aminated beads(N6)hasthehighesttoxicity
valuesinnearlyallassayresults(LDH,apoptosis,
Japanesenanotubes

necrosis,haemolytic,MTTandcellmorphology
assays).
Polystrene larex (Amine)

Zincoxide(N14)hashightoxicityvalueinLDH,
apoptosis,necrosisandinflammationassays.

Nickeloxide

Japanesenanotubes (N3)havehightoxicityvalues
inviabilityandMTTassays.
Figure: Plotofthefirstandsecond PCs based
onPCAanalysisoftoxicitydata

Nickeloxide(N12)hasshownhighlytoxicinLDH
andhaemolyticassaysresults..

Thedistinctionofthefourparticles,N6,N14,N3andN12,fromother
particlesisclear.
QualityNanoTrainingSchool

55

NP6Average Toxicity Comparison

NP14Average Toxicity Comparison

N6 hasthehighesttoxicityvaluesinnearlyallassay
results (LDH,apoptosis,necrosis, haemolytic,MTT
andcellmorphologyassays).

N14 has high toxicity value in LDH, apoptosis,

necrosis and inflammation assays.

NP12Average Toxicity Comparison

NP3Average Toxicity Comparison

N12 hasshownhighlytoxicinLDHand haemolytic

assaysresults

N3 havehightoxicityvaluesinviabilityandMTT
assays.

The toxicity results ofN6,N3,N12and N14which are likely to betoxic

QualityNanoTrainingSchool

56

Next Step:Principal Component Analysis ofDescriptors

The focus of the (Q)SAR Analysis is to identify the possible structural and
compositional properties which may contribute to the toxicity of zinc
oxide (N14), polystrene latex amine (N6), Japanese nanotubes (N3) and
nickel oxide (N12).
The theory is that there should be structural or/and compositional
distinctions for these nanoparticles from the rest of the nontoxic
particles.
PCA and clustering were then applied to the physicochemical
descriptors with the hypothesis that the toxic nanoparticles might also
be discriminated from the nontoxic particles based on the analysis of
the measured physicochemical descriptors.
If this can be proved, further analysis can be conducted to identify the
key physicochemical descriptors that lead to the observed toxicity.
QualityNanoTrainingSchool

57

Principal Component Analysis ofDescriptors

Principalcomponentanalysisofthestructuralandcompositionaldescriptorsforthe panelof
18nanoparticles(excludingBETandDTTmeasureddescriptorssincetheyare notavailablefor
thefourwetsamples).
The scaling operation was performed,3PCs
were then applied.

N5
N6

PC1PC2PC3threedimensionalplotshows

N7

N3

thatthe18nanoparticleswereclearlygrouped

N2
N1 2

intoclusters.
N1 4

Thelargestclustercontainsparticlesthat
showedlowtoxicityvalues.
JapaneseNanotubes (N3),nickeloxide(N12)
andzincoxide(N14)anddieselexhaust

Figure : PCA analysis of the physicochemical

descriptors for the panel of 18 nanoparticles
(excluding BET and DTT measured descriptors
since they are not available for the four wet
samples).

particles(N2)areoutsidethislowtoxicity
samplecluster.
There mustbeanotherreasonthatleadsto
hightoxicityvalueforN6

QualityNanoTrainingSchool

58

Results

Toexaminetheclusteringresultsmoreclearly,PC1
PC2, PC1PC3andPC2PC3twodimensionalplotsare
alsoexamined .
ItisPC2thatseparates N3(Japanesenanotubes),N12
(nickeloxide)andN14(zincoxide) fromtherest
samples,ascanbeseenfromPC1PC2andPC2PC3
plots.
QualityNanoTrainingSchool

59

Results

The contribution plot for, PC1, PC2 and

PC3 is shown in the Figure, we are mainly
interested in the contribution to PC2 (solid
red

colour),

since

it

is

PC2

that

discriminated Japanese nanotubes (N3)

nickel oxide (N12) and zinc oxide (N14)
from the rest particle samples.
The largest contributions to PC2 are: aspect
ratio measured by SEM imaging, and volume
weighted mean, uniformity, and D(0.9), all
measured by Mastersizer, as well as Ni and
Zn metal contents
Figure: PC1,PC2andPC3ContributionplotsbasedonPCA
analysisofthestructuralandcompositionaldescriptors
forthepanelof18nanoparticles

Further analysis is required.

QualityNanoTrainingSchool

60

Results

N3

Inordertodetermineexactlywhich
physicochemicaldescriptors leadto
hightoxicityvaluesofN3,N12,N14
N14

and N6,several PCanalysis were

N2

performed.

N 12

Forthe14dryparticles,PCanalysis
withBETandDTTmeasurements gave
similarresultwith the first PCanalysis
inwhich BETand DTTdatawas
Figure: Principalcomponentanalysisofthestructural

excluded.

andcompositionaldescriptors(allsuchdescriptors
includingBETandDTTmeasurements)forthe14dry
nanoparticles.
QualityNanoTrainingSchool

61

Results
Figure: PC1andPC2contributionplotsforPCA
analysisof thecompositionaldescriptors

To identify the relative

contribution of each metal
concentration to PC1 and
PC2, the contribution plot to
PC1 and PC2 is shown in the
Figure.
NicontentmadethemostimportantcontributiontoPC1,suggestingthatnickelcontent
maybethemainreasonforthehightoxicityof N12(nickeloxide).
While Zncontentmadethemostobvious contributiontoPC2, suggesting thatzinc
contentmaybethereasonfor the high toxicity of N2(dieselexhaustparticles)andN14
(zincoxide).
QualityNanoTrainingSchool

62

Table2.Thewatersolubleconcentrationsofvariousmetalsforthe18nanoparticlesamples

Theheavy metalconcentration ofeach NP

Nanotoxicology 2012,47September,Beijing

Particlenames

MetalConcentration(g/g)
Ti

Cr

Mn

Fe

Co

Ni

Cu

Zn

Cd

CarbonBlackN1

0.000

0.057

0.000

0.305

0.000

0.002

0.263

1.161

9.687

0.008

DieselExhaustN2

2.320

0.312

16.47

20.81

208.4

0.508

4.940

2.235

3181

0.387

JapaneseNanotubesN3

0.000

0.100

0.000

0.161

13.10

0.000

0.299

0.123

0.460

0.001

FullereneN4

0.000

0.084

0.000

0.000

0.000

0.000

0.534

0.142

4.837

0.000

PolystyreneLatexBeadsN5

0.000

1.172

0.000

0.000

0.000

0.000

0.000

0.625

19.56

0.000

PolystyreneLatexBeadsN6

4.368

0.779

0.343

0.199

0.000

0.006

0.440

18.23

58.71

0.010

PolystyreneLatexBeadsN7

0.000

1.018

0.000

0.000

0.000

0.000

0.000

2.588

22.323

0.000

Aluminuim OxideN8

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.714

0.000

0.000

AluminuimOxideN9

0.000

0.031

0.000

0.000

0.000

0.000

0.056

0.099

4.987

0.000

AluminuimOxideN10

0.000

0.062

0.000

0.000

0.000

0.000

0.023

0.088

0.000

0.000

CeriumOxideN11

0.000

0.004

1.058

0.128

0.000

0.000

0.616

0.191

112.9

0.007

NickelOxideN12

0.000

0.056

0.000

0.000

0.000

0.042

16963

0.286

0.000

0.038

SiliconOxideN13

0.000

0.088

0.000

0.000

0.000

0.000

14.163

0.106

0.000

0.000

ZincOxideN14

0.000

0.096

0.000

0.000

0.000

0.000

1.336

0.047

8185

9.711

TitaniumDioxideRutile N15

0.000

0.083

0.000

0.000

0.000

0.000

0.286

0.276

39.623

0.000

TitaniumDioxideAnatase N16

0.000

0.104

0.000

0.160

0.000

0.006

0.692

0.303

402.3

0.000

SilverN17

0.000

0.111

0.000

0.503

0.000

0.000

0.475

0.597

17.867

0.019

SilverN18

0.000

0.096

0.000

0.256

0.000

0.059

0.561

162.8

194.8

0.037

QualityNanoTrainingSchool

63

Conclusions

IdentifyingthefactorthatdiscriminatesN6(polystrene latexamine)fromN5
(unmodified)andN7(carboxylated)
PCAanalysis:PCAanalysisofmetalcontentsonly,orPCAanalysisofstructuraldescriptorsonly,orPCA
analysisofbothstructuralandcompositionaldescriptorstogether,couldnotdiscriminateN5,N6andN7
fromthelargestclustersofnontoxicparticlesamples.

PCAConclusion:neitherstructuraldescriptorsnormetalcontentsmadeN6differentfroN5
andN7intoxicity
Itwasatthisstageofanalysisthatwesuspectedthatthereareotherpossiblereasonsfortoxicityspecific
toN6:eitheritwasduetothechargedifference,orduetoN6beingamine,orboth.
zetapotentialforallthreepolystyrenelatexsamples.
N6(amine)
37.8,37.5,and40.3(mV),
positive
N5(unmodified)
36.2,38.8and36.8,
negative
N7(carboxylated)
54.9,55.3,and58.6.,
negative
AccordingtoVerma etal.anyparticlewithpositivechargeislikelytointeractelectrostatically sincemostcell
surfacesandotherbiologicalmembranesarenegativelychargedunderphysiologicalconditions(Verma etal.
2008).

Therefore,itismostlikelythatthelargepositivechargeofN6contributedits
observedhightoxicity,despiteitstructurallysimilartoN5andN7.
QualityNanoTrainingSchool

64

Conclusions
TheSARanalysisgeneratesthefollowingconclusions:
1. Themostlikelyreason ofhightoxicityforJapanesenanotube(N3)isitsshape(high
aspectratio).
2. Themostlikelycauseofhightoxicityforzincoxide(N14)andnickeloxide(N12)is
their highcontentsofzincandnickel.
3. Although,Aminated beads(N6)hasthehighesttoxicityvaluesinnearlyallassay
results,itwas always groupedcloselywiththeothertwopolystyrenelatexbeads,
unmodifiedbeads(N5)andcarboxylated beads(N7),whichshowednotoxicity.
Therefore,thezetapotentialofthethreepolystrene latexwas measured.The
resultsuggeststhatthelargepositivechargeofN6contributed to itsobservedhigh
toxicity.

TheworkhasshownthatSARbasedonmolecular descriptors isauseful

toolfordescribingfactorsinfluencingthetoxicityofnanoparticles.
Apanelofeighteennanoparticlesisstillconsideredtobetoosmall,
therefore,aQSARmodelthatcanbeappliedtopredictthetoxicityofnew
nanoparticles cannot bedeveloped with this small setofdata.
QualityNanoTrainingSchool

65