Environ Monit Assess (2011) 181:29–42DOI 10.1007/s10661-010-1810-z

Using condition factor and blood variable biomarkersin fish to assess water quality

Helen Sadauskas-Henrique ·Marise M. Sakuragui · Marcelo G. Paulino ·Marisa N. Fernandes

Received: 13 May 2010 / Accepted: 22 November 2010 / Published online: 9 December 2010© Springer Science+Business Media B.V. 2010

Abstract The condition factor and blood vari-ables, including erythrocyte lipid peroxidation(LPO) and the activity of antioxidant enzymessuch as superoxide dismutase (SOD), catalase(CAT), and glutathione peroxidase (GPx), intwo ecologically distinct fish species (Astyanaxfasciatus and Pimelodus maculatus) were eval-uated at five sites in the Furnas HydroelectricPower Station reservoir (Brazil) to assess waterquality. Aldrin/dieldrin, endosulfan, heptachlorepoxide, and metolachlor were detected at differ-ent concentrations in four of the sites. Conditionfactor was not directly affected by such contami-nants. A negative correlation between hematocritand heptachlor was detected in P. maculatus. Posi-tive correlations between red blood cells andheptachlor as well as an interactive effect ofmetolachlor and aldrin/dieldrin were detected inA. fasciatus. The erythrocytes of both species col-lected from the contaminated sites showed highlevels of LPO, an increase in SOD and GPx ac-tivities and a decrease in CAT activity. Although

H. Sadauskas-Henrique · M. M. Sakuragui ·M. G. Paulino · M. N. Fernandes (B)Physiological Sciences Department,Federal University of São Carlos,Rodovia Washington Luiz, km 235,13565-905 São Carlos, SP, Brazile-mail: dmnf@ufscar.br

the leukocyte number and the differential per-centage of leukocytes varied among the sites, thehematological variables, the LPO levels, and theantioxidant enzyme activities could be used toassess water quality, regardless of the differencesin the responses of the fish species.

Keywords Organochlorine · Erythrocytes ·Leukocytes · Thrombocytes ·Antioxidant enzymes · Lipid peroxidation

Introduction

Agriculture is an important activity in tropical andsubtropical regions, but these production systemsrely on the use of pesticides to control insectsor fungal diseases and to maintain productivity.The impacts of pesticides on natural fish popu-lations in aquatic ecosystems have been investi-gated, and fish biomarkers have been recognizedas very useful tools for freshwater biomonitoring(Triebskorn et al. 2007; Weber et al. 2007). Lab-oratory and field studies have revealed numer-ous organic disorders in fish that have beenexposed to xenobiotic molecules, with resultingchanges in blood variables, growth, and reproduc-tion (Cerqueira and Fernandes 2002).

The condition factor, a somatic biomarker, isindicative of health and reflects feeding conditionsas well as energy consumption and metabolism

30 Environ Monit Assess (2011) 181:29–42

(Schulz and Martins-Junior 2001; Alberto et al.2005). Toxic substances in the water may affectthe growth of fish by directly changing metabolismand increasing the energy required to maintainhomeostasis, or they can indirectly impact growthby reducing food availability.

In freshwater fish, most pesticides, metals, andother chemicals are taken up directly from thewater by the gills, due to the large volume of waterthat fish need to ventilate their gills to obtainoxygen for aerobic metabolism. Toxins can also beacquired through the intestines during the transitof contaminated food. These chemicals are thendistributed throughout the body via the blood.Blood cells are some of the first cells to come intocontact with and be affected by xenobiotics. Bloodcells also respond to changes in other tissues thathave suffered some biochemical or physiologicaldisorders due to xenobiotic exposure (Cerqueiraand Fernandes 2002; Mazon et al. 2002; Ruaset al. 2008). Changes in hematology depend onthe actions of xenobiotic molecules in biologicalsystems, the concentrations of the contaminants inthe water, the exposure time, and species sensitiv-ity. Fish erythrocytes have been useful for inves-tigations of biochemical processes at the cellularlevel, including oxidative stress and lipoperoxida-tion (Ruas et al. 2008).

Erythrocytes are essential for the transport ofgases (O2 and CO2) from respiratory surfaces totissues and vice versa, supplying metabolic needsand functioning in hydrogen buffering. Erythro-cytes are the major site for the reactive oxygenspecies (ROS) production because of their role inthe O2 transport system. Leukocytes possess highphagocytic activity, participating actively in anorganism’s defense system and they are essentialfor immunological responses in fish (Serpunin andLikhatchyova 1998). Similarly, thrombocytes areinvolved in blood clotting and organism defense(Passantino et al. 2005). Blood plasma transportsnumerous nutrient molecules, ions, and metabo-lite residues for excretion.

The main goal of this study was to evaluatethe condition factor, the blood hematological vari-ables, and some oxidative stress biomarkers inthe erythrocytes of two ecologically distinct fish

species from the reservoir of the Furnas Hydro-electric Power Station (Furnas HPS), in Brazil,to assess water quality and to establish if thesevariables can be associated with different typesof anthropogenic contamination. The oxidativestress biomarkers selected were lipid peroxida-tion (LPO) and the activity of specific antioxidantenzymes: superoxide dismutase (SOD), catalase(CAT), and glutathione peroxidase (GPx). Thephysical and chemical water characteristics werealso determined. The Furnas HPS is the largestpower station in southeastern Brazil. This site wasselected because it is not only located in a pre-dominantly agricultural region but also receivesuntreated urban water discharges from numer-ous small and medium-size cities. The lambari,Astyanax fasciatus (benthic-pelagic species), andthe mandi, Pimelodus maculatus (benthic species),are resident fish species that do not exhibit migra-tion behavior and are widely distributed through-out the reservoir, allowing for comparisonsbetween sites.

Materials and methods

Study area

Furnas HPS is located in Minas Gerais, Brazil(Fig. 1). The reservoir is the result of the dammingof the Rio Grande (250 km long) and the RioSapucaí (170 km long), and it has a 1,440 km2

overflow area with 21 million cubic meters anda perimeter of 3,500 km. The reservoir is bor-dered by 34 small- to medium-size cities, mostof which engage in intense agricultural and cat-tle farming activities. Water and fish specimenswere collected from five sites: site 1 (the referencesite) at the confluence of the Grande and SapucaíRivers [Turvo (FU10) S20!40"835"" W 46!13"232""];site 2, Guapé—FU20 (S20!44"331"" W 45!55"800"");site 3, Porto Fernandes—FU50 (S20!48"826""

W 45!40"567""), both in the Rio Grande axis;site 4, Barranco Alto—FU30 (S21!10"510"" W45!57"061""), and site 5, Fama—FU40 (S21!24"074""

W 45!49"621"") in the Rio Sapucaí axis (Fig. 1).

Environ Monit Assess (2011) 181:29–42 31

Fig. 1 Map of the Furnas Hydroelectric Power Station reservoir, Minas Gerais, Brazil, showing the sites of water and fishcollection (f illed circle): FU10 (Turvo), FU20 (Guapé), FU30 (Barranco Alto), FU40 (Fama), and FU50 (Porto Fernandes)

Fish and water collection

Lambari (A. fasciatus, n = 20/site, Wt = 37.8 ±2.6 g, Lt = 14.3 ± 0.3 cm) and mandi (P. maculatus,n = 15/site, Wt = 182.3 ± 32.9 g, Lt = 25.1 ±1.4 cm) specimens were collected along with watersamples in June (winter season) and December(summer season) of 2006. Water samples werecollected (three stations per site located 100 mapart from one another) for chemical analysisperformed according to standard methods for ex-amination of wastewater.

Water analyses

Dissolved oxygen (DO), conductivity, tempera-ture, and pH were measured in the field using a

multi-parameter water analyzer (YSI, 600XL).Alkalinity and total phosphorus were determinedas described by Golterman et al. (1978). Totalhardness and chloride concentration were deter-mined following the APHA (1992) method-ologies; ammoniacal nitrogen, nitrite, andnitrate were determined using the colorimetricmethod (Mackereth et al. 1978). Metalconcentrations were determined following stan-dards SW84603050/3051 (USEPA 1986). Thealuminum concentration was determined usingeriochrome cyanine R. Cadmium, copper, iron,and zinc concentrations were determined byatomic absorption spectrometry. The concentra-tions of pesticides in the water were determinedfollowing USEPA protocols: 2,4 dichlorophe-noxyacetic acid (USEPA 8321); alachlor, atrazine,

32 Environ Monit Assess (2011) 181:29–42

glyphosate, hexachlorobenzene, lindane (gamma-BHC), metolachlor, methoxyclor, molinate,pendimethalin, permethrin, propanil, simazine,and trifluralin (USEPA 8270); aldrin, dieldrin,chlordane, endosulfan, endrin, heptachlor, andheptachlor epoxide (USEPA 8081); bentazonand pentachlorophenol (USEPA 8151); anddichlorodiphenyltrichloroethane (DDT isomers;USEPA, 8260) using a gas chromatograph HP5980 and a mass spectrometer HP 5970 MSD.

Fish analyses

Fish were weighed and measured, and blood wastaken through the caudal vein. The relationshipbetween wet body mass (MB) and total length(Lt) was calculated as MB = aLb

t , and the con-dition factor (K) was calculated according to theequation K = MB/Lb

t , where MB is the wet bodymass (g), Lt is the total length (cm), and b is theisometry coefficient.

Blood sub-samples were used for hematologicalanalyses immediately after sampling. The hemat-ocrit (Hct, %) was determined using heparinizedcapillary tubes in a microhematocrit centrifuge.The hemoglobin concentration (Hb, g dL#1)

was determined using the cyanomethemoglobinmethod, and the red blood cell count (RBC,µL) was estimated using a modified Neubauerchamber. Mean cell volume (MCV, fL), meancell hemoglobin (MCH, pg cell#1), and meancell hemoglobin concentration (MCHC, g dL#1)

were calculated using Hct, Hb, and RBC mea-surements. Blood smears were stained using poly-chromatic differential staining (Fast Panotic LB,Laborclin). The leukocyte and thrombocyte num-bers were counted and indirectly calculated ac-cording to McKnight (1966).

The remaining blood was centrifuged, andthe plasma was removed. The erythrocytes werehemolized and centrifuged at 4!C. The super-natant aliquots were stored at #70!C, and theenzyme activities and lipid peroxidation were mea-sured spectrophotometrically (Biochrom LibraS32) at 25!C.

Lipid peroxidation was assessed by Fe2+ ox-idation in the presence of xylenol orange (fer-rous oxidation–xylenol orange) (Jiang et al. 1992).Catalase activity was determined by the de-

crease of the hydrogen peroxide (H2O2) con-centration monitored at 240 nm (Aebi 1974).Superoxide dismutase activity was measured afterhemoglobin precipitation and extraction in chlo-roform/ethanol by an indirect inhibition assay ofthe reduction of nitro blue tetrazolium (Crouchet al. 1981). Gluthatione peroxidase activity wasmeasured using DTNB reagent (Hafeman et al.1974).

Statistical analysis

Data are presented as the mean ± standard error(SEM). Principal component analysis (PCA) andhierarchical cluster analysis (HCA) were appliedto the aquatic variables to delineate geochemi-cal groups. The correlation was applied to iden-tify the interactions between the water variablesand the biological variables. Data normality wastested using the Kolmogorov–Smirnov test. Para-metric Tukey–Kramer (white and red series) andnon-parametric Kruskal–Wallis (condition factor)tests were used to compare the sites. The Mann–Whitney U test was applied to compare data fromJune and December. The accepted significance ofthe data was 5% (P < 0.05), and analyses wereconducted with XlStat 7.5 (PCA and correlations)and BioEstat v. 3.0 (ANOVA).

Results

Water

The physical and chemical variables of the waterat all sites are shown in Table 1. Temperature,dissolved oxygen, pH, and conductivity did notdiffer between sites. The N-ammonia, N-nitrite,and N-nitrate values in the water were higherat FU30 in June; chloride was higher at FU20and FU50 in June, and the iron concentra-tion was higher at FU20, FU40, and FU50 inDecember (Table 1). All of these variables, ex-cept for the iron concentration, were lower thanthe upper limits recommended by the BrazilianEnvironment National Council in water for biotapreservation (CONAMA 357/2005). Most of thepesticides that were tested were not presentat levels above the detection limits of the ana-

Environ Monit Assess (2011) 181:29–42 33

Tab

le1

Wat

erva

riab

les

from

alls

ites

ofth

ere

serv

oir

ofFu

rnas

HPS

,MG

,Bra

zil,

inJu

nean

dD

ecem

ber

2006

Var

iabl

esSi

teFU

10FU

20FU

30FU

40FU

50T

urvo

Gua

péB

arra

nco

Alto

Fam

aPo

rto

Fern

ande

sJu

neD

ecem

ber

June

Dec

embe

rJu

neD

ecem

ber

June

Dec

embe

rJu

neD

ecem

ber

Dis

solv

edox

ygen

(mg/

L)

7.14

7.63

7.20

7.00

8.16

7.76

7.03

7.75

7.43

7.40

Tem

pera

ture

(!C

)22

.624

.521

.825

.020

.825

.420

.924

.821

.624

.6pH

7.22

7.30

7.65

7.20

7.45

7.31

7.56

6.90

7.43

7.30

Con

duct

ivity

(µS/

cm)

33.0

36.5

30.0

36.5

32.0

43.0

33.0

38.6

30.0

35.4

Alk

alin

ity(m

g/L

asC

aCO

3 )29

.324

.113

.822

.515

.418

.012

.720

.013

.015

.0H

ardn

ess

(mg/

Las

CaC

O3)

24.6

2011

.721

.013

.119

.09.

023

.011

.318

.0N

-am

mon

ia(m

g/L

N)

0.2

0.10

0.10

0.2

1.0

0.10

0.49

0.05

0.10

0.02

N-n

itrite

(mg/

LN

)0.

040.

01nd

0.01

0.10

0.01

0.1

0.03

0.1

0.01

N-n

itrat

e(m

g/L

N)

0.25

nd0,

16nd

0.39

nd0.

050.

100.

29nd

Chl

orid

e(m

g/L

)0.

230.

211.

792.

10.

020.

020.

040.

031.

760.

85A

lum

inum

(mg/

L)

0.00

20.

005

0.00

60.

005

0.00

50.

002

0.00

30.

003

0.00

20.

001

Cad

miu

m(m

g/L

)nd

ndnd

ndnd

ndnd

ndnd

ndC

hrom

ium

(mg/

L)

ndnd

ndnd

ndnd

ndnd

ndnd

Cop

per

(mg/

L)

ndnd

ndnd

ndnd

ndnd

ndnd

Iron

(mg/

L)

0.03

0.03

0.28

0.30

ndnd

0.64

0.71

0.25

0.20

Zin

c(m

g/L

)nd

ndnd

nd0.

010.

010.

020.

02nd

ndA

ldri

nan

dD

ield

rin

(µg/

L)

ndnd

1.1

0.86

0.5

nd0.

003

nd1.

031.

0E

ndos

ulfa

n(µ

g/L

)nd

nd0.

80.

741.

0nd

1.0

nd0.

280.

1H

epta

chlo

rep

oxid

e(µ

g/L

)nd

nd0.

450.

40.

80.

31nd

nd0.

320.

30M

etol

achl

or(µ

g/L

)nd

nd36

1810

8.1

ndnd

2914

The

stan

dard

devi

atio

nw

asal

way

sle

ssth

an5%

ofth

em

ean

valu

ean

dth

enth

eyw

ere

omitt

ed

34 Environ Monit Assess (2011) 181:29–42

lytical methods, but different concentrations ofaldrin/dieldrin, endosulfan, heptachlor epoxide,and metolachlor were detected in FU20, FU30,FU40, and FU50 (Table 1). The concentrationwas higher than the dissolved upper limits rec-ommended by CONAMA 357/2005, which are0.005, 0.056, 0.01, and 10 µg/L, respectively, foraldrin/dieldrin, endosulfan, heptachlor epoxide,and metolachlor. PCA analyses revealed an asso-ciation (59%) between endosulfan and heptachlorin June at FU30 and aldrin/dieldrin, heptachlor,and metolachlor at FU20 and FU50. Strong asso-ciations (89%) between aldrin/dieldrin, endosul-fan, heptachlor, and metolachlor were observed inDecember at FU20 and between aldrin/dieldrin,heptachlor, and metolachlor at FU50.

The HCA analysis revealed high similarity be-tween FU20 and FU50 and between FU30 and

FU40. FU10 was not similar to any of the otherfour sites.

Condition factor

The relationship between the wet body mass (MB)

and total length (Lt) was described by the vari-ables MB = 0.001 L3.82

t , r2 = 0.87 for A. fasciatusand MB = 0.002 L3.53

t , r2 = 0.92 for P. maculatus(Fig. 2a, b), indicating positive allometric growth.The condition factor (K) was significantly lower infish collected from FU30 in June (Fig. 2c, d).

Hematological variables

Figure 3 shows the values of Hct, RBC, Hb,and the hematimetric index (MVC, MHC, andMCHC) of A. fasciatus and P. maculatus collected

Fig. 2 Relationship between total length (Lt) and bodymass (MB) (a, b); condition factor (K) (c, d) of A. fasciatusand P. maculatus from the Furnas HPS, MG, Brazil. Labels

a, b , c, d, and e indicate differences in relation to sitesFU10, FU20, FU30, FU40, and FU50, respectively. Asteriskindicates a difference between June and December

Environ Monit Assess (2011) 181:29–42 35

Fig. 3 Hematologicalvariables: hematocrit(Hct), red blood cells(RBC), hemoglobinconcentration ([Hb]),mean corpuscular volume(MCV), meanhemoglobin corpuscular(MCH), and meancorpuscular hemoglobinconcentration (MCHC)of A. fasciatus and P.maculates from theFurnas HPS reservoir,MG, Brazil, in June(unf illed bars) andDecember (f illed bars).Labels a, b , c, d, and eindicate differences fromFU10, FU20, FU30,FU40, and FU50,respectively. Asteriskindicates a differencebetween June andDecember

in June (winter) and December (summer). Nosignificant differences in the hematological valueswere found among A. fasciatus, with the excep-tion of the Hb concentration and the MCHC infish from FU40 which were lower than those offish from FU10 (in June). A positive correlationwas found between RBC and heptachlor (RBC =

199.80 + 49.1$heptachlor; r2 = 0.75P < 0.05) infish collected in June. Hematological variables inP. maculatus did not differ between the sites ineach period, but some differences were found be-tween June and December: the Hb concentrationin FU40; MCV in FU30; MCH in FU40 and FU50;and MCHC in FU30, FU40, and FU30 were all

36 Environ Monit Assess (2011) 181:29–42

different from the values obtained at other sites(P < 0.05). A negative correlation between Hctand heptachlor (Hct = 34.73 # 16.35$heptachlor;r2 = 0.81, P < 0.03) was found in June.

In A. fasciatus, the total number of leukocyteswas higher in June (P < 0.05; Table 2). The high-est and the lowest leukocyte numbers were de-tected in fish from FU30 and FU40, respectively.In December, the lowest value was observed inFU50. P. maculatus individuals from FU20, FU30,and FU50 exhibited fewer leukocytes in June(Table 2). In A. fasciatus, the lymphocyte percent-age was higher and the percentages of monocytesand neutrophils were similar in fish from FU10in both June and December. The monocyte fre-quency was lower in FU20 and FU30 (P < 0.05) inDecember. In P. maculatus, the lymphocytes werealso the most frequent white blood cell detectedin fish from FU10 in June and from FU10, FU20,and FU30 in December. The neutrophil percent-age was elevated in fish from FU40 in Decem-ber (Table 2). Special granulocytic cells (SGCs)or PAS-positive granular leukocytes were foundin all P. maculatus specimens and were elevatedin fish collected from FU30 in December. Thesecells were absent in all A. fasciatus collected inDecember.

The number of thrombocytes was lower in A.fasciatus than in P. maculatus (Table 2). Therewere no significant differences in the numberof thrombocytes in A. fasciatus from all sites;however, in P. maculatus, the thrombocyte num-ber was higher in fish from FU20 collected inJune. Positive correlations between thrombocytesand the metolachlor concentration (T = 4.96 +0.19$metolachlor, r2 = 0.96, P < 0.003) were de-tected in A. fasciatus and P. maculatus collected inJune.

Erythrocyte LPO and antioxidant enzymeactivities

The LPO concentration in the erythrocytes washigher in the contaminated sites (P < 0.05) andsignificantly higher in June (P < 0.05), exceptin FU40 (Fig. 4a, b). In A. fasciatus, the ac-tivities of SOD and GPx were higher in FU20and FU30 (June and December) and in FU50(June); CAT activity in the contaminated sites

Fig. 4 Lipoperoxidation (LPO) in the erythrocytes ofA. fasciatus (a) and P. maculatus (b) from Furnas HPSreservoir, MG, Brazil, in June (unf illed bars) and De-cember (f illed bars). Labels a, b , c, d, and e indicatedifferences from FU10, FU20, FU30, FU40, and FU50,respectively. Asterisk indicates a difference between Juneand December

was lower than that at the reference site (Fig. 5a,c, and e). A negative correlation was found be-tween aldrin/dieldrin and SOD activity (SOD =258.26 # 64.41$aldrin/dieldrin, r2 = 0.98, P <

0.05) as well as GPx activity (GPx = 1.07 #0.23$aldrin/dieldrin, r2 = 0.86, P < 0.05). Endo-sulfan showed a negative correlation with CATactivity (CAT = 18.19 # 7.21$endosulfan, r2 =0.74, P < 0.05).

In P. maculatus, the SOD and GPx activitieswere higher in all of the contaminated sites, andsignificant differences were found between Juneand December (Fig. 5b, f). CAT activity waslower in FU20, FU30, and FU50 and significantlyhigher in December (Fig. 5d). Temperature andaldrin/dieldrin exhibited a positive correlationwith GPx activity (GPx = 9.20 + 5.26$aldrin/

dieldrin + 0.40temperature, r2 = 0.98, P < 0.05)and an interactive effect was verified betweenheptachlor epoxide, alkalinity, and CAT ac-tivity (CAT = 25.63 # 11.88$heptachlor + 0.29$

alkalinity, r2 = 0.97, P < 0.05).

Environ Monit Assess (2011) 181:29–42 37

Tab

le2

Tot

alle

ukoc

ytes

,diff

eren

tiall

euko

cyte

perc

enta

ge,a

ndth

rom

bocy

tenu

mbe

r(m

ean

±SE

M)

ofA

stya

nax

fasc

iatu

san

dP

.mac

ulat

usfr

omth

ere

serv

oir

ofFu

rnas

HPS

,MG

,Bra

zil,

inJu

nean

dD

ecem

ber

2006

Var

iabl

esSi

teFU

10FU

20FU

30FU

40FU

50T

urvo

Gua

péB

arra

nco

Alto

Fam

aPo

rto

Fern

ande

sJu

neD

ecem

ber

June

Dec

embe

rJu

neD

ecem

ber

June

Dec

embe

rJu

neD

ecem

ber

Ast

yana

xfa

scia

tus

Leu

kocy

tenu

mbe

r19

.7±

2.6d

19.3

±1.

5e16

.2±

1.5c

19.0

±2.

7e27

.2±

3.3b

de14

.4±

2.0$

7.8

±0.

9ac

16.0

±2.

110

.5±

1.2c

8.9

±1.

7ab

103

µL

#1

Lym

phoc

ytes

(%)

45.2

±7.

2e32

.6±

2.9

41.7

±6.

3e47

.9±

6.6

40.7

±4,

0e26

.5±

4.7$

42.4

±6.

6e31

.4±

3.2

25.7

±4.

2abc

d31

.1±

4.1

Mon

ocyt

es(%

)27

.3±

7.0

43.7

±3.

0bcd

e$34

.5±

5.3

9.7

±08

a$25

.8±

4.5

29.3

±4.

7a34

.0±

9.5

9.7

±1.

7a$

44.1

±8.

0a25

.3±

6.2a

Neu

trop

hils

(%)

27.4

±4.

718

.0±

3.8b

cde$

23.7

±8.

942

.4±

6.9a

$33

.3±

7.6

44.2

±5.

6a22

.2±

7.9

58.2

±4.

6a$

26.4

±6.

341

.8±

5.5a

Spec

ialg

ranu

locy

tic1.

4±

0.9b

00

00.

2±

0.2

01.

4±

1.3

01.

7±

0.8

0ce

lls(%

)E

osin

ophi

ls(%

)0

5.7

±3.

0bc$

00

00

00.

8±

0.1

01.

8±

1.2

Thr

ombo

cyte

num

ber

4.2

±0.

6bc

4.8

±0.

99.

0±

1.1a

5.8

±1.

77.

5±

2.0a

d2.

9±

1.1

5.8

±1.

44.

8±

1.1

6.8

±1.

94.

4±

0.9

103

µL

#1

Pim

elod

usm

acul

atus

Leu

kocy

tenu

mbe

r39

.1±

5.3e

23.2

±3.

6c14

.5±

3.4

25.2

±3.

5c17

.6±

4.0

41.3

±4.

6abe

$28

.5±

7.3

37.4

±4.

013

.6±

1.8a

14.2

±2.

1c10

3µ

L#1

Lym

phoc

ytes

(%)

65.1

±13

.260

.2±

8.5

35.8

±3.

356

.0±

7.6

40.0

±7.

150

.6±

2.8

34.2

±9.

139

.4±

6.1

31.8

±5.

440

.5±

7.5

Mon

ocyt

es(%

)22

.0±

14.8

12.9

±1.

928

.4±

5.3

9.9

±1.

429

.0±

4.0

12.4

±0.

620

.9±

3.6

6.4

±1.

1$37

.8±

7.5

18.2

±4.

4N

eutr

ophi

ls(%

)0

00.

3±

0.1

0.5

±0.

30

00

00

0.7

±0.

6Sp

ecia

lgra

nulo

cytic

9.5

±4.

8d13

.7±

3.9d

30.4

±2.

025

.6±

7.3d

14.7

±3.

332

.0±

5.7

38.3

±7.

7a50

.0±

6.2a

b21

.7±

9.2

33.7

±10

.0ce

lls(%

)E

osin

ophi

ls(%

)3.

2±

0.5

13.2

±1.

45.

1±

0.8

8.0

±2.

916

.3±

4.6

5.0

±2.

16.

6±

3.2

4.2

±1.

08.

4±

2.3

6.9

±2.

9T

hrom

bocy

tenu

mbe

r34

.5±

12.7

14.0

±12

.0$

45.4

±17

.0d

8.7

±4.

4c$

10.4

±2.

623

.5±

20.3

bd$

7.4

±2.

3b7.

8±

4.8c

53±

1.2

14.7

±5.

510

3µ

L#1

Let

ters

a,b,

c,d,

and

ein

dica

tedi

ffer

ence

sin

rela

tion

toFU

10,F

U20

,FU

30,F

U40

,and

FU50

,res

pect

ivel

y$ I

ndic

ates

diff

eren

ces

betw

een

June

and

Dec

embe

r

38 Environ Monit Assess (2011) 181:29–42

Fig. 5 Superoxide dismutase (SOD), catalase (CAT), andglutathione peroxidase (GPx) activity in the erythrocytesof A. fasciatus (a, c, and e) and P. maculatus (b, d, andf) from the Furnas HPS reservoir, MG, Brazil, in June

(unf illed bars) and December (f illed bars). Labels a, b , c, d,and e indicate differences from FU10, FU20, FU30, FU40,and FU50, respectively. Asterisk indicates a difference be-tween June and December

Discussion

The presence of the organochlorines aldrin/dieldrin, endosulfan, heptachlor, epoxide, meto-lachlor (FU20, FU30, FU40, and FU50), and iron(FU40) dissolved in water characterize the Fur-nas HPS reservoir as contaminated. The simi-larity between sites located in the same river isdue to the predominance of certain cultures ateach riverside. The dendritic shape of the Fur-nas reservoir favor punctuates contamination pat-

terns. Fish inhabiting these areas are exposed toa mixture of contaminants that differ in concen-tration according to the time of year, a patternthat is likely due to the periodicity of pesticideapplication to the agricultural landscapes, or tothe dynamics of these contaminants in the reser-voir (water–sediment). Organochlorines have lowsolubility in water and high adsorption in the or-ganic matter, making them persistent in the en-vironment. These compounds are retained in thesediment particulate phase and are released into

Environ Monit Assess (2011) 181:29–42 39

the water depending on the characteristics of thewater.

The condition factor (K) and blood variablesof both species demonstrated the complexity ofthe fishes’ integrated responses to the physicaland chemical characteristics of their environment,which depend on the chemical concentration,mode of action, time of exposure, and season. Ingeneral, pollutants have a negative effect on Kby reducing food availability and/or by increas-ing the energy required to maintain homeostasis(Dethloff et al. 2001; Bervoets and Blust 2003);exceptions may be found in untreated domesticsewage pollution, which is usually related to highproductivity (Schulz and Martins-Junior 2001;Alberto et al. 2005). The absence of a correlationbetween the contaminants and K may suggest thatthese two parameters are not directly affected bysuch contaminants, at least in these fish species.Consistent with this hypothesis, high organochlo-rine concentrations were also found at the FU20and FU50 sites of Furnas HPS, and the K valueswere not as low as in fish from site FU30. Fur-thermore, organochlorine residues were detectedin the gills and liver of both A. fasciatus and P.maculatus from all of the contaminated sites (un-published data). The presence of unknown vari-able may explain the low K values in FU30 thatwere detected in the June sample.

Multiple contaminants in the water may resultin an interactive effect on the blood variables andother tissues (Adhikari et al. 2004; Koprucu et al.2006). The Hct, the RBc, the Hb concentration,and the hematimetric indexes of both studiedspecies were similar to those previously deter-mined for A. fasciatus (Alberto et al. 2005) and P.maculatus (Ranzani-Paiva et al. 2000). However,the variability in these measurements between thesites demonstrates the influence of environmentalcharacteristics and differences in the physiologi-cal adjustments between these two species. Thepositive correlation between RBC and heptachlorin A. fasciatus suggests an adjustment to maintainthe O2 cascade from the gills to the tissue, whilein P. maculatus, the negative correlation betweenHct and heptachlor implies a reduction of theefficiency of O2transport and the absence of anyhematological compensatory adjustment to over-come this effect. Changes in RBC, Hb concentra-

tion, and MCV are directly related to the need forO2 in oxidative metabolism and the efficiency ofO2 uptake from the environment and its transportto tissues.

The higher LPO concentration in the erythro-cytes of A. fasciatus and P. maculatus are evi-dence of oxidative stress in fish collected at theFU20, FU30, and FU50 sites in June, as these siteshave a higher concentration of organochlorine.Oxidative stress occurs when the balance betweenoxidants and antioxidants is disrupted and theexcessive generation of reactive oxygen speciesproduces LPO (Scandalios 2005).

Changes in the activity of the antioxidant en-zymes SOD, CAT, and GPx to maintain the func-tion and integrity of cells in most animals, aswell as their activation in response to exposureto pollutants, have been reported in erythrocytesand various tissues (Bainy et al. 1996; Chebabet al. 2009; Kaminski et al. 2009). The higheractivity of the antioxidant enzymes SOD, CAT,and GPx in the erythrocytes of A. fasciatus andP. maculatus collected in the contaminated siteswere not sufficient to remove ROS and neutral-ize their effects, resulting in an increase of LPO.The inhibition of CAT activity in the erythro-cytes of both species collected in the contami-nated sites contributed to higher levels of LPO.The differences of SOD, CAT, and GPx activitiesbetween the A. fasciatus and P. maculatus clearlyshowed the different susceptibilities of these en-zymes to a mixture of toxic compounds. Thisphenomenon has been previously emphasized byRuas et al. (2008), who measured the activities ofSOD, CAT, and GPx in the erythrocytes of threecichlid species living in the same contaminatedenvironment.

In vitro and in vivo studies have shown thatindividual pesticides/herbicides increase the pro-duction of ROS and enhance SOD, CAT, andGPx activities in a time- and dose-dependent man-ner, while a mixture of these toxicants may havedifferent effects. CAT is inhibited by endosulfanin chicken erythrocytes (Aggarwal et al. 2009).SOD and CAT activity increased in the liver ofrats treated with endosulfan or chlorpyriphos anddecreased in those treated with both chemicalsin combination (Chebab et al. 2009). CAT ac-tivity increased in the muscles of fish living in a

40 Environ Monit Assess (2011) 181:29–42

river that was contaminated heavily with pesti-cides/herbicides (Michelo et al. 2006).

Although specific immunity in fish is less de-veloped than that in birds and mammals, fishhave a non-specific resistance system that pro-tects them against pathogenic and environmen-tal factors (Passantino et al. 2005). Leukocytesare involved in specific and non-specific defensemechanisms, but most pollutants affect the de-fense system in fish. In general, lymphocytes arethe most numerous white blood cells (70–80%),as they are involved in antibody production andinflammatory process. These cells were relativelyscarce in both of the studied species (20–48%in A. fasciatus and 21–66% in P. maculatus),while previous studies found 70% lymphocytes inthe same species (Silva-Souza et al. 2000). Thesefindings suggest a possible immunosuppressionin fish from the Furnas HPS reservoir. Further-more, the number of monocytes and neutrophilsobserved indicates the mobilization of these cellsfrom lymphoid tissues in response to bacterialinfection. Monocytes are the precursors of tis-sue macrophages, which ingest foreign materialsand are also involved in the immune responseas antigen-presenting cells that transmit informa-tion about the ingested material to lymphocytes.Neutrophils are involved in non-specific immu-nity, migrating to the infection site and recogniz-ing, ingesting, and destroying bacteria and otherpathogens (Tavares-Dias et al. 2007). The roleof SGCs is not well established, but as they areconsidered as a type of neutrophil, they may be in-volved in inflammatory and phagocytic processes.Our results support the hypothesis that the per-centage of SGCs increases in fish stressed by lowwater quality, such as during chronic exposureto contaminants (Martins et al. 2002), as thesecells were always present in fish collected in theFurnas HPS reservoir (except in A. fasciatus inDecember).

Thrombocyte function is related to blood coag-ulation and phagocytosis as a link between innateand adaptive immunity (Passantino et al. 2005).Increases in the number of thrombocytes havebeen reported in Cyprinus carpiovar exposed tosublethal concentrations of endosulfan (1 µgL#1)

under laboratory conditions (Shafiq-ur-Rehman2006). In the present study, there was no correla-

tion of thrombocytes with the endosulfan presentin water. The low concentrations in most sites, ex-cluding FU30 and FU40 in June, or the low sensi-tivity of A. fasciatus and P. maculatus species mayhave influenced the responses of these fish. How-ever, the effect of metolachlor on the thrombocytenumber was evident in the positive correlation be-tween these variables. This response may be dueto a direct effect on thrombocyte production or toan indirect effect due to the increased gill tissuedamage (Paulino, personal communication).

Conclusion

Fish exposed to multiple contaminants in theirnatural environments present substantial variabil-ity in most physiological and biochemical vari-ables. Such responses are expected, as numerousfactors influence organic responses. Consideringthe complexity of the aquatic environment andthe coexistence of inducing and inhibiting chem-icals, this study provides evidence for the impor-tance of using several biomarkers to estimate risksin complex situations. The use of a set of bio-markers can enhance the likelihood of identifyingareas/species that are threatened by chemicals,especially in cases where differences in sensitivitymay be found among the species. Among thebiomarkers used in this study, the hematologicalvariables, the erythrocyte LPO levels, and theactivity of the antioxidant enzymes can be used toassess water quality, regardless of the differencein responses between fish species. Furthermore,the higher energy demand of fish needed to repairthe damage caused by the environment may affectreproductive activity, thereby reducing these spec-imens in the reservoir and fish diversity over anextended period.

Acknowledgements This study was part of the P&DANEEL PROGRAM of Furnas Centrais Elétricas S.A.(Proc. 0394-097-2003) and was developed by the FederalUniversity of São Carlos in conjunction with the Hydro-biology and Aquaculture Station of the FURNAS HPS.The authors participate in the INCT in Aquatic Toxicol-ogy (Proc. 573949/2008-5) from the Brazilian Council forScientific and Technological Development (CNPq). Weare thankful to Dr. A.P. Carvalho for statistical help. H.S.Henrique and M.G. Paulino acknowledge fellowships fromFURNAS and CAPES.

Environ Monit Assess (2011) 181:29–42 41

References

Adhikari, S., Sarkar, B., Chatterjee, A., Mahapatra, C. T.,& Ayyappan, S. (2004). Effects of cypermethrin andcarbofuran on certain hematological parameters andpredictions of their recovery in a freshwater teleost,Labeo rohita (Hamilton). Ecotoxicology and Environ-mental Safety, 58, 220–226.

Aebi, H. (1974). Catalase. In: H. U. Bergmayer (Ed.),Methods of enzymatic analysis (pp. 671–684). London:Academic Press.

Aggarwal, M., Naraharisetti, S. B., Sarkar, S. N., Rao, G. S.,Degen, G. H., & Malik, J. K. (2009). Effects of sub-chronic coexposure to arsenic and endosulfan on theerythrocytes of broiler chickens: A biochemical study.Archives of Environmental Contamination and Toxi-cology, 56, 139–148.

Alberto, A., Camargo, F. M., Verani, J. R., & Costa,O. F. T. (2005). Health variables and gill morphologyin the tropical fish Astyanax fasciatus from a sewage-contaminated river. Ecotoxicology and EnvironmentalSafety, 61, 247–255.

APHA (1992). Standard Methods for the Examination ofWater and Wastewater 18th (ed.). Washington, DC:American Public Health Association.

Bainy, A. C. D., Saito, E., Carvalho, P. S. M., & Junqueira,V. B. C. (1996). Oxidative stress in gill, erythrocytes,liver and kidney of Nile tilapia (Oreochromis niloti-cus) from a polluted site. Aquatic Toxicology, 34,151–162.

Bervoets, L., & Blust, R. (2003). Metal concentrations inwater, sediment and gudgeon (Gobio gobio) from apollution gradient: Relationship with fish conditionfactor. Environmental Pollution, 126, 9–19.

Cerqueira, C. C. C., & Fernandes, M. N. (2002). Gill tissuerecovery after copper exposure and blood parameterresponses in the tropical fish, Prochilodus scrofa. Eco-toxicology and Environmental Safety, 52, 83–91.

Chebab, S., Belli, N., Leghouchi, E., & Lahouel, M. (2009).Oxidative stress induced by two pesticides: Endo-sulfan and chlorpyriphos. Environnement, Risques &Santé, 8(5), 425–432.

CONAMA - Conselho Nacional do Meio ambiente (2005).Res CONAMA 357/2005. Classificação dos corposde água, diretrizes ambientais para enquadramento,condições e padrões de lançamento de efluentes.Brasília.

Crouch, R. K., Gandy, S. C., & Kimsey, G. (1981). The in-hibition of islet superoxide dismutase by diabetogenicdrugs. Diabetes, 30, 235–241.

Dethloff, G. M., Bailey, H. C., & Maier, K. J. (2001).Effects dissolved cooper on select hematologicalbiochemical and immunological parameters of wildrainbow trout (Oncorhynchus mukiss). Archives ofEnvironmental Contamination and Toxicology, 40,371–380.

Golterman, H. L., Clymo, R. S., & Ohnstad, M. A. M.(1978). Methods for Physical and Chemical Analysisof Freshwaters. IBP Handbook. 8. Oxford: BlackwellScientific Publications.

Hafeman, D. G., Sunde, R. A., & Hoekstra, W. G. (1974).Dietary selenium on erythrocyte and glutathione per-oxidase in the rat. Journal of Nutrition, 104, 580–587.

Jiang, Z. Y., Hunt, J. V., & Wolff, S. P. (1992). Ferrous ionoxidation in the presence of xylenol orange for detec-tion of lipid hydroperoxide in low density lipoprotein.Analytical Biochemistry, 202, 384–389.

Kaminski, P., Kurhalyuk, N., Jerzak, L., Kasprzak, M.,Tkachenko, H., Klawe, J. J., et al. (2009). Eco-physiological determinations of antioxidant enzymesand lipoperoxidation in the blood of White StorkCiconia ciconia from Poland. Environmental Research,109, 29–39.

Koprucu, S. S., Koprucu, K., Ural, M. S., Ispir, U., & Pala,M. (2006). Acute toxicity of organophosphorous pes-ticide diazinon and its effects on behavior and somehematological parameters of fingerling Europeancatfish (Silurus glanis L.). Pesticide Biochemistry andPhysiology, 86, 99–105.

Mackereth, F. J. H., Heron, J., & Talling, J. F. (1978).Water analysis: Some revised methods for limnologists.Cumbria: Scientific Publication.

Martins, M. L., Moraes, F. R., Fujimoto, R. Y., Nomura,D. T., & Fenerick, J. (2002). Respostas do híbrido tam-bacu (Piaractus mesopotamicus Holmberg, 1887 ma-cho x Colossoma macropomum Cuvier, 1818 fêmea) aestímulos simples ou consecutivos de captura. Boletindo Instituto de Pesca, 28, 195–204.

Michelo, S., Kennedy, C., Roy, M., Victor, W., Masahiro,Y., & Yoshimitsu, M. (2006). Pesticide/herbicide pol-lutants in the Kafue river and a preliminary investiga-tion into their biological effect through catalase levelsin fish. Japanese Journal of Veterinary Research, 54,119–128.

Mazon, A. F., Monteiro, E. A. S., Pinheiro, G. H. D., &Fernandes, M. N. (2002). Hematological and phys-iological changes induced by short-term exposureto copper in freshwater fish, Prochilodus scrofa.Brazilian Journal of Biology, 62, 621–631.

McKnight, I. M. (1966). A hematological study on themountain white. Prosopium williasonit. Journal of theFisheries Research Board of Canada, 23, 45–64.

Passantino, L., Cianciotta, A., Patruno, R., Ribaud, M. R.,Jirillo, E., & Passantino, G. F. (2005). Do fish thombo-cytes play an immunological role? Their cytoenzimaticprofiles and function during an accidental piscine can-didiasis in aquarium. Immunopharmacology and Im-munotoxicology, 27(2), 345–356.

Ranzani-Paiva, M. J. T., Silva-Souza, A. T., Pavanelli,G. C., Takemoto, R. M., & Eiras, A. C. (2000). Hema-tological evaluation in commercial fish species fromthe floodplain of the upper Paraná River, Brazil. ActaScientarium, 22, 507–513.

Ruas, C. B. G., Carvalho, C. S., Araújo, H. S. S., Espíndola,E. L. G., & Fernandes, M. N. (2008). Oxidative stressbiomarkers of exposure in the blood of cichlid speciesfrom a polluted river. Ecotoxicology and Environmen-tal Safety, 71, 86–93.

Scandalios, J. G. (2005). Oxidative stress: Molecularperception and transduction of signals triggering

42 Environ Monit Assess (2011) 181:29–42

antioxidant gene defenses. Brazilian Journal of Med-ical and Biological Research, 38(7), 995–1014.

Schulz, U. H., & Martins-Junior, H. (2001). Astyanax fas-ciatus as bioindicator of water pollution of Rio Sinos,RS, Brazil. Brazilian Journal of Biology, 61(4), 615–622.

Serpunin, G. G., & Likhatchyova, O. A. (1998). Use of theichthyohaematological studies in ecological monitor-ing of the reservoirs. Acta Veterianaria Brunensis, 67,339–345.

Shafiq-ur-Rehman (2006). Endosulfan toxicity and its re-duction by selenium: A behavioral, hematological andperoxidative stress evaluation. The Internet Journalof Toxicology, 3(1). www.ispub.com/.../endosulfan_toxicity.html.

Silva-Souza, A. T., Machado, P. M., & Almeida, S. C.(2000). Haematology of fish from Tibagi River. I.Differential white blood cell counts in Pimelodus mac-ulatus females. Boletin do Instituto de Pesca, 26, 33–39.

Tavares-Dias, M., Ono, E. A., Pilarski, F., & Moraes,F. R. (2007). Can thrombocytes participate in the re-moval of cellular debris in the circulating blood ofteleost fish? A cytochemical study and ultrastructuralanalysis. Journal of Applied Ichthyology, 23(6), 709–712.

Triebskorn, R., Telcean, I., Casper, H., Farkas, A., Sandu,C., Stan, G., et al. (2007). Monitoring pollution inRiver Mures, Romania, part II: Metal accumulationand histopathology in fish. Environmental Monitoringand Assessment, 141(1), 177–188.

USEPA (1986). Test method for evaluating solid wasteReport, no. SW846, Washington, DC.

Weber, C., Peter, A., & Zanini, F. (2007). Spatio-temporalanalysis of fish and their habitat: A case studyon a highly degraded Swiss river system prior toextensive rehabilitation. Aquatic Science, 69, 162–172.