Transactions of the Royal Society of Tropical Medicine and Hygiene (2007) 101, 660—667

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ierhea l th .com/ journa ls / t rs t

Correlation between the components of theinsulin-like growth factor I system,nutritional status and visceral leishmaniasis

Claudia M.C. Gomesa,∗, Daniel Giannella-Netob, Monica E.A. Gamac,Julio Cesar R. Pereirad, Marliane B. Campose, Carlos E.P. Corbetta

a Department of Pathology, University of Sao Paulo Medical School, Avenida Dr. Arnaldo,455 — sala 1209, CEP 01246—903 Sao Paulo, SP, Brazilb Division of Endocrinology, University of Sao Paulo Medical School, Sao Paulo, SP, Brazilc Department of Pediatrics, Federal University of Maranhao, Sao Luıs, MA, Brazild Department of Epidemiology, University of Sao Paulo, Sao Paulo, SP, Brazile Evandro Chagas Institute, Belem, PA, Brazil

Received 28 April 2006; received in revised form 6 February 2007; accepted 6 February 2007Available online 17 April 2007

KEYWORDSVisceralleishmaniasis;Malnutrition;Insulin-like growthfactor;Insulin-like growthfactorbinding-protein 3;Growth hormone;Brazil

Summary The role of the insulin-like growth factor I (IGF-I) system and nutritional status wasstudied in 241 children from a Brazilian area endemic for visceral leishmaniasis (VL). Thirty-nine children had the active form, 20 were oligosymptomatic, 38 were asymptomatic and 144were not infected. Serum concentrations of growth hormone (GH), total and free IGF-I andIGF binding-protein 3 (IGFBP3) were measured by radioimmunoassay. Nutritional status wasevaluated by anthropometric indicators and biochemical measurements. Total and free IGF-Iand IGFBP3 were significantly reduced in the active form. Z scores for total and free IGF-I and forIGFBP3 were found to be significantly lower for active VL and oligosymptomatic individuals thanfor asymptomatic individuals, but never reached values ≤2 SD. Median values of weight-for-ageZ and height-for-age Z (HAZ) scores and albumin concentration were significantly different inthe active VL group compared with the other groups. Multiple discriminant analysis selectedalbumin and HAZ score as predictors of active and oligosymptomatic VL. The lack of correlationbetween auxological data and serum concentrations of the GH/IGF axis components suggestedthat the primary cause of retarded growth in children with active VL is not dependent on IGF-I

or IGFBP3, but rather on VL intrinsic factors that might secondarily involve the GH/IGF axis.© 2007 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rightsreserved.

∗ Corresponding author. Tel.: +55 11 3081 7799; fax: +55 11 3081 7799.E-mail address: gomescla@usp.br (C.M.C. Gomes).

0035-9203/$ — see front matter © 2007 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.trstmh.2007.02.017

lrp

llZ5cav

lMcmi(pad

iast1

ocmtia

2

2

TlB1iftaacom

2

Toarea endemic for VL, were residents of the localities of

Insulin-like growth factor I and visceral leishmaniasis

1. Introduction

Visceral leishmaniasis (VL) is characterized by a specificimmunological dysfunction consequent to macrophage par-asitism by Leishmania, which generates a wide spectrum ofclinical and immunological manifestations. These may bereversed either by specific treatment, or spontaneously inimmunologically competent individuals. In Brazil, the sever-ity and incidence of VL is higher among children of 0 to 9years of age, with 75% of the cases corresponding to chil-dren younger than 5 years old. The higher susceptibilityto infection in this latter age group may be accounted forby immunological immaturity coupled with malnutrition, asobserved in children from endemic areas (Badaro et al.,1986a; Campos Jr, 1995; Dye and Williams, 1993; GrimaldiJr et al., 1980; Marzochi et al., 1994).

The clinical manifestations of VL can be grouped intothree patterns. The asymptomatic form is characterizedby positive serology to Leishmania antigens and no clinicalmanifestations. The oligosymptomatic form is also found inendemic areas, with positive serology and discrete or mildsigns and/or symptoms such as fever, mild hepatomegaly,hypergammaglobulinemia and erythrocyte sedimentationrate (Gama et al., 2004). The classical form correspondsto the fully manifested disease, characterized by severehepatosplenomegaly, fever, pancytopenia and severely com-promised health conditions. Malnutrition is evident in thisdisease, with pronounced hypoalbuminemia and hypergam-maglobulinemia (Badaro et al., 1986b).

During the course of the clinically manifested disease,the Leishmania species responsible for VL affect severalorgans, but show a marked preference for the spleen andliver, which is associated with hypertrophy and hyperpla-sia of the mononuclear phagocytic system, where parasiteproliferation occurs (Wilcoks and Manson-Bahr, 1972).

The events that take place at the onset of infectiongreatly influence the course of the disease. This initial phaseis characterized by an inflammatory process in which sev-eral mediators and tissue elements participate, includinggrowth factors and cytokines. These may be either benefi-cial to the host by contributing to parasite elimination orfavor the survival and proliferation of Leishmania, directlyor indirectly, in a hostile environment within macrophages(Barral et al., 1993; Gomes et al., 1997, 1998; Lonardoniet al., 2000). Recent in vitro and in vivo studies show theeffect of insulin-like growth factor I (IGF-I), a molecule thathas been an object of great interest due to its diversifiedaction and pleiotropism on Leishmania spp. (Gomes et al.,1997, 1998, 2001; Goto et al., 1998).

Somatic growth is thought to be regulated essentiallyby the endocrine actions of IGF-I. According to the revisedsomatomedin hypothesis, the production of IGF-I in theliver and other tissues by growth hormone (GH) stimula-tion is crucial for post-natal growth (LeRoith et al., 2001;Spagnoli and Rosenfeld, 1996). Le Roith et al. (2001) raisedthe hypothesis that the action of circulating IGF-I is mainlyrelated to the somatotropic axis. More than 90% of circulat-ing IGF-I is bound to IGF binding-protein 3 (IGFBP3), which

forms a 150 kDa protein complex with the acid-labile subunit(IGFALS) (Underwood, 1996). The synthesis of both IGF-I andIGFBP3 is GH-dependent. Low serum concentrations of IGF-I are affected by nutritional factors and occur in chronic

VRpD

661

iver diseases (Scharf et al., 1996). GH-resistant state witheduced concentrations of IGF-I has been described in somearasitic diseases (Nedic et al., 2004; Orsini et al., 2001).

IGFBP3 production occurs in a variety of tissues, but theiver is its main site of synthesis, in the sinusoid endothe-ial cells and Kupffer cells (Collet-Solberg and Cohen, 1996;immermann et al., 2000). Only IGFBP3 is part of the0 kDa binary complex (Hardouin et al., 1987). Serum con-entrations of IGF-I and IGFBP3 have been shown to havesignificant positive correlation with body height, height

elocity and body mass index (Gelander et al., 1999).Nutritional status is a predominant factor in the estab-

ishment of pre- and post-natal IGF-I concentrations.alnutrition determines a state of relative resistance to GH,aused by a reduction in the number of GH receptors in itsore severe forms, or a post-receptor defect, which results

n lower gene expression of IGF-I, especially in the liverThissen et al., 1994). Quantitative determination of IGF-Ilasma concentrations is useful for evaluating short staturend normal IGF-I concentrations generally to rule out a GHeficiency (Adan et al., 1994).

VL has a profound effect on the nutritional status ofnfected patients, with a marked reduction in fat reservesnd muscle mass. These findings appear to be related to higherum concentrations of TNF-�, frequently observed duringhe disease phase (Barral Netto et al., 1991; Harrison et al.,986).

The aim of the present study was to evaluate the rolef the IGF system in the severity of VL. To this end, serumoncentrations of total and free IGF-I, IGFBP3 and GH wereeasured. A correlation analysis was performed between

hese components of the IGF system, anthropometricalndexes and nutritional status of children from an endemicrea for VL.

. Materials and methods

.1. Study area

he study was conducted in the municipal district of Raposa,ocated in the northern region of Maranhao State (MA),razil. The area comprises a population of approximately7 088 inhabitants, of which about 38% are 0- to 14-year-oldndividuals (IBGE, 2000). This district is considered endemicor VL, with an annual incidence of 3.3 cases/1000 inhabi-ants (Gama et al., 2004). The Center for Tropical Pathologynd Social Medicine of the Federal University of Maranhaond the National Health Foundation (FUNASA, MA) have beenonducting local studies for the last 15 years, with the aimf controlling the most prevalent local diseases (VL andalaria).

.2. Study group

he 241 children studied, aged 0 to 72 months old, withr without L. (Leishmania) chagasi infection, from a known

ila Nova and Bom Viver within the municipal district ofaposa, MA, Brazil. The enrollment of participants and sam-le collection took place from April 2000 to January 2002.uring this period, meetings with community leaders of the

6

mHAt4scLc

tcTtt7apgaaatdcol

cedgces

2

2Aepw

f(

(2ciwyaa

2cMaU

ct0c(3

ir

2

TamW

oaacbaF8I2n7

2

Nobr

owveN

caa

aftegwfac

62

unicipal district of Raposa and technicians of the Nationalealth Foundation were also held for monitoring purposes.ntibody-ELISA using L. (L.) chagasi antigen was employedo identify children infected with Leishmania. The children,7.7% from Vila Nova and 52.3% from Bom Viver, were clas-ified as serologically negative for Leishmania antigens (144ontrol children), asymptomatic with positive serology foreishmania, oligosymptomatic or presenting with the clini-al disease (active VL).

The three diagnostic categories for VL were defined inhe following way. (1) There were 38 asymptomatic/infectedhildren with positive anti-Leishmania antibody ELISA. (2)here were 30 oligosymptomatic children that presentedwo or more of the following signs/symptoms in associa-ion: fever; diarrhea, with duration equal or longer than

days, without mucus or blood and negative copropar-sitological serial examinations; mucosal and cutaneousallor; mild hepatomegaly ≤5 cm at the right costal mar-in and/or interstitial pneumonitis, with no significantlterations as detected by nonspecific laboratory testing;nd positive serology at clinical evaluation regardless ofmastigote scores in bone marrow puncture aspirates orhe results of the Montenegro skin test. (3) Finally, 39 chil-ren were diagnosed with active VL, patients presentinglinical and laboratory profiles characteristic of the acuter classical forms of VL, which are well established in theiterature.

The children were enrolled in the study using cards thatontained information regarding personal identification andpidemiological, clinical and laboratory data, as well as aescription of the features of their infection by L. (L.) cha-asi. After enrollment and the collection of samples, thehildren were submitted to follow-up, with monthly clinicalvaluations and Montenegro skin testing and serology everyemester.

.3. Evaluation of nutritional status

.3.1. Anthropometryn anthropometric examination was carried out in order tovaluate the prevalence of the various forms of caloric-roteic malnutrition. This consisted of measuring bodyeight and height.

The malnutrition index was established according to theollowing Z scores: height-for-age (HAZ), weight-for-ageWAZ) and weight-for-height (WHZ).

Body weight was measured with a Salter spring scaleMinisterio da Previdencia e Assistencia Social, 1983) with5 kg maximum capacity and 100 g divisions. The height ofhildren younger than 2 years old was measured in a hor-zontal position, total length from head to heels, using aooden anthropometer (AHRTAG type). In children aged 2ears of age or over, height measurements were taken withwooden infantometer according to the method of Barros

nd Victoria (1991).

.3.2. Evaluation of albumin and ferritin serumoncentrationseasurements were carried out with a Cobas Integra 700nalyzer (Roche Diagnostic Systems, Inc., Indianapolis, IN,SA), and each measurement was run according to a spe-

tvseP

C.M.C. Gomes et al.

ific protocol. Albumin concentrations were measured byhe bromocresol green method with assay-protocol ALB,-046. The following reference values were considered:hildren (4 days—14 years old), 3.8—5.4 g/dl; adolescents14—18 years old), 3.2—4.5 g/dl; adults (>18—60 years old),.5—5.0 g/dl.

Serum ferritin concentrations were determined by anmmunoturbidimetric test, assay-protocol FERR, 0-255. Theeference values varied from 10 to 300 �g/l.

.4. Endocrine evaluation

o evaluate GH/IGF-I, axis serum concentrations of GH, totalnd free IGF-I and IGFBP3 were determined by radioim-unometric assays (Diagnostic Systems Laboratories Inc.,ebster, TX, USA).

Normal limits of total and free IGF-I, IGBP3 and GH werebtained from a group of 241 control children, who werellocated into two groups according to their chronologicalge: newborns (n = 42) and Tanner I (n = 199). Ninety-five per-ent confidence limits for total IGF-I were considered to beetween 39.40 and 255.70 ng/ml for newborns and 35.10nd 481 ng/ml for Tanner I, 10 months to 9 years of age.or free IGF-I, 95% confidence limits were between 0.26 and.56 ng/ml for newborns and 0.26 and 5.80 ng/ml for Tanner. For IGFBP3, 95% confidence limits were between 0.67 and.48 �g/m for newborns and 1.01 and 5.66 �g/ml for Tan-er I. The GH 95% confidence limits were between 0.07 and.00 ng/ml (data not published).

.5. Data analysis

utritional status was evaluated using the National Centerf Health Statistics (NCHS) reference standard, which haseen adopted by WHO as an international anthropometricaleference standard (Waterlow et al., 1977).

Z scores (i.e. how much the value deviated from the meanf the normal population in units of standard deviations)ere computed for WAZ, WHZ and HAZ scores using Epi Info,ersion 6 (CDC, Atlanta, GA, USA). A cut-off of ≤2.0 SD wasstablished for each calculated Z score in relation to theCHS normal reference.

Z scores for total and free IGF-I and IGFBP3 were alsoomputed using the normal group of children mentionedbove as a reference for mean and SD related to each vari-ble.

Means and SD were computed for continuous variables,nd median, minimum and maximum values were computedor variables that did not adhere to the normal distribu-ion by the Kolmogorov-Smirnov test. The significance ofach variable was evaluated individually within and betweenroups using ANOVA. The Kruskal-Wallis test was appliedhenever necessary. Pearson’s correlation test was used

or the correlation analysis of anthropometric, biochemicalnd endocrine data. A multivariate analysis (multiple dis-riminant analysis) was performed to choose the variables

hat might compose discriminant functions, combinations ofariables that would discriminate the study groups so as toelect variables or functions that might predict the differ-nt clinical states. The level of significance was fixed at< 0.05.

tVg1

3

Eeacjfsst(dcc(

tc1P

dirdstccab

Insulin-like growth factor I and visceral leishmaniasis

3. Results

No evidence of Chagas’ disease, Schistosoma mansoni infec-tion or American cutaneous leishmaniasis was found in themunicipal district of Raposa. The predominant age group(50%) varied from 13 to 24 months (data not shown). Withrespect to gender, distribution within the two groups studiedwas similar.

3.1. Nutritional indexes

Analysis of the nutritional profile showed that the highestnumber of children with Z scores below −2 SD was in theactive VL group. Evaluation of HAZ scores in the total popula-tion revealed a deficit of 7.1% in the children, of which 82.3%belonged to the active VL group and the remaining 17.7%were evenly distributed among the three other groups (datanot shown). A somewhat similar pattern was found for theWAZ score, with 5.6% of children from the total populationpresenting nutritional deficit, 92.8% of which were from theactive VL group and 7.2% from the oligosymptomatic group(data not shown). Regarding the WHZ score, 3.7% childrenof the total population presented nutritional deficit, 66.7%from the active VL group and 33.3% from the non-infectedgroup (data not shown).

The anthropometrical data analysis revealed that theactive VL group presented statistically significant differ-ences in median WAZ and HAZ scores (WAZ −1.55 andHAZ −1.78, respectively) compared with the asymptomatic(WAZ −0.14 and HAZ 0.29, respectively) and control(WAZ −0.16 and HAZ 0.20, respectively) groups (P < 0.05).Oligosymptomatic children presented significantly differentHAZ scores (−0.55) with respect to infected asymptomatic(0.29) and control (0.20) children. No statistically significantdifferences were found between asymptomatic groups andcontrols (Table 1).

Serum albumin and ferritin concentrations, regardedas objective and quantitative measures of nutritional

status, are also presented in Table 1. Serum albumin pre-sented statistically significant differences between active VL(2.74 ± 0.22 g/dl) and oligosymptomatic (2.97 ± 0.34 g/dl)children compared with the asymptomatic (4.22 ± 0.07 g/dl)and controls (4.36 ± 0.03 g/dl). Serum ferritin concentra-

vgsTb

Table 1 Anthropometrical and nutritional indexes for children fr

Control (n = 144) Asymptomatic (n = 3

Age (months)a 17.5 (3.0 to 60.0) 16.0 (3.0 to 49.0)Weight (kg)b 11.1 ± 0.2 10.7 ± 0.4Height (cm)b 82.1 ± 0.8 80.0 ± 1.8WHZa −0.09 (−2.65 to 1.97) −0.10 (−1.78 to 2.6HAZa 0.20 (−2.16 to 1.79) 0.29 (−3.67 to 1.3WAZa −0.16 (−1.85 to 1.96) −0.14 (−1.19 to 0.9Albumin (g/dl)b 4.36 ± 0.03 4.22 ± 0.07Ferritin (�g/dl)b 15.3 ± 0.8 18.8 ± 3.3

VL: visceral leishmaniasis; WHZ, HAZ and WAZ: weight-for-height, heiga Median (minimum/maximum).b Mean ± SEM.c P ≤ 0.05.

663

ions were found to be significantly higher in the activeL group (348.1 ± 131.4 �g/ml) compared with the otherroups (oligosymptomatic 30.8 ± 7.8 �g/ml; asymptomatic8.8 ± 3.3 �g/ml; controls 15.3 ± 0.8 �g/ml).

.2. Endocrine evaluation

valuation of the GH/IGF axis revealed significant differ-nces in serum concentrations of total and free IGF-I, IGFBP3nd, especially, of GH in the active VL and oligosymptomatichildren compared with the asymptomatic and control sub-ects, as shown in Table 2. Serum concentrations of total andree IGF-I, IGFBP3 and GH calculated for the four groupstudied are presented in Table 2. Active VL patients pre-ented statistically significant differences with respect tootal IGF-I Z score compared with asymptomatic children−1.89 vs. −1.49, P ≤ 0.05). There was also a significantifference between oligosymptomatic and asymptomatichildren (−1.89 vs. −1.49, P ≤ 0.05). The same statisticalonclusions were seen with free IGF-I Z and IGFBP3 Z scoresTable 2).

Mean serum GH concentrations were found to be sta-istically higher in the active VL group (2.1 ± 0.4 ng/ml)ompared with the other three groups (oligosymptomatic.1 ± 0.6, symptomatic 0.9 ± 0.1, and controls 0.9 ± 0.1;< 0.05) (Table 2).

The multiple correlation analysis between the endocrineata (total and free IGF-I, IGFBP3, GH) and the nutritionalndicators, such as WAZ, HAZ serum albumin and ferritin,egardless of clinical group, showed a statistically significantifference between HAZ and WAZ (r = 0.656) and albuminerum concentrations (r = 0.427) (data not shown). Similarly,here was a statistically significant correlation between theoncentrations of free IGF-I and free IGF-I Z score with theoncentrations of total IGF-I, total IGF-I Z score, IGFBP3nd IGFBP3 Z score. No significant correlation was foundetween GH or ferritin and any of the other variables.

Multiple discriminant analysis showed that three latent

ariables summarized the differences between theseroups, originating for three functions, two of which repre-ented 99.9% of the variations of the original measurements.he third function contributed little to discriminatingetween the VL clinical forms. Analysis of the correlation

om the municipal district of Raposa, Maranhao State, Brazil

8) Oligosymptomatic (n = 20) Active VL (n = 39)

22.0 (8.0 to 48.0) 24.0 (8.0 to 72.0)12.3 ± 0.6 10.7 ± 0.583.7 ± 2.0 82.8 ± 2.0

5) 0.60 (−1.43 to 1.97)c −0.63 (−4.22 to 1.55)c

0) −0.55 (−3.08 to 1.85)c −1.78 (−3.96 to 1.39)c

2) 0.01 (−2.32 to 2.44) −1.55 (−4.12 to 0.36)c

2.97 ± 0.34c 2.74 ± 0.22c

30.8 ± 7.8 348.1 ± 131.4c

ht-for-age and weight-for-age Z scores, respectively.

664 C.M.C. Gomes et al.

Table 2 Serum concentration of total and free insulin-like growth factor I (IGF-I), IGF binding-protein 3 (IGFBP3) and humangrowth hormone (GH) from uninfected control, asymptomatic, oligosymptomatic and active visceral leishmaniasis (VL) groups

Control (n = 144) Asymptomatic (n = 38) Oligosymptomatic (n = 20) Active VL (n = 39)

Total IGF-I (ng/ml)a 64.36 ± 4.41 56.62 ± 9.18 51.78 ± 22.97 39.31 ± 14.69Z (total IGF-I)b −1.44 (−1.90 to 0.53) −1.49 (−1.90 to 1.05) −1.89 (−1.96 to 1.62)c −1.89 (−1.97 to 4.39)c

Free IGF-I (ng/ml)a 0.94 ± 0.09 0.91 ± 0.20 0.59 ± 0.27 0.41 ± 0.15Z (free IGF-I)b −0.92 (−1.56 to 3.56) −1.05 (−1.55 to 4.06) −1.50 (−1.56 to 2.76)c −1.50 (−1.56 to 1.34)c

IGFBP3 (�g/ml)a 2.26 ± 0.07 2.13 ± 0.11 1.80 ± 0.24 1.21 ± 0.15Z (IGFBP3)b −0.22 (−1.73 to 7.42) −0.18 (−2.00 to 4.51) −1.01 (−1.70 to 1.89)c −1.37 (−2.25 to 5.70)c

GH (ng/ml)a 0.90 ± 0.10 0.90 ± 0.10 1.10 ± 0.60 2.10 ± 0.40c

a Mean ± SEM.b Median (minimum/maximum).c P ≤ 0.05.

Fct

ma2

raioogwvvt

4

Iwcaa

clanuwmidvVcdglt1ndb

cRi(wp1tC

irJoprsa

igure 1 Distribution of the clinical manifestations of vis-eral leishmaniasis (VL) in 241 Brazilian children according tohe discriminant functions. HAZ: height-for-age Z score.

atrix allowed for interpretation of these three functionsccording to major loadings: function 1, albumin; function, HAZ score; and function 3, WAZ score.

Figure 1 shows the distribution of all the subjects, seg-egated by the first two discriminant functions. The controlnd asymptomatic groups were found to be closely super-mposed in the upper right corner, i.e. with higher valuesf both HAZ score and serum albumin. The active VL andligosymptomatic groups, lying mainly in the left half of theraph, i.e. presenting lower values of function 1 (albumin),ere further distinguished by function 2 (HAZ), the greateralues of which suggest oligosymptomatic cases while loweralues suggest active American visceral leishmaniasis (AVL),he value of −2 SD as a probable threshold.

. Discussion

t is known that the nutritional status of individuals infectedith Leishmania spp. plays a significant role in the clini-al evolution of VL, especially among 0—5-year-old children,nd yet the determining factors that modulate such statusre still not fully understood. The present findings indi-

ccipt

ate that patients with VL present more pronounced weightoss and a deficit in HAZ and WAZ scores compared withsymptomatic or oligosymptomatic patients and individualsot infected with L. (L.) chagasi, with their Z-score val-es being significantly lower. The most frequent age groupithin the study group analyzed varied between 13 and 24onths. The anthropometric and biochemical data express-

ng the nutritional status of 241 children from the municipalistrict of Raposa revealed significantly different medianalues for WAZ and HAZ scores in the group with activeL compared with the oligosymptomatic, asymptomatic andontrol groups. Albumin serum concentrations also showedifferences between the oligosymptomatic and active VLroups and the other two, corroborating previously pub-ished findings, where it was shown that VL itself can affecthe nutritional condition of individuals (Barral-Netto et al.,991; Harrison et al., 1986; Pearson et al., 1992). A perti-ent study by Caldas et al. (2002) showed that malnutrition,eficit in height and weight, was not a risk factor in infectiony L. (L.) chagasi in Vila Nova and Bom Viver.

The deficit in the height/weight ratio, an indicator ofhronic malnutrition, observed in the municipal district ofaposa was lower than those observed in surveys carried out

n the Brazilian state of Maranhao (24%) in 1991 and in Brazil10.4%) in 1996 (Brazilian Ministry of Health), but agreesith more recent values that show a fall in the country’srevalence of height deficits. Within the period between974 and 1996, Brazilian prevalence of height deficits forhe total urban area shifted from 26.6 to 7% (Monteiro andonde, 2000).

The elevated ferritin concentrations presently observedn children with active VL agree with previous findingsegarding VL caused by Leishmania (L.) donovani (Al-urayyan et al., 1995), and are compatible with a statef chronic inflammation. In such conditions, ferritin can beroduced by several cells and body tissues, resulting in aelatively high pool of available iron that is not necessarilytored in the phagocytic mononuclear cell system (Konijn etl., 1981).

Serum albumin concentrations were clearly reduced in

hildren with active AVL. However, an important findingoncerning this condition of fully manifested disease is thenversion of the albumin/globulin ratio that results fromolyclonal activation. Globulins are highly expressed, par-icularly IgG, some specific to Leishmania and others to

tgotImp2ccbm(

awwm1aasdatc

aaaGit

ctot

ptfp

AS

FP

C

Insulin-like growth factor I and visceral leishmaniasis

autoantibodies, especially rheumatoid factor. The reductionin albumin plasma concentrations may reflect an intestinallesion or may be due to a switch to acute phase proteinsynthesis within hepatocytes (Wasunna et al., 1995).

The impact of VL on the plasma concentrations of IGF hasnot been evaluated to date and deserves greater attention,as IGF-I may be directly related to the malnutrition conditionobserved in VL.

Mean concentrations of IGF-I (total and free) and IGFBP3were significantly reduced in children of the active VL group(P < 0.05) compared with the other groups. Median valuesnormalized by age (total and free IGF-I, and IGFBP3 Z scores)for the active group were found to be lower with respectto the asymptomatic group, presenting a statistically sig-nificant difference (P < 0.05). Oligosymptomatic childrenpresented significant differences in comparison with con-trols. Z scores were low but never reached values ≤−2 SD,showing that despite being below average, they should notbe considered abnormal according to the currently accepteddiagnostic criteria for growth deficit due to GH deficiency.This suggests that the changes in the GH/IGF axis may besecondary to the effect of infection upon hepatocyte func-tion or, additionally, due to hypertrophy and hyperplasia ofthe phagocytic mononuclear cell system with the activationof macrophages, the Leishmania host cells. IGF-I synthe-sis is sensitive to nutritional status and IGF-I is negativelyregulated by caloric-proteic malnutrition (Thissen et al.,1994). In vivo studies have demonstrated that catabolicconditions such as septicemia, endotoxemia or cytokine-induced inflammation can also diminish IGF-I concentrationsin experimental animals (Fan et al., 1995; Lang et al.,1998; Lelbach et al., 2001) and humans (Lang et al., 1997).Previous studies show that both hepatic synthesis and thesecretion of IGF-I are impaired under catabolic inflammatoryconditions (Fan et al., 1995). Such reduction may be relatedto a direct effect of acute phase mediators on IGF-I synthe-sis, as IL-6, IL-1 and TNF-� are known to have an inhibitoryeffect on IGF-I biosynthesis (Lelbach et al., 2001).

The main source of endocrine IGF-I arises from the liver.Less than 5% of circulating IGF-I is found in its free form,while over 90% is bound to IGFBP3, the main carrier-proteinof IGF-I (Thissen et al., 1996). Patients with chronic ill-nesses have been shown to present reduced concentrationsof IGFBP3 due to the action of IGFBP3-specific proteases(Lelbach et al., 2001; Timmins et al., 1996). In the presentstudy, a considerable reduction in plasma concentrations ofIGFBP3 was observed in children with active AVL. However,it is thought that such reduction is not related to an increasein the activity of IGFBP3-specific proteases, as VL patientsalso present reduced free IGF-I concentrations, suggestingthat IGF-I synthesis is reduced. Hepatomegaly is an impor-tant feature of VL, where the liver displays hyperplasia ofcells of the phagocytic mononuclear cell system and thedilation of sinusoids, with highly parasitized Kupffer cells.An experimental study conducted by Lindoso et al. (2004) inhamsters challenged with L. (L.) chagasi showed a periportaland parenchymatous inflammatory infiltrate with apoptotic

Kupffer and inflammatory cells during days 15 to 45 of infec-tion. Such hepatic changes may impair the production ofthe IGFBPs, which are mainly synthesized by Kupffer cells,thereby leading to a significant reduction in available IGF-I-binding proteins.

ECeoo

665

The current finding of increased serum GH concentra-ions in the active VL group, compared with the otherroups, agrees with other published findings on infectionsf a chronic nature in individuals with compromised nutri-ional status (Naranjo et al., 2002; Thissen et al., 1999).t has been demonstrated that caloric-proteic malnutritionodulates the liver response to GH and, consequently, theroduction of liver-derived endocrine IGF-I (Naranjo et al.,002; Oster et al., 1996). There is a resulting increase inirculating GH and reduction in plasma concentrations of cir-ulating IGF-I. Such observation of resistance to GH mighte partly attributed to the lack of an inhibitory feedbackechanism due to the low concentrations of circulating IGF-I

Thissen et al., 1999).No correlation between plasma concentrations of IGF-I

nd the anthropometrical variables studied (WAZ and HAZ)ere found. Similar results were reported by other authors,ho found a poor correlation between IGF-I and anthropo-etric indexes (Ormarsdottir et al., 2001; Wilson et al.,

991). The lack of correlation between the auxological datand plasma concentrations of GH/IGF axis components (totalnd free IGF-I, and IGFBP3) observed in the present studyuggest that the primary cause of retarded growth in chil-ren with active VL is not dependent on GH or on IGF-Ixis components, but is likely to be linked to intrinsic fac-ors related to the disease conditions that might secondarilyompromise the axis as a whole.

The evaluation of liver function in individuals with VLnd its possible correlation with the GH/IGF axis showedpositive correlation between IGFBP3 Z scores and serum

lbumin. This suggests that the changes observed in theH/IGF-I axis are probably due to the nutritional deficit and

mmunological impairment that affect children with VL inhe age group investigated.

Multiple-discriminant analysis selected two functions,haracterized as albumin and HAZ score, which allow forhe separation of the active VL and oligosymptomatic formsf the disease, as well as helping to differentiate them fromhe control and asymptomatic conditions.

These findings invite further analyses regarding otherotentially pertinent variables as candidates, contributingo the effort of identifying reliable discriminatory markersor the various clinical forms of VL, work that is being dulylanned.

cknowledgement: The authors are grateful to Dr Silviaaldiva for the helpful discussions.

unding: Fundacao de Amparo a Pesquisa do Estado de Saoaulo (FAPESP 99/09721-3; 99/09974-9).

onflict of interest: None declared.

thical approval: The study was approved by the Ethicsommittees of the University of Sao Paulo and the Fed-ral University of Maranhao and fulfilled the requirementsf the Resolution of the National Council for Health 196/96n research work involving humans.

6

R

A

A

B

B

B

B

B

C

C

C

D

F

G

G

G

G

G

G

G

H

H

I

K

L

L

L

L

L

L

M

M

M

N

N

O

O

66

eferences

dan, L., Souberbielle, J.C., Brauner, R., 1994. Diagnostic mark-ers of permanent idiopathic growth hormone deficiency. J. Clin.Endocrinol. Metab. 78, 353—358.

l-Jurayyan, N.A.M., Al-Nasser, M.N., Al-Fawaz, I.M., Al Ayed, I.H.,Al Herbish, A.S., Al-Mazrou, A.M., Al Sohaibani, M.O., 1995.The haematologcal manifestations of visceral leishmaniasis ininfancy and childhoood. J. Trop. Pediatr. 41, 143—148.

adaro, R., Jones, T.C., Carvalho, E.M., Sampaio, D., Reed, S.G.,Barral, A., Teixeira, R., Johnson Jr., W.D., 1986a. New perspec-tives on a subclinical form of visceral leishmaniasis. J. Infect.Dis. 154, 1003—1011.

adaro, R., Jones, T.C., Lourenco, R., Cerf, B.J., Sampaio, D., Car-valho, E.M., Rocha, H., Teixeira, R., Johnson Jr., W.D., 1986b.A prospective study of visceral leishmaniasis in an endemic areaof Brazil. J. Infect. Dis. 154, 639—649.

arral, A., Barral-Netto, M., Yong, E.C., Brownell, C.E., Twardzik,D.R., Reed, S.G., 1993. Transforming growth factor beta as a vir-ulence mechanism for Leishmania braziliensis. Proc. Natl. Acad.Sci. USA 90, 3442—3446.

arral-Netto, M., Badaro, R., Barral, A., Almeida, R.P., Santos,S.B., Badaro, F., Pedral-Sampaio, D., Carvalho, E.M., Falcoff,E., Falcoff, R., 1991. Tumor necrosis factor (cachectin) in humanvisceral leishmaniasis. J. Infect. Dis. 163, 853—857.

arros, F.C., Victoria, C., 1991. Epidemiologia da Saude infantil: ummanual para diagnosticos comunitarios. HUCITEC/UNICEF, SaoPaulo, Brazil.

aldas, A.J., Costa, J.M., Silva, A.A., Vinhas, V., Barral, A., 2002.Risk factors associated with asymptomatic infection by Leishma-nia chagasi in north-east Brazil. Trans. R. Soc. Trop. Med. Hyg.96, 21—28.

ampos Jr., D., 1995. Clinical and epidemiological features of Kala-Azar in children. J. Pediatr. (Rio J.) 71, 261—265.

ollet-Solberg, R.F., Cohen, P., 1996. The role of Insulin-like growthfactor binding proteins and the IGBP proteases in the modulatingIGF action. Endocrinol. Metab. Clin. North. Am. 25, 591—614.

ye, C., Williams, B.G., 1993. Malnutrition, age and the risk of par-asitic disease: visceral leishmaniasis revisited. Proc. Biol. Sci.254, 33—39.

an, J., Char, D., Kolasa, A.J., Pan, W., Maitra, S.R., Patlak, C.S.,Spolarics, Z., Gelato, M.C., Lang, C.H., 1995. Alterations inhepatic production and peripheral clearance of IGF-I after endo-toxin. Am. J. Physiol. 269, 33—42.

ama, M.E., Costa, J.M., Gomes, C.M.C., Corbett, C.E., 2004. Sub-clinical form of the American Visceral Leishmaniasis. Mem. Inst.Oswaldo Cruz 99, 889—893.

elander, L., Blum, W.F., Larsson, L., Rosberg, S., Albertsson-Wikland, K., 1999. Monthly measurements of insulin-like growthfactor I (IGF-I) and IGF-binding protein-3 in healthy prepuber-tal children: characterization and relationship with growth: the1-year growth study. Pediatr. Res. 45, 377—383.

omes, C.M., Goto, H., Corbett, C.E., Gidlund, M., 1997. Insulin-like growth factor-1 is a growth promoting factor for Leishmaniapromastigotes. Acta Trop. 64, 225—228.

omes, C.M.C., Monteiro, H.P.M., Corbett, C.E.P., Gidlund, M.,Goto, H., 1998. Insulin-like growth factor-I induces phosphory-lation in Leishmania (Leishmania) mexicana promastigotes andamastigotes. J. Eukaryot. Microbiol. 45, 352—355.

omes, C.M.C., Goto, H., Magnanelli, A.C., Monteiro, H.P., Soares,R.P.S., Corbett, C.E.P., Gidlund, M., 2001. Characterization ofthe receptor for insulin-like growth factor (IGF)-I on Leishmania

promastigotes. Exp. Parasitol. 99, 190—197.

oto, H., Gomes, C.M., Corbett, C.E., Monteiro, H.P., Gidlund, M.,1998. Insulin-like growth factor I is a growth-promoting factorfor Leishmania promastigotes and amastigotes. Proc. Natl. Acad.Sci. USA 95, 13211—13216.

O

C.M.C. Gomes et al.

rimaldi Jr., G., Moriearty, P.L., Hoff, R., 1980. Leishmaniamexicana: immunology and histopathology in C3H mice. Exp.Parasitol. 50, 45—56.

ardouin, S., Hossenlopp, P., Segovia, B., Seurin, D., Portolan, G.,Lassarre, C., Binoux, M., 1987. Heterogeneity of insulin-likegrowth factor binding proteins and relationships between struc-ture and affinity. 1. Circulating forms in man. Eur. J. Biochem.170, 121—132.

arrison, L.H., Naldu, T.G., Drew, J.S., Alencar, J.E., Pearson,R.D., 1986. Reciprocal relationships between undernutrition andthe parasitic disease visceral leishmaniasis. Rev. Infect. Dis. 8,447—452.

BGE, 2000. Contagem nacional de populacoes. Superintendencia deEstudos Geograficos e Socio-Economicos, Instituto Brasileiro deGeografia e Estatıstica, Rio de Janeiro.

onijn, A.M., Carmel, N., Levy, R., Hershko, C., 1981. Ferritinsynthesis in inflammation. II. Mechanism of increased ferritinsynthesis. Br. J. Haematol. 49, 361—370.

ang, C.H., Pollard, V., Fan, J., Traber, L.D., Traber, D.L.,Frost, R.A., Gelato, M.C., Prough, D.S., 1997. Acute alter-ations in growth hormone-insulin-like growth factor axis inhumans injected with endotoxin. Am. J. Physiol. 273, 371—378.

ang, C.H., Fan, J., Wojnar, M.M., Vary, T.C., Cooney, R., 1998. Roleof central IL-1 in regulation peripheral IGF-I during endotoxemiaand sepsis. Am. J. Physiol. 274, 956—962.

elbach, A., Scharf, J.G., Ramadori, G., 2001. Regulation ofinsulin-like growth factor-I and insulin-like growth factor bind-ing protein-1, -3 and -4 in cocultures of rat hepatocytesand kupffer cells by interleukin-6. J. Hepatol. 35, 558—567.

eRoith, D., Bondy, C., Yakar, S., Liu, J.L., Butler, A., 2001. Thesomatomedin hypothesis: 2001. Endocr. Rev. 22, 53—74.

indoso, J.A., Cotrim, P.C., Goto, H., 2004. Apoptosis of Leishma-nia (Leishmania) chagasi amastigotes in hamsters infected withvisceral leishmaniasis. Int. J. Parasitol. 34, 1—4.

onardoni, M.V., Russo, M., Jancar, S., 2000. Essential role ofplatelet-activating factor in control of Leishmania (Leishmania)amazonensis infection. Infect. Immun. 68, 6355—6361.

arzochi, M.C., Marzochi, K.B., Carvalho, R.W., 1994. Visceral leish-maniasis in Rio de Janeiro. Parasitol. Today 10, 37—40.

inisterio da Previdencia e Assistencia Social, 1983. Manual paraavaliacao antropometrica do estado nutricional de criancas ate7 anos de idade. MPAS/SPR, Brazil, p. 53.

onteiro, C.A., Conde, W.L., 2000. Secular trends in malnutritionand obesity among children in the city of Sao Paulo, Brazil(1974—1996). Rev. Saude Publica 34, 52—61.

aranjo, W.M., Yakar, S., Sanchez-Gomez, M., Perez, A.U., Setser,J., LeRoith, D., 2002. Protein calorie restriction affects nonhep-atic IGF-I production and the lymphoid system: studies using theliver-specific IGF-I gene-deleted mouse model. Endocrinol. 143,2233—2241.

edic, O., Nikolic, J.A., Baricevic, I., Jovanovic, B., Ilic, N., 2004.Insulin-like growth factors in patients with liver cysts. J. Clin.Lab. Anal. 18, 299—304.

rmarsdottir, S., Ljunggren, O., Mallmin, H., Olofsson, H., Blum,W.F., Loof, L., 2001. Circulating levels of insulin-like growthfactors and their binding proteins in patients with chronic liverdisease: lack of correlations with bone mineral density. Liver 21,123—128.

rsini, M., Rocha, R.S., Disch, J., Katz, N., Rabello, A., 2001.The role of nutritional status and insulin-like growth factor inreduced physical growth in hepatosplenic Schistosoma mansoni

infection. Trans. R. Soc. Trop. Med. Hyg. 95, 453—456.

ster, M.H., Levin, N., Fielder, P.J., Robinson, I.C., Baxter, R.C.,Cronin, M.J., 1996. Developmental differences in the IGF-I sys-tem response to severe and chronic calorie malnutrition. Am. J.Physiol. 270, 646—653.

U

W

W

W

W

907—912.

Insulin-like growth factor I and visceral leishmaniasis

Pearson, R.D., Cox, G., Jeronomo, S.M., Castracane, J., Drew, J.S.,Evans, T., De Alencar, J.E., 1992. Visceral leishmaniasis: a modelfor infection-induced cachexia. Am. J. Trop. Med. Hyg. 47,8—15.

Scharf, J.G., Schmitz, F., Frystyk, J., Skjaerbaek, C., Moesus,H., Blum, W.F., Ramadori, G., Hartmann, H., 1996. Insulin-likegrowth factor-I serum concentrations and patterns of insulin-like growth factor binding proteins in patients with chronic liverdisease. J. Hepatol. 25, 689—699.

Spagnoli, A., Rosenfeld, R.G., 1996. The mechanism by whichgrowth hormone brings about growth. The relative contribu-tions of growth hormone insulin-like growth factors. Endocrinol.Metab. Clin. North. Am. 25, 615—631.

Thissen, J.P., Ketelslegers, J.M., Underwood, L.E., 1994. Nutritionalregulation of the insulin-like growth factor. Endocrinol. Rev. 15,80—101.

Thissen, J.P., Ketelslegers, J.M., Maiter, D., 1996. Use of insulin-likegrowth factor-I (IGF-I) and IGF-binding protein-3 in the diagnosisof acromegaly and growth hormone deficiency in adults. GrowthRegul. 6, 222—229.

Thissen, J.P., Underwood, L.E., Ketelslegers, J.M., 1999. Regulationof insulin-like growth factor-I in starvation and injury. Nutr. Rev.

57, 167—176.

Timmins, A.C., Cotterril, A.M., Hughes, S.C., Holly, J.M., Ross, R.J.,Blum, W., Hinds, C.J., 1996. Critical illness is associated withlow circulating concentrations of insulin-like growth factors-Iand -II, alterations in insulin-like growth factor binding proteins,

Z

667

and induction of an insulin-like growth factor binding protein 3protease. Crit. Care Med. 24, 1460—1466.

nderwood, L.E., 1996. Nutritional regulation of IGF-I and IGFBP-3.J. Pediatr. Endocrinol. Metab. 9, 3003—3012.

asunna, K.M., Raynes, J.G., Were, J.B., Muigai, R., Sherwood,J., Gachihi, G., Carpenter, L., Mcadam, K., 1995. Acute phaseprotein concentrations predict parasite clearance rate duringtherapy for visceral leishmaniasis. Trans. R. Soc. Trop. Med. Hyg.89, 678—681.

aterlow, J.C., Buzina, R., Keller, W., Lane, J.M., Nichaman, M.Z.,Tanner, J.M., 1977. The presentation and use of height andweight data for comparing the nutritional status of groups ofchildren under the age of 10 years. Bull. World Health Organ.55, 489—498.

ilcoks, C., Manson-Bahr, P.E.C., 1972. Leishmaniasis, in: Manson’sTropical Diseases, seventeenth ed. Bailliere Tindall, London, pp.117—147.

ilson, D.M., Killen, J.D., Hammer, L.D., Litt, I.F., Vosti, C., Miner,B., Hayward, C., Taylor, C.B., 1991. Insulin-like growth factor-I as a reflection of body composition, nutrition and puberty insixth and seventy grade girls. J. Clin. Endocrinol. Metab. 73,

immermann, E.M., Li, L., Hoyt, E.C., Pucilowska, J.B., Lichtman,S., Lund, P.K., 2000. Cell-specific localization of insulin-likegrowth factor binding protein mRNAs in rat liver. Am. J. Physiol.Gastrointest. Liver Physiol. 278, 447—457.