Alterations of insulin-like growth factor binding protein 3 (IGFBP-3) glycosylation in patients with breast tumours
Post on 26-Jun-2016
Embed Size (px)
(IGFBP-3)Breast tumourGlycosylationLectin-afnity chromatography
Design and methods: IGFBP-3 in serum samples from healthy women and from women with breastin terms of its concentration (IRMA), glycosylation moiety (lectin-afnity
Clinical Biochemistry 43 (2010) 725731
Contents lists available at ScienceDirect
evand tissues. Imbalances within the axis lead to aberrant signallingnetworks and over-expression of growth factors that are implicated inthe development and progression of breast cancer. Insulin-likegrowth factor binding proteins (IGFBP-1 to -6) are multi-functionalregulatory proteins that bind IGFs . They are differentiallyexpressed in cells, tissues and physiological uids. IGFBP-1, -2, -3and -4 are the most abundant IGFBPs in blood. IGFBP-3 is by far themain circulating IGFBP that maintains a circulating reservoir of IGFs,has a role in their transport and prolongs their half-life . Nativenon-glycosylated IGFBP-3 has a molecular weight of 29 kDa. In the
counterpart. It would appear that post-translational modications ofIGFBP-3 are important and are likely to affect IGFBP-3's regulatoryfunctions in breast tissue. Furthermore, IGFBPs can be cleaved byproteases into forms that possess either reduced or no afnity for IGFs.Prostate-specic antigen (PSA), cathepsin D and plasmin can cleaveIGFBP-3 and all of these proteases have been localised within breastcancer tissue .
Several large prospective and cohort studies have sought toidentify relationships between IGF-I, IGFBP-3 and breast cancer risk. Apositive association between IGF-I and breast cancer risk exists in pre-circulation IGFBP-3 exists as two glycofocontaining 3N-glycosylated sites (Asn-X-SerAsn109 and Asn172 (4, 4.5 and 5 kDa, resglycoform has two occupied N-glycosylat
Corresponding author. Tel.: +381 11 2 617 252; faxE-mail address: firstname.lastname@example.org (I. Barievi).
0009-9120/$ see front matter 2010 The Canadiandoi:10.1016/j.clinbiochem.2010.03.006lopment of the breast byc effects . The axis isolic activity within cells
glycosylated forms of IGFBP-3 are functionally similar in their abilityto bind IGFs. However, in T47D breast cancer cells glycosylated IGFBP-3 binds less to the cell surface compared to its non-glycosylatedexerting potent mitogenic and anti-apoptotialso a very important regulator of the metabThe growth hormone/insulin-likeis an essential modulator of growth anmore mannose, fucose, bisecting GlcNAc and terminal sialic acid residues.Conclusion: Our results showed that breast cancer progression causes alterations of IGFBP-3
glycosylation. The extent of changes increases with breast cancer severity. 2010 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
growth factor I (GH/IGF-I) axisd deve
45 kDa glycoform has all three . The presence of N-glycosylationsites affects the ability of the IGFBPs to adhere to the cell surfacewhichcan modify their afnity for IGF. Both non-glycosylated andResults: In patients with benign tumours the concentration and carbohydrate content of IGFBP-3 wasunaltered compared to healthy women. In patients with malignant tumours in most cases these twoparameters were unchanged, but there were women whose concentration of IGFBP-3 was reduced and itsstructure was altered. In non-surviving cancer patients the concentration of IGFBP-3 was signicantlyreduced and these molecules contained a greater amount of biantennary complex type N-glycans havingKeywords:Insulin-like growth factor binding protein-3 chromatography) and distribution of molecular species (immunoblotting).tumours was characterisedAlterations of insulin-like growth factor bpatients with breast tumours
Ivona Barievi a,, Romana Masnikosa a, Dragana Laa Institute for the Application of Nuclear Energy (INEP), University of Belgrade, Banatska 3b Institute of Oncology and Radiology of Serbia, Belgrade, Serbia
a b s t r a c ta r t i c l e i n f o
Article history:Received 13 October 2009Received in revised form 3 March 2010Accepted 10 March 2010Available online 19 March 2010
Objectives: Insulin-likedevelopment and progressiand IGF-II. Changes in the gaim of this study was to in
j ourna l homepage: www.e lsrms of 40 and 45 kDa/Thr) located at Asn89,pectively). The 40 kDaion sites whereas the
: +381 11 2 618 724.
Society of Clinical Chemists. Publishding protein 3 (IGFBP-3) glycosylation in
ndin a, Vera Golubovi b, Olgica Nedi a
11080 Belgrade, Serbia
owth factor binding protein-3 (IGFBP-3) is an important modulator off breast cancer as it regulates the amount of free, physiologically active IGF-Iosylation pattern within IGFBP-3 may affect its interaction with ligands. Theigate whether such changes occur during disease progression.
i e r.com/ locate /c l inb iochemmenopausal patients . No evidence supports a role for IGF-I inpredicting post-menopausal risk . Evidence linking IGFBP-3 tobreast cancer risk is inconsistent. The consequence of high circulatinglevels of IGFBP-3 on breast cancer risk in some studies is associatedwith increased breast cancer risk in pre-menopausal woman whereasother studies report a protective role, concluding that high IGFBP-3concentrations are associated with decreased cancer risk [7,12,14].Recent meta-analyses indicate a positive association between
ed by Elsevier Inc. All rights reserved.
726 I. Barievi et al. / Clinical Biochemistry 43 (2010) 725731circulating IGFBP-3 levels and pre-menopausal cancer risk . Incontrast, women with high IGF-I and low IGFBP-3 tend to be at highrisk for carcinoma in situ . Careful scrutiny of the circulating formsof IGFBP-3 has been suggested. Rinaldi et al. demonstrated that highlevels of functional IGFBP-3 (intact IGFBP-3 and some IGFBP-3fragments) could be associated with reduced breast cancer risk,whereas high levels of total IGFBP-3 (intact and fragmented) could beassociated with increased breast cancer risk . However, Li et al.failed to nd any association of intact, fragmented or total IGFBP-3with breast cancer risk . Espelund et al. reported that in breastcancer patients total IGFBP-3 was unchanged, whereas intact IGFBP-3was increased due to reduced IGFBP-3 protease activity . Contraryto most other studies, a recently performed large prospective studyrevealed an association between IGF-I, IGFBP-3 and breast cancer riskin post-menopausal woman . The prospective Nurses' HealthStudy II noted that both IGF-I and IGFBP-3 were not associated withbreast cancer risk in pre-menopausal patients .
The results from several studies that have investigated therelationship between IGF-I, IGFBP-3 and breast cancer risk are,therefore, unclear and contradictory. Such inconsistencies warranteda re-examination of the role of the IGF system in breast tumours. Inour current study we have measured the concentrations of IGF-I andIGFBP-3 in patients diagnosed with breast cancer (benign ormalignant) and were indicated for surgery. Our goal was not tocompare the levels of IGF-I and IGFBP-3 between pre-menopausal andpost-menopausal woman, rather we sought to determine whetherchanges in the glycosylation pattern of IGFBP-3 were involved in theoverall alteration of the circulating IGF/IGFBP system and whetherthese changes were related to malignancy. As protein glycosylationinuences stability and function of the IGF system components and aschanges in N-linked glycosylation occur during cancer development,we characterised the carbohydrate moiety of IGFBP-3 in breasttumour patient groups using eight lectin-afnity columns withdifferent carbohydrate specicities. This allowed us to identifysaccharide residues present in IGFBP-3 from patients and to comparethe glycomic prole with that characteristic for IGFBP-3 from healthywomen. Another approach was to investigate the presence andrelative abundance of different IGFBP-3 molecular species in patientswith breast tumour. Therefore, our aim was primarily focused onalterations of IGFBP-3 glycosylation during disease progression. Thesealterations may reect IGFBP-3 stability, half-life, binding capacityand, as a consequence, they may regulate IGF bioavailability.
Serum samples were obtained from healthy adult women (n=20,age 5010.2 years) and from women diagnosed with breast tumourindicated for surgery (n=53, age 5711.3 years). All womencompleted a questionnaire at interview containing details of personalcharacteristics and reproductive history. Patients were treated withchemotherapy for several months prior to surgery in the Institute ofOncology and Radiology of Serbia, Belgrade, Serbia during 2007.Patients included in this study were those indicated for surgery in theperiod JulyAugust 2008. After post-operative histopathologicalanalysis and follow up that lasted 1 year after the surgery the patientswere divided into three groups: group 1 patients diagnosed withbenign breast disease, survivors (n=8, age 5811.6 years), group2 patients diagnosed with breast cancer, survivors (n=36, age 5611.6) and group 3 non-surviving patients diagnosed with breastcancer (n=9, age 5510.3). All the patients were also consideredclustered as pre-menopausal and post-menopausal women. Thesesubgroups, however, did not show any differences in the measuredparameters. Therefore, we decided not to analyse them further as
such. 13% of patients underwent short surgical interventions(tumorectomia, up to 30 min) whereas 3% of the patients hadextensive surgery (resectio partialis mamme). 52% of patientsunderwent longer operations (Mastectomia radicalis that lastedfrom 50 to 175 min) whereas 32% of patients underwent Quad-rantectomia cum dissectioaxillae dex (surgery lasted from 70 to140 min).
The BMI of the volunteers and patients ranged between 24.5 and26.7 kg/m2. No statistical difference in the BMI between the studygroups was observed (pN0.05). Blood samples from the patients werecollected in the morning after a 12-hour fast and before the surgeryfrom the patients. The serumwas separated by centrifugation, dividedinto aliquots and stored at 20 C until use. All volunteers andpatients agreed that their sera could be used in this study according tothe current ethical standards (Helsinski Declaration of 1975, asrevised in 1983).
Determination of serum IGF-I and IGFBP-3 concentration
The IGF-I concentration was measured by RIA-IGF-I (INEP-Belgrade, Serbia) using 125I-labelled IGF-I . The assay wasstandardised against WHO reference material 87/518. The serumIGFBP-3 concentration was measured by an immunoradiometricassay IRMA-IGFBP-3 (INEP-Belgrade, Serbia) using afnity puried125I-labelled polyclonal goat anti-IGFBP-3 antibodies (DiagnosticSystems Laboratories Inc., Webster, TX, USA). The assay wasstandardised against the DSL-6600 IGFBP-3 IRMA kit .
Lectin-afnity chromatography was performed using eight columnswith the following agarose-immobilised lectins : 1) Con A (lectinfrom Canavalia ensiformis binds strongly to high-mannose (Man) typeN-glycans), 2) succinylated WGA (succinylated wheat germ agglutininhas specicity towards GlcNAc14GlcNAc, but does not bind sialicacid (Sia) residues), 3) RCA I (Ricinus communis agglutinin I is a lectinspecic for N-glycans containing a bisecting GlcNAc in a sequenceGal14GlcNAc; high afnity binding indicates the presence of corefucosylation), 4) ECL (Erythrina cristagalli lectin has the highest afnitytowards Gal14GlcNAc disaccharides, but sialylation of the latterabolishes lectin binding), 5) SNA (Sambucus nigra agglutinin is specicfor Sia26Gal and Sia26GalNAc sequences present in N-glycans),6) UEA (Ulex europaeus agglutinin is specic for terminal 12 linkedL-Fuc residues), 7) PHA-E (Phaseolus vulgaris erythroagglutinin recog-nises complex type oligosaccharides and binds with the highest afnityto biantennary chains that contain a bisecting GlcNAc residue which is14-linked to the Man in the trimannosyl core) and 8) PHA-L(Phaseolus vulgaris leukoagglutinin also binds complex type oligosac-charides, but it prefers multiantennary chains with GlcNAc16residues). All agarose-immobilised plant lectins were from VectorLaboratories (Burlingame, CA, USA). All buffers and sugar solutionswereprepared following procedures recommended by the producer and thedetails are described in Table 1.
Serum samples (0.1 mL) were pre-incubated at 4 C overnightwith 125I-IGF-I (3105 cpm) to allow 125I-IGF-I binding to IGFBP-3and then applied to the lectin columns. 125I-IGF-I binds to all IGFBPs,but 125I-IGF-IIGFBP-3 complexes are those that predominantlyinteract with lectins due to the fact that IGFBP-3 is the most abundantIGF-binding protein in postnatal circulation . The samples wererecirculated through the columns at 25 C for 1 h to ensure maximalbinding. The unbound material was washed away with 20 mL of thecorresponding buffer that was used for equilibration. The elution ofbound complexes was performed using two consecutive steps (all thedetails are in Table 1), the two specic elutions were at pH 7.5 and pH3.0. PHA-E- and PHA-L-agarose columnswere eluted in one step usingacetic acid. The radioactivity (cpm) within the collected fractions
(1 mL) was determined in a -counter. The recovery of radioactivity
extent compared to healthy women, group 1 and group 2 patients.The IGFBP-3 concentration was signicantly reduced only in group 3
ion at pH 7.5 Specic elution at pH 3.0
l--glucopyranoside, 0.2 Mannopyranoside in 0.02 M HEPES
0.2 M methyl--glucopyranoside, 0.2 Mmethyl--mannopyranoside in 0.2 M acetic acid
c in 0.01 M HEPES 0.5 M GlcNAc in 0.2 M acetic acid0.01 M HEPES 0.2 M Lac in 0.2 M acetic acid0.01 M HEPES 0.2 M Lac in 0.2 M acetic acid0.01 M HEPES 0.5 M Lac in 0.2 M acetic acid0.01 M HEPES 0.1 M Fuc in 0.2 M acetic acid
0.1 M acetic acid0.1 M acetic acid
e same cations that were used for equilibration.
727I. Barievi et al. / Clinical Biochemistry 43 (2010) 725731from the columns in all cases ranged from 97% to 100%. All theindividual samples used in this study were analysed using all eightlectin columns.
To determine the presence of different forms of IGFBP-3 withrespect to molecular mass, four separate pools of sera were prepared(one for each of the four study groups) by mixing equal volumes ofindividual sera. All these pools were separately applied to the eightcolumns (without prior incubation with 125I-IGF-I) as describedabove. The fractions (both from pH 7.5 and from pH 3.0 elutions) thatcorresponded to those with the highest binding of 125I-IGF-I(determined previously) were collected from each column, pooledand immediately neutralised using 2 M TrisHCl pH 8.9. The poolswere dialysed rst against distilled water for 3 h at room temperatureand then against 0.9% NaCl overnight at 4 C. After being concentratedto approximately 1 mL using Microcon centrifugal lter devices(10 kDa cut-off membrane, Millipore, Billerica, MA, USA) the sampleswere subjected to SDS-PAGE and immunoblotting.