Insulin-like growth factor family in malignant haemangiopericytomas: the expression and role of insulin-like growth factor I receptor

J. Pathol. 188: 69–75 (1999)

INSULIN-LIKE GROWTH FACTOR FAMILY INMALIGNANT HAEMANGIOPERICYTOMAS: THE

EXPRESSION AND ROLE OF INSULIN-LIKE GROWTHFACTOR I RECEPTOR

š 1*, . 2, 3, 4, 5 . 2

1Division of Molecular Medicine, Ruder Boškovic Institute, HR-10000 Zagreb, Croatia2Department of Cell Biology, Neurobiology and Anatomy, College of Medicine, University of Cincinnati, Cincinnati, Ohio, U.S.A.

3Department of Internal Medicine, Clinical Hospital ‘Sestre Milosrdnice’, HR-10000 Zagreb, Croatia4Department of Nuclear Medicine and Molecular Pathophysiology, Clinical Hospital Osijek, Huttlerova 4, 31000 Osijek, Croatia

5Clinical Hospital for Pulmonary Diseases Jordanovac, Jordanovac 114, HR-10000 Zagreb, Croatia

SUMMARY

Haemangiopericytoma is a rare soft tissue tumour originating from the contractile pericapillary cells. Relatively little is known aboutits molecular pathogenesis. To address this issue, the insulin-like growth factor family (IGFs) was analysed in 19 tumours collected froma human tumour bank network. Seven of the tumours were associated with severe hypoglycaemia. Of these, six were retroperitoneal andone was located in the leg. 3 out of the 19 tumours (15·8 per cent) were positive for insulin-like growth factor I (IGF I) mRNA and 11were positive for IGF II mRNA (57·9 per cent). Almost 90 per cent of haemangiopericytomas expressed IGF I receptor (IGF IR) mRNA(17 out of 19), five (26·3 per cent) expressed IGF binding protein 1 (IGF BP1), three (15·8 per cent) expressed IGF BP2, and four (21per cent) exhibited IGF BP3 mRNA. All of the 14 haemangiopericytomas examined with regard to specific receptor binding were IGFIR positive, ranging from 1·2 to 16·2 per cent. Binding was much higher in IGF I/IGF IR positive tumours (15·3&0·7) than in IGF Inegative/IGF IR positive tumours (5·1&3·3). The potential role of IGF IR as a growth promoting factor in malignant haemangio-pericytoma was studied using antisense oligonucleotides and monoclonal antibody áIR3 that specifically inhibit IGF IR synthesis oractivity. 10 ìM IGF IR antisense oligonucleotides significantly inhibited the growth of haemangiopericytoma cells in culture, by around50 per cent; monoclonal antibody against IGF IR (áIR3) also significantly inhibited proliferation. The data suggest that IGF IRmay play an important role in the genesis and progression of malignant haemangiopericytomas. Copyright ? 1999 John Wiley &Sons, Ltd.

KEY WORDS—haemangiopericytoma; insulin-like growth factors/receptors; hypoglycaemia

*Correspondence to: Professor K. Pavelic, Ruder Boškovic Insti-tute, Bijenicka 54, P.O.B. 1016, HR-1000 Zagreb, Croatia. E-mail:[email protected].

Contract/grant sponsor: Croatian Ministry of Sciences andTechnology; Contract/grant numbers: P1104, P1103, 101751.

INTRODUCTION

Haemangiopericytoma is a rare soft tissue tumouroriginating from pericytes, a small population of con-tractile pericapillary cells. These often large and slowgrowing tumours represent less than one per cent of allvascular neoplasms, the most common site being theretroperitoneum, followed by the pelvis, oesophagus andlower extremities. Hypoglycaemia is a rare metabolicmanifestation of IGF activity that has been describedin patients with mesenchymal tumours, includinghaemangiopericytoma and haemangiosarcoma.1–3 Hae-mangiopericytomas can be accompanied by insulin-independent hypoglycaemia1,3–7 that appears to bemediated by tumour-derived IGF.

The insulin-like growth factors (IGF I and IGF II)and IGF receptor (IGF IR) may play a role in tumouri-genesis. IGF I and IGF II, IGF IR and six high-affinitybinding proteins (IGBPs1–6) belong to the insulin

CCC 0022–3417/99/060069–07$17.50Copyright ? 1999 John Wiley & Sons, Ltd.

growth factor (IGF) family or to a large group ofproteins that bind insulin. These growth factors andtheir receptors exhibit anabolic,8 insulin-like9 andmitogenic activities.10,11 In addition, IGFs induce differ-entiation of fetal tissue,12 regulate tissue growth,13,14

participate in tissue regeneration15 and promotedifferentiated cellular functions.16

The IGF IR appears to be an important factor in thedevelopment of certain tumours. It belongs to a familyof transmembrane tyrosine kinase receptors and isrequired for optimal growth in vitro and in vivo. IGF IRalso appears to be required for the establishment andmaintenance of the transformed phenotype in sometumour cells.16,17 In aggregate, these observations indi-cate that in some tumours the IGF IR plays a centralrole in the pathway to cell transformation and totumourigenesis.

In vivo, a significant decrease in the number of IGFIRs in some tumour types can be accompanied bymassive cell death. Conversely, overexpression of thereceptor often appears to protect cells from apoptosis.Several human tumour types have amplified or over-expressed the genes encoding IGF IR, consistent withthe proposition that elevated levels of the IGF IR have aprotective effect against programmed cell death in vivo.18

Received 10 February 1998Revised 1 December 1998

Accepted 21 January 1999

70 K. PAVELICu ET AL.

There are abnormalities in the expression or amplifi-cation of the genes for the IGF IR and its ligands1,16

which can lead to transformation. When the IGF IRnumber falls below physiological levels, transformationis inhibited or even reversed. Associated with its ligands,IGF IR activates its two major substrates, insulin recep-tor substrate 1 and Sch,19,20 which propagate a signalcascade through ras and raf, to the nucleus.21 Since littleis known in this area about haemangiopericytoma, wehave examined the expression and role of the IGFs, andspecifically of IGF IR, in the growth of this tumourtype.

MATERIALS AND METHODS

Tissue specimens

29 haemangiopericytomas were studied by routinehistology and immunohistochemistry, as well as bystandard molecular biology methods. Frozen sectionsand paraffin-embedded tissues were collected over 10years in different institutions and obtained through thetumour bank network.22,23 Tumour sites are indicated inTable I. The majority of cases occurred in men (19males, 10 females) with a mean age at presentation of 58years for males (3–87) and 53 years for females (29–80).

Copyright ? 1999 John Wiley & Sons, Ltd.

The tissues were snap frozen in liquid nitrogen shortlyafter surgical removal and stored at "80)C. 4 ìm frozensections were cut and mounted on glass slides. A sectionfrom each paraffin block was stained with haematoxylinand eosin (H & E) to confirm the identity of the tissueanalysed. Sections adjacent to those used for H & Estaining were used for RNA analysis. Of the 29 tumours,6 were malignant haemangiopericytomas. The criteriafor determining malignancy were the occurrence ofmetastases, pleomorphism, necrosis and the presence ofgreater than 4 mitoses per 10 high-power fields.

RNA extraction

For RNA extraction, 400 ìl of tissue homogenate wasmixed with 1 ml of solution containing 4 mM guanidin-ium thyocianate, 25 mM sodium citrate, pH 7·0, 5 g/lsarkosyl and 100 mM 2-mercaptoethanol. The sampleswere vigorously shaken and the dissolved tissue fractionsplit into two tubes: 125 ìl of 2 M sodium acetate(pH 4·0), 500 ìl of water saturated phenol and 100 ìlchloroform:isoamyl alcohol (24:1) was added to eachtube. The samples were mixed vigorously, incubated for15 minutes on ice and centrifuged at 800 g for 10minutes at 4)C. The upper phase was collected andprecipitated overnight at "20)C by addition of an equal

Table I—IGFs mRNA positivity in haemangiopericytoma tissues

Casenumber Tumour site IGF I IGF II IGF-IR IGF BP1 IGF BP2 IGF BP3

1 ^ Retroperitoneum + + + " " "2 Lung " " " " " "3 Pelvis " " + + " "4 Pelvis " " " " " "5 Intracranial " " + " " "6 ^ Retroperitoneum + " + " " +7 ^ Retroperitoneum " + + " " "8 ^ Retroperitoneum " + + " " "9 Intracranial + " + " " +

10 Pelvis11 Pelvis12 Lower extremity13 Lower extremity " + + " + "14 Lung + + + + + +15 Lower extremity " + + + " "16 Retroperitoneum "17 Pelvis18 Retroperitoneum " " + " " "19 ^ Retroperitoneum " + + + " "20 ^ Lower extremity " + + " " "21 ^ Retroperitoneum " + " " "22 Pelvis23 Retroperitoneum24 Infant hand " + " + +25 Oesophagus26 Lower extremity27 Retroperitoneum +28 Lung " + + + " "29 Lung " + + " " "

Arrows indicate patients with severe hypoglycemia.

J. Pathol. 188: 69–75 (1999)

71IGF IN MALIGNANT HAEMANGIOPERICYTOMAS

volume of isopropanol. After centrifugation at 15 800 gfor 20 minutes at 4)C, the pellet was again extractedwith 150 ìl, 4 M guanidinium thiocyanate, 25 mMsodium citrate, pH 7·0, 5 g/l sarkosyl and 100 mM2-mercaptoethanol as well as 150 ìl chloroform:isoamylalcohol (24:1) and centrifuged at 15 800 g for 2 minutesat 4)C. The aqueous phase was again extracted with anequal volume of chloroform:isoamyl alcohol and centri-fuged as before. RNA was precipitated from the upperphase by adding 1/10 volume of 2 mM sodium acetateand 2·5 volumes of absolute ice cold ethanol and incu-bated for 4 hours at "20)C. The pellet obtained bycentrifugation at 15 800 g for 20 minutes at 4)C waswashed with 70 per cent ethanol and centrifuged at15 800 g for 5 minutes at room temperature. The visibleRNA pellet was dissolved in 20 ìl of sterile, deionizedwater and the concentration was determined spectro-photometrically. Endogenous ribonuclease was inhib-ited by Rnasin. All other recommended precautionswere taken to avoid ribonuclease activity.

Reverse transcriptase—PCR

For reverse transcription (42)C for 1 hour) 5 ìl dena-tured RNA (7–10 minutes, 70)C), 2 ìl of reaction bufferII (100 mM Tris HCl, pH 8·3, 500 mM KCl, PerkinElmer), 2 ìl dithiothreitol (100 mmol/l), 200 U ofMoloney murine leukaemia virus reverse transcriptase(BRL) and 30 pmol primer were used in a total volumeof 20 ìl. The predicted product size is indicated in Fig. 1.The cDNA was heat denatured (95)C, 5 minutes) beforePCR. The reaction was performed in a Perkin-Elmer-Cetus Thermal Cycler in triplicate, and alwaysaccompanied by a positive and negative (minus RNA)control. Amplification was carried out in 100 ìl volume,with the addition of Taq polymerase (1·25 U, Perkin-Elmer-Cetus), MgCl2 at a final concentration of2·5 mM, 8 ìl PCR reaction buffer II (100 mM Tris HC2,pH 8·3, 500 mM KCl, Perkin Elmer) and 120 pmol ofouter primers. The 40-fold amplification profiles under-went denaturation at 95)C for 65 seconds, annealing at42)C for 2 minutes and extension at 72)C for 2 minutes.The final extension step lasted 10 minutes. The PCRproducts were analysed by agarose gel electrophoresis(2 per cent agarose stained with 0·5 mg/ml ethidiumbromide).

IGF I binding studies

400 ìg of membrane protein was incubated for 5hours at 4)C with 125I-labelled IGF I in the presence orabsence of an excess of IGF I to determine non-specificbinding. The final incubation volume was adjusted to0·5 ml with Tris MgCl2 buffer containing 0·1 per centbovine serum albumin (Sigma). A tumour was consid-ered positive for IGF IR when the specific binding washigher than 1 per cent of total counts per 400 ìg of

24
membrane proteins.
Cultivation of haemangiopericytoma cells

Surgical specimens were collected into sterile vesselscontaining RPMI 1640 medium with 15 per cent heat-


inactivated fetal calf serum.3 Informed consent wasobtained from all patients from whom surgical speci-mens were obtained, and the study was approved bylocal ethics committees. Tumour tissue, without necro-sis, was mechanically disrupted, placed into medium, cutinto small pieces and pressed through a nylon sieve.Cells were introduced into T-75 flasks with RPMI 1640medium supplemented with 15 per cent fetal calf serum,1 per cent calf serum, 1 per cent glutamine and 20 mMHEPES. Tumour cell cultures were maintained as mon-olayers. The initial inoculum as about 2#105 cells perT-75 flask, and cells were cultivated in RPMI 1640medium supplemented with transferrin (5 ìg/ml), selenicacid (5 ng/ml) and HEPES (20 mM).

Fig. 1—Insulin-like growth factor I and II, insulin-like growth factor Ireceptor and insulin-like growth factor binding proteins 1, 2 and 3mRNA in haemangiopericytoma tissue obtained by RT-PCR. Num-bers above the bands indicate a tumour sample from the correspond-ing patient in Table I

Cell proliferation

This was determined by the MTT dye reduction assay.Haemangiopericytoma cells were grown overnight inmedium after plating in 96-well plates (2#104 cells/well)at 37)C with 5 per cent CO2. Cells were incubated inserum-free medium containing 0·1 per cent bovineserum albumin, 5 ìg/ml transferrin, 5 ng/ml selenic acid,

J. Pathol. 188: 69–75 (1999)


antibiotics and with oligonculeotides or antibodies asexperiments dictated. The MTT assay was performedafter 72 hours of treatment with antisense or senseoligonucleotides (7·5 ìM). Absorbance was measured at570 nm with an ELISA plate reader (Labsystems,Multiscan MS). Data were expressed as the percentageincrease in cell growth compared with untreatedcontrols.

Antisense and sense oligonucleotides and monoclonalantibodies

Insulin-like growth factor I receptor antisense olig-nucleotide 5* TCCTCCGGAGCCAGACTT 3* andIGF IR sense oligonucleotide 5* AAGTCTGGCTCGGAGGA 3* were synthesized as mixed phosphorothioates.The IGF IR antisense oligonucleotide was complemen-tary to codons 21–29 of the signal sequence of thehuman IGF IR. Monoclonal antibody IR3, directedagainst IGF IR was purchased from Oncogene Science(Uniondale, NY).

The primers used for each of the members of the IGFfamily were as follows: IGF I 5* CTACAGTGAAGATGCACACCATGTCC 3* and 5* CCTGTAGTTCTTGTTTCCCTGCACTCC 3* (amplified 397 bp fragment),IGF II 5* TGGGAATGCCAATGGGGAAGTCGATG 3* and IGF II 5* TCACTTCCGATTGCTGGCCATCTCTG 3* (amplified 541 bp fragment), IGF IR5* TCCTCCGGAGCCAGACTT 3* and 5* AAGTCTGGCTCCGGAGGA 3* (amplified 720 bp fragment),IGF BPI 5* ACCTCTGCACGCCCTCACCC 3* and 5*CAGGGATCCTCTTCCCATTCCAAGGGTAGA E*(amplified 446 bp fragment), IGF BP2 5* AGGTTGCAGACAATGGCTGAT 3* and 5* GTAGAAGAGATGACACTCGG 3* (amplified 497 bp fragment) IGF BP35* GCTGACTCTGCTGGTGCTGCT 3* and 5* GAGGAACTTCAGGTGATTCAGTG 3* (amplified 646 bpfragment), â-actin 5* ATCATGTTTGAGACCTTCAACACCCC 3* and 5* CATCTCTTGCTCGAAGTCCAGGGCGA 3* (amplified 317 bp fragment).

RESULTS

Tumour sites

29 cases of haemangiopericytoma were subjected tomolecular genetic and biochemical analysis. Of these tenwere located in the retroperitoneum, six in the pelvis,four in the lung and five in the lower extremities; twowere intracranial, one was located in an infant’s handand one in the oesophagus. Seven tumours were associ-ated with severe hypoglycaemia, of which six werelocated in the retroperitoneum and the other in the leg(Table I). Of the 29 cases, 6 were malignant (cases 1, 10,11, 20, 21 and 27) (Table I).

IGF mRNAs in tumour tissue

To determine whether haemangiopericytoma cellscontained mRNAs for IGF I, IGF II, IGF IR, IGFBP1, IGF BP2 or IGF BP3, total RNA was subjected to


RT-PCR analysis. Results are shown in Tables I and IIand in Fig. 1. 4 out of the 19 tumours tested (21·0 percent) exhibited IGF I mRNA and 11 showed IGF IImRNA (57·9 per cent). Almost 90 per cent of haeman-giopericytomas exhibited IGF IR mRNA (17 out of 19).IGF BP1 mRNA was expressed in five tumours (26·3 percent), IGF BP2 mRNA in three (15·8 per cent), and IGFBP3 mRNA in four (21 per cent). None of the IGFfamily of mRNAs was detected in normal tissue adjacentto tumour. Tumour cell RNA contained â-actintranscripts in all cases, as did normal adjacent tissue(positive control).

Blood levels of insulin, C-peptide, IGF I, IGF II andglucose

Those patients with malignant haemangiopericytomaand who also manifested severe hypoglycaemia wereevaluated for ciruclating insulin, C-peptide, IGF I andglucose concentrations (Table III). Patients 1 and 6 hadelevated blood levels of IGF I. All other parametersmeasured were normal. In contrast to the other hypogly-caemic patients, tumour tissue from patients 1 and 6contained IGF I mRNA (Table II). Glucose, insulin,C-peptide and IGF I in blood were normal in theremainder of this hypoglycaemic subpopulation. Alltumours tested in this subpopulation of patients werepositive for IGF II mRNA (cases 1, 7, 8, 19, 20, 21) andfor IGF IR mRNA (cases 1, 6, 7, 8, 19, 20).

Table II—Expression of IGF family mRNA in haemangioperi-cytoma tissues, normal adjacent tissue and blood

IGFs*

Tumourtissue

positivetotal (%)

Normaltissue

adjacentto tumour Blood

IGF I 3/19 0/10 2/2IGF II 11/19 0/10 2/2IGF IR 17/19 0/10 2/2IGF BP1 5/19 0/10 2/2IGF BP2 3/19 0/10 2/2IGF BP3 4/19 0/10 2/2â actin 19/19 10/10 2/2

*Obtained by RT-PCR.

Specific binding to IGF IR

To characterize further the specific IGF I binding, theeffect of incubation time and temperature on the specificbinding of radio-labelled IGF I to the membranes oftumour cells was determined. Equilibrium was rapidlyachieved at 4)C, with maximal binding between one andfive hours. At subsequent times, there was reducedbinding. Similarly, binding was diminished at 20)C and37)C. Fig. 2 shows the distribution of 14 haemangioperi-cytomas as a function of their IGF IR levels. All 14tumours showed specific binding of IGF I, ranging from1·2 per cent to 16·2 per cent. When IGF I negative/IGFIR positive tumours were compared with both IGF I

J. Pathol. 188: 69–75 (1999)


positive/IGF IR positive tumours, it was observed thatbinding was much higher in the IGF I/IGF IR positivetumours (15·3&0·7) than in the IGF I negative/IGF IRpositive tumours (5·1&3·3). Control samples (IGF Inegative/IGF IR negative) were also negative withregard to specific receptor binding.

The role of IGF IR in tumour growth

To study the potential role of IGF IR as a growth-promoting receptor in malignant haemangiopericytoma,

Table III—Glucose, insulin, C-peptide and IGF I in blood of hypoglycaemic patients bearing haemangio-pericytoma

Casenumber

Blood Tumour tissue

Glucose*(mmol/l)

Insulin(mIU/l)

C-peptide(nmol/l)

IGF I(IU/l) IGF I* IGF II† IGF IR†

1 1·2 <3 0·1 3300 + + +6 1·8 <3 0·2 2830 + N.D. +7 0·9 5 0·8 816 " + +8 2·5 14 1·0 111 " + +

19 2·0 25 1·2 215 " + +20 1·2 5 0·2 441 " + +21 1·1 5 0·2 180 " + N.D.

N.D.=not determined.*Normal values for glucose (5·2–6·2), insulin (5–25), C-peptide (0·2–1·0), IGF I (600–2200).†Determined in tumour tissue by RT-PCR.

Fig. 2—Percentage of specific binding of IGF I to IGF IR (percentage of total radioactivity). Dark columns represent both IGF I/IGFIR positive tumours and open columns represent IGF I negative/IGF IR positive. Numbers under the columns identify individualpatients (see Table I). Patients 25 and 26 were IGF I negative/IGF IR positive by means of immunochemistry


antisense oligonucleotides to IGF IR and monoclonalantibody áIR3 that specifically inhibit IGF IR activitywere used. At 10 ìM, antisense to IGF IR inhibited thegrowth of haemangiopericytoma cells by around 50 percent (control media—100 per cent). In contrast, thecorresponding sense oligonucleotides had no effect.There was also no detectable effect of antisense or senseIGF IR oligonucleotide treatment on the growth offibrosarcoma cells. Consistent with this result, áIR3monoclonal antibody inhibited haemangiopericytomaproliferation by 40 per cent (Table IV).

J. Pathol. 188: 69–75 (1999)


Table IV—Selective inhibition of cell growth by antisense and antibody targeted to IGF IR inhaemangiopericytoma and fibrosarcoma cells

Treatment Concentration

Percentage of controls*

Fibrosarcoma Haemangiopericytoma

Antisense 0·1 ìM 97 881·0 ìM 95 63†

10·0 ìM 89 49†Sense 0·1 ìM 92 97

1·0 ìM 90 10310·0 ìM 101 100

áIR 3 mAb 0·025 ìg/ml 110 950·25 ìg/ml 113 972·5 ìg/ml 108 71

10·0 ìg/ml 87 59†

*Control, media (100 per cent).†p<0·01.

DISCUSSION

In this study, more than 90 per cent of haemangioperi-cytomas expressed insulin-like growth factor I receptors.A majority of tumours expressed IGF II, whereas only afew expressed IGF I. The importance of IGF IR in thedevelopment and/or maintenance of haemangiopericyto-mas is suggested by the observation that abrogationof receptor function by antisense oligonucleotidesor anti-IGF IR monoclonal antibody áIR3 inhibitedhaemangiopericytoma tumour cell proliferation in vitro.

As a member of a large family of protein-tyrosinekinases, IGF IR can play a role in normal and abnormalproliferative processes.16 Several lines of evidence indi-cate that activation of the IGF I receptor is a conver-gence point for the transduction of mitogenic signalsinitiated by other tyrosine kinase growth factor recep-tors. Overexpression of normal receptors, includingepidermal growth factor receptor (EGF R), colony-stimulating factor I receptor (CSFIR) and insulin-likegrowth factor I receptor (IGF IR), causes ligand-dependent cellular transformation,1,16,21 suggesting thatthe absolute number of growth factor receptors may bea determinant of cellular transformation. Prager et al.25

showed that a truncated form of IGF IR can actas a dominant negative inhibitor and abrogate ligand-dependent cellular transformation and tumourigenesismediated by wild-type IGF IR, both in vitro and in vivo.

Ligand–receptor function can be abrogated by utiliz-ing monoclonal receptor antibodies such as áIR3,26 thepolyanionic compound suramine27 or phosphorylation-defective analogues of the IGF I ligand.1 A series ofexperiments based on antisense strategies against theIGF-IR RNA28–30 have unequivocally shown that adecrease in the number of IGF IRs causes a reversal ofthe transformed phenotype, as measured by colonyformation in soft agar. The reduction in the number ofIGF IRs was achieved by two basic strategies; stabletransfection of a plasmid expressing antisense RNAcomplementary to the IGF I receptor RNA, or incu-bation of cells with antisense oligonucleotides directedagainst the IGF IR RNA. The antisense strategies


inhibited or completely abrogated cell proliferation insome cell lines.16

The IGF I and IGF II stimulate cell growth ofdifferent tissues and cell lines.31 Under physiologicalconditions, IGF I and IGF II are the principal ligandswhich activate the IGF I receptor and stimulate phos-phorylation of its tyrosine kinase domains. Both IGF Iand IGF II mRNAs are expressed in some tumours andare thought to be synthesized by stromal elements. TheIGF action is modulated by insulin-like growth factorbinding proteins (IGF BPs), which act by eitherenhancing or inhibiting the action of IGFs.32 Ourresults indicate that IGF BP can be expressed byhaemangiopericytoma cells themselves.

Some patients with haemangiopericytoma manifestsevere hypoglycaemia, which is a rare event in patientssuffering from malignant tumours. The causes ofhypoglycaemia are varied; sometimes it is a consequenceof ectopic production of insulin or insulin-like growthfactors. These factors could act as autocrine stimulatorsenhancing tumour cell growth.3 Expression of growthfactors and their receptors is a known feature of humansoft tissue tumours, with malignant tumours more fre-quently expressing single or multiple factors and recep-tors. Some other growth factors and their receptors maybe an important influence on tumourigenic potential insoft tissue tumours, in particular haemangiopericyto-mas. Perosio and Brooks33 described a high frequency ofimmunoreactivity of nerve growth factor (43 per cent) inhaemangiopericytomas.

Hatva et al.34 have analysed the expression of vascularendothelial growth factor (VEGF) and its receptors, therelated placental growth factor (P1GF) and the endothe-lial receptors FLT4 and Tie in haemangioblastomasand haemangiopericytomas. VEGF mRNA was up-regulated in all of the haemangiopericytomas studiedand was highly expressed in the stromal cells of hae-mangioblastomas. VEGF is up-regulated during themalignant progression of tumours. Up-regulation of theendothelial growth factors and receptors may result inautocrine or paracrine stimulation of haemangioperi-cytomas. In addition, the same cells express genes

J. Pathol. 188: 69–75 (1999)


encoding growth factors binding to the endothelialreceptor tyrosine kinases, thus probably establishingautocrine and paracrine signal transduction to thestimulation of growth of these tumours.

Our results clearly show that IGF IR is expressed inalmost all haemangiopericytomas and that inhibition ofreceptor activity can abrogate tumour proliferationin vitro. In aggregate, our data indicate that IGF IRplays an important role in the development andprogression of malignant haemangiopericytoma.

ACKNOWLEDGEMENTS

This work was supported by Program Grant No.P1104, P1103 and Project Grant No. 101751 from theCroatian Ministry of Sciences and Technology.

REFERENCES

1. McCaulay VM. Insulin-like growth factors and cancer. Br J Cancer 1992;65: 311–320.

2. Daughaday WH, Kapadia M. Significance of abnormal serum binding ofinsulin-like growth factor II in the development of hypoglycemia in patientswith non-islet-cell tumors. Proc Natl Acad Sci U S A 1989; 86: 6778–6782.

3. Pavelic K, Vrbanec D, Marusic S, Levanat S, Cabrijan T. Autocrine tumourgrowth regulation by somatomedin C: an in-vitro model. J Endocrinol 1986;109: 233–238.

4. Pavelic K. Aplastic carcinoma in diabetic mice: hyperglycemia-suppressedproliferation rate and insulin synthesis by tumor cells. J Natl Cancer Inst1979; 62: 139–141.

5. Pavelic K, Basic I, Pavelic J. Habituation of a mammary aplastic carcinomaon diabetic conditions. J Cancer Res Clin Oncol 1980; 97: 275–287.

6. Pavelic K, Popovic M. Insulin and glucagon secretion by renal adenocarci-noma. Cancer (Philad) 1981; 48: 98–100.

7. Pavelic K, Ferle-Vidovic A, Osmak M, Vuk-Pavlovic S. Synthesis ofimmunoreactive insulin in vitro by aplastic mammary carcinoma precondi-tioned in diabetic mice. J Natl Cancer Inst 1981; 67: 687–688.

8. Oelz O, Jakob A, Froesch ER. Nonsuppressible insulin-like activity(NSILA) of human serum. Eur J Clin Invest 1970; 1: 48–53.

9. Pierson RW, Termin HM. The partial purification from calf serum of afraction with multiplication stimulating activity for chicken fibroblasts incell culture and with nonsuppressible insulin-like activity. J Cell Physiol1972; 79: 319–330.

10. Moses AC, Nissley SP, Short PA, Rechler MM, Podskalny JM. Purificationand characterization of multiplication stimulating activity. Eur J Biochem1980; 103: 387–400.

11. Dechiara TM, Efstratiadis A, Robertson EJ. A growth-deficiency pheno-type in heterozygous mice carrying an insulin-like growth factor II genedisrupted by targeting. Nature 1990; 345: 78–80.

12. Schoenle E, Zapf J, Hauri C, Steiner T, Froesch ER. Comparison of an invivo effects of insulin-like growth factors I and II and of growth hormone inhypophysectomized rats. Acta Endocrinol 1985; 108: 167–174.

13. Van Buull-Offers V, Van den Brande JL. Cellular growth in organs of dwarfmice during treatment with growth hormone, thyroxine and plasma frac-tions containing somatomedin activity. Acta Endocrinol 1982; 99: 150–160.


14. Stiles CD, Sosenko IRS, D’Ercole AJ, Smith BT. Relation of kidney tissuesomatomedin-C/insulin-like growth factor I to postnephrectomy renalgrowth in the rat. Endocrinology 1985; 1: 2397–2401.

15. Bothwell M. Insulin and somatomedin MSA promote nerve growth factor-independent neurite formation by cultured chick dorsal root ganglionicsensory neurons. Prog Clin Biol Res 1983; 118: 225–231.

16. Baserga R. The insulin-like growth factor I receptor: a key to tumor growth?Cancer Res 1995; 55: 249–252.

17. White MF, Kahn CR. The insulin signaling system. J Biol Chem 1994; 249:1–4.

18. Resnicoff M, Coppola D, Sell C, Rubin R, Ferrone S, Baserga R. Growthinhibition of human melanoma cells in nude mice by antisense strategies tothe type 1 insulin-like growth factor receptor. Cancer Res 1994; 54:4848–4850.

19. Crews CM, Erikson RL. Extracellular signals and reversible proteinphosphorylation: what to Mek of it all. Cell 1993; 74: 215–217.

20. Dickson RB, Thompson EW, Lippman ME. Regulation of proliferation,invasion, and growth factor synthesis in breast cancer by steroids. J SteroidBiochem Mol Biol 1990; 37: 305–316.

21. Yee D, Paik S, Lebovic GS, et al. Analysis of insulin-like growth factor Igene expression in malignancy: evidence for a paracrine role in humanbreast cancer. Mol Endocrniol 1989; 3: 509–517.

22. Spaventi R, Pecur L, Pavelic K, Pavelic ZP, Spaventi S, Stambrook PJ.Human tumor bank in Croatia: a possible model for a small bank as a partof the future European Tumor Bank network. Eur J Cancer 1994; 30A:419–419.

23. Pavelic K. Human tumor bank: editorial. Period Biol 1995; 97: 11–12.24. Peyrat JP, Bonneterre J, Djiane J, Dusanter R, Leroy-Martin B, Demaille

A. Insulin-like growth factor I receptor (IGF I-R) in human breast cancers:relation with estradiol and progesterone receptors (ER, PgR). Progress incancer research and therapy. In: Bresciani F, King RJB, Lippman ME,Raynaud JP, eds. Hormones and Cancer 3. New York: Raven Press, Ltd,1988; 223–226.

25. Prager D, Li H-L, Asa S, Lemed S. Dominant negative inhibition oftumorigenesis in vivo by human insulin-like growth factor I receptor mutant.Proc Natl Acad Sci U S A 1994; 91: 2181–2185.

26. Russel WE, Van Wyk JJ, Pledger WJ. Inhibition of the mitogenic effect ofplasma by a monoclonal antibody to somatomedin C. Proc Natl Acad SciU S A 1984; 81: 2389–2393.

27. Pollak M, Richard M. Suramin blockade in insulin-like growth factor Istimulated proliferation of human osteosarcoma cells. J Natl Cancer Inst1990; 82: 1349–1354.

28. Arteaga CL. Interference of the IGF system as a strategy to inhibit breastcancer growth. Breast Cancer Res Treat 1992; 22: 101–106.

29. Trojan J, Blossey BK, Johnson TR, et al. Loss of tumorigenicity of ratglioblastoma directed by episome-based antisense cDNA transcription ofinsulin-like growth factor I. Proc Natl Acad Sci U S A 1992; 89: 4874–4878.

30. Shapiro DN, Jones BG, Shapiro LH, Dias P, Houghton PJ. Antisense-mediated reduction in insulin-like growth factor I receptor expressionsuppresses the malignant phenotype of a human alveolar rhabdomyosar-coma. J Clin Invest 1994; 94: 1235–1242.

31. DeMellow JSM, Baxter RC. Growth hormone dependent insulin-likegrowth factor binding protein both inhibits and potentiates IGF stimulatedDNA synthesis in skin fibroblasts. Biochem Biophys Res Commun 1988; 156:199–204.

32. Ritvos O, Ranta T, Jalkanen J, et al. Insulin-like growth factor (IGF)binding protein from human decidua inhibits the binding and biologicalaction of IGF-I in cultured choriocarcinoma cell. Endocrinology 1988; 122:2150–2157.

33. Perosio PM, Brooks JJ. Expression of nerve growth factor receptor inparaffin-embedded soft tissue tumors. Am J Path 1988; 132: 152–160.

34. Hatva E, Böhling T, Jääskelainen M, Persico M, Alitalo K. Vasculargrowth factors and receptors in capilary hemangioblastomas and hemangio-pericytomas. Am J Path 1996; 148: 763–775.

J. Pathol. 188: 69–75 (1999)

Documents

Insulin-like growth factor family in malignant haemangiopericytomas: the expression and role of insulin-like growth factor I receptor