Main

Wilms tumor is one of the most common solid tumors of childhood, occurring with a frequency of about 1:10 000 worldwide. Recent studies showed two tumor-suppressor genes on the short arm of chromosome 11, named WT1 at 11p13 and WT2 at 11p15.5. In contrast to WT1, the precise location and sequence of WT2 is not yet clear, but is likely to be in the region 11p15.5, where the IGF-II gene also resides(1, 2).

IGF-I and IGF-II are single-chain polypeptides, structurally related to insulin, with mitogenic and insulin-like metabolic activities(3, 4). There is increasing evidence that IGF-II plays an important role as a fetal growth factor(5–8), whereas IGF-I seems to be the most prominent growth factor after birth(9–12).

Previous reports have shown that the expression of IGF-II mRNA was up to 100 times elevated in some fetal tissue and in some embryonic tumors such as Wilms tumor(13–16). We have recently reported that IGF-II is synthesized and released to a higher than normal extent by Wilms tumor tissue(17). The released IGF-II exerts systemic effects on glucose homeostasis and tissue growth of some inner organs(17). This is further evidence for the importance of IGF-II as a growth factor in the growth of embryonic tumors.

The goal of the present study was to investigate further the function of IGF-II in the growth regulation of embryonic tumor tissue. We established Wilms tumor cell lines that might provide an in vitro model to investigate autocrine and paracrine regulation of human embryonic tumor cell growth by IGF-II.

METHODS

Materials. rhIGF-I and polyclonal IGF-II antibodies were a gift from Dr. J. Zapf (Department of Medicine, University of Zurich, Switzerland). rhIGF-II was purchased from KabiGen AB (Stockholm, Sweden). Recombinant human IGFBP-3 and monoclonal IGF-I antibodies were kindly provided by Dr. R. E. Humbel (Department of Biochemistry, University of Zurich). Human transferrin, porcine insulin, sodium selenite, and mouse IgG-K1 (MOPC-21) were from Sigma Chemical Co. (St. Louis, MO). Hydrocortisone was from Calbiochem Behring Diagnostics (La Jolla, CA) and suramin from Bayer AG (Leverkusen, Germany). DMEM, Hams' F-12 medium, penicillin, streptomycin, trypsin, and EDTA were from Bioconcept (Allschwil, Switzerland). DNase I was from Boehringer (Mannheim, Germany). Mouse monoclonal immunoglobulin α-IR3 (95% IgG-K1) was purchased from BIO-TRADE (Vienna, Austria).

Cell cultures. Wilms tumor cell cultures were established from explants of nude mice heterotransplants. For one series of experiments the tumor of a single mouse was prepared exactly as previously described(17). Briefly the tumor (2-3 g), which developed in a female nude mouse 7 wk after heterotransplantation, was cut into small pieces, approximately 1 mm3 in volume. Tumor fragments were treated with 0.25% trypsin and 0.2 mg/mL DNase I at 37°C for 5 min. The supernatant with cells was collected and mixed with an equal amount of FCS, and large fragments were retreated if necessary. The collected cell solution was centrifuged at 600 × g for 5 min. Tumor cells were placed in 100-mm plastic dishes (Becton Dickinson, Plymouth, England) containing culture medium with 10% FCS and penicillin-streptomycin.

A 1:1 mixture of Hams' F-12 medium and DMEM supplemented with sodium selenite (2.9 × 10-8 mol/L), insulin (10-6 mol/L), transferrin (5 μg/mL), and hydrocortisone (10-7mol/L) was used as culture growth medium. The explants were incubated at 37°C in a 5% CO2 atmosphere and fed with fresh growth medium every 3-5 d. After reaching confluence, cultures were split 1:3 using 0.05% trypsin with 0.02% EDTA, and for experiments they were used at passages 3-6. Direct Wilms tumor and normal kidney tissue was cultured using the same methods. These cells grew in DMEM with the same supplements.

A histopathologic examination of the cell cultures after several passages in serum-free medium revealed tumor cells, but of various differentiation. There was no evidence for significant contamination with non-tumor cells. Fibroblast-like cells disappeared during the passages in serum-free cell culture medium.

For determination of concentrations of IGF-I and IGF-II in cell culture medium, approximately 2 × 106 tumor or kidney cells were separately rinsed with PBS at pH 7.2 without Ca2+ and Mg2+, plated in 100-mm plastic dishes in serum-free culture medium (as described below), and maintained at 37°C in 5% CO2. Forty-eight hours later, conditioned medium samples from tumor and kidney cell culture were collected and stored at -20°C until they were submitted to Bio-Gel P-100 chromatography.

Determination of IGF-I and IGF-II in conditioned medium. The collected medium samples were lyophilized and redissolved in 2 mol/L acetic acid/0.1% Triton X-100. The samples were chromatographed on a column of Bio-Gel P-100 (150 × 1.5 cm) in 0.5 mol/L acetic acid as described by Haselbacher et al.(18, 19). To evaporate the acetic acid, aliquots were lyophilized after elution from Bio-Gel P-100, redissolved in 0.1 mol/L PBS (pH 7.4, 0.1% BSA), and assayed in duplicate in a RIA for IGF-I and IGF-II(20).

Increasing amounts of human IGF-II or redissolved conditioned medium samples after chromatography were incubated (4°C, 24 h) with a polyclonal antiserum to human IGF-II (1:2000 final dilution) and 125I-labeled human IGF II (≈10 000 cpm) in a final volume of PBS, pH 7.4, containing 0.1% BSA as described elsewhere(20). As a modification of the method of Blum et al.(21), an excess of 25 ng/tube recombinant human IGF-I was added to each tube during the incubation to block the interaction with residual binding sites of binding proteins after acetic acid gel filtration. IGF-II and IGF-I were labeled with125 I by the chloramine-T method as described by Zapf et al.(20). rhIGF-I was used as a tracer and standard, and monoclonal antiserum to human IGF-I for IGF-I RIA as previously described(20).

Cell culture experiments. At the beginning, Wilms tumor cells were placed into 35-mm plastic dishes (≈106 cells/dish) in duplicate in culture growth medium containing 10% FCS and incubated at 37°C, 5% CO2. After overnight incubation the cells were rinsed once with serum-free DMEM and Hams' F-12 (1:1) mix medium and fed with 2 mL/dish fresh medium in the absence of FCS and in the absence of additions such as insulin, hydrocortisone, sodium selenite, and transferrin. After 24 h, the medium was replaced with fresh serum-free medium in the presence or absence of the indicated concentrations of suramin, IGF-I, IGF-II, or IGFBP-3 diluted in the appropriate medium, and the experiment was started. Thereafter, the medium was not changed throughout an experiment. In parallel, dishes of cells were fed only with serum-free growth medium and used as a control. After 48 and 96 h of incubation, the culture medium was removed, and the cells were treated with trypsin and EDTA (0.05 and 0.02%) and counted with an electronic microcellcounter (Digitana AG, Switzerland).

In two experiments, Wilms tumor cells (≈106) were placed in 24 multiwells and cultured as described above. Fresh serum-free medium containingα-IR3 (10-7 mol/L) was added 48 h after seeding and replaced every day. The cells fed only with serum-free medium were used as a control. After 24, 48, and 72 h from the addition of α-IR3, the cells were counted as described above.

Ligand binding assays for IGF cell surface receptors. Wilms tumor cells were cultured in 24 multiwells as described above. At confluent density, the cells were fed with serum-free medium without any addition in 37°C. After 1 h the cell monolayers were placed on ice and washed three times with binding medium-199, containing Earle's salts, 25 mmol/L HEPES, sodium bicarbonate, and 0.1% BSA (Sigma Chemical Co. St. Louis, MO), and then incubated in a total volume of 0.5 mL of binding medium-199 for 18 h at 0°C with 125I-labeled rhIGF-I or 125I-rhIGF-II (≈10 000 cpm), with various concentrations of either rhIGF-I, rhIGF-II, porcine insulin, α-IR3, suramin, or control IgG (MOPC-21). At the end of the incubation, cells were washed three times with medium-199 containing 0.1% BSA at 4°C and then lysed in the PBS buffer containing 1% Triton X-100 and 10% glycerol. Cell-associated radioactivity was counted in a gamma counter. Nonspecific binding was subtracted from total binding to obtain specific binding.

The Mann-Whitney test was used for statistical analysis.

RESULTS

The first question to solve in this study was whether Wilms tumor cells produce IGF-II or not. These and all the consecutive experiments were carried out in serum-free cell culture medium to exclude any effect and interaction of possible exogenous IGF. We found that Wilms tumor cell lines produced more than five times more IGF-II (41.0 ± 11.3 ng/106 cells) in 24 h than did normal kidney cells (7.4 ± 1.8 ng/106 cells). In contrast, IGF-I could hardly be detected in the culture medium, even after incubation of the cell cultures for 72 h (results not shown). Therefore, similar to our in vivo model(17), our cell cultures from Wilms tumor tissue produce and secrete IGF-II to a higher extent than do normal kidney cells.

The next series of experiments show the existence of IGF binding sites on the cell surface of Wilms tumor cells. Figures 1 and 2 prove that both forms of IGF receptors reside on the tumor cell surface. IGF-II binding is shown in Figure 1. Competitive inhibition of radiolabeled 125I-IGF-II to a significant extent was only possible-besides IGF-II itself-with suramin, but not with insulin, IGF-I, orα-IR3 antibody. Suramin has been previously shown to interact with binding of IGF-I to its receptor. Here we show that this substance also interacts with binding of IGF-II to the type II IGF receptor. Clear differences occurred, when IGF-I binding was measured as shown inFigure 2. IGF-I itself, α-IR3-antibody, IGF-II, and insulin as well as suramin competed for binding to the IGF-I receptor in a decreasing order according to their affinity to the receptor. Both suramin andα-IR3 antibody are known to competitively bind to the type I IGF receptor without a stimulatory effect. On the other hand IGF-I, IGF-II, and with very low affinity insulin exert their effects through the type I IGF receptor.

Figure 1
figure 1

Specific binding of 125I-IGF-II to Wilms tumor cell cultures. Confluent cell monolayers were incubated with125 I-IGF-II in the presence of the indicated concentrations of unlabeled IGF-II (•), IGF-I (○), insulin (▵), suramin (▾), orα-IR3 (▿), as described in “Methods.” Results are shown as the mean ± SEM of three experiments carried out in duplicates, except for a single experiment in duplicate with α-IR3.

Full size image
Figure 2
figure 2

Specific binding of 125I-IGF-I to Wilms tumor cell cultures. Confluent cell monolayers were incubated with 125I-IGF-I in the presence of the indicated concentrations of unlabeled IGF-I (•), IGF-II (○), insulin (▵), suramin (▴), or α-IR3 (▿), as described in “Methods.” Results show the mean ± SEM of three experiments carried out in duplicate, except for a single experiment in duplicate with α-IR3.

Full size image

These experiments were presupposition for determining whether the high expression of IGF-II and its release into the cell medium could be responsible for accelerated growth of Wilms tumor cells.

Figure 3 demonstrates the effect of suramin on Wilms tumor cell growth in serum-free medium. Increasing concentrations of suramin ranging from 1.2 × 10-5 to 2 × 10-4 mol/L (15.6 to 250 μg/mL) added to the Wilms tumor cell culture in the serum-free medium resulted in a dose- and time-dependent reduction of the cell number. The minimal dose of suramin in these assays that caused an inhibition of cell culture growth was 2.5 × 10-5 mol/L (31.3 μg/mL). Higher doses than 10-4 mol/L of suramin seemed to have a cytotoxic effect indicated by a decrease of cell number during incubation and a lack of growth restoration with time. However, the inhibitory effect of suramin in the nontoxic range was not definitive. It could be overcome by IGF-II. This is shown in Figure 4. To ascertain whether the suramin-induced inhibition of Wilms tumor cell growth was reversible, the cells were allowed to grow in the presence or absence of the maximal noncytotoxic concentration of suramin (10-4 mol/L). After 48 h, a clear and significant difference (p < 0.001) in cell number was shown(Fig. 4). Suramin completely blocked cell culture growth compared with the control as expected. At this time half of the cell culture samples were washed in serum-free culture medium and allowed to proliferate for a further 48 h in medium in the absence of suramin.Figure 4 shows that, after washing out the inhibitor suramin, cultures resume growing with the same growth velocity as control cells, demonstrating that the effect of suramin is fully reversible. The inhibitory effect of suramin could be competitively overcome by IGF-II. This is also shown in Figure 4. rhIGF-II at a concentration of 10-8 mol/L was added to Wilms tumor cell cultures growing in serum-free medium in the presence of 10-4 mol/L suramin. The effect of suramin on inhibition of cell growth was completely blunted, so that cell growth in the presence of suramin and IGF-II was not significantly different from control values (p = 0.721 at 48 h; p = 0.235 at 96 h). Thus, these experiments show that blocking IGF-II binding sites on the cell surface leads to inhibition of cell growth. However, it is not clear from the experiments whether endogenous production of IGF-II truly is responsible for growth stimulation. To further elucidate this, an excess of recombinant human IGFBP-3(2.5 μg/mL) to Wilms tumor cell cultures in serum-free medium was added. The results are shown in Figure 5. Cell growth in presence of exogenous IGFBP-3 was significantly lower (p < 0.0005) than in control after 48 h. However, between 48 and 96 h, the growth rate of Wilms tumor cells in the presence or absence of IGFBP-3 did not significantly differ (4.9 × 103/h versus 4.8 × 103/h, respectively). This suggested a degradation of IGFBP-3 during incubation at 37°C allowing unrestricted cell growth between 48 and 96 h. As a consequence, this experiment was repeated with the addition of a second bolus of IGFBP-3 (2.5 μg/mL) to the cell culture medium at 48 h, resulting in ongoing inhibition of cell growth (Fig. 5). The difference of cell number between these two experiments at 96 h was significant (p < 0.01). In a series of similar experiments, IGFBP-4 was added instead of IGFBP-3 with the same results (data not shown). Furthermore, it was possible to overplay the inhibitory effect of IGFBP-3 by adding an equal amount (on a molar basis) of rhIGF-I to the Wilms tumor cell cultures. Under these conditions growth of cell cultures in presence of rhIGF-I and IGFBP-3 in serum-free medium was not significantly different from control Wilms tumor cell cultures at 48 and 96 h. These data support the hypothesis that IGF-II is secreted by Wilms tumor cells and stimulates growth of the cell culture in an auto- and/or paracrine manner. The experiments, however, do not clarify the question of whether IGF-II acts through type I or through type II IGF receptors, because added IGF-I could compete with endogenous IGF-II on both binding to IGFBP-3 or type I receptor. Competition on IGFBP-3 binding would increase free endogenous IGF-II in the medium, which then could act either through type I or type II IGF-receptor.

Figure 3
figure 3

Time and dose dependence of suramin inhibition of growth of Wilms tumor cells in serum-free growth medium. Wilms tumor cells were harvested with PBS (pH 7.2) without Ca2+ or Mg2+, containing 0.05% trypsin and 0.02% EDTA, and ≈106 cells were plated in duplicate in 35-mm plastic dishes in serum-free medium. Suramin was added at different final concentrations (0-2 × 10-4 mol/L). Cells were incubated at 37°C, 5% CO2. After 48 h (○) and 96 h (•), cell number was counted as described in methods [mean ± SEM,n = 16 (8 experiments carried in duplicate)].

Full size image
Figure 4
figure 4

Reversibility of suramin-induced inhibition of Wilms tumor cell growth in serum-free growth medium. After attachment of the cells, duplicate 35-mm plastic dishes were plated with ≈106 cells in the presence or absence of suramin for 48 h. Cells of half the dishes were then washed and allowed to grow for a further 48 h in the absence of suramin. Control (○), suramin (10-4 mol/L) in the presence of rhIGF-II(10-8 mol/L, ▴), suramin (10-4 mol/L), washed out at 48 h(▵), or without washing (•). The results show the mean ± SEM of four experiments carried in duplicate.

Full size image
Figure 5
figure 5

Inhibition of growth of Wilms tumor cells in serum-free medium by IGFBP-3. Control cells (○); IGFBP-3 (2.5 μg/mL) added at time 0 (▴, n = 16); IGFBP-3 (2.5 μg/mL) added at 0 and 48 h of incubation (•, n = 4); IGFBP-3 (2.5 μg/mL) in the presence of an excess (10-7 mol/L) of exogenous rhIGF-I (▵, n = 8). The results show the mean ± SEM of two to eight experiments carried out in duplicate.

Full size image

The final set of experiments helped to distinguish through which of the two IGF binding sites endogenous IGF-II might act. Incubation of cell cultures in the presence of αIR-3, an antibody that specifically blocks the type I IGF-receptor, completely blocked cell growth (Fig. 6). This shows that growth-stimulating activity of endogenously produced IGF-II is mediated by the type I IGF receptor.

Figure 6
figure 6

Inhibition of growth of Wilms tumor cells in serum-free medium by the type I IGF receptor-blocking antibody α-IR3; control Wilms tumor cell cultures (○) and cell cultures in the presence of 10 μg/mL of α-IR3 (•). The results represent the mean ± SEM of two experiments carried out in duplicate (n = 4).

Full size image

DISCUSSION

There is increasing evidence, especially from a molecular genetic basis, that IGF-II is an important embryonic growth factor. In the human, changes on chromosome 11p15.5, the location of the IGF-II gene, clearly lead to growth abnormalities in the newborn. Maternal imprinting of the IGF-II gene has been described in the human and in the mouse(21). As a consequence uniparental paternal disomy of chromosome 11p results in increased expression of the IGF-II gene, a situation that has been described in the Beckwith-Wiedemann syndrome(22). This syndrome is characterized by neonatal gigantism, exomphalos, macroglossia, and an increased risk to develop embryonic tumors, especially Wilms tumor, during the first years of life(23). However, uniparental paternal disomy is not the only etiopathogenetic factor. Duplication of the distal arm of 11p has also been described in several patients(24, 25). On the other hand, knockout experiments of the IGF-II gene in mice led to neonatal dwarfism, supporting further the important role of IGF-II in embryonic growth regulation(26).

Wilms tumor as one of the most frequent tumors in childhood has been studied extensively during the last years on a molecular genetic level. Several reports described the importance of the same region on chromosome 11p13-15.5 in the pathogenesis of the tumor, e.g. the loss of maternal alleles or structural changes on 11p15.5(28–32), as well as deletions or point mutations of the Wilms tumor suppressor gene-1(WT-1) on chromosome 11p13. This leads to a highly elevated expression of the IGF-II gene measured by the mRNA concentration in tumor tissue(14–16), which could be reversed by the introduction of a normal chromosome 11 in tumor cell lines(33).

Not many studies describe the situation beyond the DNA or RNA level in the Wilms tumor cell. We have shown recently that the tumor not only contained high concentrations of IGF-II mRNA, but in consequence produced and released the peptide hormone IGF-II in an in vivo model(17). Wilms tumor tissue transplanted into nude mice led to elevated circulating IGF-II which exerted systemic effects on glucose homeostasis and on growth of some inner organs, especially the kidneys of mice(17). In the present study we could verify the in vivo experiments by increased production of IGF-II in monolayer cell cultures of Wilms tumors. We could establish several lines of Wilms tumor tissue which had been primarily transplanted into nude mice, but were also able to establish direct cell cultures of Wilms tumor tissue showing quantitatively and qualitatively the same results. In this study, for uniformity, we used the same tumor line for all experiments derived from Wilms tumor tissue previously transplanted into nude mice.

Our results clearly demonstrate that Wilms tumor cells produce significantly more IGF-II and that they grow faster than normal kidney cells under the same conditions. This suggests that IGF-II could stimulate cell culture growth in an auto- or paracrine fashion. Our data strongly support this idea. In addition we demonstrated that the site of action of the endogenous IGF-II is most probably the type I IGF receptor. First we showed that IGF-II binds to two IGF binding sites. This has been shown to be true in many different cell types such as muscle cells, fat cells, bone cells, fibroblasts, or chondrocytes(34, 35). The two binding sites are structurally and functionally completely different. The type I IGF-receptor is a dimeric molecule containing α- and β-subunits and is homologous to the insulin receptor(36). It binds with decreasing affinity IGF-I, IGF-II, and insulin and is competitively blocked by the antibody αIR-3(37). These well established facts are also true for Wilms tumor cells as shown inFigure 1. The type I IGF receptor is generally believed to be responsible for the transmission of the mainly growth-promoting activities of IGF-I. The type II IGF receptor is a monomer with a molecular mass of 250 kD homologous to the mannose 6-phosphate receptor. It binds only IGF-II, which is verified in Wilms tumor cells as shown inFigure 2. No clear function of this receptor could be demonstrated up to now, especially not with respect to mediation of growth-promoting effects. Our data do not change these facts. Our binding data suggest that the stimulation of cell growth by IGF-II is mediated by the type I IGF receptor. Blocking of type I IGF receptor by the antibody αIR-3 has been reported to inhibit Wilms tumor cell growth(38), which is verified in our study. In addition toαIR-3 we showed that growth of Wilms tumor cells is also stopped by the polyanion suramin(39) in a rather unspecific but reversible manner by interfering with IGF-II binding to both types I and II IGF receptors. This specifies earlier findings, that suramin inhibits binding of IGF and of other growth factors to various cell types(40, 41). We also show that absorption endogenously by Wilms tumor cell-produced IGF-II by IGF binding proteins does no longer allow cell culture growth. Our results add to the increasing evidence that in some embryonic tumors such as Wilms tumor and rhabdomyosarcoma(42) elevated IGF-II production and subsequent autocrine action is the limiting factor for tumor growth.

In summary this study shows the whole loop of elevated synthesis and secretion of IGF-II, its receptor binding, and its autocrine growth-stimulation effects in cultures of Wilms tumor cells. We conclude that IGF-II plays a crucial role in the regulation of growth of this embryonic tumor, supporting its important role as an embryonic growth factor.