The effects of beta-human chorionic gonadotrophin on the in vitro growth of bladder cancer cell lines.

The effects of human chorionic gonadotrophin (hCG) and its subunits on in vitro bladder cancer cell growth have been assessed using the a tetrazolium salt reduction assay (MTT). Intact hCG, alpha-hCG and beta-core hCG all had no effect on cell growth, while beta-hCG increased MTT reduction in all four bladder cancer lines tested. The magnitude of beta-hCG stimulation was maximal in the T24 line, which does not itself produce beta-hCG and appeared to be correspondingly lower in beta-hCG-secreting lines. The addition of antibodies to beta-hCG inhibited MTT reduction among high secretors but failed to inhibit MTT reduction in non-beta-hCG producers. These results are consistent with the poor prognosis associated with beta-hCG expression by bladder tumours in vivo and suggest an autocrine/paracrine stimulation of tumour growth by endogenously produced beta-hCG.

Human chorionic gonadotrophin (hCG) is a member of the family of glycoprotein hormones, including follicle-stimulation hormone (FSH), luteinising hormone (LH) and thyroidstimulating hormone (TSH), all of which are heterodimeric and share a common a-subunit. Each hormone has a unique P-subunit, although the f-chain of hCG exhibits 81% homology with that of LH and may have arisen from the LH gene following duplication and readthrough in the 3' direction (Fiddes and Talmadge, 1984). In primates and equids, CG originates from the conceptus and rescues the corpus luteum by binding to an LH receptor (Yoshimi et al., 1969;Braunstein et al., 1976;Bolton et al., 1980). The lone fsubunit cannot bind to the LH receptor, the intact heterodimer being required for both binding and activation (Pierce and Parsons, 1981).
The ectopic production of fl-hCG by a proportion of bladder cancers has been reported by several authors (Rodenburg et al., 1985;Dexeus et al., 1986, Iles et al., 1987 and cannot be accounted for simply by gene duplication or rearrangement. Consequently, it is likely that the control mechanisms governing fl-hCG expression are abnormal in secreting tumours (Iles et al., 1988). As a clinical marker fl-hCG may have some value for monitoring in therapy (Marcillac et al., 1993); high levels have been associated with both radio-resistance and metastases (Martin et al., 1989), but current opinion holds that the secretion is an epiphenomenon of little clinical significance (Jacobsen et al., 1990;Smith et al., 1994). However, a recent study carried out in this unit showed that 50% of T2/T3 patients had elevated urinary fl-hCG (>3.74 IU mmol-' creatinine), and 90% of these went on to develop metastases, while only 3% of the non-expressing group developed metastatic disease. In addition, survival curves for these patients, when divided into P-hCG expressors and nonexpressors, show approximately 10% and 60% survival respectively at 17 months, suggesting a valuable prognostic function (Iles et al., 1996).
We have also shown that primary cultures of normal urothelium can secrete fl-hCG (Iles et al., 1990 , 1990). These have been distributed worldwide and have been used as the First International Reference Preparation of hCG and free subunits for immunoassay and the Third International standard for intact hCG bioassay. Here we used the preparation labelled 123 of the CR series, which was first released in 1977. The purity of these preparations is well recognised. However, several degradation products of hCG metabolism have recently been recognised, the ultimate step being the formation of urinary f-core. Intermediate steps include the nicking of the f-subunit at residues 45-50. The extent of such nicking varies with each preparation and CR123 has been estimated to contain 10-16% nicked fl-chains (Birken et al., 1991). The in-house preparation of highly purified ficore used in these experiments has been previously described by Lee et al. (1991). Sheep antisera 750 and 752 were prepared in-house, following immunisation with free fl-hCG isolated from pregnancy urine. These antisera were found to react with only the free f-subunit of hCG but not intact hCG, LH, The 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT) assay (Mosman, 1983) was used to estimate cell number following exposure to intact hCG, free a-and fsubunits, urinary metabolite fl-core and to anti-,B-hCG antisera.
Microtitre plates (Coming 96-flat-bottomed wells) were seeded with 200 pl of 0.1 x 106 ml-' cell suspension (20 000 cells per well) in RPMI + 10% fetal calf serum (FCS) for 24 h before the replacement of media by fresh culture media containing test materials (0-500 ng ml-' ligands or 1:1000 dilution of antisera). Plates were then incubated for a further 36 h, followed by 1 h incubation in fresh medium, then addition of 20 ul millipore filtered MTT (5 mg ml-' in PBS).
After 4 h incubation all fluid was removed and 100 pl of acidified isopropanol (containing 0.04 M HCI) was added to each well and maintained at room temperature for 15 min to allow formazan crystals to dissolve. Absorbance measurements were carried out at 570 nm against a 630 nm blank; at least three replicates of each treatment were included. Results were expressed as optical density as a per cent of untreated controls.

Results
Stimulation by hCG and fragments Intact hCG, a-hCG and fl-core had no effect on the growth of any of the cell lines, while 25 ng ml-' ,B-hCG produced a 152% increase in MTT reduction by T24 cells; (Figure 1, Table I). 5637, SCaBER and RT112 were stimulated to a lesser extent increasing MTT reduction by 132, 112 and 116% of controls respectively (Figure 2, Table II).
Effect of anti-#-subunit antisera Concurrent addition of specific antisera with the fl-hCGcontaining media eliminated the stimulatory effect of the fi-hCG on the responding cell lines. This was best illustrated when antisera was concurrently added to escalating doses (0-50 ng ml-') of free fl-hCG in cultures of the T24 cell line (which does not secrete its own fl-hCG). The dose-response curve to the P-hCG was obliterated with co-addition of the anti-f-subunit antisera and MTT reduction matched that of controls ( Figure 3). This was not seen when non-immune control sheep serum (NSS) was used (data not shown).    Antisera to free f-subunit at a 1:1000 dilution added to cell lines SCaBER and 5637 (lines that synthesised and secreted their own fl-hCG) lowered MTT reduction to 60% and 70% of control levels respectively. However, the control NSS at the same dilution did not alter MTT reduction relative to controls. Furthermore, the T24 cell line (which does not secrete f-hCG) was unaffected by the addition of the fl-subunit antisera (Figure 4).

Discussion
The observation that 70% of established bladder cancer cell lines and normal urothelial cells established as finite lifespan cultures secrete variable quantities of,-hCG prompted our investigation into a putative biological activity of free fl-hCG (Iles, 1991). The MTT assay clearly demonstrates an increase in cell numbers following fl-hCG treatment but no such effect with intact hCG or oc-hCG (Figure 1). It is interesting to note that f-core had no effect either. This metabolite, which has a shortened carboxy terminus, fewer carbohydrate residues and several nicks in the amino acid chain (Birken et al., 1988) is found excreted in the urine. The absence of an effect with this molecule provides further evidence for the specificity of the growth effect observed with the free fl-subunit.
The effect of exogenous f-hCG on responding cell lines was inhibited in a dose-dependent manner by antisera to fi-hCG ( Figure 3). This suggests a highly specific type of interaction, possibly mediated by a receptor, as normal sheep serum produced no diminution of growth response in any of the lines. The growth response was highest in the T24 cell line, which does not itself secrete fl-hCG, while higher secretors exhibited less growth stimulation. This implies that fl-hCG producers may be self-stimulating populations in vitro (or indeed, in vivo), high level producers being incapable of further stimulation by exogenous fl-hCG. Alternatively, there might be more than one subpopulation of urothelial cells, with secretors and responders in varying proportions that determine overall production or response rates. The recently elucidated three-dimensional structure of f-hCG (Lapthorn et al., 1994) includes at its centre a distinctive arrangement of protein chain folds that is stabilised by six disulphide bonds and known as the cysteine knot motif. This motif is found in at least three growth factors: nerve cell growth factor (NGF), transforming growth factor (TGF-fl2) and platelet-derived growth factor (PDGF-BB). These molecules are able to bind to their receptors as homodimers, a possibility that could exist with fl-hCG and should be investigated further. When no exogenous /3-hCG was added, antisera against /3-hCG considerably inhibited growth in the high-producing lines such as SCaBER, with little or no growth inhibition perceived in the non-secreting T24 line ( Figure 4). This provides further evidence for the validity and specificity of the observed effects of the subunit itself and constitutes important new evidence for an autocrine/paracrine effect of f-hCG in urothelial cell carcinomas, while suggesting a possible mechanism for some of the poor prognostic associations with fl-hCG that have been reported.
In conclusion, fl-hCG (but not intact hCG, oa-hCG or 1lcore) is able to specifically increase cell growth in bladder epithelial bladder cancer lines, this effect is mediated by a specific interaction that is obliterated by anti-f,-hCG serum.
These findings could be explained if it is postulated that the free f-subunit acts as a growth factor.