Integrated pest management.

Integrated pest management, in which the conventional pesticides are augmented by one or more nonchemical control practices, has been receiving renewed interest. What is new in this revitalization of an old technique is the careful and more knowledgeable application of a variety of control techniques.

4/15 (26%) grade 1, 5/18 (28%) grade 2 and 3/5 (60%) grade 3 tumours. There was no clear relationship to tumour stage, lymphocytic infiltration or stromal content of the tumours. Clinical review one year after the 2 year period of tumour collection showed that 6/9 (66%) of patients with tumours with reduced levels of transcript had died or had disease which was not controllable by local resection and 3/9 (33%) had developed tumour re-occurrences. In comparison, in the group with normal levels of expression of TGFPi1, 3/18 (17%) had disease which was not controllable by local means, 9/18 (50%) had tumour re-occurrence and 6/18 (33%) had no evidence of disease. The association of reduced expression of TGF%1 and advanced disease was statistically significant P<0.02 (Fisher's test). Although the sample size is small, we suggest that the loss of expression of TGFP1 may be a potential marker of progressive disease or prognosis in transitional cell carcinoma and warrants further study. The TGFPI group of peptide regulatory factors is a large expanding family of multi-functional genes displaying marked homology and evolutionary conservation (Roberts & Sporn, 1990). TGFP is secreted in a latent form which is unable to bind to its receptor (Wakefield et al., 1987). Activation of this latent TGFP may be achieved in several ways including transient acidification, alkalinisation or chaotropic agents (Kryceve-Martinerie et al., 1985).
The action of TGFP is mediated through binding to specific cell surface receptors. Almost all cells, regardless of origin, bind TGFP (Roberts & Sporn, 1990). Three distinct classes of receptor with various affinities for TGFP1 and TGFPI2 have been described (Cheifetz et al., 1987) and it has been suggested that the cell specific effects of the individual forms of TGFP may be regulated by differences in the levels of receptors of different affinities present on those cells (Cheifetz et al., 1990).
The pleiotropic effects of the TGFPi family have been extensively documented (for recent reviews see Moses et al., 1990; Roberts & Sporn, 1990). The role of TGFPI in cell transformation is unclear. Most normal epithelial cells in tissue culture are growth inhibited by TGFP (Moses et al., 1985;Masui et al., 1986;Jetten et al., 1986;Kurokowa et al., 1987). In contrast many carcinoma cells show reduced inhibition by TGFPI (Wakefield et al., 1987;Lechner et al., 1983;McMahon et al., 1986) and many transformed cell lines secrete increased amounts of TGFPI (Derynck et al., 1987;Jakowlew et al., 1988;Niitsu et al., 1988) which is reflected in an increase in the steady-state levels of TGFPI mRNA in these cell lines and in tumours. Such increased TGFPI production could contribute to tumour development and progression in multiple ways via paracrine effects on neovascularisa-tion, extracellular matrix formation, chemotaxis and immunosuppression . However, changes in TGFP production and responsiveness are not demonstrated in all transformed cells. Some remain growth inhibited by TGF,B and not all transformed cells secrete increased amounts of TGFPi (Derynck et al., 1987;Wakefield et al., 1987).
There have been few studies of TGFP production by human tumours. Raised levels of TGFP mRNA were reported in breast and renal tumours (Coombes et al., 1990;Gomella et al., 1989). Similarly, TGFP1 RNA was detected in all glial tumour cells in a spectrum of cerebral malignancies (Mapstone et al., 1991). TGFP secretion and growth response of urothelial cells has not been studied in detail. The only report on bladder epithelial cells to date showed that foetal urothelial cells but not transformed urothelial cells responded to exogenous TGFP1 by a decrease in plasminogen activator activity secondary to increased transcription of PAI-I activity (Hiti et al., 1990).
Transitional cell carcinoma is the fourth most common cancer in males in the United Kingdom and the incidence is rising in both men and women (31% between 1971(31% between and 1984(31% between , OPCS, 1971(31% between -1984. Studies of the natural history of transitional cell carcinoma have identified an aggressive subset of tumours (Pryor, 1973). Identification of the molecular events involved in the genesis of transitional cell carcinoma may offer potential markers of disease progression and prognosis. As part of a study aimed at identifying some of these lesions in transitional cell carcinoma, we have examined the structure and expression of the genes encoding TGFPi1 and TGFPi2 in human urothelial cancer cell lines and transitional cell carcinomas.
We show that in bladder tumours, marked reduction or loss of expression of the gene encoding TGFPI is associated with advanced disease. No TGFP2 transcript could be detected in these tumours. In urothelial cancer cell lines, variable levels of TGFPi1 and TGFP2 mRNA were expressed with no apparent relationship between the relative amounts of these transcripts. 'cold' cup biopsy forceps and was removed from the bladder as soon as possible, trimmed of debris and a representative sample excised (including the base and attached normal tissue) for histological assessment. Tumour size ranged from 60 mg to many grams but the majority (>80%) were small and were processed as a single sample. Where biopsies of normal urothelium or carcinoma in situ were taken, the epithelial layer was dissected free of submucosa and muscle.
In these cases, urothelium from at least four biopsies from the same patient was pooled and processed together. Tissues were placed immediately at -70°C. The tissues used are shown in Table I. Tumours were graded according to W.H.O. recommendations (1973) and staged using the TNM system (UICC, 1978).

Cells and cell culture
The cell lines used were EJ (Evans et al., 1977), VM-CUB-2 (Williams, 1980), SCaBER (O'Toole et al., 1976), SD (Paulie et al., 1983) and SW1710 (Kyriazis et al., 1984) Isolation of DNA and RNA DNA and RNA were isolated from the same tumour sample by the guanidine isothiocyanate method (Maniatis et al., 1982). DNA was extracted twice with phenol, twice with phenol:chloroform and once with chloroform, ethanol precipitated and dissolved in 1 x TE prior to quantitation and use. The RNA pellet was washed in 70% alcohol, air dried and dissolved in 0.3 M sodium acetate pH 6.0 prior to precipitation with two volumes of absolute ethanol and storage at -70'C.

Southern blotting
DNA samples were digested with EcoRI (Gibco BRL, Paisley, Scotland) according to the manufacturer's instructions and the fragments separated in 0.8% agarose gels. Gels were stained with ethidium bromide and photographed prior to capillary transfer (Southern, 1975) onto Hybond-N membranes (Amersham UK). Lambda DNA digested with Hind III was used as size markers and lymphocyte DNA from normal volunteers as a normal DNA control on each gel. Blots were baked at 80°C for two hours and pre-hybridised and hybridised following the manufacturer's instructions. Following washing to high stringency (0.1% SSPE and 0.1 % SDS), blots were exposed to Hyperfilm-MP (Amersham UK) at -70°C with intensifying screens.

Northern blotting
Total cellular RNA was electrophoresed in 1 % agarose/ formaldehyde gels (modified from Thomas, 1980) and transferred by capillary blotting to Hybond-N membranes. These were pre-hybridised and probed according to the manufacturer's instructions. Some gels were stained with ethidium bromide to compare loading and RNA inetegrity with the results obtained from control probes. Blots were hyrbidised sequentially to probes for TGFP1, TGFP2 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The 1.4 kb GAPDH transcript migrated suitably close to the 2.5 kb TGFI31 message to act as a good control for RNA loading and degradation. Non-specific binding of the GAPDH probe to the 28S ribosomal band was used as a control for TGFP2 (transcripts of 4.1 kb, 5.1 kb and 6.5 kb). In addition, selected blots were hybridised to c-erbB-2 (transcript size 4.5 kb, not illustrated) to confirm the presence of intact RNA in the size range of TGFP2. Levels of transcript in tumours were compared with levels in control issues and scored as normal (+), raised (+ +) or reduced or undetectable (0). Blots were stripped by washing for two hours in 5 mM Tris-HCl pH 8.0, 2 mM Na2 EDTA and 0.1 x Denhardt's solution at 650C.

Slot blotting
Slot blots were made on Hybond-N membranes using a Schleicher and Schuell vacuum slot blotting apparatus. Each slot received 2.5 jig denatured total RNA. RNA was denatured in 50% formamide, 5% formaldehyde, 1 x SSC at 680C for 15 min. Following application of the sample, the wells were flushed with 100 gl 20 x SSPE and the membranes baked at 800C. Hybridisation was as for Northern blots. Slot blots were assessed by comparing the ratio of signals obtained with GAPDH and the gene of interest for a reference slot containing control tissue RNA with those obtained for tumour tissues.

Probes
The probes used were the 2.1 kb EcoRI fragment of phTGFb-2 (TGFP1) cloned by Dr G. Bell and kindly supplied with his permission by Dr J. Scott, the 2.3 kb EcoRI fragment of pPC-21 (TGFi2, Madisen et al., 1988) supplied by Oncogen Science (Manhasset, NY., USA) and the 1.3 kb Pst-I fragment of pRGAPDH-13 (glyceraldehyde-3-phosphate dehydrogenase, Fort et al., 1985). Probes were labelled by random priming (Feinberg & Vogelstein, 1983) and used at 106 c.p.m. ml-' of hybridisation fluid.  (Figure la). SCaBER, a cell line derived from a squamous cell carcinoma of the bladder, showed the highest levels of expression and VM-CUB-2 and JO'N the lowest levels. After allowance was made for variations in loading and RNA degradation, 5637, SW1710, EJ and SD were judged to express similar levels of TGFP1I.

Results
Levels of TGFPI2 expression were more varied. The level of expression of the three expected transcripts of 4.1 kb, 5.1 kb and 6.5 kb differed within the same bladder tumour cell line (Figure lb). Expression of the 5.1 kb and 6.5 kb transcripts was greater than that of the 4.1 kb transcript. Total expression of TGFP2 RNA also varied markedly between cell lines.
All cell lines expressed two of the three expected transcripts of 6.5 kb and 5.1 kb respectively. VM-CUB-2 appeared to express higher levels of the 6.5 kb than the 5.1 kb transcript (Figure lb) but all other lines expressed more of the 5.1 kb than the 6.5 kb transcript. This can be seen clearly for SWI710 in Figure   Expression of TGF I3 and TGFl32 in primary transitional cell carcinoma Northern and slot blots of total RNA extracted from transitional cell carcinomas were hybridised sequentially to TGFP1, TGFP2 and GAPDH probes. RNA from 38 tumours, two cases of CIS, ten matched and three unmatched field biopsies, 11 normal urothelial biopsies, two biopsies of follicular cystitis and one biopsy from a bladder with cystitis following BCG treatment were analysed by Northern and slot blots. Since large samples of normal urothelium from individuals with no history of urological symptoms are not available, samples of macroscopically normal urothelium from several sources were assessed to provide a measure of 'normal' levels of expression. These included samples from bladder tumour patients with no tumour at check cystoscopy and from prostatectomy patients. Expression of TGFI in all but one of these samples (see below) was similar and was taken as the baseline level of expression. No aberrant transcripts of TGFP1 were detected and the levels of transcript in 25 tumours were similar to those in normal urothelial controls. In 12 tumours, reduced levels of transcript were observed and in three tumours raised levels of transcript were detected. Examples are shown in Figures 2 and 3. The characteristics of these tumours are outlined and compared with grade in Table II. Although the number of G3 tumours analysed is small (five), the marked difference in the incidence of tumours with reduced expression 3/5 (60%) suggests that there is an association between high tumour grade and reduced expression of TGFPI. No relationship between the reduced levels of transcript of TGFPI and tumour stage or the amount of stroma in the tumour biopsies was detected. Three tumour re-occurrences (we have used this term for subsequent tumours or 'recurrences' in the same patient, since the relationship between initial and subsequent tumours is unclear) were subsequently analysed. In two of these, the level of expression of TGFPI was the same as that detected in the initial sample. In the third, expression was reduced in the initial biopsy and normal in the second.
Of the ten matched field biopsies, four showed similar levels of expression of TGFP to the tumour from the same bladder. These were comparable to levels detected in normal urothelium. In three cases, transcript levels were reduced in the tumour and not the field biopsy. In one patient, levels were reduced in the field biopsy but not the tumour, in one patient levels were reduced in both field and tumour biopsies and in one patient transcript levels were raised in the tumour but normal in the field biopsy. Of the three unmatched field biopsies, one showed raised levels of transcript. In addition, two biopsies of follicular cystitis (both from patients in whom transitional cell carcinomas had been resected in the past) were assessed. One of these showed reduced levels of transcript. Only one of the samples of macroscopically 'normal' urothelium showed altered levels of transcript. This sample was obtained from a patient in whom no re-occurrences were detected at check cystoscopy. Two biopsies from this patient were assessed.   Table III. These exclude the two patients with CIS, one of whom was treated with BCG and one with cystectomy, and one patient with a transitional cell tumour who was too ill for any therapy (all of whom expressed normal levels of TGFI). Correlation between reduced expression of TGFJIl and disease progression was found P = 0.02 (Fisher's test).

Discussion
The members of the TGFPi family of peptide regulatory factors have been shown to play important roles in the control of growth and differentiation of normal cells. A number of observations suggest that differences in production of, or response to TGFPi may play a role in transformation.
Here we have shown that expression of TGF,B1 and TGFP2 vary considerably in both urothelial carcinoma cell lines and tumours. Nevertheless, results indicate a correlation between decreased expression in tumours and clinical behaviour. Northern analyses of total RNA from human bladder tumour cell lines showed the expected transcripts for TGFPI and TGFPi2. The three TGFP2 transcripts may result from differential splicing and/or polyadenylation events. The 4.1 kb and the 6.5 kb messages are considered the major transcripts of TGFP2 and have been described most commonly in other cell lines (Madisen et al., 1988;Derynck et al., 1988). However, we found that the 6.5 kb and 5.1 kb transcripts were the most abundant in urothelial carcinoma cell lines and the 4.1 kb transcript was only clearly demonstrated in one cell line. Since this did not reflect mRNA degradation, it is likely that the various transcripts identified in the urothelial tumour cell lines are differentially expressed messages.
TGF,1I mRNA was expressed at higher levels than TGFP2 and although those cell lines with the highest levels of TGFPI transcript tended to have lower levels of TGFP2 transcript, there was no clear inverse relationship. Tissue and species specific differential expression of TGFPI1 and TGFP2 has been reported (Seyedin et al., 1985;Cheifetz et al., 1987;Assoian et al., 1983;Derynck et al., 1988;Wrann et al., 1987;Ikeda et al., 1987). However, results obtained using cultured cells must be interpreted with caution. In this study, the cells were harvested at semi-confluence from media containing serum. Cell density in culture has been shown to affect response to exogenous growth factors including TGFP (Ke et al., 1990)   Increased levels of RNA were also found in one unmatched field control sample and 1 sample of macroscopically normal urothelium and in one case of follicular cystitis reduced expression was found. This supported the impression that a steady state of expression of TGFP1 occurs during relatively 'controlled' growth in urothelium.
The role of TGFPI1 in transformation remains unclear although the ability of transformed cells to respond to TGF,B1 is frequently altered (Lechner et al., 1983;McMahon et al., 1986;Shipley et al., 1986;Wakefield et al., 1987). A number of studies suggest that altered expression of TGFP1 may play a part in transformation (Derynck et al., 1987;Jokowlew et al., 1988;Niitsu et al., 1988). Increased levels of TGFPIB mRNA in tumours compared to adjacent tissues has been reported in a number of tumours (Derynck et al., 1987), including breast and renal cell carcinoma (Coombes et al., 1990;Gomella et al., 1989). It has been suggested that in-creased expression of TGFP1 by non-responsive tumour cells may stimulate tumour growth indirectly via paracrine effects and may also confer an additional advantage on the tumour by suppressing the hosts immunological surveillance. We detected raised levels of transcript in only three tumours and in these there was no obvious correlation with any clinical parameter.
Reduced levels of TGFPI1 transcript have been reported in some tumour cell lines (Jakowlew et al., 1988). Undetectable or reduced levels of TGFPi1 transcript were seen in 12 bladder tumours. The significance, if any, of the association between reduced levels of transcript of TGFPI1 and high tumour grade is not clear. A larger sample size is required to clarify this point. However, this was not as striking as the apparent association of reduced transcript levels with tumours which in the relatively short period of follow up became uncontrollable by local means (P<0.02).
A similar reduction in TGFP expression has been reported in a series of breast tumours analysed by immunohistochemistry, where expression of the TGFP1 gene product was detected in only 38% (31/82) of tumours and was unrelated to stage and grade (Mizukami et al., 1990). In these breast tumours, it was observed that tumours expressing TGFPI1 were associated with a better prognosis over 2 years. Thus, a reduction in TGFP expression may be common to both aggressive breast and bladder tumours. A number of other molecular changes have been described in both tumour types. These include amplification and overexpression of ERBB2 (Coombs et al., 1989Slamon et al., 1987Slamon et al., , 1989, amplification at 1 1q13 involving INT2, HST and BCL1 Adnane et al., 1989), and loss of heterozygosity of the RB gene Varley et al., 1989).
The association of reduced expression of TGFPi1 mRNA and poor prognosis is in keeping with the known growth inhibitory activity of TGFP1 on normal epithelial cells. Loss of expression of TGFP1 might result in reduction of extracellular matrix formation, increased pericellular proteolysis and removal of the negative effect on proliferation (Sporn & Roberts, 1990). Further studies are now required to investigate the relationship of TGFPi1 transcript levels to levels of the mature active gene product in bladder tumours.
Although only a relatively small number of tumours have been examined, there appears to be a relationship between reduced levels of expression of TGFPI1 transcript and disease progression. From the information obtained in this study, the presumed loss of the inhibitory activity of TGFP1 in transitional cell carcinoma may be a late event in bladder carcinogenesis. However, this does not preclude its utility as a clinical marker of progression or prognosis.