Ultrastructure, karyology and immunology of a cell line originated from a human transitional-cell carcinoma.

A cell line (J82) was derived from a poorly differentiated, invasive, transitional-cell carcinoma, Stage T3. The cells have been propagated in vitro for 5 years and showed 100% aneuploidy and a mixed epithelial-fibroblastic morphology. The majority of cells contained 2Y chromosomes and several distinctive markers. Peripheral-blood lymphocytes from the donor of the J82 cells were tested sequentially for cytotoxicity toward autologous and allogeneic tumour cells. Autologous cytotoxicity was detected against J82 cells in early in vitro passage. Allogeneic lymphocytes from some patients with transitional-cell carcinoma were also cytotoxic to J82 cells in primary culture. However, selective cytotoxicity by lymphoid cells from bladder-carcinoma patients was not detected against J82 cells in long-term tissue culture. ImagesFig. 1Fig. 2Fig. 3Fig. 4Fig. 5Fig. 6Fig. 8Fig. 9

HUMAN transitional-cell carcinoma (TCC) exhibits diverse patterns of differentiation and tissue invasion. In general, cell lines from bladder carcinomas established in vitro retain the morphological attributes of the original tumour (Rigby and Franks, 1970;Bubenik et al., 1973;O'Toole et al., 1976;Nayak et al., 1977). TCC cell lines have been used extensively in the search for host immunity to tumour. The majority of tests reported to date have described allogeneic lymphocyte-mediated cytotoxicity to TCC cells. The results indicate that not all cell lines derived from bladder carcinomas express the antigens detected by lymphoid-cell cytotoxicity (Nayak et al., 1977;O'Toole, 1977). Further, serum antibodies have been identified in patients with squamous carcinoma of the bladder which do not react with TCC (Sofen and O'Toole, 1978).
The reactivity of autologous lympho-cytes to TCC target cells has seldom been studied (Bubenik et al., 1973) owing primarily to the difficulty of successfully culturing TCC in vitro and, secondarily, to problems of access to patients for followup. We report here a sequential study of lymphocyte cytotoxicity toward autologous and allogeneic tumour cells. The patient was tested before radiotherapy and cystectomy and followed for 3 tumour-free years.

MATERIALS AND METHODS
Tissue for cell culture was obtained 28 November, 1972, by transurethral biopsy. The patient, a 58-year-old Swedish male, had intermittent haematuria and low back pain for about 1 year, but was previously untreated. Cystoscopy showed a widespread papillary tumour with solid areas, located at the base of the bladder. The patient was given preoperative radiotherapy in a dose of 4200 rad over 43 davs. After a 3-w%Neek interval, cystectomy was performed. No metastases to lymph nodes were found and no viable tumour was observed in the bladder. The patient remains clinically tumour-free 5 years later.

Tissue culture
Tumour tissue was processed as described previously (O'Toole et al., 1976). When cell outgrowth reached confluency, cells were passaged using a trypsin (0-05%)-EDTA (0-02%) solution. Following thorough washing, cells were seeded at a concentration of 5x 105 cells per 25-cm2 flask. After 5 in vitro passages, the concentration of foetal calf serum in the culture medium was reduced to 10%. J82 cells were cryopreserved at -1 10°C at various passages, as described by O'Toole et al. (1976).
Cultures were routinely checked for mycoplasma contamination by a modification of the method described by Russell et al. (1975) and O'Toole et al. (1976). No contamination has been detected to date. HL-A typing was performed on J82 cells and autologous leucocytes by Dr Erna M6ller, Tissue Typing Laboratory, Stockholm, Sweden.

Karyotype analysis
Chromosome preparations were made at the 5th, 15th, 35th, and 48th passages. Four to 5 days after plating, cells were treated with 0 04 ,tg/ml Colcemid for 18 h and then exposed to a hypotonic solution (1:3 v/v of serum-free medium and deionized water) for 25 min at room temperature. Fluorescent banding analysis used in the quinacrine-mustard banding technique of Uchida and Lin (1974). Chromosome numbers were estimated in Giesma-stained preparations.

Electron microscopy
Cells were collected from monolayers by treatment with trypsin-EDTA solution or by scraping, and washed twice. A pellet of -107 cells was fixed in cacodylate-buffered 2% glutaraldehyde, osmicated in cacodylatebuffered 1% osmimum tetroxide, dehydrated in a water-ethanol series and then embedded in Spurr's epoxy resin. Thin sections were stained with uranyl acetate and lead citrate.

Cytotoxicity assays
Effector cells.-Defibrinated blood was mixed in 3: 1 v/v with a 3 % solution of gelatin and incubated for 1 h at 37°C in air with 5% CO2 to sediment erythrocytes. The leucocyte-plasma fraction was then added to a nylonfibre column which was incubated for 30 min at 37°C in air with 5% C02. Non-adherent cells were collected and treated with Trisbuffered NH4Cl solution at 4°C for a maximum of 10 min to lyse residual erythrocytes. The lymphoid cells were then washed x 3 and incubated at a concentration of 2 x 106/ml in Tissue Culture Medium 199 with 10% foetal calf serum for 18 h before use in cytotoxicity assays.
Target cells.-Lymphocytes from the donor of J82 cells were tested for cytotoxicity against autologous cultured cells, and the following allogeneic cell lines: T24 (Bubenik et al., 1973) and RT4 (Rigby and Franks, 1970), derived from transitional cell carcinomas. HCV-29 established by Dr J. Fogh, Sloan Kettering Institute, from urothelium; and MEL-1 derived from a metastasis of cutaneous melanoma (Unsgaard and O'Toole, 1975).
Cultures were maintained in monolayers in Medium 199 containing 10% foetal calfserum. The tissue-culture passage number (TC) when known, is given in the results.
Lymphocyte cytotoxicity was assessed by visual counting using a microplate method (O'Toole et al., 1973) or by release of 51Cr from labelled target cells (O'Toole 1977). Lymphocytes from the donor of J82 cells were compared with those from healthy donors or patients with other neoplasms or infections. Control lymphocyte donors were selected on the basis of availability only, and not on previous performance in cytotoxicity assays. The significance of differences in cytotoxicity between the patient and control lymphocyte donors was estimated by Student's t test.

RESULTS
The biopsy specimen from which the J82 cell line was derived showed transitional-cell carcinoma malignancy Grade 3 (Fig. 1). The tumour consisted of papillary exerescences with broad thickened epithelium showing moderate atypia and few mitotic figures. The tumour infiltrated into deep muscle (Stage T3).
Cell migration from the tumour explants occurred within 24 h of culture. Confluent monolayers were formed by 4 weeks, when the cells were subpassaged. J82 cells were passaged at 2-4-week intervals during the first year in culture; the

Micromorphology
The cellular micromorphology of J82 cells in monolayer culture is similar to that of normal bladder epithelium. Abundant microvilli and absence of desmosomes were notable exceptions (Fig. 3). In this respect, J82 cells resembled the line TCC SUP also derived from poorly differentiated TCC (Nayak et al., 1977), but differed from cultured squamous carcinoma of the bladder (O'Toole et al., 1976).
The cytoplasm of J82 cells is characteristic of an undifferentiated epithelial cell (Fig. 4). The varying amounts of rough endoplasmic reticulum (RER) per cell is probably a result of the metabolic activity of individual cells. Several cells contained segments of RER arranged as "confronting cisternae" (Kumegawa et al., 1968) and annulate lamellae (Kumegawa et al., 1967). Cytoplasmic fractures (Seman and Dmochowski, 1975) were evident in many J82 cells. Their significance is unknown. However, they may be a transient phenomenon of agitated RER locked in place by fixation. The abundance of normal-appearing mitochondria (Fig. 4) in the majority of cells is evidence of a high level of energy production. Lipid accumulation in the form of lipid bodies was minimal.
Desmosomes were not found in J82 cells, but microfilaments were prominent components of the cytoplasm in many cells (Figs 5 and 6). They occurred as individual filaments and in bundles. The pleomorphic nuclei were similar to those of normal bladder epithelium and several other cell lines derived from bladder carcinomas (O'Toole et al., 1976;Nayak et al., 1977;Rigby and Franks, 1970).
Chromosome analysis Chromosome number. -Metaphases studied at the 5th, 15th, 35th and 48th in vitro passages were all aneuploid. Table I and Fig. 7 show the chromosome frequency distribution at TC35 and TC48. At TC35, there was a wide frequency distribution of chromosome numbers extending from triploid to hexaploid, the modal number being 75 (22%). However, by TC48, chromosome numbers were largely restricted to the triploid area, with 4 % of cells in the tetraploid and only one (2%) in the hexaploid range. The modal number had shifted from 75 to 72 (22%) with two other peaks at 71 (16%) and 75 (14%).
In 38% of the cells at TC48, an additional minute chromosome per cell was observed. About 90 % of cells in all in vitro passages examined contained 2Y chromosomes, although metaphases prepared from the patient's phytohaemagglutininstimulated blood lymphocytes showed the normal diploid male pattern. Table  II gives  lysis by quinacrine fluorescence banding was performed at TC48 ( Fig. 8 and 9). Chromosome Groups A and B had numerical changes. Chromosomes 1 and 4 appeared normal, and Chromosomes 2, 3 and 5 were generally trisomic. C and D chromosomes had extensive numerical changes. Chromosomes 6, 7, 9, 11 and 15 were usually trisomic. In some metaphases, Chromosomes 7 and 11 had two extra homologues. Characteristically, Chromosome 13 had 5 homologues.
Groups E, F and G also had abnormal chromosome numbers. Chromosomes 17 and 18 were trisomic, 19 and 20 sometimes had 4 homologues, and 22 often showed trisomy. Chromosomes 8, 10, 14 and 21 generally appeared normal. Chromosome 21 had a bright satellite. Most of the cells had two X chromosomes.
Marker chromosomes.-At least 5, and frequently 7 or more, abnormal chromosomes have been identified in these cells on the basis of their characteristic banding pattern. These markers were not the same as the HeLa cell markers described by Nelson-Rees et al. (1974). HLA.-The patients leucocytes typed for HLA A2, AW32, B5, B12. The cell line J82 at in vitro Passage 12 typed for HLA AW32 and B5 only. Lymphocyte cytotoxicity Cytotoxicity by autologous peripheralblood lymphoid cells from the donor of J82 cells was tested on 8 occasions during a 3year period. Table III shows the results of sequential tests using a microplate assay and visual counting. At each test the number of residual target cells remaining after incubation with the patient's lymphoid cells was compared to that after incuba- tioIn with control doInors' lymnphocytes, and also wvith the number of surviving target cells incubated without effector cells (medium control). The same pattern of reactivity by J82 lymphoid cells was found using either control comparison. Before therapy the patient's lymphoid cells showed a preferential reaction against the allogeneic TCC cell line T24 as compared to HCV-29 cells. During the course of radiotherapy, after 1950 rad the patient reacted no differently from a control donor against two allogeneic cell lines derived from TCC and melanoma. Two weeks after 4200 rad the patient's lymphoid cells reacted significantly against HCV-29 cells. Three weeks after cystectomy, the patient showed no significant reaction against allogeneic cells. By 4 and 6 months after treatment, however, cytotoxicity was mainly e(1 numerous tightly packed bun(dles of filaments.

200.
restricted to atutologouis cells (J82) anid the allogenic TCC target T24. One and a half years after surgery, no significant differences in cytotoxic activity between the patient and a control donor was observed on either the autologous, or 4 different allogenic cell lines. Three years after surgery, the patient's lymphocytes were tested for cytotoxicity in a 51Cr-release assay (Table IV). At this time, no significant cytotoxicity was observed against autologous or allogeneic cells as compared with a healthy person's lymphocytes or those from 2 patients cured of melanoma. DISCUSSION (lytogenetic analysis of fresh bladdercarcinoma specimens have usually indicated a bimodal chromosome pattern (Atkin, 7 1 72 C. 01TOOLE, Z. H. PRICE, Y. OHNUKI AND 13. UNSGAARD ing patterns after short-term and longterm in vitro culture. Although this tumoutwas not karyotyped before culture, these results are consistent with invasive TCC.
In vie,", of the persistent problem of cross-contamination of cell cultures bv, other rapidly growing lines (Nelson-Rees et al., 1974) it is necessary to establish the identity of each new cell line. The presence of two Y chromosomes in the majority of J82 cells provides convincing evidence that these cells are of male origin. Duplicated Y bodies have been reported earlier in bladder carcinomas with hioh chromosome numbers (Atkin, 1973 (Bubenik et al., 1973) and RT4 (Rigby and Franks, 1970