WO1999004036A1 - Cd44 based cancer detection - Google Patents

Abstract

The human CD44 gene is composed of at least 20 exons of which 10 are standard and 10 variable. Chaotic overexpression of variable exons has been observed in tumours compared to normal cells. This invention builds on this observation and provides a diagnostic method which comprises subjecting mRNA to reverse transcriptase PCR using a primer complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene, and observing whether an amplimer is formed. Preferred primers are complementary to sequences that bridge exons 5/7 or 5/9 or 5/11.

Description

CD44 BASED CANCER DETECTION

INTRODUCTION

The human CD44 gene is composed of at least 20 exons of which 10 are consistently expressed together to encode for a peptide backbone of about 37 kD which may subsequently be glycosylated before being expressed as the "standard isoform" (CD44s) on the cell surface (1 ,2). mRNA transcripts can contain the other variably expressed exons of the gene in various combinations with the CD44s exons in a tissue-specific manner and this is determined by alternative splicing mechanisms. This results in the production of many isoforms, presumably with different functions, from the same gene (3,4). Severe disturbances in the pattern of CD44 expression have been observed in a variety of human tumours by both protein and RNA analyses. Markedly increased overall levels of CD44 transcripts and proteins have been recognised in many tumours compared to normal cells from the same tissue (5,6,7). With RT-PCR/hybridisation analysis using probes to CD44s exons, distinctive large molecular weight amplicons are visualised as well as a marked overall increase in the quantity of CD44 mRNA, specifically in tumour cell-containing samples. In many such samples a very wide range of sizes of CD44 mRNA species are evident, indicating a severely deranged array of gene transcripts. These characteristic abnormalities may be an indirect result of transcriptional deregulation which leads to over-expression of the CD44 gene. This may in turn lead to subsequent leakage through the splicing mechanisms in tumour cells, resulting in accumulation of immature or defectively processed mRNA transcript species. Accordingly, the inventors have recently observed the retention of introns in mature mRNA transcripts (8,9), an event which could result in the production of abnormal or truncated protein products.

Inappropriate expression patterns of the variable exons have been linked both to tumour growth and to metastatic potential (10, 11). When CD44 expression is analysed in tissue samples by RT-PCR and Southern hybridisation with the use of single variant exon- specific probes it is found that numerous transcripts containing different combinations of variant exons are present in tumour samples, but not in the corresponding normal tissues. This finding supports the interpretation that some loss of the normal tight regulation of CD44 RNA splicing occurs in tumour cells. The possible clinical application of these observations has been shown by the detection of abnormal CD44 expression in fresh tissue specimens from tumours of many different histogenetic origins. Furthermore, the inventors have been able to achieve non-invasive detection of malignancy by detailed analysis of CD44 expression in exfoliated cells in body fluids and waste products (12).

These findings stimulated more detailed examination of the splicing pattern of CD44 transcripts in order to determine whether such a profile could be useful as a diagnostic indicator of carcinoma in a human tissue or fluid sample.

THE INVENTION

In one aspect, the invention provides a diagnostic method which comprises subjecting mRNA to reverse transcriptase using a first primer complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene, and observing whether an extension product is formed.

In another aspect the invention provides a diagnostic method which comprises subjecting mRNA to hybridisation with a probe complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene, and observing the formation of a hybrid of said probe and said mRNA.

The starting material is mRNA which may be obtained by extraction from cells of a human or non-human mammal, or from a cultured cell line, in a manner which is not material to the invention. From the mRNA, cDNA is synthesised with reverse transcriptase using a first primer. The cDNA may be subjected to amplification by PCR under conditions which may be conventional.

For the PCR amplification, two CD44 gene primers are used. That is to say, the two oligonucleotide primers are complementary to chosen regions of the CD44 gene.

A first primer is complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene. This first primer consists of two sequences, of which one is complementary to the 3'-end of a first exon of the CD44 gene, and the other is complementary to the 5'-end of a second exon of the CD44 gene. At least one of these exons is a variable exon; the other is either a variable exon or is a standard exon which is either number 5 or number 16 of the CD44 gene. For example, the first primer may have a 5'-end complementary to the 5'-end of a variable exon, and a 3'-end complementary to the 3'-end of exon 5. Alternatively, the first primer may have a 5'-end complementary to the 5'-end of exon 16 and a 3'-end complementary to the 3'-end of a variable exon. As shown in the experimental section below, useful results are obtained when the first primer is complementary to a sequence that bridges exons 5/7 or 5/9 or 5/ 1. A second primer for the PCR amplification may be complementary to a standard exon of the CD44 gene. Or the second primer may be complementary to a variable exon. Or the second primer may (similar to the first primer) be complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene. The 5'-/3'- orientation of each of the chosen primers is tailored to that of its chosen partner for a given reaction so as to be capable of generating a PCR reaction product.

In a preferred aspect, the method of the invention involves observing whether use of these two primers in a PCR reaction results in amplification of a nucleic acid segment present in the starting cDNA. Various ways of doing this are well known in the art. For example, one of the two primers may be labelled, either with a signal moiety or in such a way that a signal moiety can be attached after completion of the PCR reaction. A signal moiety may be for example a radioisotope or a fluorescent or chemiluminescent group or a component of an enzyme system that generates light or colour. In order to separate that part of the label which has taken part in a PCR reaction from that part which has not, the other primer may be immobilised or may carry a group by which it may be immobilised after completion of the PCR reaction.

Alternatively other means are possible of observing whether a nucleic acid is amplified by the PCR reaction. For example, the product of the PCR reaction may be subjected to size separation, e.g. by gel electrophoresis. A PCR amplification product is likely to have a characteristic size, and so to be easily observable.

As shown in the experimental section below, tumour cell lines and carcinomas are rather likely to give positive results, while normal cell lines and cells are rather likely to give negative results. The method is thus potentially diagnostic for tumour or cancerous cells. Improved diagnostic discrimination may be achieved by subjecting one aliquot or several aliquots of the starting mRNA to RT-PCR using (simultaneously or separately) several different pairs of primers according to the invention. In another aspect the invention provides a diagnostic kit comprising two CD44 gene primers including a first primer complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene. FIGURE LEGENDS Figure 1.

Schematic diagram of the exons of the CD44 gene showing the positions to which the primers used for RT-PCR analysis anneal. cDNA was amplified across the entire variant exons by PCR using primers P1 and P4, complementary to exons 3 and 17 respectively. Hybridisation probes for individual variant exons were generated by PCR using exon-specific primers to amplify sequences from a human CD44 genomic clone template (C231 1 ).

Figure 2.

Schematic diagram of the "Exon-Link" assay design.

PCR amplification is performed using an standard exon anchored primer (P1 ) and a primer specific for a particular standard-variant exon junction. The overlapping primer is designed to have 18 bases complementary to the variant exon and the six 3' bases of the primer complementary to standard exon 5.

EXPERIMENTAL There follows a description of an experimental programme to evaluate the order of exon assembly at the 5'-boundary between standard and variant portions of the molecule in gastrointestinal cell lines, and to compare such patterns and specimens from human colon carcinomas and matched normal tissues.

MATERIALS AND METHODS Cell lines and tissues.

Initially a number of cell lines were studied (HT-29 and TE-1 , 7, and 12 ) for the presence of specific CD44 exon junctions. HT-29 is a colonic carcinoma cell line. Cells were routinely cultured in RPMI 1640 medium (GIBCO, BRL, Gaithersburg, MD) supplemented with 10% foetal calf serum at 37°C in an atmosphere containing 5 % CO₂.

Fresh coiorectal carcinoma tissue samples and corresponding normal colonic mucosa from surgical resection specimens were collected within 20 minutes of removal and snap frozen and stored in liquid nitrogen until use. The samples were collected as consecutive cases with no previous knowledge of CD44 status. The presence of carcinoma cells in tissues taken from colon resections was routinely confirmed by cryostat sectioning before analysis.

RT-PCR and Southern hybridisation.

Total cellular RNA was extracted by the acid guanidium phenol-chloroform method and mRNA was purified using Oligotex dT (Qiagen ). cDNA was synthesised with reverse transcriptase (RT) followed by amplification by PCR as described previously (13). Five μg of total RNA or 100 ng of poly A+ selected RNA was used for the RT reaction with the cDNA Cycle Kit (Invitrogen ). The conditions of PCR were as follows; 94°C for 5 min and 85 °C for 1 min (at which time Taq polymerase was added) followed by 30 cycles of 94°C for 1 min, 55°C for 1 min and 72°C for 2 min. The sequences of primers used for amplification across the entire variant region were as follows:

P1 ; 5'-GACACATATTGCTTCAATGCTTCAGC-3' (SEQ ID NO: 1 )

P4; S'-GATGCCAAGATGATCAGCCATTCTGGAA-S'

(SEQ ID NO: 2) (see Fig 1. for annealing positions).

10 μl of the 50 μl PCR reaction mixture was electrophoresed in a 1.2% agarose gel and visualised by ethidium bromide staining and ultraviolet illuminated photography. Samples analysed by Southern hybridisation were transferred to Hybond N+ (Amersham, U.K.) nylon membrane with 0.4 N NaOH solution overnight and hybridised with exon- specific probes made by PCR (35 cycles of 94°C 1 min, 55°C 1 min, 72°C 2 min ) using the primers listed below and a CD44 genomic clone C2311 (9) as template:

Standard probe (exons 4 and 5)

P2; 5'-CCTGAAGAAGATTGTACATCAGTCACAGAC-3'

(SEQ ID NO: 3)

AEX5; 5'-AGCAGGGATTCTGTCTGTGCTGTC-3' (SEQ ID NO: 4)

Exon 7 E 1 ; 5 ' -TTGATGAGCACTAGTGCTACAGCA-3 ' (SEQ ID NO: 5) E2; 5'-CATTTGTGTTGTTGTGTGAAGATG-3' (SEQ ID NO: 6)

Exon 8

V3; 5'-TACGTCTTCAAATACCATCTCAGC-3' (SEQ ID NO: 7) AD1 ; 5'-GGTGCTGGAGATAAAATCTTCATC-3^* (SEQ ID NO: 8)

Exon 11

T6; 5'-TCCAGGCAACTCCTA-3' (SEQ ID NO: 9) TA6; 5'-CAGCTGTCCCTGTTG-3' (SEQ ID NO: 10)

Probes were labelled with peroxidase using ECL direct nucleic acid labelling (Amersham) to produce chemiluminescent probes and detected by autoradiography. Primer annealing positions are shown in shown in Fig 1. The conditions used for hybridisation, washing and detection were those recommended by the manufacturer's protocol.

Exon-Link Assay

The linkage of which variant exon abuts upon standard exon 5 was analysed using RT-PCR. The sample preparation and PCR reactions were performed as described above except that 100 ng of poly A+ selected RNA was always used for the RT reaction. The anchor primer was P1 (see above) and the primers used for variant exon amplification were as follows:

E2; δ'-CATTTGTGTTGTTGTGTGAAGATG-S' (SEQ ID NO: 6)

5/7 5'-CACTAGTGCTCATCAAAGTGGTAG-3^* (SEQ ID NO: 11)

5/8 δ'-TGGTATTTGAAGACGTACTGGTAG-S' (SEQ ID NO: 12)

5/9 5'-CCCGTGGTGTGGTTGAAATGGTAG-3' (SEQ ID NO: 13)

5/10 5'-TGCCATTTCTGTCTACATTGGTAG-3^* (SEQ ID NO: 14)

5/11 5'-TACTAGGAGTTGCCTGGATGGTAG-3' (SEQ ID NO: 15)

5/12 5'-TGGTATGAGCTGAGGCTGTGGTAG-3' (SEQ ID NO: 16)

5/13 5'-TATGACTGGAGTCCATATTGGTAG-3' (SEQ ID NO: 17)

5/14 5'-TCTGAGAATTACTCTGCTTGGTAG-3' (SEQ ID NO: 18)

The position of annealing of primers 5/X and the design of the assay, termed the "Exon-Link" assay are depicted in Fig 2. The assay depends on a PCR amplification between an anchored primer (P1) in a standard exon of the gene transcript and a set of primers (5/X) designed to overlap the junction between standard exon 5 and any neighbouring variant exon. The assay was designed so that only six 3' bases of the 24-base primer annealed to exon 5. This prevented binding and extension of the primer in the absence of a variant exon because efficient polymerase extension is dependent upon stable annealing to the target sequence. Primers were designed to anneal to the possible junctions of exons 5/7 through to 5/14. All primers had the same 3' sequence of 6 bases, homologous to exon 5 so that the presence of a specific variant would result in an amplified product of 348bp. Products of the expected size were generated and their identity could be confirmed by hybridisation with a CD44 standard exon probe to ensure specificity of amplification. The PCR products were electrophoresed on 1.2% agarose gels and visualised by ethidium bromide staining. RESULTS

CD44 Exon-specific RT-PCR/Southern Hybridisation analysis

In accordance with the inventors previous studies (14,15) PCR amplification resulting from the use of primers P1 and P4 (primer pair P1-P4) on tumour cell lines and tissue cDNA produced a product of 482 bp in all samples representing the ubiquitously expressed standard CD44 (CD44s) transcript (exons 1-5 and 16 to 20). In tumour samples this product was accompanied by several amplicons of higher molecular weight and hybridisation with exon-specific probes confirmed that the corresponding abundant transcripts contained combinations of several variant exons (CD44v).

The expression of total CD44 transcripts in all samples was assessed by hybridisation of the blotted PCR product with a probe complementary to exons 4 and 5 (standard exon probe), (results not shown). On hybridisation with the standard probe and certain variant exon probes, many tumour samples gave such strong signals that they appeared as a dense smear on the resulting autoradiograph, suggesting the presence of a complex array of misprocessed CD44 transcripts. Over- exposure of the autoradiograph was allowed in order to detect weak signals and to allow the less abundant, larger transcript species to be identified. The tumour samples regularly contained more of the higher molecular weight transcripts than matched normal tissue samples. By the use of probes for both exon 12 and exon 8, hybridisation signals were obtained in 16 of 20 (80%) tumour samples. With the exon 12 probe, signals were detected weakly in only 3 of 20 (15%) of normal mucosas, whilst exon 8 expression was seen in 14 of 20 (70%) matched normal tissue samples. Expression of exon 7 was detected in 54% of tumour tissue samples but in only 15% of matched normal mucosae. Optimisation of CD44 exon junction analysis using oesophageal and gastrointestinal tumour cell lines

A schematic representation of the PCR-based approach is shown in Fig. 2. Initial studies were performed on a number of cell lines in order to optimise the "Exon-Link" assay conditions for experimental analysis of the identity and the pattern of assembly of variant CD44 transcripts in tissue samples. In the 3 oesophageal cell lines (TE series) and the colonic carcinoma line (HT29) every possible 5/X junction was found to be present in CD44 mRNA. A primer (E2) complementary to 3' sequences of exon 7 was used with the anchor primer P1 in a control reaction and this performed several functions. First, a P1 - 5/7 amplicon was only ever seen when the presence of exon 7 was confirmed by a P1 -E2 amplification, verifying the specificity of the 5/7 junction primer, and therefore of all junction primer design. Secondly, the fact that the P1-E2 product was always of the expected molecular weight (420bp) confirmed that CD44 exon 6 which is expressed in mice was not included in human tissue transcripts. These findings also showed that a putative exon 5 "cryptic" splice site (16) was not used in the tissues thus far tested. A plasmid construct (pBSIS) derived from peripheral blood lymphocyte mRNA containing only CD44 standard exons and therefore only exon junction 5/16, was used as a negative control template in all experiments to confirm the specificity of 5/X primer annealing. A P1-P4 positive PCR control reaction was included with each sample and this confirmed the presence of CD44 cDNA.

Exon-Link analysis of colonic tumour and normal mucosa.

Optimisation and standardisation of the assay was followed by evaluation of CD44 variant transcript assembly in fresh human colonic tissue samples from surgical resection specimens. The production of amplicons of the expected size (342 bp) confirmed the presence within this tumour type of a variety of transcripts in which exons 7, 8 and 9 abut upon the standard exon 5, whilst only the junction of exon 5 with exon 8 is detected in the normal tissue sample.

The results of the full Exon-Link analysis of 20 tumour and 20 matched normal mucosal samples are presented in Table 1 and a number of observations emerge from the data; Colonic tumour and normal tissues both displayed 80% prevalence of junction 5/8 (16 of 20 cases) and this data correlates well with the RT-PCR/Southern hybridisation results on the same samples. Together these data show that junction 5/8 is the most common in normal mucosa and may therefore be the correct, or preferred splicing option in colonic mucosa. The junctions 5/7 and 5/11 are markedly increased in colonic tumour tissues (both 17 of 20 (85%)) compared to their normal counterparts. Both are relatively uncommon in normal mucosa with scores of 5 of 20 (25%) and 6 of 20 (30%), respectively. Conversely, junction 5/9 was present in 60% of tumours and 35% of normal tissues making it less suitable as a distinguishing marker. The junction 5/10 appeared to be extremely rare in neoplastic or normal colonic tissues even though it was present in cell lines. Junctions between exon 5 and those towards the more 3' of the variant region were less prevalent in both normal and neoplastic tissue. The data show that the probability of an exon being spliced against the standard exon 5 diminishes as the exon resides at a more distal site in the pre-mRNA molecule. Junctions between exons 12 or 13 and exon 5 were present in tumour samples at between 15-30% prevalence and in normal samples at 10-15%. However, the junction 5/14 was found in 40-50% in tumour and normal samples.

DISCUSSION

The data obtained from this investigation demonstrate a general disorder involving assembly of CD44 gene transcripts in tumour cells and tissues. This work was designed to identify whether there is a tumour-defining splicing pattern which could be diagnostically useful and provide insights into the mechanisms involved in the abnormal expression of CD44 in the neoplastic process. In fact, it was found that normal tissues have preferential splicing patterns but that in tumour cells the regulation of transcript assembly is disorganised, resulting in the generation of a wide variety of unusual alternatively spliced mRNA species. In normal colonic mucosa the standard exon 5 was characteristically adjacent to variant exon 8 but in tumour tissue, transcripts in which exon 5 abuts upon variant exon 7 were common. Therefore, as exon 7 is usually removed from transcripts in normal mucosa its presence is a sign of disturbed CD44 mRNA assembly indicative of neoplasia.

From a diagnostic standpoint RT-PCR/Southern hybridisation studies have previously shown that there is both increased transcription and over-expression of the variant exons of this gene in tumour cells. However, this method is technically difficult and laborious for routine clinical use. This new assay has the advantage of being simpler, in that blotting and hybridisation are unnecessary, without compromising the specificity of detection of CD44 mRNA species. The comprehensive analysis of exon junctions at the boundary between the standard and variant regions described above has demonstrated that the combinations 5/7 and 5/11 are the most prevalent in tumour tissues and therefore of most interest. These junctions were seen in 85% of tumours and this value compares favourably with the best tumour marker data obtained with the more complicated methods previously used for CD44 analysis (15,9).

As well as the identification of a specific exon junction for use in tumour detection, the data obtained in this study also allows a number of observations to be made regarding CD44 expression. The observation that all possible junctions tested were present in the carcinoma cell lines without any typical patterns, demonstrates that the widespread disorganisation of CD44 expression and alternative splicing in such source material is not truly representative of tumours in vivo. The cell lines were useful as positive control material especially in the optimisation of the assay and the verification of all primers. However, for the clinical assessment of a new observation it is ultimately necessary to examine human tissue directly.

The sensitivity of PCR is such that the presence of relatively few molecules containing the junction under investigation, would be detected. Therefore, the absence of a signal with a given primer combination in such an assay indicates that the corresponding exon junction is extremely rare. The presence of an exon 5 to 7 junction in 10- 15% of normal colonic tissues may represent a "background" of mis- processed CD44 transcripts in some individuals. As no biological process is 100% efficient there may be considerable variation in the degree of splicing efficiency between individuals in a population.

Although splicing is more disorganised in the tumour tissue samples the data suggest certain patterns and preferences of splicing in the nuclei of colonic mucosal cells. For instance, exon 8 seems to be the most usual variant exon abutting exon 5 whereas exon 10 appears to be spliced out of the pre-mRNA efficiently in both tumour and normal tissues. It may be that this exon is the most accessible and therefore the first splicing site engaged. The data also suggest that other exons appear to be spliced together as pairs or as cassettes. For example, it has previously been documented that exons 1 1 and 12 are equally discriminatory for tumour detection in colonic mucosa by Southern hybridisation analysis (Reference 9). The exon-link data showing the markedly increased presence of the junction 5/11 compared to junction 5/12, implies that they are spliced as a unit, spliced together in ascending order. The data obtained in this work suggests some possible new avenues for tumour diagnostic applications using the abnormalities in CD44 gene expression. These include immunoassays for exon-spanning epitopes, encoded by transcripts prevalent in tumour tissues. It is a simple matter to design oligopeptides corresponding to the oligonucleotides which bridge tumour-related exon junctions. Additionally, a test for tumour-related exon junctions could be incorporated in multiplex assays for other known disorders of CD44 expression, such as intron-retention (15). Analysis of the ratios of unusual exon junctions to those typically seen in a given normal tissue, could be accomplished by automated, PCR-based analytical methods now becoming available. This could conceivably provide a rapid, clinically relevant digital result.

REFERENCES

1. Screaton, G.R., Bell, M.V., Jackson, D.G., Comeilis, F.B., Gerth, U., Bell, J.I. Genomic structure of DNA encoding the lymphocyte homing receptor CD44 reveals at least 12 alternatively spliced exons. Proc. Natl. Acad. Sci. USA, 89: 12160-12164, 1992.

2. Jalkanen S, Jalkanen M, Bargatze R, Tammi M, Butcher EC: Biochemical properties of glycoproteins involved in lymphocyte recognition of high endothelial venules in man. J Immunol 141 : 1615-1623, 1988 3. Stamenkovic, I., Aruffo, A., Amoiot, M., Seed, B. The hematopoietic and epithelial forms of CD44 are distinct polypeptides with different adhesion potentials for hyaluronate-bearing cells. EMBO J, 10: 343-348, 1991.

4. Lesley J, Hyman R, Kincade PW: CD44 and its interaction with the extracellular matrix. Adv Immunol 54: 271 -335, 1993

5. Matsumura, Y., Tarin, D. Significance of CD44 gene products for cancer diagnosis and disease evaluation. Lancet, 340: 1053-1058, 1992.

6. Heider, K-H., Dammrich, J., Skroch-Angel, P., Muller- Hermelink, H-K., Vollmers, H.P., Herrlich, P., and Ponta, H. Differential expression of CD44 splice variants in intestinal- and diffuse-type human gastric carcinomas and normal gastric mucosa. Cancer Res, 53: 4197- 4203, 1993.

7. Kaufmann, M., Heider, K-H., Sinn, H-P., von Minckwitz, G., Ponta, H., Herrlich, P. CD44 variant exon epitopes in primary breast cancer and length of survival. Lancet, 345: 615-619, 1995. 8. Matsumura, Y., Sugiyama, M., Matsumura, S., Hayle, A. J.,

Robinson, P., Smith, J. C, and Tarin, D. Unusual retention of introns in CD44 gene transcripts in bladder cancer provides new diagnostic and clinical oncological opportunities. J. Pathol., 177: 11-20, 1995. 9. Yoshida, K., Bolodeoku, J., Sugino, T., Goodison, S.,

Matsumura, Y., Warren, B. F., Toge, T., Tahara, E., and Tarin, D. Abnormal retention of intron 9 in CD44 gene transcripts in human gastrointestinal tumours. Cancer Res., 55: 4273-4277, 1995.

10. Gunthert U, Hofmann M, Rudy W., Reber S., Zoller M., Haussmann I., Matzku S., Wenzel A., Ponta H., Herrlich P. A new variant of glycoprotein CD44 confers metastatic potential to rat carcinoma cells. Cell 65: 13-24, 1991.

11. Tanabe, K.K., Ellis, L.M., Saya, H. Expression of CD44R1 adhesion molecule in colon carcinomas and metastasis. Lancet, 341 : 725- 726, 1993.

12. Sugiyama, M., Woodman, A., Sugino, T., Crowley, S., Ho, K., Smith, J., Matsumura, Y., Tarin, D. Non-invasive detection of bladder cancer by identification of abnormal CD44 proteins in exfoliated cancer cells in urine. J Clin Pathol: Mol Pathol, 48: M142-147, 1995. 13. Sugino, T., Gorham, H., Yoshida, K., Bolodeoku, J., Nargund,

V., Cranston, D., Goodison, S., and Tarin, D. Progressive loss of CD44 gene expression in invasive bladder cancer. Am. J. Pathol., 149: 873-882, 1996.

14. Gorham H, Sugino T, 1996 Distribution of CD44 mRNA in archival paraffin wax embedded tumours and normal tissues viewed by in situ hybridisation. J Clin Pathol: Mol Pathol 49,3. 147-150.

15. Yoshida K., Sugino T., Bolodeoku J, Warren BF, Goodison S, Woodman A, Toge T, Tahara E, Tarin D. Detection of exfoliated carcinoma cells in colonic luminal washings by identification of deranged patterns of expression of the CD44 gene. Journal of Clinical Pathology 49, 300-305 (1996). 16. Cooper D L. Retention of CD44 introns in bladder cancer:

Understanding the alternative splicing of pre-mRNA opens new insights into the pathogenesis of human cancers. J Pathol 177:1 -3, 1995.

TABLE 1

Exon Junction 5/7 5/8 5/9 5/10 5/11 5/12 5/13 5/14

Cell Line

TE-1 + + + + + + +

HT-29 + + + + + + +

Colonic Tissues

Tumour 17/20 16/20 12/20 4/20 17/20 3/20 6/20 8/20

(85%) (80%) (60%) (20%) (85%) (15%) (30%) (40%

Normal 5/20 16/20 7/20 3/20 6/20 2/20 3/20 10/20

(25%) (80%) (35%) (15%) (30%) (10%) (15%) (50%)

Results of Exon-Link assay on human cell lines and human colonic carcinoma tissues and corresponding normal colonic mucosa. Each tissue sample was tested for the presence of each of the possible junctions between the standard exon 5 and variant exons 7 to 14 in CD44 transcripts.

TE-1 : human oesophageal cell line and HT-29: human colon carcinoma cell line.

Claims

1 A diagnostic method which comprises subjecting mRNA to reverse transcriptase using a first primer complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene, and observing whether an extension product is formed

2 A diagnostic method which comprises subjecting mRNA to hybridisation with a probe complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene, and observing the formation of a hybrid of said probe and said mRNA

3 A diagnostic method as claimed in claim 1 or claim 2, wherein the method comprises RT-PCR using two CD44 gene primers and amplification of a nucleic acid is observed

4 A diagnostic method as claimed in claim 3 wherein one primer is labelled

5 A diagnostic method as claimed in claim 4, wherein the other primer is immobilised or immobilisable 6 A diagnostic method as claimed in any one of claims 3 to 5, wherein a second primer is complementary to a standard exon 7 A diagnostic method as claimed in any one of claims 1 to 6, wherein the first primer is complementary to a sequence that bridges exons 5/7 or 5/9 or 5/11 8 A diagnostic method as claimed in any one of claims 1 to 7, performed to diagnose tumour or cancerous cells

9 A diagnostic kit comprising two CD44 gene primers including a first primer complementary to a sequence that bridges two exons of which at least one is a variable exon of the CD44 gene 10 A diagnostic kit as claimed in claim 9, wherein a second primer is complementary to a standard exon

11. A diagnostic kit as claimed in claim 9 or claim 10, wherein one primer is labelled and the other primer is immobilised or immobilisable.