1227
Detection of melanoma cells in peripheral blood by means of reverse transcriptase and polymerase chain reaction
Only small numbers of cells from solid tumours are needed for haematogenous metastasis. Detection is difficult because existing techniques are not sensitive enough. We have used reverse transcriptase to make complementary DNA from peripheral blood messenger RNA, and the polymerase chain reaction (PCR) to amplify cDNA specific for a gene actively transcribed only in the tumour tissue type. We prepared c DNA from peripheral blood of seven patients with malignant melanoma, four patients with other metastatic cancers, and four healthy subjects, as well as from several melanoma-derived cell lines. PCR was used to amplify the gene for tyrosinase, a tissue-specific gene in melanocytes. Since normal melanocytes are not thought to circulate in peripheral blood, detection of tyrosinase transcription in peripheral blood should indicate the presence of circulating cancer cells. The method was highly sensitive and could detect a single melanoma cell from a cell line in 2 ml normal blood. Blood samples from four of the seven patients with malignant melanoma gave positive results, whereas all eight control subjects gave negative results. This method does not depend on the of characterisation cancer-specific genetic abnormalities and can be applied to any cancer for which tissue-specific genes can be identified, including epithelial cancers. It could prove useful in the diagnosis of primary or metastatic cancers, in assessing prognosis, and in detecting residual disease after treatment. Introduction
Haematogenous spread of cancer influences outcome of the disease for most patients. However, little is known about the nature of this process. Implantation of cancer cells from peripheral blood is believed to be an inefficient process,’ and for the common solid tumours few cancer cells are detected in peripheral blood by routine microscopy. Detailed study of the numbers, appearance, and biology of this important subpopulation of tumour cells has been limited by the methods available. Detection of circulating tumour cells depended on the availability of specific antibodies to tumour-specific cell-surface antigens and the use of immunocytological techniques. An example of this method is the detection of neuroblastoma cells in blood.2 Since the numbers of tumour cells in peripheral blood are very small, techniques for their detection need to be not only highly sensitive but also specific. The polymerase chain reaction (PCR)3 could enable detection of cancer cells in the
peripheral blood by amplification of specific DNA sequences. It can be applied directly, however, only when cancer-specific abnormalities in DNA-for example, the translocation forming the Philadelphia chromosome4 or point mutations in the RAS genes-have been characterised. Such abnormalities have not yet been found for most common solid cancers. We therefore tried an alternative approach, we used reverse transcriptase to prepare complementary DNA (cDNA) from peripheral blood messenger RNA and targeted tissue-specific gene transcription to identify cancer cells. Only genes that are being actively expressed should be represented in the cDNA. Since we would not expect epithelial or mesenchymal cells or melanocytes normally to be present in the peripheral blood, genes specific to these tissues might be suitable targets for the detection of malignant circulating cells. We have evaluated this approach in malignant melanoma. This tumour has a wide range of metastatic sites, most of which are reached haematogenously. Normal melanocytes are not believed to circulate, and a range of melanin biosynthetic enzymes can act as tissue-specific proteins. We chose tyrosinase, an important enzyme in melanin biosynthesis, because the genomic organisation6 and the cDNA sequenceare known. The tyrosinase gene contains several introns, so for DNA amplification by PCR we selected primers represented on different exons of the gene. PCR would be expected to amplify cDNA between the primers, but genomic DNA either would not be amplified or would generate a PCR product of much greater length.
Methods Primer sequences were devised from the published sequence for human tyrosinase cDNA7 (fig 1). The outer primers produce a PCR fragment of 284 bp and the nested primers a fragment of 207 bp. P-globin primers were devised from the published sequence:8 and Glo4=ACCCAGAGGTTCTTTGAGTC (sense) Glo5 TCTGATAGGCAGCCTGCACT (anti-sense). They are located either side of intron 2 and produce a fragment of 283 bp. Four established human melanoma-derived cell lines were used-the melanotic SK-mel-19 and SK-mel-239 and the amelanotic SK-mel-289 and M5"* cell lines. All cultures were maintained on an equal mixture of Dulbecco’s modified Eagle’s medium and RPMI 1640, supplemented with 5% fetal calf serum (Sera-Lab). Cultures were transferred by means of trypsin (0-25%) and 5 mmol/1 edetic acid (EDTA). =
ADDRESSES: Yorkshire Cancer Research Campaign Institute for Cancer Studies, St James’s University Hospital, Leeds (B. Smith, BSc, Prof P. Selby, FRCP, J. Southgate, PhD, K. Pittman, MB, C. Bradley, MD), and Department of Biochemistry and Molecular Biology, University of Leeds, Leeds, UK (G. E. Blair, PhD). Correspondence to Prof Peter Selby, Yorkshire Cancer Research Campaign Institute for Cancer Studies, St James’s University Hospital, Beckett Street, Leeds LS9 7TF, UK.
1228
DETECTION OF TYROSINASE AND 3 GLOBIN mRNA IN CANCER PATIENTS AND HEALTHY SUBJECTS
Fig 1-Diagrammatic representation of exons 1-3 of human tyrosinase gene and cDNA showing relative positions of outer primers (V) and nested primers (B7). HTYR1 sequence comprises end of exon 1 and start of exon 2. BMN = long intervening sequences (introns) not represented in cDNA or PCR products. Genomic DNA is not amplified.
HTYR1=TTGGCAGATTGTCTGTAGCC (outer, sense). HTYR2=AGGCATTGTGCATGCTGCTT (outer, anti-sense). HTYR3=GTCT1TATGCAATGGAACGC (nested, sense) HTYR4=GCTATCCCAGTAAGTGGACT (nested, anti-sense).
Samples of blood from seven patients with malignant melanoma, patients with other cancers, and four healthy controls were analysed by reverse transcription and PCR in a "blind" trial. 2-10 ml blood was centrifuged at 1000 g for 5 min, the plasma was discarded, and the samples stored at —70°C. Total RNA was extracted by the guanidinium thiocyanate/caesium chloride method’from the blood samples and from about 107 cells from the cultured lines. For reverse transcription, 10 III RNA was heated at 90°C for 4 min, cooled rapidly, and diluted to 20 III with a mixture containing final concentrations of: 1 x PCR buffer (10 mmol/1 four
*Metastatic colerectal carcinoma in subjects 8 and 11, metastatic gastnc carcinoma in subject 9, and advanced transitional cell I carcinoma of the bladder in subject 10. NO = not done because sample was too small
"tris"-HCl pH 8-4, 50 mmol/1 potassium chloride, 100 g/ml 1 mmol/1 each of dATP, dCTP, dGTP, and TTP, 8 mmol/1 magnesium chloride, 25 pmol HTYR2, 20 units ’RNAguard’ (Pharmacia), and 4 units murine moloney leukaemia virus reverse transcriptase (Pharmacia). After incubation at 37°C for 1 h, half the sample (10 Jll) was diluted to 50 1 containing final concentrations of:1xPCR buffer, 200 umol/1 of each dNTP, 16 mmol/1 magnesium chloride, 150 pmol HTYRl and HTYR2, 0 1 % ’Triton X-100’, and 1 unit Taq DNA polymerase (Promega). Each sample was overlaid with oil and heated at 95°C for 5 min; 30 cycles PCR were then carried out (95°C for 65 s, 55°C for 65 s, 72"C for 50 s). For reamplification with the nested primers HTYR3 and HTYR4,5 pl of a 1 in 100 dilution were amplified in 25 1 reaction volume for a further 30 cycles. To minimise contamination, all components of the reactions (except RNA and DNA) were prepared in a classs 100 cabinet (Cleansphere’, Gelman Sciences) and positive displacement pipettes were used. PCR products were analysed by electrophoresis on 2% agarose gels, followed by ethidium bromide staining. Blood RNA integrity was checked by reverse transcription PCR with primers for human globin mRNA (Glo4 and Glo5).
gelatin),
Results The human cDNA clones Pmel34(tyrosinase) and
JW10212 (p globin) were used to confirm that PCR with each
primers gave bands of expected sizes. With outer primers alone and 30 cycles of PCR, tyrosinase mRNA from the equivalent of 50 cells or more was easily detected in the set
Fig 2-Detection of tyrosinase mRNA in 10 ml normal (subject 15) blood samples spiked with dilutions of SK-mel-23 cells (0-104). A= PCR with HTYR1 and HTYR2 alone (30+30 cycles). B = PCR with HTYR1 and HTYR2 followed by reamplification with HTYR3 and HTYR4 (30 cycles each) S=SK-mel-23 positive control ; P= Pme]34 DNA; N=no RNA control, M = ./A///7D3 marker; M’ = 123 bp ladder marker
of
melanoma cell lines SK-mel-19, SK-mel-23, and SK-mel28. The level of detection was greatly improved by further amplification with nested primers, and RNA equivalent to less than 1 cell could be detected. No tyrosinase mRNA was detected in the cell line M5, one of the two amelanotic lines tested. However, when the method was used to detect SK-mel-23 cells added to 10 ml samples of normal blood before RNA preparation, the sensitivity was lower; the minimum number of melanoma cells detected with outer primers alone was 10°, even after reamplification for an extra 30 cycles. However, at least 10 SK-mel-23 cells per 10 ml
1229
Fig 3-Detection of tyrosinase mRNA in blood of cancer patients. Lanes D, E, M, and P=positive malignant melanoma patients; lanes C, F, and G and N (same patient) negative malignant melanoma patients; lanes A, B, K, L, 0, and Q= negative controls (2 healthy subjects and 4 other cancers); lanes H, I. and R = positive controls (tyrosinase RNA and cDNA). =
blood could be detected after
reamplification with nested
primers (fig 2). To determine the level of detection more accurately, SK-mel-19 or SK-mel-28 cells were micromanipulated under an inverted microscope with a fine, heat-drawn capillary and different numbers were added to 2 ml normal blood samples. As few as 4 SK-mel-19 cells (1 per 5 x 105 white blood cells) and 1 SK-mel-28 cell (1 per 2 x 106 white blood cells) could be detected. Tyrosinase mRNA was detected in blood samples from four of the seven patients with malignant melanoma but in none of the healthy controls or patients with other cancers (see table, fig 3). Use of 40 rather than 30 cycles in the second round of PCR gave some false-positive results
individual
(not shown). To confirm that the PCR product was homologous with tyrosinase cDNA, we used restriction enzyme digestion. cDNA from patient 3 was amplified on a large scale and digested with three restriction enzymes (Hin Fl, Pvu 2, and Dde 1), each of which has one restriction site within the amplified sequence. All enzymes produced fragments of about the predicted sizes. A small proportion of uncut DNA could be explained by errors introduced into the restriction sites by the Taq polymerase.
Discussion We have developed a non-isotopic method to detect very small numbers of melanoma cells in human blood. The results so far are encouraging in patients with metastatic melanoma and we have not yet had any false-positive results in other subjects with the experimental conditions described. The sensitivity of the test is very high and is greater than that reported for other methods.2,l3 The difference in detection levels between cell lines could reflect different extents of transcription of the tyrosinase gene. Although the number of melanoma cells circulating in patients cannot be directly measured, the encouraging results suggest that this test may be clinically relevant. Further technical developments may increase sensitivity. It is not clear whether the false-positive results seen after 40 cycles in the second round of PCR are due to very low levels of contamination, as reported, for example, by Lo et al,14 or to "illegitimate transcription"15 of the tyrosinase gene. Our method takes only 2-3 days and avoids the use of radioactive
isotopes. We believe the test should be evaluated in large numbers of patients. If its usefulness is confirmed, the test may have important applications in malignant melanoma, and perhaps in many other cancers. The diagnosis of a primary cancer may be possible by the demonstration of inappropriate cells
in the peripheral blood, thus avoiding the use of more invasive diagnostic methods. Detection of such cells in patients with widespread disease at presentation seems likely. It is not clear whether the test would be useful in patients with apparently localised disease in whom peripheral blood cells may be few or absent. Once the diagnosis has been made, the test could be useful in assessment of prognosis. It seems likely that the presence of peripheral blood cells will increase the probability of metastases. Only extensive evaluation will determine whether the presence of tumour cells in the peripheral blood is always ultimately associated with metastases or whether implantation need not be inevitable. The continued presence of circulating cells after treatment is likely to be associated with a poor outcome. The test might help to identify patients who need systemic therapy after surgery, or those whose drug regimens should be changed or discontinued because of lack of effect, thus sparing unnecessary toxicity. We propose to explore these ideas further in malignant melanoma and other cancers. They can be readily extended to any cancer for which tissue-specific genes can be identified. We thank Dr C. Goding for Pmel34; Dr J. Old for JW102; Dr P. Cox for SK-mel-19 and SK-mel-28 cell lines; and Dr I. Hart for SK-mel-23 cell line.
REFERENCES 1. Tarin D, Price JE, Kettlewell MGW, Souter RG, Vass ACR, Crossley B. Clinicopathological observations on metastases in man studied in patients treated with peritoneovenous shunts. Br Med J 1984; 288: 749-51. 2. Moss TJ, Sanders DG. Detection of neuroblastoma cells in blood. J Clin Oncol 1990; 8: 736-40. 3. Saiki RK, Gelfand DH, Stoffel S, et al. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 1988; 239: 487-91. 4. Sawyers CL, Timson L, Kawasaki ES, Clark SS, Witte ON, Champlin R. Molecular relapse in chronic myelogenous leukemia patients after bone marrow transplantation detected by polymerase chain reaction. Proc Natl Acad Sci USA 1990; 87: 563-67. 5. Lyons J. Analysis of ras gene point mutations by PCR and oligonucleotide hybridization. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR protocols: a guide to methods and applications. New York: Academic Press, 1990: 386-91. 6. Tomita Y, Takeda A, Okinaga S, Tagami H, Shibahara S. Human oculocutaneous albinism caused by single base insertion in the tyrosinase gene. Biochem Biophys Res Commun 1989; 164: 990-96. 7. Kwon BS, Haq AK, Pomerantz SH, Halaban R. Isolation and sequence
8. 9.
10.
11.
of a cDNA clone for human tyrosinase that maps at the mouse c-albino locus. Proc Natl Acad Sci USA 1987; 84: 7473-77. Lawn RM, Efstratiadis A, O’Connell C, Maniatis T. The nucleotide sequence of the human &bgr;-globin gene. Cell 1980; 21: 647-51. Carey TE, Takahashi T, Resnick LA, Oettgen HF, Old LJ. Cell surface antigens of human malignant melanoma: mixed hemadsorption assays for humoral immunity to cultured autologous melanoma cells. Proc Natl Acad Sci USA 1976; 73: 3278-82. Liao SK, Dent PB, McCulloch PB. Characterization of human malignant melanoma cell lines. I. Morphology and growth characteristics in culture. J Natl Cancer Inst 1975; 54: 1037-44. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning, a laboratory manual. Cold Spring Harbor: Cold Spring Harbor Laboratory Press,
1989. 12. Wilson JT, Wilson LB, de Riel JK, et al. Insertion of synthetic copies of human globin genes into bacterial plasmids. Nucleic Acids Res 1978; 5: 563-81. 13. Naito H, Kuzumaki N, Uchino J, et al. Detection of tyrosine hydroxylase mRNA and minimal neuroblastoma cells by the reverse transcriptionpolymerase chain reaction. Eur J Cancer 1991; 27: 762-65. 14. Lo Y-MD, Patel P, Wainscoat JS, Sampietro M, Gillmer MDG, Fleming KA. Prenatal sex determination by DNA amplification from maternal peripheral blood. Lancet 1989; ii: 1363-65. 15. Chelly J, Concordet JP, Kaplan JC, Khan A. Illegitimate transcription: transcription of any gene in any cell type. Proc Natl Acad Sci USA 1989; 86: 2617-21.
Detection of melanoma cells in peripheral blood by means of reverse transcriptase and polymerase chain reaction
Only small numbers of cells from solid tumours are needed for haematogenous metastasis. Detection is difficult because existing techniques are not sensitive enough. We have used reverse transcriptase to make complementary DNA from peripheral blood messenger RNA, and the polymerase chain reaction (PCR) to amplify cDNA specific for a gene actively transcribed only in the tumour tissue type. We prepared c DNA from peripheral blood of seven patients with malignant melanoma, four patients with other metastatic cancers, and four healthy subjects, as well as from several melanoma-derived cell lines. PCR was used to amplify the gene for tyrosinase, a tissue-specific gene in melanocytes. Since normal melanocytes are not thought to circulate in peripheral blood, detection of tyrosinase transcription in peripheral blood should indicate the presence of circulating cancer cells. The method was highly sensitive and could detect a single melanoma cell from a cell line in 2 ml normal blood. Blood samples from four of the seven patients with malignant melanoma gave positive results, whereas all eight control subjects gave negative results. This method does not depend on the of characterisation cancer-specific genetic abnormalities and can be applied to any cancer for which tissue-specific genes can be identified, including epithelial cancers. It could prove useful in the diagnosis of primary or metastatic cancers, in assessing prognosis, and in detecting residual disease after treatment. Introduction
Haematogenous spread of cancer influences outcome of the disease for most patients. However, little is known about the nature of this process. Implantation of cancer cells from peripheral blood is believed to be an inefficient process,’ and for the common solid tumours few cancer cells are detected in peripheral blood by routine microscopy. Detailed study of the numbers, appearance, and biology of this important subpopulation of tumour cells has been limited by the methods available. Detection of circulating tumour cells depended on the availability of specific antibodies to tumour-specific cell-surface antigens and the use of immunocytological techniques. An example of this method is the detection of neuroblastoma cells in blood.2 Since the numbers of tumour cells in peripheral blood are very small, techniques for their detection need to be not only highly sensitive but also specific. The polymerase chain reaction (PCR)3 could enable detection of cancer cells in the
peripheral blood by amplification of specific DNA sequences. It can be applied directly, however, only when cancer-specific abnormalities in DNA-for example, the translocation forming the Philadelphia chromosome4 or point mutations in the RAS genes-have been characterised. Such abnormalities have not yet been found for most common solid cancers. We therefore tried an alternative approach, we used reverse transcriptase to prepare complementary DNA (cDNA) from peripheral blood messenger RNA and targeted tissue-specific gene transcription to identify cancer cells. Only genes that are being actively expressed should be represented in the cDNA. Since we would not expect epithelial or mesenchymal cells or melanocytes normally to be present in the peripheral blood, genes specific to these tissues might be suitable targets for the detection of malignant circulating cells. We have evaluated this approach in malignant melanoma. This tumour has a wide range of metastatic sites, most of which are reached haematogenously. Normal melanocytes are not believed to circulate, and a range of melanin biosynthetic enzymes can act as tissue-specific proteins. We chose tyrosinase, an important enzyme in melanin biosynthesis, because the genomic organisation6 and the cDNA sequenceare known. The tyrosinase gene contains several introns, so for DNA amplification by PCR we selected primers represented on different exons of the gene. PCR would be expected to amplify cDNA between the primers, but genomic DNA either would not be amplified or would generate a PCR product of much greater length.
Methods Primer sequences were devised from the published sequence for human tyrosinase cDNA7 (fig 1). The outer primers produce a PCR fragment of 284 bp and the nested primers a fragment of 207 bp. P-globin primers were devised from the published sequence:8 and Glo4=ACCCAGAGGTTCTTTGAGTC (sense) Glo5 TCTGATAGGCAGCCTGCACT (anti-sense). They are located either side of intron 2 and produce a fragment of 283 bp. Four established human melanoma-derived cell lines were used-the melanotic SK-mel-19 and SK-mel-239 and the amelanotic SK-mel-289 and M5"* cell lines. All cultures were maintained on an equal mixture of Dulbecco’s modified Eagle’s medium and RPMI 1640, supplemented with 5% fetal calf serum (Sera-Lab). Cultures were transferred by means of trypsin (0-25%) and 5 mmol/1 edetic acid (EDTA). =
ADDRESSES: Yorkshire Cancer Research Campaign Institute for Cancer Studies, St James’s University Hospital, Leeds (B. Smith, BSc, Prof P. Selby, FRCP, J. Southgate, PhD, K. Pittman, MB, C. Bradley, MD), and Department of Biochemistry and Molecular Biology, University of Leeds, Leeds, UK (G. E. Blair, PhD). Correspondence to Prof Peter Selby, Yorkshire Cancer Research Campaign Institute for Cancer Studies, St James’s University Hospital, Beckett Street, Leeds LS9 7TF, UK.
1228
DETECTION OF TYROSINASE AND 3 GLOBIN mRNA IN CANCER PATIENTS AND HEALTHY SUBJECTS
Fig 1-Diagrammatic representation of exons 1-3 of human tyrosinase gene and cDNA showing relative positions of outer primers (V) and nested primers (B7). HTYR1 sequence comprises end of exon 1 and start of exon 2. BMN = long intervening sequences (introns) not represented in cDNA or PCR products. Genomic DNA is not amplified.
HTYR1=TTGGCAGATTGTCTGTAGCC (outer, sense). HTYR2=AGGCATTGTGCATGCTGCTT (outer, anti-sense). HTYR3=GTCT1TATGCAATGGAACGC (nested, sense) HTYR4=GCTATCCCAGTAAGTGGACT (nested, anti-sense).
Samples of blood from seven patients with malignant melanoma, patients with other cancers, and four healthy controls were analysed by reverse transcription and PCR in a "blind" trial. 2-10 ml blood was centrifuged at 1000 g for 5 min, the plasma was discarded, and the samples stored at —70°C. Total RNA was extracted by the guanidinium thiocyanate/caesium chloride method’from the blood samples and from about 107 cells from the cultured lines. For reverse transcription, 10 III RNA was heated at 90°C for 4 min, cooled rapidly, and diluted to 20 III with a mixture containing final concentrations of: 1 x PCR buffer (10 mmol/1 four
*Metastatic colerectal carcinoma in subjects 8 and 11, metastatic gastnc carcinoma in subject 9, and advanced transitional cell I carcinoma of the bladder in subject 10. NO = not done because sample was too small
"tris"-HCl pH 8-4, 50 mmol/1 potassium chloride, 100 g/ml 1 mmol/1 each of dATP, dCTP, dGTP, and TTP, 8 mmol/1 magnesium chloride, 25 pmol HTYR2, 20 units ’RNAguard’ (Pharmacia), and 4 units murine moloney leukaemia virus reverse transcriptase (Pharmacia). After incubation at 37°C for 1 h, half the sample (10 Jll) was diluted to 50 1 containing final concentrations of:1xPCR buffer, 200 umol/1 of each dNTP, 16 mmol/1 magnesium chloride, 150 pmol HTYRl and HTYR2, 0 1 % ’Triton X-100’, and 1 unit Taq DNA polymerase (Promega). Each sample was overlaid with oil and heated at 95°C for 5 min; 30 cycles PCR were then carried out (95°C for 65 s, 55°C for 65 s, 72"C for 50 s). For reamplification with the nested primers HTYR3 and HTYR4,5 pl of a 1 in 100 dilution were amplified in 25 1 reaction volume for a further 30 cycles. To minimise contamination, all components of the reactions (except RNA and DNA) were prepared in a classs 100 cabinet (Cleansphere’, Gelman Sciences) and positive displacement pipettes were used. PCR products were analysed by electrophoresis on 2% agarose gels, followed by ethidium bromide staining. Blood RNA integrity was checked by reverse transcription PCR with primers for human globin mRNA (Glo4 and Glo5).
gelatin),
Results The human cDNA clones Pmel34(tyrosinase) and
JW10212 (p globin) were used to confirm that PCR with each
primers gave bands of expected sizes. With outer primers alone and 30 cycles of PCR, tyrosinase mRNA from the equivalent of 50 cells or more was easily detected in the set
Fig 2-Detection of tyrosinase mRNA in 10 ml normal (subject 15) blood samples spiked with dilutions of SK-mel-23 cells (0-104). A= PCR with HTYR1 and HTYR2 alone (30+30 cycles). B = PCR with HTYR1 and HTYR2 followed by reamplification with HTYR3 and HTYR4 (30 cycles each) S=SK-mel-23 positive control ; P= Pme]34 DNA; N=no RNA control, M = ./A///7D3 marker; M’ = 123 bp ladder marker
of
melanoma cell lines SK-mel-19, SK-mel-23, and SK-mel28. The level of detection was greatly improved by further amplification with nested primers, and RNA equivalent to less than 1 cell could be detected. No tyrosinase mRNA was detected in the cell line M5, one of the two amelanotic lines tested. However, when the method was used to detect SK-mel-23 cells added to 10 ml samples of normal blood before RNA preparation, the sensitivity was lower; the minimum number of melanoma cells detected with outer primers alone was 10°, even after reamplification for an extra 30 cycles. However, at least 10 SK-mel-23 cells per 10 ml
1229
Fig 3-Detection of tyrosinase mRNA in blood of cancer patients. Lanes D, E, M, and P=positive malignant melanoma patients; lanes C, F, and G and N (same patient) negative malignant melanoma patients; lanes A, B, K, L, 0, and Q= negative controls (2 healthy subjects and 4 other cancers); lanes H, I. and R = positive controls (tyrosinase RNA and cDNA). =
blood could be detected after
reamplification with nested
primers (fig 2). To determine the level of detection more accurately, SK-mel-19 or SK-mel-28 cells were micromanipulated under an inverted microscope with a fine, heat-drawn capillary and different numbers were added to 2 ml normal blood samples. As few as 4 SK-mel-19 cells (1 per 5 x 105 white blood cells) and 1 SK-mel-28 cell (1 per 2 x 106 white blood cells) could be detected. Tyrosinase mRNA was detected in blood samples from four of the seven patients with malignant melanoma but in none of the healthy controls or patients with other cancers (see table, fig 3). Use of 40 rather than 30 cycles in the second round of PCR gave some false-positive results
individual
(not shown). To confirm that the PCR product was homologous with tyrosinase cDNA, we used restriction enzyme digestion. cDNA from patient 3 was amplified on a large scale and digested with three restriction enzymes (Hin Fl, Pvu 2, and Dde 1), each of which has one restriction site within the amplified sequence. All enzymes produced fragments of about the predicted sizes. A small proportion of uncut DNA could be explained by errors introduced into the restriction sites by the Taq polymerase.
Discussion We have developed a non-isotopic method to detect very small numbers of melanoma cells in human blood. The results so far are encouraging in patients with metastatic melanoma and we have not yet had any false-positive results in other subjects with the experimental conditions described. The sensitivity of the test is very high and is greater than that reported for other methods.2,l3 The difference in detection levels between cell lines could reflect different extents of transcription of the tyrosinase gene. Although the number of melanoma cells circulating in patients cannot be directly measured, the encouraging results suggest that this test may be clinically relevant. Further technical developments may increase sensitivity. It is not clear whether the false-positive results seen after 40 cycles in the second round of PCR are due to very low levels of contamination, as reported, for example, by Lo et al,14 or to "illegitimate transcription"15 of the tyrosinase gene. Our method takes only 2-3 days and avoids the use of radioactive
isotopes. We believe the test should be evaluated in large numbers of patients. If its usefulness is confirmed, the test may have important applications in malignant melanoma, and perhaps in many other cancers. The diagnosis of a primary cancer may be possible by the demonstration of inappropriate cells
in the peripheral blood, thus avoiding the use of more invasive diagnostic methods. Detection of such cells in patients with widespread disease at presentation seems likely. It is not clear whether the test would be useful in patients with apparently localised disease in whom peripheral blood cells may be few or absent. Once the diagnosis has been made, the test could be useful in assessment of prognosis. It seems likely that the presence of peripheral blood cells will increase the probability of metastases. Only extensive evaluation will determine whether the presence of tumour cells in the peripheral blood is always ultimately associated with metastases or whether implantation need not be inevitable. The continued presence of circulating cells after treatment is likely to be associated with a poor outcome. The test might help to identify patients who need systemic therapy after surgery, or those whose drug regimens should be changed or discontinued because of lack of effect, thus sparing unnecessary toxicity. We propose to explore these ideas further in malignant melanoma and other cancers. They can be readily extended to any cancer for which tissue-specific genes can be identified. We thank Dr C. Goding for Pmel34; Dr J. Old for JW102; Dr P. Cox for SK-mel-19 and SK-mel-28 cell lines; and Dr I. Hart for SK-mel-23 cell line.
REFERENCES 1. Tarin D, Price JE, Kettlewell MGW, Souter RG, Vass ACR, Crossley B. Clinicopathological observations on metastases in man studied in patients treated with peritoneovenous shunts. Br Med J 1984; 288: 749-51. 2. Moss TJ, Sanders DG. Detection of neuroblastoma cells in blood. J Clin Oncol 1990; 8: 736-40. 3. Saiki RK, Gelfand DH, Stoffel S, et al. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 1988; 239: 487-91. 4. Sawyers CL, Timson L, Kawasaki ES, Clark SS, Witte ON, Champlin R. Molecular relapse in chronic myelogenous leukemia patients after bone marrow transplantation detected by polymerase chain reaction. Proc Natl Acad Sci USA 1990; 87: 563-67. 5. Lyons J. Analysis of ras gene point mutations by PCR and oligonucleotide hybridization. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR protocols: a guide to methods and applications. New York: Academic Press, 1990: 386-91. 6. Tomita Y, Takeda A, Okinaga S, Tagami H, Shibahara S. Human oculocutaneous albinism caused by single base insertion in the tyrosinase gene. Biochem Biophys Res Commun 1989; 164: 990-96. 7. Kwon BS, Haq AK, Pomerantz SH, Halaban R. Isolation and sequence
8. 9.
10.
11.
of a cDNA clone for human tyrosinase that maps at the mouse c-albino locus. Proc Natl Acad Sci USA 1987; 84: 7473-77. Lawn RM, Efstratiadis A, O’Connell C, Maniatis T. The nucleotide sequence of the human &bgr;-globin gene. Cell 1980; 21: 647-51. Carey TE, Takahashi T, Resnick LA, Oettgen HF, Old LJ. Cell surface antigens of human malignant melanoma: mixed hemadsorption assays for humoral immunity to cultured autologous melanoma cells. Proc Natl Acad Sci USA 1976; 73: 3278-82. Liao SK, Dent PB, McCulloch PB. Characterization of human malignant melanoma cell lines. I. Morphology and growth characteristics in culture. J Natl Cancer Inst 1975; 54: 1037-44. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning, a laboratory manual. Cold Spring Harbor: Cold Spring Harbor Laboratory Press,
1989. 12. Wilson JT, Wilson LB, de Riel JK, et al. Insertion of synthetic copies of human globin genes into bacterial plasmids. Nucleic Acids Res 1978; 5: 563-81. 13. Naito H, Kuzumaki N, Uchino J, et al. Detection of tyrosine hydroxylase mRNA and minimal neuroblastoma cells by the reverse transcriptionpolymerase chain reaction. Eur J Cancer 1991; 27: 762-65. 14. Lo Y-MD, Patel P, Wainscoat JS, Sampietro M, Gillmer MDG, Fleming KA. Prenatal sex determination by DNA amplification from maternal peripheral blood. Lancet 1989; ii: 1363-65. 15. Chelly J, Concordet JP, Kaplan JC, Khan A. Illegitimate transcription: transcription of any gene in any cell type. Proc Natl Acad Sci USA 1989; 86: 2617-21.
