- Email: [email protected]
Disseminated tumor cells in the bone marrow – chances and consequences of microscopical detection methods
Cancer Letters 197 (2003) 29–34 www.elsevier.com/locate/canlet
Disseminated tumor cells in the bone marrow – chances and consequences of microscopical detection methods Peter F. Ambrosa,*, Gabor Mehesa,b, Inge M. Ambrosa, Ruth Ladensteina,c a
Children’s Cancer Research Institute, St. Anna Children’s Hospital, Kinderspitalgasse 6, A-1090, Vienna, Austria b Department of Pathology, University Medical School of Pe´cs, Vienna, Austria c St. Anna Children’s Hospital, Vienna, Austria
Abstract The detection of disseminated tumor cells (DTCs) in the hematopoetic system is important for various reasons. It is essential for tumor staging. According to the International Neuroblastoma Staging System (INSS) only the cytomorphological examination of bone marrow smears is accepted despite the fact that an infiltrate below 0.1%, can hardly be detected and even infiltrates of more than 10% are sometimes overlooked. Another important aspect is the monitoring of the disease response to cytotoxic drugs by quantifying DTCs. Moreover, bone marrow aspirates represent an ideal source to determine the genetic and biological make up of DTCs at diagnosis and during follow up. Key issues that can be tested on DTCs are: determination of the proliferation capacity, the apoptotic rate, the drug sensitivity etc. The prerequisite for such a bone-marrow diagnosis, however, is the unequivocal identification of disseminated tumor cells. Thus, in order to avoid false positive and false negative results, which are a risk in bone-marrow diagnostics, a system was developed to distinguish tumor cells from non-neoplastic cells and to facilitate the gain of insights into the biological make-up of tumor cells more easily [1,2]. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Disseminated tumor cells; Bone marrow; Fluorescence in situ hybridization; Immunofluorescence; Immuncytology; Neuroblastoma; Breast carcinoma
1. Techniques for the detection of disseminated tumor cells in the bone marrow Conventional cytology on bone marrow smears is still the only accepted procedure for detecting disseminated neuroblastoma cells (according to International Neuroblastoma Staging System (INSS)). The sensitivity of this approach is very limited as contaminations of 0.1% are virtually not detectable and even a 10% tumor cell infiltrate can be overlooked * Corresponding author. Tel.: þ43-1-40470-411; fax: þ 43-1408-7230. E-mail address: [email protected] (P.F. Ambros).
(Mehes et al., submitted for publication). The limitations of this approach are based on the nonspecific appearance of individual tumor cells. Low level bone marrow involvement can only be reliably detected when cell nests are present and even this criterion can not always be used to unequivocally discriminate tumor cells from non-tumor cells. Other approaches to detect DTCs are: immunocytochemistry or immunofluorescence, flow cytometry, in vitro culture or polymerase chain reaction (PCR) techniques. Although there are various techniques for the detection of rare or residual (left over after cytotoxic treatment) tumor cells, the immunocytological detection has become the most widely used method over
0304-3835/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3835(03)00078-8
30
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
the past few years [3]. The main reasons are that in epithelial tumors and in embryonal tumors, tumortypical translocations are present only rarely and therefore only rarely specific translocation products are available for detection. If PCR techniques are used for the detection of tumor cells without tumor – typical translocations, gene products have to be used whose expressions are more or less restricted to these cells. PCR techniques, however, can give rise to falsepositive reactions. This is because PCR techniques are highly sensitive and can detect even single mRNA molecules, which can still be found in non-neoplastic cells, or ‘free’ mRNA molecules, which do not correlate with the presence of tumor cells. Thus, tumor-specific targets have to be used to reduce/avoid these false positive reactions. GD2 synthase might represent one target fulfilling this requirement [4]. It can be concluded that the use of generally accepted guidelines and a strict quality control is crucial when these techniques are used in the clinical routine [5]. 1.1. Immunocytology – alkaline phosphatase based detection In the past years, the detection of disseminated neuroblastoma cells has been performed mainly by using anti-GD2 antibodies, whereas cells from epithelial tumors can be detected with anti-cytokeratin antibodies. Although the expression of these antigens is mostly limited to neuroblastoma and epithelial cells and therefore highly specific, the antibody detection system used in most studies, i.e. alkaline phosphatase, does not have a comparable specificity, because of the fact that hematopoetic cells can also express alkaline phosphatase. This may cause false positive results [6]. Moreover, immunological procedures can always bring about other unspecific reactions and thereby simulate tumor cell-infiltration. Even standardization of staining procedures cannot prevent all false positive reactions. For this reason, the ISHAGE-group has been aiming to combine morphological criteria with immunologically positive reactions to increase the reliability and specificity of the immunological methods [7]. Morphological criteria, however, can very often not be employed since infiltrates such as one tumor cell in 105 or even one in 106 not necessarily reveal cells to which morphological criteria can be applied.
1.2. Immunocytology – immunofluorescence based detection Immunofluorescence detection of antibodies reacting with certain cell antigens are commonly used for flow cytometric analysis of all kinds of cells, but also for the detection of DTCs. Flow cytometric detection of rare cells is limited by the cytometric system. However, when the immunological staining is directly monitored in the fluorescence microscope – either manually or in an automatic fashion – the fluorescence based detection technique also allows the evaluation of low amounts of tumor cell infiltrates. In addition to the higher sensitivity of the microscopical examination also morphological information can be gained by this technique. Crucial information on the cytoplasm/nuclear size ratio or cytoplasmic inclusions, frequently associated with auto-fluorescence, can directly be established. Importantly, fluorescence microscopical methods allow the application of a wide range of either immunological or molecular cytogenetic techniques. The applications contain the simultaneous double or multiple fluorescence visualization of two or more antigens on the same cell. In Fig 1c an example of a breast carcinoma cell double positive with Ep-CAM and cytokeratin is presented. In addition, fluorescence based detection techniques allow the sequential application of fluorescence in situ hybridization (FISH) techniques or the TUNEL technique, and thus provide important information on the genetic or functional make up of these cells. Therefore, fluorescence based techniques are used in a number of laboratories for a quick and sensitive detection of tumor cells and for further characterization of immunologically positive cells e.g. [1,2,8,9]. 1.3. Specificity and sensitivity of the combination of two methods based on fluorescence detection: immunocytology and FISH Despite the fact that the false positive rate is strongly reduced when using fluorescence techniques, false positive results due to unspecific reactions can happen [2]. Therefore, several attempts have been made to find unequivocal criteria for the discrimination of tumor cells from non-tumor cells. Genetic aberrations, which distinguish a tumor, could for example be used as discriminating markers. Method-
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
31
Fig. 1. (a) Bone marrow cells from a neuroblastoma patient stained with FITC labelled GD2 antibody. (b) The same cells were subsequently analyzed by FISH using a MYCN specfic probe (FITC) and a chromosome 2 specific probe (TRITC). Only two of the three GD2 positive cells show the tumor typical MYCN amplification, the third GD2 positive cell displayed a normal MYCN copy number (arrow). Applying this genetic verification procedure, tumor cells can be unambiguously discriminated from normal cells. (c) Double color immunofluorescence of the EpCAM (green) and cytokeratin antigen (red) of a peripheral blood sample with disseminated breast carinoma cells. (d) The identical cell after FISH as in c) using a chromosome-17 specific probe (red) and a-X specific probe. The tetrasomy of chromosome 17 clearly identifies this cell as tumor cell.
wise, however, the detection of tumor-specific genetic features is difficult. PCR-techniques, for the above mentioned reasons are, at least partly, equally unsuitable. In situ hybridization, which is in fact a very specific technique, is in most instances no screening method and therefore is not suitable for bone marrow investigations without the combined use of a further method. Therefore, we decided to combine the high sensitivity of immunocytological detection with the high specificity of in situ hybridization. Usually, simultaneous representation of the protein and the DNA information is not possible, or only with extreme difficulty. The sequential representation of both parameters on the other hand, is only possible through the exact relocation of the immunologically positive cells. The prerequisite for such a
method is a fluorescence-based detection system. There are however, devices which meet even a number of additional requirements: e.g. the RC Detect-System (MetaSystems, Germany). This detection procedure, which is usually run overnight, facilitates the close inspection of the stored images the next day, either on the monitor or under the microscope. This is made possible by the relocation device of the motorized microscope-stage. Should the nature of the target cell (tumor cell/non-tumor cell) be still unclear after this kind of thorough microscopical inspection, an additional immunological investigation or even FISH may be undertaken. After automated relocation, any additional information on the cell in question can be directly observed under the microscope. Fig. 1a shows a bone marrow aspirate from a
32
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
neuroblastoma patient after GD2 coloring (FITC marked anti-GD2 antibody in green). Only FISH could finally prove that only two out of three GD2 positive cells were tumor cells and the third GD2 positive cell (arrow) displayed a false positive reaction. In this special case the detection of the tumor typical genetic aberrations, i.e. MYCN amplification (green signals in 1b) could discriminate between a specific and an unspecific reaction of the antibody and thus lead to an unambiguous result.
2. Quantification of tumor cell infiltration and kinetics of bone marrow clearing In order to detect decrease and increase of disseminated tumor cells during the illness, it must be possible to quantify the tumor cell infiltration. And again, modern data-processing-systems, as for example the RC Detect, are equipped to identify all the cells in a sample. This quantification makes it possible to track the kinetics of bone marrow clearing. The tumor-infiltration of stage 4 neuroblastoma patients, for example, was quantified at different times during therapy. By the use of a semiquantitative method, Seeger et al. were able to identify high-risk patients who had no or only delayed cell clearing [10]. A preliminary study conducted at the St. Anna Children’s Hospital, also succeeded in showing a definite correlation between rapid – within 90 days – tumor cell clearing and favorable outcome [11]. On the other hand, for 80% of patients who had no or only delayed tumor cell clearing of the bone marrow, the prognosis was rather poor. The prognostic significance of bone marrow clearing is most thoroughly discussed in studies on leukemia [12]. And in fact, some studies do already suggest clinical consequences for the treatment of leukemia based on kinetic findings on bone marrow clearing.
3. Analysis of biological characteristics of disseminated tumor cells in the bone marrow or peripheral blood Insights into the functional characteristics of disseminated tumor cells can provide vital information on aggressiveness of a tumor. There is a
fundamental difference between detecting a proliferative tumor cell or detecting an apoptotic one. Such diametrically opposed conditions may have completely different clinical consequences. We were, for example, able to show that the decrease in proliferation activity of disseminated tumor cells during therapy is indicative of a positive response to cytotoxic treatment (Mehes et al., in press). In other cases, the situation was completely reversed: in some patients with breast carcinoma, most or indeed all disseminated tumor cells in the peripheral blood proved to be apoptotic [13]. There are, however, quite a number of additional biological features which can be assessed rapidly and reliably by modern detection systems. The detection of resistancegenes (e.g. multidrug-resistance associated protein), for example, discloses the ability of a cell to rid itself of individual cytostatic agents either slowly or quickly. The ‘homing’ behavior can also be assessed with the appropriate antigen detection. Yet apart from these applications, even the analysis of a single relevant protein on the target cell may yield important additional information on the biological make up of the tumor cell, as for example the expression of EpCAM in epithelial tumors. The expression of this antigen has been found to correlate with rather aggressive tumor behavior [14]. A cytokeratin (red) and Ep-CAM (green) double positive cell was sequentially analyzed by FISH to visualize the tumor-associated genetic aberrations. FISH using a 17 specific probe (red) revealed four signals, whereas the X-chromosome (green) was present in two copies. Yet also more basic insights can be comparatively easily gained with the help of these new methods: as for example the answer to the question whether a possible genetic heterogeneity of the primary tumor can also be found in disseminated tumor cells. Last but not the least it might be possible to find out whether disseminated tumor cells in the bone marrow show the same genetic pattern as tumor cells metastasized to organs or lymph nodes.
4. Can patients diagnosed with a localized tumor have disseminated tumor cells in the bone marrow? The use of this genetic verification system has led
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
to a new approach to diagnosing neuroblastoma. Since 1991, scientists have believed that localized/regional neuroblastomas show bone marrow infiltration in 34% of the patients [15]. Through the use of the automatic immunological plus FISH (AIPF) technique, we were able to show that in 38.5% of patients with localized/ regional tumors, cells in the bone marrow display the tumor-associated antigen GD2. These cells, however, did not display any genetic aberrations[2]. Thus, in order to exclude the possibility of this negative result being caused by loss of genetic aberration in the DTCs, we tried to detect various genetic markers which we had found in the primary tumor, in the disseminated cells as well. Since no genetic aberration could be detected, we assumed that these cells were normal cells and not tumor cells. When analyzing a greater number of samples from patients with localized disease, less than 3% of the samples were found displaying GD2 positive cells with the tumortypical genetic aberration (unpublished observation). The reason why an unspecific fluorescence staining is found in the vast majority of those cells is often due to the phagocytotic activity of the macrophages together with the internalization of neuroblastoma cells or cell fragments and thus also of GD2-molecules. With regard to the glykolipid GD2, the antigen can also be transferred to non-tumor cells. This phenomenon is known as ‘transloading’ [2]. Thus, these cells are morphologically indistinguishable from tumor cells as well as regards the antigen distribution. Therefore, other information, either immunological or genetic, have to be taken into consideration to unambiguously discriminate those cells from tumor cells. In epithelial tumors and especially in breast cancer patients, a great number (running up to 43% [16]) of immunologically positive samples with low amounts of immunologically positive cells have been observed in the bone marrow and have been interpreted as tumor cells. These data are however, opposed by a study by Fetsch et al. [17]. Apart from immunocytochemical detection, the authors also used cytomorphological criteria for the detection of possible tumor cell infiltration. Yet none of their patients with localized breast carcinoma had immunologically positive cells, that met the authors’ morphological requirements for tumor cells. The reasons that lead to these diverging results are mostly methodological (for details see also the first sub-chapter). Besides
33
immuno-histochemical detection, it is therefore important to define additional criteria in order to reliably detect disseminated tumor cells. The use of morphological features is a valuable aid (as suggested by the ISHAGE group) for discriminating nonneoplastic cells from tumor cells [7]. One must bear in mind however that the cytomorphologically-based differentiation of tumor cells from non-neoplastic cells under the conditions of a minimal cell infiltration is difficult, if at all possible to achieve. Modern techniques such as immunological bi-or even multicolor detection (as has been used for years in flow cytometry) or subsequent fluorescence in situ hybridization, are very promising methods, also to be routinely used, for the unequivocal detection of disseminated tumor cells [1,18]
5. Conclusion Nowadays, modern detection devices allow for the first time an unambiguous identification and quantification of DTCs in the hematopoetic system and, importantly, a deeper insight into the biology of these cells. This thorough characterization of tumor cells is important in many respects. First, additional immunological techniques or even the combination of immunological and genetic methods (FISH) may unequivocally disclose whether the target cell is a tumor cell at all or whether it is only a false positively reacting cell. In addition, also the biological behavior of the target cell may be analyzed, e.g. its proliferation activity. Finally, it is also possible to decide whether one deals with a vital tumor cell or whether there are apoptotic processes involved. It may be assumed that immunological fluorescence-based methods will soon be introduced to automatic detection systems since the use of fluorescence-marked antibodies and of FISHprobes will offer a number of possibilities, which could otherwise hardly, or not at all be realized.
Acknowledgements We would like to thank A. Luegmayr for excellent technical support. The help of M. Zavadil and I. Walters is greatly acknowledged. This work was
34
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
supported by the CCRI and the EC grant (QLR1-CT2002-01768 SIOPEM-R-MET).
References [1] P.F. Ambros, G. Mehes, C. Hattinger, I.M. Ambros, A. Luegmayr, R. Ladenstein, H. Gadner, Unequivocal identification of disseminated tumor cells in the bone marrow by combining immunological and genetic approaches – functional and prognostic information, Leukemia 15 (2001) 275–277. [2] G. Mehes, A. Luegmayr, I.M. Ambros, R. Ladenstein, P.F. Ambros, Combined automatic immunological and molecular cytogenetic analsis allows exact identification and quantification of tumor cells in the bone marrow, Clin. Cancer Res. 7 (2001) 1969–1975. [3] A.A. Ross, Minimal residual disease in solid tumor malignancies: a review, J. Hematother. 7 (1998) 9–18. [4] M.S. Lo Piccolo, N.K. Cheung, I.Y. Cheung, GD2 synthase: a new molecular marker for detecting neuroblastoma, Cancer 92 (2001) 924–931. [5] P.F. Ambros, I.M. Ambros, Pathology and biology guidelines for resectable and unresectable neuroblastic tumors and bone marrow examinaion guidelines, Med. Pediatr. Oncol. 37 (2001) 492–504. [6] E. Borgen, K. Beiske, S. Trachsel, J.M. Nesland, G. Kvalheim, T.K. Herstad, E. Schlichting, H. Qvist, B. Naume, Immunocytochemical detection of isolated epithelial cells in bone marrow: non-specific staining and contribution by plasma cells directly reactive to alkaline phosphatase, J. Pathol. 185 (1998) 427–434. [7] E. Borgen, B. Naume, J.M. Nesland, G. Kvalheim, K. Beiske, O. Fodstad, I. Diel, Standardization of the immunocytochemical detection of cancer cells in the BM and blood: I. establishment of objective criteria for the evaluation of immunostained cells, Cytotherapy 1(5) (1999) 377–388. [8] N.K. Cheung, G. Heller, B.H. Kushner, C. Liu, I.Y. Cheung, Detection of metastatic neuroblastoma in bone marrow: when is routine marrow histology insensitive? J. Clin. Oncol. 15 (1997) 2807–2817. [9] T. Fehm, A. Sagalowsky, E. Clifford, P. Beitsch, H. Saboorian, D. Euhus, S. Meng, L. Morrison, T. Tucker, N. Lane, B.M.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
Ghadimi, K. Heselmeyer-Haddad, T. Ried, C. Rao, J. Uhr, Cytogenetic evidence that circulating epithelial cells in patients with carcinoma are malignant, Clin. Cancer Res. 8 (2002) 2073–2084. R.C. Seeger, C.P. Reynolds, R. Gallego, D.O. Stram, R.B. Gerbing, K.K. Matthay, Quantitative tumor cell content of bone marrow and blood as a predictor of outcome in stage IV neuroblastoma: a Children’s Cancer Group Study, J. Clin. Oncol. 18 (2000) 4067– 4076. P.F. Ambros, I.M. Ambros, R. Ladenstein, A. Luegmayr, S. Rumpler, H. Kovar, H. Gadner, Tumor cell clearing in the bone marrow of neuroblastoma patients-predictor for the outcome?, Med. Pediatr. Oncol. 31 (1998) 195. J.J. van Dongen, T. Seriu, E.R. Panzer-Grumayer, A. Biondi, M.J. Pongers-Willemse, L. Corral, F. Stolz, M. Schrappe, G. Masera, W.A. Kamps, H. Gadner, E.R. van Wering, W.D. Ludwig, G. Basso, M.A. de Bruijn, G. Cazzaniga, K. Hettinger, A. van der Does-van den Berg, W.C. Hop, H. Riehm, C.R. Bartram, Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood [see comments], Lancet 352 (1998) 1731– 1738. G. Mehes, A. Witt, E. Kubista, P.F. Ambros, Circulating breast cancer cells are frequently apoptotic, Am. J. Pathol. 159 (2001) 17– 20. G. Gastl, G. Spizzo, P. Obrist, M. Dunser, G. Mikuz, Ep-CAM overexpression in breast cancer as a predictor of survival, Lancet 356 (2000) 1981–1982. T.J. Moss, C.P. Reynolds, H.N. Sather, S.G. Romansky, G.D. Hammond, R.C. Seeger, Prognostic value of immunocytologic detection of bone marrow metastases in neuroblastoma, N. Engl. J. Med. 324 (1991) 219– 226. K. Pantel, M. Otte, Occult micrometastasis: enrichment, identification and characterization of single disseminated tumour cells, Semin. Cancer Biol. 11 (2001) 327–337. P.A. Fetsch, K.H. Cowan, D.E. Weng, A. Freifield, A.C. Filie, A. Abati, Detection of circulating tumor cells and micrometastases in stage II, III, and IV breast cancer patients utilizing cytology and immunocytochemistry, Diagn. Cytopathol. 22 (2000) 323–328. G. Mehes, A. Luegmayr, C.M. Hattinger, T. Lorch, I.M. Ambros, H. Gadner, P.F. Ambros, Automatic detection and genetic profiling of disseminated neuroblastoma cells, Med. Pediatr. Oncol. 36 (2001) 205–209.
Disseminated tumor cells in the bone marrow – chances and consequences of microscopical detection methods Peter F. Ambrosa,*, Gabor Mehesa,b, Inge M. Ambrosa, Ruth Ladensteina,c a
Children’s Cancer Research Institute, St. Anna Children’s Hospital, Kinderspitalgasse 6, A-1090, Vienna, Austria b Department of Pathology, University Medical School of Pe´cs, Vienna, Austria c St. Anna Children’s Hospital, Vienna, Austria
Abstract The detection of disseminated tumor cells (DTCs) in the hematopoetic system is important for various reasons. It is essential for tumor staging. According to the International Neuroblastoma Staging System (INSS) only the cytomorphological examination of bone marrow smears is accepted despite the fact that an infiltrate below 0.1%, can hardly be detected and even infiltrates of more than 10% are sometimes overlooked. Another important aspect is the monitoring of the disease response to cytotoxic drugs by quantifying DTCs. Moreover, bone marrow aspirates represent an ideal source to determine the genetic and biological make up of DTCs at diagnosis and during follow up. Key issues that can be tested on DTCs are: determination of the proliferation capacity, the apoptotic rate, the drug sensitivity etc. The prerequisite for such a bone-marrow diagnosis, however, is the unequivocal identification of disseminated tumor cells. Thus, in order to avoid false positive and false negative results, which are a risk in bone-marrow diagnostics, a system was developed to distinguish tumor cells from non-neoplastic cells and to facilitate the gain of insights into the biological make-up of tumor cells more easily [1,2]. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Disseminated tumor cells; Bone marrow; Fluorescence in situ hybridization; Immunofluorescence; Immuncytology; Neuroblastoma; Breast carcinoma
1. Techniques for the detection of disseminated tumor cells in the bone marrow Conventional cytology on bone marrow smears is still the only accepted procedure for detecting disseminated neuroblastoma cells (according to International Neuroblastoma Staging System (INSS)). The sensitivity of this approach is very limited as contaminations of 0.1% are virtually not detectable and even a 10% tumor cell infiltrate can be overlooked * Corresponding author. Tel.: þ43-1-40470-411; fax: þ 43-1408-7230. E-mail address: [email protected] (P.F. Ambros).
(Mehes et al., submitted for publication). The limitations of this approach are based on the nonspecific appearance of individual tumor cells. Low level bone marrow involvement can only be reliably detected when cell nests are present and even this criterion can not always be used to unequivocally discriminate tumor cells from non-tumor cells. Other approaches to detect DTCs are: immunocytochemistry or immunofluorescence, flow cytometry, in vitro culture or polymerase chain reaction (PCR) techniques. Although there are various techniques for the detection of rare or residual (left over after cytotoxic treatment) tumor cells, the immunocytological detection has become the most widely used method over
0304-3835/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3835(03)00078-8
30
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
the past few years [3]. The main reasons are that in epithelial tumors and in embryonal tumors, tumortypical translocations are present only rarely and therefore only rarely specific translocation products are available for detection. If PCR techniques are used for the detection of tumor cells without tumor – typical translocations, gene products have to be used whose expressions are more or less restricted to these cells. PCR techniques, however, can give rise to falsepositive reactions. This is because PCR techniques are highly sensitive and can detect even single mRNA molecules, which can still be found in non-neoplastic cells, or ‘free’ mRNA molecules, which do not correlate with the presence of tumor cells. Thus, tumor-specific targets have to be used to reduce/avoid these false positive reactions. GD2 synthase might represent one target fulfilling this requirement [4]. It can be concluded that the use of generally accepted guidelines and a strict quality control is crucial when these techniques are used in the clinical routine [5]. 1.1. Immunocytology – alkaline phosphatase based detection In the past years, the detection of disseminated neuroblastoma cells has been performed mainly by using anti-GD2 antibodies, whereas cells from epithelial tumors can be detected with anti-cytokeratin antibodies. Although the expression of these antigens is mostly limited to neuroblastoma and epithelial cells and therefore highly specific, the antibody detection system used in most studies, i.e. alkaline phosphatase, does not have a comparable specificity, because of the fact that hematopoetic cells can also express alkaline phosphatase. This may cause false positive results [6]. Moreover, immunological procedures can always bring about other unspecific reactions and thereby simulate tumor cell-infiltration. Even standardization of staining procedures cannot prevent all false positive reactions. For this reason, the ISHAGE-group has been aiming to combine morphological criteria with immunologically positive reactions to increase the reliability and specificity of the immunological methods [7]. Morphological criteria, however, can very often not be employed since infiltrates such as one tumor cell in 105 or even one in 106 not necessarily reveal cells to which morphological criteria can be applied.
1.2. Immunocytology – immunofluorescence based detection Immunofluorescence detection of antibodies reacting with certain cell antigens are commonly used for flow cytometric analysis of all kinds of cells, but also for the detection of DTCs. Flow cytometric detection of rare cells is limited by the cytometric system. However, when the immunological staining is directly monitored in the fluorescence microscope – either manually or in an automatic fashion – the fluorescence based detection technique also allows the evaluation of low amounts of tumor cell infiltrates. In addition to the higher sensitivity of the microscopical examination also morphological information can be gained by this technique. Crucial information on the cytoplasm/nuclear size ratio or cytoplasmic inclusions, frequently associated with auto-fluorescence, can directly be established. Importantly, fluorescence microscopical methods allow the application of a wide range of either immunological or molecular cytogenetic techniques. The applications contain the simultaneous double or multiple fluorescence visualization of two or more antigens on the same cell. In Fig 1c an example of a breast carcinoma cell double positive with Ep-CAM and cytokeratin is presented. In addition, fluorescence based detection techniques allow the sequential application of fluorescence in situ hybridization (FISH) techniques or the TUNEL technique, and thus provide important information on the genetic or functional make up of these cells. Therefore, fluorescence based techniques are used in a number of laboratories for a quick and sensitive detection of tumor cells and for further characterization of immunologically positive cells e.g. [1,2,8,9]. 1.3. Specificity and sensitivity of the combination of two methods based on fluorescence detection: immunocytology and FISH Despite the fact that the false positive rate is strongly reduced when using fluorescence techniques, false positive results due to unspecific reactions can happen [2]. Therefore, several attempts have been made to find unequivocal criteria for the discrimination of tumor cells from non-tumor cells. Genetic aberrations, which distinguish a tumor, could for example be used as discriminating markers. Method-
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
31
Fig. 1. (a) Bone marrow cells from a neuroblastoma patient stained with FITC labelled GD2 antibody. (b) The same cells were subsequently analyzed by FISH using a MYCN specfic probe (FITC) and a chromosome 2 specific probe (TRITC). Only two of the three GD2 positive cells show the tumor typical MYCN amplification, the third GD2 positive cell displayed a normal MYCN copy number (arrow). Applying this genetic verification procedure, tumor cells can be unambiguously discriminated from normal cells. (c) Double color immunofluorescence of the EpCAM (green) and cytokeratin antigen (red) of a peripheral blood sample with disseminated breast carinoma cells. (d) The identical cell after FISH as in c) using a chromosome-17 specific probe (red) and a-X specific probe. The tetrasomy of chromosome 17 clearly identifies this cell as tumor cell.
wise, however, the detection of tumor-specific genetic features is difficult. PCR-techniques, for the above mentioned reasons are, at least partly, equally unsuitable. In situ hybridization, which is in fact a very specific technique, is in most instances no screening method and therefore is not suitable for bone marrow investigations without the combined use of a further method. Therefore, we decided to combine the high sensitivity of immunocytological detection with the high specificity of in situ hybridization. Usually, simultaneous representation of the protein and the DNA information is not possible, or only with extreme difficulty. The sequential representation of both parameters on the other hand, is only possible through the exact relocation of the immunologically positive cells. The prerequisite for such a
method is a fluorescence-based detection system. There are however, devices which meet even a number of additional requirements: e.g. the RC Detect-System (MetaSystems, Germany). This detection procedure, which is usually run overnight, facilitates the close inspection of the stored images the next day, either on the monitor or under the microscope. This is made possible by the relocation device of the motorized microscope-stage. Should the nature of the target cell (tumor cell/non-tumor cell) be still unclear after this kind of thorough microscopical inspection, an additional immunological investigation or even FISH may be undertaken. After automated relocation, any additional information on the cell in question can be directly observed under the microscope. Fig. 1a shows a bone marrow aspirate from a
32
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
neuroblastoma patient after GD2 coloring (FITC marked anti-GD2 antibody in green). Only FISH could finally prove that only two out of three GD2 positive cells were tumor cells and the third GD2 positive cell (arrow) displayed a false positive reaction. In this special case the detection of the tumor typical genetic aberrations, i.e. MYCN amplification (green signals in 1b) could discriminate between a specific and an unspecific reaction of the antibody and thus lead to an unambiguous result.
2. Quantification of tumor cell infiltration and kinetics of bone marrow clearing In order to detect decrease and increase of disseminated tumor cells during the illness, it must be possible to quantify the tumor cell infiltration. And again, modern data-processing-systems, as for example the RC Detect, are equipped to identify all the cells in a sample. This quantification makes it possible to track the kinetics of bone marrow clearing. The tumor-infiltration of stage 4 neuroblastoma patients, for example, was quantified at different times during therapy. By the use of a semiquantitative method, Seeger et al. were able to identify high-risk patients who had no or only delayed cell clearing [10]. A preliminary study conducted at the St. Anna Children’s Hospital, also succeeded in showing a definite correlation between rapid – within 90 days – tumor cell clearing and favorable outcome [11]. On the other hand, for 80% of patients who had no or only delayed tumor cell clearing of the bone marrow, the prognosis was rather poor. The prognostic significance of bone marrow clearing is most thoroughly discussed in studies on leukemia [12]. And in fact, some studies do already suggest clinical consequences for the treatment of leukemia based on kinetic findings on bone marrow clearing.
3. Analysis of biological characteristics of disseminated tumor cells in the bone marrow or peripheral blood Insights into the functional characteristics of disseminated tumor cells can provide vital information on aggressiveness of a tumor. There is a
fundamental difference between detecting a proliferative tumor cell or detecting an apoptotic one. Such diametrically opposed conditions may have completely different clinical consequences. We were, for example, able to show that the decrease in proliferation activity of disseminated tumor cells during therapy is indicative of a positive response to cytotoxic treatment (Mehes et al., in press). In other cases, the situation was completely reversed: in some patients with breast carcinoma, most or indeed all disseminated tumor cells in the peripheral blood proved to be apoptotic [13]. There are, however, quite a number of additional biological features which can be assessed rapidly and reliably by modern detection systems. The detection of resistancegenes (e.g. multidrug-resistance associated protein), for example, discloses the ability of a cell to rid itself of individual cytostatic agents either slowly or quickly. The ‘homing’ behavior can also be assessed with the appropriate antigen detection. Yet apart from these applications, even the analysis of a single relevant protein on the target cell may yield important additional information on the biological make up of the tumor cell, as for example the expression of EpCAM in epithelial tumors. The expression of this antigen has been found to correlate with rather aggressive tumor behavior [14]. A cytokeratin (red) and Ep-CAM (green) double positive cell was sequentially analyzed by FISH to visualize the tumor-associated genetic aberrations. FISH using a 17 specific probe (red) revealed four signals, whereas the X-chromosome (green) was present in two copies. Yet also more basic insights can be comparatively easily gained with the help of these new methods: as for example the answer to the question whether a possible genetic heterogeneity of the primary tumor can also be found in disseminated tumor cells. Last but not the least it might be possible to find out whether disseminated tumor cells in the bone marrow show the same genetic pattern as tumor cells metastasized to organs or lymph nodes.
4. Can patients diagnosed with a localized tumor have disseminated tumor cells in the bone marrow? The use of this genetic verification system has led
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
to a new approach to diagnosing neuroblastoma. Since 1991, scientists have believed that localized/regional neuroblastomas show bone marrow infiltration in 34% of the patients [15]. Through the use of the automatic immunological plus FISH (AIPF) technique, we were able to show that in 38.5% of patients with localized/ regional tumors, cells in the bone marrow display the tumor-associated antigen GD2. These cells, however, did not display any genetic aberrations[2]. Thus, in order to exclude the possibility of this negative result being caused by loss of genetic aberration in the DTCs, we tried to detect various genetic markers which we had found in the primary tumor, in the disseminated cells as well. Since no genetic aberration could be detected, we assumed that these cells were normal cells and not tumor cells. When analyzing a greater number of samples from patients with localized disease, less than 3% of the samples were found displaying GD2 positive cells with the tumortypical genetic aberration (unpublished observation). The reason why an unspecific fluorescence staining is found in the vast majority of those cells is often due to the phagocytotic activity of the macrophages together with the internalization of neuroblastoma cells or cell fragments and thus also of GD2-molecules. With regard to the glykolipid GD2, the antigen can also be transferred to non-tumor cells. This phenomenon is known as ‘transloading’ [2]. Thus, these cells are morphologically indistinguishable from tumor cells as well as regards the antigen distribution. Therefore, other information, either immunological or genetic, have to be taken into consideration to unambiguously discriminate those cells from tumor cells. In epithelial tumors and especially in breast cancer patients, a great number (running up to 43% [16]) of immunologically positive samples with low amounts of immunologically positive cells have been observed in the bone marrow and have been interpreted as tumor cells. These data are however, opposed by a study by Fetsch et al. [17]. Apart from immunocytochemical detection, the authors also used cytomorphological criteria for the detection of possible tumor cell infiltration. Yet none of their patients with localized breast carcinoma had immunologically positive cells, that met the authors’ morphological requirements for tumor cells. The reasons that lead to these diverging results are mostly methodological (for details see also the first sub-chapter). Besides
33
immuno-histochemical detection, it is therefore important to define additional criteria in order to reliably detect disseminated tumor cells. The use of morphological features is a valuable aid (as suggested by the ISHAGE group) for discriminating nonneoplastic cells from tumor cells [7]. One must bear in mind however that the cytomorphologically-based differentiation of tumor cells from non-neoplastic cells under the conditions of a minimal cell infiltration is difficult, if at all possible to achieve. Modern techniques such as immunological bi-or even multicolor detection (as has been used for years in flow cytometry) or subsequent fluorescence in situ hybridization, are very promising methods, also to be routinely used, for the unequivocal detection of disseminated tumor cells [1,18]
5. Conclusion Nowadays, modern detection devices allow for the first time an unambiguous identification and quantification of DTCs in the hematopoetic system and, importantly, a deeper insight into the biology of these cells. This thorough characterization of tumor cells is important in many respects. First, additional immunological techniques or even the combination of immunological and genetic methods (FISH) may unequivocally disclose whether the target cell is a tumor cell at all or whether it is only a false positively reacting cell. In addition, also the biological behavior of the target cell may be analyzed, e.g. its proliferation activity. Finally, it is also possible to decide whether one deals with a vital tumor cell or whether there are apoptotic processes involved. It may be assumed that immunological fluorescence-based methods will soon be introduced to automatic detection systems since the use of fluorescence-marked antibodies and of FISHprobes will offer a number of possibilities, which could otherwise hardly, or not at all be realized.
Acknowledgements We would like to thank A. Luegmayr for excellent technical support. The help of M. Zavadil and I. Walters is greatly acknowledged. This work was
34
P.F. Ambros et al. / Cancer Letters 197 (2003) 29–34
supported by the CCRI and the EC grant (QLR1-CT2002-01768 SIOPEM-R-MET).
References [1] P.F. Ambros, G. Mehes, C. Hattinger, I.M. Ambros, A. Luegmayr, R. Ladenstein, H. Gadner, Unequivocal identification of disseminated tumor cells in the bone marrow by combining immunological and genetic approaches – functional and prognostic information, Leukemia 15 (2001) 275–277. [2] G. Mehes, A. Luegmayr, I.M. Ambros, R. Ladenstein, P.F. Ambros, Combined automatic immunological and molecular cytogenetic analsis allows exact identification and quantification of tumor cells in the bone marrow, Clin. Cancer Res. 7 (2001) 1969–1975. [3] A.A. Ross, Minimal residual disease in solid tumor malignancies: a review, J. Hematother. 7 (1998) 9–18. [4] M.S. Lo Piccolo, N.K. Cheung, I.Y. Cheung, GD2 synthase: a new molecular marker for detecting neuroblastoma, Cancer 92 (2001) 924–931. [5] P.F. Ambros, I.M. Ambros, Pathology and biology guidelines for resectable and unresectable neuroblastic tumors and bone marrow examinaion guidelines, Med. Pediatr. Oncol. 37 (2001) 492–504. [6] E. Borgen, K. Beiske, S. Trachsel, J.M. Nesland, G. Kvalheim, T.K. Herstad, E. Schlichting, H. Qvist, B. Naume, Immunocytochemical detection of isolated epithelial cells in bone marrow: non-specific staining and contribution by plasma cells directly reactive to alkaline phosphatase, J. Pathol. 185 (1998) 427–434. [7] E. Borgen, B. Naume, J.M. Nesland, G. Kvalheim, K. Beiske, O. Fodstad, I. Diel, Standardization of the immunocytochemical detection of cancer cells in the BM and blood: I. establishment of objective criteria for the evaluation of immunostained cells, Cytotherapy 1(5) (1999) 377–388. [8] N.K. Cheung, G. Heller, B.H. Kushner, C. Liu, I.Y. Cheung, Detection of metastatic neuroblastoma in bone marrow: when is routine marrow histology insensitive? J. Clin. Oncol. 15 (1997) 2807–2817. [9] T. Fehm, A. Sagalowsky, E. Clifford, P. Beitsch, H. Saboorian, D. Euhus, S. Meng, L. Morrison, T. Tucker, N. Lane, B.M.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
Ghadimi, K. Heselmeyer-Haddad, T. Ried, C. Rao, J. Uhr, Cytogenetic evidence that circulating epithelial cells in patients with carcinoma are malignant, Clin. Cancer Res. 8 (2002) 2073–2084. R.C. Seeger, C.P. Reynolds, R. Gallego, D.O. Stram, R.B. Gerbing, K.K. Matthay, Quantitative tumor cell content of bone marrow and blood as a predictor of outcome in stage IV neuroblastoma: a Children’s Cancer Group Study, J. Clin. Oncol. 18 (2000) 4067– 4076. P.F. Ambros, I.M. Ambros, R. Ladenstein, A. Luegmayr, S. Rumpler, H. Kovar, H. Gadner, Tumor cell clearing in the bone marrow of neuroblastoma patients-predictor for the outcome?, Med. Pediatr. Oncol. 31 (1998) 195. J.J. van Dongen, T. Seriu, E.R. Panzer-Grumayer, A. Biondi, M.J. Pongers-Willemse, L. Corral, F. Stolz, M. Schrappe, G. Masera, W.A. Kamps, H. Gadner, E.R. van Wering, W.D. Ludwig, G. Basso, M.A. de Bruijn, G. Cazzaniga, K. Hettinger, A. van der Does-van den Berg, W.C. Hop, H. Riehm, C.R. Bartram, Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood [see comments], Lancet 352 (1998) 1731– 1738. G. Mehes, A. Witt, E. Kubista, P.F. Ambros, Circulating breast cancer cells are frequently apoptotic, Am. J. Pathol. 159 (2001) 17– 20. G. Gastl, G. Spizzo, P. Obrist, M. Dunser, G. Mikuz, Ep-CAM overexpression in breast cancer as a predictor of survival, Lancet 356 (2000) 1981–1982. T.J. Moss, C.P. Reynolds, H.N. Sather, S.G. Romansky, G.D. Hammond, R.C. Seeger, Prognostic value of immunocytologic detection of bone marrow metastases in neuroblastoma, N. Engl. J. Med. 324 (1991) 219– 226. K. Pantel, M. Otte, Occult micrometastasis: enrichment, identification and characterization of single disseminated tumour cells, Semin. Cancer Biol. 11 (2001) 327–337. P.A. Fetsch, K.H. Cowan, D.E. Weng, A. Freifield, A.C. Filie, A. Abati, Detection of circulating tumor cells and micrometastases in stage II, III, and IV breast cancer patients utilizing cytology and immunocytochemistry, Diagn. Cytopathol. 22 (2000) 323–328. G. Mehes, A. Luegmayr, C.M. Hattinger, T. Lorch, I.M. Ambros, H. Gadner, P.F. Ambros, Automatic detection and genetic profiling of disseminated neuroblastoma cells, Med. Pediatr. Oncol. 36 (2001) 205–209.