Stability of PCR Targets for Monitoring Minimal Residual Disease in Neuroblastoma
2012; Elsevier BV; Volume: 14; Issue: 2 Linguagem: Inglês
10.1016/j.jmoldx.2011.12.002
ISSN1943-7811
AutoresJanine Stutterheim, Lily Zappeij‐Kannegieter, Ingrid Øra, Peter G. van Sluis, Johannes Bras, Emmy den Ouden, Rogier Versteeg, Huib N. Caron, C. Ellen van der Schoot, Godelieve A.M. Tytgat,
Tópico(s)Virus-based gene therapy research
ResumoIn neuroblastoma (NB) patients, minimal residual disease (MRD) can be detected by real-time quantitative PCR (qPCR) using NB-specific target genes, such as PHOX2B and TH. However, it is unknown whether the mRNA levels of these targets vary either during treatment or at relapse. If marker genes are not stably expressed, estimation of MRD levels in bone marrow (BM) or peripheral blood will be hampered. We studied the stability of a panel of qPCR markers in primary tumors at diagnosis compared with i) paired metastasis (n = 7), ii) treated (n = 10), and iii) relapse (n = 6) tumors. We also compared relative expression of the targets in iv) primary tumors and BM at diagnosis (n = 17), v) BM and peripheral blood at diagnosis (n = 20), vi) BM at diagnosis and during treatment (n = 26), and vii) BM from different puncture sides (n = 110). Especially at diagnosis, PCR target expression is quite stable. Accurate quantification is possible when expression level can be related to the primary tumor; however, PCR target expression can alter on treatment and at relapse. If the median value of relative expression of a panel of PCR targets is used, most variations due to treatment and outgrowth of subclones level out, allowing for reliable application and quantification of MRD-PCR targets in NB patients. In neuroblastoma (NB) patients, minimal residual disease (MRD) can be detected by real-time quantitative PCR (qPCR) using NB-specific target genes, such as PHOX2B and TH. However, it is unknown whether the mRNA levels of these targets vary either during treatment or at relapse. If marker genes are not stably expressed, estimation of MRD levels in bone marrow (BM) or peripheral blood will be hampered. We studied the stability of a panel of qPCR markers in primary tumors at diagnosis compared with i) paired metastasis (n = 7), ii) treated (n = 10), and iii) relapse (n = 6) tumors. We also compared relative expression of the targets in iv) primary tumors and BM at diagnosis (n = 17), v) BM and peripheral blood at diagnosis (n = 20), vi) BM at diagnosis and during treatment (n = 26), and vii) BM from different puncture sides (n = 110). Especially at diagnosis, PCR target expression is quite stable. Accurate quantification is possible when expression level can be related to the primary tumor; however, PCR target expression can alter on treatment and at relapse. If the median value of relative expression of a panel of PCR targets is used, most variations due to treatment and outgrowth of subclones level out, allowing for reliable application and quantification of MRD-PCR targets in NB patients. Given that bone marrow (BM) invasion in children with neuroblastoma (NB) is present in 50% of these patients at diagnosis,1Cohn S.L. Pearson A.D. London W.B. Monclair T. Ambros P.F. Brodeur G.M. Faldum A. Hero B. Iehara T. Machin D. Mosseri V. Simon T. Garaventa A. Castel V. Matthay K.K. INRG Task ForceThe International Neuroblastoma Risk Group (INRG) classification system: an INRG Task Force report.J Clin Oncol. 2009; 27: 289-297Crossref PubMed Scopus (1232) Google Scholar many studies have been performed to define the best method for evaluating tumor load in BM.2Cheung I.Y. Feng Y. Gerald W. Cheung N.K. Exploiting gene expression profiling to identify novel minimal residual disease markers of neuroblastoma.Clin Cancer Res. 2008; 14: 7020-7027Crossref PubMed Scopus (58) Google Scholar, 3Mehes G. Luegmayr A. Hattinger C.M. Lörch T. Ambros I.M. Gadner H. Ambros P.F. Automatic detection and genetic profiling of disseminated neuroblastoma cells.Med Pediatr Oncol. 2001; 36: 205-209Crossref PubMed Scopus (15) Google Scholar, 4Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Yalcin B. Dee R. van Noesel M.M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. Detecting minimal residual disease in neuroblastoma: the superiority of a panel of real-time quantitative PCR markers.Clin Chem. 2009; 55: 1316-1326Crossref PubMed Scopus (60) Google Scholar, 5Swerts K. De M.B. Dhooge C. Brichard B. Benoit Y. Laureys G. Philippé J. Detection of residual neuroblastoma cells in bone marrow: comparison of flow cytometry with immunocytochemistry.Cytometry B Clin Cytom. 2004; 61: 9-19Crossref PubMed Scopus (32) Google Scholar, 6Swerts K. Ambros P.F. Brouzes C. Navarro J.M. Gross N. Rampling D. Schumacher-Kuckelkorn R. Sementa A.R. Ladenstein R. Beiske K. Standardization of the immunocytochemical detection of neuroblastoma cells in bone marrow.J Histochem Cytochem. 2005; 53: 1433-1440Crossref PubMed Scopus (41) Google Scholar, 7Viprey V.F. Lastowska M.A. Corrias M.V. Swerts K. Jackson M.S. Burchill S.A. Minimal disease monitoring by QRT-PCR: guidelines for identification and systematic validation of molecular markers prior to evaluation in prospective clinical trials.J Pathol. 2008; 216: 245-252Crossref PubMed Scopus (40) Google Scholar Although in current protocols BM disease is determined by morphology, complemented by immunohistochemical techniques,8Monclair T. Brodeur G.M. Ambros P.F. Brisse H.J. Cecchetto G. Holmes K. Kaneko M. London W.B. Matthay K.K. Nuchtern J.G. von S.D. Simon T. Cohn S.L. Pearson A.D. INRG Task ForceThe International Neuroblastoma Risk Group (INRG) staging system: an INRG Task Force report.J Clin Oncol. 2009; 27: 298-303Crossref PubMed Scopus (664) Google Scholar detection of tumor cells by real-time quantitative PCR (qPCR) has been shown to be more sensitive than conventional cytology, detecting tumor cells with a sensitivity of 1 in 106 normal bone marrow cells.9Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Kleijn I. Dee R. Hooft L. van Noesel M.M. Bierings M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. PHOX2B is a novel and specific marker for minimal residual disease testing in neuroblastoma.J Clin Oncol. 2008; 26: 5443-5449Crossref PubMed Scopus (74) Google Scholar, 10Viprey V.F. Corrias M.V. Kagedal B. Oltra S. Swerts K. Vicha A. Ladenstein R. Burchill S.A. Standardisation of operating procedures for the detection of minimal disease by QRT-PCR in children with neuroblastoma: quality assurance on behalf of SIOPEN-R-NET.Eur J Cancer. 2007; 43: 341-350Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar The most commonly used marker to detect minimal residual disease (MRD) using qPCR is tyrosine hydroxylase (TH).11Burchill S.A. Bradbury F.M. Smith B. Lewis I.J. Selby P. Neuroblastoma cell detection by reverse transcriptase-polymerase chain reaction (RT-PCR) for tyrosine hydroxylase mRNA.Int J Cancer. 1994; 57: 671-675Crossref PubMed Scopus (110) Google Scholar, 12Corrias M.V. Faulkner L.B. Pistorio A. Rosanda C. Callea F. Piccolo M.S. Scaruffi P. Marchi C. Lacitignola L. Occhino M. Gambini C. Tonini G.P. Haupt R. De Bernardi B. Pistoia V. Garaventa A. Detection of neuroblastoma cells in bone marrow and peripheral blood by different techniques: accuracy and relationship with clinical features of patients.Clin Cancer Res. 2004; 10: 7978-7985Crossref PubMed Scopus (34) Google Scholar, 13Miyajima Y. Kato K. Numata S. Kudo K. Horibe K. Detection of neuroblastoma cells in bone marrow and peripheral blood at diagnosis by the reverse transcriptase-polymerase chain reaction for tyrosine hydroxylase mRNA.Cancer. 1995; 75: 2757-2761Crossref PubMed Scopus (110) Google Scholar, 14Oltra S. Martinez F. Orellana C. Grau E. Fernandez J.M. Canete A. Castel V. The doublecortin gene, a new molecular marker to detect minimal residual disease in neuroblastoma.Diagn Mol Pathol. 2005; 14: 53-57Crossref PubMed Scopus (35) Google Scholar, 15Swerts K. De Moerloose B. Dhooge C. Vandesompele J. Hoyoux C. Beiske K. Benoit Y. Laureys G. Philippé J. Potential application of ELAVL4 real-time quantitative reverse transcription-PCR for detection of disseminated neuroblastoma cells.Clin Chem. 2006; 52: 438-445Crossref PubMed Scopus (24) Google Scholar, 16Tchirkov A. Paillard C. Halle P. Bernard F. Bordigoni P. Vago P. Deméocq F. Kanold J. Significance of molecular quantification of minimal residual disease in metastatic neuroblastoma.J Hematother Stem Cell Res. 2003; 12: 435-442Crossref PubMed Scopus (32) Google Scholar, 17Träger C. Kogner P. Lindskog M. Ponthan F. Kullman A. Kågedal B. Quantitative analysis of tyrosine hydroxylase mRNA for sensitive detection of neuroblastoma cells in blood and bone marrow.Clin Chem. 2003; 49: 104-112Crossref PubMed Scopus (48) Google Scholar However, the use of TH as an MRD marker in NB is hampered by its illegitimate expression in normal hematological cells.4Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Yalcin B. Dee R. van Noesel M.M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. Detecting minimal residual disease in neuroblastoma: the superiority of a panel of real-time quantitative PCR markers.Clin Chem. 2009; 55: 1316-1326Crossref PubMed Scopus (60) Google Scholar, 10Viprey V.F. Corrias M.V. Kagedal B. Oltra S. Swerts K. Vicha A. Ladenstein R. Burchill S.A. Standardisation of operating procedures for the detection of minimal disease by QRT-PCR in children with neuroblastoma: quality assurance on behalf of SIOPEN-R-NET.Eur J Cancer. 2007; 43: 341-350Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar, 15Swerts K. De Moerloose B. Dhooge C. Vandesompele J. Hoyoux C. Beiske K. Benoit Y. Laureys G. Philippé J. Potential application of ELAVL4 real-time quantitative reverse transcription-PCR for detection of disseminated neuroblastoma cells.Clin Chem. 2006; 52: 438-445Crossref PubMed Scopus (24) Google Scholar, 17Träger C. Kogner P. Lindskog M. Ponthan F. Kullman A. Kågedal B. Quantitative analysis of tyrosine hydroxylase mRNA for sensitive detection of neuroblastoma cells in blood and bone marrow.Clin Chem. 2003; 49: 104-112Crossref PubMed Scopus (48) Google Scholar, 18Cheung I.Y. Cheung N.K. Quantitation of marrow disease in neuroblastoma by real-time reverse transcription-PCR.Clin Cancer Res. 2001; 7: 1698-1705PubMed Google Scholar, 19Hoon D.S. Kuo C.T. Wen S. Wang H. Metelitsa L. Reynolds C.P. Seeger R.C. Ganglioside GM2/GD2 synthetase mRNA is a marker for detection of infrequent neuroblastoma cells in bone marrow.Am J Pathol. 2001; 159: 493-500Abstract Full Text Full Text PDF PubMed Scopus (36) Google Scholar, 20Ifversen M.R. Kågedal B. Christensen L.D. Rechnitzer C. Petersen B.L. Heilmann C. Comparison of immunocytochemistry, real-time quantitative RT-PCR and flow cytometry for detection of minimal residual disease in neuroblastoma.Int J Oncol. 2005; 27: 121-129PubMed Google Scholar, 21Kuçi Z. Seitz G. Kuçi S. Kreyenberg H. Schumm M. Lang P. Niethammer D. Handgretinger R. Bruchelt G. Pitfalls in detection of contaminating neuroblastoma cells by tyrosine hydroxylase RT-PCR due to catecholamine-producing hematopoietic cells.Anticancer Res. 2006; 26: 2075-2080PubMed Google Scholar, 22Lambooy L.H. Gidding C.E. van den Heuvel L.P. Hulsbergen-van de Kaa C.A. Ligtenberg M. Bökkerink J.P. De Abreu R.A. Real-time analysis of tyrosine hydroxylase gene expression: a sensitive and semiquantitative marker for minimal residual disease detection of neuroblastoma.Clin Cancer Res. 2003; 9: 812-819PubMed Google Scholar We have previously identified paired-like homeobox2b (PHOX2B) as a neuroblastoma-specific marker,9Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Kleijn I. Dee R. Hooft L. van Noesel M.M. Bierings M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. PHOX2B is a novel and specific marker for minimal residual disease testing in neuroblastoma.J Clin Oncol. 2008; 26: 5443-5449Crossref PubMed Scopus (74) Google Scholar which has been confirmed by others.2Cheung I.Y. Feng Y. Gerald W. Cheung N.K. Exploiting gene expression profiling to identify novel minimal residual disease markers of neuroblastoma.Clin Cancer Res. 2008; 14: 7020-7027Crossref PubMed Scopus (58) Google Scholar, 7Viprey V.F. Lastowska M.A. Corrias M.V. Swerts K. Jackson M.S. Burchill S.A. Minimal disease monitoring by QRT-PCR: guidelines for identification and systematic validation of molecular markers prior to evaluation in prospective clinical trials.J Pathol. 2008; 216: 245-252Crossref PubMed Scopus (40) Google Scholar Because the expression level of the various markers, including PHOX2B, is heterogeneous in primary NB, with >3 log difference between tumors, we subsequently reported that a panel of markers should be used for detection of MRD.4Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Yalcin B. Dee R. van Noesel M.M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. Detecting minimal residual disease in neuroblastoma: the superiority of a panel of real-time quantitative PCR markers.Clin Chem. 2009; 55: 1316-1326Crossref PubMed Scopus (60) Google Scholar It is unknown whether marker genes are up- or down-regulated during treatment or at relapse, and between the primary tumor and its metastasis. For example, it is generally believed that tumors treated with anti-GD2 immunotherapy can down-regulate the enzyme GD2 synthase (GD2S) as an escape mechanism. Similarly, it has been suggested that tumors treated with metaiodobenzylguanidine (MIBG) might down-regulate genes involved in its metabolic pathway,2Cheung I.Y. Feng Y. Gerald W. Cheung N.K. Exploiting gene expression profiling to identify novel minimal residual disease markers of neuroblastoma.Clin Cancer Res. 2008; 14: 7020-7027Crossref PubMed Scopus (58) Google Scholar where TH is a critical step. Down-regulation of marker genes will lead to an underestimation of tumor load and risk for false negative findings in the BM. Although the phenomenon of changes in marker gene expression is widely assumed, it has never been studied directly. We therefore investigated whether the expression of a panel of qPCR markers would remain stable in the process of metastasis [liver, lymph node (LN), and BM] during treatment and at relapse. This information is essential for a reliable selection of MRD-PCR targets, to prevent false negative results, as well as for accurate quantification of MRD level. Neuroblastoma cell line IMR-32 (ATCC, Manassas, VA) was maintained in Dulbecco's modified Eagle's medium, supplemented with 10% fetal calf serum, 4 mmol/L l-glutamine, 10 U/mL penicillin, and 10 μg/mL streptomycin. For isolating IMR-32 clones, cells were cultured in serial dilution in 96-well plates. Monoclonal outgrowth was observed at a concentration of 10 cells/well. Cells were harvested in 1 mL TRIzol reagent (Invitrogen, Carlsbad, CA). The following paired samples were selected from our tumor bank23Koppen A. Ait-Aissa R. Koster J. Øra I. Bras J. van Sluis P.G. Caron H. Versteeg R. Valentijn L.J. Dickkopf-3 expression is a marker for neuroblastic tumor maturation and is down-regulated by MYCN.Int J Cancer. 2008; 122: 1455-1464Crossref PubMed Scopus (29) Google Scholar: i) primary tumors and metastasis tumors (LN or liver) at diagnosis (n = 7), ii) tumors at diagnosis and tumors during treatment (at surgery) (n = 10; all patients were treated with two or three cycles of MIBG therapy and 6 of the 10 also received induction therapy), or iii) tumors at diagnosis and tumors at relapse (n = 6). In addition, paired diagnostic tumor and BM samples at diagnosis were collected (n = 17). In all, 51 different tumor samples, collected between 1991 and 2004, were selected. Diagnostic tumor samples were immediately frozen in liquid nitrogen and stored at −80°C. Sections of frozen samples were cut and stained with H&E before RNA extraction, to collect areas containing only tumor cells. The morphological classification of the H&E sections was reviewed by an experienced pathologist (J.B.) according to International Neuroblastoma Pathology Committee (INPC) criteria.24Shimada H. Ambros I.M. Dehner L.P. Hata J. Joshi V.V. Roald B. Terminology and morphologic criteria of neuroblastic tumors: recommendations by the International Neuroblastoma Pathology Committee.Cancer. 1999; 86: 349-363Crossref PubMed Scopus (550) Google Scholar The percentage of neuroblasts, ganglion cells, and Schwann cells was determined. Only samples with >60% neuroblasts were included. A total of 256 BM samples (104 at diagnosis and 152 during treatment) and 20 PB samples, collected between 1991 and 2009 from 74 patients at stage 4, were tested. Because target expression was studied, only samples that had previously been shown to be MRD-positive by qPCR4Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Yalcin B. Dee R. van Noesel M.M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. Detecting minimal residual disease in neuroblastoma: the superiority of a panel of real-time quantitative PCR markers.Clin Chem. 2009; 55: 1316-1326Crossref PubMed Scopus (60) Google Scholar were selected. The following paired samples were tested (see Supplemental Figure S1 at http://jmd.amjpathol.org): i) primary tumors at diagnosis and BM at diagnosis (n = 17), ii) BM at diagnosis and PB at diagnosis (n = 20), and iii) BM samples at diagnosis and BM during treatment (n = 26). Also, 110 bilateral BM pairs were collected and tested by qPCR and cytology. In samples that were positive by qPCR within the quantitative range (56 of the 110 BM pairs), expression levels between different sides were also compared. Written informed consent was obtained from parents or guardians for each participating child. The study was approved by the Medical Research Ethics Committee of the Academic Medical Center. For BM and PB samples, cells were isolated by NH4Cl lysis within 24 hours after collection in EDTA tubes. Total cellular RNA was extracted using TRIzol reagent (Invitrogen) according to the manufacturer's instructions. cDNA was synthesized as described previously.9Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Kleijn I. Dee R. Hooft L. van Noesel M.M. Bierings M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. PHOX2B is a novel and specific marker for minimal residual disease testing in neuroblastoma.J Clin Oncol. 2008; 26: 5443-5449Crossref PubMed Scopus (74) Google Scholar qPCR was performed in an ABI PRISM 7900 sequence detection system (PE Biosystems, Darmstadt, Germany) for six marker genes: PHOX2B; TH; dopa decarboxylase (aromatic l-amino acid decarboxylase) (DDC); cholinergic receptor, nicotinic, alpha 3 (CHRNA3); dopamine beta-hydroxylase (dopamine beta-monooxygenase) (DBH); and growth associated protein 43 (GAP43). For BM samples, the PCR markers PHOX2B, TH, DDC, CHRNA3, and GAP43 were tested, and for PB samples the PCR markers PHOX2B, TH, DDC, CHRNA3, and DBH were tested, as described previously.4Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Yalcin B. Dee R. van Noesel M.M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. Detecting minimal residual disease in neuroblastoma: the superiority of a panel of real-time quantitative PCR markers.Clin Chem. 2009; 55: 1316-1326Crossref PubMed Scopus (60) Google Scholar Reference gene β-glucuronidase (GUSB) was used for normalization [normalized CT (ΔCT) = CT GUSB − CT marker]. All qPCR experiments were performed in triplicate, and mean values were used. For the analyses, only MRD-positive PB and BM samples were used; stringent criteria for MRD positivity were applied, as described previously.4Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Yalcin B. Dee R. van Noesel M.M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. Detecting minimal residual disease in neuroblastoma: the superiority of a panel of real-time quantitative PCR markers.Clin Chem. 2009; 55: 1316-1326Crossref PubMed Scopus (60) Google Scholar, 9Stutterheim J. Gerritsen A. Zappeij-Kannegieter L. Kleijn I. Dee R. Hooft L. van Noesel M.M. Bierings M. Berthold F. Versteeg R. Caron H.N. van der Schoot C.E. Tytgat G.A. PHOX2B is a novel and specific marker for minimal residual disease testing in neuroblastoma.J Clin Oncol. 2008; 26: 5443-5449Crossref PubMed Scopus (74) Google Scholar Differences in target expression between samples were calculated as relative value (representing tumor cell infiltration), in which the marker expression in the primary tumor or BM obtained at diagnosis was set at 100%, using the following formula: 2−ΔΔCT (ΔCT primary tumor/BM − ΔCT paired sample) × 100. Stable expression was defined as <10-fold difference from 100%, which agrees with a range of 3 CT (ie, ±1.5 CT around CT of 100%), which can be considered as within the limits of normal variation.25van der Velden V.H. Cazzaniga G. Schrauder A. Hancock J. Bader P. Panzer-Grumayer E.R. Flohr T. Sutton R. Cave H. Madsen H.O. Cayuela J.M. Trka J. Eckert C. Foroni L. Zur Stadt U. Beldjord K. Raff T. van der Schoot C.E. van Dongen J.J. European Study Group on MRD detection in ALL (ESG-MRD-ALL)Analysis of minimal residual disease by Ig/TCR gene rearrangements: guidelines for interpretation of real-time quantitative PCR data.Leukemia. 2007; 21: 604-611PubMed Google Scholar For BM samples, the median relative expression was determined by taking the median of the relative values of all PCR targets and setting this at 1. Stable expression was arbitrarily defined as <10-fold difference from 1. For comparison of bilateral BM samples, the 2−ΔΔCT method was used, in which the sample with the lowest expression was set at 1. Stable expression was arbitrarily defined as <10-fold difference from 1. The paired Wilcoxon signed rank test was used to compare normalized expression levels (ΔCT) of paired tumor samples for each PCR target. The Friedman test was used to detect differences in relative normalized expression (ΔΔCT) between markers within patients. When the Friedman test was significant, the paired Wilcoxon signed rank test was used to compare the ΔΔCT for each PCR target mutually. Correction for multiple comparisons was performed by multiplying the P values of the Wilcoxon test by the number of markers tested. Statistical analyses were performed using SPSS version 15.0 software (SPSS, Chicago, IL). P values of <0.05 were considered statistically significant. To mimic the variability of different tumor cells within a tumor, we generated IMR-32 subclones via serial dilution. The normalized expression levels of the six PCR targets for 19 IMR-32 clones were determined (Figure 1). Expression of CHRNA3 and PHOX2B was within a range of 3 CT around the median CT. Also, DDC and DBH expression in the IMR clones was fairly stable (range, 10 CT). To investigate whether the expression of our MRD-PCR targets was stable during the process of metastasis, we analyzed their expression in primary tumors and in paired liver or LN metastasis (n = 7) samples taken at diagnosis. The morphological classification was identical within all pairs (Table 1). We compared the relative expression of the different markers in each metastasis with that of the primary tumor at diagnosis, which was set at 100% (Figure 2A). For PHOX2B and TH, two metastases had a slightly lower (<10-fold) expression, and for GAP43 one metastasis had a marginally higher expression, compared with the primary tumor. Overall, all MRD-PCR targets showed similar relative expression in the metastasis, around 100% of the primary tumor (Wilcoxon rank test; nonsignificant P value for all markers).Table 1Patient and Tumor Characteristics of All Tumor PairsPatient IDStageMYCN amplificationMaterialPathologyInduction therapyMetastasis tumor versus primary tumor at diagnosis824NoATPDNNoneLNPDNNone953NoATPDNNoneLNPDNNone1594YesATPDNNoneLNPDNNone1662NoATPDNNoneLNPDNNone1674sYesATUNNoneLiverUNNone2254sNoATPDNNoneLiverPDNNone5094sNoATPDNNoneLiverPDNNoneTreated tumor versus primary tumor at diagnosis4064NoATPDNNoneATPDN2× MIBG4173NoATDNNoneATGNB intermixed1× MIBG4304YesATPDNNoneATDN3× MIBG + 3× VECI4584YesATPDNNoneATPDN3× MIBG + 2× VECI5194YesATPDNNoneATPDN2× MIBG + Topo5274NoATPDNNoneATDN2× MIBG5404YesATPDNNoneATPDN2× MIBG + Topo + 4× VECI5463NoATPDNNoneATDN2× MIBG6004NoATPDNNoneATPDN2× MIBG + 8/8 rCOJEC6574YesATUNNoneATPDNRCOJECRelapse tumor versus tumor at diagnosis984sNoLiverPDNNone4MaxillaPDNNone1304YesATPDNNone4ATPDNNone1662NoLNPDNNone4MaxillaPDNNone4624NoLNUNNone4LNUNNone5714NoATPDNNone4ATPDNNone5754sNoLiverUNNone4LNUNNoneAT, abdominal; DN, differentiating neuroblastoma; GNB, ganglioneuroblastoma; LN, lymph node; PDN, poorly differentiated neuroblastoma; UN, undifferentiated neuroblastoma. Open table in a new tab AT, abdominal; DN, differentiating neuroblastoma; GNB, ganglioneuroblastoma; LN, lymph node; PDN, poorly differentiated neuroblastoma; UN, undifferentiated neuroblastoma. To determine whether treatment induces changes in expression of the PCR targets, primary tumors taken at diagnosis were compared with primary tumor at surgery after treatment (n = 10). The morphological classification remained the same during treatment in 5 of the 10 tumor pairs and became more differentiated in the other 5 tumor pairs (Table 1). All MRD-PCR targets showed a change in expression levels in several tumors after induction treatment (Figure 2B). In only 2 of the 10 tumors (N417 and N527) did all markers remain stable. For CHRNA3, the median expression level was higher (more than threefold) in the treated tumor (Wilcoxon rank test; P = 0.05). For the other targets, the change in expression level was inconsistent, both up-regulation and down-regulation were observed (see Supplemental Table S1 at http://jmd.amjpathol.org). There was no correlation between treatment (chemotherapy versus no chemotherapy) or tumor differentiation during therapy (differentiation versus no differentiation) and up- or down-regulation of PCR targets (P = 0.11 and P = 0.72, respectively; χ2 test) (see Supplemental Tables S2 and S3, respectively, at http://jmd.amjpathol.org). To determine whether the expression of our MRD-PCR targets was stable at relapse, we compared primary and relapse tumors (n = 6). In all patients, the morphological classification of the tumor at relapse was the same as at diagnosis (Table 1). All MRD-PCR targets showed either much higher or much lower expression in the relapse tumor (Figure 2C), but there was no consistent change in expression (Wilcoxon rank test; nonsignificant P value for all markers). Because the expression of the MRD-PCR targets in liver and LN metastasis at diagnosis was similar to expression in the primary tumor, we investigated whether this was true also for BM metastases. Because tumor cell infiltration in BM varies greatly among patients, we calculated for each target the relative expression in the BM compared with the primary tumor (n = 17) and calculated the median expression level of five targets, which was set at 1. Theoretically, if the expression of PCR targets is stable, all markers will have expression levels similar to the median (ie, around 1). The relative expression levels of the PCR targets were within 1 log of the median expression level (∼10-fold difference) (Figure 3A). Only the expression levels of TH varied more; these were consistently lower in diagnosis BM, compared with the other markers [P = 0.016, Friedman test; P = 0.05 (GAP43), P = 0.05 (CHRNA3), P = 0.09 (DDC), and P = 0.1 (PHOX2B), Wilcoxon rank test] (see Supplemental Figure S2A at http://jmd.amjpathol.org). In 7 of the 17 cases, the relative expression was more than threefold lower, compared with the median relative expression. This was independent of the level of tumor cell infiltration in the BM (Figure 3A). For the other markers, only 1/17 BM samples (CHRNA3 and GAP43) and 2/17 BM samples (PHOX2B and DDC) showed relative expression levels more than threefold lower than the median relative expression. We also compared the expression of PCR targets that can be used for MRD detection in both BM and PB (PHOX2B, TH, DDC, and CHRNA3) in the BM and PB diagnosis samples (n = 20). The relative expression levels in PB were similar (within 1 log) for all four targets (Figure 3B). Only DDC showed a significantly lower expression in the PB samples, compared with the other targets (P < 0.01, Friedman test; P < 0.01 for DDC versus the other three targets, Wilcoxon rank test), and the relative expression was more than threefold lower compared with the median relative expression in 7/20 samples (see Supplemental Figure S2B at http://jmd.amjpathol.org). This was independent of the tumor cell infiltration in PB. Although the expression of PCR targets seems similar in circulating tumor cells in BM and PB, it is clear from this comparative analysis that estimates of tumor load in PB and BM will not be accurate if only a single marker is used. Because MRD-PCR targets in primary tumors can be up- or down-regulated on treatment, we investigated whether this was also true for tumor cells in BM. The expression of the PCR targets in MRD-positive BM samples taken during treatment was compared with their expression in BM samples taken at diagnosis, which was set at 100% (n = 26). As expected, in most patients there was a reduction in tumor cell infiltration in the BM on treatment (data not shown). When relative expression (a proxy measure for tumor cell infiltration) in the BM taken during treatment was >0.1% of the diagnosis BM, the expression was within 1 log for most but not all PCR targets (Figure 3C). In one patient (N708), the relative expression differed by more than a factor of 10, and in five patients (N540, N604, N652, N702, and N708) the BM taken during treatment was negative for one or two PCR targets but the other PCR targets were positive (>0.1%). Remarkably, DDC was negative in all five of these samples; however DDC expression was not significantly lower (Friedman test; P = 0.32) (see Supplemental Figure S2C at http://jmd.amjpathol.org). Discrepancies in MRD positivity were more pronounced when tumor load in BM during treatment was low, compared with the diagnosis BM sample. In 10 of the 12 patients for whom all PCR markers showed a relative quantity of 10-fold. compared with the other side. In none of the BM pairs obtained during follow-up was there >10-fold difference in expression for all PCR targets, illustrating that during treatment (when BM infiltration is often lower than at diagnosis) infiltration is similar for both sides. In only 6 of the 56 BM pairs was there a >10-fold difference in relative expression for one of the PCR targets. This observation shows that if samples are taken at the same time point, there is hardly any variation in tumor load quantification between markers. For reliable PCR-based detection and quantification of MRD in NB, it is important to know the stability of expression of PCR target genes over the clinical course of the disease. Our comparative analysis of different pairs of NB tumor and BM samples shows that target expression in general is stable, although it can alter on treatment and at relapse. Our in vitro data, for which we measured MRD PCR targets in IMR-32 subclones, indeed suggest that the expression of PCR target genes can be highly heterogeneous, with >1000-fold difference in expression levels between subclones for the relatively low-expressed genes TH and GAP43. This indicates that some tumor cells have a higher or lower expression of certain targets than other tumor cells within the same tumor. One hypothesis is that only certain subclones within a tumor are able to metastasize to distant sites. Pietras et al26Pietras A. Gisselsson D. Ora I. Noguera R. Beckman S. Navarro S. Påhlman S. High levels of HIF-2alpha highlight an immature neural crest-like NB cell cohort located in a perivascular niche.J Pathol. 2008; 214: 482-488Crossref PubMed Scopus (97) Google Scholar identified a small subset of immature tumor cells within NB specimens that expressed neural crest-associated genes but lacked expression of sympathetic nervous system markers, such as TH. If tumor cells in the BM are even in part derived from this subset, then these cells could be missed when using TH for MRD detection. To address this variability in human tumors, we first investigated whether NB cells obtained from two localizations at the same time point differ with respect to expression levels of PCR target genes. It appeared that expression in liver and LN metastases was similar to their primary tumor, suggesting that the PCR target genes are stably expressed in primary tumor and its metastasis at diagnosis. In comparison of primary tumor and BM metastasis at diagnosis, however, TH expression was consistently lower in BM tumor cells, relative to expression of the other markers, suggesting that this gene is down-regulated in BM metastasis. In other paired samples taken at the same time point (BM and PB, or bilateral BM samples), expression levels of the target genes were stable in most cases, with the relative expression between the PCR targets remaining the same. Nonetheless, we observed that markers, especially TH, occasionally showed a variable expression between tumor and metastasis, so we conclude that for accurate estimation of tumor load at diagnosis at least two PCR targets should be used, preferably including the target with the highest expression in the tumor. Because the level of tumor cell infiltration in bilateral BM samples can differ, both examinations should be done for correct quantification of BM infiltration and to avoid false negative results. Next, we investigated whether the expression of the PCR targets remained stable in samples drawn at different time points of the disease (ie, during treatment and at relapse). If treatment leads to decrease or loss of expression of a PCR target in a viable cell, it would consequently lead to underestimation of tumor load or to false negative results. Although PCR targets in many treated tumors showed the same expression as in the diagnostic tumor, we also observed either up- or down-regulation of all targets in several treated and relapse tumors, or up- or down-regulation of only one or two of the five PCR targets. Although most statistical tests were nonsignificant, probably because of the small sample size tested, these up- or down-regulations could be the result of the therapy received (for example, by mechanisms that influence metabolic pathways or by differentiation of NB cells during treatment). These results emphasize once more that a panel of PCR targets is needed to overcome tumor heterogeneity. Cheung et al2Cheung I.Y. Feng Y. Gerald W. Cheung N.K. Exploiting gene expression profiling to identify novel minimal residual disease markers of neuroblastoma.Clin Cancer Res. 2008; 14: 7020-7027Crossref PubMed Scopus (58) Google Scholar previously hypothesized that MIBG therapy might cause down-regulation of genes in the metabolic catecholamine pathway, such as DDC, DBH, and TH. In our cohort, all patients were treated with MIBG therapy before induction chemotherapy, but consistent down-regulation of these genes was not observed. Furthermore, because PHOX2B is necessary for differentiation of the sympathetic nervous system, treatment that influences differentiation might result in altered expression of PHOX2B and of its target genes TH, DDC, and DBH.27Pattyn A. Goridis C. Brunet J.F. Specification of the central noradrenergic phenotype by the homeobox gene Phox2b.Mol Cell Neurosci. 2000; 15: 235-243Crossref PubMed Scopus (180) Google Scholar TH, DDC, and DBH indeed followed the same expression pattern on treatment in most tumors, but PHOX2B did not (Supplemental Table S1 at http://jmd.amjpathol.org). Nonetheless, up- or down-regulation of PCR targets did not seem to be correlated with tumor differentiation, whether or not induced by treatment. When expression levels in BM samples taken during treatment were compared with those at diagnosis, the variation between PCR targets was much higher than in BM samples taken at the same time point. Consequently, MRD quantification using the ΔΔCT method will result in different MRD levels for each PCR target. The expression level of a marker can change in all directions on treatment (ie, it may be higher, lower, or stable), and there is no indication that one of the five genes is more stable than the other four. Because the various markers change differently within a single patient sample, we calculated the median expression, to level out the variation. We assumed that using the median value would be a better reflection of the tumor load than the average expression, because average expression is more influenced by possible outliers. The present study is hampered by lack of a gold standard for molecular quantification of NB cells. The results comparing different tumors could therefore be biased by the purity of the NB biopsy. To minimize this variation, only tumor samples with >60% neuroblastic cells were included. Differences in target expression due to the purity of the tumor samples could therefore count for maximally twofold (∼1 CT). To test for the stability of PCR targets of BM infiltrated NB, we used the relative differences between PCR targets within a single patient. One could argue that the variations we observed are caused by variable efficiencies between PCR assays; however, only minimal variations between PCR targets were observed between pairs of bilateral BM samples, indicating that the PCR assays are reproducible and that therefore the variations observed in samples taken at different time points are valid. In summary, when the primary tumor is available, BM infiltration at diagnosis can be reliably quantified in most patients. Variation between expression of PCR targets is greater during treatment or at relapse, making relative quantification more difficult than at diagnosis. To be able to estimate the MRD level of infiltration, we propose using the median value of relative expression of a panel of PCR targets, because most variations due to treatment and selection of subclones will be leveled out this way, allowing more reliable longitudinal MRD monitoring in BM. Download .ppt (.02 MB) Help with ppt files Supplemental Figure S1Sample origin. BM and PB samples were collected for comparison of i) primary tumors at diagnosis and BM at diagnosis (n = 17), ii) BM at diagnosis and PB at diagnosis (n = 20), iii), BM samples at diagnosis and BM during treatment (n = 26), and iv) for comparison of qPCR with cytology in bilateral (left and right) BM samples (n = 110). Some BM samples were used for more than one analysis; thus, there was overlap between sample pools (n = 53). BM, bone marrow; PB, peripheral blood; Dx, sample taken at diagnosis; FUP, sample taken at follow up. Download .ppt (.14 MB) Help with ppt files Supplemental Figure S2Relative expression of BM samples compared to other tissues per marker. PCR targets were tested in different sample pairs: A: Bone marrow (BM) samples at diagnosis and primary tumors at diagnosis (n = 17). B: Peripheral blood (PB) samples at diagnosis and BM samples at diagnosis (n = 20). C: BM taken during treatment samples and BM samples at diagnosis (n = 26). Relative expression was calculated using the formula 2^δδCt*100 (δδCt = δCt paired sample - δCt (primary) tumor at diagnosis). PCR targets were plotted relative to the median relative expression (calculated as described under Materials and Methods) of all targets, in which the median relative expression was set at 1. Symbols represent sample pairs of individual patients. Patients are identified anonymously as N + number. Download .doc (.04 MB) Help with doc files Supplemental Table S1 Download .doc (.03 MB) Help with doc files Supplemental Table S2 Download .doc (.03 MB) Help with doc files Supplemental Table S3
Referência(s)