Multigene reverse transcription-PCR profiling of circulating tumor cells in hormone-refractory prostate cancer.
In this study we expand the analytical capabilities of immunomagnetically enriched CTC enumeration to include molecular biological characterization of CTCs. Although immunomagnetic enrichment of CTCs can provide a four to five log reduction in leukocytes, the typical range for CTC-to-leukocyte ratios is 1-10 CTCs/[10.sup.3]-[10.sup.4] leukocytes. The low number of CTCs and the leukocyte carryover during the CTC enrichment process pose significant detection restrictions in terms of the signal-to-noise ratio, which constrains the choice of genes and gene expression profiling methods. The low CTC number thus requires preamplification of the entire mRNA library to analyze a multigene reverse transcription-PCR (RT-PCR) panel without compromising the sensitivities of individual marker genes. Here we show that detection of the RNA encoded by numerous genes in CTCs is feasible. The method entails constructing and synthesizing an antisense RNA (aRNA) library from mRNA extracted from immunomagnetically enriched CTCs, which permits the analysis of numerous genes through RT-PCR. We investigated the expression of 37 genes, including the prostate-tissue-specific genes prostate-specific antigen (PSA), prostate-specific membrane antigen (PSMA), and human glandular kallikrein 2 (HK2). Also included were prostate-cancer-specific genes such as hepsin (HPN) and prostate-specific gene with homology to G protein receptor (PSGR), as well as therapy-relevant genes such as androgen receptor (AR), HER-2, BCL-2, multidrug resistance gene (MDR1), and epidermal growth factor receptor (EGFR). The ability to quantify, phenotype, morphologically image, and now profile the gene expression of CTCs allows improved biological characterization of HRPC in real time and may help expedite the development of effective patient-specific therapies.
Materials and Methods
Blood from 9 patients with prostate cancer (CaP) and 13 healthy volunteers (7 males and 6 females) was drawn into two 10-mL EDTA-containing Vacutainer Tubes (Becton Dickinson) and pooled. Of the nine patients, eight had HRPC with a sustained increase in PSA and/or a positive bone scan. The single hormone-sensitive patient initially presented with a positive bone scan, but he subsequently had a sustained complete response to androgen ablation with a negative bone scan and a PSA of 0.1 [micro]g/L. The male controls were 27-73 years of age (mean, 45 years), and the female controls were 27-61 years of age (mean, 39 years). The patients ranged in age from 60 to 81 years (mean, 74 years). The treatments, bone scan status, PSA concentrations, and CTC counts are given in Table 1. Of the nine patients, two had five longitudinal blood samples, one had four samples, three had two samples, and three had single samples. All participants signed an institutional informed consent form.
The breast cancer cell line SKBR-3 and prostate cancer cell line LNCaP were cultured in RPMI-1640 supplemented with 100 mL/L fetal calf serum.
IMMUNOMAGNETIC SAMPLE PREPARATION
Immunomagnetic enrichment of epithelial cells has been described in detail previously (10, 11). Blood samples were divided into two 7.5-mL aliquots processed within 3 h. Briefly, the enrichment step involves adding magnetic nanoparticles coated with a monoclonal antibody with affinity for epithelial cell adhesion molecule (EpCAM) to a 7.5-mL blood sample, incubating the sample, and placing the tube inside a magnetic separator for immunomagnetic separation of the magnetically labeled CTCs. After separation, the enriched CTCs from one tube are used for RNA extraction, and the second tube is used for CTC enumeration.
Enriched cells containing CTCs and contaminating leukocytes were stained with pan-anti-cytokeratin-phycoerythrin to detect epithelial cells, a pan-leukocyte marker (anti-CD45 PerCP; Becton Dickinson), and a nucleic acid dye as described previously (10). Leukocytes were defined as nucleic [acid.sup.+], CD[45.sup.+], [cytokeratin.sup.-] events with typical forward and orthogonal light scatter signals in flow cytometry. CTCs were defined and differentiated from leukocytes as nucleic [acid.sup.+], CD[45.sup.-], [cytokeratin.sup.+] events with forward light scatter and orthogonal light scatter signals at least as large as those of leukocytes.
After immunomagnetic enrichment of CTCs, the cells were lysed by adding 100 [micro]L of Trizol reagent (Trizol reagent total RNA isolation; cat. no. 15586-018; Invitrogen). During Trizol RNA isolation, coprecipitate was formulated according to the formula 100 [micro]L of 200 ng/[micro]L linear acrylamide (cat. no. 9520; Ambion) plus 25 [micro]L of pellet paint (cat. no. 69049; Novagen) and added to the aqueous phase before alcohol precipitation. RNA pellets were resuspended in 15 [micro]L of RNase-free water. Isolated RNA was treated with 1 U/[micro]L DNase I (amplification grade; cat. no. 18068-015; Invitrogen) in a 20-[micro]L reaction volume according to the manufacturer's instructions. DNase-treated RNA was purified by extraction with 180 [micro]L of TRIzol and 2 [micro]L of coprecipitate, and the pellet was resuspended in 5 [micro]L of RNase-free water. To evaluate the quantity and quality of the isolated total RNA, we electrophoresed 1 [micro]L (20%) of the RNA on a 1% denaturing agarose gel and performed a Northern blot for hybridization with a 10-pmol mixture of ribosomal 18S (5'-GCCCTCCAATGGATCCTCGTTAAAGG-3') and 28S (5'-GCTCTTCCCTGTTCACTCGCCGTTA-3') oligomer probes [sup.32]P-end-labeled with polynucleotide kinase in ExpressHyb Hybridization Solution (cat. no. 8015-1; Clontech) at 42 [degrees]C for 2 h, according to the manual for the probes. The probed blot was phosphorimaged on a Cyclone Phosphor-Imager (Packard Instruments) to determine RNA integrity (18S:28S ratio) and total mass relative to high-integrity total RNA calibrators of known mass (0.25, 0.5, 1.0, and 2.0 ng) run in parallel on the blot. The RNA mass from each sample was defined as the 7.5-mL blood-donor-equivalent for sample normalization by volume. The detailed procedure for RNA isolation can be found in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol50/issue5/.
TRANSCRIPT LIBRARY PREAMPLIFICATION BY SMART-aRNA
The SMART-aRNA procedure uses the switching mechanism at the 5' termini of RNA templates in the SMART[TM] PCR cDNA Synthesis Kit (cat. no. K1052-1; Clontech). The SMART-aRNA procedure was used because it simplifies the second-strand synthesis for constructing double-stranded cDNA libraries. The SMART primer was modified on the 3' terminus to contain a functional T7 RNA polymerase promoter for subsequent aRNA amplification reactions. In addition, a minimum number of PCR cycles was used to boost cDNA library mass before T7 RNA polymerase linear amplification with minimal negative effects on mRNA library representation. The remaining 80% of each RNA sample was used in the synthesis of first-strand cDNA followed by 10 cycles of PCR performed according to the SMART Uses Manual (part no. PT3041-1; Clontech) along with the following three modifications:
* The first-strand reaction used a total volume of 10 [micro]L (1 [micro]L of SuperScript II; cat. no. 18064-022; Invitrogen), and the volume of the SMART II oligonucleotide (10 [micro]M) was reduced from 1.0 to 0.2 [micro]L per reaction tube. This reduced the cost per reaction without compromising yield or representation.
* An oligo(dT) primer containing the following T7 RNA polymerase promoter sequence (5'-TCTAGTCGACGGCCAGTGAATTGTAATACGACTCACTATAGGGCGTTTTTTTTTTTTTTTTTTTTT-3') on its 5' terminus was substituted in an equimolar amount for the corresponding oligo(dT) primer provided as the "cDNA synthesis (CDS) primer" in the reagent set. This enabled the incorporation of T7 promoter at the 3' terminus of all cDNA.
* For assembly of the cDNA PCR amplification reactions, the entire 10-[micro]L first-strand reaction was added to 40 [micro]L of Advantage 2 PCR master mixture (cat. nos. K1910-y and K1052-1; Clontech). Advantage 2 PCR master mixture reagents were maintained in the proportions given in the instruction manual along with the addition of a second PCR primer (1 [micro]L of a 10 [micro]M stock) that was homologous to the 5' portion of the dT-T7 replacement primer composed of the sequence 5'-TCTAGTCGACGGCCAGTGAATT-3'.
We then performed 10 cycles of PCR amplification with the following cycling profile: 95 [degrees]C for 1 min followed by 10 cycles of 95 [degrees]C for 5 s, 65 [degrees]C for 15 s, and 68 [degrees]C for 6 min, with a final 20-min step at 72 [degrees]C. The PCR reaction mixture was processed through a Sephadex G-50 Quick Spin column (TE; cat. no. 1 523 023; Roche Diagnostics), and the eluate was concentrated under reduced pressure to 7 [micro]L. T7 aRNA polymerase amplification reactions using the AmpliScribe reagent set (cat. no. AS2607; Epicenter Technologies) were assembled and incubated at 37[degrees]C for 12-16 h. The aRNA was purified by use of 180 [micro]L of TRIzol with 1.25 [micro]L of coprecipitate added to the aqueous phase, and the pellet was resuspended in 10 [micro]L of RNase-free water. We denatured 20% (2 [micro]L) of the purified aRNA reaction mixture, using 1 [micro]L of 6x formamide-formaldehyde and 3 [micro]L of water at 65 [degrees]C for 15 min. The denatured mixture was electrophoresed (2% agarose gel) along with 50 and 500 ng of RNA molecular weight and mass calibrators (0.24-9.5 Kb RNA ladder; 1 [micro]g/[micro]L; cat. no. 15620-016; Invitrogen) for 15 min at 5 V/cm and then was stained with SYBR Gold (cat. no. S-11494; Molecular Probes). The quality and quantity of the aRNA library were determined by AlphaImager densitometry relative to calibrators by use of AlphaEase[TM] software, Ver. 5.04. (Alpha Innotech Corp.). Sample quantities were normalized by volume according to the equation:
Total RNA mass/7.5-mL sample x 0.015 / 3 =
1 donor-equivalent RNA
where "Total RNA mass/7.5-mL sample x 0.015" indicates that mRNA accounts for 1.5% of total RNA mass; and "3" indicates that the mean molecular weight of the aRNA library ~3-fold less than that of the total mRNA.
The detailed procedure for transcript library preamplification by SMART-aRNA can be found in the online Data Supplement.
MULTIGENE RT-PCR ANALYSIS
We reverse-transcribed 25 ng of aRNA in a 10-[micro]L reaction volume containing 50 ng of random 9mer primers, 1 [micro]L of SuperScript (Invitrogen), and 2 [micro]L of 5x reaction buffer (Invitrogen); the mixture was incubated at 25[degrees]C for 10 min, 37[degrees]C for 10 min, 42[degrees]C for 20 min, and 50[degrees]C for 60 min. For PCR, cDNA was then diluted to 1 donor-equivalent/ [micro]L. Ten donor-equivalents (10 [micro]L) per cDNA sample (mass range, 50-1300 pg) were used in each subsequent 50-[micro]L PCR reaction containing 1 U of platinum Taq (Invitrogen). We generated individual PCR kinetics curves from single PCR reaction tubes by pausing the thermocycler at 31, 35, and 40 cycles and aliquoting 15 [micro]L to a separate tube for subsequent gel densitometry analysis. The PCR program was as follows in a PE-9700 thermal cycler: 1 min at 95 [degrees]C, followed by 40 cycles of 95 [degrees]C for 15 s, 65 [degrees]C for 15 s, and 72 [degrees]C for 1 min, followed by 20 min at 72 [degrees]C. RT-PCR samples were electrophoresed on 2% agarose gels (5 V/cm for 20 min) in parallel with DNA size and mass calibrators (Low Range Quantitative DNA ladder; cat. no. 12373-031; Invitrogen), stained with ethidium bromide (5 g/L), and analyzed by an AlphaImager Gel Documentation and Image Analysis System with AlphaEase[TM] software, Ver. 5.04, to determine molecular weight and spot densitometry.
All gene-specific RT-PCR primer sets were designed to amplify within the last 500 nucleotides of each transcript's 3'-untranslated region upstream from poly(A) and optimized to give single bands of specific molecular weights (Table 2). Sequence identity was confirmed by internal oligoprobe blot hybridization.
SCALING OF RT-PCR RESULTS RELATIVE TO AN EXTERNAL CALIBRATION CURVE
Relative gene expression values of 0, 1, 2, 3, and 4 were assigned to samples based on comparisons of kinetics curve band intensities relative to an external calibration curve, prepared in parallel, of cytokeratin 19 (CK19) cDNA, as shown in Fig. 1. The lower limit of RT-PCR detection was evaluated with a CK19 in vitro-transcribed RNA construct (CK19 cRNA) that contained the 3'-most 800 nucleotides of GenBank accession no. NM_002276.
LOWER LIMIT OF DETECTION FOR THE SMART-aRNA PREAMPLIFICATION AND RT-PCR ANALYSIS SYSTEM
The lower limit of mRNA transcript detection for the entire SMART-aRNA preamplification and RT-PCR analysis system was determined in triplicate with a test sample set. The test RNA samples were composed of a tissue culture cell line RNA containing predetermined copy numbers of PSA, PSMA, CK19, AR, and HPN mRNA, which was added to a background of 2000 leukocyte-equivalents of total RNA. The test set was then subjected to SMART-aRNA preamplification and RT-PCR analysis.
CTC ENUMERATION AND RNA ANALYSIS
In the blood samples from 13 healthy donors, the number of leukocytes carried over in the immunomagnetic selection ranged from 655 to 5560 (median, 1450; mean, 1759). In the 23 samples from CaP patients, the number of leukocytes carried over ranged from 813 to 92 000 (median, 4350; mean, 12 300). The events that were classified as CTCs in blood samples from the control group ranged from 0 to 4 CTCs/7.5 mL [mean (SD), 0.8, (1.2)]. In blood samples from HRPC patients, the CTCs ranged from 4 to 283/7.5 mL (median, 89; mean, 87; Table 1). The total RNA from the 13 healthy donors as assessed by Northern blot had a mean 28S:18S band ratio of 1.4 (range, 1.2-1.8) with a mean yield of 3.5 ng (range, 0.8-11.12 ng). In CaP patients, this ratio was 1.2 (range, 0.6-1.8) with a mean yield of 7.2 ng (range, 0.8-35.12 ng). All samples subsequently produced aRNA libraries with masses closely proportional to the starting total RNA values.
CHARACTERIZATION OF SMART-aRNA LIBRARY AMPLIFICATION
The modified aRNA amplification method produced libraries with a mean 10 000-fold yield above the original mRNA mass. The original amount of mRNA was estimated to be 1.5% of the total RNA mass. The median transcript length of the libraries was 600 nucleotides (range, 550-800 nucleotides). Individual transcript lengths within each library ranged from 300 to 3000 nucleotides. We investigated the distortion effect of this modified aRNA amplification method on the relative mRNA abundance by comparing the expression of CK8, CK19, mammaglobin 1 (MGB1), mammaglobin 2 (MGB2), prolactin-inducible protein (PIP), EpCAM, PSA, and PSMA in RNA libraries obtained from tumor cell lines with and without amplification. As shown in Fig. 2, there was no dropout of any mRNA target sequence attributable to the SMART-aRNA method. In addition, we saw no dramatic differences in band intensity ratios for all eight mRNA sequences when we compared the amounts of RNA after 32 and 40 PCR cycles.
We assessed the effect of the SMART-aRNA preamplification procedure on the lower limit of detection for RT-PCR with a model system mimicking immunomagnetically enriched CTC samples. In triplicate samples, the combination of SMART-aRNA followed by RT-PCR amplification showed reproducible detection of mRNA from all six targets--PSA, PSMA, AR, HPN, EpCAM, and CK19--down to at least 50 transcripts (data not shown). We evaluated the lower limit of detection for RT-PCR with in vitro-transcribed CK19 cRNA. The results from 10 separate RT-PCR reactions showed that the CK19 cRNA was reproducibly detectable in 10 of 10 separate reactions down to 25 molecules of RNA (data not shown). RT-PCR contamination controls containing all components except cDNA sample template were run in parallel with each batch and did not show any detectable signals, indicating the absence of crossover and carryover contamination during these studies.
MULTIGENE RT-PCR PROFILING OF SMART-ARNA FROM SAMPLES FROM HEALTHY DONORS IMMUNOMAGNETICALLY ENRICHED FOR CTCS
An external calibration curve constructed with CK19 cDNA was run in parallel with samples. The band intensities from the calibration curve were used to assign relative gene expression values of 0, 1, 2, 3, and 4 to the samples shown in Fig. 1. All blood samples contained leukocytes carried over from the CTC enrichment process. The [[alpha].sub.1]-globin gene was found to be specifically expressed in leukocytes and was therefore used as the functional system control for indicating successful RNA processing and amplification. All 36 samples gave gene expression values >4 for [[alpha].sub.1]1-globin (Table 2; data not shown for patient samples). Expression of 38 genes was assessed in the aRNA libraries of the 13 healthy controls. The expression profiles of these genes and the aRNA-RT-PCR design characteristics are listed in Table 2. Expression of the genes in CTC-enriched samples from healthy controls represented the expression of these genes in the leukocytes that were carried over in the CTC enrichment procedure. We observed no expression of AR, carcinoembryonic antigen (CEA), cytokeratin 5 (CK5), CK19, EGFR, estrogen receptor-[beta] (ER-b), HK2, MGB1, MGB2, PSA, PSGR, PSMA, or tumor-associated calcium signal transducer 2 (Trop2) in any of the libraries from the controls. The expression of these genes in CTCs enriched from cancer patients could therefore be attributed to the CTCs and not the leukocytes carried over in the CTC enrichment procedure. We found no or low expression for EpCAM, macrophage-inhibitory cytokine-1 (Mic1), matrix metalloproteinase 2 (MMP2), mucin 1 (Muc1), an androgen-regulated gene (NKX3A), and topoisomerase II[alpha] (Topo2A) in controls. Expression of these genes in CTCs can therefore be attributed to CTCs if they are expressed in increased amounts. Genes more highly expressed in the controls were not used in the evaluation of CTCs with this CTC enrichment procedure. Within this group were the genes encoding for the cytoskeletal proteins cytokeratins 8, 10, and 18, which are generally restricted to nonhematopoietic cells, and genes that are potential candidates for therapeutic agents: BCL2, HER-2, multidrug resistance-associated protein (MRP), and p53. In all samples, the score for CK10 was 3, which implied a housekeeping function that potentially can be used for sample normalization.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
MULTIGENE RT-PCR PROFILING OF SMART-ARNA FROM SAMPLES FROM HRPC PATIENTS
IMMUNOMAGNETICALLY ENRICHED FOR CTCS
The relative expression of CK19, PSA, PSMA, AR, HPN, HK2, PSGR, MGB1, and MGB2 in the mRNA libraries from CTC-enriched samples obtained from 23 samples from 9 patients with metastatic CaP are shown in Fig. 3. The numbers for individual samples from Table 1 are used to indicate the relative expression of each gene. A multigene expression profile can be obtained from each patient and each sample by examining the positions of the sample numbers across all genes. CK19 was expressed in 18 of 23 (78%) samples and 6 of 9 (67%) CaP patients. PSA was expressed in 20 of 23 (87%) samples and 8 of 9 (89%) CaP patients. PSMA was expressed in 17 of 23 (74%) samples and 7 of 9 (78%) CaP patients. AR was expressed in 16 of 23 (70%) samples and 7 of 9 (78%) CaP patients. HPN was expressed in 13 of 23 (57%) samples and 6 of 9 (78%) CaP patients. It must be noted that HPN was expressed in the RNA library of one of the female controls. HK2 was expressed in 7 of 23 (30%) samples and 4 of 9 (44%) CaP patients. PSGR was expressed in 2 of 23 (9%) samples and 2 of 9 (22%) CaP patients. Unexpectedly, both "breast-specific" genes MGB1 and MGB2 scored positive in 1 of 23 (4%) and 2 of 23 (9%) CaP samples, respectively. The multigene combination of PSA and PSMA was positive in 20 in 23 (87%) samples, and addition of HPN produced positive results in 21 of 23 (91%) samples. Only one sample, patient 9, scored negative for all of the tested genes shown in Fig. 3. The genes EGFR (4 of 23; 17%), Trop2 (2 of 23; 9%), ER-b (1 of 23; 4%), CK5 (1 of 23; 4%); and CEA (0 of 23; 0%) were infrequently expressed in patient samples and were not expressed in the leukocytes. The genes EpCAM, Mic1, MMP2, Muc1, NKX3A, and Trop2A were expressed in low background amounts in leukocytes.
[FIGURE 3 OMITTED]
When the threshold was increased to 2 or higher, EpCAM scored positive in 17 of 23 (74%), Mic1 in 15 of 23 (65%), MMP2 in 0 of 23 (0%), Muc1 in 11 of 23 (48%), NKX3A in 2 of 23 (9%), and Topo2A in 5 of 23 (22%) patient samples. MDR1 expression was scored as 0, 1, or 2 in the control samples, and this expression was exceeded in 5 of 23 (22%) patient samples. The multigene combination of CK19, EpCAM, and Muc1 yielded a sensitivity of 100% (23 of 23).
SERIAL CTC AND MRNA PROFILES DURING TREATMENT OF THREE HRPC PATIENTS
Longitudinal patient profiles of CTC counts and mRNA expression for PSA, PSMA, AR, and HPN in blood samples drawn over the course of 18-26 weeks from three patients with HRPC are shown in Fig. 4. Patient 1 (Fig. 4A) was treated with long-term hormone ablation (Lupron). Patients 2 and 3 (Fig. 4, B and C, respectively) were also on long-term Lupron treatment and, in addition, were begun on chemotherapy consisting of Taxane/estramustine. The CTC values changed by ~10-fold for all three patients during this survey period, and some trends were noted as follows. For the patient treated with Lupron alone (Fig. 4A), who had four longitudinal samples drawn at 0, 5, 10, and 18 weeks, we observed a dramatic progressive increase in CTC counts from 45 up to 300 during the survey period. In contrast, the two patients treated with Lupron and adjuvant chemotherapy (Fig. 4, B and C) showed a significant increase in CTC counts followed by a decrease. For the patient in Fig. 4C, we observed a dramatic increase in CTCs after the chemotherapy regimen was discontinued. mRNA concentrations during serial sampling showed that PSMA expression paralleled CTC changes, whereas PSA expression remained high throughout the survey period. The amounts of AR mRNA in the patients represented in panels A and B of Fig. 4 remained relatively constant. AR expression was detected in blood draws from the patient represented in Fig. 4C only during treatment with Taxane/ estramustine and was not detected before or after treatment. We detected a dramatic change in HPN mRNA in the patients represented in panels B and C of Fig. 4, from high expression when the patients were untreated to complete elimination during the Taxane/estramustine treatment. HPN expression fluctuated in the blood samples from the patient represented in Fig. 4A. The changes in mRNA concentrations were independent of the number of CTCs and leukocytes.
[FIGURE 4 OMITTED]
Clinical endpoints to measure HRPC disease burden have traditionally been limited to survival. Recently, serum PSA has gained some acceptance as a viable endpoint (13), and positive signals for PSA RT-PCR, signifying the presence of CTCs in HRPC, have also been suggested to predict survival (3). We have shown that the quantity of CTCs trends with metastatic disease progression in a manner that is independent of PSA concentration in select cases (10). Our goal in this study was to determine the feasibility of profiling gene expression in CTCs with the ultimate aim of better characterizing advanced CaP biology. However, a substantial enrichment of CTCs over leukocytes is needed to overcome signal-to-noise factors. A modified T7 RNA polymerase preamplification method (SMART-aRNA) was engineered to produce aRNA CTC libraries from immunomagnetically enriched CTCs targeting the epithelial cell surface antigen EpCAM. The preamplification of mRNA produced a 10 000-fold increase in mRNA, whereas the representation of genes in the aRNA library was not compromised compared with the native, unamplified mRNA library (Fig. 2). mRNA was detectable down to amounts as low as 50 molecules (data not shown). The small linear dynamic range obtained with the semiquantitative RT-PCR detection system used in this study can be readily improved by use of "real-time" quantitative RT-PCR. This approach to profile mRNA gene expression in the CTCs present in 7.5 mL of blood is limited by a background of 103-104 leukocytes. The unavoidable leukocyte carryover translates into significant signal-to-noise restrictions that limit the genes that can be adequately evaluated. For example, the expression of p53, TS, MRP, HER-2, and BCL-2 in leukocytes precludes unambiguous detection of their overexpression in CTCs. It might, however, be possible to develop a method to subtract the leukocyte mRNA background from patient-specific CTCs based on the leukocyte profile obtained from patient-matched CTC-depleted leukocyte fractions or CTCs enriched from healthy donors. Alternatively, the leukocyte background can be significantly reduced by use of fluorescence-activated cell sorting. However, cell sorting inevitably leads to losses of CTCs and is practical only in a research setting.
Another method to overcome this limitation of gene expression is to assess the amount of mRNA through protein production. For example, although HER-2 mRNA analysis is not possible in immunomagnetically enriched CTCs because of the background expression of HER-2 mRNA in leukocytes, the production of HER-2 protein by CTCs has been assessed successfully in breast cancer patients (9). Conversely, not all proteins can be assessed on CTCs. Some secreted proteins, such as PSA, migrate out of the CTCs during permeabilization and can no longer be detected. Hence the limitations of CTC detection by flow cytometry or microscopy can be overcome by complementary aRNA analysis. This was illustrated in one of the nine patients. Case 9 was the only androgen-sensitive case in this study and had 30 CTCs/7.5 mL of blood, which was confirmed morphologically by fluorescence microscopy. However, these CTCs were negative on RT-PCR for PSA, PSMA, HK2, and PSGR, suggesting that they might express different genes reflecting their different disease biologies. However, at this time we cannot rule out the possibility that these CTCs might not be derived from the prostate. mRNA concentrations for Muc1 and Trop2A were higher in samples from patients than in control samples, suggesting that these CTCs could be derived from epithelial tissue representing another clinically occult neoplasm. CTCs ranged from 4 to 283/7.5 mL blood for 23 specimens from 9 prostate cancer patients. In contrast the CTC range in the 13 control cases was 0-4 CTCs/7.5 mL (mean, 0.8). One of the epithelial cell-specific genes (CK19, EpCAM, or Muc1) was expressed in all patient samples, whereas 21 of 23 samples from 8 of 9 patients showed expression of one of the prostate-specific genes (PSA, PSMA, or HPN). Expression of these prostate-tissue-specific genes was completely absent in controls. The relative expression of CK19, PSA, and PSMA was comparable to the values reported in the literature (14, 15). For example, the proportion of samples that were positive for HK2 expression (30%) was similar to the 25% reported by Ylikoski et al. (16). Although the proportion of samples in which PSGR expression was detectable (2 of 23; 9%) may be lower than previously published values, it is possible that PSGR is expressed in higher amounts in CTCs of patients with hormone-naive prostate disease (17, 18). HPN has not been studied extensively in patients with HRPC, and the proportion of samples with detectable expression (57%) in this report suggests that this may be a viable drug therapy target (19). Moreover, the efficacy of anti-HPN agents may be assessed by their ability to decrease HPN expression in CTCs. EGFR was expressed in only 17% of the specimens, which is low compared with the 57% reported in prostate samples from 16 patients with HRPC (20). This discrepancy may be attributable to several factors, including differential EGFR expression in CTCs vs prostate cancer cells within the organ of origin. We found a similar discrepancy in Topo2A expression (22%) (21). The expression of AR in 70% of the samples is in agreement with the results reported by Linja et al. (22), who found a correlation between androgen independence and high AR expression in 13 patients with HRPC. Many hormone-dependent features are known to be common to breast and prostate cancers, which may explain the observed expression of PSA in breast cancer (23) and the first reported expression of the breast-specific genes MGB1 and MGB2 in CTCs from CaP patients.
Expression of the AR, HPN, PSMA, and PSA genes in CTCs was surveyed serially over an ~8- to 26-week period in three patients (Fig. 4). Two of the three patients had received chemotherapy. The results shown in Fig 4 demonstrate that the CTC counts reflect disease progression more accurately than semiquantitative PSA and PSMA RT-PCR results, which showed uniformly high expression throughout the monitoring period. This may be limited by the semiquantitative RT-PCR assay used in this study, however, because a relative gene expression score of 4 for PSA and PSMA can not be differentiated from saturation in this detection system. Our data imply that previous PSA RT-PCR studies that attempted to predict survival may have been hampered by an inability to differentiate between high and low CTC counts. The results further support research into AR as a therapeutic target because the amount of AR expression was sustained and found to be minimally affected by chemotherapy.
In conclusion, we believe that gene expression profiling of CTC mRNA is feasible. Our data further validate that the immunomagnetically enriched CTCs enumerated by flow cytometry or fluorescence microscopy originated from the CaP tissue. The prostate CTCs exhibited heterogeneous expression of genes such as PSA, PSMA, and HPN. In the HRPC cases studied, AR expression appeared to be consistently increased, suggesting AR as a viable therapeutic target. The ability to quantify, phenotype, morphologically examine, and now molecularly profile CTCs could lead to improved characterization of HRPC and ultimately to development of more effective, personalized novel therapeutic strategies.
Received October 21, 2003; accepted February 5, 2004.
Previously published online at DOI: 10.1373/clinchem.2003.028563
(1.) Braun S, Pantel K, Muller P, Janni W, Hepp F, Kentenich CR, et al. Cytokeratin-positive cells in the bone marrow and survival of patients with stage I, II or III breast cancer. N Engl J Med 2000;341:525-33.
(2.) Janni W, Hepp F, Rjosk D, Kentenich C, Strobl B, Schindlbeck C, et al. The fate and prognostic value of occult metastatic cells in the bone marrow of patients with breast carcinoma between primary treatment and recurrence. Cancer 2001;92:46-53.
(3.) Halabi S, Small EJ, Hayes DF, Vogelzang HN, Kantoff PW. Prognostic significance of reverse transcriptase polymerase chain reaction for prostate-specific antigen in metastatic prostate cancer: a nested study within CALGB9583. J Clin Oncol 2003;21: 490-5.
(4.) Cristofanilli M, Reuben JM, Fritsche HA, Kau SW, Lara J, Arun B, et al. Circulating tumor cells (CTCs): a novel prognostic factor for newly diagnosed metastatic breast cancer (MBC). Proc Am Assoc Cancer Res 2003;44:1349.
(5.) Ellis WJ, Pfitzenmaier J, Colli J, Arfman E, Lange PH, Vessella RL. Detection and isolation of prostate cancer cells from peripheral blood and bone marrow. Urology 2003;61:277-81.
(6.) van't Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature 2002;415:530-6.
(7.) Van Gelder RN, von Xastrow ME, Yool A, Dement DC, Barchas JD, Eberwine JH. Amplified RNA synthesized from limited quantities of heterogeneous cDNA. Proc Natl Acad Sci U S A 1990;87:1663-7.
(8.) Gordon GJ, Jensen RV, Hsiao LL, Gullans SR, Blumenstock JE, Richards WG, et al. Using gene expression ratios to predict outcome among patients with mesothelioma. J Natl Cancer Inst 2003;95:598-605.
(9.) Hayes DF, Walker TM, Singh B, Vitetta ES, Uhr JW, Gross S, et al. Monitoring expression of HER-2 on circulating epithelial cells in patients with advanced breast cancer. Int J Oncol 2002;21: 1111-8.
(10.) Moreno JG, O'Hara SM, Gross S, Doyle G, Fritsche H, Gomella LG, et al. Changes in circulating carcinoma cells in patients with metastatic prostate cancer correlate with disease status. Urology 2001;58:386-92.
(11.) Kagan M, Howard D, Bendele T, Mayes J, Silvia J, Repollet M, et al. A sample preparation and analysis system for identification of circulating tumor cells. J Clin Ligand Assay 2002;25:104-10.
(12.) Kagan M, Trainer M, Bendele T, Rao C, Horton A, Tibbe A, et al. CellTracks cell analysis system for rare cell detection. In: Cohn GE, ed. Clinical diagnostic systems: technologies and instrumentation. Washington: SPIE, 2002;4625:20-8.
(13.) Bubley GJ, Carducci M, Dahut W, Dawson N, Daliani D, Eisenberger M, et al. Eligibility and response guidelines for phase II clinical trials in androgen-independent prostate cancer: recommendations from the Prostate-Specific Antigen Working Group. J Clin Oncol 1999;17:3461-7.
(14.) Gomella LG, Raj GV, Moreno JG. Reverse transcriptase polymerase chain reaction for prostate specific antigen in the management of prostate cancer. J Urol 1997;158:326-37.
(15.) Moreno JG, Gomella LG. Circulating prostate cancer cells detected by reverse transcription-polymerase chain reaction (RT-PCR): what do they mean? Cancer Control 1998;5:507-12.
(16.) Ylikoski A, Karp M, Pettersson K, Lilja H, Lovgren T. Simultaneous quantification of human glandular kallikrein 2 and prostate-specific antigen mRNAs in peripheral blood from prostate cancer patients. J Mol Diagn 2001;3:111-22.
(17.) Xu LL, Stackhouse BG, Florence K, Zhang W, Shanmugam N, Sesterhenn IA, et al. PSGR, a novel prostate-specific gene with homology to a G protein-coupled receptor, is overexpressed in prostate cancer. Cancer Res 2000;60:6568-72.
(18.) de Kok JB, Verhaegh GW, Roelofs RW, Hessels D, Kiemeney LA, Aalders TW, et al. DD3(PCA3), a very sensitive and specific marker to detect prostate tumors. Cancer Res 2002;62:2695-8.
(19.) Magee JA, Araki T, Patil S, Ehrig T, True L, Humphrey PA, et al. Expression profiling reveals HPN overexpression in prostate cancer. Cancer Res 2001;61:5692-6.
(20.) Autorino R, D'Armiento M, De Laurentiis M, De Placido S, Catalano G, Bianco AR, et al. Expression of epidermal growth factor receptor correlates with disease relapse and progression to androgen-independence in human prostate cancer. Clin Cancer Res 2002;11:3438-44.
(21.) Willman JH, Holden JA. Immunohistochemical staining for DNA topoisomerase II-[alpha] in benign, premalignant, and malignant lesions of the prostate. Prostate 2000;42:280-6.
(22.) Linja MJ, Savinainen KJ, Saramaki OR, Tammela TL, Vessella RL, Visakorpi T. Amplification and overexpression of androgen receptor gene in hormone-refractory prostate cancer. Cancer Res 2001;61:3550-5.
(23.) Monne M, Croce CM, Yu H, Diamandis EP. Molecular characterization of prostate-specific antigen messenger RNA expressed in breast tumors. Cancer Res 1994;54:6344-7.
(3) Nonstandard abbreviations: CTC, circulating tumor cell; HRPC, hormone-refractory prostate cancer; RT-PCR, reverse transcription-PCR; aRNA, antisense RNA; PSA, prostate-specific antigen; PSMA, prostate-specific membrane antigen; HK2, human glandular kallikrein 2; HPN, hepsin; PSGR, prostate-specific gene with homology to G-protein receptor; AR, androgen receptor; MDR1, multidrug resistance gene; EGFR, epidermal growth factor receptor; CaP, prostate cancer; EpCAM, epithelial cell adhesion molecule; CK, cytokeratin; MGB1 and MGB2, mammaglobin 1 and 2; PIP, prolactin-inducible protein; CEA, carcinoembryonic antigen; ER-b, estrogen receptor [beta]; Trop2, tumor-associated calcium signal transducer 2; Mic1, macrophage inhibitory cytokine-1; MMP2, matrix metalloproteinase 2; Muc1, mucin 1; Topo2A, topoisomerase II[alpha]; and MRP, multidrug resistance-associated protein.
S. Mark O'Hara,  Jose G. Moreno,  Daniel R. Zweitzig,  Steve Gross,  Leonard G. Gomella,  and Leon W.M.M. Terstappen  *
 Immunicon Corporation, Huntingdon Valley, PA.
 Thomas Jefferson University Hospital, Department of Urology, Philadelphia, PA.
* Address correspondence to this author at: Immunicon Corporation, 3401 Masons Mill Road, Suit 100, Huntingdon Valley, PA 19006. Fax 215-830-0751; e-mail Lterstappen@Immunicon.com.
Table 1. Patient characteristics. Sample Age, Patient no. years HTx (a) CTx (a) 1 1.1 70 + - 1.2 + - 1.3 + - 1.4 + - 2 2.1 68 + - 2.2 + + 2.3 + + 2.4 + + 2.5 + + 3 3.1 73 + - 3.2 + + 3.3 + + 3.4 + + 3.5 + + 4 4.1 69 - - 4.2 - - 5 5.1 60 + - 5.2 + - 6 6.1 86 - (a) - 6.2 - (a) - 7 7 73 + - 8 8 78 + - 9 9 74 + - Bone PSA, Patient scan (c) [micro]g/L CTCs/7.5 mL 1 + 32 39 + 38 41 + 49 124 + 107 283 2 + 47 4 + 630 218 + 630 89 + 379 70 + 562 86 3 + 43 31 + 312 183 + 437 138 + 311 23 + 431 158 4 + 1 ND (a) + 184 116 5 - 30 78 - 48 25 6 + 53 45 + 217 64 7 + 128 44 8 + 12 16 9 - 0.1 30 (a) HTx, hormone ablation therapy: +, yes; -, no. (b) CTx, chemotherapy: +, yes; -, no. (c) Bone scan: +, positive; -, negative. (d) ND, not determined. (e) Orchiectomy. Table 2. Multigene RT-PCR sequence information and gene expression in SMART-aRNA libraries from immunomagnetically enriched CTC from healthy donors. Marker GenBank gene accession no. PCR primer pairs, (a) 5'3' [[alpha].sub.1] V00491 5-CTACTTCCCGCACTT/10-TCAGCACGGT -Globin AR NM_000044 5-TGTGCAAGTGCCCAAGAT/15-GACAGACTGT CEA M29540 5-AAGATCAAGCAGACAAAA/15-AAGAGTGGATAGC CK5 NM_000424 5-CTGCCTTCCAAGTGCAGCAA/15-GATTTGAAGCAG CK19 NM_002276 5-CCAGGCGCTGATCAGCG/10-CAGAAGCCCCAG EGFR NM_005228 5-AACCTGACTGGTTAACAGCAG/15-GGGAGTCA ER-b NM_001437 5-CTGGCTCACTTGCTGAA/10-GGCATTCAGCAT HK2 XM_008996 5-CATGCAGGATGACAT/10-AGGTTCTCAG MGB1 XM_006409 5-AGCACTGCTACGCAGGC/10-AGAGAAGGTGTGGT MGB2 NM_002407 10-ACTGCTATGCAGAT/15-GTACACTGTATGCA PSA M26663 5-CACTGAGCAGAAGCTGGA/10-TGGAGGACTTCAA PSGR AF311306 5-GCTTTGGAAACAGCCTTCATC/15-TGGGCAACTGG PSMA M99487 5-AGTGAGAGACTCCAGGAC/10-AAGGCTGCAACAT TROP2 X77753 5-CTACTCTGGTGTGTCCCAAG/15-GGTACAGCTC NKX3A NM_006167 5-TTCAGCCATCAGAAGTAC/10-GTAAGGATAG MMP2 NM_004530 5-TGGCTGCCTTAGAACCTT/10-TCGGTAGGGACAT Muc1 J05582 5-TGGCAGCAGCCTCTCTTA/10-ACTGAGAAGTGTCCG EpCAM M33011 5-GCATAGGGAACTCAATGC/10-CCAAGTTCTGGAT Topo2A NM_1067 5-CCACTTCTGATGATTCTG/10-GGCTTGGTAAGA Mic1 AF019770 5-ATCCCATGGTGCTCA/10-ATCAGACCAG MDR1 AF016535 5-CCAGGCTGGAACAAAG/10-TGATGTCTCTCAC HPN NM_002151 5-AGGCGTCTACACCAA/10-GGGTCACCAT TERT NM_003219 5-ACCTGCCGTCTTCACTT/10-TGGTCACTCCAA uPA4 NM_002658 5-TGTGAGTGTAAGTGTGAG/10-GGATTGGATGAAC ER-a NM_000125 5-GTGCCTGAGACACAGA/10-CGCTGGATTCTT PIP J03460 10-CAAATTGCAGCCGTC/15-TTCCAGCCAAG Her[2.sup.NEU] M11730 5-GGAAGAGGAACAGCACTG/10-CTGACACCATTGC CK8 M34225 5-TTGAGCTCGGCCTATGG/10-CCTGCATAGCG MMP9 NM_004994 5-TCCAGTACCGAGAGAAAG/10-AAACTGGCTCCTT MRP L05628 5-TCGTCTTGGACAAAGGAG/10-CAGTTCCAGGCAG CK18 NM_000224 5-GAGTCAGAGCTGGCACAGA/10-GCTTAATGCCTCAG TS AB062290 5-CTGGCAAATGTAACTGT/10-TCCTCACTTTGTTCAT Timp2 NM_003255 5-TGCGAGTGCAAGATCAC/10-GTCCTCGATGTC BCL-2 XM_008738 5-AGTGACAGTGGATTGCAT/15-TGGAGACT Topo2b NM_001068 5-CAAGAGAGCCCCAAAAC/10-GGTGGCTCAGTA Timp1 NM_003254 5-ACCTACACTGTTGGCTGT/10-CTTCAGTTCCACT p53 AF307851 5-TCAGCCTCCGGAGTAGCT/10-AATGCAGATGTGC CK10 NM_000421 10-TTCTTTCATCTACGGTTG/15-CCTTGTAGACACC Marker PCR fragment aRNA required gene size, bp length, (b) nt [[alpha].sub.1] 451 580 -Globin AR 207 513 CEA 144 297 CK5 212 353 CK19 228 320 EGFR 200 265 ER-b 232 323 HK2 190 334 MGB1 340 408 MGB2 264 420 PSA 236 236 PSGR 281 425 PSMA 286 459 TROP2 245 334 NKX3A 246 284 MMP2 216 332 Muc1 238 361 EpCAM 275 432 Topo2A 244 356 Mic1 289 352 MDR1 215 263 HPN 101 383 TERT 234 347 uPA4 285 387 ER-a 328 382 PIP 201 226 Her[2.sup.NEU] 266 426 CK8 275 429 MMP9 215 327 MRP 300 328 CK18 275 331 TS 114 398 Timp2 219 325 BCL-2 330 440 Topo2b 279 366 Timp1 240 302 p53 243 334 CK10 196 305 Background expression (c) Marker gene 0 1 2 3 4 [[alpha].sub.1] 100 -Globin AR 100 CEA 100 CK5 100 CK19 100 EGFR 100 ER-b 100 HK2 100 MGB1 100 MGB2 100 PSA 100 PSGR 100 PSMA 100 TROP2 100 NKX3A 92 8 MMP2 92 8 Muc1 77 23 EpCAM 61 39 Topo2A 54 46 Mic1 46 54 MDR1 23 62 15 HPN 92 8 TERT 92 8 uPA4 84 8 8 ER-a 84 8 8 PIP 76 8 8 8 Her[2.sup.NEU] 61 23 8 8 CK8 46 31 8 15 MMP9 30 39 8 15 8 MRP 23 23 31 23 CK18 15 39 8 39 TS 15 15 31 39 Timp2 8 31 46 15 BCL-2 23 46 31 Topo2b 15 23 62 Timp1 8 61 31 p53 8 8 30 54 CK10 100 (a) Primer pairs are given as the truncated versions of the forward primer (P1)/reverse primer (P2), in which the 5' numbers (5, 10, and 15) indicate the number of deleted nucleotides. The complete sequences can be determined from GenBank accession numbers, primer sequences, and PCR product sizes. (b) Minimum aRNA fragment length in nucleotides (nt) required for each gene to be amplified successfully by RT-PCR with the corresponding PCR primer pair. (c) Expression of the genes in CTC-enriched blood samples from 13 healthy individuals. Expression is the result of leukocytes carried over during the procedure and is given as a percentage of the signals relative to the external calibration curve shown in Fig. 1 . (d) uPA, urouinase plasminogen activator; ER-a, estrogen receptor- [alpha]; TS, thymidylate synthetase; Timp, tissue inhibitor of metalloproteinase.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Molecular Diagnostics and Genetics|
|Author:||O'Hara, S. Mark; Moreno, Jose G.; Zweitzig, Daniel R.; Gross, Steve; Gomella, Leonard G.; Terstappen|
|Date:||May 1, 2004|
|Previous Article:||Simple technique for internal control of real-time amplification assays.|
|Next Article:||Direct visualization of cystic fibrosis transmembrane regulator mutations in the clinical laboratory setting.|