| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1
INSERM U90 and Clinical Biochemistry Laboratory, CHU Necker, 149 rue de Sèvres, 75015 Paris, France;
2
Molecular Chemistry Laboratory, Saint-Luc Clinical University, Clos Chapelle aux Champs 30, Brussels, Belgium;
3
Urology Unit and Experimental Surgery Laboratory, CHU Bicêtre, 78 rue du Général Leclerc, 94270 Le Kremlin Bicêtre, France;
a address correspondence to this author at: Laboratoire de Biochimie A, Hôpital Necker, 149 rue de Sèvres, 75015 Paris, France
The sensitive and specific detection of circulating tumor cells holds great promise for more accurate staging of cancer patients. Several reverse transcription-PCR (RT-PCR) procedures based on tissue-specific mRNA expression are now able to detect one cell derived from a given tissue among >10 peripheral nucleated blood cells [PNBCs; for a review, see (1)]. However, even for a single marker, highly discrepant results have been observed among the available clinical studies; e.g., the frequencies of positive prostate-specific antigen (PSA) RT-PCR results range from 25% (2) to 80% (3) in patients with metastatic prostate cancer (CaP), blurring the clinical relevance of these assays. Standardization and quality control in molecular diagnosis are crucial to the solution of this issue. We have previously studied factors potentially affecting RT-PCR results (4), and this current work focuses on the approaches for harvesting prostate cells among nucleated blood cells.
Since the first 1992 clinical report describing the RT-PCR detection of circulating prostatic cells in CaP, the majority of reported assays have used gradient separations to recover nucleated cells from the peripheral blood (1). We (5) and others (6) rather choose overnight hypo-osmotic red blood cell lysis as an easier and more cost-effective protocol. Because the approach used to harvest PNBCs may account for discrepancies in clinical results, we have compared a panel of nucleated blood cell separation methods.
Fresh blood from four healthy blood donors (total volume, 400 mL) was sampled in 40 x 10-mL EDTA-treated tubes (Becton-Dickinson). The LNCaP cell line, derived from a metastatic prostate carcinoma, was cultured as described (7). After trypsinization (2.5 mL/L trypsin), 10 confluent LNCaP cells were added into each of the 40 tubes, which were gently mixed. Six tubes were processed immediately by overnight hypoosmotic red blood cell lysis, as described (5)(6): Two volumes of ammonium chloride (9 g/L) were mixed with 10 mL of blood in the first three tubes (5), whereas 1.5 volumes of diethylpyrocarbonate (DEPC)-treated water was added for 5 min in the other three tubes (6). After centrifugation and removal of the supernatant, 1 mL of guanidinium thiocyanate was added to each tube, and RNA was extracted as described (8). Four commercial two-layer density gradients were utilized: LymphoPrepTM (Nycomed, d = 1.077 kg/L), Ficoll-PaqueTM (Pharmacia Biotech, d = 1.077 kg/L), PolymorphPrepTM (Life Technologies, d = 1.113 kg/L), and NycoPrepTM (Life Technologies, d = 1.068 kg/L). Two additional density gradients (d = 1.095 kg/L and d = 1.050 kg/L) were prepared by diluting Percoll (Pharmacia Biotech) in 1.5 mol/L NaCl. The amount of Percoll added to reach an isoosmotic solution was determined according to the manufacturer's instructions. Samples of 10 mL of blood mixed with LNCaP cells were laid on top of each of 10-mL gradient. This was followed by centrifugation at 450g for 20 min at 4 °C.
The efficiency of nucleated cells separation was assessed by four different methods.
In method a, the banding of [H]-thymidine-incubated LNCaP cells, aggregate-free LNCaP cells were suspended in 1 mL of phosphate-buffered saline and incubated with [H]-thymidine (0.76 TBq/mmol [6-H]-thymidine, Amersham) for 12 h under sterile conditions and continuous agitation. The total activity of the cell suspension reached 3.0 kBq. Labeled cells were then added to 10 mL of peripheral blood samples and processed in density gradient medium as described above. After centrifugation of the gradient tubes at 450g for 20 min, 500-µL fractions of the gradient were poured into scintillation vials containing 5 mL of lysis buffer (75 mmol/L NaCl, 25 mmol/L Na2EDTA, pH 8.0). [H]-thymidine ß activity was measured on a ß counter (LS8100, Beckman) after addition of the scintillation liquid (H-Ionic Fluor®, Packard).
In method b, cytological analysis, cells were collected from serial layers and cyto-centrifuged at 200g for 20 min (Cytospin II®, Shandon) and stained with May-Grunwald Giemsa (RAL Reagents).
In method c, immuno-cytochemical analysis, immunostaining of serial
layers was performed with a specific monoclonal anti-PSA antibody
(IgG1
M750 clone ER-PR8, Dako) at the final
antibody concentration of 20 mg/L. Cells were fixed for 2 h in 40
g/L p-formaldehyde in phosphate-buffered saline (pH 7.4) at
room temperature; cell smears were also labeled with a specific
monoclonal anti-PSA antibody. Positive prostate cells were visualized
by the alkaline phosphatase anti-alkaline phosphatase staining
procedure, following manufacturer's recommendations (Dako APAAP
kit®).
In method d, PSA and prostate-specific membrane antigen (PSMA) RT-PCR assays were performed as already described (5) after normalization of the cDNAs by ß-globin RT-PCR (9). Briefly, 1 µg of total RNA was used for cDNA first strand synthesis together with SuperScript II reverse transcriptase and 100 ng of oligo(dT)l2–18 as a template (Life Technologies). The ß-globin, PSA, and PSMA genes were amplified as follows: 94 °C for 1 min (2 min for the first cycle); 58 °C, 60 °C, and 62 °C, respectively, for 1 min; and 72 °C for 1 min (10 min for the last cycle) for 25 cycles. When RT-PCRs were negative, 3 µL of PCR amplimers were amplified by 25 cycles of PSA or PSMA nested-PCR in the same upper conditions (9). PCR products (15 µL) were run on ethidium bromide-stained agarose gels (20 g/L) and visualized after transillumination. Fluorescence was captured with a digital camera and integrated by using NIH Image 1.7 software. Statistical analysis was performed on the results of three independent duplicate experiments, using GraphPad Prism/StatTM software.
The results of these studies were grouped according to the four methods used.
In method a, banding density generated by prostate epithelial cells
varied from one gradient to another (Fig. 1
a). The best separation was obtained with NycoPrep (d =
1.068 kg/L) gradient, with cells collected at the interface between
platelet-enriched plasma fraction and the gradient, suggesting a close
correspondence of this value with the density of the main part of the
LNCaP cells. Whereas prostate cancer cells have been reported to be
isolated in a single band (d = 1.056 kg/L) (10) when a
mix of two different density layer gradients was used, the Percoll
dilution of equivalent density (d = 1.050 kg/L) gave a more spread
population. Our findings confirm the results of Griwatz et al.
(11), who developed a new enrichment method for circulating
epithelial cells.
|
In methods b and c, both cytological and immuno-cytochemical analyses
of the banding pattern showed, however, that prostate cells can be
isolated in fractions above the 1.068 kg/L density (Table 1
).
|
In method d, PSA and PSMA RT-PCR on the different fractions confirm those data. Moreover, because of its high sensitivity, we found RT-PCR positivity in the granulocyte layer (below 1.077 kg/L density) as well as in the peripheral blood lymphocytes and monocytes layers (1.077 kg/L).
On the one hand, density gradients are cytotoxic and can potentially
impair prostate cell viability. On the other hand, the density of
prostate cells in vivo may be more heterogeneous than the LNCaP cell
line. Thus, according to the manufacturer's recommendations, some or
even all circulating cancer cells will not be recovered at the expected
interface. Because density gradients are based on physical features of
nucleated cells, the respective concentrations of cancer and healthy
blood cells will not change their position in the gradient layers. PSMA
and PSA RT-nested PCR assays on samples treated by hypo-osmotic
lysis were consistently positive, whereas a greater variability was
observed with gradient separations (data non shown). Therefore, at any
tumor cell ratio, gradient-based optimization of the harvest of
prostate cells requires recovery of the whole fraction above the first
interface and, for some gradients, part of the third fraction. The
large volume collected (and the corresponding amount of PNBC cells)
will lower the relative concentration of epithelial cells. Furthermore,
the extraction of specific prostate messenger RNA by these preparations
is no better than that obtained by ammonium chloride procedure (Fig. 1b
).
In contrast to gradient separation methods, hypo-osmotic red blood cell lysis with ammonium chloride, which appears smoother than DEPC-treated water, fulfills the requirement for an efficient PNBC extraction mean.
Our goal was to assess a panel of conditions widely used to harvest circulating epithelial cells to delineate guidelines for extraction of cancer cells before RT-PCR. The model of LNCaP cells used in this study may not exactly mimic the behavior of prostate cancer cells in vivo, where density is more heterogeneous. The current results clearly indicate that time-consuming and expensive density gradients for blood cell separation do not give better results than easy and cost-effective red blood cell lysis.
|
Acknowledgments
This study was supported in part by grants from the Association de Recherche contre le Cancer (ARC 1367) and from the Association de Recherche sur les Tumeurs de la Prostate (ARTP).
Footnotes
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. Cristofanilli, G. De Gasperis, L. Zhang, M.-C. Hung, P. R. C. Gascoyne, and G. N. Hortobagyi Automated Electrorotation to Reveal Dielectric Variations Related to HER-2/neu Overexpression in MCF-7 Sublines Clin. Cancer Res., February 1, 2002; 8(2): 615 - 619. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
