| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Department of Biomedicine and Surgery, Division of Clinical Chemistry, University Hospital, S-581 85 Linköping, Sweden.
2
AB Sangtec Medical, S-161 02 Bromma, Sweden.
a Address correspondence to this author at: Department of Clinical Chemistry, University Hospital, S-581 85 Linköping, Sweden. Fax 46-13-223240; e-mail bertil.kagedal{at}lio.se
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: Mononucleated cells were isolated by Percoll centrifugation. RNA was isolated by each of three methods: UltraspecTM-II RNA isolation system, FastRNATM GREEN Kit, and QIAamp RNA Blood Mini Kit. cDNA was synthesized using random hexamer primers. A tyrosinase-specific product of 207 bp was amplified by PCR. As an internal standard (and competitor) we used a 207-bp cDNA with a base sequence identical to the tyrosinase target except for a 20-bp probe-binding region. The PCR products were identified by 2,4-dinitrophenol (DNP)-labeled probes specific for tyrosinase (5'DNP-GGGGAGCCTTGGGGTTCTGG-3') and internal standard (5'DNP-CGGAGCCCCGAAACCACATC-3') and quantified by ELISA.
Results: The calibration curves were linear and had a broad dynamic measuring range. A detection limit (2 SD above zero) of 48 transcripts/mL of blood was obtained from a low control. The analytical imprecision was 50% and 48% at concentrations of 1775 and 17 929 transcripts/mL (n = 12 and 14, respectively). With the cell line SK-Mel 28 added to blood and RNA extracted with the Ultraspec, Fast RNA, and QIAamp RNA methods, we found (mean ± SD) 1716 ± 1341, 2670 ± 3174, and 24 320 ± 5332 transcripts/mL of blood. Corresponding values were 527 ± 497, 2497 ± 1033, 14 930 ± 1927 transcripts/mL of blood when the cell line JKM86-4 was added. One high-risk patient was followed by repeated analysis of tyrosinase transcripts in blood. The melanoma marker 5-S-cysteinyldopa in serum and urine was within reference values, but tyrosinase mRNA was slightly increased (120–168 transcripts/mL of blood). The tyrosinase mRNA increased to 1860 transcripts/mL concomitant with the increase in 5-S-cysteinyldopa; later a spleen metastasis was found.
Conclusions: The results obtained with different RNA extraction methods illustrate the importance of quantitative methods for validation of methods. The use of QIAamp RNA improved the extraction efficiency considerably. Data from a case study suggest the assay is suitable in the follow-up of patients with high risk of developing metastases.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Since 1991, several investigations (2)(3)(4)(5) have shown that melanoma cells are present in the circulation of patients with disseminated cutaneous malignant melanoma, and that the presence of melanoma cells (read transcripts from melanoma cells) seems to be positively correlated to rapid progress of the disease (2)(4). At the time of diagnosis, 10–14% of patients with primary malignant melanoma were found to have detectable numbers of melanoma cells in blood with this technique (1)(4)(5). In cases with dissemination of cutaneous melanoma, 50–100% of the patients had melanoma cells in blood (1)(2)(4)(5). It was shown that patients with uveal melanomas had no detectable melanoma cells in their blood (3).
Several reports continue to be published on various aspects on the suitability of melanoma cells identification in blood for evaluation of tumor progression. Thus, Reinhold et al. (6) concluded that the analysis of blood samples by reverse transcription-PCR (RT-PCR)1 for tyrosinase mRNA is not suitable for the early detection of tumor progression in melanoma patients. A similar opinion was expressed by Gläser et al. (7). Farthmann et al. (8), on the other hand, concluded that RT-PCR positivity in early melanoma stages may indicate increased risk for the development of hematogenous metastases and may be of value as a progression marker.
Farthmann et al. (8) performed nested primer analysis on duplicates of each of two separated RNA preparations and obtained a definitive result in 98.4% of 123 melanoma patients examined. Their attempts to define the final results as either positive or negative illustrate the need for quantitative analysis of tumor-specific mRNA.
For semiquantitative analysis of the number of melanoma cells in peripheral blood, Brossart et al. (9) described a method combining RT-PCR and Southern blotting, and found that the amount of circulating cells correlated with the tumor burden. Attempts were also made by Curry et al. (10) to change their assay into a quantitative method by the use of an internal standard (IS). Calibration was made by comparing the quotient between the sample and IS signals with that from a selected melanoma cell line, thus indicating a specific number of cells in the patients’ blood. These "quantitative" methods, therefore, can be used when monitoring changes in the number of melanoma cells in blood.
We here present a method using internal standardization technique, according to the principles described by Lehtovaara et al. (11) for quantification of tyrosinase mRNA in blood. The suitability of the method is illustrated by comparison of the results with analysis of 5-S-cysteinyldopa in plasma and urine of a patient. The potential for further improvements of PCR methods—and perhaps the reason for the divergent results in clinical investigations—is illustrated by the finding of great variability in results when different procedures were used to extract RNA from blood samples.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To detect the tyrosinase and IS PCR products we used 2,4-dinitrophenol (DNP)-labeled probes. The tyrosinase-specific probe (5'DNP-GGGGAGCCTTGGGGTTCTGG-3') and the IS-specific probe (5'DNP-CGGAGCCCCGAAACCACATC-3') were synthesized by Scandinavian Gene Synthesis (Köping, Sweden).
cell culture
For in vitro studies, the cell lines SK-Mel 28 (ATCC) and
JKM86-4 (12) were used. SK-Mel 28 was cultured in minimal
essential medium with Hank’s salts containing 100 mL/L fetal
bovine serum (ICN Biomedicals), 2 mmol/L
L-glutamine, nonessential amino acids, 1 mmol/L
sodium pyruvate, 10 000 units/L penicillin, and 10 mg/L streptomycin.
JKM86-4 was grown in two parts of minimal essential medium with Earl’s
salts and 1 part of Leibovitz’s L-15 medium containing 150 mL/L fetal
bovine serum (ICN Biomedicals), 2 mmol/L
L-glutamine, 4.7 mg/L insulin (Sigma),
nonessential amino acids, 100 mg/L gentamicin, 10 000 units/L
penicillin, and 10 mg/L streptomycin. All reagents were from Life
Technologies. The cells were grown in humidified air with 5%
CO2.
construction of tyrosinase mrna calibrator and internal standard
The tyrosinase cDNA calibrator (TS) was a 207-bp fragment
generated by amplifying cDNA from SK-Mel 28 cells using the HTYR3 and
HTYR4 primers. The tyrosinase PCR product was subcloned in an AT vector
(Invitrogen), and the sequence was confirmed by automated fluorescence
sequencing. The TS fragment was purified using a QIAquick PCR
Purification Kit (Qiagen) and quantified by absorbance at 260 nm. The
TS was diluted to 1 000 000, 10 000, 1000, 100, and 10
molecules/25 µL and stored in aliquots at 2–8 °C. Amplification
of the TS samples in the presence of IS (see below) was used to
generate the calibration curve.
The IS was a 207-bp fragment constructed by a technique of oligonucleotide overlap extension and PCR amplification. The IS PCR amplification product was subcloned in an AT vector, and the sequence was confirmed by automated fluorescence sequencing. The IS fragment was purified using a QIAquick PCR Purification Kit and quantified by absorbance at 260 nm. The IS was included in the PCR reaction mixture at a defined concentration, 250 copies/reaction.
To achieve similar amplification efficiency between the IS and the tyrosinase target, the IS was designed to have a base sequence identical to that of the tyrosinase target except for a 20-bp probe-binding region.
patients and healthy controls
Unselected patient material was used to evaluate the precision of
the method. Blood samples from 65 patients, treated earlier for
malignant melanoma, were included in the study. Repeated samples were
taken from several patients, and a total of 226 samples were analyzed.
Healthy blood donors were used as negative controls. We also performed
case studies in five cases; one of them is reported to show the
possible use of the method. This patient underwent an axillary lymph
node dissection because of regional metastatic disease. Pulmonary x-ray
and ultrasound of the liver showed no metastases. Because the patient
was at high risk of developing systemic metastases, he started adjuvant
treatment with interferon-
2b (Introna®;
Schering-Plough).
blood collection and cell separation
Blood was collected in 5-mL Vacutainer Tubes (Becton Dickinson)
containing 0.5 mL of 0.13 mol/L sodium citrate. After mixing, 4-mL
aliquots of blood were mixed with 4 mL of phosphate-buffered saline
containing 0.13 mol/L citrate (PBS-citrate) in a 15-mL Falcon tube, and
4 mL of 67% Percoll (Pharmacia Biotech) in PBS was layered under the
blood with a syringe. The tube was centrifuged at 1000g for
30 min at 20 °C without braking. The mononuclear cell layer between
the plasma and the Percoll solution was collected and washed once with
11 mL of PBS-citrate by centrifugation at 800g for 10 min at
20 °C. The pellet was suspended in 0.25 mL of PBS-citrate.
rna extraction
RNA was extracted using the UltraspecTM-II
RNA isolation system (Biotecx). The samples were homogenized for three
30-s time periods using an Ultra Turrax (Tamro Lab) in the Ultraspec II
mixture. Chloroform was then added, and the sample was centrifuged. The
upper phase was collected, and isopropanol and
RNATacTM Resin were added. After centrifugation,
the pellet was washed twice with 750 mL/L ethanol, and then the pellet
was dried in a Savant vacuum extractor (Tectum). Finally, the sample
was resuspended in 120 µL of RNase-free water.
For comparison, RNA was also extracted with the FastRNATM GREEN Kit (Bio 101) according to the protocol from the manufacturer. Briefly, the sample was added to a FastPrepTM GREEN tube together with CRSR-GREEN, PAR, and CIA solutions. The tube was vortex-mixed twice for 1 min each, with cooling on ice between. After centrifugation, the upper phase was collected. CIA solution was added, and after another centrifugation, the upper phase was collected again. The RNA was allowed to precipitate in DIPS. After centrifugation, the pellet was washed twice with SEWS solution. Finally, the RNA was suspended in 120 µL of SAFE solution.
Alternatively, RNA was extracted from 1.5 mL of blood using the QIAamp RNA Blood Mini Kit (Qiagen) according to the manufacturer’s instructions. Briefly, the erythrocytes were lysed with EL buffer, and the remaining cells were pelleted by centrifugation. The leukocytes were then lysed with RLT buffer and homogenized with the QIAshredder. After addition of one volume of 700 mL/L ethanol, the homogenate was applied to a QIAamp RNA mini spin column. The column was washed once with RW1 buffer and twice with RPE buffer. Finally, the RNA was eluted with 2 x 60 µL of RNase-free water. We also extracted RNA from Percoll-separated cells. In this case, 4 mL of blood was used and the mononuclear cell pellet was lysed directly with the RLT buffer.
first-strand cdna synthesis
RNA (30 µL) was denatured at 70 °C for 5 min and then placed
on ice. To the RNA, 30 µL of cDNA mixture was added. The final
concentrations were as follows: 1x First Strand Buffer (Life
Technologies), 7.5 mmol/L dithiothreitol (Life Technologies), 500
µmol/L dNTPs (Pharmacia Biotech), 50 µmol/L random hexamer
(Pharmacia Biotech), 1 x 106 units/L RNasin
(Promega), and 10 x 106 U/L murine Moloney
leukemia virus reverse transcriptase (Life Technologies). A drop of
mineral oil was added to avoid condensation. First-strand cDNA
synthesis was performed at 40 °C for 45 min, and the reaction was
then heated for 5 min at 95 °C.
pcr
Twenty-five microliters of first-strand cDNA or tyrosinase
calibrator (TS) was amplified in a total volume of 100 µL containing
(final concentrations) 50 mmol/L Tris (pH 8.8), 25 mL/L glycerol, 15
mmol/L
(NH4)2SO4,
3 mmol/L MgCl2, 1 mL/L Tween 20, 0.5 mg/L
tRNA (Sigma), 800 µmol/L dNTPs (Pharmacia Biotech), 0.2 µmol/L
biotinylated HTYR3, 0.2 µmol/L HTYR4, 2.5 x
106 IS molecules/L, and 25 000 U/L AmpliTaq (PE
Applied Biosystems). The PCR was performed in a Perkin-Elmer TC 9600
thermal cycler for 35 cycles, and the PCR profile used was as
follows: 94 °C for 2 min as an initial denaturation step, followed
by 35 cycles of 30 s at 94 °C, 20 s at 60 °C, and
30 s at 72 °C. A final elongation for 10 min at 72 °C ended
the profile, and the amplification products were kept at 4 °C.
colorimetric detection
Binding buffer (100 µL; AB Sangtec Medical) was added to each
PCR tube, and 50-µL aliquots were transferred to two wells
of a streptavidin-coated microtiter plate (Labsystems Oy),
one well for measurement of the TS amplicon and one well for
measurement of the IS amplicon. The plate was placed in a shaker, and
the biotin-labeled strand was allowed to bind to the streptavidin.
After 15 min, the complementary strand was released by the addition of
50 µL of a solution containing 100 mmol/L NaOH and 300 mmol/L NaCl.
The plate was shaken for 1 min and washed six times with wash buffer
(25 mmol/L Tris-HCl, pH 7.5, 1.25 mmol/L NaCl, 2 mmol/L
MgCl2 and 3 mL/L Tween).
DNP-labeled probes (50 µL), one TS specific and one IS specific, were added to the respective samples. The probes were allowed to hybridize for 15 min at 55 °C.
After the wells were washed six times, 50 µL of anti-DNP antibody conjugated to alkaline phosphatase (AB Sangtec Medical) was added and incubated for 15 min at room temperature. The plate was washed six times, and 100 µL of substrate (4 g/L pNPP in substrate buffer; AB Sangtec Medical) was added. The absorbance at 405 nm was measured after 4, 10, and 30 min.
calculations for quantitative analysis
A blank, to which only substrate buffer was added, was included in
each experiment. The absorbance of the blank at 405 nm
(A405) was subtracted from the other
values. The subtracted values was used to calculate a ratio between the
TS DNA and the IS DNA for each sample.
A calibration curve was generated by plotting the mean ratio (n = 3) vs the TS concentration in a log-log scale, and an equation of the calibration curve was calculated using linear regression. This equation was used to calculate the tyrosinase transcript concentration of the blood samples.
analysis of 5-S-cysteinyldopa
5-S-Cysteinyldopa is produced in pigmented tissue and
increases in dissemination of malignant melanoma (13).
Therefore, 5-S-cysteinyldopa was measured in serum
(14) and urine (15) for comparison in a patient
with a high risk of developing melanoma metastases.
statistical method
The results of different RNA extraction kits were
evaluated by ANOVA.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Two hundred fifty copies of the IS were included in each
reaction. As can be seen in Fig. 1A
, the IS acts as an internal
standard when the TS is <101
copies/sample and
as a competitor at higher concentrations of TS. Table 1
shows the results for the calibration curves from 28 analyses
obtained over a period of 6 months. During this period, the same batch
of reagents for PCR and colorimetric detection was used. Within this
period, the imprecision was >70% when 
100 transcripts per reaction
were analyzed. For higher number of transcripts, the long-term
imprecision was <50%.
|
|
The ratio between the A405
values of TS and IS showed no tendency to change, and the absorbance
values were also stable over time, indicating that the reagents were
stable for at least 6 months (Fig. 2
).
|
validation of the method
For analysis of total variation during a 6-month period, RNA from
106 melanoma cells (SK-Mel 28) was prepared and
diluted 50 000-, 5000-, and 500-fold. The diluted RNA was aliquoted
and frozen. Subsequently, the different dilutions were included as
controls in the PCR. After cDNA synthesis, the sample was split into
duplicates and analyzed. For comparison, the results of these controls
were also converted to transcripts/mL of blood. Thus, the means (± SD)
of the controls were 24 (± 24) transcripts/mL of blood (n = 12),
1775 (± 882) transcripts/mL of blood (n = 16), and 17 929 (±
8601) transcripts/mL of blood (n = 14), respectively, and the
corresponding CVs were 102%, 50%, and 48%. When calculated from the
duplicates in the PCR analyses, the same mean values were obtained and
the CVs were 80%, 27%, and 27%, respectively. From the SD of 24
transcripts/mL obtained with the controls having the lowest number of
transcripts/mL, a detection limit of 48 transcripts/mL of blood was
obtained when the detection limit was defined as 2 SD above zero. The
three different dilutions corresponded to 5, 50, and 500 cells/mL of
blood. From the result of the two highest controls, 35 transcripts/cell
could be calculated for the cell line SK-Mel 28. This would give a
detection limit of 1.3 cells/mL for this cell line.
Blood samples (n = 18) from blood donors were also investigated.
None of them had a value higher than the detection limit of 48
transcripts/mL of blood. A subset of 16 samples was analyzed both by
the present method and by the method published by Smith et al.
(1). The negative samples (n = 4) had values below the
detection limit, and the positive samples had values ranging from 139
to 9140 transcripts/mL (Fig. 3
).
|
A total of 226 samples from 65 patients were analyzed, and the
imprecision was calculated from the results above the detection limit
(n = 56). The cDNA sample was split into duplicates for evaluation
of intraassay imprecision. The results (Table 2
) showed pooled CVs of 49–57% for samples containing different
number of transcripts/mL of blood. Higher intraassay CVs were obtained
for the samples (Table 2
) than for the calibrators (Table 1
). This
might be expected because the samples contain a complex matrix of
components from the blood that are not present in the calibrators.
|
comparison of different rna extraction kits
Melanoma cells were added to blood from six healthy donors to give
a concentration of 500 cells/mL. RNA was extracted by three different
kits: the Ultraspec-II RNA isolation system, the FastRNA GREEN
Kit, and the QIAamp RNA Blood Mini Kit. All of the kits used
mononuclear cells prepared from 4 mL of blood for RNA extraction. The
RNA was also extracted by the QIAamp RNA Blood Mini Kit from 1.5 mL of
whole blood. The experiment was performed with the melanoma cell lines
JKM86-4 and SK-Mel 28. Fig. 4
shows that QIAamp RNA was the best method (P <0.001
for both cell lines), giving at least 15-fold higher values than the
Ultraspec-II RNA extraction kit.
|
case study
The use of the method in monitoring melanoma patients is
illustrated by the results shown in Table 3
. A patient receiving adjuvant treatment with
interferon-2b (Introna; Schering-Plough) was monitored by
repeated analysis of 5-S-cysteinyldopa in the patient’s
blood and urine together with analysis of tyrosinase mRNA in his blood.
The 5-S-cysteinyldopa values were well within the reference
values until the last collected sample. His initial values for
tyrosinase mRNA were moderately increased (above the detection limit of
48 transcripts/mL) until the last collected sample, when it increased
markedly in a fashion similar to 5-S-cysteinyldopa. Before
that there were no clinical signs of metastases. A computerized
tomography of the abdomen was then performed, and a spleen metastasis
was found.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
For the determination of the imprecision of our method, we used unselected material from 226 samples sent to the laboratory. Of these samples, 25% had tyrosinase expression above the detection limit of 48 transcripts/mL of blood. This corresponds to 1.3 melanoma cells/mL of blood, if compared with the melanoma cell line SK-Mel 28. This cell line often has been used to denote the sensitivity of both qualitative (1)(8)(17) and quantitative methods (9)(18). Thus, the sensitivity of our method compares well with other methods as far as detection is concerned.
Great variations in the number of patients with a detectable number of melanoma cells in the blood have been reported (17). This may be attributable to several factors. Methodological differences must be taken into consideration. Most authors have used the primers designed by Smith et al. (1). However, RNA extraction methods differ. The use of whole blood in combination with phenol-chloroform extraction described by Chomczynski and Sacchi (19) is popular. Furthermore, extraction can be made from both whole blood and after density centrifugation. In the present report, we studied the extraction procedures by comparing three different commercially available methods. The Ultraspec-II RNA isolation system and the FastRNA GREEN Kit are both variants of the method of Chomczynski and Sacchi (19) and showed less recovery than the QIAamp RNA Blood Mini Kit. The QIAamp RNA kit, which uses silica-based membranes to bind RNA, gave a 15-fold higher result than the Ultraspec-II method (P <0.001). In the comparison between extraction from whole blood and after Percoll, similar separation results were obtained. To avoid the risk of losing cells from patients’ blood, we chose to use whole blood. Similar studies are needed to control the effectiveness and the variation in the RNA extraction step. These types of studies are not possible with qualitative methods but need quantitative methods.
The use of housekeeping genes to control for the variation in the RNA extraction and cDNA synthesis is not suitable when a fixed amount of blood is analyzed. Because the number of leukocytes in blood samples from healthy individuals can vary at least threefold, an extra variation is introduced if the result is correlated to a housekeeping gene. We instead used controls composed of cultured melanoma cells to evaluate the differences in the cDNA synthesis and PCR analysis. The two highest controls showed an imprecision of 50%, whereas the lower control had a CV of 102%.
We have obtained rather high overall CVs for the controls as well as the calibrators compared with other types of analytical methods. However, one has to bear in mind that this method is complex, with several steps where variation can be introduced. In addition, larger variations than in other methods have to be expected because differences that might be small at the beginning will also be amplified in the PCR. This could be one explanation for the high intraassay variation.
We also reported here the results from a patient followed for a longer period, using our PCR method together with the tumor marker 5-S-cysteinyldopa. In the first samples taken, the patient had tyrosinase expression above the detection limit, corresponding to 2–3 SK-Mel 28 cells/mL. The number of tyrosinase transcripts increased to 1860 transcripts/mL of blood before any clinical evidence of metastasis was found. This illustrates the usefulness of our method. If compared to the SK-Mel 28 cell line that we used in our in vitro experiments, 1860 transcript would correspond to ~50 cells. However, Curry et al. (10) showed with a quantitative method that the number of transcripts could vary tremendously in different cell lines.
It should be noted that serum and urine 5-S-cysteinyldopa increased at the same time as the number of transcripts (melanoma cells) in blood. Both of these changes can be considered as expression of increases in tumor burden, and when the results were reported, further clinical investigations revealed the presence of spleen metastases. These results are consistent with the prerequisite of melanotic cells, i.e., the cells must express tyrosinase. One reason why both of these methods might fail would be that the melanoma cells lack tyrosinase.
Further validation studies of methods remain to be done, and by improvement of methods, ambiguities regarding the clinical usefulness of the analysis hopefully will be resolved.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. Mitsuhashi, S. Tomozawa, K. Endo, and A. Shinagawa Quantification of mRNA in Whole Blood by Assessing Recovery of RNA and Efficiency of cDNA Synthesis Clin. Chem., April 1, 2006; 52(4): 634 - 642. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
).
(B), the ratio of the absorbance at 405 nm of TS to IS
is plotted against the number of tyrosinase transcripts to generate a
calibration curve. Each point represents a mean of triplicates;
bars, ± 1 SD. Sometimes the SD is hidden within the
symbol.
), 104 (
), 103 (
),
102 (
), and 10 (
