Clinical Chemistry 55:3

587–598 (2009) Letters to the Editor

Comparison of 7 Methods for volved diverse strategies for DNA termined by application of the
Extracting Cell-Free DNA from isolation. The experimental set-up Dunnett T3 test for pairwise multi-
Serum Samples of Colorectal involved DNA isolation from a ple comparisons (Table 1). Because
Cancer Patients 2-mL aliquot of serum in duplicate of the insufficient balance of DNA
followed by DNA quantification by in the other 2 samples, we carried
To the Editor: the fluorescent Quant-iT dsDNA out the second round of rPCR on
HS assay (Invitrogen) and a Taq- bisulfite-converted DNA in 10
The presence of cell-free DNA of
man real-time PCR (rPCR) tech- pooled samples. Among 3 methods
tumor origin in serum or plasma of
nique on the cadherin 1, type 1, E- evaluated, the NaI method exhib-
cancer patients (1 ) has triggered
cadherin (epithelial) (CDH1) gene. ited the most abundant gene copy
numerous studies to explore the
To exclude false results in the numbers of ACTB (amplicon size:
diagnostic and prognostic poten-

Downloaded from https://academic.oup.com/clinchem/article/55/3/587/5629395 by guest on 31 May 2023

CDH1 amplification due to various 115 bp), with the median of ACTB
tial of circulating DNA. This DNA,
PCR inhibitors present in DNA ex- being 7 times higher than that gen-
present only in minute concentra-
tracts, the isolated DNA was fur- erated by the QIAamp DNA blood
tions in plasma or serum, is highly
ther quantified by Taqman rPCR kit, although the difference did not
fragmented (2 ), a condition that
on bisulfite-converted DNA be- reach the statistical significance
often leads to substantial loss of
cause the procedure of bisulfite- (Table 1).
DNA of small sizes in the course of
conversion of DNA removes many The higher recovery of DNA
DNA isolation. The lack of consen-
PCR-inhibitory components such obtained with the NaI and PCI gly-
sus regarding which extraction
as proteins, EDTA, and ethanol cogen procedures was also revealed
method is better for the efficient
(EpiTech bisulfite kit, Qiagen). A on the agarose gel, which showed
capture of such DNA might be par-
fragment devoid of the CpG site of much stronger DNA signals along
tially responsible for the large dis-
the actin, beta (ACTB) gene was each of 2 lanes (electrophoretic im-
parities in the literature, which are
the target for this round of rPCR. age available on request). Interest-
reflected in reports of total concen-
All 3 assays for DNA quantification ingly, substantial amounts of small
trations of plasma or serum DNA
were carried out in duplicate, and DNA fragments were recovered
alone (3 ) or DNA integrity mea-
the DNA/gene amount in samples with these 2 methods, a result that
surement as a diagnostic or prog-
was interpolated by reference to was not achievable with the other 5
nostic tool. The ability to detect
corresponding standard curves protocols (Table 1). Furthermore,
mutated v-Ki-ras2 Kirsten rat sar-
generated by using 5 to 6 serial di- the size of these small fragments
coma viral oncogene homolog
lutions of a DNA standard. Appro- appeared to correspond to that of
(KRAS) DNA in serum has been
priate blanks were included in each nucleosomal DNA, i.e., approxi-
reported to vary with the chosen
run. For rPCR, a positive control mately 180 –220 bp or its multiples.
DNA isolation methods (4 ).
sample was run in each plate to Because circulating DNA is highly
We evaluated in parallel 7 iso-
control the interplate variation. fragmented, any isolation method
lation approaches (Table 1) by ex-
The 7 extraction methods that favors capture of fragmented
tracting cell-free DNA from 12
showed remarkable differences in DNA from serum or plasma will be
pooled sera obtained from 67 colo-
the recovery of DNA from serum useful for a variety of downstream
rectal cancer patients and grouped
(Table 1). The phenol-chloroform applications in modern clinical
on the basis of TNM tumor staging
procedure (PCI-glycogen), so- and translational research labora-
[tumor extent (T), spread to lymph
dium iodide method (NaI tories. These include efficient de-
nodes (N), and metastasis (M)].1
method), and QIAamp DNA tection of mutations and capture of
The approaches we evaluated in-
blood kit generated significantly DNA methylation markers from
higher yields of DNA, assessed by serum or plasma.
fluorescent measurement, than the In comparing the PCI-glyco-
1
A total of 67 Chinese patients with sporadic colo- other 4 methods (all P ⬍ 0.05). As- gen approach and the NaI method,
rectal cancers who underwent potentially curative sessed by rPCR targeting on CDH1 we found that the latter was not
surgical resection at the Department of Colorectal
Surgery, Singapore General Hospital, between
(amplicon size: 68 bp), the NaI only superior to the former in
February 2004 and November 2005, participated method was ranked top in the list, terms of DNA quantity, as assessed
in the present study. None of these patients had a and statistical significance (P ⬍ by 2 rounds of rPCR, but also was
known history of familial adenomatous polyposis,
hereditary nonpolyposis colorectal cancer, or other
0.05) was achieved in all pairwise simpler, more rapid, and less costly
types of cancers. None of them received preoper- comparisons except with the PCI- (data available on request).
ative chemotherapy or radiotherapy. The institu- glycogen approach. The latter, To summarize, the results of
tional review board of Singapore General Hospital
approved this study, and all study participants however, outperformed only the the present study, which involved 7
provided written informed consent. guanidine-resin procedure as de- isolation methods performed via

587
Letters to the Editor

Table 1. Yield and fragment size of serum DNA isolated by 7 methods.

DNA yield from 2 mL of serum

By rPCR targeting
ACTB on BSa
templates
By rPCR on CDH1 (relative amount Visible bands on
By fluorescent (relative amount to a to a positive electrophoresis from 1 mL
assay, ng positive control) control) of serum

Extraction method n Median (SE) n Median (SE) N Median (SE) ⬵200 bp ⬵400 bp ⬵500 bp
b c d

Downloaded from https://academic.oup.com/clinchem/article/55/3/587/5629395 by guest on 31 May 2023

PCI-glycogen 12 367.958 (94.645) 12 332.220 (119.031) 10 15.507 (8.127) Yes Yes Yes
NaI methode 12 306.040 (61.228)c 12 391.735 (89.558)f 10 17.476 (9.915) Yes Yes Yes
Guanidine-resin methodg 6 8.928 (0.364) 12 1.145 (0.352) ND No No Yes
QIAamp DNA Blood Midi kit with 12 228.915 (38.162)c 12 69.934 (14.869 10 2.247 (2.601 No No Yes
carrier RNAh
ChargeSwitch 1-mL serum kiti 12 83.165 (13.370) 12 74.978 (16.612) ND No No Yes
ZR serum DNA kitj 6 15.363 (6.580) 12 6.577 (5.749) ND No No Yes
Puregene DNA purification System 12 59.200 (11.652) 12 42.094 (19.447) No No Yes
Cell and Tissue Kitk
a
BS, bisulfite-converted; ND, not detected.
b
Phenol-chloroform method with addition of glycogen.
c
DNA yield was significantly higher than that isolated by the methods without any mark in the column (P ⬍ 0.05, by Friedman analysis followed by Dunnett T3
test for pairwise multiple comparisons), but there was no difference among marked methods.
d
CDH1 quantity was second highest among all 7 methods but statistical significance was reached only in pairwise comparison with the guanidine-resin method
(P ⬍ 0.05).
e
Sodium iodide method (protocol available on request).
f
CDH1 quantity was highest among all 7 methods and of statistical significance in all pairwise comparisons except with the PCI-glycogen procedure (P ⬍ 0.05).
g
Adapted from the method described by Wang et al. (4 ).
h
Performed according to manufacturer’s instructions except with addition of 3 ␮g of carrier RNA into each sample; catalog no. 51183, lot no. 127136152.
i
Performed according to manufacturer’s instructions, catalog no. CS 11040, lot no. 1409017.
j
Performed according to manufacturer’s instructions, catalog no. D3013, lot no. 32-200607019.
k
Performed according to manufacturer’s instructions, part no. D-5500A, lot no. GS18781.

quantification of serum DNA by 3 Employment or Leadership: S.L. Fong, Singa- (plasma/serum) of cancer patients. Cancer Me-
pore General Hospital; K.W. Eu, Singapore tastasis Rev 1999;18:65–73.
assays and examination of frag- 2. Jahr S, Hentze H, Englisch S, Hardt D, Fackelmayer FO,
General Hospital; Y. Liu, Singapore General
ment sizes of DNA isolated by elec- Hospital. Hesch RD, Knippers R. DNA fragments in the blood
trophoresis, indicated that the NaI Consultant or Advisory Role: None de- plasma of cancer patients: quantitations and evidence
procedure consistently revealed clared. for their origin from apoptotic and necrotic cells. Can-
cer Res 2001;61:1659–65.
better performance. Therefore, Stock Ownership: None declared.
3. Boddy JL, Gal S, Malone PR, Harris AL, Wain-
this procedure appears to be a suit- Honoraria: None declared.
scoat JS. Prospective study of quantitation of
Research Funding: Grant from the Na-
able method for cell-free DNA ex- tional Medical Research Council (NMRC),
plasma DNA levels in the diagnosis of malignant
traction for many downstream ap- versus benign prostate disease. Clin Cancer Res
Singapore. 2005;11:1394 –9.
plications in cancer research. Expert Testimony: None declared. 4. Wang M, Block TM, Steel L, Brenner DE, Su YH.
Role of Sponsor: The funding organizations Preferential isolation of fragmented DNA en-
played no role in the design of study, choice of hances the detection of circulation mutated K-
enrolled patients, review and interpretation of ras DNA. Clin Chem 2004;50:211–3.
Author Contributions: All authors confirmed
they have contributed to the intellectual content of data, preparation or approval of manuscript.
this paper and have met the following 3 require- Siew Lee Fong2
Acknowledgments: We would thank Bing
ments: (a) significant contributions to the concep- Zhou Li (Research Operation, Duke-NUS Ji Tuan Zhang3
tionanddesign,acquisitionofdataoranalysis,and Graduate Medical School, Singapore) for Che Kang Lim4
interpretation of data; (b) drafting or revising the his assistance in the quantification of DNA. Kong Weng Eu2
article for intellectual content; and (c) final ap-
proval of the published article.
Yanqun Liu2*
Authors’ Disclosures of Potential Conflicts of References 2
Department of Colorectal Surgery
Interest: Upon manuscript submission, all au-
thors completed the Disclosures of Potential Con- 1. Anker P, Mulcahy H, Chen Xq, Stroun M. Detec- Singapore General Hospital
flict of Interest form. Potential conflicts of interest: tion of circulating tumour DNA in the blood Singapore

588 Clinical Chemistry 55:3 (2009)

 Letters to the Editor

3
Department of Orthopedic Surgery arate detection antibodies for am- pounds were separated on the basis
Yong Loo Lin School of Medicine phetamine and methamphetamine. of retention time (methamphet-
National University of Singapore The ELISA cutoff for these drugs is amine 5.9 min; phentermine 6.6
Singapore 20 ng/g. All positive screen results are min). Two transitions were selected:
4
Department of Clinical Research confirmed by GC-MS. The quantifying transition was m/z
Singapore General Hospital We report the investigation of 150.2–91.1 and the qualifying transi-
Singapore an unconfirmed positive amphet- tion was m/z 150.2– 65.1. Ion ratios
amine result. ELISA assay of the were within ⫾20% to meet the crite-
* Address correspondence to this author at: meconium specimen in question rion for a positive result. The moni-
Department of Colorectal Surgery was positive for amphetamine but tored transition for the d5-meth-
Singapore General Hospital negative for methamphetamine. The amphetamine was m/z 155.2–92.1.

Downloaded from https://academic.oup.com/clinchem/article/55/3/587/5629395 by guest on 31 May 2023

Outram Road confirmation assay failed to detect The ELISA kit insert indicates
Singapore 169608 amphetamine, methamphetamine, that phentermine has 89% cross-re-
Fax 65-62262009 3,4-methylenedioxyamphetamine, activity at 25 ng/g with the amphet-
E-mail [email protected] or 3,4-methylenedioxy-N-methyl- amine antibody, but does not cross-
amphetamine. The specimen was react with the methamphetamine
DOI: 10.1373/clinchem.2008.110122 then analyzed with a more compre- antibody. Phentermine was fortified
hensive liquid-chromatography– at 3 concentrations (100, 200, and
tandem mass-spectrometry method 400 ng/g) and run as unknowns in
to scan for possible candidates that the amphetamine and methamphet-
Cross-Reactivity of could elicit a positive screen result. amine ELISA. The phentermine-
Phentermine with an Briefly, 0.25 g of meconium was fortified samples showed changes in
Immunoassay Designed to extracted with 1 mL of 25 mmol/L binding (B/B0) of 44%, 29%, and
Detect Amphetamine in a phosphate buffer, pH 7.3, and cen- 16% in the amphetamine assay,
Meconium Specimen trifuged. Solid-phase extraction on a whereas the methamphetamine as-
250-␮L aliquot was perfomed with a say showed negligible change in
To the Editor: mixed-mode column. Deuterated binding (B/B0) at 88%, 86%, and
analogs were added as internal stan- 85%. B/B0 is defined as the absor-
Neonates exposed to drugs of abuse dards. The drugs were eluted with 2 bance reading of the sample (B) di-
in utero can experience prenatal mL of methanol: ammonium hy- vided by the absorbance reading of
drug dependence leading to with- droxide (98:2 vol:vol), evaporated to the zero-dose standard (B0). The
drawal symptoms and a number of dryness under nitrogen, and recon- change in B/B0 for phentermine ob-
other health problems (1 ). Early de- stituted in 50 ␮L methanol. Analyses tained by using both kits is shown in
tection of exposure is critical to guide were performed with an Agilent Fig. 1. Although phentermine and
necessary treatment and improve Technologies 1200 series liquid methamphetamine have the same
outcomes for these children. Meco- chromatograph coupled to a 6410 molecular weight of 149 Da, they dif-
nium begins to form in the digestive triple quadrupole mass spectrome- fer in the position of the additional
tract at 12–16 weeks gestation. Drugs ter operated in the positive electro- methyl group. Therefore, coupling
and metabolites collect in meco- spray mode and equipped with a of the antibody through the aro-
nium beginning at about 5 months Zorbax Eclipse XDB C18 column matic ring of amphetamine, rather
gestation. Thus, meconium testing (4.6 ⫻ 50 mm, 1.8 ␮m) at 40 °C and than the nitrogen, minimizes cross-
can identify exposure to drugs dur- an ammonium-acetate:methanol reactivity with methamphetamine
ing the last 4 months of a full-term mobile phase. The specimen was but still allows cross-reactivity with
pregnancy (2 ). found to contain 191 ng/g phenter- the amphetamine antibody.
Our laboratory uses ELISA mine, an anorexiant stimulant rec- To the best of our knowledge,
reagents (Immunalysis) to detect ommended for short-term use in the this is the only reported incidence of
drugs of abuse in meconium. Poor treatment of obesity. the confirmation and quantification
specificity of immunoassay reagents We quantified phentermine by of phentermine in a meconium
for amphetamines is well character- using d5-methamphetamine as the specimen. Phentermine is a preg-
ized and as a result, specimens that internal standard. Because the mo- nancy category C drug, indicating it
test positive for amphetamines by lecular weight of methamphetamine should not be administered during
immunoassay are routinely tested and phentermine are the same, the pregnancy unless clearly needed be-
by a second analytical method to transitions for liquid chromatogra- cause no data are available to deter-
prevent false-positive results. Our phy–tandem mass spectrometry mine if it adversely affects the fetus.
ELISA screen for meconium has sep- were also the same, so the com- Although no evidence of toxicity or

Clinical Chemistry 55:3 (2009) 589