Aptamers evolved from live cells as effective
molecular probes for cancer study
Dihua Shangguan*, Ying Li†
, Zhiwen Tang*, Zehui Charles Cao*, Hui William Chen*, Prabodhika Mallikaratchy*,
Kwame Sefah*, Chaoyong James Yang*, and Weihong Tan*‡
*Center for Research at the Bio͞Nano Interface, Department of Chemistry, Shands Cancer Center, University of Florida Genetics Institute and
McKnight Brain Institute, and †Department of Pathology, Shands Cancer Center, Shands Hospital, College of Medicine, University of Florida,
Gainesville, FL 32611-7200
Edited by Mark A. Ratner, Northwestern University, Evanston, IL, and approved June 13, 2006 (received for review March 31, 2006)
Using cell-based aptamer selection, we have developed a strategy
to use the differences at the molecular level between any two
types of cells for the identification of molecular signatures on the
surface of targeted cells. A group of aptamers have been generated
for the specific recognition of leukemia cells. The selected aptamers
can bind to target cells with an equilibrium dissociation constant
(Kd) in the nanomolar-to-picomolar range. The cell-based selection
process is simple, fast, straightforward, and reproducible, and,
most importantly, can be done without prior knowledge of target
molecules. The selected aptamers can specifically recognize target
leukemia cells mixed with normal human bone marrow aspirates
and can also identify cancer cells closely related to the target cell
line in real clinical specimens. The cell-based aptamer selection
holds a great promise in developing specific molecular probes for
cancer diagnosis and cancer biomarker discovery.
cell-based selection ͉ cell imaging ͉ DNA aptamers
Understanding of human diseases at the molecular level has
been extremely challenging because of the lack of effective
probes to identify and recognize distinct molecular features of
diseases. Cancers, as well as many other diseases, are originated
from the mutations of human genes. Such genetic alterations result
in not only different behaviors of the diseased cells, but also changes
of the cells at the morphological and molecular levels. Traditionally,
cancers are diagnosed mostly based on the morphology of tumor
tissues or cells. However, these morphologic features are difficult to
be used to carry out early cancer diagnosis or to evaluate the
complex molecular alterations that lead to cancer progression (1, 2).
Therefore, molecular characteristics, especially at the proteomic
level, should be used to classify cancers because of the direct
connection between genetic features and protein expression. Cancer
diagnosis based on molecular features can be highly specific and
extremely sensitive when incorporated with proper signal transduction
and amplification mechanisms. Nonetheless, identification of
molecular signatures of a particular cancer remains a great challenge,
if not impossible, which is reflected by the fact that very few
biomarkers are available for effective cancer diagnosis.
Molecular-level differences are present between any two given
types of cells, such as normal vs. tumor cells and tumor cell type 1
vs. type 2. These differences possess great significance in aiding the
understanding of the biological processes and mechanisms of
diseases. They could also be highly useful for disease diagnosis,
prevention, and therapy. However, identifying molecular differences
between any two types of cells is not an easy task with current
technologies. For example, discovery of unknown molecular features
of diseased cells by using molecular probes is almost impractical
because most of today’s methodologies rely on known biomarkers
for the development of corresponding molecular probes, which
has been proved insufficient for addressing many emerging medical
problems. Even if the molecular-level differences can be identified,
there is still a need to validate that the specific differences are
indeed meaningful and vital for the desired biomedical property or
the disease before any real clinical applications can be benefited.
Despite all of the difficulties, a practical strategy that could
compare cancer cells with normal cells and identify the differences
at the molecular level is highly desirable to facilitating the discovery
of molecular features of cancer cells. Here, we report a systematic
approach that can conveniently circumvent the limitations of
current technologies. A new class of molecular probes termed
aptamers have been isolated and identified for the recognition of
molecular differences expressed on the surface membranes of two
different types of cells. These probes can be used to specifically
detect target cancer cells based on molecular characteristics in the
presence of other cells, leading to effective disease studies and early
diagnosis.
Aptamers are ssDNA, RNA, or modified nucleic acids. They
have the ability to bind specifically to their targets, which range from
small organic molecules to proteins (3–5). The basis for target
recognition is the tertiary structures formed by the single-stranded
oligonucleotides (6). Aptamers are obtained through an in vitro
selection process known as SELEX (systematic evolution of ligands
by exponential enrichment) (7, 8), in which aptamers are selected
from a library of random sequences of synthetic DNA or RNA by
repetitive binding of the oligonucleotides to target molecules.
Aptamers have had many important applications in bioanalysis,
biomedicine, and biotechnology (9–12). Most aptamers reported so
far have been selected by using simple targets, such as a purified
protein. Recently, aptamer selection against complex targets, such
as red blood cell membranes and endothelial cells, was also
demonstrated (13–16). Compared with molecular probes currently
available for biomarker recognition, aptamers are emerging candidates
with ample potential due to their high specificity, low
molecular weight, easy and reproducible production, versatility in
application, and easy discovery and manipulation (17). Currently,
the application of aptamers toward medical research and application
is limited because of the lack of aptamers for systems of medical
relevance. Thus, the major goal in this study is the development and
utilization of a group of effective aptamers for leukemia studies.
To identify unique molecular features of target cancer cells,
we have developed a cell-based SELEX (cell-SELEX) for the
selection of a panel of target cell-specific aptamers. A counterselection
strategy is used to collect DNA sequences that only
interact with the target cells but not the control cells. Consequently,
aptamer candidates exclusively binding to the target
cells are enriched. The membrane protein targets of the selected
aptamers represent the molecular-level differences between the
two cell lines used in this study. Not only can molecular
signatures of the cancer cells be easily discovered, but probes that
can recognize such unique features with very high affinity and
specificity are also generated at the same time. More imporConflict
of interest statement: No conflicts declared.
This paper was submitted directly (Track II) to the PNAS office.
Abbreviations: ALL, acute lymphoblastic leukemia; SELEX, systematic evolution of ligands
by exponential enrichment; TAMRA, tetramethylrhodamine anhydride.
‡To whom correspondence should be addressed. E-mail: tan@chem.ufl.edu.
© 2006 by The National Academy of Sciences of the USA
11838–11843 ͉ PNAS ͉ August 8, 2006 ͉ vol. 103 ͉ no. 32 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0602615103
tantly, the use of a panel of probes has the clear advantage over
the single-biomarker-based assays in clinical practice, providing
much more information for accurate disease diagnosis and
prognosis. At the same time, the probes recognize the targets at
their native state, creating a true molecular profile of the disease
cells. This is important in clinical application of the molecular
probes. In addition, the aptamers selected from cell-SELEX
offer valuable tools for isolating and identifying new biomarkers
of the diseased cells if desired. The development of specific
probes for molecular signatures on the cancer cell surface will
provide new opportunities in ‘‘personalized’’ medicine.
Results and Discussion
Cell-SELEX for Enrichment of Aptamer Candidates for Target Cells.
The process of our cell-SELEX is illustrated in Fig. 1, and the
detailed procedures are provided in Experimental Procedures.
Two hematopoietic tumor cell lines were chosen as a model
system for our aptamer selection because they are well studied
and consist of relatively homogeneous tumor cells. In addition,
flow cytometry analysis can be easily carried out to monitor the
selection process and to evaluate the selected aptamers for their
capability of recognizing target cells. A cultured precursor T cell
acute lymphoblastic leukemia (ALL) cell line, CCRF-CEM, was
used as the target for aptamer selection. A B cell line from
human Burkitt’s lymphoma, Ramos, was used as the negative
control to reduce the collection of DNA sequences that could
bind to common surface molecules present on both types of cells.
In our selection, a library of ssDNAs that contained a 52-mer
random sequence region flanked by two 18-mer PCR primer
sequences was used. The library was incubated with the target
cells to allow binding to take place. The cells were then washed,
and the DNA sequences bound to the cell surface were eluted.
The collected sequences were then allowed to interact with
excess negative control cells, and only the DNA sequences
remaining free in the supernatant were collected and amplified
for the next-round selection. After multiround selection, the
subtraction process efficiently reduced the DNA sequences that
bound to the control cells, while those target-cell-specific
aptamer candidates were enriched.
The progress of the selection process was monitored by using
flow cytometry. DNA products collected after each round were
labeled with FITC dye and incubated with live cells. The
fluorescence intensity of the labeled cells measured by the flow
cytometry analysis represented the binding capacity of the
enriched DNA pool to the cells. With the increasing number of
selection cycles, steady increases in fluorescence intensity on the
CCRF-CEM cells (target cells) were observed (Fig. 2A), indicating
that DNA sequences with better binding affinity to the
target cells were enriched. Nevertheless, there was no significant
change in fluorescence intensity on the Ramos cells (control
cells). These results indicate that the DNA probes specifically
recognizing unique surface targets on CCRF-CEM cells were
isolated. The specific binding of the selected pools to the target
cells was further confirmed by confocal imaging (Fig. 2B). After
incubation with a tetramethylrhodamine anhydride (TAMRA)
dye-labeled aptamer pool, the CEM cells presented very bright
fluorescence on their periphery, whereas the Ramos cells displayed
weak fluorescence.
Identification of Aptamers for the Target Cells. It usually took Ϸ20
rounds to achieve excellent enrichment of aptamer candidates.
Fig. 1. Schematic representation of the cell-based aptamer selection. Briefly,
the ssDNA pool was incubated with CCRF-CEM cells (target cells). After washing,
the bound DNAs were eluted by heating to 95°C. The eluted DNAs were
then incubated with Ramos cells (negative cells) for counterselection. After
centrifugation, the supernatant was collected and the selected DNA was
amplified by PCR. The PCR products were separated into ssDNA for next-round
selection or cloned and sequenced for aptamer identification in the last-round
selection.
Fig. 2. Binding assay of selected pool with CCRF-CEM and Ramos cells. (A) Flow cytometry assay to monitor the binding of selected pool with CCRF-CEM cells
(target cells) and Ramos cells (negative cells). The green curve represents the background binding of unselected DNA library. For CEM cells, there was an increase
in binding capacity of the pool as the selection was progressing, whereas there was little change for the control Ramos cells. (B) Confocal imaging of cells stained
by the 20th-round selected pool labeled with tetramethylrhodamine dye molecules. (Upper Left) Fluorescence image of CCRF-CEM cells. (Upper Right) Optical
image of CCRF-CEM cells. (Lower Left) Fluorescence image of Ramos cells. (Lower Right) Optical image of Ramos cells.
Shangguan et al. PNAS ͉ August 8, 2006 ͉ vol. 103 ͉ no. 32 ͉ 11839
BIOCHEMISTRYCHEMISTRY
The highly enriched aptamer pools were cloned and sequenced
by using a high-throughput genome sequencing method. The
sequences were grouped based on the homology of the DNA
sequences of individual clones with each group containing very
similar sequences.
Twenty sequences were chosen for further characterization
because they were highly abundant in their family. The binding
assays of the selected sequences with target cells were performed
by using flow cytometry. Thirteen sequences revealed obvious
binding to CCRF-CEM cells. Moreover, the binding was not
interfered with by the addition of 1,000-fold excess of starting
DNA library. Except aptamers such as sgd2, sgc4, and its
homologue sgc4a, which could recognize both CCRF-CEM and
Ramos cells (data not shown), other aptamers only recognized
the target cell line, CCRF-CEM. Ten aptamers were confirmed
to have high affinity for CCRF-CEM cells with calculated
equilibrium dissociation constants (Kd) in the nanomolar-topicomolar
range, and their Kd are listed in Table 1. CD2, CD3,
CD4, CD5, CD7, and CD45 are the surface antigens expressed
on CCRF-CEM cells, and none of our tested aptamer sequences
showed any evidence of competition with antibodies against
these antigens (data not shown). This finding indicates that the
aptamers may interact with unique surface-binding entities.
Cell-SELEX Generates High-Affinity Molecular Probes for Target Cells.
We tested the individual aptamers. As shown in Fig. 3 A and B,
aptamers sga16 (Kd ϭ 5.00 Ϯ 0.52 nM) and its homologues sgc8
(Kd ϭ 0.80 Ϯ 0.09 nM) can specifically recognize the CCRFCEM
cells with high affinity. The specificity of both aptamers
was also observed directly by using confocal imaging. Intense
fluorescence from sga16 bound on the CCRF-CEM cell surface
was observed, whereas the Ramos cells had no obvious fluorescence.
These results clearly demonstrate the great potential of
using aptamers sga16 and sgc8 as excellent molecular probes for
CCRF-CEM-cell recognition.
It is worth noting that some of our selected aptamers can
identify binding entities expressed only by a small subset of target
cells. Sequences sgc3 (Kd ϭ 1.97 Ϯ 0.29 nM), sgc6 (Kd ϭ 8.76 Ϯ
0.62 nM), and sgd3 (Kd ϭ 3.58 Ϯ 0.58 nM) were found to bind
only to a small population of the CCRF-CEM cells (Ϸ20–40%
of the cells) (the second peak in Fig. 4A and the yellow area in
Fig. 4C) with high affinity, but they did not bind to Ramos cells.
Sgc3 and sgc6 are, in fact, homologues, whereas sequence sgd3
is very different from them. Confocal imaging also confirmed
that aptamer sgc3 strongly bound to a subset of CCRF-CEM
cells (Fig. 4B). These sgc3-labeled cells showed the same forward-
and side-scatter properties as the rest of the sgc3-negative
cells in flow cytometry assays, indicating that they were viable
cells. We have also immunophenotyped the sgc3-labeled cells
and confirmed that they were CD5- and CD7-positive neoplastic
T cells rather than contaminations from other types of cells (Fig.
4C). On the other hand, the sgc3-binding CCRF-CEM cells were
CD3-negative, implying that they might represent a unique
differentiation stage or phase of the cell cycles.
The reason behind our observations might be that a subpopulation
of the target cells had unique molecular signatures
expressed on the cell surface. It is interesting but not surprising
to see that the cell-based SELEX could generate molecular
probes specific for only a fraction of the target cells. Considering
the principles of cell-SELEX, any molecular differences between
the target and control cells could lead to the selection of
aptamers that can recognize such differences, no matter whether
they are present on all or just part of the target cells.
The Selected Aptamers Can Be Used for Highly Specific Recognition of
Target Cells in Real Biological Samples. To test the feasibility of the
selected aptamers as probes for specific molecular recognition,
FITC-labeled aptamers (sgc8, sgc3, sgd3, sgc4, and sgd2) and
monoclonal antibodies were used to detect CCRF-CEM leukemia
cells mixed with normal human bone marrow aspirates. The
human bone marrow aspirates consisted of mature and immature
granulocytes, nucleated erythrocytes, monocytes, mature and
immature B cells, and T cells. As expected, sgc8, sgc3, and sgd3
only recognized cultured leukemia T cells (CCRF-CEM) (Fig. 5)
and did not bind to normal CD3-positive T cells or any other
bone marrow cells. Aptamers sgc4 and sgd2 slightly bound to
mature and immature B cells, a subset of CD3-positive T cells,
nucleated erythrocytes from the human bone marrow, and
cultured leukemia T cells (CCRF-CEM) (data not shown).
Aptamers (sgc8, sgc3, sgd3, sgc4, and sgd2) were all found to
recognize other T cell ALL cell lines, Sup-T1, Molt-4, and Jurkat,
but not all could bind to cultured B cells and AML cells (Table 2).
Recognition of tumor cells in real clinical samples by our aptamers
was also tested. Patients’ bone marrow aspirates were examined
with FITC-labeled aptamers and monoclonal antibodies. The reTable
1. Binding affinity of the selected aptamer sequences
Sequence name Kd, nM
sgc3 1.97 Ϯ 0.29
sgc6 8.76 Ϯ 0.62
sgd3 3.58 Ϯ 0.58
sgc4 26.6 Ϯ 2.1
sgc4a 229 Ϯ 38
sgc5 113 Ϯ 41
sgc7 144 Ϯ 75
sgc8 0.80 Ϯ 0.09
sga16 5.00 Ϯ 0.52
Sgd2 7.21 Ϯ 0.89
Fig. 3. Characterization of selected aptamers. (A) Flow cytometry assay for the binding of the FITC-labeled sequences sga16 and sgc8 with CCRF-CEM cells (target
cells) and Ramos cells (negative cells). The green curve represents the background binding of unselected DNA library. The concentration of the aptamers in the
binding buffer was 250 nM. (B) Flow cytometry to determine the binding affinity of the FITC-labeled aptamer sequence sga16 to CCRF-CEM cells. The nonspecific
binding was measured by using FITC-labeled unselected library DNA.
11840 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0602615103 Shangguan et al.
sults (Table 2) revealed that none of aptamers (sgc8, sgc3, sgd3,
sgc4, and sgd2) could recognize the cancer cells from B cell
lymphoma patients but all were able to bind the cancer cells from
T cell ALL patients (Fig. 6), which were closely related to the
CCRF-CEM target cells used in our cell-SELEX. The capability of
the aptamers selected in our cell-SELEX for molecular diagnosis in
clinical practice is clearly demonstrated here.
The Binding Sites of the Aptamers on the Target Cells. To test
preliminarily whether the targets of the aptamers are membrane
proteins on the cell surface, we treated CCRF-CEM cells with
proteinases such as trypsin and proteinase K for a short time before
adding the aptamer to these treated cells. As shown in Fig. 7, after
treating the cells with trypsin or proteinase K for 10 min, aptamer
sgc8, sgc3, and sgd3 lost their binding to these cells, whereas the
interactions of aptamer sgd2 and sgd4 with the cells were not
affected. It can be deduced that the binding entities of aptamer sgc8,
sgc3, and sgd3 had been removed by the proteinases, indicating that
the target molecules were most likely membrane proteins. Interestingly,
the targets of aptamers sgd2 and sgc4 were clearly not
affected by the proteinases.
In conclusion, a cell-based SELEX strategy has been developed
to generate a panel of aptamers as useful molecular probes
to reveal the molecular-level differences between any two types
of cells. The selected aptamers have then been used for the
specific recognition of diseased cells. Molecular differences
between the target and control cells could be isolated easily,
providing an effective approach to the discovery of molecular
signatures of many other diseases. More importantly, detailed
knowledge of the distinct targets on the cell surface is not needed
before the selection, which could greatly simplify the process of
molecular probe development. The entire selection process is
simple, fast, reproducible, and straightforward, and the selected
aptamers can specifically bind to target cells with Kd in the
nanomolar-to-picomolar range. Some of the aptamers can recognize
a small subset of the target cells. Target cells mixed with
normal human bone marrow aspirate can be readily distinguished.
In addition, cancer cells from clinical patients’ specimens,
which are closely related to the target cells, were also
recognized by the selected aptamers. Furthermore, the aptamers
can be used to isolate the disease-specific protein targets to
facilitate the discovery of clinically important biomarkers. Our
preliminary results have suggested that the binding sites of the
selected aptamers are most likely proteins on cell-membrane
surfaces. The development of specific probes for molecular
signatures on the cancer cell surface will allow us to define
tumors, create tailored treatment regime for more ‘‘personalFig.
4. Aptamer sgc3 only recognizes a subset of CCRF-CEM cells. (A) Flow
cytometry assay for the binding of the FITC-labeled sequence sgc3 with
CCRF-CEM cells (target cells). The green curve represents the background
binding of unselected DNA library. The second peak of the red curve represents
the sgc3-labeled subset of cells. The concentration of the aptamer in the
binding buffer was 250 nM. (B) Fluorescence confocal images of CEM and
Ramos cells stained by sgc3 labeled with TAMRA. (Left) Fluorescence images
of CCRF-CEM cells and Ramos cells. (Right) Optical images of CCRF-CEM cells
and Ramos cells. (C) Flow cytometry assay for the binding of CCRF-CEM cells to
aptamer sgc3 and monoclonal antibodies against CD5, CD7, and CD3. The
yellow area represents the sgc3-labeled subset of cells. Aptamer sgc3 selectively
bound to a subpopulation of CCRF-CEM cells, which expressed abundant
CD7 and CD5 but not CD3. The final concentration of sgc3 in the binding
buffer was 250 nM.
Fig. 5. Molecular recognition of CCRF-CEM cells and human bone marrow
cells incubated with FITC-labeled sgc8, sgc3, and peridinin chlorophyll proteinlabeled
anti-CD45 antibody. The aptamer sgc8 or sgc3 and monoclonal antibodies
were incubated with the target CCRF-CEM cells and͞or bone marrow
cells. The sgc8 (A) and sgc3 (B) were able to recognize the target leukemia cells
selectively when CCRF-CEM leukemia cells were mixed with cells from human
bone marrow aspirates.
Table 2. Using aptamers to recognize cancer cells
Cell line sgc8 sgc3 sgc4 sgd2 sgd3
Cultured cell lines
Molt-4 (T cell ALL) ϩϩϩϩ ϩϩϩ ϩϩϩϩ ϩϩϩϩ ϩϩϩϩ
Sup-T1 (T cell ALL) ϩϩϩϩ ϩ ϩϩϩϩ ϩϩϩϩ ϩϩ
Jurkat (T cell ALL) ϩϩϩϩ ϩϩϩ ϩϩϩϩ ϩϩϩϩ ϩϩϩϩ
SUP-B15 (B cell ALL) ϩ 0 ϩϩ ϩ 0
U266 (B cell myeloma) 0 0 0 0 0
Toledo (B cell
lymphoma)
0 0 ϩϩϩϩ ϩϩϩϩ ϩ
Mo2058 (B cell
lymphoma)
0 ϩϩ ϩϩ 0 ϩ
NB-4 (AML, APL) 0 0 ϩϩϩ ϩϩϩϩ 0
Cells from patients
T cell ALL ϩϩ ϩϩϩ ϩϩϩ ϩϩϩ ϩϩϩ
Large B cell lymphoma 0 0 0 0 0
A threshold based on fluorescence intensity of FITC in the flow-cytometric
analysis was chosen so that 99% of cells incubated with the FITC-labeled
unselected DNA library would have fluorescence intensity below it. When the
FITC-labeled aptamer was allowed to interact with the cells, the percentage of
the cells with fluorescence above the set threshold was used to evaluate the
binding capacity of the aptamer to the cells. 0, Ͻ10%; ϩ, 10–35%; ϩϩ,
35–60%; ϩϩϩ, 60–85%; ϩϩϩϩ, Ͼ85%; AML, acute myeloid leukemia; APL,
acute promyelocytic leukemia.
Shangguan et al. PNAS ͉ August 8, 2006 ͉ vol. 103 ͉ no. 32 ͉ 11841
BIOCHEMISTRYCHEMISTRY
ized’’ medicine, monitor the response to therapy, and detect
minimal residual diseases.
Experimental Procedures
Cell Lines and Buffers. CCRF-CEM (CCL-119, T cell line, human
ALL), Ramos (CRL-1596, B cell line, human Burkitt’s lymphoma),
Toledo (CRL-2631, B cell line, human diffuse large-cell
lymphoma), Sup-T1 (CRL-1942, T cell line, human lymphoblastic
leukemia), Jurkat (TIB-152, human acute T cell leukemia),
Molt-4 (CRL-1582, T cell line, human ALL), SUP-B15 (CRL-
1929, B lymphoblast, human ALL), and U266 (TIB-196, B
lymphocyte, human myeloma, plasmacytoma) were obtained
from American Type Culture Collection. Mo2058 (Mantle-cell
lymphoma, Epstein–Barr virus-positive cell line) and NB-4
(acute promyelocytic leukemia) were obtained from the Department
of Pathology at the University of Florida. All of the cells
were cultured in RPMI medium 1640 (American Type Culture
Collection) supplemented with 10% FBS (heat-inactivated;
GIBCO) and 100 units͞ml penicillin–streptomycin (Cellgro).
Cells were washed before and after incubation with wash buffer
[4.5 g͞liter glucose and 5 mM MgCl2 in Dulbecco’s PBS with
calcium chloride and magnesium chloride (Sigma)]. Binding
buffer used for selection was prepared by adding yeast tRNA (0.1
mg͞ml; Sigma) and BSA (1 mg͞ml; Fisher) into wash buffer to
reduce background binding. Antibodies against CD2, CD3, CD4,
CD5, CD7, and CD45 were purchased from BD Biosciences.
Trypsin and proteinase K were purchased from Fisher Biotech.
SELEX Library and Primers. The HPLC-purified library contained
a central randomized sequence of 52 nucleotides flanked by two
18-nt primer hybridization sites (5Ј-ATA CCA GCT TAT TCA
ATT- 52-nt -AGA TAG TAA GTG CAA TCT-3Ј). An FITClabeled
5Ј primer (5Ј-FITC-ATA CCA GCT TAT TCA ATT-3Ј)
or a TAMRA-labeled 5Ј primer (5Ј-TAMRA-ATA CCA GCT
TAT TCA ATT-3Ј), and a triple-biotinylated (trB) 3Ј primer
(5Ј-trB-AGA TTG CAC TTA CTA TCT-3Ј) were used in the
PCRs for the synthesis of double-labeled, double-stranded DNA
molecules. After denaturing in alkaline condition (0.2 M
NaOH), the FITC-conjugated sense ssDNA strand was separated
from the biotinylated antisense ssDNA strand with streptavidin-coated
Sepharose beads (Amersham Pharmacia Biosciences)
and used for next-round selection. The selection
process was monitored by using flow cytometry.
SELEX Procedures. The procedures of selection were as follows. The
ssDNA pool (200 pmol) dissolved in 400 ␮l of binding buffer was
denatured by heating at 95°C for 5 min and cooled on ice for 10 min
before binding. Then the ssDNA pool was incubated with 1–2 ϫ 106
CCRF-CEM cells (target cells) on ice for 1 h. After washing, the
bound DNAs were eluted by heating at 95°C for 5 min in 300 ␮l of
binding buffer. The eluted DNAs were then incubated with Ramos
cells (negative cells; 5-fold excess than CCRF-CEM cells) on ice for
counterselection for 1 h. After centrifugation, the supernatant was
desalted and then amplified by PCR with FITC- or biotin-labeled
primers (10–20 cycles of 0.5 min at 94°C, 0.5 min at 46°C, and 0.5
min at 72°C, followed by 5 min at 72°C; the Taq polymerase and
dNTPs were obtained from Takara). The selected sense ssDNA
was separated from the biotinylated antisense ssDNA strand by
streptavidin-coated Sepharose beads (Amersham Pharmacia Biosciences).
For the first-round selection, the amount of initial ssDNA
pool was 10 nmol, dissolved in 1 ml of binding buffer, and the
counterselection step was eliminated. To acquire aptamers with
high affinity and specificity, the wash strength was enhanced
gradually by extending wash time (from 1 to 10 min) and increasing
the volume of wash buffer (from 0.5 to 5 ml) and the number of
washes (from three to five). Additionally, 20% FBS and a 50- to
300-fold molar excess of genomic DNA were added to the incubation
solution. After 20 rounds of selection, the selected ssDNA pool
was PCR-amplified by using unmodified primers and cloned into
Escherichia coli by using the TA cloning kit (Invitrogen). Cloned
sequences were determined by the Genome Sequencing Services
Laboratory at the University of Florida.
Flow-Cytometric Analysis. To monitor the enrichment of aptamer
candidates after selection, the FITC-labeled ssDNA pool was
incubated with 2 ϫ 105
CCRF-CEM cells or Ramos cells in 200
␮l of binding buffer containing 20% FBS on ice for 50 min. Cells
were washed twice with 0.7 ml of binding buffer (with 0.1%
NaN3) and suspended in 0.4 ml of binding buffer (with 0.1%
NaN3). The fluorescence was determined with a FACScan
cytometer (BD Immunocytometry Systems) by counting 30,000
events. The FITC-labeled unselected ssDNA library was used as
a negative control.
The binding affinity of aptamers was determined by incubating
CCRF-CEM cells (5 ϫ 105
) on ice for 50 min in the dark with
varying concentrations of FITC-labeled aptamer in a 500-␮l volume
Fig. 6. Molecular recognition of T-ALL cells in patient bone marrow aspirates
with FITC-labeled sgc8, sgc3, sgc4, sgd2, sgd3, and R-phycoerythrin-labeled
anti-CD7 antibody. The background was measured by using FITC-labeled
unselected library. The red dots represent T-ALL cells.
Fig. 7. Binding of aptamers sgc8 and sgc3 to trypsin-treated (A) or proteinase
K-treated (B) CCRF-CEM cells. The concentration of the aptamers in the
binding buffer was 250 nM.
11842 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0602615103 Shangguan et al.
of binding buffer containing 20% FBS. Cells were then washed
twice with 0.7 ml of the binding buffer with 0.1% sodium azide,
suspended in 0.4 ml of binding buffer with 0.1% sodium azide, and
subjected to flow-cytometric analysis within 30 min. The FITClabeled
unselected ssDNA library was used as a negative control to
determine nonspecific binding. All of the experiments for binding
assay were repeated two to four times. The mean fluorescence
intensity of target cells labeled by aptamers was used to calculate for
specific binding by subtracting the mean fluorescence intensity of
nonspecific binding from unselected library DNAs (18, 19). The
equilibrium dissociation constants (Kd) of the aptamer–cell interaction
were obtained by fitting the dependence of fluorescence
intensity of specific binding on the concentration of the aptamers
to the equation Y ϭ B max X͞(Kd ϩ X), using SigmaPlot (Jandel,
San Rafael, CA).
To test the feasibility of using aptamers for recognition of
cancer cells in real biological samples, FITC-labeled aptamers
were mixed with R-phycoerythrin- or peridinin chlorophyll
protein-labeled antibodies of CD2, CD3, CD4, CD5, CD7,
CD19, and CD45, respectively, and incubated with 2 ϫ 105
cancer cells and͞or 2 ϫ 105
cells in human bone marrow
aspirates. After washing as described above, fluorescence was
determined with a FACScan cytometer (BD Immunocytometry
Systems).
Confocal Imaging of Cell Bound with Aptamer. For confocal imaging,
the selected ssDNA pools or aptamers were labeled with
TAMRA. Cells were incubated with 50 pmol of TAMRAlabeled
ssDNA in 100 ␮l of binding buffer containing 20% FBS
on ice for 50 min. Other treatment steps were the same as
described in Flow-Cytometric Analysis. Twenty microliters of cell
suspension bound with TAMRA-labeled ssDNA was dropped
on a thin glass slide placed above a ϫ60 objective on the confocal
microscope and then covered with a coverslip. Imaging of the
cells was performed on an Olympus FV500-IX81 confocal
microscope. A 5-mW, 543-nm He-Ne laser was the excitation
source for TAMRA throughout the experiments. The objective
used for imaging was a PLAPO60XO3PH ϫ60 oil-immersion
objective with a numerical aperture of 1.40 (Olympus).
Proteinase Treatment for Cells. CCRF-CEM cells (5 ϫ 106
) were
washed with 2 ml of PBS and then incubated with 1 ml of 0.05%
trypsin͞0.53 mM EDTA in HBSS or 0.1 mg͞ml proteinase K in
PBS at 37°C for 2 and 10 min. FBS was then added to quench the
proteinases. After washing with 2 ml of binding buffer, the
treated cells were used for aptamer-binding assay as described in
Flow-Cytometric Analysis.
We thank Ms. Kim Ahrens for help with cell culture and flow cytometry
analyses, Ms. Regina Shaw and Dr. William Farmerie for help with DNA
sequencing, Ms. Hui Lin for help with DNA synthesis, and Mr. Patrick
Conlon for the schematic drawing. This work was supported by National
Institutes of Health grants and a National Science Foundation Nanoscale
Interdisciplinary Research Teams grant.
1. Luo, J., Isaacs, W. B., Trent, J. M. & Duggan, D. J. (2003) Cancer Invest. 21, 937–949.
2. Espina, V., Geho, D., Mehta, A. I., Petricoin, E. F., III, Liotta, L. A. &
Rosenblatt, K. P. (2005) Cancer Invest. 23, 36–46.
3. Osborne, S. E. & Ellington, A. D. (1997) Chem. Rev. 97, 349–370.
4. Nutiu, R. & Li, Y. (2005) Angew. Chem. Int. Ed. 44, 1061–1065.
5. Wilson, D. S. & Szostak, J. W. (1999) Annu. Rev. Biochem. 68, 611–647.
6. Breaker, R. R. (2004) Nature 432, 838–845.
7. Ellington, A. D. & Szostak, J. W. (1990) Nature 346, 818–822.
8. Tuerk, C. & Gold, L. (1990) Science 249, 505–510.
9. Fang, X., Sen, A., Vicens, M. & Tan, W. (2003) ChemBioChem 4, 829–834.
10. Guo, K., Wendel, H. P., Scheideler, L., Ziemer, G. & Scheule, A. M. (2005)
J. Cell. Mol. Med. 9, 731–736.
11. Yang, C. J., Jockusch, S., Vicens, M., Turro, N. & Tan, W. (2005) Proc. Natl. Acad. Sci.
USA 102, 17278–17283.
12. Liu, J. W. & Lu, Y. (2006) Angew. Chem. Int. Ed. 45, 90–94.
13. Morris, K. N., Jensen, K. B., Julin, C. M., Weil, M. & Gold, L. (1998) Proc. Natl.
Acad. Sci. USA 95, 2902–2907.
14. Blank, M., Weinschenk, T., Priemer, M. & Schluesener, H. (2001) J. Biol. Chem.
276, 16464–16468.
15. Daniels, D. A., Chen, H., Hicke, B. J., Swiderek, K. M. & Gold, L. (2003) Proc.
Natl. Acad. Sci. USA 100, 15416–15421.
16. Wang, C., Zhang, M., Yang, G., Zhang, D., Ding, H., Wang, H., Fan, M., Shen,
B. & Shao, N. (2003) J. Biotechnol. 102, 15–22.
17. Jayasena, S. D. (1999) Clin. Chem. 45, 1628–1650.
18. Davis, K. A., Abrams, B., Lin, Y. & Jayasena, S. D. (1996) Nucleic Acids Res.
24, 702–706.
19. Davis, K. A., Lin, Y., Abrams, B. & Jayasena, S. D. (1998) Nucleic Acids Res.
26, 3915–3924.
Shangguan et al. PNAS ͉ August 8, 2006 ͉ vol. 103 ͉ no. 32 ͉ 11843
BIOCHEMISTRYCHEMISTRY