heterogeneity of 06-alkylguanine-dna-alkyltransferase …...vol. 4, 475-481, february 1998 clinical...
TRANSCRIPT
Vol. 4, 475-481, February 1998 Clinical Cancer Research 475
Heterogeneity of 06-Alkylguanine-DNA-alkyltransferase Measured
by Flow Cytometric Analysis in Blood and Bone Marrow
Mononuclear Cells1
Lili Liu, Keunmyoung Lee, Jane Schupp,
Omer N. Ko#{231},and Stanton L. Gerson2
Division of Hematology and Oncology and the Case Western Reserve
University/University Hospitals Ireland Cancer Center, Case Western
Reserve University School of Medicine, Cleveland, Ohio 44106-4937
ABSTRACT
Alkyltransferase (AGT) repairs alkylation at O6�
guanine in DNA and is a major determinant of susceptibilityto alkylating chemotherapeutic agents and carcinogens.
Using a newly developed flow cytometry assay with the
monoclonal anti-AGT antibody, mT3.1, we compared AGT
expression in single-cell suspensions with standard biochem-
ical and Western blot assays to validate the fluorescence-
activated cell sorting (FACS) method and develop potential
applications. From Chinese hamster ovary cells (CHO)
transfected with human 06-methylguanine-DNA methyl-
transferase cDNA, 6 CHO-06-methylguanine-DNA methyl.
transferase clones were isolated that expressed 0.3 to 64
fmol/p.g DNA (by biochemical assay) of human AGT. FACSyielded a linear relationship between mean fluorescence in-
tensity and both AGT activity by biochemical assay and
AGT protein by Western blot. Using this standard curve,FACS-analyzed AGT protein content in human peripheral
blood mononuclear cells (PBMCs) from normal donorsranged from 6.1 to 12.8 fmol/p.g DNA, similar to thoseobtained by biochemical assay and Western blot. This sug-
gests that the level of immunoreactive protein appears to bean accurate predictor of AGT activity in the steady state.
FACS-AGT in PBMCs from normal donors had a low indexof heterogeneity within the sample. In contrast, by FACS.
AGT analysis of human bone marrow samples and granu-locyte-colony-stimulating factor-mobilized PBMCs, AGTwas lower and had an 8-fold higher index of heterogeneitythan observed in PBMCs from normal donors. After treat-ment with 06-benzylguanine (06-bG), Western and FACS-AGT detected significant levels of AGT protein for up to
Received 8/25/97; revised 10/30/97; accepted 1 1/3/97.
The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to
indicate this fact.
I This work was supported in part by the Grants R01ES06288,
UO1CA75525, R01CA63193, and P30CA43703 from the USPHS.
2 To whom requests for reprints should be addressed, at Division ofHematology and Oncology. Department of Medicine and the Case
Western Reserve University/University Hospitals Ireland Cancer Cen-
ter, BRB-3, 10900 Euclid Avenue, Cleveland, OH 44106-4937. Phone:
(216) 368-1 176: Fax: (216) 368-1 166; E-mail:[email protected].
24 h, whereas biochemical assay showed AGT activity lessthan 5% of the basal level. Because only the biochemicalassay accurately measures net AGT activity, the AGT-FACSassay will not be useful in clinical trials to assess the efficacy
of 06-bG or other AGT inhibitors. Thus, AGT-FACS canrapidly assess the heterogeneity of steady-state AGT in sin-gle-cell suspensions and may be useful for assay in lympho-cytes, bone marrow cells, leukemic myeloma plasma cells, or
cells transfected with the AGT gene; Western blot analysis is
better for small samples such as tumor biopsies, whereasbiochemical assay is best able to measure enzyme activityand its inactivation by 06-bG or other agents.
INTRODUCTION
Certain environmental chemical agents and several chemo-
therapeutic agents used in patients with cancer are known to
alkylate DNA at the O� position of guanine, resulting in lesions
that are both mutagenic and cytotoxic. It has been observed that
AGT3 is a critical protein for repair of 06-alkylguaninc-DNA
adducts, protecting cells from the cytotoxic, mutagenie, and
carcinogenic effect of alkylating agents (1-3).
AGT is the only protein to accept an alkyl group from the
06 position in DNA on its own cysteine residue forming S-
alkylcysteine, resulting in irreversible inactivation. Regenera-
tion of AGT requires protein synthesis and takes 12-24 h (4).
Thus, the number of protein molecules present in cells equals
the number of 06-alkylguanine adducts that can be repaired (5,
6). The cellular level of this protein is a useful biomarker for
predicting cellular response or resistance to DNA damage in-
duced by alkylating agents. A number of different methods for
determination of AGT have been reported (3-8), all of which
measure mean AGT in the sample. The most specific assay is
the HPLC method that separates 06-mcthylguanine and N7-
methylguanine, allowing an internal control for the level of
substrate present in the methylated DNA (9). However, this
assay requires special equipment, is not conveniently estab-
lished in many laboratories, and might not be suitable for rapid
assay of AGT in a large number of samples. In addition, assay
for AGT activity in individual cells has been hampered by the
low amount of the protein. In this report, we describe a sensitive
flow cytometry analysis using the mouse monoclonal anti-
human AGT antibody (mT3.l). Cultured cells, human PBMCs,
and bone marrow cells were analyzed by FACS and compared
3 The abbreviations used are: AGT, alkyltransferase; FACS, fluores-
cenee-aetivated cell sorter: CHO, Chinese hamster ovary cells; MGMT,
06-methylguanine-DNA methyltransferase; PBMC, peripheral blood
mononuclear cell; G-CSF, granuloeyte-eolony-stimulating factor; MFI,
mean fluorescent intensity; 06-bG. 06-benzylguanine; IH, index of
heterogeneity; HPLC, high-performance liquid chromatography.
Research. on August 4, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
4000
BRegion PScatterogram Region R
U)C
00
CE0
C
000c�
La.
0
E 0
Region IsoPD� �.. F
0
U) 0’
�0 0�or’),
0
Cl)0� C’).
00
0 �
0
Ii�� -
476 Heterogeneity of 06-Alkylguanine-DNA-alkyltransferase
-C0)a)
I
Cl)(I)
0 D � 200 � �4bo #{149}mo 800 1doo
FSC-Height
Scatteroram Isotype
0� 2b0 400 600 800 idooFSC-Height
1� 5� � 410 10 10 10 10
FL2-Height
0 �1 i�2 �
FL2-Height
Region IsoR
-rnnui . .��;0 �o’ io2 �
FL2-Height
Fig. 1 Sufficient separation of
AGT� cell population. A, the for-
ward and right angle scatter gates
on K562 cells transfected with
human MGMT and cells were
stained with antibody mT3.l. B,histogram of gate P was pre-
sented AGT� cell population. C,histogram of gate R was predom-
inantly cellular debris and non-
permeabilized cells. D-F, K562
cells stained with mouse IgG1
isotype control. FSC, forward
scatter; SSC, side scatter; FL2,
fluorescence intensity; IsoP, iso-
type antibody, region P; IsoR,
isotype antibody, region R.
with biochemical measurement for AGT activity and Western
blotting of AGT protein. AGT-FACS identified heterogeneity in
single-cell suspensions and could be useful in clinical samples
of normal bone marrow, leukemia, mycloma, circulating tumor
cells, or to monitor MGMT-transduced hematopoictic cells dur-
ing gene therapy.
MATERIALS AND METHODS
Chemicals and Reagents. Mouse monoclonal antibody
(mT3.1) was kindly provided by Dr. Darell D. Bigner (Duke
University Medical Center, Durham, NC) and Dr. Thomas P.
Brent (Department of Pharmacology, St. Jude Hospital for Chil-
dren, Memphis, TN). 06-bG was generously provided by Dr.
Robert Mosehel (Frederick Cancer Research and Development
Center, National Cancer Institute, Frederick, MD). Stock solu-
tions were made in DMSO.
Tnansfection of Human MGMT Vector into CHO Cells.Plasmid pHM expressing human MGMT in CHO cells (CHO/
phMGMT) was constructed by inserting the human MGMT
eDNA coding sequence into the unique NcoI and BamHI re-
striction sites of a retroviral vector plasmid, pHM (10). Six �.ag
of plasmid DNA were transfected into 1.8 X 106 cells with
Lipofectaminc (Life Technologies, Inc., Gaithcrsburg, MD) fol-
lowing the manufacturer’s protocol; meanwhile, 0.6 p.g of
pSV2neo plasmid DNA was cotransfected, because this vector
plasmid lacks a neo cassette for selection. The transfected cells
were selected by addition of G418 (0.1%, w/v) into the media.
Six individual colonies were isolated that expressed variable
levels of human AGT.
Isolation Mononuclear Cells. Blood and bone marrow
samples were collected from normal donors who gave informed
consent under an Institutional Review Board approved protocol.
Mononuclear cells were isolated by Ficoll-Hypaquc discontin-
uous gradient separation. Cells at the interface were removed
and washed twice with PBS at 4#{176}C.PBMCs were also collected
from patients undergoing leukopheresis after G-CSF growth
factor treatment to increase the proportion of hematopoictic
progenitor cells in the blood. These patients subsequently un-
derwent high-dose chemotherapy with reinfusion of these pro-
genitor cells.
FACS Analysis of AGT. Cells were prepared for detec-
tion of AGT protein using a method for intracellular staining. To
develop this method, several detergents were tested to determine
which was most efficient in permeabilizing cell membranes.
Twecn-20 permeabilized cells better than either Saponin or
Triton X. Briefly, after cells were stabilized for 30 mm using 1%
paraformaldehyde, the membranes were permeabilized by incu-
bating in 1% Tween 20 for 30 mm at 37#{176}C,followed by
incubation in 2% BSAIPBS for 30 mm. Nonspecific binding
sites were blocked for 30 mm at 22#{176}Cwith 10% normal goat
serum. Human AGT antibody mT3.l (7 p.g/ml) was added, and
cells were incubated at 4#{176}Covernight. Cells were washed twice
with 2% BSAIPBS and incubated with secondary antibody [goat
anti-mouse IgG-1 (gamma) phycoerythrin conjugated] for 1 h at
4#{176}C.After washed as above, cells were resuspended in PBS +
BSA for FACS analysis. Flow cytometry was performed using
FACScan (Becton Dickinson, Mountain View, CA), equipped
with an argon laser. MFI of permeabilized cells was determined
by subtracting the nonspecific background (isotype control)
from the fluorescence intensity of cells stained with specific
AGT antibody (mT3.1). Light scatter was used for gating on
permeabilized cells. To further analyze the heterogeneity of
AGT expression, we used the parameter, IH [width at half peak
(in units of signal intensity)IMFI], based on the histogram of
fluorescent staining from each sample. A larger IH means that
the cell population has more heterogeneous expression of AGT.
Alkyltransferase Assay. The cell pellets were resus-
pended in 0.5 ml cell extract buffer [70 msi HEPES (pH 7.8), 0.1
mM EDTA, 50% glycerol, 1 mr�i Dli.’, and 25 �iM spermidinc]
and sonicated three times for 10 s at 4#{176}Cto complete cell
disruption and centrifuged at 10,000 X g for 2 mm to remove
cellular debris. The protein and DNA content were measured,
Research. on August 4, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
A
1 2 3 4 5 6
C
B
Number of CHO-MGMT clones>0
n.L)
AGT �
zea
.�‘E
.�: BQC)
�400.
300�
200�
100’
0 20 40 60 80
Clinical Cancer Research 477
Fig. 2 A, FACS analysis of humanAGT in CHO parental cell line (non-transfeeted: peaks 1 and 2) and
clones (transfeeted with human
MGMT: peaks 3-6). CHO parental
cell lines were stained either with
AGT antibody (mT3.1) or isotype
IgG with similar MFI �peak 1,
stained with isotype IgG). Clones
transfeeted with human MGMT had
a clear AGT staining signal, and
clone 3 (peak 6) expressed the high-
est level of MGMT, whereas other
clones (peaks 3, 4, and 5) had a van-
able amount of AGT. B, immunore-
active protein detected by Western
blot in CHO-hMGMT transfeetants.
The clones expressed different
amounts of human AGT. The highest
level of human AGT was detected in
clone 3. FL2, fluorescence intensity.C. the correlation between AGT ac-tivity and immunoreactive protein by
either FACS (left; r = 0.998) orWestern blot (rig/it: r 0.999).Symbols represent different CHO-
MGMT clones.
700�
0’
0 20 40 60 80
AGT activity (fmol/p.tg) DNA
and then fixed amounts of protein from each sample were
incubated with substrate DNA, which was prepared by incubat-
ing calf thymus DNA with [3H]mcthylnitrosourca. The reaction
mixture was incubated for 60 mm at 37#{176}Cand then precipitated
with 14% trichloroacetic acid at 4#{176}C.The determination of AGT
activity in cell extracts was based on the removal of [3H]mcthyl
group from 06-[3H]mcthylguaninc in 06-[methyl-3H]guaninc-
DNA. AGT activity was expressed as fmol removed 06-meth-
ylguanine/p.g DNA or fmol removed 06-methylguanine/mg
protein (9).
Western Blotting. Cell extracts were resolved by SDS-
PAGE (12% polyacrylamide) in a Bio-Rad minigel apparatus at
150 V for 1 h. Proteins were transferred onto polyvinylidene
difluoride membranes, using a Bio-Rad mini Trans-Blot cell for
1 h at 100 V. The blotted membranes were blocked with 5% dry
milk in TBS buffer and then probed for 2 h with mT3.l, which
is specific for human cellular MGMT ( 1 1). After three 5-mm
washes with TBS-Twcen 20 (0.05%), the blots were incubated
with secondary antibody, anti-mouse HRPO-anti-IgG for 1 h.
Antibody binding was visualized by ECL according to the
manufacturer’s instructions. Standard curves of samples with
known activity were included in each blot. To standardize the
Western blotting assay, K562 (chronic myclogenous leukemia)
cell line was transfected with human MGMT eDNA, and 0-50
,_Lg protein of cellular sonicate was subject to SDS-PAGE and
immunoassay. Densitometric analysis of the ECL film was
performed by a SciScan 5000 scanner using BioAnalysis soft-
ware. Regression analysis indicated that antibody-antigen curve
yielded linear plots with correlation coefficients of 0.995, when
band intensity was plotted versus fmol of AGT loaded (range,
0.9-16 fmol).
RESULTS
Identification of AGT Expression in PermeabilizedCells by FACS. After treatment with paraformaldehyde and
Tween 20, most cells were permeabilized and could be stained
with specific AGT antibody, whereas other FACS signals might
consist of platelet and cellular debris. These respective popula-
tions were identified by forward versus right angle scatter. The
scatter gates were established using K562-MGMT cells selected
to be 100% MGMT� by resistance to l,3-bis(2-chlorocthyl)-l-
nitrosourea ( 1 2). An example of the scatter gates used to detect
AGT is shown in Fig. 1 . Cells were stained with antibody mT3. 1
Research. on August 4, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
(no
0C)
A
C D
AGT’-#{216}� �- - � �
0
C3C
itFL2-Height
478 Heterogeneity of O�-Alkylguanine-DNA-alkyltransferase
PBMC by G-CSF PBMC-normal donors
Fig. 3 FACS analysis of AGT in PBMCs from normal donors and patients under going cytophenesis. A. PBMCs treated with G-CSF. B, PBMCsfrom normal donors showed a higher fluorescent staining. C, AGT protein detected by Western blot in human PBMCs. Lanes 1-4, PBMCs by G-CSF.PBMCs were collected from four patients receiving G-CSF growth factor treatment. LaneS 5-8, PBMC-normal donors. PBMCs were obtained from
tour normal donors. D, FACS analysis of mononuclear cell AGT of bone marrow. Human bone marrow had a lower MFI than the PBMCs examined
in B. FL2. fluorescence intensity.
(Fig. 1, A-C) and mouse IgG1 isotype control (D-F). Region R
contains predominantly cellular debris and nonpermeabilized
cells, confirmed by antihistone antibody staining (data not
shown). This region contains only a small number of AGT-
staining cells. In contrast, region P contains permeabilized,
AGT-positive cells staining with mT3.1. In Fig. lb. the mT3.l
staining appears uniform. and the background of nonstaining
cells is removed. This allowed us to quantitate AGT expression
in the permeabilized cell fraction.
Correlation of AGT Activity by FACS with WesternBlotting and Biochemical Assay. CHO cells containing very
low levels of AGT (<0. 1 fmol/�i.g DNA) were transfeeted with
human MGMT eDNA as described previously (10). Clones of
CHO-MGMT expressed variable AGT, ranging from 0.3 to 65
fmol/p.g DNA by biochemical assay. As shown in Fig. 14,
FACS analysis, based on MFI. had consistent results with bio-
chemical assay: clone 3 expressed the highest level of human
AGT (peak 6), whereas other clones have lesser amounts. Sim-
ilar results are also obtained by Western blot (Fig. 2B). Immu-
noreactive AGT proteins detected by Western blotting or FACS
are highly correlated with enzyme activity measured by bio-
chemical assay (Fig. 2C; r 0.998 and 0.999, respectively).
indicating that the three methods have equal specificity for
AGT.
AGT in human PBMCs and bone marrow exhibit a range
of enzyme activity. In PBMCs (ti = 6) from normal donors,
AGT activity measured by biochemical assay was 9.3 ± 2.7
fmoL/p.g DNA (range, 6.1-12.8 fmol/p.g DNA). In contrast,
PBMCs (pi 7) from leukopheresis collections obtained from
cancer patients receiving G-CSF had a much higher concentra-
tion of hematopoietic progenitors than normal blood (CD34+
cells ranged from 0.4-5%, compared with <0.1% CD34+ cells
in routine blood collections) and contained lower AGT activity
(4.6 ± 1.2 fmol!)ig DNA; range, 2.9-5.9, P < 0.002). In Fig. 3,
A and B, histograms of fluorescent staining of AGT by FACS
confirmed these differences. PBMCs from normal donors had
higher MFI (107 ± 14) than that in G-CSF mobilized-PBMCs
(38 ± 27; P = 0.015). The distribution of activity in individual
cells indicated that there is a normal distribution in most in-
stances, but that in some histograms, there appears to be a
shoulder of either high- or low-expressing cells. AGT protein
detected by FACS correlates with AGT biochemical activity
measured by the HPLC method and by Western analysis (see
also Fig. 3C), with correlation coefficients of 0.873 and 0.787,
Research. on August 4, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
100
80 #{149} #{149}
60
40
20 #{149}
Iv
0 5 10 15
AGT activity (fmoI/�tg DNA)
by HPLC
z>,
I’
AGT protein (fmol4tg DNA)
by Western blot assay
AGT activity (fmol1�1g DNA)by HPLC
Clinical Cancer Research 479
Fig. 4 The correlation of hu-
man AGT protein analyzed by
FACS, Western blot, and AGT
activity assay in human PB-
MCs. #{149},PBMCs from normaldonors; U, PBMCs from do-nors receiving G-CSF.
Tabl e 1 Comparison of heterogeneity of AGT in PBMC s from normal don ors or patientsundergoingcytophenesis
PBMCs from normal donorsPBMCs from donors receiving G-CSF
No. of samplesNo. of samples Width” at half peak MFI” IH Width” at half peak MFId IH
1 194 22 8.81 1 60 87 0.68
2 137 26 5.27 2 100 104 0.963 114 20 5.70 3 46 119 0.38
4 120 41 2.92 4 114 127 0.89
S 128 40 3.20 5 88 99 0.88
6 288 20 14.40 6 86 104 0.837 136 97 1.40
Mean 160 ± 36 38 ± 28 6.0 ± 4.4 82 ± 25” 107 ± 14’ 0.7 � 0.2”
a In units of signal intensity.
“p = 0.016.‘� P = 0.015.
“P = 0.0002.
respectively (P < 0.001 ; Fig. 4). These results demonstrate that
FACS analysis accurately measures AGT in clinic blood cell
samples and provides valuable information on the distribution of
activity in various subpopulations of cells.
With the FACS method, it is possible to assay the hctcrogc-
neity within each sample visually and mathematically. As one
measure of heterogeneity, we defined the parameter IH [width at
half peak (in units of signal intensity)IMFI]. This value is deter-
mined by two factors: the width at half peak, a measure of the
distribution of cells with positive staining; and MH, a measure of
mean level ofAGT in the cell population. Although the IH does not
directly account for subpopulations, the overlap in histograms for
these populations will tend to broaden the width at half peak. IH
values listed in Table 1 reveal that there was a much greater degree
of heterogeneity seen in the expression of AGT in the G-CSF
mobilized PBMC samples than that in the unstimulated PBMC
samples from normal donors (6.0 ± 4.4 versus 0.7 ± 0.2; P =
0.0002). As shown in Fig. 3D, analysis of the samples from bone
marrow (n = 3) indicates that their MR is even lower than that of
the G-CSF mobilized PBMC cell samples (MR. 16.8 ± 8.6) and
that they had a similar high degree of heterogeneity in the samples
(IH, 9. 1 ± 1.9).
Detection of 06-bG-inactivated AGT by FACS andWestern Blot Analysis Compared with Biochemical Assay.06-bG, a specific inhibitor of AGT, potentiatcs the cytotoxicity
of chemotherapeutic alkylating agents and has entered clinical
trials (13). Monitoring depletion of AGT by 06-bG would help
define the efficiency of drug effect in vivo. We tested whether
FACS analysis could differentiate between inactive and active
AGT. After human MGMT-CHO cells were treated with 06-bG
(25 pM) for 2, 6, and 24 h, >98% of human AGT activity was
depleted at 2 h, and little recovery was found up to 24 h,
measured by biochemical assay. However, immunorcactive
AGT protein detected by Western blot assay was unchanged at
2 h and gradually decreased (Fig. SA). FACS analysis also
revealed a gradual time-dependent decrease in fluorescent stain-
ing of cellular AGT protein after 06-bG (Fig. SB). The rate of
degradation of inactivated AGT protein measured by FACS and
Western blot assay was similar; 24 h after 06-bG, 25-35% of
baseline AGT protein was detected, which was inactive, as
proven by biochemical assay (Fig. SC). Thus, only the biochem-
ical assay accurately reflects the level of AGT activity. Western
blot and FACS analysis remain efficient methods to trace the
degradation of inactivated AGT, and this can help define con-
ditions (drug treatment or perhaps other cell stress) in which a
proportion of AGT is inactive.
DISCUSSION
FACS analysis of cellular AGT content is shown here to be
an accurate, rapid assay for single-cell suspensions, highly cor-
related with the traditional assay methods of biochemical assay
Research. on August 4, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
U
100
0 5 10 15 20 25
480 Heterogeneity of 06-Alkylguanine-DNA-alkyltransferase
Time (h) after 06b-G
0 2 6 24
AGT -*� �
.�)
0
I-
F-
C4��
0
0
ci
ci
Time afterO6-bG (hr)
A
Fig. 5 Comparison of human AGT measurement after inactivation of
AGT by 06-bG in CHO-MGMT cells. A, Western blot of inactivated
human AGT protein in CHO-MGMT transfeetants after O�-bG treat-
ment. Lrnze 1, untreated CHO-MGMT cell extract: Lanes 2-4, 2, 6, and
24 h after 06-bG treatment, respectively. B, FACS analysis of human
AGT in CHO-MGMT cells after 06-bG treatment. The high human
AGT detected by FACS decayed slowly from baseline �peak 6), 2 h
�peak 5). 6 h �peak 4). and 24 h (peak 3) after treatment but still
remained above that of nontransfeeted cells (peak 2). Peak 1. cells were
stained with isotype IgG as background. FL2. fluorescence intensity. C.comparison between AGT activity assay, Western blot, and FACS for
the rate of inactivation and degradation of human AGT after 06-bG
treatment. #{149}.FACS analysis; #{149},Western blot assay: A, HPLC meas-
urement.
and Western blot. Furthermore, as has been shown previously,
the amount of immunoreactive protein, measured previously by
Western blot, is an accurate measure of AGT activity in unper-
turbed cells (1 1). This has a number of important implications;
it suggests that the AGT-FACS assay could be used as a rapid
assay of AGT activity in a variety of samples, both laboratory
cells and clinical samples. In the clinic, samples could include
blood and marrow samples, as well as samples of malignant
cells including leukemia, lymphomas, myclomas, and disaggre-
gated tumor samples. Because only a small number of cells need
to be analyzed by FACS, it is also possible to analyze AGT in
samples too small to be measured by the conventional activity
assay. A second application is in the area of gene therapy, where
we have used FACS to detect the degree of transduced cells
(14), and in which it could be used to screen and select viral
producer cell clones and transduced, retrovirally infected cell
populations.
The AGT-FACS assay can also reveal subpopulations of
cells with different levels of AGT and the overall heterogeneity.
Thus, within each PBMC sample from normal donors, there was
a relatively uniform level of AGT protein. reflecting the rela-
tively homogeneous population of cells (about 15% monocytes
and 85% lymphocytes) and the similar levels of AGT in these
cells, as we have reported previously (15). In contrast, in PB-
MCs from G-CSF mobilized samples. as in bone marrow sam-
ples, there was much more heterogeneity in AGT levels than
that in the unstimulated PBMC samples. This difference is not
surprising, given the markedly heterogeneous cell population of
both samples, which in the case of the G-CSF mobilized cells
include activated monocytes. cosinophils. lymphocytes, CD34+
cells and other hematopoictic progenitors, and late maturing
myclocytes; and in the case of bone marrow, includes the entire
differentiation lineage of hcmatopoiesis. These results confirm
our previous observations with marrow subpopulations and
CD34+ cells, indicating that hematopoietic progenitors have
lower AGT activity than lymphocytes and monocytes (16). The
use of double or triple antibody staining techniques, while
challenging when one of the epitopes is nuclear and the others
react with surface molecules, is an appealing way to identify
specific subpopulations with either high or low AGT-FACS
levels.
We did not observe samples with an incongruity between
mean AGT activity and the level of immunoreactive protein on
Western blot, nor did we find samples in which a subpopulation
of cells lacked AGT activity but had evidence of AGT protein
by AGT-FACS or Western blot. Thus, in the steady state, the
level of immunoreactive protein appears to be an accurate
predictor of AGT activity. Furthermore, this suggests that the
normal breakdown of AGT, which appears to involve ubiquiti-
nation and proteosome digestion ( 17), does not result in excess
immunorcactive protein relative to active protein. Given the
long half-life of this protein, in excess of 24 h, this result may
not be surprising but reiterates the suggestion that the level of
MGMT mRNA is the major determinant of AGT activity in
cells (18).
Finally, our data show that after treatment with 06-bG,
only the biochemical assay accurately measures active AGT.
Although the AGT is degraded more rapidly after inactivation,
the monoclonal antibody we used. mT3. 1 , and most if not all
Research. on August 4, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Clinical Cancer Research 481
other antibodies react with the inactive as well as active protein
(19). Accordingly, both AGT-FACS and Western blot assays
failed to distinguish active protein from inactive protein. We
have found that the time course of loss of immunoreactive
protein after 06-bG exposure varies between cells and that it
matches the AGT-FACS measurement of AGT content, al-
though new AGT is being synthesized by the cells at a slow but
constant rate (20). This indicates that the AGT-FACS assay will
not be useful in clinical trials to assess the efficacy of 06-bG or
other AGT inhibitors, and that the biochemical assay will need
to be used unless an antibody is developed that reliably distin-
guishes between the active and inactive forms of AGT.
In summary, AGT-FACS is a highly sensitive technique to
rapidly and accurately analyze AGT content in single-cell sus-
pensions and to assess the degree of heterogeneity that exists
within that population. Potential applications include: detection
of tumor cells with high AGT amenable to AGT inactivation
therapy; gene therapy detection of transduced cells; and analysis
of subpopulations by multiparameter surface or internal marker
analysis.
ACKNOWLEDGMENTS
We thank Deborah A. Corey for assistance with the establishment
of this method.
REFERENCES
1. Dumenco, L. L., Allay, E., Norton, K., and Gerson, S. L. The
prevention of thymic lymphomas in transgenic mice by human O6�
alkylguanine-DNA AGT. Science (Washington DC), 259: 219-222,
1993.
2. Dolan, M. E., Mosehel, R. C., and Pegg, A. E. Depletion of mam-malian 06-alkylguanine-DNA AGT by 06-benzylguanine provides a
means to evaluate the role of this protein in protection against carcino-genie and therapeutic alkylating agents. Proc. Natl. Acad. Sci. USA, 87:
5368-5372, 1990.
3. Liu, L., Allay, E., Dumenco, L. L., and Gerson, S. L. Rapid repair of
O�-methylguanine-DNA adducts protects transgenic mice from N-
methylnitrosourea-induced thymic lymphomas. Cancer Res., 54: 4648-
4652, 1994.
4. Kroes, R. A., and Erickson, L. C. The role of mRNA stability andtranscription on the 06-methylguanine-DNA methyltransferase expres-sion in Mer+ human tumor cells. Carcinogenesis (Lond.), 16: 2255-2257, 1995.
5. Pegg, A. E., Mammalian 06-alkylguaninc-DNA AGT: regulation andimportance in response to alkylating carcinogenic and therapeutic
agents. Cancer Res., SO: 61 19-6129, 1990.
6. Pegg, A. E., Dolan, M. E., Moschel, R. C. Structure, function andinhibition of 06-alkylguanine-DNA AGT. Prog. Nucleic Acid Res. Mol.
Biol., 51: 167-223, 1995.
7. Kyrtopoulos, S. A., Souliotis, V. L., Ambatzi, P., Pangalis, G., Bou-siotou, V., Itaritopoulos, G., and Davaris, G. Novel, sensitive assays for
06-alkylguanine and its repair and their application to studies of the
molecular epidemiology of this lesion in human populations. IARC Sci.
PubI., 105: 78-82, 1991.
8. Wu, R. S., Hurst-Calderone, S., and Kohn. K. W. Measurement of
06-alkylguanine-DNA alkyltransferase activity in human cells and tu-
mor tissues by restrictive endonuelease inhibition. Cancer Res.. 47.
6229-6235, 1987.
9. Gerson, S. L.. Trey. J.. Miller, K., and Berger, N. A. Comparison of
06-alkylguanine-DNA AGT activity based on cellular DNA content inhuman, rat and mouse tissues. Careinogenesis (Lond.), 7: 745-749.
1986.
10. Liu, L., Lee, K., Wasan, E., and Gerson, S. L. Differential sensi-tivity of human and mouse AGT to 06-benzylguanine using a transgenie
model. Cancer Res., 56: 1880-1885, 1996.
1 1. Brent, T. P., von Wronski, M. A., Pegram, C. N., and Bigner, D. D.
Immunoaffinity purification of human 06-alkylguanine-DNA AGT us-
ing newly developed monoclonal antibodies. Cancer Res.. 50: 58-61,
1990.
12. Allay. J. A., Dumenco. L. L., Liu, L., and Gerson, S. L. Retroviraltransduction and expression of the human alkyltransferase eDNA pro-
vides nitrosourea resistance to hematopoietic cells. Blood, 85: 3342-
3351, 1995.
13. Gerson. S. L.. and Willson, J. K. V. 06-Alkylguanine-DNA alkyl-transferase: a target for the modulation of drug resistance. HematoliOneol. Clin. N. Am., 2: 431-450, 1995.
14. Davis, B. M., Reese, J. S., Koc, 0. N.. Sehupp, J. E., and Gerson,
S. L. Selection for G156A MGMT transduced hematopoietie progeni-tors and protection form lethality in mice treated with 06-benzylguanineand BCNU. Cancer Res., 57: 5093-5099, 1997.
15. Gerson, S. L., Trey, J. E., Miller, K., and Berger. N. A. Or’-Alkylguanine-DNA alkyltransferase activity in human myeloid cells.
J_ Clin. Invest., 76: 2106-21 14, 1985.
16. Gerson, S. L., Phillips. W., Kastan, M., Dumenco, L. L., andDonovan, C. Human CD34+ hematopoietie progenitors have low. cy-tokine-unresponsive 06-alkylguanine-DNA alkyltransferase and are
sensitive to 06-benzylguanine plus BCNU. Blood, 88: 1649-1655.
1996.
17. Srivenugopal, K. S., Yuan, X-H., Friedman, H. S., and Ali-Osman,F. Ubiquitination-dependent proteolysis of 06-methylguanine-DNAmethyltransferase in human and murinc tumor cells following inactiva-
tion with 06-bcnzylguanine or I .3-bis(2-ehloroethyl)- I -nitrosourea.
Biochemistry, 35: 1328-1334, 1996.
18. Gerson, S. L., Allay. E., Vitantonio, K., and Dumenco, L. L.
Determinants of 06-alkylguanine-DNA alkyltransferase activity in hu-
man colon cancer. Clinical Cancer Res., 1. 519-525, 1995.
19. Ayi. T-C., Oh, H-K., Lee. T. K-Y., and Li, B. F. L. A method forsimultaneous identification of human active and active-site alkylated
06-mcthylguanine-DNA methyltransferase and its possible applicationfor monitoring human exposure to alkylating carcinogens. Cancer Res.,
54: 3726-3721, 1994.
20. Pieper, R. 0., Futseher, B. W., Dong, Q., and Erickson, L. C.
Effects of streptozotocin/bisehloroethylnitrosourea combination therapyon O�-methylguanine-DNA methyltransferase activity and mRNA 1ev-
els in HT-29 cells in vitro. Cancer Res., 5/: 1581-1585, 1991.
Research. on August 4, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1998;4:475-481. Clin Cancer Res L Liu, K Lee, J Schupp, et al. marrow mononuclear cells.measured by flow cytometric analysis in blood and bone Heterogeneity of O6-alkylguanine-DNA-alkyltransferase
Updated version
http://clincancerres.aacrjournals.org/content/4/2/475
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://clincancerres.aacrjournals.org/content/4/2/475To request permission to re-use all or part of this article, use this link
Research. on August 4, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from