Biology of Blood and Marrow Transplantation 11:890-902 (2005)
䊚 2005 American Society for Blood and Marrow Transplantation
1083-8791/05/1111-0006$30.00/0
doi:10.1016/j.bbmt.2005.07.008
Cytomegalovirus Immune Reconstitution Occurs
in Recipients of Allogeneic Hematopoietic Cell
Transplants Irrespective of Detectable
Cytomegalovirus Infection
Ghislaine Gallez-Hawkins,1 Lia Thao,1 Simon F. Lacey,2 Joybelle Martinez,2 Xiuli Li,1
Anne E. Franck,1 Norma A. Lomeli,3 Jeff Longmate,3 Don J. Diamond,2 Ricardo Spielberger,4
Stephen J. Forman,4 John A. Zaia1
1
Division of Virology, 2Laboratory of Vaccine Research, and 3Division of Information Sciences, Beckman Research
Institute, and 4Department of Hematology and Bone Marrow Transplantation, City of Hope National Medical
Center, Duarte, California
Correspondence and reprint requests: John A. Zaia, MD, Beckman Research Institute of the City of Hope, Division
of Virology, 1500 E. Duarte Rd., Duarte, CA 91010 (e-mail: [email protected]).
Received June 6, 2005; accepted July 13, 2005
ABSTRACT
The question of when immune reconstitution of cytomegalovirus (CMV)–specific CD8 T cells occurs after
hematopoietic cell transplantation and, more specifically, to which CMV targets this immunity is likely to be
directed remains poorly understood. The dependence of immune reconstitution on CMV reactivation is even
less clear. To better understand these events, 44 CMV-seropositive HLA-A*0201 subjects were followed up at
approximately days 40, 90, 120, 150, 180, and 360 after hematopoietic cell transplantation for CMV immunity
as measured by 2 types of assays: (1) an HLA-A*0201 tetramer-binding assay for both CMV pp65 (pp65) and
immediate-early 1 (IE-1) or (2) intracellular cytokine interferon ␥ responses induced by pp65 or IE-1– derived
peptides. To verify the reliability of IE-1–specific assays relative to the pp65-based assays, a pilot study first
compared the development of IE-1–specific immunity in a subgroup by using multiple HLA-A*0201–restricted
peptides, and then these recipients were followed up for 1 year for immunologic function and for CMV
infection. The IE-1–specific response occurred to each of the 3 HLA-A*0201–restricted peptides studied
(IE-1-256, -297, and -316), and there was no predominant IE peptide response. However, the immunodominant HLA-A*0201–restricted pp65 peptide was recognized significantly more frequently than these IE-1
peptides. When this was compared with the occurrence of CMV infection, the overall immune reactivity, as
measured by the mean or median number of CD8ⴙ T cells reactive to either pp65 or IE-1 peptides by
intracellular cytokine or tetramer binding assay, was not significantly different in those with and without CMV
infection. For patients who demonstrated reconstituted immunity to CMV at 1 year, all were reconstituted by
6 months, and the timing of the first observed immune reactivity to either of the pp65 or the IE peptides was
not different in those with and without detectable CMV infection.
© 2005 American Society for Blood and Marrow Transplantation
KEY WORDS
Cytomegalovirus ● Hematopoietic cell transplantation
Immune reconstitution ● Tetramer
INTRODUCTION
Quantitative assessment of HLA-restricted T-lymphocyte recognition of specific antigens currently analyzes cells by using methods that can enumerate the
binding of recombinant HLA-peptide antigen complexes or the specific induction of intracellular cyto890
●
Intracellular cytokine responses
●
kines (ICCs). These methods have been applied to the
analysis of risk for cytomegalovirus (CMV) complications after allogeneic hematopoietic cell transplantation (HCT; see review [1]). Because CMV seems to
encode more than 200 polypeptides, the immune response to CMV is complex and involves multiple protein targets, and the peptide sequences of the HLA-
CMV ICC and Tetramer Binding
specific restriction elements are known for only a few
of the CMV proteins [2]. The most abundant viral
protein, CMV pp65, is the apparent immunodominant target of the immune response [3-5], and most
analyses of the protein-specific immune response after
HCT have focused on CMV pp65.
For CMV pp65, the immune response in HCT
recipients has been well described [5-9]. In HLA-A*0201–
restricted responses, the pp65-495 (NLVPMVATV)
peptide is known to induce a CMV pp65 protein–
specific ICC response [4,5,10], and when this peptide
is used to fold recombinant tetrameric major histocompatibility complex molecules, a remarkable number of T cells can often be enumerated [8,11-14]. The
cellular immune response to the CMV pp65 protein is
therefore robust and has been used to characterize the
immune response to CMV in other HLA contexts,
such as B7, A1, and B35 [2,8].
By using the binding of CMV pp65–specific HLApeptide complexes to CD8 cells as a measure of CD8
immunity, it has been reported that levels reaching 10
⫻ 106 CD8 tet⫹ cells per liter are associated with
protection from CMV disease in HCT recipients [15]
and that use of ganciclovir for treatment of CMV
infection is significantly less if there are ⱖ2 ⫻ 106
CD8 tet⫹ cells per liter [14]. Few targets of CMV
immunity have been used in these assessments, and,
because of the potential for CMV to selectively escape
from immune detection, it has been suggested that the
apparent immunodominance of CMV pp65 might
lead to overinterpretation of this protein as an important target of protective immunity [16]. Rather, immunity to other CMV proteins, especially those made
at immediate early times after infection, might be
protective. A recent report in CMV-seropositive
heart/lung transplant recipients indicated that immunity to CMV IE-1 and not to CMV pp65 was more
protective [17].
Less is known about the immune response to
CMV IE-1, which is one of the first proteins expressed
after virus reactivation. CMV IE-1 is clearly a target of
T-cell immunity after CMV infection, and an early
report [18], confirmed by others [19], described the
IE-1-316 (VLEETSVML)–specific peptide response
in cells from 6 of 18 HLA-A*0201 subjects by using an
interferon (IFN)–␥ enzyme-linked immunospot assay.
The IE-1-315 peptide and some of its variants [19,20]
and IE-1-354 [21] have been reported to trigger cytotoxic responses as well. In the primary immune
response to CMV infection in infants, it is known that
IE-1-316 –specific responses are a major component
of cellular immune reactivity [22]. In addition, we
have reported that, using transgenic HLA-A*0201
mice immunized with CMV IE-1 DNA, there is CMV
IE-1–specific cytotoxic T-lymphocyte recognition of
several peptides, including IE-1-297, IE-1-316, and
IE-1-256 [23].
BB&MT
The purpose of this study was to describe the
onset and durability of CMV immune reconstitution
in allogeneic HCT as measured by T-cell immune
recognition of CMV pp65 and CMV IE-1 and to
assess the effect of CMV reactivation on quantitative
measures of such cellular immunity. We found that, as
anticipated, the CMV pp65 response was significantly
more robust than the CMV IE-1 responses, but in
certain recipients, IE-1 responses were higher than
pp65-specific responses. Contrary to conventional
wisdom, however, the onset of detectable CMV immunity was not dependent on documented CMV viremia or DNA-emia, and there was no statistically significant difference between CMV immunity to CMV
pp65 and CMV IE-1 in those with and without detectable CMV reactivation. This indicates that occult
CMV reactivation is sufficient to induce reconstitution of immunity in HCT recipients.
PATIENTS AND METHODS
Study Patients
Allogeneic HCT recipients, including related sibling
donors and recipients of unrelated donor transplants at
risk for CMV infection because of donor and/or recipient CMV seropositivity, were enrolled at the City of
Hope Comprehensive Cancer Center with the approval
of the institutional review board. During 2001 to 2004,
44 recipients were studied. Within this group, a pilot
subgroup of 11 consecutive HLA-A*0201 subjects and
their donors, when available, was evaluated for comparative responses to IE-1 peptides and the CMV pp65-495
peptide; this was done with frozen, density-gradient purified leukocytes. After this, a group of 33 consecutive
HLA-A*0201 subjects were analyzed for evidence of
immunity to IE-1 and CMV pp65 by using fresh whole
blood samples. Donor samples (unrelated donors not
included) were drawn before the administration of granulocyte colony-stimulating factor and later at the time of
cell harvest. Recipient blood samples were collected at
day 40, 90, 120, 150, 180, and 360 after transplantation
for ICC and tetramer-binding assays, and the CMV
reactivation was monitored at day 21 after HCT and
twice weekly until day 100 by using a shell vial assay and
quantitative polymerase chain reaction (PCR). After day
100, CMV was monitored in patients at high risk because of graft-versus-host disease (GVHD) or immunosuppressive medication. Subjects with reactivation of
CMV infection as determined by 1 positive shell vial
sample (CMV blood culture [BC]) or 2 consecutive positive PCR assays (CMV PCR) were treated with preemptive ganciclovir for 6 weeks as previously described
[24]. For the purpose of analysis, recipients were placed
into a CMV group, meaning that they had at least 1
positive CMV blood shell vial culture or 1 PCR-positive
plasma test, or into a no-CMV group, meaning that
891
G. Gallez-Hawkins et al.
these conditions were not met despite frequent viral
surveillance. All patients underwent ⱖ90% of the scheduled viral and immunologic surveillance.
of the assay was 200 gc/mL of plasma. No PCR
inhibition was detected in samples when the EXO
gene was introduced into the PCR mixture, as described by Limaye et al. [26].
Transplantation Protocol
The HCT protocol was essentially performed as
described by Bensinger et al. [25], but the conditioning regimen was modified according to patient age
and diagnosis. Briefly, the disease-specific conditioning regimens consisted of high-dose chemotherapy
with or without total body irradiation and were administered before transplantation. Most recipients received peripheral blood stem cells collected after
treatment of the donor with subcutaneous granulocyte
colony-stimulating factor (16 ␮g/kg/d for 4 days). For
some recipients, marrow was used for transplantation
and was collected from the donor by standard techniques on the day of infusion.
Isolation and Preservation of Peripheral Blood
Mononuclear Cells
Peripheral blood mononuclear cells from heparintreated whole blood were isolated by using Histopaque1077 (Sigma, St. Louis, MO) density gradients, washed
with 1⫻ phosphate-buffered saline (PBS), and cryopreserved in aliquots of 3 to 5 ⫻ 106 cells per milliliter in
90% fetal bovine serum (HyClone, Logan, UT) and
10% dimethyl sulfoxide (American Type Culture Collection, Manassas, VA). The plasma was collected by
centrifugation, filtered through a 0.45-␮m Acrodisc
(Pall Corp., Ann Arbor, MI), and stored at ⫺20°C
until DNA extraction.
Quantitative PCR
Quantitative PCR was performed with DNA extracted from 200-␮L plasma samples by using the
QIAamp DNA Blood Minikit (Qiagen, Valencia CA)
and resuspended in 200 ␮L of elution buffer. A glycoprotein B (gB) CMV DNA sequence was amplified
by using 1 picomole of the forward primer 5=CTGGCCAGGCCCAAGAC-3=, 1 picomole of the reverse
primer 5=CGGCCATTTACAACAAACCG3=, and 1
picomole of the probe 5=-FAM-CCCATGAAACGCGCGGCA-TAMRA (Applied Biosystems, Foster City, CA) in a 30-␮L reaction that contained the TaqMan Universal PCR Master Mix
(Applied Biosystems) and 10 ␮L of extracted DNA.
The PCR cycles were set according to the manufacturer’s protocol: 2 minutes at 50°C and 10 minutes at
95°C, followed by 40 cycles of 15 seconds at 95°C and
1 minute at 60°C, and the data were collected and
analyzed on an ABI Prism 7900HT Sequence Detection System (Applied Biosystems). Serial dilutions
(100-106 genome copies [gc]) of the plasmid containing the amplified sequence (pDCMVgB) were used
for determination of a standard curve. The sensitivity
892
Peptides
Peptides were synthesized at the City of Hope
peptide-synthesis core facility (IE-1 peptides) or in the
laboratory of D.J.D. (pp65-495 and human immunodeficiency virus [HIV]– 468) by using an ABI 432A
(PE Biosystems, Foster City, CA) or a Pioneer (Perseptive Biosystems Inc., Framingham, MA) peptide
synthesizer and standard Fmoc protocols with purification to 90% by high-performance liquid chromatography. They were dissolved in 10% dimethyl sulfoxide/water to a concentration of 5 mmol/L and used
at a final concentration of 25 ␮mol/L for T-cell stimulation. The following peptides were used: IE-1-256
(ILDEERDKV), IE-1-297 (TMYGGISLL), IE-1316 (VLEETSVML), pp65-495 (NLVPMVATV; as a
positive control), and HIV-468 (ILKEPVHGV; as a
negative control).
IFN-␥ ICC Detection
For the first group of 11 subjects, frozen samples
from each patient were thawed at 37°C and washed
with cold RPMI with 10% fetal bovine serum, and
aliquots of approximately 1 ⫻ 106 peripheral blood
lymphocytes (PBLs) were stimulated with individual
peptides (pp65-495, IE-1-256, IE-1-297, and IE-1316) in separate tubes or an IE-1 peptide mixture in 1
tube, as indicated in the experiment. The positive
stimulation control was phytohemagglutinin (Remel,
Lenexa, KS), and the negative control was HIV peptide. The ICC assay was adapted from Dunn et al. [27]
for frozen cells. After 1 hour of incubation at 37°C in
5% carbon dioxide, 1 ␮L (stock 5 mg/mL) of brefeldin
A (Sigma), a cytokine secretion inhibitor, was added to
the cells and further incubated overnight.
When fresh blood samples were used for the second group of 33 subjects, 200 ␮L of whole blood was
treated as described previously but was incubated at
37°C for only 4 hours. Samples were then washed with
1⫻ PBS and 0.5% bovine serum albumin (Sigma) and
stained with 5 ␮L of anti-CD8 conjugated to streptavidin-phycoerythrin (CD8-PE; Pharmingen, San Diego, CA) for 20 minutes in the dark. The cells were
fixed and permeabilized for 20 minutes by using the
Cytofix/Cytoperm Kit (Pharmingen) and stained with
1 ␮L of anti–IFN-␥ conjugated with streptavidinallophycocyanin (APC) for 30 minutes at 4°C in the
dark. Samples were washed twice with 1⫻ Cytofix
Wash Solution (BD Biosciences, Pharmingen, San
Jose, CA) and analyzed with the fluorescence-activated cell sorter (FACS) FACSCalibur (Becton Dickinson, San Jose, CA). The lymphocytes were gated on
BB&MT
Table 1. Demographics of the 11 Allogeneic Stem Cell Transplant Subjects
Patient
No.
Diagnosis
CMV Serostatus
(Donor/Recipient)
ⴙ
53
ALL Phⴙ
D /R
65
76
CML
ALL
81
ⴙ
Hematopoietic
Stem Cell
Source
HLA
CMV PCR
(Genome
Copies/mL)*
CMV BC
Antiviral
Treatment
Grade of GVHD
(Months after
HCT)
A0101
A0201
neg
neg
No
Dⴙ/Rⴚ
Dⴙ/Rⴙ
UD BM
UD BM
A0201
A0201
A1101
A6901
neg
neg
neg
neg
No
No
Pre–B-ALL
Dⴙ/Rⴙ
Sibling PBSC
A0206
A2402
neg
neg
No
91
ALL
Dⴙ/Rⴙ
Sibling PBSC
A0201
A2601
neg
neg
No
98
CML
Dⴙ/Rⴙ
Sibling PBSC
A0205
A1101
neg
neg
No
54
Biphenotypic
leukemia
Dⴚ/Rⴙ
UD BM
A0101
A0201
1416-10,234 (37)
Yes (69)
GCV
70
AML
Dⴙ/Rⴙ
Sibling PBSC
A0203
A0206
4755-12,331 (41)
Yes (41)
GCV
93
ALL
Dⴚ/Rⴙ
Sibling PBSC
A0201
A3001
202-1838 (39)
Yes (55)
GCV
GVHD III: gut
and liver (2)
94
Hodgkin
disease
CML
Dⴙ/Rⴙ
UD PBSC
A0201
A0301
Yes (50)
GCV
Dⴙ/Rⴙ
Sibling PBSC
A0201
A6803
neg
No
GVHD: gut and
liver (2)
None
105
neg
205 (72)
GVHD of the
mouth (1-8)
None
GVHD II: skin
and gut (1);
liver (5); skin
and eyes (5)
GVHD: liver (4)
GVHD II: gut (1);
liver (4)
GVHD: mouth
(1); liver (4)
CMV colitis;
GVHD III: gut
(1); mouth,
eyes (10)
GVHD II: gut (1);
stomatitis (2)
CSA: 0-6 mo MMF:
day 28 to 6 wk
PSE 3 d; FK506 0-3
mo; PSE 6-7;
MMF 6-9
Clinical Outcome
Alive
Alive
Dead: multiorgan
failure (20 mo)
CSA 4-5 mo; MMF
4-5 mo; PSE 4-12
mo
CSA 0-1 mo
Alive
PSE 15 d; CSA 15 d
Alive
PSE 3 d; MMF 1-4
mo
Alive
CSA 0-15 mo; PSE
1-6 mo; MMF 812 mo
CSA 2-3 mo; MMF
2-3 mo; PSE 2-3
mo
CSA 2-3 mo; MMF
2-3 mo
Alive
Alive
Alive
Alive
Alive
ALL indicates acute lymphoblastic leukemia; CML, chronic myelocytic leukemia; AML, acute myelogenous leukemia; UD, unrelated donor; PBSC, peripheral blood stem cells; BM, bone marrow; GCV,
ganciclovir; BC, blood culture; GVHD, graft-versus-host disease; CSA, cyclosporine; PSE, prednisone; MMF, mycophenolate mofetil; FK506, tacrolimus; Ph⫹, philadelphia chromosome positive;
neg, the result was negative for CMV infection.
*Number of days after HCT to first positivity are shown in parentheses.
893
CMV ICC and Tetramer Binding
Sibling PBSC
Medication
>1 mg/kg/d
G. Gallez-Hawkins et al.
the basis of forward and side scatter, and a minimum
of 50 000 events were analyzed per sample. The reported data are the values obtained after subtraction of
background levels observed with HIV peptide stimulation (range, 0%-0.5%). On the basis of analysis of
CMV-negative donors, a limit of detection of 1 ⫻ 105
cells per liter was established.
HLA-A*0201 Tetramer Binding
The HLA-A*0201 tetramers were prepared as described previously [8] by using the individual CMV
peptides to fold the HLA-A*0201 heavy chain and
␤2-microglobulin, which were then biotinylated and
conjugated with streptavidin-APC (Pharmingen). The
samples from each patient were thawed at 37°C,
washed with 1⫻ PBS with 0.5% bovine serum albumin, and transferred into polystyrene round-bottomed FACS tubes (Becton Dickinson). Aliquots were
then individually labeled with 0.5 to 1 ␮g of tetpp65495, tetIE1-297, and tetIE1-316 and incubated for
1 hour on ice in the dark. The samples were washed
and stained with 5 ␮L of anti-CD8 conjugated to
streptavidin-phycoerythrin, incubated for 20 minutes, washed again, and, finally, analyzed by FACS. A
European multicenter evaluation of this method has
been established and validated as a reproducible routine diagnostic assay for the enumeration of antigenspecific CD8 T cells [28].
Statistical Analyses
Data were analyzed with GraphPad Prism software (GraphPad Software, Inc., San Diego, CA). The
Mann-Whitney U test was used to derive P values for
comparing data between groups.
RESULTS
Clinical Course of CMV Infection for 1 Year
after HCT
Among the initial 11 recipients studied (Table 1)
to establish whether we could readily detect reconstitution of cellular immunity to CMV IE-1, 5 subjects
(patients 54, 78, 93, 94, and 105) had at least 1 CMVpositive blood sample by shell vial or had a qualifying
CMV DNA quantitative PCR–positive assay in blood
plasma. In addition, subject 54 had disease diagnosed
as CMV colitis by biopsy. CMV positivity in blood
was determined in 4 of 5 subjects for each assay. Of
note, patient 54’s donor and the donor for patient 93
were CMV seronegative: this may have explained the
prolonged CMV infection detected in these recipients. Six of these subjects had no detectable CMV
infection (no-CMV group).
GVHD grade II or less occurred in 2 subject in the
CMV group and in 6 subjects in the no-CMV group,
and GVHD grade greater than II occurred in 2 sub894
jects in the CMV group and in 0 subjects in the
no-CMV group. Thus, as anticipated, GVHD was
more severe in the CMV group, and the exposure to
prednisone therapy at a dose of ⱖ1 mg/kg/d was
greater in this group (Table 1). Ganciclovir was used
only in the CMV group, but there was a clinical
decision not to treat patient 105, who was CMV positive by quantitative PCR only. The overall survival
(5/6 versus 5/5) and disease relapse (none) rates for
these subjects at 1 year after HCT were the same, and
GVHD grade III and CMV reactivation were the
main clinical parameters that differentiated the 2
groups.
Frequency of IFN-␥ Response to CMV Peptide
Stimulation in the Pilot CMV Group and
No-CMV Group
To evaluate the immune reactivity to CMV in
HLA-A*0201 subjects, cryopreserved PBLs were
stimulated with peptides derived from CMV pp65 and
CMV IE-1 proteins. To ensure the functionality of
the stimulated cells, only the samples that responded
to phytohemagglutinin stimulation were reported.
During the year of observation, the percentage of
specimens obtained from subjects in the 2 groups was
the same (80% and 77.7%), and all specimens were
analyzed for reactivity to pp65-495, IE-1-256, IE-1297, and IE-1-316 peptide stimulation. Figure 1 describes the overall frequency of IFN-␥–positive T-cell
specimens in the 2 groups. The CMV group had a
higher overall frequency of positive specimens reactive to pp65-495, IE-1-256, IE-1-297, and IE-1-316
peptide stimulation. It was noted that the immune
reconstitution to CMV after HCT was characterized
by recognition directed to multiple IE-1 peptides in
all subjects of the CMV group but that the response to
IE-1 peptides in the no-CMV group was generally
negative or low.
Kinetics of ICC/IFN-␥–Positive Cells after HCT in
the CMV Group and No-CMV Group
To better understand the occurrence of immune
reactivity to these antigens, blood samples were
collected at days 40, 90, 120, 150, 180, and 360 after
HCT, and the results were expressed as the total
number of IFN-␥–reactive cells ⫻ 105 per liter
(Figure 2). The IFN-␥–positive cells in the CMV
group stimulated with pp65-495 peptide (Figure
2A) peaked at day 120 and 150 after HCT, reaching
median levels of 2.75 ⫻ 107 cells per liter, and
peptide-responsive cells were still present at 1 year
(day 360 median, 3.12 ⫻ 107 cells per liter). The
levels of pp65-495–specific IFN-␥–positive cells in
the no-CMV group reached a peak (median, 1.8 ⫻
106 cells per liter) at day 150 and decreased to 3.0 ⫻
105 cells per liter after 1 year (Figure 2C). With the
CMV ICC and Tetramer Binding
Figure 1. ICC/IFN-␥–positive samples expressed as a percentage of
all samples tested in patients with CMV reactivation or without
CMV reactivation (no-CMV group). After overnight stimulation
with individual CMV peptides (pp65-495, IE-1-256, IE-1-297, and
IE-1-316), samples were analyzed for IFN-␥ by using an intracellular cytokine (ICC) assay as described in “Patients and Methods.”
The detection limit was 0.01% of CD8⫹/IFN-␥–positive cells or 1
⫻ 105 cells per liter after subtraction of background data by using an
HIV peptide.
nonparametric Mann-Whitney 2-tailed test, there
was no significant difference by using the mean ICC
level of each subject and the highest peak reached
between those with and without CMV infection
(compare Figure 2A and C; P ⫽ .08).
The IFN-␥–positive cells stimulated with IE-1256, IE-1-297, and IE-1-316 were reported as the
sum of the total number of cells ⫻ 105/L for the
evaluation of the response to CMV IE-1, as shown
in Figure 2B and D. As indicated by the difference
in the y-axis range, the levels of CMV IE-1/IFN␥–positive responsive cells were a log lower than the
CMV pp65–specific levels and did not seem to increase until late after HCT. The median CMV IE-1
immune response in the CMV group was 3.8 ⫻ 106
cells per liter at 1 year (Figure 2B) compared with 4
⫻ 105 cells per liter in the no-CMV group (Figure
2D). The level of ICC response to CMV pp65 as
compared with IE-1 peptides was significantly different (Mann-Whitney test; P ⫽ .03) in the CMV
reactivation group, but there was no such difference
in the no-CMV group. In summary, a detectable
number of T cells were reactive to CMV pp65 and
IE-1 in the CMV group, but because the sampling
was small, there was no significant difference between the CMV infection group and the no-CMV
group. For this reason, a larger cohort of subjects
was analyzed (see below).
Relationship between CMV Reactivation by PCR
and Levels of CMV-Reactive Cells by ICC in the
CMV Reactivation Group
The question of what effect CMV reactivation
has on CMV immune reconstitution was addressed
by comparing the timing to the 2 events. CMV
BB&MT
reactivation started at a median of 55 days after
HCT, and, as shown in Figure 3, the immune responses after this infection varied for individual
recipients, with no predictable expansion of T cells
reactive to any specific CMV peptide or peptides of
pp65 or IE-1. To summarize the observations in
this group of subjects, the presence of an ICC response to pp65-495 did not automatically imply
reactivity to CMV IE-1 peptides (Figure 3; subjects
54 and 105), nor did the absence of a brisk response
to CMV pp65 mean that there would not be a
robust CMV IE-1 response (Figure 3; subjects 70
and 94). Despite the administration of immunosuppressive drugs, there was evidence of a high level of
pp65-495–reactive CD8 cells in the presence of
prolonged CMV infection (Figure 3; subject 93).
The prolonged CMV infection was also found in
subject 54, and the low levels of CD8 responses to
pp65 or IE-1 corresponded to heightened immunosuppression with prednisone and mycophenolate
mofetil (Table 1). Of note, there was a simultaneous
response to all 3 IE-1 peptides within the same
sample in all subjects at some time after CMV
infection, but this was never observed in the noCMV group (data not shown).
Assessment of the Immune Response by
Tetramer-Binding Assay
HLA-A*0201 tetramers labeled with the APC
molecule were prepared with pp65-495, IE-1-297,
and IE-1-316 peptides, and the binding of these
tetramers was evaluated on the same samples tested
for IFN-␥ ICC (Figure 4). The CMV reactivation
group was defined by high levels of CD8⫹/tetpp65495 cells that appeared at day 120 after HCT (median, 6.36 ⫻ 107 cells per liter) and was maintained
throughout the year (day 360 median, 8.23 ⫻ 107
cells per liter; Figure 4A). In contrast, the number
of cells binding to CD8⫹/tetIE-1-297 and CD8⫹/
tetIE-1-316, expressed as a sum, steadily increased
to a median of 3.7 ⫻ 106 cells per liter at day 360
(Figure 4B). Of note, the median levels of CD8⫹/
tetpp65-495 and CD8⫹/tetIE-1 at day 40 after
HCT in the CMV reactivation group were similar
(median, 5 ⫻ 106 and 7.4 ⫻ 106 cells per liter,
respectively). However, upon CMV reactivation,
the immune cells specific for CMV pp65 increased
to 7.58 ⫻ 107 cells per liter by day 150, whereas the
cells specific to CMV IE-1 decreased to 1.8 ⫻ 106
cells per liter. Of note, the no-CMV group developed a peak median CD8⫹/tetpp65-495 cell count
at a later time (median, 2.28 ⫻ 107 cells per liter at
day 180; this decreased to 4 ⫻ 105 cells per liter by
1 year; Figure 4C). Although higher at day 40 in the
CMV reactivation group, the levels of CD8⫹/
tetIE-1 remained the same in both groups (Figure
4B and 4D) throughout subsequent observation in
895
G. Gallez-Hawkins et al.
Figure 2. Levels of CMV-specific CD8⫹/IFN-␥–positive cells at various times after HCT. CMV peptide–specific CD8⫹ cells were stimulated
in the CMV-reactivation group (A and B) and in the no-CMV group (C and D). A and C, Cells stimulated with pp65-495. B and D, Sum of
cells stimulated with IE-1-256, IE-1-297, and IE-1-316. The number of IFN-␥–positive cells, expressed as x times 1 ⫻ 105 cells per liter, is
shown after subtraction of background data by using HIV peptide stimulation. The median value was calculated at each time point and is shown
on the graph as a continuous line.
the year after transplantation. The tetramer-binding data therefore indicate that the CMV-specific
CD8⫹ immune cells are present in reasonably high
quantity in HCT subjects during hematopoietic reconstitution whether or not there is detectable
CMV reactivation.
Further Evaluation of pp65 and IE-1 Peptide Pools
in HLA-A*0201 Subjects
On the basis of this experience indicating that
ICC-positive cells could be observed after stimulation with individual IE-1 peptides, the follow-up
protocol used fresh whole blood cells from HCT
subjects stimulated with an IE-1 mixture of these
peptides specific for HLA-A*0201: namely, CMV
IE-1-256, IE-1-297, and IE-1-316 and the pp65495 peptide. Figure 5 shows the results of 23 subjects with CMV reactivation (Figure 5A and C) and
of 10 subjects (Figures 5B and D) in the no-CMV
896
group followed up for 1 year. IFN-␥ responsiveness
to CMV pp65 peptide (Figures 5A and B) and to
IE-1 (Figure 5C and D) was compared between
groups. With use of a contingency table (Fisher
exact test) to test whether CMV reactivation was
predictive for the presence of either CMV pp65–
reactive or CMV IE-1–reactive cells at any time
after transplantation, there were no significant differences between groups. The CMV group displayed more CMV pp65–reactive T cells (compare
Figure 5A and B) and CMV IE-1–reactive T cells
(compare Figure 5C and D), but these levels were
not significantly different.
Time to First Evidence of CD8 IFN-␥–Positive
CMV-Specific T Cells after HCT
We next asked whether there was a difference in
the time to first demonstration of immune reconsti-
CMV ICC and Tetramer Binding
Figure 3. CMV immunity and CMV infection. CMV reactivation is shown for each individual (patients 54, 70, 93, 94, and 105) as a bold line
at the top of the graph showing the time to PCR positivity after HCT. The number of pp65-495–specific /IFN-␥–positive cells is shown in
(A), and IE-1–specific peptides consisting of IE-1-256, IE-1-297, and IE-1-316 stimulation are shown in (B). The number of IFN-␥–positive
cells, expressed as x times 1 ⫻ 105 cells per liter (y-axis), is shown after subtraction of background data by using HIV peptide stimulation. The
time after HCT for each subject is on the x-axis, and, when available, the number of cells is shown for the donor before and after granulocyte
colony-stimulating factor treatment (PreG and PostG, respectively).
tution in the CMV and no-CMV groups. For this, we
defined immune reconstitution as the first observation
of at least 1 ⫻ 105/L CD8 IFN-␥ responsive to pp65495 peptide (Figure 6A) or to IE-1 peptide mixture
(Figure 6B). There was no difference between subjects
with and without CMV infection in the time to first
evidence of T-cell immunity by either pp65 or IE-1
BB&MT
responses. An important observation is that the time
to immune reconstitution was at least as early and
complete in subjects with no detectable CMV infection as in the subjects in whom CMV reactivation was
seen. One hundred percent immune reconstitution
was never attained in the subjects who had documented CMV infection. In all recipients, however, if
897
G. Gallez-Hawkins et al.
Figure 4. Time course of tetramer-positive T cells after HCT. The number of tetramer-positive cells specific to pp65-495 (A and C) and the
sum of tetramer-positive cells specific to IE-1-297 and IE-1-316 (B and D) are shown. The left panels (A and B) represent results for subjects
with documented CMV reactivation, and the right panels (C and D) represent results in subjects with no documented CMV infection. The
number of IFN-␥–positive cells is shown as x times 1 ⫻ 105 cells per liter, and the median value is calculated at each time point and shown
on the graph as a continuous line.
immune reconstitution was to occur during the first
year after HCT, then it was present by 6 months.
DISCUSSION
In this work, the evaluation of CMV immunity in
HCT HLA-A*0201 subjects relied on 2 assays (the
ICC/IFN-␥ and tetramer-binding assays) and on reactivity to 2 major CMV proteins (CMV pp65 and
CMV IE-1). This is the first comparative study that
used both of these antigens to characterize the time of
CMV immune reconstitution after HCT. The tetramer-binding assay has been widely used because it
permits the quantification of cytotoxic T lymphocytes
in a simple assay by using flow cytometry [8,11,
13,14,18,29-39]. It is useful for measuring cellular
immune responses to CMV pp65 because it targets
898
mainly 1 immunodominant epitope, the pp65-495 of
the HLA-A*0201 allele. Other epitopes encompassing
CMV pp65 have been described for other HLA alleles, thus suggesting that CMV pp65 is a major target
for immune responses [5]. In contrast, the CMV IE-1
protein presents multiple peptides in the same HLA
context and therefore requires multiple tetramer reagents. This places added requirements on the tetramer technology for determining the immune status
of an individual, but multiplex tetramer reagents have
been developed to address this. Our report shows that
the median value of CD8⫹/tet⫹ cells binding to
tetpp65-495 in subjects with documented CMV infection after HCT, expressed as the concentration of
cells per microliter, was 10-fold higher than that in
patients with no such CMV infection. Also, the pp65495 peptide–specific response was consistently higher
CMV ICC and Tetramer Binding
Figure 5. Levels of CMV-specific CD8⫹/IFN-␥–positive cells at various times after HCT. An ICC assay on fresh blood from 33 HLA-A*0201
subjects stimulated with CMV-pp65 peptide (A and B) or a mixture of IE-1–specific peptides (CMV IE-1-256, -297, and -316; C and D) is
shown. A and C, Number of cells obtained in recipients with detectable CMV reactivation (23 subjects). B and D, Results from those with no
detectable CMV infection (10 subjects).
than the IE-1–specific response. Of note, when compared with the ICC assay at peak time, only 43% of
pp65-495-tet⫹ cells were expressing IFN-␥ (compare
Figure 2A and Figure 4A); this suggests that more
than half of the tet⫹ cells were not functional and
agrees with the results of others [1,32,37].
During CMV reactivation, the CMV IE-1 protein
is the first protein to be expressed in infected cells,
and, therefore, it should be part of the immune response during immune reconstitution. By using the
peptides uncovered with the HLA-A*0201 transgenic
mouse model, namely, the CMV IE-1-256, -297, and
-316 peptides [23], PBLs from HCT recipients were
stimulated and tested by ICC/IFN-␥ assay at various
times up to 1 year after HCT. After detectable CMV
reactivation, all 3 CMV IE-1 peptides stimulated
Figure 6. Percentage of HCT recipients with CMV-specific T-cell immune reconstitution. The cumulative incidence of IFN-␥–positive cells
stimulated with pp65 (A) or a mixture of IE-1 peptides (B) is shown at the indicated times after HCT. The recipients with detectable CMV
infection (CMV group) are shown as a continuous line, and those with no detectable CMV infection (no CMV) are shown as a hatched line.
The log-rank test showed no significant differences between the 2 curves in (A) and (B).
BB&MT
899
G. Gallez-Hawkins et al.
PBLs in all 5 subjects at some time between day 40
and day 360 after HCT. There was no indication that
one peptide use was more prominent than the others
in the CMV reactivation group; therefore, a CMV
IE-1 peptide mixture was used in the follow-up study
of 33 subjects to determine the detection of immune
cellular reactivity. Moreover, in the HLA-A*0201
context, the immune response in the CMV reactivation group to CMV IE-1 was always lower than that to
CMV pp65. The highest median value was observed
at day 360 (Figure 2B), thus suggesting that there may
be additional antigenic stimulation and expansion after CMV reactivation. The CMV IE-1 ICC cell count
was still low at day 180 even though the median time
to the first day of CMV reactivation was day 55 after
HCT.
The pp65 response increased briskly between days
100 and 180 after HCT and remained high at 1 year,
especially after CMV infection. In the no-CMV
group, there was a low-level tet⫹ response to both
pp65-495 and IE-1 during the same period, and this
population of T cells never expanded, remaining at
very low levels at 1 year. It is likely that exposure to
CMV antigen is required for these CD8 expansions
and that, in the no-CMV group, if there was an initial
CMV reactivation state, it was then limited, never
reached detectable levels in blood, and was never
sufficient to lead to expansion of CD8-specific cells.
Nevertheless, on the basis of these immunologic
observations, it seems that many allogeneic HCT
recipients at risk for CMV undergo occult CMV
reactivation even when there is no documented infection. The fact that the patients with no detectable
CMV infection had the earlier and the most complete
immune reconstitution to CMV suggests a previously
unrecognized phenomenon of subclinical CMV reactivation that is shown only by the occurrence of
CMV-reactive T cells.
The question is why there is a reduced response to
CMV IE-1 compared with pp65. The immune system
may not be exposed as much to the IE protein during
the reactivation process as it would be during a primary infection. Indeed, in congenital and postnatal
CMV infection, IE-1–specific responses dominated by
1 year of age, regardless of the specificity of initial
responses [22]. This response to the CMV IE-1 gene
is a typical response to a primary infection and is in
contrast to what is seen in adults with chronic infection. In HLA-A*0201 HCT subjects, although the
response to CMV pp65 always predominates, the response to CMV IE-1 was detectable and present in the
CMV group. It is noteworthy that 3 subjects in the
CMV reactivation group showed ICC positivity to all
3 CMV IE-1 peptides simultaneously at day 360, but
this was never seen in the no-CMV group. These
results show for the first time a multipeptidic IE-1–
specific immune response within the same blood sam900
ples. By tetramer-binding assay, we know that CD8
cells are directed toward IE-1-297 and IE-1-316 in
equal numbers in both the CMV reactivation and
no-CMV groups, and when they are shown combined
in Figure 4B and D, there is still no significant difference between groups. Therefore, although the number of CMV IE-1 immune cells is small, they are
present in sufficient number and seem to respond to
documented or occult CMV reactivation. It remains
to be determined whether this response contributes to
protection from progressive infection.
The CMV-specific T-cell immunity observed in
the recipients with no CMV infection could have been
passively transferred in the graft, and this would account for the increased numbers of T cells at day 40.
In this relatively small study, the percentage of recipients with ⱖ1 ⫻ 105 IFN-␥ CD8 cells at this early
time after HCT was not significantly different. However, of note, the number of immune subjects in the
no-CMV group continued to increase despite the absence of detectable CMV infection. This suggests that
the transfer of CMV-reactive T cells might protect
from detectable CMV infection, perhaps by control of
virus replication after reactivation, and this justifies
attempts to augment the transfer of such cells at the
time of HCT to test such a hypothesis. The manipulation of donor cells either by in vitro expansion or by
in vivo stimulation with a vaccine may prevent CMV
reactivation through such adoptive immune therapy.
ACKNOWLEDGMENTS
The authors thank Brenda Williams and the staff
of the General Clinical Research Center for sample
preparation; Allison Sano, Kathryn Patane, and the
staff of the bone marrow transplantation unit for their
work in patient recruitment and obtaining samples;
and Drs. Mark Davis and Pat Roth for providing the
HLA-A*0201 and ␤2-microglobulin expression constructs obtained under a material transfer agreement
with Beckman-Coulter. Finally, we thank Wengang
Chen in the pathology department for his construct
pDCMV-gB that was used in the quantitative CMV
PCR. This study was supported in part by US Public
Health Service grant nos. RO1 AI58148 (J.A.Z.), PO1
CA 30206 (S.J.F.), and RO1 AI43267 and CA77544
(D.J.D); by grant nos. 6122-01 (S.F.L.) and 6116-98
(D.J.D) from the Leukemia and Lymphoma Society;
and by grant no. MO1 RR-43 from the General Clinical Research Center branch of the National Center
for Research Resources, National Institutes of Health.
REFERENCES
1. Lacey SF, Diamond DJ, Zaia JA. Assessment of cellular immunity to human cytomegalovirus in recipients of allogeneic stem
cell transplants. Biol Blood Marrow Transplant. 2004;10:433-447.
CMV ICC and Tetramer Binding
2. Elkington R, Walker S, Crough T, et al. Ex vivo profiling of
CD8⫹-T-cell responses to human cytomegalovirus reveals
broad and multispecific reactivities in healthy virus carriers.
J Virol. 2003;77:5226-5240.
3. Greenberg PD, Reusser P, Goodrich JM, Riddell SR. Development of a treatment regimen for human cytomegalovirus
(CMV) infection in bone marrow transplantation recipients by
adoptive transfer of donor-derived CMV-specific T cell clones
expanded in vitro. Ann N Y Acad Sci. 1991;636:184-195.
4. Wills MR, Carmichael AJ, Mynard K, et al. The human cytotoxic T-lymphocyte (CTL) response to cytomegalovirus is
dominated by structural protein pp65: frequency, specificity,
and T-cell receptor usage of pp65-specific CTL. J Virol. 1996;
70:7569-7579.
5. Diamond DJ, York J, Sun JY, Wright CL, Forman SJ. Development of a candidate HLA A*0201 restricted peptide-based
vaccine against human cytomegalovirus infection. Blood. 1997;
90:1751-1767.
6. Solache A, Morgan CL, Dodi AI, et al. Identification of three
HLA-A*0201-restricted cytotoxic T cell epitopes in the cytomegalovirus protein pp65 that are conserved between eight
strains of the virus. J Immunol. 1999;163:5512-5518.
7. Engstrand M, Lidehall AK, Totterman TH, Herrman B, Eriksson BM, Korsgren O. Cellular responses to cytomegalovirus in
immunosuppressed patients: circulating CD8⫹ T cells recognizing CMVpp65 are present but display functional impairment. Clin Exp Immunol. 2003;132:96-104.
8. Lacey SF, Gallez-Hawkins G, Crooks M, et al. Characterization of
cytotoxic function of CMV-pp65-specific CD8⫹ T-lymphocytes
identified by HLA tetramers in recipients and donors of stem-cell
transplants. Transplantation. 2002;74:722-732.
9. Mohty M, Mohty AM, Blaise D, et al. Cytomegalovirus-specific
immune recovery following allogeneic HLA-identical sibling
transplantation with reduced-intensity preparative regimen.
Bone Marrow Transplant. 2004;33:839-846.
10. Gallez-Hawkins G, Lomeli NA, Li X, et al. Kinase-deficient
CMVpp65 triggers a CMVpp65 specific T-cell immune response in HLA-A*0201.Kb transgenic mice after DNA immunization. Scand J Immunol. 2002;55:592-598.
11. Aubert G, Hassan-Walker AF, Madrigal JA, et al. Cytomegalovirus-specific cellular immune responses and viremia in recipients of allogeneic stem cell transplants. J Infect Dis. 2001;
184:955-963.
12. Chen FE, Aubert G, Travers P, Dodi IA, Madrigal JA. HLA
tetramers and anti-CMV immune responses: from epitope to
immunotherapy. Cytotherapy. 2002;4:41-48.
13. Engstrand M, Tournay C, Peyrat MA, et al. Characterization of
CMVpp65-specific CD8⫹ T lymphocytes using MHC tetramers in kidney transplant patients and healthy participants.
Transplantation. 2000;69:2243-2250.
14. Gratama JW, van Esser JW, Lamers CH, et al. Tetramer-based
quantification of cytomegalovirus (CMV)-specific CD8⫹ T
lymphocytes in T-cell-depleted stem cell grafts and after transplantation may identify patients at risk for progressive CMV
infection. Blood. 2001;98:1358-1364.
15. Cwynarski K, Ainsworth J, Cobbold M, et al. Direct visualization of cytomegalovirus-specific T-cell reconstitution after allogeneic stem cell transplantation. Blood. 2001;97:1232-1240.
16. Manley TJ, Luy L, Jones T, Boeckh M, Mutimer H, Riddell
SR. Immune evasion proteins of human cytomegalovirus do not
prevent a diverse CD8⫹ cytotoxic T-cell response in natural
infection. Blood. 2004;104:1075-1082.
BB&MT
17. Bunde T, Kirchner A, Hoffmeister B, et al. Protection from
cytomegalovirus after transplantation is correlated with immediate early 1-specific CD8 T cells. J Exp Med. 2005;201:10311036.
18. Khan N, Cobbold M, Keenan R, Moss PA. Comparative analysis of CD8⫹ T cell responses against human cytomegalovirus
proteins pp65 and immediate early 1 shows similarities in precursor frequency, oligoclonality, and phenotype. J Infect Dis.
2002;185:1025-1034.
19. Prod’homme V, Retiere C, Imbert-Marcille BM, Bonneville M,
Hallet MM. Modulation of HLA-A*0201-restricted T cell responses by natural polymorphism in the IE1(315-324) epitope
of human cytomegalovirus. J Immunol. 2003;170:2030-2036.
20. Retiere C, Prod’homme V, Imbert-Marcille BM, Bonneville M,
Vie H, Hallet MM. Generation of cytomegalovirus-specific
human T-lymphocyte clones by using autologous B-lymphoblastoid cells with stable expression of pp65 or IE1 proteins: a tool
to study the fine specificity of the antiviral response. J Virol.
2000;74:3948-3952.
21. Frankenberg N, Pepperl-Klindworth S, Meyer RG, Plachter B.
Identification of a conserved HLA-A2-restricted decapeptide
from the IE1 protein (pUL123) of human cytomegalovirus.
Virology. 2002;295:208-216.
22. Gibson L, Piccinini G, Lilleri D, et al. Human cytomegalovirus
proteins pp65 and immediate early protein 1 are common
targets for CD8⫹ T cell responses in children with congenital
or postnatal human cytomegalovirus infection. J Immunol.
2004;172:2256-2264.
23. Gallez-Hawkins G, Villacres MC, Li X, Sanborn MC, Lomeli
NA, Zaia JA. Use of transgenic HLA A*0201/Kb and HHD II
mice to evaluate frequency of cytomegalovirus IE1-derived peptide usage in eliciting human CD8 cytokine response. J Virol.
2003;77:4457-4462.
24. Zaia JA, Schmidt GM, Chao NJ, et al. Preemptive ganciclovir
administration based solely on asymptomatic pulmonary cytomegalovirus infection in allogeneic bone marrow transplant
recipients: long-term follow-up. Biol Blood Marrow Transplant.
1995;1:88-93.
25. Bensinger WI, Martin PJ, Storer B, et al. Transplantation of
bone marrow as compared with peripheral-blood cells from
HLA-identical relatives in patients with hematologic cancers.
N Engl J Med. 2001;344:175-181.
26. Limaye AP, Huang ML, Leisenring W, Stensland L, Corey L,
Boeckh M. Cytomegalovirus (CMV) DNA load in plasma for
the diagnosis of CMV disease before engraftment in hematopoietic stem-cell transplant recipients. J Infect Dis. 2001;183:
377-382.
27. Dunn HS, Haney DJ, Ghanekar SA, Stepick-Biek P, Lewis DB,
Maecker HT. Dynamics of CD4 and CD8 T cell responses to
cytomegalovirus in healthy human donors. J Infect Dis. 2002;
186:15-22.
28. Heijnen IA, Barnett D, Arroz MJ, et al. Enumeration of antigen-specific CD8⫹ T lymphocytes by single-platform, HLA
tetramer-based flow cytometry: a European multicenter evaluation. Cytometry B Clin Cytom. 2004;62:1-13.
29. Hebart H, Rauser G, Stevanovic S, et al. A CTL epitope from
human cytomegalovirus IE1 defined by combining prediction
of HLA binding and proteasomal processing is the target of
dominant immune responses in patients after allogeneic stem
cell transplantation. Exp Hematol. 2003;31:966-973.
30. Komatsu H, Sierro S, Cuero A, Klenerman P. Population
901
G. Gallez-Hawkins et al.
31.
32.
33.
34.
35.
902
analysis of antiviral T cell responses using MHC class I-peptide
tetramers. Clin Exp Immunol. 2003;134:9-12.
Villacres MC, Lacey SF, Auge C, Longmate J, Leedom JM,
Diamond DJ. Relevance of peptide avidity to the T cell receptor for cytomegalovirus-specific ex vivo CD8 T cell cytotoxicity. J Infect Dis. 2003;188:908-918.
Ouyang Q, Wagner WM, Wikby A, et al. Large numbers of
dysfunctional CD8⫹ T lymphocytes bearing receptors for a
single dominant CMV epitope in the very old. J Clin Immunol.
2003;23:247-257.
Gratama J, Cornelissen J. Clinical utility of tetramer-based
immune monitoring in allogeneic stem cell transplantation.
BioDrugs. 2003;17:325-338.
La Rosa C, Wang Z, Brewer JC, et al. Preclinical development
of an adjuvant-free peptide vaccine with activity against CMV
pp65 in HLA transgenic mice. Blood. 2002;100:3681-3689.
Foster AE, Gottlieb DJ, Sartor M, Hertzberg MS, Bradstock
KF. Cytomegalovirus-specific CD4⫹ and CD8⫹ T-cells fol-
36.
37.
38.
39.
low a similar reconstitution pattern after allogeneic stem cell
transplantation. Biol Blood Marrow Transplant. 2002;8:501-511.
Villacres MC, Lacey SF, La Rosa C, et al. Human immunodeficiency virus-infected patients receiving highly active antiretroviral therapy maintain activated CD8⫹ T cell subsets as a
strong adaptive immune response to cytomegalovirus. J Infect
Dis. 2001;184:256-267.1
He XS, Rehermann B, Boisvert J, et al. Direct functional
analysis of epitope-specific CD8⫹ T cells in peripheral blood.
Viral Immunol. 2001;14:59-69.
Szmania S, Galloway A, Bruorton M, et al. Isolation and expansion of cytomegalovirus-specific cytotoxic T lymphocytes to
clinical scale from a single blood draw using dendritic cells and
HLA-tetramers. Blood. 2001;98:505-512.
Gandhi MK, Wills MR, Okecha G, et al. Late diversification in
the clonal composition of human cytomegalovirus-specific
CD8⫹ T cells following allogeneic hemopoietic stem cell
transplantation. Blood. 2003;102:3427-3438.