| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of
*
Pathology and
Infectious Diseases, Case Western Reserve University, Cleveland, OH 44106; and
University Hospital of Ulm, Section of Endocrinology, Ulm, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1012 different T cell clones with unique
receptors, T cells with specificity for any given Ag tend to occur at
very low frequencies (<<1/1000) even for clonally expanded memory cell
populations and, hence, have been generally inaccessible to direct
measurement. The classic approach used to establish clonal sizes of Ag-specific T cells has been limiting dilution analysis (LDA).3 The results obtained by LDA, however, have recently become controversial when new flow cytometry-based technical developments (tetramer and/or intracytoplasmic cytokine staining) that can directly visualize Ag-specific T cells in the high frequency range (>1/1,000) yielded fundamentally different results. While, for example, LDA suggested precursor frequencies of virus-specific CD8 cells to be in the range of 1/4000–1/10,000 (2), tetramer staining, intracytoplasmic staining, and ELISPOT analysis, in confirmation of each other, showed that these frequencies can be as high as 10% of all CD8 cells in the acute phases of the infection (3, 4, 5). This discrepancy between the indirect LDA data on which much of our present understanding of T cell population kinetics in general is based and the results obtained by methods suited for directly measuring Ag-specific T cells raised renewed interest in the most basic parameters that define immune function and memory in vivo, the clonal sizes and cytokine effector functions of memory T cells in freshly isolated material. By permitting a new look at T cells (1), the recently gained ability to monitor memory cells directly should help to resolve many of the controversial issues about T cell function in vivo, issues at the very heart of our understanding of immunological protection and of deficiency states such as the one caused by HIV infection.
Without the ability to visualize memory T cells directly, it is still unknown whether they are generally long-lived or short-lived, whether they require Ag to persist, or whether their frequencies fluctuate as a result of cross-reactive stimulations and/or cytokine-driven bystander reactions (6, 7). Evidence for all of these conflicting scenarios has been generated in murine models. Particularly during acute viral infections like HIV, the massive expansions of virus Ag-reactive T cells could result in a compensatory decline in the number of memory cells specific for third-party Ags as a result of the homeostatic control that keeps the overall number of T cells in the body constant (8, 9, 10). The opposite is conceivable as well, that long-lived memory cells specific, e.g., for tetanus toxoid (TT) or purified protein derivative of Mycobacterium tuberculosis (PPD) expand clonally under the influence of bystander cytokine during such infections (11, 12, 13). Furthermore, to understand the various states of T cell immunodeficiency, it is critical to know whether they are the result of reduced numbers of Ag-specific memory cells (frequencies), impaired function of these cells (reduced cytokine output per cell), a change in memory cell function (Th1/Th2 switch), or possibly a combination of these (14). Cytokine ELISPOT analysis (15) might provide a unique tool with which to address these questions.
Naive T cells do not produce type 1 or type 2 effector cytokines (such
as IFN-
or IL-5), and for Ag to induce them to differentiate into
memory cells producing IFN- or IL-5 requires several days
(16); therefore, when freshly isolated T cells contained
within PBMC are challenged with an Ag and subjected to IFN- or IL-5
ELISPOT assays of 24–48-h duration, as we did, it should be only the
memory cells that differentiated in vivo that produce the cytokine.
Moreover, because the duration of such assays (24–48 h) is too brief
for in vitro proliferation and/or differentiation to occur, the number
of spots detected should reflect both the frequency of memory cells
present in the cell isolate and the commitment to produce IFN- or
IL-5 imprinted on them in vivo. Furthermore, with the cytokine being
continuously captured around the secreting T cell for the entire
duration of the assay, the size of the spots produced should reflect
the total cytokine produced by the individual cell; hence, the assay
directly measures the biological function of these cells. Should the
above postulates hold, the cytokine ELISPOT assay would provide a
powerful tool for studies that characterize memory T cells based on
their frequencies in freshly isolated PBMC and the cytokine output of
individual cells; this is what we set out to establish in the first set
of experiments reported in this work.
We recently introduced technical improvements in ELISPOT analysis that include membranes that improve the signal-to-noise ratio (17) and computerized image analysis (18). We report in this work the first formal evidence that this modification of the assay is suited to characterize the frequency, cytokine signature, and cytokine output of individual memory cells specific for recall Ags. These advances enabled us to characterize, for the first time, the recall Ag-specific T cell memory cell pool in healthy individuals and HIV patients at the highest conceivable resolution and at the single cell level. This article is intended to establish the feasibility of monitoring Ag-specific CD4 memory in the low frequency range, not to provide a complete analysis of the range of responses exhibited by highly variable human population.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Peripheral blood was obtained by venipuncture from 22
HIV-positive patients (4 women, 18 men, ages between 27 and 66 years)
from the Special Immunity Unit at the University Hospitals (Cleveland,
OH). All patients were under HAART therapy at the time of testing. The
CD4 counts ranged from 150 to 1070 cells/ml, as specified in Fig. 4
C. Viral load was determined by PCR (Amplicor;
Hoffmann-LaRoche, Nutley, NJ) and is shown in Fig. 4C. The
patients were categorized according to the Centers for Disease Control
and Prevention (CDC) classification: 14 belonged to group 1, 7 to group
2, and 1 to group 3; because there were no statistically significant
differences in the data we obtained while testing their recall
Ag-specific memory response, the data were pooled. The 19 healthy
individuals that we tested (10 women, 9 men, between 22 and 45 years of
age) were members of our laboratory and adjoining laboratories. PBMC
were isolated from 10 to 30 ml of heparinized blood by standard Ficoll
density-gradient centrifugation (using Isoprep, Robbins Scientific,
Sunnyvale, CA; for T cell separation, 100 ml blood was processed). All
studies were performed under the approval of the Institutional Review
Board for Human Investigation at the University Hospitals of
Cleveland.
|
ImmunoSpot plates (Cellular Technology, Cleveland, OH) were
coated overnight at 4°C with capture Abs dissolved in PBS that were
specific for either IFN- (mAb M700A-E, 6 µg/ml; from Endogen,
Woburn, MA,) or IL-5 (mAb TRFK5, 5 µg/ml; hybridoma was obtained from
American Type Culture Collection, Manassas, VA, and the Ab was grown
and purified in our laboratory). The plates were then blocked with BSA
(10 g/L in PBS) for 1 h and washed three times with PBS. Cells
were plated in complete RPMI medium (94% RPMI + 5% ABO serum + 1%
L-glutamine). RPMI was from BioWhittaker (Walkersville,
MD); ABO serum was from Gemini Bioproducts (Calmasas, CA) and was heat
inactivated at 56°C for 30 min. When T cell lines were tested (see
Fig. 1), the indicated numbers of T cells were plated per well along
with a constant number of 300,000 autologous T cell-depleted APC (these
were obtained by E-rosetting (4) or plate adherence). For
the experiments shown, we used an IFN--producing human-alloreactive,
CD4+ T cell line (PAN) and an
IL-5-producing hemagglutinin peptide-specific human
CD4+ line. The APC were irradiated with 3000 rad
of rays from a cobalt source. Similar autologous APC were used for
testing purified T cell fractions (see Fig. 3):
CD3+ and CD4+ cells were
obtained by negative selection, passing PBMC though affinity columns
Human T-Cell Enrichment and Human CD4 Subset Column Kit (both from R&D
Systems, Minneapolis, MN). The efficacy of enrichment was controlled by
FACS analysis, staining with labeled anti-CD4, anti-CD8, and
anti-CD3 Abs (all from PharMingen, San Diego, CA). The enrichment
for the desired phenotype was between 87 and 94%. When unseparated
PBMC were tested, 3 x 105 freshly isolated
cells were plated. At the initiation of the activation culture, Ags
were added as specified in the figures. TT was from Connaught
(Swiftwater, PA; 1/100), PPD from Evans Medical (Langhurst, U.K.; 10
µg/ml), Standardized Mite DF from Bayer (Elkhart, IN; 200 AU/ml), and
PHA from Sigma (St. Louis, MO; 10 µg/ml). Control wells contained
responder PBMC with medium alone. After 24 h for IFN- assays,
or after 48 h for IL-5 assays, the plates were washed and
biotinylated detection Abs were added for 12 h, at 4°C. The
secondary Abs specific for IFN- were mAb M701 from Endogen (3
µg/ml) and for IL-5, mAb 18522D from PharMingen; they were in 10
g/L BSA/PBS with 0.5% Tween (Sigma). Subsequently, streptavidin-HRP
(Dakopatts, Glostrup, Denmark; 1/2000 dilution) in PBS/BSA/Tween was
added for 2 h at room temperature. The spots were developed using
AEC (Pierce Pharmaceuticals, Rockford, IL) solution: 10 mg/ml AEC
dissolved in 1 ml N,N-dimethyl formamide (Fischer
Scientific, Fair Lawn, NJ), of which 1 ml was freshly diluted into 30
ml of 0.1 M sodium-acetate buffer (pH 5). This solution was then
filtered and mixed with 15 µl
H2O2 to provide the final
AEC development solution, of which 200 µl plated per well. The plates
were developed for 10–20 min for IFN- assays, and 45–60 min for
IL-5 assays; the reaction was stopped by rinsing with tap water when
clear spots became visible macroscopically. The plates/membranes were
air dried overnight before the plates were subjected to image
analysis.
|
|
We used a Series 1 ImmunoSpot Image Analyzer (Cellular
Technology) specifically designed for the ELISPOT assay. Digitized
images were analyzed for the presence of areas in which color density
exceeds background by an amount set on the basis of the comparison of
experimental wells (containing T cells, APC, and Ag) and control wells
(containing T cells and APC only); see, e.g., Fig. 2, A vs
B. After background and noise subtraction and separation of
the areas formed by touching and partially overlapping spots,
additional criteria of spot size and circularity are applied to gate
out speckles and artifacts caused by spontaneous substrate
precipitation and nonspecific Ab binding. Objects that do not meet
these criteria are gated out (the hatched lines in Fig. 2, D
and E), and areas that meet them are recognized as spots and
counted. The spot size distribution (see Fig. 2, D and
E; Fig. 3
, C and D; and Fig. 6A), a built-in function, is based on the array of spot
sizes in a given well sorted according to distinct size categories, and
these results were plotted in above figures.
|
|
, FITC-labeled IFN- Ab (Becton
Dickinson, San Jose, CA) was used; for staining of IL-5, FITC-labeled
TRFK-5; and for CD3, peridinin chlorophyl protein-labeled
anti-human CD3 Abs (Becton Dickinson). The isotype-matched control
mAbs were from Becton Dickinson. The samples were analyzed on a FACScan
flow cytometer (Becton Dickinson). As a protein transport inhibitor, we
used Monensin or brefeldin A (PharMingen). Statistical analysis
The overall significance of the differences seen for HIV
patients in Figs. 4 and 5 was calculated by the Mann-Whitney Rank Sum
Test, and for Fig. 6, using the t test.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and IL-5 ELISPOT assays measure the accurate frequencies of
recall Ag-specific CD4 memory cells with a detection limit of 1/300,000
cells
Although ELISPOT analysis has frequently been performed in
settings of T cell recall assays, it has not been stringently
established whether the assay accurately measures the true frequency of
CD4 memory T cells in freshly isolated PBMC in the low frequency range.
We first tested human CD4+ T cell lines and
freshly isolated PBMC activated with mitogen because the high frequency
of cytokine-producing T cells under these conditions permitted us to
directly compare the resolution of intracytoplasmic FACS staining and
ELISPOT assays. The percentage of T cells producing IFN- and IL-5
was first established by intracytoplasmic FACS analysis (Fig. 1, A and B, and in
Fig. 1C the numbers in parentheses). When these activated T
cells were serially diluted with inactivated T cells of the same type,
the detection limit of intracytoplasmic staining was reached at
1/5,000 (data not shown). In the ELISPOT assays, however, a linear
relationship between the cell numbers plated and the spot numbers
detected was seen down to 1/300,000 cells, the lowest dilution tested.
Moreover, the number of cytokine spots detected closely matched the
number of cytokine-producing T cells plated after correction for the
frequency of cytokine-producing cells (as established by
intracytoplasmic staining). When, for example, 123 activated T cells of
the PAN line were plated per well (of which 23% were
IFN- positive by flow cytometry), together with 300,000 APC, then 27
± 2 spots were detected, that is, 22% of the plated T cells
generated IFN- spots. Although both the ELISPOT assay and
intracytoplasmatic cytokine staining gave equal results for detecting
per cell cytokine production, the ELISPOT assay was able to
detect a single cell in the presence of 300,000 bystander cells where
intracytoplasmic staining could not, having reached a functional
detection limit at 1 in 5,000.
We next attempted to detect recall Ag-induced cytokine production in
freshly isolated PBMC. When PBMC of healthy donors were tested with
intracytoplasmic cytokine staining for IFN- and IL-5 production
induced by TT and PPD, no signal over background was seen. In contrast,
very clear signals were obtained in ELISPOT assays (Fig. 2B provides an example of
TT-stimulated PBMC; the spontaneous cytokine production of these PBMC
is shown in Fig. 2A). By using criteria of spot size and
density analogous to gating in FACS analysis, image analysis of
ELISPOTS permitted the objective evaluation of the results to
distinguish between background and signal, count the numbers of spots,
and determine their size distribution (Fig. 2, D and
E).
The spectrum of spot sizes induced by TT (or PPD) was comparable with
those seen in mitogen-stimulated T cell lines or PBMC (Fig. 2, D vs E, and sample wells B vs
C). After verifying with T cell lines that these spots were
produced by individual T cells (see above), the amount of cytokine
produced in total and by individual T cells could be derived from the
size histogram. Based on this, the similarity in the spot size
distributions found in the recall Ag-activated PBMC and T cell lines
suggested that the cytokine spots from the PBMC population also
reflected cytokine produced by individual T cells. This notion was
further confirmed when the PBMC activated by TT and PPD were serially
diluted, as the resulting plot of the number of PBMC plated against the
number of IFN- and IL-5 spots detected conformed to a linear
function that passed through the origin (Fig. 1D). These
data establish that IFN- and IL-5 ELISPOT assays are suitable for
direct visualization of cytokine-producing CD4 memory cells in the low
frequency range (IL-4 and IL-10 data are not shown, as the production
of these cytokines was not confined to memory cells).
Recall Ags induce IFN- and/or IL-5 production only in sensitized
individuals
We next tested whether the IFN- and the IL-5 induced by the
test Ags in PBMC are secreted by Ag-specific memory T cells. Most
healthy individuals that we tested produced IFN- and IL-5 when
challenged with TT, consistent with them all being vaccinated with TT.
In contrast, PPD and dust mite mix (DM) Ags induced cytokine production
only in individuals that had a history of vaccination or allergy,
respectively (examples are shown in Table I). In most PPD-responsive individuals,
IFN- production was seen in the absence of Ag-induced IL-5
production, a result consistent with highly polarized type 1 immunity.
The type 1 nature of immune memory to PPD was confirmed by detecting,
additionally, IL-2 but not IL-4 spot formation (data not shown). DM Ag
induced IL-5 (and IL-4) production with variable numbers of IFN-
spots. DM-sensitized individuals showed, therefore, type 2 polarized or
mixed memory.
|
and/or IL-5 by recall
Ag-activated T cells
The production of type 1 cytokines can be inhibited by type 2
cytokines and vice versa. It therefore seemed important to test whether
the frequencies of IFN- and of IL-5 memory cells that we detected in
PBMC in ELISPOT assays are the uninhibited, true frequency of recall
Ag-specific T cells, or whether the activation of one type of memory
cells is masked by the activity of the other type. The presence of both
types of memory cells in healthy donors that were vaccinated with PPD
and also allergic to DM permitted us to directly address this question.
Unpurified PBMC (Fig. 3A) and
purified T cells were tested (with purified autologous plate adherent
cells functioning as APC) (Fig. 3B) for the frequency of
cells producing IFN- and IL-5 after challenge with the individual
Ags, PPD and DM, and with a combination of both. PPD alone induced a
unipolar IFN- response; DM Ag triggered a mixed response. When
challenged with both Ags simultaneously, the number of IFN- and IL-5
spots detected corresponded to the sum of the IFN- and IL-5 spots
that the individual Ags induced. The activation of the type 1 and type
2 memory cells in the culture was, therefore, mutually uninhibited. The
results obtained with purified T cells (>95% enrichment) and with
bulk PBMC was, when corrected for the frequency of CD3 cells present in
the test populations, identical. Further cell separations showed that
these cytokine-producing cells as well as the TT-reactive cells are
CD4+ (data not shown). The production of IFN-
and/or IL-5 per cell was also unaffected by the concurrent production
of the other type of cytokine (Fig. 3, C and D).
The findings provide additional evidence that true frequencies of
cytokine-producing cells are established by the IFN- and IL-5
cytokine ELISPOT assays and that they are directly visualizing the
memory cells while avoiding the caveats of more complex bioassays in
use for low frequency analysis. We attribute this lack of
cross-inhibition to the fact that both types of memory cells are
resting when plated, not secreting cytokine, such that their initial
activation occurs in the absence of the potentially inhibitory cytokine
of the other type. Before the induced cytokines can establish
inhibitory effects, the plate-bound cytokine is developed.
The frequency of memory cells producing IFN- and/or IL-5 in
healthy individuals is stable over time
Because it has been a matter of debate whether the clonal sizes of
Ag-specific memory T cells in an individual are stable over time
(6, 7), we repeatedly tested the recall responses to TT
and PPD of four healthy donors over a period of up to 4 yr, measuring
IFN- and IL-5 production. The frequencies of cytokine spots were
within a tight range over this time period (Table II). After having established the
validity of the cytokine ELISPOT approach for monitoring individual CD4
memory cells in PBMC and after having first gained insight by testing
healthy individuals, we set out to study the frequency and single cell
cytokine output of recall Ag-specific memory cells in HIV patients.
|
and
IL-5 shows no evidence of type 2 bias in HIV patients on HAART therapy
We tested 22 HIV patients whose CD4 cell counts were between 150
and 1070 cells/ml and all of whom were on HAART therapy at the time of
testing. The numbers of IFN- spots induced in their PBMC by TT were
significantly decreased (Mann-Whitney Rank Sum Test, p
= 0.006) compared with the 18 healthy donors that were tested (Fig. 4A). The apparent reduction in
IL-5 production (Fig. 4B) was not statistically significant,
however (p = 0.143). These data are consistent
with a moderate TT-specific Th2 bias in the HIV patients (19, 20). When the production of these cytokines was measured for the
entire T cell pool by stimulating cells with mitogen (PHA), the
frequency of IFN--producing cells was slightly increased
(p = 0.083), while that of the IL-5-producing
cells was in the same range (Fig. 4B). At the polyclonal
level, when tested at single cell resolution, we therefore found no
type 2 bias in the response of these patients.
The frequency of TT-reactive memory cells in HIV patients is inversely defined by the viral load
The patients were subdivided into groups with
<103,
103–104, and
>104 copies of HIV-RNA/ml of blood. The group
with <103 copies (n = 12) had
frequencies of IFN--producing TT-reactive cells indistinguishable
from those of healthy individuals (p = 0.615).
Patients with a viral load exceeding 103
(n = 12), however, had considerably lower, barely
detectable numbers of TT-reactive IFN--producing memory
cells (p < 0.001). This reduction in
frequencies did not reflect an overall reduction in CD4 cell numbers in
the patients because the frequencies shown (Fig. 4, A and
B) have been normalized for CD4 cells and because the high
viral loads did not correlate (r2 = 8.406)
with low CD4 cell numbers in patients (Fig. 4C). Moreover,
the reduction in TT-reactive CD4 cells was selective, as PHA-induced
IFN- production was essentially unaffected by the high viral load
(p = 0.328).
HIV patients that are PPD skin test negative show IFN- responses
to PPD
Only bacillus Calmette-Guérin-vaccinated healthy donors
displayed IFN- responses to PPD (Fig. 5A, p =
0.001). Nine of the twenty-two patients showed a significant PPD recall
response (Fig. 5B, p < 0.001).
The frequency of the patients’ IFN--producing memory cells was
in the range of 70–400 cells per million CD4 cells, slightly but not
statistically significantly reduced when compared with the 10
vaccinated healthy individuals that we tested
(p = 0.064). While all PPD-responsive healthy
donors tested positive in the PPD:Mantoux skin test, none of the four
PPD-responsive patients that we succeeded to test was skin test
positive (data not shown). As in the healthy donors, the PPD recall
response in the HIV patients was dominated by IFN--producing memory
cells; two of the patients, however, in addition to showing a high
frequency IFN- response, also had IL-5-producing memory cells, a
response type that was not seen in healthy individuals. The 13 patients
that did not respond to PPD fell into the full spectrum of high,
intermediate, and low viral loads (data not shown). Their lack of
responsiveness, therefore, could be attributed either to the
immunodeficiency that was seen for TT in patients with high viral
loads, or to the lack of sensitization to mycobacterial Ags. These two
possibilities cannot be readily distinguished for the PPD response,
because none of the HIV patients have been bacillus
Calmette-Guérin vaccinated, and their sensitization to
mycobacterial Ags was environmental. (This distinction can be made for
the tetanus toxoid response, however, because all HIV patients have
been vaccinated with TT.)
Reduced IFN- and IL-5 production per cell in a subpopulation of
HIV patients
ELISPOT assays measure, over the entire test period (24–48 h in
our set up), the amount of cytokine secreted by individual T cells
physiologically stimulated by encountering Ag; hence, where there is
decreased cytokine production in immunodeficiency such as occurs in
HIV, these assays can determine whether it was caused by a decrease in
the frequency of the responding cells or a reduction in the cytokine
output of the individual cells. In 17 of the 22 patients, the
production of IFN- and IL-5 per cell in response to stimulation with
TT, PPD, and PHA was indistinguishable from healthy individuals;
however, five of the patients displayed spots of dramatically reduced,
although still clearly detectable, size after TT, PPD, and PHA
challenge when tested in parallel, under identical assay conditions.
When studied by the t test, the changes in spot sizes <0.05
mm2 reached statistical significance
(p = 0.0375), as did the reduction of spots
that exceeded this size (p = 0.0001). Fig. 6A shows representative spot
size histograms for these two response types, one from a patient who
responded as healthy individuals do and one who displayed the reduced
spot size. Enlarged details of wells are also shown to illustrate both
response types (Fig. 6, C and D; the medium
control for 6C is shown in 6B). In these two
groups, the frequencies of cytokine-producing cells were in a
comparable range; it was the cytokine production per cell that
differed. To date, this reduced cytokine production per cell was seen
only in HIV patients; we encountered it neither in transplant patients
(n = 30), which were on immunosuppressive therapy
(21), nor in other patient groups (n >
100 to date) that we are in the process of studying and that include
individuals with type 1 diabetes, multiple sclerosis, and tumors. The
lower production of cytokines in HIV-infected patients did not
correlate with viral load, CD4 count, or any other observable feature
(data not shown).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In previous experiments, in which tetramer technology was used to study
Ag-specific CD8 responses in HIV and other viral infections, it came as
a major surprise that the numbers of the virus-specific cells can rise
to frequencies as high as 10% of all CD8 cells during acute infection:
data obtained with both ELISPOT (23) and intracytoplasmic
cytokine staining (24) confirmed these results for high
frequency CD8 cells. Similar information has not yet been obtained for
Ag-specific CD4 cells. Our testing of CD4 cell lines in ELISPOT and
intracytoplasmic staining in parallel showed that both assays provide
essentially identical frequency measurements in mitogen-stimulated cell
populations. Although we found that intracytoplasmic cytokine staining,
like FACS analysis in general, had a detection limit of 1/5000
(25), we showed that the IFN- and IL-5 ELISPOT approach
provides accurate frequency measurements down to frequencies as low as
1 in 300,000 cells (Fig. 1C). Moreover, by performing
experiments with purified cells, we established that the cytokine
signals obtained when testing PBMC for recall Ag responses were derived
from T cells (Fig. 3). The cytokine responses were present only in
sensitized individuals (Table I), as is consistent with our cytokine
ELISPOT findings in experiments comparing responses in sensitized with
those of naive or control-immunized mice (17, 18). This
observation is also consistent with the fact that naive T cells do not
produce IFN- or IL-5 (16). Importantly, while several
cell lineages other than T cells can produce IFN- and IL-5, no such
nonspecific responses were induced by the test Ags in the recall assays
or by the cytokines released by the Ag-stimulated T cells. The linear
relationship between the number of primed T cells present in the
cultures and of the number of spots detected (Fig. 1, C and
D), despite the presence of a constant number of APC, shows
that only cognate cytokine production from activated T cells was
detected. The exact number of T cells that expressed these cytokines
was therefore detected at all cell dilutions in the ELISPOT assay down
to 1/300,000, thus providing accurate frequency measurements for type 1
and type 2 cells producing IFN- and IL-5, respectively (Fig. 1).
Enabled by this high resolution, we report in this work the first
direct visualization of the recall Ag-specific, CD4 memory cell pool in
healthy individuals and in HIV patients on HAART therapy.
Memory cells reactive to TT and PPD in healthy individuals (and in most
HIV patients) occurred in the frequency range 1/5,000–1/100,000 CD4
cells (Fig. 4). Despite considerable interindividual variations, which
should not come as a surprise given the fully MHC-disparate genotypes
and different Ag histories of the test population, the frequency of
Ag-specific memory cells in healthy individuals was remarkably constant
over time. We tested four donors repeatedly over time periods of up to
4 yr and found that the frequencies of memory cells reactive to TT and
PPD that produced IFN- and IL-5 were essentially unchanged (Table II). These data might contribute to the ongoing discussions on the
population kinetics of memory cells. It has been shown in animal models
and proposed also for humans that, during infections, TNF-
and IL-15
cause clonal expansions in third-party, Ag-specific memory cell
populations (11, 12, 13). The healthy (non-HIV-infected)
donors experienced several episodes of flu-like infections during the
testing period, during which they displayed high IFN- activity in
medium control wells (data not shown). After recovery, this spontaneous
production ceased and the frequencies of Ag-specific spots returned to
what they were before. These data support the notion that memory cell
populations, once their clonal size has established itself after
immunization, remain stable in the absence of specific stimulation by
the Ag.
Although the TT-specific memory cell pool in most healthy test persons
comprised both cells producing IFN- and IL-5 (Table I, Fig. 4), the
PPD-specific memory cells produced IFN-, but not any IL-5 (Table I,
Fig. 5). Conversely, DM Ag induced a highly polarized IL-5 response,
which in some donors occurred without any IFN- production. These
polarized response phenotypes are not the result of dominant cytokine
response types overriding or masking less dominant ones, but reflect
the true frequencies of memory cells committed to the production of the
respective cytokine (Fig. 3). Based on studies of long-term cultured T
cell clones (which can lose differentiation in cell culture) and on
assays that are not suited to defining the cellular sources of the
individual cytokine, it has been widely held that the clear-cut
polarization of T cell responses along the Th1/Th2 paradigm seen in the
mouse does not hold for humans (26). Our data, obtained at
single cell resolution, suggest that IFN- and IL-5 can be expressed
by human memory T cells in a mutually exclusive manner, favoring
emerging concepts that the expression of type 1 and type 2 cytokine
genes is individually regulated in T cells (27, 28, 29).
Indeed, when we performed two-color ELISPOT assays simultaneously
detecting IL-5 and IFN-, we found that these cytokines were not
produced by Th0-type cells that coexpress both, but that IFN- and
IL-5 were produced by different CD4 cells (manuscript in
preparation).
We applied this high resolution analysis to memory cell repertoires in HIV patients in the hope of gaining new insights into the basic parameters defining Ag-specific T cell immunity, which were still controversial. It was important to study these parameters because they could change as a consequence of both HIV infection and therapy. For example, the production of T cells is reduced in untreated patients, as is the life span of the T cells (30, 31), yet, with antiviral treatment, the generation of precursor cells increased again and exceeded the normal rate, while the t1/2 of T cells remained strongly reduced (32, 33). After treatment, there is an initial increase in the numbers of memory T cells, and only later do frequencies of T cells with a naive phenotype return to approximately the normal range, at which time CD8 cell numbers fall (34, 35). Moreover, the frequency of recall Ag-specific memory T cells in the blood may change if they are redistributed between the extralymphoid compartments and the recirculating pool, as has been postulated; additionally, depletion/exhaustion and regeneration may contribute to the overall T cell repertoire in ways that are still debated (36, 37). It is also unclear how the massive expansion of the virus-specific CD8 cells affects the CD4 memory cell pool specific for third-party Ags, including those reactive to recall Ags, as is whether a bias toward type 1 or type 2 immunity in patients exists. All of these problems emphasize the need for high resolution analysis.
Previous attempts to establish recall Ag-specific T cell memory in
freshly isolated cellular material from patients relied on readout
systems that do not measure the frequency of the responding cells
(ELISAs, RT-PCR, proliferation); other studies utilized T cell clones
and in vitro expanded cell lines, which also do not provide data on the
frequencies or state of differentiation/function of Ag-specific cells
in vivo (reviewed in Ref. 38). LDAs were performed
(39) to determine the precursor frequency of TT-reactive
IL-2-producing T cells (the limitations of which are specified above
and in 1). In one study, the numbers of PPD-specific,
IFN--producing T cells were measured by ELISPOT (40).
Most of these studies came to the conclusion that T cell reactivity
declines with the progression of HIV infection, but they left open the
questions of how T cell function correlates with virus load and whether
this loss of function is a result of decreased frequencies or decreased
cytokine output, i.e., whether it is a function of T cell exhaustion,
the dilution of these cells by virus-specific CD8 cells, or their
suppressed state. Although many of these changes seem to be reversible
after HAART therapy (34, 35), the extent to which specific
immune reactivity is restored remains unclear. All 22 patients that we
studied were undergoing this therapy, so it was important for us to
establish how the reactivity of individual T cells to recall Ags in
these patients compares with that of healthy individuals.
All patients that had a low viral load (<103
copies/ml) had frequencies of TT-specific IFN--producing memory
cells in a range indistinguishable from healthy individuals, although
they had decreased frequencies of IL-5-producing memory cells
(Fig. 4). All of these patients were revaccinated with TT, as
is common practice for the cohort we studied. The lack of IL-5 (type 2)
memory to TT might have resulted from vaccinating/boosting concurrent
with the ongoing HIV infection that, through IL-12 induction, should
bias the differentiation of T cells toward the expression of type 1
cytokines.
In the patients with a viral load exceeding 103
copies/ml, the TT-reactive T cells producing IFN- occurred in
10-fold-reduced numbers, and IL-5-producing cells were barely
detectable (Fig. 4, A and B). Although reduced
IFN- production in mitogen-stimulated PBMC has been reported in
cases in which the disease has progressed (41), we
observed a direct correlation between viral load and recall Ag-induced
T cell function, but not to mitogen-driven cytokine production (Fig. 4, A and B). Because in patients the number of
virus-specific CD8 cells can expand considerably, it is possible that
the recall Ag-specific T cells have been diluted out.
Although the patients with the low viral load showed T cell response
similar to healthy individuals, it might be premature to conclude that
this signifies fully restored immune competence, as suggested by our
data on the PPD responses in the HIV patients. Nine of the patients had
PPD-reactive IFN--producing memory cells (Fig. 5). Either
cross-reactive stimulation with nonmycobacterial infections might prime
these T cells or opportunistic infections with mycobacteria (including
M. tuberculosis and Mycobacterium avium) might
specifically prime these PPD-reactive memory cells; future studies will
be required to distinguish between these possibilities. Irrespective of
how this PPD reactivity was acquired, these patients were PPD skin test
negative, while the vaccinated, and only the vaccinated, healthy
individuals were skin test positive, even though some had PPD-reactive
T cell frequencies equal to or lower than those of the patients. The
loss of skin test reactivity, which represents a PPD Ag-specific, type
1, T cell-mediated, delayed-type hypersensitivity (DTH) reaction, in
HIV infection is a known phenomenon thought to result from impaired T
cell function (42). Our data show that the DTH deficiency
of HIV patients to PPD neither results from the absence of PPD-reactive
memory nor from impaired ability to secrete IFN-. Whether a
counterregulatory, type 2 response contributes to this deficiency is a
point that still needs to be elucidated: unlike any of the healthy
donors, two of the HIV patients showed IL-5 responses to PPD. DTH is a
complex inflammatory response that depends not only upon T cells
secreting IFN-, but upon their ability to kill, to respond to
chemokines, and upon the response of the nonlymphoid compartment to T
cell-mediated inflammation (43), among others. Apparently,
one of these downstream responses is impaired in HIV. Our data suggest
that PPD ELISPOT recall assays might have diagnostic value. It could
show that patients have been exposed to M. tuberculosis, a
pathogen that represents a major infectious complication in HIV disease
and one that presently cannot be diagnosed reliably because of the
frequency of false-negative skin tests (42); if
recognized, this infectious complication can readily be prevented or
treated. Alternatively, it might signify infection with M.
avium, which is also an infectious complication requiring
treatment.
Five of the HIV patients showed a dramatic reduction in the cytokine
output of individual cells (Fig. 6), which, as opposed to reduced
numbers of T cells producing normal amounts of cytokine per cell,
constitutes a distinct mechanism for immunodeficiency. The impaired
cytokine production of individual T cells was seen after T cell
activation induced by TT, PPD, and mitogen (PHA and anti-CD3). The
frequency of T cells induced by these stimuli was, however, in the
range of HIV patients with normal cytokine output per cell (Fig. 6).
These data suggest that there is an overall immunodeficiency present in
these patients that may emerge as a result of, e.g., a burst of
antiviral immune response resulting in cytokine-mediated immune
suppression. In the five patients that displayed this response type
(which we have not seen in other states of immunodeficiencies or
diseases; see above), we found no correlation with viral load, CD4 cell
counts, or therapy. Studies of larger cohorts and longitudinal testing
will be required to understand more closely this phenomenon in
particular and T cell memory in general when taking into account the
complexity of the disease afforded by the genetic variability of the
host and the virus and of various environmental influences. The initial
observations reported in this work demonstrate the feasibility of
studying clonal sizes, cytokine signatures, and the cytokine output of
individual Ag-specific T cells, and provide early insights into these
issues. High resolution analysis in the low frequency range should
facilitate progress in defining the function of the immune system in
healthy individuals and patients. Moreover, it should be an invaluable
tool with which to study with unprecedented accuracy the effectiveness
of vaccination against HIV and other infectious organisms by directly
visualizing the specific T cell response induced.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Magdalena Tary-Lehmann, Institute of Pathology, School of Medicine, Case Western Reserve University, 10900 Euclid Avenue, BRB 928, Cleveland, OH 44106-4943. E-mail address:
3 Abbreviations used in this paper: LDA, limiting dilution analysis; DM, dust mite Ag; DTH, delayed-type hypersensitivity; ELISPOT, enzyme-linked immunospot; HAART, highly active antiretroviral therapy; PPD, purified protein derivative; TT, tetanus toxoid.
Received for publication September 23, 1999. Accepted for publication January 21, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and interleukin 5-producing cells as a predictive marker for renal allograft failure. Transplantation 66:219.[Medline]
and IL-2-producing cells in the CD8+ T cell subset. Eur. J. Immunol. 28:3630.[Medline]
secretion. AIDS Res. Hum. Retroviruses 10:115.[Medline]
is related to disease progression in HIV infection: Evidence from a cohort of 347 HIV-infected individuals. AIDS Res. Hum. Retroviruses 13:1039.[Medline]
This article has been cited by other articles:
|
|
 |
|
  B. D. Forrest, M. W. Pride, A. J. Dunning, M. R. Z. Capeding, T. Chotpitayasunondh, J. S. Tam, R. Rappaport, J. H. Eldridge, and W. C. Gruber Correlation of Cellular Immune Responses with Protection against Culture-Confirmed Influenza Virus in Young Children Clin. Vaccine Immunol., July 1, 2008; 15(7): 1042 - 1053. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Barbey, E. Pradervand, N. Barbier, and F. Spertini Ex Vivo Monitoring of Antigen-Specific CD4+ T Cells after Recall Immunization with Tetanus Toxoid Clin. Vaccine Immunol., September 1, 2007; 14(9): 1108 - 1116. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. An, B. Poon, R. H. T. Fang, K. Weijer, B. Blom, H. Spits, I. S. Y. Chen, and C. H. Uittenbogaart Use of a Novel Chimeric Mouse Model with a Functionally Active Human Immune System To Study Human Immunodeficiency Virus Type 1 Infection Clin. Vaccine Immunol., April 1, 2007; 14(4): 391 - 396. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J K Kalsi, J Grossman, J Kim, P Sieling, D W Gjertson, E F Reed, F M Ebling, M Linker-Israeli, and B H Hahn Peptides from antibodies to DNA elicit cytokine release from peripheral blood mononuclear cells of patients with systemic lupus erythematosus: relation of cytokine pattern to disease duration Lupus, July 1, 2004; 13(7): 490 - 500. [Abstract] [PDF]
|
|
|
|
 |
|
 C. M. Constantin, E. E. Bonney, J. D. Altman, and O. L. Strickland Major Histocompatibility Complex (MHC) Tetramer Technology: An Evaluation Biol Res Nurs, October 1, 2002; 4(2): 115 - 127. [Abstract] [PDF]
|
|
|
|
 |
|
 N. S. Butler, M. M. Monick, T. O. Yarovinsky, L. S. Powers, and G. W. Hunninghake Altered IL-4 mRNA Stability Correlates with Th1 and Th2 Bias and Susceptibility to Hypersensitivity Pneumonitis in Two Inbred Strains of Mice J. Immunol., October 1, 2002; 169(7): 3700 - 3709. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. McLay, E. Leonard, S. Petersen, D. Shapiro, N. S. Greenspan, and J. R. Schreiber {gamma}3 Gene-Disrupted Mice Selectively Deficient in the Dominant IgG Subclass Made to Bacterial Polysaccharides. II. Increased Susceptibility to Fatal Pneumococcal Sepsis Due to Absence of Anti-Polysaccharide IgG3 Is Corrected by Induction of Anti-Polysaccharide IgG1 J. Immunol., April 1, 2002; 168(7): 3437 - 3443. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. P. Fontenot, L. A. Maier, S. J. Canavera, T. B. Hendry-Hofer, M. Boguniewicz, E. A. Barker, L. S. Newman, and B. L. Kotzin Beryllium Skin Patch Testing to Analyze T Cell Stimulation and Granulomatous Inflammation in the Lung J. Immunol., April 1, 2002; 168(7): 3627 - 3634. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Lieberman, P. Shankar, N. Manjunath, and J. Andersson Dressed to kill? A review of why antiviral CD8 T lymphocytes fail to prevent progressive immunodeficiency in HIV-1 infection Blood, September 15, 2001; 98(6): 1667 - 1677. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Hesse, A. Y. Karulin, B. O. Boehm, P. V. Lehmann, and M. Tary-Lehmann A T Cell Clone's Avidity Is a Function of Its Activation State J. Immunol., August 1, 2001; 167(3): 1353 - 1361. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M. Pelfrey, R. A. Rudick, A. C. Cotleur, J.-C. Lee, M. Tary-Lehmann, and P. V. Lehmann Quantification of Self-Recognition in Multiple Sclerosis by Single-Cell Analysis of Cytokine Production J. Immunol., August 1, 2000; 165(3): 1641 - 1651. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, the
y-axis on the left) and PHA (
, the
y-axis on the right). On the right, the
data for the patients were subdivided according to the viral load, and
the number of patients in each group is indicated. The SD among
replicate wells was <12%. C, The viral load vs CD4
cell numbers in the HIV patients we tested is shown.
) was
performed in parallel; the IL-5-producing T cell line NBHAC25 was also
tested (
; that induced by TT, 
; TT-induced IL-5 spot formation
is represented by 
; PPD did not induce IL-5 production. These
Ag-induced responses were undetectable by intracytoplasmic cytokine
staining performed in parallel (data not shown).
) spot formation in the presence of PPD or DM
alone, or in the presence of both Ags simultaneously (PPD + DM), as
specified. The percentage of CD3-positive cells in each test cell
population is indicated. Cytokine production in the medium control (<5
spots/well) has been subtracted and SD for replicate wells are shown.
The data show that the cytokine-producing cells are CD3+,
that their numbers are directly proportional to the percentage of CD3
cells present in the culture, and that the production of IFN-
