| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Malaria Program and
Pathology Division, Naval Medical Research Center, Silver Spring, MD 20910
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The mechanism of action of GM-CSF as an immune enhancer has been studied in the transfected tumor model. Its function is thought to involve a paracrine (local) effect of the cytokine at the injection site that recruits a mixed cellular infiltrate including APCs, eosinophils, and T and B cells capable of recognizing tumor Ags at metastatic sites (1, 18, 19). However, it is not certain that the same mechanism is operating when GM-CSF is delivered by other means. A GM-CSF recombinant protein could exert its effects locally at the site of injection or in distant tissues due to circulating levels of GM-CSF in blood or lymph. When pGM-CSF is used to enhance DNA immunization, somatic cells at the site of injection could be transfected and make GM-CSF protein, the plasmid could leave the injection site and transfect distant cells, or APCs moving through the injection site could be transfected (20).
Understanding the mechanism of action of GM-CSF is important both to optimize its immune enhancing effects and for reasons of safety. Recombinant human GM-CSF has been used for several years to stimulate granulocyte production from the bone marrow of neutropenic patients. However, systemic toxicity (21) and flaring of autoimmune disease have been reported (22, 23). If pGM-CSF had effects at distant anatomical sites, toxicity and autoimmunity might limit the usefulness of GM-CSF as a vaccine enhancer. However, if GM-CSF acted locally at the site of vaccine injection, unwanted distant effects would be avoided. Accordingly, before proceeding to clinical trials in humans with pGM-CSF, we initiated studies to define its mechanism of action.
Using our P. yoelii DNA vaccine model, we now show that i.m. injection of GM-CSF plasmid leads to a highly localized influx of cells with APC phenotype, but without T cells or B cells. The immune-enhancing effect requires admixing of GM-CSF plasmid with Ag-encoding plasmids and cannot be replicated by injecting GM-CSF plasmid i.v. or at distant i.m. sites.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The plasmids used in this study have been previously described (17, 24). Briefly, the P. yoelii circumsporozoite protein (PyCSP)-encoding plasmid, PyCSP1012, denoted pPyCSP, was created by PCR amplification of the DNA sequence encoding PyCSP from the plasmid nkCMVintPyCSP.1 (24) and ligated into the plasmid VR1012 (25). The plasmid encoding murine GM-CSF was produced in the VR1019 plasmid, a derivative version of the VR1012 plasmid with the addition of a leader element from rat preproinsulin II (26). As a negative control, we used a VR1019 plasmid encoding the murine GM-CSF point-mutated at two bases, which has lost the ability to bind to the GM-CSF receptor and lacks bioactivity (17, 27). Plasmids for immunization were purified by double cesium banding and diluted in normal saline. Endotoxin levels were less than 0.6 EU/mg.
Animals and immunization procedure
All experiments were conducted according to published regulations (28). Female BALB/cByJ mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice received their first immunization at 4–6 wk of age. Unless otherwise specified in the text, mice were immunized twice at 6-wk intervals. Immediately before injection, plasmids were mixed and administered in a total volume of 50 µl in PBS into the tibialis anterior or gastrocnemius muscle of each leg, using a 29-gauge needle (Becton Dickinson, Franklin Lakes, NJ). Unless otherwise specified, each immunizing dose contains 50 µg PyCSP plasmid and 15 µg GM-CSF plasmid or equal amounts of their respective control plasmids, p1012 and p1019, at each site. In some studies, mice were immunized i.v. with 100 µg PyCSP plasmid and 30 µg GM-CSF plasmid in a total volume of 100 µl in PBS.
Synthetic peptides and recombinant proteins
Peptides corresponding to PyCSP57–70 aa (KIYNRNIVNRLLGD) or PyCSP280–288 aa (SYVPSAEQI) were used for in vitro T cell studies (29, 30). Peptides were kindly provided by Dr. G. P. Corradin, University of Lausanne, Epalinges, Switzerland. A recombinant PyCSP protein lacking the secretion and anchor signal, produced in yeast (kindly provided by Dr. Anthony Stowers, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD), was used as solid phase Ag for ELISA.
Lymphocyte proliferation assay
Spleen cells were cultured in complete medium in 96-well flat-bottom plates at a concentration of 2.5 x 105 cells/well at 37°C and 5% CO2. Complete medium was DMEM (Life Technologies, Gaithersburg, MD) supplemented with 10% FCS (Life Technologies), 200 U/ml penicillin, 200 µg/ml streptomycin, and 200 mM L-glutamine. After 5 days in culture with peptide PyCSP57–70 aa at a final concentration of 2 µM, cells were pulsed for 16 h with 1 µCi tritiated thymidine, and radioactivity incorporated into DNA was measured by scintillation counting.
IFN-
assay
Ag-specific IFN- secreting cells were evaluated by
enzyme-linked immunospot (ELISPOT) in a modification of a previously
published assay (31). Briefly, 96-well
nitrocellulose-backed microtiter plates (Millipore, Bedford, MA) were
coated overnight at 4°C with 10 µg/ml IFN- mAb(clone RMMG-1,
Biosource, Camarillo, CA) in PBS (pH 7.2) and then blocked with 5% BSA
in PBS. Two-fold serial dilutions of a single-cell suspension, starting
from 5 x 105 cells/well, were incubated
overnight with or without peptide at 37°C in a humidified 5%
CO2 incubator. The plates were then washed with
0.025% Tween 20 in PBS and overlaid with 1 µg/ml biotinylated
anti-IFN- Ab (clone R46A2, Lee Biomolecular, San Diego, CA) for
2 h at room temperature. After washing with 0.025% Tween 20 in
H2O, spots corresponding to the position of
cytokine-secreting cells were developed by addition of
avidin-conjugated alkaline phosphatase (PharMingen, San Diego, CA)
and 5-bromo-4-chloro-3-indolyl phosphate-nitroblue tetrazolium
(Kirkegaard and Perry, Gaithersburg, MD). The dilution of cells
producing 
50 spots/well was used to calculate the number of
cytokine-secreting cells per 1 x 106
cells.
Serum Abs
Abs reactive with the PyCSP yeast recombinant protein were assayed by ELISA as previously described (32).
Histology
For histology and immunohistochemistry, groups of 20 mice were injected with plasmids in both gastrocnemii, and the injected muscles were removed from two mice daily for 7 days and at day 14 after injection. Standard hematoxylin and eosin (H&E) staining was conducted to assess cell morphology and inflammatory infiltrate in the injected muscles. Immediately after removal, muscles were fixed for 24 h in 10% formalin, after which each sample was embedded in paraffin. Sections 6 µm thick at 10 evenly spaced levels were cut throughout each muscle and adhered onto Superfrost Plus slides (Fisher Scientific, Fair Lawn, NJ) which were stained with Mayer’s H&E (Sigma, St. Louis, MO) following manufacturer’s standard protocol. All histological assessment was conducted by two independent observers in a blinded manner. Grading of muscle infiltrates was done semiquantitatively using the following scale: 0 = no infiltrate; 1+ = one small cell cluster; 2+ = two small or moderate size cell clusters; 3+ = extensive, multifocal cell infiltration.
Immunohistochemistry
Muscles were snap frozen by overlaying with Histo-Prep tissue-embedding medium (Fisher Scientific) and immersing in liquid nitrogen-cooled isopentane. All samples were stored at -70°C until analysis. Serial 6-µm-thick frozen sections from each muscle were adhered to Superfrost Plus slides, fixed in ice-cold acetone at -20°C for 10 min, air-dried, and rinsed in distilled water to remove embedding medium. Staining was conducted following standard procedures (33). Briefly, sections were sequentially incubated for 20 min with 0.05% H2O2 in PBS and with 5% normal goat serum (Sigma) in PBS. Samples were then incubated for 2 h with primary Abs, CD11b (M1/70), Mac-3 (M3/84), CD11c (HL3), IAd/Ed (2G9), GR-1 (Ly-6G), B7-1 (16-10A1), CD4+ (RM4-5), CD8+ (53-6.7), and B220 (RA3-6B2) (all from PharMingen) in accordance with manufacturer’s instructions, followed by a 1-h incubation with 5 µg/ml biotinylated goat anti-rat or anti-hamster secondary Ab (PharMingen). After a 30-min incubation step with streptavidin-peroxidase (PharMingen), Ag-Ab reactions were developed using 3,3'-diaminobenzidine (Dako, Carpinteria, CA) as substrate. Slides were washed twice with PBS between each incubation step. All reagents were added in a volume of 50 µl, and the incubations were conducted at room temperature in a humid chamber. Grading of tissue staining was done semiquantitatively using the following scale: 0 = no cells stained, 1+ = fewer than 10% cells stained, 2+ = 10–50% stained, 3+ = 50–100% cells stained.
Analysis of CD11c+ DC in draining lymph nodes by flow cytometry
Expression of DC surface markers in the draining lymph nodes of mice injected with pGM-CSF or p1019 was quantitated by flow cytometry. Plasmids were injected bilaterally into the gastrocnemius muscles, and the popliteal lymph nodes were removed at days 4, 8, 12, and 20. Single-cell suspensions were prepared at 1 x 106 cells in 1% BSA in PBS and stained with 1 µg/ml of anti-CD11c (HL3)-CD11b (M1/70), -MHC II (2G9), and -B7.1 (16–10A1) (PharMingen). After 45 min incubation on ice, cells were washed twice in PBS containing 0.1% BSA and 0.02% NaN3. At least 100,000 cells were analyzed on a FACScan (Becton Dickinson, San Jose, CA) using CellQuest software (Becton Dickinson).
Statistical analysis
An unpaired, two-tailed Student t test was used for evaluating statistical significance between the immunization groups.
One-way ANOVA with Scheffé post hoc tests was used to evaluate differences among the mean immune responses in the groups of mice with CD11c+ indexes of 3, 2, 1, or 0.
ANOVA was performed in SPSS for Windows, version 8 (SPSS, Chicago, IL).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Torres et al. (34) reported that plasmid DNA
injections into mouse muscles resulted in Ag-specific immune responses
even when the muscles were surgically removed minutes after injection.
One possible explanation for those findings is that a small amount of
plasmid might have leaked into the blood stream after i.m. injection,
leading to a systemic immune response. To address this possibility, we
injected plasmid pGM-CSF into mice separately or together with pPyCSP
via the i.m. or i.v. routes and evaluated their immune responses.
Injections were at week 0 and 6 with a total of 100 µg pPyCSP and 30
µg pGM-CSF or equal amounts of corresponding control plasmids, p1012
and p1019, as described in Fig. 1
.
Experiments were performed on 3 separate occasions. Data from a
representative experiment are shown in Fig. 1.
|
in response to the defined CD8 T
cell epitope, PyCSP280–288 aa, at 3 and 8 wk
after the final immunization is presented in Fig. 1
A.
Intramuscular coinjection of pPyCSP + pGM-CSF (group 1) significantly
enhanced the response at both times compared with pPyCSP + p1019
control injections (group 2) (p < 0.05).
Injection of pPyCSP i.m. combined with an i.v. injection of either
pGM-CSF (group 6) or p1019 (group 7) resulted in a transient boost to
the response at 3 wk after injection which disappeared by week 8
(p < 0.05).
The T cell proliferation response to the PyCSP57–70
aa CD4 T cell epitope in the same experiment is shown in
Fig. 1B. Again, i.m. injection of the mixture of pPyCSP +
pGM-CSF (group 1) gave an enhanced stimulation index (SI) at 3 (SI
= 11) and 8 (SI = 10) wk after the last immunization
(p < 0.05). Interestingly, i.v. injection with
pPyCSP + pGM-CSF (group 3) gave a much stronger response at the early
time point (SI = 48), compared with i.v. pPyCSP + p1019 control
injection (group 4) (SI = 5) (p < 0.05),
but this response diminished substantially (SI = 3) by 8 wk. The
serum Ab responses to the recombinant PyCSP protein 3 wk after
the second immunization are shown in Fig. 1C. As noted
for T cell responses to the 280–288 aa epitope, i.m. coinjection of
pPyCSP + pGM-CSF elicited the highest titers with almost a 2-log
difference compared with the corresponding control injections (group 2)
(p < 0.05). No difference in Ab titers was
observed when an i.m. injection of pPyCSP was given combined with an iv
dose of either the pGM-CSF (group 6) or the p1019 control plasmid
(group 7).
In summary, these data indicate that i.v. pGM-CSF may result in
nonspecific immune enhancement due to the DNA backbone (Fig. 1, A and C) as well as in specific enhancement
elicited by the encoded GM-CSF (Fig. 1B). Yet, the specific
enhancement effects of i.v. pGM-CSF appear to be short lived as
compared with the prolonged enhancement observed with i.m. pGM-CSF.
Separating the site or time of i.m. injection with pPyCSP and pGM-CSF abrogates enhancement of immune responses
Next, experiments were designed to test whether colocalization of
pGM-CSF with pPyCSP in muscles was required for immune enhancement.
Specifically, we wished to compare the effects of a single i.m.
injection with mixed plasmids in one muscle with 1) injections of each
plasmid into separate limbs, 2) injections of each plasmid into
adjacent muscles within the same lymphatic drainage area, or 3)
consecutive injections of pPyCSP and pGM-CSF plasmids given 10 min
apart in the same muscle. Representative data from one of four
experiments are shown in Table I. All
mice received a constant total amount of DNA split between three
different muscles: the right tibialis anterior; the left tibialis
anterior; and the right gastrocnemius. Some groups received pGM-CSF and
others received p1019 as a negative control. In groups 1–4, each of
the three muscles was injected with a total volume of 50 µl. Groups 5
and 6 differed from the other groups in that they received two
injections of 25 µl given 10 min apart into the right tibialis
anterior. An attempt was made to place the second injection in the same
location as the first. Six weeks after injection, mice were bled for
serum Abs, and spleens were harvested for T cell studies. Ab assays
were performed on pooled serum, and T cell studies were performed on
individual animals.
|
shows that coinjection of pPyCSP and pGM-CSF gave a substantial
boost to the Ab response to PyCSP, as well as to the IFN- responses
to the CD4 (PyCSP57–70 aa) and CD8
(PyCSP280–88 aa) T cell epitopes (group 1
vs group 2). T cell proliferation to the 57- to 70-aa epitope was also
enhanced in group 1 (data not shown). Comparison of responses in groups
1–4 indicates that separating the pPyCSP and pGM-CSF injection sites
into muscles on separate limbs (groups 1 vs 3) or into separate muscles
on the same limb (group 1 vs group 4) completely eliminated the immune
enhancing effects of pGM-CSF. Statistical analysis of these data using
Student’s t test indicated that for each T cell epitope,
the responses from group 1 were significantly higher than those of
groups 2–6 (p < 0.02 for both CD4 and CD8 T
epitopes). Interestingly, although there is no significant difference
in responses between groups 3 and 4 (p > 0.1),
which received pGM-CSF separate from pPyCSP, both these groups had
lower T cell responses to the 57- to 70-aa epitope than group 2, which
received no pGM-CSF (p < 0. 02). Responses to
the 280- 288-aa epitope in groups 3 and 4 were also lower than in group
2 but not statistically different (p > 0.1).
Thus, it appears that the physical separation of the pGM-CSF and pPyCSP
injection sites decreased T cell and Ab responses to pPyCSP (groups 3
and 4 vs group 1). Interestingly, the data from groups 5 and 6 showed that when the pPyCSP and pGM-CSF are injected 10 min apart into the same muscle, there is only a limited enhancing effect (small increase in Ab and no change in T cell responses). This result was reproduced three of five times. In two experiments, the second injection of pGM-CSF did enhance the T cell and Ab responses to PyCSP (data not shown). Thus, even within the same muscle, admixing the two plasmids in the same syringe is the optimal way of boosting by GM-CSF.
Kinetics of cell infiltration into muscle after injection of pGM-CSF
To investigate the local effects of plasmid-expressed GM-CSF in
muscle, mice were injected once in both gastrocnemius muscles with
pGM-CSF + pPyCSP, with the GM-CSF plasmid in decreasing amounts (30, 3,
and 0.3 µg; in Fig. 2, groups 1, 2, and
3, respectively). Control mice were injected with pGM-CSF + p1012
plasmid (group 4), mutated pGM-CSF + pPyCSP (group 5), p1019 + pPyCSP
plasmid (group 6), p1019 + p1012 (group 7), or saline (group 8). The
gastrocnemius muscle was used in these studies because of the ease with
which it can be removed and handled. The injected muscles were removed
at different time points and examined by histology and
immunohistochemistry.
|
.
Data represent the average infiltration index for four muscles at each
time point, with SDs ranging between 0 and 0.6. Coinjection i.m. of
control plasmids p1019 + pPyCSP (group 6) or p1019 + p1012 (group 7)
led to small infiltrates which persisted for 7 days. A similarly
small infiltrate was seen after injection of pPyCSP plasmid mixed with
a plasmid encoding the murine GM-CSF gene with two point mutations that
lead to production of GM-CSF without bioactivity (group 5). However, a
significantly larger infiltrate was seen when plasmid encoding the
native GM-CSF was injected either alone (data not shown), with pPyCSP
(group 1) (p < 0.002) or with control p1012
(group 4) (p < 0.006). Doses of 30 or 3 µg of pGM-CSF +
pPyCSP produced infiltrates similar in duration and size (groups 1 and
2) which were significantly larger than 0.3 µg pGM-CSF (group 3)
(p < 0.02). The latter dose showed infiltrates
that were indistinguishable from pPyCSP alone (not shown) or control
plasmid injections (group 4, 5, or 6) (p > 0.1
in all cases). Interestingly, in most cases, the infiltration induced
by pGM-CSF was found at one area within the muscle but with several
foci of inflammation in the vicinity (Fig. 3A). This infiltrate reached
its largest size at days 3, 4, and 5; diminished by day 7; but was
still detected at day 14 after injection. Muscles excised 1 or 8 h
after injection with pGM-CSF showed little infiltration which was
seemingly limited to the needle path and was not different from that
seen with PBS or control plasmid (data not shown).
|
To identify the cells recruited by pGM-CSF to the site of
injection, muscle sections were analyzed by immunohistochemistry using
Abs with specificities for different cell markers. The results from
muscles analyzed at day 3 after plasmid injection are presented in
Table II. The predominant cell type found
in all infiltrates after injection with any of the DNA plasmids was a
population with a macrophage phenotype, staining strongly with Abs to
CD11b+ (Fig. 3B),
Mac-3+ and
IAd/Ed+. Cells with this
staining were observed throughout the entire 14 days of the study (data
not shown). Appreciable numbers of granulocytes, as detected by Abs
specific to GR-1 (Fig. 3C), were also present in all
infiltrated muscles, regardless of the plasmid used for injection
(Table II). H&E staining revealed that
these granulocytes were mostly neutrophils, with few eosinophils (not
shown). Perhaps the most striking finding was the presence of cells
expressing the DC marker CD11c in muscles injected with 30 or 3 µg of
pGM-CSF (Fig. 3D) but not in muscles injected with 0.3-µg
doses or lower of pGM-CSF or with any of the control plasmids (Table II). These CD11c+ cells were detected only on
days 3, 4, and 5 after pGM-CSF injection, appearing later and
disappearing earlier than the cells with CD11b+,
Mac-3+,
IAd/Ed+, or GR-1
phenotypes. Serial sections from muscles displaying
CD11c+ staining did not present any reactivity
with B7-1 or CD40-specific Abs (not shown), which are both markers of
APC activation and DC maturation. Neither was there staining with CD4-,
CD8-, or B220-specific Abs in any of the injected muscles (not shown),
indicating the absence of T or B lymphocytes in the infiltrates.
Immunostaining of normal spleen sections with anti-CD4, -CD8,
-B220, B7-1, or -CD40 Abs, however, displayed a typical reactivity (not
shown). No reactivity was observed with isotype control primary Abs in
any muscle (Fig. 3E). Thus, these data indicate that the
effect of pGM-CSF is both to increase the infiltrate size and duration
(as seen in H&E sections) and to participate in the recruitment of
CD11c+ cells.
|
Immature
CD11c+B7.1-CD40-
DC are recruited into muscles by GM-CSF but are detected for only a
short period after injection. Either these cells are dying or they must
be migrating out of the muscle tissue, possibly to become mature DC. To
investigate whether these DC leave the muscle, the draining popliteal
lymph nodes were analyzed by flow cytometry. Thus, increases in numbers
of CD11c/CD11b, CD11c/B7.1, or CD11c/MHC II double-positive mature DC
were assessed at different time points in the draining lymph nodes of
mice injected with pGM-CSF compared with control p1019. Significant
increases were found in the percentages of CD11c/CD11b
(p < 0.01) and CD11c/B7.1
(p < 0.06) but not those of CD11c/MHC II
(p > 0.1) (data not shown) double-positive
cells at days 4 (Fig. 4), 8, and 12 after
i.m. injection with pGM-CSF. No significant differences were found at
day 20 after injection (data not shown). Fig. 4 shows the FACS profiles
from representative animals at day 4 after injection with pGM-CSF or
p1019, where a 1.5- to 3.0-fold difference in CD11c/B7.1 or CD11c/CD11b
double-positive cells was consistently found between the two
immunization groups.
|
To determine the lowest dose of pGM-CSF that would boost immune responses and to assess whether the immune enhancement effect could be dissociated from the presence of the i.m. infiltrate, groups of mice were injected with decreasing amounts (60, 6, or 0.6 µg) of GM-CSF plasmid mixed with 50 µg pPyCSP. Injections were split into both gastrocnemius muscles, each muscle receiving 30, 3, or 0.3 µg of pGM-CSF together with 25 µg pPyCSP. Control groups of mice received injections with 60 µg control plasmids, and/or 50 µg pPyCSP alone, also split bilaterally. Three days after injection, two mice from each dose group were sacrificed, and their injected muscles (two from each mouse) were studied for cellular infiltrates using immunohistochemistry and H&E staining. Other mice were maintained for 6 wk, when their T cell and Ab responses to pPyCSP were measured.
The immune responses from mice injected with pPyCSP plus varying doses
of pGM-CSF are shown in Fig. 5. Data
shown are typical of three separate experiments using five mice per
group. Fig. 5A shows the numbers of spleen cells producing
IFN- in an ELISPOT assay after in vitro exposure to the
PyCSP280–88 aa peptide, CD8 epitope. Compared
with control injections with pPyCSP alone or admixed with p1019 control
plasmid, a dose of 0.6 µg of pGM-CSF or higher boosted the response
to the CD8 (p < 0.01) epitope. Fig. 5B shows the T cell proliferation responses to the
PyCSP57–70 aa CD4 epitope. As noted for
the IFN- results, higher amounts of pGM-CSF lead to more robust
proliferative responses. Stimulation indexes obtained from mice
injected with 60, 6, or 0.6 µg of pGM-CSF were significantly higher
than those evoked by the control injections with 60 µg p1019 + 50
µg pPyCSP (p < 0.01). An increase in Ab
responses occurred at doses of 60 or 6 µg pGM-CSF but not at lower
doses (Fig. 5C). In summary, there was an association,
albeit imperfect, between enhancement of immune responses and the
presence of CD11c+ cells in the injected muscles.
Infiltrates containing CD11c+ cells were detected
with doses of 3 µg pGM-CSF per muscle but not with doses of 0.3 µg
or lower. Likewise, enhancement of Ab responses was observed with
bilateral doses of 3 but not with 0.3 µg pGM-CSF. However, both CD4
and CD8 T cell responses to defined peptide epitopes were enhanced when
mice were given a bilateral dose of 0.3 µg of pGM-CSF, despite the
absence of CD11c+ in muscles injected with this
dose. A one-way ANOVA indicated, however, that the T cell responses of
mice presenting i.m. infiltrating CD11c+ cells
(injected with 30 or 3 µg pGM-CSF) were significantly different from
the T cell responses of those animals lacking these infiltrating cells
(0.3 µg pGM-CSF or control plasmid injections)
(p < 0.005).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our study is the first description of the immunohistology in muscle associated with pGM-CSF injection.4 We found large and multifocal conglomerates of cells that were initially MHC class II+Mac-3+CD11b+CD11c-B7.1-CD40-. During the peak of the infiltrate, a subset of cells expressing CD11c+ appeared. Numerous GR1+ cells that were mostly neutrophils with few eosinophils were also present within this infiltrate, but there were no T cells or B cells at any time point examined.
The infiltrates we identified in muscles injected with pGM-CSF plasmid are different from those observed in the skin surrounding irradiated tumor vaccines either transduced with the GM-CSF gene (18, 19, 35) or infected with viral vectors expressing GM-CSF (36, 37). Whereas plasmid encoded GM-CSF seems to have induced a relatively homogenous infiltrate of APCs and neutrophils in muscle, the tumor-expressed GM-CSF generated infiltrates in skin that included T cells, B cells, and eosinophils. The differences with regard to plasmid GM-CSF may be due to the complexity of the antitumor immune response at the site of injection, to the intrinsic cell composition of each tissue (muscle vs skin), or perhaps to the Th1 immunomodulatory effects of the DNA altering the local cytokine milieu (38, 39). At present, it is unclear whether the GM-CSF in our system is directly acting to recruit the infiltrating cells or is working through the induction of other lymphokines or chemokines (36).
Our results show that after pGM-CSF injection, the incoming cells are a mixture of macrophages, neutrophils, and immature DC. It has been well established that in vitro culture of bone marrow or CD34+ PBMC with GM-CSF and IL-4 leads to the outgrowth of immature DC with a CD11b+CD11c-/lowB7.1-/lowB7.2-/lowCD40- phenotype, which can mature into CD11b+CD11c+B7.1+B7.2+CD40+ DC after a variety of inflammatory stimuli (40, 41). It is possible that after i.m. pGM-CSF injection, in vivo expressed GM-CSF may have similar effects on circulating DC precursors recruited into the injection site, with DC maturation resulting from the low grade inflammatory response seen after any DNA injection. Immature DC are especially efficient at Ag uptake, whereas mature DC are more adapted for Ag presentation (40, 41). The appearance in muscle of CD11c+ cells lacking the activation markers B7.1 or CD40 on days 3–5 after pGM-CSF injection, followed by an increase of CD11c+CD11b+B7.1+ cells in lymph nodes on days 4–12, is consistent with egression of these immature DC from the muscle and migration to the draining lymph nodes, where they become mature DC (42).
Ongoing efforts to isolate and purify the cells from the i.m. pGM-CSF infiltrates as well as from the corresponding draining lymph nodes should shed more light on the functional attributes of these cells.
The pGM-CSF dose titration experiments showed that infiltrate size and appearance of CD11c+ cells are associated with enhancement of Ab responses to an admixed malaria DNA vaccine. Yet, T cell responses were boosted at low doses of pGM-CSF where no CD11c+ cells are detected. It is possible, however, that the detection of the relatively scarce CD11c+ cells may have been missed by our sectioning procedure in infiltrates of small size and reduced depth.
The plasmid backbone of the DNA vaccines used in this study is of bacterial origin and contains some immunostimulatory nonmethylated CpG sequences (38, 39). It is unlikely, however, that the inflammatory reaction we observe with pGM-CSF as well as the adjuvant effect of this plasmid is due to its CpG content. Injection of several different control plasmids not only evoked smaller infiltrates lacking CD11c+ DC but also elicited statistically lower immune responses than did pGM-CSF. The most appropriate control for immunostimulatory effects of DNA is the plasmid encoding murine GM-CSF with two point mutations (not at CpG sites), producing inactive GM-CSF protein (17, 27). Injections with this plasmid also gave small infiltrates that lacked DC and did not enhance the immune responses. Thus, although we acknowledge that in some circumstances CpGs can have immunostimulatory activity, we believe that this cannot explain the effects of pGM-CSF. Our findings suggest that events occurring in the muscle after pGM-CSF injection may be contributing to the immune enhancement. Direct transfection of the infiltrating APC with the coinjected pPyCSP plasmid or improved cross-priming by APCs taking up PyCSP immunogens from myocytes would be predicted to contribute to a better immune response (43, 44, 45). However, this premise has two important caveats. Firstly, the enhancement of immune responses by pGM-CSF may be multifactorial and some of these factors may be independent of DC (10). Secondly, the importance of Ag production by the plasmid-transfected muscle is still controversial. It has been reported that Ag expression by nonlymphoid somatic cells may be responsible for the majority of antigenicity of an OVA-encoded DNA vaccine (45). However, surgical ablation of injected muscles minutes after receiving plasmid was reported to have no effect on immunogenicity, implying that the muscle itself may not be the site of Ag expression during DNA vaccination (34). Leakage of plasmid from the muscle into the blood or the lymph, or direct transfection of an APC in transit through the injected muscle, might represent the true mode of immunization (46). Notably, Klinman et al. (47) showed that migratory cells at the site of vaccination were indispensable for induction of primary immunity as well as immunological memory. Our results indicate that leakage of pGM-CSF into the bloodstream does not contribute significantly to immune enhancement, because an i.v. injection of pGM-CSF combined with pPyCSP administered either i.v. or i.m. evoked at best only transient immune responses.
Separation of injections with pGM-CSF and pPyCSP, either into two adjacent muscles in the same leg or into muscles in opposite legs gave interesting results. We reasoned that if plasmid can leak into the lymphatic circulation, separating the injection of each plasmid into two different muscles drained by the same lymph nodes could result in a response equivalent to that evoked by injection of both plasmids into the same muscle. However, not only was there no boosting of responses when pGM-CSF and pPyCSP were given into separate muscles, as previously noted for rabies-encoding plasmids (12), but there was a decrease in T cell responses to pPyCSP. This is not unprecedented, because weaker T cell responses from PBMC have been reported to occur when recombinant GM-CSF protein was injected s.c. preceding peptide immunization in humans (48). It is possible that the production of GM-CSF after plasmid injection into one muscle was able to draw away circulating APCs from the location where pPyCSP was injected, thereby decreasing the immune responses.
As a second part of this experiment, the i.m. injections with pPyCSP and pGM-CSF were given 10 min apart in the same muscle. We expected to reproduce the boosting effects of a single injection of admixed plasmids. Instead, we found that the second injection with pGM-CSF had very little effect. This lack of response could be due to several factors. Perhaps the first injection transiently changes the character of the muscle so that the second injection given within minutes of the first is not as effective. The results of an experiment in progress support this idea. We injected pGM-CSF i.m. 3 days before injecting pPyCSP into the same muscle. Specific Abs were boosted when pGM-CSF and pPyCSP injections were separated by several days or when the plasmids were coinjected. This 3-day time lag is similar to that found by Disis et al. (6) using recombinant GM-CSF protein to boost protein Ags. However, Disis found very poor immunogenicity when she mixed recombinant GM-CSF protein together with protein Ag, in sharp contrast to our results with plasmid immunization. We think the kinetics of protein production from plasmid-transfected cells as well as the multifocal and highly localized character of the infiltrates seen after pGM-CSF injection could account for this difference. There is still much to be learned about the mechanisms and timing of boosting by GM-CSF.
The exclusively local activity of i.m. pGM-CSF is encouraging for its safe use as a vaccine adjuvant. It would seem extremely unlikely that pGM-CSF injected into one site would enhance an ongoing pathological T cell response in a different anatomic location. Thus, pGM-CSF plasmid merits serious consideration as a DNA vaccine enhancer in humans.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Walter R. Weiss, Malaria Program, Naval Medical Research Center, 503 Robert Grant Avenue, Forest Glen Annex, Silver Spring, MD 20910.
3 Abbreviations used in this paper: DC, dendritic cells; ELISPOT, enzyme-linked immunospot; PyCSP, Plasmodium yoelii circumsporozoite protein; H&E, hematoxylin and eosin; SI, stimulation index.
4 At the time our manuscript was submitted, Kusakabe et al. published a study (49 ) in which the presence of S-100+ DC were detected in muscle by immunohistology.
Received for publication March 14, 2000. Accepted for publication July 12, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and accumulation of dendritic cells at vaccination sites in vivo. Cancer Immunol. Immunother. 48:123.[Medline]
This article has been cited by other articles:
|
|
 |
|
  L. Qin, J. R. Greenland, C. Moriya, M. J. Cayabyab, and N. L. Letvin Effects of Type I Interferons on the Adjuvant Properties of Plasmid Granulocyte-Macrophage Colony-Stimulating Factor In Vivo J. Virol., October 1, 2007; 81(19): 10606 - 10613. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Skountzou, F.-S. Quan, S. Gangadhara, L. Ye, A. Vzorov, P. Selvaraj, J. Jacob, R. W. Compans, and S.-M. Kang Incorporation of Glycosylphosphatidylinositol-Anchored Granulocyte- Macrophage Colony-Stimulating Factor or CD40 Ligand Enhances Immunogenicity of Chimeric Simian Immunodeficiency Virus-Like Particles J. Virol., February 1, 2007; 81(3): 1083 - 1094. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Chavan, K. A. Marfatia, I. C. An, D. A. Garber, and M. B. Feinberg Expression of CCL20 and Granulocyte-Macrophage Colony-Stimulating Factor, but Not Flt3-L, from Modified Vaccinia Virus Ankara Enhances Antiviral Cellular and Humoral Immune Responses. J. Virol., August 1, 2006; 80(15): 7676 - 7687. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. F. McCutchan, K. C. Grim, J. Li, W. Weiss, D. Rathore, M. Sullivan, T. K. Graczyk, S. Kumar, and M. R. Cranfield Measuring the Effects of an Ever-Changing Environment on Malaria Control Infect. Immun., April 1, 2004; 72(4): 2248 - 2253. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. B. Ismael, D. Sekkai, C. Collin, D. Bout, and M.-N. Mevelec The MIC3 Gene of Toxoplasma gondii Is a Novel Potent Vaccine Candidate against Toxoplasmosis Infect. Immun., November 1, 2003; 71(11): 6222 - 6228. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. A. Siddiqui, T. Phillips, H. Charest, R. B. Podesta, M. L. Quinlin, J. R. Pinkston, J. D. Lloyd, M. Paz, R. M. Villalovos, and J. Pompa Induction of Protective Immunity against Schistosoma mansoni via DNA Priming and Boosting with the Large Subunit of Calpain (Sm-p80): Adjuvant Effects of Granulocyte-Macrophage Colony-Stimulating Factor and Interleukin-4 Infect. Immun., July 1, 2003; 71(7): 3844 - 3851. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Pullarkat, P. P. Lee, R. Scotland, V. Rubio, S. Groshen, C. Gee, R. Lau, J. Snively, S. Sian, S. L. Woulfe, et al. A Phase I Trial of SD-9427 (Progenipoietin) with a Multipeptide Vaccine for Resected Metastatic Melanoma Clin. Cancer Res., April 1, 2003; 9(4): 1301 - 1312. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Gri, E. Gallo, E. Di Carlo, P. Musiani, and M. P. Colombo OX40 Ligand-Transduced Tumor Cell Vaccine Synergizes with GM-CSF and Requires CD40-Apc Signaling to Boost the Host T Cell Antitumor Response J. Immunol., January 1, 2003; 170(1): 99 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Mwangi, W. C. Brown, H |
