granulocyte-macrophage colony-stimulating factor: an effective … · 2012. 3. 8. ·...

Granulocyte-Macrophage Colony-Stimulating Factor: An Effective Adjuvant for Protein and Peptide-Based Vaccines

By Mary L. Disis, Helga Bernhard, Faith M. Shiota, Susan L. Hand, Julie R. Gralow, Eric S. Huseby, Steven Gillis, and Martin A. Cheever

The current studies evaluate granulocyte-macrophage colony-stimulating factor (GM-CSF) as a vaccine adjuvant. An important issue for developing vaccine therapy for human malignancy is identifying adjuvants that can elicit T-cell responses t o proteins and peptides derived from "self" tumor antigens. GM-CSF, in vitro, stimulates the growth of antigen-presenting cells such as dendritic cells and macrophages. Initial experiments examined whether GM-CSF injected into the skin of rats could affect the number or character of antigen presenting cells, measured as class II major histocompatability complex expressing cells, in lymph nodes draining the injection site. Intradermal (id) inoculation of GM-CSF every 24 hours for a total of five inoculations resulted in an increase of class II+ fluorescing cells that peaked at the fourth inoculation. Subcutaneous (sq) inoculation resulted in an increase of class II+ fluorescing cells that peaked following the second inoculation, then decreased over time. Using this schema for "conditioning" the inoculation site, GM-CSF was administered id or sq for five injec-

N APPROACH to the treatment of human malignancy A is the use of cancer vaccines to stimulate patient immunity . Recent identification of the proteins and peptides that constitute tumor antigens',' has rekindled new interest in the generation of such vaccines. Given that T cells recognize peptide fragments of proteins presented by major histocom- patibility complex (MHC) molecules, much interest has fo- cused on formulating vaccines using antigenic peptides as the vaccinating antigens.' The use of synthetic peptides as the immunizing antigens in cancer vaccines offers many advantages: relatively easy construction and production, chemical stability, and a lack of infectious or oncogenic p~tent ia l .~ Perhaps the most important advantage is the theoretical ability to manipulate the immune response through the use of defined immunogenic epitopes. Protective immunity generated by vaccination depends on the ability of the vaccine to elicit the appropriate immune response to eradicate the antigen bearing cell or elicit protective immunity against the antigen. The use of peptide vaccines will allow the expansion of T-cell populations specific for defined immuno-

From the Department of Medicine, Division of Oncology, Univer-

Submitted December 12, 1995; accepted February 27, 1996. Supported by National Institutes of Health (NIH) Grants No.

SROICA 54561-05, 5ROICA 57851-04, and 5R37CA 30558-14 (for F.M.S, S.L.H, E.S.H, and M.A.C.) and NIH Grant No. 5T32 CA09.515-IO (for J.R.G.). HB is supported by The Deutsche Forschungsgemeinschaft (Bel 579/1 I ) .

Address reprint requests to Mary L. Disis, MD, University of Washington, Division of Oncology, Box 356527, Room BB1321, Seattle, WA 98195-6227.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with I8 U.S.C. section 1734 solely to indicate this fact.

sity of Washington, and the Corixa Corporation, Seattle, WA.

0 1996 by The American Society of Hematology. 0006-4971/96/8801-0032$3.0O/O

tions and a foreign antigen, tetanus toxoid (tt), was given at the beginning or the end of the immunization cycle. Id immunization was more effective than sq at eliciting tt specific immunity. In addition, GM-CSF id, administered as a single dose with antigen, compared favorably with complete Freund's adjuvant (CFA) and alum in eliciting tt specific antibody and cellular immunity. We have shown that immunity to rat neu (c-erbB-2) protein, an oncogenic self protein, can be generated in rats by immunization with peptides derived from the normal rat neu sequence plus CFA. The current study demonstrates that rat neu peptides inoculated with GM-CSF could elicit a strong delayed type hypersensitivity reaction (DTH) response, whereas peptides alone were non- immunogenic. GM-CSF was as effective as CFA in generating rat neu specific DTH responses after immunization with a neu peptide based vaccine. Soluble GM-CSF is a potent adjuvant for the generation of immune responses t o foreign proteins as well as peptides derived from a self tumor antigen. 0 1996 by The American Society of Hematology.

genic epitopes, thus allowing a more specific and, perhaps, a more effective immune response. Unfortunately, peptides are weak immunogens.

A potential problem in immunizing humans to weakly immunogenic proteins, such as tumor antigens, or peptides, is the lack of effective and potent adjuvants that will allow the augmentation of antigen specific immune responses. Pep- tide immunizations have elicited specific immune responses when injected with powerful adjuvants in animals, but many of these adjuvants are quite toxic for repeated use in hu-

Different adjuvants may function in many ways: affecting the character and number of antigen presenting cells in the inoculation environment, acting as a depot to prolong antigedantigen presenting cells (APC) exposure, or affecting the pathway by which proteins are processed.' The use of cytokines as vaccine adjuvants is appealing because their effect may allow for a choice of which arm of the immune response is enhanced, they may act as chemoattrac- tants for immune cells, and may further augment the vaccine's protective effects.

The studies outlined here evaluate the use of granulocyte- macrophage colony-stimulating factor (GM-CSF) as an adjuvant. GM-CSF has been known to play a role in the augmentation of the immune response for nearly a decade. Morrissey et a1 demonstrated GM-CSF augmented the primary in vitro immune response to sheep red blood cells by murine spleen cells.' Splenic adherent cells, which were incubated overnight with GM-CSF and antigen, were more efficient APC than those cells incubated with antigen alone. GM-CSF was found to increase the level of interleukin-1 secretion and Id antigen secretion. GM-CSF has been shown, in vitro, to stimulate the growth of APC such as dendritic cells and macrophages. GM-CSF has also been shown to increase the immunogenicity of tumors in animal model^.^ Transfecting the cytokine gene into tumor cells, and injecting the transfected cells subcutaneously, results in the production of GM-CSF by the tumor cell itself and enhances its immunoge-

202 Blood, VOI 88, NO 1 (July l) , 1996: pp 202-210

GM-CSF AS A VACCINE ADJUVANT 203

nicity. The increased immune response is presumed to be mediated by the role GM-CSF plays in the maturation and function of antigen presenting cells such as dendritic cells and macrophages.” Further studies, using GM-CSF encapsu- lated in microspheres mixed with irradiated tumor cells and injected subcutaneously also demonstrated increased tumor immunogenicity.” These encouraging studies with whole tumor cell preparations led us to question whether GM-CSF could be useful in augmenting specific antigen responses to soluble protein and peptides. We present an immunization strategy using GM-CSF as an adjuvant to increase the immunogenicity of antigen in both a foreign antigen and tumor antigen system.

MATERIALS AND METHODS

Animals. Rats used in this study were Fischer strain 344 (CDF (F-344)KrlBR) (Charles River Laboratories, Portage, MI). Animals were maintained at the University of Washington Animal Facilities under specific pathogen free conditions and routinely used for experimental studies between 3 and 4 months of age.

Experiments presented here evaluate GM- CSF as an adjuvant in two different antigen systems in rats: a foreign antigen, tetanus toxoid (tt); and peptides derived from a self antigen, rat neu protein. Preservative free tetanus toxoid (Lederle, American Cyanamid Corporation, Pearl River, NY) was diluted in sterile phos- phate-buffered saline (PBS) and stored in aliquots at -70°C. Nine peptides derived from the amino acid sequence of rat neu were constructed, 5 from the ECD of the protein and 4 from the ICD of the protein. ECD peptides were p45-59, p98-112, p323-337, p332- 349, and p433-447. ICD peptides were p781-795, p788-802, p932- 948, and pl171-1185. These peptides, 15 to 20 amino acids in length, were shown in previous studies to elicit T-cell immune responses in Fischer rats when injected subcutaneously (sq) with a classical adjuvant, complete Freund’s adjuvant (CFA).’’ The peptides were synthesized and purified by H. Zebroski (University of Washington, Seattle, WA), then dissolved in PBS, pH 7.4, to give 2 mg/mL stock solutions. Before aliquoting, peptides were sterile filtered, then stored at -70°C. Two single peptides that could elicit peptide-specific T cells, p45-49 and p788-802 were used in preliminary studies. Multiple peptide combinations were used as immunizing antigens in later studies. Purified recombinant murine GM-CSF (kindly sup- plied by Immunex Corporation, Seattle, WA) was used as an adjuvant at a dose of 5 or 10 pg per inoculation.

Preliminary experiments defined the need to pre- pare the inoculation site with GM-CSF, theoretically to stimulate Langerhan’s cell growth and differentiation. Initial injections with either GM-CSF or PBS without an antigen to access immune cell trafficking were performed over five cytokine inoculations, each 24 hours apart. Rats’ backs were shaved and a template drawn to ensure inoculation would be at the same site each day. In the groups that received GM-CSF, the cytokine was administered at a dose of 10 pg per injection in a total volume of 50 pL. Animals in the control group received injections of 50 pL of sterile PBS. Groups received adjuvant intradermally (id) or sq. Each group consisted of three experimental animals (GM-CSF) and two control animals (PBS). Groups of animals were killed 24 hours after 2, 3,4, and 5 inoculations. Lymph nodes draining the injection site were analyzed for infiltrating cells. Animals showed no adverse effects to repeated injection at the same site.

Subsequent immunization protocols with antigen were based on the skin preparation strategy defined above. Animals were immunized with GM-CSF or PBS at the same site for five inoculations. In the tetanus experiments, tt was mixed with the adjuvants at a

Antigedadjuvant.

Immunizations.

concentration of 3 Lf units per animal and administered with the first two injections or the last two injections. Animals were assessed for tt specific delayed type hypersensitivity reaction (DTH) responses, as an in vivo measure of cell mediated immunity, 20 days after the fifth injection. Serologic responses were evaluated 24 days after the fifth injection. In the experiments evaluating peptide specific responses, 250 pg of the rat neu peptides were administered with the first two of five cytokine or PBS injections. As with the tt immunizations, the peptides were mixed with the GM-CSF or PBS at the time of injection and inoculated together. Peptide specific DTH responses were assessed 20 days after the fifth injection. Immu- nizations were administered either id or sq.

The final series of experiments examined the necessity of multiple dose GM-CSF administration by evaluating immunization with a single dose of GM-CSF mixed with antigen, similar to the use of a “classical” adjuvant. These experiments also compared GM-CSF, as an adjuvant, with two other standard adjuvants, CFA (Sigma Immunochemicals, St Louis, MO) and Alum (Pierce, Rockford, IL). Tetanus toxoid was mixed with the adjuvants (5 pg GM-CSF, 1:l with CFA, and 1:l with Alum) and administered as a one time immunization. CFA and Alum with tt were administered sq. GM- CSF with tt was administered id or sq. Animals were assessed for DTH response and antibody immunity as described. Studies of a multiple peptide vaccine for the generation of rat neu specific immunity” demonstrated that T-cell immunity could be generated with three in vivo immunizations with immunogenic rat neu peptides in CFA. Experiments described here duplicate these studies comparing CFA with GM-CSF as an adjuvant. Animals were inoculated with multiple peptides vaccines (mix of 5 rat neu ECD peptides or a mix of 4 rat neu ICD peptides at 100 &peptide) in CFA (sq) or 5 pg GM-CSF (id) once every 20 days for three immunizations. Twenty days after the third immunization, DTH responses were measured as described.

Cells derived from the draining lymph nodes of rats injected id or sq with GM-CSF or PBS without any antigen were stained in suspension with MRC-OX-6 (Rat Ia; Serotec Ltd, Oxford, UK). Briefly, cells were harvested, washed, and resus- pended in PBS-I% bovine serum albumin (BSA) and seeded into microtiter plates at a find concentration of 2 to 3 x 10s cells per well. Cells were incubated on ice with the unconjugated monoclonal antibody at saturating concentrations for 45 minutes, washed twice with 150 pL PBS-I% BSA and stained with FITC conjugated F(ab’)’ rabbit antimouse IgG for an additional 30 minutes as the second step antibody. The second step FITC conjugate has been shown to have reduced cross reaction with rat (Serotec). Finally, all cells were treated by washing twice with PBS-1% BSA (Sigma Chemical Co, St Louis, MO) and fixed with 1 % paraformaldehyde. Negative controls consisted of cells labeled with the second step antibody alone. Fluo- rescence analyses were performed by flow cytometry. Two micron fluoresbrite calibration grade microspheres (Polyscience, War- rington, PA) are run on a daily basis to standardize the instrument.

96-well Im- mulon 4 plates (Baxter SP; Dynatech Laboratories, Redmond, WA) were incubated overnight at 4°C with tt diluted 1:1,000 in carbonate buffer (equimolar concentrations of Na2C03 and NaHCO, pH 9.6) altemating rows with buffer alone. After incubation, all wells were blocked with PBS-1% BSA, 100 pL per well for 1 hour at room temperature. The plate was washed with PBS-0.5% Tween and experimental sera was added at the following dilutions: 1:25 to 1:3,200 or 6,400. The sera was diluted in PBS-I% BSA-I% FBS-25 pg/mL mouse IgG-0.01% NaN3 and then serially into PBS-I% BSA. Fifty microliters of diluted sera was added per well and incubated for 1 hour at room temperature. Each experimental sera was added to a well with tt and a well without tt. Sheep antirat Ig F(ab’), horseradish peroxidase (HRP) was added to the wells at a 1:5,000 dilution in

Flow cytometric analysis.

ELISA for rat tetanus spec$c antibody responses.

204

20 -

0

DISK ET AL

PBS SQ

I I I I

PBS-I% BSA and incubated for 45 minutes at room temperature (Amersham Co, Arlington Heights, IL). Antibody class assays were performed similarly with rabbit antirat IgG F(ab’), and sheep antirat IgM HRP F(ab’), antibodies as the second step antibody at a concentration of 1:5,000 (Serotec). Following the final wash, TMB (Kirke- gaard and Perry Laboratories, Gaithersburg, MD) developing reagent was added. Tetanus toxoid immune control rat sera was derived by immunizing rats with tt 3 LFU and CFMFA for a priming immunization and two boosts each 20 days apart. This positive control rat sera was analyzed on each plate. When the optical density (OD) at 650 nm reached 0.3 at the 1: 100 sera dilution of the “hyper- immunized” sera, the plate was developed. The mean and standard deviation of all controls in each experimental group show the in- terassay variability. Color reaction was read at an OD of 450 nm. The OD of each serum dilution was calculated as the OD of the tt coated wells minus the OD of the PBS-1% BSA coated wells. Sera derived from two nonimmunized animals was run on each plate as an assessment of native tt response.

Antigen specificity was confirmed by analyzing experimental sera for antibody responses to ova albumin. In these analyses, plates were incubated overnight at 4°C with purified ova albumin protein at 10 pg/mL concentration in carbonate buffer alternating with rows of buffer alone. Antibody evaluation then proceeded as is described for tetanus.

Change in ear thickness on exposure to an antigen is a measurement of DTH response in rodents.” Twenty days after the final inoculation baseline ear thickness was measured in each animal using a dual thickness gauge (Mitutoyo Corporation, Japan). Immediately following the baseline measurement, the left ear was treated epicutaneously with a carrier solvent consisting of a 1:l mix of acetone and dibutyl pthalate. Ten microliters of the carrier solvent was applied to the front of the ear and 10 pL was applied to the back of the ear. The right ear of each animal was treated with the carrier solvent and antigen with 10 pL of the carrier diluted antigen mix applied to the front of the ear and 10 pL applied to the back of the ear. Animals in the tt immunization series were tested with 0.1 Lf U/mL of tt, animals immunized with the single rat neu peptides were tested with 125 pg/mL of their immunizing peptide, either p45 or p788, and animals immunized with multiple rat neu peptides were tested with 1 pg/mL of purified rat neu protein. DTH response as a measure of ear thickness was measured at 48 hours and calculated as the difference in the thickness of the experimental ear compared with control.

DTH responses.

RESULTS

Subcutaneous or intradermal injection of GM-CSF for five inoculations increases the percent of class 11 MHC+ cells in regional Lymph nodes. Initial experiments examined whether local application of GM-CSF into a site where antigen would be introduced would be effective in the recruitment of cells with the potential ability to generate an immune response. Sq and id administration was studied. Class I1 MHC+ cells were used as a marker for the presence of APC. GM-CSF is known to upregulate class I1 surface expression on APC such as dendritic cells (DC) and macrophage^.'^.'^ Cytokine was administered by both routes every 24 hours for a total of five inoculations to assess the length of time needed to recruit APC or affect the ability of APC to process antigen. Daily administration of GM-CSF increased the percent positive class I1 MHC expressing cells in regional lymph nodes. The results with id and sq administration dif- fered (Fig 1). GM-CSF id resulted in a marked increase in fluorescence of class 11 expressing cells in regional lymph

A

u

8 Lj t5 E

B

8 B p E E

Class 11: Lymph Node

20 I

PBS ID o l I I 1 I

1 2 3 4 5

Number of Injections

100 Class II: Lymph Node

1

Number of Injections

Fig 1. Subcutaneous or intradermal injection of GM-CSF for five inoculations increases the percent of class II MHC+ cells in regional lymph nodes. Rata were inoculated with 10 pg of GM-CSF id or sq each day for five inoculations, each 24 hours apart. A control group was injected with PBS. Draining lymph nodes were procured 24 hours after the 2, 3, 4, and 5 injections (3 animah/experimental group, 2 animalslcontrol group). Cells were evaluated for class II MHC expression by flow cytometry. (A) Draining lymph nodes from id GM-CSF and PBS injected animals. The data represents the mean percent fluorescence and standard deviation of determinations on three experimental or two control animals. Three thousand to 5,000 lymph node cells were counted at each evaluation. (B) The same analysis performed on lymph nodes of rats who received GM-CSF or PES sq.

nodes later in the injection cycle, after four inoculations (A). This is shown for a representative animal from each group in Fig 2. GM-CSF sq induced an increase in class TI expressing cells in draining lymph nodes early, within two injections, but after four injections was no different from control (B). Figure 3, FACS profiles from representative animals, demonstrated GM-CSF sq resulted in an early increase in class I1 expressing cells, which resolved over time. Animals immunized with PBS alone were used as a control. The total number of harvested cells from the draining lymph nodes from GM-CSF or PBS inoculated animals was not signifi- cantly different.

Intradermal injection of GM-CSF for five inoculations is


0: a 0: a 0: n

GM-CSF 3 D n

Y

Interdermal e t f Fig 2. Interdermal injection

of GM-CSF results in an increase

four injections. Rats were inoculated with 10 p g of GM-CSF id each day for five inoculations, each 24 hours apart. A control

1.00 J/FITC L O

1009 . I 3lFITC L O

1.00 . I 1009 .1 3,FIlC L O J f l I l C LO in class II expressing cells after . I

0: a a: a 3090,3890 .

group was injected with PES. pBs s 9 s Draining lymph nodes were ana- rnterdemal f f L

0

lyzed 24 hours after 2.3, 4, and 5 injections for class II MHC expression by flow cytometry. FACS analysis from a representative animal in each group is

.(

1 m m m . I I900 J R I T C LO 3,FITC LO

i8.e .1 J R I T C LO

imam . I J / r I l C LO

. I

shown here. 2 injecuonr 3 injeetfons 4 InJeCtlons 5 Injections

an effective adjuvant for eliciting antibody responses to tetanus toxoid when the antigen is administered with the last two injections of cytokine. GM-CSF administration had an effect on class I1 expressing cells in regional lymph nodes. The augmentation of class I1 expressing cells was more marked in the id immunized animals and occurred later in the inoculation cycle. To determine whether this effect translated into an increased ability to induce an immune response, rats were inoculated with GM-CSF sq versus id. Tetanus toxoid was injected concurrently with the last two GM-CSF inoculations. All animals who received GM-CSF id with tt developed significant antibody responses compared with PBS controls (Fig 4). Anti-tt antibody titers of animals receiving sq immunization were lower (Fig 4). No antibodies were detected with the negative control antigen, ova albumin.

Intradermal injection of GM-CSF for jive inoculations is an effective adjuvant for eliciting antibody responses to tt when the antigen is administered with thejrst two injections of cytokine. To determine whether adding antigen only after the site has been “conditioned” with GM-CSF is necessary, or whether co-administration of antigen and cytokine

GM-CSF Subcutaneous

PBS Subcutaneous

8

r . imma

J/FITC L O . I

2 Injections 3 injections

01 a

early in the cycle would suffice, rats were inoculated with GM-CSF id versus sq, with tt injected concurrently with the first two of five adjuvant inoculations. All six animals in the id GM-CSF adjuvant group developed a marked tt specific antibody response (Fig 5) . None of the animals in the id PBS group developed vigorous antibody responses to tt. Ani- mals who were immunized by the sq route developed anti- tt antibodies, but of a lower titer than the id immunized animals (Fig 5) . No antibodies were detected with the negative control antigen, ova albumin. Prolonged “conditioning” of the vaccination site with GM-CSF before the introduction of antigen did not appear to be critical in the generation of an immune response. In both regimens, antigen administered late (Fig 4) or early (Fig 5) , the id route of immunization appeared to be the most effective in generating significant anti-tt antibody responses. The remainder of the experiments concentrated on GM-CSF and antigen administered id as the optimal route of vaccination.

Intradennal GM-CSF, as an adjuvant, elicits IgG antibody responses to tt. Class of antibody elicited was analyzed for animals immunized with GM-CSF or PBS id for five injec-

4 injeaiorrr 5 injealons

Fig 3. Subcutaneous injection of GM-CSF results in an increase in class II expressing cells after two injections. Rats were inoculated with 10 p g of GM- CSF sq each day for five inoculations, each 24 hours apart. A control group was injected with PES. Draining lymph nodes were analyzed 24 hours after 2, 3, 4, and 5 injections for class II MHC expression by flow cytometry. FACS analysis from a representative animal in each group is shown here.

206 DlSlS ET AL

-0.2-1 , I I I I I I I

Antibody Titer

Fig 4. Intradermal injection of GM-CSF for five inoculations is en effective adjuvant for eliciting antibody responses to tt when the antigen is administered with the last two injections of cytokine. Rats were injected wRh OM-CSF or PBS id or sq for f i e inoculations. Tetanus toxoid was added to the cytokine or PBS inoculation at a concentration of 3 Lf units during the last two inoculations in all groups. Three animals received GM-CSF as an adjuvant and two animals received PBS as an adjuvant along with the tt in both the id or sq group. Three control animals that were not immunized ere shown as an example of a native rat response to tetanus. Twenty-four days after the fifth inoculation, sera was obtained from each animal. Teta- nus toxoid responses were determined by ELISA. Results are depicted as the mean and standard deviation of the antibody response of each experimental group at each sera dilution. Positive control sera from an animal with high tt specific antibody was run on each plate and is expressed as the mean and standard deviation of all experimental plates in this analysis.

tions with tt administered with the first two inoculations. Data from analysis of three animals is shown as an example (Fig 6). Animals who received GM-CSF as the immunizing adjuvant developed IgG tt specific antibodies. PBS control animals did not develop significant tt specific antibody responses.

GM-CSF, administered in the same manner as a “classical” adjuvant, is effective in eliciting antibody responses to tt. Experiments depicted in Fig 5 suggest that prolonged “conditioning” of the injection site with GM-CSF did not appear to be necessary for the induction of immune responses. Most standard adjuvants are mixed with antigen and administered as a single dose per vaccination cycle, not administered repeatedly without antigen for days after the immunization. To determine the efficacy of GM-CSF as an adjuvant in relationship to other standard adjuvants as well as to evaluate the use of GM-CSF in a more standard vaccination regimen, we immunized rats with tt and CFA, Alum, GM-CSF idsq, or PBS. Immunizations were administered once, with no prolonged dosing of GM-CSF alone, and animals were evaluated for antibody response 24 days after the inoculation. Figure 7 demonstrates that GM-CSF id, again, generates a higher titer anti-tt antibody response than GM- CSF sq, despite being administered as a one time dose. Sig- nificant anti-tt antibody response in CFA and Alum immu-

nized animals is still detectable at a 1:6,400 sera dilution. Although they show a lower titer response, anti-tt antibodies are still above baseline in the GM-CSF id immunized animals at 1:6,400. No antibodies were detected with the negative control antigen, ova albumin.

GM-CSF, as an adjuvant, elicits DTH responses to tt similar to those seen in animals immunized with a standard adjuvant. Cell mediated immunity was assessed in the animals previously described by the measurement of tt specific DTH. All animals immunized with an adjuvant developed DTH responses to tt. GM-CSF immunized animals had responses similar to those seen in animals immunized with a standard adjuvant and tt (Fig 8). Animals immunized with PBS and tt developed weak to no DTH response.

Intradermal GM-CSF is a potent adjuvant for the generation of an immune response to peptides derived from a self tumor antigen. Experiments previously described show that GM-CSF id is an effective adjuvant for the generation of immune responses to foreign antigens. We questioned whether the same regimen would be as effective in augmenting immune responses in tumor antigen systems with peptide based vaccines. In initial experiments, two peptides derived from the rat neu protein sequence, p45 and p788, were chosen for analysis. Peptide $5 is derived from the extracellular domain of the neu protein and p788 is derived from the intracellular domain. This animal system, a model

1.6 , -

8 0.8 -

0 6 -

0.4 - 0.2 -

0 -

b ..... . ..... pBs

..... + ....

...A ....

Antibody Titel

Fig 5. Intradermal injection of GM-CSF for five inoculations is an effective adjuvant for eliciting antibody responses to tt when the antigen is administered with the first two injections of cytokine. Rats were injected with OM-CSF or PBS id or sq for five inoculations. Six animals were included in each experimental group. This figure represents data collected from two separate experiments. Tetanus toxoid was added to the cytokine or PBS inoculation at a concentration of 3 Lf units during the first two inoculations. Three control animals that were not immunized are shown as an example of a native rat response to tt. Twenty-four days after the f i ih GM-CSF or PBS inoculation, sera was obtained from each animal. Tetanus toxoid responses were determined by EUSA. Results are depicted as the mean and standard deviation of the antibody response of each experimental group at each sera dilution. Positive control sera from an animal with high n specific antibody was run on each plate and is expressed as the mean and standard deviation of al l experimental plates in this analysis.


'1 0 POlVClOnal

OM-CSF BS Adjuvant

B

OD

I PBS 8s Adjuvant

Fig 6. Intradermal GM-CSF, as an adjuvant, elicits IgG antibody responses t o tt. Rats were injected with GM-CSF or PBS id for five inoculations. tt was added t o the cytokine or PBS inoculation at a concentration of 3 Lf units during the first two inoculations. An example of antibody class analysis is shown here for three GM-CSF and three PBS immunized animals. A control animal that was not immunized is shown as an example of a native rat response to tetanus. Twenty-four days after the fifth GM-CSF or PBS inoculation, sera was obtained from each animal. Tetanus toxoid responses were determined by ELISA. Results are shown at a sera dilution of 1:lOO. (A) The predominant IgG antibody responses of animals immunized with tt and GM-CSF id as adjuvant. (B) The responses of the PBS immunized and control animal.

for evaluating the generation of immunity to the HER-Uneu oncogene protein product in human adenocarcinomas, allows the development of an immune response to rat neu, a self protein, when CFA is used as an adjuvant.I2 Animals immunized with p42 and p788 in CFA will generate peptide specific T-cell responses (M.L.D., unpublished observations, 1995). Initial experiments evaluated the 5-day regimen of GM-CSF inoculation with antigen administered on the first 2 days of the cycle. Id administration of GM-CSF for five inoculations with peptides administered with the first two injections induced vigorous peptide specific DTH responses (Fig 9A). Nonexistent or weak peptide specific responses were measured in rats who received peptide in PBS.

Protein specific rat neu cellular immune responses could be induced with a vaccine composed of multiple neu derived peptides administered with CFA sq after three in vivo immunizations.'* We compared this regimen with rat neu peptides inoculated id with GM-CSF. Vaccines consisting of 5 ECD peptides or 4 ICD peptides were administered once every

1.4 1 1.2 - I -

0.8 - n 0.6 -

0.4 -

0.2 7

- CFA - ALUM

..... 0 ..... GMCSF,d

Antibody Tiler

Fig 7. GM-CSF, administered in the same manner as a "classical" adjuvant, is effective in eliciting antibody responses t o tt. Rats were injected with CFA sq, Alum sq, GM-CSF id or sq (5 pg) and PBS sq with tt at a concentration of 3 Lf units. Immunizations were administered on one day only with no repeated administration of GM-CSF. Six animals were included in each experimental group. This figure represents data collected from two separate experiments. Three control animals that were not immunized are shown as an example of a native rat response t o tt. Twenty-four days after the immunizations sera was obtained from each animal. Tetanus toxoid responses were determined by ELISA. Results are depicted as the mean and standard deviation of the antibody response of each experimental group at each sera dilution. Positive control sera from an animal with high tt specific antibody was run on each plate and is expressed as the mean and standard deviation of all experimental plates in this analysis.

CFA Alum GM-CSFid GM-CSFsq PRS

Rat Adjuvant Groups

Fig 8. GM-CSF, as an adjuvant, elicits DTH responses t o tt similar t o those seen in animals immunized with a standard adjuvant. Rats were injected with CFA sq, Alum sq, GM-CSF id or sq (5 pg) and PBS sq with tt at a concentration of 3 Lf units. Immunizations were administered on one day only with no repeated administration of GM-CSF. Six animals were included in each experimental group. This figure represents data collected from two separate experiments. Twenty days after immunization, a DTH response to tt was measured in the immunized rats. Antigen was applied t o rat ear and responses measured at 48 hours. Ear swelling of experimental compared with control ear was measured. Results are depicted as the mean and standard deviation of measurements taken from each experimental group.

208 DlSlS ET AL

20 - B

T

15 - l- l l

CPA + GMCSP+ CFA + GMCSF + ECDpcpfidcs ECDpptidcs ICDpptidcs ICDpptidcs

Rat Adjuvant Gmup

Fig 9. Intradermal GM-CSF is a potent adjuvant for the generation of immune reaponses to peptides derived from a wlf tumor antigen. (A) Rats were immunized with rat neu peptides p45 and p788 id in either GMCSF or PBS. Thesa peptides had previously been shown to generate T-cell responses in Fischet rats. Immunizing peptide (250 pg) was added to adjuvant and injected with the first two of five cytokine or PBS immunizations. Either GMCSF or PES alone ware administered for the remaining 3 days. Peptide specific DTH responses wera measured 20 days after the fifth inoculation. These results represent the mean and standard deviation of measurements taken from three animals in each experimental group. (6) R a t s were immunizad with multiple peptides derived from rat neu protein, ai- ther a vaccine composed of a group of 5 ECD peptides or 4 ICD peptides. Animals were administered GM-CSF and antigen id or CFA and antigen sq. Immunizations wera administered on one day only with no repeated administration of OM-CSF. The immunizations ware repeated every 20 days for a total of three vaccinations. Twenty days after the final immunization, a DTH response to rat neu protein was measured. Antigen was applied to rat ear and responses assessed at 48 hours. Ear swolllng of expodmental compared with control aar was measured. Resub are dapicted as the mean and standard deviation of measurements taken from each experimental group.

20 days with GM-CSF or CFAAFA for three inoculations. GM-CSF, id, was as effective as CFA as an adjuvant for eliciting rat neu protein specific DTH responses (Fig 9B). Thus, GM-CSF is an effective adjuvant for both protein and peptide based vaccines.

DISCUSSION

GM-CSF has many characteristics that would make it an excellent potential vaccine adjuvant. One mechanism by which GM-CSF is hypothesized to augment an immune response is its function as a growth and differentiation factor

for epidermal Langerhan’s cells,“ a class of dendritic cell. These cells are potent APC and thought to be the operative cell type in the initiation of T cell antigen interaction.” GM- CSF has been shown to stimulate growth of other APC in- cluding blood born DC and macrophage^.^^^'^^^^ DC, them- selves, have been used as vaccine adjuvants and are capable of inducing significant immune responses when injected along with antigen.” Using local injection of GM-CSF to similarly expand the number of APC would be expected to enhance vaccine immunogenicity. Another benefit to the use of GM-CSF as a vaccine adjuvant is its ease of use in humans. GM-CSF has been used extensively as a hematopoietic growth factor in patients undergoing chemotherapy, and repeated administration of the cytokine is generally well toler- ated.21,22 Finally, cytokines as adjuvants may influence the character of the immune response generated by influencing the subset of CD4’ helper cell activated, TH1 versus TH2.

Initial investigations, presented here, concentrated on de- fining the most appropriate method and route of GM-CSF administration for use as an adjuvant. GM-CSF has been described not only as a growth and differentiation factor for epidermal Langerhan’s cells, but also as a chemoattractant for these APCS.’~ It was unknown how long a local injection site would need to be “conditioned” with GM-CSF to recruit andor grow APC. It is also unknown how long GM- CSF would have to be administered to enhance the efficacy of the recruited cells. A study by Kaplan et a12’ addressed evaluating Langerhan’s cell recruitment after GM-CSF administration in patients with leprosy. Investigators evaluated the effect of GM-CSF injected into leprosy lesions either id or sq daily for 10 days. They noted recognizable Langerhan’s cells present in the dermis at 48 hours and reaching peak numbers at 4 days when the cytokine was administered id. No Langerhan’s cell infiltration was noted when the cytokine was administered sq. This observation led us to study the effect of GM-CSF administration over a 5-day period. The current studies examined both sq and id routes of administration. Although sq administration did not show increased Langerhan’s recruitment in studies by Kaplan et al,” it was a very effective route of immunization with GM-CSF transfected tumors? The sq route has the theoretical advantage of acting as a depot site for the cytokine, thus allowing a longer half-life in the vicinity of antigen deposition. The id route has the advantage of delivering a DC growth factor directly to the skin compartment where DC reside. Once DC are exposed to antigen and activated, they migrate to regional lymph nodes to activate T cells. Using class I1 expressing cells in lymph nodes draining the injection site as a measure of APC, we noted an effect on these cells in both id and sq immunized groups as compared with controls; however, the kinetics of that effect varied depending on the route of administration used. Sq administration resulted in either an early influx of “APC” into regional lymph nodes or upregulation of class I1 on cells present in regional lymph nodes with marked differences occurring between the experimental and control groups within the first three injections and then returning to baseline. Id administration resulted in a later effect peaking at the fourth and fifth injections, similar to the results Kaplan et a1 demonstrated in humans. We evaluated


whether this difference in class I1 expressing cells in draining lymph nodes translated into functional differences in antigen presentation between the two routes of administration.

Functional experiments evaluating GM-CSF as an adjuvant in a foreign antigen (tt) system were initiated adminis- tering antigen late in the cytokine regimen. This time course was tested first based on the assumption that the site for induction of immunity might need to be “conditioned” by inducing the differentiation and growth of APC. The observation that id GM-CSF administration resulted in later changes in class I1 expressing cells “APC,” was supported by that assumption. Although tt specific antibody responses were generated in both id and sq immunized groups, responses were more consistent in the id immunized animals. This data implies that the changes in class I1 expressing cells observed with GM-CSF administration did have important functional effects. To test the necessity of site “conditioning” with GM-CSF, we next evaluated the co-administration of GM-CSF and tt early in the cytokine cycle, injecting antigen with the first two of five inoculations of GM-CSF or PBS. Early antigen administration by both the id and sq routes was effective. The id route of administration appeared, again, to be more effective than sq at eliciting tt specific antibody responses. The upregulation of expression of class I1 MHC on DC has been associated with their maturation and the ability to augment a mixed lymphocyte reaction, but not necessarily with antigen presentati~n.’~.~~.’’ The efficacy of the id route in early antigen administration may represent loading of antigen on immature DC and subsequent presentation. The continued efficacy later in the cycle may reflect continuing recruitment of immature APC to the site of antigen deposition.

The adjuvant properties of GM-CSF were compared with those of two “standard” adjuvants, CFA and Alum. In addition, vaccination classically involves the administration of the antigen with the adjuvant as a one time dose. Repeated administration of adjuvant after the introduction of the immunizing antigen is not the norm. Experiments were de- signed not only to compare the use of GM-CSF with other adjuvants, but to tailor the immunizing regimen to one more commonly used in human immunizations: single dose of antigen in adjuvant. GM-CSF was as effective as CFA or Alum in generating DTH response to tt. GM-CSF id was effective in eliciting antibody immunity to titers similar to those seen with the standard adjuvants. Studies are currently underway optimizing the dose of cytokine needed to stimulate an effective immune response.

There are many adjuvants effective in eliciting antibody and T-cell responses to foreign proteins. A more critical issue for developing vaccine therapy for human malignancy is whether adjuvants can elicit T-cell responses to proteins or peptides derived from self tumor antigens. The recent identification of specific tumor antigens has renewed interest in the formulation of a new generation of antigen-directed cancer vaccines. As more of these antigens are elucidated, it becomes clear that a growing number of putative tumor antigens are self proteins.’ Investigations in our laboratory involve the development of a rat system for evaluating immunity to rat neu as a model for studying immunity to human

HER-2heu. Studies have identified several peptide epitopes, derived from the normal rat neu protein sequence, capable of eliciting T-cell responses in rats when immunized with CFA.” Although peptides are generally considered weak immunogens, and these peptides were derived from a self protein, cell mediated immune responses to these peptides were generated by priming with intradermal GM-CSF as a vaccine adjuvant. We presumed even greater responses would ensue following boosting experiments. Using a multiple peptide vaccine known to generate rat neu specific responses in the rat, and CFA as an adjuvant, we administered an initial immunization and two boosting immunization to the animals. GM-CSF and the peptide based vaccine administered id was as effective as CFA and the peptide vaccine administered sq in generating cell mediated immunity.

The current study identifies GM-CSF as an excellent adjuvant for the elicitation of immunity to both foreign proteins and tumor antigen derived peptides. The cytokine is able to aid in the generation of both antibody and cell mediated immune responses. Studies shown here show the effective- ness of GM-CSF in the initiation of an immune response, without cytokine gene transfection or depot formulation for prolonged exposures at the site of antigen deposition. In addition, GM-CSF compares favorably with other standard adjuvants. The recent observations of an immune response directed against GM-CSF i t ~ e l f , ~ ~ , ~ ’ when administered at higher concentrations over time, theoretically may make the intermittent use of standard doses of GM-CSF, as in the current studies, the most effective way to use the cytokine to enhance immune responses. GM-CSF has been used extensively in humans and has shown minimal toxicity even after repeated injections. The use of GM-CSF in humans, as an adjuvant for peptide-based vaccines, may allow the realization of an effective and long-lived immune response to immunogenic epitopes.

ACKNOWLEDGMENT

We wish to acknowledge Kent Slaven for animal technical support and Kevin Whitham for manuscript preparation.

REFERENCES 1. Pardoll DM: A new look for the 1990s. Nature 369:357, 1994 2. Houghton AN: Cancer antigens: Immune recognition of self

and altered self. J Exp Med 180:1, 1994 3. Milich DR: Synthetic peptides: Prospects for vaccine develop-

ment. Semin Immunol 2:307, 1990 4. Amon R, Horwitz RJ: Synthetic peptides as vaccines. Curr

Opin Immunol 4449, 1992 5. Claassen E, de Leeuw W, de Greeve P. Hedriksen C, Boersma

W: Freund’s complete adjuvant: An effective but disagreeable for- mula. Res Immunol 143:478, 1992

6. Edelman R: An update on adjuvants in clinical trial. AIDS Res Hum Retrovirus 8:1409, 1992

7. Allison AC: Adjuvants and immune enhancement. Int J Tech- no1 Assess Health Care 10:107, 1994

8. Momssey PJ, Bressler L, Park LS, Alpert A, Gillis S: Granulo- cyte-macrophage colony-stimulating factor augments the primary antibody response by enhancing the function of antigen presenting cells. J Immunol 139: 1 1 13, 1987

9. Dranoff G, Jaffee E, Lazenby A, Golumbek P, Levitsky H,

210 DlSlS ET AL

Brose K, Jackson V, Hamada H, Pardoll D, Mulligan RC: Vaccina- tion with irradiated tumor cells engineered to secrete murine granulocyte-macrophage colony-stimulating factor stimulates potent, specific, and long-lasting anti-tumor immunity. Proc Natl Acad Sci USA 90:3539, 1993

10. Reid CD, Stackpoole A, Meager A, Tikerpae J: Interactions of tumor necrosis factor with granulocyte-macrophage colony-stimulating factor and other cytokines in the regulation of dendritic cell growth in vitro from early bipotent CD34+ progenitors in human bone marrow. J Immunol 149:2681, 1992

11. Golumbek FT, Azhari R, Jaffee EM, Levitsky HI, Lazenby A, Leong K, Pardoll DM: Controlled release, biodegradable cytokine depots: A new approach in cancer vaccine design. Cancer Res 535841, 1993

12. Disis ML, Grolow JR, Bemhard H, Hand SL, Rubin WD, Chewer MA. Peptide based, but not whole protein, vaccines elicit immunity to HER-2/neu, an oncogenic self-protein. J Immunol 1996 (in press).

13. 't Hart BA, Kraakman E, Jonker M: An ear swelling reaction for in vivo monitoring of inflammatory T cell activity in collagen- arthritic rats. J Immunol Methods 176:179, 1994

14. Zecher R, Scheicher C, Wagener S, Reske-Kunz AB, Reske K: Modulation of accessory cell function of immortalized bone marrow-derived macrophages by granulocyte/macrophage colony-stimulating factor. Med Microbiol Immunol Berl I82:153, 1993

15. Gosselin ET, Wardwell K, Rigby WF, Guyre PM: Induction of MHC class I1 on human polymorphonuclear neutrophils by granulocyte/macrophage colony-stimulating factor, IFN-gamma, and IL- 3. J Immunol 151:1482, 1993

16. Caux C, Dezutter-Dambuyant C, Schmitt D, Banchereau J: GM-CSF and TNF-alpha cooperate in the generation of dendritic Langerhans cells. Nature 360:258, 1992

17. Steinman RM: The dendritic cell system and its role in immunogenicity. Annu Rev Immunol 9:271, 1991

18. Kat0 T, Inaba K, Ogawa Y, Inaba M, Kakihara K, Shimizu S, Ikehara S, Sudo T, Muramatsu S: Granulocyte-macrophage colony-

stimulating factor enhances macrophage accessory function in con A-stimulated T-cell proliferation. Cell Immunol 130:490, 1990

19. Bemhard H, Disis ML, Heimfeld S, Hand S, Gralow JR, Cbeever MA: Generation of immunostimulatory dendritic cells from human CD34' hematopoietic progenitor cells of the bone marrow and peripheral blood. Cancer Res 55:1099, 1995

20. Flamand V, Somasse T, Thielemans K, Demanet C, Leo 0, Urbain J, Moser M: Vaccination with tumor-antigen-pulsed dendritic cells induces in vivo resistance to a B cell lymphoma. Adv Exp Med Biol 329:611, 1993

21. Stem AC, Jones TC: The side-effect profile of GM-CSF. Infection 2:S124, 1992

22. Armitage JO: The use of granulocyte-macrophage colony- stimulating factor in bone marrow transplantation. Semin Hematol 29:14, 1992

23. Kaplan G, Walsh G, Guido LS, Meyn P, Burkhardt RA, Ab- alos RM, Barker J, Frindt PA, Fajardo 'IT, Celona R: Novel responses of human skin to intradermal recombinant granulocyte/macrophage-colony-stimulating factor: Langerhans cell recruitment, keratinocyte growth, and enhanced wound healing. J Exp Med 175:1717, 1992

24. Streilein JW, Grammer SF: In vitro evidence that Langerhans cells can adopt two functionally distinct forms capable of antigen presentation to T lymphocytes. J Immunol 143:3925, 1989

25. Romani N, Koide S, Crowley M, Witmer-Pack M, Living- stone AM, Fathman CG, Inaba K, Steinman RM: Presentation of exogenous protein antigens by dendritic cells to T cell clones. Intact protein is presented best by immature, epidermal Langerhans cells. J Exp Med 169:1169, 1989

26. Chen TT, Levy R Induction of autoantibody responses to GM-CSF by hyperimmunization with an id-GM-CSF fusion protein. J Immunol 154:3105, 1995

27. Ragnhammar P, Friesen HJ, Frodin JE, Lefvert AK, Hassan M, Osterborg A, Mellstedt H: Induction of anti-recombinant human granulocyte-macrophage colony-stimulating factor (Escherichia coli-derived) antibodies and clinical effects in nonimmunocompro- mised patients. Blood 84:4078, 1994

granulocyte-macrophage colony-stimulating factor: an effective … · 2012. 3. 8. ·...

Documents