Idiotypic protein-pulsed dendritic cell vaccination in multiple myeloma

Authors


Abstract

Idiotypic protein (Id) produced by myeloma cells is clone-specific and may be a suitable tumor-specific antigen for immune targeting. Advances in dendritic cell (DC) technology suggest the opportunity for using this potent antigen presentation system to deliver myeloma Id to the autologous host to elicit anti-tumor immune responses. We have generated DCs from adherent PBMCs from 6 patients with IgG myeloma. These cells were pulsed with the autologous Id and a control vaccine, KLH, and re-infused i.v. back to the patients on 3 separate occasions. Immune responses to KLH and autologous Id were measured and clinical responses monitored. We found that all treatments were well tolerated without any side effects. All patients developed both B- and T-cell responses to KLH, suggesting the integrity of the host immune system to mount immune responses to an antigen delivered using our vaccination strategy. Id-specific responses were also observed. PBMC proliferative responses to Id were observed in 5 of the 6 patients following treatment. In 2 patients, the responses were associated with the production of IFN-γ. There were also increases in cytotoxic T-cell precursor frequencies for Id-pulsed autologous targets in 3 patients. B-cell responses characterized by the production of anti-Id IgM occurred in 3 and anti-Id IgG in 4 of the 5 evaluated patients. In 1 patient, a modest (25%) but consistent drop in the serum Id level was observed. Id-pulsed DC vaccination can therefore elicit potentially useful anti-myeloma immune responses in patients with multiple myeloma. Int. J. Cancer 83:215–222, 1999. © 1999 Wiley-Liss, Inc.

Multiple myeloma (MM) is a malignant disease commonly affecting elderly individuals and characterized by accumulation in the bone marrow of mature plasma cells. Although it is possible to produce a period of disease remission using combination chemotherapy, the disease is essentially incurable. Therefore, novel approaches are much needed.

The malignant cells in MM, like the other B-cell malignancies, undergo immunoglobulin gene re-arrangement and secrete idiotypic protein (Id), which is usually monoclonal IgG or IgA. Unlike B-cell lymphoma, Id produced by myeloma cells is secreted into the serum and the malignant cells express very little surface Id. Since Id is clone-specific, it represents an ideal tumor-specific antigen for immune targeting.

A successful immunological approach to tumor vaccination relies on the optimal processing and presentation of tumor antigens to the host immune system. Dendritic cells (DCs) are powerful antigen-presenting cells equipped with the necessary co-stimulatory molecules and antigen presentation properties needed for the initiation of an effective primary immune response (Larsen et al.,1994; Hathcock etal., 1994). Animal studies have demonstrated that DCs pulsed with tumor peptides can be used to induce protective immunity against tumor challenge (Flamand et al., 1994). Clinical and immunological responses have also been reported in various tumor-bearing patients treated with antigen- or cell lysate–pulsed DCs (Hsu et al., 1996; Nestle et al., 1998; Rosenberg et al., 1998). DCs can be expanded in vitro from CD34+ cells (Caux et al., 1992; Egner and Hart, 1995) or from the differentiation of peripheral blood monocytes using IL-4 and granulocyte-macrophage colony-stimulating factor (GM-CSF) (Sallusto and Lanzavecchia, 1994; Romani et al., 1994).

Although Id is a weak antigen, various laboratory and clinical studies on B-cell lymphoma suggest that Id or idiotypic peptides are immunogenic to autologous hosts (Campbell et al., 1988, 1990; George et al., 1988; Kwak et al., 1992; Wen and Lim, 1997). In MM, the immunogenicity of Id (Osterborg et al., 1995, 1998; Kwak et al., 1995; Wen et al., 1998a) and idiotypic peptide (Wen et al., 1998b) has also been demonstrated. Idiotype-responsive T cells have been demonstrated in the peripheral blood of patients with stage I disease (Osterborg et al., 1995) and the immune repertoire of a normal individual (Kwak et al., 1995). We have shown that Id-pulsed DCs could prime the immune system of a patient with refractory MM (Wen et al., 1998a). In addition, Id given in combination with GM-CSF to patients with MM induced type I T-cell responses (Osterborg et al., 1998). All of these results suggest a potential for immune targeting in patients with stable or early disease. In the present study, we have administered Id-pulsed DCs to 6 patients with early-stage/early-relapse MM. Anti-Id immunity as well as clinical effects were monitored.

MATERIAL AND METHODS

Patients

Six IgG MM patients were enrolled following informed consent and local ethical approval. The clinical characteristics of these patients are shown in Table I. Three patients had previously received systemic anti-myeloma therapy. None of the patients received any systemic therapy at the time of recruitment into the study. Each patient received 3 i.v. DC vaccinations, each 2 weeks apart. Following vaccination, patients were followed up at regular intervals and immune responses to vaccines monitored closely.

Table I. CLINICAL CHARACTERISTICS OF PATIENTS BEFORE VACCINATION
PatientAgeSexPreviousInterval from lastNumber of
number (years)chemotherapy chemotherapy to vaccination vaccinations
  1. ABCM, adriamycin, BCNU, cyclophosphamide and melphalan; C/VAMP, cyclophosphamide, vincristine, doxorubicin and methylprednisolone.

00149MNoneNot applicable3
00289F5 courses of oral melphalan6 months2
   12 courses of dexamethasone 
   6 courses of dexamethasone 
00364FNoneNot applicable3
00484FNoneNot applicable3
00563F14 courses of melphalan and prednisolone3 weeks3
00657F5 courses of ABCM3 years3
   4 courses of C/VAMP
   High-dose melphalan with stem cell rescue

Isolation of Id

To isolate and purify the myeloma Id from serum, we employed diethylaminoethyl (DEAE) chromatography. The protein was more than 95% pure as determined by SDS-PAGE. A panel of control Id was also isolated using the same methodology from the sera of 4 other IgG myeloma patients. The control Id was included in subsequent experiments to determine the specificity of the immune responses. Fab fragments were prepared from the purified Id by digesting the Id (2.5% solution in PBS, to which cysteine and EDTA were added to produce a final concentration of 0.01 M and 0.002 M, respectively, with 1 mg of crystalline mercuripapain/100 mg of Id). The mixture was incubated for 18 hr at 37°C and the reaction stopped by adding iodoacetamide (to a final concentration of 0.01 M). The preparation was then purified by running the mixture through a protein A Sepharose column (Pharmacia, Milton Keynes, UK) and collecting the unbound fraction to remove undigested Id and Fc fragments.

Generation of DCs

Each patient underwent steady-state leukapheresis, to enable the harvesting of a large number of PBMCs. Fresh PBMCs (40 ml) were obtained from each patient. Cells were washed and resuspended in RPMI 1640 + 10% (v/v) FCS (GIBCO, Paisley, UK). PBMCs were incubated at 37°C in 5% CO2 for 2 hr. Non-adherent cells were removed by gentle washes. Adherent cells were then cultured in RPMI 1640 + 10% FCS, supplemented with GM-CSF (800 U/ml) (Sandoz, Leeds, UK) and IL-4 (500 U/ml) (Schering-Plough, Welwyn Garden City, UK) for 7 days. Cells were pulsed with Id (200 μg/ml) and keyhole limpet hemocyanin (KLH; Sigma, St. Louis, MO) (50 μg/ml) on days 1 and 6 of culture.

Proliferation assays

Fresh PBMCs were seeded in 96-well flat-bottomed microtiter plates (Becton Dickinson, San Jose, CA) at 2 × 105 per well in 100 or 200 μl of RPMI 1640 supplemented with 10% human AB serum, 1 mM glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin and 25 mM HEPES (referred to as complete medium). KLH, Id or control Id was added to cultures in various final concentrations, and the cultures were incubated at 37°C in 5% CO2. Cells were pulsed on day 4 with 0.5 μCi [methyl-3H]-thymidine per well and harvested 18 hr later; [methyl-3H]-thymidine incorporation was measured by liquid scintillation counting. Results from triplicate cultures are given as stimulation index (SI) + SD. SI was calculated by dividing the mean radioactivity for a triplicate of stimulated cells by that of unstimulated cells. For Id-specific proliferation, the results are given as relative stimulation index (RSI), which was calculated by dividing the SI obtained from Id stimulation by those obtained from the panel of control Id stimulation. An SI or RSI in excess of 2 was considered positive. Supernatants were harvested after 72 hr and assayed for the production of IFN-γ using a commercially available ELISA kit (Genzyme, Cambridge, MA). All results were confirmed on at least 2 separate occasions.

Cytotoxic T-lymphocyte assays

Cytotoxic T-lymphocyte (CTL) precursor assays were carried out as described by Wen et al. (1998a). Briefly, PBMCs obtained before and immediately after the 3 vaccinations were incubated in 96-well round-bottomed microtiter plates at varying concentrations (5,000, 10,000, 20,000 and 50,000/well, 21 wells per cell concentration) in a volume of 150 μl of complete medium in the presence of Id (50 μg/ml) or control Id (50 μg/ml) and IL-2 (30 U/ml) for 1 week. Cells were then assayed for their cytotoxic ability for various targets in a standard 4-hr 51Cr-release assay. Target cells consisted of autologous DCs and autologous CD3-depleted PBMCs loaded with either Id or control Id. Target cells were labeled with sodium [51Cr] chromate for 1 hr and mixed with pre-cultured effector cells. The frequency of CTLp was estimated by standard limiting dilution assays; wells with counts above mean + 3SD were scored as positive. Results are shown as CTLp/106 PBMCs.

ELISA measurements of antibodies

Antibody responses to vaccination were measured by ELISA. Microtiter plates were coated with Fab fragments from Id, Fab fragments from the panel of isotype control Id or KLH (all at 50 μg/ml) overnight. Pre- and post-vaccination serum samples were serially diluted and allowed to bind to the target proteins. The binding of anti-Id IgM antibodies and anti-KLH IgM antibodies was detected by mouse anti-human IgM antibodies, followed by horseradish peroxidase–conjugated goat anti-mouse IgG antibodies. The binding of anti-KLH and anti-Id IgG antibodies was detected using biotinylated goat anti-human IgG antibodies, followed by peroxidase-conjugated streptavidin.

RESULTS

Vaccinations

DCs were successfully generated in all 6 patients. However, 1 patient (patient 004) had a significantly lower cell number. The proportions of cells expressing CD1a positivity were not significantly different from those obtained from normal donors using a similar culture protocol (Coleman et al., 1997). Cultured DCs were decanted from the flasks and adherent cells dislodged by vigorous shaking with PBS. Pooled cells pulsed with both Id and KLH were washed 3 times in PBS and divided into 3 equal aliquots, to be administered respectively on each occasion. One aliquot was re-infused fresh and consisted of fresh cultured DCs resuspended in PBS/4% human albumin solution. The other 2 aliquots, resuspended in cooled PBS containing 4% human albumin solution and 7.5% DMSO, were frozen using the Planer rate-controlled freezer at 1°C/min. Frozen cells were stored in the vapor phase of a liquid nitrogen vessel. When required for re-infusion, cells were rapidly thawed at 37°C, diluted 5-fold with PBS and immediately injected i.v. over 15 min. Cell recovery and Trypan blue viability were >90%.

Each vaccine was administered i.v. over 15 min with i.v. chlorpheniramine, 10 mg. Five patients received 3 vaccines each and one patient (patient 002) 2 vaccines. The number and characteristics of DCs infused with each treatment varied in each patient (Table II). FCS was used in the generation of DCs because our preliminary results suggested that it is superior to human AB serum and the yields of DCs were more consistent. In addition, FCS may provide non-specific helper function to the vaccine. All treatments were well tolerated with no significant side effects observed in any of the patients throughout the treatment period.

Table II. CHARACTERISTICS OF VACCINES
 Cell dose each 
Patient vaccination% cells expressing surface marker
number (×106)CD1aCD11cCD14CD80CD86
00147.5193804513
00234.4155220229
003383671186051
0043.5456705018
005404157305718
006891350474353

Immunological responses to control vaccine, KLH

To check the integrity of the immune system and the vaccination approach, a control vaccine consisting of DCs pulsed with KLH was also administered to each patient. For this purpose, we checked the T-cell proliferative responses to KLH and the production of anti-KLH IgM and IgG antibodies following each treatment. No significant proliferative responses to KLH were observed prior to treatment in all 6 patients. However, following the first vaccination, PBMCs from 5 of the 6 patients proliferated in response to in vitro KLH restimulation (Fig. 1). Only 1 patient developed the proliferative responses after the second vaccination. Proliferative responses became stronger with each treatment. CTL responses to KLH were not determined.

Figure 1.

Proliferative responses of PBMCs to in vitro KLH restimulation before and after each treatment in each patient (results expressed as SI ± SD. 3H-thymidine uptake of unstimulated triplicate wells, in all cases, was <500 cpm (SI = cpm from culture wells with KLH/cpm from culture wells without KLH).

B-cell responses to the vaccine were also observed in all 6 treated patients. Anti-KLH IgM (Fig. 2a) and anti-KLH IgG antibodies (Fig. 2b) were detected in the serum from all 6 patients after Id-pulsed DC treatment. Antibodies were detected as early as after the first vaccination, and IgG titers increased with each treatment.

Figure 2.

B-cell responses to KLH, determined by the production of (a) anti-KLH IgM antibodies and (b) anti-KLH IgG antibodies in each patient at various serum dilutions before and after each treatment (figures represent 1 of 2 similar results obtained in 2 separate experiments).

Id-specific proliferative responses

We then measured Id-specific proliferative responses of PBMCs derived from the patients before and after each vaccination. Relative to responses to a panel of control Id, only patient 002 demonstrated PBMC proliferative responses prior to vaccination (Fig. 3). No proliferative responses were observed in the other 5 patients. Following the first vaccination, patients 001 and 002 developed positive proliferative responses (as judged by RSI) in a dose-dependent manner. Positive proliferative responses were observed in 4 of the 5 patients who could be monitored after the second and third treatments. No increase in 3H-thymidine uptake was observed in patient 006 after any of the vaccinations. In general, the proliferative responses to Id, when observed, were of a lower magnitude compared with those observed with KLH restimulation.

Figure 3.

PBMC proliferative responses to in vitro Id stimulation before and after each treatment in each patient (results expressed as RSI ± SD. 3H-thymidine uptake of unstimulated triplicate wells, in all cases, was <500 cpm (RSI = SI from Id stimulation/SI from control Id stimulation).

Proliferative responses to the panel of control Id were also observed in patient 004 after each treatment, but they were of a lower magnitude than those with autologous Id. In patient 006, although no specific responses to autologous Id were observed, PBMCs isolated after the third vaccination exhibited proliferative responses to both autologous and control Id, and the magnitude of the response to each stimulant was not significantly different.

B-cell responses to autologous Id

B-cell responses to autologous Id were evaluated in 5 of the 6 treated patients. Antibodies in patient 002 were not measured because there was insufficient Id for the preparation of Fab fragments. Anti-Id IgM and IgG were measured in the serum in these patients before and after each treatment. Despite the presence of circulating Id, IgM (Fig. 4a) and IgG (Fig. 4b), antibodies directed at autologous idiotypic Fab fragments were detected in the serum of 3 of the 5 and 4 of the 5 treated patients, respectively. Antibodies did not bind to the panel of control idiotypic Fab fragments. Anti-Id IgM antibodies were not detected in patient 003, and both anti-Id IgM and IgG antibodies were detected in patient 006.

Figure 4.

(a) B-cell responses to Id and control Id, determined by the production of anti-vaccine IgM antibodies. (b) Anti-vaccine IgG antibodies in each patient at various serum dilutions before and after each treatment (figures represent 1 of 2 similar results obtained in 2 separate experiments).

Id-specific IFN-γ production

Production of IFN-γ in the culture supernatant following in vitro restimulation of PBMCs with autologous Id was determined in all 6 patients before and after treatments. This was compared to the cytokine levels in the test culture supernatants using a panel of control Id. No significant IFN-γ was detected in any of the supernatants when PBMCs were stimulated in vitro with the panel of control Id, either before or after vaccination. Similarly, no IFN-γ was secreted into the culture supernatants derived from all 6 patients before treatment when PBMCs were stimulated with the autologous Id. In contrast, significant levels of IFN-γ were produced in the culture supernatants from patients 001 and 003 after vaccination (Fig. 5). IFN-γ was detected in these 2 patients only after the third treatment.

Figure 5.

Production of IFN-γ by PBMCs upon restimulation in vitro with autologous Id before and after treatment (results expressed as mean ± SD, figures represent 1 of 2 similar results obtained in 2 separate experiments).

CTL induction

To determine the ability of the treatment to induce CTLs, we evaluated CTLp frequency in the PBMCs using 3 autologous targets: Id-pulsed DCs, Id-pulsed T-depleted PBMCs and control Id-pulsed T-depleted PBMCs. Increased CTLp frequency for autologous Id-pulsed targets was observed in 3 patients, patients 001, 004 and 006 (Table III). In general, CTLp frequencies were higher for Id-pulsed autologous DCs than T-depleted PBMCs. The specificity of the CTLs was suggested by the lack of cytotoxicity for autologous PBMCs pulsed with the control panel of Id. Interestingly, increased CTLp frequency after vaccination was observed in patient 006, in whom there was no detectable T-cell proliferative response or Id-specific antibodies.

Table III. FREQUENCY OF CTL PRECURSORS USING DIFFERENT TARGETS (EXPRESSED AS CTL PRECURSORS/106 PBMCs)
 AutologousAutologousControl
Patient Id-loaded DCs Id-loaded PBMCs Id-loaded PBMCs
numberPrePostPrePostPrePost
  1. n.t., not tested.

001315n.t.n.t.n.t.n.t.
0026256n.t.n.t.n.t.n.t.
003660000
004236201200
005660000
00602201000

Durability of immune responses

The proliferative responses and IFN-γ production were determined in patients 001 and 003 8 and 4 months, respectively, after vaccination. In both patients, the PBMCs were still able to proliferate specifically upon in vitro restimulation with autologous Id and to produce IFN-γ. The magnitudes of the proliferation, however, were lower than when determined immediately following the third vaccination (Fig. 6). Similarly, lower levels of IFN-γ were produced by the PBMCs from both patients (data not shown).

Figure 6.

Durability of T-cell responses to autologous Id in patients 001 and 003 (results expressed as RSI ± SD).

Clinical effects

Only 1 patient (patient 001) demonstrated a minor but persistent reduction in the serum Id level, from a pre-vaccination level of between 20 to 21 g/l to a post-vaccination level of 15 to 17 g/l. The response occurred 2 weeks following completion of the treatment and has remained for 13+ months after starting vaccination. Although disease progression was not observed in the other 5 patients during the period of vaccination, the disease progressed in patients 005 and 006 5 and 2 months, respectively, after completion of the vaccination. Patient 002 could not be evaluated for response because she died of an intercurrent chest infection after the second vaccination. Serum levels of Id in patients 003 and 004 have remained constant 8+ months after completion of the vaccination.

DISCUSSION

The immune system provides potent effector mechanisms which possess memory and can respond to rechallenge by the same antigen. If a tumor-specific antigen could be targeted to produce anti-myeloma effect, immunotherapy would be an ideal therapeutic approach. Active myeloma could be induced into remission using chemotherapy, and disease relapse could be prevented by long-term tumor immunosurveillance. Although Id is tumor-specific, immunogenic and present in abundance, immune responses in the autologous host, especially in patients with active disease, are seldom observed, suggesting that the failure to elicit immune responses may be due to ineffective antigen presentation. In this study, we have therefore examined the ability of autologous DCs pulsed with Id to function as a vaccine when infused i.v. to a group of patients with early-stage/early-relapse MM.

The immune system of patients with MM is often impaired, due to a combination of previous chemotherapy and the disease process. Therefore, we included a control vaccine in the study. Although not completely comparable to Id, KLH is a strong immunogen which served as an immunological tracer molecule and allowed us to determine whether the host immune system was able to respond to the vaccination approach, thereby providing some indicators of the integrity of both the host immune system and our immunization strategy. Results obtained from the control vaccines indicated that KLH delivered via DCs was capable of eliciting both proliferative and antibody responses in patients with MM. In keeping with a previous study involving B-cell lymphoma patients (Hsu et al., 1996), both proliferative and antibody responses to KLH occurred in most patients as early as 1 to 2 weeks after a single infusion of DCs. Responses were strong and increased with progressive rounds of vaccination.

In contrast with responses to KLH, the proliferative responses to autologous Id were much weaker in most of the patients, being between 20 to 50 times less intense; they were observed in only 5 of the 6 patients. Proliferative responses were, however, Id-specific, as judged by the positive RSI. In 4 of the 6 patients, they did not occur in response to the panel of control Id, and in 1 patient, in whom PBMC proliferation occurred when challenged with the control Id, the magnitude of the non-specific proliferation was lower than that observed with autologous Id. Proliferative responses were associated with the production of detectable levels of IFN-γ in only 2 of these 5 patients, supporting a recent study (Osterborg et al., 1998) indicating a lack of correlation between antigen-specific cell proliferation and cytokine production. Of possible significance, proliferative responses were generally stronger in patients who had not previously received any chemotherapy than in those previously treated. Interestingly, one patient made a good proliferative response despite a low vaccination dose. In addition, proliferative and cytokine responses induced in these patients could still be detected in 2 patients 4 and 8 months after completion of the vaccination.

B-cell responses to Id were also measured. We observed the production of specific anti-Id IgM in 3 and IgG in 4 of the 5 evaluated patients. In general, PBMC proliferation predicted the development of antibodies. This finding is in contrast with a previous study involving Id-pulsed DC vaccination of B-lymphoma patients, in which only cellular immune responses were induced (Hsu et al., 1996). The ability to detect anti-Id antibodies in the face of high circulating Id was surprising. One possible explanation may be the low affinity of the anti-Id antibodies.

We also measured the effects of treatment on CTL precursor frequencies for different targets. Increased CTL precursor frequency was observed in 3 of the 6 patients. Increased CTL precursor frequency was observed using both autologous Id-pulsed DCs and T-depleted PBMCs, though frequencies were generally lower with the latter targets, probably reflecting the inferior antigen uptake and presentation by PBMCs compared to DCs. The use of EBV-transformed autologous lymphoblastoid cells as targets might have produced higher CTL frequencies than those obtained using PBMCs. Patient 006, who did not show any T-cell proliferation or cytokine release and B-cell responses following vaccination, was among the 3 patients showing induction of Id-specific CTLs. While induction of Id-specific CTLs is interesting, its clinical significance in the context of lysis of fresh autologous myeloma cells remains to be determined since fresh autologous myeloma cells were not used as targets in these assays.

Although our study involved only 6 patients, a period of stable disease was observed in all treated patients, suggesting, at least in the short term, that Id-pulsed DC vaccination did not exacerbate the disease. In addition, 1 patient showed a consistent 25% reduction in the serum Id level. Obviously, only a randomized study can determine whether or not such modest reduction in the Id level translates into improved overall survival.

In conclusion, our results indicate that Id-pulsed DCs can be administered safely to patients with MM to produce anti-Id immune responses. However, there is poor correlation between the various immunological parameters measured.

Ancillary