Immunological Consolidation of Ovarian Carcinoma Recurrences with Monoclonal Anti-Idiotype Antibody ACA125: Immune Responses and Survival in Palliative Treatment¹: See The Biology Behind: K. A. Foon and M. Bhattacharya-Chatterjee, Are Solid Tumor Anti-Idiotype Vaccines Ready for Prime Time? Clin. Cancer Res., 7:1112–1115, 2001.

Ovarian cancer affects 10–25% of patients with gynecological carcinoma, thus representing the third most common malignancy of the genital tract. Because >50% of ovarian cancer patients are stage III (International Federation of Gynecologists and Obstetricians) and 20% are stage IV at the time of diagnosis (1), this tumor entity is associated with an extremely unfavorable prognosis. Because the success of cytostatic regimes after failure of first-line therapy is limited, the development of more suitable therapeutic methods becomes necessary.

Within the last few years, different immunotherapeutic strategies based on immunization with tumor-specific antibody constructs (e.g., bispecific antibodies), immunogenic peptides, or DNA vaccines have been developed (2, 3). Alternative concepts include the application of genetically modified tumor cells or fibroblasts for the expression of cytokines or costimulatory molecules as well as dendritic cells for the effective presentation of immunogenic peptides in the context of MHC and the activation of cellular immune responses (3).

Another promising strategy, the vaccination with anti-idiotypic antibodies (Ab2,³) is based on the immune network approach of Nils Jerne (4). According to this hypothesis, the variable antigen-binding regions of antibodies (Ab1) contain idiotypic determinants that are immunogenic and induce the formation of so-called anti-idiotypic antibodies (Ab2). Some of these Ab2 (“internal-image”) antibodies are able to functionally mimic the three-dimensional structure of the original antigen. Thus, selective immunization with Ab2 could induce a specific immune reaction directed against the original antigen (5, 6, 7, 8, 9).

Thus far, several clinical studies using murine Ab2 have demonstrated the induction of antitumoral humoral and cellular immune responses leading to improved clinical responses and tumor regression. These results, concerning patients with advanced colorectal carcinoma (7, 10, 11, 12, 13) and malignant melanoma (14), have been confirmed by our own studies in patients with advanced ovarian cancer (15, 16, 17, 18, 19). The Ab2 approach is especially suitable for ovarian cancer, because both the primary sequence and the structure of the tumor-associated antigen CA125 have not been described thus far (20).

Therefore, we have generated a monoclonal murine Ab2 called mAb ACA125 [anti-id OC125 (Ab1), IgG1k], which functionally imitates the CA125-antigen and induces humoral as well as cellular anti-CA125 immunity in vitro and in animal trials (15, 19, 21, 22, 23, 24, 25).

A clinical Phase I study was performed to evaluate dose, toxicity, and immunological competence of monoclonal anti-idiotype ACA125 for treatment of ovarian cancer patients (15). Eighteen patients with recurrent ovarian carcinoma received 3–24 injections of mAb ACA125, and 12 of 18 patients developed specific anti-anti-idiotypic antibodies (Ab3). Side effects were limited to local pain (WHO grade I) at vaccination sites and abdominal pain (WHO grade III) in one single case (15).

The present Phase I/II study was conducted to evaluate the effectiveness of anti-idiotype vaccination with ACA125 in consolidating the third-line therapy results of patients with platinum-pretreated recurrences. Documented parameters were both overall survival and the individually induced Ab3 titer as a marker for a specific immune response.

Forty-two patients with advanced/recurrent epithelial ovarian carcinoma were included in the present study. The average age of patients was 58.6 years. All patients had tumors that strongly expressed CA125. Staging and follow-up of the course of disease were conducted routinely by physical examinations, computer tomography of the abdomen, and chemical and hematological controls as well as by measurements of serum CA125 concentrations before every immunization with ACA125. Additionally, the induction of specific immune responses was monitored by detection of Ab3s directed against the anti-idiotype ACA125 (Ab3) and the nominal antigen CA125 (Ab1′) in selected patients’ sera.

All patients had received platinum-containing combination regimes for first-line therapy and had an average of two preceding chemotherapeutic regimes before antibody vaccination. In cases of clinically relevant progression under antibodytherapy, patients were treated with follow-up chemotherapy, although immunotherapy with mAb ACA125 was continued after completion of chemotherapy. Patients were immunized with a mean of 12.6 applications (range, 2–46) of monoclonal antibody ACA125 (Tables 1 and 2).

The following protocol was applied in this clinical trial. Treatment was started with four immunizations at 2-week intervals and then monthly application of ACA125. Patients were vaccinated at indicated time points with 2 mg of alum-precipitated complete anti-idiotypic antibody ACA125 injected deeply into the gluteal muscle. The 2-mg dose was chosen because it had resulted in effective immunization with few toxicities in a previous Phase I study (15).

Generation and production of anti-idiotypic antibody ACA125 has been described in detail elsewhere (21). Briefly, BALB/c mice were immunized with murine monoclonal antibody OC125 (CIS; Bio International, Gif-Sur-Yvette, France) directed against the tumor-associated antigen CA125. Splenic cells of immunized mice were fused with myeloma cells, yielding an ACA125-producing hybridoma cell clone that was adapted to serum-free medium. Large-scale production of mAb ACA125 was performed in a hollow-fiber system with yields of 10–15 mg of antibody per day. Mab ACA125 was purified from culture supernatant by affinity chromatography with protein G-Sepharose and the preparation was checked by SDS-PAGE analysis (PHAST-System; Pharmacia Biotech, Uppsala, Sweden) to give >95% purity. The final product for vaccination contained 2 mg of Ab2 IgG1 in sterile, pyrogen-, polynucleotide-, mycoplasm-, virus-, and retrovirus-free PBS-buffered solution. The lympho-cytotoxical properties as well as sterility, pyrogen content and general safety of this antibody were tested before starting the study.

Anti-allo- and anti-isotypic HAMAs in patients’ sera were determined by a commercially available ELISA (Medac; Hamburg, Germany). These unspecific antibodies the reflect general immunocompetence of the individual patient toward mAb ACA125.

Sera from immunized patients with positive HAMA responses were tested for specific Ab3s. To prevent from interference by HAMAs, these antibodies were eliminated by adsorption to mouse IgG agarose (Sigma, Deisenhofen, Germany) and then by centrifugation.

Purified supernatants were allowed to bind to ACA125 F(ab)₂ coated onto microtiter plates (1 μg/ml). Complete ACA125 (1 μg/ml) and subsequently goat antimouse IgG (Fc-specific, peroxidase-labeled; Dianova, Hamburg, Germany) was added. After appropriate washings, color development from 2,2-azino-di-[3-ethyl-benzthiozolinsulfonat(6) substrate was monitored at 405/492 nm.

As the assay was calibrated with OC125 F(ab)₂ (Ab1), Ab3 concentrations in patients’ sera are given in arbitrary units/ml (arb.U/ml) corresponding to 1 ng/ml of the Ab1 antibody.

Ab3 antibodies were purified from patients’ sera by affinity chromatography. Sera were preincubated with mouse IgG agarose for HAMA-elimination. Afterward, 1 ml of HAMA-eliminated serum was pipetted on a column consisting of ACA125 bound to N-hydroxy-succinimide-activated Sepharose (HiTrap; Pharmacia Biotech). Bound Ab3 were eluted from the column with glycin hydrochloride [0.05 m (pH 2.7)] and immediately neutralized to pH 7.0 with Tris hydrochloride [1 m (pH 8.0)]. The Ab3 concentration in the final volume of 1 ml (corresponding to 1 ml of patient sera) was quantified by absorbance at 280 nm.

IgG isotypes were determined in the hyperimmune sera of 10 patients at different stages of immunization (4th–35th immunization). Sera were purified (see above), and IgG subclasses were quantified by commercially available ELISA (PeliClass ELISA; CLB, Amsterdam, the Netherlands).

Hyperimmune sera of 16 patients who had developed Ab3 titers >100,000 arb.U/ml were purified (see above) and tested for binding to a CA125-positive (OVCAR57+) and a CA125-negative (OVCAR29−) human ovarian carcinoma cell line. Purified sera from patients before starting immunotherapy and at different stages of immunization were incubated with 2 × 10⁵ cells (1 h; 4°C). After washing FITC-conjugated rabbit antihuman IgG (DAKO, Hamburg, Germany) was added (30 min; 4°C). A negative control included cells that were incubated with FITC-conjugated secondary antibody alone. Flow cytometric (FACS) analysis was performed on FACS Calibur (Becton Dickinson, Heidelberg, Germany).

Additionally, the cells stained with patients’ antibodies and FITC-labeled secondary antibody were analyzed by fluorescence microscopy. To amplify the fluorescence signal, cells were incubated further with mouse anti-fluorescein antibody (Fluor-Amp Kit; Oncor) for 15 min at room temperature and then stained with antimouse IgG FITC-labeled antibody (Fluor-Amp Kit; Oncor) for 15 min at room temperature.

To verify the specificity of induced antibody responses, different competition formats were performed. Plates coated with ACA125 F(ab)₂ were incubated with mixtures of fixed amounts of purified patients’ sera and various dilutions of the inhibitor Ab1 OC125 F(ab)₂ or the CA125 antigen. Bound antibodies were detected with goat antihuman IgG (Fc-specific, peroxidase-labeled) as described above. The percentage of competition was calculated on the basis of the binding of the antibodies without competitor.

Inhibition by OC125 F(ab)₂ represents the antibodies directed against the antigen-binding region of ACA125, whereas the inhibition by CA125 indicates the existence of antibodies binding simultaneously to the anti-idiotype and to the original antigen.

Additionally, competition of binding of Ab1 OC125 to CA125-positive human ovarian carcinoma cell line by purified patients’ sera was measured by flow cytometric analysis. Briefly, mAb OC125 (1 μg/ml) was mixed with Ab3 purified from patients’ hyperimmune sera or equally treated preimmune sera (no detectable Ab3) and incubated with CA125-positive human ovarian carcinoma cells (OVCAR57+), as mentioned above. After washing, cells were stained with PE-labeled rabbit antimouse IgG (DAKO, Hamburg, Germany) and analyzed on FACS Calibur (Becton Dickinson). Competition was detected by decreased fluorescence intensity compared with OVCAR57+ cells stained with Ab1 OC125 alone. An irrelevant monoclonal mouse antibody 14C5 Ab1 (1 μg/ml; Ref. 26) binding to OVCAR57+ cells was used instead of Ab1 OC125 as a negative competition control.

The cytolytic activity of immunized patients’ PBLs against human ovarian cancer cells was examined by a standard europium-release cytotoxicity assay. Briefly, the CA125-expressing human ovarian carcinoma cell line (OAW-42) and one CA125-negative cell line (SK-OV3) were labeled with europium chloride by electroporation and used as target cells. PBLs from patients (preimmune and postimmune; 1 × 10⁷ cells/ml) were mixed with different numbers of target cells (at an effector:target cell ratio of 100:1 to 3:1) and incubated for 4 h at 37°C. The supernatant was collected and the amount of europium released from tumor cells was determined fluorometrically. The percentage of lysis was calculated as follows: % lysis = 100 × (experimental release – spontaneous release) / (maximum release – spontaneous release).

This Phase I/II study is the first to examine the properties of an Ab2 imitating the tumor-associated antigen CA125 in third-line therapy of a larger collective of patients with platinum-pretreated recurrences of ovarian cancer. The primary aim of this study was the evaluation of both the clinical effectiveness of ACA125 as represented by overall survival. Furthermore, the significance of ACA125-induced immunity, as indicated by anti-anti-idiotype response, as a prognostic marker for survival was to be assessed.

Severe negative effects of the vaccination were not observed during application or in the follow-up phase. Thus far, the antibody has been applied 530 times with a maximum of 46 injections in one individual. No serious allergic reactions could be detected within a control period of up to 56 months. In some cases, repeated immunizations with mAb ACA125 resulted in pain limited to the injection sites, with a maximum of discomfort after 48 h, accompanied by local infiltrates resembling a delayed-type hypersensitivity-response. Generally, the monthly application of mAb ACA125 was well tolerated by the patients.

The development of specific anti-anti-idiotypic Ab3 immune responses associated with ACA125 vaccination was measured in patients’ sera before every injection of the vaccine. Hyperimmune sera of 27 of 42 patients (64.2%) showed an increase of HAMAs (>100 ng/ml)) with a mean of 48.754 ng/ml (maximum, 1,000,000 ng/ml) during ACA125 treatment (Tables 1 and 2). Ab3s directed against the antigen-binding regions of mAb ACA125 (Ab3) were detected in 28 of 42 patients (66.7%), whereas Ab3 responses were negative before treatment in all patients (Tables 1 and 2). Two patients produced a specific Ab3 response (>1,000 arb.U/ml) without a measurable anti-isotype (HAMA) reaction (Table 1). In contrast, one patient with a positive HAMA response (>100 ng/ml) developed no specific Ab3-titers (Table 1). Apart from these three women, all other patients showed uniform iso- and idiotype reactivity. The IgG subclass composition of the Ab3 response was evaluated in selected patients (n = 10), and a predominant IgG1- and IgG2-type reaction could be observed (Table 3). In addition, no correlation between number of applications, maximum Ab3-reactivity, and serological CA125concentration could be shown.

The specificity of the immune responses induced after vaccination could be demonstrated by competition experiments showing that binding of patients’ Ab3s to the anti-idiotype ACA125 could be successfully inhibited by the idiotypic mAb OC125 and the CA125 antigen, respectively (Fig. 1). Furthermore, purified sera from 16 patients who developed Ab3-titers >100,000 arb.U/ml were tested for specific binding to a CA125-expressing human ovarian carcinoma cell line (OVCAR57+; Fig. 2). Ab3 preparations from 11 of 16 patients bound to CA125-positive cells as indicated by increased fluorescence intensity compared with negative control cells, whereas no reactivity with a CA125-negative ovarian carcinoma cell line was observed. Equally treated preimmune sera from the same patients showed no binding to CA125-expressing as well as CA125-negative ovarian carcinoma cells. Fig. 3, A–D show representative FACS analysis of one patient before and at different stages of anti-idiotype vaccination. During the course of immunization, the number of Ab3s with anti-CA125 reactivity (Ab1′) increases, which could be detected as a shift toward higher fluorescence intensities; whereas no binding of corresponding preimmune sera to CA125-positive cells was observed. Binding of unprocessed patients’ postimmune sera to CA125-positive cells could be successfully demonstrated in some cases, with one example (pre- and postimmune serum of patient 16) presented in Fig. 3, E–F. However, the use of unprocessed sera often resulted in high background staining of antigen-negative and -positive cells, which could be overcome by affinity-purification of sera.

Additionally, binding of Ab1 OC125 to CA125-positive cells could be partially inhibited by purified patients’ sera (Fig. 4, A–C). The degree of competition varied between individual patients, which could be explained by different amounts of Ab1′ present in the tested Ab3-preparations. No competition of OC125-binding to CA125-positive cells could be detected after addition of equally prepared preimmune sera (Fig. 4,D). To further verify the specificity of Ab3 antibodies induced by ACA125-vaccination, a control experiment with an irrelevant monoclonal mouse antibody (14C5 Ab1; no CA125-specificity) was performed. It could be shown that Ab3 from patients’ hyperimmune sera did not inhibit the binding of 14C5 Ab1 to CA125-expressing ovarian carcinoma cells (Fig. 4 E), indicating that the Ab3 antibodies specifically recognize the CA125-antigen as did the Ab1 OC125.

In the first 18 patients, cell-mediated cytotoxicity from PBLs against CA125-expressing (OAW-42) and nonexpressing (SK-OV3) human ovarian cancer cell lines was evaluated. PBL-mediated lysis of CA125-expressing tumor cells increased in 9 of 18 cases; following the vaccination with the anti-idiotype, measured cell-kill changed from 19.6 ± 11.7% to 52.7 ± 13.6% (minimum, 27%; maximum, 70%) at the effector:target cell ratio of 100:1 (Fig. 5). Lysis of non-CA125-expressing tumor cells was not influenced at 16.7 ± 12.7% after vaccination (Fig. 5).

Cell-mediated lysis was accompanied by the induction of Ab3 in eight of nine patients. The one Ab3-negative patient whose PBLs also induced cell-mediated lysis had received only three vaccinations before evaluation, whereas all others had undergone at least five courses. Because of the correlation of Ab3 induction and cell-mediated lysis, we continued to evaluate merely humoral responses in subsequent patients.

Forty-two patients received a mean of 12.6 applications of vaccine mAb ACA125 for induction of anti-CA125 immunity and consolidation of recurrent ovarian cancer. Before starting immunotherapy, the patients were treated with an average of 2.1 chemotherapeutic regimes (Refs. 1, 2, 3, 4, 5; Table 1). The course of immunization was followed up for 1 to a maximum of 56 months, depending on the individual patient. The survival of the whole ACA125-treated collective was 14.9 ± 12.9 months (Table 2). In the case of newly developing clinically relevant progression under ACA125-therapy, patients obtained additional chemotherapy, and immunization was continued after completion of chemotherapy [on average 1.1 (0–6) treatments]. Humoral immune responses, measurable before follow-up chemotherapy continued to be detectable during and after the chemotherapeutic treatment.

The evaluation of surrogate marker anti-anti-idiotype reactivity in patients indicated a specific anti-anti-idiotypic immune response in 28 of 42 patients. The survival of patients with a positive immune response was 19.9 ± 13.1 months in contrast with 5.3 ± 4.3 months in patients without detectable Ab3-titers (Table 2). The difference between both collectives (responder − nonresponder) was highly significant according to Kaplan-Meier lifetable analysis (P < 0.0001; Fig. 6). There were no significant differences between the two subgroups concerning age, Karnowsky Index, and number of chemotherapeutic regimens before the initiation of ACA125-vaccination (Table 1).

Earlier treatment consisted of a mean of 2.11 (1–5) previous therapies in the Ab3-positive group and 2.0 (1–4) in the Ab3-negative group. The distribution of complete and partial responses (CR/PR), no change (NC) and cases of progression (P) at the time of study entry showed a spectrum of 4 CRs, 5 PRs, 13 NCs, and 6 Ps in the Ab3 positive group and 1 PR, 9 NC, and 4 Ps in the Ab3-negative group (Table 1).

The mean number of ACA125-applications was higher in the Ab3-positive group (16.8 ± 10.8 applications; median, 14.5) as compared with Ab3-negative patients (4.2 ± 3.0, median: 3.0). However, the responder group already had developed detectable Ab3-titers after 4.4 ± 4.6 applications (median, 3.0; Table 2). Thus, the number of applications required for the induction of an immunological response is comparable with the total application number in the nonresponder group.

Clinical response to the chemotherapies after anti-idiotype vaccination proved to be NCs in most cases for both groups. Survival data were assessed from the beginning of anti-idiotype vaccination. Table 1 shows the individual status at entry to the study, total number of ACA125 applications and number of applications required for detectable Ab3-titers, survival from study entry (in days), HAMA concentration (prior/maximal), Ab3 responses (prior/maximal), number of previous therapies, and number of and best response to following cytotoxic regimens.

Despite the generally encouraging tendencies of improved survival in advanced ovarian carcinoma (with a mean of 38 months after paclitaxel-platinum combination therapy), recurrences still occurred in most of the patients. Although treatment may be successfully attempted with platinum-reinduction or other chemotherapeutic regimens (27), the limits of cytotoxic treatment are obvious in the therapy of platinum-resistant recurrences or of patients with primary progression under first-line chemotherapy. Therefore, a main goal of the current research is the development of new therapeutic strategies, e.g., in the field of tumor immunology, which might serve eventually for the consolidation of initial therapeutic responses.

In first pre-clinical and clinical Phase I studies, the ability of the monoclonal anti-idiotype ACA125 to provoke humoral as well as cellular immune responses against the tumor-associated antigen CA125 without serious side effects was demonstrated (15, 23, 28).

The present clinical study was conducted to assess whether this immunological effect positively influences the survival of patients with recurrent ovarian cancer after repeated chemotherapeutic treatment. In 66.7% of treated patients (28 of 42 patients) a specific immune response after vaccination with anti-idiotype ACA125 was induced. This rate of response is similar to the results obtained with other Ab2s, representing different antigenic structures for the treatment of melanoma or colorectal or bronchial carcinoma (6, 10, 11, 12, 13, 14).

Additionally, in the first 18 treated patients, PBL-mediated cytotoxicity against CA125-expressing and -nonexpressing human ovarian cancer cell lines was evaluated. In nine of these patients, an increased lysis of CA125-expressing tumor cells was observed. This was accomplished by anti-anti-idiotypic responses in eight of these nine patients, all of whom had undergone at least five ACA125-vaccinations.

A more detailed characterization of the induced polyclonal Ab3 response revealed the presence of Ab1-like antibodies with CA125-specifity in purified sera from 11 of 16 tested patients (Ab3-titer, >100,000 arb.U/ml). This could be demonstrated by competition of Ab3-binding to Ab2 ACA125 by CA125-antigen as well as by direct binding to a human CA125-expressing ovarian carcinoma cell line. Although binding to CA125-positive cells could also be detected in untreated postimmune sera in some of the tested patients, additional purification steps (HAMA elimination and Ab2-affinity purification of sera) as used in the present study has proven to be beneficial for several reasons. In some patients, the untreated pre- and postimmune sera gave a high background staining of CA125-negative as well as CA125-positive cells, which could be blocked only partially. This problem could be solved by testing affinity-purified serum preparations containing only those antibodies which arose during ACA125-immunization and no more interfering serum components.

Despite high Ab3-titers in the purified sera, 5 of 16 tested patients showed no measurable reactivity with CA125-positive cells. On the basis of this observation, it could be concluded that the determination of polyclonal Ab3 responses does not necessarily correlate with detectable anti-CA125 reactivity in some cases. However, it must be taken into consideration that induced anti-CA125 antibodies could have been bound to CA125expressing tumor cells or to circulating CA125 antigen so that the concentrations of free antibodies is below the detection level of our assay system. Because of these principal difficulties, the simple and reliable determination of the total amount of Ab3s seems to be more suitable as a surrogate marker for routine monitoring of specific immune response after vaccination with ACA125.

Whereas the mean survival of all ACA125-treated patients was 14.9 months, the discrimination between Ab3 positive (66.7%) and Ab3 negative (33.3%) patients resulted in a statistically significant higher mean survival of 19.9 months in the Ab3-positive group as compared with 5.3 months in the nonresponder group.

The improvement of survival of the Ab3-responder group cannot be explained exclusively by the higher number of ACA125-applications (16.8 ± 10.8) in contrast with a mean of 4.2 ± 3.0 immunizations (median, 3.0) in the Ab3-negative group. It has to be considered that the patients of the responder group developed measurable Ab3 responses already after 4.4 ± 4.6 applications (median, 3.0). Therefore, nonresponding patients received a sufficient number of applications to develop an immune response. Thus, the difference of survival might be influenced by the induced immune responses because of the vaccine ACA125.

Objective tumor response was detectable in only one patient, who showed a regression in different sites of intra-abdominal paracolic tumor masses and inguinal lymph node metastases (patient 5). Although the Karnowsky status and the number of previous therapies do not differ between both groups, and the rate of stable disease and progressive patients under previous chemotherapy in both collectives is comparable, previous complete responses (n = 4) were only seen in the Ab3-positive group, demonstrating an influence of tumor load on immune response and survival.

According to the present results, vaccination with a suitable Ab2 offers an effective way to induce specific immunity against a primarily nonimmunogenic tumor-associated antigen such as CA125, although cancer patients are immunologically compromised in general. Additionally, the anti-idiotype concept is safe with negligible side effects and a positive influence on the outcome of disease.

As the collective selected in this study contains patients with different preceding treatments for ovarian cancer as well as a high percentage of patients in progression who received anti-idiotype therapy as last-line treatment of their disease, a randomized Phase III trial for immunological consolidation after first-line therapy seems to be justified.