RNA-Mediated Immune Modulation Conjunction with Double-Stranded

Immunologic Control of Tumors by In Vivo Fc γ Receptor-Targeted Antigen Loading in Conjunction with Double-Stranded RNA-Mediated Immune Modulation This information is current as of October 2, 2021. Adrian Bot, Dan Smith, Bill Phillips, Simona Bot, Constantin Bona and Habib Zaghouani J Immunol 2006; 176:1363-1374; ; doi: 10.4049/jimmunol.176.3.1363

http://www.jimmunol.org/content/176/3/1363 Downloaded from

References This article cites 44 articles, 18 of which you can access for free at: http://www.jimmunol.org/content/176/3/1363.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on October 2, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Immunologic Control of Tumors by In Vivo Fc␥ Receptor-Targeted Antigen Loading in Conjunction with Double-Stranded RNA-Mediated Immune Modulation

Adrian Bot,1* Dan Smith,*† Bill Phillips,*† Simona Bot,* Constantin Bona,‡ and Habib Zaghouani§

Despite the expression of non-self or neo-epitopes, many tumors such as lymphoid malignancies or cancers induced by oncogenic viruses are able to gradually overcome the immune defense mechanisms and spread. Using a preclinical model of hematological malignancy, we show that Ig-associated idiotypic determinants are recognized by the immune system in a fashion that results in immune deviation, allowing tumor progression and establishment of metastases. Using gene-targeted mice, we show that anti- ؉ ؉ idiotypic MHC class I-restricted immunity is promoted by ITAM motif (ITAM )Fc␥R, but kept in check by ITIM motif (ITIM ) Downloaded from Fc␥RIIB-mediated mechanisms. In addition to interfering with the functionality of ITIM؉ Fc␥R, effective anti-idiotypic and antitumoral immunity can be achieved by Fc␥R-targeted delivery of epitope in conjunction with administration of stimulatory motifs such as dsRNA, correcting the ineffective response to idiotypic epitopes. The immune process initiated by Fc␥R-mediated targeting of epitope together with dsRNA, resulted in control of tumor growth, establishment of immune memory and protection ,against tumors bearing antigenic variants. In summary, targeted delivery of MHC class I-restricted epitopes via ITAM؉ Fc␥R in conjunction with use of TLR-binding immune stimulatory motifs such as dsRNA, overcomes suboptimal responses to idiotypic http://www.jimmunol.org/ determinants and may constitute a novel approach for the treatment of a broad range of malignancies. Finally, the results shed light on the mechanisms regulating the idiotypic network and managing the diversity associated with immune receptors. The Journal of Immunology, 2006, 176: 1363–1374.

large variety of cancers acquire mechanisms to circum- Previous reports show that malignancies caused by oncoviruses vent or avoid the immune effectors potentially capable are associated with continuous expression of non-self TAA (10– A of removing malignant cells, ranging from immune ig- 12), ruling out Ag loss as universal mechanism of immune escape. norance to deviation or active interference with immunity (1–3). Furthermore, despite the quasi-intact T cell repertoire, previous 2 by guest on October 2, 2021 Although many tumor-associated Ags (TAA) are of self-nature studies have suggested that in the case of TAA that are of non-self and thus relatively poor immunogens, a significant category of origin, a specific T cell response is present but inadequate, for cancers comprises tumor cells that, at various stages, express non- example dominated by T2 cells (13, 14), and unable to clear tu- (or neo-) self epitopes of viral origin (oncoviruses, EBV, human moral cells. Thus, immune deviation rather than Ag-loss or toler- papillomavirus, human T cell leukemia virus type-1), idiotypic ance, may be a causative factor for the lack of control of tumors markers (B or T cell lymphomas, myeloma, and certain leukemias) expressing non-self or Ags. or cryptic epitopes (products of alternate mRNA splicing). Most A particular case is represented by lymphoid malignancies that malignant B and T cell myelomas and lymphomas are monoclonal express neo-self Ags in the form of Id borne by TCR or Ig recep- and thus express Id on the Ig receptors or TCRs, stochastically tors. It is not clear to what extent such determinants are recognized generated during somatic rearrangement of the V, D, and J or what the nature of immune response is during the progression of genomic segments (4), with subsequent point mutations within ϩ complementarity-determining regions (CDRs) or addition of nu- Id lymphoid malignancies. For example, despite the neo-self na- cleotides. Id may represent B epitopes, Th or Tc epitopes, or both, ture of such determinants, the immune response against Id ex- and in certain conditions, are able to induce anti-idiotypic re- pressed by malignant cells may occur (15–17) but in a suboptimal sponses (5–9). fashion, in particular in later stage disease, resulting in a failure of the endogenous defense mechanisms to control the tumoral process. In support of this possibility, clinical studies conducted over *Alliance Pharmaceuticals and †MultiCell Immunotherapeutics, San Diego, CA the last decade using Id in combination with KLH or GM-CSF, or 92121; ‡Mount Sinai School of Medicine, New York, NY 11029; and §University of infusion of Id-pulsed immature dendritic cells (DC), showed that Missouri, Columbia, MO 65212 although numerous patients mounted anti-Id immunity (i.e., Ab, Received for publication June 1, 2005. Accepted for publication November 8, 2005. Th, and/or CTL), the magnitude was generally reduced, the im- The costs of publication of this article were defrayed in part by the payment of page mune profile dominated by T2 cells and the clinical impact rela- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. tively modest (17, 18), somewhat moderating the optimism gen- 1 Address correspondence and reprint requests to Dr. Adrian Bot at the current ad- erated by earlier data (19). In general, despite the acknowledged dress: Mannkind Corporation, 28903 North Avenue Paine, Valencia, CA 91355. capability of CTL immune responses to remove cells that express E-mail address: [email protected] new Ags and the documented circumstantial evidence on genera- 2 Abbreviations used in this paper: TAA, tumor-associated Ag; CDR, complementarity-determining region; DC, dendritic cell; NP, nucleoprotein; HA, hemagglutinin; tion of anti-Id cytotoxic immunity in mice (20, 21) and humans IC, immune complex. (22), it is not known whether neo-self MHC class I-restricted Id

generated stochastically can play a significant role in the immune- retained binding to Fc␥R, which is critical in effective targeting of APC mediated containment of malignant or nonmalignant lymphoid (23, 25). The recombinant Ig peptides were obtained by cell culture and proliferations. purified using affinity chromatography (23, 25, 28, 29). The strain of influenza virus used in this study was A/WSN/32 H1N1, From a different point of view, Igs have been used as carriers for with mouse as permissive host. The virus was grown on Madine Darby B and T cell epitopes, for the purpose of induction of prophylactic Bovine Kidney carcinoma cells and purified, and titers were measured by and therapeutic responses in preclinical models (23, 24). That conventional techniques (35). Synthetic dsRNA (pA:pU), previously char- method has been largely promoted by the fact that peptides have a acterized as a T1 adjuvant (36), was obtained from Sigma-Aldrich. SP2/0 myelomatous cells, of MHC class Iϩ of H-2d background, widely rather poor immune activity, due to their suboptimal pharmacoki- used as partners for cell fusions in generating B cell hybridomas were netic profile. The alternative, ex vivo peptide loading of APC, previously described (37). Tumorigenic cells expressing secreted or re- implies individualized therapies and thus, it is more difficult to tained Ags were obtained by transfection, selection, and subcloning of implement therapeutically on a large scale. It has been shown that SP2/0 cells with plasmids carrying H chain and L chain of IgNP, IgHA, the recombinant Ig, carrying an MHC class II-restricted epitope de- IgG2b backbone, or whole-length influenza virus NP (encompassing a nu- clear targeting motif) (38). Cells were characterized as previously derived from influenza virus within CDR3 of the H chain, effectively scribed and the production of Ag confirmed by standard immunochemistry targets DC via Fc␥R, resulting in enhanced presentation to Th cells techniques. The mouse breast carcinoma cell line 4T-1 of BALB/c back- as compared with nominal peptide (24, 25). In addition, recombi- ground were purchased from American Type Culture Collection, trans- nant Igs carrying B cell epitopes from the circumsporozoite or HIV fected with whole length NP, selected, subcloned, and characterized similarly. All cell lines were tested for tumorigenicity in BALB/c mice and Ags resulted in induction of neutralizing, anti-idiotypic responses upon retrieval were shown to retain Ag expression. Cells were cultured at (26, 27). Work using engineered recombinant Ig encompassing 37°C under 5% CO2, FCS-free HL-1 medium (BioWhittaker) supple- MHC class I-restricted epitopes showed that such epitopes may be mented with HEPES and antibiotics. processed by proteasome and presented via MHC class I by trans- Downloaded from Tumor challenge, monitoring, and immunotherapy fected cells to Tc (28). Overall, although these data support the concept that Id (or in general, Ig-associated Id) are being pro- The mice were challenged by s.c. injection into the lateral back area, with ϫ 5 cessed, presented, and recognized by immune cells, it is still un- 5 10 SP2/0 myeloma cells or 4T-1 breast carcinoma cells expressing TAA or TAAϪ control cells in 200 ␮l of DMEM. When the primary tumor clear whether Ig bearing class I-restricted epitopes can elicit a Tc became clinically detectable (day 12 or before if the tumor reached a vol- response. For example, transfectomas expressing recombinant Ig ume of 0.5 ml) the mice were randomized into treatment groups. The with a class I epitope from the nucleoprotein (NP) of influenza treatment regimen was promptly initiated and repeated twice every 5 days, http://www.jimmunol.org/ virus were capable of inducing, in certain conditions, a CTL re- consisting of s.c. homolateral injection of 50 ␮g of pA:pU, 50 ␮gofIg ␮ ␮ sponse, whereas the purified Ig was unable to do so (28, 29). It has peptides, or a combination of 50 g of pA:pU plus 50 g of Ig peptides. ␥ The evolution of the primary tumor, closely paralleling the clinical pro- been thus concluded that the Fc R-initiated processing pathway gression, was monitored on a regular basis and the size of the tumor mea- does not result in presentation of Id via MHC class I, due to a lack sured using a caliper (the volume of quasispherical tumors was estimated of intersection with the endogenous pathway of processing and using the formula 4/3␲((aϩb)/2)3, in which a and b are two measured presentation. More recent studies encompassing Ig carrying perpendicular diameters expressed in cubic centimeters. Mice displaying significant morbidity associated with terminal disease were euthanized. epitopes outside the idiotypic region (30) or immune complexes Cell-based immunization was conducted using ex vivo Ag-loaded DC (IC), suggested otherwise and in the latter case, showed a thera- ϩ ϫ 6 from animals having the same MHC haplotype. CD11c DC (2 10 by guest on October 2, 2021 peutic benefit in a preclinical tumor model as well as pinpointing cells/ml) purified by magnetic sorting according to manufacturer’s instruc- the role of various classes of Fc␥R in this process (31). Neverthe- tions (Miltenyi Biotec) were pulsed for 14 h with 50 ␮g/ml IgNP or an ␮ less, the practicality and potential safety profile of IC as therapeu- equivalent amount on a molar basis of NP peptide (1 g/ml), in the presence or absence of 100 ␮g/ml dsRNA (pA:pU). After washing, these Ag- tic agents is uncertain. loaded DC were injected into recipient mice at 5 ϫ 105 cells/mouse i.p. in In the present study, we systematically approach the question as 200 ␮l of DMEM. to whether idiotypic class I-restricted determinants are immuno- Measurement of T cell response genic, using the model of a dominant NP epitope from the influenza virus and leveraging the potential of certain TLR ligands to For ELISPOT analysis, spleens were harvested and cell suspensions pre- promote APC activation. Further, we delineate the parameters reg- pared using a standard technique involving removal of RBC by hypotonic lysis. Splenocytes were incubated at 5 ϫ 105 cells/well (and serial, two- ulating the magnitude and nature of the MHC class I-restricted fold dilutions) in “complete” HL-1 medium (containing 10% FBS) with 10 response against the foreign epitope inserted within an Ig, in ad- ␮g/ml peptide (NP or HA) or medium only, in anti-IFN-␥, anti-IL-2, or dition to the possibility of using such constructs to induce antitu- anti-IL-4 (BD Pharmingen) precoated plates (MAHA S4150; Millipore), moral responses capable of controlling malignant processes. subsequently blocked with BSA. The cells were incubated for 72 h at 37°C and under 5% CO2. The assay was developed by first washing the cells, then serial incubation with 2 ␮g/ml biotin-conjugated anti-cytokine Abs Materials and Methods (BD Pharmingen), followed by addition of streptavidin-conjugated HRP Mice (1/1000 v/v) and AEC (3-amino-9-ethyl-carbazole) insoluble substrate Ϫ Ϫ Ϫ Ϫ (Sigma-Aldrich). The data corresponding to the number and average size of BALB/c, Fc␥R / (32), and Fc␥RIIB / mice (33) on a BALB/c back- spot forming colonies were acquired using an automated system equipped ground, and C.B10 congenic mice bearing the H-2b MHC haplotype on a with a camera (Navitar) and assisted by ImagePro Plus software (Media BALB/c genetic background were purchased from The Jackson Laboratory Cybernetics). and maintained at Alliance Pharmaceutical according to Institutional An- Cytokine production was measured after ex vivo stimulation of spleno- imal Care and Use Committee and National Institutes of Health policies, in cytes, obtained as discussed, with 10 ␮g/ml peptide and in the presence of pathogen-specific free conditions. For all experiments, 8- to 12-wk-old 5 U/ml rIL-2 for 5 days. The concentration of IFN-␥ in the supernatants female mice were used. was measured by ELISA (BioSource International). The procedure was Reagents and cell lines repeated following another round of stimulation with mitomycin-treated feeder cells, in presence of peptide and rIL-2 (same concentration as dis- The dominant MHC class I (Kd)-restricted influenza virus NP epitope cussed), for 5 additional days. (147–155, TYQRTRALV) and the MHC class II (I-Ed)-restricted hemag- For infection, mice were mildly anesthetized with isoflurane and treated 3 glutinin (HA) epitope (110–120, SFERFEIFPKE) were previously de- by nasal instillation with a sublethal inoculum of 10 TCID50 (tissue cul- scribed and well characterized (25, 34). Recombinant idiotypic Igs, IgNP ture infectious dose 50%) of WSN virus. The cytotoxic assay was carried and IgHA, bearing the NP and HA epitopes, respectively, were engineered by using effector cells from mice immunized with influenza virus Ags, by replacing the H chain CDR3 segments of a mouse IgG mAb with an obtained following ex vivo stimulation with 10 ␮g/ml NP and 5 U/ml rIL-2 original specificity against the hapten arsonate (23, 28). This manipulation for 5 days. The effector cells were coincubated, at various ratios, with The Journal of Immunology 1365 peptide-coated or uncoated M12 (Kdϩ, B cell lymphoma) target cells for Results 5 h. Cell supernatants were tested for LDH release, an indicator of cellular dsRNA-mediated activation of APC promotes induction of IFN- damage, using a kit according to the manufacturer’s instructions (Cytotox ␥ 96, nonradioactive cytotoxicity assay kit; Promega). -producing T cells against an Ig-borne epitope ϩ Effective antiviral and antitumoral immunity is thought to require CD8 T cell separation, adoptive transfer, and FACS analysis generation of IFN-␥ producing T cells and MHC class I-restricted ϩ CD8 T cells were isolated by magnetic selection from spleens harvested cytotoxic cells. We studied the possibility of inducing MHC class from mice that underwent tumor rejection subsequent to immunotherapy. I-restricted T cell responses by targeted Ag delivery via Fc␥R The magnetic selection was conducted using beads coated with anti-CD8 mAbs according to the manufacturer’s instructions (Miltenyi) and the pu- expressed on professional APC. To circumvent the safety concerns rity of CD8ϩ T cells (Ͼ95%) confirmed by flow cytometry (FACS) using posed by polyvalent Ag Ab IC, we used a molecule comprising the a FACSCalibur instrument (BD Biosciences). Cells were resuspended at IgG2b backbone with a defined influenza virus-derived Kd-re- ϫ 6 ␮ 10 10 /100 l and infused i.v. into syngeneic mice before challenge with stricted epitope NP (147–155) (34) inserted within the CDR3 of tumor cells (100 ␮l/recipient). For the characterization of the degree of activation of tumor-infiltrating the H chain IgNP (28). lymphocytes, tumor masses were retrieved, collagenase digested, and cell We first assessed whether ex vivo epitope targeting of APC ϩ suspensions prepared. The expression of CD25 was tested in CD3 tumor- results in cross-processing and enables DC to trigger class I-re- infiltrating lymphocytes, by using two-color FACS analysis, with reagents stricted immunity. To this aim, CD11cϩ DC isolated from sec- from BD Pharmingen. ondary lymphoid organs were pulsed with IgNP or molar equiv- Statistical analysis alent amount of NP peptide, washed, and adoptively transferred C E Comparative analysis of the ELISPOT results was conducted by applying into naive BALB/c mice. As depicted in Fig. 1, and , IgNP or the t test, with values of p calculated accordingly. In addition, the log-rank NP peptide-pulsed APC elicited IL-2- and IL-4-producing NP Downloaded from test was used to analyze the tumor progression data. 147–155-reactive T cells, but no significant IFN-␥-producing T http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 1. MHC class I-restricted immunity in mice immunized with DC pulsed with recombinant IgNP and synthetic dsRNA. CD11cϩ DC were isolated from spleens of naive female BALB/c mice by magnetic sorting and pulsed overnight with equimolar amount of NP 147–155 peptide (NP) or recombinant IgNP. In some cases, the DC were copulsed with synthetic dsRNA (pA:pU). The DC were washed and adoptively transferred to syngeneic mice (5 ϫ 105 cells equally divided in two inoculi, delivered by s.c. and i.p. injections; n ϭ 3 recipient mice/group). Two weeks after the transfer, the T cell response to NP peptide was measured by ELISPOT analysis, using splenocytes incubated ex vivo with or without cognate peptide. The data were acquired and analyzed using an automated ELISPOT analysis system. The results were displayed as the number of cytokine producing spot forming colonies p Ͻ 0.05) relative ,ء p Ͻ 0.01 and ,ءء) SFC) in spleen (A, C, and E), and error bars represent mean Ϯ SEM (n ϭ 3 mice/group). Significant difference) to naive control group, as assessed by the t test. In addition, the average size (area) of cytokine producing colonies is depicted, proportional to the secondary expansion of specific T cell populations, with or without (Nil) stimulation with NP peptide (B, D, and F) expressed in relative arbitrary units (AU). 1366 IMMUNOTHERAPY WITH RECOMBINANT Ig AND dsRNA cell immunity. We hypothesized that the activation state of the DC Strategies that allow effective and safe in vivo loading of APC may have been limiting their capability in inducing IFN-␥-produc- circumvent the need for cumbersome ex vivo cell manipulation. ing Tc subsequent to processing of IgNP. To address this question, Thus, we next tested the magnitude and profile of T cell response we took advantage of the previous observation that synthetic resulting from in vivo targeted delivery of the class I-restricted NP dsRNA activate professional APC, resulting in rapid induction of epitope, by injection of recombinant IgNP into naive Kdϩ BALB/c IL-12 and TNF-␣ (36). Copulsing of DC with IgNP and pA:pU mice. Consistent with ex vivo pulsing data, administration of IgNP promoted substantial generation of IFN-␥-producing, NP 147– to naive mice resulted in the generation of peptide-specific IL-4- 155-specific T cells upon adoptive transfer into naive BALB/c and IL-2-producing T cells, with a very limited IFN-␥ component mice (Fig. 1A), with decreased induction of IL-4-producing T cells (Fig. 2A). Coadministration of IgNP and pA:pU greatly enhanced relative to DC pulsed with IgNP alone (Fig. 1E). In addition, the generation of IFN-␥- and IL-2-producing NP 147–155-specific Fc␥R-mediated delivery of NP via recombinant Ig, in conjunction T cells with a limited impact; however, on IL-4-producing cells with DC activation by pA:pU, afforded an increased expansion of (Fig. 2A). In contrast, the administration of IgNP together with, or IFN-␥-producing, specific T cells upon peptide restimulation, re- without pA:pU to CB10 congenic mice sharing the H-2b MHC flected into a larger size of spot forming colonies (Fig. 1B). As haplotype with BALB/c, failed to induce NP 147–155 peptide spe- expected, pulsing of DC with pA:pU alone or transfer of DC in- cific immunity (data not shown). cubated only with media, did not result in generation of NP 147– The IFN-␥-producing T cells, triggered by IgNP administration 155-specific T cell responses. together with pA:pU, maintained their ability to produce IFN-␥ Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 2. Tc1 and cytotoxic immunity in mice immunized with recombinant IgNP bearing a class I-restricted Id. A, Naive female BALB/c mice were injected s.c. with IgNP alone or in combination with synthetic dsRNA (pA:pU). The T cell response against class I-restricted NP peptide was measured 2 wk later by ELISPOT analysis. The results were represented as number of cytokine-producing spot forming colonies (SFC)/spleen (average Ϯ SE, n ϭ 4 mice/group) after background subtraction. Internal control data (inset), representing average SFC/spleen for each cytokine, correspond to ex vivo p Ͻ 0.05 relative to mice immunized with IgNP alone, as assessed by t test. B, The ,ءء .(incubation with culture medium alone (y-axis is logarithmic capability of NP-specific T cells to produce IFN-␥ was assessed by repeated ex vivo stimulation with NP plus rIL-2, using splenocytes from animals immunized with IgNP, IgNP ϩ pA:pU, or from naive mice. The concentration of IFN-␥, subsequent to first round (Ⅺ) and second round (f) of peptide stimulation, was measured by ELISA (results expressed in picogram per milliliter as mean Ϯ SEM of quadruplicate samples. C, To assess induction of CTL, mice were primed with IgNP alone or in combination with pA:pU and challenged 14 days later with infectious influenza virus, strain A/WSN/32 H1N1. Shortly after infection (4 days), the mice were sacrificed and the NP peptide-specific T cell response measured by ELISPOT analysis and cytotoxicity. The results were expressed as the number of IFN-␥-producing spot forming colonies in spleen along with the percentage of specific lysis at an E:T ratio of 50:1, after the subtraction of background against uncoated target cells (mean Ϯ SEM; n ϭ 4 mice/group). Control mice were not primed, but infected with influenza virus 4 days before sacrifice. The Journal of Immunology 1367

upon repeated ex vivo stimulation with peptide and rIL-2, while in cells. The s.c. inoculation of SP2/0-IgNP tumor cells into BALB/c the presence of syngeneic APC (Fig. 2B). In contrast, the T cells mice resulted in progressive development of a primary tumor, me- induced by IgNP alone, failed to acquire the in vitro ability to tastasis, and death, similar to injection of nontransfected SP2/0 produce IFN-␥ when treated similarly. Nevertheless, NP 147–155- (Fig. 3A). We assessed whether immunotherapy of tumor-bearing specific T cells primed in vivo with IgNP rapidly (within 4 days) mice with IgNP with or without pA:pU had any effect on evolution acquired the capability to produce IFN-␥ upon infection with in- of tumor process. Mice displaying primary tumors of at least 0.5 cc fluenza virus (Fig. 2C). The highest secondary expansion of IFN- were randomized into different treatment groups and received a ␥ -producing T cells was measured shortly after infection of mice series (5-day interval starting on day 12 or whenever tumor size primed with a combination of IgNP and pA:pU (Fig. 2C). This was reached 0.5 cc) of homolateral s.c. injections of 50 ␮g of pA:pU mirrored by more effective, early expansion of specific CTL in (Fig. 3B), 50 ␮g of IgNP (Fig. 3C), or 50 ␮g of IgNP plus 50 ␮g these mice (Fig. 2C) together with significantly decreased pulmo- of pA:pU (Fig. 3D). Separate injection of IgNP or pA:pU resulted nary virus titers (5-fold), showing that the generation of CTL ef- in a slight, nonstatistically significant slow-down in the growth of fectors through cross-priming is regulated by the degree of APC the primary tumor, with inexorable progression to serious morbid- activation. ity followed by death. Tumor remission in naive, pA:pU or IgNP- Thus, dsRNA stimulation of APC was key to the induction of treated mice was extremely rare (ϳ10% or less). In contrast, the IFN-␥-producing T cell immunity against an MHC class I-recombination of IgNP and pA:pU resulted in significant control of stricted epitope, incorporated within a recombinant Ig. primary tumor growth, induction of complete tumor regression, dsRNA-mediated cross presentation of recombinant Ig induces and prevention of serious morbidity and mortality in 60% of IFN-␥-producing T cells effective against tumor growth treated mice (Fig. 3D). The rest of the IgNP plus pA:pU treated Downloaded from mice displayed nonprogressing tumor disease with only ϳ20% We next tested whether induction of MHC class I-restricted T cells by in vivo targeted delivery of a model tumor-associated epitope to progressing to advanced disease resulting in moribund status. A Fc␥Rϩ APC can control a tumor process in a preclinical model. similar trend was noted when the animals were treated with 10- ␮ The tumor model used was based on the observation that s.c. fold lower doses of IgNP and pA:pU (5 g), although most of the inoculation of BALB/c mice with Kdϩ SP2/0 myelomatous cells animals displayed stable primary tumor size rather than complete results in development of large primary tumor mass, progressing to regression, consistent with a dose-effect relationship (data not http://www.jimmunol.org/ metastasis to the major internal organs (liver, spleen), and signif- shown). icant mortality within 4 wk followed by death within 6–8 wk, To assess whether the beneficial effect of IgNP plus pA:pU ex- concordant with the previously reported tumorigenic potential of tends beyond myelomatous to epithelial carcinoma cells that ex- hybridomas derived from SP2/0 cells (39). SP2/0 cells stably trans- press, rather than secrete, the TAA, we used the same immuno- ϩ fected with a plasmid expressing IgNP were previously shown to therapy protocol in mice inoculated s.c. with Kd breast cancer process and present the MHC class I-restricted NP 147–155 cells (4T-1) permanently transfected with a plasmid expressing the epitope in context of Kd (28, 29) and were used as model tumor whole NP of the influenza virus. As shown in Fig. 3E, injection of by guest on October 2, 2021

FIGURE 3. Effect of immunotherapy with a recombinant idiotypic Ig (IgNP) and synthetic dsRNA (pA: pU) in mice bearing tumors. Female BALB/c mice were challenged s.c. with syngeneic SP2/0 tumor cells or 4T-1 breast carcinoma cells expressing NP (SP2/0- IgNP) (A–D) or 4T-1-NP (E and F). When the primary tumor became clinically detectable (day 12 or before if the tumor reached a volume of 0.5 cc) the mice were randomized into treatment groups (n ϭ 10 mice/group). The treatment consisted in s.c. homolateral injection of pA:pU (B), IgNP (C), or a combination of pA:pU ϩ IgNP (D and F) every 5 days, for a total of three times. Mice inoculated with tumor cells but untreated with IgNP or pA:pU were used as controls (A and E). The evolution of the primary tumor, closely paralleling the clinical status, was monitored and represented as mean Ϯ SEM of the volume (cubic centimeter). D,We represented the tumor evolution of the entire group (ࡗ) vs the subgroup that recovered completely from tumoral disease, which represented 60% of the animals (Ⅺ). Sig- -compared with untreated con (ءء) niﬁcant difference trols was assessed by log-rank test. 1368 IMMUNOTHERAPY WITH RECOMBINANT Ig AND dsRNA

BALB/c mice with 4T-1-NP breast cancer cells resulted in pro- APC, in vivo, in the absence of immunotherapy, illustrating the gressive growth of the primary tumor, followed by relative stabi- impact of cross-priming in this case. In contrast, the mice that lization of the disease 3–4 wk after inoculation. When immuno- underwent tumor rejection showed peptide-specific T cells com- therapy was initiated in mice displaying clinical tumors, the prising significant IFN-␥- and IL-2-producing subsets, along with progression of the primary tumor was significantly curbed, with IL-4 secreting cells (Fig. 4). Characterization by FACS analysis of substantial amelioration of the overall clinical status over a period tumor-infiltrating lymphocyte revealed a local infiltration with ␥␦ of 4 wk or more (Fig. 3F). No benefit in terms of tumor control T cells (20–25%) and ␣␤ T cells (45–50%, with equal distribution was observed when the immunotherapy was conducted in non-NP- between the CD4ϩ and CD8ϩ T cell subsets), irrespective of expressing 4T-1 or SP2/0 tumor-bearing mice (data not shown). whether the mice were treated. In addition, the treated mice dis- Together, these data indicate that in vivo Fc␥R-mediated Ag load- played higher frequency of IL-2R␣ϩ lymphocytes within tumors ing along with activation of APC may be used to control tumor (on average, 5.9% in treated vs 0.8% in nontreated mice). In con- processes of diverse cell lineage origin. trast to the mice treated with the combination of IgNP and pA:pU, Overall, these results suggest that the ineffective control of mice treated with IgNP or pA:pU separately and unable to reject MHC class I epitope-expressing tumors may be due to reduced the tumor, presented an IL-4-biased profile, largely similar to non- access of Ag to APC (immune ignorance) and/or to a limited de- treated mice (data not shown). Thus, in this model, immunotherapy gree of APC activation resulting in immune deviation or lack of with Fc␥R-targeted epitope and APC activator (pA:pU) enabled differentiation to effector cells. To study this aspect, we measured the induction of Tc1 immunity and tumor rejection. the T cell reactivity to NP 147–155 peptide in spleens of nontreated mice that failed to control the tumor process and mice undergoing remission subsequent to immunotherapy with IgNP Protection of mice that overcame tumor growth upon treatment Downloaded from plus pA:pU. Interestingly, nontreated mice inoculated with SP2/ with recombinant Ig and dsRNA against subsequent tumor cell 0-IgNP cells developed a specific response confined to Tc2 cells challenge (Fig. 4), similar to treated mice that failed to resolve the tumor. Induction of persisting immunity, with capability to counteract This development shows that TAA is transferred from tumor to mechanisms of immune escape (such as Ag-loss) is of paramount importance to successful immunotherapy. Thus, we next tested whether the immune response subsequent to immunotherapy with http://www.jimmunol.org/ IgNP plus pA:pU that elicited tumor rejection, conferred protection against subsequent challenge with homologous tumor. Mice that underwent remission were challenged with a tumorigenic dose of SP2/0-IgNP cells and followed for 1 mo. As shown in Fig. 5A, the challenged mice were completely protected against secondary challenge, whereas the control mice developed tumors with a 100% rate. Interestingly, mice that underwent immunotherapy and recovered from the primary SP2/0-IgNP tumor were completely by guest on October 2, 2021 protected against the challenge with SP2/0 myeloma cells expressing Ag variants (whole NP, IgHA encompassing an MHC class II-restricted influenza virus HA epitope, or IgG2b backbone) (Fig. 5, B, D, and E). In addition, such animals that recovered from primary tumor disease were protected against “loss-of-Ag” variant of the original tumor (untransfected SP2/0 myeloma cells), indi- cating that the repertoire of antitumoral T cells expanded during the process of tumor rejection gradually involving T cells specific for SP2/0 myeloma-associated epitopes in addition to the model TAA (NP). Ex vivo cultured T cells from animals resistant against multiple tumor variants display a long lasting, increased production of IFN-␥, IL-2, and IL-4 concordant with expanded frequency of antitumor effector cells, but hampering the efforts to define a specificity pattern (data not shown). Nevertheless, the mice resistant to SP2/0 myeloma antigenic variants, lacked protection against the Kdϩ epithelial carcinoma cell line 4T-1 (Fig. 5F). Thus, there was no statistically significant difference between the evolution of 4T-1 carcinoma in naive vs mice recovered from SP2/0 tumor subsequent to immunotherapy. To evaluate the role of CD8ϩ T cells in this process of acquired broad immunity against FIGURE 4. The cytokine profile of NP-specific, MHC class I-restricted SP2/0 Ags, adoptive transfer experiments were performed in mice T cells in mice that underwent tumor rejection subsequent to immunother- challenged with SP2/0 or SP2/0-IgNP cells. In both cases, CD8ϩ apy with IgNP and pA:pU. Female BALB/c mice were challenged s.c. with T cells transferred from mice recovering from SP2/0-IgNP tumor syngeneic SP2/0-IgNP tumor cells and subsequently underwent immuno- upon immunotherapy with IgNP plus pA:pU, negatively interfered therapy with IgNP and pA:pU as described in Materials and Methods. Five with the progression of tumoral process in recipient mice (Fig. weeks after tumor cell challenge, the T cell reactivity to NP 147–155 peptide was measured by ELISPOT analysis, using splenocytes from success- 5G). fully treated mice, or untreated tumor-bearing control mice (A–C). The Together, these data indicate that the expansion of the antitumor results are expressed as mean Ϯ SEM of spot forming colonies (SFC)/ T cell repertoire was limited to SP2/0-associated Ags, to which the spleen (n ϭ 4 mice/group) following ex vivo peptide stimulation (f)or immune system was exposed during the rejection process initiated incubation with culture medium only (Nil) (Ⅺ). by immunotherapy. The Journal of Immunology 1369

FIGURE 5. Protection of mice against secondary challenge with Ag variant tumor cells. Female BALB/c mice bearing SP2/0-IgNP tumors were treated with IgNP plus pA:pU and subsequently challenged with homologous tumor cells (n ϭ 4 mice/group) (A), SP2/0 cells expressing native NP (B), Ag-loss variant SP2/0 (C), SP2/0-IgHA (D), SP2/ 0-IgG2b (E), or breast cancer cells 4T-1 (F). The size of the tumors (cubic centimeters) expressed over time, as mean Ϯ SEM. Mice that underwent previous immune therapy and tumor rejection (closed symbols) and control mice (open symbols) are shown. G, Depicts the effect of the adoptive transfer of CD8ϩ T cells separated from spleens of mice that underwent tumor rejection on tumor growth in BALB/c recipients challenged with

SP2/0 (interrupted line) or SP2/0-IgNP Downloaded from tumor cells (continuous line). Mice challenged with tumor cells but not infused with CD8ϩ T cells were used as controls (open symbols). The results were expressed for tumor size as mean Ϯ SEM (n ϭ 4 mice/group) over time. Signiﬁcant /compared with untreated http://www.jimmunol.org (ءء) difference controls is assessed by log-rank test.

Contrasting roles of ITAMϩ Fc␥R and ITIMϩ Fc␥RIIB in tumor (Fig. 6). In contrast, in the absence of functional ITIMϩ controlling the antitumor T cell immunity elicited by Fc␥RIIB, the treated mice mounted significantly stronger NP pep- recombinant Ig and dsRNA tide-specific T cell responses along with effective control of the Various Fc␥R isoforms may be differentially involved in the re- tumoral process. In fact, the expansion of cytokine-producing (in by guest on October 2, 2021 sponse to IgNP plus pA:pU, Fc␥R plus Fc␥RI/RIII bear activating particular IFN-␥) T cells recognizing the MHC class I-restricted Ϫ Ϫ ITAM motifs and Fc␥RIIB carries inhibitory ITIM motifs. For NP peptide was more substantial in treated Fc␥RIIB / mice as example, DC express Fc␥RI and Fc␥RIIB receptors and it is not compared with wild-type mice that underwent similar immuno- clear how they participate following the response initiated by therapy (Figs. 6 and 7). The ELISPOT analysis showed that in Fc␥R targeted delivery of an MHC class I-restricted peptide. To addition to the frequency, the secondary clonotypic expansion of address this issue, we induced tumors by inoculating SP2/0-IgNP NP 147–155-specific T cells were substantially and differentially Ϫ Ϫ Ϫ Ϫ myeloma cells into Fc␥R / or Fc␥RIIB / mice, along with influenced by Fc␥RIIB and Fc␥Rϩ receptors (Fig. 7). Fc␥R competent BALB/c mice. The mice underwent a similar To test whether ITAM and ITIMϩ Fc␥R are similarly involved treatment (5-day interval starting with day 12 or whenever tumor in the regulation of the immune response to MHC class II-re- size reached 0.5 cc) with IgNP plus pA:pU, as in the previous stricted epitopes borne by CDR of IgG, we measured the immune experiments. At 4 wk after tumor initiation, disease evolution in response to APC loaded with IgHA using a protocol similar to that ␥ Ϫ/Ϫ ␥ treated Fc RIIB mice was similar to that of treated Fc R-com- described for Fig. 1. As shown in Fig. 8, the Fc␥R expression petent mice, with substantial and statistically significant control of profile on APC influenced profoundly the magnitude of peptide- tumor growth and prevention of morbidity, as opposed to non- specific immune response subsequent to adoptive transfer of ex treated control wild-type mice (Fig. 6A). In stark contrast, mice ϩ vivo IgHA-pulsed APC into naive BALB/c mice. In the absence of deficient in ITAM Fc␥R that underwent treatment with IgNP ϩ ITIM Fc␥R expression, the resulting HA-specific response was plus pA:pU failed to control the tumor growth, similar to non- ␥ Ϫ/Ϫ significantly enhanced, mirroring the pattern observed in the case treated wild-type or Fc R mice (Fig. 6A). Even in the absence Ϫ/Ϫ Ϫ Ϫ of IgNP. Thus, the Fc␥RIIB APC loaded with IgHA and co- of treatment, there was a trend of Fc␥RIIB / mice to control the Ϫ Ϫ treated with pA:pU triggered overall enhancement of both Th2 and tumoral growth, in contrast to wild-type or Fc␥R / mice. That ␥ was complemented by increased frequency of IL-2ϩ NP-specific T IFN- -producing Th1 cells (Fig. 8). In contrast, the magnitude of ␥ Ϫ/Ϫ the immune response was diminished considerably in the absence cells in untreated Fc RIIB mice carrying SP2/0-IgNP tumors ϩ (Fig. 6C). of functional ITAM Fc receptors, or when the Ag was HA pep- Measurement of the magnitude and profile of the NP 147–155- tide alone rather than IgHA. ϩ specific T cell response yielded significant differences between the Thus, the expression of ITAM Fc␥R subunit (essential com- tumor-bearing mice expressing competent or defective Fc␥R genes ponent of the Fc␥RI and Fc␥RIII) is essential for the Fc␥R tar- upon immunotherapy with IgNP ϩ pA:pU. In the absence of geted immunotherapy to be effective in this model of MHC class ϩ ITAMϩ Fc␥R, the immunotherapy failed to trigger substantial ex- I-restricted TAA. In addition, these results show that ITAM and pansion of cytokine producing T cells (except, to a certain extent, ITIMϩ Fc␥R regulate, in opposing fashion, the T cell response to IL-4ϩ T cells) along with inability to suppress the growth of the class I- and class II-restricted epitope borne by CDR. 1370 IMMUNOTHERAPY WITH RECOMBINANT Ig AND dsRNA Downloaded from http://www.jimmunol.org/

FIGURE 6. Differential role of ITAMϩ and ITIMϩ Fc␥R in the control of tumor rejection by immunotherapy with recombinant IgNP and synthetic dsRNA. Fc␥RϪ/Ϫ,Fc␥RIIBϪ/Ϫ, and wild-type (wt) BALB/c mice were challenged s.c. with syngeneic SP2/0-IgNP tumor cells and subsequently underwent immunotherapy with IgNP and pA:pU as described in Materials and Methods. The evolution of the tumoral process was monitored and 4 wk later, the mice were sacrificed and the T cell response quantified by ELISPOT analysis. The evolution of the tumoral process is represented comparatively (A), as the primary tumor volume at 4 wk after challenge (cubic centimeters) with mean Ϯ SEM (n ϭ 8 mice/group). Mice of similar genetic background, not treated with IgNP and pA:pU, were used as controls. B–D, The results of the ELISPOT analysis subsequent to NP peptide stimulation (f) or incubation with cell by guest on October 2, 2021 culture medium alone (Ⅺ) are shown in spot forming colonies (SFC)/spleen at mean Ϯ SEM (n ϭ 8 mice/group): IFN-␥ (B), IL-2 (C), and IL-4 (D). compared with values corresponding to untreated (ءء) compared with values corresponding to wild-type mice; significantly lower (ء) Significantly higher controls of the same genetic background (A) or wild-type mice (B–D).

Discussion Tc1/Tc2 denomination to illustrate IFN-␥/IL-2- and IL-4-produc- The current study was undertaken to evaluate the potential thera- ing T cells, respectively). In contrast, coactivation of DC by pA:pU IFN-␥ peutic effect of recombinant IgG carrying an idiotypic, class I-re- enabled induction of Tc without interfering with the induction stricted epitope for Ag-based immunotherapy of tumors. In addi- of TcIL-2, but resulting in slight diminution of the TcIL-4 population tion, the present work addressed the mechanisms regulating the (Fig. 1). This profile was essentially reproduced by direct immu- response against idiotypic T cell epitopes, in general. nization with IgNP (Fig. 2): in contrast to IgNP alone, coadmin- ␥ Using models based on transplantable TAAϩ tumors of lym- istration of pA:pU promoted induction of TcIFN- cells specific for phoid and carcinoma origin, it was shown that recombinant IgNP the Id NP, with a stable cytokine profile (Fig. 2B), shifting the (carrying a class I-restricted NP determinant) (28) induced a strong T1/T2 balance in favor of the former (Fig. 2A, inset). Interestingly, immune response mirrored by a therapeutic effect (Fig. 3). How- copriming with IgNP and pA:pU enabled a more rapid expansion ␥ ever, this effect occurred only if IgNP was coadministered in con- of NP-specific TcIFN- and acquisition of cytolytic function shortly junction with a potent activator of APC such as synthetic dsRNA after infection with influenza virus (Fig. 2C). In addition, mice (pA:pU) (38), resulting in a majority of mice recovering from the bearing SP2/0-IgNP tumors and undergoing tumor rejection sub- disease in the case of the SP2/0 myeloma tumor or with a stabi- sequent to immunotherapy with IgNP plus pA:pU displayed in- lized disease in the case of the 4T-1 carcinoma (Fig. 3). No sig- creased numbers of NP-specific Tc-producing IFN-␥ and IL-2, nificant therapeutic effect was provided by either IgNP or dsRNA compared with tumor-bearing untreated mice (Fig. 4). In contrast, alone (Fig. 3); in addition, there was no effect of IgNP plus pA:pU untreated mice (Fig. 4) or those failing to control the tumoral pro- on NP-negative SP2/0 tumor (data not shown). Measurement of cess subsequent to immunotherapy showed a predominance of immune response elicited by IgNP-pulsed DC in BALB/c mice TcIL-4 specific for the NP epitope. Together, these data show that showed that, although the recombinant Ig was more immunogenic coactivation of APC using a ligand for a TLR enables an optimal than a molar equivalent amount of NP peptide (Fig. 1), the re- immunotherapeutic effect of the recombinant idiotypic Ig, result- sponse was dominated by IL-4- and IL-2-producing T cells (for ing in differentiation of naive Tc to a stage encompassing TcIFN-␥ simplicity and with the caveat that the functional profile of MHC with effector function relative to the tumoral process expressing class I-restricted T cells is quite heterogenous, we will use the secreted or retained Ag. Finally, in the absence of APC activation The Journal of Immunology 1371

FIGURE 7. Cytokine profile of T cells from tumor- bearing Fc␥RϪ/Ϫ,Fc␥RIIBϪ/Ϫ, and wild-type BALB/c mice subsequent to immunotherapy with IgNP and pA: pU. Mice were challenged s.c. with syngeneic SP2/0- IgNP tumor cells and subsequently underwent immunotherapy with IgNP and pA:pU as described in Fig. 6 and in Materials and Methods. Four weeks later, the mice were sacrificed and the T cell response quantified by ELISPOT analysis. Representative pictures of wells corresponding to splenocytes ex vivo incubated with NP peptide or cell culture only (Nil) are shown, along with control wells (a–f) corresponding to splenocytes from tumor cell challenged, untreated wild-type mice (Fc␥R competent).

by TLR ligand, the MHC class I-restricted Id is still recognized, mune escape mechanisms deployed by genetically unstable tumor but the response is of limited magnitude and dominated by Tc2 cells. In addition, beyond effects on primary tumor, a successful Downloaded from (TcIL-4) and unable to mediate in vivo clearance of IgNP or NP- immunotherapeutic strategy should mobilize immune effectors that expressing cells. have the capability to mediate a body-wide immune surveillance In light of the fact that these tumor cells produce immunogenic and curb metastatic disease. To address these questions, we tested Ig, one may have expected that pA:pU alone would have resulted whether mice that recovered from SP2/0 tumoral disease subse- in protective immunity against the tumor process. To address this quent to immunotherapy can deal with further tumorigenic chal- question, mice challenged with tumor cells were injected with

lenges comprising TAA variants (such as Ag-loss mutants). The http://www.jimmunol.org/ pA:pU alone. As shown in Fig. 3B, there was no significant ben- striking, complete protection against subsequent ectopic challenge eficial effect on the tumor progression, conferred by pA:pU alone. with antigenic variants of SP2/0 but not a different tumor cell line Our interpretation was that the tumor process alone does not en- (Fig. 5, A–F), demonstrates that secondary immunity against SP2/0 sure optimal presentation of tumor-associated Ag in this system, TAA determinants occurred in an effective fashion, in the animals despite the production of IgNP by tumor cells. This was further recovering from primary tumor due to treatment with IgNP plus supported by the modest NP-specific response in nontreated, tu- pA:pU. Despite the fact that we cannot rule out at this point the mor-bearing mice similar to the modest Tc2 response in treated involvement of additional immune effector mechanisms, adoptive mice that failed to control the tumoral process (Fig. 4). Finally, transfer experiments strongly suggested a role for CD8ϩ T cells

based on the clear correlation between induction of T1 response by guest on October 2, 2021 and tumor regression (Fig. 4), we inferred that there are two lim- recognizing additional MHC class I-restricted, SP2/0-derived iting factors related to the antitumoral response in this system: first, epitopes (Fig. 5G). Overall, these data show that during the im- the suboptimal Ag processing/presentation of endogenous tumor mune effector process elicited by immunotherapy with the recom- Ag and secondly, the status of innate immunity. Thus, to ensure a binant idiotypic Ig (Fig. 4D), a significant process of epitope significant impact on tumor progression, an immunotherapeutic spreading from the NP-determinant to additional epitopes borne by strategy must address both. the tumor cells occurred, resulting in lasting protection against A crucial parameter of any immunotherapeutic strategy is to ectopic tumors comprising antigenic variants. initiate pleiotropic effector mechanisms and/or induce immune In addition to the level of APC activation determined by ex- cells recognizing an increased number of TAAs, counteracting im- posure to a TLR ligand, the response to the class I-restricted

FIGURE 8. Induction of an MHC class II-restricted peptide (HA) by DC pulsed with recombinant IgHA and costimulated with synthetic dsRNA. CD11cϩ DC were isolated from spleens of Fc␥RϪ/Ϫ,Fc␥RIIBϪ/Ϫ, and wild-type BALB/c mice by magnetic sorting and pulsed overnight with equimolar amount of recombinant IgHA (A) or HA 110–120 peptide (HA) (B). In addition, the DC were copulsed with synthetic dsRNA (pA:pU), washed, and adoptively transferred to syngeneic mice (5 ϫ 105 cells equally divided into two inoculi, delivered by s.c. and i.p. injections (n ϭ 3 recipient mice/group). Two weeks after the transfer, the T cell response to HA peptide was measured by ELISPOT analysis, using splenocytes incubated ex vivo with or without cognate peptide. The data were acquired and analyzed automatically. The results were displayed as the number of cytokine-producing spot forming colonies (SFC)/spleen (mean Ϯ SEM) for the three cytokines measured (IFN-␥, IL-2, and IL-4) after the subtraction of the background corresponding to cells compared with values (ءء) than values corresponding to wild-type mice; significantly higher (ء) incubated with culture medium alone. Significantly lower corresponding to wild-type mice, as assessed by t test. 1372 IMMUNOTHERAPY WITH RECOMBINANT Ig AND dsRNA determinant borne by IgNP was significantly and differentially reg- in conjunction with therapeutics addressing other pathogenic ulated by ITAMϩ and ITIMϩ Fc receptors. More specifically, in- factors. ϩ tact functionality of ITAM Fc␥R was critical for effective con- On a theoretical level, it becomes evident that although Fc␥R- ϩ tainment of the TAA tumoral process by immunotherapy with mediated internalization of Idϩ Ig results in processing and pre- IgNP plus pA:pU (Fig. 6A). This result was paralleled by a de- sentation of MHC class I-restricted epitopes, the nature of immune creased magnitude of immune response (both Tc1 and Tc2) against response depends on the context of immunization (e.g., degree of NP in tumor-bearing treated mice defective in the ITAM Fc␥R ϩ APC activation, serving as a “checkpoint”). This explains apparent subunit (Fig. 6, B–D). In contrast, although functionality of ITIM discrepancies with a previous report showing failure of induction ␥ Fc RIIB receptor was not required for effective tumor control by of cytolytic immunity by IgNP (28) and cautions against exclusive idiotypic immunotherapy (Fig. 6A), mice defective for Fc␥RIIB use of a single readout for the measurement of MHC class I re- showed an elevated Tc1 immunity against the MHC class I-responses. Instead, this study and a previous study (42) unravel a stricted NP epitope (Figs. 6, B and C, and Fig. 7). Strikingly, a great heterogeneity in regard to the phenotype and multiplicity of similar pattern was noticed in the case of an MHC class II-re- Tc subsets, with distinct function and potential, coexisting rather stricted epitope (HA 110–120 of influenza virus) borne by the ϩ than being mutually exclusive and thus in accordance to recently CDR3 of the H chain of IgG (Fig. 8). Although ITAM Fc receptors were essential for the generation of Th immunity against HA, proposed models (42, 43). In the absence of significant APC ac- ITIMϩ Fc␥RIIBϪϪ mice showed strongly enhanced Th1 and Th2 tivation, idiotypic T cell determinants are being processed and pre- immunity upon immunization with IgHA. Together, these data sented but the immune response is limited and dominated by T2 show that beyond the receptor-mediated internalization resulting in cells. This study sheds light on the importance of a second check-

ϩ Downloaded from more effective Ag processing and presentation in context of MHC point regulating this process, which is dependent on ITIM Fc class I or class II, ITAMϩ and ITIMϩ Fc receptors control the receptors. magnitude and quality of response against Id, likely by contribut- ing to the regulation of APC maturation and function (40, 41). Further, ITIMϩ Fc receptors constitute a checkpoint keeping under control the response against T cell Id. Additional studies are war- ranted, to outline elucidate whether selective targeting of ITAMϩ http://www.jimmunol.org/ Fc␥R or inhibition of ITIMϩ Fc␥R circumvents the need to si- multaneously activate APC in context of active immunotherapeutic approaches. Together, these results have direct implications in regard to the immunotherapy of Idϩ malignancies such as B cell lymphomas, myelomas, and some lymphocytic leukemias, using autologous idiotypic Ag. First, optimal cell based therapy with ex vivo pulsed DC must encompass activated APCs, and use of TLR ligands such by guest on October 2, 2021 as synthetic dsRNA or CpG motifs may accomplish this goal. This finding may explain why previous clinical studies with ex vivo pulsed immature DCs, despite achieving the goal of inducing mod- erate anti-Id responses, showed quite a limited clinical impact (17). Secondly, use of IgG as a vector is a far more optimal means to deliver epitopes to the APC for processing and presentation in the context of MHC class I or II molecules than is use of peptide epitopes, per se, for active efficient immunotherapy. This confirms prior observations with MHC class II epitopes (23, 25) and extends them to Tc epitopes; in both cases, Fc␥Rs bearing the ␥ subunit (ITAMϩ) mediate both cellular internalization and the amplification of subsequent APC function. Further, recombinant Ig may be safer and more practical than polyvalent IC for in vivo immuno- FIGURE 9. Schematic description of a model integrating the role of ϩ ϩ ␥ therapy. Besides potential use of autologous Id for treatment of ITAM and ITIM Fc R and of APC in the regulation of immune response to Id and in the homeostasis of the idiotypic network. In summary, lymphatic malignancies, engineered recombinant Ig bearing the administration of Ig carrying class I- or class II-restricted T cell epitopes (such as class I-restricted) derived from TAA may be epitopes results in generation of low magnitude Tc2 or Th2 responses used in combination with ligands for TLRs, such as synthetic respectively, unless there is: 1) coactivation of APCs that handle the pro- dsRNA, for the treatment of solid tumors (e.g., carcinomas) ex- cessing and presentation of the immunogen or 2) an impaired function of pressing TAA. In addition, this type of agent may be used to build ITIMϩ Fc␥R. In these cases, there is an elevation of both T1 and T2 up or prime immunity against conserved epitopes associated with components of the immune response against the T cell determinant on Ig. infectious agents of great concern for public health, such as influ- These observations suggest that idiotypic diversity resulting from immune enza virus. Finally, interfering with the function of ITIMϩ Fc re- repertoire generation and expansion is being dealt with in a fashion de- ϩ ϩ ceptors by various means (e.g., blocking Abs, peptidomimetics, pendent on a balance between the activity of ITAM and ITIM Fc␥R, antagonists interfering with downstream ITIM-dependent cell sig- along with the degree of activation of APCs. The default status of the ϩ idiotypic network is quiescent, with anti-idiotypic responses limited by naling) may greatly enable immunotherapy with Id autologous or ϩ ITIM Fc␥R and lack of APC activation. This checkpoint ensures limited recombinant Ig, or even enable immune systems to mount effective negative interference of immune defense mechanisms with the immune anti-Id responses as monotherapy. Conversely, selective interfer- ϩ repertoire diversification unavoidably associated with generation of new Id. ence with the function of ITAM Fc receptors in disorders medi- Thus, targeting of this checkpoint may offer novel therapeutic strategies to ated by IC (e.g., lupus nephritis) or amplification of the activity of deal with malignancies comprising, but not limited to, certain lymphoid ϩ ITIM Fc receptors (40) may represent a viable strategy alone or cancers. The Journal of Immunology 1373

The existence of the idiotypic network was proposed more than 3. Fehe´rvari, Z., and S. Sakaguchi. 2004. CD4ϩ Tregs and immune control. J. Clin. three decades ago (6). Since then, advances in molecular mecha- Invest. 114: 1209–1217. 4. Brack, C., M. Hirama, R. Lenhard-Schuller, and S. Tonegawa. 1978. Molecular nisms of generation of immune receptor diversity (4) have shed construction (recombination, mutation) of Ag rec A complete immunoglobulin light on the nature of Id (4, 44). Besides the practical implications, gene is created by somatic recombination. Cell 15: 1–14. this current work may unravel a mechanism explaining the ho- 5. Oudin, J., and M. Michel. 1963. Une nouvelle forme d’allotypie des globulines gamma du serum de lapin, apparenment lieé a la fonction et a la spećificite´ meostasis of the idiotypic network or management of the extraor- anticorps. C. R. Hebd. Seances Acad. Sci. (Paris) 257: 805–808. dinary idiotypic diversity of the Ig receptors. Thus, rather than 6. Jerne, N. K. 1974. Towards a network theory of the immune system. Ann. Im- being excluded from the repertoire via immune-mediated censor- munol. Inst. Pasteur (Paris) 125 C(1–2): 373–389. 7. Cazenave, P. A. 1977. Idiotypic-anti-idiotypic regulation of antibody synthesis in ship, new Id (essential neo-self Ags) are actually immune moni- rabbits. Proc. Natl. Acad. Sci. USA 74: 5122–5125. tored (Id processed and presented without significant response in a 8. Bona, C., and W. E. Paul. 1979. Cellular basis of regulation of expression of fashion resulting in clonal deletion, by virtue of the two check- idiotype. I. T-suppressor cells specific for MOPC 460 idiotype regulate the expression of cells secreting anti-TNP antibodies bearing 460 idiotype. J. Exp. Med. points discussed previously). This finding implies that in a default 149: 592–600. mode, the idiotypic network has a reduced activity level dominated 9. Sakato, N., M. Semma, H. N. Eisen, and T. Azuma. 1982. A small hypervariable by regulatory/suppressive functions and, in contrast, appropriate segment in the variable domain of an immunoglobulin light chain stimulates formation of anti-idiotypic suppressor T cells. Proc. Natl. Acad. Sci. USA 79: APC activation may actually determine a transition to a high ac- 5396–5400. tivity mode consisting in elevated anti-Id responses. This model 10. Furumoto, H., and M. Irahara. 2002. Human papilloma virus (HPV) and cervical (Fig. 9) is corroborated by previous evidence in humans showing cancer. J. Med. Invest. 49: 124–133. 11. Brećhot, C. 2004. Pathogenesis of hepatitis B virus-related hepatocellular carci- immunization with Id resulting in a B cell, Th or Tc cell response noma: old and new paradigms. Gastroenterology 127(5 Suppl 1): S56–S61. (reviewed in Ref. 17), implying that antigenic Id are not censored 12. Young, L. S., and P. G. Murray. 2003. Epstein-Barr virus and oncogenesis: from but rather managed in a way to minimize a negative impact on the latent genes to tumours. Oncogene 22: 5108–5121. Downloaded from 13. Sheu, B.-C., R.-H. Lin, H.-C. Lien, H.-N. Ho, S.-M. Hsu, and S.-C. Huang. 2001. diversity of the immune repertoire. Moreover, the Ig-VH region, Predominant Th2/Tc2 polarity of tumor infiltrating lymphocytes in human cer- and especially the CDR2-FR3-CDR3 area, encompasses multiple vical cancer. J. Immunol. 167: 2972–2978. MHC class I- and II-restricted epitopes, frequently mirrored by 14. Ghim, S. J., J. Sundberg, G. Delgado, and A. B. Jenson. 2001. The pathogenesis of advanced cervical cancer provides the basis for an empirical therapeutic vac- low level T cell reactivity in multiple myeloma patients (45), mak- cine. Exp. Mol. Pathol. 71: 181–185. Ϫ ing the occurrence of Id Ig an extremely improbable event. It thus 15. Yi, Q., I. Eriksson, W. He, G. Holm, H. Mellstedt, and A. Osterborg. 1997. becomes evident that, due to the apparently high frequency of an- Idiotype-specific T lymphocytes in monoclonal gammopathies: evidence for the http://www.jimmunol.org/ presence of CD4ϩ and CD8ϩ subsets. Br. J. Haematol. 96: 338–345. tigenic Id in the human immune repertoire, a censorship or nega- ϩ 16. Yi, Q., A. Osterborg, S. Bergenbrant, H. Mellstedt, G. Holm, and A. K. Lefvert. tive selection of Id lymphocytes would heavily interfere with the 1995. Idiotype-reactive T-cell subsets and tumor load in monoclonal gammopa- diversity of the immunologic potential, suggesting that this ho- thies. Blood 86: 3043–3049. 17. Yi, Q. Vaccines for Hematological Malignancies. In Handbook of Cancer Vac- meostatic mechanism is of great importance to the functionality of cines. M. A. Morse, T. M. Clay, and H. K. Lyerly, eds. Humana Press, Totowa, the immune system. NJ, p. 425–449. An interesting but deleterious consequence of the homeostatic 18. Yi, Q., R. Desikan, B. Barlogie, and N. Munshi. 2002. Optimizing dendritic mechanism keeping the functionality of the idiotypic network “in cell-based immunotherapy in multiple myeloma. Br. J. Haematol. 117: 297–305. 19. Kaminski, M. S., K. Kitamura, D. G. Maloney, and R. Levy. 1987. Idiotype check,” and thus the very integrity of the immune repertoire, is the

vaccination against murine B cell lymphoma: inhibition of tumor immunity by by guest on October 2, 2021 fact that monoclonal malignancies bearing class I-restricted Id free idiotype protein. J. Immunol. 138: 1289–1296. 20. Thrush, G. R., A. W. Butch, and S. P. Lerman. 1989. CD8 suppressor cell activity epitopes are responded against poorly (i.e., a modest Tc2, noncy- and its effect on CD4 helper cell-dependent growth of SJL/J B-cell lymphomas. totoxic response). This finding is concordant with the observation Cell. Immunol. 122: 555–562. that many malignant lymphatic tumors carry in fact MHC class 21. Chakrabarti, D., and S. K. Ghosh. 1992. Induction of syngeneic cytotoxic T lymphocytes against a B cell tumor. III. MHC class I-restricted CTL recognizes I-restricted Id (45) and illustrate that immune escape by Ag loss is the processed form(s) of idiotype. Cell. Immunol. 144: 455–464. not a frequent event in such diseases. Conversely, this model sup- 22. Dhodapkar, M. V., J. Krasovsky, and K. Olson. 2002. T cells from the tumor ports the use of immunotherapeutic approaches based on the in- microenvironment of patients with progressive myeloma can generate strong, tumor-specific cytolytic responses to autologous, tumor-loaded dendritic cells. duction of anti-Id responses. Proc. Natl. Acad. Sci. USA 99: 13009–13013. In conclusion, the present study shows that recombinant Ig car- 23. Zaghouani, H., R. Steinman, R. Nonacs, H. Shah, W. Gerhard, and C. Bona. rying disease-associated MHC class I-restricted epitopes are prom- 1993. Presentation of a viral T cell epitope expressed in the CDR3 region of a self ising therapeutic tools. In addition, it delineates factors that must immunoglobulin molecule. Science 259: 224–227. 24. Zanetti, M., F. Rossi, P. Lanza, G. Filaci, R. H. Lee, and R. Billetta. 1992. be dealt with to optimize their therapeutic use, enhancing in par- Theoretical and practical aspects of antigenized antibodies. Immunol. Rev. 130: allel our understanding on how idiotypic diversity is being man- 125–150. ϩ ϩ 25. Brumeanu, T. D., W. J. Swiggard, R. M. Steinman, C. A. Bona, and aged via APC and the balance between ITAM and ITIM Fc H. Zaghouani. 1993. Efficient loading of identical viral peptide onto class II receptors. molecules by antigenized immunoglobulin and influenza virus. J. Exp. Med. 178: 1795–1799. Disclosures 26. Billetta, R., M. R. Hollingdale, and M. Zanetti. 1991. Immunogenicity of an engineered internal image antibody. Proc. Natl. Acad. Sci. USA 88: 4713–4717. A. Bot, D. Smith, and B. Phillips are among the inventors on a patent 27. Li, S., V. Polonis, H. Isobe, H. Zaghouani, R. Guinea, T. Moran, C. Bona, and application, “Compositions and methods to initiate or enhance antibody P. Palese. 1993. Chimeric influenza virus induces neutralizing antibodies and and major histocompatibility class I- or class II-restricted T cell responses cytotoxic T cells against human immunodeficiency virus type 1. J. Virol. 67: by using immunomodulatory, noncoding RNA motifs.” The patent appli- 6659–6666. cation has been filed by Alliance Pharmaceuticals, and the license is held 28. Zaghouani, H., Y. Kuzu, H. Kuzu, T. D. Brumeanu, W. J. Swiggard, R. M. Steinman, and C. A. Bona. 1993. Contrasting efficacy of presentation by by MultiCell Therapeutics. A. Bot, D. Smith, and B. Phillips were previous major histocompatibility complex class I and class II products when peptides are employees of Alliance Pharmaceuticals, and D. Smith and B. Phillips are administered within a common protein carrier, self immunoglobulin. Eur. J. Im- recent employees of Astral Incorporated, which has been acquired by munol. 23: 2746–2750. MultiCell Therapeutics. 29. Kuzu, Y., H. Kuzu, H. Zaghouani, and C. Bona. 1993. Priming of cytotoxic T lymphocytes at various stages of ontogeny with transfectoma cells expressing a chimeric Ig heavy chain gene bearing an influenza virus nucleoprotein peptide. References Int. Immunol. 5: 1301–1307. 1. Dunn, G. P., L. J. Old, and R. D. Schreiber. 2004. The immunobiology of cancer 30. Wallace, P. K., K. Y. Tsang, J. Goldstein, P. Correale, T. M. Jarry, J. Schlom, immunosurveillance and immunoediting. Immunity 21: 137–148. P. M. Guyre, M. S. Ernstoff, and M. W. Fanger. 2001. Exogenous antigen tar- 2. Houghton, A. N., and J. A. Guevara-Patino. 2004. Immune recognition of self in geted to Fc␥RI on myeloid cells is presented in association with MHC class I. immunity against cancer. J. Clin. Invest. 114: 468–471. J. Immunol. Methods 248: 183–194. 1374 IMMUNOTHERAPY WITH RECOMBINANT Ig AND dsRNA

31. Kalergis, A. M., and J. V. Ravetch. 2002. Inducing tumor immunity through the 39. Brodeur, B. R., P. Tsang, and Y. Larose. 1984. Parameters affecting ascites tu- selective engagement of activating Fc␥ receptors on dendritic cells. J. Exp. Med. mour formation in mice and monoclonal antibody production. J. Immunol. Meth- 195: 1653–1659. ods 71: 265–272. 32. Takai, T., M. Li, D. Sylvestre, R. Clynes, and J. V. Ravetch. 1994. FcR␥ chain 40. McGaha, T. L., B. Sorrentino, and J. V. Ravetch. 2005. Restoration of tolerance deletion results in pleiotrophic effector cell defects. Cell 76: 519–529. in lupus by targeted inhibitory receptor expression. Science 307: 590–593. 33. Nakamura, A., T. Yuasa, A. Ujike, M. Ono, T. Nukiwa, J. V. Ravetch, and 41. Dhodapkar, K. M., J. L. Kaufman, M. Ehlers, D. K. Banerjee, E. Bonvini, T. Takai. 2000. Fc␥ receptor IIB-deficient mice develop Goodpasture’s syndrome S. Koenig, R. M. Steinman, J. V. Ravetch, and M. V. Dhodapkar. 2005. Selective upon immunization with type IV collagen: a novel murine model for autoimmune blockade of inhibitory Fc␥ receptor enables human dendritic cell maturation with glomerular basement membrane disease. J. Exp. Med. 191: 899–906. IL-12p70 production and immunity to antibody-coated tumor cells. Proc. Natl. 34. Rotzschke, O., K. Falk, K. Deres, H. Schild, M. Norda, J. Metzger, G. Jung, and Acad. Sci. USA 102: 2910–2915. H. G. Rammensee. 1990. Isolation and analysis of naturally processed viral pep- 42. Sad, S., R. Marcotte, and T. R. Mosmann. 1995. Cytokine-induced differentiation tides as recognized by cytotoxic T cells. Nature 348: 252–254. of precursor mouse CD8ϩ T cells into cytotoxic CD8ϩ T cells secreting Th1 or 35. Bot, A., S. Bot, and C. A. Bona. 1998. Protective role of ␥ interferon during the Th2 cytokines. Immunity 2: 271–279. recall response to influenza virus. J. Virol. 72: 6637–6645. 43. Bot, A., K. A. Smith, and M. von Herrath. 2004. Molecular and cellular control 36. Wang, L., D. Smith, S. Bot, L. Dellamary, A. Bloom, and A. Bot. 2002. Non- of T1/T2 immunity at the interface between antimicrobial defense and immune coding RNA danger motifs bridge innate and adaptive immunity and are potent pathology. DNA Cell Biol. 23: 341–350. adjuvants for vaccination. J. Clin. Invest. 110: 1175–1184. 44. Victor-Kobrin, C., and C. A. Bona. 1989. The molecular origin of regulatory 37. Ko¨hler, G., and C. Milstein. 1976. Derivation of specific antibody-producing idiotypes. Int. Rev. Immunol. 5: 117–129. tissue culture and tumor lines by cell fusion. Eur. J. Immunol. 6: 511–519. 45. Hansson, L., H. Rabbani, J. Fagerberg, A. O¨ sterborg, and H. Mellstedt. 2003. 38. Wang, P., P. Palese, and R. E. O’Neill. 1997. The NPI-1/NPI-3 (karyopherin ␣) T-cell epitopes within the complementarity-determining and framework regions binding site on the influenza a virus nucleoprotein NP is a nonconventional nu- of the tumor-derived immunoglobulin heavy chain in multiple myeloma. Blood clear localization signal. J. Virol. 71: 1850–1856. 101: 4930–4936. Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021