Antiidiotypic response against murine monoclonal antibodies reactive with tumor-associated antigen TAG-72

Immunogenicity of CAR-T Cell Therapeutics: Evidence, Mechanism and Mitigation

Chimeric antigen receptor T cell (CAR-T) therapy demonstrated remarkable success in long-term remission of cancers and other autoimmune diseases. Currently, six products (Kymriah, Yescarta, Tecartus, Breyanzi, Abecma, and Carvykti) are approved by the US-FDA for treatment of a few hematological malignancies. All the six products are autologous CAR-T cell therapies, where delivery of CAR, which comprises of scFv (single-chain variable fragment) derived from monoclonal antibodies for tumor target antigen recognition is through a lentiviral vector. Although available CAR-T therapies yielded impressive response rates in a large number of patients in comparison to conventional treatment strategies, there are potential challenges in the field which limit their efficacy. One of the major challenges is the induction of humoral and/or cellular immune response in patients elicited due to scFv domain of CAR construct, which is of non-human origin in majority of the commercially available products. Generation of anti-CAR antibodies may lead to the clearance of the therapeutic CAR-T cells, increasing the likelihood of tumor relapse and lower the CAR-T cells efficacy upon reinfusion. These immune responses influence CAR-T cell expansion and persistence, that might affect the overall clinical response. In this review, we will discuss the impact of immunogenicity of the CAR transgene on treatment outcomes. Finally, this review will highlight the mitigation strategies to limit the immunogenic potential of CARs and improve the therapeutic outcome.

[Mechanisms of resistance and escape to CAR-T cells]

CAR-T cells represent a new approach to anti-tumor cellular immunotherapy allowing to combine the recognition of tumor antigens on target cells and the activation, proliferation and cytotoxic capacity of T lymphocytes. Impressive clinical results have been obtained with CAR-T cells targeting the CD19 antigen in relapsing or refractory B cell malignant lymphomas or acute lymphoblastic leukemias, with complete response rates of 40 to 90%. However, 30 to 50% of responding patients in B malignancies will escape treatment secondarily, and the effectiveness of these approaches in solid tumors remains limited. Different mechanisms of primary resistance and/or escape to CAR-T cells have been described. This review aims to describe these mechanisms and explore potential ways for optimization. We will see that the initial response and its long-term persistence depends on several parameters: the functional characteristics of the CAR-T cells in vivo, the expression of targeted antigens on tumor cells, the development of a immunosuppressive microenvironment. Or of an immune response directed against the CAR molecule. In solid tumors in particular, the specificity of the antigen target and the "homing" of CAR-T cells in the tumor site are additional elements to consider. A better knowledge of mechanisms of resistance will help to improve the clinical outcomes by either modulating the construction and the production of CAR-T cells and/or to combine them with other immunotherapeutic approaches to better control the tumor microenvironment.

Immunogenicity of CAR T cells in cancer therapy

Patient-derived T cells genetically reprogrammed to express CD19-specific chimeric antigen receptors (CARs) have shown remarkable clinical responses and are commercially available for the treatment of patients with certain advanced-stage B cell malignancies. Nonetheless, several trials have revealed pre-existing and/or treatment-induced immune responses to the mouse-derived single-chain variable fragments included in these constructs. These responses might have contributed to both treatment failure and the limited success of redosing strategies observed in some patients. Data from early phase clinical trials suggest that CAR T cells are also associated with immunogenicity-related events in patients with solid tumours. Generally, the clinical implications of anti-CAR immune responses are poorly understood and highly variable between different CAR constructs and malignancies. These observations highlight an urgent need to uncover the mechanisms of immunogenicity in patients receiving CAR T cells and develop validated assays to enable clinical detection. In this Review, we describe the current clinical evidence of anti-CAR immune responses and discuss how new CAR T cell technologies might impact the risk of immunogenicity. We then suggest ways to reduce the risks of anti-CAR immune responses to CAR T cell products that are advancing towards the clinic. Finally, we summarize measures that investigators could consider in order to systematically monitor and better comprehend the possible effects of immunogenicity during trials involving CAR T cells as well as in routine clinical practice.

Safety, tumor trafficking and immunogenicity of chimeric antigen receptor (CAR)-T cells specific for TAG-72 in colorectal cancer

Background

T cells engineered to express chimeric antigen receptors (CARs) have established efficacy in the treatment of B-cell malignancies, but their relevance in solid tumors remains undefined. Here we report results of the first human trials of CAR-T cells in the treatment of solid tumors performed in the 1990s.

Methods

Patients with metastatic colorectal cancer (CRC) were treated in two phase 1 trials with first-generation retroviral transduced CAR-T cells targeting tumor-associated glycoprotein (TAG)-72 and including a CD3-zeta intracellular signaling domain (CART72 cells). In trial C-9701 and C-9702, CART72 cells were administered in escalating doses up to 10¹⁰ total cells; in trial C-9701 CART72 cells were administered by intravenous infusion. In trial C-9702, CART72 cells were administered via direct hepatic artery infusion in patients with colorectal liver metastases. In both trials, a brief course of interferon-alpha (IFN-α) was given with each CART72 infusion to upregulate expression of TAG-72.

Results

Fourteen patients were enrolled in C-9701 and nine in C-9702. CART72 manufacturing success rate was 100% with an average transduction efficiency of 38%. Ten patients were treated in CC-9701 and 6 in CC-9702. Symptoms consistent with low-grade, cytokine release syndrome were observed in both trials without clear evidence of on target/off tumor toxicity. Detectable, but mostly short-term (≤14 weeks), persistence of CART72 cells was observed in blood; one patient had CART72 cells detectable at 48 weeks. Trafficking to tumor tissues was confirmed in a tumor biopsy from one of three patients. A subset of patients had ¹¹¹Indium-labeled CART72 cells injected, and trafficking could be detected to liver, but T cells appeared largely excluded from large metastatic deposits. Tumor biomarkers carcinoembryonic antigen (CEA) and TAG-72 were measured in serum; there was a precipitous decline of TAG-72, but not CEA, in some patients due to induction of an interfering antibody to the TAG-72 binding domain of humanized CC49, reflecting an anti-CAR immune response. No radiologic tumor responses were observed.

Conclusion

These findings demonstrate the relative safety of CART72 cells. The limited persistence supports the incorporation of co-stimulatory domains in the CAR design and the use of fully human CAR constructs to mitigate immunogenicity.

Construction of di-scFv through a trivalent alkyne–azide 1,3-dipolar cycloaddition

Heterofunctional azide and alkyne PEG-linkers have been synthesized and site specifically conjugated to scFv via a reactive thiol functionality; two scFv were coupled by copper catalyzed 1,3-dipolar cycloaddition to make divalent scFv (di-scFv) with an inter-scFv distance defined to provide divalent binding; antigen binding was maintained for the di-scFv construct and increased several times compared to that of the parent scFv; the cycloaddition reaction reported herein represents an important ligation strategy to covalently link macromolecular proteins and retain sensitive structural conformations.

Pharmacokinetics and biodistribution of 177Lu-labeled multivalent single-chain Fv construct of the pancarcinoma monoclonal antibody CC49

PURPOSE

Lutetium-177 (177Lu) is a radionuclide of interest for radioimmunoimaging (RII) and radioimmunotherapy (RIT) on account of its short half-life (161 h) and the ability to emit both beta and gamma radiation. Single-chain Fv (scFv) constructs have shown advancement in cancer diagnosis and therapy due to the pharmacokinetics advantage and seem to be intriguing tools in oncology. The objective of this study was to evaluate the pharmacokinetics and biodistribution characteristics of the 177Lu-labeled tetravalent scFv of CC49 MAb and intact CC49 IgG in vivo.

METHODS

Conjugation and labeling conditions of multivalent scFv with 177Lu were optimized without affecting integrity and immunoreactivity. For this purpose, multivalent scFv constructs {dimer, sc(Fv)2; tetramer, [sc(Fv)2]2} of the MAb CC49 were expressed as secretory proteins in Pichia pastoris. The purified scFv constructs and IgG form of CC49 were conjugated with a bifunctional chelating agent, ITCB-DTPA, and labeled with 177Lu. The comparative biodistribution, blood clearance, and tumor-targeting characteristics of 177Lu-labeled tetravalent [sc(Fv)2]2 construct of CC49 MAb and intact CC49 IgG were investigated in the athymic mice bearing LS-174T xenografts.

RESULTS

Approximately, 90% of 177Lu incorporation was achieved using ITCB-DTPA chelator, and the labeled immunoconjugates maintained integrity and immunoreactivity. Blood clearance studies demonstrated an alpha half-life (t1/2alpha) of 177Lu-labeled [sc(Fv)2]2 and IgG of CC49 at 4.40 and 9.50 min and a beta half-life (t1/2beta) at 375 and 2,193 min, respectively. At 8 h post administration, the percent of the injected dose accumulated/gram (%ID/g) of the LS-174T tumor was 6.4+/-1.3 and 8.9+/-0.6 for 177Lu-labeled [sc(Fv)2]2 and IgG of CC49, respectively, in the absence of L-lysine. The corresponding values were 8.0+/-0.6 and 8.4+/-1.2 in the presence of L-lysine. Renal accumulation of [sc(Fv)2]2 was significantly (p<0.005) reduced in the presence of L-lysine.

CONCLUSION

The results of this study demonstrate that the ITCB-DTPA conjugation and 177Lu-labeling of scFvs are feasible without influencing the antibody characteristics. 177Lu-labeled [sc(Fv)2]2 showed faster clearance and equivalent tumor uptake at 8 h compared with its IgG form, with a markedly reduced renal uptake in the presence of L-lysine. Therefore, 177Lu-labeled [sc(Fv)2]2 may be a potential radiopharmaceutical for the treatment of cancer.

Grafting of "abbreviated" complementarity-determining regions containing specificity-determining residues essential for ligand contact to engineer a less immunogenic humanized monoclonal antibody

Murine mAb COL-1 reacts with carcinoembryonic Ag (CEA), expressed on a wide range of human carcinomas. In preclinical studies in animals and clinical trials in patients, murine COL-1 showed excellent tumor localization. To circumvent the problem of immunogenicity of the murine Ab in patients, a humanized COL-1 (HuCOL-1) was generated by grafting the complementarity-determining regions (CDRs) of COL-1 onto the frameworks of the variable light and variable heavy regions of human mAbs. To minimize anti-V region responses, a variant of HuCOL-1 was generated by grafting onto the human frameworks only the "abbreviated" CDRs, the stretches of CDR residues that contain the specificity-determining residues that are essential for the surface complementarity of the Ab and its ligand. In competition RIAs, the recombinant variant completely inhibited the binding of radiolabeled murine and humanized COL-1 to CEA. The HuCOL-1 and its variant showed no difference in their binding ability to the CEA expressed on the surface of a CEA-transduced tumor cell line. Compared with HuCOL-1, the HuCOL-1 variant showed lower reactivity to patients' sera carrying anti-V region Abs to COL-1. The final variant of the HuCOL-1, which retains its Ag-binding reactivity and shows significantly lower serum reactivity than that of the parental Ab, can serve as a prototype for the development of a potentially useful clinical reagent.

Effects of humanization and gene shuffling on immunogenicity and antigen binding of anti-TAG-72 single-chain Fvs

One major constraint in the clinical application of murine monoclonal antibodies (MAbs) is the development of a human antimurine antibody response. The immunogenicity of MAbs can be minimized by replacing nonhuman regions with corresponding human sequences. The studies reported in our article were undertaken to analyze the immunoreactivity and the immunogenicity of the CC49 single-chain antibody fragments (scFvs): (i) an scFv construct comprised of mouse CC49 VL and VH (m/m scFv), (ii) a light chain shuffled scFv with human VL (Hum4 VL) and mouse CC49 VH (h/m scFv), and (iii) a humanized scFv assembled from Hum4 VL and CC49 VH complementary determining regions (CDRs) grafted onto a VH framework of MAb 21/28' CL (h/CDR scFv). The CC49 scFvs competed for an antigen binding site with CC49 IgG in a similar fashion in a competition radioimmunoassay and were able to inhibit the binding of CC49 IgG to the antigen completely. The immunogenicity of CC49 scFvs was tested using sera with antiidiotypic antibodies to MAb CC49 obtained from patients treated by CC49 IgG in clinical trials. All tested sera exhibited the highest reactivity to the m/m scFv. However, the sera demonstrated differential reactivities to h/CDR scFv and h/m scFv. Replacement of the mouse chain in h/m scFv and h/CDR scFv decreased or completely averted serum reactivity. Our studies compared for the first time the antigen binding and immunogenicity of different scFv constructs containing the mouse, CDR grafted or human variable chains. These results indicate that the humanized CC49 scFv is potentially an important agent for imaging and therapeutic applications with TAG-72-positive tumors.

Antitumor effect of the idiotypic cascade induced by an antibody encapsulated in poly(d,l-lactide-co-glycolide) microspheres

A major difficulty encountered during development of antibody vaccines is their weak immunogenicity. In this study, a monoclonal antibody CS20.5 to human breast cancer antigen CA15.3 was coencapsulated in poly(d,l-lactide-co-glycolide) microspheres with monophosphoryl lipid A. The antitumor effect of this formulation was investigated in a murine model. The induced Ab2 biologically mimics antigen as it competed with CA15.3 for the same idiotope on Ab1. Ab3 induction was also observed. After five sequential administrations of encapsulated antibody, mice showed statistically significant tumor regression. These results indicate that this formulation may serve as a potential treatment for breast cancer.

Antibody constructs for radioimmunodiagnosis and treatment of human pancreatic cancer

Pancreatic cancer (PC) is a common disease that is seldom cured. Current approaches to the treatment of PC are not effective because the non-specific nature of both chemotherapy and external beam radiation results in toxicity to normal tissue. Monoclonal antibodies (MAbs) can be used as selective carriers for delivering radionuclides, toxins, or cytotoxic drugs to malignant cell populations. Therefore, MAb-technology has led to a significant amount of research in targeted therapy. Targeted therapy would generally allow the concentration of cytotoxic agents in tumors and would markedly lessen the toxicity to normal tissues, which limits the dosage and effectiveness of systemically administered drugs. A variety of MAbs are being pre-clinically evaluated for the diagnosis and treatment of PC. Novel recombinant antibody constructs hold a promising future in both the diagnosis and treatment of cancer. By genetic-engineering methods, several high affinity antibody fragments with optimum tumor targeting properties, such as higher functional affinity (divalent and multivalent scFvs) and blood residence time (good tumor localization with high radiolocalization index), have been generated. Animal models have permitted the in vivo assessment of these antibody-based reagents, therapeutic/diagnostic radionuclide, radiolabeling conditions, and efficacy of administration regimes. For PC, immunoscintigraphy using MAbs has taken new strides. The use of MAbs and their fragments for radioimmunoguided surgery and therapy of PC has shown encouraging results at preclinical levels and warrants further attention.

Removal of amphipathic epitopes from genetically engineered antibodies: production of modified immunoglobulins with reduced immunogenicity

Several approaches have been developed to reduce the human immune response to nonhuman antibodies. However, chimeric antibodies and humanized antibodies often have decreased binding affinity. We described a new approach for reducing the immunogenicity of chimeric antibodies while maintaining the affinity. This approach seeks to prevent the recognition of murine immunogenic peptides from the antibody variable region by human lymphocytes. Putative immunogenic epitopes in the variable region are identified and subjected to site directed mutagenesis to make them human and/or to break the amphipathic motifs. The R3 antibody, which blocks the epidermal growth factor (EGF) receptor, was used as a model system to test this approach. Four segments containing possible amphipathic epitopes were found in the heavy variable domain using the program AMPHI. Six amino acids within two of these segments were substituted by the corresponding residues from a homologous human sequence. No mutations were made in the murine light variable domain. Experiments in monkeys suggested that the "detope" R3 antibody was less immunogenic than its chimeric analogue. A search for possible amphipathic epitopes in the Kabat database revealed the presence of conserved patterns in the different families of variable region sequences, suggesting that the proposed method may be of general applicability.

Relative position of the hexahistidine tag effects binding properties of a tumor-associated single-chain Fv construct

Hexahistidine tag (His-tag) is the most widely used tag for affinity purification of recombinant proteins for their structural and functional analysis. In the present study, single chain Fv (scFv) constructs were engineered form the monoclonal antibody (MAb) CC49 which is among the most extensively studied MAb for cancer therapy. For achieving efficient purification of scFvs by immobilized metal-ion affinity chromatography (IMAC), a His-tag was placed either at the C-terminal (scFv-His6) or N-terminal (His6-scFv) of the coding sequence. Solid-phase radioimmunoassay for scFv-His6 showed only 20-25% binding whereas both His6-scFv and scFv (no His-tag) showed 60-65% binding. Surface plasmon resonance studies by BIAcore revealed the binding affinity constant (KA) for His6-scFv and scFv as 1.19 x 10(6) M(-1) and 3.27 x 10(6) M(-1), respectively. No K(A) value could be calculated for scFv-His6 due to poor binding kinetics (kon and koff). Comparative homology modeling for scFv and scFv-His6 showed that the C-terminal position of the His-tag partially covered the antigen-binding site of the protein. The study demonstrates that the C-terminal position of His-tag on the CC49 scFv adversely affects the binding properties of the construct. The results emphasize the importance of functional characterization of recombinant proteins expressed with purification tags.

Radioimmunotherapy: designer molecules to potentiate effective therapy

The evolution of monoclonal antibody forms for radioimmunotherapy and other antibody-based applications has been driven by a series of problems that each new form has introduced. Ehrlich was the first to present the concept that antibodies could be exploited in such a manner. Four decades were required before technological advances allowed the exploration of the potential of antibodies for radioimaging and radioimmunotherapeutic applications. Advances in DNA technology have led to the ability to tailor and manipulate the immunoglobulin molecule for specific functions and in vivo properties. This article discusses the use of monoclonal antibodies for radiotherapy with an emphasis on the problems that have been encountered and the subsequent solutions.

Human Anti-Animal Antibody Interferences in Immunological Assays

Purpose: The scope and significance of human anti-animal antibody interference in immunological assays is reviewed with an emphasis on human anti-animal immunoglobulins, particularly human anti-mouse antibodies (HAMAs).

Issues: Anti-animal antibodies (IgG, IgA, IgM, IgE class, anti-isotype, and anti-idiotype specificity) arise as a result of iatrogenic and noniatrogenic causes and include human anti-mouse, -rabbit, -goat, -sheep, -cow, -pig, -rat, and -horse antibodies and antibodies with mixed specificity. Circulating antibodies can reach gram per liter concentrations and may persist for years. Prevalence estimates for anti-animal antibodies in the general population vary widely and range from &lt;1% to 80%. Human anti-animal antibodies cause interferences in immunological assays. The most common human anti-animal antibody interferent is HAMA, which causes both positive and negative interferences in two-site mouse monoclonal antibody-based assays. Strategies to prevent the development of human anti-animal antibody responses include immunosuppressant therapy and the use of humanized, polyethylene glycolylated, or Fab fragments of antibody agents. Sample pretreatment or assay redesign can eliminate immunoassay interferences caused by anti-animal antibodies. Enzyme immunoassays, immunoradiometric assays, immunofluorescence, and HPLC assays have been designed to detect HAMA and other anti-animal antibodies, but intermethod comparability is complicated by differences in assay specificity and lack of standardization.

Conclusions: Human anti-animal antibodies often go unnoticed, to the detriment of patient care. A heightened awareness on the part of laboratory staff and clinicians of the problems caused by this type of interference in routine immunoassay tests is desirable. Efforts should be directed at improving methods for identifying and eliminating this type of analytical interference.