Antibodies: At The Nexus of Antigens and Cancer Vaccines

Abstract

This review describes the development of monoclonal antibodies and the inception of their use in cancer therapy, their impact on defining cancer biomarkers, and their structural utility in new cancer vaccine development.

There is no doubt that monoclonal antibodies (mAbs) are the most successful and important therapeutic strategy developed in the past 2 decades for the treatment of cancer [1]. By the third year after Kohler and Milstein first described the method for generating mAbs, The Wistar Institute was leading the way to make antibody therapy a clinical reality. Sears et al [2] reported one of the first clinical trials that indicated the potential efficacy of monoclonals. The relative importance of this and related work from The Wistar Institute was the basis for the awarding of 2 patents to The Wistar Institute in October 1979 and April 1980 for making mAbs against tumor and viral antigens. These were the first patents awarded for showing the utility of clinically useful mAbs. Since that time, multiple mAbs have received approval from the US Food and Drug Administration (FDA) for the treatment of various solid tumors and hematological malignancies, including trastuzumab (ERBB2) for breast cancer, cetuximab (EGFR) for colon cancer, rituximab (CD20) for non-Hodgkin lymphoma, and bevacizumab (VEGF) for tumor vasculature. A large number of highly promising mAbs targeting host immune response, such as anti-CTLA-4 (ipilimumab) and anti-PD-1 (pembrolizumab), received FDA approval for treatment of various tumors.

Wistar Institute investigators showed that antibodies could define antigens as biomarkers [3]. More generally, cancer biomarkers are defined as molecules produced either by the tumor itself or by the microenvironment in response to the presence of cancer or other associated conditions, such as inflammation. These early antibodies serve as the prototypes for understanding the mechanisms by which tumor-associated carbohydrate antigens (TACAs) are expressed, determining their structure and biological function, and ultimately learning how to target them with therapy. Among TACAs, blood group antigens were identified as having an important role (Table 1).

Table 1.

Structures of the H, A, B, Le^a, Le^b, X, and Y Determinants and Enzymes Involved in Their Biosynthesis

Specificity	Glycosyltransferase(s)	Structure of Determinant
H	α-2-L-fucosyl transferase (H/Se) genes	Fucα1–2Galβ1-R
A	α-3-N-acetyl-D-galactosaminyl transferase (A gene)	GalNAcα1–3(Fucα1–2)Galβ1R
B	α-3-D-galactosyl-transferase (B gene)	Galα1–3(Fucα1–2)Galβ1-R
Lea	α-4-L-fucosyl-transferase	Galβ1–3(Fucα1–4)GlcNAcβ1-R
Leb	α-2-L-fucosyl and α-4-L-fucosyl transferases	Fucα1–2Galβ1–3(Fucα1–4)GlcNAcβ1-R
X	α-3-L-fucosyl-transferase (X gene)	Galβ1–4(Fucα1–3)GlcNAcβ1-R
Y	α-2-L-fucosyl- and α-3-L-fucosyl transferases	Fucα1–2Galβ1–4(Fucα1–3)GlcNAcβ1-R
Sialyl LeX	α-2-3-sialyltransferase	NeuAcα2–3Galβ1–4(Fucα1–3)GlcNAcβ1-R
Sialyl Lea	α-2-3-sialyltransferase	NeuAcα2–3Galβ1–3(Fucα1–4)GlcNAcβ1-R

Blood group antigens are associated with the outcomes of [4] and risk for [5] some cancers. Identification of blood group–related CA19-9 as the first carbohydrate tumor-specific antigen was a milestone in tumor antigen and biomarkers discovery because it represented a new class of cancer-specific molecules [6]. Of no less importance are gangliosides, particularly GD2, which are associated with cell attachment and regulation of signaling with anti-GD2 [7]. It is impressive that, despite the emergence of highly powerful genomic, proteomic, epigenetic, and other methods for discovery in cancer, the most useful biomarkers and potential therapeutic targets are those discovered at almost 30 years ago.

Antibodies also provide a technological platform to design new immunogens [8, 9]. These antibodies led the way to the development of surrogates of nominal antigens [10, 11] and peptide mimetics, especially for TACAs [8, 12–14]. Technologically, antibodies were used to develop anti-idiotypic antibodies [10, 11] or to screen random peptide phage libraries to identify carbohydrate-mimetic peptides (CMPs) [12]. Knowing the crystal structure of antibodies [15] also permits the integration of drug design and pharmacophore design concepts and their application to vaccine design [9, 13].

THE BEGINNING OF ANTIBODY THERAPY

The successful use by Kohler and Milstein of somatic cell hybridization between immune B cells and murine myeloma cells resulted in monoclonal hybridomas secreting unique mAbs in unlimited quantities. Somatic cell hybridization was originally used in our laboratories by using inactivated Sendai virus as a cell-fusing agent [16] to analyze genome interaction in heterokaryons [17]. Introduction of polyethylene glycol for somatic cell fusion resulted in standardization of the technology of hybridoma formation [18]. Using polyethylene glycol–induced hybridoma formation, a panel of murine mAbs detecting tumor-associated antigens with specificities for melanoma, astrocytoma, colorectal carcinoma, and breast carcinoma was established. All antitumor mAbs were selected from murine splenocytes after immunization with human tumor cell lines [3, 19, 20]. Some of these antigen-antibody pairs are listed in Table 2. A large panel of mAbs was used to define specific binding to antigens of tumor cells, including gastrointestinal cancer– and melanoma-associated antigens [3, 19–21]. Using such antibodies, antigenic structures such as SA-Le^a and SA-LeX were demonstrated to be overexpressed in many cancers. The gastrointestinal tumor–specific mAbs were shown to have antitumor biological activity in vitro by detection of complement-dependent cytotoxicity and antibody-dependent cellular cytotoxicity (ADCC) [22] and inhibition of tumor growth in vivo [22, 23].

Table 2.

Wistar Institute–Defined Antigens, by Monoclonal Antibody

MAb	Antigen	Specificity	Patent	Status or Use
CA19-9	Sialyl-Lea	Gastrointestinal tumors	4 471 057	Licensed, marketed
Lewis a, b	Lea, Leb	Gastrointestinal tumors	4 607 009	Forensic, diagnostic
ME36.1	GD2/GD3	Melanoma	4 849 509	BB-IND 3003, phase I
CO17-1A	EpCAM	Gastrointestinal tumors		Approved in Germany
BR55-2	LeY	Adenocarcinoma	4 971 792	BB-IND 2709, phase I, II
425	EGFR	Adenocarcinoma, astrocytoma		BB-IND 2893, phase I, II

Abbreviations: EGFR, epithelial growth factor receptor; BB-IND, biologics–investigational new drug number.

The relevance of the antibodies listed in Table 2 is that the antigenic moieties that these antibodies recognize are well defined. These particular antibodies were novel at the time, as indicated by the filing and awarding of patents protecting these antibodies as therapeutics along with their associated targets as diagnostic tools (Table 2). mAbs against Le^a, Le^b, and Le^ab were used by investigators in forensic medicine. CA19-9 is structurally identical with the sialylated Le^a-carbohydrate determinant. A diagnostic assay using CA19-9 is now established as one of the most extensively used tools to measure clinical end points in pancreatic and gastrointestinal cancers worldwide. An immunodiagnostic assay that quantitatively measures the presence of SA-Le^a antigen in human serum and plasma received FDA approval as the only validated assay for monitoring patients with pancreatic cancer. Furthermore, the discovery and structural elucidation of SA-Le^a [24] and the study of molecular mechanisms involving this structure promoted the understanding of a variety of diseased and normal states. Because the binding of SA-Le^a on tumor cells to E-selectin expressed on endothelial cells is responsible for its function in hematogenous spread, an opportunity emerged to interrupt the attachment of extravasated tumor cells to vasculature and the metastasis process, using novel therapeutic compounds such as glycomimetic agents [25].

Some of the antibody-antigen pairs formed the basis for clinical trials with therapeutic agents under investigational new drug applications (Table 2) submitted to the FDA to test their clinical feasibility. mAb CO-171A was one of the first murine antibodies approved by FDA for clinical trials. mAb CO-171A underwent extensive phase 1/2 trials and a study led by G. Riethmuller (which involved a patient in Munich with colorectal carcinoma) for the prevention of recurrence, resulting in its approval and marketing in Germany. Centocor later decided not to continue the application process. mAb 425 (EGFR specific), after positive findings from a phase 1 trial, was studied by Dr Luther Brady (Hahneman Hospital, Philadelphia, Pennsylvania) as a delivery system for radioactive iodine (I¹²⁵) in the treatment of astrocytomas.

More importantly, these mAbs were the forerunners of some of the FDA-approved antibodies, and they are the gold standards for others presently being evaluated in the clinic. The antibody 425 taught that EGFR could be targeted to inhibit EGFR signaling with cetuximab (Erbitux) and panitumumab (Vectibix), representing 2 of the presently 12 mAbs approved by the FDA for treatment of patients with cancer. Catumaxomab, a mouse trifunctional bispecific antibody against CD3 × EpCAM, is approved in the European Union for use in patients with malignant ascites generated by an EpCAM-positive tumor [26]. The anti-GD2 antibody 14.18 [27] is now in phase 3 clinical studies in patients with neuroblastoma. Several anti-Lewis Y (LeY)-reactive antibodies have been also tested in the clinic. Most recently, the safety, tolerability, and therapeutic efficacy of the LeY-specific humanized mAb MB311 was evaluated in patients with malignant effusions in a phase 2 clinical trial [28]. What is amazing, from a development perspective, was the relatively short period between antibody discovery, through preclinical assessment; antibody production; and clinical trial testing. In the shortest possible time, the Wistar team coordinated preclinical testing with simultaneous establishment of the good manufacturing practices (GMP)–compliant facility (in Malvern) to produce clinical-grade biologicals. The establishment of an antibody production facility under GMP was extraordinary in an academic institution.

LINKING ANTIGEN DISCOVERY WITH THERAPEUTICS

The discoveries of sialylated, fucosylated lacto-type, and neolacto-type carbohydrate structures as antigens for these mAbs were accomplished with the aid of analytical methods, such the immunostaining of thin layer chromatograms, and structural analysis, using mass spectrometry and nuclear magnetic resonance imaging (Tables 1 and 2). The role of the LeY antigen in cancer treatment and prevention has been extensively studied. LeY is a type 2 blood group antigen overexpressed on the surface of adenocarcinoma cells, either as glycoproteins or glycolipids. LeY is a difucosylated oligosaccharide whose terminal structure is catalyzed by fucosyltransferases, such as FUT1 (α1,2) and FUT4 (α1,3; Table 1) [29]. Of particular interest is that, early on, it was observed that the γFc receptor associated with BR55-2 could influence mechanisms of cytotoxicity [30]. In vitro experiments suggested that murine immunoglobulin G3 (IgG3), IgG2a, and IgG2b isotypes mediated ADCC with murine macrophages, while IgG1 was less efficient [30]. Interestingly enough, murine IgG2a was more efficient than other isotypes at mediating ADCC [30]. Further study demonstrated an independent regulation of anti-LeY mAb–induced effector functions of ADCC and cytokine release through different subclasses of FcγRs. While ADCC activity depends on activation of FcγRI, tumor necrosis factor α release seemed to be triggered by activation of FcγRII [31].

Much like adenocarcinoma-reactive antibodies, it was shown that melanoma cells treated with mAbs suppressed the growth of melanoma tumors in xenograft models [19]. Furthermore, inhibition of metastases of a human melanoma xenograft, using mAb specific to the GD2/GD3 ganglioside ME36.1 indicated that the significant level of killing of tumor cells that express GD2, GD3, or both suggests the importance of multiple specificity toward tumor antigens (ie, binding of an antibody to ≥2 tumor-associated antigens) [7, 32]. This observation suggested that mAbs targeting multiple tumor antigens might prove beneficial as therapeutics. This consideration has lead to a broader concept of developing multivalent carbohydrate-based vaccines to induce antibodies to the multiple TACAs found on cancer cells. Overexpression of >1 tumor-specific antigenic determinant, especially glycans belonging to the same type of oligosaccharide family, is observed in cancer cells [20]. TACAs are pan-targets on tumor cells because they are collectively and intimately involved in cell-death-signaling pathways and antibodies that bind to ≥2 TACAs are hypothesized to be more-effective therapeutics.

TACAs play a significant role in tumor cell survival, proliferation, and adhesion and migration. The LeY antigen plays a positive role invasion and metastasis of cancer cells. However, the mechanisms by which LeY enhances invasion and tumor metastasis are still largely unknown. It has been realized that LeY regulates the phosphorylation and expression of some molecules involved the expression of cell-cycle-related factors through EGFR/PI3K signaling pathways to promote cell proliferation [33]. Such studies highlight the potential of targeting signaling pathways and the role that glycans can play in signaling events. Inhibition of PI3K/AKT signaling is a paradigm for targeted cancer therapy [34]. PI3K/AKT affect the signaling pathway associated with angiogenesis, tumorigenesis, stress sensitivity, and inhibition of apoptosis as mediated through phosphatidylinositol-3,4-bisphosphate (PIP2)/phosphatidylinositol-3,4,5-triphosphate (PIP3; Figure 1). Receptor activation recruits PI3K to the inner cell membrane via phosphorylated receptor tyrosine kinases, activated RAS, or G protein β and γ subunits. AKT kinase catalyzes the phosphorylation of PIP2 to PIP3 (Figure 1). This lipid anchors AKT protein kinase family members to the membrane for phosphorylation and activation by PDK1. PTEN phosphatase antagonizes PI3K-AKT signaling by converting PIP3 back to PIP2. PI3K-AKT signaling inhibits apoptosis and regulates cell growth, survival, and proliferation, particularly in cancer and tumor cells.

Figure 1.

Postulated signaling pathways underlying the function of focal adhesion kinase (FAK). Anti-LeY and ganglioside antibodies have the potential to inhibit tumor-related signal transduction related to survival, cell motility, and proliferation through FAK. FAK localized at focal contact sites communicate with tumor-associated carbohydrate antigen–expressing molecules recruits a variety of signaling molecules. FAK activation results in phosphorylation at Tyr-397 and recruitment of Src family kinases into a signaling complex consisting of FAK, Src, and p130^Cas. Phosphorylation of focal adhesion– and/or actin-associated substrates triggers the binding of adaptor molecules, leading to the modulation of small GTPases, ACK, ERK2/MAP kinase, and JNK/SAP kinase cascades. Coordinated and localized stimulation of these cascades influences focal contact turnover and actin cytoskeleton dynamic addition to expression of motility- and invasion-associated proteins such as matrix metalloproteinases.

LeY can also influence the biological behavior of a tumor cell as an important component of integrins α5 and β1 by various signal pathways, such as cell adhesion and migration, that might involve focal adhesion kinase (FAK), activation of β₁ integrin, and PI3K/AKT signaling. Antibodies such as BR55-2, directed against LeY side chains of the ErbB receptor family, are capable of inhibiting ErbB-mediated signaling [35], while other anti-LeY mAbs mediate the mitogen-activated protein kinases (MAPK) signaling pathway [36]. The extracellular-signal-regulated kinase extracellular-signal-regulated kinases (ERK) in the MAPK signaling pathway has been the subject of intense research leading to the development of pharmacologic inhibitors for the treatment of cancer. ERK is a downstream component of an evolutionarily conserved signaling module that is activated by the RAF serine/threonine kinases and associated with motility (adhesion and migration; Figure 1). Consequently, much effort is directed toward the EGFR-RAS-RAF-MEK-ERK signaling network to identify novel target-based approaches for cancer treatment. The identification of the LeY antigen provided an early glimpse of the potential to probe and target the underlying signaling pathways related to tumor-related signal transduction.

Much like LeY, GD2 ganglioside can facilitate cell attachment [37]. Gangliosides isolated from tumor cells promote collagen-stimulated platelet aggregation and ATP secretion and enhance platelet adhesion to immobilized collagen. Gangliosides interact with a number of cell-surface receptors, including integrin receptors. Anti-GD2 mAb affects α2β1 integrin-mediated platelet adhesion to collagen and phospho-tyrosine signaling of FAK [38]. Likewise, LeY carried by integrin αvβ3 plays a critical role in the attachment of cells and activates integrin αvβ3/FAK signaling [39]. Such results indicate that gangliosides and LeY enhance adhesion of tumor cells to extracellular matrix by upregulating integrins mediated through FAK (Figure 1), thereby providing insight into how these TACAs may be involved in metastasis [38] and emphasizing the importance of targeting these moieties.

ANTIBODIES AS TEMPLATES FOR CANCER VACCINES

The Wistar Institute was, in toto, a biotechnology powerhouse before biotechnology was popularized as an industry. At its heart was vaccine discovery driven by advances in understanding underlying biological and immunological mechanisms of the various diseases of interest and applying state-of-art technology as needed. Its role in mAb technology and its impact on the biotechnology field is well known [40] and embodied by the founding and emergence of Centocor. Among its patent portfolio was one of the original patents on anti-idiotypic antibodies, which described the immune response to virus induced by anti-idiotype antibodies (US patent 4731237 A), and a patent on the use of anti-idiotypic antibodies to induce responses to GD2 (US patent 5854069).

The idea that anti-idiotypic antibodies have a potential immunotherapeutic role came of age at The Wistar Institute [10, 11]. While the treatment of patients with leukemia with mouse mAbs was being pursed [41], patients with gastrointestinal cancer treated with mAbs were shown to develop anti-idiotypic antibodies [10]. Immunization of patients with colorectal carcinoma with an anti-idiotypic antibody further suggested the utility of molecular mimicry in treating cancer [42] and that triggering of the idiotypic antibody cascade can be associated with a favorable clinical response to antibody-based therapy in patients with cancer [10].

The development of therapeutics for cancer also necessitates understanding the structural nature of the target, as well as the therapeutic agent. Such structural data form the basis for translational studies, bringing a laboratory finding to the clinic. Crystallized and modeled antibodies included anti-GD2 mAb ME36.1 [15] and anti-LeY mAb BR55-2 [43]. The extension of anti-idiotypic antibodies is the development of peptide mimics either through a structural approach or screening of random peptide phage libraries for surrogate peptide mimics recognized by the mAb [12, 44]. While such mimics can perhaps find utility as inhibitors of metastases [25], they have also shown utility as cancer vaccines [8, 14, 45, 46].

We have shown that CMPs of LeY, as identified by screening with BR55-2 and ME36.1 mAbs [12], can induce responses that mediate both humoral and cellular responses to cancer cells in vivo [8, 45, 46]. The intent of these CMPs is to induce polyspecific antibodies that can recognize multiple TACAs. Among mechanisms of tumor cell growth inhibition was the observation that CMP-induced antibodies could mediate apoptosis via antibody internalization [45]. Quick internalization of mAbs into cytoplasm is advantageous for enhanced cytotoxic effects of toxins or cytotoxic drugs conjugated with mAbs. Anti-GM3 and anti-SA-Le^a antibodies are internalized, while anti-GD3, anti-GD2, and anti-GM2 antibodies are known to accumulate on the cell surface, bringing a new aspect to inducing TACA responses in cancer vaccine strategies, which typically rely on complement-dependent cytotoxicity or ADCC.

The concept that T cells of the immune system could recognize TACA upon immunization with TACA surrogates was also developed [14, 46]. Both anti-idiotypes and CMPs were shown to activate TACA-specific CD4⁺ and CD8⁺ T cells that targeted TACA-expressing cancer cells [14, 46], providing a structural interpretation for the mechanism involved [47]. Effective targeting of LeY as an antigen by antibodies has also facilitated development of novel therapeutics with engineered T cells [48, 49]. Human chimeric antigen receptor–expressing T cells redirected to LeY have been described [49]. These T cells caused established LeY-expressing tumors to regress in a xenotransplantation model [48], while in vitro studies showed that LeY-specific T cells induced lysis of LeY-expressing tumor targets of both epithelial and hematological origin [48].

The importance of antibodies for research and clinical application also demands knowledge of their structure. This information can be further used for protein engineering, modifying the antigen, and identifying an epitope and has been facilitated by various methods, including X-ray crystallography and computational modeling. The structural information about both BR55-2 and ME36.1 mAbs provided a technological approach of using reverse engineering of an antigen to develop a better immunogen by considering how an antibody recognizes and binds to the nominal antigen (Figure 2) [9, 13]. The steps in developing CMPs for vaccines are illustrated in Figure 2. The first step is to identify antibodies or receptors with functional properties against tumor cells. These receptors are then used to screen a random peptide library for converting TACAs into CMPs. After screening and identification of CMPs, we then used crystal structures to further design CMPs. Briefly, the crystal structure of ME36.1 was analyzed in the context of comparing GD2 binding and CMP binding, using a molecular docking approach [13]. A CMP called P10 was identified from a random peptide library screen, using ME36.1 mAb [12], and was subsequently shown to generate immune responses (step 3 in Figure 2) in mice that inhibited tumor growth in vivo (step 4 in Figure 2) [14].

Figure 2.

General scheme of converting tumor-associated carbohydrate antigens (TACAs) into peptide-based vaccine. It is important to start with a lectin(s) or antibody with functionality, but not all carbohydrate-mimetic peptides (CMPs) selected will induce the desired response. Computer modeling based on crystallographic information can lead to CMPs with improved fidelity to TACA structures – reverse engineering of an immunogen. This approach was used to develop a CMP, P10s, that is in clinical trials. Abbreviation: mAb, monoclonal antibody.

Based upon docking calculations with ME36.1, P10 formed a minimal number of hydrogen bonds with ME36.1, compared with GD2 binding to ME36.1. Based on the hydrogen bond interaction between GD2 and the CDRs of ME36.1, a new CMP, P10s, was developed [13]. Conformational and docking calculations suggested that P10s would form an increased number of hydrogen bonds with ME36.1 that are in common with the GD2 hydrogen bond interaction pattern with ME36.1. We observed that P10s did indeed induce higher-titer antibodies to the target antigen and antigen-expressing tumor cells than the parent CMP, P10 [13]. These studies suggest that, for carbohydrate mimics, pharmacophore-based design is superior over the conformational approach undertaken for other peptide mimics.

The CMP P10s has entered the clinic in a phase 1 study of high-risk patients with breast cancer [50]. In the phase 1 dose-escalation trial, 2 cohorts of patients with stage IV breast cancer were immunized, using a 3 × 3 cohort design. P10s was synthesized with the pan–T-cell epitope PADRE, formulated at 300 and 500 µg per injection with Montanide ISA 51 VG and administered subcutaneously. Immunized subjects mounted an anti-P10s response, and antibodies of obtained after immunization mediated cytotoxicity of human breast cancer cell lines, including a cell line de novo resistant to trastuzumab, by as yet unknown mechanism(s). In addition, admixing serum from subjects immunized with P10s vaccine with docetaxel improves the efficiency of cell death by reducing the 50% lethal concentration of the drug [50]. Our next steps are to demonstrate that the addition of the vaccine to chemotherapy regimens (Chemovax) can enhance pathologic complete response (pCR) rates in subjects who typically display low pCR rates, to address how a new agent that induces cytotoxic antibodies might increase pCR rates in combination with chemotherapy and to examine how understanding these mechanisms may help to identify selective targets and subjects for therapy.

SUMMARY

The fundamental basis of antibody-based therapy of tumors dates back to the original observations of antigen expression and immunogenicity elicited by tumor cells through serological techniques of mAb production at The Wistar Institute in the 1970s. The hybridoma technology involved the immunization of mice with human cancer in the hope of generating antisera with some degree of cancer specificity. Most importantly, this approach yielded recognition of carbohydrate cell-surface structures as tumor-associated antigens and paved the way to using antibodies and antigen surrogates in immune therapy strategies.

The ensemble of the results and the achievements proved that the synergy between various disciplines, such as immunology, structural biology, cancer biology, and oncology, was a successful approach before the multidisciplinary team concept in science was developed and pursued. These achievements fueled a venue that became a center of gravity for outstanding cancer research investigators where discoveries that set the foundations for the new wave of clinically applicable agents and long-lasting contribution to the cancer research and patient care were made. This legacy continues to facilitate investigations to meet the urgent need for new and improved vaccines. Defining the nexus of antigens and cancer vaccines required synergy between various established disciplines. This intentional synergy proved a successful approach before the multidisciplinary team concept in science was developed and pursued.