TCR-like antibodies in cancer immunotherapy

Tumor antigens are grouped into several categories according to their origins and specificity. The first category is oncovirus antigens, which include Epstein-Barr nuclear antigen 1-3 (EBNA 1-3), latent membrane protein 1 (LMP1), and LMP2 derived from Epstein-Barr virus (EBV) [54], hepatitis B virus X protein (HBX) from hepatitis B virus (HBV) [55, 56], core nonstructural protein 3 (NS3) and nonstructural protein 5A (NS5A) from hepatitis C virus (HCV) [57], type E5, E6, and E7 proteins from human papillomavirus (HPV) [58], viral transactivator (Tax) from human T cell leukemia-lymphoma virus (HTLV) [59], latency-associated nuclear antigen (LANA), virus active G protein-coupled receptor homolog (vGPCR), and virus IFN-inducible factor (vIRF-1) from Kaposi sarcoma-associated herpesvirus (KSHV) [60], structural protein PP65 from cytomegalovirus (CMV) [61], and group-specific antigen (gag) and pol reading frame 468 (Pol468) from human immunodeficiency virus (HIV) [62]. The oncoviruses can cause many diseases, including Burkitt’s lymphoma (BL), non-Hodgkin’s B cell lymphoma (NHL), nasopharyngeal carcinoma (NPC), hepatocellular carcinoma (HCC), cervical cancer, adult T cell leukemia (ATL), primary effusion lymphoma (PEL), Kaposi’s sarcoma (KS), and Merkel cell carcinoma (MCC). The oncovirus antigens are highly tumor-specific, as they are unique to the oncoviruses and are not shared by normal human tissues. However, viral infections cause only about 10–15% of all human cancers, and some healthy individuals do not develop cancer even with the infection of an oncovirus [60, 63, 64]. Hence, the oncovirus antigens are of limited use in the clinic.

The second group of tumor antigens involves chromosome/gene mutations in cancer cells [65, 66]. These mutations include chromosomal translocation, loss, duplication, and loss or point mutation of nucleic acids in the exons, introns, or regulatory regions of genes [67]. These mutations can lead to the expression of truncated proteins, fusion proteins, or neoantigens that are unique to cancer cells, such as beta-catenin S37F in melanoma [68], alpha-actinin-4 K122N in lung cancer [69], heat shock protein 70 kilodalton-2 (hsp70-2) F293I in renal cancer [70], Kirsten rat sarcoma viral oncogene (K-ras) G12D in colon cancer [71], myeloid differentiation primary response 88 (MYD88) L265P in hairy cell leukemia [72], and B cell receptor-Abelson murine leukemia viral oncogene homolog 1 (BCR-ABL) fusion protein in chronic myeloid leukemia [73]. These antigens are tumor cell-specific. However, some types of cancer have a high burden of genetic mutations, whereas other types of cancers may not; in addition, many genetic mutations are unique to the tumor cells of individual patients [74, 75]. Hence, this group of tumor antigens is difficult to target with the current adoptive cellular therapy strategy.

The third group of tumor antigens is the cancer-testis antigens, which are overexpressed in multiple types of tumor cells of patients [76, 77]. In healthy donors, this group of antigens is expressed only in immune-privileged organs, such as the testis or placenta. Because the immune-privileged organ cells do not express MHC alleles, TCRs that recognize the peptide/MHC complex derived from this group of antigens will not damage the normal tissue cells [78]. Moreover, high-affinity TCRs targeting cancer-testis antigens can be isolated from the peripheral blood of normal donors because of the absence of cancer-testis antigens in the peripheral blood [79, 80]. Hence, this group of tumor antigens, including New York esophageal squamous cell carcinoma-1 (NY-ESO-1), melanoma-associated antigen A (MAGE-A), and synovial sarcoma X (SSX), comprises the largest number in current clinical trials [81, 82].

The fourth group of tumor antigens involves antigens with minimal or limited expression in normal cells, such as carcinoembryonic antigen (CEA), melanoma antigen recognized by T cells 1 (MART-1), and tyrosine kinase 10 [83‐85]. Targeting these antigens can damage normal tissues, and sophisticated technology is needed for the future development of immunotherapy against these antigens [86, 87]. This group also includes antigens derived from non-essential organs, such as CD19 and CD20 from B cells [88]. Targeting these antigens can cause non-fatal damage to normal tissue, which medical interventions can cure [89, 90].

Importantly, about 95% of the aforementioned tumor antigens are intracellular proteins, and very few tumor-specific antigens are extracellular [91]. Thus, to target tumors through tumor-specific antigens, a novel strategy must be developed.

Because intracellular proteins can be digested into small peptides in the proteasome of a cell, which can be conjugated with MHC molecules in the endoplasmic reticulum (ER) and transported to the tumor cell surface, the peptide/MHC complex on the tumor cell surface has been deemed as a tumor-specific antigen [92]. MHC class I molecules are expressed on the surface of all nucleated cells, and numerous studies have demonstrated the feasibility of targeting tumors through the recognition of the peptide/MHC complex on the cell surface [85, 93, 94].

In 1981, Wylie and Klinman conducted the first study of a TCR-like antibody [95]. To study the immune response to influenza, they injected influenza virus and the virus-infected cell line PR8-L929 into C3H/HeJ and BAL6.K mouse strains. They found that approximately one third of the virus-specific antibodies reacted to viral hemagglutinin (HA) or neuraminidase. The remainder of virus-specific antibodies recognized antigens found on the surface of virus-infected PR8-L929 cells but not on the virion or uninfected cells. It was later found that the MHC participated in the recognition of viral antigens by the antibodies [96]. Similar results have been found in mouse cells transformed with simian virus antigen (SV40), murine cytomegalovirus (MCMV) pp89 (168–176) peptides, vesicular stomatic virus (VSV), and EBV [97‐99]. It was demonstrated that mouse MHC conformational epitopes are peptide-specific. The monoclonal antibody (mAb) 34.4.20 recognized the VSV nucleoprotein (52–59) peptide on mouse H-2Kb but not ovalbumin (OVA) (257–264), MCMV pp89 (168–176), or influenza nucleoprotein (Y345–360) peptides on the same MHC allele [98]. Although these studies did not test the cytotoxic effect of TCR-like antibodies, they provided clear evidence that TCR-like antibodies generated in the mouse B cells can specifically bind to the peptide/MHC complex on the cell surface.

In 2000, Chames and colleagues reported the first TCR-like antibody targeting human tumor antigens [20]. Using the phage library technique, they isolated a human antibody directed against the EADPTGHSY peptide encoded by MAGE-A1 and presented by the HLA-A1 molecule. MAGE-1 is a cancer-testis gene overexpressed in multiple cancers but with restricted expression in the testis of a healthy person [100]. The phage Fab antibody bound to the HLA-A1 molecule complexed with the MAGE-A1 peptide but not to the HLA-A1 molecule complexed with other peptides, indicating the specificity of the antibody. Furthermore, the TCR-like antibody bound to the MAGE-1+/HLA-A1+ melanoma cells, indicating that the phage library-derived Fabs could recognize the native complex displayed on the surface of tumor cells. Compared to mouse hybridoma technology, the phage library screening is structure-dependent, fast, and cost-effective. This technique was subsequently explored in the study of TCR-like antibodies against peptide/MHC complexes derived from other tumor antigens, such as telomerase catalytic subunit [27], glycoprotein 100 (gp100) [23, 24], mucin 1 (MUC1) [28], human telomerase reverse transcriptase (hTERT) [27], NYESO-1 [29], MART-1 [34], preferentially expressed antigen in melanoma (PRAME) [45], tyrosinase [38], and WT1 [15]. TCR-like antibodies targeting virus epitopes derived from HTLV [46, 47], influenza [48], HIV [50, 52], and CMV [53] were also developed through the phage library strategy.

Early studies of these phage library-derived Fab antibodies focused on the use of antibodies as tools to detect the expression levels of peptide/MHC complexes on the tumor cell surfaces. To develop therapeutic strategies with phage library-derived TCR-like antibodies, researchers have used the CAR strategy by ligating the heavy chain variable (VH) and light chain variable (VL) region of the phage library-derived Fab antibody with the intracellular domain of CD3 molecules. The first TCR-like CAR-T strategy was developed in 2001 by ligating the VH and VL of the Fab antibody targeting the melanoma cells expressing MAGE-A1 and HLA-A1 [21]. The Fab recognizing the MAGE-A1 EADPTGHSY peptide/MHC complex on the melanoma cell surface was fused to the Fc (epsilon)RI-gamma molecule and retrovirally transduced into normal T cells. The transduced primary human T lymphocytes bound to the MAGE-A1 peptide/MHC complexes and responded to native MAGE-A1+/HLA-A1+ target cells by specific cytokine production of interferon gamma (IFNγ) and tumor necrosis factor alpha (TNFα). These T cells could also lyse MAGE-A1+/HLA-A1+ target cells but not control MAGE-A1-/HLA-A1+ or MAGE-A1+/HLA-A1- tumor cells, indicating that the lysis of tumor cells via TCR-like antibodies was HLA-restricted and antigen-dependent. In a later study, the phage library-purified antibodies were further mutated through a combination of light (L) chain shuffling, heavy (H) chain-targeted mutagenesis, and in vitro selection of phage display libraries to be higher affinity (Fab-Hyb3) [22]. A functional study of Fab-Hyb3 found that the mutated TCR-like CAR-T mediated better recognition of the antigen on the tumor cell surface, indicating that the affinity of TCR-like antibodies dramatically affected the killing ability of the antibodies. The CAR-T technology has since been employed in several other TCR-like antibody studies, including those of gp100, minor histocompatibility antigen 1H (HA-1H), and WT1 [11, 25, 43, 44].

In 2006, Wittman and colleagues started to use the TCR-like antibody as a typical antibody therapy to mediate ADCC and CDC effects against tumors [31]. To target an HLA-A2-restricted peptide derived from human chorionic gonadotropin beta (hCG-β), which is overexpressed in over 90% of breast cancers, they developed a mouse IgG2a mAb (termed 3.2G1) via the hybridoma technique. The 3.2G1 antibody recognized the GVLPALPQV peptide from hCG-β presented by the HLA-A2 molecule and specifically stained the cells in a peptide- and antibody concentration-dependent manner. Staining of human tumor lines with the 3.2G1 TCR-like antibody also demonstrated the antibody’s ability to recognize endogenously processed peptides from the breast cancer cell line MDA-MB-231. Moreover, 3.2G1 antibody mediated CDC and ADCC against the human breast carcinoma MDA-MB-231 cell line in vitro and inhibited tumor implantation and growth in nude mice. These results provided valid evidence for the development of novel therapeutic antibodies that specifically kill tumors via recognition of peptide/MHC complexes. Since then, several TCR-like antibodies have been developed via the hybridoma strategy to mediate ADCC, CDC, or ADCP effects against tumor cells. These include TCR-like antibodies targeting peptide/MHC complexes derived from tumor protein 53 (TP53) [36], macrophage migration inhibitory factor (MIF) [40], proteinase 3 (PR1) [41], and WT1 [15, 44]. In addition to ADCC and CDC effects, the mouse hybridoma-derived TCR-like antibodies can also be utilized therapeutically to detect the expression of peptide/MHC complexes on the tumor cell surface as phage library-derived Fab antibodies [30, 32, 33, 37, 49, 51].

Because antibodies can be conjugated with toxins to deliver specific cytotoxic effects into cells, Denkberg and colleagues generated a conjugation molecule with a TCR-like antibody in 2003 [16]. In their study, a single-chain HLA-A2 molecule complexed with a common antigenic T cell HLA-A2-restricted epitope derived from the gp100 was used to immunize HLA-A2 transgenic mice. A phage display library was constructed from the immunized mice, and a recombinant single-chain fragment variable (scFv) antibody that could bind to the gp100 IMDQVPFSV peptide/MHC complex with a high affinity in the nanomolar range was isolated. When fused to a very potent cytotoxic effector molecule in the form of a truncated bacterial toxin, the TCR-like antibody could specifically kill antigen-presenting cells (APCs) in a peptide-dependent manner. In 2008, Epel and colleagues employed the same technology to fuse a truncated form of Pseudomonas exotoxin A with the phage-derived TCR-like antibody that specifically targets the FLRNFSLML peptide/HLA-A2 complex derived from TCR gamma alternate reading frame protein (TARP) [35]. The fusion molecule exhibited specific cytotoxic activity on breast and prostate cancer cells that correlated with their TARP and HLA expression patterns and inhibited the growth of human breast tumor cells in nude mice. These results demonstrated the power of the TCR-like antibody conjugation approach to generate novel targeting molecules to eliminate tumor cells with the unique specificity observed in cytotoxic CD8+ T cells [101]. In the same year, a TCR-like antibody targeting MART-1 conjugated with immunotoxin was also developed for anti-melanoma therapy [26].

TCR-like antibodies can also induce tumor cell death directly after binding to the peptide/MHC complex on the tumor cell surface [102]. In 2006, Verma and colleagues generated two TCR-like antibodies (RL4B and RL6A) that recognized peptides derived from hCG-β and human p68 RNA helicase. They found that two TCR-like antibodies destroyed tumor cells independently of immune effector mechanisms, such as ADCC and CDC. TCR-like antibodies mediated the apoptosis of tumor cells through selective and specific binding to p68 RNA helicase YLLPAIVHI and hCG-β GVLPALPQV peptide/HLA class I complexes, which triggered the activation of c-Jun N-terminal kinases (JNKs) and intrinsic caspase pathways. This signaling was accompanied by the release of mitochondrial cytochrome c and apoptosis-inducing factor. The apoptosis induced by the TCR-like antibodies was completely inhibited by soluble MHC tetramers loaded with relevant peptides and by inhibitors for JNKs and caspases. Thus, their study suggested the existence of a novel mechanism of TCR-like antibodies in the mediation of tumor cell destruction, in addition to ADCC and CDC. This mechanism would appear to be especially important due to the absence or tolerance of immune cells in cancer patients [103‐105].

The major functions of TCR-like antibodies include the detection of peptide/MHC complexes, CAR-T strategy, ADCC, CDC, ADCP, immunotoxin targeting, and direct induction of tumor cell death. TCR-like antibodies can also be used as a block to prevent the recognition of normal tissue cells by self-reacting T cells in autoimmune diseases. In an experimental allergic encephalomyelitis mouse model, Aharoni and colleagues developed several monoclonal antibodies that bound to the complex of myelin basic protein (BP) peptide on mouse I-As [106]. The antibodies blocked the proliferative response of in vitro cultured T cells to the BP peptide/I-As complex without affecting the T cell response to an irrelevant peptide derivative from tuberculin on the same allele. The antibodies also inhibited experimental allergic encephalomyelitis in H-2s mice. Hence, antibodies directed specifically to the autoantigen/MHC complex may offer a highly selective and effective treatment in autoimmune diseases. Moreover, in 2004, Held and colleagues generated a high-affinity (Kd = 60 nM) antibody that specifically recognized the NY-ESO-1 (157–165) but not NY-ESO-1 (157–167) or a cryptic NY-ESO-1 (159–167) peptide/HLA-A2 complex. In a dose-dependent manner, the antibody blocked the recognition of NY-ESO-1/HLA-A2-positive tumor cells by NY-ESO-1 (157–165) peptide-specific CD8+ T cells [29].

Due to the clinical prevalence of cancers, most studies of TCR-like antibodies have been conducted in the field of cancers. The major functions of TCR-like antibodies have been explored in two areas—the detection and measurement of the expression of tumor-specific peptide/MHC complexes on the tumor cell surfaces and the mediation of cytotoxicity against tumor cells. The detailed molecular mechanisms of TCR-like antibodies are summarized as follows (Fig. 2):

It is generally believed that, because of the repetitive antigen stimulation and in vivo selection process of hybridoma technology, TCR-like antibodies isolated using this technology have relatively high binding affinity (low nanomolar range) compared with the moderate to average binding affinity (~ 50–300 nM) of phage-derived TCR-like antibodies [31, 41, 53]. However, phage library-derived TCR-like antibodies of high affinity in the low nanomolar range have also been successfully isolated from second-generation libraries and by in vitro affinity maturation [22, 109]. TCR-like antibodies derived from both technologies have been evaluated in pre-clinical studies.

The greatest advantage of TCR-like antibodies is their ability to target intracellular tumor antigens with minimal in vitro manipulation. The TCR-T adoptive cell therapy can also target intracellular antigens but requires a much more complicated preparation process [78]. In the traditional TCR-T adoptive cell therapy, the peripheral blood or tumor infiltration lymphocytes from a cancer patient must be isolated by apheresis. The lymphocytes are activated for 1 to 3 days to be transduced by TCR-containing lentivirus, retrovirus, or transposon vectors. The transduced T cells are then expanded to a large number (1 × 10⁹) before infusion back into the patient. The entire procedure takes about 3 to 4 weeks and is technically demanding, expensive, and time-consuming without the guarantee of success [94]. In addition, the transduced antigen-specific TCRs may mismatch with endogenous wild-type TCRs, as both TCRs exist in the same T cells [110, 111]. TCR-like antibodies, however, are relatively easy to prepare and store and used as off-the-shelf. Through the binding of the Fab region to the peptide/MHC complex, the Fc region of the TCR-like antibody can bind to the Fc gamma receptors (FcγR) expressed by patients’ NK cells, monocytes, or macrophage cells and activate these cells to kill tumors.

CAR-T is a specific form of tumor immunotherapy that equips the T cells with the tumor surface antigen-specific antibody and CD3 signaling pathway [112]. The recognition of tumor surface antigen by the antibody can trigger the CAR-T cell activation and the killing of tumor cells. The clinical success of CD19 CAR-T cells has proved their dramatic effect against tumors [113‐115]. There are several reports of converting the TCR-like antibodies, especially the phage library-derived Fab antibodies, into CAR vectors [11, 25, 43]. T cells transduced with TCR-like antibody-derived CARs can specifically lyse tumor cells, indicating the therapeutic effectiveness of TCR-like antibody CAR-T cell therapy. Because of the lack of tumor-specific biomarkers on the surface of tumor cells, the traditional CAR-T therapy has achieved little success in solid tumors [116]. We envision that the TCR-like antibody CAR-T cell therapy could have specific value for solid tumors, as it targets intracellular tumor-specific antigens.

The checkpoint antibody strategy is a significant step in the history of humanity’s fight against cancer [117]. The molecular mechanism of this strategy is that the checkpoint antibody can reverse the immune suppression of tumor antigen-specific T cells that pre-exist in the patient’s body so that they may target the cancer cells [103]. The success of CTLA-4 and PD-1 checkpoint antibody therapy in the clinic has confirmed this mechanism [118]. However, checkpoint antibody therapy is effective in only about 20–30% of patients when used individually and 40–60% of patients when used in combination [119, 120]. These low rates indicate there may be a lack of tumor antigen-specific T cells at the tumor site, which hampers the effect of the therapy. TCR-like antibody therapy, however, does not depend on the existence of tumor antigen-specific T cells in the patient’s body and can activate the normal immune cells to target the tumor cells through ADCC, CDC, or ADCP [7, 18]. Combining TCR-like antibodies with checkpoint antibodies in future clinical studies may further improve the responses of patients.

Vaccine therapy is a longtime developed idea in the field of cancer immunotherapy, preceding the CAR-T cellular therapy and checkpoint antibody therapy [121]. The concept of using the host’s own immunity to fight cancers in the long-term has attracted significant interest from the scientific community. However, only two vaccines have currently been approved to treat cancer patients, and most tumor vaccines have shown poor clinical results, leading to their failure to secure approval from the US Food and Drug Administration (FDA) [122, 123]. It is hypothesized that the effect of a tumor vaccine is dependent on the development of memory immunity of tumor-specific T cells, and the tumor environment is usually plagued with immune-suppressive molecules [124]. Thus, it is difficult to induce a strong anti-tumor effect by the vaccine strategy. Moreover, the vaccine strategy is time-consuming and may take several months to develop tumor antigen-specific T cells. TCR-like antibodies, however, do not depend on the existence of tumor antigen-specific T cells and can take effect immediately after administration.

TCR-like antibodies, as new tools in the cancer immunotherapy field, have just begun to attract attention from the scientific community. By combining their fine specificity to recognize the peptide/MHC complexes of T cells with the biological and pharmacological properties of an antibody, TCR-like antibodies may have broad applications in the clinic. However, there are also several hurdles that must be overcome to achieve clinical success with the TCR-like antibodies.

First, TCR-like antibodies are MHC-restricted, which means that they are effective only for a certain group of patients expressing the tumor-specific antigen on a specific MHC allele. With HLA-A2 as the most common MHC allele in cancer patients, many tumor-specific peptides associated with this allele have been found [91]. Other HLA alleles, however, still lack tumor-specific peptides, which hamper the development of TCR-like antibody therapy. Further identification of less-common MHC-associated peptides will help solve this problem.

Second, the downregulation or absence of peptide/MHC complexes on the tumor cell surface is a common mechanism of tumor cells to evade immune surveillance [125]. TCR-like antibodies, like TCR-T therapy, may suffer from this effect. However, reports showed that some chemicals, cytokines, or radiation therapy can upregulate the expression of MHC and activate the MHC signaling pathway [126, 127]. Thus, TCR-like antibodies may combine with other therapies to achieve the best results. Furthermore, the affinity of TCR-like antibodies is generally higher than the affinity of in vitro synthesized TCRs [43]. The affinity of TCR-like antibodies can also be easily mutated to a higher affinity via molecular technology [22]. This will render antibodies more capable of recognizing the peptide/MHC molecule at extremely low levels.

Third, the immune-suppressive environment is a hurdle for the TCR-like antibody immunotherapy. Tumor cells reside in hidden sites to prevent the access of T cells, generate a hypoxic environment, and secrete a large amount of immune-suppressive cytokines, such as interleukin 10 (IL-10), transforming growth factor beta (TGF-β), or other molecules that cause the T cells, NK cells, macrophages, or monocyte to experience anergy or death [124, 128]. In addition, there are many suppressive immune cells around the tumor cells, which dampen the anti-tumor immune response [129, 130]. Thus, TCR-like antibodies may bind to the peptide/MHC complex on the tumor cell surface but might not mediate tumor destruction. Combining the TCR-like antibody therapy with other immune suppression-reversion therapy could help solve this problem. Examples may include the adoptive transfer of freshly expanded NK cells, monocytes, or macrophages in combination with TCR-like antibody therapy, or the combination of anti-PD-1 or anti-CTLA-4 antibody therapy. One advantage of TCR-like antibodies is that they can easily penetrate the tumor environment and they do not require the existence of tumor antigen-specific T cells at the tumor site. Furthermore, some of the TCR-like antibodies can induce tumor cell death directly through binding to the peptide/MHC complex [39, 102].