1 Introduction

The pathogenesis of cancer is a multiple-step process [1–4], taking from 10 to 20 years or more for the manifestation of cancer from a single initiated cell. The host immune system holds the cancer at the stealthy “elimination and equilibrium” phases for 90 % of its development history, with outgrowth being the consequence of “escape” from immune surveillance [1, 2]. Both the innate and adaptive immune systems participate in the battle against the cancer development, although adaptive immunity plays the major role [1, 5]. Human adaptive immunity comprises B cell and T cell responses, the function of which is to recognize and kill the “non-self” invaders. In addition to microbes, cancer cells can also be recognized as “non-self” and eliminated by the host immune system [1]. The cancer-specific antigens that induce such immune responses can be divided into the following five classes: (1) mutated antigens that are produced by mutated genes such as EGFRvIII [6] and the p53 and Ras families [7]; (2) cancer–germ line antigens, also known as cancer–testis antigens, which are expressed in germ cells with immune privilege and expressed in certain cancers, such as the GADE, CTAG, SSX and MAGE gene families [7]; (3) viral proteins, expressed by viruses such as hepatitis B virus (HBV), Epstein–Barr virus (EBV) and human papilloma virus (HPV); (4) over-expressed proteins that are expressed at very high levels in some cancers but at very low levels in some normal tissues, including EpCAM- and prostate-specific antigen (PSA); and (5) differentiation antigens, which are expressed by particular cell types, such as GP100 and melan-A in melanoma [7].

Three factors have been attributed to the transition of cancer development from the “equilibrium” to the “escape” phase [1]: (1) the accumulation of mutations that cause the cancer to grow more aggressively and become “invisible” to the host immune system [1, 5]; (2) the development of host immune tolerance that leads T cells to become unresponsive and unable to contain the cancer; and (3) the development of a cancer suppressive microenvironment that results in immune effector cell dysfunction [1, 8, 9]. Recently, breakthroughs in anticancer immunotherapy have been achieved by direct delivery of immunity, i.e., the administration of engineered T cells [such as tumor infiltrating lymphocytes (TILs) and chimeric antigen receptor modified T cells (CARTs)] and antibodies (eAbs) [such as bispecific antibody (BsAb) capable of retargeting T cells and immune checkpoint blocking antibodies (Ab) for unleashing T cells] [10–14]. Unprecedented success has been achieved using this approach in terms of the proportion of the treated patients experiencing complete remission, or even cure, indicating that this strategy overcomes the barrier of immune suppression, at least partially. However, these successes are limited, with a substantial proportion of the treated patients experiencing only transient remission or complete failure to respond. The perspectives presented here summarize these promising technologies, define the mechanisms that form the bottleneck to these immunotherapeutic advances and propose the synthetic immunity (SI) strategy as a systematic approach to further overcome the cancer-related suppression mechanism and bring cancer immunotherapy to a new level of efficacy.

2 Recent breakthroughs in cancer immunotherapy

Recently, encouraging progress has been reported in anticancer clinical trials of eAbs and T cell engineering technologies. As the foundation of the SI strategy, these technologies are summarized briefly as follows:

2.1 Monoclonal antibodies (mAbs)

mAbs have been used for treating cancers for decades because they are able to bind the targets on cancer cells with high specificity and affinity [15, 16]. In addition to inducing antibody-dependent cell-mediated cytotoxicity (ADCC) [17], mAbs can also be used to destroy cancer cells through the delivery of conjugated toxins and isotopes. The application of traditional anticancer antibodies has been well summarized in recent reviews and is therefore not described here [15, 16]. Recently, a class of so-called immune checkpoint blockades, including the anti-programmed death protein-1 (PD-1) and anti-cytotoxic T lymphocyte-associated antigen 4 (CTLA-4), has demonstrated promising results in clinical trials of the treatment of multiple solid cancers, such as melanomas, prostate cancers and even the extremely aggressive non-small cell lung cancers [14, 18]. PD-1 and CTLA-4, which are expressed by immune cells such as T cells, dendritic cells (DC), nature killer (NK) cells, negatively regulate T cell activation by stimulating the intracellular signaling pathways leading to cellular inactivation upon binding to their ligands PD-LI/2 and B7, respectively. Several cancers express these ligands as an escape mechanism to inactivate CTLs and avoid destruction. Monoclonal antibodies (mAbs) that block these ligand–receptor interactions overcome this escape mechanism and retain the capacity of CTLs to destroy the cancer cells. These results confirm the presence of multiple inactivated anticancer CTL clones in patients’ tumors, which could be reactivated to perform the function of immune surveillance and herald a new era in antibody-mediated anticancer immunotherapy. Many clinical trials are being conducted to determine the efficacy of this approach in other cancers [14, 18]. These mAbs have demonstrated moderate off-target effects and are undergoing optimization [7, 19]. To date, the US FDA has approved the use of an antibody against CTLA-4 for the treatment of melanoma [20].

2.2 Tumor infiltrating lymphocytes (TILs)

TILs are collected from surgical tumor samples, amplified in vitro and reinfused into the patient. In a recent study, a group of 93 patients with metastatic melanomas were treated with this technique; 50 % responded to the treatment, and 20 % survived for 64–109 months [11, 21]. More recently, the same group of researchers reported that a woman with cholangiosarcoma responded well to the infusion of TILs containing a high proportion of mutated ERBB21P-specific CD4+CTLs [22]. This observation suggests that TILs are also effective for the treatment of solid cancers of epithelial cell origin, which account for approximately 90 % cancer deaths in USA. These results further confirm that cancer-specific CTL clones exist in vivo and are able to destroy the cancer cells following in vitro expansion and reinfusion into the patients. Because TILs are autologous T cells, this approach is associated with few side effects. However, to prolong the persistence of the TILs in vivo, patients have to undergo a “preparative lymphodepletion” procedure, consisting of chemotherapy alone or in combination with whole-body irradiation prior to TIL infusion, and the administration of the T cell growth factor IL-2 treatment after cell delivery. These results are in accordance with histological studies demonstrating that the presence of TILs in cancer tissues is associated with improved survival in patients with prostate, breast, colorectal and ovarian cancer or melanoma [11]. Clinically, however, although TILs are highly effective in treating patients with melanomas, this approach achieves only limited efficacy in patients with renal cancers and is ineffective for the treatment of other tumors [21, 23]. The mechanism underlying these disparate responses is not known at present. More clinical trials are in progress in an attempt to expand these early successes [21] although the technique used to scale-up TIL production remains to be optimized.

2.3 Chimeric antigen receptor modified T cells (CARTs)

CARTs are generated by replacing the extracellular portion of the TCR of the CTL with a single-chain antibody (mAb); therefore, the chimeric antigen receptor is also called a TCR-like antibody [24]. This changes the T cell specificity such that an Ab–antigen interaction occurs instead of the usual receptor–ligand (TCR-MHC-peptide) interaction, leading to enhanced target killing activity. Currently, CARTs are generated by transducing T cells with lentiviral, retroviral or transposon vectors expressing the chimeric antigen receptor. The transduced T cells are then expanded in vitro before being infused back into the patient [25]. Compared with the earlier technology, such as the recombinant TCR (rTCR), which is a natural TCR mimic, CARTs have many advantages, including the following: (1) CARTs can be more easily produced using well-established Ab engineering technology; (2) CARTs are also able to target he unconventional B cell epitopes and kill tumor cells in both MHC-dependent and independent manners, while MHC presentation is required by rTCRs; (3) CARTs have a much higher binding affinity and hence higher therapeutic efficacy than CTLs and rTCRs [26]; (4) the third generation CARTs, which are constructed with elements capable of generating strong intracellular signals, exhibit greatly enhanced cancer cell killing efficiency [27, 28]. In addition to hematological malignancies, CARTs are also effective in melanoma and synovial cell sarcoma [21]. The first case of successful cancer treatment using this technology was reported in a patient with advanced B lymphoma treated with a CART targeting CD19. The patient underwent a partial remission that lasted for 32 weeks after the CART infusion [29]. The efficacy of CARTs is further confirmed by their ability to eradicate kilograms of leukemia in a few weeks in patients whose cancers were refractory to chemotherapy [30]. In a recent report, CART therapy resulted in complete remission in 55 patients out of a group of 75 with B cell malignancies [31]. Multiple CART clinical trials are being conducted for the treatment of sarcoma, kidney and other epithelial cancers [21]. However, CARTs can induce severe side effects, including the massive overproduction of cytokines known as “cytokine storm”, which can be fetal and a related death was reported in an earlier trial [32]. CART technology is also limited by high cost and difficulty in scaling-up production. To overcome these obstacles, an RNA-based CART has been developed recently, in which mRNA is used to replace the viral vector to generate shorter-lived CART populations. This modified CART technology, which has reduced off-target toxicity, reduced production costs and improved the ease of scale-up, is currently being tested in several clinical centers [25, 32].

2.4 Bispecific antibodies (BsAbs)

BsAbs, also known as bispecific T cell engagers (BiTEs) [10, 33], are a class of antibodies capable of binding two antigens simultaneously. They were made initially by chemically mediated cross-linking of two antibodies [34] or quadroma technology and are currently generated using antibody-engineering technology [35, 36]. When used in cancer immunotherapy, BsAbs function as TCRs, i.e., forming a synapse between the T cell and cancer cell to initiate killing of the target [37]. BsAbs are highly effective in anticancer therapy. The typical BsAb blinatumomab, which binds to both CD3 on T cells and CD19 on B cells, was shown to result in regression of B cell malignancies in all seven patients treated at a daily dose as low as 90 µg (in a 60-kg human) [10]. However, it is associated with two drawbacks: (1) With a half-life of only 2–3 h in circulation, a minipump is required for continuous drug delivery to maintain an effective concentration throughout the course of the treatment; (2) it kills both normal B cells and their precursors, although the loss of normal B cells is not lethal and its consequences are manageable. Furthermore, this problem can be avoided by using tumor-specific targets [38]. The first EU FDA-approved BsAb, catumaxomab, has demonstrated excellent therapeutic results in clinical trials for the treatment of malignant ascites [39]. Catumaxomab is a trifunctional BsAb; in addition to binding the T cell CD3 and the tumor epithelial cell adhesion molecule EpCAM, it includes an Fc region capable of recruiting DCs, NKs and macrophages expressing FcγR receptors. It has been demonstrated that this achieves a coordination of adaptive and innate immune elements, and consequently, the induction of the long-term memory immune responses desired for cancer immunotherapy [40]. Catumaxomab is effective in treating ascites resulting from almost all abdominal cancers overexpressing EpCAM, including gastric, pancreatic and ovarian cancers [41]. Compared to the TIL and CART technologies, BsAbs have many other advantages that will be summarized later.

In summary, these successes represent breakthroughs in the long history of human anticancer strategies. Importantly, these successes are achieved not by the induction of immunity, but by the delivery of immunity in the form of engineered Abs or T cells that mediate the therapeutic effects. In contrast, only a small fraction of the numerous cancer therapeutic vaccine clinical trials have demonstrated objectively determined effects [11]. These observations suggest that there are significant obstacles to the generation of effective vaccine responses in cancer patients and that the delivery of immunity is the most effective way to overcome cancer-related immune suppression. However, the successes to date are far from satisfactory, with responses seen only in a subset of patients, and many experiencing only an incomplete and transient regression or a complete failure to respond [10, 11, 14, 18, 21, 29]. In order to break down the bottleneck and make further progress in the field of cancer immunotherapy, it is important to understand the mechanisms underlying the breakthroughs and remaining obstacles.

3 Cancer heterogeneity at the center of immune escape

It is suggested that patients have demonstrated different responses to the treatments because of the heterogeneity of TIL populations comprising cells in different differentiation states, which are inactivated through different mechanisms; therefore, it is likely that different strategies will be required to restore their cellular function [8]. This hypothesis explains the incomplete response to treatment with in vitro expanded TILs and the immune checkpoint inhibitors that work by relieving the suppression of TILs, but not the success of the CART and BsAb technologies. CARTs are made by recombinant vector-mediated expression of target molecules in T cells derived from peripheral blood mononuclear cells (PBMC), while BsAbs work by conferring in vivo CD3-positive T cells with cancer-specific cytolytic (CTL) function. Thus, both T cell types are derived from active T cells capable of killing cancer cells without the need for an activation step. Given that all four treatment strategies target only one cancer antigen, it is reasonable to hypothesize that the genetic heterogeneity of cancers is the major mechanism underlying the incomplete responses and is also responsible for the failure of therapeutic vaccination.

Cancer and virus heterogeneity is one of the major mechanisms underlying their escape and treatment failure. With almost no exceptions, an HIV infection comprises multiple clones [42, 43], and a cancer, although derived from one single cell, comprises multiple cell lineages that continue to evolve [2, 3, 44–47]. Actually, a cancer is a micro-ecosystem subject to selection pressure and evolutionary processes similar to those that shape ecosystems in nature as described by Darwin [3, 4, 48, 49]. In an unfavorable environment, many cells die and only the fittest survive and thrive. Selection pressure comprises many factors [50, 51], including hypoxia due to poor circulation resulting from the rapid growth of cancers, and acidity resulting from aerobic glycolysis—the so-called Warburg effect [52, 53]—that occurs in almost all solid cancers (Supporting Information Box 1). Immune editing is another important selection factor [1, 5], where the host immune surveillance eliminates the emerging cancer cells with strong antigenicity, leaving only weakly antigenic cells that escape detection and elimination. Therapeutic drugs are also powerful selection factors [3]; most of the current targeting treatments, including small molecules (e.g., tyrosine kinase inhibitors) [54–57], antibodies [15, 16] and engineered T cells [10, 12, 22, 58], target only one cell lineage [54, 55]. Consequently, only partial or transient remission is achieved, and although some treatments do result in a longer-term remission, relapse occurs eventually [48, 57, 59]. One such example is the interesting observation that in multiple cancer patients treated with anti-PD-L1, the preexisting lesions regressed while new lesions emerged [60]. The fundamental mechanism underlying these phenomena is clear—only the targeted lineage is destroyed, while the untargeted lineages survive and grow continuously. Consequently, a cancer can evolve before and after treatment, and metastatic cancers can differ from the parental tumor [2, 44, 48, 57]. This hypothesis is schematically illustrated in Fig. 1a. The same mechanism may be responsible for the failure of therapeutic vaccination. Conceivably, a therapeutic vaccine will stimulate immune responses that kill the targeted cancer cell lineages, but not the untargeted lineages, which will continue to thrive. However, the reasons for the failure of therapeutic vaccination may be much more complicated. This hypothesis is supported by a recent report that a vaccine designed to protect against HIV infection resulted in an increase in the risk of HIV infection, with some vaccine-induced antibodies promoting the infection [61–63]. Other studies also revealed unexpected immune responses targeting particular subsets of DCs that induced immune tolerance [64], or inhibition of tumor-specific CTLs [65, 66].

Fig. 1
figure 1

The synthetic immunity strategy. (a) A cancer comprises multiple cell lineages and mutation-derived new lineages continually emerge. Treatments, including small molecule drugs, antibodies and T cell-based immunotherapies and therapeutic vaccines, destroy only the targeted cell lineage(s) while allowing the untargeted cells to continue growing, and rendering these treatments ineffective [44, 55]. Red circles indicate the drug-targeted cell lineage; blue and black circles, the untargeted lineages; green circles, newly emerged cell lineages. (b) A minicircle can be used to produce mAb or BsAb; the mAb can kill cancer cells by inducing antibody-dependent cellular cytotoxicity (ADCC), while the BsAb can bind cellular surface antigens of the T cell and tumor cell simultaneously and trigger tumor cell killing; (c) A personalized composition of minicircles encoding mAbs and BsAbs (i.e., the eAb-vectors) as determined by reference to the patient’s cancer epitope-tree (Fig. 2a) are used to transfect T cells, which express mAbs and BsAbs in vivo to kill tumor cells via ADCC or to retarget the resting T cell to kill the tumor cell selectively

Full size image

It has been well documented that genetic variations, including single-nucleotide polymorphisms (SNP), insertion, deletion and inversion, play important roles in cancer initiation, promotion, progression, metastasis and response to anticancer therapies. A large number of genes are involved, including oncogenes, tumor suppressor genes and genes involved in DNA repair, drug metabolism, J-V-D recombination and signal transduction pathways. For example, the efficacy of the antibody trastuzumab (Herceptin®) in breast cancer treatment is greatly influenced by the FcγRIIIA-158 polymorphism [65]. These will not be described in detail because this class of genomic alterations occurs before cancer initiation and is not within the scope of present review of cancer heterogeneity and immunotherapy.

4 Synthetic immunity (SI) for further breakthroughs

It is apparent that only technologies capable of coping effectively with the heterogeneity and dynamics of cancers will lead to further breakthroughs. Accordingly, a SI strategy is proposed to achieve this goal. The principle of SI involves the generation of a panel of engineered antibodies (eAbs), mAbs and BsAbs, aiming at killing all the cell lineages of a cancer with no possibility of escape. BsAbs confer T cells with cancer-specific CTL function [10], and mAbs disable virus infections or induce the death of cancer and virus-infected cells through Ab-induced ADCC [15, 16]. To ensure that none of the cell lineages escape, the eAbs are designed under the guidance of an epitope-tree illustrating the clonal antigen architecture of the cancer; this issue will be addressed later.

The power of SI will be enhanced further by a series of technologies including (1) optimization of the binding affinity and avidity of the eAbs, for either effector or cancer targets, to enhance functional efficiency compared with that of the natural Abs or TCRs [21]; (2) using technologies such as vaccination to increase the number of BsAb-retargetable T cells and hence the cancer-specific CTLs with predetermined efficiency (e.g., EBV-specific CTLs [67] or cytomegalovirus (CMV)-specific CTLs) [68]; and (3) using gene and cell engineering technologies to tap into the unmined riches of T cell epitopes to expand the immunotherapeutic repertoire of the SI. This possibility is illustrated by a recent anti-HIV vaccination study [42], in which monkeys were vaccinated using a virus vector expressing a gene from the simian immunodeficiency virus (SIV), which is the counterpart of human HIV. This resulted in CD8+ T cell responses to a wide range of SIV epitopes presented by both MHC-I and MHC-II, covering 66 % of the viral protein, which is threefold greater than that achieved by conventional vaccination. Importantly, these unconventional CTL responses cleared the SIV infection, which was resistant to the conventional vaccine-induced MHC-1-restricted CD8+ T cells [42]. Interestingly, it has also been demonstrated that the anti-melanoma effects of anti-CTLA-4 were largely mediated by broadening of the T cell repertoire [69]. These findings are important because they suggest that the immune system may be more plastic than previously thought and that SI may allow exploitation of a far greater number of cancer- and pathogen-specific peptide epitopes than those targeted by natural T cells. A complete understanding of the technology to exploit this class of antigens may lead to more powerful immunotherapy [7, 70].

In summary, SI represents an effective strategy to overcome the obstacles faced by current immunotherapies and therapeutic vaccination and to elevate the treatment of cancers and infectious diseases to a new level of efficacy.

5 DNA vector-based SI

SI can be administered in different ways. A combination of engineered antibodies (eAbs) and CARTs, for example, can be used to generate the same set of effector eAbs and T cell clones as those induced by vaccination. Alternatively, eAbs alone will also achieve this by using mAbs to mimic neutralizing Abs and BsAbs to generate effector T cells (Supporting information Box 2).

The present perspectives propose another solution, that is, to produce both types of eAbs using optimized non-viral DNA vectors, minicircles, either ex vivo or in vivo (Fig. 1b). Minicircles are a class of DNA vectors encoding almost solely the gene expression cassette and are generated by eliminating the bacterial backbone DNA elements from standard plasmids [71]. Plasmid backbone DNA is associated with multiple detrimental effects, such as gene silencing and the induction of inflammatory responses to unmethylated CpG motifs. In the absence of plasmid backbone DNA, minicircles express high levels of gene products both in vitro and in vivo [73]. Currently, there are multiple technologies capable of producing high-quality minicircles with ease [71–73, 75]; consequently, minicircles are widely used by researchers in the biomedical community and are commercially available worldwide.

Compared to antibody-based SI, minicircles have several advantages: (1) it has a huge cost advantage in that minicircle production, purification, storage and transportation incur only a minor fraction of the costs associated with the use of antibodies; (2) technically, it will be easier to generate a library comprising thousands of minicircles encoding the whole spectrum of anticancer eAbs; (3) for the same reason, it will be easier to scale-up minicircle production to meet the demands of the huge patient populations, even at the level of millions. As such, minicircle-based SI has the potential to be available to all patients worldwide.

Other advantages of minicircle-based SI include its safety and flexibility. Many US FDA-approved non-viral DNA vectors are already available to the market, confirming the safety of minicircles [76]. In addition to the expression of eAbs, minicircles can be used to provide more functions of SI; for example, through expression of short hairpin RNA (shRNA) to block the pathways leading to T cell apoptosis and anergy, it will be possible to generate longer-lived and unsuppressible T cell populations [77].

Although non-viral DNA vectors have many advantages compared to viral vectors, including safety and cost, the problem of delivery has prevented their widespread use. However, the breakdown of this barrier has begun in the form of the emergence of technologies for the transfection of cells with nucleic acids both in vitro [74] and in vivo [78, 79] that are being evaluated in clinical trials. Although B lymphocytes are the sole cell type capable of antibody production, the feasibility of ectopic production by other cell types has already been tested [80]. In addition to PBMC or derived products such as dendritic cell/cytokine-induced killer cell (DC-CIK) for ex vivo applications, many other cell types, such as muscle or liver sinusoid endothelial cells, have the potential to mediate the expression of eAbs by minicircles. Furthermore, the use of cell types with different life spans provides a convenient control to the duration of eAb expression. We have constructed a series of minicircles encoding anticancer BsAb and determined their capacity in mediating cancer cell killing in vitro and in treating B lymphoma in vivo (unpublished data). Nevertheless, there remain rooms for optimization of technology in minicircle DNA delivery, either ex vivo or in vivo, especially in transfection efficacy, nanomaterial toxicity and targeting accuracy.

In summary, minicircle-based SI has the potential to be developed as a safe, effective and affordable immunotherapeutic strategy that will play a major role in curing cancers and infections.

6 Future directions

In terms of mechanistic studies, it is critically important to catalog all somatic mutations to help identify the altered signaling pathways involved in the pathogenesis of cancers [81]. For immunotherapy, however, it is important to catalog all the cancer epitopes and to determine the epitope-tree mapping cancer clonal antigen architecture (Fig. 2a). This map will guide the selection of an optimal set of “trunk or major branch” epitopes to ensure that the panel of eAbs will lead to a cure by killing all the cancer cell lineages without severe toxicity (Figs. 1c, 2b, c). Conceivably, this systematic approach is an effective way to reveal the “Achilles’ heel” of cancers and to organize a powerful and effective attack. Fortunately, epitope-tree construction will be facilitated by its resemblance to a phylogenic tree and can be constructed using the exome/genome sequences of the paired cancer and normal samples; initial progress in this area has already been made [81]. Currently, the International Cancer Genome Consortium (ICGC) has recorded almost nine million somatic mutations (Release 16, June 4, 2014), while The Cancer Genome Atlas (TCGA) reports more than three million (as of January 23, 2014) [82]. The total number of cancer epitopes will be markedly fewer, because only a small percentage of the mutations in these databases will fall into the protein encoding regions, and only a fraction of the mutated proteins will be immunogenic. Furthermore, in most studies documenting malignant heterogeneity, it has been shown that the driver mutation genes are present at the trunk of the tree with few exceptions [2, 44]. Approximately 140 driver mutations have been identified [2], although most of the epitopes may come from passenger mutations [1, 7]. The subsequent step, which involves the construction of the library of antibodies targeting the mutated epitopes (Fig. 2b), will be more challenging. Although the technology to develop antibodies targeting B cell epitopes, which are usually cell surface antigens, is mature, the high throughput technology required to generate TCR-like antibodies, which interact with intracellular antigens presented by MHC systems, is not fully developed. However, this goal should become achievable once a worldwide collaboration is established similar to those that have led to the success of the human and cancer genome projects.

Fig. 2
figure 2

Synthetic immunity and the cancer epitope-tree. (a) Epitope-tree of a cancer. In this model, the tumor comprises six cell lineages (P1–6) with five driver mutations (M1–4, M6) and six tumor-specific epitopes (E1–6). Each driver mutation starts a new cell lineage with an additional growth advantage. A new cell lineage, for example M1E2/P6, can also start with the loss of a driver mutation. Consistent with the high ratio of passenger to driver mutations, which can be as high as 2,000:1, most of the mutated antigens are derived from passenger mutations and only a few from driver mutations [31, 44, 45]; therefore, the two biomarkers are not necessarily linked. E, epitope; M, driver mutation; N, normal cell; Circle, cell lineage; Triangle, epitope; the number inside indicates the sequential number of either mutation (circle) or epitope (triangle). (b) The four steps of the SI strategy: (1) construct a cancer epitope library from cancer mutation databases; (2) translate this into an eAb library; (3) construct a library of minicircles encoding the eAbs; (4) use the patient’s cancer sample to determine a patient-specific epitope-tree and, with reference to this information, select a personalized minicircle set. (c) Deliver the minicircles via ex vivo methods (e.g., PBMC) or in vivo to express the mAb and BsAbs that mediate the destruction of the targeted cancer or virus-loaded cells

Full size image