Programmable cancer treatments: Engineering biology approaches for living cures

Abstract

Living cures are cell‐based, programmable therapies that integrate the latest learnings in immunology and synthetic biology. Although Adoptive Cell Therapies (ACTs) have transformed the treatment landscape of haematological malignancies by harnessing the powerful anti‐tumour properties of immune cells, commercialisation and ensuring access is challenging. Their application in solid tumour treatment has been hindered by the immunosuppressive tumour microenvironment (TME) and its associated physical barriers. Conversely, bacterial immunotherapies offer cost‐effective solutions by utilising tumour‐colonising bacteria that trigger localised inflammatory responses within the TME. The authors briefly examine advancements in ACT and propose bacterial immunotherapies as an alternative or complementary treatment modality with potential use either as standalone therapies or in conjunction with other treatments.

Living cures are cell‐based, programmable therapies that integrate the latest learnings in immunology and synthetic biology. Bacterial immunotherapies offer cost‐effective solutions by utilising tumour‐colonising bacteria that trigger localised inflammatory responses within the tumour microenvironment.

1. INTRODUCTION

Cancer is a devastating disease that primarily manifests as solid tumours, abnormal masses of heterogenous cells that grow uncontrollably and establish their own host‐independent microenvironment. The prevailing standard of care comprises surgical resection, followed by chemotherapy or radiotherapy, both of which are associated with debilitating side effects. Immunotherapy has emerged as an alternative therapeutic approach, focusing on targeted strategies to train the patient's immune system to identify and eradicate malignant cells. This approach has resulted in prolific development of new therapies, such as monoclonal immune checkpoint inhibitor antibodies [1, 2]. A particularly exciting area that combines the latest learnings in immunology and synthetic biology is the use of live cells, either human‐derived or bacterial, as cancer therapeutics. These living cures are programmable through engineering biology techniques to enhance their interaction with the immune system. Adoptive Cell Therapies (ACTs) have leveraged the potent anti‐tumour properties of cytotoxic immune cells to develop highly targeted therapies that revolutionised the treatment landscape of haematological malignancies. However, the use of ACT in the treatment of solid tumours, which represent 90% of all diagnosed cancers [3], has yet to yield positive outcomes, partly due to the immunosuppressive nature of the tumour microenvironment (TME) and its associated physical barriers. Moreover, industrial manufacturing of adoptive therapies remains expensive, restricting access only to a small proportion of the world's population. Bacterial immunotherapies in contrast are cost‐effective treatments that harness the solid tumour‐colonising properties of certain bacteria to induce a local, broad inflammatory immune response that stimulates the recruitment of endogenous cytotoxic immune cells into TME. Whilst the first bacterial immunotherapies are attributed to William B. Coley [4], who in 1891 performed foundational experiments by injecting different mixtures of live‐attenuated Serratia marcescens and Streptococcus pyogenes into patients' tumours, in the past decade, the bacterial immunotherapeutic space has seen a resurgence spurred by advancements in our understanding of tumour immunology and ability to read and manipulate bacterial genetic information. In this review, we discuss recent work and emerging trends in the living cures space for cancer treatment. We briefly summarise advancements and ongoing challenges in both adoptive mammalian cell and bacterial immunotherapies, contrasting the different attributes of these two alternative approaches and exploring the potential transformative impact the combination of adoptive cell and bacterial therapeutic strategies can have to maximise efficacy and patient benefit by blending the malleable and tumour‐colonising properties of bacteria with the powerful responses of adoptive cells.

2. ADOPTIVE CELL THERAPIES

In cancer immunotherapy, ACT predominantly refers to the use of autologous (from the same patient) or allogeneic (from a healthy donor) immune cells to target tumour cells [5, 6]. Whilst the use of various immune cells has been explored (e.g. lymphocytes, dendritic cells, or macrophages), adoptive T cell therapy (Figure 1) currently stands out as the most prevalent therapeutic approach [7]. T cell therapy can be categorised into three main applications: tumour‐infiltrating lymphocytes (TIL), chimaeric‐antigen receptor T cells (CAR‐T), and T cell receptor T cells (TCR‐T).

FIGURE 1.

Adoptive cell therapies. Adoptive T cell therapy relies on the targeted recognition of cancer cells by natural or engineered T cell Receptors (TCRs) or engineered Chimeric Antigen Receptors (CARs). They can be classified into three major categories: (1) Tumour‐infiltrating lymphocytes can be isolated from the patient’s tumour and re‐introduced into the patient; (2) or the TCR of tumour‐antigen reactive lymphocytes could be sequenced, its expression optimised, and transduced into T‐cells (TCR‐T) prior to infusion. The last approach (3) involves using synthetic receptors (CARs) that can recognise tumour‐associated or tumour‐specific antigens. This approach has been successful in the treatment of blood cancers and is actively being researched for its use in solid tumours. CAR, Chimeric Antigen Receptors; TCR, T cell Receptors.

The autologous treatment of TIL involves the identification and ex vivo expansion of naturally occurring tumour‐specific cytotoxic T cells, followed by their re‐infusion into the patient. The expansion, mediated by exogenous addition of IL‐2 [8, 9], results in a pool of cells with multiple T cell Receptor (TCR) clones, providing an effective response against cancer heterogeneity. However, natural infiltration of T cells inside solid tumours can be limited in many cancers [10] and recovered T cells are often dysfunctional and exhausted by the immunosuppressive nature of the tumour microenvironment [11]. Moreover, the personalised manufacturing that is required for TIL therapy presents significant logistical hurdles. This, together with the short half‐life of infused TILs [12], has resulted in slow progress of TIL therapies to the market, with the first Food and Drug Administration approval only having recently been granted [13]. Biological engineering‐based approaches have been developed to address such limitations: the TCR and Chimaeric Antigen Receptor (CAR) technologies.

TCR‐T, either autologous or allogeneic, are engineered to express a TCR specific for tumour‐specific antigens (TSA), molecular signatures only found in tumours, or tumour‐associated antigens (TAA), molecular signatures that are over‐represented in tumour tissues [14]. TCRs are heterodimeric proteins, each having a unique variable domain and a constant domain [15]. The unique variable regions can recognise specific peptides presented by mammalian cells, leading to T cell activation. The TCR‐T therapy process [16] involves isolation of lymphocytes from either healthy donors or cancer patients. The antigen‐responsive lymphocytes are isolated upon expansion and their corresponding TCR is identified through sequencing techniques [17, 18]. After selection of the most promising TCRs, the TCR molecules can undergo optimisation [19] and are loaded onto T cells for validation and therapeutic lead selection and development.

CARs are engineered synthetic receptors [5] that function to direct autologous or allogeneic T cells (CAR‐T), or in some instances also Natural Killer cells (NK, [20]) or macrophages [21], to recognise and eliminate cells expressing and displaying a specific antigen, triggering potent anti‐tumour responses. They generally rely on four main components [5, 22]. The extracellular binding domain, built of a single‐chain variable domain derived from a monoclonal antibody, is the element responsible for specific antigen recognition. It is linked to the hinge region, a flexible element that connects the binding domain with the transmembrane domain, a poorly characterised region that anchors the CAR to the T cell. Finally, the intracellular signalling domain [23] provides the activating signal(s) upon ligand engagement. The signalling domain is the most engineered element of the CAR, having undergone up to four different generations of improvement by adding up to two co‐stimulatory elements (e.g. CD28 and/or 4‐1BB) and nuclear factors that induce cytokine production upon stimulation (4th generation, T cell redirected for antigen‐unrestricted cytokine‐initiated killing (TRUCK), [24]). A fifth generation of CAR‐Ts was engineered to contain additional membrane receptors that respond to antigen stimulation [25].

Prior to administering therapeutic T cells, patients must undergo treatment aimed at reducing the number of recipients’ lymphocytes, known as lymphodepletion, which is critical for the ability of donor cells to engraft and expand in the patient. However, it leads to toxicities and immune suppression [26, 27], and there is currently no agreement on the combinations of drug and treatment regimes, with cyclophosphamide and fludarabine being the most commonly used compounds [28, 29]. Additionally, ACTs themselves have been linked to significant toxicities in the clinic. Most common adverse effects are those associated with supraphysiological cytokine production (also known as cytokine‐release syndrome) and immune effector cell‐associated neurotoxicity syndrome [30]. Some of the toxicities are thought to be mediated by IL‐6 and current management involves administering IL‐6 receptor blockade (tocilizumab and corticosteroids) [31, 32]. Further unpredictable toxicities can be caused by on‐target off‐tumour T cell activation, triggered by recognition of TAAs in healthy tissues [5]. Engineering approaches have been proposed for enhancing the safety profile by adding small molecule‐responsive switches [33] or by manipulating the affinity of the CAR‐T cell antigen binding domain, which must remain in a range that allows target binding but prevents CAR‐T disengaging from dying tumour cells, causing trogocytosis (excision of antigen from cancer cell to CAR‐T [34]) or exhausting the T cells.

Cancer is a dynamic disease that evolves as it progresses, contributing to tumour heterogeneity [35]. This leads to post‐treatment relapses caused by antigen escape, where target antigen expression is abolished or reduced to levels below the T cell threshold of activation [5, 36]. This effect has been observed to be more pronounced in CARs, with reports of 30%–70% relapse in patients with acute lymphoblastic leukaemia, where 10%–30% had lost CD19 [37]. A strategy to overcome this limitation has been explored using bi‐specific CARs [38, 39, 40]. A recent example of CD19/CD20 bi‐specific CARs [41] introduced into naive and memory T‐cells [42] resulted in safe and effective responses against non‐Hodgkin lymphoma.

While ACTs have been successfully deployed in the treatment of haematological malignancies, efficacy in the treatment of solid tumours remains elusive [43, 44]. The challenges in developing T cell therapies for solid tumours stem from the immunosuppressive nature of the tumour microenvironment that prevents T cell expansion and persistence as well as tumour‐associated physical barriers, such as cancer‐associated fibroblasts [45]. Ongoing research efforts to overcome these barriers include creating CARs that rely on the co‐administration or production of immunostimulatory cytokines [22, 23] and expression of enzymes that degrade the tumour‐associated extracellular matrix, such as heparinase [46].

Finally, manufacturing challenges remain a barrier to widespread adoption of these therapies. Processes are costly, time‐consuming, laborious, and can affect the functionality of the final therapeutic cellular product [47, 48, 49], with high batch‐to‐batch variability that is unavoidable due to the use of cells from different donors to initiate manufacturing. For autologous treatments in particular, patients generally possess low numbers of lymphocytes due to cancer‐induced immunosuppression and previous treatments (e.g. with chemotherapy) and experience high vein‐to‐vein treatment times between T cell harvest and therapy injection.

3. BACTERIAL IMMUNOTHERAPIES

Bacterial immunotherapies (Figure 2) leverage the latest advancements in synthetic biology to develop cost‐effective approaches that take advantage of the tumour‐homing and immunostimulatory capabilities of some bacterial strains to reprogramme the TME from immunosuppressive ‘cold’ to triggering a ‘hot’ immune response. These therapies build upon the observation that bacteria administered to hosts bearing tumours can accumulate preferentially in the TME in ratios of up to 10,000:1 when compared to healthy tissues such as the liver or spleen [50, 51, 52]. It is thought that such refined targeting mechanisms derive from the leaky vasculature in the TME and the ability of bacteria to exploit the unique features of the tumour, such as low pH [53], enhanced nutrient availability [54], or a hypoxic core [55]. When established in the TME, shedding of microbe‐associated molecular patterns (MAMPs), such as flagella or lipopolysaccharide (LPS) [56, 57, 58], can trigger a broad stimulation of the immune system, resulting in the infiltration of immune effector cells such as NK cells or T cells (e.g. CD8⁺) [58]. These findings have led to the development of several bacterial strains as cancer therapeutics, including Clostridium [55], Listeria [59], Escherichia coli (e.g. Nissle strain) [60] and Salmonella [61]. Pathogenicity and bacteraemia are prevented by introducing attenuating mutations that limit the bacterial ability to replicate in healthy tissue, restricting growth to the tumour. Selection of appropriate attenuating mutations is critical as over‐attenuated microorganisms will fail to elicit an appropriate immune response and under‐attenuated bacteria can lead to toxicity. Obligate anaerobes, such as Clostridium, which can only survive in the necrotic regions of the tumour, require attenuation by the removal of any potentially produced bacterial toxins, such as the α‐toxin in C. novyi‐NT [62], the most advanced Clostridium tested in humans, or nine gene clusters with high sequence similarity to the Streptolysin S operon from S. pyogenes in Clostridium sporogenes‐NT [63]. Facultative bacteria, such as Escherichia or Salmonella, colonise the tumour through more complex mechanisms and thus require additional attenuating mutations. The Nissle strain, a naturally occurring probiotic isolated in 1917, is the most researched E. coli strain for cancer treatment, including clinically tested strain variants [64]. Salmonella enterica is perhaps the most researched species overall for cancer treatment, with the auxotrophic strains S. Typhimurium VNP20009 (deletion mutant of purI, involved in purine biosynthesis, and msbB, involved in LPS biosynthesis) [51] and S. enterica Typhi ZH9 (deletion mutant of aroC, involved in aromatic acid biosynthesis, and ssaV, involved in bacterial replication) [65] having been safely administered to humans. An alternative approach to ensuring lack of pathogenicity and viable bacterial replication is complete inactive bacteria, known as bacterial ghosts, which have been constructed through the inducible expression of lysin E from φX174 [66], yielding the bacterial cell envelope which retains all the inflammatory elements of the bacterium but no intracellular material (e.g. genomic information). Decoration of the membrane with selected cargo proteins fused to outer membrane proteins prior to bacteriolysis can be used to refine the immunostimulatory response [67]. In some instances, bacterial ghosts have been loaded with chemotherapies (e.g. doxycycline [68]) or plasmid DNA for transfer into macrophages [69].

FIGURE 2.

Bacterial immunotherapies. Bacterial therapies are cost‐effective treatments that leverage the tumour‐colonising properties of certain bacterial species, which can result in up to 10,000:1 ratios of enrichment in the malignant tissue. In the tumour, the established microbe can shed a series of immunostimulatory molecules that reprogramme the properties of the tumour microenvironment and promote the infiltration of anti‐tumour immune cells. Ease of engineering and high loading capacity of these bacteria allow the production of synthetic payloads, either on a plasmid or integrated into the chromosome, that can be delivered precisely into the tumour after sensing of tumour‐associated metabolic or physicochemical signatures. Risk of adverse events is minimised by introduction of attenuating mutations or the construction of bacterial ghosts. Bacterial therapies could provide the required immunostimulatory components, either through shedding of inflammatory molecular signatures or recombinant payloads, that promote infiltration and sustain the activity of ACT. ACT, Adoptive Cell Therapies.

The new generation of bacterial immunotherapies builds upon the foundational technology of attenuated bacteria by creating improved synthetic strains with enhanced tumour‐colonising capabilities and safety profile. This can be achieved by controlling the expression of essential genes [70] or payload production [71] through specific promoters activated by the differential attributes of the tumour versus healthy tissue or after bacteria reach a certain population threshold (i.e. quorum sensing, [72]). Such approaches can suffer on‐target off‐tumour misfiring, necessitating the exploration of external triggers, such as the arabinose‐inducible promoter (P_BAD, [57, 73]), the lac promoter [74], light‐inducible promoters [75] or heat‐inducible systems [76]. These complex circuit control strategies are in their infancy and have yet to demonstrate utility in the human context but hold great promise for enhanced precision and specificity of bacterial immunotherapies. While individual promoter‐regulated circuits can be affected by leaky expression, combination of one or more gene circuits into Boolean genetic logic gates is poised to overcome such limitations [77]. Another mitigating approach has been to increase the bacterial tumour tropism by overexpressing bacterial motility components [78], encapsulating the microbes in polysaccharide capsules [74], or displaying tumour‐targeting peptides or antibody‐like molecules on the bacterial surface [79, 80, 81, 82].

Such precise bacterial therapeutics can contain engineered payloads to selectively deliver therapeutic molecules directly into the tumour, pairing cargo with target cell biology. The host invasive properties of intracellular bacteria, such as Salmonella or Listeria, have been leveraged to deliver nucleic acids, plasmid DNA or RNA, in a process known as bactofection [83, 84]. Therapeutic DNA is loaded into strains as plasmids that are released into the host cytosol upon bacteriolysis and transported to the cell nucleus. These plasmids, which typically cannot replicate in the host cell, encode TSA [83, 85] or silencing RNA [86, 87, 88, 89] downstream of strong viral promoters. However, the delivery of such molecules can be inefficient, mainly due to the rate‐limiting step of nuclear entry. RNA bactofection overcomes this obstacle [90]. A recent publication by Singer et al. demonstrates that bactofection can effectively deliver not just small silencing or epitope‐encoding RNAs but also large functional sequences, and in this case, an oncolytic virus under bacterial control, showcasing the enormous potential of bactofection as a targeted delivery vehicle for a range of RNA therapeutics [91].

Approaches whereby therapeutic proteins are expressed directly by live‐attenuated bacterial carriers have also garnered increasing attention in the recent years, with most examples using S. enterica and E. coli Nissle as bacterial chassis. Salmonella has been highlighted [54, 73] as a unique chassis due to its tendency to accumulate in tumours, its programmability, and its cell invasive nature. The initial steps of the Salmonella invasion cycle rely on the use of a needle‐like structure known as SPI‐1 [92], a type III secretion system that injects effector proteins into the host cell and can be repurposed for the delivery of heterologous payloads [93, 94, 95, 96, 97]. Internalised Salmonella establishes inside a Salmonella‐containing vacuole (SCV) that protects itself from intracellular defence mechanisms [98, 99] and enables replication [100]. Induction of bacterial lysis after Salmonella's establishment in the SCV has been shown to allow payload delivery in murine breast cancer models [73]. Escape from the vacuole can be forced by expression of Listeriolysin O [101], a pore‐forming cytolysin from Listeria monocytogenes, or by using the sifA mutant strain [102], which causes disruption of the vacuole during its maturation. Examples of intracellular payloads which have been explored include intrabodies [103] and bacterial toxins, including chimeric anticancer toxins [104]. Non‐invasive strains have also been engineered to inject protein payloads into eukaryotic cells through heterologous type III secretion systems [105] or, more recently, contractile injection systems [106] and can enable payload deployment throughout the TME by fusing targets with secretion signal peptides [107] or through synchronous release triggered by quorum sensing [60, 108, 109, 110]. Examples of extracellular payloads include cytokines [107, 110, 111, 112, 113], antibody‐like molecules [108, 114] and immunostimulatory proteins (e.g. FlaB [57]).

Engineered bacteria overcome the manufacturing and logistical drawbacks of engineered mammalian cell therapies, offering a relatively straightforward fermentation manufacturing process [115, 116] with the potential for lyophilisation to produce off‐the‐shelf therapies stable at standard refrigerator (2°C–8°C) or even room temperatures [117], as is well established for probiotic and bacterial vaccine products, or to produce bespoke therapies in as little as 3–4 days compared to the 3–5 weeks required for autologous ACT [118]. However, most examples of bacterial therapies in the literature do rely on the use of bacterial plasmids to incorporate engineered functions, which can lead to sub‐optimal final product performance derived from inherent cell‐to‐cell variability [119] and increased metabolic burden [120, 121, 122]. Moreover, a large portion of these plasmids rely on the use of antibiotic resistance cassettes [123] or addiction strategies [124, 125, 126] that can be incompatible with use in the clinical setting. Integration of synthetic cassettes into the chromosome is a strategy deployed to circumvent these limitations, although insulation from the surrounding genomic context and minimisation of polar effects are critical to achieve robust strain performance [127]; and positional effects (i.e. gene location in the genome) can impact the total payload yield [127, 128, 129, 130].

4. FUTURE PROSPECTS

Both ACTs and bacterial immunotherapies have harnessed recent advancements in immunology and engineering biology to bring new treatment opportunities to cancer patients. Adoptive cell therapies have achieved success in the treatment of haematological malignancies but to date yielded disappointing results in solid tumours which represent the vast majority of cancers. Bacterial immunotherapies are ideally positioned to address this area of unmet need due to their immense engineering potential and intrinsic tumour targeting properties, offering alternative approaches.

The recent advances in engineering biology capabilities open up opportunities not just for new monotherapies but also for the integration of engineered bacterial therapies with current standards of care and other classes of immunotherapies in development. Such bacteria‐coupled regimes could, for example, restore immune fitness in immune‐compromised patients ahead of other treatments, potentially addressing the major issue of insufficient responders or refractory patients common to many therapies. Bacteria‐dependent immune reprogramming could also directly synergise with therapies like ACT in various ways. For example, colonisation of live‐attenuated bacteria in tumours before adoptive cell infusion could induce profound changes in the TME immune landscape due to exposure and shedding of MAMPs, or through expression of engineered cytokines and chemokines, to promote and sustain engineered immune cell infiltration and activity. Additional bacterial payloads could also degrade tumour‐associated physical barriers to improve immune cell penetration. Safety profiles of ACT could be enhanced by coupling their activity to bacterial therapies that either produce an activating payload, ensuring synthetic immune cells are only triggered in appropriate tissues and thus reducing on‐target off‐tumour effects or that produce a safety switch to prevent the development of serious adverse effects. Ease of engineering and high circuit loading capacity of bacterial strains can also provide opportunities for the introduction of both such circuits on a single strain, providing a highly dynamic control of living cures therapies. Finally, the ability of bacterial therapies to colonise tumours independently of the need for specific targeting molecules can be leveraged to develop ACT treatment regimens that respond to the bacteria instead of the tumour. These beacon approaches, whose initial activity is independent from specific TAA or TSA antigens, have the potential to overcome the tumour heterogeneity hurdle and offer the opportunity to develop off‐the‐shelf bacterial and adoptive cell combination therapies that can be manufactured and distributed ahead of time, thereby reducing overall treatment costs and timelines.

Bacterial cancer therapies have thus far achieved clinical translation only in the form of Bacillus Calmette‐Guérin treatment for bladder cancer, but other live attenuated strains such as Prokarium's Salmonella Typhi ZH9 (NCT06181266) or T3 Pharmaceuticals' Yersinia enterocolitica (NCT05120596) in clinical development promise to establish regulatory pathways that will facilitate adoption of engineered bacterial therapies in the future. Due to the simple, scalable and inexpensive fermentation manufacturing processes of bacterial therapies, this transformative approach promises to democratise access to life‐altering treatments to patients worldwide, making significant strides towards equitable healthcare access on a global scale.

AUTHOR CONTRIBUTIONS

Marc Biarnes Carrera: Writing – original draft. Alexandra Sevko: Writing – review & editing. Nicholas Glanville: Writing – review & editing. Livija Deban: Writing – review & editing.

CONFLICT OF INTEREST STATEMENT

At the time of submission, all authors were full time employees of Prokarium Ltd and received funding from Prokarium Ltd in the form of salary. This does not alter adherence to journal policies.

Biarnes Carrera, B. , et al.: Programmable cancer treatments: engineering biology approaches for living cures. Eng. Biol. 8(2-3), 31–40 (2024). 10.1049/enb2.12032

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no new data were created or analysed in this study.