Date: 4th June 2020
Cancer immunotherapies modulate and leverage the host immune system to treat cancer. The past decade has seen historical advancements in such therapeutics however, with an estimated 6.9 million people dying of cancer in 2018 worldwide, there is still much work to be done in the fight against the disease. Now scientists have used non-pathogenic bacteria as a versatile platform for the development of a living biotherapeutic for the treatment of cancer.
Scientists from Synlogic Inc – a clinical stage company applying synthetic biology to microbes in order to create living medicines – have engineered a bacterial strain (E coli Nissle), referred to as SYNB1891, to clear tumours and stimulate antitumor immunity in preclinical models of cancer.
The team led by Jose Lora initially chose E coli Nissle (EcN) as their bacterial chassis; a non-pathogenic bacteria with established safety profiles in humans and high serum sensitivity, it is susceptible to a broad range of antibiotics. In addition to this, with a defined genomic landscape that is permissive to genome engineering, this strain was an ideal choice to work with.
Previous work had shown that EcN can colonise a wide-range of tumours upon intratumoural delivery. So to test EcN as an oncology therapeutic vector the team injected immunocompetent mice harbouring various different tumour types with the non-engineered bacteria. The bacteria showed a rapid expansion in all tumour types from 100-1000 fold, and the EcNs reached a steady state between 24-72 hours, remaining within the tumour itself. They also persisted in the tumour for up to 14 days and were metabolically active during that period. Interestingly during this period, the team also observed a significant delay in tumour growth compared to saline injection controls, suggesting that these bacterial already conveyed a moderate level of anti-tumour activity.
With EcN established as a viable oncolytic targeting system in mouse models, the researchers next wanted to use the system to deliver a sustained and relevant immunological therapeutic payload. They chose to engineer the bacteria with a STING (STimulator of INterferon Genes) agonist. The STING pathway plays a critical role in the initiation of an anti-tumour immune response via the activation of APCs (antigen presenting cells), in the production of type 1 interferons (IFNs) and the presentation of tumour antigens. The pathway bridges innate and adaptive immunity making it a good potential therapeutic target, and so the team engineered the bacteria to produce cyclic di-AMP (CDA) – a stimulator of the STING pathway. By monitoring IFNs which are the cytokines that help regulate both innate and adaptive immune systems the team could assess the bacteria’s function.
The scientists engineered the bacteria with three different CDA-producing enzymes; one in particular showed high expression in vitro so this enzyme was carried forward into an in vivo setting.
When EcNs expressing this enzyme were injected into mice bearing tumours which were typically refractory to immunotherapy (they were poorly inflamed) there was an increased in production of CDA resulting in an increase of IFNβ1 (a type 1 IFN) induction when compared to non-engineered EcN. More importantly, a significant decrease in tumour growth could be observed by eight days post-treatment initiation. Whilst innate cytokines were seen in the first 24 hours of treatment with either engineered or non-engineered EcNs, only the engineered version showed a shift in expression to T-cell associated cytokines at 8 days post dose initiation, suggesting that only the engineered version could induce T-cell antitumor immunity in vivo.
There are strict safety and regulatory guidelines that clinical therapeutics must adhere to and a biotherapy should be stable during the manufacturing process. To address this the team, introduced biocontainment controls by engineering a dual safety mechanism, through two distinct auxotrophy genes, thyA and dapA. By generating this as a double mutant strain of EcN, this prevented extra-tumoural and intratumoural proliferation respectively
In addition, the team removed antibiotic resistance genes which ensured the bacteria could be cleared by available antibiotics if required. An additional and final level of control was inferred by the addition of inducible promoters to the CDA enzyme so that enzyme production could be induced only when required. The group tested a range of inducible promoters and found that the hypoxic-driven fumarate-and-nitrate reductase (PfnrS) promoter provided the highest level of payload induction without the requirement for any chemical addition as tumour microenvironments are already hypoxic.
The modified EcNs harbouring these mutations and genetic elements were named SYNB1891.
The team initially tested the treatment by applying SYNB1891s to murine bone marrow-derived dendritic cells (BMDCs) in vitro, and showed that this resulted in increased synthesis of inflammatory cytokines (which activate the immune system) relative to the application of EcN controls.
The team next wanted to determine whether phagocytosis was an important mechanism for SYNB1891s so they inhibited this process using an inhibitor cytochalasin D. They observed that pre-treatment of BMDCs or macrophages with the inhibitor and subsequent exposure to SYNB1891s resulted in a significant reduction in the number of bacterial cells observed within the phagocytic cells, and subsequently lead to a reduction in the expression of IFNβ1.
Whilst this data strongly supported SYNB1891 as a biotherapeutic, the real test was to evaluate the system in vivo, therefore a single SYNB1891 injection into melanoma tumours of mice was delivered and the pharmacokinetics and pharmacodynamics were assessed.
SYNB1891 treatment significantly delayed tumour growth and resulted in undetectable tumours in some animals. Intratumoural production of CDA was also detected, confirming the functionality of the bacteria. In addition, Type I IFNs and proinflammatory cytokines were induced in a dose-dependent manner. In addition as it was designed, the dapA (auxotrophy) mutation allowed the bacteria to be cleared over time in the tumour microenvironment, and SYNB1891s were not detected in the blood at any time (assessed by colony forming unit (CFU) assay).
Overall, the data suggested that SYNB1891 treatment could function in a pharmacological dose-dependent manner, remained localised to the tumour, and could induce localised inflammation and had anti-tumour properties.
The treatment showed a strong “one-off” short-term response but cancer immunotherapies need to be effective over long period of time. So the team increased the injection regime to three doses in a week and then monitored the mice for up to 60 days.
SYNB1891 treatment resulted in a dose-dependent, significant delay in tumour growth compared to saline injection controls, and achieving durable tumour rejections in 30–40% of the animals. Injection of non-engineered EcN did lead to some tumour reduction but never in tumour rejection. Furthermore, whilst SYNB1891 treatment lead to an increased mouse survival rate of 40% at 30 days compared with 0% of saline controls (at 14 days), and 0% of EcN controls (at 18 days), treatment of the mice with a STING agonist resulted in only a 10% survival rate after 30 days.
An addition tumour type was tested, B cell lymphoma, and here at the highest dose of SYNB1891 treatment around 80% of the mice showed tumour rejection. The team showed that this was dependent on T cells, and in particular CD8+ T cells were required for the efficacy of the treatment. Finally, SYNB1891-treated animals which remained tumour-free for 60 days, were re-challenged with the B cell lymphoma, and remarkably all cured mice remained tumour-free after re-challenge – suggesting the formation of protective immunological memory.
In this study the team have exploited synthetic biology techniques and drug development criteria to design and create a highly engineered bacterial strain capable of localised, targeted STING activation and a product suitable for manufacturing and evaluation in human clinical trials.
The resulting biotherapeutic, SYNB1891, can produce high levels of CDA and is able to induce potent type I IFN production in a phagocytosis-dependent manner in mouse APCs. They also extended the mouse work and showed that SYNB1891 could also induce the STING pathway in human APCs making this a viable translatable system in the future as a clinical therapeutic.
From a safety perspective, the fact SYNB1891 remained localised to the tumour site, and therefore the cytokine response was also restricted to the tumour makes this a strong candidate for a biotherapy. Furthermore, the presence of a dual biocontrol mechanism makes for a very attractive clinical tool.
Indeed, intratumoural administration of SYNB1891 is currently being evaluated as a monotherapy in an ongoing Phase 1 dose escalation clinical trial (NCT04167137) in patients with advanced solid tumors or lymphoma. Synlogic are hoping to release data from the monotherapy arm of this study later this year.
The harnessing and engineering of bacteria to deliver drugs and therapeutics is a growing area of exploitation. Whether it is engineering the organism to produce molecules that can stimulate the immune system as we’ve seen here, or as we saw a few weeks ago the engineering of bacterial secretion (injection) systems such that cargo is released under a light-controlled molecular switch it is hoped that this biohijacking of organisms will provide fruitful translations into a clinical setting.
This work presented here, is already on the clinical path and we will be waiting with anticipation for the data to emerge.
For more information please read the press release from Synlogic or find them on our synthetic biology company map page.
Leventhal, D. S., A. Sokolovska, N. Li, C. Plescia, S. A. Kolodziej, C. W. Gallant, R. Christmas, J.-R. Gao, M. J. James, A. Abin-Fuentes, M. Momin, C. Bergeron, A. Fisher, P. F. Miller, K. A. West and J. M. Lora (2020). “Immunotherapy with engineered bacteria by targeting the STING pathway for anti-tumor immunity.” Nature Communications 11(1): 2739.