As mentioned in the previous section, BCG has been used for decades in the treatment of superficial bladder cancer. Instillation of BCG into the bladder allows the bacteria to contact the urothelium and bladder cancer cells themselves on the luminal surface. The mycobacterial fibronectin attachment protein interacts with host fibronectin through integrin α5β1mediating attachment (19, 20, 34, 35). Fibronectin is a part of the extracellular matrix (ECM) distributed on normal and malignant urothelium. The fibronectin attachment protein of BCG has a high affinity receptor for the collagen domain of fibronectin stimulating the internalization of BCG into tumor cells. Fibronectin serves as a bridging molecule to both urothelial cells and BCG working with other molecules like heparin sulphate containing proteoglycans that interact with the mycobacterial heparin-binding hemagglutinin adhesin (20, 34). In other words, fibronectin is not mandatory but rather enhances adherence of BCG to the bladder wall. Data suggests that the addition of anti-fibronectin antibodies can impair BCG binding to the murine bladder. In addition, the clinical effects of BCG therapy can be related to the degree of fibronectin expression (34).

Following internalization, BCG is broken down into various peptides and proteins causing phenotypic alterations within the cancer and bladder cells that trigger the host immune system. These peptides are processed and associate with major histocompatibility complex (MHC) class II molecules that get expressed on the surface of professional antigen presenting cells (APCs) or MHC class I molecules on the surface of urothelial tumor cells (18, 20, 34, 35). The release of pro-inflammatory cytokines such as interleukin (IL)-1, IL-2, IL-6, IL-8, IL-12, tumor necrosis factor (TNF)-α, and interferon (IFN)-γ also occurs following antigen presentation. Cytokines cause further recruitment of immune cells such as CD8+ and CD4+ T lymphocytes, macrophages, and neutrophils to induce a Th1 cytotoxic response and enhance cancer cell recognition (20, 22, 34).

IL-2, IL-6, IL-8 and IFN-γ are particularly important because they can “fine tune” the generation of APCs and activation of T-cells which contribute to antitumor effects. IL-6 stimulates T cell proliferation, macrophage maturation, cytotoxic T-cell differentiation, and activation of resting T cells via interactions with IL-2 and IL-2 receptor expression. In cancer, IL-6 induces natural killer (NK) cell proliferation which contributes to the MHC-nonrestricted cytotoxic activity of those cells. Dendritic cells (DCs) and macrophages produce IL-8 which has strong chemotactic properties that attract T lymphocytes, neutrophils, and other innate effector cells to induce antitumor effects (24, 34).

More recently, scientists recognize the importance of IL-17. IL-17 induces neutrophil infiltration in the bladder producing chemokines such as MIP-1α and GRO-α for CD4+ T-cell migration and generates important Th1-cell responses (23, 34, 35). Luo et al. show that this Th1-cell response is vital for BCG treatment particularly to produce IFN-γ. While IFN-γ appears to be a late responder in the cytokine milieu, it has a role in driving the activation of CD8+ cytotoxic T lymphocytes (20, 36). CD4+ T-cells and cytotoxic CD8+ T cells mediate the Th1 cell response by IL-2, TNF, IL-12, and IFN-γ secretion. NK cells drive anti-tumor immunity through Th2 responses via IL-4, IL-5, IL-6, and IL-10. This immune cascade from these cytokines causes direct anti-tumor activity (20). The cytokine milieu during BCG therapy is of recent interest. There is evidence suggesting that urinary IL-2 levels as well as IL-6, IL-8, IL-12, TNF-α, and IFN-γ are promising markers in predicting BCG clinical responses (19, 22) suggesting cytokines as a potential treatment for cancer. High-dose IL-2 for example is approved for the treatment of metastatic renal cell cancer and melanoma with many others like IL-12 and IFN-γ under investigation (37). Since relapse typically occurs within the first five years of treatment with BCG (34), combination treatments with cytokines could enhance efficacy and side effect reduction.

As discussed in the previous section, Coley and physicians over the years explored the use of S. pyogenes in the treatment of cancer. While Coley used heat-inactivated bacteria for his studies, live S. pyogenes has several adaptations that provide the possibility for future cancer treatments such as the selective recognition and attachment to oncofetal fibronectin via streptococcal collagen like protein 1 (Scl1) (38–40), as well as a variety of immunomodulatory proteins and enzymes.

The Scl1 protein is anchored to the cell wall of S. pyogenes in homotrimeric organization (40). It consists of a central collagen-like (CL) region comprised of Gly-Xaa-Yaa repeats, a non-collagenous N-terminal V-domain, and a C-terminal cell wall-associated region (38, 39). The CL region and V-domain have significant implications in cancer. The central CL domain allows for a triple helical formation that mimics that of mammalian collagen. This similarity supports direct interactions with collagen binding integrins such as α₂β₁ and α₁₁β₁ facilitating internalization by human cells (38, 40–42). The V-domain is important in the stabilization of a triple helical conformation of the CL domain (43, 44). It directs Scl1 recognition and binding to the cellular fibronectin type III repeats, the extracellular domains A (EDA) and B (EDB) – also known as oncofetal fibronectin – deposited by cancer associated fibroblasts (CAFs) (39, 40, 45, 46). Solid tumors have extensive ECM deposition due to stromal CAFs (47–49). These CAFs deposit EDA/EDB-containing fibronectins in the tumor microenvironment (TME) where Scl1 mediates adhesion by GAS (39).

There is some indication that Scl1 may modulate other aspects of host immunity such as neutrophil infiltration and neutrophil extracellular trap formation. Neutrophil extracellular traps (NETs) have gained recognition for their role in inflammatory disease and cancer. NETs occur when activated neutrophils release their intracellular contents including DNA, histones and granules into the tumor microenvironment or circulation. NETs can promote the migration and invasion of cancer cells, mediate the epithelial-mesenchymal transition, awaken dormant cancer cells, and facilitate metastasis (50–52). While NETs are detrimental in cancer progression, they provide necessary protection against bacterial infection (53). However, bacteria deploy various virulence factors to evade NETs. S. pyogenes has several mechanisms for this (54–56). While limited research shows the specific mechanistic role of Scl1 in NET inhibition, Dohrmann et al. show that scl1 mutant strain of GAS is more susceptible to phagocytic killing by neutrophils and induce more NETs in comparison to wild-type or complemented strains. Evidence from this study suggests that surface Scl1 can hinder the release of myeloperoxidase (MPO) from NETs and directly protect bacteria from cathelicidin LL-37, a molecule that induces pore formation in bacterial membranes (54).

The M1 protein of GAS can elude NET killing by promoting proinflammatory conditions and evading phagocytosis. M proteins facilitate GAS adherence to fibronectin which can cause an immune cascade of neutrophil infiltration leading to NET release. However, the M1 protein confers resistance to NET killing by cathelicidin LL-37 release following phagocytic degranulation (57). GAS strains also secrete streptolysin O (SLO), a toxin belonging to the cholesterol-dependent cytolysins family that promotes resistance to neutrophil clearance (56). SLO binds to the cholesterol lipid rafts in the eukaryotic cell membrane and forms large pores. Typically, this is not cytocidal unless in large quantities (58). Rather, SLO blocks neutrophil oxidative burst and IL-8 responsiveness leading to resistance of neutrophil killing (56). Although the benefits of SLO in cancer responsiveness are limited, Hall et al. report that SLO can reduce breast cancer invasion and cell growth by activating the ErbB1 kinase, a protein encoded by the epidermal growth factor receptor gene involved in cell signaling and proliferation (59). Therefore, future cancer studies involving SLO and molecular signaling will be beneficial.

DNase expression is another way that bacteria, including GAS, can escape neutrophil killing. DNase Sda1 is a secreted endonuclease that provides protection against the host innate immune response. Specifically, Sda1 can degrade the DNA component of NETs as well as that seen in the extracellular traps of macrophages. Evidence suggests that Sda1 may suppress the TLR9-mediated immune response by degradation of CpG-rich DNA (55, 60). Therefore, targeting NETs via bacterial mechanisms such as a DNase-based therapy may be a useful approach in inhibiting NET formation and release during cancer.

S. typhimurium is extensively studied for its potential application in cancer treatment. For one, S. typhimurium is a facultative anaerobe, allowing it to adapt to both the hypoxic and nonhypoxic zones within a tumor. Second, proper flagella constructs, active motor function, or signal transduction proteins are necessary for tropism to the tumor region where the bacterium can encounter an aspartate receptor, serine receptor, or a galactose or ribose receptor (25). From here, S. typhimurium uses a type-3 secretion system to infect the cancerous cells, colonize the tumor, and activate antitumor immunity (25, 61). However, the tumor cells are not directly destroyed by the bacteria in this manner. Rather, antigen presenting cells (APCs) present bacterial antigens targeted by anti-salmonella T cells (61).

APCs can activate the adaptive immune response against S. typhimurium in different ways. Salmonella bacteria are gram negative with a lipopolysaccharide (LPS) on the outer membrane. LPS is a strong microbial associated molecular pattern (MAMP) molecule that can activate toll-like receptor 4 (TLR4), stimulating the release of different proinflammatory cytokines like IL-6, IL-8, TNF-α, and matrix metallopeptidase (MMP)-9 (27, 62). This overexpression of IL-6 plays a major role in inflammatory diseases and cancer (63). The bacterial flagellum can interact with TLR5 and mediate innate immunity in a similar way, causing elevated cytokine expression and subsequent activation of nuclear factor kappa B (NF-κB) and the Janus Kinase/STAT3 pathway, promoting the maturation of dendritic cells and enhancement of anti-tumor immunity (27, 62). The upregulation of pro-inflammatory cytokines can also lead to the production of IFNγ which can induce chemokines CXCL9 and CXCL10 to recruit NK cells, neutrophils, macrophages, and T cells to the tumor and stimulate acquired immunity for a tumor-specific response (25, 61). Alternatively, Salmonella can induce robust immune responses to APCs via connexin 43 (Cx43) to active CD8+ T cells and enhance anti-tumor immunity. Cx43 is a ubiquitous protein that forms gap junctions in a variety of cell types. Salmonella can induce the upregulation of Cx43 and thus enhance gap junction communication, leading to processing of tumor peptides by DCs (25, 64). In the instance of cancer therapy, increased gap junction communication can enhance drug delivery by allowing a drug to penetrate through the complex cellular network seen in tumors (25).

Salmonella accumulation in the tumor can also stimulate inflammasome association and tumor apoptosis. Briefly, during inflammasome assembly, a MAMP or damage-associated molecule pattern like LPS interacts with a pattern recognition receptor such as TLR4, leading to activation of transcription factors that encode pro-IL-1β and pro-IL-18 (25, 65, 66). From here activation of various proteins, NLR family pyrin domain containing 3 (NLRP3) being the most famous, forms an inflammasome with or without an apoptosis associated speck-like protein containing a caspase recruitment domain (ASC). ASC recruits pro-caspase 1 or caspase-8 which leads to the maturation and secretion of IL-1β, IL-12, or Gasdermin-D to induce pyroptosis or apoptosis (66). Inflammasome formation can further lead to the activation and interaction of macrophages (65, 66). Macrophages as well as dendritic cells can also be important producers of IL-1β and TNF-α, suggesting that cytokine therapy in addition to Salmonella therapy could be of potential value in cancer therapy (67).

Autophagy is another mechanism important for cancer progression. Autophagy is a cellular metabolic recycling process that allows cancer cells to survive the nutrient deprived, hypoxic tumor microenvironment. In pancreatic cancer, autophagy is hyperactive leading to reduced apoptosis and increased proliferation of cancer cells (68–71). The protein kinase B (AKT)/mammalian targets of rapamycin (mTOR) pathway is a significant regulator of autophagy consisting of a cascade of autophagy-related (Atg) proteins encoded by ATG genes (69, 71). PDAC cells tend to maintain increased levels of ATG expression leading to higher degrees of basal autophagy activity. This hyperactivity promotes enhanced inflammation and PDAC progression particularly in KRAS-driven tumors (69, 70). Salmonella can downregulate AKT/mTOR signaling leading to activated autophagic pathways and thus inhibit tumor growth in certain cancers (25, 71). In pancreatic cancer, the role of autophagy in progression is context dependent, with autophagy generally considered to aid tumor progression once an established tumor has formed. However, greater induction of autophagy has also been shown to promote an autophagic cell death (68). Yet autophagy inhibition induces T-cell and neutrophil infiltration into the TME which causes subsequent pro-tumorigenic inflammation (72). This dual role of autophagy makes the process a challenging therapeutic target. Because of the capacity of Salmonella to regulate both apoptosis and autophagy in addition to the aforementioned anti-tumor immune activation, attenuated Salmonella is of great interest for bacterial-mediated cancer therapy. A study conducted by Wang et al. showed that combination treatment of hydroxychloroquine, a prominent agent that decreases autophagy activity, and Salmonella VNP20009 was associated with longer survival, shorter tumor growth, and enhanced treatment responses in melanoma (73). Therefore, a combination treatment of an autophagy inhibitor like hydroxychloroquine with Salmonella could overcome the limitations of standard monotherapy.

Listeria is a gram-positive intracellular facultative anaerobe which is known for its association with foodborne illness in humans (74). However, over the past few decades, Listeria is being investigated in vaccine development, particularly against cancer. Listeria monocytogenes (Lm) elicits strong innate and adaptive immune responses making it a desirable candidate in cancer therapy.

Because Listeria localizes within the cytosol, it can directly activate APCs allowing antigen presentation without disseminated infection (27, 74–76). Lm entry into a variety of host cells is complex and involves bacterial cell surface internalins, internalin A (InlA) or internalin B (InlB), which mediate Lm entry via receptor-mediated endocytosis. However, Lm developed mechanisms to evade killing and survive in the cytosol. In the blood, Lm expresses the transcription factor PrfA which activates genes encoding virulence factors like listeriolysin O (LLO), a cholesterol-dependent pore-forming cytolysin similar to SLO discussed previously (56, 74). LLO forms a β-barrel-pore-like structure in the phagosomal membrane allowing Lm to escape into the cytosol. Here, Lm secretes peptides that get degraded by proteosomes and loaded onto MHC class I molecules on APCs resulting in the induction of CD4+ and cytotoxic CD8+ T cell responses (74, 75).

Alternatively, retinoic acid-inducible gene I (RIG-I) recognizes Lm RNA which can lead to the transcription of IFN genes (74). The cyclic GMP-AMP synthase (cGAS) – stimulator of interferon genes (STING) pathway further contributes to transcription of IFN genes (77). Briefly, Lm exposes genomic DNA recognized by cGAS (74, 77). cGAS binds to the dsDNA and forms dimers resulting in the synthesis of 2’3’ cyclic GMP-AMP (cGAMP) which binds to STING and interacts with trafficking factors to pass through the ER-Golgi complex. STING then recruits TANK-binding kinase I (TBK1) which phosphorylates STING and interferon regulatory factor 3 (IRF3) which leads to the expression of type I IFNs to induce pro-inflammatory mediators and priming of CD8+ T cells in the tumor microenvironment (TME) (74, 77). These CD8+ T cells are specific to the tumor associated antigens (TAAs) found within the TME and secrete a variety of cytokines and cytotoxic molecules to induce apoptosis in the tumor (78).

One of the most critical aspects of bacterial mediated cancer therapy is the capability of bacteria to localize specifically in the tumor environment (17). Lm is capable of directly infecting myeloid derived suppressor cells (MDSCs) found within primary tumors. MDSCs in cancer patients can inhibit both T-cell and NK responses. Listeria can combat this by infecting MDSCs within the TME and spreading into tumor cells (79, 80). From there, Lm can induce an immunostimulatory phenotype on the MDSCs promoting the production of IL-12 and improving T-cell and NK responses (27, 76, 80). Lm can also directly kill tumor cells by activating NADP+ oxidase and producing reactive oxygen species and activating Listeria-specific cytotoxic T lymphocytes (CTL) (27, 75, 76, 81). Kim et al. show that this CTL response contributes to tumor cell regrowth and shrinkage in breast cancer models, making them an important player in cancer mediated therapy (81). The capacity of Lm to both directly and indirectly kill cancer cells allows multiple applications and opportunities for bacterial mediated cancer therapy.

Clostridium is a gram-positive, spore-forming, obligate anaerobic bacteria associated with severe human disease. Clostridium butyricum, C. tetani, C. histolyticum, C. beijerinckii, and C. acetobulyicum are being observed as anti-cancer agents. One of the best studied Clostridium species in cancer therapy is the bioengineered C. novyi. Because of the spore-forming capabilities, Clostridium is of much interest. Spores can encounter the hypoxic tumor regions, germinate, and become active, leading to tumor regression. Furthermore, the vegetative bacteria cannot spread into normal tissues due to higher oxygen levels between tissues providing tumor specificity and localization to the hypoxic tumor microenvironment (17). Of great interest in cancer therapy, Clostridia strains can be engineered to express cell suicide genes – such as bacterial enzyme cytosine deaminase - or eukaryotic host molecules like TNF-α (61). For example, C. novyi- non-toxic (NT) expressing RecA, a protein responsible for DNA repair and maintenance, induces host immunity and the expression of pro-inflammatory cytokines and chemokines leading to tumor regression (61).

Several mechanisms exist for Clostridium to provide these tumor destructive forces. Clostridium has three lipid-degrading proteins: phospholipase C and two lipases. These enzymes alter the structure of lipid bilayers and membrane permeability of the tumor cells leading to direct cellular cytotoxicity. Phospholipase C can also activate inflammatory responses which can induce anti-tumor immunity through CD8+ T cells. Alternatively, Clostridium can modulate innate immune responses. Clostridium spores can induce strong inflammatory response involving pro-inflammatory cytokines such as IL-6 which signals immune infiltration by different APCs such as DCs. These DCs can release danger signals that recruit tumor associated antigen specific T cells that induce immunogenic cell death of the tumor cells (17).

A major drawback to bacterial mediated cancer therapy is the risk of disseminated infection in immunocompromised patients. Live-attenuated vaccines like BCG or Listeria are potent in their abilities to activate the innate and adaptive immune systems and induce anti-tumor responses (34, 74). Yet, immunocompromised individuals have a propensity for adverse effects. During BCG therapy, some patients require additional prophylactic pharmacotherapies to combat this issue. While this can help reduce the risk of potential infection, the Food and Drug Administration (FDA) is critical about proper antibiotic stewardship (20). Engineering less virulent bacteria could ameliorate this risk but efficacy wanes. Heat-killed Lm strains for example are favorable cancer vaccine candidates for immunocompromised individuals. However, these heat-killed strains do not induce a robust CD8+ T cell response adequate for anti-tumor effects. Killed but metabolically active Lm vaccines, those which contain attenuated pathogens via photochemical inactivation but still retain metabolic activity for sufficient induction of immune responses (82), can counter this problem but are still considered experimental and not in a clinical trial phase (74).

Proper dosage of agents also poses a challenge. The balance between tolerable, efficacious, and toxic is subtle. Many current and proposed bacterial therapies are not long-lasting and require maintenance to be effective. For example, BCG-induced immunity in bladder cancer is short in comparison to the lifelong risk of tumor recurrence with around 30-50% of patients failing to respond to therapy or relapsing within the first five years of treatment (20, 34). To combat this issue, there are efforts to induce a more robust immune response using bacterial components like the BCG cell wall (BCG-CW). However, the BCG-CW has a negative surface charge making interacting with the negatively charged glycosaminoglycans along the urothelium of the bladder poor (34, 35).

Sporulation therapy suffers a similar disadvantage with administration and dosage. Elimination of the phage carrying α-toxin that contributes to the toxicity associated with Clostridium makes C. novyi-non-toxic (NT) safe for use in cancer patients. Injection of spores into the hypoxic regions of the tumor promotes germination and subsequent regression of the tumor while minimizing the harmful effects of infection. However, if only a partial response is given, the tumors can regrow and become more resistant to C. novyi-NT therapy (17). The first human study involving intratumoral injection of C. novyi-NT spores (NCT01924689) resulted in three patients developing severe reactions such as sepsis and gas gangrene. Others had manageable toxicities such as rash at the maximum tolerated dose (83). Since the tumor only retains a small fraction of spores, a large dose is necessary; however, this poses the risk of side effects. Moreover, unless spores are delivered to the hypoxic regions of the tumor, the body clears the spores and germination does not occur (17). While this may limit toxicity, the treatment may be less effective in targeting tumors without necrotic areas (16).

Current bacterial mediated cancer therapies rely on a combination of bacteria or bacterial components with chemotherapy or radiation. Combination therapies to such a degree create increased toxicity and side effects which can be detrimental to immunocompromised individuals (17). Therefore, developing ways of delivering effective but safe doses without the need for combination bacteriolytic therapy is necessary to achieve the desirable results without adverse consequences.

Although bacterial mediated cancer therapy does pose several limitations as discussed in the previous section, current advances in recombinant technology and bioengineering allow the manipulation of bacteria and bacterial components to combat these issues. Clostridia strains can be engineered to express cell suicide genes – such as bacterial enzyme cytosine deaminase - or eukaryotic host molecules like TNF-α (61). For example, C. novyi- non-toxic (NT) expressing RecA, a protein responsible for DNA repair and maintenance, induces host immunity and the expression of pro-inflammatory cytokines and chemokines leading to tumor regression (61).

Engineering attenuated bacteria like Salmonella and Listeria to be less virulent can manage both dose toxicity and the risk of disseminated infection. The three most recognized attenuated strains in Salmonella therapy are VNP20009, AR-1, and ΔppGpp. VNP20009 contains a deleted purI gene and a deleted msbB gene to reduce the potential for septic shock and increase antibiotic susceptibility (25). Serovar VNP20009 shows premise with an increased capability to colonize tumors without serious toxicity or side effects shown in a phase I clinical trial (61) in addition to quick clearance from the bloodstream during experimental trial (25).

One of the main mechanisms theorized for the success seen in AR-1 is its ability to stimulate the progression of tumor cells from the G0/G1 phase of the cell cycle to the S/G2/M phase. Because most cytotoxic chemotherapies only recognize and kill tumors in the S/G2/M phase, this can increase the treatment sensitivity in resistant tumors (25).

Serovar ΔppGpp is a relA and SpoT double mutant defective in ppGpp synthesis, an important protein that plays a role in Salmonella drug susceptibility (25). ΔppGpp has implications in inflammasome activation, cytokine secretion, and heme, iron metabolism, and cytoskeletal structure within the tumor (25, 84). A recent study conducted by Yun et al. shows that tumor tissues from mice treated with ΔppGpp exhibit decreased cytoskeletal components increased iron and heme metabolism suggesting that bacteria proliferation can disrupt tumor tissues while also providing free heme and iron within the tumor environment to enhance bacterial growth (84) allowing for the potential of adequate bacterial colonization without the need for maintenance therapy.

Deletion of actA and inlB in Listeria produces the live-attenuated double-deleted (LADD) strain which is currently the primary candidate for vaccines employed in clinical trials because it exhibits reduced liver toxicity and rapid clearance from the liver and spleen (74). Alternatively, prfA gene deletions can severely attenuate Lm strains as seen in the XFL-7 serovar. Partial virulence can be restored to an immunogenic level with the addition of a plasmid encoded PrfA cloned with a tumor-associated antigen (TAA). The Δdal/Δdat (Lmdd) strain, deficient in the genes necessary to build peptidoglycan and dependent on alanine co-administration, can also regain partial virulence with the use of a plasmid without the addition of alanine. Both Lmdd and XFL-7 strains are currently being used in clinical trials (74).

Several combination therapies have been proposed using these modified organisms. For example, Listeria combined with the autophagy inhibitor chloroquine shows inhibited tumor growth, increased accumulation of Lm within the tumor, and higher incidences of cell apoptosis (85). Currently, the M.D. Anderson Cancer Center in collaboration with BioMed Valley Discoveries, Inc and Merck Sharp & Dohme LLC is investigating the antitumor effects of combination treatment of pembrolizumab and C. novyi-NT (NCT03435952) followed by management of Doxycycline on malignant neoplasms of the breast, digestive organs, urinary tract, head and neck, and genital organs (86). While these show promise, there is limited evidence suggesting increased efficacy in metastatic PDAC patients.

As discussed previously, combination therapies pose a risk for higher toxicity. To combat this issue, fusion-based vaccines are in development. Fusion-based vaccines utilize a bacterial protein fused with a TAA to enhance delivery of the TAA into the TME subsequently stimulating APC maturation and pro-inflammatory cytokine secretion and inducing a more robust anti-tumor response. One of the most notable fusion-based vaccines, CRS-207, uses the Listeria protein ActA fused with the TAA mesothelin. Overexpression of mesothelin is associated with many different cancer cell types including pancreatic cancer (74). Although the initial phases of CRS-207 proved safe and efficacious for PDAC patients, during phase II after receiving priming-boosts with GVAX and cyclophosphamide (Cy), patients showed no differences in mesothelin-specific CD8+ T cells responses in comparison to patients receiving Cy/GVAX alone or gemcitabine alone (74, 87). Currently, additional trials are attempting further prime-boosting using checkpoint inhibitors like αPD-1 and αCTLA-4 as well as indoleamine dioxygenase inhibitors (74). However, at this current time, there is little evidence suggesting the efficacy in treatment resistant patients.

Many new bacterial therapies have been proposed within the last decade with several showing potential in clinical trials. Still with the proposed risks associated with bacterial mediated cancer therapy, there is a need to enhance efficacy and improve overall patient response. Nano delivery of bacterial components, TLR receptors, and enhancing the microbiota are just a few routes of improvement.

Nanoparticles (NP) are gaining recent recognition for their ability to ameliorate toxicities and enhance drug delivery. Evidence indicates that NPs can increase drug delivery into the tumor while reducing bacterial toxicity on host cells (88). One group utilizes an R8-liposome which acts as an enveloped virus to improve BCG-CW delivery. This provides successful inhibition of tumor growth because the R8-liposome has a cell penetrating peptide that promotes the entry of BCG-CW into the cytoplasm of the bladder cancer cells (20, 35). In PDAC, NPs are designed as liposomes, micelles, and more recently mesoporous silica (MSNPs) to deliver drugs to the pancreatic tumor microenvironment (88, 89).

Currently, only a few drugs exist using nanotechnology. Albumin-bound paclitaxel (PTX) nanocarrier (Abraxane) is the first new drug for PDAC in over a decade. PTX improves the overall efficacy of gemcitabine by inhibiting the expression of cytidine deaminase in the stroma, thus preventing gemcitabine degradation (90). Meng et al. show that the effects of PTX and gemcitabine could be enhanced when packaged in a mesoporous silica nanoparticle (MSNP). Co-delivery of PTX and gemcitabine via MSNP exhibits more pronounced tumor shrinkage and inhibition of tumor growth with little evidence of toxicity (90). Liposomal irinotecan (Onivyde) was approved by the FDA in 2015 for use in metastatic PDAC. A combination treatment of Onivyde and 5-fluorouracil(5-FU)/Leucovorin shows increased survival averages of two months with a three-month delay in tumor growth in comparison to individuals who receive 5-FU/Leucovorin alone (88).

However, liposomal formulations have limitations. A major issue is systemic drug leakage and a loss of selectivity over time due to exposure of biomolecules in fluids like the blood (88, 89). These molecules form a biomolecular corona that has the capacity to reduce targeting capabilities of the NP (91). Liu et al. show that by loading irinotecan across a coated lipid bilayer in a MSNP carrier there is significantly less drug leakage and therefore less toxicity in comparison to the liposomal carrier. Moreover, because there is less drug leakage, there is a higher concentration of drug delivered to the tumor, thus improving drug efficacy (89).

Still, chemotherapy has toxicity and potential harsh side effects, impacting quality of life and leading to dose reductions and treatment delays. While bacteria can pose a threat to immunocompromised individuals, isolated bacterial proteins could produce positive results without these types of detrimental effects. Listeria and Salmonella both have intrinsic capabilities to kill cancer cells directly (27, 74, 76). Therefore, packaging a protein like flagellin or a specific bacterial antigen in a tumor could induce a cytotoxic T cell response directed specifically at the tumor without the toxicities associated with cancer drugs. A bacterial antigen in a NP carrier could also mitigate the risk of disseminated infection as active bacteria are not being injected into a tumor.

Another alternative could be the use of TLR agonists. TLR agonists may be beneficial in a combination treatment with BCG, for example. TLRs play a role in the innate and adaptive immune responses in urothelial cells with reduced levels suggested in bladder tumors. The addition of a TLR agonist could prevent relapse associated with BCG treatment (20). Because there are a variety of TLRs and associated pathogen associated molecular patterns (PAMPs), like flagellin and TLR5 (27), TLR agonists could provide customization towards the bacteria in use and corresponding desired anti-tumor response.

The tumor microbiome is gaining recognition for a potential role in influencing responses to cancer immunotherapy. The microbiome is found to be a predictive marker for immune checkpoint blockade (ICB) response (92–94). Responders to treatments not only show differences in microbiota composition in comparison to non-responders but they also show immune profiles associated with increased anti-tumor immunity (92, 93). Several studies demonstrate that in melanoma, patients who had positive responses to anti-PD-L1 therapy had a higher abundance of the members of the Ruminococcaceae family, as well as species of Bifidobacterium longum and Enterococcus faecium. Moreover, a bacterial transfer from these patients into germ-free mice generated higher anti-tumor immune profiles similar to the responders and better responses to anti-PD-L1 (92, 95). A phase 1 clinical trial (NCT03353402, Sheba Medical Center) conducted by Baruch et al. demonstrated that following fecal microbiota transplantation, recipients had an up-regulation of immune-related gene sets related to IFN-γ signaling and Th1-type immune responses (96, 97). These patients also showed an abundance of bacteria taxa that are associated with responses to anti-PD-1 and CD8+ T cell activation. These results suggest that fecal microbiota transplantation could be able to affect immunotherapy response such as that to anti-PD-1 therapy (97).

However, there is evidence to suggest that probiotic bacteria can trigger both favorable and unfavorable responses to ICB. Many patients take probiotics for health benefits. Several probiotic strains described previously are being investigated for anti-cancer benefits. Lactobacillus, for example, can exert immunomodulatory effects like increased production of IL-2 and IL-12, antioxidant activity like production of superoxide dismutase and glutathione, and preventing DNA damage, thus, leading to decreased inflammation and tumor growth. There is also evidence suggesting a role in regulating the metabolic activity of indigenous microbiota which can modulate host immune responses and inflammatory pathways (98).

One of the better studied mechanisms is the ability of Lactobacillus and Bifidobacterium to augment NK cell activity (98, 99). A study conducted by Ogawa et al. demonstrates that oral administration of L. casei ssp. casei in a combination treatment with dextran increases NK cell activity in healthy individuals. There are also significant increases in IL-12 production suggesting that dietary supplementation of Lactobacillus could enhance NK cell activity and lead to lower cancer risk, suggesting a protective advantage to individuals predisposed to cancer (98, 99).

Probiotics can also affect the activities of APCs and subsequent T cell activation. Probiotic bacteria have a variety of MAMPs on the bacterial surface. Lipoteichoic acid, LPS, and flagellin can interact with different TLRs on the surface of DCs which can modulate T cell polarization and lead to antitumor effects (99). Sivan et al. found that DCs isolated from tumor-bearing mice in a melanoma model had an upregulation of genes involved in CD8+ T cell activation, DC maturation, chemokine-mediated recruitment of immune cells to the TME, and type 1 interferon signaling after oral administration of Bifidobacterium. results were comparable to mice receiving a programmed cell death protein 1 ligand 1 (PD-L1) therapy, a checkpoint blockade. The combination treatment of oral Bifidobacterium and PD-L1 nearly abolished tumor growth (94).

Yet, diet can have a significant impact on the benefit of probiotic usage. A study conducted by Spencer et al. compared ICB responses in mice receiving high-fiber or low-fiber diets with or without probiotic supplementation. Mice receiving high-fiber and no probiotic usage achieved the best responses. However, mice receiving low-fiber diets or probiotic supplementation alone had poor ICB responses with lower microbial diversity and suppressed intratumoral T cell responses (100). This suggests that while probiotic bacteria can provide a protective effect in some cases, tailored probiotics composed of the specific strains from treatment responders could benefit more than a typical over-the-counter probiotic supplement for patients receiving ICB therapy.