AI-2 quorum sensing controlled delivery of cytolysin-A by tryptophan auxotrophic low-endotoxic Salmonella and its anticancer effects in CT26 mice with colon cancer

Graphical abstract

Highlights

• •Attenuated auxotrophic Salmonella Typhimurium was utilized for the anti-tumor effect.
• •Reduction of endotoxicity was rendered by lipid A modification.
• •Regulated expression of ClyA at higher cell density under QS signaling.
• •Engineered Salmonella Typhimurium enhanced safety and tumor invasion.
• •ClyA expressing Salmonella reduced 67 % of tumors in mouse CT 26 cancer model.

Abstract

Introduction

The limitations of conventional cancer therapies necessitate target-oriented, highly invasive, and safe treatment approaches. Hence, the intrinsic anti-tumor activity of Salmonella can offer better options to combat cancers.

Objectives

This study aims to utilize attenuated Salmonella and deliver cytolytic protein cytolysin A (ClyA) under quorum sensing (QS) signaling for precise localized expression in tumors but not in healthy organs.

Methods

The therapeutic delivery strain was imposed with tryptophan auxotroph for selective colonization in tumors by trpA and trpE deletion, and lipid-A and O-antigen were altered by pagL and rfaL deletions using lambda red recombination method. The strain was transformed with the designed QS-controlled ClyA expression vector which was validated by western blot. The in vivo passaged therapeutic strain was used for treatment four times at a weekly interval, with a dose of 5 × 10⁶ CFU/mouse for cancer therapy.

Results

The attenuated strain induced minimal endotoxicity-related cytokines TNF-α, IL-1β, and IFN-γ and exhibited adequate colonization despite earlier exposure in mice. The QS-controlled ClyA expression was confirmed by western blot from bacterial cultures grown at different cell densities. The results demonstrated that the in vivo passaged strain preferentially colonized the tumor after vacating the spleen, liver, and lung, leaving no outward histological scars. The anti-cancer effect of the designed construct was evaluated in the murine CT26 colon cancer model. The expression of ClyA increased tumoricidal activity by 67 % compared to vector control.

Conclusion

Hence, the anti-tumor effect of the engineered Salmonella strain was improved by ClyA expression via QS activation after achieving the threshold bacterial cell density. Further, immunohistochemical staining of the tumor and other organs corroborated the QS-controlled tumor-specific expression of ClyA. Overall, the results imply that the developed anti-cancer Salmonella has low endotoxicity and QS-controlled expression of ClyA as beneficial safety elements and supports recurrent Salmonella inoculation by O-antigen deficiency.

Introduction

Cancer is a burgeoning global burden with a high mortality rate. Among all types, colorectal cancer is the third most diagnosed cancer worldwide and the second leading cause of cancer-related deaths [1]. Current therapeutic practices such as chemotherapy, radiotherapy, and immunotherapy have insufficient specificity and elicit side effects, while cancer resection has a limited success rate [2], [3], [4]. The limited success with conventional treatments necessitates the exploration of more reliable and target-oriented therapeutic approaches. Bacteria-mediated cancer treatment (BMCT) has garnered attention as an alternative strategy to overcome the inherent challenges of conventional cancer treatment [5]. Researchers have explored the anti-tumor potential of Salmonella enterica serovar Typhimurium (ST) [6], and such intrinsic tumor-targeting characteristics and tumoricidal activities have been exploited through various modulations to improve its efficiency [7], [8], [9], [10].

The attenuated Salmonella strains VNP20009 and A1-R are the two most widely studied with anti-cancer potential [11], [12]. However, these strains need further improvement due to their limited anti-tumor effects in humans and infection-associated adverse effects [12], [13], [14]. They also showed nonsynonymous single nucleotide polymorphisms due to mutation in nucleoside synthesis [15], [16]. These setbacks hindered their use for anti-cancer implementation. We previously published a tryptophan auxotrophic Salmonella with high tumor-specific colonization due to nutrient dependency [17]. To improve safety and support repeated inoculation for onco-therapy, we further attenuated the strain in this study. Reduction of endotoxicity is a critical safety parameter for a live-attenuated therapeutic strain, such that recurrent inoculation and long-term persistence of live Salmonella in the host is not toxic. PhoP/PhoQ-induced lipase called PagL is a lipid A 3-O-deacylase that modifies lipid A by removing one or more acyl groups from lipopolysaccharides (LPS). The LPS is a potent inflammatory stimulator and inductor of lipid A alteration via pagL deletion to reduce endotoxicity, making it an ideal candidate for Salmonella attenuation [18], [19], [20]. Salmonella is a ubiquitous microbe with prevalent infectivity in humans and livestock. The presence of pre-existing immunity is the main challenge in implementing Salmonella as a cancer therapeutic delivery vector [21]. The O-antigen is an important virulence factor targeted by host immune systems. It is a determinant factor of the humoral response against invading bacteria [22]. To overcome pre-existing antibodies against the therapeutic delivery strain, it is beneficial to delete the O-antigen. Collectively, Salmonella attenuation via surface modification and nutrient dependency promises reduced virulence, safe administration, and target-oriented delivery.

Salmonella’s anti-tumor response can be further upgraded via bioengineering and delivery of heterologous anti-cancer agents [23]. Cytolysin A (ClyA) is a pore-forming, 34 kDa toxic protein secreted by gram-negative bacteria such as E. coli, Salmonella enterica serovar Typhi, and Paratyphi A [24]. It is cytotoxic to mammalian cells by damaging the plasma membrane and releasing intracellular contents [24], [25]. A barrier to applying ClyA-expressing Salmonella in cancer treatment is its toxicity to nontumoral reticuloendothelial organs. Therefore, controlled expression of ClyA specifically in tumors is crucial. Quorum sensing (QS) is a cell density-dependent cell-to-cell communication mechanism that switches on QS-regulated gene expression only at high cell density. Salmonella possesses QS signal molecules like autoinducer (AI)-2, which are produced by conversion of S-adenosyl homocysteine by the AI-2 synthase LuxS [26]. AI-2-mediated QS regulates T3SS-related, flagella-related, and oxidative stress-response genes [27]. In addition, AI-2 positively controls the expression of phase 2 flagellin (FljB) by the QS signaling mechanism [28].

In the present study, a plasmid with an fljB promoter was constructed and subsequently cloned clyA for QS-controlled expression. The tryptophan auxotrophic Salmonella can proliferate to high cell density only in tumors but not in healthy organs. The high cell density can switch on QS and clyA expression through the QS-controlled fljB promoter. Thus, tumor-specific clyA expression is conceptualized. In addition, the benefits of lipid A alteration by pagL deletion and O-antigen removal by rfaL deletion at the prospect of low endotoxicity and avoidance of pre-existing antibodies were investigated. The findings suggest that the current anti-cancer Salmonella strain has an array of safety benefits conferred by alteration in lipid A and O-antigen. Additionally, QS-regulated toxin delivery is an ideal strategy for safe tumor-targeted therapeutic delivery.

Materials and methods

Development of therapeutic strain

Bacterial strains and plasmids

All bacterial strains were routinely grown in Luria Bertani (BD, Sparks, MD, USA) medium with agitation at 37 °C using appropriate antibiotics as selection markers when applicable. All of the bacterial strains and plasmids used in the present study are listed in Table 1. In this study, an attenuated auxotrophic Salmonella enterica serovar Typhimurium (ST) strain was developed with the deletion of pagL and rfaL genes for cancer therapy. For this, the tryptophan auxotrophic Salmonella JOL2514 strain (ΔtryA ΔtrypE) was engineered to develop the Salmonella JOL2936 strain (ΔtryA ΔtrypE Δasd ΔpagL ΔrfaL) applying the λ red recombination technique described elsewhere with modifications [20]. This recombineering approach inserts a chloramphenicol resistance (cat^R) gene into the chromosome by replacing the target gene. Briefly, the parent auxotrophic ST strain was transformed with a helper plasmid, pKD46, and induced to express recombinase with L-arabinose for homologous recombination. The linear DNA cassette of the cat^R gene flanked by an asd gene homologous sequence was amplified from pKD3 and electroporated (Harvard Apparatus, USA) in pKD46-transformed Salmonella. The asd deleted mutant colonies were screened by plating on LB media containing chloramphenicol. Colonies were confirmed by inner primers and transformed with pCP20 plasmid to eliminate the FRT-flanked cat^R through flippase production. The cat^R deletion was confirmed by flanking primers, as listed in Table 2, and further proceeded to include pagL and rfaL deletion in the same procedure.

Table 1.

List of bacterial strains and plasmids.

Bacteria/Plasmid	Genotypic characteristics	References
S. Typhimurium
JOL2514	ΔtrpA ΔtrpE	Lab stock
JOL2844	ΔtrpA ΔtrpE Δasd	This study
JOL2952	JOL2844 with pJHL90	This study
JOL2940	ΔtrpA ΔtrpE Δasd ΔrfaL	This study
JOL2953	JOL2940 with pJHL90	This study
JOL2936	ΔtrpA ΔtrpE Δasd ΔrfaL ΔpagL	This study
JOL2954	JOL2936 with pJHL90	This study
JOL2954P3	Thrice in vivo passaged JOL2954 in CT26 tumor-bearing mice	This study
JOL2955	JOL2936 with pJHL90 – ClyA	This study
JOL2955P3	Thrice in vivo passaged JOL2955 in CT26 tumor-bearing mice	This study
S. Typhi	(Source of clyA amplification)	Lab stock
E. coli
DH5α	E. coli F-Φ80dlacZΔM15Δ (lacZYA-argF) U169recA1 endA1 hsdR17(rk-mk+) phoA supE44 thi1 gyr A96 relA1λ-	Lab stock
BL21(DE3)	F–, ompT, hsdSB (rB–mB–), dcm, gal, λ(DE3)	Lab stock
JOL2858	DH5α carrying pET28a (+) – clyA	This study
E. coli 232	F – λ – φ80Δ(lacZYA-argF) endA1 recA1 hadR17 deoR thi-1 glnV44 gyrA96 relA1 ΔasdA4	Lab stock
Plasmids
pJHL65	asd + vector, pBR ori, β-lactamase signal sequence-based periplasmic secretion plasmid, 6xHis, high copy number	[54]
pJHL90	Plasmid with QS promoter prepared by replacing Ptrc from pJHL65	This study
pJHL90 – clyA	pJHL90 containing clyA to express under QS promoter	This study
pET28a (+)	IPTG-inducible expression vector; Kanamycin resistance	Lab stock
pET28a (+) – clyA	pET28(+) carrying clyA for protein expression	This study
pKD46	oriR101-repA101ts; encodes λ red genes (exo, bet, gam); native terminator (tL3); arabinose-inducible promoter for expression (ParaB); bla	[39]
pKD3	oriR6Kgamma, bla (ampR), rgnB (Ter), catR, FRT	[39]
pCP20	Helper plasmid contains a temperature-inducible flp gene for removing the FRT flanked chloramphenicol gene	[55]

Table 2.

List of primers.

Gene	Primer	5′– 3′ Sequences		References
Gel deletion
asd-pKD3	Sense	TGAAGGATGCGCCACAGGATACTGGCGCGCATACACAGCACATCTCTTTGGTGTAGGCTGGAGCTGCTTC		[20]
	Antisense	TATCCGGCCTACAGAACCACACGCAGGCCCGATAAGCGCTGCAATAGCCAATGGGAATTAGCCATGGTCC
pagL-pKD3	Sense	AATTTTAAATATGTTAGCCGGTTAAAAATAACTATTGACATTGAAATGGTGTGTAGGCTGGAGCTGCTTC		This study
	Antisense	CGGTGATTAATTACTCCTTCAGCCAGCAACTCGCTAATTGTTATTCAACTATGGGAATTAGCCATGGTCC
rfaL-pKD3	Sense	TTTGGAAAGATTCATTAAAGAGACTCTGTCTCATCCCAAACCTATTGTGGGTGTAGGCTGGAGCTGCTTC		[20]
	Antisense	CCTGATGATGGAAAACGCGCTGATACCGTAATAAGTATCAGCGCGTTTTTATGGGAATTAGCCATGGTCC
asd-inner	Sense	CATGGTAGAGGAGCGCGATT		[20]
	Antisense	TACCGCCCACAAAGGTCTTC
asd-outer	Sense	GCGACGGAAATGATTCCCTT		[20]
	Antisense	AAGCTACCCTTAAAGAATAGCC
palL-inner	Sense	CAGATCTCTTTTGCTGCGGG		[20]
	Antisense	AAAAGCCCCAAAGTTCCAGC
pagL-outer	Sense	TGGATGTGCCTGAACAACACT		[20]
	Antisense	TTAGCCTCCCTGTCGCCATA
rfaL-inner	Sense	ACAAGTTTAGGACTTCGCTGCC		[20]
	Antisense	CAGAATGGTATTATGCGGACCG
rfaL-outer	Sense	GCA GCG TTT CGA GGA ACA AA		[20]
	Antisense	TCG TAT CGG TTG ATA CCG GC
clyA amplification
	Sense	GGATCC ATG ACC GGA ATA TTT GCA GAA
	Antisense	GTCGAC TCA GAC GTC AGG AAC CTC

Preparation of plasmid containing a quorum sensing promoter

The plasmid pJHL65 containing the bla secretion signal sequence under the P_trc promoter was used as a parental source for pJHL90 preparation. This plasmid encodes the asd gene for Darwinian selection retention. The plasmid pJHL90 was constructed by replacing the P_trc promoter with the QS promoter. For this, the QS promoter of the fljB gene from ST was selected and chemically synthesized (Cosmogenetech, Seoul, South Korea). Plasmid pJHL65 was digested with BamHI followed by partial digestion with AseI. The synthesized QS promoter treated with the same restriction enzymes was ligated with the digested pJHL65 to prepare pJHL90 (Fig. 2A). The developed plasmid was confirmed both by digestion with restriction enzymes and PCR amplification. The engineered plasmid was cloned with the clyA gene amplified from Salmonella enterica serotype Typhi with the aid of restriction enzymes, BamHI and SalI. Thus, the developed plasmid was used to transform JOL2936 into JOL2955.

Fig. 2.

Cloning and in vitro evaluation of cytolysin A expression under the quorum sensing (QS) promoter. A. pJHL90 was constructed by replacing the P_trc promoter in pJHL65 with the QS promoter. The clyA gene was inserted in the MCS region of pJHL90 using restriction enzymes, BamHI and SalI. B. Growth curve. Overnight grown cultures of WT Salmonella Typhimurium (ST) and attenuated ST JOL2955 were sub-cultured in LB broth. Bacterial growth was measured by optical density (OD) at 600 nm every hour for 10 h. C. Expression of cytolysin A protein under QS promoter. The bacterial lysate prepared at 0.5, 1.5, 1.8, and 2.0 OD₆₀₀ values was evaluated for ClyA expression by western blot. M = protein marker and +ve = positive control, cytolysin A protein extracted from E. coli BL21 (DE3). D. Hemolytic activity on mice erythrocytes. The suspension of 2 % erythrocytes was incubated for 4 h with protein extracts and centrifuged to detect hemolysis. Hemolysis by JOL2955 was evaluated at OD₆₀₀ values of 0.5 and 1.5. The tubes labeled C and VC denote control and vector control, respectively.

Preparation and characterization of the therapeutic strain

Attenuated Salmonella Typhimurium, JOL2936, was transformed with plasmid, pJHL90, containing clyA by electroporation (Harvard Apparatus). The PCR-confirmed strain was used as a cancer therapeutic strain. The growth curve analysis of the mutant strain was performed in vitro in 20 mL of LB broth. For this, overnight grown cultures of wildtype JOL401 (WT) and developed JOL2955 strains were sub-cultured in LB broth. Bacterial growth was assessed by measuring the optical density (OD) at 600 nm every hour for 10 h.

Cell lines

A murine colorectal carcinoma cell line, CT26, was purchased from American Type Culture Collections (ATCC; CRL-2638: RRID: CVCL_7256) and maintained in Dulbecco’s modified Eagle’s medium (DMEM, Lonza, Basel, Switzerland) supplemented with 10 % fetal bovine serum (FBS, Gibco, NY, USA), 100 units/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin B (Lonza, Walkersville, MD, USA) at 37 °C in a humidified atmosphere containing 5 % CO₂. The human colorectal adenocarcinoma cell line HT29 was procured from the Korean Cell Line Bank (KCLB; KCLB:30038) and maintained in Roswell Park Memorial Institute Medium 1640 (RPMI 1640; Thermo Scientific, Waltham, MS, USA) with 10 % FBS and 1 % broad-spectrum antibiotics.

Animals and ethics statement

Six-week-old, specific pathogen-free female BALB/c mice procured from Samtako, Osan, Korea, were maintained on a standard feeding regimen with antibiotic-free food and water ad libitum at the Animal Housing Facility of the College of Veterinary Medicine, Jeonbuk National University. Animal experiments were conducted under the guidelines of the Korean Council on Animal Care and the Korean Animal Protection Law, 2007: Article 13 (https://koreananimals.org/animal-protection-law-2007/), with the approval of the Jeonbuk National University animal ethics committee (JBNU-2021–027). Animal euthanasia was performed at the end of the experiments following the standard protocol. Animals were anesthetized with chloroform and then euthanized through cervical dislocation.

Expression and purification of protein

ClyA protein was expressed in E. coli BL21 (DE3) following the standard cloning procedure described elsewhere. Briefly, clyA was cloned into pET28a (+), and the E. coli DH5-α strain was transformed with the cloned plasmid. The engineered plasmid was extracted, PCR confirmed and used to transform E. coli BL21 (DE3). The transformed bacterial culture was induced with 1 mM IPTG for 4 h at 37 °C to express the protein, and the expression was confirmed by SDS-PAGE. Further, the expression was confirmed by western blot using His-Tag HRP mouse monoclonal IgG1 (H-3) (Santa Cruz Biotechnology, Dallas, TX, USA). The obtained protein was purified by nickel-NTA column chromatography (Takara, Shiga, Japan) using the urea lysis method. The purified recombinant protein was inoculated in a rabbit to produce polyclonal antibodies, as described elsewhere [20].

Controlled expression of ClyA

Western blot for expression of ClyA under QS promoter

The therapeutic strain (JOL2955) grown overnight at 37 °C was subjected to growth at absorbances of 0.5, 1.5, 1.8, and 2.0 at OD₆₀₀. Cells were harvested at 4 °C by centrifugation at 3,500 rpm for 15 min. Collected cells were washed and resuspended in PBS maintaining an equal OD₆₀₀ value of 1 for all samples. The cell suspensions were sonicated for 10 min with a 10 s interval at 40 % amplitude. The lysate was centrifuged at 4 °C for 15 min at 3,500 rpm. The supernatant was collected, mixed with an equal volume of 2 × SDS sample buffer, and run on a 12 % gel. Then, the protein was transferred to a PVDF membrane (Immobilon^R-P; Millipore, Tullagreen, Ireland), followed by 2 h incubation with 3 % BSA at room temperature to block unsaturated sites [20]. The membrane was treated overnight with primary antibodies (1:500 dilution), followed by interaction with HRP-conjugated anti-rabbit IgG antibodies (1:6,000 dilution) (SouthernBiotech, Birmingham, AL, USA) for 1 h. Finally, the images were developed by adding DAB substrate and documented for further analysis (Cytiva, Marlborough, MA, USA).

Hemolytic assay

The hemolytic action of ClyA was evaluated with fresh mice erythrocytes. Blood was collected from the orbital sinus, stabilized with heparin and erythrocytes were isolated following the procedures as described elsewhere with some modification [29]. Briefly, PBS was added in a 1:2 ratio and centrifuged at 1,000g for 10 min and 4 °C to isolate red blood cells (RBCs). The collected RBCs were washed thrice with PBS and finally diluted with PBS to prepare a 2 % RBC solution. Bacterial lysate of JOL2955 was harvested at cell densities of 0.5 and 1.5 OD₆₀₀, and protein was extracted following the procedure mentioned earlier. The same method was replicated for JOL2954 at a 1.5 OD₆₀₀ value. The RBC solution, 500 μL, was added to the collected protein supernatant (100 μL), while PBS was added for the negative control group. The tubes were incubated for 4 h at 37 °C and centrifuged at 1,000g. The images for RBC hemolysis in the tubes were captured and recorded.

Assessment on the impact of attenuation

Effects of pre-existing immunity

The effects of pre-existing antibodies as a consequence of priming with the rfaL intact Salmonella strain (JOL2952) were compared with the rfaL mutated strain (JOL2953). A total of 25 female BALB/c mice were equally divided into five groups. Initially, two groups were selected and inoculated intraperitoneally with 1 × 10⁶ CFU/mouse JOL2952 or JOL2953 [30]. Mice were sacrificed three days after inoculation to evaluate the bacterial colonization in a non-primed state. Two groups of mice received the same concentration of JOL2952 as priming with the O-antigen-positive strain, and the remaining group was inoculated with JOL2953 as priming with the O-antigen-deficit strain. The group of mice primed with JOL2952 was inoculated with JOL2952 and the other group received JOL2953 after 14 days from priming via the intraperitoneal route. Likewise, mice primed with JOL2953 were injected with JOL2953 14 days after priming. The spleen and liver were collected after three days of injection to ascertain bacterial colonization after priming. Bacterial colonies were enumerated by plating the tissue homogenates in Brilliant Green Agar (BGA) (BD Difco™).

Real-time PCR (RT-PCR) for inflammatory cytokines

Sixteen six-week-old BALB/c mice were divided into 4 groups and injected intraperitoneally at 1 × 10⁶ CFU/mouse with JOL401 (WT), JOL2953, or JOL2954. The control group received an equal volume of PBS (100 μL). Mice were sacrificed at 3 dpi to examine the endotoxicity induced by the attenuated Salmonella strain. The spleen was harvested, and the total RNA was extracted following the manufacturer’s instructions (RiboEx kit, GeneAll Biotechnology, South Korea). The cDNA was synthesized from purified RNA (Elpis Biotech, Daejeon, South Korea). The expression of inflammatory cytokines TNF-α, IL-1β, and IFN-γ at the mRNA level was quantified with ABI (Step One Plus, Applied Biosystems, USA) using an SYBR green PCR master mix (Elpis Biotech). Melting peak analysis was performed to confirm the absence of contamination and the specificity of the PCR amplification. Transcriptional levels were normalized to that of β-actin, and the 2^-ΔΔCT method was executed to determine the changes in mRNA levels [31].

Hematoxylin and eosin (H&E) staining

Histopathological assessment was performed to evaluate the extent of tissue distortion caused by Salmonella infection in mice. The spleen and liver of JOL401-, JOL2953-, and JOL2954-infected animals were excised on 3 dpi. Organs were fixed in 10 % formalin and embedded in paraffin. Tissues were excised from anterior, medial, and posterior region of organs and each region were sectioned 2 μm thick using a microtome (Thermo Shandon Finesse 325, Thermo Scientific) and stained with hematoxylin and eosin (H&E) [32]. The stained sections were examined for visual pathological signs by a light microscope (Axio Imager 2, Zeiss, Germany).

In vivo passage and its effects

In vivo passage

Six-week-old BALB/c mice were implanted with 5 × 10⁶ CT26 cells via a subcutaneous route at the right flank to induce tumors that then were grown for 12 days. The tumor-bearing mice (n = 2) were administered with attenuated bacterial strain JOL2955 at 5 × 10⁶ CFU through the intraperitoneal route (IP). On the third day, mice were sacrificed, and the tumors were collected, homogenized, and plated on BGA. The strain was confirmed by PCR and passaged to obtain the final therapeutic strain JOL2955P3 at three passages. The same procedure was replicated to acquire JOL2954P3 from its parent strain, JOL2954, as a vector control.

Adhesion, invasion, and intracellular survival of the passaged strain

Adhesion and invasion of the bacterial strains were conducted in CT26 and HT29 cells, as described elsewhere [20]. Briefly, in vivo passaged and non-passaged JOL2955 strains were allowed to interact with cells for 30 min at a multiplicity of infection (MOI) of 10 and washed three times with PBS to remove non-adherent cells. The adhered bacteria were counted by plating on BGA after treating cells with 0.25 % Triton X-100. For the invasion assay, bacteria were incubated with cells for 2.5 h to invade the cells. The non-infected bacteria were eliminated by 150 μg/mL gentamycin, and internalized bacteria were counted by plating on BGA after treating cells with 0.25 % Triton X-100. For intracellular survival, cells were further incubated for 12 h post-gentamycin treatment. After incubation, monolayers were lysed with 0.25 % Triton X-100 to release intracellular bacteria, serially diluted, and plated on BGA for enumeration.

Characterization of therapeutic strain

Wound healing assay

CT26 cells were cultured in 12-well plates and allowed to grow until they reached 70 % confluence. A confluent monolayer of cells was scratched at the middle by the tip of a 200 μL pipette to create a gap. Then, the cells were washed and exposed to JOL2955P3, JOL2954, or media alone as a control for 2.5 h. Extracellular bacteria were eliminated by applying gentamycin in the culture media. Cells were monitored for 48 h to examine the healing of the scratches, and microscopic images were collected using a phase-contrast microscope (Leica, Leitz Park, Wetzlar, Germany) [17]. The experiment was conducted in triplicate.

Apoptosis

CT26 cells were cultured in a 12-well plate at 5 × 10⁵ cells/well. Cells were infected with the mutants at 10 MOI for 24 h. Cells were harvested, and apoptosis induced by the mutants was analyzed using a fluorescein isothiocyanate (FITC) Annexin V Apoptosis Detection Kit I (BD Pharmingen, USA) following the manufacturer’s instructions. For this, cells were labeled with propidium iodide (PI) and annexin V-FITC and read using a MACSQuant analysis system (Miltenyi Biotec, Bergisch Gladbach, Germany) and a fluorescence-activated cell sorting (FACS) reader to differentiate live and apoptotic cells.

Adaptability of passaged and non-passaged ST

The adaptable potentiality of passaged and non-passaged ST in the tumor was analyzed in CT26 tumor-bearing mice. CT26 cells (5 × 10⁶) were injected subcutaneously into the right flank of six-week-old female BALB/c mice to develop tumors for 12 days. Mice (n = 5) were challenged with passaged and non-passaged JOL2955 at 5 × 10⁶ CFU intraperitoneally [33]. Three mice per group were sacrificed at 3-, 7-, and 14-days post-infection (dpi). The tumors were collected and homogenized in PBS using a mechanical homogenizer (IKA ULTRA-TURRAX, Merck, Darmstadt, Germany), serially diluted, and plated in BGA to enumerate the invading colonies.

Safety assessment of the therapeutic strain

The safety of the designed therapeutic strain was evaluated in six-week-old female BALB/c mice. Mice (n = 5) were inoculated with 5 × 10⁷ CFU JOL2955P3; ten times of therapeutic dose, via the IP route and the same dose of the WT strain were compared for survivability. Animals were monitored up to 35 dpi for the adverse effect of the therapeutic construct. Another group of mice (n = 5) was administered 5 × 10⁶ CFU JOL2955P3 regularly for four weeks at a weekly interval. Mice receiving PBS in placebo were regarded as the naïve control (n = 3). Mice were sacrificed two weeks after the final dose to evaluate gross changes of the vital organs. Histopathological examination of tissue sections was carried out to determine the deterioration caused by ST following the protocol described previously. For this, mice from each naïve and treated group were sacrificed to harvest the spleen, liver, and lung. Collected organs were preserved in a formaldehyde solution before the process. Samples were embedded in paraffin, sectioned, and stained with hematoxylin and eosin to examine the histopathology and extent of damage to the organs.

Invasion of the therapeutic strain in tumors

Invasion of JOL2955P3 in the tumors and vital organs was performed in tumor-bearing mice. For in vivo invasiveness of the developed strain, the tumor was induced with subcutaneous inoculation of CT26 cells (5 × 10⁶) on the right flank of six-week-old female BALB/c mice (n = 15). Tumor growth was monitored for 12 days, and mice were injected with 5 × 10⁶ CFU of JOL2955P3 via the IP route. Tumor, spleen, liver, and lung samples were collected from five mice per group sacrificed at 3, 7, and 14 dpi and subjected to bacterial quantification by plating on BGA. For this, the samples were homogenized in PBS as described previously using a mechanical homogenizer and were plated at 10-fold serial dilution. The colonies were enumerated, and representative colonies were confirmed for ST carrying clyA by PCR.

Assessment of anti-tumor effect

Tumor therapy of the designed construct

CT26 cells (5 × 10⁶) were subcutaneously implanted into the right flank of BALB/c mice to grow a CT26 syngeneic tumor [33]. The tumor size was recorded; on day 12, mice were randomly divided into 3 groups of 6 mice. The first group of mice was administered four weekly treatments of 5 × 10⁶ CFU (100 μL) JOL2955P3 via the IP route, and the second group received similar treatment with JOL2954 carrying an empty plasmid as a vector control. The remaining mice were untreated placebo controls and were given 100 μL sterile PBS. Mice were monitored daily, and the sizes of the tumors were measured regularly using a Vernier caliper. Tumor volume was calculated using the formula: tumor volume, V = ½ (length × width²) [34]. Mice were euthanized at 35 days post-treatment to excise the tumor, spleen, liver, and lung. Gross morphological changes in the collected organs were noted. The size and weight of the tumor and spleen were recorded. Tumor samples were preserved and fixed in 10 % formalin for histopathological examination.

Quantification of cytokines and cell cycle regulatory genes

The expression of cytokines and cell cycle regulatory genes at mRNA in the tumors was measured by RT-PCR. Briefly, tumors treated with JOL2955P3 and the vector control were harvested at the end of the experiment. The total RNA was extracted (GeneAll) and converted to cDNA (Elpis Biotech) following the manufacturer’s instructions. The expression of cytokines such as TNF-α, IL-1β , IFN-γ, IL-6, and IL-10; cell cycle regulatory genes, namely p21, p27, p53, Bcl2, Bcl xL, and Bax; and vascular endothelial growth factor (VEGF), were quantified with ABI (Applied Biosystems) using the SYBR green PCR master mix (Elpis Biotech). Transcriptional levels were normalized against that of β-actin, and the 2^-ΔΔCT method was executed to determine the changes in mRNA levels [31].

Tumor-associated macrophage (TAM) markers

Mice were sacrificed at the end of the experiment and tumors were collected. Connective tissues were removed from the tumor and rinsed with RPMI-1640. Tissues were chopped into small 1–2 mm³ pieces, followed by digestion with 0.05 % collagenase IV in RMPI-1640 containing 2 % FBS at 37 °C for 2 h with shaking at 15 min intervals. The digested cell suspension was filtered with a 70 μm cell strainer. Repeated filtration was performed to remove residual undigested tumor tissue. Cells were collected by centrifugation at 300g for 10 min and suspended in RPMI-1640 supplemented with 2 % FBS. Tumor-associated macrophages (TAM) were isolated by density gradient centrifugation using Ficoll-Paque plus density gradient media (density of 1.077 g/mL) (Cytiva, Uppsala, Sweden) by centrifugation at 300g for 30 min at room temperature [35]. A distinct band of TAM formed between the two layers was harvested. Cells were washed with media and tested for viability using Trypan blue dye. Then, cells were stained with TAM markers, CD68-PE, iNOS-FITC, and CCL2-APC, according to the manufacturer’s instructions (Miltenyi Biotec). The labeled cells were analyzed using the MACSQuant analysis system (Miltenyi Biotec).

Histopathological analysis and immunohistochemical staining

Tumor samples fixed in formalin were subjected to standard tissue processing with paraffin embedding. Tissues were sectioned from three different regions of tumor (outer, mid, and inner region) and deparaffinized as described earlier. Histopathological analysis of tumor sections was performed by H&E staining [32]. Immunohistochemical staining (IHC) of the tissue sections of the tumor, spleen, and liver was conducted following the standard protocol to examine the localization and expression of ClyA by JOL2955P3 [20]. Briefly, specimens were deparaffinized by xylene and subsequently hydrated with ethanol gradient and distilled water. The sections were heated in citrate buffer (pH 6) at 100 °C for 30 min to retrieve antigens followed by the addition of 0.3 % methanolic H₂O₂ to inhibit peroxidase activity. The samples were consecutively labeled with primary antibodies raised in a rabbit (1:200 dilution) and treated with HRP-conjugated goat anti-rabbit antibodies (1:3,000 dilution) (SouthernBiotech). Finally, a 3, 3′ diaminobenzidine (DAB) substrate (Sigma-Aldrich, Steinheim am Albuch, Germany) was added to develop color, and images were documented.

Statistical analysis

All data were analyzed using GraphPad Prism 9.0 software (GraphPad, San Diego, CA, USA). All the collected data were confirmed to have a normal distribution by performing a Shapiro-Wilk normality test. An unpaired t-test was conducted for evaluating the influences of pre-existing antibodies, effect of in vivo passage on cells, and fitness in tumor. One-way analysis of variance (ANOVA) with Tukey’s multiple comparison tests was conducted to determine statistical differences among the treated and control groups for endotoxicity assessment, apoptosis evaluation and other experiments. P values < 0.05 were considered statistically significant. All data are expressed as mean ± SEM.

Results

Preparation of the attenuated Salmonella strain

Construction of attenuated Salmonella strains was accomplished by the λ red recombination method (Fig. 1). ST tryptophan auxotroph JOL2514 (ΔtrpA ΔtrpE) was selected as a parental strain for the preparation of attenuated strains. The strain was engineered by deletions of pagL and rfaL to modify the lipid A structure and eliminate O-antigen ligase. Further, the strain was asd-mutated to impose Darwinian selection on the retention of an asd⁺ plasmid that encodes foreign antigens. Therefore, the selective pressure on the attenuated strain was complemented with the asd⁺ plasmid vector. The engineered construct, JOL2936 (ΔtrpA ΔtrpE ΔpagL ΔrfaL Δasd), was confirmed for respective gene deletion using the corresponding inner and flanking primers (Supplementary Fig. 1).

Fig. 1.

Schematic experimental layout of the development of the therapeautic strain and its evaluation for anti-tumor effects.

Construction of plasmid and preparation of therapeutic strain

A plasmid, pJHL90, was created from pJHL65 by substituting the QS promoter for the P_trc promoter. The expression of Phase 2 flagellin (FljB) is a proven QS-regulated phenotype [28]. Hence, the fljB promoter was employed as a QS promoter in pJHL90 by replacing the P_trc promoter in pJHL65. The incorporation of the QS promoter was confirmed by PCR amplification. Consequently, the engineered plasmid, pJHL90, was used to clone clyA under the QS promoter (Fig. 2A) and was confirmed by PCR amplification. The clyA-positive cloned plasmid was electroporated to JOL2936, and the therapeutic construct was designated as JOL2955. The engineered JOL2955 strain was adapted to the tumor environment by repeated in vivo passage in CT26 tumor-bearing BALB/c mice. Thus, the engineered strain was passaged thrice in tumor-bearing mice, and the final strain was designated as JOL2955P3.

Proliferation of the attenuated JOL2955 strain

The therapeutic strain, JOL2955P3, and WT strain were compared for growth patterns in LB broth for 10 h from overnight grown cultures. The therapeutic strain showed no apparent differences from the wild-type strain at any of the test times. Both strains reached the stationary phase after 8 h (Fig. 2B).

Expression of ClyA under the QS promoter

The expression of ClyA under the QS promoter was investigated by western blotting of bacterial lysate harvested at different cell densities (Fig. 2C). The protein expression was not detected at 0.5 OD₆₀₀ but it was detected at 1.5 OD₆₀₀ and above. The results show that the expression switched on with an increase in bacterial cell density. The expression of ClyA was also determined by its functional character hemolysis on mice erythrocytes. Hemolysis was observed in protein extract collected at 1.5 OD₆₀₀ but absent at 0.5 OD₆₀₀ (Fig. 2D). Collectively, the results imply that the ClyA expression was regulated by the QS signaling mechanism, which allows expression only at higher cell densities. Thus, the cytolytic activity of the strain could depend on the cell density level in host organs.

Influence of pre-existing antibodies

The ubiquitous prevalence of Salmonella results in frequent exposure and imparts a higher chance of pre-existing antibodies. The therapeutic strain was attenuated by rfaL deletion to escape from the interference of antibodies from prior exposure. Bacterial colonization of O-antigen-deficit JOL2953 (ΔtrpA ΔtrpE Δasd ΔrfaL) and O-antigen-positive JOL2952 (ΔtrpA ΔtrpE Δasd) Salmonella strains in the spleen and liver were compared before and after Salmonella exposure. Then, the two were inoculated to test the interference by pre-existing antibodies. Bacterial recovery of the rfaL intact strain from the spleen (p = 0.0002) and liver (p < 0.0001) of primed mice was significantly reduced, whereas the rfaL-deleted strain remained comparable (Fig. 3A). It is anticipated that the O-antigen-deficit strain escaped pre-existing antibody interference and was recovered with a nominal difference after priming. Moreover, a nonsignificant difference was observed between pre and post-priming of O-antigen-deficit Salmonella (JOL2953).

Fig. 3.

Attenuated Salmonella influences bacterial colonization and impairs endotoxicity. A. Effect of priming on the colonization of Salmonella. Salmonella infection was ascertained in the presence of pre-existing antibodies induced by priming intraperitoneally with O-antigen-positive (JOL2952) and -negative (JOL2953) strains (1 × 10⁶ CFU/mice). Primed mice were injected with O-antigen-positive and O-antigen-deficient (JOL2953) Salmonella strains. Mice were sacrificed at 3 days post inoculation to enumerate bacterial CFU in the spleen and liver at primed and non-primed states. Data were analyzed by unpaired t-test. B. The expression of inflammatory cytokines, TNF-α, IL-1β, and IFN-γ. Cytokine expression was measured by RT-PCR in the spleen collected at 3 days post-inoculation. Data were analyzed by Tukey’s multiple comparison tests using one-way ANOVA. The error bars denote the SEM. ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. D. Histopathological examination of the spleen and liver. Hematoxylin and eosin staining was performed to analyze the effect of Salmonella infection in the organs. The investigation showed prominent tissue dispersion of red pulp with diffused lining between red and white pulp indicated by the black arrow and infiltration of inflammatory cells denoted by the red arrow in the spleen infected with wildtype (WT) and JOL2953. Infiltration of macrophages in the liver is indicated by the green circle. Noticeably altered tissue architecture was not demonstrated between JOL2954 and the PBS-inoculated group. Images were obtained at 200× magnification. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Assessment of endotoxicity

The pagL deletion-based lipid-A alteration was carried out in the therapeutic strain to reduce endotoxicity. The engineered therapeutic strain showed a substantial reduction in the induction of inflammatory cytokines such as TNF-α (p = 0.0005), IL-1β (p < 0.0001), and IFN-γ (p = 0.0002) compared to WT. The reduction of TNF-α was 2.7-fold, and IL-1β was 2.3-fold, while that of IFN-γ was 3.5-fold compared to WT treatment. Likewise, TNF-α (p = 0.0183) and IFN-γ (p = 0.0452) were reduced by 2-fold compared to pagL intact JOL2953 (Fig. 3B). As a result, the role of pagL in endotoxicity was supported by higher cytokine production in mice treated with the pagL intact strain. Histopathological examination by H&E staining demonstrated prominent tissue distortion in the spleen infected with WT and JOL2953. Immune cell infiltration was noted in the liver infected with WT and JOL2953 as a sign of inflammation (Fig. 3C). A considerable reduction in inflammatory cytokines and minimal histopathological changes in JOL2954-treated mice elucidated the reduced endotoxicity of the pagL-deficient therapeutic strain.

In vitro fitness in cancer cell lines

The attenuated strain was subjected to in vivo tumor to tumor passaging thrice in tumor-bearing mice. The finally retrieved strain, JOL2955P3, was investigated to evaluate the alteration of bacterial ability to adhere to and invade cancer cells. The passaged strain exhibited enhanced adhesion, invasion, and intracellular survival in both murine (CT26 cells) and human (HT29) cancer cell lines (Fig. 4A). Both adhesion and invasion increased by nearly 2-fold in CT26 cells while the increments were more than 3-fold in HT29 cells. These findings demonstrate the improvement of the biological fitness of the bacterial strain on the cancer cells.

Fig. 4.

Evaluation of adhesion, invasion, and intracellular survival and the cytolysin A effect. A. JOL2955 was passaged in vivo in CT26 tumor-bearing mice thrice to obtain JOL2955P3 and the strain was compared to its original, JOL2955P0, for adhesion, invasion, and intracellular survival evaluation in CT26 and HT29 cells. Data were analyzed by the unpaired t-test. *p < 0.05 and **p < 0.01 where the error bars denote the SEM. B. Wound healing assay. A confluent monolayer of CT26 cells was scratched in the middle to create a gap between cells. Cells were washed and treated with JOL2955P3, vector control (VC), or media alone as a control (C) for 2.5 h. Extracellular bacteria were eliminated by applying gentamycin in the culture media. Cells monitored for 48 h revealed the gap closure in C and VC groups, as denoted by the black circle. Cell proliferation was inhibited in the scratched regions in the treated group. C. Apoptosis induced by the strains. CT26 cells were treated overnight with JOL2955P3 and VC. Cells were stained with propidium iodide (PI) and annexin V-FITC and evaluated by a fluorescence-activated cell sorting (FACS) reader to differentiate live and apoptotic cells. D. Histogram of apoptosis induced by attenuated strains. The cells were categorized into live (PI^- annexin V^-), early apoptosis (PI^- annexin V⁺), and late apoptosis (PI⁺ annexin V⁺) based on their stage of apoptosis corresponding to staining. Data were analyzed by Tukey’s multiple comparison tests using one-way ANOVA, *p < 0.05, **p < 0.01, and ****p < 0.0001, where the error bars denote the SEM.

In vitro anti-cancer effects elicited by therapeutic strain

The cytolytic activity of the designed therapeutic strain was examined in CT26 cells. The healing of scratches in the confluent monolayer was monitored to evaluate cell proliferation and migration for 48 h. The regenerative capacity was investigated and revealed that the gap closure was significantly inhibited in cells treated with JOL2955P3 (Fig. 4B). Cells treated with null vector control (JOL2954P3) also showed wound healing inhibition, but the activity was considerably higher in the active strain (JOL2955P3). In parallel, the induction of apoptosis by the JOL2955P3 strain was investigated in CT26 cells. Early apoptotic and late apoptotic cells were detected by annexin V and PI staining. A significant number of early apoptotic cells, i.e., annexin V⁺ and PI^-, were observed for samples treated with the vector control (p < 0.0001) and therapeutic strain (p < 0.0001). Late apoptosis (annexin V⁺ and PI⁺) in CT26 cells infected with JOL2955P3 was higher than that of the VC group (p = 0.0038). This might be a reflection of the killing effect of ClyA produced by the therapeutic strain (Fig. 4C & 5D). In the control group, cells were live and were not stained by annexin and PI. These in vitro results demonstrated that the therapy has anti-cancerous potential.

Improved tumor invasion by in vivo passage

The increase in tumor invasiveness of the passaged strain compared to the non-passaged strain was evaluated in CT26 tumor-bearing mice. The bacterial colonization of in vivo passaged strain in the tumor was substantially higher than the non-passaged strain. In the initial days after infection, the bacterial load was comparable, but it declined with time; at 14 dpi, only the passaged strain was present in the tumor (Fig. 5A). Thus, the tumor-targeting ability of the strain was enhanced by in vivo passage and the strain was subsequently used in the study for anti-cancer therapy.

Fig. 5.

Safety assessment of the attenuated Salmonella Typhimurium JOL2955P3 strain. A. Fitness of JOL2955 in the tumor after three successive in vivo passages in CT26 tumor-bearing mice. The syngeneic CT26 BALB/c mice were inoculated with JOL2955, where the strain was recovered from the tumor and subsequently passaged to obtain JOL2955P3. The colonization of the passaged strain was compared with that of its parental strain in the tumors of CT26 mice at 3, 7, and 14 days post-infection (dpi) using the unpaired t-test. B. Survival curve for BALB/c mice challenged with 5 × 10⁷ CFU/mL by an intraperitoneal route (IP). C. Colonization of Salmonella Typhimurium JOL2955P3 in the tumor, spleen, liver, and lung. Tumor-bearing mice were inoculated with 5 × 10⁶ CFU/mL bacteria via the IP route. Five mice were sacrificed at each time point. The organs were harvested, homogenized, and plated in BGA plates to enumerate the bacterial colonies that invaded the organs. Data were analyzed by Tukey’s multiple comparison tests, one-way ANOVA, *p < 0.05 and ****p < 0.0001, where the error bars denote the SEM. D. Histopathological changes in the spleen, liver, and lung. The organs were harvested 2 weeks after the final dose and fixed in paraffin. The tissue sections were deparaffinized and stained with hematoxylin and eosin for histopathological analysis. The effect on the organs was investigated and compared with organ samples of mice receiving PBS as a control. No notable changes were investigated in the organs of challenged mice. Images were obtained at 100 × magnification.

Safety assessment and localization of the attenuated auxotrophic strain in vital organs

Safety is the prime interest of researchers when developing therapeutic bacterial strains. A safety assessment of the attenuated auxotrophic JOL2955P3 was performed in mice following systemic infection and compared with the WT strain. No mortality was recorded in mice inoculated with the attenuated strain, but WT-infected mice could not survive for more than 6 days (Fig. 5B). The therapeutic strain’s safety was confirmed by testing it at a dose ten times higher than the therapeutic level. This ensured that even in cases of individual variations or unexpected reactions, the intended dose remains within a safe range. Localization of the attenuated therapeutic strain in the tumor, spleen, liver, and lung was analyzed from the perspective of safety and tumor specificity. Although a similar bacterial load was determined in the tumor and healthy organs at 3 dpi, the relative colonization of the therapeutic strain was considerably reduced in the spleen, liver, and lung at 7 dpi (p < 0.0001) and nullified at 14 dpi compared to the tumor (Fig. 5C). The persistence of the therapeutic strain, even after 14 days in the tumor but not in healthy organs, manifests the tumor specificity. Gross morphological changes in the spleen, liver, and lung of mice administered 5 × 10⁶ CFU once a week for 4 consecutive weeks were evaluated. There were no remarkable changes compared to naïve mice. Histopathological examination of tissue sections by H&E staining also showed healthy tissue topology (Fig. 5D).

Anti-cancer therapy of JOL2955P3

The in vivo anti-tumor effect of the designed therapeutic strain was investigated in a CT26 mouse colon cancer model. During Salmonella therapy, tumor growth was measured for up to 2 weeks after the fourth inoculation. In mice treated with JOL2955P3, substantial tumor regression was recorded (Fig. 6). Although both strains showed tumor reduction, a more significant reduction was observed in the JOL2955P3-treated group than in the VC group. Such a substantial tumor regression in the therapeutic strain treated group was noticeable from 14 days post treatment where tumor volume was nearly half of the remaining groups. At the end of the experiment, the tumor volume regressed by 80 % in treated mice compared to the control group (p = 0.0003), while the reduction was 38 % in the VC group (p = 0.0141) (Fig. 6D). The developed therapeutic strain was estimated to possess 67 % better potential than its vector control (p = 0.0094). Mild splenomegaly was noted in the placebo mice, whereas the spleens of the treated mice were comparable with those of naïve mice, showing minimal gross changes (Fig. 6F).

Fig. 6.

In vivo anti-tumor effects of JOL2955P3. A. Schematic schedule for the cancer treatment using Salmonella Typhimurium JOL2955P3. The tumor was induced in six-week-old BALB/c mice with subcutaneous (SC) injection of 5 × 10⁶ cells/mL CT26 cells. Animals were challenged intraperitoneally with the strains, and the tumor volume was measured by a Vernier caliper for up to 35 days post-treatment. B. Gross morphological appearance of the tumors in mice sacrificed at the end of the experiment. The red circle denotes the position and size of the tumor. C. Gross size and appearance of tumors and spleens harvested from respective groups. D. The volume (V) of a tumor was measured by applying the tumor volume formula, V = ½ (length × width²). The weights of the E. tumors and F. spleens. Mice were sacrificed at the end of the experiment, and the tumors and spleens were harvested and weighed. Data were analyzed by Tukey’s multiple comparison tests, one-way ANOVA, ns = not significant, *p < 0.05, **p < 0.01, and ***p < 0.001, where the error bars denote the SEM. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Effect of JOL2955P3 treatment on cytokine expression in tumors

The influence of Salmonella treatment on the tumor microenvironment was evaluated for selected cytokine expression by RT-PCR. The mRNA levels of IFN-γ (p = 0.0477) and IL-6 (p = 0.0313) were upregulated in JOL2955P3-treated mice (Fig. 7A). Although the VC treatment elevated the levels of these cytokines, the difference was not remarkable. The expression of IL-1β was significantly increased in both JOL2955P3 (p = 0.0020) and VC (p = 0.0152) groups compared to the placebo group; the expression in the JOL29533P3 group was 2-fold higher than that of the placebo group. For both groups, cell cycle regulatory genes were examined to evaluate treatment effects on tumor cell cycle regulation. The p21 expression was upregulated only in mice treated with therapeutic strains. The expression of Bax was upregulated while VEGF was downregulated in both groups. Among the 12 tested cancer-related cytokines and cell-cycle regulatory genes, six had significant changes upon Salmonella treatment; of them, expression of IL-1β and p21 was significantly higher than that with VC treatment.

Fig. 7.

Effects of treatment in tumors. A. Expression of cytokines and cell cycle regulatory genes. Relative in vivo expression of A. cytokines and B. apoptosis-related genes. RNA was extracted from the tumors and the expression of cytokines and apoptosis-related genes was quantified by RT-PCR. B. Expression of tumor-associated macrophage (TAM) markers in tumors. TAMs were isolated from tumors by density gradient centrifugation using Ficoll-Paque plus density gradient media. Isolated TAMs were stained with CD68-PE, iNOS-FITC, and CCL2-APC and quantified by FACS. A higher number of marker-positive cells was noted in tumors from non-treated mice, while positive cells were fewer in tumors from mice treated with JOL2955P3. Data were analyzed by Tukey’s multiple comparison tests, one-way ANOVA, *p < 0.05, **p < 0.01, and ***p < 0.001, where the error bars denote the SEM. C. Hematoxylin and eosin (H&E) staining of tumors. Paraffin-fixed tumor tissues were sectioned, deparaffinized, and stained with H&E. The necrotic lesions in treated samples are denoted by red arrows. D. Immunohistochemical (IHC) observation of tumors, spleens, and livers to confirm the expression of cytolysin A (ClyA). Tissue sections were treated with anti-ClyA rabbit polyclonal antibodies and developed color with DAB. The brown spots marked inside the circle indicate the expression of ClyA in the tumor region of treated samples. The absence of brown spots in the spleen and liver in treated mice denotes the controlled expression of ClyA. The samples from the vector control and placebo groups also lack brown spots. Images were acquired at 200× for H&E and at 400× magnification for IHC. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Infiltration of tumor-associated macrophages

The TAMs level in the tumors was investigated by FACS using TAMs makers. Tumor tissue was found with a high percentage of CD68, iNOS, and CCL2 TAMs in control mice. Those in the treated tumor were substantially reduced by 50–60 % in the VC group and 85–89 % in the JOL2955P3 group. The therapeutic strain decreased TAMs by about 60–70 % compared to the VC group (Fig. 7B). The reductions in CD68, iNOS, and CCL2 expressed in the TAMs in tumors of the engineered bacteria-treated mice suggest modulation of immune cells in TMEs to enhance the efficacy of immunotherapy. Such a process could relieve TME from immunosuppression and improve anti-tumor immunity.

Histopathological analysis

Histopathological examination of tumor sections by H&E staining demonstrated prominent necrotic lesions as an effect of Salmonella therapy. The therapeutic strain induced a larger region of necrosis compared to Salmonella with an empty vector (Fig. 7C). The controlled expression of ClyA under QS promoter in the tumor, spleen, and liver was examined by IHC staining using rabbit anti-ClyA polyclonal antibodies. Clear brown foci were noted in the tumor sections of JOL2955P3-treated mice, but such foci were absent in the spleen and liver of the same group of mice (Fig. 7D). This confirmed confinement of the strain with higher cell density in the tumor to produce ClyA activated by the QS promoter. Of note, there were no brown foci of ClyA expression in the control and VC groups.

Discussion

Salmonella has been engineered and utilized as a therapeutic agent after attenuation or as a vector to deliver anti-tumor molecules [10] due to its intrinsic tumor colonization properties. Cytolysin A is a cytolytic toxin that lyses erythrocytes and diverse nucleated cells through transmembrane pore formation [24]. Cytotoxicity has been demonstrated for anti-cancer therapy by constitutive expression via plasmid [36], controlled expression under the regulation of a highly hypoxia-inducible promoter [37], and photothermally-programmed expression [38]. As a combination therapy, ST ΔppGpp expressing ClyA under arabinose inducible pBAD promoter was studied in a mouse model [39]. The clinical application of ClyA or any such toxin demands tumor-specific scrutinized expression because of its cytolytic activity on normal cells. The controlled expression via photothermal activation requires an additional approach and complications. The hypoxia-inducible promoter can express the toxin specifically in the tumor, but the stringent hypoxic expression may not occur in the vasculated tumor region, where the predominant population of viable cancer cells exists. Intriguingly, ClyA expression under QS signaling control can address these issues. In this study, a plasmid, pJHL90, containing the QS promoter for toxin expression was developed (Fig. 2A). Being a facultative anaerobe and a tryptophan auxotroph, the developed therapeutic strain colonized well in the tumor. Consequently, the cell density and QS signaling molecule (AI-2) concentration increased in the tumor. Thus, the QS-regulated ClyA expression was activated specifically in the tumor without any additional/external stimuli. As a proof of principle, in vitro ClyA expression was detected only at higher bacterial cell density but not at the initial growth phase (Fig. 2C). Furthermore, the hemolytic activity was detected only in culture supernatant at high OD (Fig. 2D), demonstrating QS-regulated ClyA.

Auxotrophs for tryptophan pressurize the bacteria toward tumors [17] as they are a niche full of nutrients [40]. The tryptophan gradient has been documented to be two-fold higher in the tumor than in the serum of the CT26 colon cancer mouse model [41]. This could be due to preferential bacterial colonization in the tumor. Excessive tryptophan facilitates tumor immune escape by inducing immunosuppression through its metabolite, kynurenine, via apoptosis of effector immune cells [42]. The auxotroph competes for tryptophan consumption and indirectly could increase immune surveillance. Therefore, ST auxotrophic for tryptophan was selected as the precedent strain in the study. The widely studied cancer therapeutic ST strains VNP20009 and A1-R were created by mutagenesis. Non-targeted random mutagenesis caused undesired mutations. Broadway et al. reported Suwwan deletion (108 kb) and 50 nonsynonymous single-nucleotide polymorphisms in VNP20009 [15]. Furthermore, the VNP20009 strain underwent a phase I clinical trial and was terminated due to its weak anti-tumor effect and dose-dependent toxicity [12]. Therefore, to fill the gap, a reliable and well-defined strain was engineered by the highly efficient and stable λ red recombineering approach to delete the target genes [43]. The strain created by this method is precise, stable, and has little chance of reverting to its earlier form due to its permanent deletion. Since live Salmonella is used as a delivery/therapeutic agent, safety is the prime concern. An ideal onco-therapeutic Salmonella is anticipated with low endotoxicity to colonize specifically in the tumor but not healthy organs of the reticuloendothelial system (RES). For this purpose, a tryptophan auxotrophic strain was engineered and attenuated by pagL and rfaL deletion.

The ubiquitousness of Salmonella and poor hygiene practices increase the risk of Salmonella infection. Consequently, pre-existing antibodies may impede the treatment. Predominantly, the antibodies are developed against surface carbohydrate antigens [22]. Therefore, O-antigen was eliminated by rfaL deletion in the therapeutic strain. A considerable number of therapeutic strains was recovered from the spleen and liver of mice primed with O-antigen-positive and negative Salmonella (Fig. 3A). Thus, a significant avoidance of pre-existing antibody interference was observed, where the results were found to be in accordance with a previous report [44]. Moreover, a large proportion of O-antigen-deficit Salmonella (JOL2953) escaped from pre-existing antibodies upon priming and reinfection. This trait is envisaged to be useful to avoid pre-existing antibodies and also to have a considerable number of therapeutic Salmonella in tumors upon repeated inoculation. The pagL-deleted mutant exhibited reduced expression of endotoxicity-related pro-inflammatory cytokines such as TNF-α, IL-1β, and IFN-γ (Fig. 3B). Thus, the RT-PCR results confirmed the low endotoxicity of the therapeutic strain. The result is consistent with pagL mutation-based low endotoxicity reported previously [19]. The low endotoxicity might be due to the modified interaction between lipid A and TLR-4. Consequently, it might activate the TLR4-TRAM-TRIF pathway that induces type I interferon and immune activation instead of the TLR4-TIRAP-MyD88 pathway, which causes inflammation via TNF-α, IL-1β, and IFN-γ production [45]. Thus, the pagL-truncated Salmonella deciphered the compromised endotoxicity via reduced expression of inflammatory cytokine. Therefore, pagL- and rfaL-deleted tryptophan auxotrophic Salmonella was developed as an anti-cancer therapeutic delivery vehicle.

The attenuated tryptophan auxotrophic ST possesses reduced fitness in healthy tissues due to tightly regulated nutrient metabolism and no free tryptophan. To further improve preferential colonization in tumors, the strain was in vivo passaged thrice. As anticipated, the retrieved strain developed improved adhesion and invasion both in murine and human cancer cell lines (Fig. 4A). The improved intracellular survival of the passaged strain is projected as the consequence of the higher invasion. In addition, as in vivo confirmation, bacterial accumulation in the tumors increased, confirming the enhanced fitness of the in vivo passaged strain. The repeated passage could have induced bacterial memory in the form of transcriptional reprogramming to sense the tumor environment, as recurring exposure led to efficient adaptation [46]. In addition, nutrient abundance and environmental stress help to establish bacterial memory response [47].

The final therapeutic strain was developed by genetic engineering to arrive at multifactorial benefits including the default anti-tumor potential, low endotoxicity, pre-existing antibody evasion, and simultaneous delivery of cytolytic ClyA in a controlled fashion. The in vitro results of the wound healing assay and apoptosis FACS analysis revealed a substantial cancer cell growth inhibitory potential and apoptosis induction by both strains, but JOL2955P3 was found to be better than the vector control due to its ClyA expression (Fig. 4C). Studies have shown that the engineered ST has robust anti-cancer activities. Implementation of such therapeutic bacteria necessitates a thorough safety investigation. The 100 % survival depicted by the survival curve and the absence of outward histological changes in the spleen, liver, and lung support the safe administration and low endotoxicity of the therapeutic strain (Fig. 5B & 5D). The inoculated bacteria was retained for up to 14 days in the tumor but was not recovered from the healthy organs (Fig. 5C). This result illustrates the tumor specificity of the therapeutic strain and the ability to sustain and grow to high cell density only in tumors. This phenotype further supports the concept of QS-based tumor-specific expression.

In order to achieve tumor-specific colonization and low endotoxicity and to evade pre-existing antibodies, Salmonella was engineered. These attenuations could compromise the virulence against tumor cells, where collateral impairment allowed pagL and rfaL to be interlinked with many virulence traits [48]. For this reason, ClyA under QS regulation was incorporated to compensate for lethality and developed the therapeutic strain. The expression of ClyA by the therapeutic strain enhanced its tumoricidal activity by 67 % compared to the strain without it (Fig. 6). The tumor microenvironment (TME) has a decisive role in the proliferation of tumors particularly by enrichment of TAMs. Considerable decreases of the CD68, iNOS, and CCL2 TAM populations were observed (Fig. 7B). These reductions of TAMs by Salmonella treatment provide essential cues to suppress the tumor microenvironment. TAMs promote cancer proliferation by invasion, angiogenesis, and tumor cell growth in addition to suppressing T-cell response [49]. As a proof of concept, the reduction of tumor-promoting TAMs might be a factor for tumor therapy. A detailed understanding of the modulation of TAMs by Salmonella invasion could be a new strategic approach against a tumor.

As an additional confirmation of tumor regression, RT-PCR was used to measure the levels of cancer-related cytokines and genes associated with the cell cycle. The IL-1β expression was upregulated in tumors treated with JOL2955P3 (Fig. 7A). A previous report states that IL-1β has a pivotal role in suppression of colon cancer [50]. IL-6 is a pleiotropic cytokine that has both pro-inflammatory and anti-inflammatory functions. IL-6 in the tumor microenvironment supports enhanced trafficking of CD8⁺ T cells into tumors [51], and the activated CD8⁺ T cells kill tumor cells [52]. Upregulated expression of IL-6 in treated tumors contributed to cytolytic activation. In addition, the enhanced production of IFN-γ in treated mouse tumors might also activate CD8⁺ T lymphocytes [53]. Cyclin-dependent kinase inhibitor p21 promotes cell cycle arrest mainly in association with p53 and displays a tumor-suppressive function [54]. The mild upregulation of p21 in treated tumors could also be one of the reasons for tumor suppression. Studies have shown that p21 has an apoptotic role, independent of p53 [55]. Similarly, the upregulated expression of a proapoptotic gene, Bax, may promote apoptosis in treated animals [56]. VEGF is a proangiogenic factor that helps in vascularization and angiogenesis to promote cancer [57]. The suppression of VEGF production in mice treated with a therapeutic strain supports halting the angiogenesis and controls tumor growth.

Histopathological analysis of tumor sections performed by H&E staining demonstrated prominent necrotic regions in treated mice. In comparison with VC mice, a more widespread area of necrosis was noted in JOL2955P3-treated mice, which supports the increased tumoricidal activity of the therapeutic strain (Fig. 7C). It is clear from such pronounced tumoricidal effects of the therapeutic strain that ClyA is responsible. The QS-regulated ClyA expression was demonstrated in vitro by western blot and a hemolytic assay. However, the applicability of the concept in a realistic scenario demands in vivo confirmation. Therefore, IHC analysis was performed using the ClyA-specific rabbit antibodies. The results clearly showed the brown spot indication of ClyA expression only in tumors and not in other healthy organs (Fig. 7D). Thus, the in vitro and in vivo results exemplified the QS-regulated tumor-specific expression of toxins in mice.

Conclusion

In this study, the intrinsic anti-tumor ability of Salmonella was utilized, and its tumor specificity and penetrating caliber was enhanced by tryptophan auxotroph. The safety of the delivery strain was assured with reduced endotoxicity imposed by pagL mutation, and selective tumor colonization was enhanced via subsequent in vivo passage. It was demonstrated that the auxotrophic ST strain is avirulent and selectively colonized in the syngeneic CT26 tumor. It safely delivered ClyA under the tight regulation of the QS signaling mechanism. This engineering approach ensured the expression of the cytolytic protein in the tumor microenvironment only at high bacterial cell density. In conclusion, the delivery of the cytolytic proteins under the control of quorum sensing via the attenuated ST strain was an effective and safe therapeutic strategy against cancer.

Compliance with Ethics Requirements

All animal experiments performed in this study were according to the methods approved by the Jeonbuk National University Animal Ethics Committee (CBNU2021-027) under the guidelines of the Korean Council on animal care and the Korean Animal Protection Law, 2007, Article 13.

CRediT authorship contribution statement

Ram Prasad Aganja: Conceptualization, Investigation, Methodology, Data curation, Formal analysis, Writing – original draft, Writing – review & editing. Chandran Sivasankar: Writing – review & editing. John Hwa Lee: Conceptualization, Resources, Funding acquisition, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

The following are the Supplementary data to this article: