Bacillus velezensis inhibits azoxymethane/dextran sulphate sodium induced colitis associated colorectal cancer via the small molecule HeLM

Introduction

The incidence of inflammatory bowel disease (IBD), having grown significantly over the past five decades, appears to be reaching an apex in Europe and North America, with a prevalence of approximately 1 in 100 (1,2). However, the global burden of the disease is still increasing as more nations shift to an urban and “western” lifestyle (2). The most common of the subtypes of IBD, ulcerative colitis (UC), is characterised by chronic inflammation that is limited to the colon and rectum. It is the result of an abnormal immune response to the colonic microbiome, in genetically susceptible individuals (3).

A diagnosis of UC significantly increases the risk of developing colorectal cancer (CRC), such that after thirty years from the time of diagnosis, the risk of CRC is 18%, with this being listed as the cause of death for 15% of all UC patients (4,5). The risk of developing this type of cancer, termed colitis associated cancer (CAC), directly correlates with the severity and chronicity of the colitis (6). Symptoms of the disease classically start around the second decade of life, as such, patients are exposed to cumulative years of inflammation from an early starting point, subsequently, leading to a cancer diagnosis at a younger age than de novo CRC seen in the wider population (7). Moreover, the tumour biology of CAC has distinct genetic variances when compared to de novo CRC. These include early TP53 mutations as well as chromosomal instability, which explains, in part, the more aggressive characteristics of CAC (8). Consequently, CAC is diagnosed in a younger cohort, at an advanced stage, with poor response to chemotherapy, and with a worse prognosis (9,10).

Medical therapy has improved considerably over the past two decades as biologics and small-molecule therapies have come on board, acting on specific inflammatory pathways (11). Clinical trials using experimental faecal microbial transplantation (FMT) to modulate the microbiome are also showing promise for further therapeutic options (12). Nonetheless, 52 weeks after initiating therapy, deep remission is only achieved in 44% of patients (13). Indeed, approximately two thirds of patients will either fail to respond to, or lose response to, medical therapy, necessitating high rates of surgical resection (14,15). Outside of major resectional surgery, there is no known cure for UC (16). The goal, therefore, of medical therapy must be twofold: (I) to improve the quality of life by limiting the symptoms of disease; (II) to achieve long-lasting biological remission to reduce the risk of inflammatory-related malignant transformation.

Our group has recently established that the oral administration of a strain of Bacillus velezensis (EHv5), isolated from healthy human faecal samples, significantly reduced colonic inflammation in a murine model of acute UC (17). The mechanism of action is via a secondary metabolite produced during sporulation—HeLM. It remains to be seen if this holds true in the chronic setting. Our hypothesis is as follows: that sustained administration of this strain of Bacillus velezensis will show efficacy in the setting of chronic colitis, sufficient to reduce the risk of CAC. Accordingly, we will report on a pre-clinical trial of EHv5 using dextran sodium sulphate (DSS) to induce a chronic colitis in mice that have an increased propensity to develop colorectal neoplastic transformation due to the administration of the carcinogen azoxymethane (AOM). This will be tested against both an isogenic mutant of the strain that does not produce HeLM, as well as positive and negative controls. We will measure two primary end-points: (I) the amelioration of colonic inflammation over a 3-month period, as measured by a variety of clinical and biochemical parameters; and (II) the prevention or reduction in CAC as measured endoscopically, histologically, and by total tumour burden. We present this article in accordance with the ARRIVE reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-24-610/rc).

Methods

Bacillus velezensis strains

Two strains of Bacillus velezensis are reported in this study, both having been characterised elsewhere (18,19). (I) EHv5^WT is a wild-type strain that produces “HeLM”; (II) EHv5^HeLM‑ is an isogenic mutant of the above, it carries an insertional disruption of surfactin synthetase (srfAA) gene that is responsible for the non-ribosomal synthesis of surfactin, the majority component of HeLM, and crucial for micelle formation (19).

Interventions

Interventions, listed below, were administered via intra-gastric (I.G.) gavage using a single-use soft tip polypropylene feeding tube (20G, Linton Instruments) in 200 µL volumes per mouse every Monday, Wednesday, and Friday throughout the experiment.

• EHv5^WT—a suspension of EHv5 in phosphate-buffered-solution (PBS). Bacteria were grown in brain heart infusion broth (BHIB) (37 °C, 8 h), and then further sub-cultured in 200 mL of BHIB (16 h, 37 °C). The culture, consisting of (~80%) vegetative cells and (~20%) spores, was harvested (10,000 g, 10 min) and suspended in PBS for a concentration of 5×10⁹ (CFU)/mL.
• EHv5^HeLM−—a suspension of the isogenic mutant EHv5^HeLM− in PBS, prepared as above.
• Negative and positive control—PBS only.

Animals

Thirty-five specific pathogen-free 10-week-old female BALB/cN mice (Charles River, UK) were maintained in standard laboratory cages (5/cage) and conditions (18–21 °C; 12-hour light/dark cycles) with ad libitum chow (5LF5, LabDiet) and drinking water. Each cage was randomly assigned to their respective groups (n=10), with the negative control of 5 (Figure 1). Due to the prolonged nature of the experiment, female-only cages were used to avoid in-group violence that is typical with prolonged experimentation. All animals were sacrificed on Day 80. If any subjects reached the experimental limits necessitating euthanasia, or died, during the experiment, this was recorded and all results were then analysed on an intention to treat (complete case) basis. All experimental work on animals conformed to the Guide for the Care and Use of Laboratory Animals, Eighth Edition (2011). The experiment went through ethical review at the Institution’s Animal Welfare Ethical Review Body (AWERB) and was approved by the UK Home Office (Project Licence Number: 70/8276).

Figure 1 Schematic of the experiment. On the Y axis, a breakdown of each group, number of mice per group, and sex ratio. On the X axis is the timeline of the experiment with the timepoints for each investigation and cycle of colitis. AOM, azoxymethane; CACI, chronic colitis associated with cancer activity index; DAI, Disease Activity Index; DSS, dextran sulphate sodium; IP, intraperitoneal; PBS, phosphatebuffered-solution.

Induction of chronic colitis and CAC

This murine model was first described in 1996 (20). Briefly, AOM is a potent procarcinogen that is administered via an intraperitoneal injection. It is metabolised by CYP450 in the liver to methylazocymethanol (MAM). MAM is then excreted via the biliary system and taken up by colonic epithelial cells. MAM is a highly reactive alkylating agent that induces O⁶-methylguanine adducts in DNA, leading to transition point mutations (21,22). DSS, is toxic to colonic mucosa and produces a colitis that morphologically resembles UC (23). This is believed to be driven by immunogenic stimulation by the colonic microbiome that is presenting to the now-exposed epithelium. The combination of a single AOM dose and DSS induced colitis results in an average tumorigenesis time of less than 8 weeks. In the absence of inflammation AOM is unlikely to induce CRC, indeed one study using repeated and high-dose injections of AOM found it took up to 52 weeks for CRC to develop (24).

After a 7-day acclimatazation period, each mouse was weighed and inspected to establish their starting body weight and to ensure no obvious health defects. On Day 1, under general anaesthesia (isoflurane), each mouse was given a single intraperitoneal injection of 12 mg/kg AOM (Sigma-Aldrich). A baseline colonoscopy was performed, ensuring no aberrant colonic pathology. On Day 10 of the experiment, drinking water was replaced with 2.5% DSS [(w/v)] (colitis grade 36,000–50,000 Da, MP Biomedicals). This was made fresh every 2 to 3 days, and lasted for a period of 7 days, whereupon it was replaced with standard drinking water. The mice then entered a period of recovery. This represents one cycle of colitis. A proportion of 2.5% DSS was given twice during the experiment (weeks 2 and 7), with the experiment ending in week 12 (Figure 1). Previous work suggested that exceeding two cycles of 2.5% DSS would risk reaching the severity limits of our licence and result in an underpowered study.

Clinical assessment of colitis severity

The Disease Activity Index (DAI) (25) was calculated every 2–3 days, starting on Day −1. Microscopic rectal blood was assessed with a guaiac-based faecal occult blood test (Hema-Screen, Immunostics). To reflect the chronic and oncogenic nature of this experiment, a new scoring system was devised—the chronic colitis-associated with cancer activity index (CACI). This was measured in parallel with the DAI to assess its utility. For the score matrix, see Table 1 and Figure S1. For both scoring systems, colitis was defined as a significantly increased score compared to the negative control. On top of the DAI and CACI scores, the time to recovery (TtR) was also measured (days). Recovery was defined as when the clinical score was no longer significantly different from the negative control. Endoscopic evaluation of colitis was also undertaken, with the Mouse Colitis Endoscopic Index being assessed at regular intervals throughout the experiment (27). More detailed methodology on endoscopy can be found below.

Biochemical assessment of colitis

Faeces were collected via a fresh catch method and stored at −20 °C for later work. The heterodimer protein S100A8/S100A9 (calprotectin) is a pro-inflammatory compound released upon neutrophil apoptosis, and is widely used in clinical practice as an objective measurement of colitis severity (28). Following the manufacturers’ protocol, a two-site sandwich enzyme-linked-immuno-sorbent-assay (ELISA) was used for the direct measurement of calprotectin (measured at 450 nm) of each faecal sample (Immunodiagnostik AG & SpectraMax Plus384). Data was analysed using the Four Parameter Algorithm method to generate ng/mL quantifications of calprotectin.

Haematinics

Cardiac puncture was performed immediately after culling and 1–2 mL of blood was taken and mixed in paediatric K3EDTA tubes. These serum samples were processed to obtain a Complete/Full Blood Count (Royal Veterinary College, UK). Reference ranges for BALB/c mice have been published previously (29), with anaemia defined as a Hb less than 12 g/dL and microcytosis less than 44.3 fL.

Assessment of CRC

Every four weeks, the mice would undergo general anaesthesia to facilitate colonoscopy, starting on Day −1. Briefly, mice were anaesthetised using inhaled isoflurane via nose cone to an approximate (MAC) value of 1.0 (30). In the supine position, faecal matter was cleared with water enemas using a 18G soft-tip polypropylene feeding tube. A 10 cm 0°, 1.9 mm rigid endoscope with a portable light source (Karl Storz & ClaraMed) was used to examine the colon under direct vision. The endoscope was modified with an additional channel to provide further water irrigation, allowing the colon to be examined under water. The (MCEI) was calculated during the procedure. A phone-endoscope adaptor (ClaraMed) was used to capture images. Extensive testing before the experiment on deceased mice demonstrated that with experience, colonoscopy could reach 8 cm distally along the length of the colon, up to the caecal pouch. Tumour burden was assessed as being present or not. If present, the number of tumours was counted. On the final day of the experiment, the animals were culled, the colorectum harvested, and the number of tumours counted, with the total volume of tumour measured using callipers. The colon was “Swiss-rolled” and fixed in 10% (v/v) neutral buffered formalin (NBF) for histological processing with haematoxylin and eosin (H&E) staining. The tumours were analysed under the microscope (Olympus CKX41) looking for the highest grade of neoplasia, as per the Royal College of Pathologist’s standard (G049). These scores were categorised as either: (I) no abnormality; (II) low grade dysplasia; (III) high grade dysplasia/adenocarcinoma. The depth of invasion was also recorded using the Union for International Cancer Control’s (UICC) TNM8 definitions. Any visible metastasis at the time of harvesting the colon was recorded and histologically processed with H&E staining.

Statistical analysis

All animals were included in the analysis. GraphPad Prism (9.4.1, Dotmatics) was used for analysis. Two-way repeated measures ANOVA was used to determine the effect of the different treatment groups over time for the DAI, CACI scores. Sphericity was assumed to be present (per the Greenhouse & Geisser method). Tukey’s post hoc multiple comparisons analysis was performed and data is graphed as the mean, unless otherwise stated, the positive control was used as the comparator. One-way ANOVA was used for the macro and microscopic assessments of tumour burden, RT-qPCR and haematinics. Kaplan-Meier survival analysis was employed for the polyp detection during endoscopy. The significance threshold was 0.05.

Results

Severity of colitis (CACI)

The first cycle of 2.5% DSS was given on Day 10. Both the DSS-only and EHv5^HeLM- groups, when compared to the negative control, showed a significant rise in their CACI score on Day 15, both scoring 2.6 [P=0.002, 95% confidence interval (CI): −3.199 to −1.112 & P=0.003, 95% CI: −3.355 to −1.045]. EHv5^WT did not reach significance until Day 18 with a score of 2.4 (P=0.002, 95% CI: −3.378 to −0.6222), Figure 2A. On Day 20, the severity of colitis peaked, with scores for the DSS-only group being 4.8, EHv5^HeLM− was similar with a score of 4.1 (P=0.69). In contrast, EHv5^WT scored 1.7, which was significantly lower than the DSS-only group (P=0.001, 95% CI: 0.9675 to 3.855). Indeed, during this first cycle, the EHv5^WT group remained significantly different from the DSS-only group throughout, demonstrating substantial protection. The TtR for EHv5^WT was shorter (7 days) compared to DSS-only and EHv5^HeLM− (both 13 days). The second cycle of 2.5% DSS started on Day 48 and resulted in a more chronic and severe colitis. After 48 hours, the DSS-only and EHv5^HeLM− groups demonstrated significant colitis, with scores of 1.8 and 1.9 (P=0.03, 95% CI: −3.020 to −0.1352 and P=0.017, 95% CI: −3.117 to −0.2833) respectively. On Day 53, the EHv5^WT group had a significantly raised CACI score of 1.3 (P=0.01, 95% CI: −1.956 to −0.2347). On Day 60, the peak CACI for the DSS-only, EHv5^HeLM−, and EHv5^WT were 5.4, 5.5 (P>0.99, 95% CI: −3.659 to 3.548), and 3.2 (P=0.22, 95% CI: −1.008 to 5.497) respectively. The DSS-only and EHv5^HeLM− groups continued to demonstrate colitis for the rest of the experiment, and never “recovered”, hence their TtR was >32 days. In contrast, EHv5^WT TtR was 22 days. During this second cycle, the EHv5^WT group was statistically non-different from the DSS-only group for 8 days, suggesting that whilst there was still a therapeutic effect, the level of amelioration was less protective during this second cycle when compared to the first cycle.

Figure 2 Clinical scores throughout the experiment. For both graphs, α indicates the initiation of DSS (days 10 and 48) whilst Ω shows the cessation. (A) Graph showing CACI scores over time. On day-80 (end of experiment) the mean score for EHv5^WT was 1.2 which was significantly different (P=0.047) from the positive control which scored 4.6. This protection from chronic colitis was not seen in the mutant (EHv5^HeLM−), which scored 4.3. (B) The DAI score over time. At the end of the experiment, EHv5^WT was significantly different from the positive control (P=0.02) with a score of 0.6 compared to 3.1. There was no protective effect seen in the mutant group. CACI, chronic colitis associated with cancer activity index; DAI, Disease Activity Index; DSS, dextran sodium sulphate.

Severity of colitis (DAI)

The DAI and CACI were well correlated during the first cycle, but meaningfully diverged during the second cycle of DSS. The latter cycle demonstrated the DAI over-estimating the severity in mice with cancer, and under-estimating the mice without. During the first cycle, the peak DAI score for the DSS-only and EHv5^HeLM- groups on Day 20 were 4.2 (P<0.001, 95% CI: −6.054 to −2.391) and 4.6 (P<0.001, 95% CI: −6.291 to −2.909) respectively. EHv5^WT peak score was 1.9 on Day 18 (P=0.009, 95% CI: −2.759 to −0.4412), Figure 2B. Similar to the CACI system, EHv5^WT remained significantly different to the positive control throughout this first cycle. TtR for EHv5^WT was also shorter (7 days) compared to DSS-only and EHv5^HeLM− (both 15 days). After the second cycle, all groups had a peak score on Day 55, with the DSS-only scoring 6.8, EHv5^HeLM− 5.7 (P=0.44, 95% CI: −1.345 to 3.123), and EHv5^WT 2.2 (P<0.001, 95% CI: 2.079–7.077). Indeed, EHv5^WT had a significantly improved DAI score throughout the second cycle, signifying protection from colitis. According to the DAI, EHv5^WT never statistically diverged from the negative controls, as such, the TtR was 0 days. The DSS-only and EHv5^HeLM− groups TtR were both 29 days.

Severity of colitis (endoscopic)

All four groups had normal examinations on Day −1. In week 4, the DSS-Only group had a mean MCEI score of 4.5. EHv5^HeLM- had a similar score of 4.7 (P=0.96, 95% CI: −1.436 to 1.036). EHv5^WT score was calculated as 2.8 (P=0.001, 95% CI: 0.6763–2.724), Figure 3A. The negative control was 0. In week 8, all mice receiving DSS had worse scores. DSS-only scored 6.8. Similarly, EHv5^HeLM− scored 6.9 (P=0.99, 95% CI: −1.551 to 1.351). EHv5^WT scored 4.6 (P=0.002, 95% CI: 0.8102–3.590). The negative control mean was 0.2. Week 12 demonstrated a more marked difference between groups, with DSS-only having a mean score of 6.8, EHv5^HelM− 6.5 (P=0.96, 95% CI: −1.698 to 2.298). In stark contrast, EHv5^WT demonstrated significant mucosal healing with a mean score of 1.8 (P<0.001, 95% CI: 3.202–6.798).

Figure 3 Endoscopic and biochemical evaluation of colitis. (A) MCEI scores throughout the experiment. A score of 0 indicates no colitis, 9 is the maximal severity score. EHv5^WT was significantly protected compared to the negative control. (B) Faecal calprotectin levels taken after the cycle of 2.5% DSS. After both cycles, EHv5^WT, whilst elevated, was significantly lower than the positive control. *, P≤0.05; ****, P≤0.0001; ns, non-significant. DSS, dextran sodium sulphate; MECI, mouse endoscopic colitis index.

Severity of colitis (biochemical)

On Day 23, the positive control had a mean faecal calprotectin of 42.0 ng/mL. EHv5^HeLM− had a similar mean of 32.1 ng/mL (P=0.41, 95% CI: −8.524 to 28.30). In contrast, EHv5^WT had a mean of 19.1 ng/mL (P=0.01, 95% CI: 4.460–41.29), suggesting that colitis after the first cycle was significantly reduced, Figure 3B. The negative control was 8.1 ng/mL. On Day 60, all groups receiving DSS had higher faecal calprotectin levels compared to the first cycle. The positive control’s mean was 71.1 ng/mL. EHv5^HeLM− averaged 62.2 ng/mL (P=0.49, 95% CI: −9.445 to 27.38). EHv5^WT was significantly less than the positive control, with a mean of 30.3 ng/mL (P<0.001, 95% CI: 22.43–59.26). The negative control’s mean was 6.5 ng/mL.

Haematinics

The negative control had a mean Hb of 14.4 (range, 13.3–14.9) g/dL; DSS-only was 11.6 (range, 8.4–16.5) g/dL (P=0.15, 95% CI: −0.7451 to 6.345); EHv5^HeLM− was 12.3 (range, 9.2–17.6) g/dL (P=0.33, 95% CI: −1.420 to 5.670); EHv5^WT was 14.7 (range, 9.9–17.4) g/dL (P=0.99, 95% CI: −3.920 to 3.170), Figure 4A. When analysing for the presence of anaemia, 78% and 75% of the DSS-Only and EHv5^HeLM− mice were anaemic, respectively (P=0.005 & P=0.009), whereas 25% of the EHv5^WT had anaemia (P=0.97), Figure 4B. None of the negative controls demonstrated anaemia. The mean MCV was 44.75 fL for the negative control, 40.5 fL for DSS-only, 41.33 fL for EHv5^HeLM−, and 45.75 fL for EHv5^WT, Figure 4C. When analysis for the presence of microcytosis, the DSS-only and EHv5^HeLM− groups were significantly different from the negative control (P=0.008 & P=0.03, respectively), whereas EHv5^WT was protected, Figure 4D. All samples that were anaemic were found to have microcytic anaemia, strongly suggesting this phenomenon to be an iron deficiency anaemia driven by tumour-related blood loss.

Figure 4 Haematological evaluation. (A) Hb (g/dL) levels taken at the end of the experiment. The red line at 12 g/dL indicates the lower boundary for “normal”—i.e., below 12 indicates anaemia. There was no significant difference in Hb between the groups. (B) The percentage of mice in each group with anaemia, analysing the results this way shows that the positive control and the mutant had significant rates of anaemia, a finding not seen in the EHv5^WT group. (C) MCV (fL) values. The red line at 44 fL is the boundary, below which signifies microcytosis, no significant difference between the groups were found. (D) The percentage of mice in each group with microcytosis, with the EHv5^WT group being non-significantly different to the negative control. *, P≤0.05; **, P≤0.01; ns, non-significant. Hb, haemoglobin; MCV, mean corpuscular volume.

CRC burden

There were no tumours seen during the first endoscopy session. By week 4, no tumours were seen in EHv5^WT, however, 40% of the mice in both the DSS-only group and EHv5^HeLM- demonstrated tumour growth (P=0.10 for both) when analysed against the negative control group. Week 8 demonstrated that 90% and 80% of the mice had tumours in the DSS-only (P<0.001, 95% CI: −1.240 to −0.3961) and EHv5^HeLM− (P<0.001, 95% CI: −1.149 to −0.3052) groups respectively; 20% of the EHv5^WT group had tumours (P=0.66, 95% CI: −0.6039 to 0.2402), with none in the positive control group. At 12 weeks, 90% of mice in both the DSS-only group (P<0.001, 95% CI: −1.302 to −0.5162) and the EHv5^HeLM− (P<0.001, 95% CI: −1.211 to −0.4253) had visible tumour growth. In contrast, 30% of EHv5^WT mice demonstrated tumour growth (P=0.26, 95% CI: −0.6656 to 0.1202), with none in the positive control group. These results are charted as a Kaplan-Meir survival plot, Figure 5A,5B. When performing the statistical analysis on individual curves, EHv5^WT had significantly fewer “tumour events” than the DSS-Only group (P<0.001), suggesting that the group was well protected from CAC. EHv5^HeLM− did not confer any protection, with similarity observed with the DSS-Only group (P=0.66). Endoscopic evaluation was assessed for accuracy at the end of the experiment by comparing the final endoscopic detection of tumours against the presence of tumours in the opened colon under 4× magnification (Olympus CKX41). This demonstrated an endoscopic diagnostic sensitivity of 79% and a specificity of 100%.

Figure 5 Colorectal cancer burden. (A) Kaplan-Meier plot showing the probability of tumour detection during endoscopy throughout the experiment. The EHv5^WT curve was significantly different to the positive control (P<0.001). (B) Captured images during endoscopy, the top row shows normal mucosa, with all subsequent images demonstrating polyps (white star). (C) The mean total size of tumour(s) (mm) per group, with the EHv5^WT demonstrating a significantly reduced mean size of tumour compared to the positive control. (D) The total number of polyps counted at the end of the experiment, grouped into categories. There was no significant difference from “1–2” to “7–8”, there was, however, significant difference in the category “9+”. (E) Representative photos of the opened colorectum. The top left photo shows normal mucosa, as the photos progress there is an increasing tumour burden. Tumour growth identified with a white star. *, P≤0.05; ****, P≤0.0001; ns, non-significant. AOM, azoxymethane; DSS, dextran sodium sulphate.

On opening the colorectum, tumour burden was quantified by size (total size, aggregated, of all tumours), which ranged from 2 to 55 mm. EHv5^WT had a mean tumour size of 2.5 (range, 0–8) mm, which was significantly less than the positive control (P=0.049, 95% CI: −0.06969 to 32.33), which had a mean tumour size of 16.3 (range, 0–50) mm, Figure 5C. There was no difference between EHv5^HeLM−, with a mean of 22.6 (range, 0–55) mm, and the positive control (P=0.98, 95% CI: −18.33 to −13.93). There were no tumours seen for the negative control. Furthermore, the total number of polyps/tumours was counted and grouped categorically. This showed that the DSS-only and EHv5^HelM− groups were largely represented in the categories signifying a high tumour burden (e.g., “7–8” & “9+”). In contrast for EHv5^WT if tumours were present, they were more likely to be represented in the lower categories, with significantly less tumour burden in the 7–8 and 9+ categories (P=0.015 & P<0.001), Figure 5D. Figure 5E shows representative photos of differing colorectal tumour burdens.

Histological analysis

There were no abnormalities seen in the negative control. Of the abnormal samples, none penetrated the muscularis mucosa (i.e., the maximum stage observed was intramucosal adenocarcinoma), a feature of this murine model that has been noted elsewhere (31). The number presenting with normal histology was low in the DSS-only and EHv5^HeLM− groups, with 10% of the positive control samples demonstrating no abnormalities and none of the EHv5^HeLM− samples being normal (P=0.85, 95% CI: −0.2363 to 0.4363). In contrast, 40% of the EHv5^WT group had no abnormalities, although this did not reach significance (P=0.54, 95% CI: −0.8917 to 0.2917), Figure 6A,6B. LGD was seen in 10% of the positive control, 20% of EHv5^HeLM− (P=0.97, 95% CI: −0.6081 to −0.4081), and 40% of EHv5^WT (P=0.54, 95% CI: −0.8917 to 0.2917). (HGD/AC) was seen in 80% of the positive control samples, 80% of EHv5^HeLM− (P>0.99, 95% CI: −0.4858 to 0.4858), and only in 20% of EHv5^WT (P=0.04, 95% CI: 0.02983–1.170). On secondary analysis, the level of efficacy EHv5^WT confers against developing HGD/AC was non-significantly different from the negative control group (P=0.59, 95% CI: −0.6483 to 0.2483), which was not observed in the other groups.

Figure 6 Histological analysis. (A) Histology grouped into either “no abnormality”, “low grade dysplasia”, or “high grade/adenocarcinoma”. This was determined as per the Royal College of Pathologists standards for colorectal cancer. EHv5^WT provided significant protection from developing high grade dysplasia/adenocarcinoma compared to the positive control. (B) Representative histology slides, stained with hematoxylin and eosin, taken at 10× magnification. For the negative control (a), EHv5^WT group (b), DSS-only (c), and EHv5^HeLM− (d), tumour growth is highlighted with a white star. *, P≤0.05; ***, P≤0.001; ****, P≤0.0001; ns, non-significant. DSS, dextran sodium sulphate; HGD.

Discussion

The pathogenesis of CAC follows a commensurate version of the classic report by Vogelstein (32). In the case of IBD, inflammation is driving the carcinogenesis via the inflammation-dysplasia-carcinoma sequence (8). The inflammatory microenvironment associated with IBD causes tumorigenesis via three established mechanisms: (I) increased reactive oxygen (and nitrogen) species that lead to direct DNA damage and mutagenesis, initiating the process; (II) the activation of pro-survival and anti-apoptotic pathways in the colonic epithelium, contributing to tumour promotion; (III) local environmental changes (angiogenesis, migration and invasion of tumour cells) that support progression and metastasis (33). These three stages are believed to be, in large part, mediated by nuclear factor-κB (NF-κB) coordinated inflammation (34,35). Inflammatory cytokines and microbial antigens initiate several signalling cascades that lead to the translocation of NF-κB to the nucleus with subsequent upregulation of target genes (e.g., the cytokines TNF-α and IL-6) (35). These inflammatory cytokines are pivotal in the pathogenesis of IBD. Moreover, NF-κB has also been shown to induce the production of the anti-apoptotic proteins Bcl-1 and Bcl-X_L, as well as chemokines and proteases that drive tumour invasion and metastasis (36,37). In CAC these mutations, notably TP53, occur particularly early when compared to de novo CRC, in part explaining the more advanced nature of CAC (8,38,39). Accordingly, if one can achieve a cessation of, or at least a significant reduction in, the severity and duration that the colon is in a state of inflammation, this carcinogenic sequence can be arrested and the risk of CRC significantly reduced (40). Accordingly, a large multicentre study has found that the use of biologics was associated with a lower risk of advanced CAC in patients with UC (41). Interestingly, this beneficial effect was not found in patients with CD, potentially illuminating a separating mechanism of cancer progression in CD. There are several pathways that can be pharmacologically manipulated to interrupt this inflammation cascade, with a recent meta-analysis suggesting there may be benefit in utilising combinations of different therapies (42).

We have recently shown that the administration of EHv5^WT, isolated from healthy human subjects, successfully ameliorates acute colitis (17). We were able to demonstrate that the mechanism of action is mediated through the bioactive secondary metabolite HeLM. This bacterially derived small molecule (~10 nm) is a compound made from a family of three non-ribosomal cyclic lipopeptides (surfactin, iturin, and fengycin) that form micelles once the critical micelle concentration (CMC) is reached (18). To prove that the molecule was causative, the mutant EHv5^HeLM− was created using an insertional disruption of surfactin synthetase (srfAA). This gene encodes for the synthesis of surfactin (19). Without surfactin, HeLM is never assembled (19). Importantly, EHv5^HeLM− conveyed no protection against colitis, convincingly showing HeLM to be mechanistic. To understand the mode of action, cell culture work with Human Embryonic Kidney (HEK) cells showed that HeLM is a potent agonist of Toll-like receptor (TLR) 2, with activity present at low concentrations (0.00075 ng/mL). Gene expression analysis elucidated that this specific TLR2 agonism results in an anti-inflammatory IL-10-dependent phenotype (17). The significance of this TLR2-mediated anti-inflammatory response has been documented elsewhere (43,44). Interestingly, mice treated with EHv5^HeLM− failed to demonstrate this increase in gene expression for IL-10. Moreover, we observed a secondary effect of the HeLM molecule via the inhibition of Gram-negative lipopolysaccharide (LPS) to bind to TLR4 (17). LPS is a potent pathogenic endotoxin and is known to mediate colitis and sepsis (45). The HeLM-mediated inhibition of LPS/TLR4 binding was only evident at concentrations of 0.075 µg/mL and above. We hypothesise this effect is a result of the surfactant properties of the HeLM molecule. Interestingly, treatment with EHv5^WT or purified HeLM did not appear to disrupt the faecal microbiome of the subjects. The HeLM molecule is gaining interest in the medical research community, with similar findings being described in other (non-colitic) animal models (46-50).

The course of colitis can last for several months to years for patients with IBD, even longer when one includes the median 3.2-month delay in diagnosis reported in high-income countries (51,52). The leading question therefore remains—would the success seen with EHv5^WT in the acute setting translate into the chronic, and moreover, would the effect be large enough to reduce the risk of CAC?

Strikingly, the administration of EHv5^WT provided a meaningful reduction in the severity of colitis throughout the entirety of the experiment. This is demonstrated by significantly reduced DAI and CACI scores, both in terms of peak score, but also in the time taken to recover (i.e., the length of time in a colitic state). Furthermore, faecal calprotectin and endoscopic evaluation which were taken at regular intervals, both revealed a significant reduction in the severity of inflammation when compared to the positive control. The degree of amelioration against chronic colitis was such that it translated into a significant protection against the development of CAC. This is demonstrated by the stark difference in rates of high-grade dysplasia seen in the EHv5^WT group compared to the positive control. Regular endoscopic evaluation for tumorigenesis also demonstrated a significant reduction in a number of polyps at every time point. Indeed, 20% of the EHv5^WT group had high-grade dysplasia or adenocarcinoma on histological analysis, compared to 80% in the positive control. This is further reflected by the significant increase in microcytic anaemia that was found in the DSS-only and EHv5^HeLM− groups, presumably secondary to the increased cancer burden. These results also confirm our previous findings that the secondary metabolite HeLM appears to be causative, as EHv5^HeLM− conferred no protection, and was otherwise identical to the positive control for both severity of colitis and rates of tumour growth.

The next finding this study raises pertains to the utility of the DAI in the setting of chronic colitis in mice with tumours. The DAI regards: (I) stool consistency (normal, loose, diarrhoea); (II) weight loss (0%, 1–5%, 5–10%, etc.); and (III) bleeding per rectum (none, microscopic, macroscopic)—with the range of scores for each category being 0–4. This scoring system was first published in 1993 for an acute study of 7 days (25). We believe there are three shortfalls in applying this method to our three-month chronic CAC model. Firstly, as the cancerous polyps form, bleeding becomes a feature of the model, not a finding. As such, using haematochezia, in either its micro or macroscopic form, as a surrogate marker for inflammation could artificially lead one to conclude that inflammation is more severe than the reality might be. Parallels can be drawn to medicine in that rectal bleeding is a screening tool used for CRC, rather than a qualitative measurement for colitis. Secondly, measuring weight loss fails to capture failure to thrive (normal growth). The mice used in our study were 10–12 weeks old at the start of the experiment, indeed, this age range is commonly used in the literature. Laboratory mice, including BALB/c are not considered mature adults until 3–6 months, with growth continuing past 20 weeks of age (53). As such, for prolonged experiments like this, one expects the mice to gain weight, as demonstrated by the negative control. It is possible; indeed, our findings show this, that the DAI could score “0” for no weight loss, but in reality, the subjects are significantly losing weight when compared to their negative control counterparts—a “failure to thrive”. As such, we accounted for this in the new scoring system. Thirdly, to remove the subjective nature of assessing stool consistency, we employed a validated and economically viable method to assess the water content of the stool, which has been described elsewhere (26). This provides three grades of severity of diarrhoea using the capillary action of a paper towel to measure the water content of the stool. In our experience, there are two further clinical signs that signify severe colitis such that following them there is a high risk of mortality—(bloody) diarrhoea staining around the anus and rectal prolapse. As such, these were included as a further data point to measure the severity of colitis. As an apparent, or pseudo, rectal prolapse can occur on normal defecation in healthy mice for a brief moment of time, we used any time greater than 5 seconds as a cut-off for abnormal. When analysing the results from this experiment, we found a significant difference between the DAI and CACI for all the groups. Indeed, if the DAI were the only measurement, the EHv5^WT group would have been statistically similar to the negative control throughout the experiment, perhaps over-emphasising their level of protection from colitis. Our findings suggest that CACI is more reflective of the severity of chronic inflammation in the setting of CAC, and correlates well with faecal calprotectin measurements. Given the longstanding use of DAI for this animal model, we believe it should continue to be recorded for comparison to previous studies, although consideration should be given to this novel, and more specific, scoring system.

Conclusions

In summary, this pre-clinical murine model has demonstrated that the Bacillus velezensis strain EHv5^WT significantly ameliorates the severity of colitis in the chronic setting. Importantly, the anti-inflammatory effect is large enough that the risk of developing CAC was significantly reduced. We have reproduced our previous findings, with the employment of an isogenic mutant, showing that HeLM is the mechanistic molecule delivering the clinical effect. Finally, we have devised a new clinical scoring metric, CACI, that provides a more representative reflection of the severity of chronic colitis in the setting of CRC. EHv5 may have potential as an adjunct in the pharmacological armamentarium to treat UC, and a clinical trial is currently being developed to assess this.

Acknowledgments

We wish to thank Dr Eve Fryer (consultant pathologist (gastrointestinal and hepatic pathology), Oxford University Hospitals NHS Trust) for her guidance on the grading of histological specimens.

Footnote

Reporting Checklist: The authors have completed the ARRIVE reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-24-610/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-24-610/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-24-610/prf

Funding: This study was supported by funding from the UK Medical Research Council (MRC) under grant number MR/R026262/1 to S.M.C. Doctoral fellowship of Dr Edward Horwell was paid for by Ashford and St Peter’s NHS Foundation Trust (UK).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-24-610/coif). E.H. reports grant money for his PhD provided by Ashford & St Peter’s NHS Foundation Trust hospital. S.M.C. reports grant from UK Medical Research Council (MRC) under grant No. MR/R026262/1, and he is CEO of SporeGen Ltd. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All experimental work on animals conformed to the Guide for the Care and Use of Laboratory Animals, Eighth Edition (2011). The project was granted approval by the UK Home Office (Project Licence Number: 70/8276) as well as the institution’s Animal Welfare and Ethical Review Body (AWERB).

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Lewis JD, Parlett LE, Jonsson Funk ML, et al. Incidence, Prevalence, and Racial and Ethnic Distribution of Inflammatory Bowel Disease in the United States. Gastroenterology 2023;165:1197-1205.e2. [Crossref] [PubMed]
2. Wang R, Li Z, Liu S, et al. Global, regional and national burden of inflammatory bowel disease in 204 countries and territories from 1990 to 2019: a systematic analysis based on the Global Burden of Disease Study 2019. BMJ Open 2023;13:e065186. [Crossref] [PubMed]
3. Ordás I, Eckmann L, Talamini M, et al. Ulcerative colitis. Lancet 2012;380:1606-19. [Crossref] [PubMed]
4. Eaden JA, Abrams KR, Mayberry JF. The risk of colorectal cancer in ulcerative colitis: a meta-analysis. Gut 2001;48:526-35. [Crossref] [PubMed]
5. de Campos Silva EF, Baima JP, de Barros JR, et al. Risk factors for ulcerative colitis-associated colorectal cancer: A retrospective cohort study. Medicine (Baltimore) 2020;99:e21686. [Crossref] [PubMed]
6. Rutter M, Saunders B, Wilkinson K, et al. Severity of inflammation is a risk factor for colorectal neoplasia in ulcerative colitis. Gastroenterology 2004;126:451-9. [Crossref] [PubMed]
7. Loftus EV Jr, Sandborn WJ. Epidemiology of inflammatory bowel disease. Gastroenterol Clin North Am 2002;31:1-20. [Crossref] [PubMed]
8. Baker AM, Cross W, Curtius K, et al. Evolutionary history of human colitis-associated colorectal cancer. Gut 2019;68:985-95. [Crossref] [PubMed]
9. Yaeger R, Paroder V, Bates DDB, et al. Systemic Chemotherapy for Metastatic Colitis-Associated Cancer Has a Worse Outcome Than Sporadic Colorectal Cancer: Matched Case Cohort Analysis. Clin Colorectal Cancer 2020;19:e151-6. [Crossref] [PubMed]
10. Maryńczak K, Włodarczyk J, Sabatowska Z, et al. Colitis-Associated Colorectal Cancer in Patients with Inflammatory Bowel Diseases in a Tertiary Referral Center: A Propensity Score Matching Analysis. J Clin Med 2022;11:866. [Crossref] [PubMed]
11. Binienda A, Fichna J, Salaga M. Recent advances in inflammatory bowel disease therapy. Eur J Pharm Sci 2020;155:105550. [Crossref] [PubMed]
12. Haifer C, Paramsothy S, Kaakoush NO, et al. Lyophilised oral faecal microbiota transplantation for ulcerative colitis (LOTUS): a randomised, double-blind, placebo-controlled trial. Lancet Gastroenterol Hepatol 2022;7:141-51. [Crossref] [PubMed]
13. Alipour O, Gualti A, Shao L, et al. Systematic review and meta-analysis: real-world data rates of deep remission with anti-TNFα in inflammatory bowel disease. BMC Gastroenterol 2021;21:312. [Crossref] [PubMed]
14. Buhl S, Steenholdt C, Rasmussen M, et al. Outcomes After Primary Infliximab Treatment Failure in Inflammatory Bowel Disease. Inflamm Bowel Dis 2017;23:1210-7. [Crossref] [PubMed]
15. Roda G, Jharap B, Neeraj N, et al. Loss of Response to Anti-TNFs: Definition, Epidemiology, and Management. Clin Transl Gastroenterol 2016;7:e135. [Crossref] [PubMed]
16. Liu S, Eisenstein S. State-of-the-art surgery for ulcerative colitis. Langenbecks Arch Surg 2021;406:1751-61. [Crossref] [PubMed]
17. Horwell E, Vittoria M, Hong HA, et al. A Family of Cyclic Lipopeptides Found in Human Isolates of Bacillus Ameliorates Acute Colitis via Direct Agonism of Toll-Like Receptor 2 in a Murine Model of Inflammatory Bowel Disease. Dig Dis Sci 2024;69:3729-41. [Crossref] [PubMed]
18. Ferreira WT, Hong HA, Hess M, et al. Micellar Antibiotics of Bacillus. Pharmaceutics 2021;13:1296. [Crossref] [PubMed]
19. Ferreira WT, Hong HA, Adams JRG, et al. Environmentally Acquired Bacillus and Their Role in C. difficile Colonization Resistance. Biomedicines 2022;10:930. [Crossref] [PubMed]
20. Okayasu I, Ohkusa T, Kajiura K, et al. Promotion of colorectal neoplasia in experimental murine ulcerative colitis. Gut 1996;39:87-92. [Crossref] [PubMed]
21. Nyskohus LS, Watson AJ, Margison GP, et al. Repair and removal of azoxymethane-induced O6-methylguanine in rat colon by O6-methylguanine DNA methyltransferase and apoptosis. Mutat Res Genet Toxicol Environ Mutagen 2013;758:80-6. [Crossref] [PubMed]
22. Delker DA, McKnight SJ 3rd, Rosenberg DW. The role of alcohol dehydrogenase in the metabolism of the colon carcinogen methylazoxymethanol. Toxicol Sci 1998;45:66-71. [Crossref] [PubMed]
23. Solomon L, Mansor S, Mallon P, et al. The dextran sulphate sodium (DSS) model of colitis: An overview. Comp Clin Path 2010;19:235-9. [Crossref]
24. Kobaek-Larsen M, Ritskes-Hoitinga M. Review of colorectal cancer and its metastases in rodent models: Comparative aspects with those in humans. Comparative Medicine 2000;50:16-26. [PubMed]
25. Murthy SN, Cooper HS, Shim H, et al. Treatment of dextran sulfate sodium-induced murine colitis by intracolonic cyclosporin. Dig Dis Sci 1993;38:1722-34. [Crossref] [PubMed]
26. Yim SK, Kim SW, Lee ST. Efficient Stool Collection Methods for Evaluating the Diarrhea Score in Mouse Diarrhea Models. In Vivo 2021;35:2115-25. [Crossref] [PubMed]
27. Koelink PJ, Wildenberg ME, Stitt LW, et al. Development of Reliable, Valid and Responsive Scoring Systems for Endoscopy and Histology in Animal Models for Inflammatory Bowel Disease. J Crohns Colitis 2018;12:794-803. [Crossref] [PubMed]
28. D'Amico F, Nancey S, Danese S, et al. A Practical Guide for Faecal Calprotectin Measurement: Myths and Realities. J Crohns Colitis 2021;15:152-61. [Crossref] [PubMed]
29. Santos EW. Hematological and biochemical reference values for C57BL/6, Swiss Webster and BALB/c mice. Braz J Vet Res Anim Sci 2016;53:138. [Crossref]
30. Navarro KL, Huss M, Smith JC, et al. Mouse Anesthesia: The Art and Science. ILAR J 2021;62:238-73. [Crossref] [PubMed]
31. Tanimura Y, Fukui T, Horitani S, et al. Long-term model of colitis-associated colorectal cancer suggests tumor spread mechanism and nature of cancer stem cells. Oncol Lett 2021;21:7. [PubMed]
32. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61:759-67. [Crossref] [PubMed]
33. O'Connor PM, Lapointe TK, Beck PL, et al. Mechanisms by which inflammation may increase intestinal cancer risk in inflammatory bowel disease. Inflamm Bowel Dis 2010;16:1411-20. [Crossref] [PubMed]
34. Karin M. NF-kappaB as a critical link between inflammation and cancer. Cold Spring Harb Perspect Biol 2009;1:a000141. [Crossref] [PubMed]
35. Viennois E, Chen F, Merlin D. NF-κB pathway in colitis-associated cancers. Transl Gastrointest Cancer 2013;2:21-9. [PubMed]
36. Karin M, Lin A. NF-kappaB at the crossroads of life and death. Nat Immunol 2002;3:221-7. [Crossref] [PubMed]
37. Karin M, Greten FR. NF-kappaB: linking inflammation and immunity to cancer development and progression. Nat Rev Immunol 2005;5:749-59. [Crossref] [PubMed]
38. Willenbucher RF, Aust DE, Chang CG, et al. Genomic instability is an early event during the progression pathway of ulcerative-colitis-related neoplasia. Am J Pathol 1999;154:1825-30. [Crossref] [PubMed]
39. Laurent C, Svrcek M, Flejou JF, et al. Immunohistochemical expression of CDX2, β-catenin, and TP53 in inflammatory bowel disease-associated colorectal cancer. Inflamm Bowel Dis 2011;17:232-40. [Crossref] [PubMed]
40. Nardone OM, Zammarchi I, Santacroce G, et al. Inflammation-Driven Colorectal Cancer Associated with Colitis: From Pathogenesis to Changing Therapy. Cancers (Basel) 2023;15:2389. [Crossref] [PubMed]
41. Seishima R, Okabayashi K, Ikeuchi H, et al. Effect of Biologics on the Risk of Advanced-Stage Inflammatory Bowel Disease-Associated Intestinal Cancer: A Nationwide Study. Am J Gastroenterol 2023;118:1248-55. [Crossref] [PubMed]
42. Ahmed W, Galati J, Kumar A, et al. Dual Biologic or Small Molecule Therapy for Treatment of Inflammatory Bowel Disease: A Systematic Review and Meta-analysis. Clin Gastroenterol Hepatol 2022;20:e361-79. [Crossref] [PubMed]
43. Gaddis DE, Maynard CL, Weaver CT, et al. Role of TLR2-dependent IL-10 production in the inhibition of the initial IFN-γ T cell response to Porphyromonas gingivalis. J Leukoc Biol 2013;93:21-31. [Crossref] [PubMed]
44. Netea MG, Van der Meer JW, Kullberg BJ. Toll-like receptors as an escape mechanism from the host defense. Trends Microbiol 2004;12:484-8. [Crossref] [PubMed]
45. Candelli M, Franza L, Pignataro G, et al. Interaction between Lipopolysaccharide and Gut Microbiota in Inflammatory Bowel Diseases. Int J Mol Sci 2021;22:6242. [Crossref] [PubMed]
46. Wang Y, Tian J, Shi F, et al. Protective effect of surfactin on copper sulfate-induced inflammation, oxidative stress, and hepatic injury in zebrafish. Microbiol Immunol 2021;65:410-21. [Crossref] [PubMed]
47. Gan P, Jin D, Zhao X, et al. Bacillus-produced surfactin attenuates chronic inflammation in atherosclerotic lesions of ApoE(-/-) mice. Int Immunopharmacol 2016;35:226-34. [Crossref] [PubMed]
48. Subramaniam MD, Venkatesan D, Iyer M, et al. Biosurfactants and anti-inflammatory activity: A potential new approach towards COVID-19. Curr Opin Environ Sci Health 2020;17:72-81. [Crossref] [PubMed]
49. Gao Z, Zhao X, Yang T, et al. Immunomodulation therapy of diabetes by oral administration of a surfactin lipopeptide in NOD mice. Vaccine 2014;32:6812-9. [Crossref] [PubMed]
50. Zhang Y, Liu C, Dong B, et al. Anti-inflammatory activity and mechanism of surfactin in lipopolysaccharide-activated macrophages. Inflammation 2015;38:756-64. [Crossref] [PubMed]
51. Jayasooriya N, Baillie S, Blackwell J, et al. Systematic review with meta-analysis: Time to diagnosis and the impact of delayed diagnosis on clinical outcomes in inflammatory bowel disease. Aliment Pharmacol Ther 2023;57:635-52. [Crossref] [PubMed]
52. Vasudevan A, Gibson PR, van Langenberg DR. Time to clinical response and remission for therapeutics in inflammatory bowel diseases: What should the clinician expect, what should patients be told? World J Gastroenterol 2017;23:6385-402. [Crossref] [PubMed]
53. The Jackson Laboratory - Body Weight Information for BALB/cJ (Accessed 01.04.2024). Available online: https://www.jax.org/jax-mice-and-services/strain-data-sheet-pages/body-weight-chart-000651#