Not quite a cure yet: unlocking the unfulfilled promise of live biotherapeutics for disease treatment
Andrea Verdugo-Meza, Han M. Chiang, Emeran A. Mayer, Deanna L. Gibson
TL;DR
Live microbial treatments have shown mixed results in treating diseases, with inflammation and delivery challenges limiting their effectiveness.
Contribution
The paper identifies inflammation and delivery issues as key barriers to live microbial therapies and emphasizes the need for interdisciplinary collaboration.
Findings
Intestinal inflammation is negatively correlated with clinical responsiveness to probiotics.
Escherichia coli Nissle 1917 is effective in maintaining remission but not in treating active ulcerative colitis.
Shelf life and delivery strategies significantly impact microbial viability in the gut.
Abstract
Few live microbial therapeutics have demonstrated conclusive or reproducible clinical efficacy. Cochrane meta-analyses of probiotic interventions across multiple clinical trials, analyzed using a rank-biserial correlation test, revealed that intestinal inflammation is negatively correlated with clinical responsiveness. This is exemplified by Escherichia coli Nissle 1917, which shows efficacy in maintaining remission comparable to frontline therapies, yet fails to demonstrate clear benefit in the treatment of active ulcerative colitis, an inflammatory bowel disease defined by chronic intestinal inflammation. Beyond inflammation as a key barrier to efficacy, inadequate shelf life and delivery strategies further compromise microbial viability and functional persistence in the gastrointestinal tract. Here, we highlight both the challenges and emerging opportunities in the field of live…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2| Study | Intention of intervention | Sample size (n/N) | Daily dose | Pre-treatment or additive therapy | Treatment comparison | Outcome | Primary/Secondary endpoint | Result of the intervention | Weaknesses |
|---|---|---|---|---|---|---|---|---|---|
|
| To induce remission | 70/90; 24 for 40 mL dose; 23 for 20 mL dose; 23 for 10 mL dose | 40, 20 or 10 mL x 10^8 (enama) | NA | EcN vs. Placebo | −/+ | DAI ≤2/time to remission, DAI = 0, histological healing | No differences in EcN vs. Placebo (ITT p = 0.4430); relapse rate in placebo 35% compared to 43.5%, 47.8%, and 36.4% remission in EcN 40 mL, 20 mL and 10 mL, respectively. Suggested dose-dependent efficacy in inducing remission (PP, p = 0.0443) | EcN dosage-dependency identified in PP was not observed in the ITT |
|
| To induce remission | 50/100; 25 in Ciprofloxacin/EcN; 25 in Placebo/EcN | 5–50 × 10^9 (capsule) | 5-ASA (≥64%) and/or thiopurines (≤32%) | Placebo/Placebo, Ciprof/Placebo, Placebo/EcN, and Ciprof/EcN (1 week/7 weeks) | − | CAI ≤4 (CRP and hemoglobin levels)/number of patients withdrawn from the study | No benefit in EcN use as an additional therapy to conventional treatment; a greater % of patients in the placebo group reached remission (P < 0.02) with remission rates of 80% for placebo/placebo, 72% for Cipro/placebo, 60% Cipro/EcN, and 41% placebo/EcN | Standard therapies were allowed to be modified by physicians throughout the study |
| Mirsepsai-Lauridsen HC. et al. Sci Rep. 2016 | To induce remission | 50/100, 25 in Ciprofloxacin/EcN; 25 in Placebo/EcN | 5–50 × 10^9 (capsule) | Conventional therapies, except systemic steroids or TNF-α inhibitors | Placebo/Placebo, Ciprof/Placebo, Placebo/EcN, and Ciprof/EcN | − | Fecal calprotectin <200 mg/kg/colonization with | Reduction of fecal calprotectin was not superior than results in the placebo group (P < 0.05); remission rates of 61% for placebo/placebo, 57% for Cipro/placebo, 41% Cipro/EcN, and 18% placebo/EcN | |
| Park SK. et al. Korean J Gastroenterol. 2022 | To induce remission | 58/133 | 5 × 10^9 (capsule) | 5-ASA | EcN and 5-ASA vs. placebo and 5-ASA | −/+ | IBDQ >16/Mayo score decrease of 3 points, for clinical remission and 1 point for endoscopic | No differences in patients reaching primary endpoint (ITT, p = 0.86); 47% for placebo and 44.8% for EcN; secondary endpoints, including clinical remission and response rates, were better for EcN group | Did not specify 5-ASA dosage (merely mentioned that the dosage was stable throughout the trial) |
|
| To induce and to maintain remission | 57/116 | 5 × 10^10 (capsule) | 1 week of gentamicin and 2 weeks of hydrocortisone acetate enema |
| + | Maintenance: Time to relapse and rate of relapse/Induction: rate and time to remission | Efficacy in maintaining remission after exacerbation of UC comparable to 5-ASA; relapse rate 73% in 5-ASA and 67% in EcN (significant equivalence p = 0.0059); remission induced in 75% in 5-ASA and 68% in EcN (significant equivalence p = 0.0508) | The statistics performed barely met the equivalence margin cut-off set by the authors (but was still able to statistically show that EcN was equivalent to 5-ASA) |
|
| To maintain remission | 50/120 | 5 × 10^10 (capsule) | NA | EcN vs. 5-ASA | + | Compare clinical remission (≤4 CAI)/compare relapse rates, relapse-free times and global assessment | Efficacy in maintaining remission and preventing relapse was comparable to 5-ASA (ITT p = 0.12); relapse rate 11.3% in 5-ASA and 16% in EcN | |
|
| To maintain remission | 162/327 | 5–50 × 10^9 (capsule) | NA | EcN vs. 5-ASA | +/− | Relapse prevention; relapse identified when CAI>6 or +3 points in CAI or endoscopic index >4/general wellbeing, and calculation of a quality of life index | Efficacy and safety in maintaining remission comparable to 5-ASA; relapse rate 37% in 5-ASA and 45.1% in EcN (ITT significant equivalence, p = 0.013); no significant differences observed in secondary endpoints | |
|
| To maintain remission | 24/34 | At least 5 × 10^10 (capsule) | NA | EcN vs. 5-ASA | + | Relapse prevention/quality of life index | Efficacy in maintaining remission comparable to 5-ASA; relapse rate of 30% in 5-ASA and 25% in EcN, no statistical analysis reported | No statistics were reported in the study |
| Oh GM. et al. Korean J Gastroenterol. 2021 | To maintain remission | 94 | 5 × 10^9 (capsule) | 90.4% on 5-ASA ±30.9% in thiopurine | None | −/+ | Changes in fecal calprotectin/Mayo score, body weight, BMI, and hemoglobin | IBDQ scores no different between EcN and placebo groups (PP p = 1, ITT p = 0.86); effective for symptom reduction when used with other therapies (P = 0.031), but not for fecal calprotectin (P = 0.653) | Retrospective study |
| Bodini G. et al. J Gastrointestin. Liver Dis. 2023 | To maintain remission | 49/94 | First month 2 capsules; second month 1 capsule | 31% on biologics, 8% on immuno-suppressants and 4% on steroids | None | −/+ | EcN’s effect on Fecal calprotectin values in clinical remission | Effective in maintaining remission after 2 months of concomitant treatment (P < 0.005); >60% patients relapsed post 6 months cessation | No prospective control group |
|
| EcN is equivalent to 5-ASA in remission maintenance, but EcN is not superior to placebo in inducing remission |
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsInflammatory Bowel Disease · Gut microbiota and health · Probiotics and Fermented Foods
Overview
Probiotics and live biotherapeutic products (LBPs) differ fundamentally in their definitions, intended purposes, and regulatory classifications. While probiotics are dietary supplements or food additives aimed at promoting general or gut health in otherwise healthy populations, LBPs are developed and regulated as pharmaceutical drugs for the prevention or treatment of specific diseases under the authority of the U.S. Food and Drug Administration (FDA) (Heavey et al., 2022). Consequently, LBPs are restricted to microbial strains with demonstrated therapeutic efficacy in defined disease contexts, criteria that most conventional probiotics do not meet. Despite the FDA’s efforts in 2010 to delineate LBPs through a more rigorous regulatory pathway, many over-the-counter probiotics continue to make health claims targeting diseased populations, and several have been evaluated in clinical trials. There are only two approved LBPs: both are for Clostridia difficile infection (Bland et al., 2025 #734). Here we present our perspective after reviewing clinical trial and meta-analysis data for both probiotics and LBPs. This led to our conclusion that microbial therapeutics have shown limited efficacy in treating chronic intestinal inflammation. Given the same beneficial bacteria could be either a probiotic or a LBP, depending on the regulatory path taken, we extend our findings of probiotics tested in clinical trials to postulate why microbial therapeutics have yet to fulfill their promise and outline strategies for advancing the field to realize their full clinical potential.
Intestinal inflammation undermines the promise of live biotherapeutic products
Severe ambiguity surrounding the effectiveness of probiotics in clinical trials led us to question whether intestinal inflammation itself might act as a barrier to success in disease populations. However, before addressing this biological constraint, we must first acknowledge the challenges arising from the absence of a stringent regulatory framework for probiotics. These challenges arose because industry players approached regulatory approvals as dietary supplements instead of rigorous trials like those for drugs, given the cost, effort, and required expertise. The consequence of this was evident when we reviewed over 1,700 clinical studies on probiotics testing Lactobacillus spp. and Bifidobacterium spp. Revealing notable gaps in the reported data and a lack of rigor in trial design. Despite retrieving 581 clinical trials for Bifidobacteria (clinicaltrial.gov), only 31% specified essential information such as the clinical phase under investigation (Figure 1A). The lack of strain-level identification was striking, and tracing studies through all stages, from phase 1 to phase 4, proved nearly impossible. For example, of the 35 phase 3 studies identified, only 12 studies specified the strain of bacteria being used (drug class), and although 23 studies had a “completed” status, their results were not available, revealing a well-known publication bias for positive findings.
The ineffectiveness of LBPs is reflected in ambiguous clinical trials and meta-analyses. (A) Randomized controlled trials registered in clinicaltrials.gov for Bifidobacterium spp., showing the inconsistent trackability through clinical phases 1 to 4, and the lack of detailed reporting on the Bifidobacterium type strains used and the status of the studies in each phase. (B) Pie charts showing the conclusions of 63 meta-analyses retrieved from Cochrane Library, on the effect of probiotics (no strain distinction) on disease treatment and/or prevention, highlighting that the few studies with positive effects are associated with a lack of underlying intestinal inflammation. (C) The presence of intestinal inflammation in the target health condition is negatively correlated with the percentage of meta-analyses concluding probiotics are effective. Due to the non-normal distribution of the data, a Mann-Whitney U Test was used to test for significant differences between the 2 groups, and the effect size is determined using the Rank biserial correlation test. (D) Pie chart showing the efficacy of probiotics in IBD meta-analyses. Of the 63 meta-analyses identified, 10 meta-analyses specifically investigated IBD. No life-threatening adverse effects were reported. Authors AV and HC searched ClinicalTrials.gov for clinical trials using keywords like “Bifidobacteria.” They reviewed details on LBP strains, conditions studied, and trial statuses. Similarly, the authors AV and HC systematically retrieved all Cochrane reviews related to the keyword “live biotherapeutic product” (0 results) or “probiotic” from the Cochrane Library (https://www.cochranelibrary.com/). The researchers verified the quality of the results and extracted the authors’ conclusions from each meta-analysis. Figure 1 was created with GraphPad Prism and R.
Given these data gaps, we turned to meta-analyses for clarity, anticipating that they would meet the benchmark of methodological rigor. Of the 63 meta-analyses found in the Cochrane Library (Supplementary Table S1), only 22 provided conclusive results, while 41 highlighted the heterogeneity of the findings, which statistically hindered the authors from reaching conclusions, leading to our classification as ineffective outcomes. It is important to note that 4% of analyses did not recommend the use of probiotics for the target disease, with one meta-analysis pinpointing harmful side effects (Supplementary Table S1), supporting that probiotics are not always safe in every patient population. Most strikingly, among the 30% of meta-analyses that reported LBPs to be effective, the overwhelming majority (89.5%) were conducted in populations without underlying intestinal inflammation (Figure 1B). These findings suggest that while LBPs may exhibit limited efficacy in treating active disease, more favorable effects in contexts where inflammation is absent or minimal could be apparent. This pattern was evident in studies on bacterial vaginosis, where 60% of meta-analyses demonstrated a positive clinical effect, as well as in disease-prevention trials involving otherwise healthy participants, where 68% of meta-analyses reported efficacy (Figure 1B). A rank-biserial correlation test supported the observation that the presence of intestinal inflammation negatively correlated with clinical outcomes, whereas the lack of intestinal inflammation correlated with positive clinical outcomes (Figure 1C). A single meta-analysis (including 7 studies) concluded that probiotics were clinically effective for inflammatory bowel disease (IBD) and may slightly improve the induction of remission. However, the authors reported very low-quality evidence of included studies and remarked on insufficient study information, failure to define remission, lack of blinding of participants, incomplete and selective reporting, poor quality of data and bias across the studies, in addition to no specific probiotic across studies (Kaur et al., 2020). Despite one positive analysis, approximately 90% of Cochrane meta-analyses conclude that probiotics, including those with LBP claims, lack clinical efficacy in treating IBD, a disease with chronic intestinal inflammation (Figure 1D). Given the limited efficacy of probiotics and LBPs in inflammatory disease contexts, intestinal inflammation appears to be a key biological barrier that undermines their therapeutic potential.
To explore further the observation that inflammation is a barrier to probiotic efficacy, we focused on the clinical evidence for Escherichia coli Nissle 1917 (EcN) treating ulcerative colitis (UC), a form of IBD. EcN is an extensively studied probiotic but has been tested in several clinical trials like it were an LBP. Of the 10 clinical trials reported, 6 focused on maintenance of remission, with the others testing induction of remission in active UC patients often comparing or adjunct to the frontline therapy, 5-aminosalicylic acid (5-ASA; Table 1). A meta-analysis including 5 of these trials concluded EcN is ineffective at inducing remission in patients with active UC, and any therapeutic benefit appears restricted to remission maintenance (Losurdo et al., 2015) (Table 1). This means that the more inflamed the gut environment is, like in active UC, the less likely EcN can confer any clinical benefit. A plausible explanation is EcN’s poor ability to persist and thus colonize the inflamed intestine. Still, to be effective even in remission, EcN still requires continuous high-dose daily administration with relapse rates between 16-45% and >60% 6 months post-cessation (Table 1). This is supported by data revealing EcN detectability drops precipitously post-cessation (Joeres-Nguyen-Xuan et al., 2010), observed in adults (Tannock et al., 2011) as well as infants (Cukrowska et al., 2002).
These data highlight a systemic challenge in microbial therapy, the inability to overcome colonization resistance and achieve durable engraftment, particularly under inflammatory conditions. This limitation explains why the American Gastroenterology Association (AGA) and the European Crohn’s and Colitis Organization (ECCO) do not endorse any probiotic, including EcN, as a primary therapy for IBD (Su et al., 2020; Harbord et al., 2017). EcN exemplifies this broader issue affecting microbiome-based therapeutics, including transient colonization, high dosing requirements, and limited efficacy during active disease. Still, EcN has been recognized as a potential complementary medicine to maintain remission in UC patients by ECCO but not AGA, given its general safety profile and potential to add benefit (Su et al., 2020; Torres et al., 2019).
The translational roadblocks to live biotherapeutic product success
Colonization is a major challenge that jeopardizes the success of any microbial therapy, probiotics or LBPs. Indeed, several research groups have demonstrated that probiotics show poor engraftment into the host’s microbial ecosystem (Zmora et al., 2018). While ecological theories of colonization resistance, where LBPs must compete with native gut microbes for successful niche establishment, are well known, other hurdles arise even earlier during the manufacturing and formulation stages. First, the microbe must remain viable throughout the production process and remain stable during storage. Ideally, LBPs should have a long shelf life at room temperature without loss of potency. This maintenance of long shelf life and stability at ambient temperatures can be achieved through means like microencapsulation. However, from our review of meta-analyses (Supplementary Table S1), we found that the vast majority of studies did not report the use of probiotic formulations, including microencapsulation, needed to extend shelf life. Second, the viable product must then withstand the harsh gastric acid exposure when taken by the subject, to reach the small or large intestine, the final destination to exert its beneficial effects. Enteric-coated capsule technologies designed to be acid-resistant can significantly improve the delivery of intact microbes to the intestine. Other novel solutions include the use of hydrogels that allow the formation of probiotic microcolonies, offering protection and favouring probiotic bacterium behaviour for increased colonization (Liu et al., 2025). Surprisingly, very few studies reported the use of enteric-coated capsules for the delivery of the microbial therapy (Supplementary Table S1).
Another challenge is the need for mitigation strategies against adverse microbial adaptations, as microbes can rapidly evolve and mutate within their host. Reports have identified mutations linked to antibiotic-resistance genes in probiotics isolated from patients with bacteremia (Yelin et al., 2019). This has been observed in patients treated simultaneously with antibiotics and probiotics, where the latter supported the growth of vancomycin-resistant strains (Montassier et al., 2021). Furthermore, since LBPs can modulate the gut microbiome as part of their therapeutic mechanism, other vital considerations, such as inter-individual microbiome differences, must be accounted for while selecting the most suitable LBP for the patient. Such appreciation for inter-individual variabilities is increasingly relevant as the age of “personalized medicine”, or individualized treatment approaches, has taken center stage.
Finally, the loose regulatory framework used for probiotics has been permissive for many health claims not supported by high-quality evidence. Moreover, there is a lack of consistency in global regulatory frameworks, as evidenced by differences between the FDA in contrast to the Australian Therapeutic Goods Administration, which has tailored and updated guidelines for the quality of LBPs. The scope of microbiome-based therapies is moving towards the use of the LBPs framework instead of general probiotics, but this could be as messy and as unreliable as probiotics without a harmonized approach.
Turning microbes into functional medicine
Is bioengineering the future?
What is needed to turn microbes into functional medicines that are reliable and clinically meaningful for the plethora of clinical challenges we face? Certainly, LBPs may need to be tailored to the patient, considering their host-associated factors, including the immunological status and their individualized microbiomes. This specific approach may require bioengineering microbes to target a specific desired effect. For example, engineered LBPs like Lactococcus lactis secreting IL-10 or diverse strains of Lactobacillus spp. Secreting superoxide dismutase both showed promising results in pre-clinical models of colitis (del Carmen et al., 2014; Watterlot et al., 2010), but still failed to translate into clinically relevant interventions in clinical trials (Clinicaltrials, 2009). A trailblazing company, Synlogic, took their LBP SYNB 1934, an engineered variant of EcN for treating phenylketonuria, from pre-clinical studies to clinical phase 3 (Clinicaltrials, 2023). Despite the initial promise, the trial was terminated because the study was unlikely to meet the primary endpoint (Synlogic, 2024). Nevertheless, this trial lights the path towards the potential for rationalized bioengineering of LBPs.
A key takeaway from these translational failures is that the inflammatory environment in many chronic diseases may not be conductive for live microbial therapies to act effectively, resulting in diminished efficacy and inconsistent clinical outcomes. One approach to overcoming this challenge is to genetically modify the probiotic to be tolerant or have a survival advantage in the inflamed environment. Our group has begun to address this by bioengineering EcN to have a fitness advantage during inflammation. Although EcN has been shown to promote multiple intestinal benefits in mostly rodent or tissue culture studies (Figure 2A), clinical studies in IBD patients have indicated no measurable and consistent differences in effectiveness when compared to placebo or even first-line therapy during active disease (Table 1). Based on the observation that intestinal inflammation is correlated with the lack of clinical efficacy of probiotics (Figure 1), we hypothesized that the intestinal inflammation prevents enough engraftment during colitis, limiting EcN’s ability to elicit its beneficial effects (Gibson et al., 2018). To overcome EcN’s colonization limitations during inflammation, we introduced genetic tools to support its persistence and expansion during colitis (Figure 2B), resulting in a more consistent protection, as we have shown in pre-clinical models (Verdugo-Meza et al., 2024). Understanding how EcN::ttr translates into clinical outcomes will be key in the next-generation of bioengineered LBPs.
Engineering E. coli Nissle 1917 as an example of a rationalized LBP. (A) Summary of the characteristics or fitness factors and proposed mechanisms of action of the LBP E. coli Nissle 1917 (EcN; serotype O6:K5:H1); Phenotypically, E. coli Nissle 1917 (EcN) possesses an atypical O6 antigen with only one saccharide unit as a result of a point mutation in the O6 antigen polymerase wzy gene, rendering EcN serum-sensitive as the O antigen is critical for bacterial virulence (Grozdanov et al., 2002; Sonnenborn and Schulze, 2009). EcN is also capable of forming biofilms, which consists of F1C fimbriae, F1A fimbriae, cellulose, and secretory phospholipids (Sonnenborn and Schulze, 2009). EcN is notable for its ability to defend against a variety of pathogenic (or potentially pathogenic) organisms through means such as nutrient competition, outperformance/inhibition of biofilm formation, and (to a certain extent) secretion of microcins H47 and M (Sonnenborn and Schulze, 2009; Fang et al., 2018; Hafez et al., 2010; Ma et al., 2023). EcN K5 capsule and flagellin can also enhance toll-like receptor (TLR)-4 and −5 as well as human β-defensin 2 production (respectively), both of which can enhance pathogen defense and gut homeostasis (Hafez et al., 2010). EcN also possesses immunomodulating properties. Through unknown mechanisms, EcN can prevent tumor necrosis factor (TNF)-α induced interleukin (IL)-8 production in HCT15 cells, reduce the levels of IL-2 and IL-5, and increase the levels of immunoglobulin (Ig) A and M (Sonnenborn and Schulze, 2009). EcN lipopolysaccharide (LPS) has been shown to downregulate chemokine CCL24 expression (and thus, inflammation) in human peripheral blood mononuclear cells (Güttsches et al., 2012). EcN increases bloodstream naïve and memory CD4+ T cells (without influencing gut epithelial T cell populations) and regulates the activation/progression and apoptosis of γδ T cells, enabling an immune response adequately strong to target inflammation-triggering agents (e.g., bacterial antigens) while controlling the extent of gut inflammation (Sonnenborn and Schulze, 2009). EcN can also influence gut metabolites and epithelial barrier integrity. Serotonin, which is involved in gut motility and signaling, was previously shown to be elevated by EcN under non-inflammatory conditions through an unknown mechanism (Nzakizwanayo et al., 2015); later experiments showed that EcN outer membrane vesicles (OMVs) can modulate serotonin levels, particularly during inflammation (Olivo-Martínez et al., 2024). In mice with dextran sodium sulfate (DSS)-induced colitis, OMVs reduce the expression of cyclooxygenase 2 (COX2), an enzyme that synthesizes prostaglandin, and modulates cytokine levels (Olivo-Martínez et al., 2024). EcN OMVs have also been shown to increase the expression of tight junction (TJ) proteins zonula occludens (ZO)-1 and claudin-14 while reducing claudin-2 expression, all of which strengthen the gut barrier (Olivo-Martínez et al., 2024). EcN can also increase the expression of another TJ protein called ZO-2 (Olivo-Martínez et al., 2024), though the specific mechanism is unclear. (B) Summary of EcN::ttr mechanism of persistence; EcN::ttr is created by bio-engineering EcN with a Salmonella typhimurium-derived operon called the ttr operon, which consists of five genes: ttrS, ttrR, ttrB, ttrC, and ttrA (Price-Carter et al., 2001). The ttr operon-derived enzyme tetrathionate reductase enables S. typhimurium to anaerobically respire using tetrathionate, an inflammatory metabolite produced from thiosulfate and reactive oxygen species (that infiltrates into the gut during inflammation). Though it remains elusive how EcN can specifically detect tetrathionate, tetrathionate presence initiates a cascade of reactions within EcN::ttr, starting with the autophosphorylation of the membrane-bound protein TtrS followed by the activation of the TtrR protein through a phosphate translation event (Price-Carter et al., 2001). Activated TtrR induces the transcription and assembly of the tetrathionate reductase enzyme, which catalyzes the reduction of tetrathionate into thiosulfate (Price-Carter et al., 2001). Tetrathionate reduction drives anaerobic respiration and boosts ATP production (Price-Carter et al., 2001), allowing EcN::ttr to thrive within inflammatory conditions. Created in BioRender. Chiang, H. (2025) https://BioRender.com/rod4eh5.
Bioengineered LBPs may provide promises of the next-generation of microbial medicines, but several safety challenges must be addressed. For example, bioengineered LBP design should account for known risks like horizontal gene transfer, environmental biocontainment, and growth control in the patient’s ecosystem with the addition of kill switches (Rottinghaus et al., 2022). More studies are needed to know if these are real or just perceived risks to understand if we must address these challenges to support the safe use of engineered LBPs. Potential risks can be addressed by adhering to FAO/WHO guidelines for characterizing administered microbes, including assessing antibiotic resistance and metabolic activity (such as D-lactate production or bile salt deconjugation), testing for toxin production, and checking for hemolytic activity. It is important to evaluate genetic stability through whole genome sequencing, test for virulence factors, and ideally evaluate horizontal gene transfer capacity, including phage sensitivity. By employing this approach, the negative ecological impact on the host’s microbiome and side effects such as systemic microbial translocation and abnormal immune responses in vulnerable patient populations may be reduced.
Strategies to enhance the efficacy of live biotherapeutic products
Even if bioengineered or otherwise optimized LBPs ultimately demonstrate clinical efficacy, significant manufacturing challenges remain. Chief among these is ensuring that LBPs reach the intestine intact and viable. While simple enteric-coated capsules can protect against gastric acid degradation, microencapsulation should become standard practice in the field to maintain viability prior to the patient consuming the product. This technique has long been shown to preserve the stability of lyophilized microbes during storage and to enhance overall product viability. Microencapsulation, often employing alginate-based matrices supplemented with cryoprotectants, provides an effective means of stabilizing bacterial formulations and extending shelf life (Razavi et al., 2021). Moreover, multi-layered encapsulation systems can offer additional protection while enabling controlled release of viable microbes throughout the small and large intestine. Beyond these established approaches, next-generation encapsulation strategies, such as engineered polymer “cages,” are being developed to achieve spatially targeted and temporally controlled microbial release along the gastrointestinal tract (Finbloom et al., 2023).
Finally, there is an imperative need for the regulatory systems built exclusively for LBPs, including bioengineered LBPs. This regulatory system should standardize how we evaluate the health claims for LBPs. An example of this could be guidelines for clinical trials that account for the survival, engraftment and clearance of the evaluated LBP. In addition, given that diet exerts a major influence on microbial composition, function and metabolite production, reliable assessment of patients’ dietary status could become routine in clinical trials testing LBPs in patient populations. If industry moves towards employing already well-researched methods of microencapsulation of LBPs, then we need to assume that the dose of the drug is the dose delivered into the GI tract. This may mean dose escalation challenges will be important to consider, since a therapeutic dose may in fact be far less than typically used with most current products. This could be critical for susceptible populations like pre-term infants with necrotizing enterocolitis, shown to be susceptible to sepsis with probiotic treatment (Chiang et al., 2021). Lastly, the regulatory system should be adaptable as we keep gaining knowledge about host-microbiome interactions in health and disease.
Translating promise into practice
Over the past 5 years, investments in microbiome-based therapeutics have soared, reflecting immense optimism in their potential. Yet, stakeholders have been largely disappointed by underwhelming clinical outcomes. The initial wave of investor enthusiasm is now confronting a reality check, as efficacy results remain inconsistent and frequently disappointing (Zamecnik, 2023). Notable examples include consortia LBPs for IBD treatment, such as SER-287 (Seres Therapeutics) and VE202 (Vedanta Biosciences). Both advanced to Phase 2 clinical trials but were ultimately terminated after failing to meet primary endpoints. Despite these setbacks, the field continues to evolve rapidly driven by advances in engineered LBPs, precision microbiome approaches integrated with predictive modelling (Markets Ra, 2025; Eisenstein, 2020). Looking ahead, meaningful progress will depend on developing microbial strains or consortia capable of durable colonization and mechanistic alignment with disease pathophysiology.
By addressing key limitations in clinical evidence, quality control, and patient stratification, the field may finally achieve targeted and reproducible therapeutic outcomes. This may include rational genetic modifications to enhance LBP function in specific disease contexts and integrating advanced formulation technologies to ensure both targeted delivery and shelf-life stability. For many chronic inflammatory diseases, where treatment options remain limited to broad immunosuppression, LBPs offer a transformative paradigm, one that leverages their immunomodulatory properties and host-microbe synergy to restore health. Implementing rational yet practical strategies to enhance clinical effectiveness may ultimately allow LBPs to fulfill their long-promised therapeutic potential.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Bland C. M.Love B. L.Jones B. M. (2025). Human microbiome: Impact of newly approved treatments on C. difficile infection. Am. J. Health Syst. Pharm. 82, 174–183. 39230353 10.1093/ajhp/zxae 249 · doi ↗ · pubmed ↗
- 2Chiang M. C.Chen C. L.Feng Y.Chen C. C.Lien R.Chiu C. H. (2021). Lactobacillus rhamnosus sepsis associated with probiotic therapy in an extremely preterm infant: pathogenesis and a review for clinicians. J. Microbiol. Immunol. Infect. 54, 575–580. 10.1016/j.jmii.2020.03.029 32307246 · doi ↗ · pubmed ↗
- 3Clinicaltrials (2009). A phase 2a randomized, placebo-controlled, double-blind, multi-center dose escalation study, to evaluate the safety, tolerability, pharmacodynamics and efficacy of AG 011, in subjects with moderately active ulcerative colitis. Available online at: https://clinicaltrials.gov/study/NCT 00729872.
- 4Clinicaltrials (2023). A phase 3, double-blind, placebo-controlled, randomized withdrawal study to evaluate the efficacy and safety of SYNB 1934 in patients with PKU (SYNPHENY-3).
- 5Cukrowska B.Lod Inová-Zádn IkováR.Enders C.Sonnenborn U.Schulze J.Tlaskalová-HogenováH. (2002). Specific proliferative and antibody responses of premature infants to intestinal colonization with nonpathogenic probiotic E. coli strain nissle 1917. Scand. J. Immunol. 55 (2), 204–209. 10.1046/j.1365-3083.2002.01005.x 11896937 · doi ↗ · pubmed ↗
- 6del Carmen S.Martín Rosique R.Fau - Saraiva T.Saraiva T.Fau - Zurita-Turk M.Zurita-Turk M. (2014). Protective effects of lactococci strains delivering either IL-10 protein or c DNA in a TNBS-induced chronic colitis model. J. Clin. Gastroenterol. 48, S 12–S 17. 10.1097/MCG.0000000000000235 25291117 · doi ↗ · pubmed ↗
- 7Eisenstein M. (2020). Early investments powering the ascent of microbiome therapeutics biopharma dealmakers. Nature. 10.1038/d 43747-020-01178-x · doi ↗
- 8Fang K.Jin X.Hong S. H. (2018). Probiotic Escherichia coli inhibits biofilm formation of pathogenic E. coli via extracellular activity of Deg P. Sci. Rep. 8 (1), 4939. 10.1038/s 41598-018-23180-1 29563542 PMC 5862908 · doi ↗ · pubmed ↗