Bacteriophages, or phages for short, may be too small to see without an electron microscope, but they have enormous therapeutic potential. Given the bacteria-killing capabilities of phages, the most obvious therapeutic opportunity is in the fight against bacterial infections. Indeed, this opportunity was recognized more than 100 years ago.

In 1919, microbiologist Felix d’Hérelle, a co-discoverer of phages, participated in the administration of what was probably the first phage therapy.1 It was received by a 12-year-old boy who had severe dysentery and was near death in the Hôpital des Enfants-Malades in Paris. A single dose cleared the boy’s symptoms.

After this early success, d’Hérelle continued to explore phage therapy’s potential. For example, he participated in a clinical trial against cholera and reported favorable results. But phage therapy soon faded in importance. Its antibacterial potential was eclipsed by antibiotics, which emerged in the 1940s.

Today, phage therapy is being prepared for a comeback—and not just for antibiotic-resistant bacterial infections, but also for a host of other diseases. Emerging possibilities for phage therapy were highlighted at the most recent Bacteriophage Therapy Summit. At this event, experts described how phages form key phageomes in the body. They also discussed how to enhance therapeutic bactericidal properties using CRISPR engineering. Another approach employs a non-GMO process that naturally evolves phages to create off-the-shelf products targeting a range of bacterial infections.

Additionally, rather than killing bacteria, nonlytic phages can be engineered to ferry therapeutic payloads into bacteria as treatments for cancer as well as inflammatory and metabolic diseases. However, considerable bottlenecks in both the design and manufacturing remain. To overcome those challenges, new artificial intelligence (AI)-mediated methodologies are helping to move phage therapy from an obscure process into a promising tool.

Phageome interactions

About 1 × 1031 viruses inhabit the earth. And as a recent review has noted, most of the viruses that account for this mind-boggling number are phages.2 This review also points out that phages form phageomes, much like bacteria form human microbiomes. Phageomes and microbiomes both colonize particular environments, and it appears that phageomes, like microbiomes, play roles in areas such as the oral cavity, lung, skin, and gut.

Dylan Lawrence
Dylan Lawrence, PhD
Senior Scientist, Pfizer

“In normal, healthy adults, the bacteriophage population is relatively stable, containing phages that are not actively lysing their hosts and are generally just replicating along with the bacteria,” says Dylan Lawrence, PhD, one of the authors of the review. In 2019, when the review was published, Lawrence was earning his doctorate at Washington University at St. Louis. Today, he is a senior scientist at Pfizer. “Recent research,” he notes, “has primarily focused on tailed-phages and Microviridae. Phages can be lytic or exhibit other life cycles such as lysogenic (nonreplicating), pseudolysogenic (plasmid like), or chronic. Healthy gut phages are largely nonlytic.”

It has been established that phages can play different roles depending on their lifecycle patterns. But there is still much to be learned about phages. “The largest current challenge with studying phages in the gut is associating them with their potential bacterial hosts,” Lawrence suggests. “Many current methods for microbiota studies are unable to definitively link phages to bacterial hosts. While phages may have some direct interactions with human hosts, we believe their primary effect on health stems from their interactions with bacteria.”

Lawrence notes that Pfizer’s interest in phages extends beyond basic research. “There’s still much to be learned about phages as applied to diseases,” he says, “and I remain hopeful for future breakthroughs.” He is also optimistic that phage therapy will become increasingly acceptable to regulators.

Optimized phage cocktails

The scale of the threat to public health posed by antibiotic resistance was indicated by a 2022 article in The Lancet.3 “[This article noted] that more than 1.2 million global deaths in 2019 were directly attributed to antibiotic-resistant bacterial infections,” relates Paul Kim, PhD, chief development officer, Locus Biosciences. “Furthermore, these types of infections were also predicted to be associated with almost 5 million deaths in total.”

Locus is developing phage-based products for the precise removal of pathogenic bacterial infections as well as targeting constituents of the microbiome that contribute to common diseases such as inflammatory bowel disease and cancer. The company does this by creating optimized engineered phage cocktails that target specific bacteria.

Locus Biosciences lab
Credit: Locus Biosciences

“We employ global sourcing of bacteria and environmental samples, which are a rich source of diverse phages, as the starting point for our phage discovery efforts,” Kim explains. “To optimize phage cocktails, company scientists employ long- and short-read DNA and RNA sequencing, bioinformatics tools, cloud computing, high-throughput liquid-handling robotics with machine vision, automated biobanking, and machine learning (including natural language processing and regression models trained on large-scale genotypic and phenotypic datasets).”

According to Kim, Locus uses advanced synthetic biology techniques to identify optimal insertion site-promoter configurations, and then engineers genetic payloads, such as CRISPR-Cas3 or bacteriocins (antimicrobial peptides), into each phage genome to enhance its bactericidal properties relative to the wild-type phage. “This approach,” Kim summarizes, “uses the target bacteria as a pharmaceutical factory of sorts to produce antibacterials in situ at the site of infection.”

The company is conducting multisite, randomized, controlled clinical trials for its lead program, LBP-EC01, a genetically enhanced phage cocktail used for the treatment of urinary tract infections caused by antimicrobial resistant strains of Escherichia coli. “We also have preclinical programs that are advancing rapidly toward the clinic targeting Klebsiella pneumoniae and Pseudomonas aeruginosa to treat antibiotic-resistant infections,” Kim points out. “In addition, we are working on an asset in preclinical development that targets bacteria suspected to be a key driver of inflammation in Crohn’s disease.”

Naturally evolved superphages

According to Amanda Burkardt, CEO, Phiogen, there are, in general, two approaches to phage therapy. The first involves naturally occurring phages, and the second involves synthetically engineered phages. Burkardt says that although the first approach has been shown to be effective in personalized treatments, it has lacked scalability. She adds that although the second approach promises scalability (by using phages with broad host ranges instead of painstakingly assembled phage cocktails), it faces regulatory hurdles due to the GMO status of engineered phages.

Amanda Burkardt
Amanda Burkardt
CEO, Phiogen

Phiogen is working on phage therapies that are not only effective, but that also combine the regulatory acceptability of the first approach and the scalability of the second. The company is using a platform that is designed to enable the discovery and natural evolution of phages that offer enhanced therapeutic capabilities such as expanded host ranges and improved anti-biofilm activities. Such phages are better able to overcome bacterial resistance.

Burkardt asserts that Phiogen is developing high-coverage, off-the-shelf products that can treat a wide range of bacterial infections for broad population sizes: “By optimizing the stability and efficacy of our phages, we are developing biotherapeutic products that can effectively clear infections and prevent reinfection.”

The company’s platform executes a six-step procedure: discovery, genetic profiling, bioclimatic screening, directed evolution, precision phage formulation, and preclinical testing. Discovery involves the mass capture and initial screening of phages. Genetic profiling involves the analysis of each phage’s host range, lytic activity, and genetic makeup to ensure, in Burkardt’s words, the “selection of the most safe and promising candidates.” Bioclimatic screening involves lead optimization via the simulation of real-world infections in a “patient on a plate” to ensure that phages work in the “microenvironment where these infections naturally occur.” Directed evolution involves Phiogen’s patented Directed Evolution Device, which allows the company to “train” phages for the desired antibacterial properties. Formulation involves the design, production, and testing of different phage combinations to ensure that in each one, the phages synergize together as a single drug product. Finally, preclinical testing involves the validation of end products in humanized animal models.

“This is a rigorous test cascade,” Burkardt declares. “It allows us to de-risk our lead candidates, reducing the need for unnecessary trials and failed translations outcomes.”

The company’s lead candidate, which targets invasive E. coli, is in the final stages of preclinical testing and is expected to start clinical trials next year. This candidate includes phages that have been used in over 11 FDA-approved emergency use cases “with remarkable outcomes,” Burhardt discloses. “For our E. coli product, we have seen nearly 100% bacterial clearance and symptom resolution, with some patients remaining infection-free for over two years.”

For the future, Burkardt envisions a world where clinicians reach for phage products before they reach for antibiotics. For now, she is focused on ensuring that phage therapy is “widely recognized and adopted as a mainstream medical treatment” that “provides hope where antibiotics fail.”

Phages as gut delivery vehicles

The gut microbiome must be suitably diverse and its constituents present in the right numbers and proportions if human health is to be preserved. Consequently, when the gut microbiome falls out of balance, therapeutics may restore it by suppressing some constituents and/or promoting others. These approaches may involve genetic manipulations that change how constituents of the microbiome function. For example, therapeutics are being developed that target dysfunctional gut bacteria using nonlytic phages as delivery vehicles.

“We engineer bacteriophages to deliver uniquely designed DNA, which we coin TBX Cargos, to target bacteria in our own microbiome,” says Adi Elkeles, CEO, Trobix Bio. “Each one of our TBX Cargos holds the needed information to enable the target bacteria to become a port for producing biotherapeutics, which in turn act to treat specific indications.”

Adi Elkeles
Adi Elkeles
CEO, Trobix Bio

Rather than killing bacteria, the company’s approach uses phages for a patient’s benefit. Elkeles elaborates, “Our TBX technology combines proprietary molecular engineering, synthetic biology, and computational algorithms to engineer phages, enabling precision delivery of unique DNA cargo to target microbiome E. coli strains. We engineer phage tails to develop a cocktail of phages that target the diversity of E. coli human microbiome strains and pack phage heads with uniquely designed TBX DNA cargo to control specific microbiome activities.

“Most important, each of our TBX DNA cargos contain a CRISPR-Cas3 ‘protection system’ that enables target bacteria that have our TBX DNA cargos to survive a lytic phage attack and propagate, while target bacteria that do not carry TBX DNA cargos die. This capability effectively drives in vivo selection of desired bacterial population for a robust and durable therapeutic effect.”

The company is advancing its orally administered biotherapeutics to treat cancer as well as inflammatory and metabolic diseases. “For example,” Elkeles points out, “TBX301 drives production of a well-known cytokine, interleukin-10, to locally prevent and treat colitis that is induced as a severe side effect of many immune checkpoint treatments, without the systemic immunosuppression side effects of current colitis treatments.”

Although the company’s candidates are in preclinical stages, Elkeles is looking optimistically to the future. “I believe that bacteriophages represent a huge potential for translational medicine that has not been realized to date,” he insists. “Our TBX platform technology holds immense potential to develop products that will offer local production of biotherapeutics for local treatment.”

TBX platform graphic
Trobix Bio develops precision microbiome oncology therapeutics using its TBX platform. TBX enables the engineering of phage tails, for targeting specific bacteria, and phage heads, for delivering DNA cargo. In bacteria, the DNA cargo can control microbiome activities and protect against lytic phage attack. The company’s lead products are designed to treat the life-threatening gastrointestinal side effects of cancer drugs.

Phage therapies and AI

Johan Wikstrom, the CEO of Tolka AI Therapeutics, cautions that while there have been many successful applications of phage therapy, they have typically been personalized by skilled scientists. “This approach,” he says, “won’t scale.” Wikstrom believes that the challenge of scaling phage therapy can be overcome with AI and automation technology. “We need automation to replace human labor and software to replace human decision-making,” he maintains. “Using AI, software that can learn, we can develop an antibacterial that improves the more we use it.”

Tolka AI Therapeutics is focusing on both the design and manufacturing of phages for therapy. “Given the recent explosion in genomic information, we see more use cases for AI in phage therapy,” Wikstrom elaborates. “For example, you could use AI to predict the likely host range of a given phage (how many bacterial strains it effectively infects) using only the sequenced genome. Combining AI and phage therapy is an emerging concept, and we will see many applications in the coming years.

“Several biotechs have struggled to scale the traditional hand-crafted phage therapy to modern manufacturing standards. They have almost exclusively tried to fit phage therapy into the one-size-fits-all mold of antibiotics. We are building an automated platform to isolate, purify, sequence, characterize, and manufacture phages that match a specific patient’s infection. Each patient will receive a unique ‘cocktail’ of phages, alleviating future resistance concerns. If the bacteria grow resistant to one phage cocktail, we will update it on the fly.”

Thus, Wikstrom believes employing AI in these processes will be key to future successes: “We focus on scaling our technology by replacing human labor with automation and human decisions with AI.” The company is currently focusing on preclinical studies in which its phage therapies target Mycobacterium abscessus, a complex and difficult-to-treat pathogen.

A promising future

Near its start, this article cited a review2 that discussed the role of phages in human health. So, it seems appropriate to end this article by quoting what the review concluded about phage therapy’s future: “[Phage therapy] is only just beginning to be realized. … It is clear by the surge in publications and popular press involving phages that we are in the midst of a phage renaissance, and much ongoing work may have dramatic implications for human health and disease. Many areas of human health have been neglected when it comes to phage research, but this is being gradually remedied, and we are entering an unprecedented age of discovery.”

 

References
1. Aswani VH, Shukla SK. An Early History of Phage Therapy in the United States: Is It Time to Reconsider? Clin. Med. Res. 2021; 19: 82–89. DOI: 10.3121/cmr.2021.1605.
2. Lawrence D, Baldridge MT, Handley SA. Phages and Human Health: More than Idle Hitchhikers. Viruses 2019; 11: 587. DOI: 10.3390/v11070587.
3. Antimicrobial Resistance Collaborators. Global Burden of Antimicrobial Resistance in 2019: A Systematic Analysis. Lancet 2022; 399: 629–655. DOI: 10.1016/S0140-6736(21)02724-0.

 

 

Previous articleBetter Biomarkers for Better Medicines
Next articleValidating CRISPR Preclinical Models with Multiomics