This week, we’re going over a new phage therapy case report. This report marks the first time engineered phages have been used to treat a human patient, and the first time a Mycobacterium infection has been treated with phage therapy.

The study was published yesterday in Nature Medicine by Rebekah Dedrick, Carlos Guerrero-Bustamente and colleagues in Graham Hatfull’s phage lab at U Pitt, alongside Helen Spencer, an MD at Great Ormond Street Hospital (GOSH) in London, UK, her colleagues, and Chip Schooley of UCSD’s IPATH.

Quick snapshot

What was the patient dealing with?

A 15-year-old girl with cystic fibrosis in the UK, who’d had a double lung transplant, was diagnosed with disseminated Mycobacterium abscessus subspecies massiliense, a form of non-tuberculous Mycobacteria (NTM).

M. abscessus was cultured in her sputum, she had lesions on her liver and skin, and her skin continued to develop new infected nodules, even with antibiotics.

She was sent home 7 months post-transplant with a palliative care plan in place, and given a 1% chance of survival (most patients don’t recover when NTM grows post-transplant).

What was done?

A personalized phage therapy approach was used to treat the patient. A large phage library was screened against the patient’s isolate, and a three-phage cocktail (two of which were engineered to lyse the strain more efficiently) was given to the patient over at least 32 weeks.

What was the outcome?

The patient improved substantially, and went on to resume normal activities. Most of her Mycobacterium infection was eliminated. No adverse effects of the phages were observed.

The nitty gritty: how was this done?

Identifying the bacteria

The patient’s strain was isolated 1 month post-transplant and found to be Mycobacterium abscessus, subspecies massiliense. It was resistant to all antibiotics tested: clarithromycin, amikacin, tobramycin, ciprofloxacin, moxifloxacin, cefoxitin, cotrimoxazole, doxycycline and linezolid.

What is M. abscessus?

M. abscessus is a species of non-tuberculous Mycobacteria (NTM). NTM frequently colonize CF patients, and is commonly associated with poor prognosis due to antibiotic resistance.

Choosing and preparing the phages

The strain was sent to Graham Hatfull’s Mycobacterium phage genetics lab. A collection of >10,000 phages isolated by undergraduate students in the SEA-PHAGES program was exploited (representatives were screened against the clinical strain). They also screened >100 environmental isolates, and cultured the patient’s strain with a pool of unsequenced phages to enrich for active phages.

Three were found to have lytic activity on the patient’s strain:
Phage Muddy - killed efficiently (lytic)
Phage ZoeJ - killed inefficiently (temperate)
Phage BPs - killed inefficiently (temperate)

Phage characterization

All three phages were found to be siphoviridae, but all were from unrelated families and showed little genetic similarity.

Engineering to improve phage efficiency

The two temperate phages were each genetically engineered to lose their lytic repressor gene using the BRED technique, developed in the Hatfull lab by Laura Marinelli and colleagues in 2008.

BRED engineering successfully made one phage, ZoeJ, an efficient killer, but the other, BPs, was still producing turbid plaques. For this phage, host range mutants were isolated that could kill more efficiently (genetic analysis showed single base changes in a portal gene of the phage). The three phages were combined in a cocktail.

Testing the cocktail in vitro

The three-phage cocktail was assessed for phage resistance by culturing with the patient’s strain in vitro. Survivor colonies, which grew at a low frequency, were tested individually for phage resistance, and all retained sensitivity to at least one phage in the cocktail.

How were the phages administered?

First, a topical test-dose of the phage was administered to the sternal wound. After 24 h observation, and no ill-effects, the three-phage cocktail (10^9 PFU per phage) was administered intravenously every 12 h for at least 32 weeks. After noticing that the sternal wound, which had received a topical dose of phages, began to close first, phages were also administered topically to the other skin lesions.

What about endotoxin levels?

Endotoxin levels were not a concern in this case, as Mycobacteria do not produce lipopolysaccharide (LPS).

Were antibiotics administered concurrently?

Yes: the patient remained on the same antibiotics regimen throughout phage therapy (amikacin, imipenem/cilastatin, tigecycline, bedaquiline, clofazimine).

What was the outcome?

After 9 days, the patient was discharged, but continued taking IV phage therapy at home. Over the next 6 months, her sternal wound (the lung transplant incision site) healed, her infected skin nodules substantially resolved (though some skin nodules remained), and her lung and liver function improved.

M. abscessus could not be cultured from her blood or sputum at any time post-phage threatment, although it could still be cultured from the remaining skin nodules.

The patient was able to go back to school and resume her normal life, although according to a BBC news article from this week, her phage treatment continues to this day. In addition, the Hatfull lab is working to prepare a fourth phage to add to the cocktail, to help eradicate the remaining M. abscessus.

Were there any adverse or immune reactions?

No adverse effects, other than feeling sweaty and flushed at first, were observed. Immune reactions were minimal (low cytokine response), and although a weak anti-phage protein antibody response was detected, it was not enough to neutralize the phages.

Did the phages replicate in the body?

The authors tracked phage concentrations (using both plaque assays and digital PCR) at various body sites, and found substantial replication of phages (10 PFU/mL to 10^9 PFU/mL) in the blood within about 4 days.

Where did the phages end up?

They tracked phages in serum, feces, wounds and sputum over several months. By plaque assay, phages were first detected in the serum (1 day post-initial phage treatment), and then in the feces (days 4-6) and in wounds (days 3-5), followed by a spike of phages in the sputum (day 9), then about a month of undetectable phages, then two days where phages were again detected in serum. dPCR showed similar results, albeit resulting in generally lower numbers compared with plaque assays.

Was phage resistance an issue?

Although some phage resistance was observed in vitro, phage resistance was not detected in vivo. Not all the skin nodules were resolved, but wound swab isolates showed no phage resistance.

What was the process of getting approval to try phage therapy?

(For detailed info, see the Supplementary Material for the paper)

• The process began with a discussion with the UK Medicines and Healthcare Regulatory Agency (MHRA). Medical need was established.
• It was decided that the phage cocktail would be treated as an unlicensed product from a non-GMP research facility. A pharmaceutical importer licensed to import unlicensed drugs was identified, and documentation was completed according to MHRA requirements (manufacturing details, certificate of analysis).
• The UK Health and Safety Executive (HSE) was contacted to establish the Genetically Modified Organism (GMO) status of the cocktail. It was concluded that the cocktail would NOT be considered a genetically modified microorganism since phage DNA was only deleted, and none was added.
• The phage cocktail was tested for sterility at the hospital’s in-house certified cell therapy labs.
• The GOSH Drugs and Therapeutic Committee, Ethics Committee and clinical team put together a plan to manage the patient’s risk of anaphylaxis and immune response to bacterial lysis (including initial use of a topical test dose, an initially low IV dose, and the immediate availability of an ICU bed, if needed).
• Other risks, such as to the environment (considered low) and to phage research, were also discussed with the GOSH Ethics Committee. These discussions led the team to commit to publishing the results of the study regardless of the clinical outcome.
• The patient and her family gave informed consent for the treatment.

Conclusions

This study shows the safe treatment of a multi-drug resistant, disseminated, non-tuberculous Mycobacterium infection in a post-transplant patient with CF using a three-phage cocktail. Phages against the patient’s strain were uncommon, but could be identified with extensive searches. The treatment substantially reduced the presence of M. abscessus, phage replication was clearly demonstrated, and the condition of a patient who was not expected to live was substantially improved.

There are two important firsts associated with this study:

• It is the first to show use of phage therapy to treat Mycobacterium in a human patient
• It is the first to show use of engineered phages to treat any human infection

Looking into the future: could this cocktail treat other patients?

The phage cocktail was not effective on other M. abscessus strains the lab tested, suggesting against its general use in other patients. However, the Hatfull lab did show that they could find other phages in their collection that were effective against these strains. This suggests that a personalized approach could be promising for this species, given a large enough phage library. However, the paper mentions another patient with a similar infection, who unfortunately died before they could find appropriate phages, emphasizing the inherent difficulty of using personalized phage therapy for critically-ill patients infected with this species.

Further reading

• Learn more about the human story behind the paper in this article by STAT News reporter Eric Boodman
• Learn more about the phage therapy work Graham Hatfull’s lab is doing at the University of Pittsburgh
• Learn more about the SEA-PHAGES program for undergraduate phage research, including how to get your university involved
• Learn more about the BRED technique for engineering phages

Main source for this issue:

Rebekah M. Dedrick, Carlos A. Guerrero-Bustamante, Rebecca A. Garlena, Daniel A. Russell, Katrina Ford, Kathryn Harris, Kimberly C. Gilmour, James Soothill, Deborah Jacobs-Sera, Robert T. Schooley, Graham F. Hatfull & Helen Spencer. Engineered bacteriophages for treatment of a patient with a disseminated drug-resistant Mycobacterium abscessus (2019). Nature Medicine 25, 730–733.

Link: https://www.nature.com/articles/s41591-019-0437-z
DOI: 10.1038/s41591-019-0437-z
Supplementary Information: PDF