Personalized bacteriophage therapy to treat spinal drug-resistant Pseudomonas aeruginosa infection


Ethics approval

The patient has consented to our use of the samples and data in research and to the publication of the clinical data, photo and video shown here. He also gave written informed consent to receive phage therapy in accordance with CARE guidelines and the principles of the Declaration of Helsinki. The Hospices Civils de Lyon ethics committee concluded that phage therapy was ethically justified in this clinical situation (19-162).


Phagograms were performed as follows: 15 mL of prewarmed (45–50°C) LB containing 0.75% (w/v) agar (LB soft-agar) was inoculated with 200 µL of culture of a night of each bacterial strain tested and poured into 90 mm round Petri dishes. After solidification of LB soft agar at room temperature, 5 µL drops of 10-fold serially diluted stock suspensions of each phage were placed on top of the LB soft agar layer, with the most concentrated suspension on top and the more diluted. at the bottom of the Petri dish. After drying the drops at room temperature, the Petri dishes were incubated aerobically at 37° C. overnight. The next day the Petri dishes were visually checked for areas of lysis.

Phage characterization

The phagogram was performed with > 100 P. aeruginosa phages from the phage collection of the University of Lausanne to find active phages on the patient’s strain. The genomic DNA of three active phages (named vB_PaeP_4029, vB_PaeP_4032 and vB_PaeP_4034) has been fully sequenced. Phage genomic libraries were prepared with an optimized protocol and standard Illumina adapter sequences, and sequencing was performed with Illumina technology, NovaSeq 6000 (2 × 150 bp read mode) at Eurofins Genomics Germany GmbH (Ebersberg , Germany). Reads were assembled into contigs using the PATRIC v3.6.12 pipeline for assembly with the parameters “trim read before assembly=TRUE”, “min contig length=1000”, and “min contigcover=100” and the contigs were annotated using PATRIC v3. 6.12 pipeline for annotation following “bacteriophage recipe” with default settings ( The genomes of the three phages were 72,063 bp long. Using blastn (, the nearest neighbor of vB_PaeP_4029 was phage vB_PaeP_PYO2 (Genbank accession number MF49236.1) with 100% coverage and 99, 72% identities. The nearest neighbor of vB_PaeP_4032 and vB_PaeP_4034 was phage PEV2 (Genbank accession number KU948710.1) with 100% coverage and 99.94% identity and 100% coverage and 99.95% identities, respectively. Phage vB_PaeP_4029 had 99.86% identity over 99% coverage with phage vB_PaeP_4032 and 99.73% identity over 100% coverage with phage vB_PaeP_4034. The vB_PaeP_4032 and vB_PaeP_4034 phages differed by five single nucleotide polymorphisms. The phages, which resembled podoviruses on electron microscopy (EM) micrographs (Fig. 3A–C) were further classified into the family Schitoviridaegender Litunavirus based on their genomic sequences. The lytic nature of the three phages was verified using the PhageAI repository (, which confirmed their suitability for use in a phage therapy protocol. Moreover, the total absence on the three genomes of genes linked to the 31,552 genes encoding known and predicted virulence factors listed in the Virulence Factors Database (VFDB, main.htm) has confirmed their good security profile. The phagogram and plating efficiency tests performed as previously described16 on the patient and the producing strain of the three unpurified phages showed lytic activity (Fig. 3D–F and Table 1).

Phage production

The production of the pAPIs, in accordance with a monograph of the Belgian Pharmacopoeia describing the production process and the QC system for incorporation into magistral preparations, was carried out at the laboratory of the Queen Astrid military hospital in Brussels in collaboration with pharmacists at the Croix-Rousse hospital (Hospices Civils de Lyon)4.16, and under the supervision of the National Agency for the Safety of Medicines and Health Products (ANSM). It should be noted that the monograph received on January 10, 2018 a formal positive opinion from the Belgian Minister of Public Health requested from the Federal Agency for Medicines and Health Products (FAMHP). It was designed by representatives of the Queen Astrid Military Hospital located (QAMH) in Brussels, the FAMHP and Sciensano, formerly known as the Belgian Scientific Institute for Public Health. Here the three P. aeruginosa phages were produced using P. aeruginosa host strains PAO1 (for phages vB_PaeP_4029 and vB_PaeP_4034) and ATCC® 15442™ (for phage vB_PaeP_4032), according to the Phage API Monograph. Phages were propagated using the double agar overlay method in growth media without animal products with an alternative protein source (APS-LB broth, BD) following the work instructions (WI) developed by QAMH. The phages were mixed with the corresponding bacterial host strains at the expected multiplicity of infection of the order of 10−3–5 × 10−4 and 0.6% APS-LB agar at 45°C in the final volume of 12 ml and spread on 12 mm square plates filled with 55 ml of 1.5% APS-LB agar. Plates were incubated overnight at 37°C. The top layer was scraped from the plates and centrifuged at 6000× g for 20 mins. The supernatants obtained were filtered through 0.45 μm and 0.22 μm polyether sulfone (PES) filters. Additionally, phages were pelleted by centrifugation at 35,000 ×g for 60 mins. The resulting phage pellet was resuspended in 10 times less volume of Dulbecco’s phosphate buffered saline (DPBS) to obtain purified phage stocks at high titers, i.e. in the range of 1010–11 pfu/mL. Next, the phage production process continued in the cleanroom facility where the phage stocks were packaged as APIs by diluting to a final concentration of 109–10pfu/mL in DPBS, followed by filtration through 0.22 µm PES filters, endotoxin purification using EndoTrap®(Lionex, Germany) and final filtration through 0.22 µm medical grade polyvinylidene fluoride filters. Samples of each of the three phage APIs were sent to Sciensano for QC testing, including determination of pH, endotoxin level (EU/mL), and bioburden. pH values ​​for phage APIs ranged from 7.29 to 7.35, endotoxin content, endotoxin concentration ranged from 134 to 3400 IU/mL, and no bacterial growth was observed in any of the samples. Based on the above results, the three phage APIs have been approved by Sciensano and found to be safe for various human applications taking into consideration the endotoxin level limits for each specific route of administration.4.16.

Temporary authorization for the use of cefiderocol

We have obtained from the ANSM a Temporary Authorization for Use (ATU) of cefiderocol, a new cephalosporin evaluated in clinical trials17which has shown activity against the patientP. aeruginosa strain (MIC = 1 mg/L, microdilution in broth).

Summary of reports

Further information on the research design can be found in the summary of nature research reports linked to this article.


Comments are closed.