Compounds with a graded row

Ordered by rung, strongest first. The rung rates the study design; the chip beside it says what the study measured. Both are needed.

1. Bacteriocins

   animal in vivo Direct outcome Not approved for this condition

   Who was studied

   female BALB/cJRj mice aged 6 weeks, 6 mm excisional dorsal wounds inoculated with luciferase-tagged Staphylococcus aureus Xen31, derived from the multidrug-resistant clinical isolate ATCC 33591; topical gel applied daily for up to 7 days

   Bacterial burden was followed as bioluminescence and fell sharply and durably after treatment with a hydroxypropyl cellulose gel, so this is an infection outcome in a living animal against a resistant strain. The gel was a combination - garvicin KS 5 mg/mL and micrococcin P1 0.1 mg/mL together with penicillin G 5 mg/mL - so the result belongs to the mixture and not to a bacteriocin alone, systemic physiology was not assessed, and no bacteriocin has an infection outcome measured in a person.

   Source PubMed Central 1 primary source read for this row

2. Melittin

   animal in vivo Direct outcome Not approved for this condition

   Who was studied

   immunosuppressed BALB/c mice with peritoneal infection by extensively drug-resistant Acinetobacter baumannii, MRSA, or KPC-producing Klebsiella pneumoniae; melittin 2.4 mg/kg intraperitoneally every 12 hours, the highest sub-lethal dose

   Melittin inhibited all three resistant pathogens in culture (MIC 8-16 mcg/mL for XDR A. baumannii, 8-32 for MRSA, 32 for KPC-producing K. pneumoniae), but in infected mice repeated dosing changed neither survival nor peritoneal bacterial load, while colistin and vancomycin did. The limit is toxicity: 50% haemolysis occurred at 0.44 mcg/mL and the intraperitoneal LD50 in mice was 4.95 mg/kg, so the concentrations that kill these organisms in a dish cannot be reached in a living animal.

   Source PubMed Central 1 primary source read for this row

3. Proline-rich Antimicrobial Peptides (PrAMPs)

   animal in vivo Direct outcome Not approved for this condition

   Who was studied

   mice with bacteraemia caused by multidrug-resistant Acinetobacter baumannii HUMC1, 6-8 animals per group; ARV-1502 given intravenously at 1.25, 2.5 or 5.0 mg/kg twice daily for 3 days

   Survival and tissue bacterial density were both measured: all ARV-1502 monotherapy groups survived while untreated animals and those given imipenem/cilastatin alone all died, with dose-dependent falls in organ bacterial counts and no systemic toxicity noted. ARV-1502 is an engineered analogue of the insect proline-rich peptides drosocin and pyrrhocoricin, not a natural PrAMP; the natural members - Bac7(1-35), oncocin Onc72, Api137 - have in vivo results only against antibiotic-susceptible strains, and no PrAMP has been given to a person with a resistant infection in any study on file.

   Source PubMed Central 2 primary sources read for this row

4. Cecropins

   in vitro Mechanistic only Not approved for this condition

   Who was studied

   one polymyxin B- and colistin-resistant clinical isolate of Acinetobacter baumannii (strain L1), with a susceptible reference strain D41 as comparator; broth MIC and bactericidal assays only, no living subject

   Recorded, and not evidence for this condition

   Cecropin P1 inhibited the polymyxin- and colistin-resistant isolate at 2.0 mcg per assay (2.0 microM), a concentration close to that needed for the susceptible strain, showing the resistance mechanism did not protect the organism from this peptide. That is one isolate in broth; no cecropin has been given to an animal or a person with a resistant infection in any study on file, and the mouse work in the wider cecropin literature used engineered cecropin-like analogues against susceptible strains.

   Source PubMed Central 1 primary source read for this row

5. Defensins

   in vitro Mechanistic only Not approved for this condition

   Who was studied

   30 multidrug-resistant nosocomial clinical isolates - six each of Staphylococcus aureus, Enterococcus faecium, Pseudomonas aeruginosa, Stenotrophomonas maltophilia and Acinetobacter baumannii, most resistant to at least three antibiotic classes; bacterial suspensions only, no living subject

   Recorded, and not evidence for this condition

   Human beta-defensin 3 killed these resistant isolates within 1-20 minutes at 4-16 mcg/mL in low-salt buffer, but at 150 mM phosphate it needed 64 mcg/mL against A. baumannii and was inactive against S. aureus, and 20% human serum abolished its killing of the Gram-negative strains - so the killing shown here does not survive body-like conditions. Defensins is a family: the one defensin actually administered in a resistant-organism-free sepsis model, human neutrophil peptide 1, made outcomes worse (80% mortality at 10 mg/kg versus 10% in controls), the favourable survival data belong to rhesus theta-defensin RTD-1, a different molecule, and the ABSSSI trial often cited here tested brilacidin, a synthetic defensin mimetic. No study on file administered a defensin to an animal or a person with a resistant infection.

   Source PubMed Central 2 primary sources read for this row

6. Lactoferricin

   in vitro Mechanistic only Not approved for this condition

   Who was studied

   three multidrug-resistant, ESBL-producing enteroaggregative Escherichia coli clinical isolates; broth MIC/MBC, time-kill and membrane-permeation assays, plus Galleria mellonella wax moth larvae. No mammal or human subject.

   Recorded, and not evidence for this condition

   Lactoferricin (17-30), a 14-residue fragment of lactoferricin rather than the full peptide, inhibited and killed these resistant isolates at 32 microM with under 5% haemolysis at four times that concentration. What was measured was killing in broth, plus survival of infected wax moth larvae; no study on file has given lactoferricin to a mammal or a person with a resistant infection, so nothing here measured treatment of a resistant infection and the link to it is an argument from mechanism, not a result. The one lactoferrin-derived peptide taken into humans, hLF1-11, was a safety and tolerability trial that measured no infection outcome at all.

   Source PubMed Central 2 primary sources read for this row

7. LL-37 (Cathelicidin)

   in vitro Mechanistic only Not approved for this condition

   Who was studied

   seven clinical Acinetobacter baumannii isolates, four of them multidrug-resistant; broth MIC, time-kill and biofilm assays only, no living subject

   Recorded, and not evidence for this condition

   Native human LL-37 failed to inhibit any of the seven isolates at the highest concentration tested, with MICs above 250 mcg/mL, while the comparator peptide WAM-1 inhibited them at 166-250 mcg/mL; the authors also note that human serum diminishes LL-37's antibacterial activity. This is a negative culture result, and no study on file gave LL-37 to an animal or a person with a resistant infection - the human LL-37 trials on file are in venous leg ulcers and diabetic foot ulcers and measured healing and surface colony counts, not resistant-organism cure.

   Source academic.oup.com 1 primary source read for this row

8. Magainins

   in vitro Mechanistic only Not approved for this condition

   Who was studied

   109 bacterial isolates tested by CLSI broth microdilution: 46 from diabetic foot infections (including 8 MRSA) and 63 selected for resistance mechanisms - VRSA, VISA and hVISA, VanA and VanB vancomycin-resistant enterococci, ESBL-producing Enterobacteriaceae, KPC, NDM, IMP-1 and VIM-2 carbapenemase producers, and MDR Acinetobacter and Pseudomonas. No living subject.

   Recorded, and not evidence for this condition

   Pexiganan, a synthetic magainin analogue, inhibited these resistant isolates in culture with an overall MIC50 of 16 mcg/mL and MIC90 of 32 mcg/mL, with no loss of potency against the resistant phenotypes; everything here was measured in broth. Pexiganan is the one magainin taken into people, in mildly infected diabetic foot ulcers rather than in confirmed resistant infection: two phase 3 equivalence trials against oral ofloxacin led to FDA refusal in 1999 because equivalence was not superiority, and the repeat placebo-controlled phase 3 trials OneStep-1 and OneStep-2 (NCT01590758, NCT01594762) missed their primary endpoint in 2016, so it has never been approved. No study on file measured treatment of a resistant infection in an animal or a person with any magainin.

   Source PubMed Central 3 primary sources read for this row

9. Nisin

   in vitro Mechanistic only Not approved for this condition

   Who was studied

   106 staphylococcal isolates from periprosthetic joint infection, including 26 MRSA and 27 methicillin-resistant Staphylococcus epidermidis; broth microdilution only, no living subject

   Recorded, and not evidence for this condition

   Native nisin A inhibited the methicillin-resistant subsets in culture at MIC50 4 and MIC90 8 mcg/mL for MRSA, and MIC50 2 and MIC90 4 mcg/mL for methicillin-resistant S. epidermidis. This was measured in broth against isolates taken from patients, not in patients; no study on file gave nisin to an animal or a person with a resistant infection, and nisin's long record as a food preservative is a food-additive safety record, not evidence of treating infection.

   Source PubMed Central 1 primary source read for this row

10. Temporins

    in vitro Mechanistic only Not approved for this condition

    Who was studied

    20 carbapenemase-producing Klebsiella pneumoniae clinical isolates, KPC and metallo-beta-lactamase producers, plus reference strains; broth MIC assays only, no living subject

    Recorded, and not evidence for this condition

    The peptide that inhibited these carbapenem-resistant isolates at 6.25-25 microM was a lipidated temporin L analogue, and native unmodified temporin L was not tested in this work, so the result belongs to the analogue rather than to a natural temporin. Everything was measured in broth: no study on file gave any temporin, native or modified, to an animal or a person with a resistant infection.

    Source PubMed Central 2 primary sources read for this row