Discovery of a novel series of Penicillin-Binding Protein 3 inhibitors as monotherapy for Pseudomonas aeruginosa infections: Rational design of biochemical potency and bacterial permeation

Thomas Durand-Réville

14 April 2019 - ECCMID

2

Disclosures

• Thomas Durand-Réville: Full-time Employee; Share Holder; Entasis Therapeutics.

3

Addressing a Significant Unmet Medical Need: Pseudomonas aeruginosa (P.a.)

European CDC estimated in 2009, ~800,000 extra hospital days and 10,000 extra deaths were caused by infections with carbapenem-resistant P. aeruginosa1

P. aeruginosa causes a variety of infections, including nosocomial pneumonia, BSI, IAI, UTI, SSI– 570,000 US cases estimated in 2016– 14% of P. aeruginosa infections were multidrug-

resistant (MDR) in 20142

– P. aeruginosa MDR rates are expected to reach ~30% by 2040

Critical priority 1 (WHO pathogen list, 2017 Report)

Serious Threat level (CDC AR Threats, 2013 Report)

(1) European Centre for Disease Prevention and Control/European Medicines Agency Joint Working Group (2012). ECDC/EMEA Joint Technical Report. The bacterial challenge: time to react. http://www.ecdc.europa.eu/en/publications/Publications/0909_TER_The_Bacterial_Challenge_Time_to_React.pdf (ECDC 3)

(2) Decision Resources Group (2015): Hospital Treated Gram-negative Infections.

Annual Cases of P. aeruginosa by Region

(in millions)

0.000

0.500

1.000

1.500

2.000

Total UnitedStates

Europe Japan

2010 2020 (Est)2010 2020 (Est)

Decision Resources Group (2012): “Pseudomonas aeruginosaInfections: The Emergence of Multidrug-Resistant Strains CreatesUntapped Market Opportunity for Novel Anti pseudomonals”.

4*Reversibility of β-lactamase inhibition first exemplified with avibactam in: Ehmann D. et al., PNAS (2012) 109, 11663

Towards a Novel Class of Gram-negative Agents: Diazabicyclooctanone (DBO)

Tazobactam: hydrolyzed by -lactamases

ETX2514: stable to -lactamase hydrolysis hydrolysis

carbamoylation

reversible

acylation

irreversiblehydrolysis

X

-lactamase(active)

-lactamase(inactive)

-lactamase(active)

• In addition to acting as potent inhibitors of diverse β-lactamases, certain DBO analogs have intrinsic antibacterial activity

• Unique mechanism of inhibition and action of this class demonstrated by ETX2514*

• Several other DBO analogs have also exhibited PBP2 activity (i.e. nacubactam, zidebactam, NXL-105) but this feature has not yet been optimized

GOAL: Discover single I.V. DBO (no BLI required) to treat infections caused by P. aeruginosa, including MDR strains

5

NXL-105 model in P.a. PBP2

SER

ASP409*

NXL-105PBP2 Inhibitor

P.a. PBP2 acylation rate k(on) (M-1.s-1) 5,200

P.a. PBP3 acylation rate k(on) (M-1.s-1) 11

P.a. PBP1a acylation rate k(on) (M-1.s-1) 2

P.a. MIC (ARC6347, OXA-486, PDC-24) (mg/L) 0.25

P.a. MIC90 (N=302) (mg/L) 1

NXL-105 Shows Excellent Activity against P. aeruginosa In Vitro

*ASN in P.a. PBP1a, PBP1b and PBP3

PBP2 selectivity rationalized by key salt bridge with ASP409

6

NXL-105 is not Efficacious against P. aeruginosa In Vivo

5.93

8.298.04

6.42 6.66

3.28

2

3

4

5

6

7

8

9

10

Pre-treatment vehicle only 6.25 50 250 levofloxacin160 mg/kgq24h (PO)

Log

(CFU

/g)

Stasis

NXL-105 (mg/kg q3h)

NXL-105

70% fT>MIC

95% fT>MIC 100% fT>MIC

Murine neutropenic thigh infection model MDR P. aeruginosa Clinical Isolate (ARC6347: OXA-486, PDC-24)

MIC (imipenem) > 4 mg/L, MIC (NXL-105) = 0.25 mg/L

• Despite excellent P.a. MICs in vitro and exposures of up to 100% fT/MIC, NXL-105 is devoid of in vivo efficacy in a murine infection model

• NXL-105 also suffers from high frequency of resistance (2x10-5 to 7x10-6 at 4xMIC, with >64-fold MIC increase) and severe inoculum effect

• Hypothesis: Is this driven by PBP spectrum and can we improve on that? (Yes!)

7

New Xray Co-crystal Structure (P.a. PBP3) Unveils Key Features and Helps Guide Analog Design

Methyl displaces

water

Water deep

in pocket

ASN351

TYR357 defining

pocket volume

Key hydrogen

bond to ASN351

ETX’678 model in P.a. PBP3

ETX’054 Xray in P.a. PBP3

• Hydrogen bond with ASN351

• Small lipophilic group displaces water deep in binding pocket

• TYR357 defining pocket volume

Key P.a. PBP3 Medchem design principles:

NXL-105 ETX’991 ETX’678 ETX’054

R1 group

P.a. PBP2 acylation rate k(on) (M-1.s-1) 5,200 131 110 <8

P.a. PBP3 acylation rate k(on) (M-1.s-1) 11 230 610 582,000

8

ETX’054 X-ray in P.a. PBP3

ASN351

ETX’054 Displays Key In Vitro Differentiation Compared to NXL-105

NXL-105 ETX’054

Structure

P.a. PBP2 acylation rate k(on) (M-1.s-1) 5,200 <8

P.a. PBP3 acylation rate k(on) (M-1.s-1) 11 582,000

P.a. PBP1a acylation rate k(on) (M-1.s-1) 2 2,310

P.a. MIC (ARC6347, OXA-486, PDC-24)

(mg/L)0.25 0.25

Frequency of resistance at 4x MIC 2 x 10-5 <1 x 10-9

Inoculum effect

(MIC shift between 105-106 cfu/mL)4x none

MBC/MIC Ratio 8 1ETX’054 is a novel P.a. PBP3/PBP1a inhibitor

TYR357

9

• PK profile similar to other members of DBO class

• Robust activity achieved in neutropenic murine thigh efficacy model

• High MIC90 against recent P.a. clinical isolates required further investigation

ETX’054 is Efficacious In Vivo Against a P. aeruginosa Clinical Isolate

Compound NP.a. MIC (mg/L)

Min Max MIC50 MIC90

ETX’054 599 0.12 >32 8 32

Imipenem 599 0.06 >32 2 32

5.69

8.77

6.69

4.09

3.614.01

2

3

4

5

6

7

8

9

10

Pre-treatment vehicle only ETX'054 12.5mg/kg q3h

ETX'054 50mg/kg q3h

ETX'054 250mg/kg q3h

levofloxacin160 mg/kgq24h (PO)

Stasis

Log

(CFU

/g)

44% fT>MIC

56% fT>MIC

70% fT>MIC

ETX’054

Murine neutropenic thigh infection model MDR P. aeruginosa Clinical Isolate (ARC6347: OXA-486, PDC-24)

MIC (imipenem) > 4 mg/L, MIC (ETX’054) = 0.25 mg/L

10

ETX’054 Antibacterial Activity is Unaffected by β-lactamases in a P. aeruginosa Isogenic Panel

• -lactams lose their activity in presence of -lactamases as depicted below for piperacillin, ceftazidime and imipenem

• In contrast, ETX’054 maintains activity in the presence of all 4 classes of -lactamases tested

• ETX’054 also retains activity against recent CAZ/AVI-resistant KPC-3 variants

• Both low permeation and efflux are confirmed to contribute to high P.a. MIC90

MIC (mg/L) against P. aeruginosa isogenic strains expressing individual β-lactamases

parent Class A Class B Class C Class D

Compound PA01 CTX-M-15 KPC-2 SHV-2a TEM-1 NDM-1 VIM-1 VIM-2 AmpC P99 OXA-10 OXA-23 OXA-40 OXA-48 OXA-58

piperacillin 4 >256 >256 >256 >256 128 >256 128 256 128 >256 256 256 256 >256

ceftazidime 2 64 >64 32 4 >64 >64 >64 32 64 4 16 2 2 2

imipenem 1 1 32 1 1 16 64 64 1 1 1 16 32 64 8

ETX‘054 4 4 8 4 4 8 4 4 8 8 8 4 4 8 4

11

Optimized Cell-based Porin Over-expression Assay

• E. coli K-12 genetically modified for more sensitive and uniform response to inducer

• Deleted native porins (ompF, ompC and ompA)

• Deleted tolC to test porins with/without efflux

• Heterologous porin expression (10 P. aeruginosa,1 A. baumannii and 2 K. pneumoniae porins)

• Increasing porin concentration in the OM to increase OM permeability

• Tuning OM permeability with selective inducer in a controlled fashion

• Read out using fold-change in MIC, independent of activity level

[L-arabinose] a [porin] a permeability

[L-arabinose]

Iyer R. et al., ACS Infect. Dis. (2017) 3, 310

Incorporation of Uptake Assay in Screening Cascade: An Unprecedented Approach

L-arabinose (µM)

12

Generating Structure-Porin Permeation Relationships (SPPR) to Improve Cellular Potency

Compound ETX’054 ETX’678 ETX’991

Structure

Porin permeation none++,

multiple

+++,

multiple

P.a. PBP3 acylation

rate k(on) (M-1.s-1)582,000 610 230

0 3 0 6 0 9 0 1 2 0 1 5 0

E T X - 0 1 2 9 9 1

E T X - 0 1 2 6 7 8

E T X - 0 2 0 0 5 4

0 3 0 6 0 9 0 1 2 0 1 5 0

0 . 0 3 1 2 5

0 . 0 6 2 5

0 . 1 2 5

0 . 2 5

0 . 5

1

E T X - 0 1 2 9 9 1

E T X - 0 1 2 6 7 8

E T X - 0 2 0 0 5 4

OprDOprF

[L-arabinose], µM

Fold

-ch

ange

in

MIC ETX’054

ETX’991

ETX’678

ETX’054

ETX’991

ETX’678

• Striking differences in uptake with simple structural modifications

• Hydrogen-bond donor appears required at R1 to effectively permeate through several porins

• Design also informed by carbapenem SPPR

• Using steered MD simulation to model porin permeation at the atomistic level

13

Conclusions and Future Directions

• Entasis is redefining antibacterial design by incorporating definition of the molecular drivers of compound uptake• Unique multidisciplinary approach combining med chem, in vitro/in vivo biology and in silico tools

• We discovered a novel class of non-β-lactam PBP3 inhibitors• Lead compounds maintain activity in the presence of all 4 classes of β-lactamases tested• Using structure-based drug design, selectivity shifted from PBP2 to PBP3/PBP1a• Translated into robust in vivo activity in murine infection model• Uptake and efflux identified as features to increase potency against recent MDR P. aeruginosa clinical isolates

• The team is continuing optimization of both multi-porin permeation and target activity in parallel• Lead Optimization program sponsored by CARB-X• Candidate to be selected at end of 2019

14

Acknowledgements

• Novexel• AstraZeneca Infection Discovery

• Pharmaron• JMI Laboratories• Neosome Life Sciences