In vitro activity of cefepime/zidebactam (WCK 5222) against Gram-negative bacteria

Objectives: Diazabicyclooctanes (DBOs) inhibit class A, class C and some class D β-lactamases. A few also bind PBP2, conferring direct antibacterial activity and a β-lactamase-independent ‘enhancer' effect, potentiating β-lactams targeting PBP3. We tested a novel DBO, zidebactam, combined with cefepime.

Methods: CLSI agar dilution MICs were determined with cefepime/zidebactam in a chequerboard format. Bactericidal activity was also measured.

Results: Zidebactam MICs were ≤2 mg/L (mostly 0.12–0.5 mg/L) for most Escherichia coli, Klebsiella, Citrobacter and Enterobacter spp., but were >32 mg/L for Proteeae, most Serratia and a few E. coli, Klebsiella and Enterobacter/Citrobacter. The antibacterial activity of zidebactam dominated chequerboard studies for Enterobacteriaceae, but potentiation of cefepime was apparent for zidebactam-resistant isolates with class A and C enzymes, illustrating β-lactamase inhibition. Overall, cefepime/zidebactam inhibited almost all Enterobacteriaceae with AmpC, ESBL, K1, KPC and OXA-48-like β-lactamases at 1 + 1 mg/L and also 29 of 35 isolates with metallo-carbapenemases, including several resistant to zidebactam alone. Zidebactam MICs for 36 of 50 Pseudomonas aeruginosa were 4–16 mg/L, and the majority of AmpC, metallo-β-lactamase-producing and cystic fibrosis isolates were susceptible to cefepime/zidebactam at 8 + 8 mg/L. Zidebactam MICs for Acinetobacter baumannii and Stenotrophomonas maltophilia were >32 mg/L; potentiation of cefepime was frequent for S. maltophilia, but minimal for A. baumannii. Kill curve results largely supported MICs.

Conclusions: Zidebactam represents a second triple-action DBO following RG6080, with lower MICs for Enterobacteriaceae and P. aeruginosa. Clinical evaluation of cefepime/zidebactam must critically evaluate the reliance that can be placed on this direct antibacterial activity and on the enhancer effect as well as β-lactamase inhibition.

Introduction

Diazabicyclooctanes (DBOs) are among the most promising new β-lactamase inhibitors.¹ The first member of the class, avibactam, is already marketed in combination with ceftazidime and is under investigation combined with aztreonam^1,2 while a second analogue, relebactam, is now in Phase III development combined with imipenem/cilastatin.³ Avibactam and relebactam act solely as inhibitors of class A, class C and some class D β-lactamases at clinical concentrations, though avibactam does directly inhibit the growth of many Escherichia coli strains at concentrations a little above the 4 mg/L routinely used in the MIC tests. Avibactam MICs for other species are higher.

Some developmental DBOs have greater direct antibacterial activity. RG6080/OP0595 (Meiji, Fedora, Roche) not only has similar β-lactamase inhibitory activity to avibactam, but also has MICs of ∼1–4 mg/L for most E. coli, Klebsiella, Enterobacter and Citrobacter spp., contingent on attacking PBP2.^4,5 Proteeae and non-fermenters are resistant, with MICs >32 mg/L. Like mecillinam⁶—another PBP2-targeting agent—RG6080 also synergizes or ‘enhances’ the activity of PBP3-targeted β-lactams against many E. coli, Klebsiella spp. and Enterobacter spp., regardless of whether these produce β-lactamases. The enhancer effect is retained against some strains and mutants with resistance to the antibacterial action of OP0595 and these additional activities allow β-lactam/RG6080 combinations to achieve in-vitro activity against many Enterobacteriaceae with metallo-β-lactamases (MBLs), even though these evade inhibition by DBOs.^4,5,7

In the present study, we characterized the activity of a second DBO with direct antibacterial activity, zidebactam (WCK 5107; Wockhardt, Aurangabad, India; Figure 1), tested in combination with cefepime, which is its intended clinical partner β-lactam (see e.g. https://clinicaltrials.gov/ct2/show/NCT02707107). The cefepime/zidebactam combination is also known by the code number WCK 5222.

Figure 1.

Structure of zidebactam.

Materials and methods

Isolates

Isolates (n = 269) were reference submissions to Public Health England from UK diagnostic laboratories, or were collected during resistance surveys. The distribution of resistance mechanisms by species is shown in Table 1. Isolates were identified using API20E or API20NE strips (bioMérieux, Marcy-l’Étoile, France) or by MALDI-TOF MS (Maldi-Biotyper; Brüker, Bremen, Germany), with the exception that Acinetobacter baumannii were identified by the PCR detection of bla_OXA-51-like.⁸ Carbapenemase genes were identified by PCR or sequencing;⁹ other mechanisms were inferred from phenotype and (where available) genotype data.

Table 1.

Species and genera represented in Enterobacteriaceae groups

Species	Resistance mechanisms
	ESBL	AmpC	KPC	OXA-48- like	MBLa	susceptible controls
E. coli	10	10	10	5	10	5
Klebsiella	10	5	10	10	10	5
Enterobacter/Citrobacterb	10	10	10	0	10	5
Serratia		5				5
Proteeaec	4	5			5	5

a

20 with NDM enzymes and 15 with VIM types.

b

12 Citrobacter freundii and 33 Enterobacter spp.

c

13 M. morganii, 4 Providencia spp. and 2 Proteus spp.

Susceptibility testing

The MICs of cefepime (US Pharmacopoeial Convention, Rockville, MD, USA) were determined by CLSI agar dilution¹⁰ in a chequerboard format with zidebactam (Wockhardt) included at 0.06–8 mg/L. Comparator antibiotics were tested in parallel and comprised: piperacillin (Sigma–Aldrich, Poole, UK) with 4 mg/L tazobactam (Wockhardt), ceftazidime (Sigma–Aldrich) alone and with 4 mg/L avibactam (Wockhardt), and meropenem (Sequoia Research Products, Pangbourne, UK).

Killing curves

Bacteria were grown overnight, with shaking, in Mueller–Hinton broth at 37°C then diluted 1000-fold into 100 mL of fresh warm broth. Incubation was continued, with shaking, for 90 min to bring the cells into the early log phase. The cultures were then divided into 10 mL volumes and antibiotics or combinations were added, with incubation continued as before. This point was defined as T₀, and a single count was performed, representing the starting point for all curves with that strain. Further counts were performed on all cultures at T + 1, T + 2, T + 4, T + 6, T + 8 (non-fermenters only) and T + 24 h. Counts were by the Miles and Misra method and ‘bactericidal’ is used in the classical sense as meaning ‘causing some initial reduction in bacterial counts’, irrespective of the extent or duration of these reductions.

Results

Antibacterial activity of zidebactam

The great majority (92 of 102) of isolates of E. coli, Enterobacter spp. and Citrobacter spp. were susceptible to zidebactam at ≤1 mg/L, with 86 of 102 MICs clustered from 0.12 to 0.5 mg/L (Table 2). MICs for the Klebsiella spp. were more bimodally distributed, with 40 of 58 values from 0.12 to 2 mg/L and 16 of 58 at ≥32 mg/L. Trailing endpoints and surviving colonies made reading difficult, particularly with Klebsiella spp. Zidebactam MICs also were bimodal for Serratia spp., but with most (7 of 10) values ≥32 mg/L. All Proteeae (n = 19) were resistant, with MICs ≥32 mg/L. No relationship was apparent between zidebactam MICs and the β-lactamase phenotypes and genotypes for which the Enterobacteriaceae were selected for inclusion in the study.

Table 2.

MIC distributions of zidebactam by species and, for P. aeruginosa, resistance mechanism

	Number of isolates with indicated MIC (mg/L)
	≤0.06	0.12	0.25	0.5	1	2	4	8	16	32	>32
E. coli (n = 50)	3	28	12	5	1				1
Klebsiella spp. (n = 58)		3	17	17	2	1	1	1		13	3
Enterobacter and Citrobacter spp. (n = 52)		11	20	10	2		1			1	7
Serratia spp. (n = 10)			1	1				1		7
Proteeae (n = 6)											6
P. aeruginosa (n = 50)
β-lactam-susceptible controls (n = 10)						3a	5	1			1
AmpC derepressed (n = 10)							2	5	3
MBL producers (n = 10)								6	2	1	1
up-regulated efflux (n = 10)								3	2	5
cystic fibrosis, mixed mechanisms (n = 10)						1a	2	2	3	1	1
A. baumannii (n = 30)											30
S. maltophilia (n = 10)											10

a

MIC ≤ indicated value.

In the case of P. aeruginosa, MICs for 36 of 50 isolates were in the range 4–16 mg/L. The median values for AmpC- and MBL-producing isolates (8 mg/L) were one doubling dilution higher than for the susceptible controls (4 mg/L) and the median for the increased efflux isolates was a further 2-fold higher, at 16 mg/L. Zidebactam MICs for A. baumannii and Stenotrophomonas maltophilia were universally >32 mg/L.

Combination activity of cefepime/zidebactam: Enterobacteriaceae

At 1 mg/L (EUCAST's susceptible breakpoint, http://www.eucast.org) unprotected cefepime inhibited only 6 of 33 ESBL producers, 26 of 35 AmpC hyperproducers, 4 of 5 K1 hyperproducers, 7 of 15 with OXA-48-like enzymes and none of those with KPC (n = 30) or MBLs (n = 35) (Table 3). The addition of zidebactam increased these proportions markedly, so that cefepime/zidebactam at 1 + 1 mg/L was active against all 33 Enterobacteriaceae with ESBLs, all 35 with hyperproduced AmpC enzymes, all five with hyperproduced K1 enzyme (n = 5), all 15 with OXA-48-like carbapenemases, 29 of 30 with KPC enzymes and 29 of 35 with MBLs. The sole KPC isolate that was resistant at 1 + 1 mg/L was an Enterobacter cloacae that was inhibited by zidebactam alone at 4 mg/L and by cefepime/zidebactam at 8 + 2 or 4 + 4 mg/L. Much of this gain in spectrum reflected the direct antibacterial activity of zidebactam, which anyway inhibited many E. coli, Klebsiella, Enterobacter and Citrobacter spp. isolates at 1 mg/L (above, Table 2).

Table 3.

MIC distributions for cefepime/zidebactam and comparator agents in relation to resistance groups and zidebactam concentrations

Antibiotic and inhibitor (the concentration of which is given in terms of mg/L)	Inhibited by zidebactam alone (n)	Number of isolates with indicated MIC (mg/L)
		≤0.03	0.06	0.12	0.25	0.5	1	2	4	8	16	32	64	128	256	>256
Control Enterobacteriaceae, without cephalosporin-hydrolysing β-lactamases or carbapenemases (n = 25)
FEP	–	12	7	4	2
FEP/ZID, 0.06	1	16	3	3	2
FEP/ZID, 0.12	9	10	1	4	1
FEP/ZID, 0.25	13	8	3	1
FEP/ZID, 0.5	14	11
FEP/ZID, 1	14	11
FEP/ZID, 2	14	11
FEP/ZID, 4	14	11
FEP/ZID, 8	14	11
TZP, 4	–				4	1		11	6	3
CAZ	–		1	9	10	3	2
CAZ/AVI, 4	–	5	4	6	7	2	1
MEM	–	14	8	3
ESBL-producing Enterobacteriaceae (n = 33)
FEP	–					2	4	8	3	1	4	1	3		1	6
FEP/ZID, 0.06	0		1	6	4	4	7	3	2	2	2	2
FEP/ZID, 0.12	16	4	2	1	3	2	1	2		1	1
FEP/ZID, 0.25	24	5	1	1				1		1
FEP/ZID, 0.5	27	3	1	1			1
FEP/ZID, 1	27	5				1
FEP/ZID, 2	27	5			1
FEP/ZID, 4	27	5		1
FEP/ZID, 8	27	5	1
TZP, 4	–						1	6	11	6	3	1	1		1	3
CAZ	–					1	3	3	2		2	3	6	9		4
CAZ/AVI, 4	–	1	2	7	15	6	1	1
MEM	–	20	11	2
K. oxytoca, hyperproducing K1 β-lactamase (n = 5)
FEP	–					1	3		1
FEP/ZID, 0.06	0			2	1	2
FEP/ZID, 0.12	0		4	1
FEP/ZID, 0.25	0	5
FEP/ZID, 0.5	3	2
FEP/ZID, 1	3	2
FEP/ZID, 2	3	2
FEP/ZID, 4	3	2
FEP/ZID, 8	3	2
TZP, 4	–														1	4
CAZ	–					4	1
CAZ/AVI, 4	–			3	2
MEM	–	3	2
AmpC β-lactamase-producing Enterobacteriaceae
FEP	–	2	2	5	4	7	6	6	3
FEP/ZID, 0.06	0	5	7	11	3	2	4	3
FEP/ZID, 0.12	4	14	7	2	1	4	3
FEP/ZID, 0.25	15	9	5	2	4
FEP/ZID, 0.5	23	10	2
FEP/ZID, 1	25	8	2
FEP/ZID, 2	25	8	2
FEP/ZID, 4	25	8	2
FEP/ZID, 8	27	6	2
TZP, 4	–				2			1		5	7	6	6	4	3	11
CAZ	–						1	1	1	1	5	6	9	10	1
CAZ/AVI, 4	–	1	3	3	11	14	3
MEM	–	13	14	4	4
KPC β-lactamase-producing Enterobacteriaceae
FEP	–							1	1	3	5	2	3	7	2	6
FEP/ZID, 0.06	1				1			3	7	1	8	4	1	2	1	1
FEP/ZID, 0.12	6				2	4	1		3	7	1	2	2	1	1
FEP/ZID, 0.25	17	3		1	1	1			1		2		3	1
FEP/ZID, 0.5	24	1	1		1							2		1
FEP/ZID, 1	26	1	1	1									1
FEP/ZID, 2	27	1	1							1
FEP/ZID, 4	28	2
FEP/ZID, 8	28	2
TZP, 4	–												1	2	2	25
CAZ	–								1	7	3	7	7	1	1	3
CAZ/AVI, 4	–	6	1	6	5	7	3	1	1
MEM	–						1	2	5	6	4	8	4b
OXA-48 β-lactamase-producing Enterobacteriaceae
FEP	–				3	1	3					1	5	1		1
FEP/ZID, 0.06	1		2	1	1	2		1	2	4		1
FEP/ZID, 0.12	5	2	2	1		1		2	1		1
FEP/ZID, 0.25	6	5		1	2				1
FEP/ZID, 0.5	11	2	2
FEP/ZID, 1	11	4
FEP/ZID, 2	11	4
FEP/ZID, 4	12	3
FEP/ZID, 8	12	3
TZP, 4	–													2	6	7
CAZ	–			1	3	2		1	3				3	2
CAZ/AVI, 4	–		1	5	7	2
MEM	–		1			3	6	2	1			2
MBL-producing Enterobacteriaceae
FEP	–				1				1	2	4	4	7	3	5	8
FEP/ZID, 0.06	0				1		1	1	3	7	1	2	8	4	5	2
FEP/ZID, 0.12	5	1		1	1	1	4	1		6	2	3	3	2	5
FEP/ZID, 0.25	20	1			1		1		3	2	1		1		4	1
FEP/ZID, 0.5	25			2	1	1	1	1					1		3
FEP/ZID, 1	26		1		1	1	1	1					1		3
FEP/ZID, 2	27				1	1	1	1					1		3
FEP/ZID, 4	27				1	1	1	1					1		3
FEP/ZID, 8	27				2		1	1					1		3
TZP, 4	–											1	1	1	7	25
CAZ	–									1			1	4	3	26
CAZ/AVI, 4	–						1			1	5	3	1			24
MEM	–							1	9	5	3	9	5	3
Control P. aeruginosa
FEP	–				1		2	3	2	1	1
FEP/ZID, 4	7						3
FEP/ZID, 8	9					1a
TZP, 4	–					1		1	3	4			1
CAZ	–					2	2	3	2		1
CAZ/AVI, 4	–					4a	2	4
MEM	–	1	1	2	5		1
P. aeruginosa, derepressed for AmpC β-lactamase
FEP	–								1	1	2	5		1
FEP/ZID, 4	1						1		2	4	1	1
FEP/ZID, 8	7						1	1			1
TZP, 4	–											1	1	1	3	4
CAZ	–									1	1	1	4	2	1
CAZ/AVI, 4	–						1	2	3	3	1
MEM	–					2	6	1	1
P. aeruginosa, with MBLs
FEP	–										1	2		1	1	5
FEP/ZID, 4	0									1		1	2	1	1	4
FEP/ZID, 8	6					2a									1	1
TZP, 4	–											1	4		2	3
CAZ	–											2	1	1		6
CAZ/AVI, 4	–											2	1	1		6
MEM	–											2	1	7b
P. aeruginosa, with up-regulated efflux
FEP	–									2	6	2
FEP/ZID, 4	0					1a				5	4
FEP/ZID, 8	2						1	1	1	3	2
TZP, 4	–											4	2	1	2	1
CAZ	–									5	1	2	2
CAZ/AVI, 4	–								1	8	1
MEM	–										1	6	3
P. aeruginosa, cystic fibrosis isolates
FEP	–												2	1	5	2
FEP/ZID, 4	3					1a						2	2	2
FEP/ZID, 8	5					2a		2					1
TZP, 4	–									1					2	7
CAZ	–														4	6
CAZ/AVI, 4	–									3		1	2	1	3
MEM	–									2	4	2	2
A. baumannii, susceptible controls
FEP	–						2	2	1
FEP/ZID, 4	0						1	3	1
FEP/ZID, 8	0							5
TZP, 4	–					4a				1
CAZ	–						2	2	1
CAZ/AVI, 4	–							2	2	1
MEM	–				4	1
A. baumannii, OXA carbapenemases
FEP	–										3	12	4			1
FEP/ZID, 4	0									2	10	5	3
FEP/ZID, 8	0							1		2	9	6	2
TZP, 4	–											2	0	1	4	13
CAZ	–									1	2		2	7	2	6
CAZ/AVI, 4	–								1	1	6	5	2	1		4
MEM	–								1	1	1	11	4	2b
A. baumannii, metallo (NDM) carbapenemases
FEP	–															5
FEP/ZID, 4	0														1	4
FEP/ZID, 8	0														1	4
TZP, 4	–															5
CAZ	–															5
CAZ/AVI, 4	–															5
MEM	–												1	4
S. maltophilia
FEP	–							1		2	2	3	2
FEP/ZID, 4	0						1	1	3	2	2	1
FEP/ZID, 8	0					1a		1	3	2	2	1
TZP, 4	–										1		3		1	5
CAZ	–					1		1		1	1	1	1	1	2	1
CAZ/AVI, 4	–					1		1		1	1	1	2		3
MEM	–												1	9b

AVI, avibactam; CAZ, ceftazidime; FEP, cefepime; MEM, meropenem; TZP, piperacillin/tazobactam; ZID, zidebactam.

a

MIC ≤ indicated value.

b

MIC ≥ indicated value.

The β-lactamase inhibitory activity and enhancer effects of zidebactam became evident for a minority of Enterobacteriaceae with high MICs for the DBO, taken here as MIC ≥16 mg/L, which are listed in Table 4. Strong, dose-dependent synergy was seen for all zidebactam-resistant Enterobacteriaceae isolates with class A β-lactamases, including ESBLs (which were mostly CTX-M types based on higher MICs for cefotaxime than ceftazidime) and KPC types, with cefepime MICs of 2 to >256 mg/L reduced below 1 mg/L even by zidebactam at ≤1 mg/L. The sole ‘zidebactam-resistant’ (MIC >32 mg/L) representative with an AmpC enzyme (Serratia marcescens SE01046) had only intermediate resistance to cefepime, with an MIC of 2 mg/L, reduced to ≤0.03 mg/L by zidebactam at 1 mg/L. Good cefepime/zidebactam synergy was seen for two zidebactam-resistant isolates with OXA-48 carbapenemase, but this oxacillinase has little activity against cefepime¹¹ and it is most likely that the synergy reflected inhibition of co-produced ESBLs, which were not further characterized in this study. Potentiation of cefepime by zidebactam was variable for the zidebactam-resistant metallo-carbapenemase producers, being at least 8-fold for two Klebsiella pneumoniae (H113980340 and H112240413), one Morganella morganii (H092540314) and one Providencia stuartii (H124880510), all of which were susceptible to cefepime/zidebactam at 2 + 1 mg/L, but weak or absent for all three Providencia rettgeri (H123140552, H123560843 and H124880511) and the one E. coli (H130680324), where the cefepime/zidebactam MIC remained >64 + 8 mg/L.

Table 4.

Combination behaviour against Enterobacteriaceae with zidebactam MICs ≥16 mg/L and cefepime MICs ≥2 mg/L

Specimen ID	Species and mechanism	MIC of zidebactam (mg/L)	Cefepime MIC (mg/L) with zidebactam at:
			0	0.06	0.12	0.25	0.5	1	2	4	8
SE01046	S. marcescens, AmpC	>32	2	1	0.25	0.25	≤0.03	≤0.03	≤0.03	≤0.03	≤0.03
H053420099	K. pneumoniae, CTX-M 9 group	>32	64	32	16	8	0.125	≤0.03	≤0.03	≤0.03	≤0.03
NCTC 13465	K. pneumoniae, CTX-M-25	>32	16	1	0.5	0.06	≤0.03	≤0.03	≤0.03	≤0.03	≤0.03
Mei 1	K. pneumoniae, ESBL	>32	2	0.06	0.125	≤0.03	≤0.03	≤0.03	≤0.03	≤0.03	≤0.03
SE06031	M. morganii, CTX-M 1 group	>32	4	0.25	0.06	≤0.03	≤0.03	≤0.03	≤0.03	≤0.03	≤0.03
H053460141	Proteus spp., ESBL	>32	>256	32	8	2	1	0.5	0.25	0.125	0.06
LN09056	Proteus mirabilis, ESBL	>32	>256	1	0.25	0.125	0.06	≤0.03	≤0.03	≤0.03	≤0.03
H092260700	Klebsiella spp., OXA-48 + ESBL	>32	64	8	2	0.25	0.06	≤0.03	≤0.03	≤0.03	≤0.03
H112860135	Klebsiella spp., OXA-48 + ESBL	>32	>256	8	2	0.125	0.06	≤0.03	≤0.03	≤0.03	≤0.03
H131480242	M. morganii, ESBL	>32	>256	>256	>256	256	128	≤0.03	≤0.03	≤0.03	≤0.03
H124240625	K. pneumoniae, KPC + SHV	>32	256	128	64	64	32	0.125	0.06	≤0.03	≤0.03
H114600525	Enterobacter aerogenes, KPC	>32	64	16	8	4	0.06	≤0.03	≤0.03	≤0.03	≤0.03
H113980340	K. pneumoniae, NDM, ATM-R	>32	256	64	32	8	0.25	0.25	0.25	0.25	0.25
H112240413	K. pneumoniae, VIM, ATM-R	>32	4	2	2	1	0.5	0.5	0.5	0.5	0.25
H130680324	E. coli, NDM, ATM-R	16	>256	256	256	256	256	256	256	256	256
H092540314	M. morganii, NDM, ATM-I	>32	64	8	8	4	1	1	1	1	1
H123140552	P. rettgeri, NDM, ATM-R	>32	>256	>256	256	256	256	256	256	256	256
H123560843	P. rettgeri, NDM, VEB, CMY-14 ATM-R	>32	>256	256	256	256	256	256	256	256	256
H124880510	P. stuartii, NDM, ATM-S	>32	16	16	16	16	2	2	2	2	2
H124880511	P. rettgeri, NDM, ATM-S	>32	64	64	64	64	64	64	64	64	64

ATM-S/I/R, aztreonam susceptible, intermediate or resistant, based on prior testing by BSAC methodology and taken as an indicator of ESBL/AmpC presence or absence in MBL-producing isolates.

Ceftazidime/avibactam, tested as a comparator, was active against all ESBL, K1, AmpC, OXA-48 and KPC strains at its 8 + 4 mg/L EUCAST and FDA breakpoint. Its MICs were higher than for cefepime/zidebactam, largely owing to the lack of direct antibacterial activity by avibactam; more critically, almost all (33/35) MBL producers were resistant to ceftazidime/avibactam, even at 8 + 4 mg/L. The other comparators had very limited activity against this highly resistant strain collection. Unprotected ceftazidime was only active against control strains, K1-enzyme-hyperproducing K. oxytoca and those isolates that had OXA-48-like enzymes, but lacked ESBLs. Non-susceptibility rates to piperacillin/tazobactam (8 + 4 mg/L) were >90% among isolates with AmpC, K1, OXA-48-like, KPC enzymes of MBLs; meropenem resistance was near universal among the MBL- and KPC-producing isolates, though MICs for many with OXA-48-like enzymes remained around the CLSI and EUCAST susceptible breakpoints of 1 and 2 mg/L.

Combination activity of cefepime/zidebactam: non-fermenters

At a concentration of 8 mg/L, the antibacterial activity of zidebactam dominated combination results for P. aeruginosa, with 33 of 50 isolates inhibited by the DBO alone (Table 2). Largely owing to this, 9 of 10 isolates with derepressed AmpC, 8 of 10 with MBLs, 8 of 10 with up-regulated efflux and 9 of 10 cystic fibrosis isolates were susceptible to cefepime/zidebactam at 8 + 8 mg/L. Even at 4 mg/L, zidebactam increased the proportion of strains counting as susceptible to cefepime (MIC ≤8 mg/L) from 2 of 10 to 8 of 10 for AmpC hyperproducers, 2 of 10 to 6 of 10 for efflux strains and from 0 of 10 to 4 of 10 for cystic fibrosis isolates, although 9 of 10 MBL producers remained resistant.

Cefepime MICs for A. baumannii isolates with OXA carbapenemases were mostly reduced by one doubling dilution by zidebactam at 4 or 8 mg/L, with modal values falling from 32 to 16 mg/L (Table 2); MICs for susceptible controls or those with NDM MBLs were not reduced. MICs for S. maltophilia isolates were reduced by zidebactam: without the DBO only 2 of 10 isolates were susceptible to cefepime at 8 mg/L, but this proportion rose to 7 of 10 with zidebactam present at 4 or 8 mg/L.

Killing curves

Killing curves were determined with two isolates each of K. pneumoniae, E. coli and P. aeruginosa, all producing NDM metallo-carbapenemases. In each case, the test strains per species were chosen to include one susceptible to zidebactam and one resistant, though the differential was much greater in the Enterobacteriaceae pairs than for the P. aeruginosa pairs (see Figure 2 and its legend). A single A. baumannii strain with OXA-23 carbapenemase was also tested; this, like all members of its species, was highly resistant to zidebactam. Cefepime MICs were ≥256 mg/L for all these organisms.

Figure 2.

Killing curves for Gram-negative bacteria with NDM carbapenemases by cefepime, zidebactam and their combinations. (a) K. pneumoniae H113840625 with cefepime MIC 256 mg/L, zidebactam MIC 0.25 mg/L and meropenem MIC 32 mg/L. (b) K. pneumoniae H113980340 with cefepime MIC 256 mg/L, zidebactam MIC >32 mg/L and meropenem 32 mg/L. (c) E. coli H131020913 with cefepime MIC >256 mg/L, zidebactam MIC 0.25 mg/L and meropenem MIC 64 mg/L. (d) E. coli H130480324 with cefepime MIC >256 mg/L, zidebactam MIC 16 mg/L and meropenem MIC 32 mg/L. (e and f) P. aeruginosa H130680310 with cefepime MIC >256 mg/L, zidebactam MIC 8 mg/L and cefepime MIC and meropenem MIC >64 mg/L. (g and h) P. aeruginosa H131800691 with cefepime MIC >256 mg/L, zidebactam MIC 32 mg/L and cefepime MIC and meropenem MIC 64 mg/L. (i) A. baumannii H104940508 with OXA-23 carbapenemase with cefepime MIC >256 mg/L, zidebactam MIC >32 mg/L and meropenem MIC 32 mg/L. FEP, cefepime; ZID, zidebactam.

Both the zidebactam-susceptible (H113840625 MIC 0.25 mg/L; Figure 2a) and, more surprisingly, the zidebactam-resistant (H113980340, MIC >32 mg/L; Figure 2b) NDM-positive K. pneumoniae were killed by zidebactam at 4 mg/L, though the extent of killing was reduced for the resistant organism (1.5 log maximum after 4 h exposure versus 3 log). The cefepime/zidebactam combinations (1 + 4 and 8 + 4 mg/L) achieved 3–4 log kills for both organisms and it is notable that the zidebactam-resistant K. pneumoniae H113980340 was likewise susceptible to cefepime/zidebactam combinations in MIC tests (Table 4). For the two NDM-positive E. coli [H131020913, zidebactam MIC 0.25 (Figure 2c) and H130680324 MIC 16 mg/L (Figure 2d)], killing simply tracked MICs. Thus, for the zidebactam-susceptible organism, zidebactam and its cefepime combinations achieved extensive killing, whereas, for the resistant strain, neither zidebactam nor its combinations achieved a significant kill at the concentrations studied. Corresponding with this result, and unlike for K. pneumoniae H113840625, there was no hint of an enhancer effect for cefepime/zidebactam in MIC combination studies for this E. coli strain (Table 4).

Zidebactam MICs were 8 and 32 mg/L for the two NDM-positive P. aeruginosa strains [H130680310 (Figure 2e and g) and H131800691 (Figure 2f and h), respectively]. There was some suppression of growth for the more susceptible strain with zidebactam alone at 8 mg/L or cefepime/zidebactam at 16 + 8 mg/L, whereas the more resistant strain was unaffected. A 2–4 log bactericidal effect was achieved within 8 h for both strains with cefepime/zidebactam at higher concentrations (Figure 2g and h), though only once the zidebactam was present at MIC (32 mg/L). The A. baumannii strain (H104940508) with the OXA-23 enzyme, was highly resistant to zidebactam (MIC >32 mg/L); cefepime/zidebactam at 8 + 4 mg/L had little effect, but cefepime/zidebactam at 16 + 8 mg/L did achieve bacteriostasis, a result in keeping with the MIC of 32 + 8 mg/L.

In most cases where cefepime/zidebactam achieved substantial killing there was overnight regrowth. Nevertheless, where examined, the organisms remained susceptible in repeat MIC tests with cefepime/zidebactam and did not represent resistant mutants.

Discussion

Zidebactam represents a second DBO with multiple activities, acting not only as a β-lactamase inhibitor but also as a direct antibacterial and exerting an enhancer effect with PBP3-targeting β-lactams. Key differences from RG6080 are that (i) the MICs of zidebactam for susceptible Enterobacteriaceae are lower, typically falling into the 0.12–0.5 mg/L range rather than 1–4 mg/L, and (ii) zidebactam alone inhibited many P. aeruginosa at 4–8 mg/L, whereas MICs of OP0595/RG6080 are consistently >32 mg/L for this species. Proteeae, most Serratia, A. baumannii and S. maltophilia remained resistant, exactly as with RG6080. The antibacterial activity of zidebactam is believed to depend on binding to PBP2, as with RG6080;¹² it is uncertain if the lower MICs of zidebactam reflect increased target affinity, a more favourable balance of permeation and efflux, or combination of all the three, or other factors. Raised zidebactam MICs (typically 16–32 mg/L versus 4–8 mg/L) for P. aeruginosa were associated with strains known to have up-regulated efflux, indicating that the molecule does not entirely evade this mechanism. Otherwise, however, no association was seen between the MICs of zidebactam and the resistance mechanisms for which the isolates were selected. This is in keeping with experience that raised MICs of RG6080 were associated primarily not with ‘conventional’ β-lactam resistance mechanisms, but with mutations that activate the stringent response, thereby compensating for inactivation of PBP2.¹³ Similar types of mutation can confer resistance to mecillinam, which also targets PBP2.¹⁴ The fact that PBP2 itself remains unaltered means that the enhancer effect can remain even when the antibacterial activity has been lost.¹⁵

Despite its low MICs, zidebactam is better suited for development in combination than as a single agent, owing (again like OP0595/RG6080) to a high frequency of mutational resistance (M. Patel, Wockhardt, personal communication). Cefepime has been chosen as a partner agent, based on: (i) its broad spectrum and good safety record; (ii) its wide range of licensed indications; (iii) its relative stability to AmpC enzymes, which can mutate to resist DBO inhibition;¹⁶ and (iv) an enhancer effect being most likely with agents, such as cefepime, that target PBP3.⁴ Even at 1 + 1 mg/L (i.e. below any likely breakpoint for a high dosage formulation), cefepime/zidebactam was active against almost all Enterobacteriaceae with AmpC, ESBL, K1, OXA-48 and KPC β-lactamases and the great majority (29 of 35) of those with MBLs. Even when zidebactam itself lacked activity, the combination retained activity against Enterobacteriaceae with class A and C β-lactamases, which is in keeping with kinetic data showing that zidebactam inhibits these enzymes.¹⁷ Activity was also retained against both zidebactam-resistant klebsiellas with OXA-48 carbapenemase, though—given cefepime’s stability to OXA-48¹⁸—it is most likely that this result reflected inhibition of co-produced ESBLs rather than of OXA-48 itself. Combination activity was more variable against the small number of zidebactam-resistant Enterobacteriaceae with MBLs, but the observation of strong synergy between cefepime and zidebactam for several of these organisms, notably K. pneumoniae H113980340, P. stuartii H124880510 and M. morganii H092540314 supports the view of an enhancer effect and/or the inhibition of co-produced ESBLs. Potentiation against S. maltophilia was widespread and may reflect either an enhancer effect or, more probably, inhibition of the L-2 cephalosporinase, which confers resistance to cefepime.¹⁹

The killing curves, done with pairs of NDM-carbapenemase-positive zidebactam-susceptible and -resistant E. coli, K. pneumoniae and P. aeruginosa largely supported the MIC data with the notable exceptions that zidebactam achieved some killing of the ‘zidebactam-resistant’ K. pneumoniae strain H113980340. Moreover cefepime/zidebactam achieved equally extensive killing of this strain as of its zidebactam-susceptible counterpart (H113840625), whereas there was minimal killing of the NDM-positive zidebactam-resistant E. coli H130480324 by cefepime/zidebactam This variability recapitulates that seen in MIC studies here and previously with OP0595-resistant strains and mutants;^5,7 though it should be added that zidebactam-resistance (Table 2) and the lack of an enhancer effect (M. Patel, Wockhardt, personal communication) seem exceptional in E. coli. Such variation may reflect the diversity of different mutations that can underlie resistance to PBP2-targeted DBOs, though precise relationships remain uncertain.

In conclusion, these findings further illustrate the expanding potential of the DBO class. The first member of the class to enter clinical use, avibactam, has been successfully used, combined with ceftazidime, for infections due to Gram-negative bacteria with KPC carbapenemases,²⁰ though these were poorly represented in Phase III trials. Zidebactam and RG6080 extend this potential by adding direct antibacterial activity and an enhancer effect, contingent on binding to PBP2, with zidebactam having lower MICs for Enterobacteriaceae and P. aeruginosa than RG6080. The result is that β-lactam combinations based on these DBOs have an in-vitro spectrum that includes many MBL-producing Enterobacteriaceae—with 80% of these organisms susceptible at 1 + 1 mg/L in the case of cefepime/zidebactam. Even MBL-producing P. aeruginosa were mostly susceptible to cefepime/zidebactam at 8 + 8 mg/L, though MICs for A. baumannii with OXA carbapenemases were higher. Only clinical trials and experience will reveal the extent to which these additional potentials are realized and, until then, some uncertainty will remain about the risk for selection of resistance to the antibacterial effect of these DBOs and strain-to-strain variability in the enhancer effect.

Acknowledgements

These data were presented, in part, at Microbe 2016, Boston, MA, USA (Abstract Sun-439).

We are grateful to Wockhardt Ltd for financial support of these studies. We also are grateful to Drs Mahesh Patel and Sachin Bhagwat of Wockhardt for many helpful discussions and debates.

Funding

Wockhardt Ltd provided financial support for these studies.

Transparency declarations

D. M. L.: advisory boards or ad-hoc consultancy (Accelerate, Achaogen, Adenium, Allecra, AstraZeneca, Auspherix, Basilea, BioVersys, Centauri, Discuva, Meiji, Pfizer, Roche, Shionogi, Tetraphase, VenatoRx, Wockhardt, Zambon and Zealand); paid lectures (AstraZeneca, Cepheid, Merck and Nordic); and relevant shareholdings in Dechra, GSK, Merck, Perkin Elmer and Pfizer amounting to <10% of portfolio value. All other authors: no personal interests to declare; however, PHE’s AMRHAI Reference Unit has received financial support for conference attendance, lectures, research projects or contracted evaluations from numerous sources, including: Achaogen Inc., Allecra Antiinfectives GmbH, Amplex, AstraZeneca UK Ltd, Becton Dickinson Diagnostics, BSAC, Cepheid, Check-Points B.V., Cubist Pharmaceuticals, Department of Health, Enigma Diagnostics, Food Standards Agency, GlaxoSmithKline Services Ltd, Henry Stewart Talks, IHMA Ltd, Merck Sharpe & Dohme Corp., Meiji Seika Kiasya Ltd, Momentum Biosciences Ltd, Nordic Pharma Ltd, Norgine Pharmaceuticals, Rempex Pharmaceuticals Ltd, Rokitan Ltd, Smith & Nephew UK Ltd, Trius Therapeutics, VenatoRx and Wockhardt Ltd.