Effect of sub-inhibitory dose of cefotaxime on multidrug resistant Staphylococcus haemolyticus isolates

Critical care of neonates involves substantial usage of antibiotics and exposure to multi-drug resistant (MDR) nosocomial pathogens. These pathogens are often exposed to sub-MIC doses of antibiotics which might result in a range of physiological effects. There-fore, to understand the outcome of sub-inhibitory dosage of antibiotics on Staphylococcus populations, nasal swab specimens were collected from 34 neonates admitted to the Sick Newborn Care Unit between 2017-2018, a total of 41 non-repetitive isolates were included in this study. Staphylococcus haemolyticus was the prevalent species (58.54%) with high nonsusceptibility to cefotaxime (CTX) (79.16%), gentamicin (87.50%), and meropenem (54.17%). Biofilm forming abilities of S. haemolyticus isolates in the presence of sub-optimal CTX (30μg/mL), the predominantly prescribed β-lactam antibiotic, were then determined by crystal violet assays and extracellular DNA (eDNA) quantitation. CTX was found to significantly enhance biofilm production among the non-susceptible isolates (p-valueWilcoxin test0.000008) with increase in eDNA levels (p-valueWilcoxin test0.000004). Additionally, no changes in non-susceptibility were observed among populations of two MDR isolates, JNM56C1 and JNM60C2 after >500 generations of growth in the absence of antibiotic selection in vitro. These findings demonstrate that sub-MIC concentration of CTX induces biofilm formation and short-term non-exposure to antibiotics does not alter non-susceptibility among S. haemolyticus isolates.


Introduction
Across the world, approximately 2.4 million children lose lives in the first month of birth each year and India contributes majorly to it. India has a neonate mortality rate of 21.7% [1] and 20-41% of preterm neonates admitted to tertiary care hospitals succumb to sepsis [2] often ascribed to multidrug resistant (MDR) bacteria [3]. Among the causal pathogens, members of the Enterobacteriaceae and Moraxellaceae families dominate, inspite of that, Coagulase-negative staphylococci (CoNS) have been found to contribute significantly [4,5].
When bacterial populations are exposed to antibiotics, resistance often emerges quickly due to mutations or horizontal gene transfer [6]. Additionally, CoNS are also known for causing device-related infections by forming biofilms [7]. Biofilms, in general, have been reported to have considerably higher resistance to antimicrobial agents as compared to planktonic cells [8] and several studies have found that sub-optimal doses of antibiotics can induce biofilm development in a wide range of species [9,10,11].
The Sick Newborn Care Units (SNCU) are known for considerable use of antibiotics especially third generation cephalosporins, aminoglycosides and penems to a certain extent [12]. This is one of the contributing factors resulting in selection of resistant nosocomial bacteria [13]. In addition, empiric use of the same antibiotics against non-susceptible pathogens among neonates requiring longer duration of stay due to prematurity, might play a vital role in enhancing biofilm-formation. Further, it is also unclear as to how many generations would it take after discontinuation of an antibiotic for a resistant bacterial population to become susceptible again [14].
To address each of these questions, we focused on colonization. Nasally colonizing Staphylococcus spp. among the admitted neonates were characterized to identify the predominant nosocomial multidrug resistant species. The next set of experiments were carried out to understand if sub-optimal dosage of the most prescribed group of antibiotics, namely, cefotaxime, a third-generation cephalosporin, results in enhanced biofilm formation in the laboratory. Finally, short-term 15-day evolution experiments were carried out to evaluate if discontinuation of the antibiotic could result in loss of non-susceptibility.

Species identification and susceptibility profiling
Among the 34 neonates, the average number of days spent in the SNCU was 7.1.
Respiratory distress after birth was the most common clinical symptom (n=17; 50%) in the cohort and only 3 preterm along with 4 term born neonates had higher than normal levels of were non-susceptible to AK. Out of 5 CTX sensitive isolates all were susceptible to AK and MRP and 3 were susceptible to GEN ( Figure 1B, Supplementary table 2). All the other Staphylococcus spp. isolates (n=17) were susceptible to AK and MRP, 5 were nonsusceptible to GEN and 3 to CTX. 13 (54.17%) of S. haemolyticus isolates harbored both the mecA and blaZ genes and another 8 (33.33%) had only the mecA gene. Two CTX susceptible isolates with MIC of 8 µg/mL (JNM17C1 and JNM51C1) were also found to harbor the resistance genes ( Figure 1C).

Biofilm enhancement and eDNA release among S. haemolyticus isolates
All except for one S. haemolyticus isolate (JNM50C1a), grown in TSBglu and TSBNaCl were identified to be biofilm producers after 24 hours. The growth media strongly influenced biofilm formation and a total of 14 out of 24 (58.33%) isolates produced biofilms in TSBglu whereas, 22 (91.67%) isolates formed biofilms in TSBNaCl. In the presence of sub-inhibitory concentration of CTX (30μg/mL), biofilm production was enhanced significantly (p-valueWilcoxin test-0.000008) among all the non-susceptible isolates compared to the controls. Given that eDNA release has been implicated in S. haemolyticus biofilm formation, whether increased biofilm formation also resulted in an increase in quantities of eDNA in the biofilms was next evaluated. Biofilm forming sessile cells were harvested and removed by centrifugation and the cell-free nucleic acid in the supernatant was harvested and resolved on 0.8% agarose gels ( Figure 2B). Significant increase in eDNA levels were observed in the TSBNaCl (p-valueWilcoxin test-0.000004) and TSBCTX (p-valueWilcoxin test-0.000004) treated groups (Supplementary table 4).

Whole genome sequencing and resistome mapping of ancestral populations
To generate the resistome profiles of two MDR S. haemolyticus ancestral populations (JNM56C1 and JNM60C2), both non-susceptible to CTX, GEN and MRP, paired-end whole genome sequencing was carried out. Iterative de novo and reference guided assembly (NC_007168) resulted in alignment of >93% of error-corrected reads. JNM56C1 was determined to be a multi locus sequence type (ST) 38 with a chromosome length of 25,54,979 bp (CP063753) at an average sequencing depth of 310. A total of 6 antimicrobial resistance genes were annotated namely, AAC(6')-Ie-APH(2'')-Ia, blaZ, dfrG, mecA, msrA and mphC.

No change in susceptibility in the absence of antibiotic selection
It has been shown in S. aureus strains using in vitro curing assays that harboring SCCmec imposes a fitness cost [22]. However, if a similar cost is attached with nonsusceptibility in S. haemolyticus, a decay of non-susceptibility in the absence of antibiotic selection needs to be tested. To understand this, JNM56C1 and JNM60C2, two completely

Discussion
Increase in the emergence of MDR pathogens is the major cause of mortality in neonatal sepsis due to the ineffectiveness of first line of antibiotics recommended by the World Health Organization [23]. Neonates admitted to the SNCU, or Newborn Intensive Care Units (NICU) are prone to anticipatory broad spectrum antibiotic therapies [24], possess an immature immune system [25] and have been observed to be colonized by pathogens [26,27] which make them extremely susceptible to infections. Hence, this study was undertaken to understand the role of antibiotic use and disuse on multi-resistant Staphylococcus spp. that frequently colonize neonate nares in an SNCU with high usage of β-lactam antibiotics. MDR S. haemolyticus, an emerging pathogen known for multi-resistance [28,29] was observed to be the most frequent colonizer with 79.16% non-susceptibility to CTX and >256 µg/mL MIC value in each case.
To understand if an in vitro equivalent of the therapeutic dose of CTX (30 µg/mL) [18] could affect the physiology of high MIC resistant colonizers, we chose to delineate biofilm formation which is known to accentuate antibiotic resistance and colonization. S. haemolyticus lacking the PIA operon has been shown to form biofilms in vitro under different growth conditions [17,30,31,32] and the same were used to identify the biofilm-formers. A total of 23 out of 24 isolates were identified to be biofilm formers inspite of the absence of PIA. However, we found TSBNacl to be a better media for in vitro production compared to TSBglu, in contrast to the previous studies [17,31]. Escalation of biofilm formation in the presence of insufficient antibiotics has been observed in a variety of species through a variety of mechanisms [9,10,11], inspite of that, not much work has been done to understand biofilm development in S. haemolyticus and there exists only a single report where nosocomial clones were found to show enhanced biofilm formation on glass and polystyrene surfaces in the presence of ¼ MIC of three antibiotics namely, oxacillin, vancomycin, and linezolid [32].Therefore, this study is one of the first showing increase in biofilm formation and eDNA release among MDR, nosocomial S. haemolyticus isolates in the presence of sub-inhibitory concentration of CTX.
Evolved antibiotic resistance is a costly affair for many species of bacteria and often result in decreased competitive fitness in the absence of selection pressure [33]. Nevertheless, studies have also exhibited that often bacteria can tackle such trade-offs by way of mutations that compensate the changes [34] and in a long-term evolution experiment comprising of an ancestral streptomycin resistant Escherichia coli population by Lamrabet et al, it was clearly demonstrated that after 50, 000 generations of growth in the absence of antibiotic there was no change in susceptibility [35]. Similarly, in our short-term evolution experiments (>500 generations) we observed no changes in the MIC values of CTX and MRP in the evolved clones as compared to both the ancestral populations and minor changes in 1 clone per population in case AK and GEN. However, we believe that a better approach would have been whole genome sequencing of the evolved clones instead of a candidate gene approach to gauge the loss of resistance genes if any in these experiments.

Bacterial DNA isolation and species identification
Genomic DNA was isolated from 2 subcultured isolates per specimen using the QIAamp DNA extraction mini kit (Qiagen, Hilden, Germany). Amplification and Sanger sequencing of 16SrRNAgene were carried out using the S-D-Bact-0008-c-S-20/ S-D-Bact-1391-a-A-17 primer pairs [15] (Eurofins Scientific, Bengaluru, India) to identify the species. One colony per subject was included in the study unless different species were identified.

Quantification of biofilms
The biofilm forming ability of Staphylococcus isolates was determined by a modified crystal violet assay method as described previously [17]. Briefly, 96well polystyrene, flatbottom microtiter plates were filled with 180μL of tryptic soy broth (TSB) (Himedia labs) and 20μL bacterial cells grown to a Macfarland score of 0.5 in brain heart infusion broth (BHI) (Himedia labs) were added and incubated at 37 °C for 24 hours (hrs) statically. Biofilm forming capacities of all isolates were determined in TSB, TSB with 1% glucose (TSBglu), TSB with 3% NaCl (TSBNaCl), and TSB with 30μg/mL CTX (TSBCTX) which is an in vitro equivalent of the therapeutic dosage [18]. After 24 hrs, planktonic cells were removed, adherent cells were fixed with 99% methanol (Finar chemicals, Ahmedabad, India) for 10 mins, and plates were washed once with 1× PBS (Sigma-Aldrich,St. Louis,United States) and air-dried for 10 mins. Modified crystal violet assays were performed, and the absorbance

Genomic DNA isolation and whole-genome sequencing
Two isolates (JNM56C1 and JNM60C2) were subcultured in the presence of 30μg/mL CTX. Total DNA from both were purified using the QIAamp DNA extraction mini kit (Qiagen, Germany) and were subjected to paired end whole genome sequencing (2 ×300bp) on an Illumina HiSeq2500 platform (Illumina, San Diego, USA). Both de novo and reference guided assembly was carried out using Velvet and Bowtie2 respectively [19,20] to build genomes as described previously [21].

Short-term evolution experiment
The isolates grown in the presence of 30μg/mL CTX formed the ancestral population for each. Generation time for both were calculated from growth curves. The two populations were serially passaged for >500 generations in triplicate for 15 days in fresh Luria-Bertani

Statistical analyses
Normality across datasets was evaluated using the Kolmogorov-Smirnov test. To identify significant differences among experimental conditions tested, the Wilcoxon test was performed using GraphPad Prism version 9.1.2 (GraphPad Software, La Jolla, California, USA). A p-value of<0.05 was considered to be statistically significant.

Conclusions
To summarize, this study highlights the high nasal carriage rates of MDR S.
haemolyticus isolates among neonates admitted to the SNCU, demonstrates that these isolates have a tendency towards enhanced biofilm formation upon exposure to insufficient quantities of CTX and further reveals no change in susceptibility to both commonly used β-lactams and aminoglycosides when grown in the absence of selection pressure in the short-term. All these findings further reinforce the idea that a definite means of reducing antibiotic resistance is antibiotic stewardship.
Author Contributions: MC ran all the laboratory experiments, analyzed the data and wrote the first draft of the manuscript. TB ran the initial screening experiments and standardized the biofilm assays. MB was the clinician who was involved in specimen collection. BB conceptualized all the experiments, carried out all the next generation sequencing analyses, collated all the data and wrote and reviewed the final draft.