Evaluation of matrix-assisted laser desorption ionization-time of flight mass spectrometry in identification of methicillin resistance in clinical isolates of Staphylococcus aureus

INTRODUCTION

Antimicrobial resistance is an emerging threat globally.^[1] Early identification of antimicrobial resistance will help clinicians provide more precise antimicrobial therapy to patients. Various phenotypic and genotypic tests are available to identify antimicrobial resistance and thus guide antimicrobial therapy. Each type of available test for detecting antimicrobial resistance has its own set of limitations. For example, the available phenotypic tests (e.g., disk diffusion, epsilometer gradient diffusion, and broth macro-and micro-dilution) used for performing antimicrobial susceptibility testing (AST) are labor-intensive and time-consuming. Moreover, these phenotypic tests are affected by incubation conditions such as temperature, pH, and salt concentration of media, and thus can fail to identify the heterogeneous strains that may be present in clinical isolates.^[2] Similarly, molecular methods such as polymerase chain reaction are also available, but such tests are expensive and also require technical expertise.^[3]

Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) has replaced most of the phenotypic tests for the identification of bacteria, including the fastidious microorganisms. MALDI-TOF MS technology is based on proteomics, i.e., it studies proteins and peptides. The identification of microorganisms is based on the study of highly conserved ribosomal protein sequences. Laser rays are directed at the sample, which leads to the ionization of the ribosomal proteins and peptides present in the sample. The ions are then activated by the electric field, and they traverse through the flight tube under vacuum. The ions are then separated based on their mass-charge (m/z) ratio. The ions having a higher m/z will reach the detector first, followed by the ions with a lower m/z ratio. The time taken for the ions to reach the detector is measured, and it is converted into a mass spectrum. This mass spectrum is plotted as ion abundance (intensity of signal) on the y-axis and m/z value on the x-axis. The generated mass spectrum is then compared with its database. A similar principle is exploited to identify drug resistance. m/z ratio of certain proteins, which may be responsible for any drug resistance mechanism, will also be generated as “peaks” in the mass spectrum. These peaks can then be analysed manually to check for the type of drug resistance associated with a particular isolate.

Owing to its advantages, such as cost-effectiveness, speed, and accuracy of results, and fully automated technology, studies are now being planned to incorporate the use of this technology in the identification of drug resistance in microorganisms.^[4] Moreover, no special sample preparation is needed, and the m/z peaks can be studied from the m/z spectrum formed and stored in the analyzer during the routine step of isolate identification.^[5,6] Furthermore, in MALDI-TOF MS, the cost of reagents is very less, making the per-test and recurrent costs very reasonable.

Furthermore, as discussed above, MALDI-TOF MS is based on the detection of expressed peptides/proteins.^[7] Signals based on mass: Charge (m/z) spectra of organisms differ from each other in terms of intensity of the signal, absence/presence of a particular peak, or shifting of the peak. Loss of a particular peak is due to a lack of expression of a protein. This lack of expression of a particular protein could be because of a particular mutation in the genes. Shift in the peak in signal is related to point mutations in genomes, leading to a change in amino acid, which in turn increases/decreases the mass of the ions, hence the altered m/z spectra.^[8]

The use of the MALDI-TOF system in predicting various antimicrobial resistances in different bacterial isolates is being extensively studied using different methods. The various studies which have used MALDI-TOF to predict antimicrobial resistances have used one of the following methods; (i) either using various software,^[9] (ii) or using different assays in Biflex III (Bruker Daltonik GmbH, Bremen, Germany) platform,^[10] (iii) or by comparing m/z profiles of microorganisms in the presence/absence of particular antimicrobial agent,^[11] or (iv) through machine learning models.^[12]

All the above-mentioned methods either require a Biflex III platform or costly software; thus, it might not be very feasible for the laboratories using a different MALDI-TOF platform (i.e., other than Biflex III), for example, Vitek MS (MALDI-TOF MS Technology, bioMerieux, USA) as used in our study. To the best of our knowledge, this is the first study that visually identifies specific m/z peaks, the presence/absence of which can give a clue about methicillin resistance in Staphylococcus aureus strains on a platform other than Biflex III.

Methicillin-resistant S. aureus (MRSA) is not only difficult to treat but is also associated with higher morbidity and mortality.^[12] To reduce the serious infections caused by MRSA, it is of utmost importance to know the status of methicillin susceptibility of S. aureus isolates as soon as possible.^[7] Resistance to methicillin in S. aureus is linked to production of an altered penicillin binding protein, which imparts resistance against all the beta-lactam antibiotics. The two genes mecA and mecC, which code for these enzymes, are carried on genomic islands known as Staphylococcal Cassette Chromosome mec.^[13] The use of cefoxitin disc as a surrogate marker for identifying methicillin resistance in S. aureus has been proposed by the Clinical and Laboratory Standards Institute (CLSI), M100, 17^th edition, 2007, providing detailed methodology and interpretation based on inhibition zone diameters.^[14]

The present study was planned to identify S. aureus as MRSA or Methicillin Sensitive S. aureus (MSSA) by analyzing their m/z peaks as seen on MALDI-TOF MS while comparing it with the methicillin susceptibility/resistance status by evaluating cefoxitin minimum inhibition concentration (MIC)^[13] using standard methods of Vitek-2 Compact and also by measuring zone diameter around cefoxitin disc using disc diffusion test as proposed by CLSI M100 17^th edition, 2007.^[14]

MATERIALS AND METHODS

This study was done in the department of microbiology of a tertiary care hospital in North India. The research was conducted in accordance with the Declaration of Helsinki. Informed consent was taken from the participants. The study was planned and conducted as follows:

RESULTS

A total of 329 bacterial isolates identified as S. aureus from various samples were received during the study. The sample distribution of these isolates is as follows: 188 pus, 79 blood, 41 urine, 8 BAL, 6 cerebrospinal fluid, 4 tissue, and 3 synovial fluid. Of these, 156 (47.4%) were found to be MSSA, and the remaining 173 (52.6%) were MRSA as identified by the disk diffusion method and then confirmed by Vitek-2 Compact. The status of cefoxitin resistance as seen on disc diffusion AST and Vitek-2 compact of all the isolates was in 100% concordance.

One hundred each of the MRSA and MSSA were randomly selected for further analysis.

The m/z peaks of the selected hundred isolates of MRSA, MSSA, and S. aureus ATCC strains are given in Table 1. The m/z peaks corresponding to the highest signals are only mentioned Table 1.

The m/z peaks of S. aureus 25923 ATCC include: 2153, 3007, 3876, 4490, 4512, 4814, 5033, 5508, 6890, and 9628. The specific m/z peaks in the spectra of MRSA isolates were found at 2417, 4822, and 5545. Specific m/z peaks in spectra of MSSA isolates were found at 2190 and 2229. Some m/z peaks were found in the majority of both MRSA and MSSA isolates. These include 3007, 5033, and 6890. These common peaks were the same as those found in the S. aureus ATCC 25923 strain. Table 1 also shows the distribution of each m/z peak in percentages of MRSA and MSSA isolates. The P-value, as shown in Table 1, shows the level of significance with which different m/z peaks were associated with either MRSA or MSSA isolates.

The mass spectrum of one each of MRSA, MSSA isolate, and control S. aureus ATCC 25923 strain is given as Figures 1-3, respectively.

Close

Figure 1:

Mass spectra (m/z value on x axis and signal intensity on y axis) of one methicillin-resistant Staphylococcus aureus (MRSA) isolate (representative).

Export to PPT

Close

Figure 2:

Mass spectra (m/z value on x axis and signal intensity on y axis) of one methicillin-sensitive Staphylococcus aureus (MSSA) isolate (representative).

Export to PPT

Close

Figure 3:

Mass spectra (m/z value on x axis and signal intensity on y axis) of Staphylococcus aureus American type culture collection (ATCC) 25923.

Export to PPT

The mass spectra of one MRSA and one MSSA representative isolate are shown in Figure 4 for comparison of specific m/z peaks between the two.

Close

Figure 4:

Comparative mass spectra of one representative each of Methicillin sensitive Staphylococcus aureus (MRSA) and Methicillin resistant Staphylococcus aureus (MSSA) isolates comparing m/z peaks between MRSA and MSSA isolates. Image showing comparison between m/z spectra of MRSA and MSSA. (i) Line “a” shows a peak at 2190 present in 87% of MSSA and 14% of the MRSA isolates. (ii) Line “b” shows peak at 2229, present in 81% of MSSA and only 17% of MRSA isolates (iii) Line “c” shows peak at 2417 present specifically in 83% of MRSA and absent in all MSSA isolates (iv) Line “d” shows peak at 4822 present in 100% of MRSA and 13% of MSSA isolates. (v) Line “e” shows peak at 5545 present in 72% of MRSA and 11% of MSSA isolates.

Export to PPT

The sensitivity, specificity, positive predictive value, and negative predictive value (NPV) of different m/z peaks for MRSA and MSSA are given in Table 2.

DISCUSSION

The m/z spectra obtained in the MALDI-TOF MS analyzer during the identification of an isolate are very clear, so that the m/z peaks can be easily visualized. If the m/z peaks that indicate various resistance mechanisms, for example, cefoxitin resistance in S. aureus, are known to us, we can easily search for such peaks in the m/z spectra generated and thus can get an idea about the resistance of the tested isolate. In the present study, the m/z peaks were visualized by two different observers. Both observers gave their independent readings, and the readings obtained from one observer were kept hidden from the other observer. When the inter-observer variability was analyzed, it was found to be nil.

MALDI-TOF MS, apart from bacterial species identification, can be used for identifying resistance in bacterial isolates. There are various tested methods available to identify antimicrobial resistance in Enterobacterales, in various Gram-positive cocci, non-fermenting Gram-negative rods, anaerobes, and Mycobacteria.^[11] All the methods either require technical expertise or expensive software.^[2] In the present study, we visually analyzed the m/z spectra peaks for the possibility of differentiating the methicillin-resistant and sensitive S. aureus.

According to a study done on the Biflex III (Bruker Daltonik GmbH, Bremen, Germany) platform, the predominant peaks of S. aureus were at m/z 6889 and 9627, corresponding to a stress response protein and DNA binding protein, respectively. Similar peaks at m/z 6890 and 9627 are seen in the spectra of S. aureus ATCC 25923 in our study [Table 1 and Figure 3]. Of the three m/z peaks seen commonly in both MSSA and MRSA in our study (3007, 5033, and 6890), two have already been reported as species-specific for S. aureus (m/z 5031 and 6887 with mass difference of m/z ±5).^[8]

m/z at value of 3006 ± 2 has been shown to correspond to the expression of delta toxin peptide, related to agr gene (accessory gene regulator) of Staphylococcus, which is associated with the virulence in S. aureus.^[17] Few studies have shown the m/z peak at 3006 ± 2 to be specific and unique for MRSA strains^[18] whereas few other studies have shown it to be present in S. aureus irrespective of its susceptibility to methicillin.^[12] Even in our study, the m/z value of 3007 was seen in most of the S. aureus isolates, including methicillin-resistant and methicillin-susceptible strains [Table 1 and Figures 1-3].

m/z value of 2417 was seen statistically significant in MRSA isolates (83/100; P < 0.0001) and in none of the MSSA isolates in our study, with specificity of 100% and 83% sensitivity [Table 2 and Figure 1]. Various studies have also found the m/z value of 2415 ± 5 to be significantly associated with the mecA gene in S. aureus, the presence of which denotes methicillin resistance in S. aureus.^[4] A similar result has been seen in other studies. Studies have shown phenol-soluble protein toxin (PSM-mec) to be produced by up to 50% of MRSA isolates. The biological function of PSM-mec is not well known, nor is it known to be associated with methicillin resistance, but it is seen to be produced only by MRSA isolates.^[6]

Various other studies done on identifying MSSA and MRSA have shown peaks in m/z spectra obtained from MALDITOF that can strongly differentiate the above two types of strains, but may not be able to detect all of the MRSA/MSSA. m/z of 5545 as a differentiating marker for MRSA, as seen in our study (89% specificity), has also been reported previously^[19] as a marker of MRSA [Table 1 and Figure 1]. m/z peak of 4822 seen in 100% of the MRSA and 13% of MSSA in our study, with a NPV of 100% and 100% sensitivity (P < 0.0001), has not yet been reported as an identifying peak for MRSA in any of the published literature. More studies with larger sample sizes are needed to support this finding.

Similarly, for MSSA, there are some peaks, the presence of which can give a clue towards the methicillin-susceptible status of the S. aureus isolate. Peak value of 2190 and 2229 was seen significantly (P < 0.0001) in MSSA (87%, 81%, respectively) isolates as opposed to MRSA isolates (14% and 17%, respectively) [Table 1 and Figure 2]. Similar peaks at m/z 2194 and 2232 have been reported to be MSSA-specific markers in a previous publication.^[19]

Variation in the m/z peaks seen in MRSA and MSSA could indicate that different peaks can be seen in MALDI-TOF MS for MRSA and MSSA isolates; thus, a large database is required for the correct identification of the same.^[19] Future multicentric studies on larger numbers of isolates can aid in further enhancing the validity and reproducibility of our results. More importantly, implementing rapid AST will boost the diagnostic stewardship and, in turn, the antimicrobial stewardship program of any health care center. Thus, MALDI-TOF could prove to be a turning point for the critically ill patients.

CONCLUSIONS

The MALDI-TOF MS system is an affordable and technically easy system that has been established for identifying microorganisms using its elaborate database based on m/z ratio. Studies are now being done to exploit this technology for interpreting antimicrobial resistance in different bacterial species. Although there are a few limitations in this methodology, too, since there can be variations in the peaks obtained in m/z spectra associated with resistance to various drugs, more studies with larger cohorts are needed to build a good database. Besides, a bacterium may be multidrug resistant, and when resistance to different drugs is present together, it may alter the m/z spectra differently, thus specific m/z peaks need to be established to identify the specific antimicrobial resistance correctly. The results of the present study are encouraging, and the number of isolates tested is pretty large, but it is a single-center study; thus, before putting these results into clinical practice, we need to test their reproducibility and have multi-centric studies to further validate the results.