The Effect of Inoculum Size on Antimicrobial Susceptibility Testing of Mycobacterium tuberculosis

ABSTRACT Phenotypic drug susceptibility testing (DST) requires a standardized amount of inoculum to produce reproducible susceptibility results. The most critical step in the application of DST in Mycobacterium tuberculosis isolates is the preparation of the bacterial inoculum. In this study, the effect of bacterial inoculum prepared in various McFarland turbidities on primary antituberculosis drug susceptibility of M. tuberculosis strains was investigated. Five standard ATCC strains (ATCC 27294 [H37Rv], ATCC 35822 [izoniazid-resistant], ATCC 35838 [rifampicin-resistant], ATCC 35820 [streptomycin-resistant], ATCC 35837 [ethambutol-resistant]) were tested. Inoculums of McFarland standard of 0.5, 1, 2, 3, and 1:100 dilutions of 1 McFarland standard of each strain were used. The effect of inoculum size on DST results was determined by the proportion method in Lowenstein-Jensen (LJ) medium and nitrate reductase assay (NRA) in the LJ medium. In both test methods, the increase in inoculum size did not affect the DST results of the strains. On the contrary, DST results were obtained more rapidly as a result of the use of dense inoculum. DST results obtained in all McFarland turbidities were found to be 100% compatible with the recommended amount of inoculum, 1:100 dilution of 1 McFarland standard (inoculum size of gold standard method). In conclusion, the use of a high amount of inoculum did not change the drug susceptibility profile of tuberculosis bacilli. Minimizing manipulations during the inoculum preparation phase of susceptibility testing, this outcome will decrease the need for equipment and make the test application easier, particularly in developing countries. IMPORTANCE During DST application, it can be challenging to evenly homogenize TB cell clumps with lipid-rich cell walls. These experiments must be carried out under Biosafety Level-3 (BSL-3) laboratory conditions with personal protective equipment and taking safety precautions because the procedures applied at this stage cause the formation of bacillus-laden aerosols and carry a serious risk of transmission. Considering this situation, this stage is important given that it is not possible to establish a BSL-3 laboratory in poor and developing countries. Reducing the manipulations to be applied during the preparation of bacterial turbidity will minimize the risk of aerosol formation. Perhaps there will be no need to do these steps for susceptibility tests in these countries or even in developed countries.

most critical step in the application of DST in M. tuberculosis isolates is the preparation of mycobacterial inoculum. Clumping of TB bacilli causes uncertainty in the bacillus distribution in the inoculum and the number of bacilli per milliliter. This is a critical step in terms of test reproducibility (4). Phenotypic DST requires a standardized amount of inoculum to produce reproducible susceptibility results. In the documents published by the Clinical and Laboratory Standards Institute (CLSI) in TB susceptibility tests, it is recommended to use the inoculum between McFarland standard of 0.5 to 1 (5)(6)(7). In Etest studies, it is recommended to use a McFarland standard of $3 bacterial density (8). During DST application, it can be challenging to evenly homogenize TB cell clumps with lipid-rich cell walls. These experiments must be carried out under Biosafety Level-3 (BSL-3) laboratory conditions with personal protective equipment and taking safety precautions because the procedures applied at this stage cause the formation of bacillus-laden aerosols and carry a serious risk of transmission (9,10). Considering this situation, this stage is important given that it is not possible to establish a BSL-3 laboratory in poor and developing countries.
The number, distribution, and viability of organisms in the inoculum have a significant impact on DST results. To obtain reliable DST results, the inoculum preparation must be properly standardized (11). However, there are not enough studies on the effect of inoculum size on the drug susceptibility of M. tuberculosis strains. It has been previously shown that the inoculum size of 0.1 to 3 mg does not affect the streptomycin (STR) susceptibility of M. tuberculosis. However, experimental study data are insufficient (12).
However, resistance in M. tuberculosis isolates is usually due to spontaneous mutations. Therefore, the denser the amount of inoculum, the easier it will be to detect resistant isolates in susceptibility testing (13). For this purpose, the effects of bacterial inoculums with various McFarland turbidities on susceptibility tests were investigated in Löwenstein-Jensen (LJ) medium with standard ATCC strains using both the proportion method and nitrate reductase assay (NRA).

RESULTS
In the study, the tests for all bacterial densities were completed on the 21st day for the proportion method in the LJ medium. DST results with inoculum sizes used for all strains were 100% consistent with the recommended inoculum size (1:100 dilution of 1 McFarland standard inoculum) ( Table 1). On the other hand, microcolonies were observed in 2 and 3 McFarland standard inoculums in LJ medium containing EMB in the ATCC-STR strain set. However, the number of these microcolonies was less than 1% of the number of colonies in the control tubes. Therefore, the categorical agreement of the strain remained the same. In the NRA, susceptibility results for all inoculum sizes, including the recommended inoculum size, were 100% consistent. DST results obtained with NRA are shown in Table 2 (Table 3).

DISCUSSION
Both World Health Organization (WHO) and CLSI, recommend 1:100 and 1:10,000 dilutions from 1 McFarland standard (CLSI previously recommended using McFarland standard of 0.5) as the amount of inoculum in susceptibility methods performed on both LJ medium and Middlebrook 7H10-11 agar (5-7). However, in the Etest susceptibility method, $3 McFarland standard inoculum is recommended (8).
There are also phenotypic colorimetric susceptibility testing methods. In these methods, there are differences between bacterial inoculums. In the NRA test performed on LJ medium, 0.2 mL of 1 McFarland standard turbidity is directly inoculated on antibiotic media. Also, 0.2 mL from 1:10 dilution of 1 McFarland standard was inoculated on growth control media (14,15). In the NRA test performed directly from the clinical sample, the same dilution and amount of inoculum are used. (16,17). In these studies, the number of bacteria inoculated on antibiotic media was used higher than in the control medium. When the NRA test is performed on microplates, a 1:10 dilution of 1 McFarland turbidity is used, and 100 mL was inoculated in each well of 96 microplates (18). In the Resazurin Microtiter Assay (REMA), 1:20 and 1:10 dilutions of 1 McFarland standard are used and 100 mL was inoculated in each well of 96 microplates (19,20). In malachite green microtube assay (MGMT), 1:5 dilution of 1 McFarland standard (21, 22) and 1:10 dilution of 1 McFarland standard in crystal violet decolorization assay (CVDA) is used (23). As in these studies, there are differences between bacterial inoculum in the used methods.
Resistance in M. tuberculosis isolates depends on the presence of mutants. The average mutation rate for resistance to INH, RIF, EMB, and STR is 2.56 Â 10 28 , 2.25 Â 10 210 , 1 Â 10 27 , and 2.95 Â 10 28 , respectively (24). Therefore, the use of high-density inoculum may facilitate the detection of these mutants. However, the bacterial density used in the study was approximately 10 6 26 Â 10 6 CFU/mL (McFarland no 0.5-3). These mutant isolates could not be detected because the tested bacterial density was far below the rate of resistant bacteria due to mutation. In other words, the number of bacterial inoculums to show false-drug resistance due to these mutants could not be reached.
Because TB bacilli tend to cluster together, it is difficult to count bacteria in liquid culture (25). In the study, the bacterial concentration in the inocula was determined as done   (25). Colony counting on agar plates is often used to determine the number of cells in a bacterial suspension. This requires an incubation period of at least 3 weeks. Although this method is time-consuming, it is not easily applicable to TB bacillus due to the uncertain growth of single cells. It is also difficult to make direct microscopic counts due to the clustering of TB bacilli and the inability to clearly distinguish between dead and live bacilli (26,27). Considering these disadvantages, in the 1950s, bacillus weight per unit volume or nitrogen weight per unit volume were often used to determine the amount of TB bacillus (27). The relationship between the wet weight of the bacillus in the TB bacillus suspension and the turbidity of that suspension has been explained before and has allowed us to compare our study results with the limited existing studies in the literature. It has been reported that the weight of 1 mg of dry   TB bacillus is equivalent to approximately 3.10 8 TB bacillus with repetitions performed 6 times and considering that the bacilli contain 85% water, 1 mg of wet TB bacillus is equivalent to approximately 4.10 8 bacilli (28). It was shown by Youmans (29) that inoculum size between 0.1 and 1.0 mg did not affect the bacteriostatic activity of STR for TB bacillus. Additionally, 4 years later, Youmans and Williston together reported that 0.01 to 0.25 mg inoculum size did not affect the STR susceptibility test results (30). Previous studies were supported by showing that inoculum size between 0.1 and 3.0 mg, which is a larger inoculum range, does not affect the STR sensitivity of TB bacilli (12). Contrary to STR, it has been reported that there is an inverse relationship between the bacteriostatic effect of para-aminosalicylic acid (PAS) and the inoculum density of TB bacillus. Thus, drug susceptibility is affected by the number of bacteria in the inoculum (31). In addition, it has been reported that the size of the inoculum, which contains 2.5 to 3.7 mg of wet TB bacillus and its 10 21 to 10 27 dilutions, affects the apparent sensitivities of STR, INH, and PAS (32). In summary, studies have shown that the STR susceptibility results of a minimum of 0.01 mg and a maximum of 3 mg of wet TB bacillus are the same and are not affected by the amount of inoculum.
Only STR, PAS, and INH were known as anti-TB drugs in the years 1945 to 1955 when the effect of the amount of inoculum on drug susceptibility results of TB bacilli was investigated. Other anti-TB drugs have not yet been discovered and their activities on TB have not been determined (33). Was the "inoculum effect," a widely accepted laboratory phenomenon (34), also seen between primary anti-TB drugs and M. tuberculosis strains? Did the increase in inoculum density cause a significant increase in the MIC values of antibiotics? The results of our study based on these questions also show that there was no change in the DST results of TB bacilli from the recommended standard inoculum amount to 3 McFarland turbidity. However, although the generation time of bacilli was not affected as a result of using dense inoculum, DST results were obtained faster. Presumably, this allows earlier observation of bacterial communities formed due to the high bacterial count. To talk about the phenomenon of the "inoculum effect," obviously higher inoculum amounts should be used. Jung et al. claimed that the problems experienced during the homogenization of TB bacillus clusters and the problems caused by unequal fragmentation will be reflected in the DST results, cause an increase in MICs, and therefore lead to the inoculum effect (35). However, no such effect was found in our study. This may be because MIC was not investigated in our study and only one critical concentration was used. Jung et al. (35) stated that there may be an increase in MIC values. However, if this increase does not affect the categorical group of susceptible or resistant strains, they cannot be determined by the proportion method.
The formation of microcolonies is another important result of our study. The formation of microcolonies already in the presence of EMB has been reported previously (36). However, it is difficult to decide whether these microcolonies represent resistant or partially resistant mycobacteria or an overgrowth of susceptible organisms following drug degradation (5). Because LJ medium is an egg-based medium, it must be heated at a high temperature to solidify. This situation may cause degradation and decrease in stability of antibiotics added to LJ medium for proportion method. Therefore, the critical concentrations of drugs vary in different media using the same method. For example, RIF is tested with the proportion method at a concentration of 1.0 mg/mL in 7H10 agar medium, while it is tested at a concentration of 40 mg/mL in LJ medium. Likewise, while STR is tested at a concentration of 2 mg/mL in 7H10 agar medium, it is tested at a concentration of 4 mg/mL in LJ medium. However, the same is not true for EMB. While it is used at a concentration of 5 mg/mL in 7H10 Agar medium, it is used at a concentration of 2 mg/mL in LJ medium. Inconsistent DST results may occur following the degradation of EMB, depending on factors such as the heating time and heating temperature of the LJ medium, as well as the difference between their critical concentrations. In this case, it can result in the formation of microcolonies in the presence of EMB. It has already been reported by the World Health Organization (WHO) that the phenotypic DST for EMB is not reliable and reproducible, with inconsistent drug susceptibility results (3).
Because resistance in M. tuberculosis isolates usually occurs due to spontaneous mutations, the higher the amount of inoculum, the easier it will be to detect resistant isolates in the susceptibility test. Furthermore, because the increase in bacterial density (high McFarland values as used in the Etest method) does not change the drug susceptibility profile of TB bacilli, the procedures to be applied during the preparation of the inoculum will be reduced and both the laboratory workload and the risk of transmission will be minimized.
As a result, TB is more common, especially in poor and developing countries and it is not possible to provide BSL-3 laboratory conditions in these countries. Reducing the manipulations to be applied during the preparation of bacterial turbidity will minimize the risk of aerosol formation. Perhaps there will be no need to do these steps for susceptibility tests in these countries or even in developed countries. Further studies are needed on this subject by increasing the number of clinical isolates. Preparation of bacterial inoculum. During the preparation of bacterial inoculum, all experiments were carried out in a Class II type B microbiological safety cabinet in the BSL-3 laboratory, with personal protective equipment (3M Versaflo, TR-3001) taking safety precautions. Inoculums of bacterial strains were prepared using fresh cultures grown on LJ medium. After transferring colonies from fresh cultures to tubes containing eight to 10 sterile glass beads and 3 to 5 mL of saline, the tubes were tightly closed and wrapped with parafilm. Afterward, it was mixed at 2,500 rpm with a vortex (Onilab MX-S, City of industry, CA 91748 USA) for about 1 min, then the tubes were kept in an upright position at room temperature for about 30 to 60 min to precipitate aerosols and large particles. Then, the supernatant was transferred to another tube and the separated supernatant was adjusted to McFarland standard of 0.5, 1, 2, and 3 for each strain by the densitometer (BIOSAN Medical-Biological Research & Technologies, Riga, Latvia). In addition, for the recommend standard inoculum size, the McFarland standard of 1 was used to prepare 1:100 dilution (3).

MATERIALS AND METHODS
Preparation of antibiotic solutions. In this study, primary antibiotics INH, RIF, STR, and EMB were tested. Powder forms of all antibiotics (Sigma-Aldrich) were used. INH, STR, and EMB were dissolved in sterile distilled water and RIF was dissolved in methanol. Stocks were prepared at a final concentration of 4,096 mg/mL for each antibiotic. Antibiotics were stored in aliquots at 280°C until use.
Preparation of LJ medium for the proportion method. LJ medium (Merck, Germany) was prepared according to the manufacturer's recommendations. After adding eggs, antibiotics were added to the medium, with final concentrations of 0.2 mg/mL, 40 mg/mL, 4 mg/mL, and 2 mg/mL for INH, RIF, STR, and EMB, respectively. Then, the 5 mL medium was dispensed into screw cap tubes. Similarly, antibiotic-free growth control tubes were also prepared. The tubes were heated at 85°C for 40 to 50 min in a slanted position. The prepared media were stored at 14°C. The storage period did not exceed 4 weeks (3).
Preparation of LJ medium for nitrate reductase assay. LJ medium was prepared as described above. In addition, KNO 3 was added to the medium at a final concentration of 1,000 mg/mL. Antibiotic concentrations were the same as those used in the proportion method. For each strain, five antibiotic-free growth control tubes were also prepared.
Proportion method on LJ medium. The LJ proportion method is illustrated in Fig. 1. In the study, five sets were prepared for each strain. Each set consisted of a growth control and LJ media containing four primary anti-TB agents. Each set also contained McFarland standards of 0.5, 1, 2, 3, and 1:100 dilution of 1 McFarland standard inoculum. A 1:100 dilution of 1 McFarland standard inoculum was the recommended inoculum size. In addition, 100 mL of prepared inoculums for each strain were inoculated into LJ mediums with and without antibiotics (Fig. 1). LJ tubes were incubated at 37°C for 21 days. For contamination, it was checked every 3 days during the first week, then weekly. If sufficient growth was observed in the growth control tube at the end of the incubation, the test was terminated. The 1% proportion approach is used to evaluate the results. Accordingly, the culture is considered resistant if at least 1% of the population demonstrates resistance. The strain is classified as susceptible if the difference between the number of colonies on the anti-TB agent-containing medium and the number of colonies on the control medium is less than 1% (3).
Nitrate reductase assay on LJ medium. The NRA method is illustrated in Fig. 1. The inoculums prepared in this study were applied simultaneously for the NRA. Additionally, in the NRA, five antibiotic-free growth control tubes were used for each strain. And 100 mL of prepared inoculums for each strain were inoculated into LJ mediums with and without antibiotics (Fig. 1). After inoculation, the tubes were incubated at 37°C for 5, 7, 10, 14, and 21 days. After 5 days of incubation, 500 mL of Griess reagent (one part of 50% [vol/vol] concentrated hydrochloric acid, two parts of 0.2% [wt/vol] sulfanilamide, and two parts of 0.1% [wt/vol] n-1-naphthyl-ethylenediamine dihydrochloride) was added to one of the growth control tubes. If purple/violet color formation was observed in the growth control tube, 500 mL of Griess reagent was added to the antibiotic tubes, the test was terminated and the susceptibility results were evaluated. The susceptibility results were evaluated according to the color transformation in the growth control tube. If a purple/violet color appeared in the growth control tube after the addition of reagent in the tubes containing antibiotics, the isolate was considered resistant to the tested antibiotic. If no color was observed, the isolate was considered susceptible to the tested antibiotic. If purple/violet color formation was not observed in the growth control tube, no reagent was added to the antibiotic tubes, and incubation was continued. The same procedures were performed on the other growth control tubes on the 7th, 10th, 14th, and 21st days of incubation (14).