This experiment was approved by the institutional review board (IRB) of the Chang Gung Memorial Hospital (CGMH, IRB 98-0675B), and informed consent was obtained from all of the patients who were hospitalized at the 1000-bed Chiayi CGMH in southern Taiwan from 2010 to 2012. In total, 43 cases of bacterial infectious spondylitis were enrolled, and the bacterial species were identified at the Department of Anatomic Pathology and Laboratory Medicine, Chang Gung Memorial Hospital, Chiayi Branch. Among these, six patients with infectious spondylitis infected by S. aureus were retrospectively analyzed for factors including age, gender, infection site; moreover, comorbid underlying chronic conditions, such as diabetes mellitus, hypertension, chronic liver disease, chronic renal insufficiency, chronic obstructive pulmonary disease, and malignancy, were investigated. Additionally, infectious pathogens, empiric antibiotics, the number of operative debridements and reconstructions, the duration of hospitalization, and the in-hospital mortality rate were evaluated.

Six infectious spondylitis-associated S. aureus isolates were identified in the hospital laboratory and the university. The bacteria were first analyzed by coagulase testing and Gram staining. Furthermore, S. aureus was identified by PCR amplification of the S. aureus-specific clfA, 16S rRNA, and nuc genes, as previously described [12] [13] . The genotype of each isolate was determined by SmaI-digested PFGE analysis, according to a previously reported method [14] . Briefly, whole-cell embedded agarose plugs were digested with the restriction endonuclease, SmaI (New England Biolabs, Ipswich, MA, USA). The DNA fragments were resolved by a CHEF DR-III apparatus (Bio-Rad, Hercules, CA, USA). The isolates were defined as a subgenotype for those with ≤ 3 fragment differences and as a genotype for those with more than 3 fragment differences. The MLST types for each isolate were determined according to the method described by Enright et al. [15] and by analysis of MLST databases (http://saureus.mlst.net).

The susceptibility of the six S. aureus isolates to the following antimicrobials was examined: AMP (10 μg), CEF (30 μg), CIP (5 μg), CLI (2 μg), ertapenem (10 μg), ERY (15 μg), IPM (10 μg), MEM (10 μg), OXA (1 μg), oxytetracycline (30 μg), PEN (10 units), TET (30 μg), and SXT (1.25 μg for trimethoprim and 23.75 μg for sulfamethoxazole) (Becton Dickinson, Spark, MD, USA). Susceptibility analysis was performed using the disc diffusion method and the guidelines of Clinical and Laboratory Standards Institute (CLSI) [16] . The S. aureus isolates BCRC10781 and BCRC15211 were used as OX-susceptible and resistant reference strains, respectively.

SCCmec types I-IV were identified by multiplex PCR amplification [17] [18] . If not groupable into types I-IV, isolates were grouped into SCCmec type V or VT (or VII) by the PCR detection of ccrC (ccr5) homologues [19] - [21] . Genes for virulence factors PVL, Hlg, TSST-1, ETA, and ETB were identified by simplex and multiplex PCR amplifications using primers described elsewhere [22] [23] . Accessory gene regulator (agr) typing was also performed, as described previously [24] .

Human alveolar basal epithelial cell line A549 was routinely cultured in RPMI-1640 medium with 10% FBS. After the transfer of 5 × 10⁵ A549 cells into each well of a 24-well plate (Becton, Dickinson and Company), the isolate (1 × 10⁷ cfu) was added to reach MOI = 20. The mixtures were incubated for 1 or 4 h at 37˚C and then washed twice with PBS. The cells were treated with 0.25% trypsin/0.53 mM EDTA for 5 min until cell detachment, and then 800 µl of RPMI-1640 medium with 10% FBS was added. The cells were then pelleted at 3000 rpm for 10 min. After washing with PBS, the cells were stained with propidium iodide (PI) and annexin V using an Annexin V-FITC Apoptosis Detection Kit (Cat. No. AVK250, Strong Biotech Corporation, Taiwan) for 10 - 15 min in the dark. The fluorescence of PI and annexin V in the cells was measured by flow cytometry, and the data were analyzed using WinMDI software.

S. aureus-associated infectious spondylitis was only observed in male patients with an infection in the lumbar spine (five patients) and thoracic spine (one patient) With the exception of one patient without any underlying disease, diabetes mellitus and renal disease were found in four of the patients, followed by hepatitis C infection and cellulitis in three patients, and gastric ulcer, septic shock, and arterial disease in two patients. S. aureus infection was not correlated with the duration of hospitalization, previous amputation or mortality. With a change in the antibiotic use, patient 3, who had eight underlying diseases, died. Patients 2 and 5, who had MRSA infections, received two and four spine debridements and reconstructions, respectively.

Among 43 bacterial infectious spondylitis, M. tuberculosis and S. aureus accounted for 30.2% (13 cases) and 41.9% (19 cases), respectively (Table 1). The remaining species included coagulase-negative Staphylococcus,

A: anterior; P: posterior; L: lumbar; T: thoracic; L + S: lumbar and S segment.

E. corrodens, E. coli, K. pneumonia, Micrococcus spp., Peptostreptococcus spp., Prevotella spp., Proteus mirabilis and S. enterica serogroup D. Anterior operations (72.1%) were more prevalent than posterior operations (23.3%). M. tuberculosis and MRSA caused more anterior operations, while MSSA caused more posterior operations. Infection sites differed among species that mainly infected a single site in the lumbar (58.1%) or thoracic (32.6%) spine. More lumbar infections were observed for other species (91.7%, 11/12), followed by MSSA (75%), M. tuberculosis (69.2%), and MRSA (30%). Infectious spondylitis occurred more in males than in females (1.5:1), with a ratio of 2.25 for M. tuberculosis, 1.5 for MRSA (1.5:1) and 1.67 for MSSA, but was equally distributed for other species.

All six S. aureus isolates were resistant to PEN and AMP (Table 2). Among the 5 MRSA isolates, SCCmec type IIIA isolates 2 and 3 were resistant to all 11 antimicrobials examined, while the MRSA SCCmec type IIIA-like isolate 6 was only resistant to 7 of the tested antimicrobials and lacked PCR products for C and F of the normal SCCmec IIIA type. Additionally, SCCmec type IV isolate 5 was less resistant than the other three isolates. While the MRSA SCCmec type IIIA and IIIA-like isolates were resistant to at least one of the two carbapenems (ertapenem and IMP), the SCCmec type IV isolate appeared to have a reduced susceptibility to ETP but was susceptible to two other carbapenems (IMP and MEM).

MLST analysis revealed that MSSA isolates 1 and 4 belonged to ST959 and ST15, respectively, and four MRSA isolates belonged to ST239 (isolates 2 and 3), ST30 (isolate 5) and ST59 (isolate 6) (Table 3). In addition to being identical in ST239 and antimicrobial patterns, pulsotype analysis demonstrated that isolates 2 and 3 differed with respect to pulsotypes (Figure 1(a)). Plasmid analysis revealed that isolates 1 to 4 carried one plasmid of 50-kb or a less than 6.6-kb plasmid, and isolates 5 and 6 lacked a plasmid (Figure 1(b)). The agr type analysis indicated that all MRSA SCCmec type IIIA and IIIA-like isolates belonged to agr 1, and the other isolates belonged to a separate agr type.

Among the virulence genes examined, all of the isolates carried leukocidin genes; lukED was identified in isolates 1 to 4, and pvl was observed in isolates 5 and 6 (Table 3). The hemolysin genes hla, hld and hlg encoding α, δ, and γ-hemolysin, respectively, were identified in all of the isolates, while β-hemolysin was only identified in isolate 6 (Table 3). Within the two hlg types, five isolates carried hlg-2. The exfoliative toxin genes eta and etb were not found, while the toxic shock syndrome toxin gene, tst, was only observed in MRSA isolate 5.

^aP: penicillin, AMP: ampicillin, OX: oxacillin, CF: cephalothin, TE: tetracyclin, OT: oxytetracycline, CC: clindamycin E: erythromycin, SXT: trimethoprim/ sulfamethoxazole, ETP: ertapenem, IMP: imipenem and MEM: meropenem. The susceptibility to ETP, IMP, and MEM was determined by Disc diffusion and minimum inhibitory concentration (MIC) methods; the number in the parenthesis is the MIC value (μg/mL).

^ahla, hlb, hld, and hlg (or hlg-2): hemolysin genes α, β, δ and γ, respectively; tst: toxic shock syndrome toxin 1; agr: accessory gene regulator. eta and etb was not found in any isolate.

M 1 2 3 4 5-1 5-2 6 M M1 1 2 3 4 5-1 5-2 6 M2

Infectious spondylitis can result from distant infections, such as respiratory tract. Therefore, we used the human alveolar basal epithelial cell line A549 to investigate the cytotoxicity of S. aureus to epithelial cells. Although six S. aureus isolates resulted in different cell death rates of cell line A549 by necrosis, early and late apoptosis, all clinical isolates caused cell death, with over a 40% cell death rate for isolates 4, 5, and 6 at the 4-h incubation point, whereas the rate of cell death did not differ among the bacteria at the 1-h incubation point (Figure 2). While isolates 2 and 3 resulted in the lowest cell death rates (similar to that caused by E. coli pir116), the remaining four isolates showed increased necrosis compared with the control S. aureus ATCC25923. Analysis of the cell death types between 1 hour and 4 hours demonstrated that isolates 4 - 6 caused more cell death than did other isolates, and all six clinical isolates increased necrosis more than late apoptosis, with a ratio larger than 1.27 compared with E. coli pir116 (0.53) and S. aureus ATCC 25923 (0.34) (Table 4).