Isolation, identification & characterization of Proteus penneri - a missed rare pathogen

Genus Proteus currently has four species namely P. mirabilis, P. vulgaris, P. penneri, and P. hauseri, and three unnamed genomospecies, Proteus 4, 5, and 6 (single-strain P. myxofaciens isolated from gypsy moth)1–3. P. penneri (formerly P. vulgaris biogroup 1) was recognized as a new species in 1982 but its role in clinical infection remains cryptic23. Most infections encountered usually are caused by P. mirabilis (70-90%) followed by P. vulgaris and rarely by P. penneri which is misidentified as P. mirabilis since both are indole negative on routine biochemical testing. P. penneri has ability to cause major infectious diseases and nosocomial outbreaks2 and carries similar pathogenic determinants to P. mirabilis and P. vulgaris3.

P. penneri usually infects urinary tract, blood, abdominal wound, groin, neck and ankle and has been isolated mostly from urine (50%), wound and soft tissue exudates (25%), and blood cultures (15%)3–5. In a few individual case reports it has been isolated from subcutaneous abscess, urosepsis in a case with diabetes mellitus and epidural abscess, etc5–9. Due to lack of awareness, and inability of routine bacteriology laboratories to identify P. penneri isolates there are no reports on the isolation and identification of P. penneri from the country nor on its characterization, hence we carried out this study to isolate P. penneri from various clinical specimens and to characterize these isolates.

Material & Methods

Our laboratory receives several thousand samples comprising approx. 8000 urine, approx. 5000 pus and body fluids per year. Clinical specimens comprising urine, pus and body fluids received from ICUs, wards and out patients departments of Sanjay Gandhi Post-Graduate Institute of Medical Sciences, Lucknow, a tertiary health care center in north India were collected and cultured routinely for bacterial isolations, and identification in the department of microbiology over a decade (1997-2007). The clinical specimens were first screened microscopically by Gram's stain, then cultured on blood agar (aerobically and anaerobically), MacConkey agar and Robertson cooked meat broth for 48 h at 37°C in 5-10 per cent CO₂. Urine specimen were examined microscopically as wet mount and then its semiquantitative cultures were done on cystine lactose electrolyte deficient (CLED) agar medium (Hi-Media Laboratories, India) and sheep blood agar media and incubated at 37°C for 18-24 h.

All the isolates that grew from urine (10⁵ cfu/ml), pus and body fluids were at first subjected to standard routine biochemical tests and provisionally diagnosed as Proteus species based on the findings of non-lactose fermenting colonies with (or without) swarming on blood agar plate, characteristic fishy smell, Gram negative pleomorphic bacilli with active motility and that reduced nitrate, produced catalase but not oxidase, fermented glucose (with acid and gas), sucrose and mannitol but not lactose, usually did not utilize citrate but hydrolyzed urea and produced phenyl pyruvic acid, were identified as Proteus species3510.

Indole negative Proteus species as identified above were then picked up whenever possible during the study period stocked and maintained in nutrient agar stabs and later identified by putting extended biochemical tests as laid down in the manual for the identification and differentiation of P. penneri isolates10. The extended, simple differentiating biochemical tests are shown in the Table, among these additional production of ornithine decarboxylase and fermentation of salicin and maltose were included in the present study. The most reliable single test and the absolute biochemical criteria used was failure of production of the enzyme ornithine decarboxylase (after 5-7 days of incubation) by P. penneri (0%) in contrast 100 per cent strains of P. mirabilis which produces this enzyme10. Further, P. penneri isolates are non-fermenter of salicin and non-utilizer of citrate but acidify sucrose and maltose10. Antibiotic sensitivity tests were done in all eight isolates soon after their isolation against 18 antimicrobial drugs viz., ampicillin, ampicillin plus clavulanic acid, amikacin, gentamicin, cotrimoxazole, ciprofloxacin, cefuroxime, ceftazidime, chloramphenicol, cefotaxime, nalidixic acid, norfloxacin, tetracycline, cefodroxil, meropenem, ceftriaxone, cephalexin and ticarcillin by modified Stokes’ disc diffusion method11. Additionally, Dienes phenomenon was tested later on in five isolates (as three isolates 4U, 5U and 8P were lost to subcultures) using freshly prepared blood agar culture plates as per standard procedures12 for further typing and comparison of P. penneri isolates. Screening for β-lactamase production was tested by nitrocefin method using standard procedure13. Briefly, a disc of nitrocefin (Becton Dickinson, USA) made wet with sterile distilled water was put over the colonies of an overnight culture plate and examined after 30 min for development of pink colour in the disc if β-lactamase was produced while no colour was developed in non-producers of this enzyme. β- lactamase production was done on the same five isolates of P. penneri which were tested for Dienes phenomenon. The results of antibiograms of isolates and Dienes phenomenon were compared and utilized for the characterization and typing of P. penneri into strains.

Results & Discussion

Among 61 indole negative Proteus species subjected to detailed biochemical analysis, eight (13.1%) isolates of P. penneri were identified (Table). Majority of the P. penneri infected patients were males (5 of the 8 cases) and belonged to a wide age range (3-73 yr). In all the eight cases, P. penneri was the sole pathogen as it was isolated in pure culture and all cases had some underlying disease (Table). Five of the eight isolates of P. penneri showed β-haemolysis on sheep blood agar. Biochemically the strains were quite homologous except a few reactions, e.g. one strain was citrate positive and H₂S negative. P. penneri are known to be non-producer of the enzyme ornithine decarboxylase and are uniformly salicin negative but acidify maltose. These characters were strictly adhered to and were also seen among all the isolates in this study.

Table Clinical characteristics, predisposing factors and biochemical identification of eight isolates of P. penneri

Antibiogram were different for different isolates and all were multi-drug resistant (MDR) with resistance to 6-8 drugs and even to 12 drugs in one isolate (strain 5). Strain No. 1 and 2 had quite similar antibiogram except for being sensitive to chloramphenicol. Strain-3 was also nearly similar to first two except being sensitive to co-trimoxazole, nalidixic acid and norfloxacin. Among the β-lactams, ampicillin proved to be the least efficient and all isolates were resistant to ampicillin, first and second-generation cephalosporins (cephalexin and cefuroxime). Third generation cephalosporins (ceftazidime, cefotaxime and ceftriaxone) showed better although heterogeneous, in vitro sensitivity. β-lactamase production was seen in three of the five isolates tested. The antimicrobial drug sensitivity tests done in the present study and by other authors too were according to modified Stokes’ disc diffusion method, being prevalent that time while CLSI guidelines came much later in the year 2005-06 and were adopted gradually.

Swarming was not seen in the first strain-1 on primary isolation while growth and swarming was poor in the strain-4. On the basis of Dienes phenomenon (done on the 5 isolates) it was observed that four strains (strains-1, 2, 3, and 5) were distinct from one another as swarming edges did not merge with one another but showed demarcation line of non-identity while strain-4 could not be typed due to its poor growth and swarming. Typing by Dienes phenomenon was more discriminative since strains-1 and 2 had identical antibiogram except in sensitivity to single drug namely chloramphenicol but were distinct Dienes types. Three isolates were lost due to long period of study.

Most bacteriologists are largely unaware of species P. penneri and even if aware no attempts have been made to find its clinical significance. Hence, there is a scarcity of large series1415 reports on the isolation of P. penneri from infected patients. Another reason is that P. penneri was recognized as a new species in 1982 only. In the present study, among indole negative Proteus species isolated, only 13.1 per cent (8/61) isolates were P. penneri. In a study from Taiwan among all positive urine cultures from hospitalized patients with urinary tract infection/urinary calculi an incidence of 2.1 per cent for P. penneri was reported14. Another study from Korea15 focusing on bacteraemia due to tribe Proteeae isolated 132 strains, but only two were P. penneri. In the present study, P. penneri was isolated from post-operative patient's urine, drain fluid or pus and who had some underlying disease besides diabetic foot ulcer, which was found to be novel presentation.

Since P. penneri may be non-swarming on first isolation, extra vigilance is required. Infections due to P. penneri (and also P. mirabilis and P. vulgaris) could be more pronounced, persistent and difficult to eradicate owing to being an invader pathogen with several virulence factors16 besides swarming and haemolysin and production of proteolytic enzymes. The major mechanism of antimicrobial drug resistance is determined chromosomally317 with hyperproduction of the chromosomally encoded beta-lactamase, occasionally due to plasmids. Besides P. penneri strains are naturally resistant18 to penicillin G, amoxicillin, cephalosporins (i.e. cefaclor, cefazoline, cefuroxime and cefdinir), oxacillin, most of macrolides but are naturally sensitive to aminoglycosides, carbapenems, aztreonam, quinolones, sulphamethoxazole and co-trimoxazole. Most of isolates in the present study were MDR including resistance to amoxy-clavulanic acid combination (except one isolate). In one study, P. penneri was found to be more resistant to the penicillins and cephalosporins than P. mirabilis and mostly in patients with urogenital infections18. This further emphasizes the need for species level identification of P. penneri. In the present study, beta-lactamase production was seen in three of the five isolates; however, all eight P. penneri isolates were uniformally sensitive or moderately sensitive to ceftazidime, cefadroxyl and meropenum and could be treated successfully except the first case who had post-urethral valve.

Typing by antibiogram found P. penneri to be highly hetrogenous MDR strains and hence was of limited value. Moreover, strain differentiation by antibiogram and Dienes phenomenon did not show concordance. Typing by Dienes phenomenon was more discriminative as it differentiated strains even if it differed in resistance with a single antibiotic, e.g. strains-1 and 2 and is still employed19.

Hence efforts should be made to identify P. penneri infections by using a few additional biochemical tests in a large series studies by ribotyping and pulse field gel electrophoresis (PFGE) in order to find its incidence, predisposing host factors and the clinical significance of this missed bacterium. Further, patients with underlying diseases going for abdominal or urogenital surgeries should be examined carefully for P. penneri infections because of their MDR nature with potential to spread drug resistant plasmids nosocomially.