Bacteremia Due to Moraxella atlantae in a Cancer Patient

Abstract

A gram-negative alkaline phosphatase- and pyrrolidone peptidase-positive rod-shaped bacterium (CCUG 45702) was isolated from two aerobic blood cultures from a female cancer patient. No identification could be reached using phenotypic techniques. Amplification of the tRNA intergenic spacers revealed fragments with lengths of 116, 133, and 270 bp, but no such pattern was present in our reference library. Sequencing of the 16S rRNA gene revealed its identity as Moraxella atlantae, a species isolated only rarely and published only once as causing infection. In retrospect, the phenotypic characteristics fit the identification as M. atlantae (formerly known as CDC group M-3). Comparative 16S rRNA sequence analysis indicates that M. atlantae, M. lincolnii, and M. osloensis might constitute three separate genera within the Moraxellaceae. After treatment with amoxicillin-clavulanic acid for 2 days, fever subsided and the patient was dismissed.

CASE REPORT

In April 2001 a 31-year-old female patient was admitted to the oncology department with a history of intermittent fever for more than 1 week and with complaints of right hypochondrial pain and rectal cramps. The patient had a rectal adenocarcinoma that was diagnosed 1 year previously, with recent massive liver metastasis which had been treated with an anterior resection and with palliative chemotherapy (irinotecan and 5-fluorouracil) for 10 months. Clinical examination revealed a relapse of the rectal cancer at the anastomosis and a hepatomegaly of 8 cm. X-ray examination of the thorax revealed no infiltrates. Three pairs of blood cultures and a urine culture were carried out, with the latter being negative. Streptococci of the viridans group were isolated from one aerobically incubated culture bottle. A gram-negative rod, susceptible to all antibiotics tested and later identified as Moraxella atlantae (CCUG 45702), was isolated from the two other aerobic bottles. The anaerobic cultures remained negative. The patient was treated with amoxicillin-clavulanic acid (875 mg) two times a day, resulting in the disappearance of the fever after 2 days, whereafter the patient was dismissed. A computed tomography of the abdomen 7 days later showed an abscess around the rectum, which was drained with evacuation of 100 ml of serous fluid. Culture of this fluid remained negative. The patient died 4 months later as a result of her underlying disease.

The blood culture bottles (FA aerobic and SN anaerobic bottles; Organon Teknika, Boxtel, The Netherlands) were inoculated with 10 ml of blood and incubated in the Bact/Alert 3D system (Organon Teknika). Gram staining showed short, plump, rod-shaped to coccoid gram-negative rods. On tryptic soy agar plus 5% sheep blood (Becton Dickinson, Erembodegem, Belgium), colonies were small and grayish, with twitching motility and corroding activity. The isolate grew well on MacConkey agar, in contrast to most other Moraxella sp. It was nonmotile. Catalase and oxidase were positive. No sugars were acidified, either fermentatively or oxidatively. No growth was observed on triple sugar iron agar. Nitrate reduction and production of indol, urease, phenylalanine deaminase, lysine and ornithine decarboxylases, and arginine dihydrolase were negative. Using Rosco Diatabs (Taastrup, Denmark), tributyrin and proline aminopeptidase were negative, but alkaline phosphatase was positive and pyrrolidone peptidase was strongly positive. The code obtained by API 20 NE (BioMérieux, Marcy L'Etoile, France) strips was 0000004, corresponding to Moraxella sp. (82.3%). No identification was obtained when the ID 32 GN (BioMérieux) strips were used. This phenotypic profile did not initially lead to an identification, but in retrospect it was found to be consistent with the genotypic identification as M. atlantae.

Susceptibility was tested with the disk diffusion technique. No criteria are available for interpretation of the results, but because of the large inhibition zones, the strain was considered to be susceptible to all tested antibiotics, namely ampicillin, cotrimoxazole, cefuroxime, gentamicin, colimycin, temocillin, and ciprofloxacin.

Since no initial phenotypic identification could be obtained, the isolate was subjected to genotypic identification. Initially, identification with tRNA-PCR (13) in combination with capillary electrophoresis (1, 12) failed because no entries for M. atlantae were present in our tRNA-fingerprint library. However, since this M. atlantae isolate had tRNA spacer fragments with lengths of 116, 133, and 270 bp, a pattern which is different from all other members of the genus Moraxella and from those obtained for all bacterial species tested thus far, addition of this fingerprint to the library should enable future identification by means of tRNA-intergenic spacer-length polymorphism analysis. Two M. atlantae culture collection strains (CCUG 10707 and CCUG 31324) were shown to have tRNA spacer fragments with identical lengths.

16S rRNA gene amplification and sequencing was carried out as published before (11). Briefly, the complete 16S rRNA gene was amplified, followed by sequencing reactions using the Big Dye Terminator Sequencing kit (Applied Biosystems, Foster City, Calif.) and analysis of the obtained fragments on the ABI 310 capillary electrophoresis apparatus (Applied Biosystems). Total gene assembling of the obtained fragments, alignment, and clustering were done with GeneBase (Applied Maths, Kortrijk, Belgium). The obtained sequence (1,354 bp) was compared to all known sequences in the GenBank by Blast (National Center for Biotechnology Information, Bethesda, Md.; http://www.ncbi.nlm.nih.gov/blast/index.html). The Blast search resulted in a similarity of almost 99% with the only M. atlantae sequence present (GenBank number AF005191).

However, the sequences obtained for the M. atlantae strain of this study and for the GenBank M. atlantae strain contained two regions with substantial differences. Therefore, two M. atlantae culture collection strains (CCUG 10707 and CCUG 31324) were sequenced. This revealed other differences in the same regions. Those differences appeared to be nonrandom. For region 1 (Escherichia coli position 201 to 218), sequences a (TTTWGGGTTC) and b (GCGAGAGCTTT) were observed. Sequence a was present in strain CCUG 10707 and the strain of this case report, and sequence b was observed for the GenBank entry AF005191. Strain CCUG 31324 apparently carried different alleles from either one or both sequences, since the ambiguities in the sequence obtained for this strain corresponded to a mixture of both sequences found in the other strains (Fig. 1). For region 2 (E. coli position 455 to 478), two possible sequences were seen, with strain CCUG 10707 and the GenBank entry AF005191 having one possible sequence and CCUG 31324 and our clinical isolate having the other. Other points of difference were found at E. coli position 381 (either C or G), E. coli position 668 (either A or G), E. coli position 848 (either C or T), and E. coli position 1136 (either A or G).

FIG. 1.

Alignment of two variable regions of the 16S rRNA gene sequences of four M. atlantae strains. IUB codes: S, C or G; Y, C or G; K, G or T; W, A or T; R, A or G. Sequences: 1, M. atlantae CCUG 10707 (GenBank no. AJ417491); 2, M. atlantae CCUG 31324 (GenBank no. AJ417492); 3, M. atlantae CDCA1922 (GenBank no. AF005191); 4, M. atlantae CCUG 45702, the clinical strain of this study (GenBank no. AJ313278); 5, E. coli ATCC 11775^T (GenBank no. X80725).

In fact, all four strains have highly identical 16S ribosomal DNA sequences which at some regions are different mixtures of a few basic themes, probably reflecting past recombination events (5, 9). Furthermore, strain CCUG 31324 appears to have alleles with different sequences in a region between E. coli positions 201 and 218 whereby one or more cistrons have sequence a while one or more others have sequence b.

The 16S rRNA sequences of two M. atlantae isolates were compared with those of other moraxellae (including a more complete sequence of Moraxella lincolnii obtained in this study), Psychrobacter immobilis and Psychrobacter phenylpyruvicus, and the Acinetobacter calcoaceticus type strain (Fig. 2). Psychrobacter and Acinetobacter are genera which belong to the Moraxellaceae as well. A dendrogram of 16S rRNA gene sequences was constructed using the GeneBase software package. First, pairwise alignment using the unweighted pair group method with arithmetic mean was carried out with a gap penalty of 100%, a unit gap cost of 20%, and an ambiguity cost of 50% of the mismatch cost. Subsequently, global alignment with A. calcoaceticus as the outgroup was carried out on the region corresponding to positions 60 through 1334 of the 16S rRNA gene of E. coli, with costs as described above. Finally, a similarity matrix of the aligned sequences was constructed by global alignment homology calculation and a gap penalty of 20%. The neighbor-joining method was used to construct the dendrogram based on this similarity matrix (7).

FIG. 2.

Cluster analysis of 16S rRNA gene sequences of M. atlantae and other Moraxellaceae. Cluster analysis was based on the neighbor-joining method, with A. calcoaceticus as the outgroup. The bar indicates 1% homology divergence. The numbers in parentheses are references.

This revealed that in fact the species Moraxella osloensis, M. lincolnii, and M. atlantae could be classified as separate genera, based on a less than 95% sequence relatedness to each other, to the other moraxellae, and to Psychrobacter. The similarity percentages were 91 for M. atlantae versus M. lincolnii, 92 for M. atlantae versus M. osloensis, 89 for M. atlantae versus the other moraxellae, 93 for M. lincolnii versus M. osloensis, and 89 for M. atlantae versus P. immobilis and P. phenylpyruvicus. The sequence differences between the two Psychrobacter species amounted to 5.7% as well. The suggestion that these species could be classified as separate genera is in correspondence to earlier reports based on 16S rRNA gene sequencing (8), DNA-DNA hybridization (10), and DNA transformation (2).

M. atlantae is an unusual and only rarely isolated bacterium, formerly known as CDC group M-3. Bøvre et al. (2) described five strains isolated from blood cultures, all with the tendency to spread on blood agar plates, but without mentioning the clinical importance. The strain of the only case report thus far (3) was isolated from a blood culture of a 25-year-old patient suffering from systemic lupus erythematosis and had the same biochemical and phenotypic characteristics as the strain from the present case report.

M. atlantae is a rare opportunistic pathogen that is apparently susceptible to most common antibiotics. This—together with the difficulties encountered in most laboratories in the identification of gram-negative nonfermenters—may cause possible underestimation of its occurrence. For example, the Culture Collection of the University of Göteborg (CCUG) harbors 12 M. atlantae strains, of which 10 were isolated since 1981. Ten are from blood, one from pleural fluid, and one from a dog bite wound. Unambiguous identification of this organism is possible by means of 16S rRNA gene sequencing, tRNA-PCR, and phenotypic characteristics. Short, nonmotile, gram-negative rods forming small colonies, possessing twitching motility and corroding activity on blood agar, capable of growing on MacConkey agar, oxidase and catalase positive, not acidifying sugars, negative for nitrate reduction, urease, acetate assimilation, and tributyrin hydrolysis, and positive for pyrrolidone peptidase can be considered M. atlantae.

Nucleotide sequence accession numbers.

The sequences of the 16S rRNA genes from M. atlantae and M. lincolnii obtained in this study were deposited under GenBank numbers AJ313278 and AJ417490, respectively.