Erythema caused by a localised skin infection with Arthrobacter mysorens

In June early summer, a nine-year old boy spent four hours in a forest digging out a bicycle track to ride his mountain bike. He returned home with a dirty shirt in particular at the right side of the chest, very close to the right acromastium. Since he felt a localised pruritus there, he had intensively scratched the region, thereby contaminating the skin with forest soil. A small erythema with an average diameter of one centimetre and a clear-cut red edge above the right acromastium was apparent on the following day. His mother suspected a potential insect or tick bite, although no tick could be found. The patient had no previous erosion. The then conducted Borrelia-specific ELISA was negative for IgM antibodies but positive for IgG antibodies. An immunoblot (Recomblot Borrelia, Mikrogen, Germany) with the patient's serum revealed a B. burgdorferi sensu lato-specific, IgG antibody response to p100, p41, BmpA, OspC (weakly positive), p41, and p18 but no IgM-specific antibody response could be detected. This finding was consistent with a Borrelia infection at an advanced stage (> 6 months after infection) or a residual of an earlier infection, as a symptom-free patient may also have a similar Borrelia-specific antibody response based on longtime persistent antibodies. Clinical findings in this stage are typically those of advanced neuroborreliosis (progressive encephalomyelitis etc.), acrodermatitis chronica atrophicans or Lyme arthritis. Since the patient did not display any symptoms corresponding to these clinical syndromes, a residual, asymptomatic infection was suspected and no specific treatment was initiated. These findings argue against a Borrelia infection as the cause for the patient's symptoms, as EM caused by Borrelia burgdorferi s. l. represents a very early stage of infection [2, 5, 6] and would probably lead to a specific IgM antibody response. To exclude a possible re-infection, a second serum sample parallel to the first sample was tested 4 weeks later. However, no significant serological changes could be observed. Within one week, the small erythema spread laminarly, exhibiting a red edge with a paler, faded centre, and the patient showed symptoms resembling an EM caused by B. burgdorferi (Figure 1) as an early symptom of Lyme disease [1, 4]. In distinction to a typical EM with its regular, oval form, the shape of the erythema was nearly rectangular and displayed "protuberances" at the edge of the lesion (Figure 1). Thus, the clinical picture, together with the serological findings did not clearly correlate with the early clinical manifestation of Lyme disease, namely EM [2, 3]. Since the patient complained about a persistent pruritus, which is atypical in EM (where minimal pruritus is only occasionally described at an early stage), a dermal mycosis was suspected, leading to the prescription of a topical antimycotic drug solution with the active component Econazol (Janssen Cilag GmbH, Neuss, Germany) by the family physician. The treatment proved to be ineffective since the erythema showed clinical progression and started to cover the complete right chest, neck and shoulder. It still presented a clear erythematous edge and a faded centre (Figure 1).

At this point, the family consulted the Institute of Medical Microbiology (Giessen, Germany) and as part of a staged diagnostic approach, three swabs of the skin were taken simultaneously from different localisations; the first one from the initial focus of infection, the second one from the edge of the erythema and the third one approximately 50 millimetres outside the visible edge. The swabs were cultured for fungi and bacteria. The culture on sheep blood agar plates from swab one and two revealed many yellow-pigmented small colonies at room temperature after 48 hours of incubation. After 72 to 96 hours, the colonies showed a frayed edge indicating motile bacteria (Figure 2). These colonies could not be demonstrated in the culture of the third swab from the outside of the erythema. Additionally, few Staphyloccus epidermidis colonies, as members of the indigenous skin flora, could be found in all the three swab samples. No fungal growth could be evidenced even after prolonged incubation of 4 weeks, which argues against the presence of dermatophytes. The yellow-pigmented colonies did not grow at 37°C in the presence or absence of CO₂. The partial sequencing of the 16 S rRNA gene and data base analysis using the BLAST algorithm (NCBI) and the SepsiTest™ BLAST http://​www.​sepsitest-blast.​de identified the isolate as Arthrobacter mysorens (GeneBank accession number HM751094). Biochemical assays using API systems (BioMerieux, Nuertingen, Germany) demonstrated only the genus Arthrobacter but not the species. Antibiotic susceptibility testing via agar diffusion method and E-Test, classified the isolate to be susceptible against amoxicillin, doxycycline, cefuroxime, ceftriaxone, and cefotaxime (MIC's < 0.064 mg/ml). The patient was thereafter successfully treated with oral amoxicilline for one week (3 × 1 g per day). The skin erythema showed significant clinical improvement and was no longer detectable within two days. Subsequent swab analysis remained negative for Arthrobacter mysorens. To further determine the role of A. mysorens an immunoblot analysis using a whole cell extract of A. mysorens was performed. The patient's serum showed a specific immune response against a ~53-kDa and a ~32-kDa protein, respectively (Figure 2D), which demonstrates that the immune system responded to A. mysorens with a type-specific response. A control Borrelia serology after 6 months demonstrated a loss of the OspC and p18 bands and a weakened intensity of the BmpA band in the IgG-immunoblot, thus indicating a diminishing immune response and supporting the initial theory of a residual antibody response reflecting a past infection. The genus Arthrobacter includes a heterogenous group of aerobic, gram-positive, catalase-positive, non-fermentative coryneform bacteria that are widely distributed in the environment where their main habitat is the soil. Cells are able to resist desiccation and starvation [7]. From a review of the literature only about 42 documented cases of Arthrobacter species isolated from clinical samples could be retrieved [8].