Within the genus Mycobacterium, non-pigmented rapidly growing mycobacteria (NPRGM) are among the most common species of nontuberculous mycobacteria isolated in clinical mycobacteriology laboratories. Most members of this group have been described as causing human infection,1 including nosocomial disease. The most common species implicated in these infections are M. fortuitum, M. chelonae and M. abscessus; in many other cases, these bacteria are not clinically significant. Other members of this group, such as M. peregrinum or M. mucogenicum, are found less often.
Despite the importance of NPRGM as human pathogens, few in vitro studies have investigated the pathogenic mechanisms of these microorganisms and the relationship between these mechanisms and virulence. One study has shown differences in growth characteristics and colony phenotype between pathogenic and non-pathogenic strains of Mycobacterium abscessus.2 In another study using a strain of M. smegmatis as a negative control, Bermudez et al. demonstrated that M. avium invades HEp-2 monolayers.3 Herein, we report the results of a study evaluating the relationships between cellular invasiveness, clinical significance, and colony phenotype of NPRGM strains isolated from human samples, together with collection strains.
The study was carried out with collection strains and clinical strains of NPRGM. The clinical strains were isolated from samples processed for mycobacterial culture in the Mycobacteriology Laboratory of Fundación Jiménez Díaz (Madrid, Spain). Strains were identified according to commonly recommended schemes, using biochemical tests and PCR-restriction enzyme analysis (PRA)1. The clinical significance of the strains was assessed by reviewing the patients’ clinical charts according to internationally accepted criteria.4
For colony phenotype and fibroblast microcolony study, we followed our previously described protocol. To investigate invasion in HEp-2 monolayers, the experiment was developed modifying the method described by Bermudez et al.3 Results were statistically analyzed using a contingency table and the chi-square test.
The 18 collection strains used in the study pertained to 16 different species, and the 74 strains obtained from clinical samples belonged to 5 different species (3 strains of M. abscessus, 24 M. chelonae, 33 M. fortuitum, 6 M. mucogenicum and 8 M. peregrinum). Twenty-four of these strains were considered clinically significant.
In the fibroblast microcolony assay, 35 strains presented rough colonies and only 12 strains showed elongated colonies.
All strains but one infected HEp-2 monolayers. Fifteen strains showed counts of 1 to 1000CFU/mL and 45 strains showed counts of 1000 to 10 000CFU/mL. Twenty-nine strains had counts of 10 000 to 100 000CFU/mL and 2 strains, both of them M. chelonae, had counts of more than 100 000CFU/mL.
The statistical analysis showed no relationship between a rough or smooth phenotype and the capacity to invade fibroblasts or HEp-2 cells. Furthermore, no relationships could be established between the clinical significance of the isolates, and the capability for intracellular penetration or the specific colony phenotype.
The presence of a rough or smooth colony phenotype in conventional agar culture has been suggested as a potential sign of pathogenicity. Several studies performed with M. abscessus2,5 and M. avium6 using in vitro and in vivo models have shown that strains producing rough colonies are pathogenic, whereas those producing a smooth variant are not. Nevertheless, these studies are limited by the small number of strains analyzed. In a preliminary report performed with a small number of clinical isolates, we found that fibroblast invasiveness was not related to the capability of a strain to cause human disease.7 Although in that experiment fibroblast invasiveness seemed to appear in strains isolated from severe disease (bacteremia), in our present study performed in a large number of strains, this preliminary result was not confirmed.
We found that smooth and rough colonies appeared as often in clinically significant strains as in non-significant ones. Therefore, the phenotype does not seem to be associated with the capacity of a strain to produce human infection. No relationship was established between clinical significance and cellular invasiveness in either cell culture model. Only 5 of the 24 clinically significant strains showed elongated microcolonies; hence this characteristic does not seem to be directly related to the capability to produce human infection.
Several experiments have demonstrated that mycobacteria can also invade epithelial cells that are non-professional phagocytes, such as HEp-2 cells.3,8 Again, our data indicate that there is no relationship between the capacity to invade HEp-2 cells and the clinical significance of the strain. Our study, using a large number of strains from several species, demonstrated that invasiveness for fibroblast and HEp-2 cells was not related to the clinical significance of the strains; therefore, this capability seems to be a pathogenic factor of minor importance in the development of human disease.
This study was supported by a grant from the Fondo de Investigaciones Sanitarias (FIS PI030146). Martín-de-Hijas NZ was funded by the Fundación Conchita Rábago de Jiménez Díaz. Kinnari TJ was funded by the Academy of Finland, Paulo Foundation and Proteesisäätiö Foundation.
Conflicts of interestAll authors declare no conflict of interest.
This work was presented as a poster in the 17th ECCMID (Munich, 2007). Poster P1998.
