TEXT

Top Abstract Text References

Acanthamoeba keratitis is a chronic progressive infection that has been increasingly identified worldwide. The condition currently is successfully managed medically in most patients (8), but it can progress to perforation of the corneal ulcer if untreated or mismanaged medically.

Many drugs have been suggested to have in vitro activity against Acanthamoeba spp. (5). The first successful medical cure involved topical therapy with a combination of the aromatic diamidine propamidine isethionate (Brolene) and neomycin (11). Subsequent studies confirmed the efficacy of this treatment (1), but it later was found to be successful in only half of the patients. One possible explanation for the failures was found in the observation that resistance to propamidine developed in a temporal series of Acanthamoeba isolates from a patient (2).

The patient had been wearing soft contact lenses and presented with a 3-week history of bilateral keratitis. Culture of corneal scrapes from each eye, with incubation at 37°C, yielded prolific numbers of Acanthamoeba organisms. Topical administration was initiated with the previously successful combination of propamidine (0.1% [wt/vol]) and neomycin (0.5% [wt/vol]) (11). Neomycin was withdrawn because of contact hypersensitivity. This led to recrudescence of the infection in both eyes, with Acanthamoeba being isolated and cultured at 37°C (2). Treatment continued with propamidine and with an arsenical (R6/56, an atoxyl derivative), but no control of the infection was achieved, and fulminant keratitis resulted. Resistance to the diamidine was associated with reduction in the temperature at which optimum growth and replication occurred (5). Further isolates at this stage would not replicate at 37°C, but did at 25°C; these isolates proved to be resistant to both propamidine and the arsenical. The infection eventually was halted by surgery on both eyes with repeated, bilateral corneal grafts (2).

The present study addresses the question of whether the infection in this patient and the subsequent development of propamidine resistance during therapy was associated with a single Acanthamoeba strain, or whether it resulted from a mixed infection that included both propamidine-sensitive and -resistant strains. The answer to the question depends on the experimental ability to distinguish different strains. The question was investigated previously by Kilvington et al. (6). This group compared restriction fragment length polymorphisms of mitochondrial DNA (mtRFLP) from Acanthamoeba obtained from the patient before therapy (isolate AcPHL/7a) and after therapy (isolates AcPHL/7b and -7c). Because the mtRFLPs for the three isolates were identical, the authors concluded that the propamidine resistance developed in a single strain due to the therapy. However, identical mtRFLPs often are found in different Acanthamoeba strains. For example, in three studies, samples of 8, 13, and 33 strains had only 4, 7, and 11 different RFLP phenotypes (4, 6, 12). Thus, it was possible that the RFLP analysis failed to identify the sensitive and resistant isolates as different strains. Therefore, we reexamined the issue by using a test with greater strain specificity.

The genes coding for 18S rRNA (18S rDNA) of Acanthamoeba have relatively high interstrain variation in DNA sequences (3, 9). In addition, strains can differ in the number of 18S rDNA alleles and the presence or absence of rDNA introns. These three types of differences provide an excellent basis for differentiating among individual strains of this genus. In a sample of 53 strains representing 12 lineages referred to as sequence types, the 18S rDNAs differed in 52 (98%) of the isolates (3, 7, 9).

We compared the 18S rDNA base sequences from three isolates, H30 (AcPHL/7a), H31 (AcPHL/7b), and H32, from the right eye of the patient studied by Kilvington et al. (6). The methods used for DNA isolation, PCR amplification, and 18S rDNA sequencing are described elsewhere (3, 9, 10). H30, a January isolate, was collected before therapy, and H31, a February isolate, was collected 4 weeks after therapy (6). The March isolate from the original investigation was missing; H32 (a July isolate) was used instead.

Drug sensitivity testing originally used a mixed trophozoite and cyst culture and a double-dilution technique in a tube (2). At that time, H30 was sensitive to propamidine at a MIC of 1.5 µg/ml and a minimum cysticidal concentration (MCC) of 12.5 µg/ml. The posttherapy isolate H31 and the March isolate were not cultivatable at 37°C, but grew at 25°C and were resistant to propamidine (MIC, 12.5 µg/ml; MCC, 50 µg/ml).

In the present study, which included H30, H31, and H32 (a July isolate), the sensitivities of trophozoites and cysts were tested separately by a microtiter method (5). Trophozoites and cysts of H30 both were sensitive to propamidine at a minimum trophozoite amoebicidal concentration (MTAC) of 3.2 µg/ml and an MCC of 12.5 µg/ml (5). H31 and H32 were resistant to propamidine at an MTAC of 25 µg/ml and MCCs of 50 and 100 µg/ml, respectively. Similar values were obtained for isolates retested after the molecular studies. All isolates grew at 25°C. H30 grew optimally at 37°C, but neither H31 nor H32 grew at 37°C.

Initial sequencing results indicated that more than one 18S rDNA sequence was present in each of the stock cultures of H30, H31, and H32. This was evidence for more than one strain in each culture or for more than one rDNA allele in each strain. To distinguish between these possibilities, single amoebae were cloned from the stock cultures. The 18S rDNA was obtained by PCR from the amoeba clones and then cloned in bacterial hosts. Sequencing of the cloned 18S rDNA revealed that three alleles, RnsA, RnsB, and RnsC (Table 1), were present in all amoebae in all clones from H30, H31, and H32. The observation that propamidine-sensitive and -resistant isolates had identical alleles is strong evidence that all isolates belong to the same strain.

Analysis of the three Acanthamoeba keratitis isolates examined here strongly supports the previous work of Kilvington et al. (6). Both laboratories concluded that propamidine-resistant amoebae obtained during drug therapy were from the same strain as amoebae obtained prior to therapy. Thus, it is probable that resistance was due to a genetic or physiological change that occurred during therapy. The possibility that resistance developed because of preferential multiplication of resistant amoebae from a second strain originally present in a mixed infection is very unlikely. There was no evidence for a second strain in any of the isolates. The existence of multiple alleles in the infectious strain was uncommon for Acanthamoeba, but cell cloning established clearly that all three alleles are present in all amoebae in each of the isolates. Thus, only one strain was isolated from the infection. The data cannot rule out the possibility that a drug-resistant mutant amoeba or subpopulation was present at the time of infection. If this was so, however, it is clear that a significant increase in propamidine resistance was a response to therapy.