Acanthamoeba Keratitis in Non–ContactLens Wearers in India: DNA Typing–Based Validation and a Simple Detection Assay

Objectives  To establish that the protozoan Acanthamoeba isone of the causative organisms associated with non–contact lens–relatedkeratitis in the Indian population and to develop a simple and sensitive diagnosticassay for clinical testing.

Design  DNA sequencing of nuclear 18S and 26S ribosomal DNA motifs was performedand compared with the reference Acanthamoeba strains,to establish the genetic identity of the putative amoeba isolates obtainedfrom the corneal scrapings of non–contact lens–wearing patientswith keratitis. Ribosomal DNA typing of clinical corneal scrapings from thepatients with keratitis was performed by means of a simple agarose gel–basedmultiplex polymerase chain reaction assay, to detect the cases of Acanthamoeba keratitis.

Results  The ribosomal DNA analysis of 15 putative amoeba isolates obtained fromthe corneal scrapings of 14 patients with keratitis and 1 from the patients'environment established the isolates to be pathogenic formsof Acanthamoeba belonging to type T4 ribosomal DNA genotype. Multiplexpolymerase chain reaction assay was specific and sensitive enough to detectas low as 5 pg of Acanthamoeba DNA. Its utility asa reliable diagnostic assay was demonstrated directly with the use of 34 additionalcorneal scrapings.

Conclusions  Acanthamoeba is one of the causative organismsof keratitis in Indian patients with no history of contact lens usage. Moreover,the Acanthamoeba infection can be easily detectedin the clinical samples by means of the simple multiplex polymerase chainreaction assay based on ribosomal DNA typing.

Clinical Relevance  This study suggests the need and means to determine the incidence andprevalance of Acanthamoeba keratitis in India andelsewhere. Moreover, the polymerase chain reaction assay would help in earlyand definitive diagnosis, leading to better prognosis of Acanthamoeba keratitis condition.

Acanthamoeba keratitis hasbeen described primarily in reports from developed countries, with severalstudies suggesting soft contact lens wear as the greatest risk factor.^1-4 Incontrast, the reports from India and other developing countries are few andhave mainly been in non–contact lens wearers.⁵ Thisapparent low incidence of Acanthamoeba keratitisin India has resulted from the belief that the disease is related mainly tocontact lens wear—a factor usually absent in most cases of keratitisfrom this part of the world—and also from the unavailability of simpleand sensitive diagnostic tools for its clinical detection. A consequence ofthis low reporting is the skeptical acceptance of Acanthamoeba as a pathogenic organism by the medical community as well as the healthauthorities in India. Although our group⁵ reporteda number of cases, all of these were based on fluorescence microscopy of thecorneal scrapings and culture on nonnutrient agar with Escherichia coli overlay. These techniques involving culture establishmentand fluorescence microscopy require considerable experience and sophisticatedinfrastructure that are not available to most ophthalmologists, and thus dataremain lacking from other parts of India. In comparison, direct detectionmethods based on routine light microscopy of smears and histologic preparations,although widely available to clinicians, are not only less sensitive but alsounreliable. It has been estimated that 60% to 70% of cases of Acanthamoeba keratitis are misdiagnosed by such methods.⁶ Thesituation thus calls for newer, cost-effective, sensitive, simple, and reliablediagnostic tools that can be easily integrated in a small to medium-sizedclinical setup, leading to more realistic estimates of the disease incidenceas well as helping early diagnosis and treatment of the disease for betterresponse from the patients.

In this study, we provide DNA-based evidence that Acanthamoeba is associated with keratitis in Indian patients with nohistory of contact lens usage. In addition, we report a simple, rapid, andsensitive agarose gel–based multiplex polymerase chain reaction (PCR)assay for reliable detection of Acanthamoeba in clinicalsamples.

Twenty-two amoeba isolates obtained from the corneal scrapings of non–contactlens–wearing patients who had nonbacterial and nonfungal keratitis wereused for DNA analysis–based genetic identity. These comprised 15 isolatesfrom India (14 from patients with keratitis and 1 from potable water fromthe home of one of the patients), 1 each from Pakistan and Argentina, 4 AmericanType Culture Collection (ATCC) strains, and 1 standard strain of Acanthamoeba culbertsoni. Of the 14 keratitis samples from India, 12(Table 1) were collected at L.V. Prasad Eye Institute, Hyderabad, and the other 2 were from Sankara Nethralaya,Chennai. All isolates were grown axenically as monolayers in PYG (proteasepeptone–yeast–glucose) medium.

Initial standardization of a PCR-based diagnostic assay was performedwith the samples used for molecular validation described in the precedingparagraph. Subsequently, the assay was tested for its reliability and utilityin clinical diagnosis directly on corneal scrapings from 34 patients diagnosedand treated at L. V. Prasad Eye Institute from January 1, 2000, through October31, 2001. These samples were selected from among the 2213 cases of microbialkeratitis with no history of contact lens usage. All the 34 corneal scrapingsamples were collected in duplicate, of which one set was stored in phosphate-bufferedsaline and the other set was directly investigated for the causative organismby our laboratory-specific protocol⁵ that includesfluorescence microscopy. The latter investigation showed 25 corneal scrapingsto be culture positive for Acanthamoeba and 9 forbacteria or fungus.

The clinical data and results of microbial investigations of all theamoeba-positive samples, except for those obtained from outside India, aregiven in the Table 1. The studywas approved by the institutional ethics committee of L. V. Prasad Eye Institute.

For ribosomal DNA (rDNA) analysis, totalDNA was extracted from the amoebic cultures and corneal scrapings by meansof the UNSET (urea–NaCl–SDS–EDTA–Tris hydrochloride)lysis buffer method.⁷ To isolate DNA from cultures,amoebae were harvested from 4- to 5-day-old, 5-mL confluent cultures (approximately1 × 10⁶ amoebae, containing approximately 9:1 ratio of trophozoitesand cysts) by centrifugation at approximately 1000g for5 minutes. The harvested cells were washed twice with 5 mL of phosphate-bufferedsaline and resuspended in 0.5 mL of UNSET lysis buffer for DNA isolation.The aqueous lysate was extracted twice with 0.5 mL of phenol–chloroform–isoamylalcohol (25:24:1). The DNA was finally precipitated with 0.1 vol of 3M sodiumchloride and 2 vol of ethanol and dissolved in 200 µL of 1× TEbuffer (10mM Tris hydrochloride, 1mM EDTA, pH 8.0). For corneal scraping samples,a small amount of corneal tissue (approximately 0.2-0.5 mm) scraped by meansof a surgical blade (No. 15 on Bard Parker handle) with the patient undertopical anesthesia was collected in 1 mL of phosphate-buffered saline andstored at –20°C until testing. For DNA isolation, the stored cornealscraping samples were pelleted, lysed in 0.5 mL of UNSET lysis buffer, andthen processed as described in the preceding sentences, except that the organic-phaseextraction was done only once and final DNA was dissolved in only 50 µLof 1× TE buffer. All centrifugations for DNA isolation were carriedout at 8000g for 8 to 10 minutes in a centrifuge(Biofuge Fresco; Heraeus Instruments, Osterode, Germany).

Each sample was used to amplify 5 genus- or pathotype-specific genomicregions based on the published literature, of which 2 targets specific to18S rDNA⁸ and 26S rDNA⁹ giving463–base pair (bp) and 126-bp fragments, respectively, were found tobe most promising for characterization as well as diagnosis of Acanthamoeba. Primer sequences used to amplify 18S rDNA domain wereas follows: 5′-GGCCCAGATCGTTTACCGTGAA-3′ and 5′-TCTCACAAGCTGCTAGGGGAGTCA-3′(oral communication, Thomas J. Byers, PhD, Ohio State University, Columbus), whereas those for 26S rDNA target were as follows: 5′-GGAGCTCCCACGGGAGGCC-3′and 5′-TGGACCGCGTGAGGCTGCGGCT-3′, as described by Lai et al.⁹ Amplified products were sequenced on an automatedDNA sequencer (ABI-3700; Applied Biosystems, Foster City, Calif) and alignedwith those of reference sequences (identified by BLAST [basic local alignmentsearch tool] search and retrieved from the EMBL [European Molecular BiologyLaboratory] database) using ClustalX (available at: http://www-igbmc.u-strasbg.fr/BioInfo/ClustalX/Top.html). The sequence data were then used to construct phylogenetic treesto infer the genetic identities of the amoeba isolates analyzed in the presentstudy. Moreover, to ascertain the confidence values of the rDNA sequence–basedinferences regarding the genetic identities, sequence data were subjectedto bootstrap analysis.¹⁰ For the latter purpose,the aligned sequences were resampled 100 times by means of SEQBOOT softwareand the resampled data set was then used to compute genetic distances (DNADIST)and consensus distance trees showing genetic relationships by means of unweightedpair group method using arithmetic averages followed by CONSENSE programs.All programs were run with the use of PHYLIP package 3.6, which also containssimplified documentation files of each of these sequence analysis programsused in the study (available at: http://evolution.genetics.washington.edu/phylip.html).

For rDNA typing of isolates as well as for diagnostic assay, the PCRswere done in 20-µL reactions using 1 and 3 µL of template DNA(for amoebae and corneal scraping samples, respectively). Each PCR contained1 U of Taq DNA polymerase (AmpliTaq Gold; PerkinElmer,Inc, Boston, Mass), 2pM of each primer, 200µM dNTPs (deoxyribose nucleotidetriphosphates), and 1.5mM magnesium chloride and was amplified for 35 three-stepcycles of 94°C, 61°C, and 72°C, each for 1 minute, followed byfinal extension of 72°C for 5 minutes in an MJR PTC-200 thermocycler (MJResearch Inc, Waltham, Mass). In each PCR amplification, the cycling profilewas preceded by one heating step of 94°C for 10 minutes to activate the Taq DNA polymerase. To develop a direct gel-based assayfor Acanthamoeba detection in clinical samples, primersfor the 2 rDNA targets were tested individually and also in combination, initiallywith DNA from culture-proved and "Type" isolates of Acanthamoeba and then with the use of DNA from 34 corneal scraping samples frompatients with keratitis. The amplified products were resolved and visualizedon ethidium bromide–stained 1.5% agarose gel. In all the experiments,proper negative (water in place of template DNA) and positive (DNA from Acanthamoeba castellanii, ATCC 50370) controls were used.Moreover, specificity of the assay was tested with DNA from 2 Acanthamoeba-related protozoa, Balamuthia mandrillaris and Hartmanella vermiformis (ATCC 30966),and human leukocytes, Pseudomonas aeruginosa, Aspergillus species, and herpes simplex virus. The sensitivitywas determined by testing 10-fold dilutions of A castellanii DNA.

All the rDNA sequences specific to 22 amoeba isolates analyzed in thestudy were deposited in GenBank (accession numbers AF-534135 to AF-534179).The rDNA-based phenogram (Figure 1)showed all the Indian isolates, including the environmental one, and thosefrom Argentina and Pakistan, to be Acanthamoeba havingtype T4 sequence, the type found most commonly associated with the pathogenicisolates causing keratitis in contact lens wearers.^7,8,11 Thegenetic identities were supported by very high bootstrap values that indicatedthe robustness and reliability of the rDNA sequence data obtained in the studyfor establishing the genetic affiliation of the analyzed amoeba isolates.Within themselves, the Indian isolates showed considerable variation, suggestingmultiple Acanthamoeba species, as was also supportedby their detailed morphometric analysis (data not shown). These results arein conformity with the original work of Gast et al,⁷ whichdescribed the rDNA sequence types and showed that the T4 type sequence characterizesa heterogeneous group of pathogenic isolates of Acanthamoeba comprising many different species.

The attempts to develop a reliable diagnostic test led to the multiplexPCR assay having high sensitivity (5 pg of Acanthamoeba DNA) and specificity for the agarose gel–based detection of Acanthamoeba in clinical samples (Figure 2). The assay, being based on 2 markers (18S and 26S rDNA),provides relatively more reliable detection than the one described earlierbased on only 18S rDNA^8,12 andalso overcomes the limitations of using different amplification conditionsfor sensitivity and specificity.⁸ The 18S rDNAprimers that amplify a 463-bp amplicon from Acanthamoeba also produced a similar-sized amplicon for Balamuthia and Hartmanella, the 2 closely related protozoa,under the conditions used in our studies. However, in the multiplex assay(Figure 2) wherein primers specificto both 18S and 26S rDNAs were used together, the 463-bp fragment was notamplified from the related protozoa and only the 126-bp fragment was seenfor Balamuthia. In contrast, a double-band patterncomprising 463 and 126 bp was obtained only with Acanthamoeba. The Acanthamoeba-specific 2-fragment phenotypemay not indicate the presence of Balamuthia in thetest sample if the latter is also present as an additional protozoan, althoughkeratitis cases having such mixed protozoa infections are yet to be defined.Experiments done to test the diagnostic utility of the assay, using 34 clinicalcorneal scraping samples and DNA from 22 Acanthamoeba isolatesand many nontarget organisms as negative controls (see earlier), resultedin only 1 false-negative finding, establishing the high specificity (>98%)of the assay. On the other hand, the 2-fragment phenotype on the agarose gelspecifying Acanthamoeba was seen for 24 of the 25culture-proved, Acanthamoeba-positive corneal scrapingsamples, showing high accuracy of positive identification (96%), and for noneof the 9 corneal scrapings of patients with bacterial or fungal keratitis.

Perusal of the clinical data of the patients from whom the study sampleswere drawn did not suggest any bias with respect to the sex and age of thepatients (Table 1) and the incidenceof Acanthamoeba keratitis. Notably, all the patientsbelonged to relatively poor strata of society, none wore contact lenses, andfor the few cases with known history, mechanical injury or trauma was foundto be the main predisposing factor.

Results presented herein conclusively prove that Acanthamoeba is one of the causative organisms of keratitis in non–contactlens wearers from India. Ribosomal DNA typing–basedgenetic identity analysis showed a high degree of genetic diversity amongthe Acanthamoeba isolates(Figure 1), and all of these carried the T4 signatures that definethe pathogenic forms.⁷ These results emphasizethat ophthalmologists and general physicians in India and other developingcountries should be aware of Acanthamoeba as oneof the causes of keratitis in non–contact lens wearers, and thus shouldtake immediate measures for its early diagnosis and management. For the latterpurpose, the simple PCR-based test described herein holds great promise, especiallyafter such technology has become widely available in most parts of India andelsewhere. Even creation of a new setup for PCR-based testingis much easier and cost- and time-effective than the other relatively reliableapproaches of fluorescence microscopy using calcofluor white staining⁵ and culture-based methods.

Awareness about the disease and availability of an easy detection toolwould not only help in early detection and better management of Acanthamoeba keratitis, a condition that presently largely goes undetectedand uncured at most centers, but would also help greatly in obtaining realisticestimates of disease incidence. Thus, the diagnostic and epidemiologic valueof the multiplex PCR cannot be overemphasized for countries like India wheresuch data are lacking.

Correspondence: Ramesh K. Aggarwal, PhD, Centre for Cellular andMolecular Biology, Uppal Road, Tarnaka, Hyderabad 500 007, India (rameshka@ccmb.res.in).

Submitted for publication November 14, 2002; final revision receivedOctober 31, 2003; accepted February 25, 2004.

This study was supported by the Department of Biotechnology, Governmentof India, New Delhi.

Hajib N. Madhavan, PhD, Vision Research Foundation, Sankara Nethralaya,Chennai, India, provided 2 clinical isolates of Acanthamoeba; Thomas J. Byers, PhD, Ohio State University, Columbus, provided 18SrDNA primer sequences and ATCC-typed strains of Acanthamoeba; Gerald L. McLaughlin, PhD, Indiana University School of Medicine,Indianapolis, provided clinical isolates of Acanthamoeba from Argentina and Pakistan; Govind S. Visvesvara, PhD, Centers forDisease Control and Prevention, Atlanta, Ga, provided the culture of B mandrillaris; and Prashant Garg, MS, Cornea Services,L. V. Prasad Eye Institute, Hyderabad, India, provided the clinical data ofthe patients.