Metagenomic Deep Sequencing for the Diagnosis of Corneal and External Disease Infections

Advances in diagnostic metagenomics have potential to transform how we diagnose ocular infections and care for patients. Metagenomic deep sequencing (MDS), an unbiased approach that interrogates all genomes in a clinical sample, has been shown to enhance detection of common and unusual pathogens from the intraocular fluid of patients with infectious uveitis and other systemic infections. 1Doan T. Wilson M. R. Crawford E. D. et al. Illuminating uveitis: metagenomic deep sequencing identifies common and rare pathogens. Genome Med. 2016; 8: 90Crossref PubMed Scopus (113) Google Scholar, 2Wilson M. R. Naccache S. N. Samayoa E. et al. Actionable diagnosis of neuroleptospirosis by next-generation sequencing. N Engl J Med. 2014; 370: 2408-2417Crossref PubMed Scopus (546) Google Scholar, 3Graf E. H. Simmon K. E. Tardif K. D. et al. Unbiased detection of respiratory viruses by use of RNA sequencing-based metagenomics: a systematic comparison to a commercial PCR panel. J Clin Microbiol. 2016; 54: 1000-1007Crossref PubMed Scopus (117) Google Scholar The frequency and potential visual impact of infections of the cornea and adnexa offer the possibility to further use this technology. Unlike most systemic infections, the external location of the cornea, sclera, and conjunctiva uniquely allow for direct visualization of the nidus of infection and easier access for tissue sampling. Despite this advantage, conventional diagnostics are imperfect. Corneal culture, our present gold standard, has a sensitivity of approximately 60%. 4McLeod S. D. Kolahdouz-Isfahani A. Rostamian K. et al. The role of smears, cultures, and antibiotic sensitivity testing in the management of suspected infectious keratitis. Ophthalmology. 1996; 103: 23-28Abstract Full Text PDF PubMed Scopus (138) Google Scholar Likewise, the definitive diagnosis of infectious scleritis is frequently challenging and can mimic inflammatory scleritis. Although high-throughput sequencing appears ideal as a diagnostic assay for ocular surface infections, the microenvironment of the ocular surface also includes numerous species of normal flora, which pose inherent diagnostic challenges to this highly sensitive technology. The purpose of this proof-of-concept study is to compare the performance of MDS to present standard microbiologic testing for the diagnosis of infections of the cornea, sclera, and conjunctiva. Institutional review board approval was obtained. This single-center, retrospective case series recruited patients with clinical examinations consistent with infectious keratitis, scleritis, or conjunctivitis (Fig 1), who yielded a positive conventional diagnostic test result. The study adhered to the tenets of the Declaration of Helsinki. Written informed consents were obtained from all patients. Corneal ulcers and scleral nodules were scraped per standard clinical practice for routine diagnostic studies, including direct inoculum onto blood, chocolate, potato-flake, and non-nutrient agar with Escherichia coli overlay. Separate swabs were sent for viral polymerase chain reaction (PCR). Swabs of the affected cornea or sclera and unaffected control conjunctiva were obtained for MDS. Samples were stored at −20°C within 15 minutes of sample collection and transferred to long-term storage at −80°C within 48 hours. MDS was performed as previously described. 5Gonzales J. A. Hinterwirth A. Shantha J. et al. Association of ocular inflammation and rubella virus persistence. JAMA Ophthalmol. 2019; 137: 435-438Crossref PubMed Scopus (26) Google Scholar Briefly, total RNA was extracted from the swabs and reverse transcribed to double-stranded complementary DNA. The complementary DNA was converted to Illumina libraries (San Diego, CA) and amplified with 16 PCR cycles. The sample was sequenced on the Illumina HiSeq 4000 or NovaSeq using 150-nucleotide paired-end sequencing. Analysis of sequenced data was made using a computational pipeline developed in-house to classify sequencing reads and identify potential pathogens by aligning to the National Center for Biotechnology nucleotide database. Because the ocular surface is nonsterile, for the cornea and scleral infections, the taxa identified from the control contralateral unaffected eye were used for background subtraction. In the case where the contralateral eye control was not suitable (bilateral infection, aqueous sample from deep corneal infection, or no contralateral eye specimen), a water control on the same sequencing run was used for background analysis. An organism was considered diagnostic by MDS if it is was a pathogen known to cause ocular infections and if it represented the most abundant reads in the sample after background subtraction. Human infectious agents detected by our pipeline were manually inspected using Bowtie26Langmead B. Salzberg S. Fast gapped-read alignment with Bowtie 2. Nature Methods. 2012; 9: 357-359Crossref PubMed Scopus (26088) Google Scholar align against full-length reference genomes. Geneious (Biomatters, Ltd, New Zealand) was used for visualization. All organisms identified by culture or PCR were also detected by MDS. Metagenomic deep sequencing, as a single assay, was able to identify parasitic, fungal, bacterial, and viral infections. The identified pathogenic organisms ranged in size from smaller genomes, such as herpes simplex virus-1 and adenovirus, to larger genomes, such as Acanthamoeba and Aspergillus species. In 8 of 9 samples, the pathogens identified with conventional approaches yielded the highest number of sequencing reads with MDS, allowing for easy thresholding of signal compared with background sequencing noise. In case 3, Purpureocillium lilacinum was the second most abundant organism identified in the sample. Auricoccus indicus, with the most reads, is not known to be associated with ocular infections. This bacterium is not listed in the University of California San Francisco’s mass spectrometry’s database for identifiable organisms. In this case series, all patients were recruited from the University of California San Francisco. Eight of the 9 patients were referred with progressive disease despite treatment (Fig 1). In 8 of 9 cases, a definitive pathogen was eventually identified with viral PCR or standard culture techniques. The single patient with bilateral adenoviral conjunctivitis was diagnosed by classic clinical features. The mean time from disease onset to diagnosis was 37 days (ranging from 0 to 125 days, n = 9). Further, when atypical organisms were the cause of the infection (fungal or parasitic), the length of time to diagnosis further increased to 62 days (ranging from 25 to 125 days, n = 5). The consecutive processing time required from sample to diagnosis for MDS is approximately 5 to 7 days. Although the wait time for MDS is fairly similar to that of standard culture, the cost is noticeably higher. At present, the base reagent and sequencing costs of MDS for a single patient (2 swabs each) range from 300 to 1000 depending on the extent of parallel library processing and the type of sequencing machine used. It is anticipated, with time, costs of sequencing will continue to decrease. Especially with atypical infections, MDS has the potential to be more efficient and economical. As we investigate the diagnostic potential of MDS for ocular surface infections, a few challenges are notable. When MDS yields a diagnostic result not obtained by the current gold standard methods, independent confirmation of this result with another assay in a Clinical Laboratory Improvement Amendments–certified laboratory is currently necessary. With sensitivities approximating 60% for routine diagnostics, this will be a fairly frequent occurrence. Another challenge is how best to determine when a commensal organism, for example, Staphylococcus aureus, is indeed the causative organism of an infection. Subtracting the organisms present in the nonaffected control swabs (Table S1, available at www. aaojournal. org) may be too stringent, and other comparative sequence analysis algorithms may need to be explored. In summary, for the diagnosis of corneal, scleral, and conjunctival infections, this proof-of-concept case series demonstrates MDS can replicate the identification of causative organisms as determined by current gold standard microbiologic and molecular testing. As with any diagnostic assay, clinical correlation and assessment of treatment response remain imperative. Download. docx (. 22 MB) Help with docx files Table S1