Bacterial Biofilm Diversity in Contact Lens-Related Disease: Emerging Role of Achromobacter, Stenotrophomonas, and Delftia

Abstract

Purpose.: Multi-species biofilms associated with contact lens cases and lenses can predispose individuals to contact lens-related inflammatory complications. Our study used culture-independent methods to assess the relationship between the severity of contact lens-related disease and bacteria residing in biofilms of contact lens cases and lenses.

Methods.: Contact lens cases and lenses from 28 patients referred to the West Virginia University Eye Institute and diagnosed as having mild keratitis, keratitis with focal infiltrates, or corneal ulcers were processed and evaluated for bacterial composition based on 16S ribosomal RNA gene sequencing. Cases and lenses from nine asymptomatic contact lens wearers were processed in a manner similar to controls. Relationships between disease severity, bacterial types, and bacterial diversity were evaluated statistically.

Results.: Disease severity and presenting visual acuity correlated with an increase in the diversity of bacterial types isolated from contact lens cases. A significant difference also was observed in the number of bacterial types associated with the three clinical groups. Achromobacter, Stenotrophomonas, and Delftia were prevalent in all disease groups, and Achromobacter and Stenotrophomonas were present in one asymptomatic control. Scanning electron microscopy revealed that Achromobacter and Stenotrophomonas formed a biofilm on the surface of contact lenses.

Conclusions.: Culture-independent methods identified an association between disease severity and bacterial diversity in biofilms isolated from cases and lenses of patients with contact lens-related corneal disease. Achromobacter, Stenotrophomonas, and Delftia were predominant bacteria identified in our study, drawing attention to their emerging role in contact lens-related disease.

Introduction

Contact lens wear is a major predisposing factor to microbial keratitis and infectious corneal ulcers. ¹ Microbes can reside on lenses and within lens cases, often in association with multi-species biofilms that serve as reservoirs for the establishment of eye infections. ^2–4 From 24–81% of contact lens storage cases are contaminated with microbial biofilms, with the frequency of contamination increasing in wearers suffering from microbial keratitis (reviewed by Szczotka-Flynn et al.⁵). Contamination of contact lenses is less frequent and associated with fewer organisms than lens cases, but correlates more closely with organisms that cause corneal infections. ^3,6,7 Eye infections are initiated when lens-ocular surface interactions modify and/or compromise the epithelial surface, allowing contaminating microbes to adhere and invade. ^8–12 Despite improvements in contact lens materials, solutions, and wearing habits, epidemiological studies find the incidence of corneal infections has not changed among contact lens wearers in the last two decades. ^13,14 The sustained incidence of corneal infections indicates that microbes are able to adapt to clinical modifications designed to interfere with corneal infections. 

Effective treatment of contact lens-related keratitis and corneal ulcers is dependent on the identification of the infectious or non-infectious origin of the problem. Culture dependent methodologies identified Pseudomonas aeruginosa as the most common pathogen in contact lens-related infections, followed by Serratia marcescens, Staphylococcus aureus , Acanthamoeba spp ., and Fusarium spp. ^13,15,16 Inaccurate characterization and treatment of the offending pathogen may result in prolonged infection, permanent damage to ocular tissues, decreased vision, and in extreme cases removal of the infected tissue. ¹⁷ A major difficulty in identifying the origin of contact lens infections is that the contaminating organism often is not cultivable. ^18–20 As a result, there is reduced reliance on culturing in the diagnosis of contact lens infections, and the current recommended treatment for infections lacking a confirmed microbial diagnosis is antibiotics that target P. aeruginosa (Bacterial Keratitis Preferred Practice Pattern [PPP], 2011, American Academy of Ophthalmology, http://one.aao.org/CE/PracticeGuidelines/PPP_Content.aspx?cid=0f20807f-bc61-4f11-b570-01ae26990edb). Consequently, the etiology of the majority of contact lens-related infections remains unknown. 

The goal of our study was to gain further insight into the etiology of contact lens-related disease by using culture-independent, 16S ribosomal RNA (rRNA) gene analysis to characterize bacterial populations on contact lens cases and lenses. Our patient population was individuals with problematic anterior segment disease referred to the West Virginia University (WVU) Eye Institute for additional treatment. When contaminating bacteria were identified using 16S rRNA gene analysis, a correlation was observed between the diversity of bacterial types isolated from contact lens cases and disease severity. The bacterial genera identified most often in our study were Achromobacter, Stenotrophomonas, and Delftia, whereas Pseudomonas was identified rarely. Achromobacter, Stenotrophomonas, and Delftia exhibit physiological properties similar to Pseudomonas, but have a different antibiotic resistance pattern, which might contribute to their potential emerging role in establishing biofilms associated with contact lens-related disease. 

Methods

Subject Populations and Evaluations

Soft contact lenses and cases were obtained from 28 patients referred to the WVU Eye Institute over a five-year period for contact lens-related anterior segment disease. Patients were examined by a single clinician and assigned to one of three categories independently of bacterial culturing: (1) mild diffuse keratitis, defined as the presence of stromal inflammatory cells without an area of dense infiltration or epithelial defect; (2) keratitis with focal infiltrates, defined by the presence of stromal inflammatory cells and areas of dense accumulations of cells presenting as a focal opacity; or (3) corneal ulcer, defined as a large epithelial defect with associated suppurative infiltrate and severe generalized anterior segment inflammation. Contact lenses and cases obtained from nine asymptomatic soft contact lens wearers working in the medical university environment served as controls. Cases and lenses were frozen and maintained at −20°C until the time of analyses. All procedures were performed in accordance with the guidelines of the Declaration of Helsinki using Institutional Review Board approved protocols, and informed consent was obtained from each subject. 

DNA Extraction and 16S rRNA Gene Sequence Analysis

Biofilms were released from contact lens cases by scraping and sonication, and then combined with solutions within cases, and subjected to DNA extraction and purification procedures using the Microbial DNA Isolation Kit (Mo Bio Laboratories, Inc., Carlsbad, CA). DNA was extracted and purified from contact lenses by placing lenses directly in tubes provided by the Microbial DNA Isolation Kit, which then were subjected to the same DNA isolation procedures used for contact case samples. The bacterial content of each sample (viable and non-viable) was characterized using 16S rRNA gene sequencing and analysis, as described previously. ²¹ Briefly, bacterial 16S rRNA genes were PCR amplified using the universal bacterial 16S rRNA primers (forward, 5′-GAGTTTGATYMTGGCTCAG, and reverse, 5′-GAAGGAGGTGWTCCADCC). PCR reactions yielding products of ∼1500 base pairs (bp) were cloned using the TOPO TA Cloning Kit for Sequencing (Invitrogen, Carlsbad, CA), ligation products were electroporated into Escherichia coli (TOP10 Chemically Competent cells; Invitrogen), and transformants were plated. Following overnight incubation, 50–100 colonies containing inserts were isolated from each sample, cultured, and PCR amplified using M13 primers (forward, 5′-TGTAAAACGACGGCCAGT, reverse 5′-CAGGAAACAGCTATGAC), and products of the appropriate size were sequenced by a commercial facility. 

For 16S rRNA gene analysis, sequences were scanned first for a segment of the original primer sequence (ATCAAACT) near the expected position, and the match was confirmed by eye. The initial part of each sequence containing uncalled bases (N) was removed, the distal part (vector sequence) was trimmed at the primer, and the sequence was reversed to standard orientation. Chimeric sequences were screened using ChimeraSlayer, ²² and six chimeric sequences were identified and removed. Sequences were classified by BLASTN using the legacy executable blastall (NCBI) against a local database, as well as by using the NCBI server with the GenBank non-redundant database. Bacterial types were determined based on >98% sequence identity between the 16S rRNA gene PCR product and reference database. Bacterial classification used custom Python code (Python 2.6), and the heatmap was generated using matplotlib and Python, as described previously. ²¹  

Scanning Electron Microscopy (SEM)

Biofilm formation on contact lenses was examined by SEM on randomly selected contact lenses from patients with mild keratitis or keratitis with focal infiltrates. Lenses were cut in half, and one half was examined by SEM, and the other half was processed for 16S rRNA gene sequencing as above. For SEM, lenses were fixed in 3% glutaraldehyde in Dulbecco's PBS (DPBS; Hyclone, Logan UT) for 1 hour, then resuspended in 50% glycerol/50% deionized (DI) water and shipped to Marshall University (Huntington, WV) for analysis. Comparisons were made between post-fixing lenses in 2% osmium tetraoxide (OsO₄; Electron Microscopy Sciences, Hatfield, PA) in DI water for 2 hours or not, but no difference in image quality was evident. All lenses were dehydrated in a graded ethanol series: 50%, 70%, 85%, 90%, and 95%, followed by two changes in 100% absolute ethanol. Lenses were cut in half again before mounting, and the inside (corneal touching) or outside surface of the lens was oriented upward for imaging. Samples were dried immediately in the vacuum chamber of a sputter coater (Hummer 6.2; Anatech, Ltd, Union City, CA) and coated with <10 nm of gold/palladium (Ladd Research, Williston, VT). Lenses were imaged with a JEOL 5310-LV Scanning Electron Microscope (Japan Electron Optics Laboratory, Commerce, MI) and digital images were captured with Printerface software (GW products; SOQUELEC Ltd). 

Statistical Analysis

Statistical differences between groups were determined using a non-parametric Kruskal-Wallis H test for overall significance, which was followed by a Mann-Whitney U test for significance between two groups. Correlations between variables were tested using the Spearman rank-order coefficient (rho) or a Pearson parametric coefficient (r), as indicated. Statistical calculations were performed using Statistica (Statsoft, Tulsa, OK). 

Results

Patient Evaluations

Patients diagnosed with contact lens-related corneal disease were divided into three clinical diagnosis groups: (1) mild diffuse keratitis, (2) keratitis with focal infiltrates, or (3) corneal ulcer, as determined through retrospective chart analysis (Table 1). Patients diagnosed with mild keratitis or keratitis with focal infiltrates received prophylactic antibiotics, whereas patients diagnosed with corneal ulcer were cultured and treated with intensified fortified antibiotics, as well as daily follow-up. Presenting visual acuity was 20/20 in asymptomatic controls and the mild keratitis group, which worsened to counting fingers or hand motion in patients with severe corneal ulcers. A statistically significant (P < 0.001) difference in presenting visual acuity was observed when all groups were compared (Fig. 1), with more severe clinical categories being associated with poorer visual acuity (rho = 0.89, P < 0.001). The asymptomatic control and mild keratitis groups did not differ significantly in presenting visual acuity, but other comparisons of visual acuity between groups were significant (P < 0.05), clinically validating the groupings. 

Table 1. 

 

Patient Demographic Data

Patient	Age	Sex	Infected Eye	Presenting Visual Acuity	Location of Corneal Defect	Compliance History
Clinical group 1: Mild keratitis
2	21	M	Right eye	20/20	NA*	Slept in lenses
7	33	F	Left eye	20/20	NA	Mascara brush injury, slept in lenses
25	24	F	Left eye	20/25	NA	Lens overuse, chemical irritant exposure
27	54	F	Left eye	20/25	NA	Dry eye
47	18	M	Left eye	20/25	NA	Poor compliance with cleaning solutions
Clinical group 2: Keratitis with focal infiltrates
32	37	M	Both eyes	20/20 & 20/25	Multiple peripheral	Unknown
46	31	F	Left eye	20/25	Multiple peripheral	Unknown
52	22	M	Left eye	20/30	Several peripheral	Slept in lenses, poor lens hygiene
56	23	M	Right eye	20/30	Multiple peripheral	Lens overuse
28	22	F	Left eye	20/40	Multiple central & peripheral	Slept in lenses
62	23	F	Left eye	20/40	Several central	No lens wear for 2 weeks, inserted lens before presentation
24	29	F	Right eye	20/50	Multiple central & peripheral	Slept in lenses
15	19	F	Left eye	20/60	Few central	Lenses worn, stored 1 year then worn again
Clinical group 3: Corneal ulcer
10	48	F	Right eye	20/400	Inferior paracentral	Unknown
9	19	F	Left eye	Count fingers	Central	Unknown
57	51	F	Right eye	Hand motion	Central	Unknown
73	15	M	Right eye	Hand motion	Central	Lens overuse
Asymptomatic controls
26	25	F	NA	20/20	NA	Unknown
48	26	M	NA	20/20	NA	Unknown
55	49	F	NA	20/20	NA	Unknown
68	27	M	NA	20/20	NA	Unknown
69	27	F	NA	20/20	NA	Unknown
74	33	F	NA	20/20	NA	Unknown
75	24	M	NA	20/20	NA	Unknown
76	40	F	NA	20/20	NA	Unknown
77	32	M	NA	20/20	NA	Unknown

*  NA, not applicable.

Figure 1. 

 

Declining visual acuity is associated with contact lens-related corneal disease. In our study 17 patients presenting with contact lens-associated corneal disease and having lens cases that produced 16S rRNA gene PCR products were categorized based on disease severity as having mild keratitis, keratitis with focal infiltrates, or corneal ulcer. Presenting visual acuity was assessed for all patients and nine asymptomatic controls. *Indicates a significant difference (P < 0.05) in visual acuity between each of the three clinical groups. **Indicates an overall significant difference (P < 0.001) in visual acuity when all four groups were compared.

Relationship between Bacterial Contamination of Contact Lens Cases and Disease Severity

To assess the poly-bacterial nature of biofilms associated with contact lens-related disease, PCR amplification and sequencing of cloned 16 rRNA genes were used to identify bacteria in contact lens cases of our patient group. Of the 28 cases examined, 17 (61%) were positive for a 16S rRNA gene PCR product. The lack of PCR product in the remaining 11 cases is presumed to relate to the absence of PCR amplifiable bacterial DNA in these samples. Contact lens case biofilms and solutions collected in an identical manner from nine asymptomatic contact lens wearers served as controls. Of these samples, one (11%) produced a 16S rRNA PCR product. The increase in PCR amplifiable products from cases of patients with keratitis and corneal ulcers, as compared to control cases, is consistent with bacterial contamination of lens cases being associated with acquisition of contact lens-related disease. 

When the bacterial composition of the 18 PCR-positive contact lens case samples (17 disease and one control) was characterized, 38 different bacterial types were identified. The heatmap in Figure 2 shows the number of clones identified for each bacterial type in each sample and represents a semi-quantitative assessment of the frequency of bacterial types within the case. Variations in the total number of sequences obtained per sample reflect the number of clones subjected to sequencing and the number of clones that produced readable sequences. When the number of bacterial types identified by 16S rRNA gene sequencing was related to disease severity, a significant (P < 0.001) difference between the clinical groups was observed (Fig. 3A), with more severe clinical categories associated with an increased number of bacterial types (rho = 0.82, P < 0.001). Statistical comparisons of individual clinical groups found the number of bacterial types in the mild keratitis group to be significantly (P < 0.05) less than that of the corneal ulcer group. The asymptomatic control group also was found to have fewer bacterial types (P < 0.05) than each of the clinical groups. No other statistical differences in bacterial types were observed among the clinical groups. Independently of clinical groupings, a positive correlation (r = 0.5, P < 0.001) was observed between increasing numbers of bacterial types and decreased visual acuity. Consistent with the dynamics and complexity of the disease process, the relationship between disease severity and bacterial diversity was not absolute, as evident in one individual with normal visual acuity having a large number of bacterial types and three individuals with poorer visual acuity having a limited number of bacterial types (Fig. 3B). 

Figure 2. 

 

Identification of bacterial types in lens cases of patients diagnosed with contact lens-related corneal disease. Solutions and biofilms from contact lens cases of 17 patients with corneal disease and one asymptomatic control were evaluated for bacterial composition based on 16S rRNA gene sequencing. Patients were grouped by disease category and ranked further within categories based on worsening visual acuity (determined by increasing log arc minute values). Patient reference numbers are indicated at the top of the heatmap. Numbers within colored boxes indicate how many clones of a specific bacterial type were in each sample, and the color of the box reflects bins for observed counts. The total number of clones sequenced for each sample is indicated at the bottom of the heatmap. C, asymptomatic control.

Figure 3. 

 

Relationship between bacterial types and severity of disease. (A) The number of bacterial types identified by 16S rRNA gene sequencing was related to clinical groupings. *Indicates a significant difference (P < 0.05) in bacterial types between the mild keratitis group and corneal ulcer group. **Indicates a significant difference (P < 0.001) in bacterial types when all groups were compared. (B) Pearson correlation analysis (r = 0.50, P < 0.001) was used to examine the relationship between visual acuity of individual patients and the number of bacterial types identified by 16S rRNA gene sequencing in their contact lens cases. Graph allows recognition of outliers of the visual acuity and bacterial diversity relationship.

When specific bacterial types in contact lens cases were compared, unexpected bacteria emerged as predominant residents. Among the 38 identified bacterial types, Achromobacter had the highest frequency of occurrence (P < 0.05) with detection in 13 of 17 (76%) disease-associated contact lens cases. Stenotrophomonas was second (P < 0.05) with detection in 12 of 17 (71%) disease cases. Delftia and bacteria in the Enterobacteriaceae family were third (P < 0.05) with detection in seven of 17 (41%) disease cases. Enterobacteriaceae is a large family of gram-negative, glucose-fermenting bacteria found frequently in the gut, and in our studies included bacteria from the genera Enterobacter, Serratia, Escherichia, Ewingella, and Shigella. Notably, P. aeruginosa, recognized to be the most frequent cause of contact lens-related eye infections, was detected in only two of 17 cases. Achromobacter and Stenotrophomonas also were detected in the one PCR-positive control case at a relative frequency of 18:1, respectively, indicating that these bacteria can reside in contact lens cases in the absence of clinical symptoms. 

Together these results demonstrate that the diversity of bacterial types increases with disease severity, and that increased bacterial diversity correlates with a negative impact on visual acuity. In addition, three unexpected bacterial types, Achromobacter, Stenotrophomonas, and Delftia were identified as the predominant contaminating bacteria in contact lens cases in our study. 

Comparison of Bacterial Types Associated with Cases and Lenses

To assess the relationship between lens and case contamination in our patient group, we used 16S rRNA gene sequencing to examine bacterial contamination on lenses from 13 of the 17 lens cases analyzed in Figure 2. When two lenses were provided for analysis, bacterial results for both lenses were combined, and these results are compared to bacteria identified in the cases from the same patient in Table 2. Of the 13 contact lens samples, seven (54%) were positive for 16S rRNA gene PCR products. Lenses from asymptomatic controls were negative for PCR products. Comparisons found bacterial types on lenses to be highly similar with those in the respective lens case for patients 47, 46, 62, 57, and 73, albeit the relative percentage of the bacterial types in lenses and cases varied. The two exceptions to this relationship were: (1) Patient 2, diagnosed with mild keratitis, where bacterial types in the case were limited to Achromobacter (91%) and Stenotrophomonas (9%), while multiple diverse types, including Achromobacter and Stenotrophomonas, were detected on lenses; and (2) Patient 9, diagnosed with a corneal ulcer, where greater bacterial diversity and a different spectrum of bacteria were detected on the lenses as compared to the case. Interestingly, the spectrum of bacteria on the lenses of patient 9 reflected gut flora, whereas the case bacteria reflected more closely skin or oral flora. 

Table 2. 

 

Bacterial Types on Contact Lenses in Comparison to Cases

Patient	Clinical Group	Case	% Bacteria	Lens	% Bacteria
2	Mild keratitis	Achromobacter	91 (11)*	Achromobacter	34 (32)
		Stenotrophomonas	9	Stenotrophomonas	22
				Enterobacteriaceae	19
				Staphylococcus	6
				Xanthomonas	3
				Burkholderia	3
				Eubacterium	3
				Peptostreptococcus	3
				Selenomonas	3
				Veillonella	3
47	Mild keratitis	Achromobacter	64 (22)	Achromobacter	57 (23)
		Stenotrophomonas	36	Stenotrophomonas	43
46	Keratitis with focal infiltrate	Stenotrophomonas	93 (14)	Stenotrophomonas	94 (35)
		Achromobacter	7	Achromobacter	6
62	Keratitis with focal infiltrate	Achromobacter	79 (48)	Achromobacter	66 (94)
		Enterobacteriaceae	8	Enterobacteriaceae	18
		Delftia	6	Delftia	10
		Stenotrophomonas	6	Stenotrophomonas	6
9	Corneal ulcer	Staphylococcus	21(42)	Enterobacteriaceae	35 (23)
		Streptococcus	14	Veillonella	22
		Stenotrophomonas	10	Selenomonas	9
		Lactobacillus	7	Campylobacter	9
		Finegoldia	7	Dialister	9
		Peptoniphilus	5	Stenotrophomonas	4
		Peptostreptococcus	5	Peptostreptococcus	4
		Anaerococcus	5	Xanthomonas	4
		Pseudomonas	2	Rhizobiales	4
		Corynebacterium	2	Leptotrichia	4
		Dialister	2
		Megasphera	2
		Achromobacter	2
		Aeromonas	2
		Gemella	2
		Abiotrophia	2
		Facklamia	2
		Granulicatella	2
		Oxalobacter	2
57	Corneal ulcer	Achromobacter	53 (47)	Stenotrophomonas	56† (48)
		Stenotrophomonas	36	Achromobacter	35
		Delftia	6	Delftia	6
		Enterobacteriaceae	4	Enterobacteriaceae	2
73	Corneal ulcer	Achromobacter	38 (45)	Delftia	38 (42)
		Delftia	38	Stenotrophomonas	36
		Stenotrophomonas	18	Achromobacter	26
		Staphylococcus	4
		Enterobacteriaceae	2

*  Parentheses enclose number of sequences analyzed per sample.

†  Only left lens was analyzed.

Collectively, comparisons of bacteria associated with lenses and cases of individual patients using 16S rRNA gene sequencing found contamination of lenses to be less frequent than cases, as evident by the lack of recovery of PCR products from six of 13 PCR product-positive cases, and that the relationship between bacterial types on lenses and cases varied among patients independently of clinical diagnosis. 

SEM Analysis of Contact Lenses

To assess if bacteria identified by 16S rRNA gene sequencing exist within a biofilm, two lenses were selected randomly and cut in half, and one half was analyzed by 16S rRNA gene sequencing, while the other half was processed for SEM analysis. The 16S rRNA gene sequence analysis identified Achromobacter (57%) and Stenotrophomonas (43%) on the lenses of patient 47, diagnosed with mild keratitis. Achromobacter and Stenotrophomonas are both gram-negative rods, and SEM analysis revealed the presence of a dense film of rod-shaped bacteria on the inside and outside surfaces of the lens from patient 47 (Fig. 4A). The 16S rRNA gene sequence analysis of the lenses of patient 46 also identified Achromobacter (6%) and Stenotrophomonas (94%), and SEM analysis revealed a biofilm containing rod-shaped bacteria (Fig. 4B). These results confirm that Achromobacter and Stenotrophomonas are able to form biofilms on the surface of contact lenses of patients diagnosed with keratitis. 

Figure 4. 

 

Bacterial biofilm formation on lenses from patients with contact lens-related corneal disease. To assess that ability of bacterial populations identified by 16S rRNA gene analyses to form a biofilm on contact lenses, randomly selected contact lenses from (A) patient 47, diagnosed with mild keratitis, or (B) patient 46, diagnosed with keratitis with focal infiltrates were analyzed by scanning electron microscopy, and this was related to bacterial types identified on lenses. Lenses were cut in half to analyze the inside (corneal touching) and outside surfaces. The right panel shows a higher magnification of the area enclosed in the white square in the left panel. White arrows point to bacteria. Scale bars represent 1 or 5 μM for low magnification, and 0.5 or 1 μM for higher magnification.

Discussion

This study used 16S rRNA gene sequencing to examine bacterial populations present in cases and on lenses of patients diagnosed with contact lens-related corneal disease. The 16S rRNA gene analysis was chosen for our study to enhance the likelihood of characterizing the bacterial diversity in biofilms associated with specific disease states. Patients were categorized based on disease severity as having mild keratitis, keratitis with focal infiltrates, or corneal ulcers, and we observed a significant decrease in presenting visual acuity and a significant increase in bacterial diversity as the severity of corneal disease increased. 

Among the patients in our study, a 16S rRNA gene PCR product was obtained from 17 of 28 cases (61%). The lower than anticipated recovery of PCR products can be related to several factors. First, it is possible, particularly in milder cases of keratitis, that the disease might be caused by non-infectious agents or toxic products, and hence no bacterial DNA would be present. Second, sources other than contact lenses, such as saliva, tap water, or a fingertip, might have introduced the offending microbe into the eye. Third, it is possible that disease might have resulted from fungi, amoebae, or a viral source, which are recognized to contaminate from 4–59% of contact lens cases, ¹⁶ but would not be detected by our analysis due to the specificity of the PCR amplification primers for bacterial 16S rRNA genes. The sensitivity of PCR amplification of 16S rRNA is in the range of 5–50 bacteria, and is primer- and bacterial species-dependent. ²³ Our results are dependent on the efficiency of biofilm extraction, efficiency of DNA extraction from the isolated biofilm, and efficiency of PCR product cloning, the latter relates semi-quantitatively to the frequency of a bacterial species in a sample. It should be noted that we have not determined the efficiency of recovery of PCR product relative to number of bacteria per sample in our analysis, which precludes our ability to evaluate the sensitivity of our analysis relative to other methods of bacterial detection. 

Interestingly, when 16S rRNA gene analysis was performed on the 17 PCR-positive samples, a diverse and unexpected panel of bacteria was identified. A total of 38 different bacterial types was identified in lens cases, which was significantly higher than that obtained from asymptomatic controls. In most patients, the prevalence and diversity of gram-negative bacteria increased with disease severity, but in contrast with previous studies (reviewed by Liesegang, ¹³ and Hall and Jones¹⁶), relatively few gram-negative Pseudomonas were detected in our samples. Gram-positive bacterial diversity was evident in lens cases of two patients, 28 and 9, who presented with a clinical diagnosis of keratitis with focal infiltrates and corneal ulcer, respectively. The similar gram-positive diversity but different clinical diagnosis of these two patients may be explained by the time of collection of samples relative to the natural history of the disease. For example, patient 28 might be anticipated to have a corneal ulcer with increasing time in the absence of clinical intervention. In this regard, corneal disease caused by multi-species biofilms is dynamic and host-dependent, and consistent with this complexity the relationship between bacterial diversity and disease severity was not absolute in our study. However, it is known from studies of other biofilm infections, such as periodontitis, that disease severity increases with bacterial diversity, ²⁴ which was the statistical trend of our study. The converse of this relationship is that limited bacterial diversity will be detected in non-disease continuous wear lenses, as was reported previously. ¹¹ Based on our findings, and consistent with other biofilm infections, the severity of contact lens-related disease is predicted to increase with bacterial diversity. This notion is of clinical importance because infections caused by multiple bacterial species are problematic to treat. 

In general, bacterial contamination on lenses was detected less frequently than within cases in our study, which is consistent with previous bacterial culturing studies. ^3,16,25 Also consistent with the premise that lens cases serve as a source of contaminating bacteria leading to eye infections, in five of seven comparisons highly similar bacterial types were identified on lenses and cases. The corneal ulcer treatment regimen at the WVU Eye Institute includes the culturing of eye scrapings by the clinical laboratory. One of the four corneal ulcer patients in our study (patient 9) was culture-positive, and had a multi-species profile of Pseudomonas, diphtheroids, and coagulase-negative Staphylococci. These same bacteria were among the profile of 19 bacterial types identified in the lens case of this patient using 16S rRNA sequencing, documenting concurrence between the two methods of analysis. While the goal of our study was not to compare bacterial profiles of culture-dependent and culture-independent approaches, 16S rRNA gene analysis draws attention to the likely underestimate of bacterial species linked to contact lens-related disease when culture-dependent approaches are used. The 16S rRNA gene analysis is labor-intensive and costly, which limited the number of samples analyzed in our study and currently precludes its routine use in the diagnosis of eye disease. However, future implementation of such procedures in clinical laboratories might prove essential for the diagnosis and treatment of persistent eye infections associated with complex biofilms. 

The prevalence of Achromobacter and Stenotrophomonas in cases and lenses in our study was unexpected. Both bacteria are gram-negative, aerobic, non-glucose-fermenting rods, which can be confused clinically with Pseudomonas. ^26,27 Like Pseudomonas, Achromobacter and Stenotrophomonas are opportunistic pathogens, but differ from Pseudomonas in being less virulent and having different antibiotic resistance profiles. Achromobacter is considered an emerging pathogen and has been found as a contaminant in solutions, including disinfectant solutions. ^28,29 A limited number of Achromobacter corneal infections have been reported and are characterized as slow progressing recurrent infections, associated with a localized infiltration that can progress to cause corneal ulcers. ^30–32 Achromobacter maintains resistance to aminoglycosides, first generation cephalosporins, and variable susceptibility to fluoroquinolones. ³² Notably, these antibiotics are included in the regimen for treatment of eye infections that lack an identified causative organism (Bacterial Keratitis PPP, 2011, American Academy of Ophthalmology). The number of reported Stenotrophomonas eye infections also is limited, and the major predisposing factor to Stenotrophomonas infections is prior exposure to antibiotics. ²⁷ Stenotrophomonas often is a co-infecting organism, and has limited invasiveness and high antibiotic resistance, but unlike Achromobacter is resistant to carbapenems and can develop additional resistance during an infection. ³³ The third most frequent contaminant of lens cases in our studies was Delftia, also a gram-negative, non-glucose-fermenting opportunistic pathogen resistant to aminoglycosides and β-lactam antibiotics. A recent study in Japan identified gram-negative, non-glucose-fermenting bacteria as the most frequent contaminants of contact lens cases in a student population. ³⁴ In these studies, Stenotrophomonas was identified as the most frequent contaminant, followed by Delftia, Pseudomonas, and then Achromobacter. Similarly, a study comparing bacterial contamination of commercially available contact lens solutions found the rate of bacterial contamination to vary with different solutions, but again the most prevalent contaminating bacteria were Delftia, Stenotrophomonas, and Achromobacter. ³⁵ It also should be noted that the one asymptomatic control case in our study that produced a PCR amplifiable 16S rRNA gene product was contaminated with Achromobacter and Stenotrophomonas. In this regard, the low (11%) frequency of detection of bacterial contamination in asymptomatic controls in our study, as compared to previously reported frequencies of contamination in asymptomatic controls of 24–81%, ⁵ might relate to the efficiency of biofilm or DNA extraction in our analysis, which would compromise the lower limit of detection. 

While contamination is required for an infection, it must be accompanied by some form of corneal compromise for infection to be established. The most common compromising factor for contact lens corneal infections is overnight lens wear. ³⁶ Of the 17 patients in our study, seven were documented as having slept in their lenses or having lens overuse. A second threat to corneal infections is biofilm formation, which provides bacterial populations with resistance to antiseptics and antibiotics. ⁴ Interestingly, gram-negative, non-fermenting bacteria, such as Pseudomonas, Achromobacter, Stenotrophomonas, and Delftia have a propensity to form biofilms. ^5,34 The unexpected outcome of our study was that Achromobacter, Stenotrophomonas, and Delftia, rather than Pseudomonas, were the predominant bacteria associated with patients having contact lens-related disease. As an explanation for this outcome: (1) the source of these bacteria may relate to their contamination of contact lens solutions; (2) the survival of these bacteria in contact lens specimens can be explained by their ability to form biofilms, as was visualized for Achromobacter and Stenotrophomonas in SEM analysis; and (3) the prevalence of these bacteria over Pseudomonas in our studies can be explained by their resistance to antibiotics used empirically to treat corneal disease, noting that our patient pool was referral cases who required additional treatment. In turn, the reason that biofilms containing Achromobacter, Stenotrophomonas, and Delftia remain less of a threat in eye disease than P. aeruginosa biofilms is that these bacteria lack the arsenal of virulence factors that make P. aeruginosa infections particularly problematic. 

In summary, difficulties in culturing bacteria from individuals with contact lens-related corneal disease have led to less reliance on the identification of causative organism in the treatment of eye disease. Instead, empiric administration of antibiotics that target the most prevalent cause of severe eye infections, P. aeruginosa , is recommended and has proved clinically to be successful and economic. When 16S rRNA gene sequencing was used to characterize bacterial populations in cases and lenses of patients referred to the WVU Eye Institute for treatment of contact lens-related anterior segment disease, diverse (generally non-Pseudomonas) bacteria were identified, and the presence and severity of disease correlated with the complexity of the bacterial biofilm. Three clinically rare gram-negative bacteria, Achromobacter, Stenotrophomonas, and Delftia, emerged as frequent inhabitants of cases and lenses of patients in our study. These bacteria maintain biochemical and metabolic properties similar to P. aeruginosa, can survive in contact lens solutions, and may form the basis for biofilms that can persist and cause contact lens-related problems in the presence of prophylactic antibiotic therapy. Our study provides further support for the premise that contact lens case biofilm development and progression are associated with the presence and severity of contact lens disease, and importantly highlights the potential of the biofilm composition to adapt to antibiotic resistance pressures. 

Acknowledgments

Clinical assistance was provided by Michael McAllister and Scott Jamerson. Technical assistance was provided by LaKeisha Hall in the development of 16S rRNA gene analysis procedures. SEM analysis was performed by Michael Norton, David Neff, and Francesca Karle from Marshall University Department of Chemistry (Huntington, WV) in conjunction with the Marshall University Molecular and Biological Imaging Center. 

Footnotes

 Supported by an Unrestricted Challenge Grant from Research to Prevent Blindness to West Virginia University.

Footnotes

 Disclosure: L. Wiley, None; D.R. Bridge, None; L.A. Wiley, None; J.V. Odom, None; T. Elliott, None; J.C. Olson, None

Footnotes

¹  These authors contributed equally to the work presented here and should therefore be regarded as equivalent authors.