Antimicrobial Action of Biguanides on the Viability of Acanthamoeba Cysts and Assessment of Cell Toxicity

Abstract

Purpose.: To assess dose- and concentration-dependent rates of biguanides on the viability of Acanthamoeba cysts isolated from severe ulcerative keratitis, and to correlate cysticidal activites with cytotoxic profiles in corneal and endothelial cells.

Methods.: Cysticidal activities of polyhexamethylene biguanide and chlorhexidine digluconate were evaluated in the Acanthamoeba castellanii strain and clinical isolates of Acanthamoeba spp obtained from two severe and recurrent cases of ulcerative keratitis. The molecular characterization of protozoa used in the experimental assays was performed by sequencing reactions of the 18S rDNA gene. Acanthamoeba cysts were exposed at different dosages and concentrations of both biguanides; the application of double-biguanides was also evaluated. Automated cell viability assessment of cysts was performed using the trypan blue dye exclusion method. Cytotoxicity assays of biguanides were conducted using primary cultures of endothelial cells alone or in coculture with Acanthamoeba cysts. Human corneal epithelial cells were used as a comparative pattern to assess the toxicity of biguanide compounds. Cell viability was measured using both quantitative and qualitative methods. Statistical analyses were applied to the data.

Results.: The in vitro study showed that all dosages, concentrations, and combinations of biguanides tested had a cysticidal effect on Acanthamoeba spp strains tested compared with control cultures not exposed to any antimicrobials; the difference in response was statistically significant. The use of both biguanides in combination demonstrated the best cysticidal effect. The use of isolated biguanides was associated with greater cytotoxic effects than with biguanides used in combination. Chlorhexidine digluconate used alone tended to have greater cytotoxicity than polyhexamethylene biguanide. Furthermore, the double-biguanide application had a statistically significant decrease in the deleterious effect on endothelial cells at higher dosage and concentration. Quantitative and qualitative analyses demonstrated the toxic effect of biguanide compounds on the viability of corneal epithelial cells, under single or in combination usage.

Conclusions.: We demonstrated that the combined use of biguanides had greater cysticidal activity than individual drug application as well as a possible protective effect on endothelial cells. The biguanide compounds tested were able to induce corneal epithelial cell death in time and concentration-independent fashions. Findings support the hypothesis concerning the cysticidal effect and the differential patterns of toxicity expressed by polyhexamethylene biguanide and chlorhexidine digluconate on the endothelial and corneal cells.

Introduction

Amoebae of the genus Acanthamoeba are free-living protozoa with wide dispersion in aquatic and terrestrial environments. ¹ Besides the planktonic form, Acanthamoeba spp are able to adhere to microbial biofilms providing symbiotic and pathogenic interactions. ² The life cycle of the protozoan comprises a trophozoitic stage, which is characterized by locomotion, proliferation, and feeding, and a resistant encysted stage. ³ Certain Acanthamoeba species are associated with human disease, of which keratitis is the most common pathology. ⁴  

Acanthamoeba keratitis (AK) is a painful and sight-threatening corneal disease, the pathogenesis of which is characterized by three primary biologic events: trophozoite adhesion to the epithelial surface, invasion of the anterior stromal layer, and degradation of stromal extracellular matrix components. ^5,6 Despite a lack of standardization, the therapeutic regimen applied to AK has employed the topical application of chemical antimicrobial compounds, such as biguanides and diamidines, periods varying from weeks to months. ⁷ However, encysted stage of the protozoan promotes resistance to chemical therapy, and recurrence of infection is commonly observed. ^{8 –10} As a result, failure of topical therapy is common. ⁵ Patients who do not respond to topical therapy or who experience recurrence of the disease are often subjected therapeutic or optical keratoplasty. ^11,12  

The therapeutic regimen for AK has not been standardized in Brazil, and monotherapy has been the first-line procedure in the treatment of the infection. Diamidine compounds are not approved for ocular therapeutic use in Brazil, although biguanides have been approved for disinfection by Brazilian regulatory agency (Brazilian Health Surveillance Agency, Anvisa). The purpose of this study was to compare the efficacy of chlorhexidine (CLX), a bisbiguanide antiseptic, and polyhexamethylene biguanide (PHMB), a polymeric compound with disinfectant and antiseptic properties, on the viability of Acanthamoeba cysts, and to correlate the viability of protozoa with the concentration and dose-dependent effect of each biguanide evaluated. Furthermore, the toxicity of both biguanides at higher concentration was investigated using in vitro assays in endothelial and corneal cells in order to determine the cytotoxicity of single versus combined biguanide use. 

Materials and Methods

This study was approved by the local ethics committee (the Federal University of São Paulo [UNIFESP] 0344/12HE) and was conducted in full compliance with the current biosecurity standards. Acanthamoeba cysts were obtained from corneal tissue surface of two female patients, contact lens wearers, whose primary clinical symptoms were contact lens wear intolerance, pain, photophobia, and tear production. The clinical isolates of Acanthamoeba spp were designated as isolate 01 and isolate 02. The degree of corneal infection from both clinical cases was categorized as severe, using criteria described previously by Vital et al. ¹³ Collection of clinical samples was conducted in accordance with the tenets of the Declaration of Helsinki, and prior informed consent was obtained. Primary isolation of Acanthamoeba cysts was obtained on nonselective agar medium, used in the laboratory diagnostic routine to isolate free living amoeba from corneal samples, as described previously. ^14,15 The reference strain Acanthamoeba castellanii from the American Type Culture Collection (ATCC 30011) was used as experimental control. The average diameter of Acanthamoeba cysts from isolate 01 and isolate 02 was 17.35 and 16.55 μm, respectively, providing preliminary morphologic information to characterize Acanthamoeba cysts from isolate 01 and isolate 02 as members of group II, according the classification defined by Pussard and Pons. ¹⁶ Molecular characterization of clinical isolates of Acanthamoeba spp was based on the nearly complete 18S rRNA gene sequence. The reference strain A. castellanii (genotype T4, ATCC 30011) and Milli-Q ultrapure water (Milli-Q; Millipore GmbH, Eschborn, Germany) were used as positive and negative controls, respectively. Genomic DNA was extracted with an UltraClean Tissue and Cells DNA Isolation Kit (MoBio Laboratories, Carlsbad, CA) and the nucleic acid content was quantified by spectrophotometry at 260 nm using a BioPhotometer (Eppendorf, Hamburg, Germany). The purity of the DNA extracted was standardized in 1.8 considering the ratio of A260/A280. PCRs were performed using five sets of universal eukaryotic primers, as shown in Table 1. Amplicons were generated in a 50 μL reaction volume containing 1× Q5 Hot Start High-Fidelity Master Mix (New England Biolabs, Beverly, MA), both forward and reverse primers (0.2 μM) and DNA template at final concentration of 1 ng/μl. The standardized amplification profile was performed with an initial denaturation step at 98°C for 2 minutes followed by 30 cycles at 98°C for 30 seconds (denaturation), 60°C for 30 seconds (annealing), 74°C for 1 minute (extension), and a final extension at 74°C for 6 minutes. PCR products to be sequenced were purified with the PureLink PCR Purification Kit (Invitrogen Life Technologies, Carlsbad, CA). Both strands of each amplicon were sequenced and data were processed with the BioEdit Sequence Alignment Editor version 7.2.0 (Ibis Biosciences, Carlsbad, CA). ²¹ The nucleotide sequences obtained were compared to the GenBank data library with BLAST software. ²² The 18S rRNA gene sequences were deposited in the GenBank database under accession numbers KF318460 to KF318462. 

Table 1. 

 

Universal Eukaryotic Primer Sets Used for Amplification and Sequencing Reactions of 18S rRNA Gene

Primer Designation	Sequence, 5′–3′	Amplicon, bp	References
Euk-A	AACCTGGTTGATCCTGCCAGT	680	17, 18
570R	GCTATTGGAGCTGGAATTAC
373F	GATTCCGGAGAGGGAGCCT	1200	19, 20
1262R	GAACGGCCATGCACCAC
570F	GTAATTCCAGCTCCAATAGC	870	18, 19
1137	GTGCCCTTCCGTCAAT
570F	GTAATTCCAGCTCCAATAGC	960	18
1055R	CGGCCATGCACCACC
1055F	GGTGGTGCATGGCCG	730	17, 18
Euk-B	GATCCTTCTGCAGGTTCACCTAC

Acanthamoeba Axenization Assay

Acanthamoeba cysts were subjected to an axenization process aseptically by picking up a piece of agar culture containing amoebae cysts from each corneal sample and transferring it to a tissue culture flask containing 5 mL of peptone-yeast extract-glucose (PYG) broth medium. ²³ After 48 hours of incubation at 25°C, the excystment process was observed followed by growth of trophozoites in a monolayer. Cell supernatants were discarded and encystment of adherent trophozoites was performed in fresh PYG medium culture. Viable Acanthamoeba cysts were harvested and washed in diluted saline solution twice by centrifugation at 405g for 10 minutes. The washing process was followed by cell counting and quantitative standardization of both total and viable cysts, determined by trypan blue staining in a countess automated cell counter (Invitrogen Life Technologies), following the manufacturer's instructions. 

Viability Assays of Biguanides on Acanthamoeba Cysts

Assays were designed to evaluate both dose- and concentration-dependent activities of PHMB and CLX in Acanthamoeba cysts. Both PHMB and CLX eye drop solutions were purchased from Ophthalmos Pharmaceutical Laboratory (São Paulo, SP, Brazil). In experimental assays, the biguanide solutions were administered in the same manner used in the therapeutic clinical regimen; thus, drops of both chemical compounds were considered as dosage volumes. Assays were carried out in two 24-well plates: microplate A, for cysticidal assays with PHMB and microplate B, for cysticidal assays with CLX. Eighteen and a half microliters of each cyst inoculum were added to each well of the 24-well plates containing 500 μL of fresh PYG medium. Plates were maintained at 25°C for 48 hours and monitored daily for excystment or amoeba growth. Dose- and concentration-dependent activities of PHMB and CLX in Acanthamoeba cysts were carried out as described above. 

Microplate A (PHMB): 1 drop at 0.02%; 2 drops at 0.02%; 1 drop at 0.04%; 2 drops at 0.04%; 1 drop at 0.02% associated with 1 drop of CLX at 0.02%; 1 drop at 0.04% associated with 1 drop of CLX at 0.04%, and 2 drops at 0.04% associated with 2 drops of CLX at 0.04%. The microplate wells containing cysts inoculum of ATCC 30011 strain or clinical isolates of Acanthamoeba spp in the absence of CLX and PHMB were considered as positive experimental controls. 

Microplate B (CLX): 1 drop at 0.02%, 2 drops at CLX a 0.02%, 1 drop at 0.04%, 2 drops at 0.04%, 1 drop at 0.02% associated with 1 drop of PHMB at 0.02%, 1 drop at 0.04% associated with 1 drop of PHMB at 0.04%, and 2 drops at 0.04% associated with 2 drops of PHMB at 0.04%. The pattern of positive experimental controls was the same described to microplate A. 

The volume of a drop of PHMB instilled into each well was measured as 22 μL, which corresponded to final concentrations of 0.0008% and 0.0016% for PHMB solutions assayed at concentrations of 0.02% and 0.04%, respectively. Moreover, the final concentrations of each well assayed with two drops of PHMB at 0.02% and 0.04% were 0.0016% and 0.0032%, respectively. In the experimental assays with CLX, the volume of one drop instilled into each well was 35 μL, which corresponded to final concentrations of 0.001% and 0.002% for CLX solutions provided at 0.02% and 0.04%, respectively. In addition, final concentrations of well plates assayed with two drops of CLX solutions at 0.02% and 0.04% were 0.0024% and 0.0048%, respectively. After 48 hours at 25°C, cysts suspension of both microplates containing PHMB and CLX were harvested, and cell viability from each well was measured twice employing the trypan blue exclusion method with a countess automated cell counter (Invitrogen Life Technologies), following the manufacturer's instructions. 

Cytotoxicity Assays in Human Umbilical Vein Endothelial Cells (HUVECs)

HUVECs were obtained as previously described. ²⁴ In order to evaluate the effect of biguanides and the presence of Acanthamoeba cysts on cell viability, HUVECs (2 × 10⁴) were seeded on 96-well microtiter plates coated with 2% gelatin and incubated for 48 hours at 37°C in a 5% CO₂ humidified incubator in RPMI 1640 supplemented with 10% fetal bovine serum (FBS), containing a dilution of 1:4 of 0.04% PHMB, 0.04% CLX, or their association, in the presence and absence of both clinical isolates (isolate 01 and isolate 02) and ATCC 30011 Acanthamoeba cysts, in the proportion of 2:1 (cysts:HUVECs). Assays were carried out in triplicate, and the cytotoxic activeties of PHMB and CLX were measured using the PrestoBlue cell viability reagent (cat#A13262; Invitrogen Life Technologies) according to the manufacturer's instructions. Measurements of cytotoxicity were based on resazurin reagent that functions as a metabolic indicator in living cells. After 48 hours under PHMB and CLX activity and co-incubation with Acanthamoeba cysts, the supernatant contents were harvested from each well and adherent HUVEC cells were incubated for 1 hour with PrestoBlue reagent (cat#A13262; Invitrogen Life Technologies) diluted in fresh RPMI medium (1:10), in order to provide the resazurin reduction assay. Measurements were recorded as the difference between the absorbance at 570 and 600 nm. In both assays, values obtained were normalized with a blank well without cells, and expressed in function of non-treated control cells, which were incubated in absence of biguanides and Acanthamoeba cysts, regarded as 100% viable cells. For fluorescence staining, HUVEC cells were incubated, as described above, with biguanides and/or Acanthamoeba cysts, and stained with acridine orange (AO, 2 μg/mL) and propidium iodide (PI, 2 μg/mL). Differential interference contrast (DIC) images were acquired at 100× of magnification using an inverted Nikon-Ti microscope (Nikon, Shinjuku, Tokyo, Japan) and were compared with control cells to analyze positive nucleus staining and endothelial cell morphology. 

Cytotoxicity Assays in HCECs

Human corneal epithelial cells (HCECs) were purchased from Gibco-Invitrogen Corp. (Grand Island, NY) and cultured in keratinocyte serum free medium (K-SFM) supplemented with bovine pituitary extract and recombinant human epidermal growth factor (Gibco-Invitrogen Corp.). In order to evaluate the effect of biguanides on cell viability, HCECs (2 × 10⁴ cells) were seeded on 96-well microtiter plates and incubated for 12, 24, and 48 hours at 37°C in a 5% CO₂ humidified incubator in K-SFM medium, containing a dilution of 1:4 of 0.02% PHMB, 0.02% CLX, or their association, as well as 0.04% PHMB, 0.04% CLX, or their association. Assays were carried out in triplicate, and the cytotoxic activities of PHMB and CLX were measured using the PrestoBlue cell viability reagent (cat#A13262; Invitrogen Life Technologies) according to the manufacturer's instructions. Measurements of cytotoxicity were based on resazurin reagent that functions as a metabolic indicator in living cells. After incubation for 12, 24, and 48 hours, the supernatant contents were harvested from each well and adherent HCEC cells were incubated for 1 hour with PrestoBlue reagent diluted in fresh K-SFM medium (1:10) in order to provide the resazurin reduction assay. Measurements were recorded as the difference between the absorbance at 570 and 600 nm. Values obtained were normalized with a blank well without cells, and expressed in function of non-treated control cells, which were incubated in absence of biguanides, regarded as 100% viable cells. For fluorescence staining, HCEC cells were incubated, as described above, with biguanides and stained with AO (2 μg/mL) and PI (2 μg/mL). Photomicrographs were acquired at 100× of magnification by epifluorescence using a Nikon eclipse Ti-S microscope equipped with a Nikon DS-Ri1 digital camera (Nikon). Images were individually captured, transported to a computer with Nikon NIS element software (Nikon) and compared with control cells to analyze positive nucleus staining and cell morphology. 

Statistical Analyses

For data analyses of the in vitro assays, the Kolmogorov-Smirnov test was used to verify the normality among samples. The mean value of all three measurements was calculated and compared. From the results of this test, the nonparametric Kruskal-Wallis test was applied in order to compare data of three or more groups from both cysticidal and cytotoxicity assays. The comparison of results in pairs was analyzed by the nonparametric Jonckheere-Terpstra test. Data were analyzed using the statistical package SPSS for Windows, version 19.0 (SPSS, Chicago, IL) and R-project version 2.13.1 (R Project for Statistical Computing, Wirtschaftsuniversität, Vienna, Austria; http://www.r-project.org). Continuous variables were presented as mean ± SD. In all cases, the significance level for rejection of the null hypothesis was set at 5%. 

Results

The molecular characterization of both clinical isolates of Acanthamoeba spp was based on a fragment of 2200 base pairs of the 18S rRNA gene. Isolate 01 (GenBank accession number KF318460) revealed 99% of sequence identity to A. castellanii CDC:0786:V042 strain (ATCC 50493, GenBank accession number U07403), while isolate 02 (GenBank accession number KF318461) showed 99% of sequence identity to Acanthamoeba rhysodes BMC:0685:116 strain (ATCC 50368, GenBank accession number U07406). The results of morphologic aspects associated with the 18S rDNA sequence analysis demonstrated that the reference strain and both clinical isolates assayed were characterized within the Acanthamoeba genus as members of group II and genotype T4. 

The in vitro experiment of cysticidal activities of biguanides, alone and in combination, showed no statistically significant difference between the ATCC 30011 strain and clinical isolates of Acanthamoeba spp, with the exception of one experimental group (Table 2). The instillation of PHMB in conjunction with CLX, considering two drops at concentration of 0.04%, resulted in the highest rate of cysticidal activity in both ATCC strain and clinical isolates (Fig. 1). The percentage of viable cysts in the other experimental groups showed a wide range of variation, which could be related to the differential resistance patterns among Acanthamoeba cysts assayed (Fig. 1). 

Figure 1. 

 

Susceptibility of Acanthamoeba cysts to different profiles of PHMB and CLX. ATCC 30011: A. castellanii strain from ATCC. Isolate 01: clinical isolate of Acanthamoeba spp number 01. Isolate 02: clinical isolate of Acanthamoeba spp number 02.

Table 2. 

 

Percentual Viability of Acanthamoeba Cysts After 48 Hours in Contact With Different Concentrations and Doses of Both Individual and Associated Instillations of PHMB and CLX

Biguanide Concentration (Dose)	Cyst Viability, %			P §
	ATCC 30011*, Mean ± SD	Isolate 01†, Mean ± SD	Isolate 02‡, Mean ± SD
PHMB 0.02% (1 drop)	35.5 ± 13.4	25.0 ± 11.3	62.0 ± 2.8	0.156
PHMB 0.02% (2 drops)	33.0 ± 14.1	20.0 ± 0.0	49.0 ± 14.1	0.148
PHMB 0.04% (1 drop)	33.0 ± 25.5	15.5 ± 7.8	50.0 ± 8.5	0.276
PHMB 0.04% (2 drops)	26.5 ± 4.9	10.0 ± 9.9	36.5 ± 6.4	0.102
CLX 0.02% (1 drop)	30.0 ± 31.1	26.5 ± 4.9	57.0 ± 2.8	0.18
CLX 0.02% (2 drops)	16.0 ± 1.4	38.5 ± 54.4	23.0 ± 0.0	0.555
CLX 0.04% (1 drop)	17.5 ± 6.4	9.5 ± 4.9	43.0 ± 2.8	0.123
CLX 0.04% (2 drops)	0.5 ± 0.7	9.0 ± 7.1	13.0 ± 0.0	0.171
PHMB 0.02% + CLX 0.02% (1 drop)	23.3 ± 11.3	9.5 ± 1.7	30.8 ± 5.6	0.06
PHMB 0.04% + CLX 0.04% (1 drop)	22.8 ± 12.6	21.8 ± 19.0	17.0 ± 4.6	0.944
PHMB 0.04% + CLX 0.04% (2 drops)	1.5 ± 1.3	8.5 ± 5.4	11.5 ± 9.0	0.030‖
Positive experimental control	82.0 ± 7.7	92.0 ± 2.9	93.5 ± 4.7	0.075

*   A. castellanii from ATCC. Initial amount of viable cysts in 18.5 μL = 4.7 × 10⁵.

†  Clinical isolate of Acanthamoeba spp number 01. Initial amount of viable cysts in 18.5 μL = 3.2 × 10⁵.

‡  Clinical isolate of Acanthamoeba spp number 02. Initial amount of viable cysts in 18.5 μL = 6.0 × 10⁵.

§  Kruskall-Wallis test.

 ‖ Statistically significant.

The statistical analyses pair-to-pair showed the susceptibility of Acanthamoeba cysts to both PHMB and CLX, with significant differences in comparison with experimental control groups (Table 3). CLX 0.04% at dosage of one or two drops, as well as two drops of CLX 0.02%, were as effective as the association of CLX and PHMB at any dosages in terms of its cysticidal activity, demonstrating the susceptibility of Acanthamoeba reference strain and the clinical isolates to the cysticidal effect of high doses of CLX or the association of both biguanides compounds. When data from the application of different doses and concentrations of the same biguanide in Acanthamoeba cysts were compared, no statistical difference between ATCC 30011 strain and isolate 01 was observed (Table 4). However, when the cysticidal activity of one drop of CLX 0.02% and two drops of CLX 0.02% (P = 0.001); one drop of CLX 0.02% with two drops of CLX 0.04% (P < 0.0001), and one drop of CLX 0.04% with two drops of CLX 0.04% (P = 0.005) were comparatively analyzed in isolate 02, a statistically significant difference was observed. Significant differences related to isolate 02 were also demonstrated in the comparative statistical analysis between different doses and concentrations of PHMB and CLX and between different concentrations and doses of both PHMB and CLX with combination (Table 4). 

Table 3. 

 

Comparative Statistical Analysis of Cysticidal Activity Between Experimental Groups of Acanthamoeba Cysts Submitted to PHMB and CLX and Experimental Control Groups, Which Cysts Were Cultivated on the Fresh PYG Broth Medium Without PHMB or CLX

Biguanide Concentration (Doses)	ATCC 30011*, P‖	Isolate 01†, P ‖	Isolate 02‡, P ‖
PHMB 0.02% (1 drop) and experimental control group	0.013§	0.002§	0.001§
PHMB 0.02% (2 drops) and experimental control group	0.008§	0.001§	<0.0001§
PHMB 0.04% (1 drop) and experimental control group	0.008§	<0.0001§	<0.0001§
PHMB 0.04% (2 drops) and experimental control group	0.002§	<0.0001§	<0.0001§
CLX 0.02% (1 drop) and experimental control group	0.004§	0.003§	<0.0001§
CLX 0.02% (2 drops) and experimental control group	<0.0001§	0.021§	<0.0001§
CLX 0.04% (1 drop) and experimental control group	<0.0001§	<0.0001§	<0.0001§
CLX 0.04% (2 drops) and experimental control group	<0.0001§	<0.0001§	<0.0001§
PHMB/CLX 0.02% (1 drop) and experimental control group	<0.0001§	<0.0001§	<0.0001§
PHMB/CLX 0.04% (1 drop) and experimental control group	<0.0001§	<0.0001§	<0.0001§
PHMB/CLX 0.04% (2 drops) and experimental control group	<0.0001§	<0.0001§	<0.0001§

*   A. castellanii from ATCC.

†  Clinical isolate of Acanthamoeba spp number 01.

‡  Clinical isolate of Acanthamoeba spp number 02.

§  Statistically significant.

 ‖ Jonckheere-Terpstra test statistical analysis.

Table 4. 

 

Comparative Statistical Analysis, Two by Two, of the Cysticidal Activity Between Different Concentrations and Doses Among Each Biguanide (1), Between Biguanides (2), and Between Each Biguanide in Combination (3) in the Reference-Strain and Clinical Isolates of Acanthamoeba spp

Association of Biguanides (Doses)		ATCC 30011*, P‖	Isolate 01†, P‖	Isolate 02‡, P‖
1	PHMB 0.02% (1 drop) and PHMB 0.02% (2 drops)	1.000	1.000	0.655
	PHMB 0.02% (1 drop) and PHMB 0.04% (1 drop)	1.000	1.000	0.748
	PHMB 0.02% (1 drop) and PHMB 0.04% (2 drops)	1.000	0.996	0.024
	PHMB 0.02% (2 drops) and PHMB 0.04% (1 drop)	1.000	1.000	1.000
	PHMB 0.02% (2 drops) and PHMB 0.04% (2 drops)	1.000	1.000	0.703
	PHMB 0.04% (1 drop) and PHMB 0.04% (2 drops)	1.000	1.000	0.607
	CLX 0.02% (1 drop) and CLX 0.02% (2 drops)	0.989	0.999	0.001§
	CLX 0.02% (1 drop) and CLX 0.04% (1 drop)	0.996	0.989	0.559
	CLX 0.02% (1 drop) and CLX 0.04% (2 drops)	0.466	0.987	<0.0001§
	CLX 0.02% (2 drops) and CLX 0.04% (1 drop)	1.000	0.735	0.135
	CLX 0.02% (2 drops) and CLX 0.04% (2 drops)	0.978	0.715	0.897
	CLX 0.04% (1 drop) and CLX 0.04% (2 drops)	0.958	1.000	0.005§
2	PHMB 0.02% (1 drop) and CLX 0.02% (1 drop)	1.000	1.000	0.999
	PHMB 0.02% (1 drop) and CLX 0.02% (2 drops)	0.903	0.998	<0.0001§
	PHMB 0.02% (1 drop) and CLX 0.04% (1 drop)	0.939	0.995	0.178
	PHMB 0.02% (1 drop) and CLX 0.04% (2 drops)	0.247	0.993	<0.0001§
	PHMB 0.02% (2 drops) and CLX 0.02% (1 drop)	1.000	1.000	0.975
	PHMB 0.02% (2 drops) and CLX 0.02% (2 drops)	0.958	0.980	0.020§
	PHMB 0.02% (2 drops) and CLX 0.04% (1 drop)	0.978	1.000	0.997
	PHMB 0.02% (2 drops) and CLX 0.04% (2 drops)	0.336	1.000	0.001§
	PHMB 0.04% (1 drop) and CLX 0.02% (1 drop)	1.000	1.000	0.991
	PHMB 0.04% (1 drop) and CLX 0.02% (2 drops)	0.958	0.916	0.015
	PHMB 0.04% (1 drop) and CLX 0.04% (1 drop)	0.978	1.000	0.991
	PHMB 0.04% (1 drop) and CLX 0.04% (2 drops)	0.336	1.000	<0.0001§
	PHMB 0.04% (2 drops) and CLX 0.02% (1 drop)	1.000	0.992	0.116
	PHMB 0.04% (2 drops) and CLX 0.02% (2 drops)	0.999	0.753	0.607
	PHMB 0.04% (2 drops) and CLX 0.04% (1 drop)	1.000	1.000	0.995
	PHMB 0.04% (2 drops) and CLX 0.04% (2 drops)	0.636	1.000	0.046§
3	PHMB/CLX 0.02% (1 drop) and PHMB/CLX 0.04% (1 drop)	1.000	0.988	0.210
	PHMB/CLX 0.02% (1 drop) and PHMB/CLX 0.04% (2 drops)	0.409	1.000	0.013§
	PHMB/CLX 0.04% (1 drop) and PHMB/CLX 0.04% (2 drops)	0.411	0.978	0.951
	PHMB 0.02% (1 drop) and PHMB/CLX 0.02% (1 drop)	0.989	0.984	0.001§
	PHMB 0.02% (1 drop) and PHMB/CLX 0.04% (1 drop)	0.985	1.000	<0.0001§
	PHMB 0.02% (1 drop) and PHMB/CLX 0.04% (2 drops)	0.137	0.975	<0.0001§
	PHMB 0.02% (2 drops) and PHMB/CLX 0.02% (1 drop)	0.998	0.999	0.098
	PHMB 0.02% (2 drops) and PHMB/CLX 0.04% (1 drop)	0.997	1.000	0.001§
	PHMB 0.02% (2 drops) and PHMB/CLX 0.04% (2 drops)	0.206	0.999	<0.0001§
	PHMB 0.04% (1 drop) and PHMB/CLX 0.02% (1 drop)	0.998	1.000	0.069
	PHMB 0.04% (1 drop) and PHMB/CLX 0.04% (1 drop)	0.997	1.000	<0.0001§
	PHMB 0.04% (1 drop) and PHMB/CLX 0.04% (2 drops)	0.206	1.000	<0.0001§
	PHMB 0.04% (2 drops) and PHMB/CLX 0.02% (1 drop)	1.000	1.000	0.994
	PHMB 0.04% (2 drops) and PHMB/CLX 0.04% (1 drop)	1.000	0.998	0.082
	PHMB 0.04% (2 drops) and PHMB/CLX 0.04% (2 drops)	0.496	1.000	0.008§
	CLX 0.02% (1 drop) and PHMB/CLX 0.02% (1 drop)	1.000	0.969	0.005§
	CLX 0.02% (1 drop) and PHMB/CLX 0.04% (1 drop)	1.000	1.000	<0.0001§
	CLX 0.02% (1 drop) and PHMB/CLX 0.04% (2 drops)	0.320	0.955	<0.0001§
	CLX 0.02% (2 drops) and PHMB/CLX 0.02% (1 drop)	1.000	0.555	0.947
	CLX 0.02% (2 drops) and PHMB/CLX 0.04% (1 drop)	1.000	0.972	0.997
	CLX 0.02% (2 drops) and PHMB/CLX 0.04% (2 drops)	0.962	0.509	0.629
	CLX 0.04% (1 drop) and PHMB/CLX 0.02% (1 drop)	1.000	1.000	0.545
	CLX 0.04% (1 drop) and PHMB//CLX 0.04% (1 drop)	1.000	0.998	0.007§
	CLX 0.04% (1 drop) and PHMB/CLX 0.04% (2 drops)	0.929	1.000	0.001§
	CLX 0.04% (2 drops) and PHMB/CLX 0.02% (1 drop)	0.623	1.000	0.117
	CLX 0.04% (2 drops) and PHMB/CLX 0.04% (1 drop)	0.651	0.997	0.999
	CLX 0.04% (2 drops) and PHMB/CLX 0.04% (2 drops)	1.000	1.000	1.000

*   A. castellanii from ATCC.

†  Clinical isolate of Acanthamoeba spp number 01.

‡  Clinical isolate of Acanthamoeba spp number 02.

§  Statistically significant.

 ‖ Jonckheere-Terpstra test statistical analysis.

The cytotoxicity assays showed a toxic effect of CLX and PHMB in endothelial cells (Table 5). Interestingly, the association of both biguanide compounds seemed to show a statistically significant synergic activity, and HUVEC viability was at the same rate as the experimental control group, without biguanide contact (Table 5). Qualitative data from endothelial cells subjected to coculture with Acanthamoeba cysts and the toxic activities of PHMB and CLX were assessed by bright field and fluorescent microscopy procedures following nucleus staining (Fig. 2). As shown in Figure 2A, coculture procedure observations showed that Acanthamoeba cysts from ATCC 30011 and isolate 02 were able to induce a mild necrotic activity in endothelial cells, observed by a low rate of PI nucleus staining. On the other hand, nucleus condensation observed by AO staining suggested induction of an apoptotic process. These findings suggest the hypothesis that the cell death process may be mediated via a contact-dependent mechanism induced by Acanthamoeba cysts. Both PHMB and CLX showed cytotoxic effect on HUVEC. However, results suggest that CLX caused cell death by necrosis and that PHMB may induce apoptosis (Figs. 2B, 2C). Interestingly, the association between PHMB and CLX, at concentration of 0.04%, appeared to decrease the amount of cell death when compared with the cytotoxic effect presented by single instillation of biguanides (Figs. 2C, 2D). 

Figure 2. 

 

Coculture of Acanthamoeba cysts and biguanides-induced cell death in HUVEC. (A) Coculture of Acanthamoeba cysts with HUVECs. (B) Coculture of Acanthamoeba cysts and HUVEC cells treated with 22 μL (1 drop) of PHMB at 0.04%. (C) Coculture of Acanthamoeba cysts and HUVEC cells treated with 35 μL (1 drop) of CLX at 0.04%. (D) Coculture of Acanthamoeba cysts and HUVEC cells treated with 22 μL (1 drop) of PHMB at 0.04% and 35 μL (1 drop) of CLX at 0.04%. After 48 hours of contact with cysts and biguanides, HUVECs were harvested with fresh RPMI medium following staining method with AO and PI fluorescent reagents. Nuclei of cells whose membranes are intact stain green while those of necrotic cells stain red. Blue arrows show viable endothelial adherent cells. Magnification: ×100.

Table 5. 

 

Percentual of HUVEC Cell Viability in Coculture With Acanthamoeba Cysts Under 48 Hours Exposure to PHMB and CLX

Biguanides (Doses)	Cell Viability, %				P *
	HUVEC Mean ± SD	HUVEC/ATCC 30011 Mean ± SD	HUVEC/Isolate 01 Mean ± SD	HUVEC/ Isolate 02, Mean ± SD
PHMB 0.04% (1 drop)	5.7 ± 2.3	4.0 ± 3.6	4.3 ± 0.6	7.7 ± 1.2	0.305
CLX 0.04% (1 drop)	0.7 ± 1.2	2.7 ± 1.2	2.7 ± 1.5	5.3 ± 1.2	0.153
PHMB 0.04% (1 drop) + CLX 0.04% (1 drop)	119 ± 0.6	10.0 ± 3.5	84.3 ± 2.5	0.7 ± 1.2	0.015†
Control group	99.8 ± 5.2	31.6 ± 0.5	102.7 ± 4.8	8.2 ± 4.0	<0.0001†

 HUVEC metabolic activity was measured using PrestoBlue reagent.

*  Kruskall-Wallis test.

†  Statistically significant.

The cytotoxicity assays in HCECs showed a toxic effect of CLX and PHMB and association of both biguanide compounds (Fig. 3). There was no statistically significant difference between the concentrations of each biguanide used in the assays, and between CLX and PHMB or their association, at all times evaluated. After 12, 24, or 48 hours of incubation with biguanides, the number of metabolically active viable cells was slight or negligible in comparison to controls (P < 0.001).Qualitative data from corneal epithelial cells subjected to the toxic activities of PHMB and CLX were assessed by bright field and fluorescent microscopy following nucleus staining (Fig. 4). The toxic effect of both CLX and PHMB on HCEC was observed. Results suggest that CLX caused cell death by necrosis, indicated by staining with PI. The absence of PI staining was noticed in PHMB-treated cells. Interestingly, the association between PHMB and CLX showed results similar to PHMB treatment. Bright field images of PHMB-treated cells show darker cells, and shadow areas suggestive of cell agglomerating and cell debris. Results observed at different times were similar, demonstrating the cytotoxic effect of biguanide compounds as early 12 hours after treatment. 

Figure 3. 

 

Cell viability of HCEC after exposure to PHMB and CLX compounds. HCEC cells were incubated at different times with PHMB, CLX, or association of both biguanides (1:4 in culture medium). Cell viability was evaluated by PrestoBlue and presented as percentage of non-treated controls, considered as 100% viable cells. Data are presented as mean ± SD. All treatments showed a P less than 0.001 in comparison with control.

Figure 4. 

 

Biguanide induced cell death in HCEC. (A) After 12 hours; (B) after 24 hours; (C) after 48 hours. HCEC cells were treated with PHMB and/or CLX at 0.02% and 0.04% (1:4 in culture medium), and incubated for 12 to 48 hours. After incubation with biguanides, HCECs were harvested with fresh K-SFM medium following staining method with AO and PI fluorescent reagents. Nuclei of cells whose membranes are intact stain green while those of necrotic cells stained red. Blue arrows show adherent viable epithelial cells. Magnification: ×100.

Discussion

The cysts of Acanthamoeba spp are composed of an ectocyst, an external cellulosic layer, and an endocyst, a thinner, internal fibrillar layer, which together provide amoebic resistance to physical and chemical compounds. ²⁵ The prolonged use of biguanides and/or diamidines for the treatment of painful recurrent and progressive ulcerative amoebic keratitis has frequently been associated with topical anesthetic abuse with its toxic complications. ^{20 –22} In addition, in the later therapeutic stage of prolonged AK, other serious side effects, such as glaucoma, cataract, and iris atrophy have been described. ^{26 –28}  

An upward trend in the number of new cases of Acanthamoeba infection on the corneal surface has been observed. ¹⁴ Moreover, the occurrence of severe ulcerative disease in our referral center in São Paulo, Brazil, has led to greater attention and concern by ophthalmologists. The therapeutic management of Acanthamoeba keratitis spp includes the topical use of individual or associated biguanides and diamidines. ^{29 –31} The treatment of amoebic infection has been initiated with instillation of PHMB or CLX 0.02% hourly or topical application of diamidines. ^31,32 The therapeutic application of diamidines has not been properly regulated by the competent Brazilian oversight agencies, and biguanides at higher concentrations of 0.04% have been proposed as an alternative therapy in clinical cases of greater severity that have been associated with recurrence of keratitis and resistance to chemical treatment. ³³ Furthermore, the application of physical agents (e.g., UV-A) has been proposed as an alternative technology to inactivate protozoa, and in vitro and in vivo experimental studies have shown conflicting results. ^34,35  

There is no consensus about the efficacy of biguanides against Acanthamoeba cysts and trophozoites with respect to the toxic effects on the cells, extracellular matrix components, glycoproteins, and proteoglycans of the anterior and posterior corneal layers. ^{36 –38} Our study was conducted in order to correlate the dose- and concentration-dependent in vitro effect of biguanides on loss of viability of Acanthamoeba cysts and at the same time, provide data on toxic effects of PHMB and CLX compounds, alone or in combination, on the endothelial cells. The results shown in this study corroborate previous descriptions of the efficacy of both individual CLX as of PHMB on the loss of viability of Acanthamoeba cysts. ^39,40 Different concentrations and dosages of single or associated biguanides were evaluated against Acanthamoeba cysts, and the findings indicate that PHMB 0.04% associated with CLX 0.04%, at dosage of two drops of both, has an effective cysticidal profile against ATCC 30011 strain and clinical isolates of Acanthamoeba spp. The results presented in this study suggest the occurrence of differential resistance patterns among cysts obtained from two cases of severe ulcerative keratitis, and isolate 02 seemed to be more resistant to the cysticidal activity of PHMB and CLX than isolate 01. The difference of susceptibility to biguanide compounds may be caused by carbohydrate and protein contents in the composition of cyst wall of A. castellanii and A. rhysodes, ⁴¹ which nucleotide sequences were most closely related to isolate 01 and isolate 02, respectively. Our findings showed that choice of both the pattern of dosage and concentration of biguanides used may be relevant to the effectiveness of the therapeutic regimen chosen for the treatment of AK. 

The HUVEC cells have been commonly used in cytotoxicity and cell viability assays due to their high proliferation capacity and sensitivity to changes in environmental conditions. ^{42 –44} For example, during systemic biodistribution of chemical compounds (i.e., antibiotics and chemotherapeutic compounds) the vascular endothelium corresponds to one of the first barrier cells exposed to the drugs. Endothelial cells also represent a major target for adhesion and invasion of microorganisms and have, therefore, been used for the experimental assays as in vitro cellular models of infection and cell interaction. ^42,43 When comparing the action of PHMB and CLX in the viability of HUVECs in different experimental groups, it was observed by spectrophotometric assays with PrestoBlue cell viability reagent that the experimental group composed of HUVEC cells in individual contact with PHMB or CLX in the absence of coculture with Acanthamoeba spp, demonstrated significantly reduced cell viability, showing the toxic effect of the biguanides in endothelial cells. However, when PHMB and CLX were used together, there was, interestingly, no deleterious effect on these cells, as if the combination of biguanides might provide a protective effect on HUVECs. The screening cytotoxicity assays using fluorescence microscopy with vital dyes correlated well with quantitative data, which indicated a decrease in the cytotoxic activity of PHMB in combination with CLX both at higher concentration and dosage when compared with other experimental groups. The quantitative and qualitative data in cellular assays also demonstrated a slightly smaller cytotoxic effect of the compound PHMB when compared with CLX, which results were statistically significant. 

The cytotoxicity of biguanides observed on HCEC cells correlated with experimental results reported previously. ^45,46 The different patterns of PI staining demonstrated by fluorescence microscopy assays suggested that CLX and PHMB could display cytotoxic activity by distinct mechanisms. Cell death by necrosis was observed in HCEC cells with loss of membrane integrity after treatment with a CLX containing solution. ⁴⁵ Dutot and collaborators demonstrated that PHMB could be able to induce apoptosis through activation of the cell death receptor P2X7, with no alterations on cell membrane integrity. ⁴⁶ The toxicity profile of CLX and PHMB was compared between HUVEC and HCEC because endothelial cells from adnexal tissue of corneal epithelium could be exposed to biguanides in the topical application during the treatment of AK, while corneal epithelial cells represent the first barrier of cell line to be in contact with chemotherapeutic compounds in the corneal surface. Our findings demonstrated that HCEC seemed to be more susceptible to toxic effects of biguanides than HUVEC. This observation could be related with intrinsic physiologic aspects of each cell line, for example, the demand for FBS to provide the viability of HUVECs and the absence of this supplementation in the HCEC growth. Kim et al. demonstrated the protective effect of FBS against the cytotoxicity of ocular therapeutic drug. ⁴⁷  

Efforts to provide a standard and safe therapeutic regimen for AK based on inactivation of cysts with minimal toxicity to the host cells, suggests the possibility of development of more experimental assays based on cytotoxic activities of biguanides in the corneal tissue of standard animal models. In conclusion, we were able to demonstrate by in vitro assays that the dosage of two drops of PHMB and CLX in combination, at concentration of 0.04%, could induce the cysticidal activity on Acanthamoeba spp independent of the degree of protozoan resistance. Furthermore, our studies demonstrate that double-biguanide application could provide synergistic activity with improved endothelial cell viability. 

Acknowledgments

Supported by grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP Grant 11/51626-1 [FRSC] and FAPESP Grant 08/53969-0 [DF]) and the National Council for Scientific and Technological Development (CNPq Grant 311612/2012-1), a postdoctoral fellowship from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (PNPD/CAPES [LCCC]), and a fellow from the Young Researchers at Emerging Centers Program (FAPESP Grant 2012/15603-0 [FRSC]). 

Disclosure: C.S.P. Mafra, None; L.C. Carrijo-Carvalho, None; A.M. Chudzinski-Tavassi, None; F.M.C. Taguchi, None; A.S. Foronda, None; F.R.S. Carvalho, None; D. de Freitas, None