Background/aims The aim of this study was to correlate the various forms of Acanthamoeba on ex vivo confocal microscopy (EVCM) with in vivo confocal microscopy (IVCM) and findings from cultured positive cases of Acanthamoeba keratitis.
Methods Acanthamoeba live, dead and empty cysts, and live trophozoites were prepared in vitro and inoculated into porcine cornea using a sterile 26-gauge needle and examined ex vivo using the Heidelberg Retina Tomograph II/Rostock Corneal Module. IVCM images from 12 cultured positive Acanthamoeba cases, obtained using the same instrument, were compared with EVCM findings. Phase contrast images were also obtained to compare with both EVCM and IVCM findings. The change in cyst morphology with depth was evaluated by imaging the same cysts over a defined cornea depth measurement.
Results EVCM morphologies for live cysts included four main types—hyper-reflective central dot with hyper-reflective outer ring, hyper-reflective central dot with hyporeflective outer region, stellate shaped hyper-reflective centre with hyporeflective outer region and hyper-reflective round/polygonal shaped cyst; one main type for dead cysts—hyper-reflective central dot with hyporeflective outer region; two main types for empty cysts— hyper-reflective central dot with hyper-reflective outer ring/hyporeflective outer region; and one main type for trophozoites—large coarse speckled area of heterogeneous hyper-reflective material. Matching IVCM images show good correlation with EVCM. Cyst morphology altered when imaged at different depths.
Conclusion EVCM demonstrated the various forms of Acanthamoeba cyst and trophozoites can be used as a reference to identify similar structures on IVCM.
http://dx.doi.org/10.1136/bjophthalmol-2022-321402
Acanthamoeba keratitis is a sight-threatening disease and in vivo confocal microscopy has been shown to be a useful tool in diagnosing the infection with both high sensitivity and specificity but this is dependent on observer experience.
The limitation with current in vivo confocal microscopy technology is the difficulty in distinguishing host cellular structures from Acanthamoeba cysts and trophozoites. Ex vivo confocal microscopy findings of the various life cycle stages of Acanthamoeba identified in this study can be used as a reference to aid the identification of the various trophozoite and cyst morphologies seen in vivo.
Improving the accuracy of identifying Acanthamoeba cysts and trophozoite-like structures seen on in vivo confocal microscopy by comparing to an ex vivo reference standard will improve diagnostic certainty, enabling the instigation of prompt antiamoebic treatment, with the potential of reducing visual loss in patients with Acanthamoeba keratitis.
Acanthamoeba spp are opportunistic pathogens of humans and can potentially cause a blinding keratitis.1 Early stages of Acanthamoeba keratitis (AK) can resemble other forms of infective keratitis, which can make establishing the correct diagnosis difficult, resulting in diagnostic delay and potentially poorer visual outcome.1–3
Current diagnostic modality for AK include culture, PCR and in vivo confocal microscopy (IVCM). Although culture is still considered the ‘gold’ standard, the effectiveness of culture is suboptimal with a sensitivity rate ranging from 31% to 55%.4 5 IVCM has been shown to be a promising tool in diagnosing AK, yielding sensitivity and specificity values of 56%–100% and 84%–100%.5–7 The diagnostic criteria of AK using IVCM include the recognition of specific cystic and trophozoite-like structures with single file presentation and clustering of cystic objects to be a strong predictor of poorer visual outcome.8 9 In validating the structures seen on IVCM, previous studies have correlated IVCM findings with ex vivo imaging of the organism grown on cultured plates obtained from corneal scrape or contact lens solution of the same patients10–12 or from directly examining a suspension of cultured Acanthamoeba trophozoites.13 The limitation with these approaches is that imaging findings of Acanthamoeba grown on a culture plate may look different to when it is in the cornea.
The aim of this study was to inoculate porcine corneas with the various forms of Acanthamoeba and to correlate what we see on ex vivo confocal microscopy (EVCM) with imaging findings from 12 culture positive AK cases on IVCM.
The strains used in this study were Acanthamoeba castellanii (ATCC 50370) and Acanthamoeba polyphaga (ATCC 30461). Trophozoites and cysts were produced as previously described.14 15
Dead Acanthamoeba trophozoites and cysts were prepared by exposing them to 0.02% (v/v) polyhexamethylene biguanide for 1 hour. After incubation the cells were visually inspected by phase contrast microscopy (x100) and cell death was confirmed by inoculating the cysts and trophozoites back into growth medium and confirming that they were non-viable. The cells were then pelleted by centrifugation at 500 × g for 5 min.
Cysts were inoculated into Ac#6 medium for 4 hours to facilitate maximal excystation (>90%) before exposure to N-lauryl-sarcosine 0.5% (w/v) to lyse trophozoites.16 The remaining cyst walls were pelleted by centrifugation at 1000 x g for 5 min and inspected microscopically to confirm >90% empty cyst.
Porcine eye globes were obtained from freshly slaughtered pigs from a local abattoir and processed as defined previously.17 Briefly, the eye globes were immersed in sterile Povidone-iodine (0.5% w/v) for 2 min to facilitate decontamination before neutralisation in sodium thiosulphate (0.1%, w/v). The globes were transferred to 0.1% (w/v) gentamicin-phosphate-buffered saline (PBS) solution for 15 min to remove any residual bacterial contamination before transfer to sterile saline (PBS) solution until processing.
The following Acanthamoeba samples were injected into different eyes sequentially: dead trophozoites, live trophozoites, dead cyst, live cysts and empty cysts at a concentration of 1×105 cells / mL. Samples were inoculated using a sterile 26-gauge needle inserted tangentially into the middle stroma of the cornea without applying pressure to the syringe to avoid pockets of liquid forming in the stroma. The needle was then withdrawn slowly and the sample was inoculated into the needles path by capillary action. Globes were imaged by EVCM immediately after inoculation with Acanthamoeba.
EVCM was performed using the Heidelberg Retina Tomograph II with the Rostock Corneal Module (HRT II/RCM, Heidelberg Engineering, Dossenheim, Germany). A sterile TomoCap (Heidelberg Engineering) was mounted over the objective of the microscope and GelTears (0.2% w/w carbomer 980, Bausch & Lomb, UK) was used as a coupling agent between the disposable cap and the lens objective. A drop of 1% Carmellose sodium was put on the surface of the cornea of the porcine eye inoculated with Acanthamoeba and the whole eye was held secure on a standard retort stand by a three prong clamp (figure 1). Prior to inoculating the corneas with Acanthamoeba, baseline IVCM images of an intact porcine cornea and a cornea after a stab incision were obtained to act as a control. The instrument was then brought into contact of the cornea corresponding to the area where the Acanthamoeba was inoculated and multiple volume (a series of 40 images over 80 µm depth) scans starting from the superficial epithelium all the way to the deep stroma were recorded. EVCM images were assessed qualitatively and classified based on the morphology features detected.
Ex vivo scanning process. Ex vivo confocal microscopy imaging of the cornea of the porcine eye globe with the Heidelberg Retina Tomograph II/Rostock Corneal Module laser confocal microscope. The image shows the eye globe held in place using a three-pronged clamp attached to a retort stand with the cornea in contact with the TomoCap during the scanning process.
Twelve cases, with matching morphological features to those seen on EVCM, were selected from a cohort of cultured positive AK cases that were published in a previous study.8 AK staging was classified by a corneal specialist on recruitment into the following: (1) epitheliitis, (2) epitheliitis with perineural infiltrates, (3) anterior stromal disease, (4) deep stromal disease and (5) ring infiltrate.
The Acanthamoeba morphologies identified with EVCM were used to validate the various cyst and trophozoite-like features seen on IVCM. All available EVCM images were reviewed by one experienced observer (SH). In total, this equated to approximately 500 images were reviewed for each porcine cornea. Images were reviewed anteroposteriorly from the epithelium to deep stroma. One best image corresponding to each Acanthamoeba form was identified and then used as a reference for validating structures seen on IVCM images. In addition, phase contrast images of the various forms of Acanthamoeba were produced and compared with the EVCM and IVCM images. To evaluate how cyst morphology changes with increasing depth, a series of images obtained from a single volume stack, tracking the same cysts, separated by 2 μm, were obtained from both EVCM and IVCM.
The same cell suspensions prepared for the inoculation of the corneal tissue were used to generate the phase contrast images. Samples of each morphology were placed onto microscope slides and overlaid with a glass coverslip. The cells were then viewed on a Zeiss Primovert inverted microscope (Zeiss, Cambridge, UK) at a magnification of x400 and the images were acquired on a Canon EOS M50 digital camera using a Zeiss P95-T2 DSLR 1.6x trinocular microscope adapter.
The cornea of the control porcine eye showed the appearance of the intact corneal epithelium as polygonal shaped cells with hyper-reflective cell border, with part of the stroma containing a corneal nerve seen within its layers. The second control cornea with a tangential incisional wound to the epithelium demonstrates an area of hyporeflectivity surrounded by hyper-reflectivity in the superficial and basal epithelium, respectively (online supplemental figure 1).
EVCM of live Acanthamoeba cysts demonstrated four main cyst morphologies: hyper-reflective central dot with hyper-reflective outer ring, hyper-reflective central dot with hyporeflective outer region, stellate shaped hyper-reflective centre with hyporeflective outer region and hyper-reflective round/polygonal shaped cyst (figure 2A). Representative IVCM cases with matching cyst morphologies are shown in figure 2B–E and the patient demographics are shown in table 1.
Morphological classification of live cysts observed with ex vivo confocal microscopy, in vivo confocal microscopy and phase contrast imaging. (A) Ex vivo confocal microscopy. (B) In vivo confocal microscopy of case 1. (C) In vivo confocal microscopy of case 2. (D) In vivo confocal microscopy of case 3. (E) In vivo confocal microscopy of case 4. (F) Phase contrast images. Labels: hyper-reflective central dot with hyper-reflective outer ring (star), hyper-reflective central dot with hyporeflective outer region (arrow head), stellate shaped hyper-reflective centre with hyporeflective outer region, (arrow), hyper-reflective round/polygonal shaped cyst (dotted arrow), signet ring (cross).
The same cyst preparation was viewed under phase contrast microscopy at x 400 and the cyst wall is seen as refractile and the double cyst wall is clearly visible (figure 2F). The trophozoite and their cytoplasmic vacuoles show up as hyper-reflective areas inside the cyst wall that are visible on both EVCM and IVCM but they are not refractile under phase contrast. The EVCM images closely resemble those from the IVCM images apart from in case 4 (figure 2E) where the signet ring morphology (hyper-reflective outer ring with a grey/dark centre) was seen on IVCM but not ECVM. Tracking the same cysts with increasing depth revealed the cyst morphology to change depending on which part of the cyst was imaged and this was shown in both ex vivo and in vivo imaging (online supplemental figure 2). In both cases, three types of morphology were seen over a change in depth of 11 µm.
Under EVCM, the main cyst morphology identified is the hyper-reflective central dot with hyporeflective outer region (figure 3A) but no outer ring is seen. Hyper-reflective round/polygonal shaped cysts are also seen but are less common and in general, the cyst size appears much smaller compared with the live cysts. Furthermore, the hyporeflective outer region in dead cysts appears much darker than live cysts. Phase contrast microscopy showed similar findings of a non-refractile outer cell wall, a ruptured cell membrane leading to leakage of the cytoplasm, and the shrinking of the cell with a small refractile dot inside the cyst demonstrating the remainder of the trophozoite (figure 3F). Representative IVCM cases, with similar cyst morphology to the dead cysts of EVCM, are shown in figure 3B–E (case number 5–8—table 1). The dark hyporeflective outer rings are similar to the images from EVCM, and it is apparent other cyst morphologies are present on IVCM, such as hyper-reflective round/polygonal shaped cyst, among the dead cysts.
Morphological classification of dead cysts observed with ex vivo confocal microscopy, in vivo confocal microscopy and phase contrast imaging. (A) Ex vivo confocal microscopy. (B) In vivo confocal microscopy of case 5. (C) In vivo confocal microscopy of case 6. (D) In vivo confocal microscopy of case 7. (E) In vivo confocal microscopy of case 8. (F) Phase contrast image. Labels: hyper-reflective central dot with hyper-reflective outer ring (star), hyper-reflective central dot with hyporeflective outer region (arrow), hyper-reflective round/polygonal shaped cyst (dotted arrow).
EVCM images of empty cysts show the outerwall to be visible in some cells but they are mainly hyporeflective (figure 4A). There are also small hyper-reflective dots inside the cysts probably representing the remaining cellular material of the trophozoite. Although this morphology was also seen in EVCM live cysts, it is much more prevalent in empty cysts. Corresponding IVCM images from two patients (case 9 and 10), with predominantly these features are shown in figure 4B,C and their corresponding clinical characteristics are shown in table 1. Phase contrast image shows visible mainly non-refractile cyst wall with a small amount of cellular material that has been left behind inside the cyst by the excysting trophozoite (figure 4D).
Morphological classification of empty cysts observed with ex vivo confocal microscopy, in vivo confocal microscopy and phase contrast imaging. (A) Ex vivo confocal microscopy. (B) In vivo confocal microscopy of case 9. (C) In vivo confocal microscopy of case 10. (D) Phase contrast image. Labels: hyper-reflective central dot with hyporeflective outer region (arrow head), hyper-reflective round/polygonal shaped cyst (dotted arrow), stellate shaped hyper-reflective centre with hyporeflective outer region (arrow).
Under EVCM, individual cells and their cell membrane cannot be visualised, instead trophozoites appear as a large coarse speckled area of heterogeneous hyper-reflective material. These hyper-reflectivity dots are indicative of the food vacuoles inside the cytoplasm of the trophozoites (figure 5A). Representative IVCM images from 2 cases (case 11 and 12) show similar aggregates of hyper-reflective dot like material with no apparent discernible cell boundary seen (figure 5B, C).
Appearance of trophozoites on ex vivo confocal microscopy, in vivo confocal microscopy and phase contrast imaging. (A) Ex vivo confocal microscopy indicating an irregular area of heterogeneous hyper-reflectivity (boundary demarcated by arrows). (B) In vivo confocal microscopy of case 11 showing an area of heterogeneous hyper-reflectivity that is similar to 4A (arrows). (C) In vivo confocal microscopy of case 12 showing two areas of heterogeneous hyper-reflectivity (arrows) and hyper-reflective round/polygonal shaped cysts (dotted arrow). (D) Phase contrast image showing multiple trophozoites and their food vacuoles, with some cells possessing acanthopodia.
Similarly, the phase contrast image demonstrated live trophozoites that are amoeboid in shape with intact non-refractile cell membrane and the cytoplasm is full of refractile food vacuoles (figure 5D). When the trophozoites are joined together, similar to both EVCM and IVCM, the cell boundaries between cells are not easy to discern.
In this study, we have identified various forms of Acanthamoeba morphology using EVCM and correlated these findings with culture positive AK cases on IVCM. Various IVCM studies have been published describing the range of cyst morphologies seen in AK but there is a paucity of data corroborating IVCM findings with actual Acanthamoeba organism examined ex vivo using the same confocal microscope. The evaluation of live cysts under EVCM demonstrated four main types of cyst morphology and we found good correlation with representative culture positive cases seen on IVCM. Our EVCM findings of stellate shape hyper-reflective cysts,10 cysts with outer hyper-reflective walls,10 11 cysts with outer hyporeflective wall, are in agreement with previous ex vivo and in vivo correlation studies.
Although all four cyst morphologies have been reported in previous IVCM studies,8 9 other phenotypes such as ‘signet ring’, which was detected in case 4, and ‘coffee-bean shaped’ were not detected on EVCM. We found the morphology of the cyst altered depending on how the cyst was sectioned during image acquisition in terms of depth, and this might partly explain the range of cyst morphologies reported in previous IVCM studies. This assumption is further supported by the presence of a more uniform cyst morphology seen on phase contrast when the cyst images were obtained only at one particular depth.
We found dead Acanthamoeba cysts to be smaller, without a hyper-reflective ring, and darker hyporeflective outer region, when compared with live cysts. This was evident from the phase contrast images which showed non refractile cell wall, ruptured cell membrane and shrinkage of the cysts. The comparative IVCM image showed among the presumed dead cysts, morphologies consistent with live cysts are also present within the same image. Publications on how cyst morphology change with treatment are limited: Li et al18 described the cysts either disappeared or formed hollow structures with anti-amoebic treatment but there was no direct ex vivo correlation to confirm this assumption. Therefore, it is difficult to ascertain whether the hollow structures described were actual dead or unviable cysts or it was partly related to how the cysts were sectioned on image acquisition. The morphology of the dead cysts found in this study appears to be different to the hollow structures described by Li et al18 but as we specifically cultured this morphology for imaging; we believe what we found is an accurate description.
The main phenotype seen in empty cysts on EVCM was the central hyper-reflective dot with an outer hyporeflective wall. This appearance is likely to be related to the refractile residual cellular material remaining after the trophozoite has excysted as shown in the phase contrast image. Moreover, even though this phenotype was seen among live cysts, the area surrounding the central hyper-reflective dot was much darker and homogeneous compared with those seen in live cysts. There were one or two cysts that had a stellate/polygonal hyper-reflective centre, similar to those seen in live cysts, and this probably reflects the trophozoite failing to excyst from the cell.
In a previous ex vivo study using cultured samples from eight eyes of seven patients, Yamazaki et al13 have shown trophozoites to be amorphous, highly reflective, high contrast objects with no walls and a mean size of 25.4 µm. However, it is unclear how the authors deduce the objects seen were in fact trophozites by imaging a suspension of the cultured medium.13 Furthermore, they were not able to detect trophozoites in the cornea, possibly because of the difficulty in distinguishing pathological features from trophozoites. Shiraishi et al12 have also identified trophozoites to be highly reflective, pleomorphic, organisms with presumed acanthopodia when they imaged the surface of culture plate. Similarly, Matsumoto et al11 showed trophozoites had homogeneous intense, highly reflective, multiform structures that were generally larger than 100 µm in diameter on IVCM. However, when they imaged the culture plate of culture positive cases on light microscopy, they found the size to range between 25 and 50 µm, which is more in keeping with the size found by other investigators. The authors cited the difference in size could be explained by the behaviour of the organisms in living tissue or due to different pathogenic strains. In a previous retrospective case–control study, trophozoite-like images, as defined by hyper-reflective objects with spiny surface structures suggestive of acanthopodia, were one of the morphologies that was found to be pathognomonic of AK.9 In contrary to these studies, we found the appearance of aggregation of trophozoites on imaging as a coarse speckled area of hyper-reflective material with no distinctive separation of cell boundary seen. This was confirmed by phase contrast microscopy that showed amoeboid trophozoites presented with non-refractile cell membrane and multiple refractile food vacuoles within the cytoplasm. Moreover, we did not identify any isolated hyper-reflective objects consistent with previous definition of trophozoite-like images on EVCM.12 13 We believe inoculating cultured trophozoites directly into the cornea and imaging them in an ex vivo manner, using the same confocal microscope as in vivo, provides the best evidence of defining the appearance of trophozoites in the cornea. The lack of defined cell boundary and the way the trophozoites aggregate on the cornea explain why it is difficult to identify them on imaging.19
There are several limitations in this study. Even though the various types of cysts and trophozoites were cultured using established methodology, it is possible that more than one morphologies are present within each culture medium, thereby giving rise to mixed morphologies on imaging. The way the organism was inoculated into the cornea is not the same as how the infection is acquired in vivo so it is possible that cyst morphologies could look different. That said, we have shown good agreement in cyst morphology between EVCM and IVCM so we believe the method of inoculation would not have affected the outcome. Other factors that may affect the cysts and trophozoite appearance include we used cadaver porcine cornea, the corneas were imaged straight after inoculation, the interaction of the host innate immune system with Acanthamoeba may induce different morphologies to those seen ex vivo, and EVCM images were acquired from untreated porcine corneas and then compared with IVCM images from treated patients. The main strength in this study is we have imaged the Acanthamoeba organism ex vivo, using the same confocal microscope, and correlated the cysts and trophozoite morphologies with culture confirmed cases of AK in vivo. This, we believe, is a robust way of confirming the various cyst and trophozoite-like features seen on IVCM are indeed representative of Acanthamoeba in the cornea. The correlation of the EVCM images taken in the absence of cytopathic effect induced by drug treatment and the cells of the innate immune system with IVCM further supports the reliability of the EVCM model.
In conclusion, we have described and correlated the appearance of Acanthamoeba on EVCM with IVCM, and found good agreement in the phenotypes identified on imaging. Although recent IVCM studies have found certain morphological features are more sensitive in confirming a diagnosis and are indicative of a less favourable visual outcome,8 9 the relationship between the prevalence of a specific cyst morphology with anti-AK treatment, and how these cystic features change with time, are not known. The changes in the prevalence of the various cyst morphologies with treatment may be a useful prognostic indicator and further research is needed to elucidate the relationship between cyst morphology and treatment.
Ethical approval was obtained from the Research Ethics Committee, Moorfields Eye Hospital NHS Foundation Trust, London, UK—Reference Number: ROAD 15/042 and University Life Sciences Ethics Committee—LSEC/201920/WH/110 (LSEC/201920/WH/97).
Contributors All authors contributed equally to the work and the writing of the manuscript. All authors (NA, WH and SH) were involved in all aspects of article conception, writing, critical revision and review and approval. WH is the guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Online: ISSN 1468-2079 Print: ISSN 0007-1161 Copyright © 2022 BMJ Publishing Group Ltd. All rights reserved.