Complement Cascade 8 - Alpha and Calpain-2 in Extracellular Vesicles of Human Vitreous as Biomarkers of Infectious Endophthalmitis

Abstract

Purpose: Extracellular vesicles (EVs) are messenger pigeons of the cells that communicate about cellular microenvironment. In this study, we evaluated the expression of C8α and calpain-2 in EVs from vitreous of patients with bacterial endophthalmitis to assess its utility as a diagnostic marker.

Methods: EVs were isolated from vitreous of patients with bacterial endophthalmitis (culture positive and culture negative) and noninfectious control by exosome isolation reagent and characterized, and the levels of C8α and calpain-2 was assessed by enzyme-linked immunosorbent assay in isolated EVs and direct vitreous. The receiver operating characteristic curve was generated to assess the diagnostic performance.

Results: Scanning electron microscopy (SEM) and dynamic light scattering (DLS) confirmed the presence of EVs having a diameter (nm) of 275.2 ± 93, 92 ± 22, and 77.28 ± 12 in culture-positive (CP), culture-negative (CN), and control respectively. The expression level (ng/mL) of C8α in the EVs obtained from CP was 144 ± 22 and CN was 31.2 ± 9.8, which was significantly higher (P < 0.01) than control 3.7 ± 2.4. Interestingly, C8α is not expressed directly in the vitreous of CN and controls. Calpain-2 was significantly downregulated (P ≤ 0.0001) in CP (0.94 ± 0.16) and CN (0.70 ± 0.14) than control. The sensitivity and specificity of 1 for C8α and calpain-2 in the EVs implied that its diagnostic accuracy was significant.

Conclusions: This study showed that the EV proteins C8α and calpain-2 could be suitable diagnostic markers for endophthalmitis. However, the presence of C8α in the EVs of CN samples but not in direct vitreous promises EVs as the future of diagnostics.

Translational Relevance: Expression levels of EV-calpain-2 and EV-C8α could diagnose CN bacterial endophthalmitis.

Introduction

Endophthalmitis is an ophthalmic emergency characterized by infection and inflammation in the inner coats of the eye often resulting in vision loss despite prompt treatment.^1,2 Although endophthalmitis can be caused by trauma or via hematogenous spread, postoperative cases have been on the rise because of increased cataract surgeries in the aging population.^3,4 Although Staphylococcus spp. is the major causative gram-positive pathogen, Pseudomonas aeruginosa dominates the gram-negative profile of organisms, especially in India.⁵ Because the onset of infection is rapid and presentation of patients to the institute is delayed, a quicker and more accurate diagnosis with appropriate treatment initiation is crucial. However, culture negativity of the pathogens, which as reported by our institute crosses 70% of the total endophthalmitis cases diagnosed, presents a major challenge in managing this condition.⁵ This calls for development of an advanced diagnostic method that could possibly improve diagnosis or prognosis. 

Extracellular vesicles (EVs) have long been applauded for their diagnostic and therapeutic potential because they are crucial regulators of cellular communication and signaling via their molecular cargo composed of nucleic acids, metabolites, and proteins.^6,7 These vesicles range from 30 to 1000 nm in diameter and are categorized into exosomes, microvesicles, and apoptotic bodies based on their size, function, and biogenesis.^8,9 They may help in host immune response or may trigger worse infection depending on their packed molecular cargo.¹⁰ We have previously investigated their role in Staphylococcus aureus– and Pseudomonas aeruginosa–infected retinal cells (ARPE-19) and mice (C57BL/6) model, revealed the role of intravesicular proteins such as cathepsin D, complement component 5a, complement component 8α (C8α), calpain-2, tenascin and caveolins in the pathogenesis of endophthalmitis.^11–13 Although all three subunits, C8α, C8β, and C8γ were present in the EVs during S. aureus and P. aeruginosa endophthalmitis mouse models, with significantly (P < 0.05) higher expression levels of C8β during infection; interestingly, C8α was present only in infection and not in the control, making it an ideal marker to explore for the diagnosis of endophthalmitis, especially in culture-negative cases. However, the proteome profiling studies could not hypothecate the differential abundance of the proteins present in EVs, and future studies could possibly provide insight into the mechanism of EV cargo packaging. 

Furthermore, C8α is an important part of the membrane attack complex (MAC) along with other complement proteins such as C5b, C6, C7, and C9. Assembly of MAC drives lysis and death of the infected cells.¹⁴ Additionally, calpains are heterodimers that are calcium-dependent cysteine proteases and are largely implicated in several inflammation-associated diseases. Calpain-1 and calpain-2 are the predominant isoforms that are majorly studied for their involvement in bacterial clearance and neutrophil migration. Calpain-2 predominates in macrophages, which is responsible for its hyperactivation and drives NFκB-associated inflammatory pathway.^15,16 However, its increased level is also associated with synaptic plasticity and neurodegeneration.¹⁷ In this study, we quantitatively assessed both C8α and calpain-2 expression in EVs released in the vitreous of patients with microbiologically confirmed, as well as clinically presumed, culture-negative cases to assess its feasibility in the diagnosis of endophthalmitis. The expression levels of EV-calpain-2 and EV-C8α could diagnose bacterial endophthalmitis, especially in culture negative cases. 

Material and Methods

Ethics Statement

This study was performed at our tertiary eye care center in South India and was approved by the Institutional Ethics Board (Ref No. LEC-BHR-P-04-21-619). All procedures were performed in accordance with the tenets stated in the Declaration of Helsinki following the patient's approval on the informed consent form. Patient details such as age, gender, endophthalmitis type (exogenous or endogenous), visual acuity, microbiological workup, and treatment plan were collected from the in-house electronic medical records. 

Subjects and Vitreous Sampling

A total of 35 vitreous samples from patients presenting to our institute between September 2020 and December 2022 were included in the study. These included 15 culture positive for bacteria and 10 samples that were culture negative but presumed clinically as endophthalmitis. Additionally, 10 vitreous samples from patients undergoing vitrectomy for noninfectious retinal conditions were included as controls. Undiluted vitreous 300 µL from these patients was collected in sterile syringes directly from the operating room and transferred to 1.5mL Eppendorf in a sterile environment before processing for microbiological workup according to our institutional protocol.⁵ 

Inclusion Criteria

For controls, vitreous (200-300 µL) from patients undergoing vitrectomy for any noninfectious retinal disorders such as macular hole and retinal detachment was collected. Vitreous samples of patients clinically presumed with bacterial endophthalmitis were collected for routine diagnostics, and after microbiological workup the samples were divided into two groups, culture positive and culture negative. Culture-negative samples included in the study were clinically presumed as bacterial endophthalmitis and confirmed for the presence of pathogen either by direct examination of the samples stained by Gram's stain or by polymerase chain reaction (PCR) for 16S rRNA gene. 

Exclusion Criteria

The patients diagnosed with any ocular infections or complications excluding macular hole or retinal detachments were not considered as controls. The presumed infectious endophthalmitis vitreous samples were processed as per our institutional protocol, which includes inoculation of samples on sabouraud dextrose agar (SDA) and potato dextrose agar media and PCR of the internal transcribed spacer (ITS) region, as well as to check the presence of fungus. The samples positive for fungus were not included in our study. Additionally, patients with mixed and viral infections were also excluded in the study. 

Microbiological Workup

Vitreous sample (approximately 100 µL) was directly inoculated onto various media such as blood agar, chocolate agar, brain heart infusion broth and potato dextrose agar were incubated at 37°C for 24 to 48 hours. Bacteria grown was further identified on the basis of morphological characteristics and VITEK-2 compact system. Additionally, vitreous was also examined by direct microscopy after Gram staining. 

Sample Processing and EV Isolation

Approximately 200 µL of vitreous was processed for further experiments after routine diagnostics. Vitreous was spun in a centrifuge at 10,000 g for 15 minutes to remove cellular debris and immediately filtered with 0.22 µm filter to remove larger sized EVs. Exosome isolation reagent (Invitrogen, Carlsbad, CA, USA) was used to isolate EVs from the vitreous according to manufacturer's instructions. Briefly, 100 µL of reagent was added to 100 µL vitreous and incubated at 4°C overnight (∼12 hours) to precipitate EVs and spun in a centrifuge at 10,000 g for 90 minutes to pellet down smaller EVs. The pellet was resuspended in 100 µL of 1X PBS, and remaining vitreous and isolated EVs were immediately transferred to −80° C until further processing. 

Characterization of EVs Derived From Vitreous Sample

Dynamic Light Scattering

Determination of the size of the EVs was done using freshly isolated EVs diluted in sterile water (1:1000), and 10 spectra were recorded for each EV fraction (three biological replicates, three technical replicates) at 25°C using light scattering analyzer (Nicomp Nano Z3000 ZLS; Entegris, Billerica, MA, USA). 

Scanning Electron Microscopy

For SEM (Evo 18, Carl Zeiss), 4% glutaraldehyde (Sigma-Aldrich Corp., St. Louis, MO, USA) was used to fix the 10 µL of isolated EVs for about an hour, washed in PBS, and dehydrated with various concentrations of ethanol (40%, 60%, 80%, 96%–98%), followed by drying at room temperature for 24 hours and gold–palladium sputtering before imaging and analysis. 

Colorimetric Assay

The levels of human calpain-2 (MBS2887078) and complement component 8-alpha polypeptide chain (C8α) (MBS924799) were estimated using an enzyme-linked immunosorbent assay kit with monoclonal antibodies according to the manufacturer's instructions (MyBioSource, San Diego, CA, USA). Total protein concentration of vitreous fluid and extracellular vesicles isolated from vitreous was estimated by bicinchoninic acid (BCA), and all samples were lysed by radioimmunoprecipitation assay (RIPA) buffer and diluted using dilution buffer provided in the kit to equalize concentrations. Lysis is a crucial step to ensure the complete release of intravesicular contents into the solution for the antigen-specific-antibody binding.¹⁸ To the 96-well antibody-coated plate, 50 µL of the sample and 100 µL horse radish peroxidase conjugate was added and incubated for one hour. After the washing step, chromogenic solution was added, and the plate was incubated for 10 minutes, which was followed by the addition of stopping buffer. The optical density (OD) was measured using UV-Vis spectrophotometer (BT 2000 Microkinetics) at 450 nm. The concentration of vitreal proteins and extracellular vesicles proteins (Calpain-2 and C8α) was then determined using the standard curve generated. 

Statistical Analysis

Mean range was used to illustrate the data because non-normal distribution of data was seen in all the data. Unpaired t-test was performed for DLS and significance of the cytokine concentration difference was assessed by nonparametric Mann-Whitney test. The diagnostic performance of the proteins was evaluated using a receiver operating characteristic (ROC) curve analysis, and area under the curve (AUC) was calculated. The significance of AUC was assessed by logistic regression. Additionally, the sensitivity and specificity values were evaluated and the maximum value of Youden's index was used as a criterion for selecting the optimum cut-off point. P value <0.05 was considered statistically significant. 

Results

Demographic Features of Study Patients

A total of 15 patients clinically diagnosed as bacterial endophthalmitis (two endogenous, nine posttraumatic, and four postoperative cases) were included in the study. The average age of the cohort was 36.93 ± 19.4 years (range 6-66 years), and 80% (n = 12) were males. Similarly, the mean age of the controls was 52.62 ± 17.12 years (range 18–70 years), and the group included nine males. Among 10 patients with culture-negative endophthalmitis, the average age was 50.78 ± 6.93 years (range 1-74 years) and 50% were males. Of 10 samples, four were smear positive for bacteria, six were PCR positive for eubacteria, and among them, two samples were both smear and PCR positive. Microbiological workup of culture positive samples identified S. aureus (n = 3), P. aeruginosa (n = 3), Staphylococcus epidermidis (n = 4), Bacillus spp. (n = 3) and Streptococcus pneumonia (n = 2) in the test group. Nine out of ten patients with culture negative endophthalmitis were diagnosed with postoperative endophthalmitis, and one was diagnosed with endogenous endophthalmitis. 

Clinical Outcome

All patients in the culture-proven study group (n = 15) were treated by a combination of vitrectomy and intravitreal antibiotics (vancomycin/ceftazidime) at least twice, but the visual outcome in 14/15 patients remained poor (≥20/200), and one eye underwent evisceration. Similarly, patients in the culture-negative study group (n = 10) were treated by a combination of vitrectomy and antibiotics at least once, but the visual outcome in 8/10 patients remained poor (≥20/200) patients. 

Characterization of EVs by DLS and SEM

The isolated EVs were characterized according to size and morphology using DLS and SEM. DLS revealed the size of EVs in control was 97.5 ± 32.5 nm, in culture positive was 181.1 ± 49.8 nm and in culture negative was 125 ± 44.38 nm which correspond to the size of exosomes (50–200 nm). Additionally, SEM also confirmed the presence of EVs and the size obtained was similar to the average size obtained by the DLS (Fig. 1). 

Figure 1.

 

Characterization of EVs by SEM and DLS. 10µL of isolated EVs derived from control (A), culture positive (B) and culture negative (C) were fixed with glutaraldehyde and subjected to SEM to confirm their presence and morphology. (D) Hydrodynamic size distribution profile of isolated EVs measured by DLS. Mean ± SEM of the diameter sizes measured. All data are represented as the mean ± SEM.

Figure 2.

 

EV-Calpain-2 expression levels in the human vitreous. The scatter plot represents the expression levels of calpain-2 in patients’ vitreous samples in the study group—Control, bacterial culture positive and culture negative. Statistical analysis was performed using the unpaired t-test by comparing control. ****P < 0.0001

Figure 3.

 

EV-C8α levels in the study group. The scatter plot represents the mean C8α concentration in EVs derived from patients’ vitreous samples in the study group—Control; bacterial culture positive and culture negative. Statistical analysis was performed using the unpaired t-test by comparing control. **P < 0.01, ***P < 0.001, ****P < 0.0001

Evaluation of Calpain-2 Levels in EVs and Direct Vitreous

The human calpain-2 level was significantly low in EVs derived from culture positive bacterial (0.94 ± 0.64) endophthalmitis patients in comparison to controls (4.54 ± 1.94, n = 10) (P < 0.0001). Additionally, calpain-2 levels in EVs obtained from culture-negative but clinically presumed bacterial endophthalmitis was also significantly low compared to EVs derived from culture-proven bacteria (0.70 ± 0.14) (P < 0.0001). We also quantified calpain-2 directly in the vitreous of controls (4.92 ± 0.59), culture-positive (0.79 ± 0.10), and culture-negative (1.23 ± 0.76) samples that correlated to the concentrations in the EVs Figure 2. 

Evaluation of C8α in EVs and Direct Vitreous

C8α levels were significantly high in EVs isolated from culture proven bacterial (144.0 ± 22.1) endophthalmitis patients in comparison to controls (4.5 ± 2.8, n = 10) (P < 0.0001). Interestingly, C8α levels in EVs obtained from culture-negative but clinically presumed bacterial endophthalmitis (31.20 ± 9.8, n = 10) (P < 0.01) was also significantly high compared to EVs in comparison to controls. C8α levels in vitreous of culture-positive bacterial endophthalmitis was around 137 ± 25.86, which was extremely significant (P < 0.001) in comparison to control (1.20 ± 1.20). However, surprisingly, in the vitreous of culture-negative samples, the C8a level was 0.25 ± 0.7. Although antibody specific for C8α was used, it could possibly bind to complete C8 protein as well Figure 3. 

ROC Curve Analysis Reveals the Promising Diagnostic Potential of EV-Calpain-2 and EV-C8α

To evaluate the diagnostic potential of EV-calpain-2 and EV-C8α as diagnostic marker for culture-negative bacterial endophthalmitis ROC curve was determined. The analysis for the EV-Calpain-2 in a culture negative group showed a threshold of 0.8ng/mL, with a sensitivity and specificity of 1 (95% CI = 0.83-1). The analysis showed an AUC of 1 with a significance of P < 0.001 (Fig. 4). Additionally, the EV-C8α analysis of a culture positive vitreous samples demonstrated a sensitivity of 0.88 and specificity of 0.87 (95% CI = 0.83–1) with the threshold of 0.18 ng/mL (P < 0.001). The positive predictive value was 0.8, whereas the negative predictive value was 0.93. EV-C8α level in the EVs of culture negative samples was well above the threshold value in 70% (7 of 10) samples implying the potential to be used as a marker for culture-negative endophthalmitis (Fig. 4). 

Figure 4.

 

ROC curve and 95% confidence bounds for calculating the cutoff value for EV-calpain-2 and EV-C8α, and assessing their as a diagnostic marker for culture negative bacterial endophthalmitis. ROC curve for (A) Calpain-2, control versus culture-positive, and (B) C8a, control versus culture-positive, were plotted using R software. Positive likelihood ratio (LR+), negative likelihood ratio (LR−), sensitivity, and specificity are reported for the best cutoff values found in each ROC.

Discussion

Culture negativity of the pathogens hinders prompt diagnosis, especially in endophthalmitis, which further delays treatment initiation and puts visual recovery at stake. With nearly 50% of samples being culture negative the world over^19,20 and more than 70% of culture-negative endophthalmitis cases in our institute alone,⁵ there is an earnest need to develop an alternative diagnostic method targeting culture-negative cases. EVs have gained significant attention in the field of diagnostics because of their potential as biomarkers for various diseases. In our previous in vitro and in vivo studies, we reported that calpain-2 was upregulated, whereas C8α was present exclusively in the EVs of infected samples.^11–13 In this study, we validated these markers in direct vitreous and EVs isolated from the vitreous samples diagnosed with bacterial endophthalmitis. A total of 70 samples from 35 patients (10 vitreous controls, 10 EV controls, 15 culture positive vitreous, 15 EVs from culture positive, 10 culture negative vitreous, and 10 EVs from culture-negative samples) were used in this study. After the EV isolation, expression levels of calpain-2 and C8α was assessed by enzyme-linked immunosorbent assay and ROC curve was generated to evaluate their diagnostic potential in bacterial endophthalmitis especially for culture negative cases. The results showed that the EV-calpain-2 levels (ng/mL) were significantly higher in control (4.54 ± 1.94) in comparison to culture negative (0.94 ± 0.64) samples and the similar level of expression was observed in direct vitreous of control (4.92 ± 0.59) and culture negative (1.23 ± 0.76) samples. 

Interestingly, the results were not in accordance with our in vivo studies where the calpain-2 levels were higher in infection when compared to control. Calpain-2 is a calcium-activated protease that is majorly involved in activation of NF-κB pathway playing a major role in acute inflammation.²¹ Additionally, they are also associated with neurodegeneration and lower levels of calpain-2 is linked to neuroprotection. We hypothesize that the lower cargo of calpain-2 in the EVs and in vitreous could attribute to its neuroprotective role in the humans.²² Furthermore, regulating excess inflammation could also preserve retinal architecture thereby preventing vision loss. Complement component is a core part of immune system, and C8α is a part of MAC that is responsible for pathogen killing and controlling infection.²³ Although EV-C8α levels were significantly higher in culture-negative (31.20 ± 9.8) when compared to control (4.5 ± 2.8), in the direct vitreous, however, the difference between culture-negative (0.25 ± 0.7) and control (1.20 ± 1.20) was not significant. As reported by Boukouris et al. (2015)²⁴, the body fluids consist of diverse and vast majority of the proteins which makes it increasingly difficult to assess biomarkers. Although this discrepancy has no reasoning yet, the extracellular vesicles cargo is relatively smaller which explains the differential C8α level in vitreous and EVs and this highlights the potential of EVs in diagnosis of culture negative endophthalmitis. Although, several complement proteins have been explored as potential biomarkers for ocular diseases, such as C3 upregulation is strongly associated with age-related macular degeneration, and C7 is reported to discern diabetic retinopathy cases with control; however, C8 is not explored in any ocular diseases especially in infections. The higher expression of C8α could signify the active participation of the host immune system in resolving the infection. Additionally, the significant differential expression of both the proteins (calpain-2 and C8α) in the EVs of culture-negative and control samples further promises their potential to be developed as a diagnostic marker. Another noteworthy finding in our in vivo study of S. aureus endophthalmitis model was the downregulation of complement component 5 (C5), and the breakdown of which into C5a and C5b is a critical step for the initiation of MAC.^23,25–28 However, in this study we did not find any significant difference in the C8α expression in S. aureus endophthalmitis vitreous samples versus any other bacterial vitreous samples. 

Deshmukh et al.²⁹ have reported the potential of cytokines as biomarkers for infectious culture negative endophthalmitis. However, instead of a panel of cytokines, assessment of single protein EV-C8α is a good alternative since it is relatively cost effective, rapid and feasible to predict the presence of endophthalmitis in over 70% of the cases. The limitation of our study is the inclusion of lesser sample size because of which the subgroup analysis of protein (calpain-2 and C8α) expression levels of different bacterial species could not be performed because, when grouped, the sample size is statistically inadequate. Analysis of protein level in gram-positive versus gram-negative revealed a slight increase of C8α in the gram-negative, although there was no significant difference in the expression levels of both proteins. Additionally, performing Western blot would have further validated the specificity of the C8α antibody; however, because of the limited availability of the patient vitreous samples, this additional step could not be performed. Overall, EV has been a swiftly emerging field, and research has extended to every possible area to use its diagnostic and therapeutic potential. Bacterial endophthalmitis, especially culture-negative cases, have seen a dramatic increase because of increasing use of empirical therapy and lack of ocular microbiology facilities. Because of its rapid manifestation, endophthalmitis is considered an ophthalmic emergency, and the extensive involvement of the retina often leads to irreversible vision loss. A study at our institute previously reported that more than 70% of total endophthalmitis cases are culture negative, which delays diagnosis further.⁵ Our study highlights that with sensitivity and specificity of 1 with 95% class interval, EV-C8α could be a promising quick diagnostic marker for culture-negative bacterial endophthalmitis. Although with current advancements, isolation of pure EVs is still challenging and time consuming, with the rapid advancement of research on EVs, especially exosomes, it is not too far to fully realize and use their potential in the clinics. 

Acknowledgments

Supported by DST-SERB (J.J.) (File Number: CRG/2019/004386) India and Hyderabad Eye Research Foundation (HERF). 

Disclosure: D. Rudraprasad, None; V. K., None; J. Nirmal, None; M.H. Ali, None; J. Joseph, None