Particle-Induced Endophthalmitis: Possible Mechanisms of Sterile Endophthalmitis after Intravitreal Triamcinolone

Abstract

Purpose.: Wedetermined the cause of sterile endophthalmitis after intravitreal triamcinolone acetonide (IVTA).

Methods.: The clinical records of 21 eyes that had sterile endophthalmitis after IVTA were reviewed. Cytologic studies were performed, and cytokine and chemokine profiles in the aqueous humor were determined. The effects of a direct contact of triamcinolone acetonide (TA) particles with cultured lens epithelial cells (B3) or retinal pigment epithelial cells (ARPE-19) on cytokine production were determined. Noncontact culture studies were performed in a Boyden chamber; 11-deoxycortisol was used for nonbioactive particles.

Results.: Inflammation appeared one day after the IVTA, and the incidence of severe sterile endophthalmitis was 13.0% (6/46), which was reduced to 4.3% (2/47) after switching to preservative-free TA. Most cells found in the aqueous humor were granulocytes. The concentrations of IL-6 (447.8–<5000 pg/mL) and IL-8 (47.2–<5000 pg/mL) were increased significantly, but IL-1β, IL-10, IL-12p70, and TNF-α were not significantly changed. In B3 cells, the level of IL-8 increased significantly when the cells made direct contact with TA particles (1.33 times that of PBS, P < 0.05, Student's t-test), but not in noncontact cultures. Similar reaction pattern was found in ARPE-19 cells. These effects also were observed when the cells were exposed to 11-deoxycortisol.

Conclusions.: Sterile endophthalmitis after IVTA is characterized by an immediate granulocytic infiltration and an increase of IL-6/IL-8 in the aqueous humor. Preservatives may be a factor causing this condition, and mechanical/rheologic stress by particles presumably might be a new causative factor by increasing IL-6/IL-8.

Introduction

Intravitreal triamcinolone acetonide (IVTA) has been used to treat various ocular diseases. ¹ Although IVTA is effective, several complications have been associated with its use. One of the major problems of IVTA use is the development of endophthalmitis. In some cases, pathogenic microorganisms cannot be found by conventional methods, and the inflammation subsides spontaneously with fairly good visual prognosis. ^2–13 These cases are classified as having sterile endophthalmitis. However, IVTA also can be associated with the development of infectious endophthalmitis. The early clinical symptoms and signs of these two types of endophthalmitis are similar, and it is difficult to differentiate them at an early stage. This then makes it difficult to prescribe the appropriate treatment, namely no treatment for sterile endophthalmitis and vigorous treatment for infectious endophthalmitis. Identifying the type of endophthalmitis is important because sterile endophthalmitis has a somewhat benign course, while infectious endophthalmitis can lead to vision-threatening complications. 

It has been speculated that sterile endophthalmitis after IVTA is related to preservatives contained in commercially available triamcinolone acetonide (TA). ^3–13 Indeed, the incidence of sterile endophthalmitis after IVTA was reported to have dropped after switching to preservative-free triamcinolone (PFTA). ^4,8,10,13 Regardless, there still is an incidence of sterile endophthalmitis even with PFTA. ^4,8–10 Another possible cause is contamination of endotoxin, because there was an epidemic cluster of sterile endophthalmitis. ^3–10,12 Nevertheless, there still are sporadic occurrences of sterile endophthalmitis. Therefore, it is assumed that an unknown mechanism other than these two factors is associated with this disease. ^9,10  

Many reports indicate there is a risk of sterile endophthalmitis in eyes with a history of intraocular surgery or multiple injections. ^2–5,7,10 In these eyes, TA particles often were observed in the anterior vitreous and freely attaching to intraocular tissue, suggesting that interaction of particles and intraocular tissue might have some role in disease progression. It is accepted widely that mechanical or rheologic stress, including particles, induces immediate inflammation in various tissues. ^14–18 Of importance is that the mechanical effect on the induction of inflammatory cytokines is augmented strongly in the presence of particles. ¹⁶ These facts lead to the thesis that particles of TA might induce intraocular inflammation. In our study, we reviewed cases of sterile endophthalmitis after IVTA and examined intraocular samples. Then, inflammatory response to mechanical stress on lens epithelial cells, retinal cells, and retinal pigment epithelial cells was evaluated using an in vitro assay. It is possible that this type of inflammation may be a unique and new clinical entity of intraocular inflammation. 

Subjects and Methods

The procedures used in our study were approved by the Institutional Review Board of the Kagoshima University Hospital, Kagoshima, Japan. The procedures conformed to the tenets of the 1989 Declaration of Helsinki, and a written informed consent for the treatment and the study was collected from every participant. 

The medical records of subjects diagnosed with sterile endophthalmitis after IVTA were reviewed. All cases had been examined between January 2009 and December 2011. Sterile endophthalmitis was diagnosed when inflammation was detected in the anterior chamber and/or vitreous cavity by slit-lamp biomicroscopy within 1 week after the IVTA. Most importantly, no organisms were detected in the aqueous humor and vitreous samples by gram staining, and standard cultures of intraocular specimens did not grow any pathogenic organisms. If specimens had not been collected, patients were defined as having sterile endophthalmitis when the inflammation cleared with only topical medications within one month. 

Clinical symptoms, findings, systemic conditions, treatments, and final results were studied. IVTA was performed for diabetic retinopathy, retinal vein occlusion, or age-related macular degeneration according to our previously described method. ^19,20 When visual acuity (logMAR) at the last visit decreased 0.2 or more in comparison with that pre-injection, it was categorized as “improved.” If it increased 0.2 or more, it was categorized as “deteriorated.” Intermediate cases were categorized as “stable.” 

Injections were performed according to the accepted technique. ^2,10,20 TA (4 mg, Kenacort-A; Bristol-Meyers, Tokyo, Japan) was injected through the pars plana. TA was injected without removing preservatives before September 2010. Based upon the national survey of sterile endophthalmitis after IVTA in Japan, the Institutional Review Board (IRB) of Kagoshima University Hospital recommended to use TA without preservatives (PFTA). ² Therefore, preservatives were removed with a 0.22 μm filter (Millex GP; Millipore, Cork, Ireland) and PFTA was administered after October 2010. On the day after surgery, each eye was examined by a specialist. If signs of endophthalmitis were found, such as moderate (cell 2+) or more severe inflammation in the anterior chamber/vitreous, or severe conjunctival injection, the patient was examined further for endophthalmitis. The aqueous humor was obtained by paracentesis for bacterial culture where possible. The remaining specimen was kept at −80C for further analysis. When this retrospective study was approved by the IRB, five specimens were available for cytokine/chemokine analysis. Among them, four specimens were used for cytokine/chemokine analysis with patients' agreement. 

Cytological and Cytokine/Chemokine Analysis

For cell smear studies, the cytospin technique was used to maximize the number of cells in the aqueous samples. ²¹ Approximately 0.1 mL of fluid samples with an equal amount of cytospin collection fluid (Thermo-Shandon, Pittsburg, PA) was added to the cytospin slides. The slides were assembled with filter cards, and the delivery chamber was secured by a metal clip (Thermo Scientific, Tokyo, Japan). After spinning at 1800 rpm for 2 minutes, the slides were removed from the cytospin chamber and fixed in 70% alcohol. The fixed specimen was stained with hematoxylin and eosin, and examined under a microscope (BX51; Olympus, Tokyo, Japan) at ×100 magnification. 

For cytokine and chemokine assays, the concentration of inflammatory molecules (IL-1β, IL-6, IL-8, IL-10, IL-12p70, and TNF-α) were measured using a cytometric bead array (BD Biosciences, San Diego, CA), according to the manufacturer's instructions. ^19,22 Samples of aqueous humor obtained from eyes without endophthalmitis were used as controls. The detection limits for IL-1β, IL-6, IL-8, IL-10, IL-12p70, and TNF-α were 7.2, 2.5, 3.6, 3.3, 1.9, and 3.7 pg/mL, respectively. All concentrations less than the detection level were assigned a value of zero in the subsequent analysis. ^19,22  

Cell Culture Studies

ARPE-19 cells and B3 cells were obtained from ATCC (Manassas, VA). ARPE-19 cells were used to represent retinal pigment epithelial cells, and B3 cells to represent lens epithelial cells. R28 cells to represent neural retinal cells were donated by Gail M. Siegel, PhD (The State University of New York, Buffalo, NY). These cells were cultured according to our previous method. ^21,23 Briefly, cells were cultured in modified Eagle medium (MEM) in cell culture dishes and cultured for 24 hours in the serum-free medium before the following studies were performed. 

Effects of Contact of Cultured Cells with TA Particles

The effects of direct contact of the cultured cells with the TA particles on the expression of cytokines and chemokines were evaluated in a double chamber system as reported in detail (Fig. 1). ²⁴ This was done with a culture plate membrane filter system (Cell cultureware; BD Falcon, Tokyo, Japan) based on the Boyden chamber (BD Biosciences). For positive control, a suspension of 11-deoxycortisol (11-DC; Sigma-Aldrich, St. Louis, MO) was made, which contained particles of same size as TA without steroid activity. ²⁵ A suspension of TA or 11-DC at a final concentration of 0.08 to 0.8 mg/mL was added directly onto the cell layer to examine the effects of contact of the TA or 11-DC particles on the expression of cytokines in the cultured cells. These cultures were called contact cultures. In other experiments, the TA or 11-DC particles were added onto the protective filter insert that prevented direct contact between the particles and cells. These were called noncontact cultures. The same volume of the supernatant of TA containing the preservatives was added to evaluate the effects of the preservatives. The commercial Kenacort-A TA preparation was allowed to sit for 60 minutes, and the supernatant was collected. Culture medium with the same volume of PBS as that of the TA suspension was used as a control. The cell viability was evaluated after 24 hours of culture with the WST-1 assay kit (Dojindo, Kumamoto, Japan), according to the manufacturer's instructions. 

Figure 1

 

Schematic drawing of direct contact cell culture (A) and noncontact cell culture (B). Cells were incubated in direct contact or without contact with drug particles to evaluate mechanical stress-induced cytokine production. Direct contact was blocked by a membrane.

Effects of Lipopolysaccharide

Because common contaminants of drugs are lipopolysaccharides (LPS; Sigma-Aldrich), different amounts of LPSs were added to the MEM to test its effect on the cultured cells. Our preliminary studies showed that TA and/or bevacizumab contained less than 20 pg/mL of LPS, and thus cells were cultured with MEM containing 1000 to 100,000 pg/mL of LPS. After 24 hours, the cytokine and chemokine analyses were performed as described. 

Statistical Analysis

The concentrations of the inflammatory cytokines and chemokines in the supernatant of cultured cells were compared with the Mann-Whitney U test. All of the in vitro data are presented as the means ± SEMs, and the significance of differences were determined by the Student's t-tests. P < 0.05 was considered to be significant. 

Results

One day after IVTA, a careful slit-lamp examination showed that moderate to severe inflammation was present in the anterior chamber and/or vitreous in 21 eyes, which were considered to be potential sterile endophthalmitis cases (Fig. 2, Table 1, and see Supplementary Material and Supplementary Table S1). The control samples were collected before IVTA from some patients who did not suffer endophthalmitis (see Supplementary Material and Supplementary Table S1). Cultures of specimens from 11 of these eyes grew no pathogenic organisms. 

Figure 2

 

Representative photograph of the anterior eye segment with sterile endophthalmitis the day after the intravitreal triamcinolone acetonide injection. Numerous cells are present in the anterior chamber together with fibrin.

Table 1. 

 

Clinical Findings of Sterile Endophthalmitis after Intravitreal Triamcinolone Acetonide

Patient No.	Time to Presentation, h	Pain	AC Reaction	Hypopyon	Vitreitis	Treatment	Cultures	Visual Acuity, logMAR
								Before Injection	Final Visit
1	<24	N	Y (3+)	N	Y	Topical	N.P.	0.3	0.15
2	<24	N	Y (3+)	Y	Y	PPV	Negative	1.22	1.1
3	<24	N	Y (3+)	N	Y	Topical	N.P.	0.22	0.1
4	<24	N	Y (2+)	Y	Y	Topical	Negative	2	1.4
5	<24	N	Y (3+)	Y	Y	Topical	Negative	0.7	0.52
6	<24	N	Y (3+)	Y	Y	Topical	Negative	1.05	1.15
7	<24	N	Y (3+)	N	Y	Topical	Negative	1.3	N.R.
8	<24	N	Y (2+)	N	N	Topical	Negative	0.4	0.82
9	<24	N	Y (2+)	N	N	Topical	N.P.	0.82	1
10	<24	N	Y (2+)	N	Y	Topical	Negative	1	0.52
11	<24	N	Y (2+)	N	N	Topical	Negative	1.1	1.1
12	<24	N	Y (3+)	Y	Y	PPV	Negative	1.1	0.7
13	<24	N	Y (3+)	Y	Y	Topical	N.P.	1	0.82
14	<24	N	Y (3+)	Y	N	Topical	N.P.	0.52	0.15
15	<24	N	Y (3+)	Y	Y	Topical	N.P.	0.3	0.3
16	<24	N	Y (2+)	N	Y	Topical	N.P.	0.7	0.15
17	<24	N	Y (2+)	N	N	Topical	Negative	1	0.52
18	<24	N	Y (2+)	N	N	Topical	N.P.	1.05	0.7
19	<24	N	Y (2+)	N	N	Topical	Negative	0.22	0.22
20	<24	N	Y (2+)	N	Y	Topical	N.P.	0.3	0.82
21	<24	N	Y (2+)	N	N	Topical	N.P.	0.7	0.4

 AC, anterior chamber; PPV, pars plana vitrectomy; Topical, topical medication drops; N.P., not performed; N.R., not recorded.

The major complaint of these patients was mild to severe blurring of the eyes, but no ocular pain. Eye drops containing antibiotics and corticosteroid were administered, and the inflammation subsided in 19 cases within 1 week. Two cases with severe inflammation were treated by vitrectomy. The final BCVA was improved in 8/20 (40%) eyes diagnosed with sterile endophthalmitis, stable in 10/20 (50%) eyes, and deteriorated in 2/20 (10%) eyes. The final visual acuity of one case was not recorded (Table 1). 

The incidence of sterile endophthalmitis was 28.3% (13/46) injected with preservative-containing TA before September 2010, and the incidence decreased to 17.0% (8/47) after switching to preservative-free TA suspension. When the definition of sterile endophthalmitis was changed to the presence of hypopyon or severe visual disturbances as was used in earlier reports, ^4,7,8,12,13 the incidence was 13.0% (6/46) with preservative-containing TA suspension and 4.3% (2/47) with PFTA. 

Cytology and Cytokine Analysis

Cytologic analyses were performed on the aqueous humor from cases 2, 3, 5, and 8. Most of the cells were mature granulocytes (Fig. 3), and lymphocytic cells were seen rarely. Crystal-shaped particles that resembled TA particles were not detected in these samples. 

Figure 3

 

Analyses of aqueous humor of eyes with sterile endophthalmitis after IVTA. Numerous cells are present and most are granulocytes. Cells were collected by the cytospin method. (A) Case 8. (B) Case 3. Original magnification: ×400.

The concentrations of the cytokines detected in the aqueous humor are summarized in Table 2. The concentrations of IL-6 and IL-8 in eyes with inflammation were significantly higher than those without inflammation (Mann-Whitney U test, P = 0.02, both). The differences for the other four cytokines between eyes with and without inflammation were not significant. The levels of IL-6 and IL-8 were very high in samples from cases 19 and 21 that had severe inflammation. In contrast, the concentration of other cytokines was not significantly different from the controls or was less than the detection level. The ratio of IL-6/IL-8 ranged from 2.2 to 9.5. 

Table 2. 

 

Aqueous Humor Levels of Cytokines/Chemokines

Case	IL-12p70, pg/mL	TNF-α, pg/mL	IL-10, pg/mL	IL-6, pg/mL	IL-1β, pg/mL	IL-8, pg/mL	IL-6/IL-8
Sterile endophthalmitis after IVTA
18	0	4.2	0	447.8	0	47.2	9.5
19	0	0	6.1	>5000	0	>5000	–
20	0	0	0	941.7	0	420.6	2.2
21	0	0	6.4	>5000	0	4904.8	–
Control samples
Control 1	0	0	0	6.5	0	6.1
Control 2	0	0	0	87.3	0	28.5
Control 3	0	0	0	9.6	0	16.5
Control 4	0	0	0	4.3	0	10.4
P value*	–	0.32	0.13	0.02	–	0.02

*  Mann-Whitney U test.

In Vitro Studies

Although the trend of each experiment was the same, the absolute concentrations of the different cytokines varied in each experiment. The concentration of IL-6 ranged from 23 to 170 pg/mL in B3 cells and from 20 to 400 pg/mL in ARPE-19 cells. The IL-8 ranged from 20 to 600 pg/mL in B3 cells and from 20 to 700 pg/mL in ARPE-19 cells. The concentrations of the other cytokines were below the detectable levels. The differences of the cytokines were expressed as a ratio of the measured concentration to that of the control for each experiment. In R28 cells, none of the cytokines and chemokines was increased significantly. 

In the contact culture of B3 cells (with no filter), IL-8 was significantly higher with TA (1.33 × control, Student's t-test, P < 0.05) or 11-DC (1.48 × control, Student's t-test, P < 0.05) than the control. IL-6 decreased significantly with TA (0.53 × control, Student's t-test, P < 0.05), but not with 11-DC or preservatives. While, no significant difference for the other four cytokines (IL-1β, IL-10, IL-12p70, or TNF-α) was found even with any of TA, 11-DC, or preservatives compared to the control (Fig. 4A). On the other hand, in the noncontact cell culture, no significant difference was found even in IL-8; however, IL-6 decreased significantly with TA (Fig. 4B). 

Figure 4

 

Concentrations of cytokines/chemokines in cultured B3 cells. (A) Contact culture. IL-8 is significantly higher with TA or with 11-DC. (B) Noncontact culture. *P < 0.05 Student's t-test.

In the contact culture of ARPE-19 cells, IL-6 and IL-8 were significantly high with TA (2.13 × control and 2.14 × control, respectively, Student's t-test, P < 0.01, Fig. 5A). In the contact culture with 11-DC, IL-6, and IL-8 also were significantly high (2.03 × control and 1.68 × control, respectively, Student's t-test, P < 0.05). The other four cytokines did not show any significant change or the change was less than the detectable level. Preservatives did not have any significant effect in each assay, while in the noncontact cell culture no significant difference was found in any of the cytokine/chemokines examined (Fig. 5B). 

Figure 5

 

Concentration of cytokines/chemokines in cultured ARPE-19 cells. (A) Contact culture. IL-6 and IL-8 are significantly higher with TA or 11-DC. (B) Noncontact culture. *P < 0.05, **P < 0.01, Student's t-test.

Increase of IL-8 caused by B3 cells in the contact culture with TA or 11-DC showed clear dose dependency (Fig. 6). Increase of IL-6 and IL8 caused by ARPE-19 cells in the contact culture with TA or 11-DC showed clear dose dependency (Fig. 7). In the contact culture with TA, 11-DC, or preservatives, no apparent cell damage was evaluated with WST-1 assay even at the highest doses (Fig. 8). 

Figure 6

 

The IL-8 increase by TA particles in B3 cells is dose-dependent. IL-8 was increased by B3 cells in contact culture with TA or 11-DC in a dose-dependent manner.

Figure 7

 

The IL-6 or IL-8 increase by TA particles in ARPE-19 cells is dose-dependent (A). IL-6 and IL-8 are increased by ARPE-19 cells in contact culture with TA (A) or 11-DC particles (B) in a dose-dependent manner. **P < 0.01 Student's t-test (n = 4).

Figure 8

 

WST-1 assay for cell viability in contact culture. There was no significant change of cell viability of B3 cells or ARPE-19 cells even at the highest dose (0.8 mg/mL) of TA or 11-DC (n = 6). Prsv, preservative.

Effects of LPS

In cells exposed to LPS, the concentrations of IL-6 and IL-8 in the aqueous humor were increased by cultured B3 cells and ARPE-19 cells in a dose-dependent way. The increase was not found when the dose of the LPSs was <1 ng/ml. None of the other four cytokines was changed by exposure to LPS (Fig. 9). 

Figure 9

 

IL-6 and IL-8 in cultured cells stimulated by LPS. (A) B3 cells. (B) ARPE-19 cells. IL-6 and IL-8 are increased by LPS in a dose-dependent manner, but it is not increased at doses < 1 ng/ml. **P < 0.01, by Student's t-test.

Discussion

In our study, the major clinical characteristics were similar to those reported previously; however, the incidence was higher than has been reported. ^2–13 We examined every eye with slit-lamp biomicroscopy the day after IVTA, and it is likely that eyes with a mild or more severe inflammatory reaction all were included among positive cases. In earlier studies, sterile endophthalmitis was defined as severe inflammation with hypopyon or severe visual disturbances. ^{3,4,7,8,12,13} This stricter definition made the incidence of sterile endophthalmitis lower than that in our cases. However, the incidence of sterile endophthalmitis would be 13% with preservative-containing TA and 4.3% with PFTA using the same criteria as the earlier reports. The real incidence of sterile endophthalmitis might be higher when less strict criteria are used, and also if all cases were examined by slit-lamp the day after the injection. 

Our cytologic studies showed that non-granulomatous cells infiltrated the anterior chamber within 24 hours in all of our cases. Even though the visual disturbances were severe on presentation in some of our cases, the visual prognosis was good in all as in the earlier reports. ^2–13 Sterile inflammation also has been reported after intravitreous bevacizumab, but it was reported that invading cells were granulomatous cells, which differs from our findings. ²⁶ Therefore, sterile endophthalmitis after IVTA may be a different clinical entity from that after intravitreous bevacizumab. 

To the best of our knowledge, this is the first report on performing cytologic and cytokine/chemokine profile analyses of aqueous humor of sterile endophthalmitis after IVTA. It was reported that TA particles were found in aqueous samples from sterile endophthalmitis after IVTA and presumed that some of the floaters were TA particles. ⁶ In our cases, TA particles were not seen in the aqueous humor, and most of the cells were granulocytes indicating that there was an acute granulocytic reaction after the IVTA. It is difficult to explain why no TA particle was detected in aqueous humor in our study. It can be assumed as follows. Even though TA particles were present, they were comparatively smaller than granulocytes in number and they might be overlooked among granulocytes. Additionally, TA particles become degraded and fragile in the eye. As a result, they might be lost during the procedures of cytologic analysis. 

Analyses of the cytokine profile of the aqueous humor in Behcet's disease showed that IL-10 was significantly higher in cases of infectious endophthalmitis, and IL-6 was much higher than IL-8. Then, the IL-6/IL-8 ratio ranged from 29 to 60. ^27,28 In contrast, IL-6 and IL-8 were increased significantly in our cases of sterile endophthalmitis, and the IL-6/IL-8 ratio ranged from 2.2 to 9.5. Therefore, the inflammation in eyes with sterile endophthalmitis might be caused by a different mechanism than that for other types of uveitis. IL-8 probably had a significant role in this process because the IL6/IL-8 ratio was lower than that of other endophthalmitis, and IL-8 is a potent chemoattractant for granulocytes. ²⁹  

Several explanations have been proposed for the cause of sterile endophthalmitis. One involves a contamination of the TA by LPS endotoxin. In pharmaceutic industries, it is possible to have contamination by LPS during the production processes and in the final products. Although LPSs are linked within the bacterial cell wall, they are liberated continuously into environments and LPSs are found almost everywhere. In our in vitro study, LPS significantly increased IL-6 and IL-8, especially in B3 cells, thus LPS can be a candidate for the inducer of sterile endophthalmitis. However, our findings showed that LPS is not necessarily an important factor in sterile endophthalmitis after IVTA. We measured the concentration of LPS in four TA solutions that caused sterile endophthalmitis and found LPS to be less than 20 pg/mL, which is much less than the concentration required to upregulate the cytokines. Earlier studies also did not detect bacterial endotoxin in their vials of TA. ^10,12  

Another possible mechanism for the sterile endophthalmitis after IVTA is the benzyl alcohol preservative contained in most commercial TA. ^3–12 This is supported strongly by the decrease in the incidence of sterile endophthalmitis after switching from preservative-containing TA to PFTA in our study and earlier studies. ^8,10,11,13 Nonetheless, the issue still remains; if the preservatives were the only reason, the incidence should fall to zero with PFTA. However, sterile endophthalmitis still occurs; 2.5% as reported by Stepien et al. ¹⁰ and 3.5% by Lorenzo Carrero et al., ⁹ and 4.3% in our cases using the same criteria for diagnosing sterile endophthalmitis. 

Therefore, we explored other mechanisms that might cause sterile endophthalmitis after IVTA. It has been shown that mechanical and/or rheologic stress can induce inflammation in some tissues, and that exposure of cells to particles can lead to an upregulation of inflammatory cytokines in vitro. ^14–18 Good clinical example are gouty arthritis, in which soluble uric acid does not cause any inflammation. ¹⁸ We found that IL-8 was increased significantly in B3 cells exposed to TA particles, and IL-6 and IL-8 were increased in ARPE19 cells. In addition, significant increases were not found in cells in no-contact cultures. Of importance was our finding that exposure to 11-DC particles of the same size as TA with no steroid action also upregulated the inflammatory cytokines. Thus, it is highly likely that IL-8 was induced by mechanical stress from the particles rather than by a steroid action. In the eye in situ, it would be difficult for many TA particles to make direct contact with RPE cells, although the TA particles in the vitreous can contact the lens directly. Therefore, it may be possible that the TA particles increased the mechanical and/or rheologic stress on the lens epithelial cells, which induced the expression of inflammatory cytokines. Above all, IL-8 is considered to be particularly important because it is a potent trigger of granulocytic infiltration, which is compatible with our cytologic findings of the aqueous humor. This is supported by the fact that eyes with a history of intraocular surgery or multiple injections have a higher risk of sterile endophthalmitis. ^2–5,7,10 In these eyes, injected TA particles can move easily to the retrolental space and stimulate the production of IL-8 by the lens epithelial cells. 

Earlier, Szurman et al. reported that direct contact between TA particles and cultured cells caused cellular damage, including apoptosis. ^30,31 IL-8 can be produced after apoptosis by the scavenging monocytes in vivo. ^32–34 However, IL-6 and IL-8 were significantly upregulated in B3 and ARPE-19 cells without the presence of macrophages in our in vitro study. This would indicate that mechanical stress itself could be the causative factors for the initial step. Several studies have suggested that the increased cytokine production induced by mechanical stress is induced by integrin receptors, p38, and NF-κB dependent pathway. ^14–18 A similar mechanism might have been involved in our eyes. 

From our findings and those of earlier reports, we suggested that there are three major mechanisms involved in the development of sterile endophthalmitis after IVTA (Fig. 10). First, epidemiologic findings suggested that preservatives in the TA suspension is an important factor. Second, contamination by LPS might be a potential factor, and contamination by endotoxins always should be monitored carefully because there have been at least four reports of clusters of sterile endophthalmitis. ^3,4,10,12 Third, a new factor that should be considered is mechanical stress caused by the particles in the TA suspension. We may call all of them as a part of “toxic anterior segment syndrome.” 

Figure 10

 

Hypothetical schematic drawing of sterile endophthalmitis after intravitreal triamcinolone.

Even if TA is free from a toxic preservative or LPS, the TA particles can induce IL-8. The increase of IL-8 by B3 cells was significant, but less than 200% of the control, although it might be much greater in some eyes in vivo. It is likely due to the result of cellular chain reactions with positive feedbacks. Once granulocytes are recruited into the intraocular space by IL-8, pro-inflammatory molecules, such as IL-6 and IL-8, can be produced by these cells, which would augment this reaction continuously and exponentially, which can increase the concentration of inflammatory cytokines greatly as well. This phenomenon was well studied in diseases, such as psoriasis. ³⁵ In their study, the mixed culture of neutrophils and fibroblasts with inflammatory cytokine augmented the production of IL-8 more than 100 times in comparison with monoculture of these cells alone even though the number of cells were equal. ³⁵ In clinical conditions, this reaction is likely to be associated with additional increase of inflammatory cells. Thus, the great increase of IL-8 would be possible as was found in aqueous humor. When the TA particles begin to degrade and be absorbed, the mechanical stress will be reduced. In addition, steroids are potent inhibitors of inflammation, and the sterile inflammation might subside spontaneously within a short period as was found in the clinical course. Thus, sterile endophthalmitis after IVTA could be caused by more than one factor. Even though each factor alone might not be sufficient, a combination of several synergistic factors could have caused the sterile endophthalmitis. More importantly, the reaction is dependent on the sensitivity of each eye. For example, a lower threshold might be possible especially in uveitic eyes because eyes with a history of uveitis are reported to have a greater probability of suffering sterile endophthalmitis after IVTA. ⁷  

Our study has several limitations. The number of eyes that underwent cytokine profile analyses was not large. Because the primary purpose of collecting aqueous humor was to search for pathogenic microorganisms, the amount of aqueous humor that could be used for a cytokine analyses was limited. Second, we measured six different inflammatory cytokines, and other unexamined biomolecules might have had important roles in the sterile endophthalmitis. Third, based upon the pathogenic theory of disease, intraocular injection of particles should cause sterile endophthalmitis. However, rat eyes injected with TA particles did not have endophthalmitis in our preliminary studies. Nonetheless, injection of IL-8 alone can cause granulocyte-dominant endophthalmitis in rat eyes, and its clinical course is similar to sterile endophthalmitis after IVTA. ³⁶ Therefore, the increase of IL-8 is likely an inducer of sterile endophthalmitis. Fourth, the size of TA/control particles and their homogeneity were not evaluated in our study, although they were reported to be important in behavior in vitro. ³⁷ Since our study is a retrospective study, the size or the homogeneity of particles injected in the eyes could not be re-evaluated. However, this issue should be evaluated carefully in the next study. Finally, the present rate of sterile endophthalmitis still seems high no matter what criteria we used. It might be due to the fact that we always see the patients next day after an IVTA injection, while most retina specialists may not believe it necessary to do so. Therefore, we may observe cases provocatively that are subclinical without symptoms significant enough that would have prompted the patients to seek care after the IVTA injection. This possibility should be noted when the present results are interpreted. 

Our study showed sterile endophthalmitis occurs the day after IVTA with a fair visual prognosis. The incidence might be higher than was reported if there were to be a more careful examination or more sensitive criteria. The TA particles in the suspension may be one cause of this disease, because intraocular particles can induce mechanical or rheologic stress of lens epithelial cells, which in turn will upregulate the synthesis of IL-6 and IL-8. These cytokines can be the cause of the inflammation. Thus, the mechanical stress induced by TA particles may be a new causative factor for sterile endophthalmitis. This information would be important not only to physicians considering IVTA, but also to researchers developing new drug delivery systems. 

Supplementary Materials

pdf S1 

Footnotes

 Supported by a grant from the Research Committee on Chorioretinal Degeneration and Optic Atrophy, Ministry of Health, Labor, and Welfare, Tokyo, Japan; and by a grant-in-aid for Scientific Research from the Ministry of Education, Science, and Culture of the Japanese Government, Tokyo, Japan. The authors alone are responsible for the content and writing of this paper.

Footnotes

²  These authors contributed equally to the work presented here and should therefore be regarded as equivalent authors.

Footnotes

 Presented in part at the XXVIIIth Meeting of the Club Jules Gonin, Reykjavik, Iceland, June 23, 2012, and at the 45th Annual Scientific Meeting of the Retina Society, Washington DC, October 6, 2012.

Footnotes

 Disclosure: H. Otsuka, None; H. Kawano, None; S. Sonoda, None; M. Nakamura, None; T. Sakamoto, None