Chemokines in aqueous humour before and after intravitreal triamcinolone acetonide in eyes with macular oedema associated with branch retinal vein occlusion

Authors


Hiroshi Kunikata, MD
Department of Ophthalmology
and Visual Science
Tohoku University Graduate
School of Medicine
1-1 Seiryo-machi, Aoba-ku
Sendai 980-8574
Japan
Tel: + 81 22 717 7294
Fax: + 81 22 717 7298
Email: kunikata@oph.med.tohoku.ac.jp

Abstract.

Purpose:  To determine the aqueous humour levels of chemokines before and after an intravitreal injection of triamcinolone acetonide (IVTA) in eyes with macular oedema associated with a branch retinal vein occlusion (ME-BRVO).

Design:  Single-centre, prospective, consecutive interventional case series.

Participants:  Seventeen eyes of 17 consecutive patients with ME-BRVO who underwent IVTA were studied. Seven eyes without retinal vascular disease served as control.

Intervention:  All patients with ME-BRVO underwent IVTA.

Main outcome measures:  The optical coherence tomographically determined foveal thickness (FT) and the aqueous humour levels of inflammatory chemokines of the C-C subfamily, including eotaxin, monocyte chemotactic protein-1 (MCP-1), macrophage inflammatory protein-1α (MIP-1α), β (MIP-1β), and RANTES was determined before the IVTA (baseline) and at 1 week after the IVTA.

Results:  At the baseline, only MCP-1 and MIP-1β were detected in the aqueous, and MIP-1β was significantly higher in eyes with a ME-BRVO than in controls (p = 0.004). The level of both of these chemokines was not correlated with the FT (p = 0.654 and p = 0.608, respectively). One week after IVTA, the FT was significantly decreased (p < 0.001), and the levels of MCP-1 and MIP-1β were also significantly reduced (p < 0.001 and p = 0.044, respectively). The decrease in the FT was correlated with the decrease in only MIP-1β (r = 0.58, p = 0.020).

Conclusions:  Alterations of the aqueous level of MIP-1β reflect the improvement of the macular oedema after IVTA in eyes with ME-BRVO. This indicates that the steroid-dependent ME-BRVO was closely related with the level of MIP-1β.

Introduction

Macular oedema because of a branch retinal vein occlusion (ME-BRVO) is a relatively common retinal vascular condition that causes severe visual disturbances. Although different approaches, e.g., pars plana vitrectomy, have been used to treat the oedema (Opremcak & Bruce 1999; Tachi et al. 1999; Noma et al. 2009), macular oedema is important because it is difficult to treat satisfactorily.

An intravitreal injection of triamcinolone acetonide (IVTA) has been reported to reduce the ME associated with the BRVO (Chen et al. 2004; Inoue et al. 2004; Jonas et al. 2004; Lee & Yang 2004). However, an elevation of the intraocular pressure and a recurrence of the macular oedema have been reported as IVTA-associated side-effects. In addition, it has been shown that the effects of IVTA are only transient in resolving the ME-BRVO (Cakir et al. 2008; Kogure et al. 2008; Vasconcelos-Santos et al. 2008).

Although the pathogenesis of ME-BRVO has not been determined, an increase in the rigidity of a crossing artery resulting from atherosclerotic disease has been suggested to cause a compression of the underlying vein, resulting in turbulent blood flow, endothelial damage, and thrombus formation (Fekrat 1999). Atherosclerosis is a chronic low-grade inflammatory disease with a distinct pro-inflammatory cytokine pattern (Huber et al. 1999; Whitman et al. 2000; Pinderski et al. 2002). Also, based on the breakdown of the blood:retinal barrier by damage to the tight junctions of retinal capillary endothelial cells (Silva et al. 1995), fluid flow from the occluded vessels into the tissue can lead to ME. By these mechanisms, inflammation and vascular dysfunctions interact and stimulate each other and cause retinal ischemia which induces the expression of vascular endothelial growth factor (VEGF) (Aiello et al. 1995).

Several studies have shown that the vitreous levels of several cytokines, including VEGF and interleukin-6 (IL-6) were increased in eyes with BRVO-ME. The levels of these cytokines are associated with the severity and the prognosis of the BRVO-ME (Noma et al. 2005, 2006; Shimura et al. 2008). There may also be some interactions between the intraocular cells and various cytokines. For example, Tojo et al. (2009) reported that porcine vitreous-derived cells are stimulated by IL-1α, IL-1β, and TNF-α to induce VEGF and IL-6 (). However, it is not easy for ophthalmologists to evaluate the vitreous levels of various cytokines in a routine examination. On the other hand, the aqueous humour is relatively easy to aspirate and examine even in an outpatient clinic, and the aqueous humour levels of VEGF and IL-6 have been reported to be correlated with the vitreous levels and with the severity of the ME-BRVO (Noma et al. 2008).

Chemokines are multifunctional mediators that can recruit leucocytes to the sites of inflammation and promote inflammation (Struyf et al. 2003; Abu El-Asrar et al. 2006). The vitreous levels of some chemokines, including monocyte chemotactic protein-1 (MCP-1), macrophage inflammatory protein (MIP)-1α, and MIP-1β, have been reported to be affected by different retinal diseases including ischaemic retinopathy, diabetic retinopathy, proliferative retinopathy, and retinal vein occlusions (Yoshida et al. 2003; Zeng et al. 2005; Abu El-Asrar et al. 2006; Funk et al. 2009). A recent study demonstrated the importance of IL-6 and MCP-1 in the pathogenesis of clinically significant macular oedema (Roh et al. 2009).

In this prospective study, we focused on one of the chemokine ligand subgroups of CCL1 to CCL5, including CCL1/eotaxin, MCP-1 (CCL2/MCP-1), MIP-1α (CCL3/MIP-1α), MIP-1β (CCL4/MIP-1β), and CCL5/RANTES. The aqueous levels of these chemokines in eyes with a ME-BRVO were measured before and after IVTA. In addition, the relationships between the alteration of aqueous level of each chemokine and the IVTA-induced reduction of macular oedema in BRVO were investigated.

Patients and Methods

Participants

Seventeen eyes of 17 patients (eight men and nine women) with decreased visual acuity because of a ME-BRVO who were treated with IVTA between February 2007 and August 2007 were studied. The ages of the patients ranged from 49 to 81 with a mean of 69.2 ± 9.15 (±standard deviation). The duration of the BRVO ranged from 30 to 360 days with a mean of 166 ± 102 days. The inclusion criteria were clinically detectable macular oedema because of BRVO, foveal thickness (FT) >250 μm, and decimal visual acuity <0.8. The exclusion criteria were prior: vitreous surgery, IVTA, intravitreal antivascular endothelial growth factor (VEGF), ocular inflammation, and vitreoretinal or optic nerve diseases. All patients who were scheduled to undergo pars plana vitrectomy had an IVTA injection 1 week before the vitrectomy. The levels of chemokines in the anterior chamber were measured in samples of aqueous humour collected just before the IVTA, i.e. 1 week before the vitrectomy. A second aqueous humour sample was collected 1 week after the IVTA just before the vitrectomy. Four patients with an epiretinal membrane (ERM) and three with a macular hole (MH) who underwent vitrectomy were studied as controls. Aqueous humour samples were collected just before the vitrectomy. All patients were followed in the ophthalmology clinic of the Tohoku University Hospital, Sendai, Japan.

The indications for the IVTA were visual disturbances with clinically detectable ME because of a BRVO of more than 1-month duration. All IVTA injections were performed at the Tohoku University Hospita.

The collection of aqueous humour samples was approved by Institutional Review Board (IRB) of Tohoku University Graduate School of Medicine. The procedures conformed to the tenets of the Declaration of Helsinki, and an informed consent was obtained from all patients after the purpose and procedures of the IVTA were explained.

Intravitreal injection of triamcinolone acetonide

A topical antibiotic, 0.3% gatifloxacin hydrate ophthalmic solution (Gatiflo; Senju Pharmaceutical Co, Ltd, Osaka, Japan), was applied four times/day in both eyes for at least 3 days before the IVTA. The intravitreal injection was performed after topical 0.4% oxybuprocaine (Benoxil; Santen Pharmaceutical Co, Ltd, Osaka, Japan) and povidone iodine diluted eightfold. The triamcinolone acetonide (TA; Kenacort-A, Bristol-Meyers Squibb, Tokyo, Japan) diluent was replaced with balanced salt solution after Millipore filtration, and the volume was adjusted so 0.1 ml containing 4 mg TA. The TA was injected into the vitreous cavity without a prior parathentesis. For all eyes, a 27-gauge needle on a 1-ml syringe was inserted through the sclera 3.5 mm posterior to the limbus.

Measurement of MCP-1, MIP-1α, and MIP-1β levels

Undiluted aqueous samples were collected just before the IVTA, and another sample was collected 1 week after the IVTA just before the vitrectomy. The samples of aqueous humour (about 200 μl) were collected in sterile tube and were rapidly frozen to −30°C. Samples of aqueous humour from eyes with an ERM or MH were collected as controls at the time of vitreous surgery and were rapidly frozen in the same way.

The concentrations of eotaxin, MCP-1, MIP-1α, MIP-1β, and RANTES were measured by Luminex® 100 (Luminex, Austin, TX, USA) at Kyushu University in a masked way. The minimum concentration of these factors in the aqueous humour that could be detected by the Luminex system was 50 pg/ml for eotaxin, 0.1 ng/ml for MCP-1, 100 pg/ml for MIP-1α, 100 pg/ml for MIP-1β, and 150 pg/ml for RANTES.

Measurement of FT

The FT was determined by optical coherence tomography (OCT3000; Zeiss–Humphrey Ophthalmic Systems, Dublin, CA, USA) just before and 1 week after the IVTA. The retinal thickness of the central fovea was defined as the distance between the inner limiting membrane and the retinal pigment epithelium, and was automatically calculated by the software of the OCT3000 computer. A macular thickness map was made with the OCT retinal mapping program from the six radial scans that intersected at the fovea. This program calculates the mean retinal thickness in nine regions, the 1000-μm central area, and the four quadrants in an inner and outer ring. The foveal thickness was defined as the value of the 1000 μm central area.

Statistical analyses

The data are presented as the mean ± standard deviations. The significance of the differences between the pre- and post-IVTA data was assessed by the Wilcoxon signed-ranks test. The significance of the differences in the concentration of the chemokines between eyes with a ME-BRVO and control subjects was assessed by the Mann Whitney’s U-test. The Spearman’s coefficient of correlation by rank was calculated to determine the correlation between aqueous humour level of chemokines and the clinical parameters. A p-value of <0.05 was considered to be statistically significant.

Results

Alteration of FT after IVTA

Before the IVTA, the mean FT was 461 ± 168 μm in the patients with a ME-BRVO. All but one case showed a regression of the ME after the IVTA, and the FT was significantly decreased to 272 ± 84.5 μm (Wilcoxon signed-ranks test, p = <0.001). The degree of decrease of the FT was strongly correlated with the FT at baseline (= 0.91, p = <0.001; Spearmans correlation; Fig. 1).

Figure 1.

 Relationship between the decrease in the foveal thickness (FT) and FT at the baseline, i.e. before intravitreal triamcinolone acetonide. The decrease in FT was strongly correlated with the FT at baseline (Spearman’s coefficient of correlation; r = 0.91, p = 0.001).

Aqueous humour levels of eotaxin, MCP-1, MIP-1α, MIP-1β, and RANTES at baseline

The aqueous humour level of MCP-1 in patients with a BRVO-ME was 1.30 ± 1.04 ng/ml (range 0.384–4.50ng/ml) which was higher than that in the controls (0.785 ± 0.286 ng/ml; range 0.474–1.31 ng/ml), but the difference was not significant (p = 0.216; Mann Whitney’s U-test; Table 1). On the other hand, the mean aqueous humour level of MIP-1β in eyes with a ME-BRVO was 169 ± 26.5 pg/ml (range 134–278 pg/ml), which was significantly higher than the 124 ±26.1 pg/ml (range 100–162 pg/ml) in the controls (p = 0.004; Mann Whitney’s U-test). The level of MCP-1 was not significantly correlated with that of MIP-1β (p = 0.352; Spearmans correlation). In addition, both MCP-1 and MIP-1β were not significantly correlated with the FT at the baseline (p = 0.654, p = 0.608, respectively). The levels of both MCP-1 and MIP-1β at baseline were not significantly correlated with the duration of the BRVO (p = 0.101, p = 0.737, respectively). Eotaxin, MIP-1α, and RANTES were not detected in any eyes, including the eyes with an ERM and MH.

Table 1.   Foveal thickness (FT) and aqueous humour levels of chemokines before and after intravitreal infection of triamcinolone acetonide (IVTA).
 ControlBRVO-MEp-Value
Baseline1 week after IVTAB versus I
  1. B = BRVO-ME at Baseline, I = BRVO-ME 1 week after IVTA, MCP-1 = monocyte chemotactic protein-1, MIP = macrophage inflammatory protein.

  2. * Wilcoxon signed-ranks test.

FT (μm)461 ± 168272 ± 84.5<0.001*
Eotaxin (pg/ml)NDNDND
MCP-1 (ng/ml)0.785 ± 0.2861.30 ± 1.040.671 ± 0.508<0.001*
MIP-1α (pg/ml)NDNDND
MIP-1β (pg/ml)124 ± 26.1169 ± 26.5156 ± 30.90.044*
RANTES (pg/ml)NDNDND

Alterations of levels of chemokines in aqueous humour after IVTA

The levels of MCP-1 and MIP-1β in the aqueous humour were significantly decreased after IVTA (p = <0.001 and p = 0.044, respectively; Table 1), and the aqueous humour level of MCP-1 after IVTA was significantly correlated with the level at baseline (= 0.68, p = 0.006; Spearmans correlation; Fig. 2). The aqueous level of MIP-1β after IVTA was also significantly correlated with that at baseline (= 0.58, p = 0.021; Spearmans correlation; Fig. 3).

Figure 2.

 Alterations of aqueous humour levels of MCP-1 after intravitreal injection of triamcinolone acetonide (IVTA). The aqueous humour level of MCP-1 after IVTA was correlated with that at baseline level (Spearman’s coefficient of correlation, r = 0.68, p = 0.006).

Figure 3.

 Alterations of the aqueous humour levels of MIP-1β after intravitreal injection of triamcinolone acetonide (IVTA). Aqueous level of MIP-1β after IVTA was also correlated with that at baseline level. (Spearman’s coefficient of correlation, r = 0.58, p = 0.021).

Relationship between FT and alterations of levels of chemokines in aqueous humour

The decrease of MCP-1 and MIP-1β was determined by subtracting the levels after the IVTA from that at baseline. The improvement of FT was determined by subtracting the FT after IVTA from the FT at baseline. The decrease in MIP-1β was correlated with the FT at baseline and the improvement of FT (r = 0.60, p = 0.016 and = 0.58, p = 0.020, respectively; Spearmans correlation; Fig. 4). On the other hand, the decrease in MCP-1 was not significantly correlated with the FT at baseline and the improvement of FT (p = 0.682 and p = 0.852, respectively).

Figure 4.

 Relationship between the decrease in MIP-1β and decrease in the foveal thickness (FT). Decrease in MIP-1β was significantly correlated with the decrease in the FT (Spearman’s coefficient of correlation, r = 0.58, p = 0.020).

Discussion

Our results showed that IVTA reduced the FT and was effective in resolving the ME-BRVO. In addition, our results showed that the aqueous humour levels of MIP-1β in patients with BRVO-ME were significantly higher than that of the control eyes even before IVTA. One week after IVTA, the level of MIP-1β was significantly decreased.

The aqueous humour level of MCP-1 was higher than that of the control eyes but the level was not statistically significant. Although the level of MCP-1 in the aqueous humour of patients with central retinal vein occlusion (CRVO) was significantly higher than that of controls, the MCP-1 level in eyes with BRVO was not significantly higher (Funk et al. 2009). Thus, our findings on BRVO were similar to their findings.

The new finding of our study was that the aqueous humour levels of MCP-1 and MIP-1β were significantly reduced in the eyes with ME-BRVO after IVTA. In addition, the decrease in MIP-1β in the aqueous humour was correlated with the FT at baseline and the improvement of FT after IVTA. Because we measured the chemokines in the aqueous humour sample collected 1 week before vitrectomy and a second sample collected just before the vitrectomy, the vitrectomy could not have influenced our results. So, our findings must be because of the IVTA.

There were three patients who had increased MIP-1β levels of more than 10 pg/ml with decreased FT of more than 100 μm. Although the reason for this relationship was not determined, the decrease in the FT was <200 μm after IVTA in these three patients. So, if the decrease of FT is <200 μm after the IVTA, an increase of the MIP-1β level would be expected and possible. This suggests that MIP-1β-independent factors could be related with improvement of ME-BRVO after IVTA. Thus, an improvement of the FT after IVTA might be caused by MIP-1β-dependent and MIP-1β-independent pathways. To determine the exact mechanism, we will need an antibody for MIP-1β.

Although the exact mechanism that leads to the reduction of the ME-BRVO by IVTA has not been determined, the results of recent studies have presented several possible mechanisms. For example, BRVO in the rat retina results in a rapid and transient increase in the expression of VEGF (Rehak et al. 2009). Also, the down-regulation of inward rectifying K+ (Kir) 4.1 channels and aquaporin-1 and -4, the mislocation of Kir4.1 protein, and the osmotic swelling of the Mueller cells have also been observed in rats with BRVO (Rehak et al. 2009). These changes can contribute to the development of ME. Intravitreal injection of triamcinolone acetonide downregulates VEGF, which may prevent a decrease in occludin and also inhibits an increase in glial fibrillary acidic protein expression in Mueller cells (McAllister et al. 2009). Intravitreal injection of triamcinolone acetonide was also shown to prevent the osmotic swelling of Mueller cells by opening of K+ and Cl in the Mueller cell membrane (Uckermann et al. 2005). These events may contribute to a reduction in the blood:retinal barrier breakdown that occurs in BRVO and promote the resolution of the retinal oedema. The changes in the aqueous levels of VEGF and IL-6 after IVTA in eyes with a BRVO have already been investigated, and the IL-6 independent VEGF secretion may have contributed to the persistence of ME-BRVO after IVTA (Park & Ahn 2008). In our patients, MIP-1β was found to play key roles in the reduction of the ME after IVTA, thus a MIP-1β-dependent mechanism probably contributed to the persistent ME.

MIP-1β is a member of the C-C subfamily of chemokines and has pro-inflammatory activity. It is also a potent pyrogen activator and is implicated in the inflammatory reaction of the brain to ischaemia (Gourmala et al. 1999). Mononuclear cell chemoattractants such as the C-C chemokines are known to be expressed in ischaemic areas, and may regulate the recruitment of monocytes and lymphocytes (Frangogiannis 2004). Thus, it may not be surprising that MIP-1β was expressed in eyes with ME-BRVO because the BRVO leads to retinal ischaemia and inflammation. The other C-C chemokines, including eotaxin, MIP-1α, and RANTES, were not detected in the aqueous humour, but these chemokines must be linked to MIP-1β. A possible explanation for their absence is that the levels of these cytokines were too low to be detected in the aqueous humour, and may be present in the vitreous of eyes with ME-BRVO.

It is interesting that the aqueous level of MCP-1 was not significantly correlated with the reduction of the ME-BRVO in our study, even though an earlier study showed that the vitreous levels of MCP-1 was significantly correlated with the central retinal thickness in patients with diabetic macular oedema (Funatsu et al. 2009). Other studies showed that the level of MCP-1 in the aqueous humour was significantly elevated in eyes with a CRVO and might be related to the pathogenesis of macular oedema induced by CRVO (Funk et al. 2009). So, MCP-1 might not be directly related with severity of ME-BRVO but with diabetic mellitus and ME-CRVO. Therefore, MCP-1 is probably expressed more than MIP-1β in eyes with more severe ischaemia. The increased expression and release of MCP-1 was also reported to be an important cause of photoreceptor degeneration and apoptosis following a retinal detachment (Nakazawa et al. 2006, 2007a,b). Another possibility is that the expression of MCP-1 is independent of the TA. In fact, IVTA did not always fully reduce the ME, and the pathogenesis of ME-BRVO may include an IVTA non-sensitive mechanism which is related to the MCP-1 expression.

The interval from the onset of the ME to the IVTA was also correlated with the baseline visual acuity in patients with CRVO but not with BRVO (Scott et al. 2009). The vitreous level of VEGF was strongly correlated with duration of BRVO (Shimura et al. 2008), and the aqueous level of VEGF was also correlated with the vitreous levels and also to the severity of the ME-BRVO (Noma et al. 2008). The visual acuity and aqueous humour levels of both MCP-1 and MIP-1β at the baseline were not significantly correlated with the duration of the BRVO in this study (p = 0.316, p = 0.101 and p = 0.737, respectively). The aqueous levels of both MCP-1 and MIP-1β were also not correlated with the FT at baseline, which indicated that both chemokines were not directly related to the mechanism of the acute and chronic ME.

Taken together, our most important finding was that a ME-BRVO that is TA-sensitive can be evaluated by the aqueous level of MIP-1β. Although further investigations are required, focus should be placed on MIP-1β in the future to determine the mechanism of the IVTA-induced reduction of the ME.

Acknowledgments

This article was partially presented at the Annual Meeting of the Japanese Society of Clinical Ophthalmology, Kyoto, October 2007. Supported in part by research grants from the Ministry of Education, Culture, Sports, Science and Technology, Tokyo, Japan.

Ancillary