Upregulation of Chemokine Expression in the Retinal Vasculature in Ischemia–Reperfusion Injury

Abstract

purpose. To evaluate chemokine expression at various retinal sites after ischemia–reperfusion injury, using reverse transcription–polymerase chain reaction (RT-PCR) analysis of selected tissue obtained by laser capture microdissection.

methods. Retinal ischemia was produced in Lewis rats by increasing intraocular pressure for 75 minutes. At 3, 6, 12, and 24 hours after reperfusion, RT-PCR was used to measure the levels of monocyte chemoattractant protein (MCP)-1, macrophage inflammatory protein (MIP)-1α, MIP-1β, interleukin (IL)-8, and interferon-γ–inducible 10-kDa protein (IP-10) mRNA expression in the ganglion cell layer (GCL), inner nuclear layer (INL), outer nuclear layer (ONL), and retinal vessels, after laser capture microdissection of these retinal layers. These chemokines were further localized by immunohistochemical methods, using antibodies specific to MCP-1 and MIP-1α. Leukocyte infiltration into the retina was detected with immunostaining for leukocyte common antigen.

results. Ischemia–reperfusion induced expression of MCP-1, MIP-1α, and MIP-1β mRNA in the retinal vessels 3 hours after reperfusion. Six hours after reperfusion, expression of these chemokines and IL-8 mRNA was seen in the GCL and INL. Twelve hours after reperfusion, IP-10 mRNA expression was seen in the GCL and INL. Immunoreactive MCP-1 and MIP-1α were detected in the GCL, INL, and the retinal vessels 24 hours after reperfusion. No chemokine mRNA expression or immunoreactivity was detected in the ONL at any time. Leukocyte infiltration was noted at 12 hours, increasing markedly 24 hours after reperfusion.

conclusions. Ischemia–reperfusion retinal injury results in generation of highly chemotactic agents, initially in the retinal vasculature, then in the other inner retinal layers. Such differential chemokine expression may play a role in leukocyte recruitment and selective leukocyte infiltration in the inner retina, leading to retinal damage primarily localized to the ganglion cells and other inner neuronal structures.

The retinal ischemia–reperfusion injury model has been studied extensively and is an ideal model for studying ischemia-induced neural damage. The model is reproducible and consistently quantifiable by histologic and functional criteria. In this model, leukocytes are thought to play a major role in inducing retinal damage by impeding blood flow and by producing superoxide radicals and inflammatory cytokines. ^{1 2 3 4 5} The trafficking of leukocytes into the ischemic retina requires generation of two broad classes of molecules: leukocyte chemotactic factors, including members of the chemokine family, and various cell adhesion molecules, especially the selectins and integrins. ^{6 7 8} Recently, the role of adhesion molecules in leukocyte recruitment has been intensively investigated in this model. ^{9 10} However, such studies have not been extended to detect expression of chemokines and their distribution in such retinal injury. 

Chemokines are a rapidly expanding family of chemotactic cytokines, of which there are more than 30 recognized members to date. These chemokines are subdivided into four groups, known as CXC, CC, C, and CX3C chemokines, depending on the positions of conserved cysteine residues in their amino acid sequences. ^{11 12 13 14} Moreover, these chemokines are also distinguished on the basis of their ability to cause migration of different leukocyte populations. Those CXC chemokines that contain the glutamic acid-leucine-arginine (ELR) peptide motif, including interleukin-8 (IL-8) and granulocyte chemotactic peptide-2, are potent mediators of neutrophil chemotaxis, whereas those chemokines that have no ELR peptide motif, including interferon-γ-inducible 10-kDa protein, (IP-10) and platelet factor-4 are not chemoattractants for neutrophils. The CC chemokines are thought to act mainly on monocytes. However, this subdivision of the chemokines based on their action on cells in vitro is becoming increasingly invalid, because more cell types, such as natural killer cells and B and T lymphocytes, are also attracted by these chemokines. Moreover, the functions of the chemokines overlap in their chemotactic activities on various cells. ¹¹  

In vitro and in vivo models suggest that chemokines play a role in a variety of inflammatory conditions. Similarly, chemokine expression in the ischemic retina could play an important role in the homing and directional migration of leukocytes into the inner retina, which is the primary site of ischemic injury. In the present study, we used the novel technique of laser capture microdissection, to examine chemokine expression in a retinal ischemia–reperfusion model, particularly in a time-dependent manner, in the retinal layers. This method allows detection of chemokine mRNA expression in the ganglion cell layer (GCL), inner nuclear layer (INL), outer nuclear layer (ONL), and retinal vessels. 

Materials and Methods

Experimental Retina Ischemia–Reperfusion Injury

A total of 30 female Lewis rats, each weighing 150 to 175 g (Charles River Laboratory, Wilmington, MA), were used for all the studies. All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

The animals were anesthetized with ketamine and xylazine were administered intramuscularly. Retinal ischemia was induced with a slight modification of the previously described technique. ¹⁵ The right anterior chamber was cannulated with a 25-gauge infusion needle connected by silastic tubing (Dow-Corning, Midland, MI) to a physiological saline reservoir. The intraocular pressure (IOP) of the cannulated eye was raised to 110 cm Hg for 75 minutes by elevating the saline reservoir. At the end of 75 minutes, the cannula was removed, and reperfusion of the retina vessels was confirmed by ophthalmoscopy. Animals were killed by injection of an overdose of sodium pentobarbital at various times after reperfusion. Six rats each were killed at 3, 6, 12, and 24 hours after reperfusion. Six animals without ischemia–reperfusion injury served as the control. 

Immunohistochemical Analysis for Infiltration of Leukocyte

To analyze the infiltration of leukocytes in retinal ischemia–reperfusion injury, we used an immunohistochemical method, with an avidin-biotin complex kit (Vectastain ABC; Vector Laboratories, Burlingame, CA) and used the antibody for leukocyte according to the manufacturer’s protocol. All steps were performed at room temperature unless otherwise stated. All incubation steps were conducted in a moist chamber. Briefly, cryosections were fixed in 4% paraformaldehyde at 4°C for 15 minutes, and incubated in 3% hydrogen peroxide and methanol at 4°C for 10 minutes to block endogenous peroxidase activity. After the reaction was blocked with 2% bovine serum albumin and 1% normal rabbit serum, the primary antibody, mouse anti-rat monoclonal antibody for leukocyte common antigen (LCA)/CD45 (1:100; Serotec, Oxford, UK) was applied to the sections at 4°C overnight. The sections were incubated with biotinylated rabbit anti-mouse IgG antibody (Dako, Carpinteria, CA) for 60 minutes, then with avidin coupled to peroxidase for 30 minutes. Amino-9-ethyl-carbazole was used as a chromogen, and the slides were counterstained with either methyl green or hematoxylin. Between each step, the sections were washed three times with phosphate-buffered saline. Three rats were used for each time point. Two areas of each retina (n = 3), superior and inferior, 1 mm from the optic disc were selected for the cell count. Cells displaying positive staining for LCA were counted as leukocytes. 

Laser Captured Microdissection

The retinal layers were dissected by laser capture microdissection (LCM) to analyze chemokine expression in the GCL, INL, ONL, and retinal vessels. Three rats were killed at each time point. The eyes were enucleated, quickly frozen, and cut into 10-μm-thick sections. These sections were fixed in 75% ethanol for 45 seconds and rinsed in RNase free water for 30 seconds. The sections were stained with hematoxylin for 60 seconds, rinsed in RNase-free water twice for 60 seconds, and stained with eosin for 15 seconds. Slides were immediately dehydrated in graded ethanol solutions (70%, 95%, 100%) for 5 seconds each, and finally cleared by two incubations in xylene for 2 minutes each. The sections were laser microdissected to obtain 300 cells each of GCL, INL, and ONL and 100 cells of retinal vessels using an LCM system (Robot-MicroBeam; PALM Microlaser Technologies, Bernried, Germany; Fig. 1 ). The transfer film cap containing LCM-derived cells was inserted into a 500-μL microcentrifuge tube containing denaturing solution (RNAeasy Mini Kit; Qiagen Inc., Valencia, CA). The quantity of RNA obtainable from 300 cells was below the detection limit of the most sensitive method (Ribogreen nucleic acid dye binding assay; Molecular Probes, Eugene, OR). Therefore, the amount of the extracted RNA was estimated by PCR amplification for the transcript of the ubiquitously expressed gene GAPDH. 

Semiquantitative Reverse Transcription–Polymerase Chain Reaction

RNA was isolated from microdissected cells using the RNAeasy Mini Kit. Relative reverse transcription–polymerase chain reaction (RT-PCR) analysis was performed using chemokine-specific primers. ^{16 17} Reverse transcription was performed with commercial reverse transcriptase (Sensiscript RT; Qiagen Inc.), according to the manufacturer’s protocol. PCR was performed with Taq DNA polymerase (AmpliTaq Gold; Applied Biosystems, Foster City, CA). The primers used in these experiments are shown in Table 1 . The 50-μL reaction mixture consisted of 2 μL cDNA, 2 μL of sense and antisense primer, 200 μM each of deoxynucleotide, 5 μL 10× PCR buffer, 4 μL 25 mM MgCl₂, and 1.25 U Taq DNA polymerase. Conditions for each chemokine amplification were as follows: preheating for 10 minutes at 94°C; denaturation for 1 minute at 94°C, and extension for 1minute at 72°C; annealing temperatures were 54°C for GAPDH, 56°C for IL-8, 58°C for IP-10, and 60°C for monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein (MIP)-1α and -1β. Each PCR product was separated by gel electrophoresis on 2% agarose containing ethidium bromide and analyzed under ultraviolet light. Three rats were used for each time point. The amount of RT-PCR products was corrected with GAPDH expression to determine the level of chemokine using NIH Image software (available by ftp at zippy.nimh.nih.gov/ or at http://rsb.info.nih.gov/nih-image; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). The widely different levels of expression of these genes makes it necessary to use different cycles to obtain optimum expression—for example, 37 cycles for MCP-1, MIP-1α, and MIP-1β; 39 cycles for IP-10 and IL-8; and 30 cycles for GAPDH. The level of expression was then compared within the time course (between 0 and 24 hours) for one particular gene, but was not compared between genes. In a particular series of experiments (between 0 and 24 hours), the highest expression is then arbitrarily normalized to 100%. 

Immunohistochemical Analysis for Chemokines

To detect the presence of chemokines in ischemia–reperfusion retina, immunohistochemical studies were performed with an avidin-biotin complex kit (Vectastain ABC; Vector Laboratories), with antibodies for MCP-1 and MIP-1α, according to the manufacturer’s protocol. The procedure used for this chemokine staining was the same as that described above for LCA staining. The primary antibodies used were rabbit anti-rat MCP-1 (1:100; Peprotech, Rocky Hill, NJ), rabbit anti-rat MIP-1α (1:100; Chemicon, Temecula, CA), rabbit anti-rat MIP-1β (Biosource, Camarillo, CA), and rabbit anti-rat IP-10 (Cell Science, Norwood, MA). No positive staining was obtained using mouse anti-human IL-8 (Chemicon) for IL-8 staining. The secondary antibodies were biotinylated goat anti-rabbit IgG (Vector Laboratories) and biotinylated goat anti-mouse IgG (Dako). 

Results

Leukocyte Infiltration in Retinal Ischemia–Reperfusion

There were no LCA-positive cells in the retina at 3 and 6 hours after reperfusion. Twelve hours after reperfusion, a few positive cells could be detected in the INL (Fig. 2B) . At 24 hours after reperfusion, several positive cells could be detected in the nerve fiber layer and GCL (Fig. 2C) . A few positive cells were also present in the inner plexiform layer and INL. Infiltration of leukocytes into the retina after reperfusion began to occur at 12 hours after reperfusion and was markedly increased at 24 hours after reperfusion (Fig. 3) . 

Time Course of Chemokine mRNA Expression in Retinal Ischemia–Reperfusion

Relative levels of mRNA for MCP-1, MIP-1α, MIP-1β, IP-10, and IL-8 were measured by semiquantitative RT-PCR at various time points after reperfusion in the GCL (Fig. 4) , INL (Fig. 5) , ONL, and retinal vessels (Fig. 6) . The microdissection samples consisted of cells of each layer and other adjacent structures (Fig. 1) . 

In the GCL, the chemokines were undetectable in the control retinas and in the experimental retina 3 hours after reperfusion. MCP-1 and MIP-1α expression was noted at 6 hours, declining to normal levels by 24 hours after reperfusion. MIP-1β and IL-8 expression was detected at 6 hours, declining to normal levels by 12 hours after reperfusion. IP-10 expression increased markedly at 12 hours, declining to normal levels at 24 hours after reperfusion. In the INL, the expression pattern of all chemokine mRNA was similar to the expression noted in the GCL. In the retinal vessels, MCP-1 expression was detected at 3 hours, declining to normal levels by 12 hours after reperfusion. MIP-1α and MIP-1β expression appeared at 3 hours, declining to normal levels at 6 hours after reperfusion. IL-8 and IP-10 were not detected at any time points in the retinal vessels. No chemokines were detectable at any time in the ONL. 

Localization of MCP-1, MIP-1α, MIP-1β, and IP-10 Protein in Retinal Ischemia–Reperfusion

Immunoreactivity for MCP-1 and MIP-1α were not evident in normal retina. Immunoreactivity for MCP-1 (Fig. 7A) and MIP-1α (Fig. 7B) was observed in the GCL, INL, and retinal vessels at 24 hours after reperfusion. Relatively weak staining for MIP-1β (Fig. 8A) was present in the RPE and photoreceptor inner segment. Intensely positive stain was present for IP-10, especially in the GCL (Fig 8B) 24 hours after reperfusion. No staining was revealed for control, unaffected retinas for all chemokines (Fig. 8C) . At 3, 6, and 12 hours after reperfusion, no immunoreactivity was detected for these chemokines. 

Discussion

Leukocytes are thought to play a central role in postischemia neural damage by generating superoxide radicals and inflammatory cytokines at the site of infiltration. ^{1 2 3 4 5} Infiltration of such cells is regulated by a multistep process that includes adhesion molecule expression by endothelial cells and generation of chemoattractants. ^{18 19} However, chemokines play crucial roles in the transendothelial migration of leukocytes. ²⁰ The present study revealed chemokine expression initially in the retinal vasculature in a retinal ischemia–reperfusion model. Subsequently, expression of several chemokines was noted in the GCL and the INL before the arrival of leukocytes into the inner retina. The leukocyte chemoattractant chemokines expressed included MCP-1, MIP-1α, MIP-1β, IL-8, and IP-10. These findings suggest that initial expression of chemokines in the retinal vasculature could play a role in the recruitment of leukocytes along with transmigration of these cells to the inner retina. 

However, these recruited leukocytes are not the major source to secrete these chemokines. The reasons to arrive to this conclusion are as follows: First, the major increase of leukocytes did not start until after 12 hours, whereas the gene expression for these chemokines was maximum at 6 hours in the GCL and in INL (Figs. 4 and 5) . Second, although immunoreactivity for these chemokines was present at the 24-hour time point (possibly representing the delayed translation of genes present in the 6- and 12-hour periods), the localization was concentrated in GCL and therefore, did not coincide with the location of leukocytes that spread to several retinal layers. 

In previous studies, Szabo et al. ²¹ demonstrated that retinal leukocyte infiltration increases after 90 minutes of ischemia and 24 hours of reperfusion. Sixty minutes of ischemia and 24 hours of reperfusion did not lead to infiltration. Hangai et al. ⁵ reported that 2 hours of ischemia led to neutrophil migration in the retina, but no leukocyte infiltration was observed before 3 hours of reperfusion. Recently, acridine orange digital fluorography allowed visualization of leukocytes and evaluation of leukocyte dynamics in this model. ^{9 10 22} By this technique, Tsujikawa et al. ²³ demonstrated that leukocyte rolling and accumulation were observed at 4 hours after reperfusion, peaking at 12 and 24 hours after reperfusion. Our present study using immunohistochemical analysis revealed infiltration of leukocytes at 12 hours, which was increased at 24 hours after reperfusion, similar to that reported by Tsujikawa et al. In our study, chemokine mRNA expression in the GCL and INL appeared before 12 hours after reperfusion, and MCP-1 and MIP-1α protein were observed at 24 hours. Such data suggest that the leukocyte infiltration in this model could be mediated by chemokine expression in the retinal vasculature and in the inner retinal layers. 

As chemoattractants for leukocytes, the chemokines play an important role in the directional migration of leukocytes through tissues. Recent studies have also revealed the importance of chemokines in enhancing the avidity of leukocyte integrins for locally expressed adhesion molecules. ²⁴ Chemokines can induce the rapid conversion of lymphocyte or monocyte rolling to firm adhesion through integrin activation under physiological conditions. ^{25 26 27} Such studies indicate that chemokine expression contributes to the initial step of leukocyte attachment to vascular endothelium, which is then followed by transmigration into the surrounding tissues. In the present study, MCP-1, MIP-1α, and MIP-1β expression occurred first in the retinal vessels, followed by expression of these and IL-8 and IP-10 in the GCL and INL. These findings suggest that chemokine expression in the retinal vessels may contribute to the adhesion of leukocytes to vascular endothelium with subsequent transmigration to the inner retinal layers. For leukocytes to migrate to areas of inflammation, there must be a soluble chemoattractant gradient in which cells move from a low concentration to a high concentration. After leukocyte binding to endothelium through complex interaction of adhesion molecules and chemokines and their receptor, migration of leukocyte to the GCL and INL may occur through leukocyte chemotaxis toward the chemokine gradient present in GCL and INL. 

LCM is a rapid method of obtaining pure cell populations from specific microscopic regions of tissue sections. With careful manipulation, LCM-derived cells can be used to isolate sufficient quantities of intact total RNA, DNA, and protein for specific analyses, including RT-PCR, cDNA microarray analysis, differential gene expression profiling, loss of heterozygosity, and two-dimensional PAGE analysis of proteins. ^{28 29 30 31} We selectively collected cells of each retinal layer and the retinal vessels using LCM. Use of such technique, as shown in the present study, can enable detection of various inflammatory molecules and expression of genes involved in retinal ischemia. 

Furthermore, the present study clearly documented the role of vascular endothelium in the initial step in recruitment of leukocytes and subsequent retinal damage. It must be indicated that the Lewis rats are the only rat strain used for this study, and the observations made may therefore be limited to Lewis and similar rat strains. Therapeutic agents directed specifically against such chemokine expression may be found useful in preventing ischemia-mediated retinal damage. 

 Supported in part by Grants EY03040 and EY13253 from the National Eye Institute and by an unrestricted grant from Research to Present Blindness.

 Submitted for publication December 19, 2002; revised April 11, 2003; accepted April 30, 2003.

 Disclosure: N. Jo, None; G.-S. Wu, None; N.A. Rao, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Figure 1.

 

Retinal sections showing before (A) and after (B) laser capture microdissection (hematoxylin and eosin staining). Three hundred cells each from the ganglion cell layer (GCL), inner nuclear layer (INL), and outer nuclear layer (ONL), and 100 cells from the retinal vessels per eye were pooled in a single tube. G, GCL; I, INL; O, ONL; V, retinal vessel.

Table 1.

 

 

Primer Sequences for RT-PCR

Table 1.

 

Primer Sequences for RT-PCR

Chemokine	Primer
IL-8	5′-CTCCAGCCACACTCCAACAGA-3′
	5′-CACCCTAACACAAAACAGAT-3′
IP-10	5′-TTCCTGCAAGTCTATCCTGTCCGC-3′
	5′-TTTGCCATCTCACCTGGACTGC-3′
MCP-1	5′-CCTGTTGTTCACAGTTGCTGCC-3′
	5′-TCTACAGAAGTGCTTGAGGTGGTTG-3′
MIP-1α	5′-CCCTTGCTGTTCTTCTCTGCAC-3′
	5′-GCATTCAGTTCCAGCTCAGTGATG-3′
MIP-1β	5′-TCTGCGATTCAGTGCTGTCAGC-3′
	5′-GATTTGCCTGCCTTTTTTGGTC-3′
GAPDH	5′-TGATGACATCAAGAAGGTGGTGAAG-3′
	5′-TCCTTGGAGGCCATGTAGGCCAT-3′

Figure 2.

 

Immunohistochemical staining for leukocyte common antigen in control retina (A), at 12 hours (B), and 24 hours (C) after reperfusion in the ischemic retina. Original magnification, ×40.

Figure 3.

 

Number of LCA-positive cells counted per square millimeter of inner retina at 3, 6, 12, and 24 hours after reperfusion. Data are the mean ± SD of counts in three retinas examined at each time point. Leukocyte infiltration into the inner retina was noted at 12 hours and had increased markedly at 24 hours after reperfusion.

Figure 4.

 

(A) The electrophoresis pattern of PCR products for MCP-1, MIP-1α, MIP-1β, IL-8, and IP-10 in the ganglion cell layer (GCL). (B) Time course for expression of various chemokine mRNAs in the GCL after reperfusion. The relative level of mRNA expression was quantified and corrected for the levels of GAPDH mRNA expression. Because the chemokines are run with different cycles—37 cycles for MCP-1, MIP-1α, and MIP-1β; 39 cycles for IP-10 and IL-8; and 30 cycles for GAPDH—the maximum value in the time course for any particular gene is normalized to 100%. M, molecular weight marker; N, normal retina.

Figure 5.

 

(A) The electrophoresis pattern of PCR products for MCP-1, MIP-1α, MIP-1β, IL-8, and IP-10 in the INL. (B) Time course for expression of various chemokine mRNAs in the INL after reperfusion. The relative level of mRNA expression was quantified and corrected for the levels of GAPDH mRNA expression. The cycles run for the chemokines were the same as those in Figure 4 , and maximum intensity in the time course for each chemokine was arbitrarily set at 100%. M, molecular weight marker; N, normal retina.

Figure 6.

 

(A) The electrophoresis pattern of PCR products for MCP-1, MIP-1α, and MIP-1β in retinal vessels. (B) Time course for expression of various chemokine mRNAs in retinal vessels after reperfusion. The relative level of mRNA expression was quantified and corrected for the levels of GAPDH mRNA expression. The cycles run for the chemokines were the same as those in Figure 4 , and maximum intensity in the time course for each chemokine was arbitrarily set at 100%. IL-8 and IP-10 were not detected at any time point in the vessels. M, molecular weight marker; N, normal retina.

Figure 7.

 

Immunohistochemical staining for MCP-1 (A) and MIP-1α (B) at 24 hours after reperfusion in the ischemic retina. Note positive immunoreactivity in the GCL, INL, and retinal vessels. Original magnification, ×40.

Figure 8.

 

Immunohistochemical staining for MIP-1β (A), IP-10 (B), and unaffected retina staining for MCP-1 (C). Note the low-intensity positive staining in the RPE and inner segment for MIP-1β, the intense positive staining in the GCL and INL for IP-10, and no staining in the control unaffected retina. Original magnification, ×40.

 

The authors thank Tim T. Lam, PhD, for advice and assistance in the development of the animal model.