Astrocyte and microglial activation in the lateral geniculate nucleus and visual cortex of glaucomatous and optic nerve transected primates

Animal surgical procedures

Two rhesus and two cynomolgus monkeys underwent laser ablation of the trabecular meshwork [10,50] in one eye, which resulted in ipsilateral elevated IOP for periods of two to eight weeks (Table 1). A standard clinical argon laser and slit lamp delivery system was used to produce a series of 75 to 250 focal lesions to the trabecular meshwork. The laser was set to deliver a 50 μm spot diameter at 1–1.5 W with a 0.5 s duration. Two additional animals had spontaneous unilateral elevation of IOP. These animals had developed unilateral secondary glaucoma due to persistent inflammation following intracameral injection of different experimental drugs. IOP was monitored every three to ten days under intramuscular (IM) injection of 10 mg/kg ketamine anesthesia (3–5 mg/kg supplement as needed) using a minified Goldmann applanation tonometer [51] (Haag-Streit, Koniz, Switzerland). These measurements were occasionally backed up by measurements with a handheld applanation tonometer (Tono-pen XL; Mentor O & O, Norwell, MA). Tonopen measurements were converted to actual mmHg based on a standard calibration curve [52]. IOP was measured with the monkey lying prone in a head holder. If IOP was not consistently above 30–35 mmHg, additional laser treatments were performed until stable ocular hypertension was achieved. IOP was checked every three to ten days thereafter to assure stability. Additional laser treatment or IOP lowering therapy was applied as needed to maintain IOP at the desired level. If the IOP was higher than the protocol target range or if there was any sign of discomfort due to elevated IOP, the monkeys were treated topically once or twice daily with a single drop of one or more of the following until the desired IOP was achieved: 0.5% timolol maleate in gel-forming vehicle (Timoptic-XE; Merck & Co., Whitehouse Station, NJ), 0.2% brimonidine tartrate (Alphagan; Allergan, Irvine, CA), 2% dorzolamide hydrochloride (Trusopt; Merck & Co.), and 2 μg prostaglandin F2 alpha-isopropyl ester (donated by Pharmacia Corp, Peapack, NJ). If necessary, 5 mg/kg IM acetazolamide sodium (Ben Venue Laboratories, Bedford, OH) was given once or twice daily. The opposite eye served as a normal control eye.

Quantification of vision loss experienced by animals with unilateral ocular hypertension was calculated using an “integral” value, defined as the number of days of IOP measurements >30 mmHg multiplied by the mean IOP value for that period of time [53]. Vision loss by ONT was quantified by the total number of days after ONT surgery and before euthanasia. A summary of these values is shown in Table 1.

Six additional rhesus monkeys underwent unilateral optic nerve transection [54] (Table 1). Animals were given an initial IM injection of 10–15 mg/kg ketamine, which was followed by intubation, then general anesthesia under 1%–3% isoflurane gas. After surgery, monkeys were treated with systemic benzathine and 30,000 U/kg procaine penicillin for five days and 1mg/kg IM methylprednisolone acetate for three weeks, tapering to 0.1 mg/kg for 1 more week. As an analgesia, 0.1 mg/kg IM injection of buprenorphine was given for three days. Two animals were euthanized with 1 ml/4.5 kg intravenous Euthasol (each milliliter contains 390 mg of pentobarbital and 50 mg of phenytoin) at two weeks following the surgery, and the remaining four animals were euthanized at four weeks following ONT surgery.

Tissue processing

All twelve animals were euthanized and perfused intracardially with 750 ml phosphate buffer saline (PBS; 9 g/l NaCl, 3.96 g/l NaH₂PO₄-H₂O, and 22.68 g/l HNa₂O₄P-7H₂O), followed by one liter of 4% paraformaldehyde and again with 200–300 ml PBS following 25 mg/kg IV pentobarbital anesthesia. The brain was removed and placed in a solution of 10% sucrose in phosphate buffer and shipped on ice packs by overnight courier to the University of British Columbia. Upon arrival, area V1 and LGN sections were dissected from the rest of the brain. Area V1 was blocked such that it could be gently flattened tangentially between two glass slides before being frozen on dry ice and stored at −80 °C. The LGN was blocked in the coronal plane, frozen on dry ice and stored at −80 °C as previously described [23].

CO histochemistry and immunohistochemistry

Tissue blocks were cut tangentially at 50 μm on a freezing microtome. Alternate 50 μm sections were stained for CO activity using previously-published protocols [23,55]. Briefly, 20 mg of diaminobenzidine (DAB; Sigma-Aldrich, St. Louis, MO) was dissolved in 50 ml of distilled water. Once dissolved, 50 ml (0.1 M, pH 7.2) phosphate buffer (PB), 2 g sucrose, 30 mg cytochrome C (Sigma-Aldrich) derived from horse heart and 20 mg of catalase (Sigma-Aldrich) derived from bovine liver were added to the DAB solution. Then, 5 ml of 1% nickel ammonium sulfate was added dropwise followed by approximately 3 ml of 1% cobalt chloride until the solution appeared slightly opaque. Tissue sections were placed into 12-well plates filled with 1.5 ml of the cytochrome oxidase solution and incubated at 37 °C for 30–45 min. Upon completion of the reaction, sections were washed three times for 5 min each in PB. Then the sections were mounted onto glass slides and air dried overnight. Mounted sections were then dehydrated in a series of graded alcohols and cleared in xylene. Mounting medium, Permount (Fisher Scientific, Waltham MA) was applied onto sections and then coverslipped.

Monoclonal antibodies against human CD11b (MAC-1; Cederlane Labs, Burlington, VA) and glial fibrillary acidic protein (GFAP; Chemicon, Temecula, CA) were used to identify activated microglial and astrocytes, respectively. Free-floating sections were incubated in 3% normal horse serum (NHS) for 1 h to block nonspecific binding, and then washed three times for 5 min each in PB before incubating in primary antibodies. Primary antibodies against MAC-1 and GFAP were used at a 1:300 dilution in 3% NHS in 0.3% Triton X-100 (TX-100) made in PB. After rinsing, sections were incubated at 4 °C in primary antibody for 36 to 48 h with agitation. They were then rinsed three times for 5 min each in PB before incubating in 0.1% secondary antibody (biotinylated anti-mouse made in horse) in 3% NHS and 0.3% TX-100 in PB for 2 h at room temperature. Next, sections were placed in a solution of avidin biotin complex (Vector Laboratories, Burlingame, CA) for 1 h. Visualization of the antibody binding was undertaken using a glucose oxidase driven DAB as described previously [23,55]. Sections were mounted on gelatin-coated slides, air-dried, dehydrated, cleared in xylene and coverslipped with Permount (Fisher Scientific). Negative control sections were processed after the primary antibody was replaced with a solution containing 3% NHS and 0.3% TX-100.

Data collection and analysis

Images of processed LGN and V1 sections were captured using standard digital camera settings and 1×, 20×, 40×, and 100× objective lenses attached to an upright, bright-field compound microscope (Nikon, Tokyo Japan). Serial sections were aligned using the pattern of blood vessels as fiduciary landmarks. High power images of GFAP and MAC-1 immunoreactivity (Figure 1D-I, Figure 2D-I, Figure 3D-I, and Figure 4D-I) were taken using a 100× oil objective lens. A semiquantitative analysis of the immunoreactivity was assessed using three microscope fields in each of three tissue sections for a total of nine microscope fields. The intensity of immunoreactivity in the treated eye layers was scored in a semiquantitative manner. Background immunoreactivity was represented with a minus sign (−). One plus (+) represented moderate immunoreactivity, while two pluses (++) represented significant immunoreactivity (see Table 1). The diameters of the immunoreactive cell profiles were measured using a 40× objective and 10× eyepieces and expressed as mean±SD (Figure 5). A paired t-test (Minitab Statistical Software, State College, PA) was used to compare the measurements of cell profile diameters for GFAP and MAC −1 populations and a two sample t-test (Minitab Statistical Software) was used to compare the GFAP populations between ONT and OHT tissues.

Animals with optic nerve transection

CO histochemistry revealed light CO staining in layers 1, 4, and 6 of the LGN, a pattern that confirmed that the transected optic nerve (right) was contralateral to the processed hemisphere (left) in five of the six brains (Figure 1A,D,E). CO staining in layers 2, 3, and 5, the ipsilateral layers of the LGN, demonstrated darker CO staining and represented normal levels of CO activity in both the 14 day and the 28 day post ONT brains. In one animal (AP75), the hemisphere ipsilateral to the transected nerve was processed; CO histochemistry revealed light CO staining in the ipsilateral layers of the LGN (layers 2, 3, and 5). The relative density of the CO staining in the treated magnocellular and parvocellular layers of the LGN was equally reduced after ONT. CO histochemistry of tangential sections through V1 revealed a series of lightly and darkly staining ocular dominance bands in layer 4C (Figure 2A,D,E) and blobs in layer 2/3 (not shown), reflecting the transsynaptic loss of geniculocortical afferent activity associated with the transected nerve. This pattern of CO staining in the LGN and V1 is consistent with earlier studies [1,4,5,7,8] in which animals were subjected to ONT or MD by lid suture.

Astrocytic reactivity was assessed with GFAP immunoreactivity. All ONT brains processed for GFAP immunoreactivity demonstrated a pattern of robust labeling in the treated eye layers of the LGN (n=4) and the treated ocular dominance bands of area V1 (n=5; Figure 1B,F,G; Figure 2B,F,G, and Figure 6B,C). GFAP immunoreactivity revealed cytoplasmic labeling of astrocyte somata and their highly ramified processes (Figure 1F and Figure 2F). The mean diameter of the GFAP immunoreactive cell bodies in the parvocellular layers of the LGN was 14.2 µm (±0.4 SD). The mean diameter of the GFAP immunoreactive profiles in V1 was 11.8 µm (±0.9 SD). GFAP immunoreactivity also resulted in strong labeling of highly ramified processes that surrounded what were presumed to be unlabeled neuronal profiles.

LGN layers receiving input from the intact optic nerve and the ocular dominance bands in V1 associated with the normal eye were significantly less immunoreactive for GFAP. At high power, GFAP immunoreactivity in the normal eye layers/bands demonstrated sparse labeling of thin, astrocytic processes and few somatic profiles (Figure 1G and Figure 2G).

Microglial activation was assessed by immunoreactivity to MAC-1. MAC-1, also known as CD 11b, is a component of the complement receptor C3b. All ONT brains processed for MAC-1 immunoreactivity demonstrated robust labeling in the treated layers/bands in the LGN (n=2) and V1 (n=3; Figure 1C,H, Figure 2C,H, and Figure 5B,C). Both the cytoplasm and processes of microglial cells displayed robust immunoreactivity. The diameter of the immunoreactive profiles in the LGN and V1 were 10.5 µm (±0.4 SD) and 9.5 µm (±0.4 SD), respectively. In the normal eye layers/bands of the LGN and V1, few immunoreactive cell bodies or processes were present (Figure 1I and Figure 2I,).

We performed a paired t-test to verify that, based on cell diameter, the population of GFAP cells was distinct from the population of MAC-1 immunoreactive cells in the LGN of ONT brains (p<0.05). This test also confirmed that GFAP and MAC-1 positive cells were from separate populations in V1 of ONT brains (p<0.05; Figure 6A).

Animals with unilateral ocular hypertension

Vision loss by unilateral OHT caused a downregulation of CO activity in the LGN layers innervated by the OHT eye in all six brains processed. Each animal experienced different degrees of IOP elevation, ranging from 35 to 51 mmHg. Furthermore, the durations of elevated IOP also differed among the animals, ranging from 14 to 58 days. The calculated integral values, which reflect the relative severity of OHT, ranged from 672 to 2,632 (Table 1). The fluctuations in CO density between the treated and normal eye layers of the LGN were of similar relative contrast among all six brains. The example shown in Figure 3A is from an animal in whom the ipsilateral hemisphere was processed, and thus the downregulation in CO activity was evident in layers 2 (magnocellular) and layers 3, 5 (parvocellular) of the LGN (Figure 3D,E). The example shown in Figure 5A is from an animal in whom the contralateral hemisphere was processed, and thus the downregulation in CO activity was evident in layers 1, 4, and 6. CO histochemistry of tangential sections through V1 revealed a series of lightly and darkly stained ocular dominance bands in layer 4C (Figure 4A), reflecting a transsynaptic reduction in activity of the geniculocortical afferent pathway associated with the OHT eye. The observed pattern of CO staining in the LGN and V1 was consistent with the results after vision loss by ONT as described above and consistent with earlier studies after short-term and long-term unilateral OHT [21-23].

Astrocyte activation after unilateral OHT was assessed with GFAP immunoreactivity. All OHT brains processed for GFAP immunoreactivity demonstrated robust immunostaining in the treated eye layers of the LGN (n=3), and moderate to robust immunostaining in the treated ocular dominance bands of V1 (n=5; Figure 3B, Figure 4B, and Figure 5A). Higher power images of GFAP immunohistochemistry demonstrated labeling of astrocytic processes often surrounding what were presumed to be unlabeled neuronal profiles (Figure 3F and Figure 4F). The mean diameter of the GFAP immunoreactive cell bodies in the parvocellular layers of the LGN was 14.0 µm (±0.3 SD), while the mean diameter of the GFAP immunoreactive profiles in V1 was 11.1 µm (±1.1 SD). The normal eye layers/bands in the LGN and V1 displayed background levels of immunoreactivity for GFAP; this was evident particularly at low power (Figure 3B, Figure 4B, and Figure 5A). At high power, GFAP immunoreactivity in the normal eye layers and ocular dominance bands displayed occasional thin, astrocytic processes that were labeled, but these processes rarely outlined presumably unlabeled neuronal profiles (Figure 3G and Figure 4G). We performed a two-sample t-test to determine, based on cell diameter, whether the GFAP immunoreactive cells in the OHT and the ONT brains were from the same population. The tests demonstrated that, based on cell diameter, GFAP immunoreactive cells in the LGN of the ONT and OHT brains were from the same population (p=0.595). This was also confirmed for the GFAP immunoreactive cells in V1 of the ONT and OHT (p=0.476; Figure 6B).

MAC-1 immunoreactivity was at background levels in the LGN (Figure 3C and Figure 5A) and V1 (Figure 4C) of the animals with unilateral OHT. There were no fluctuations in immunoreactivity between treated and normal eye layers of the LGN or in the ocular dominance bands in V1; there was a uniform staining throughout the LGN and V1 of all animals tested (n=3; Figure 3H,I and Figure 4H,I).