Higher Expression of Serotonin 5-HT_2A Receptors in the Postmortem Brains of Teenage Suicide Victims

All study victims were diagnosed by means of the Structured Clinical Interview for DSM-III-R (13) and the Diagnostic Evaluation After Death (14); diagnoses were based on the information obtained from interviewing at least one family member and from all available medical records. Discrepancies were resolved through a consensus conference between two senior psychiatrists. Subjects were considered to be suicides only if the manner of death was determined to be suicide by the medical examiner. The comparison groups consisted of normal subjects who were verified as free from mental illnesses by use of these same diagnostic procedures. This study was approved by the institutional review board at our facility.

Tissues were homogenized in 10 vol of hypotonic medium (5 mM of Tris and 0.1% of EDTA; pH=7.5), by using a polytron setting of 9 for 30 seconds, and centrifuged at 1,000 g for 10 minutes at 4°C. The resulting supernatant was centrifuged at 30,000 g for 15 minutes at 4°C. The pellet thus obtained was washed twice with 10 vol of hypotonic buffer and centrifuged. The final pellet was suspended in 2–5 ml of incubation buffer (50 mM of Tris, 120 mM of sodium chloride, 5 mM of potassium chloride, 1 mM of magnesium chloride, and 0.05% of ascorbic acid; pH=7.4), depending on the size of the membrane pellet. This suspension was used for the determination of the 5-HT_2A receptor level.

The receptor binding assay was carried out in triplicate in plastic tubes containing incubation buffer, [¹²⁵I]LSD, ranging from 0.25–3.00 nM (five to six different concentrations), and a 40-μl cortical membrane suspension with or without 1 μM of ketanserin in a total incubation volume of 100 μl and incubated at 37°C for 1.5 hours. The incubation was terminated by rapid filtration over glass microfiber filters and washed three times with 5 ml of ice-cold 50-mM Tris buffer (pH=7.7) containing 0.01% bovine serum albumin. The filters were dried and then counted by a gamma counter. Specific binding was defined as the difference between the binding observed in the presence or absence of 1 μM of ketanserin and ranged from 75% to 80%, depending on the concentration of the ligand.

The mRNA content of the 5-HT_2A receptors was measured by use of quantitative reverse transcription polymerase chain reaction, as described earlier (15), with internal standards and these amplification primers: forward, 772–779 bp (5′ATCAAGTCACTCCAGAAAGAAGCT); reverse, 1153–1176 bp (5′TGAAAAGGCTGACCTATAGGTCTT).

Internal standard templates were generated by site-directed mutagenesis to introduce a BglII restriction site midway between the amplification primers so that the digestion of internal standard amplicon would generate two fragments of approximately equal size (208 and 197 bp). The internal primer sequence was as follows: bp=958–981, 5′AAGGCATGCAAGATCTGGGCATC. The italicized bases indicate the BglII restriction site, while bold and italicized bases [bold not shown; “ATC”] indicate the mutation site. To quantitate 5-HT_2A receptor mRNA, various amounts of cRNA were added to a l-μg mixture of total RNA and RNA–cRNA and reverse transcribed by using 200 U of Moloney’s murine leukemia virus reverse transcriptase. The transcription was carried out in the presence of random hexamer, 50 mM of Tris hydrochloride (pH=8.3), 74 mM of potassium chloride, and 3 mM of magnesium chloride in a volume of 20 μl. The incubation was carried out at room temperature for 10 minutes then at 37°C for 1 hour; the contents were heat denatured at 95°C for 5 minutes. The aliquots were amplified with 1 μM of primers, 200 mM of deoxynucleotide triphosphate, 1.5 mM of magnesium chloride, 50 mM of Tris hydrochloride (pH=9.0), 20 mM of ammonium sulfate, 1 μCi of [³²P]cytidine 5′-triphosphate, and 2.5 U of Hot Tub DNA polymerase (Amersham Pharmacea Biotech, Piscataway, N.J.) in a total volume of 100 μl. The mixture was amplified for 30 cycles: 94°C for 45 seconds, 60°C for 1 minute; 72°C for 1 minute, and 72°C for 15 minutes. Fifty-one microliters of aliquots were digested with 30 U of BglII overnight at 37°C in 50 mM of Tris hydrochloride (pH=8.0), 10 mM of magnesium chloride, and 100 mM of sodium chloride in a volume of 60 μl. The competitive polymerase chain reaction was analyzed by agarose gel electrophoresis in triplicate (1.5% weight/volume) in 0.5 × Tris-borate-EDTA buffer.

To quantitate the amount of product, bands stained with ethidium bromide were removed and counted. Data are presented as the counts incorporated into the amplified cRNA standard, divided by the counts incorporated into the corresponding subunit mRNA amplification product, versus the counts from a known amount of internal standard (cRNA) added to the test sample.

Tissues were homogenized in homogenizing buffer containing 5 mM of Tris hydrochloride (pH=7.4), 0.1% of EDTA, 2 μM of leupeptin, 0.5 mM of phenylmethylsulfonyl fluoride, 1.45 μM of pepstatin, 0.2 U/ml of aprotinin, and 2 mM of dithiothreitol. The homogenate was centrifuged at 3,000 rpm for 10 minutes at 4°C. The supernatant was recentrifuged at 32,000 rpm for 15 minutes at 4°C. The resulting pellet was resuspended in homogenizing buffer.

The samples (30 μg of protein in each lane) were resolved on 7.5% polyacrylamide gel and then blotted for 2 hours on enhanced chemiluminescence membranes. The membranes were then incubated overnight with 5-HT_2A receptor antibody (1:5000 dilution; Pharmingen, San Diego) and thereafter with a horseradish peroxidase-linked secondary antibody (antimouse immunoglobulin G [IgG], 1:3000 dilution) for 5 hours. The signal was detected by treating the blots with Western blot detection reagent (Amersham Pharmacia Biotech, Piscataway, N.J.) and exposing them to enhanced chemiluminescent membrane. To determine the immunolabeling of β-actin, the membranes were stripped and incubated with β-actin (monoclonal β-actin antibody) at a dilution of 1:3000 for 5 hours and with secondary antibody (antimouse IgG) at a dilution of 1:5000 for 2 hours. The optical densities of the bands were quantified by using a densitometer (Loats Image Analysis System, Westminster, Md.), and values were corrected with the optical density of β-actin protein in each sample.

Brain tissue samples that were flash frozen after autopsy and kept at –80°C were dropped into ice-cold fixative (4% paraformaldehyde in 0.1 M of phosphate-buffered saline; pH=7.4) and rotated for 2 hours at 4°C. Tissues were kept in fixative at 4°C for 24 hours, embedded in 30% sucrose in phosphate-buffered saline, and refrozen on dry ice for sectioning.

Twenty-micron sections were rinsed in phosphate-buffered saline for 2 days at 4°C and blocked with RPMI medium 1640 (Gibco-BRL, Grand Island, N.Y.) for 30 minutes. This was followed by rinsing in 2% normal serum for 30 minutes and 1% bovine serum albumin in phosphate-buffered saline for 30 minutes before incubation in 5-HT_2A receptor antibody (2 μg/ml) diluted in 1% bovine serum albumin in phosphate-buffered saline for 18 hours at 4°C, followed by 2 hours at room temperature, while rotating. Sections were rinsed twice in 1% bovine serum albumin in phosphate-buffered saline for 30 minutes, then incubated for 60 minutes (at room temperature) with the corresponding secondary antibody conjugated to l nm of colloidal gold particles and diluted in 1% bovine serum albumin in phosphate-buffered saline (concentration of 1:200). Sections were then rinsed several times for 30 minutes with 1% bovine serum albumin in phosphate-buffered saline, followed by distilled water. For gold particles to be visible with light microscopy, they were silver-enhanced for 12 minutes and rinsed repeatedly with distilled water. For comparison sections, from which nonspecific labeling was determined, an identical protocol was followed, except that 1% bovine serum albumin in phosphate-buffered saline was substituted for the primary antibody. Sections were then mounted onto microscope slides, air dried, and counterstained with toluidine blue.

For quantification purposes, an oil immersion objective of 100× was coupled with a 1.6× magnification tube; therefore, the number of immunogold particles was clearly discernible. When there is dense immunolabeling, immunogold particles sometimes overlap. Particles are counted by using a Samba-2000 image analysis system (Imaging Products International, R&M BioMetrics, Inc., Chantilly, Va.). The grain-counting program first calculates the average size of a silver-enhanced gold particle, then quantifies the number of gold particles on the basis of their average size. For each area, gold particles were also quantified manually by an experimenter who was blind to the diagnosis to ensure the reliability of the computerized counts. An interclass correlation coefficient (ICC) with 95% confidence intervals (CIs) for the reliability of computer-generated counts, as compared to manual counts, for gold particle quantification were as follows: pyramidal cells: ICC=0.9177 (95% CI=0.8194–0.9625) for normal subjects and ICC=0.9364 (95% CI=0.8531–0.9725) for suicide victims; neuropil: ICC=0.9015 (95% CI=0.7839–0.9551) for normal subjects and ICC=0.9054 (95% CI=0.7814–0.9591) for suicide victims.

With the Samba image analysis, 120×120-μm sampling frames were randomly placed in each cortical layer. For pyramidal cells, only cells that met the following criteria were measured: 1) presence of a nucleolus, 2) having a pyramidal shape, and 3) presence of an apical dendrite. The average number of gold particles per 100 μm² was calculated in the cell soma of at least 10 randomly selected cells in each cortical layer for each of three to five 20-μm sections of the prefrontal cortex. The mean coefficients of error for average number of gold particles per cell was 2.8% (SD=0.5%) for the normal subjects and 3.2% (SD=0.6%) for the suicide victims. To estimate gold immunodensity in the neuropil, we averaged at least 10 measurements taken in randomly selected 100-μm² cell-free regions of each cortical layer of three to five 20-μm sections. The mean coefficients of error for average number of gold particles per 100 μm² of cell-free neuropil region were 1.3% (SD=0.2%) for the normal subjects and 1.1% (SD=0.3%) for the suicide victims. Nonspecific labeling was determined by counting the number of gold particles per 100 μm² in sections in which the primary antibody was omitted. Background labeling was corrected by subtracting nonspecific labeling on pyramidal cells from the total labeling on pyramidal cells and likewise by subtracting nonspecific labeling in the neuropil from total labeling in the neuropil.

Data analyses were performed by using the SPSS 8.0 statistical software package (SPSS, Inc., Chicago). [¹²⁵I]LSD binding, protein, and mRNA 5-HT_2A receptor levels were compared between suicide victims and normal subjects by using analysis of covariance, in which race was used as covariate. Group differences between normal subjects and suicide victims with or without a history of mental disorders were tested in one-way analysis of variance, followed by Bonferroni’s multiple comparison procedure where significant main effects were present. The relationships among postmortem interval, age, and 5-HT_2A receptor measures were determined by Pearson product-moment correlation analyses. Comparisons of 5-HT_2A receptor measures with gender or race were performed with independent t tests. Age and postmortem interval were compared between suicide victims and normal subjects with independent t tests. The level of significance was alpha ≤0.05.

The demographic characteristics, postmortem intervals, and toxicology of the suicide victims and normal subjects are shown in Table 1. There were no significant differences in age or postmortem interval between the suicide victims and the normal subjects. We did not observe any significant effects of age on [¹²⁵I]LSD binding (r=0.24, df=28, p=0.49), protein expression of 5-HT_2A receptor immunolabeling (r=0.26, df=28, p=0.16), or mRNA expression levels (r=0.11, df=28, p=0.56) in the prefrontal cortex. Similarly, there were no significant effects of postmortem interval on [¹²⁵I]LSD binding (r=0.11, df=28, p=0.55), mRNA expression levels (r=0.17, df=28, p=0.37), or protein expression levels (r=0.19, df=28, p=0.31) in the prefrontal cortex. There were no significant correlations between gender and [¹²⁵I]LSD binding (r=0.07, df=28, p=0.71), mRNA levels (r=0.03, df=28, p=0.88), or protein expression (r=0.03, df=28, p=0.88) in the prefrontal cortex. As in the prefrontal cortex, we did not find any significant effects of age, postmortem interval, or gender on [¹²⁵I]LSD binding, mRNA expression, or protein levels of 5-HT_2A receptors in the hippocampus or nucleus accumbens. Although we did not observe any significant correlation between age and 5-HT_2A receptors, such a relationship could not be ruled out, since the age range of our subjects was so narrow (i.e., 13 to 19 years).

The group of suicide victims was composed of one Hispanic, two blacks, and 12 whites; the group of normal subjects was composed of 10 blacks and five whites. No significant effect of race was observed on any of the measures of 5-HT_2A receptors in either the prefrontal cortex, hippocampus, or nucleus accumbens.

Earlier studies have examined 5-HT_2A receptors by means of radioligand binding techniques with different radioligands. Since a higher level of 5-HT_2A receptor binding sites has been reported in the brains of adult suicide victims, it was of interest to examine whether similar changes were present in the prefrontal cortex of teenage suicide victims. We determined 5-HT_2A binding sites by Scatchard analysis, using [¹²⁵I]LSD as the ligand and a saturation isotherm and Scatchard plot for a normal subject and a suicide victim, as shown in Figure 1. After analyzing the binding data, we observed that the mean B_max value of [¹²⁵I]LSD binding was significantly greater in this brain area of teenage suicide victims than in this area of normal subjects, without any significant difference in K_D values (Table 2).

We studied earlier the distribution of 5-HT_2A receptor mRNA in various areas of the human brain and observed that 5-HT_2A receptor mRNA levels are highly expressed in various cortical areas, the hippocampus, amygdala, and nucleus accumbens (14); therefore, we next quantitated 5-HT_2A receptor mRNA in the prefrontal cortex, hippocampus, and nucleus accumbens obtained from teenage suicide victims and normal subjects, using the quantitative reverse transcription polymerase chain reaction technique. A representative autoradiogram of the gel electrophoresis of 5-HT_2A receptors in the prefrontal cortex is given in Figure 2A. As expected, we observed the amplification product arising from the mRNA template at 405 bp and the corresponding digestion products arising from the cRNA at 208+197 bp. As shown in Table 2, we found that the mean 5-HT_2A mRNA levels were significantly greater in the prefrontal cortex of the 15 teenage suicide victims than in the 15 normal subjects. We also found that the mean mRNA levels in the hippocampus of the suicide victims were significantly higher than in the normal subjects, as shown in Table 2.

To determine if the higher levels of 5-HT_2A mRNA observed in the prefrontal cortex and hippocampus were generalized abnormalities or were specific to certain relevant areas of the brain, we determined 5-HT_2A receptor mRNA levels in the nucleus accumbens, which also is rich in 5-HT_2A receptor mRNA (14). However, the mean 5-HT_2A receptor mRNA levels in the nucleus accumbens of suicide victims were not significantly different from those of the normal subjects, as shown in Table 2, indicating that the higher 5-HT_2A receptor mRNA in suicide victims is specific to certain brain areas, such as the prefrontal cortex and hippocampus.

To further examine if the changes in the gene transcription (mRNA levels) of 5-HT_2A receptors also occur at the translational level, we sought to determine the immunolabeling of 5-HT_2A receptor protein, using a specific antibody. A representative autoradiogram showing a Western blot of 5-HT_2A receptors is depicted in Figure 2B, which shows that the levels of 5-HT_2A receptor protein are much higher in the prefrontal cortex from three teenage suicide victims than in the brains of three normal subjects. Comparison of the mean 5-HT_2A receptor protein levels in the prefrontal cortex of 15 suicide victims and 15 normal subjects revealed that the protein expression levels of 5-HT_2A receptors were significantly higher in these suicide victims than in the normal subjects (Table 2). The mean 5-HT_2A protein levels were also significantly higher in the hippocampus, but not in the nucleus accumbens, of the suicide victims compared with the normal subjects (Table 2), suggesting that abnormal 5-HT_2A protein levels are specific to the prefrontal cortex and hippocampus and that this larger amount is not a generalized abnormality, since these levels appeared to be normal in the nucleus accumbens of the suicide victims.

In accordance with previously published reports (16–18) making use of immunogold labeling, we found the expression of 5-HT_2A receptors to be most dense in pyramidal neurons in layers III, V, and VI and their apical dendrites (Figure 3). Immunogold labeling for 5-HT_2A receptors was also observed in a small subgroup of γ-aminobutyric acid (GABA)-ergic neurons, and a moderate expression was also found in glial cells. To localize the changes in the expression density of 5-HT_2A receptors at the cellular level, we quantified the number of immunogold particles on pyramidal cell somata and in soma-free regions of the neuropil. Because of its localization in only a subset of GABA-ergic interneurons, the expression density of 5-HT_2A receptors was not quantified on this cell type. Figure 3 shows low-magnification photomicrographs of 20-μm sections of Brodmann’s area 9 from a normal teenage subject and a teenage suicide victim that were immunogold-labeled for 5-HT_2A receptors. Higher magnification of layer V pyramidal neurons shows a higher expression density of 5-HT_2A receptors in the teenage suicide victim than in the normal subject. The mean expression levels of 5-HT_2A receptors were significantly higher in the pyramidal cells of layer V of the teenage suicide victims (N=9) than in the normal subjects (N=9), whereas no changes in expression levels of 5-HT_2A receptors were found in the pyramidal cells of other cortical layers (layers III and VI) or in the surrounding neuropil (i.e., layers IV, V, and VI) (Figure 4). The mean values of positive neurons of 5-HT_2A receptor (number of gold particles/100 μm²) in cortical layers in the normal subjects and the suicide victims were as follows: pyramidal cells (layer III—normal subjects: mean=537.23 gold particles/100 μm², SD=143.51, suicide victims: mean=502.92, SD=85.91; layer V—normal subjects: mean=341.75, SD=54.06, suicide victims: mean=542.73, SD=82.34; layer VI—normal subjects: mean=442.87, SD=74.75, suicide victims: mean=445.62, SD=78.69); neuropil (layer IV—normal subjects: mean=476.90, SD=107.19, suicide victims: mean=437.85, SD=78.69; layer V—normal subjects: mean=436.62, SD=91.75, suicide victims: mean=381.27, SD=58.70; layer VI—normal subjects: mean=413.82, SD=68.65, suicide victims: mean=396.56, SD=70.80). Thus, it appears that the higher expression of 5-HT_2A receptors in the postmortem brains of teenage suicide victims is restricted to the pyramidal cells of the somata of cortical layer V.

Several studies (7) have suggested an association between mental disorders and serotonin receptors in postmortem brains of adult suicide victims. Therefore, we examined the association between 5-HT_2A receptors and mental illness in teenage suicide victims. Five suicide victims had no history of mental illness, eight had a history of mental illness (depression, adjustment disorder), and two had a history of alcohol abuse. Although there were no significant differences in any of the 5-HT_2A receptor measures, as shown in Figure 5 and Figure 6, in teenage suicide victims with or without mental illness in any of the brain areas studied, the mean mRNA (attomoles/μg of total RNA) in the prefrontal cortex tended to be higher in the suicide victims with mental illness (mean=1197.00 attomol/μg, SD=604.11) than in the suicide victims with no mental illness (mean=883.80 attomol/μg, SD=372.74). Because of the small group size, we could not examine the relationship between specific mental illness, such as depression, and 5-HT_2A indices. Our results, however, do indicate that 5-HT_2A receptor expression is higher in the prefrontal cortex of suicide victims, independent of a history of mental disorders.