MRNA expression patterns and distribution of white matter neurons in dorsolateral prefrontal cortex of depressed patients differ from those in schizophrenia patients

Acquisition, storage, and fixation of brain tissue 

Brain samples from the dorsolateral prefrontal cortex (area 9) of the left hemisphere of 18 patients suffering from major depressive disorder and 18 nonpsychiatric control subjects were used in the present study. Brain tissue was obtained from the University of California, Irvine, Brain Bank and Brain Tissue Repository of the Center for Neuroscience, University of California, Davis. The diagnoses were performed by Board-certified psychiatrists using DSM-IV criteria (American Psychiatric Association 1994). Among the patients, 10 were affected by major depression and 8 were affected by bipolar illness. Brains of control subjects were matched to those of the depressed patients by age, gender, and autolysis time (Table 1). Control subjects had no clinical history of neurologic or psychiatric disease or of substance abuse. None of the patients were chronic alcoholics or suffered from other potentially relevant conditions. Medication and clinical state at the time of death are listed in Table 1. After autopsy, brains were cooled to 4°C, cut into coronal slices slightly less than 1 cm thick, flash-frozen between two supercooled aluminum plates, and stored at −85°C, as previously described (Jones et al 1992).

Table 1. Characteristics of Matched Pairs

	Depressed Samples										Controls
	Pair	Gender	Age	PMI	Onset	Cause of Death	Syndrome tod	Perscription Medication tod	Diagnosis	Gender	Age	PMI	COD
	1	F	72	21	Unk	Suicide (OD)	Dep	Trazadone, Halcion, Benadryl	Maj Dep	F	70	20.5	Cardiac
	2	F	56	29	20s	Suicide (Asphyxia)	Dep	Wellbutrin, Tegretol, Xanax, Halcion, Paxil	Bipolar	F	60	24	Cardiac
	3	M	19	18	Teens	Suicide (Asphyxia)	Dep	No data	Maj Dep	M	18	22	Drowning
	4	M	22	9	21	Suicide (Asphyxia)	Dep	Mellaril, Benzotropine, Lithonate	Bipolar	M	19	6.5	Trauma
	5	M	58	24	57	Suicide (Asphyxia)	Dep	Trazadone, Prozac, Desyrel	Maj Dep	M	58	26	Cardiac
	6	M	53	17.3	20s	Suicide (OD)	Dep	Buspar, Prozac, Ascendin, Tofranil, Anafranil,	Maj Dep	M	55	18	Cardiac
	7	M	52	28	21	Suicide (Hemorrhage)	Dep	None	Bipolar	M	54	25	Cardiac
	8	M	49	31	34	Accidental OD	Dep	Zoloft, Xanax, Ambien, Amytryptyline	Maj Dep	M	50	29	Cardiac
	9	M	69	11.3	38	Trauma	Unk	None	Bipolar	M	72	15	Cancer
	10	M	46	27	20s	Cardiac	Dep	Depakote	Maj Dep	M	45	21	MI
	11	M	49	27	27	Suicide (Asphyxia)	Dep	None	Maj Dep	M	52	23	Lung Ca
	12	M	52	16	17	Cardiac	Unk	Ambien, Wellbutrin, Prozac, Risperdol	Maj Dep	M	50	16.5	Ulcer
	13	F	48	37	38	Suicide (OD)	Dep	Celexa, Neurontin	Maj Dep	F	47	41.25	Breast Ca
	14	F	63	22	20	Cardiac	Unk	Neurontin, Prozac, Pamelor	Bipolar	F	64	27	Colon Ca
	15	F	68	25.5	30	Aspirated	Dep w/ psychosis	Paxil, Dilantin	Bipolar	F	68	25	Ov Ca
	16	M	59	15.5	30s	Cardiac	Unk	Eskalith	Bipolar	M	58	15.3	Abd Ca
	17	M	69	28.5	22	Natural	Dep	None	Bipolar	M	70	27	ASHD
	18	M	39	27.5	20s	Suicide (Asphyxia)	Dep	None	Maj Dep	M	44	23	Cardiac

OD, overdose; Dep, depression; Maj, major; MI, myocardial infarction; Unk, unknown; Ca, cancer; Ov, ovarian; Abd, abdominal; ASHD, artheriosclerotic heart disease; tod, time of death; PMI, post mortem interval.

For in situ hybridization histochemistry, small blocks were cut from the left dorsolateral prefrontal cortex from each brain. Blocks were raised to approximately 4°C, and placed in cold 4% paraformaldehyde in 0.1 mol/L phosphate buffer (pH 7.4) overnight, infiltrated with 30% sucrose in 0.1 mol/L phosphate buffer, refrozen in dry ice, and kept at −85°C until sectioning.

In situ hybridization procedure and quantitative analysis 

For in situ hybridization experiments, serial sections 50-μm thick were cut on a sliding microtome and collected in groups of six. One group was processed for standard Nissl staining. The remaining sections were stored in 4% paraformaldehyde in 0.1 mol/L phosphate buffer for at least 7 days and processed for in situ hybridization histochemistry. All the 18 pairs were used for the analysis of expression of GAD67, CAMKII-α, and TBR1 mRNAs. For the analysis of BDNF mRNA expression only nine pairs were used (specifically, pairs 1, 3, 5, 6, 8 [major depressed], and pairs 2, 4, 7, 9 [bipolar patients]).

Free-floating sections were hybridized with [α-33P]UTP-labeled antisense (or sense) riboprobes transcribed from linearized cDNA templates.

α-type II calcium/calmodulin dependent protein kinase (CAMKII-α) riboprobes were transcribed from 350 nucleotide monkey cDNA, as previously described (Benson et al 1991a).

GAD67 riboprobes were transcribed from 2.7 kb human GAD cDNA (provided by Dr. A. Tobin).

BDNF riboprobes were transcribed from a 384-bp genomic fragment of monkey BDNF (Isackson et al 1991).

TBR1 riboprobes where transcribed from a 1.2 kb human TBR 1 cDNA (provided by Dr. J.L.R. Rubenstein).

Free-floating sections were prepared for in situ hybridization by washing in 0.1 mol/L glycine in 0.1 mol/L phosphate buffer (pH 7.4) followed by two washes in 0.1 mol/L phosphate buffer (pH 7.4) and two washes in 2 × saline sodium citrate (SSC; sodium chloride/sodium citrate solution pH 7.0). Sections were then incubated in the hybridization solution containing 50% formamide, 10% dextran sulfate, 0.7% Ficoll, 0.7% polyvinyl pyrolidone, 0.5 mg/mL yeast tRNA, 0.33 mg/mL yeast tRNA, 0.33 mg/mL denatured herring sperm DNA 20 mmol/L dithio-threitol (DTT), and 5.0 × 105 cpm/mL of the 33P-radiolabeled antisense or sense probe. Hybridization was carried overnight at 60°C in a humid chamber. Proteinase K treatment was omitted from the prehybridization washes to reduce fragility of the sections.

After hybridization, sections were washed twice in 4 × SSC at 60°C, digested with 20 μg/mL of ribonuclease A (pH 8.0) for 30 min at 45°C, and washed through descending concentrations of SSC to a final stringency of 0.5 × SSC. Sections were mounted on gelatin-coated slides, dried, and exposed to Amersham β-max film for 4–15 days. After development of the film, sections were lipid-extracted in chloroform, dipped in Kodak NTB2 emulsion (diluted 1:1 with water), exposed for 8 days at 4°C, developed in Kodak d-19, fixed, and counterstained with thionin.

Film autoradiograms of hybridized sections were quantified by taking optical density measurements using a microcomputer imaging device (MCID/M5; Imaging Research, St. Catharine’s Ontario, Canada). Mean gray-levels recorded by the system in each sample and reported as integrated optical density (IOD) value were converted to units of radioactivity (nCi/g) based on a set of 14C radioactive standards (Amersham) exposed on each film. At least six samples within each cortical layer were taken in at least three sections for each brain. The level of the background labeling as measured over the white matter was subtracted from each individual measurement in all layers, in each brain.

NADPH staining and neuronal counting in the white matter 

For NADPH staining, a randomly selected group of seven pairs of brains were analyzed, using a previously published protocol (Akbarian et al 1996a). Serial sections of 30 μm thickness were cut and collected in 0.1 mol/L Tris buffer (pH 8.0). Sections were repeatedly washed in 0.1 mol/L Tris buffer (pH 8.0) then transferred to a reaction solution that contained 1.2 mmol/L NADPH-d, 30 mmol/L L-malic acid (tetrasodium salt), 0.3-mmol/L nitro blue tetrazolium, 0.5% Triton X-100, and 1.2% dimethyl sulfoxide. Sections were incubated at 37°C in total darkness for up to 3 hours, then they were washed several times in 0.1 mol/L phosphate buffer (pH 7.4), mounted on gelatin coated slides, air dried, cleared in xylene, and cover slipped. At least six sections were processed from the prefrontal cortex of each brain.

For neuronal counting in the white matter, camera lucida drawings were made at six times magnification, showing gray matter-white matter borders and profiles of blood vessels as internal landmarks. The position of each neuron was plotted on the drawings, by reference to the internal landmarks. After plotting the subcortical white matter was subdivided into successively deeper 1000 μm wide compartments. Compartment 1 was the most superficial and compartment 6 was deepest. Compartment 6 was 3000 μm wide. The number of neurons for each compartment was counted, and the area of each compartment was measured using NIH Image software, available on line. Three sections were analyzed for each brain.

Statistical analysis 

For each brain and cortical layer, the mean values of GAD67, CAMKII-α, BDNF, and TBR1 mRNA levels were calculated, and the mean value and SE for each of the two cohorts were calculated from the 36 single means (for BDNF 18 single means). Statistical significance between the mean values of depressed and controls was determined by using repeated measures analysis of variance (RMANOVA), using the two cohorts as groups and cortical layers as variates. Repeated measures analysis of variance was performed separately on major depressed subjects versus controls and on bipolar subjects versus controls, using four cohorts as groups and cortical layers as variates. In addition, Student two-tailed paired t test was used to determine significance of differences between the six cohorts for each cortical layer separately.

In situ hybridization 

Figure 1 shows the general appearance of the hybridization signal obtained with antisense riboprobes for GAD67, CAMKII-α, BDNF, and TBR1 mRNAs in the dorsolateral prefrontal cortex of controls. In general, all the probes showed labeling of all six layers with a laminar pattern specific for each mRNA. Sense controls revealed only nonspecific background labeling.

Figure 1. Film autoradiograms from sections through the prefrontal cortex of control subjects showing the distribution of the 67 KD form of glutamic acid decarboxylase (GAD67), α-type II calcium/calmodulin dependent protein kinase (CAMKII-α), T-brain–1 (TBR1), and brain derived neurotrophic factor (BDNF) mRNAs. (A) GAD67. Note the numerous intensely labeled cells, especially in layers II and IV. (B) CAMKII. Labeling is more intense than for any other mRNA studied, especially in layers II and IIIa. (C) TBR1. Labeling is diffuse and present in all layers, being higher in layer VI. (D) BDNF. Labeling is weak and diffuse, but in layer V there is a line of cells with a slightly stronger hybridization signal. Scale bar is 1 mm.

GAD67 

Film and emulsion autoradiograms of sections from depressed and control samples hybridized with the antisense GAD67 riboprobe showed numerous intensely labeled cell bodies that could easily be distinguished from unlabeled cells (Figure 1A). The concentration of GAD67 mRNA in the cortex, as seen in the films and as determined by optical density measurements on film autoradiograms, was highest in layers II and IV.

No significant difference in the level of expression of GAD67 mRNA was observed between depressed subjects and controls. After segregating the samples into major depressed and bipolar cohorts, no significant differences were noted ([Figure 2] bipolar patients F [1,72] = 1.06 p > .3; major depressed patients F [1108] = 0.21 p > .6). In the bipolar patients, GAD67 expression showed a tendency to increase, but did not reach statistical significance. The level of expression between the major depressed subjects and the controls was very similar.

Figure 2. Histograms showing variations in the level of expression of the 67 KD form of glutamic acid decarboxylase (GAD67), α-type II calcium calmodulin dependent protein kinase (CAMKII-α), T-brain 1 (TBR1), and brain derived neurotrophic factor (BDNF) mRNAs in patients with major depression and with bipolar disorder. Note that CAMKII-α and TBR1 mRNA expression was significantly increased in patients with bipolar disorder.

CAMKII-α 

Labeling for CAMKII-α mRNA was more intense than for any other mRNA studied (Figure 1B). CAMKII-α labeling pattern was typically diffuse, reflecting the presence of high densities of the mRNA in dendrites in addition to somata Burgin et al 1990, Benson et al 1991a. The densest hybridization occurred in layers II and IIIa, followed by layers IV and VI. The lowest level of CAMKII-α expression was detected in layers I and V.

Using RMANOVA statistical analysis, CAMKII-α mRNA was found to be increased in the prefrontal cortex of both bipolar and major depressed patients compared with controls (bipolar patients F [1,98] = 10.29 p < .01; major depressed patients F (1126) = 5.717 p < .05). Two-tailed paired t test analysis revealed that the increased levels observed by RMANOVA analysis were significantly increased in all six layers of the cortex in the bipolar patients. Specifically, the percentage increase was: layer I, 29%; layer II, 21%; layer IIIa, 27%; layer IIIb, 36%; layer IV, 28%; layer V, 33%; layer VI, 33% (Figure 2). Two-tailed paired t test analysis failed to detect a statistically significant difference in any layers in the major depressed patients compared with the controls. The statistically significant difference detected by applying the RMANOVA test was due to the increase in the number of measurements that occurs by grouping all the layers when applying this test. Overall, in the major depressed patients, although a trend toward increased expression of CAMKII-α mRNA was detected, this did not reach a statistically significant level.

TBR1 

Labeling for TBR1 mRNA was diffuse and present in all layers, but highest in layer VI (Figure 1C). Repeated measures analysis of variance analysis revealed that the mean density of TBR1 mRNA was increased in prefrontal cortex of both bipolar and major depressed groups compared with controls (Bipolar F[1,84] = 15.99 p < .001; major depressed F[1108] = 6.743 p < .05). Two-tailed paired t test analysis on individual layers showed a statistically significant increase in layers III, IV, V, and VI of bipolar patients. Specifically, the percentage increase in bipolar patients was: layer III, 25%; layer IV, 34%; layer V, 34%; layer VI, 48%.). Two-tailed paired t test analyses failed to detect a statistically significant difference in any layers in the major depressed patients compared with controls. In the major depressed patients, however, there was a trend toward increased expression of TBR1 mRNA, although it did not reach statistical significance.

BDNF 

Labeling for BDNF mRNA was in general weak and diffuse reflecting the diffuse distribution of BDNF expressing cells in the cortex. There was a line of large cells in layer V with a slightly stronger hybridization signal than in any other layer (Figure 1D).

The level of expression for BDNF was generally reduced in the depressed cohort overall and in both major depressed and bipolar patients compared with controls; however, these changes did not reach statistical significance (bipolar patients F [1,36] = 1.8 p > .1; major depressed patients F [1,48] = 1.48 p > .2).

Distribution of NADPH-positive cells in the white matter 

In the subcortical white matter of the dorsolateral prefrontal cortex, NADPH-positive neurons were multipolar and stained intensely dark blue or black (Figure 3).

Figure3. Nicotinamide-dinucleotide phophate-diaphorase (NADPH)-diaphorase positive cells in the white matter of dorsolateral prefrontal cortex. (A) Example of NADPH-diaphorase positive cells distribution in the white matter of the prefrontal cortex of a control subject. (B) Example of a NADPH-diaphorase positive cell. Scale bar is 1 mm in A and 8 μm in B.

There was no difference in the density of NADPH positive cells in the white matter compartments of depressed subjects and controls. Similarly, no statistical difference was observed after subdividing the depressed patients into major depressed and bipolar in comparison with controls (Figure 4).

Figure 4. Histogram showing the density of nicotinamide-dinucleotide phosphate-diaphorase (NADPH)-diaphorase positive cells in the six compartments of the white matter of depressed patients and controls. No significant differences were observed in the distribution of these cells between control subjects and depressed patients.