The major cellular polyamines, putrescine,' spermidine, and spermine are polyvalent cations that are ubiquitously distributed in mammalian cells (Pegg and McCann, 1982) . Polyamines are involved in many diverse cellular and physiological processes including cellular growth and differentiation (Pegg and McCann, 1988), regulation of nucleic acid and protein synthesis (Slotkin and Bartolome, 1986), and stabilization of lipids . Polyamines have multiple functions within the CNS, including roles in brain development (Fozard et al ., 1980; Bell and Slotkin, 1988), nerve growth and regeneration (Kauppila, 1992), response to brain injury and stress (Dienel and Cruz, 1984; Paschen et al ., 1988), brain metabolism (Seiler and Bolkenius, 1985), regulation of ionic flux and neuronal ion channels (lqbal and Koenig, 1985 ; Scott et A., 1993), and modulation of several neurotransmitter receptors in the brain (Koenig et al ., 1989; Wasserkort et al ., 1991 ; Williams et al ., 1991). As might be expected from the regulatory nature of these compounds, polyamine synthesis is a tightly controlled process (see Fig . 1) . Putrescine is formed from the enzymatic decarboxylation of ornithine by ornithine decarboxylase (ODC; EC 4.1 .1 .17) . S-Adenosylmethionine decarboxylase (SAMDC ; EC 4.1 .1 .5) catalyzes the decarboxylation of S-adenosylmethionine to form decarboxylated S-adenosylmethionine, the donor of aminopropyl groups for spermidine and spermine synthesis . In addition, spermine may be metabolized back to spermidine, spermidine to putrescine, and putrescine to GABA through the polyamine interconversion pathway . 

The major cellular polyamines, putrescine,' spermidine, and spermine are polyvalent cations that are ubiquitously distributed in mammalian cells (Pegg and McCann, 1982) . Polyamines are involved in many diverse cellular and physiological processes including cellular growth and differentiation (Pegg and McCann, 1988), regulation of nucleic acid and protein synthesis (Slotkin and Bartolome, 1986), and stabilization of lipids . Polyamines have multiple functions within the CNS, including roles in brain development (Fozard et al ., 1980; Bell and Slotkin, 1988), nerve growth and regeneration (Kauppila, 1992), response to brain injury and stress (Dienel and Cruz, 1984; Paschen et al ., 1988), brain metabolism (Seiler and Bolkenius, 1985), regulation of ionic flux and neuronal ion channels (lqbal and Koenig, 1985 ; Scott et A., 1993), and modulation of several neurotransmitter receptors in the brain (Koenig et al ., 1989; Wasserkort et al ., 1991 ; Williams et al ., 1991). As might be expected from the regulatory nature of these compounds, polyamine synthesis is a tightly controlled process (see Fig . 1) . Putrescine is formed from the enzymatic decarboxylation of ornithine by ornithine decarboxylase (ODC; EC 4.1 .1 .17) . S-Adenosylmethionine decarboxylase (SAMDC ; EC 4.1 .1 .5) catalyzes the decarboxylation of S-adenosylmethionine to form decarboxylated S-adenosylmethionine, the donor of aminopropyl groups for spermidine and spermine synthesis . In addition, spermine may be metabolized back to spermidine, spermidine to putrescine, and putrescine to GABA through the polyamine interconversion pathway . 

The key rate-limiting enzyme of the interconversion pathway is spermidine/ spermine N-acetyltransferase . All three of the abovementioned enzymes are highly regulated, inducible enzymes with a high turnover rate (Pegg and McCann, 1988). Induction of the polyamine system may occur in response to a variety of stimuli, such as hormones, growth factors, tumor promoters, cerebral ischemia, mechanical and thermal brain injury, neurotoxin insult, neuronal deafferentation, and seizure activity (Pajunen et al ., 1978 ; Dienel and Cruz, 1984: Crozat et al ., 1992; Paschen, 1992). Neurochemical studies of the influence of aging on brain polyamine levels in the rodent have shown that polyamine levels decline precipitously during postnatal development to low levels in mature brain (Janne et al ., 1964; Suorsa et al ., 1992) . However, to our knowledge, only sparse information is available regarding polyamine levels in normal human brain (Perry et al ., 1967 ; Shaw and Pateman, 1973; Sturman and Gaul], 1975, McAnulty et al ., 1977; Chaudhuri et al ., 1983) . Previous studies have been limited either by the age range of the samples (fetal and infant brain, McAnulty et al ., 1977; Chaudhuri et al ., 1983), or by the number of cases (n = 2; Shaw and Pateman, 1973) or brain areas (whole brain, Perry et al ., 1967; occipital lobe, Sturman and Gaull, 1974) examined. The present study describes the regional distribution of putrescine, spermidine, and spermine and the influence of aging (1 day to 103 years) on polyamine levels in neurologically and neuropsychiatrically normal human brain.

MATERIALS AND METHODS

Autopsied brain was obtained from 57 subjects who died without evidence of neurological or psychiatric disease or brain pathology (mean postmortem time, 12 .3 } 0.8 h, mean -+- SEM) . At time of autopsy, each brain was dissected sagittally into two equal halves . One half-brain was used for routine histopathological analyses, whereas the other half was frozen at -80°C for biochemical analysis . Brain sample preparation, polyamine derivatization with 9-fluorenylmethyl chloroformate (FMOC), and reversephase HPLC analyses were conducted based on previous methods (Einarsson et al ., 1983 ; Sabri et al ., 1989), with modifications . Brain tissue was stored at -80°C before use, thawed on ice, and prepared by sonicating at a concentration of 400 mg of tissue/ml for 15 s in distilled H2O on ice. An aliquot (20 p1) was removed for protein determinations (Coomassie Blue protein assay, Sigma) . The samples were deproteinized by mixing an equal volume of sonicate and 10% (wt/vol) trichloroacetic acid (TCA) so that the final TCA concentration was 5% (vol/vol) . The samples were vortexed for 10 s before centrifugation at 4°C at 12,000 rpm for 30 min in an Eppendorf microfuge. The supernatant was collected, the pH adjusted to neutral with NaOH and filtered (0 .2 pm) . The samples were derivatized by adding 0.1 ml of supernatant to a tube containing 1 ml of borate buffer (0 .1 M, pH 9 .6), 1 ml of acetone, and 10 p,1 of internal standard, 100 fig/ml solution of 1,6-hexanediamine (1,6- DAH) . All tubes were vortexed before the addition of 100 p1 FMOC in acetone (0 .01 M made fresh) . 

The tubes were vortexed for 30 s and the derivatization allowed to proceed for 10 min at room temperature. After this time, 2.0 ml of hexane/ethyl acetate (1 :1, vol/vol ) were added and the tubes vortexed for 30 s. The upper solvent layer containing the polyamine derivatives was removed and the extraction process repeated . The solvent was dried in a rotary vacuum extractor and the derivatized polyamines were reconstituted in 1 ml of HPLC-grade ethanol. The derivatives were filtered (0 .2 pm) before HPLC analysis. The efficiency of the derivatization and extraction procedures was checked using radiolabeled standard, which indicated that >93% of label was extracted in the upper phase (data not shown) . A 10-pl sample was injected onto the HPLC column (5 pm, ODS 2) . The equipment consisted of a binary solvent delivery system equipped with a gradient controller and integrator (Spectraphysics) . A fluorometer (Perkin-Elmer LS30) was used to monitor the elution of polyamine derivatives from the column (excitation wavelength 254 nm, emission wavelength 316 nm) . The derivatives were separated using 20% A (consisting of 3 ml glacial acetic acid, 1 ml triethylamineper L of distilled H2O, pH 4.2)/80% B (methanol) for 7 min, 10% A/90% B from 7.2 to 22 min, then 20% A/80% B for 18 min, at a flow rate of 1 .5 ml/min. Putrescine and polyamine peaks were identified by comparison of retention time with authentic standards. Stock solutions of putrescine, spermidine, spermine, and 1,6-DAH were prepared at 100 Fig/ml and used to construct standard curves (0-10 ng of putrescine, 0-50 ng of spermidine, 0-20 ng of spermine, and 10 ng 1,6-DAH injected) .

Our data demonstrate that brain polyamine levels are not increased as a consequence of death-associated hypoxia. As shown in Fig. 3, a heterogeneous regional distribution of all three polyamines was observed in adult brain (mean age, 61 ± 6 years; one-way ANOVA, p < 0 .05) . Of the 10 regions examined, mean putrescine levels were high in temporal and occipital cortex (1 .0 and 1 .2 nmol/mg of protein, respectively) and low in cerebellar cortex and thalamus (0.3 and 0.5 nmol/mg of protein, respectively) . Highest spermine levels were present in cerebellar, occipital, and temporal cortices (1 .9-3 .4 nmol/mg of protein) and lowest in hippocampus( 1 .1 nmol/mg of protein), whereas spermidine was most concentrated in white matter and thalamus (20 and 9.3 nmol/mg of protein) with relatively low levels in frontal, insular, and cerebellar cortices (3 .7- 4.7 nmol /mg of protein) . Figure 4 shows the influence of aging on mean polyamine levels in occipital cortex of 57 neurologically and neuropsychiatrically normal subjects, aged 1 day to 103 years, who were grouped into six age groups with mean ages of 6 weeks (n = 6 ; range, 21 h to 3 .6 months ; 6 .3 +_ 2 weeks, mean ± SEM), 1 year (n = 9 ; range, 8 months to 2 years 10 months ; l year 5 months - 2.4 months), 19 years (n = 7 ; range, I I 28 years; 18 .6 ± 2.6 years), 41 years (n = 9 ; range, 30-48 years; 40.6 ± 2 .2 years), 61 years (n = 12; range, 52-69 years; 61 .4 ± 1 .5 years) ; 80 years (n = 14 ; range, 70-103 years ; 80.2 ± 2.6 years) . The mean postmortem intervals of the age groups were not significantly different (one-way ANOVA, p > 0.05) . The mean freezer storage times of the two youngest age groups (4 years) were about one-half that of the adults (8 years) .

This discrepancy may be due to differences in the polyamine recovery efficiencies and sensitivities of the different methodologies used. A heterogeneous and discrete distribution of all three polyamines was observed in autopsied human brain. Our finding that putrescine levels were highest in occipital and cerebral cortices, putamen, and hippocampus, and lowest in the cerebellar cortex and thalamus, can be compared with a previous animal (mongolian gerbil, Paschen et al ., 1988) study in which putrescine levels were lower in cerebellar cortex than in hippocampus or parietal cortex . In adult human brain, spermidine levels were markedly higher in white matter and thalamus than in the other brain regions examined (see Fig. 3) ; however, unlike an earlier report (Shimizu et al ., 1964), substantial concentrations of spermidine were present in gray matter . Our regional distribution data are consistent with the results of earlier studies on spermidine and spermine distribution in human brain, in which levels of spermine were found to be highest in the visual (occipital ) and cerebellar cortices, and spermidine levels were greatest in richly myelinated areas (Shaw and Pateman, 1973 ; McAnulty et al ., 1977) . In human brain the activity of SAMDC, the key enzyme of spermidine and spermine biosynthesis, is high in occipital and other cortical regions and caudate but barely detectable in white matter (Morrison et al ., 1993) . The reason for the low SAMDC activity and relatively high spermidine and spermine levels in white matter is unknown but could be explained by specific uptake of spermidine and spermine into white matter by transport processes similar to those previously described in various cell types (Khan et al ., 1990, 1993 ; Toninello et al ., 1992) . 

ODC activity and putrescine levels in rodent (Anderson and Schanberg, 1972) and nonhuman primate (Sturman and Gaul], 1975) brain are highest in early life and decrease to low adult levels. In the human, Sturman and Gaull (1974) reported that putrescine levels in fetal brain are approximately eightfold higher than in mature brain . Although we did not observe any precipitous reduction in putrescine levels during postnatal development, putrescine concentration was, in fact, higher in the youngest (6 weeks) age group compared with adult brain. Our observation of markedly increased spermidine in human brain from birth to adulthood is consistent with a previous study performed in rhesus monkey (Sturman and Gaul], 1975) but differs from the profile seen in rodent brain (Shimizu et al ., 1964; Shaskan et al ., 1973; Suorsa et al ., 1992) in which levels of all three polyamines decrease after birth . Because spermidine and spermine have stimulatory effects on glutamate receptor function (Ransom and Stec, 1988 ; Ransom and Deschenes, 1990; Williams et al ., 1991), these polyamines may be involved in human postnatal developmental processes in which the polyamine-activated NMDA-preferring glutamate receptor is implicated, such as regulation of neuronal survival, synaptic reorganization, connectivity, and plasticity (for review, see McDonald and Johnston, 1990) . The essential role polyamines have in brain development has been demonstrated in rodent models (Fozard et al ., 1980 ; Slotkin et al ., 1983 ; Bell and Slotkin, 1988) in which inhibition of polyamine biosynthesis resulted in fetal death or retarded brain development. 

Although the precise roles of polyamines on developing neurons are not fully understood, several studies have now associated polyamines with neurotrophic functions, such as enhancement of neuronal survival and control of axonogenesis, neurite elongation, and synaptogenesis (Slotkin and Bartolome, 1986; Abe et al ., 1993; Chu et al ., 1994) . Our observation that polyamine levels are maintained throughout adulthood suggests functions for these cations in mature brain other than those associated with rapid growth . These may include regulatory functions such as posttranscriptional modification of proteins (Folk et al ., 1980), modulation of G-protein activity (Bueb et al ., 1992), regulation of brain receptor function (Ransom and Stec, 1988 ; Koenig et al ., 1989; Wasserkort et al., 1991 ; Williams et al ., 1991), and control of synaptic and neuronal activity (Wedgewood and Wolstencroft, 1977; Iqbal and Koenig, 1985 ; Scott et al ., 1993) . In our study, the concentrations of spermidine and spermine determined in various regions of adult human brain are in the range 55-1,410 and 25-120 MM, respectively . Experimental studies indicate that both spermidine and spermine potentiate NMDA-receptor responses and modulate Ntype Caz+ channel function at concentrations of 0.2- 200 pM (Pullan et al ., 1990; Carter, 1994) . This suggests potential roles in the human brain for polyamines in the phenomenon of long-term potentiation and the processes of learning and memory (Morris et al ., 1986 ; Collingridge and Bliss, 1987; Mondadori et al ., 1988) . In this regard, spermine has recently been shown to facilitate the generation of long-term potentiation in the dentate gyrus of rodent brain in vivo (Chida et al ., 1992) .