INTRODUCTION
The striatum is a subcortical forebrain gray matter compartment of the basal ganglia comprised of neurons with relatively homogeneous morphology (the principal medium-sized spiny neurons) but diverse connections and neurotransmitters (Crittenden and Graybiel, 2011; Haber et al., 2012). The striatum is divided into the caudate nucleus and putamen that comprise the dorsal striatum, and the nucleus accumbens and olfactory tubercle that comprise the ventral striatum (Heimer et al., 1997; Haber et al., 2012). The synaptic interactions within the striatum and interconnected brain regions function in sensory-motor integration, planning and initiation of somatic movements, cognition, and some types of social-emotional and species-typical behaviors (Heimer and Van Hoesen, 2006; Crittenden and Graybiel, 2011; Baez-Mendoza and Schultz, 2013; Shepherd, 2013; Smith et al., 2014). Anatomical principles of organization of the striatum include the differentiation of the dorsal striatum into at least two primary compartments designated as striosomes (patches) and extrastriosomal matrix (Crittenden and Graybiel, 2011), the differentiation of the nucleus accumbens into core and shell regions (Heimer et al., 1997), and the serial and parallel arrangements of input-output connectivity (Alexander et al., 1986; Heimer and Van Hoesen, 2006; Haber et al., 2012). Striosomes in cat, macaque, and human caudate nucleus are organized as complex reticular networks (Groves et al., 1988; Desban et al., 1989; Manley et al., 1994; Mikula et al., 2009); specifically, in cynomolgus monkey a typical striosome segment is ~335 μm in diameter by ~1 mm in length that then can issue collateral channels (Mikula et al., 2009). Neuronal populations in striosomes and matrix are believed to differ with respect to their time of embryogenesis (van der Kooy and Fishell, 1987), afferent and efferent connectivity with cerebral cortex, substantia nigra and globus pallidus (Goldman-Rakic, 1981; Gerfen, 1984, 1985; Gimenez-Amaya and Graybiel, 1991), expression of neurotransmitters or neurotransmitter-related enzymes (Graybiel and Ragsdale, 1978; Gerfen and Young, 1988; Martin et al., 1991a), neurotransmitter receptors (Martin et al., 1993), and intracellular second messenger effector systems (Fotuhi et al., 1993; Goto et al., 1994). However, more recent studies do not support the concept of a dual striosome-matrix striatofugal system in non-human primates (squirrel monkeys) because elegant single-axon tracings show that axons of striosomal neurons target globus pallidus and substantia nigra (Levesque and Parent, 2005).
The development of the neurochemical organization of the basal ganglia in rhesus monkey and rat is vulnerable to early postnatal experience (Martin et al., 1991b; Robbins et al., 1996; Thorsell et al., 2006). The normal compartmental arrangements of peptidergic neurotransmitters of striatal output neurons, the localization of neuropeptides within striatal synaptic targets (globus pallidus and substantia nigra), and the dopaminergic afferents to the striatum from substantia nigra are all profoundly altered in adult rhesus monkeys that experienced severe sensory/social deprivation during their first year of infancy (Martin et al., 1991b). Rhesus monkeys deprived of ample infant-mother and infant-peer interactions during early development manifest psychosocial abnormalities and motor impairments, including withdrawal and fearfulness, lack of play, apathy, learning deficits, stereotypic movements, and self-injurious behaviors; as adults such monkeys do not show appropriate parental behavior and are indifferent and abusive to their offspring (Vicedo, 2009). Because monkeys that experienced social/sensory deprivation have neurochemical alterations in their basal ganglia (Martin et al., 1991b), and because the behavioral disabilities induced by social/sensory deprivation of infants primates persist into adulthood (Vicedo, 2009) and suggest dysfunction of the basal ganglia, we have hypothesized that the normal postnatal development of the structural and functional organization of the basal ganglia, particularly the dorsal striatum, is partly determined by social environment and experience (Martin et al., 1991b). However, the normal maturational events that occur in the striatum of primates during postnatal development must be understood before this theory of experience-dependent neurodevelopment of the basal ganglia can be substantiated, and before the possible cellular and molecular mechanisms responsible for the mutability of basal ganglia organization by early postnatal experience can be understood. This study was initiated to evaluate the temporal and spatial development of striatal cytology and chemical neuroanatomy in normal socially reared infant rhesus monkeys.
DISCUSSION
The major contribution of this study is the demonstration that the rhesus monkey striatum has a prolonged postnatal maturation extending over a year. This maturation was reflected by changes in cytology and chemical neuroanatomy. In many instances the remodeling was dramatic and involved reversal of chemical patterns in the different compartments of the striatum. In general, several gradual developmental events were observed in the maturing striatum, including reductions in cellular density, increased innervation by dopaminergic afferents, an initial high expression of LENK (in matrix) followed by a subsequent down-regulation of ENK expression, and a progressive increase in SP expression (in matrix). Developmental gradients were apparent in the dorsal-ventral, medial–lateral, and anterior–posterior axes of the striatum.
The development of the striatum is an example of neural pattern formation (i.e., how groups of cells assemble, form appropriate patterns of connectivity, and establish their neurotransmitter/receptor phenotypes to form functional neuronal ensembles) because it is divided into different cellular compartments. The cellular compartments were discernible even with basic histological staining methods. The Nissl-identified cellular islands in the caudate nucleus of rhesus monkey have been shown before in fetal monkeys (Goldman-Rakic, 1981) and in 2-month-old, 1-year-old, and adult monkeys (Goldman-Rakic, 1981, 1982). They are not entirely equivalent to striosomes (Graybiel and Ragsdale, 1978) because the cellular islands are fewer in number than histochemically- and immunohistochemically identified striosomes (Selemon et al., 1994), but many Nissl-identified cellular islands register with striosomes (Selemon et al., 1994). Goldman-Rakic (1981) also showed elegantly that in infant rhesus monkeys the cellular islands are largely free of prefrontostriatal innervation, while the surrounding matrix has dense innervation from frontal cortex. We also observed a subcompartmental organization within striosomes, prominently in the caudate nucleus. Subsets of SP- and LENK-immunoreactive striosomes possessed center-low and surround-high levels of immunoreactivity. This subcompartmental organization of striosomes has been seen before in rhesus monkeys by visualizing SP and SP receptor immunoreactivity (Jakab et al., 1996) and in humans by detecting neurotensin receptor binding (Faull et al., 1989), SP immunoreactivity (Holt et al., 1997), LENK immunoreactivity (Manley et al., 1994; Prensa et al., 1999), and limbic system-associated membrane protein immunoreactivity (Prensa et al., 1999).
The neurogenesis in the primate striatum has been nicely documented in the distant and recent literature. Studies in rhesus monkey fetus show that neostriatal neurons are generated during a period lasting about 50 days (gestational days 36–85 of a 165 day gestation) with maximal neuronal proliferations occurring at about fetal day 45 (Brand and Rakic, 1979, 1980). The human striatum emerges embryonically from the lateral ganglionic eminence (Pauly et al., 2014) with the caudate being identified at the 13 mm stage (~6.5 weeks) and the putamen recognized at the 26 mm stage (~8.5 weeks; Humphrey, 1968). Human striatal neurons originate from PAX6-negative lateral ganglionic eminence precursors, and the immature striatal neurons express the postmitotic marker FOXP1 already at 50 days post-fertilization (Pauly et al., 2014). Limited neurogenesis persists in adult striatum of squirrel monkeys (Bedard et al., 2002) and humans (Ernst et al., 2014). Interestingly, the adult generated striatal neurons in humans are interneurons (Ernst et al., 2014), but in squirrel monkeys they are projection neurons rather than interneurons (Bedard et al., 2006).
Cells within striosomes and matrix appear to have distinct developmental programs, because neurons within different compartments of the neostriatal mosaic are generated at different times during gestation. Neurons that reside within striosomes become postmitotic before neurons that reside within the matrix in rat (Marchand and Lajoie, 1986; van der Kooy and Fishell, 1987). The striatum in rhesus monkey appears to follow a similar histogenetic pattern regarding the differential birth of striosome and matrix neurons (Brand and Rakic, 1979). In addition, the establishment of efferent connections of early born striatal neurons with midbrain targets may allay developmental cell death within the rodent neostriatum (Fishell and van der Kooy, 1991). However, the developmental events that underlie striatal cell fate and the formation and maintenance of neuronal phenotype expression and afferent/efferent connectivity of distinct populations of striatal neurons are unclear. Important questions about striatal development concern the relative contributions of genetic, extrinsic epigenetic, and environmental determinants in the generation of its compartmentalization (Fishell and van der Kooy, 1987; van der Kooy and Fishell, 1987; Martin et al., 1991b). These questions apply to the prenatal stages of compartment formation and to the postnatal period when the striatum undergoes extensive maturational changes for an extended duration during infancy as shown here. Gender might also be an important variable in this regard, but because of our small sample size, gender differences were not examined. The present study shows, for the first time, that at birth the rhesus monkey striatum has an immature cytologic and chemoarchitectonic organization and that the postnatal maturation of the striatum is accompanied by prominent temporal and spatial changes in the expression of several neuronal phenotypes.
Experimental data on the development of the chemical neuroanatomy of the non-human primate striatum is sparse, despite the prominent anatomical representation of the striatum in primates and its functional role in complex motor and non-motor behaviors including sensorimotor integration, motor control, cognition, and social-emotional behaviors such as mood and addiction, and its role in several human neurological disorders (Crittenden and Graybiel, 2011; Baez-Mendoza and Schultz, 2013). A few available studies of the striatum in baboons show prenatal and/or postnatal changes in adrenergic, cholinergic, and dopaminergic ligand binding sites by autoradiography (Slesinger et al., 1988; Lowenstein et al., 1989) as well as sodium-dependent high affinity choline uptake binding sites (Lowenstein et al., 1987). However the normal spatiotemporal development of connectively, cell adhesion molecules, synapses, neurotransmitters, receptor proteins, and signal transduction molecules in the striatum of non-human primates is mostly unknown. One study used infant rhesus monkeys at 2, 8, 10, and 31 days of age to show that mRNA expression of forkhead motif transcription factor gene FOXP2 is enriched in striosomes and that expression declines with maturation (Takahashi et al., 2008). We demonstrate in this study three new ideas regarding the primate striatum. First, the mature anatomical organization of the dorsal and ventral striatum in primate is established postnatally, because the cellular and chemoarchitectonic differentiation of the dorsal striatum in rhesus monkey continues significantly throughout the first postnatal year. Second, striatal maturation is topographically ordered, because the putamen, caudate nucleus, and nucleus accumbens reach adult-like chemoarchitecture at different times. This idea is best illustrated by comparing the graphs of TH, LENK, and calbindin immunoreactivities for caudate (Figure 6), putamen (Figure 7), and nucleus accumbens (Figure 9). The major developmental increases in TH are skewed to the left for caudate and to the right for putamen and the major increases in LENK in putamen occur later than in caudate. The maturation of the nucleus accumbens has a maturation tempo similar to caudate nucleus for TH, Calbindin, and LENK. Third, striosomes and matrix compartments mature postnatally at different rates. Based on these results, we conclude that primate striatum undergoes marked prolonged remodeling during postnatal development and speculate that development of normal functions within the striatum occurs postnatally and that the normal maturation of primate striatum is contingent on multiple cellular and molecular events that are specified genetically and epigenetically by environmental factors. This developmental theory of striatal postnatal maturation would be in line with the finding that striatal organization is profoundly abnormal in isolation-reared rhesus monkeys presenting with severe psychopathology (Martin et al., 1991b).
The striatum of non-primates (mostly, rat, cat, and rabbit) also is immature at birth and undergoes changes postnatally in organization (Graybiel et al., 1981). Thus like primates, the mature organization of the striosomes-matrix pattern in adult rodents, rabbits, and cats is established postnatally, but the rate of morphological and neurochemical development is very accelerated in non-primates compared to primates. In non-primates, the expressions of many neuronal and non-neuronal molecular phenotypes are developmentally regulated within the striatum. For example, in the developing non-primate striatum, ontogenetic changes occur in: the levels of enzymatic activities for acetylcholinesterase, TH, glutamic acid decarboxylase, and choline acetyltransferase (McGeer et al., 1971); the histochemical localizations of acetylcholinesterase (Graybiel and Ragsdale, 1980; Graybiel et al., 1981) and dopamine (Loizou, 1972; Olson et al., 1972; Tennyson et al., 1972; van der Kooy, 1984); the immunocytochemical localizations of SP, enkephalin, neurotensin, TH (Graybiel et al., 1981; Goedert et al., 1985; Boylan et al., 1990; Zahm et al., 1990); the expression of mRNA for neurotransmitter synthesis enzymes (Greif et al., 1992) and primary transcripts for neurotransmitters (Tecott et al., 1989; Haverstick et al., 1990; Cimino et al., 1991); the autoradiographic localizations of binding sites for SP (Quirion and Dam, 1986), opiates (Kent et al., 1981; van der Kooy, 1984), dopamine (Murrin and Zeng, 1989), acetylcholine (Nastuk and Graybiel, 1985), and other neuropeptides not normally abundant in adult striatum (e.g., oxytocin and vasopressin; Petracca et al., 1986; Tribollet et al., 1989); the expression of genes that encode messenger RNA for dopamine receptors (Guennoun and Bloch, 1991; Mack et al., 1991); and glial-derived extracellular matrix glycoproteins (Steindler et al., 1988). Importantly, it is thought that the striatum in cat is adult-like by postnatal day 21 (Graybiel and Ragsdale, 1983), and the rodent striatum has a similar developmental timing with maturity of SP and neurotensin patterns achieved at about postnatal day 20 (Zahm et al., 1990). However, a recent study of the mesocorticolimbic system in rat revealed a protracted postnatal maturation (Yetnikoff et al., 2014). The comparatively rapid maturation of the striatum in non-primates contrasts with our results showing that the duration of maturation of the rhesus monkey striatum extends to 12 months of age and beyond. Thus the developmental timing of cellular differentiation of the striatum is extraordinarily different among primates and non-primates. The retention of more immature patterns of chemoarchitecture for longer periods of time during postnatal development (neoteny) may result in a greater vulnerability of basal ganglia organization to experience and environmental stressors in infants and toddlers.
Because of the uniqueness of this animal cohort and the general lack of information on basic morphometric parameters in developing rhesus monkeys, we measured brain weights and striatal volumes. Our findings were very informative. The ratio of brain weight to body weight showed a steep progressive decline early in life, while a postnatal change in striatal volume was modest. The early postnatal decrease in brain weight to body weight ratio in rhesus monkeys sharply contrasts with observations in human infants. With humans there is a rapid rise until about 2 years of age followed by a slow gradual decline until about 13–15 years of age (Dekaban, 1978). Recent MRI studies also reveal major differences in the postnatal maturation of gray matter to white matter ratios in infant rhesus monkeys and humans (Sakai et al., 2012). Infant chimpanzees are also very distinct from infant rhesus monkeys in the maturation of gray matter and white matter, but chimpanzees have some similarities to human infants (Sakai et al., 2012). We also report that striatal volume increases postnatally in rhesus monkeys. The finding is interesting because caudate volume determined by MRI was shown typically to decrease with child development, but in autistic children caudate volume increases in association with the presence of repetitive behaviors (Langen et al., 2009). These distinctions in the postnatal development of monkey, ape, and human brains have important relevance to the evolutionary biology of the brain (Sakai et al., 2012) and may be possibly relevant to the vulnerability of the higher primate brain to early childhood environmental stress (Bogart et al., 2014). Recent imaging studies have shown that early life stress, such as traumatic adverse childhood events, is associated with reduced caudate volume in adults (Cohen et al., 2006) and reduced nucleus accumbens activation and adolescent depression (Goff et al., 2013).
We also studied aspects of striatal cytology during postnatal development. Neuronal density in striatum was found to decline during the first year of life. This reduction of cell density occurred in matrix and striosome compartments. It may be related to the modest increase in striatal volume that was shown to occur during this time. More likely, the loss of neuronal density could also be related to the elimination of cells through normal developmental programmed cell death (Martin, 2001). We found prominent evidence for naturally occurring programmed cell death through apoptosis in the striatum of infant monkeys. We used Nissl staining to identify apoptotic cells. This approach is simple, but reliable, and we have shown that cells with a Nissl-staining pattern of apoptosis are usually positive for TUNEL and cleaved caspase-3 (Al-Abdulla et al., 1998; Martin et al., 1999; Northington et al., 2001; Lok and Martin, 2002; Mueller et al., 2005). Apoptosis that is detected by Nissl staining also registers with internucleosomal fragmentation of DNA detected by Southern blotting (Lok and Martin, 2002). This apoptosis data is useful because it demonstrates that another component of the postnatal remodeling of the monkey striatum involves cell death, in addition to the changes in neuropeptide localization and abundance.
Our current work is likely to be relevant to human infants because it may be reflective of the neoteny of the higher primate brain. It is known that the human prefrontal cortex undergoes extensive synaptic remodeling at puberty and beyond adolescence (Petanjek et al., 2011). The adult human dorsal striatum has a prominent striosome-matrix organization, as well as additional characteristics that could be peculiar to the human striatum (Holt et al., 1997; Prensa et al., 1999). The adult human nucleus accumbens also has a fascinating complexity with features not seen in monkeys (Prensa et al., 2003). Moreover, babies that are born prematurely show MRI abnormalities in the ventral striatum (Shi et al., 2013) and abnormalities in frontostriatal connectivity (Duerden et al., 2013). The ontogeny of the human striatum has not been studied comprehensively, and, indeed, very few studies have been done. Most studies to date have focused on the fetal brain (13–40 weeks gestation) and have shown in striatum patchy distributions with striosomal enrichments of acetylcholinesterase (AChE) and muscarinic cholinergic binding sites (Nastuk and Graybiel, 1985), AChE and NADPH-diaphorase (Sajin et al., 1992), calbindin (Letinic and Kostovic, 1996), dopamine D1 receptor and dynorphin mRNA (Brana et al., 1996), glutamate receptor scaffold kinase anchoring protein (Ulfig et al., 2001), and synaptophysin (Ulfig et al., 2000). A study of postnatal human brain revealed an inversion from an immature pattern of calbindin-enriched striosomes to a mature pattern of calbindin-enriched matrix sometime between term gestation to 18 months of life (Letinic and Kostovic, 1996). Another study showed that the immature pattern of AChE-enriched patches is still present at 3 months of age (Graybiel and Ragsdale, 1980). Thus, the human striatum also undergoes major prenatal and postnatal remodeling in its chemical neuroanatomy, but the extent is unknown, begging the question of whether the postnatal pattern formation and neurogenesis in this brain region, known to function in complex behaviors and shown to be involved in several motor and behavioral disorders (Crittenden and Graybiel, 2011; Baez-Mendoza and Schultz, 2013), is vulnerable to environmental influences and stress through epigenetic mechanisms.