- Email: [email protected]
The percentage of interneurons in the dorsal striatum of the rat, cat, monkey and human: A critique of the evidence
Basal Ganglia 3 (2013) 19–24
Contents lists available at SciVerse ScienceDirect
Basal Ganglia journal homepage: www.elsevier.com/locate/baga
The percentage of interneurons in the dorsal striatum of the rat, cat, monkey and human: A critique of the evidence Dorothy E. Oorschot * Department of Anatomy, Otago School of Medical Sciences, and the Brain Health Research Centre, University of Otago, Dunedin, New Zealand
A R T I C L E I N F O
A B S T R A C T
Article history: Received 1 October 2012 Accepted 7 November 2012
In the published literature on mammals, there are various answers to the question: What percentage of the total population of neurons within the dorsal striatum is represented by interneurons? The percentage of interneurons is reported to be 26% for the human, 23% or 4–6% for the monkey, 5–6% or <3% for the rodent, 4–5% for the cat and 5–10% for the mammal. Based on a critique of the literature, including the use of stereological approaches, caution is needed when citing historical non-stereological data on the percentage of interneurons in the dorsal striatum of the cat, monkey and human. In these species, the vast majority of neurons in the dorsal striatum are spiny projection neurons and a smaller number are interneurons, yet the precise percentage is currently unknown. Hence, in the absence of stereological data on the absolute number of all neuronal subtypes for the cat, monkey and human, it is currently unknown if there is a species difference, or not, for the percentage of interneurons in the dorsal striatum of the human and monkey versus the cat and rat. Modern stereological data indicate that the percentage of interneurons in the rat dorsal striatum is <3%. More specifically, in the Sprague-Dawley rat, 0.8% of all neurons are somatostatin/neuropeptide Y/GABA interneurons, 0.4% are cholinergic interneurons, 0.6% are GABA/parvalbumin interneurons and 0.5% are GABA/calretinin interneurons. Current stereological data also indicate that the percentage of interneurons in the dorsal striatum is <1% for cholinergic interneurons in the Sprague-Dawley rat and human. ß 2012 Elsevier GmbH. All rights reserved.
Keywords: Dorsal striatum Interneurons Cholinergic Parvalbumin Somatostatin Calretinin
1. Introduction The mammalian dorsal striatum is a major input nucleus of the basal ganglia and is primarily involved with motor and associative functions [1]. The dorsal striatum in rodents is a single nucleus called the caudate-putamen. In higher vertebrates the dorsal striatum is comprised of the caudate nucleus and the putamen, which are partitioned by the internal capsule. Within the dorsal striatum of mammals, the vast majority of neurons are considered to be medium-sized spiny projection neurons [2]. The number of interneurons in the dorsal striatum is thus generally considered to be small. Compared to the striatal spiny projection neurons, the striatal interneurons are aspiny [2]. In spite of a relatively small number of aspiny interneurons, their role(s) are likely to be functionally profound [2]. In the published literature on mammals, there are various answers to the question: What percentage of the total population
* Corresponding author at: Department of Anatomy, Otago School of Medical Sciences, University of Otago, P.O. Box 913, Dunedin, New Zealand. Tel.: +64 3 479 7379; fax: +64 3 479 7254. E-mail address: [email protected]. 2210-5336/$ – see front matter ß 2012 Elsevier GmbH. All rights reserved. http://dx.doi.org/10.1016/j.baga.2012.11.001
of neurons within the dorsal striatum is represented by interneurons? A difference in prevalence between species has been indicated by some [3–10] but not by others [11,12]. Species differences could be important, for example, in determining the validity of animal models of human pathology. The percentage of interneurons in the dorsal striatum is reported to be 26% for the human [4,13,14], 23% or 4–6% for the monkey [3,11,12], 5–6% for the rat and mouse (i.e. rodent [3]) or <3% for the rat [15], 4–5% for the cat [4,16], and 5–10% for the mammal [2,17]. The early evidence for these percentages is summarised by Roberts et al. [4] (see their Table 5), which is included in this current review as Table 1. These reports raise the question: For the percentage of interneurons in the dorsal striatum is there a species difference or not between the human, monkey, cat and rat? A critique of the evidence is discussed in the next sections. Four major types of aspiny interneurons have been identified within the normal dorsal striatum of mammals [2,10]. These are the large cholinergic interneurons, and the medium-sized GABA/ parvalbumin interneurons, GABA/calretinin interneurons, and somatostatin/neuropeptide Y/nitric oxide synthase/GABA interneurons (see Fig. 1). The cholinergic, GABA/parvalbumin, and somatostatin/neuropeptide Y/GABA interneurons innervate the
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
20
Table 1 The early evidence cited by Roberts et al. [4] on the percentage of aspiny neurons (i.e. interneurons) in the dorsal striatum of the human, monkey (Macaca fascicularis), cat and rat (Sprague-Dawley). Note that this evidence is based on non-stereological data and may be misleading for the human, monkey and cat (see the text for further details). In the Sprague-Dawley rat, the percentage of all the aspiny interneurons is half the value indicated here (i.e. <3%, see Table 2). Species
Medium spiny (%)
Medium aspiny (%)
Large aspiny (%)
Humana Monkeyb Catc Ratb
<76 76 95–96 95–96
>23 23 3–4 4–5
1 1 1 1
a b c
Data from Graveland et al. [13]. Data from Graveland and DiFiglia [3]. Data from Kemp and Powell [16].
spiny projection neurons (see reviews by Kawaguchi et al. [2] and Oorschot [18]). Whether the GABA/calretinin neurons innervate the spiny projection neurons remains to be investigated. A critique of the evidence for the percentage of each type of interneuron in the dorsal striatum of the rat, monkey and human is also discussed in the next sections. Investigation of the percentage of interneurons, and the percentage of each type of interneuron, requires quantitative data
[(Fig._1)TD$IG]
from stained sections of the dorsal striatum. The limitations of some of the methods that have been used are discussed below. 2. A critique of the early evidence for the cat, monkey and human To obtain quantitative (i.e. stereological) data from sections on the number and percentage of neurons, one must be confident that all of the cells of interest are observable (stained) and can be distinguished from other cell types [19,20]. One must also know the three-dimensional extent of the entire region that contains the neurons (i.e. the dorsal striatum in this instance). This ensures that all of the neurons can be sampled with equal probability and permits the measurement of the absolute number of the specific neurons of interest [19,20]. The use of these stereological principles will now be discussed for the relevant literature. In the cat, the percentage of 4–5% for all interneurons in the dorsal striatum is derived from Golgi stained sections [16]. A fundamental characteristic of Golgi impregnation is that it outlines only 5–10% of the neurons present in a given structure [21,22]. This limitation restricts considerably its value in regard to the quantitative evaluation of the relative proportion of various neuronal types [21,22]. Such a limitation may be even stricter since it is possible that certain classes of cells are more susceptible
Fig. 1. Light micrographs of interneurons in the dorsal striatum of the Sprague Dawley rat. (A) Somatostatin (SOM) interneurons (from Ref. [33]), (B) a calretinin (CALR) interneuron (from Ref. [36]), (C) parvalbumin (PARV) interneurons (from Ref. [35]), and (D) cholinergic (ChAT) interneurons (from Ref. [34]). In (A), the purple interneurons have been in situ hybridised for somatostatin mRNA and the tissue, including the nucleus of the hybridised interneurons, has been counterstained with neutral red. Superimposed on the section are a number of unbiased counting frames that were used to count these somatostatin interneurons using the optical disector method. In (B), a calretinin interneuron has been double-immunostained for calretinin and for bromodeoxyuridine (a nuclear label for detecting recently generated neurons during development). In (C) parvalbumin interneurons have been immunostained for parvalbumin. In (D) cholinergic interneurons (arrowheads) have been immunostained for choline acetyltransferase (ChAT). Scale bars: (A) 50 mm, (B) 25 mm, (C) 100 mm and (D) 45 mm. Used with permission.
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
to impregnation than others [21]. It is therefore of uncertain value to state, as indicated by Kemp and Powell [16], that 95–96% of striatal neurons are of the spiny type and 4–5% of striatal neurons are of the aspiny interneuronal type in the cat [21]. Graveland et al. [13] examined Golgi stained sections from the striatum of 20 adult humans. They found that neurons with spinerich dendrites were the most frequent type and suggest, from their observations, that the proportion of medium-sized aspiny neurons may be apparently greater in the human than rodents, cats and monkeys (see their Abstract and p. 326). They also comment that aspiny neurons of this type impregnate with the Golgi stain with variable frequency from case to case, but in some of the tissue sections these neurons comprise about 25% of the total (see their p. 320). These comments form the basis of >23% for medium-sized aspiny neurons in the human striatum in the summary of the early evidence by Roberts et al. [4] (see Table 1). They also form the basis for the statement by Bernacer et al. [14,23] that the human striatum contains projection neurons that represent about 74% of the total neuronal population and interneurons that account for the remaining 26%. It seems evident, however, that 26% is likely to be an upper limit and that an average across all sections analysed by Graveland et al. [13] would likely yield a lower overall percentage of interneurons. It is also evident that due to the capricious nature of Golgi staining, the percentage of 26% is of uncertain reliability. Graveland and DiFiglia [3] measured the frequency of mediumsized neurons with indented nuclei in resin-embedded 0.5 mm sections of the monkey, rat and mouse dorsal striatum. Mediumsized neurons with indented nuclei were identified as aspiny interneurons and medium-sized neurons with unindented nuclei were identified as spiny projection neurons. The proportion of striatal neurons with nuclear indentations was greater in the monkey (23%) than in the mouse and rat (4–5%). There is a potential limitation of this method of categorising cells in the primate. Neurons with nuclear indentations have been shown in the human to express specific immunohistochemical markers (i.e. substance P) of spiny projection neurons. More specifically, Hutcherson and Roberts [24] observed that most of the substance P-labelled neurons in the human striatum had an unindented nucleus typical of medium-spiny projection neurons in other species. Yet a small percentage (8%) of labelled neurons had one shallow nuclear indentation but otherwise had similar morphology (see their Fig. 2B). These findings may be relevant to the results in the monkey of Graveland and DiFiglia [3] because neurons with only one shallow nuclear indentation comprised 13% of all neurons in the caudate and 10% of all neurons in the putamen (yielding an average of 12%). Hence, based on the data of Hutcherson and Roberts [24] there is the possibility that the percentage of true aspiny interneurons in the dorsal striatum of the monkey could be lower at no more than 11% (i.e. 23%–12%).
21
Schroder et al. [25] and Lange et al. [26] measured the absolute number of medium-sized (called ‘small’) and large neurons in the human striatum of males and females. This study was completed before modern stereological methods were available (see Section 3). However, the robust stereological principles outlined at the start of this section were applied, as well as a variation of the disector method of modern stereology (see below) and correction for shrinkage [21,25–27]. The absolute number of medium-sized neurons averaged 100,000,000 for males and 105,000,000 for females. For the large striatal neurons, the averages were 670,000 (male) and 570,000 (female). Since the vast majority of large striatal neurons are likely to be cholinergic interneurons, this yields a percentage for presumptive cholinergic interneurons in the human dorsal striatum of 0.67% (male) and 0.54% (female) [25,26]. These percentages are similar to the Sprague-Dawley rat (0.4%, see text below and Table 2). Hence, stereological studies have yielded a similar result that less than 1% of all striatal neurons are likely to be cholinergic neurons in the human and SpragueDawley rat. In summary, due to the capricious nature of Golgi staining, the percentage of interneurons in the dorsal striatum in the cat of 4–5% [16] and in the human of >23% [13] (as summarised by Roberts et al. [4], see Table 1), is of uncertain reliability. Hence, caution is needed when citing these percentages. In the primate, measurement of the percentage of indented (interneuronal) nuclei of 23% in the dorsal striatum [3] (see Table 1) may be of restricted value since this definition of an interneuron may be ambiguous to some extent. Stereological studies [25,26] have yielded a percentage for presumptive cholinergic interneurons in the human dorsal striatum of 0.7% (male) and 0.5% (female). 3. A critique of more recent evidence for the rat, monkey and human Over the past 17 years, the absolute number of interneuronal subtypes with the rat striatum has been quantified using modern stereological methods. A key feature of modern stereological methods is that the methods employed (i.e. the optical disector or the disector methods) are not dependent on features of neurons that cannot be changed (like variable shape and size) [19,27–30]. Older stereological methods assume that objects (e.g. neurons) have a particular size or shape, which can be approximated using a mathematical model such as the method of Abercrombie (1946) (see Fig. 3 in Ref. [27]; [31,32]). This approach introduces a bias unless the objects exactly match the model. A decision on whether the shape and size assumptions are met is usually dependent on comparison with a modern stereological method to estimate the absolute number of neurons [27,32] (or alternatively with the very inefficient method of serially sectioning the entire structure and counting the number of neurons in every section and avoiding
Table 2 The absolute unilateral number and the percentage of each known type of neuron in the dorsal striatum of the Sprague-Dawley rat. These data indicate that <3% of neostriatal neurons in the Sprague-Dawley rat are likely to be aspiny interneurons. Neuronal type
Average absolute number
Percentage (%) to two decimal places
Rounded %
Medium-sized spiny projection neurons Aspiny somatostatin/GABA interneurons* Aspiny cholinergic interneurons* Aspiny GABA/parvalbumin interneurons* Aspiny GABA/calretinin interneurons* * Total for aspiny interneurons
2,791,000a 21,300b 12,200c 16,900d 13,200e 63,600
97.77 0.75 0.43 0.59d 0.46 2.23
97.8 0.8 0.4 0.6d 0.5 2.2–2.3
a b c d e *
Data from Oorschot [37]. Data from West et al. [33]. Data from Oorschot et al. [34]. Data from Luk and Sadikot [35]. See text for further explanation of this percentage. Data from Rymar et al. [36]. This average absolute number contributes to the total average absolute number for all aspiny interneurons that is indicated in this Table.
22
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
double counting, see [27] for further details). The measurement of absolute number using modern stereological methods also overcomes problems that can arise from using density measures, like number per unit area or number per unit volume, as the definitive data ([19,27–30], see also further discussion below). In summary, the use of modern stereological methods, and the measurement of absolute number, minimises any biases that can arise either from counting cells (i.e. neurons) in sections of the striatum or through the use of density measures [19,21,25,27–30]. To identify the neuronal subtypes, immunohistochemical markers to specific proteins (e.g. parvalbumin) or in situ hybridisation to a specific mRNA (i.e. somatostatin) has been used (Fig. 1). These modern stereological studies have revealed (Table 2) that there are, unilaterally, 21,300 somatostatin/ neuropeptide Y/GABA interneurons [33], 12,200 cholinergic interneurons [34], 16,900 GABA/parvalbumin interneurons [35] and 13,200 GABA/calretinin interneurons in the Sprague-Dawley rat [36]. As far as I am aware, these are the only studies that have been published for this rat strain. It needs to be acknowledged, however, that the vast majority, but not all, of the dorsal striatum was analysed for the estimate of GABA/parvalbumin interneurons ([35], i.e. from 2.2 mm anterior to bregma to 2.6 mm posterior to bregma; see their Fig. 1 and further text below). The entire dorsal striatum was analysed for the other studies [33,34,36]. Hence, the absolute number of GABA/parvalbumin interneurons may be an underestimate. When compared to the absolute unilateral number of mediumsized striatal projection neurons of 2,791,000 in the SpragueDawley rat [37], these data indicate that less than 3% (i.e. 2.2–2.3%) of striatal neurons are interneurons in the Sprague-Dawley rat [7,15,36,38] (Table 2). More specifically, these stereological data indicate that, in this rat strain, 0.8% (0.75%, to two decimal places) of all neurons are somatostatin/neuropeptide Y/GABA interneurons, 0.4% (0.43%) are cholinergic interneurons, approximately 0.6% (0.59%) are GABA/parvalbumin interneurons and 0.5% (0.46%) are GABA/calretinin interneurons [15] (Table 2). These data also indicate that approximately 1.80% (to two decimal places) are medium-sized (i.e. non-cholinergic) interneurons in the rat dorsal striatum (Table 2). Importantly, these data indicate that the percentage of medium-sized aspiny interneurons in the dorsal rat striatum is even lower than the finding of 4–5% for medium-sized aspiny neurons in the Sprague-Dawley rat by Graveland and DiFiglia [3]. The data in Table 2 are from young adult Sprague-Dawley rats. Either males [35–37], females [33] or both sexes [34] were investigated. Using the modern optical disector and fractionator methods, the number of unilateral striatal interneurons for young adult male Wistar rats is reported by Larsson et al. [39] to be 14,355 neuropeptide Y (i.e. somatostatin) interneurons, 6803 cholinergic interneurons and 16,597 parvalbumin interneurons. However, these data are likely to be underestimates because only the dorsal striatum between 1.7 mm anterior to bregma and 1.4 mm posterior to bregma was analysed. Based on Paxinos and Watson [40], the dorsal striatum starts at 2.2 mm anterior to bregma and tapers to 3.8 mm posterior to bregma. As far as I am aware, this is the only relevant study that has been published for this rat strain. Hence, additional robust stereological studies are now needed in rats to confirm (i.e. replicate) or refute the current data and to ascertain whether there are similarities or differences between strains, genders, and ages. The absolute numbers and percentages indicated in Table 2 may be minimum levels due to variability in the intensity or sensitivity of immunohistochemical staining [9]. This seems unlikely for the dorsal striatum of the normal rat, since the intensity of immunostaining in each labelled interneuron is invariably strong and unambiguous when appropriate and consistent fixation is
used. In injured or diseased tissue from the rat or other species, the measured number may be a minimum or maximum number since the injury or disease could affect the expression of the specific protein of interest. Kawaguchi et al. [2], in Trends in Neurosciences, cite the work of Kita and co-workers [41,42] as evidence that 3–5% of striatal interneurons are parvalbumin interneurons in the male, young adult Sprague-Dawley rat. Kita and Kitai [41] report that 3–5% of rat striatal neurons stain intensely for the synthetic enzyme of GABA, glutamate decarboxylase (GAD). Their cell counts were taken only from the striatum located proximal to the lateral ventricle and from four 1 mm2 areas on the surface of sections per animal. Hence their non-stereological count may not be representative of the entire striatum nor the section depth. Kita et al. [42] observed that most striatal neurons immunoreactive for parvalbumin were strongly immunoreactive for GAD, and vice versa, and thus inferred from the data of Kita and Kitai [41] that 3–5% of rat striatal neurons are immunoreactive for parvalbumin. However, Kita et al. [42] also report that the frequency of parvalbumin interneurons is strongly dependent on their location within the entire striatum. Hence, the sampling of a specific location within the striatum for the cell counts [41] may have yielded the higher, non-stereological percentage of 3–5%. Parent et al. [43] cite the work of Celio [22] as evidence that the percentage of parvalbumin-immunoreactive interneurons in the rat dorsal striatum is 2%. The specific rat strain that was used for this part of the study is not specified by Celio [22], with four strains (i.e. Wistar, ZUR-Siv, Sprague-Dawley, and Long-Evans) used from both genders in this 100-page paper. Celio counted the number of parvalbumin-positive neurons in every 2nd 50 mm vibratome section of a whole sagittal series through the dorsal striatum. The parvalbumin-positive neurons ‘account for less than 2% of striatal neurons as inferred from counting the total number of neurons in adjacent cresyl-violet stained sections’ [22]. The absolute number of parvalbumin-immunoreactive interneurons in the dorsal striatum was not stated by Celio [22]. Less than 2% is closer to the finding of approximately 0.6% [15,35] discussed above (see Table 2). Wu and Parent [6] compared the density of calretinin, parvalbumin and somatostatin (i.e. NADPH-diaphorase) positive neurons in the dorsal striatum of the rat (i.e. female young adults of the Sprague-Dawley strain), squirrel monkey, and human. They report that the density of these three striatal interneurons was much higher in the monkey than in the rat and human (see also Cicchetti et al. [44]). Unfortunately, comparative neuronal density measures across species are inconclusive when they are not corrected for processing-induced shrinkage [19,21,25,27–30]. For example, a higher apparent neuronal density per mm2 could occur because the tissue of one species shrinks relatively more during processing through the same dehydration schedule. This may occur because the striatum in a particular species has a different tissue composition with respect to the ratio of white matter to grey matter. The white matter tract, the internal capsule, is located within the dorsal striatum in the rat but is separate from the caudate nucleus and putamen in the monkey and human. This known species difference in tissue composition within the dorsal striatum may contribute to the report [6] of a lower density of interneurons in the rat. A lower density in the human compared to the monkey [6] could also be due to the effects of postmortem delay on the intensity of immunostaining. A slightly higher density in the human compared to the rat may have occurred because 40 mm sections were used for the rat (and monkey) and 50 mm sections for the human. Due to these methodological limitations, the comparative density measures of Wu and Parent [6] may be misleading.
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
Deng et al. [12] used immunolabelling of the various neuronal types to define the frequency of each type in the caudate and putamen of the rhesus monkey. Counts of NeuN+ neurons were used to determine the total number of striatal neurons per unit area. The number of projection neurons and interneuronal subtypes were also counted per unit area and compared with the total number of NeuN+ neurons per unit area to calculate the frequency as a percent of all neurons. Implicit to this approach is that the assumption that the immunostained sections for each respective antibody (i.e. NeuN, calbindin, choline acetyltransferase, parvalbumin, calretinin, somatostatin) responded very similarly to the immunostaining processing (e.g. dehydration), such that the degree of shrinkage or swelling of the striatum is equivalent across all sections and antibodies used. This assumption is required in the interpretation of the data for cell number per unit area [27]. Since the Abercrombie method [31] was used, an assumption is also required that the shape and size of the neurons matched this mathematical model. For the caudate nucleus, Deng et al. [12] observed that 95% of the neurons were calbindin-positive and likely to be spiny projection neurons. Since less than 1% of calbindin neurons are also somatostatin-positive in the human, it was estimated that 94% of neurons in the caudate nucleus of the monkey are spiny projection neurons [12]. For the interneurons, 1.1% were cholinergic interneurons, 1% were parvalbumin interneurons, 1.9% were somatostatin interneurons and 2% were calretinin interneurons [12]. For the putamen, it was reported that 0.9% of the neurons were cholinergic interneurons, 2.4% were parvalbumin interneurons, 1.2% were somatostatin interneurons and 1.5% were calretinin interneurons [12]. More lightly labelled calbindin neurons, and calbindin-negative patches, were observed in the putamen which yielded a lower frequency of 83% for calbindin-positive neurons. This is likely to under-represent the percentage of spiny projection neurons in the putamen, such that the percentage of these neurons in the caudate nucleus was used for the putamen [12]. Of note, since calretinin is expressed by a subset of projection neurons and large interneurons, and by medium-sized interneurons, in the monkey striatum, the percentage of medium-sized calretinin interneurons was obtained from the percentage of spiny projection neurons (of 94%) and the percentage of other interneurons (of 4– 5%) to yield 1–2% are medium-sized calretinin interneurons. In summary, the total percentage of all interneurons in the dorsal striatum of the monkey was 6% for the caudate nucleus and 6% for the putamen [12]. To confirm or refute these results, a modern stereological study is now needed in which the absolute number of all neuronal subtypes in the dorsal striatum of the monkey is measured. 4. Conclusions It is evident that caution is needed when citing historical nonstereological data on the percentage of interneurons in the dorsal striatum of the cat, monkey and human. In these species, it is evident that the vast majority of neurons in the dorsal striatum are spiny projection neurons and a smaller number are interneurons, yet the precise percentage is currently unknown. Hence, in the absence of stereological data on the absolute number of all neuronal subtypes for the cat, monkey and human, it is currently unknown if there is a species difference, or not, for the percentage of interneurons in the dorsal striatum of the human and monkey versus the cat and rat. Future research is required to measure the absolute number of each interneuronal subtype in the dorsal striatum of the cat, monkey and human using specific immunohistochemical and histochemical markers and modern stereological methods.
23
Modern stereological data indicate that the percentage of interneurons in the rat dorsal striatum is <3%. More specifically, these data indicate that, in the Sprague-Dawley rat, 0.8% of all neurons are somatostatin/neuropeptide Y/GABA interneurons, 0.4% are cholinergic interneurons, approximately 0.6% are GABA/ parvalbumin interneurons and 0.5% are GABA/calretinin interneurons (i.e. a total of 2.2–2.3%, see Table 2). Current stereological data also indicates that the percentage of interneurons in the dorsal striatum is <1% for cholinergic interneurons in the SpragueDawley rat and human. Acknowledgement The editorial assistance of Mrs. Liping Goddard is gratefully acknowledged. References [1] Tepper JM, Abercrombie ED, Bolam JP. Basal ganglia macrocircuits. Progress in Brain Research 2007;160:3–7. [2] Kawaguchi Y, Wilson CJ, Augood SJ, Emson PC. Striatal interneurons: chemical, physiological and morphological characterization. Trends in Neurosciences 1995;18:527–35. [3] Graveland GA, DiFiglia M. The frequency and distribution of medium-sized neurons with indented nuclei in the primate and rodent neostriatum. Brain Research 1985;327:307–11. [4] Roberts RC, Gaither LA, Peretti FJ, Lapidus B, Chute DJ. Synaptic organisation of the human striatum: a postmortem ultrastructural study. Journal of Comparative Neurology 1996;374:523–34. [5] Desjardins C, Parent A. Distribution of somatostatin immunoreactivity in the forebrain of the squirrel monkey: basal ganglia and amygdala. Neuroscience 1992;47:115–33. [6] Wu Y, Parent A. Striatal interneurons expressing calretinin, parvalbumin or NADPH-diaphorase: a comparative study in the rat, monkey and human. Brain Research 2000;863:182–91. [7] Tepper JM, Bolam JP. Functional diversity and specificity of neostriatal interneurons. Current Opinion in Neurobiology 2004;14:685–92. [8] Kreitzer AC. Physiology and pharmacology of striatal neurons. Annual Review of Neuroscience 2009;32:127–47. [9] Tepper JM, Tecuapetla F, Koos T, Ibanez-Sanodval O. Heterogeneity and diversity of striatal GABAergic interneurons. Frontiers in Neuroanatomy 2010;4:150. http://dx.doi.org/10.3389/fnana.2010.00150. [10] Rice MW, Roberts RC, Melendez-Ferro M, Perez-Costas E. Neurochemical characterization of the tree shrew dorsal striatum. Frontiers in Neuroanatomy 2011;5:53. http://dx.doi.org/10.3389/fnana.2011.00053. [11] Groves PM. A theory of the functional organization of the neostriatum and the neostriatal control of voluntary movement. Brain Research Reviews 1983;5:109–32. [12] Deng Y-P, Shelby E, Reiner A. Immunohistochemical localization of AMPA-type glutamate receptor units in the striatum of rhesus monkey. Brain Research 2010;1344:104–23. [13] Graveland GA, Williams RS, DiFiglia M. A Golgi study of the human neostriatum: neurons and afferent fibers. Journal of Comparative Neurology 1985;234: 317–33. [14] Bernacer J, Prensa L, Gimenez-Amaya JM. Distribution of GABAergic interneurons and dopaminergic cells in the functional territories of the human striatum. PLoS One 2012;7:e30504. [15] Oorschot DE. Cell types in the different nuclei of the basal ganglia. In: Steiner H, Tseng KY, editors. Handbook of basal ganglia structure and function. London: Academic Press/Elsevier; 2010. p. 63–74. [16] Kemp JM, Powell TPS. The structure of the caudate nucleus of the cat: light and electron microscopy. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 1971;262:383–401. [17] Thomas TM, Smith Y, Levey AI, Hersch SM. Cortical inputs to m2-immunoreactive striatal interneurons in rat and monkey. Synapse 2000;37: 252–61. [18] Oorschot DE. The domain hypothesis: a central organising principle for understanding neostriatal circuitry? In: Miller R, Wickens JR, editors. Conceptual advances in brain research, brain dynamics and the striatal complex. Reading, UK: Gordon and Breach; 2000. p. 151–63. [19] Gundersen HJG. Stereology of arbitrary particles. A review of unbiased number and size estimators and the presentation of some new ones, in memory of William R. Thompson. Journal of Microscopy 1986;143:3–45. [20] West MJ. New stereological methods for counting neurons. Neurobiology of Aging 1993;14:275–85. [21] Pasik P, Pasik T, DiFiglia M. The internal organization of the neostriatum in mammals. In: Divac I, Gunilla R, Oberg E, editors. The neostriatum. Oxford: Pergamon; 1979. p. 5–36. [22] Celio MR. Calbindin D-28k and parvalbumin in the rat nervous system. Neuroscience 1990;35:375–475.
24
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
[23] Bernacer J, Prensa L, Gimenez-Amaya JM. Cholinergic interneurons are differentially distributed in the human striatum. PLoS One 2007;2:e1174. [24] Hutcherson L, Roberts RC. The immunocytochemical localization of substance P in the human striatum: a postmortem ultrastructural study. Synapse 2005; 57:191–201. [25] Schroder KF, Hopf A, Lange H, Thorner G. Morphometrical-statistical structure analysis of human striatum, pallidum and subthalamic nucleus. Journal fur Hirnforschung 1975;16:333–50. [26] Lange H, Thorner G, Hopf A, Schroder KF. Morphometric studies of the neuropathological changes in choreatic diseases. Journal of Neurological Sciences 1976;28:401–25. [27] Oorschot DE. Are you using neuronal densities, synaptic densities or neurochemical densities as your definitive data? There is a better way to go. Progress in Neurobiology 1994;44:233–47. [28] Swaab DF, Uylings HBM. Density measures: parameters to avoid. Neurobiology of Aging 1987;8:574–6. [29] West MJ. Stereological methods for estimating the total number of neurons and synapses: issues of precision and bias. Trends in Neurosciences 1999;22: 51–61. [30] Boyce RW, Dorph-Petersen K-A, Lyck L, Gundersen HJG. Design-based stereology: introduction to basic concepts and practical approaches for estimation of cell number. Toxicologic Pathology 2010;38:1011–25. [31] Abercrombie M. Estimation of nuclear populations from microtome sections. Anatomical Record 1946;94:239–47. [32] Raadsheer FC, Oorschot DE, Verwer RWH, Tilders FJH, Swaab DF. Age-related increase in the total number of corticotropin-releasing hormone neurons in the human paraventricular nucleus in controls and Alzheimer’s Disease: comparison of the disector with an unfolding method. Journal of Comparative Neurology 1994;339:447–57. [33] West MJ, Østergaard K, Andreassen OA, Finsen B. Estimation of the number of somatostatin neurons in the striatum: an in situ hybridization study using the optical fractionator method. Journal of Comparative Neurology 1996; 370:11–22.
[34] Oorschot DE, Zhang R, Wickens JR. Absolute number and three-dimensional spatial distribution of rat neostriatal large interneurons: a first and second order stereological study. Proceedings of the Xth International Congress of Stereology 1999;87. [35] Luk KC, Sadikot AF. GABA promotes survival but not proliferation of parvalbumin-immunoreactive neurons in rodent neostriatum: an in vivo study with stereology. Neuroscience 2001;104:93–103. [36] Rymar VV, Sasseville R, Luk KC, Sadikot AF. Neurogenesis and stereological morphometry of calretinin-immunoreactive GABAergic interneurons of the neostriatum. Journal of Comparative Neurology 2004;469:325–39. [37] Oorschot DE. Total number of neurons in the neostriatal, pallidal, subthalamic, and substantia nigral nuclei of the rat basal ganglia: a stereological study using the Cavalieri and optical disector methods. Journal of Comparative Neurology 1996;366:580–99. [38] Oorschot DE, Tunstall MJ, Wickens JR. Local connectivity between striatal spiny projection neurons: a re-evaluation. In: Nicholson L, Faull RLM, editors. The basal ganglia VII. New York: Plenum Press; 2002. p. 421–34. [39] Larsson E, Lindvall O, Kokaia Z. Stereological assessment of vulnerability of immunocytochemically identified striatal and hippocampal neurons after global cerebral ischemia in rats. Brain Research 2001;913:117–32. [40] Paxinos G, Watson C. The rat brain in stereotaxic coordinates, 4th ed., San Diego: Academic Press; 1998. [41] Kita H, Kitai ST. Glutamate decarboxylase immunoreactive neurons in rat neostriatum: their morphological types and populations. Brain Research 1988;447:346–52. [42] Kita H, Kosaka T, Heizmann CW. Parvalbumin-immunoreactive neurons in the rat neostriatum: a light and electron microscopic study. Brain Research 1990;536:1–15. [43] Parent A, Fortin M, Cote P-Y, Cicchetti F. Calcium-binding proteins in primate basal ganglia. Neuroscience Research 1996;25:309–34. [44] Cicchetti F, Prensa L, Wu Y, Parent A. Chemical anatomy of striatal interneurons in normal individuals and in patients with Huntington’s disease. Brain Research Reviews 2000;34:80–101.
Contents lists available at SciVerse ScienceDirect
Basal Ganglia journal homepage: www.elsevier.com/locate/baga
The percentage of interneurons in the dorsal striatum of the rat, cat, monkey and human: A critique of the evidence Dorothy E. Oorschot * Department of Anatomy, Otago School of Medical Sciences, and the Brain Health Research Centre, University of Otago, Dunedin, New Zealand
A R T I C L E I N F O
A B S T R A C T
Article history: Received 1 October 2012 Accepted 7 November 2012
In the published literature on mammals, there are various answers to the question: What percentage of the total population of neurons within the dorsal striatum is represented by interneurons? The percentage of interneurons is reported to be 26% for the human, 23% or 4–6% for the monkey, 5–6% or <3% for the rodent, 4–5% for the cat and 5–10% for the mammal. Based on a critique of the literature, including the use of stereological approaches, caution is needed when citing historical non-stereological data on the percentage of interneurons in the dorsal striatum of the cat, monkey and human. In these species, the vast majority of neurons in the dorsal striatum are spiny projection neurons and a smaller number are interneurons, yet the precise percentage is currently unknown. Hence, in the absence of stereological data on the absolute number of all neuronal subtypes for the cat, monkey and human, it is currently unknown if there is a species difference, or not, for the percentage of interneurons in the dorsal striatum of the human and monkey versus the cat and rat. Modern stereological data indicate that the percentage of interneurons in the rat dorsal striatum is <3%. More specifically, in the Sprague-Dawley rat, 0.8% of all neurons are somatostatin/neuropeptide Y/GABA interneurons, 0.4% are cholinergic interneurons, 0.6% are GABA/parvalbumin interneurons and 0.5% are GABA/calretinin interneurons. Current stereological data also indicate that the percentage of interneurons in the dorsal striatum is <1% for cholinergic interneurons in the Sprague-Dawley rat and human. ß 2012 Elsevier GmbH. All rights reserved.
Keywords: Dorsal striatum Interneurons Cholinergic Parvalbumin Somatostatin Calretinin
1. Introduction The mammalian dorsal striatum is a major input nucleus of the basal ganglia and is primarily involved with motor and associative functions [1]. The dorsal striatum in rodents is a single nucleus called the caudate-putamen. In higher vertebrates the dorsal striatum is comprised of the caudate nucleus and the putamen, which are partitioned by the internal capsule. Within the dorsal striatum of mammals, the vast majority of neurons are considered to be medium-sized spiny projection neurons [2]. The number of interneurons in the dorsal striatum is thus generally considered to be small. Compared to the striatal spiny projection neurons, the striatal interneurons are aspiny [2]. In spite of a relatively small number of aspiny interneurons, their role(s) are likely to be functionally profound [2]. In the published literature on mammals, there are various answers to the question: What percentage of the total population
* Corresponding author at: Department of Anatomy, Otago School of Medical Sciences, University of Otago, P.O. Box 913, Dunedin, New Zealand. Tel.: +64 3 479 7379; fax: +64 3 479 7254. E-mail address: [email protected]. 2210-5336/$ – see front matter ß 2012 Elsevier GmbH. All rights reserved. http://dx.doi.org/10.1016/j.baga.2012.11.001
of neurons within the dorsal striatum is represented by interneurons? A difference in prevalence between species has been indicated by some [3–10] but not by others [11,12]. Species differences could be important, for example, in determining the validity of animal models of human pathology. The percentage of interneurons in the dorsal striatum is reported to be 26% for the human [4,13,14], 23% or 4–6% for the monkey [3,11,12], 5–6% for the rat and mouse (i.e. rodent [3]) or <3% for the rat [15], 4–5% for the cat [4,16], and 5–10% for the mammal [2,17]. The early evidence for these percentages is summarised by Roberts et al. [4] (see their Table 5), which is included in this current review as Table 1. These reports raise the question: For the percentage of interneurons in the dorsal striatum is there a species difference or not between the human, monkey, cat and rat? A critique of the evidence is discussed in the next sections. Four major types of aspiny interneurons have been identified within the normal dorsal striatum of mammals [2,10]. These are the large cholinergic interneurons, and the medium-sized GABA/ parvalbumin interneurons, GABA/calretinin interneurons, and somatostatin/neuropeptide Y/nitric oxide synthase/GABA interneurons (see Fig. 1). The cholinergic, GABA/parvalbumin, and somatostatin/neuropeptide Y/GABA interneurons innervate the
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
20
Table 1 The early evidence cited by Roberts et al. [4] on the percentage of aspiny neurons (i.e. interneurons) in the dorsal striatum of the human, monkey (Macaca fascicularis), cat and rat (Sprague-Dawley). Note that this evidence is based on non-stereological data and may be misleading for the human, monkey and cat (see the text for further details). In the Sprague-Dawley rat, the percentage of all the aspiny interneurons is half the value indicated here (i.e. <3%, see Table 2). Species
Medium spiny (%)
Medium aspiny (%)
Large aspiny (%)
Humana Monkeyb Catc Ratb
<76 76 95–96 95–96
>23 23 3–4 4–5
1 1 1 1
a b c
Data from Graveland et al. [13]. Data from Graveland and DiFiglia [3]. Data from Kemp and Powell [16].
spiny projection neurons (see reviews by Kawaguchi et al. [2] and Oorschot [18]). Whether the GABA/calretinin neurons innervate the spiny projection neurons remains to be investigated. A critique of the evidence for the percentage of each type of interneuron in the dorsal striatum of the rat, monkey and human is also discussed in the next sections. Investigation of the percentage of interneurons, and the percentage of each type of interneuron, requires quantitative data
[(Fig._1)TD$IG]
from stained sections of the dorsal striatum. The limitations of some of the methods that have been used are discussed below. 2. A critique of the early evidence for the cat, monkey and human To obtain quantitative (i.e. stereological) data from sections on the number and percentage of neurons, one must be confident that all of the cells of interest are observable (stained) and can be distinguished from other cell types [19,20]. One must also know the three-dimensional extent of the entire region that contains the neurons (i.e. the dorsal striatum in this instance). This ensures that all of the neurons can be sampled with equal probability and permits the measurement of the absolute number of the specific neurons of interest [19,20]. The use of these stereological principles will now be discussed for the relevant literature. In the cat, the percentage of 4–5% for all interneurons in the dorsal striatum is derived from Golgi stained sections [16]. A fundamental characteristic of Golgi impregnation is that it outlines only 5–10% of the neurons present in a given structure [21,22]. This limitation restricts considerably its value in regard to the quantitative evaluation of the relative proportion of various neuronal types [21,22]. Such a limitation may be even stricter since it is possible that certain classes of cells are more susceptible
Fig. 1. Light micrographs of interneurons in the dorsal striatum of the Sprague Dawley rat. (A) Somatostatin (SOM) interneurons (from Ref. [33]), (B) a calretinin (CALR) interneuron (from Ref. [36]), (C) parvalbumin (PARV) interneurons (from Ref. [35]), and (D) cholinergic (ChAT) interneurons (from Ref. [34]). In (A), the purple interneurons have been in situ hybridised for somatostatin mRNA and the tissue, including the nucleus of the hybridised interneurons, has been counterstained with neutral red. Superimposed on the section are a number of unbiased counting frames that were used to count these somatostatin interneurons using the optical disector method. In (B), a calretinin interneuron has been double-immunostained for calretinin and for bromodeoxyuridine (a nuclear label for detecting recently generated neurons during development). In (C) parvalbumin interneurons have been immunostained for parvalbumin. In (D) cholinergic interneurons (arrowheads) have been immunostained for choline acetyltransferase (ChAT). Scale bars: (A) 50 mm, (B) 25 mm, (C) 100 mm and (D) 45 mm. Used with permission.
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
to impregnation than others [21]. It is therefore of uncertain value to state, as indicated by Kemp and Powell [16], that 95–96% of striatal neurons are of the spiny type and 4–5% of striatal neurons are of the aspiny interneuronal type in the cat [21]. Graveland et al. [13] examined Golgi stained sections from the striatum of 20 adult humans. They found that neurons with spinerich dendrites were the most frequent type and suggest, from their observations, that the proportion of medium-sized aspiny neurons may be apparently greater in the human than rodents, cats and monkeys (see their Abstract and p. 326). They also comment that aspiny neurons of this type impregnate with the Golgi stain with variable frequency from case to case, but in some of the tissue sections these neurons comprise about 25% of the total (see their p. 320). These comments form the basis of >23% for medium-sized aspiny neurons in the human striatum in the summary of the early evidence by Roberts et al. [4] (see Table 1). They also form the basis for the statement by Bernacer et al. [14,23] that the human striatum contains projection neurons that represent about 74% of the total neuronal population and interneurons that account for the remaining 26%. It seems evident, however, that 26% is likely to be an upper limit and that an average across all sections analysed by Graveland et al. [13] would likely yield a lower overall percentage of interneurons. It is also evident that due to the capricious nature of Golgi staining, the percentage of 26% is of uncertain reliability. Graveland and DiFiglia [3] measured the frequency of mediumsized neurons with indented nuclei in resin-embedded 0.5 mm sections of the monkey, rat and mouse dorsal striatum. Mediumsized neurons with indented nuclei were identified as aspiny interneurons and medium-sized neurons with unindented nuclei were identified as spiny projection neurons. The proportion of striatal neurons with nuclear indentations was greater in the monkey (23%) than in the mouse and rat (4–5%). There is a potential limitation of this method of categorising cells in the primate. Neurons with nuclear indentations have been shown in the human to express specific immunohistochemical markers (i.e. substance P) of spiny projection neurons. More specifically, Hutcherson and Roberts [24] observed that most of the substance P-labelled neurons in the human striatum had an unindented nucleus typical of medium-spiny projection neurons in other species. Yet a small percentage (8%) of labelled neurons had one shallow nuclear indentation but otherwise had similar morphology (see their Fig. 2B). These findings may be relevant to the results in the monkey of Graveland and DiFiglia [3] because neurons with only one shallow nuclear indentation comprised 13% of all neurons in the caudate and 10% of all neurons in the putamen (yielding an average of 12%). Hence, based on the data of Hutcherson and Roberts [24] there is the possibility that the percentage of true aspiny interneurons in the dorsal striatum of the monkey could be lower at no more than 11% (i.e. 23%–12%).
21
Schroder et al. [25] and Lange et al. [26] measured the absolute number of medium-sized (called ‘small’) and large neurons in the human striatum of males and females. This study was completed before modern stereological methods were available (see Section 3). However, the robust stereological principles outlined at the start of this section were applied, as well as a variation of the disector method of modern stereology (see below) and correction for shrinkage [21,25–27]. The absolute number of medium-sized neurons averaged 100,000,000 for males and 105,000,000 for females. For the large striatal neurons, the averages were 670,000 (male) and 570,000 (female). Since the vast majority of large striatal neurons are likely to be cholinergic interneurons, this yields a percentage for presumptive cholinergic interneurons in the human dorsal striatum of 0.67% (male) and 0.54% (female) [25,26]. These percentages are similar to the Sprague-Dawley rat (0.4%, see text below and Table 2). Hence, stereological studies have yielded a similar result that less than 1% of all striatal neurons are likely to be cholinergic neurons in the human and SpragueDawley rat. In summary, due to the capricious nature of Golgi staining, the percentage of interneurons in the dorsal striatum in the cat of 4–5% [16] and in the human of >23% [13] (as summarised by Roberts et al. [4], see Table 1), is of uncertain reliability. Hence, caution is needed when citing these percentages. In the primate, measurement of the percentage of indented (interneuronal) nuclei of 23% in the dorsal striatum [3] (see Table 1) may be of restricted value since this definition of an interneuron may be ambiguous to some extent. Stereological studies [25,26] have yielded a percentage for presumptive cholinergic interneurons in the human dorsal striatum of 0.7% (male) and 0.5% (female). 3. A critique of more recent evidence for the rat, monkey and human Over the past 17 years, the absolute number of interneuronal subtypes with the rat striatum has been quantified using modern stereological methods. A key feature of modern stereological methods is that the methods employed (i.e. the optical disector or the disector methods) are not dependent on features of neurons that cannot be changed (like variable shape and size) [19,27–30]. Older stereological methods assume that objects (e.g. neurons) have a particular size or shape, which can be approximated using a mathematical model such as the method of Abercrombie (1946) (see Fig. 3 in Ref. [27]; [31,32]). This approach introduces a bias unless the objects exactly match the model. A decision on whether the shape and size assumptions are met is usually dependent on comparison with a modern stereological method to estimate the absolute number of neurons [27,32] (or alternatively with the very inefficient method of serially sectioning the entire structure and counting the number of neurons in every section and avoiding
Table 2 The absolute unilateral number and the percentage of each known type of neuron in the dorsal striatum of the Sprague-Dawley rat. These data indicate that <3% of neostriatal neurons in the Sprague-Dawley rat are likely to be aspiny interneurons. Neuronal type
Average absolute number
Percentage (%) to two decimal places
Rounded %
Medium-sized spiny projection neurons Aspiny somatostatin/GABA interneurons* Aspiny cholinergic interneurons* Aspiny GABA/parvalbumin interneurons* Aspiny GABA/calretinin interneurons* * Total for aspiny interneurons
2,791,000a 21,300b 12,200c 16,900d 13,200e 63,600
97.77 0.75 0.43 0.59d 0.46 2.23
97.8 0.8 0.4 0.6d 0.5 2.2–2.3
a b c d e *
Data from Oorschot [37]. Data from West et al. [33]. Data from Oorschot et al. [34]. Data from Luk and Sadikot [35]. See text for further explanation of this percentage. Data from Rymar et al. [36]. This average absolute number contributes to the total average absolute number for all aspiny interneurons that is indicated in this Table.
22
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
double counting, see [27] for further details). The measurement of absolute number using modern stereological methods also overcomes problems that can arise from using density measures, like number per unit area or number per unit volume, as the definitive data ([19,27–30], see also further discussion below). In summary, the use of modern stereological methods, and the measurement of absolute number, minimises any biases that can arise either from counting cells (i.e. neurons) in sections of the striatum or through the use of density measures [19,21,25,27–30]. To identify the neuronal subtypes, immunohistochemical markers to specific proteins (e.g. parvalbumin) or in situ hybridisation to a specific mRNA (i.e. somatostatin) has been used (Fig. 1). These modern stereological studies have revealed (Table 2) that there are, unilaterally, 21,300 somatostatin/ neuropeptide Y/GABA interneurons [33], 12,200 cholinergic interneurons [34], 16,900 GABA/parvalbumin interneurons [35] and 13,200 GABA/calretinin interneurons in the Sprague-Dawley rat [36]. As far as I am aware, these are the only studies that have been published for this rat strain. It needs to be acknowledged, however, that the vast majority, but not all, of the dorsal striatum was analysed for the estimate of GABA/parvalbumin interneurons ([35], i.e. from 2.2 mm anterior to bregma to 2.6 mm posterior to bregma; see their Fig. 1 and further text below). The entire dorsal striatum was analysed for the other studies [33,34,36]. Hence, the absolute number of GABA/parvalbumin interneurons may be an underestimate. When compared to the absolute unilateral number of mediumsized striatal projection neurons of 2,791,000 in the SpragueDawley rat [37], these data indicate that less than 3% (i.e. 2.2–2.3%) of striatal neurons are interneurons in the Sprague-Dawley rat [7,15,36,38] (Table 2). More specifically, these stereological data indicate that, in this rat strain, 0.8% (0.75%, to two decimal places) of all neurons are somatostatin/neuropeptide Y/GABA interneurons, 0.4% (0.43%) are cholinergic interneurons, approximately 0.6% (0.59%) are GABA/parvalbumin interneurons and 0.5% (0.46%) are GABA/calretinin interneurons [15] (Table 2). These data also indicate that approximately 1.80% (to two decimal places) are medium-sized (i.e. non-cholinergic) interneurons in the rat dorsal striatum (Table 2). Importantly, these data indicate that the percentage of medium-sized aspiny interneurons in the dorsal rat striatum is even lower than the finding of 4–5% for medium-sized aspiny neurons in the Sprague-Dawley rat by Graveland and DiFiglia [3]. The data in Table 2 are from young adult Sprague-Dawley rats. Either males [35–37], females [33] or both sexes [34] were investigated. Using the modern optical disector and fractionator methods, the number of unilateral striatal interneurons for young adult male Wistar rats is reported by Larsson et al. [39] to be 14,355 neuropeptide Y (i.e. somatostatin) interneurons, 6803 cholinergic interneurons and 16,597 parvalbumin interneurons. However, these data are likely to be underestimates because only the dorsal striatum between 1.7 mm anterior to bregma and 1.4 mm posterior to bregma was analysed. Based on Paxinos and Watson [40], the dorsal striatum starts at 2.2 mm anterior to bregma and tapers to 3.8 mm posterior to bregma. As far as I am aware, this is the only relevant study that has been published for this rat strain. Hence, additional robust stereological studies are now needed in rats to confirm (i.e. replicate) or refute the current data and to ascertain whether there are similarities or differences between strains, genders, and ages. The absolute numbers and percentages indicated in Table 2 may be minimum levels due to variability in the intensity or sensitivity of immunohistochemical staining [9]. This seems unlikely for the dorsal striatum of the normal rat, since the intensity of immunostaining in each labelled interneuron is invariably strong and unambiguous when appropriate and consistent fixation is
used. In injured or diseased tissue from the rat or other species, the measured number may be a minimum or maximum number since the injury or disease could affect the expression of the specific protein of interest. Kawaguchi et al. [2], in Trends in Neurosciences, cite the work of Kita and co-workers [41,42] as evidence that 3–5% of striatal interneurons are parvalbumin interneurons in the male, young adult Sprague-Dawley rat. Kita and Kitai [41] report that 3–5% of rat striatal neurons stain intensely for the synthetic enzyme of GABA, glutamate decarboxylase (GAD). Their cell counts were taken only from the striatum located proximal to the lateral ventricle and from four 1 mm2 areas on the surface of sections per animal. Hence their non-stereological count may not be representative of the entire striatum nor the section depth. Kita et al. [42] observed that most striatal neurons immunoreactive for parvalbumin were strongly immunoreactive for GAD, and vice versa, and thus inferred from the data of Kita and Kitai [41] that 3–5% of rat striatal neurons are immunoreactive for parvalbumin. However, Kita et al. [42] also report that the frequency of parvalbumin interneurons is strongly dependent on their location within the entire striatum. Hence, the sampling of a specific location within the striatum for the cell counts [41] may have yielded the higher, non-stereological percentage of 3–5%. Parent et al. [43] cite the work of Celio [22] as evidence that the percentage of parvalbumin-immunoreactive interneurons in the rat dorsal striatum is 2%. The specific rat strain that was used for this part of the study is not specified by Celio [22], with four strains (i.e. Wistar, ZUR-Siv, Sprague-Dawley, and Long-Evans) used from both genders in this 100-page paper. Celio counted the number of parvalbumin-positive neurons in every 2nd 50 mm vibratome section of a whole sagittal series through the dorsal striatum. The parvalbumin-positive neurons ‘account for less than 2% of striatal neurons as inferred from counting the total number of neurons in adjacent cresyl-violet stained sections’ [22]. The absolute number of parvalbumin-immunoreactive interneurons in the dorsal striatum was not stated by Celio [22]. Less than 2% is closer to the finding of approximately 0.6% [15,35] discussed above (see Table 2). Wu and Parent [6] compared the density of calretinin, parvalbumin and somatostatin (i.e. NADPH-diaphorase) positive neurons in the dorsal striatum of the rat (i.e. female young adults of the Sprague-Dawley strain), squirrel monkey, and human. They report that the density of these three striatal interneurons was much higher in the monkey than in the rat and human (see also Cicchetti et al. [44]). Unfortunately, comparative neuronal density measures across species are inconclusive when they are not corrected for processing-induced shrinkage [19,21,25,27–30]. For example, a higher apparent neuronal density per mm2 could occur because the tissue of one species shrinks relatively more during processing through the same dehydration schedule. This may occur because the striatum in a particular species has a different tissue composition with respect to the ratio of white matter to grey matter. The white matter tract, the internal capsule, is located within the dorsal striatum in the rat but is separate from the caudate nucleus and putamen in the monkey and human. This known species difference in tissue composition within the dorsal striatum may contribute to the report [6] of a lower density of interneurons in the rat. A lower density in the human compared to the monkey [6] could also be due to the effects of postmortem delay on the intensity of immunostaining. A slightly higher density in the human compared to the rat may have occurred because 40 mm sections were used for the rat (and monkey) and 50 mm sections for the human. Due to these methodological limitations, the comparative density measures of Wu and Parent [6] may be misleading.
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
Deng et al. [12] used immunolabelling of the various neuronal types to define the frequency of each type in the caudate and putamen of the rhesus monkey. Counts of NeuN+ neurons were used to determine the total number of striatal neurons per unit area. The number of projection neurons and interneuronal subtypes were also counted per unit area and compared with the total number of NeuN+ neurons per unit area to calculate the frequency as a percent of all neurons. Implicit to this approach is that the assumption that the immunostained sections for each respective antibody (i.e. NeuN, calbindin, choline acetyltransferase, parvalbumin, calretinin, somatostatin) responded very similarly to the immunostaining processing (e.g. dehydration), such that the degree of shrinkage or swelling of the striatum is equivalent across all sections and antibodies used. This assumption is required in the interpretation of the data for cell number per unit area [27]. Since the Abercrombie method [31] was used, an assumption is also required that the shape and size of the neurons matched this mathematical model. For the caudate nucleus, Deng et al. [12] observed that 95% of the neurons were calbindin-positive and likely to be spiny projection neurons. Since less than 1% of calbindin neurons are also somatostatin-positive in the human, it was estimated that 94% of neurons in the caudate nucleus of the monkey are spiny projection neurons [12]. For the interneurons, 1.1% were cholinergic interneurons, 1% were parvalbumin interneurons, 1.9% were somatostatin interneurons and 2% were calretinin interneurons [12]. For the putamen, it was reported that 0.9% of the neurons were cholinergic interneurons, 2.4% were parvalbumin interneurons, 1.2% were somatostatin interneurons and 1.5% were calretinin interneurons [12]. More lightly labelled calbindin neurons, and calbindin-negative patches, were observed in the putamen which yielded a lower frequency of 83% for calbindin-positive neurons. This is likely to under-represent the percentage of spiny projection neurons in the putamen, such that the percentage of these neurons in the caudate nucleus was used for the putamen [12]. Of note, since calretinin is expressed by a subset of projection neurons and large interneurons, and by medium-sized interneurons, in the monkey striatum, the percentage of medium-sized calretinin interneurons was obtained from the percentage of spiny projection neurons (of 94%) and the percentage of other interneurons (of 4– 5%) to yield 1–2% are medium-sized calretinin interneurons. In summary, the total percentage of all interneurons in the dorsal striatum of the monkey was 6% for the caudate nucleus and 6% for the putamen [12]. To confirm or refute these results, a modern stereological study is now needed in which the absolute number of all neuronal subtypes in the dorsal striatum of the monkey is measured. 4. Conclusions It is evident that caution is needed when citing historical nonstereological data on the percentage of interneurons in the dorsal striatum of the cat, monkey and human. In these species, it is evident that the vast majority of neurons in the dorsal striatum are spiny projection neurons and a smaller number are interneurons, yet the precise percentage is currently unknown. Hence, in the absence of stereological data on the absolute number of all neuronal subtypes for the cat, monkey and human, it is currently unknown if there is a species difference, or not, for the percentage of interneurons in the dorsal striatum of the human and monkey versus the cat and rat. Future research is required to measure the absolute number of each interneuronal subtype in the dorsal striatum of the cat, monkey and human using specific immunohistochemical and histochemical markers and modern stereological methods.
23
Modern stereological data indicate that the percentage of interneurons in the rat dorsal striatum is <3%. More specifically, these data indicate that, in the Sprague-Dawley rat, 0.8% of all neurons are somatostatin/neuropeptide Y/GABA interneurons, 0.4% are cholinergic interneurons, approximately 0.6% are GABA/ parvalbumin interneurons and 0.5% are GABA/calretinin interneurons (i.e. a total of 2.2–2.3%, see Table 2). Current stereological data also indicates that the percentage of interneurons in the dorsal striatum is <1% for cholinergic interneurons in the SpragueDawley rat and human. Acknowledgement The editorial assistance of Mrs. Liping Goddard is gratefully acknowledged. References [1] Tepper JM, Abercrombie ED, Bolam JP. Basal ganglia macrocircuits. Progress in Brain Research 2007;160:3–7. [2] Kawaguchi Y, Wilson CJ, Augood SJ, Emson PC. Striatal interneurons: chemical, physiological and morphological characterization. Trends in Neurosciences 1995;18:527–35. [3] Graveland GA, DiFiglia M. The frequency and distribution of medium-sized neurons with indented nuclei in the primate and rodent neostriatum. Brain Research 1985;327:307–11. [4] Roberts RC, Gaither LA, Peretti FJ, Lapidus B, Chute DJ. Synaptic organisation of the human striatum: a postmortem ultrastructural study. Journal of Comparative Neurology 1996;374:523–34. [5] Desjardins C, Parent A. Distribution of somatostatin immunoreactivity in the forebrain of the squirrel monkey: basal ganglia and amygdala. Neuroscience 1992;47:115–33. [6] Wu Y, Parent A. Striatal interneurons expressing calretinin, parvalbumin or NADPH-diaphorase: a comparative study in the rat, monkey and human. Brain Research 2000;863:182–91. [7] Tepper JM, Bolam JP. Functional diversity and specificity of neostriatal interneurons. Current Opinion in Neurobiology 2004;14:685–92. [8] Kreitzer AC. Physiology and pharmacology of striatal neurons. Annual Review of Neuroscience 2009;32:127–47. [9] Tepper JM, Tecuapetla F, Koos T, Ibanez-Sanodval O. Heterogeneity and diversity of striatal GABAergic interneurons. Frontiers in Neuroanatomy 2010;4:150. http://dx.doi.org/10.3389/fnana.2010.00150. [10] Rice MW, Roberts RC, Melendez-Ferro M, Perez-Costas E. Neurochemical characterization of the tree shrew dorsal striatum. Frontiers in Neuroanatomy 2011;5:53. http://dx.doi.org/10.3389/fnana.2011.00053. [11] Groves PM. A theory of the functional organization of the neostriatum and the neostriatal control of voluntary movement. Brain Research Reviews 1983;5:109–32. [12] Deng Y-P, Shelby E, Reiner A. Immunohistochemical localization of AMPA-type glutamate receptor units in the striatum of rhesus monkey. Brain Research 2010;1344:104–23. [13] Graveland GA, Williams RS, DiFiglia M. A Golgi study of the human neostriatum: neurons and afferent fibers. Journal of Comparative Neurology 1985;234: 317–33. [14] Bernacer J, Prensa L, Gimenez-Amaya JM. Distribution of GABAergic interneurons and dopaminergic cells in the functional territories of the human striatum. PLoS One 2012;7:e30504. [15] Oorschot DE. Cell types in the different nuclei of the basal ganglia. In: Steiner H, Tseng KY, editors. Handbook of basal ganglia structure and function. London: Academic Press/Elsevier; 2010. p. 63–74. [16] Kemp JM, Powell TPS. The structure of the caudate nucleus of the cat: light and electron microscopy. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 1971;262:383–401. [17] Thomas TM, Smith Y, Levey AI, Hersch SM. Cortical inputs to m2-immunoreactive striatal interneurons in rat and monkey. Synapse 2000;37: 252–61. [18] Oorschot DE. The domain hypothesis: a central organising principle for understanding neostriatal circuitry? In: Miller R, Wickens JR, editors. Conceptual advances in brain research, brain dynamics and the striatal complex. Reading, UK: Gordon and Breach; 2000. p. 151–63. [19] Gundersen HJG. Stereology of arbitrary particles. A review of unbiased number and size estimators and the presentation of some new ones, in memory of William R. Thompson. Journal of Microscopy 1986;143:3–45. [20] West MJ. New stereological methods for counting neurons. Neurobiology of Aging 1993;14:275–85. [21] Pasik P, Pasik T, DiFiglia M. The internal organization of the neostriatum in mammals. In: Divac I, Gunilla R, Oberg E, editors. The neostriatum. Oxford: Pergamon; 1979. p. 5–36. [22] Celio MR. Calbindin D-28k and parvalbumin in the rat nervous system. Neuroscience 1990;35:375–475.
24
D.E. Oorschot / Basal Ganglia 3 (2013) 19–24
[23] Bernacer J, Prensa L, Gimenez-Amaya JM. Cholinergic interneurons are differentially distributed in the human striatum. PLoS One 2007;2:e1174. [24] Hutcherson L, Roberts RC. The immunocytochemical localization of substance P in the human striatum: a postmortem ultrastructural study. Synapse 2005; 57:191–201. [25] Schroder KF, Hopf A, Lange H, Thorner G. Morphometrical-statistical structure analysis of human striatum, pallidum and subthalamic nucleus. Journal fur Hirnforschung 1975;16:333–50. [26] Lange H, Thorner G, Hopf A, Schroder KF. Morphometric studies of the neuropathological changes in choreatic diseases. Journal of Neurological Sciences 1976;28:401–25. [27] Oorschot DE. Are you using neuronal densities, synaptic densities or neurochemical densities as your definitive data? There is a better way to go. Progress in Neurobiology 1994;44:233–47. [28] Swaab DF, Uylings HBM. Density measures: parameters to avoid. Neurobiology of Aging 1987;8:574–6. [29] West MJ. Stereological methods for estimating the total number of neurons and synapses: issues of precision and bias. Trends in Neurosciences 1999;22: 51–61. [30] Boyce RW, Dorph-Petersen K-A, Lyck L, Gundersen HJG. Design-based stereology: introduction to basic concepts and practical approaches for estimation of cell number. Toxicologic Pathology 2010;38:1011–25. [31] Abercrombie M. Estimation of nuclear populations from microtome sections. Anatomical Record 1946;94:239–47. [32] Raadsheer FC, Oorschot DE, Verwer RWH, Tilders FJH, Swaab DF. Age-related increase in the total number of corticotropin-releasing hormone neurons in the human paraventricular nucleus in controls and Alzheimer’s Disease: comparison of the disector with an unfolding method. Journal of Comparative Neurology 1994;339:447–57. [33] West MJ, Østergaard K, Andreassen OA, Finsen B. Estimation of the number of somatostatin neurons in the striatum: an in situ hybridization study using the optical fractionator method. Journal of Comparative Neurology 1996; 370:11–22.
[34] Oorschot DE, Zhang R, Wickens JR. Absolute number and three-dimensional spatial distribution of rat neostriatal large interneurons: a first and second order stereological study. Proceedings of the Xth International Congress of Stereology 1999;87. [35] Luk KC, Sadikot AF. GABA promotes survival but not proliferation of parvalbumin-immunoreactive neurons in rodent neostriatum: an in vivo study with stereology. Neuroscience 2001;104:93–103. [36] Rymar VV, Sasseville R, Luk KC, Sadikot AF. Neurogenesis and stereological morphometry of calretinin-immunoreactive GABAergic interneurons of the neostriatum. Journal of Comparative Neurology 2004;469:325–39. [37] Oorschot DE. Total number of neurons in the neostriatal, pallidal, subthalamic, and substantia nigral nuclei of the rat basal ganglia: a stereological study using the Cavalieri and optical disector methods. Journal of Comparative Neurology 1996;366:580–99. [38] Oorschot DE, Tunstall MJ, Wickens JR. Local connectivity between striatal spiny projection neurons: a re-evaluation. In: Nicholson L, Faull RLM, editors. The basal ganglia VII. New York: Plenum Press; 2002. p. 421–34. [39] Larsson E, Lindvall O, Kokaia Z. Stereological assessment of vulnerability of immunocytochemically identified striatal and hippocampal neurons after global cerebral ischemia in rats. Brain Research 2001;913:117–32. [40] Paxinos G, Watson C. The rat brain in stereotaxic coordinates, 4th ed., San Diego: Academic Press; 1998. [41] Kita H, Kitai ST. Glutamate decarboxylase immunoreactive neurons in rat neostriatum: their morphological types and populations. Brain Research 1988;447:346–52. [42] Kita H, Kosaka T, Heizmann CW. Parvalbumin-immunoreactive neurons in the rat neostriatum: a light and electron microscopic study. Brain Research 1990;536:1–15. [43] Parent A, Fortin M, Cote P-Y, Cicchetti F. Calcium-binding proteins in primate basal ganglia. Neuroscience Research 1996;25:309–34. [44] Cicchetti F, Prensa L, Wu Y, Parent A. Chemical anatomy of striatal interneurons in normal individuals and in patients with Huntington’s disease. Brain Research Reviews 2000;34:80–101.