|Home | About | Journals | Submit | Contact Us | Français|
The common marmoset, Callithrix jacchus, is of growing importance for research in neuroscience and related fields. In the present work, we describe a combined histological and magnetic resonance imaging (MRI) atlas constructed from the brains of two adult female marmosets. Histological sections were processed from Nissl staining and digitized to produce an atlas in a large format that facilitates visualization of structures with significant detail. Naming of identifiable brain structures was performed utilizing current terminology. The histological sections and a simplified schematic atlas are available online at http://udn.nichd.nih.gov/brainatlas_home.html.
The common marmoset, Callithrix jacchus, is of growing importance for research in neuroscience and related fields, including pharmacology, ethology, genetics, endocrinology, and reproductive biology. Their relatively small body size (adults range from 14 to 18 cm and weigh 250–500 g), and social organization in small family groups lend themselves to being kept in manageable captive housing facilities. Their lissencephalic brain is smaller than that of the rhesus macaque (Figure 1), making brain preparation and experimentation less cumbersome than with larger primate brains. Yet, their brains share a number of anatomical features found in other primates, with a well-defined lateral fissure, temporal lobes, and internal structures that are very similar to that of other primates, including humans.
For readers less familiar with the marmoset and research based on this primate, we offer the following section as an overview. Marmosets and tamarins (family Callitrichidae) are small New World primates. Most live in small family groups, in which the father and older offspring play a major role in caring for the infants, which are typically born as twins. Marmosets have high reproductive capacity, reaching breeding age at around 18 months, and females can produce two litters per year. The common marmoset, Callithrix jacchus, is probably the best known, and the most widely used in research, spanning a wide range of fields. These include behavioral pharmacology (e.g., Cagni, et al., 2009), reproductive physiology (e.g., Haig, 1999; Saltzman and Abbott, 2009), genetics (e.g., Datson, et al., 2007; Raveendran, et al., 2008), transgenics (e.g., Sasaki, et al., 2009; Schatten and Mitalipov, 2009), comparative psychology (e.g., Hook and Rogers, 2008), and social behavior (e.g., Barbosa, et al., 2009; Clara, et al., 2008 ; Zahed, et al., 2008).
Notably, a draft sequence of the marmoset genome has recently been released (http://www.hgsc.bcm.tmc.edu/project-species-p-Marmoset.hgsc?pageLocation=Marmoset). Perhaps of greatest relevance to the marmoset brain atlas we are presenting here, there is an extensive literature on the use of common marmosets in neuroscience research. Some examples of this rich literature include chemoarchitecture of calcium-binding proteins (Bourne, et al., 2007), architecture of frontal cortex (Burman, et al., 2009), sex differences in brain anatomy (Reinius, et al., 2008), auditory physiology (Wang, 2007; Wang, et al., 2008), cortical neurotransmitters (Walker, et al., 2009), parietal cortex anatomy and physiology (Qi, et al., 2002; Burish, et al., 2008), visual cortex anatomy (Rosa, et al., 2009), early developmental manipulations and gene expression in the brain (Law, et al., 2009; Pryce, 2008), and marmosets as models for studying human neuromuscular and autoimmune diseases (Eslamboli, 2005; Hohjoh, et al., 2009; ‘t Hart and Massacesi, 2009).
Several neuroanatomical overviews of the marmoset brain have been published (Osman Hill, 1957; Hershkovitz, 1977; Prado-Reis and Erhart, 1979a,b; Saavedra and Mazzuchelli, 1969). Descriptions of marmoset cerebral cortex have also been published (Brodmann, 1909; Peden and von Bonin, 1947). Until very recently, only one comprehensive brain atlas had been published, and is now out of print and difficult to obtain (Stephan, Baron, and Schwerdtfeger, 1980). In 2008, a new stereotaxic atlas became available (Palazzi and Bordier, 2008). This atlas presents brain sections stained for acetylcholine esterase (AChE). Although the plates in the printed atlas are quite small to show much detail, the accompanying Compact Disc permits enlarging the figures in considerable detail. AChE staining leaves out much of the cellular detail that permits a researcher to reliably identifying a structure under the microscope, and distinguishing cellular masses (nuclei) with the clarity that Nissl material permits. Most recently, motivated by the creation of a public internet database for histological brain sections (http://brainmaps.org, Miluka, et al., 2007), Tokuno, et al. (2009) constructed a database of histological sections of the marmoset brain that were converted to digital images and which are available online (http://marmoset-brain.org:2008).
Magnetic resonance imaging (MRI) has established itself as the most widely used modality for imaging the whole brain in vivo. MRI excels in neuroimaging because of its intrinsic soft-tissue contrast, high resolution, and ability to follow in vivo processes longitudinally. The marmoset has been increasingly studied with MRI (Figure 2). However, to date, there are no studies relating the soft tissue contrast obtained with MRI with the cytoarchitectonic information provided by histology. The goal of the present work is to present a combined histological and MRI-based atlas of the marmoset brain that may serve as a reference for studies in contemporary neuroscience research, including electrophysiology, gene expression studies (e.g., through the use of immunocytochemistry) and functional magnetic resonance imaging. We also concluded, through our own experience and careful reading of existing stereotaxic atlases, that distortions created by histological preparation and inter-brain variability in the dimensions and spatial location of internal brain structures made reliance on the positional information found in a stereotaxic atlas problematic. Our goal here is to present an atlas with good quality histology, in a large format that facilitates visualizing detailed structure, and using terminology that reflects current conventions.
The complete set of Nissl images upon which the atlas is based may be found at http://udn.nichd.nih.gov/brainatlas_home.html. For the present atlas, each Nissl image was cropped to present only one hemisphere, and structures were labeled with Adobe Photoshop Elements 2.0. Each plate is identified by an ‘A’ or ‘P’ number to designate anterior or posterior coordinates respectively, as one would expect to find in a stereotaxic atlas. These designations are meant only as plate identifiers, and not as accurate stereotaxic levels. Structures were identified by reference to existing atlases, in particular, the marmoset atlas by Stephan, et al. (1980). However, their use of different nomenclature, together with several cell clusters in our female’s brain that appeared to have no equivalent in the male brain as used in the atlas by Stephan, et al., made identification of a number of structures problematic. We took a conservative approach, and wherever doubt existed about the identity of a particular cell cluster, no identification was made.
The brain atlas is presented both as a series of Nissl histology plates (Figures 3–31), and a series of MRI images (Figures 32–60) matched as closely as possible to the Nissl plates. Each Nissl plate shows a 40 micron-thick section. The distance between sections is 480 microns. The MRI images are 66 micron-thick. To improve the signal-to-noise ratio of the MRI while preserving gray-to-white matter contrast, the fixed brain was soaked for one week in a 10% formalin solution doped with 5 mM gadopentetate dimeglumine (see Experimental Procedure below). This MRI contrast agent reduces both T1 and T2 relaxation time constants of the fixed tissue, allowing the MRI to be acquired in a shorter period of time. A T2-weighted sequence was chosen to enhance the MRI contrast between gray and white matter under the conditions of a shorter repetition time, and excellent contrast between gray and white matter structures were obtained. While similar contrast could have been obtained using a T1-weighted sequence, as it is commonly used for in vivo MRI, a much longer acquisition time would have been necessary. Even though the contrast in the MRI images was different than the contrast in either the corresponding Weil or Nissl histological plates, the spatial resolution of 66 microns was sufficient to allow clear delineation of all areas identified in the histological sections. The images of the sections have been cropped to highlight the detailed histology of deep brain structures. While this excludes the cortex, cortical maps now tend to be made on a variety of evidence from anatomical, neurochemical, and functional studies, rather than from histology alone (Rosa and Tweedale, 2005). Examination of the histology underlying the atlas indicates that the basic neuroanatomical organization described in earlier publications of the marmoset brain is present in our atlas brain. The Nissl images used for the atlas are also available online (http://udn.nichd.nih.gov), showing the complete section in each instance, along with matching Weil-stained sections for myelinated fibers. Eventually, the MR images, along with higher-resolution Nissl and Weil images, will be available online.
While there is, at present, no detailed description of internal differences in the brains of male and female marmosets, most primate brain atlases employ the brain of a male of the species. Given that the female brain is at least as interesting from a research standpoint as the male brain, we chose to use the brain of a female for our atlas. We chose not to produce a stereotaxic atlas for several reasons. In our experience, the coordinates presented in atlases, at least for the marmoset and the squirrel monkey, with which we have extensive experience, are not very accurate. This may be because of variability in the location and shape of brain structures across different individual monkeys, or because of spatial distortions introduced during fixation, histological sectioning and staining. As Gergen and MacLean (1962, p. 5) state in their squirrel monkey atlas, ‘…one would have a 50% chance of coming within 0.5 mm of a point described in the present atlas.’ This is not to say that the use of the stereotaxic frame and an atlas developed for stereotaxic research doesn’t have continuing value in neuroscience research. Rather, our focus has been to provide a reference tool useful for those doing immunocytochemical and other types of anatomical studies for which stereotaxic coordinates are of little value. The authors agree with the statement by C. N. Woolsey that an atlas based on the actual brain histology is preferable to line drawings, ‘…since line drawings necessarily force one to make arbitrary decisions concerning boundaries between structures when, without experimental verification, the boundaries may be uncertain or, in fact, not at all sharply definable’ (p. vi, in Emmers and Akert, 1963).
Some nuclei identified in Stephan, et al (1980) were not clearly identified in our brain sections, and, as consequence, fewer structures in our atlas are identified. It is unclear whether this is due to the fact that we used a female brain whereas the earlier atlas used a male brain, or, more likely, to the possibility that we were more conservative in labeling only structures that were clearly delineated in our histological material. Our use of English equivalents for nomenclature is a departure from most of the older brain atlases, although English terminology is used in more recent atlases (e.g., Paxinos, et al., 2000). The history of anatomical nomenclature suggests that attempts to standardize the naming of anatomical structures using Latin terms, as laid out in the BNA (Basle Anatomical Nomenclature, e.g., Jamieson, 1916), has not been without controversy. The original list of terms was derived from human anatomical treatises, so it is not surprising that attempts to transfer these terms to the brain anatomy of other species will not meet universal acceptance. We chose the English terms used for the rhesus macaque brain illustrated at the http://brainmaps.org website because of the agreement in internal positioning of what appear to be the equivalent structures in the marmoset brain used in our atlas (Table 1).
Experiments were performed in two adult female common marmosets, two years in age, 500 g body weight, from our breeding colony. All procedures followed the guidelines of the NIH Animal Care and Use Committee.
One of the female marmosets was deeply anesthetized with sodium pentobarbital (100mg/kg) and perfused through the heart with PBS followed by 10% formalin. The brain was then sent to Neuroscience Associates of Knoxville, TN, who prepared it for histological analysis. The brain was cut in the coronal (frontal) plane at 40 microns, every sixth section stained for Nissl granules with thionine and the adjacent section stained for myelinated fibers with the Weil technique. Only Nissl-stained sections are used in the present atlas. The actual distance between the Nissl-stained sections was 240 microns. We selected every other section to produce an atlas with sections spaced every 480 microns. The mounted sections were photographed at the NIH (Medical Arts and Photography Branch). The equipment used was a Nikon Multiphot optical bench with Zeiss Luminar 100 mm lens, and scanned with a Better Light 6100 scan back driven by Better Light Viewfinder 5.3 software. The final images were saved as arrays of 6000×8000 pixels in Adobe Photoshop 6.0. Some additional re-touching (brightness and contrast) was done with Adobe Photoshop Elements 2.0.
MRI images were made in the fixed brain of another female marmoset. The marmoset was sacrificed under deep anesthesia (sodium pentobarbital, 100 mg/kg, iv) via perfusion fixation through the ascending aorta with 500 mL of ice-cold 4.0% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The brain was removed and further fixed by immersion in the paraformaldehyde solution for 2 hours at 4°C. It was then stored in 10% neutral buffered formalin doped with 5 mM gadopentetate dimeglumine (Magnevist, Berlex Laboratories, Wayne, NJ, USA) contrast agent for one week. MRI was performed on a 7T/30 cm USR/AVIII MRI scanner (Bruker-Biospin, Inc., Ettlingen, Germany) equipped with a 15 cm gradient set capable of 450 mT/m strength (Resonance Research, Inc., Billerica, MA, USA). The brain was imaged in a 3.5 cm inner diameter transmit/receive birdcage coil (Bruker-Biospin) with a T2-weighted, 3D multi spin-echo sequence (TE = 13, Number of Echoes = 2, TR = 350 ms, FOV = 34 mm × 26 mm × 22 mm, Matrix = 512 × 395 × 335) in about 13 hours. The displayed coronal sections from this data set are 66 μm thick with an in-plane resolution of 66 μm.
The nomenclature and abbreviations used in our atlas have come, almost exclusively, from brainmaps.org (http://brainmaps.org/index.php?action=viewslides&datid=1) (Table 1). The coronal sections of a male rhesus macaque posted there meet the criteria of good quality histology, and identification of sufficient structures to be useful.
The authors wish to thank Ms. Julie Mackel and Ms. Deborah Bernhards for technical assistance. This research was supported by the Intramural Research Program of the NIH, NICHD and NINDS.