Our research is focusing to understand mechanisms underlying memory, formation and evolution of the brain. For one approach to understand these questions, we are studying the genes that are expressed in specific areas of the primate neocortex. Using differential display method, we obtained three genes that showed marked differences within areas of the primate neocortex. Our second approach is to understand informational processing in the brain underlying learning behaviors by examining gene expression. The third approach is to study voles of CNTF in the nervous system.

I. Genes expressed in specific areas of the neocortex

The neocortex, most remarkably evolved in primates, plays the major role in higher functions of the brain. It is known to be divided into distinct functional and anatomical areas and has been a matter of debate what extent the area of the neocortex are genetically and environmentally deter-mined. It is also puzzling why, during the evolution of mammals, the neocortex was most markedly expanded while the number of the genes in the mammal was little changed. To access these questions, we studied gene expression within different areas of the neocortex.

1) In collaboration with Professor Hiroyuki Nawa (Nigata university), we used the DNA macroarray technique to examine gene expression in the areas of human prefrontal, motor and visual cortexes. We found almost all the genes among 1088 genes examined showed only less than a factor of two in the difference of their expressions. Only one gene showed more than three fold difference and another one was between two and three fold difference within the three areas. These results suggest that the genes that are expressed among the different areas of the human neocortex are very similar. However, the question remained whether there are any genes that show marked difference within areas of neo-cortex.

2) In order to answer this question, we employed differential display methods and found at least two genes that indicated area specific expressions.
i) One gene, designated occ1, is specifically expressed in the occipital cortex, particularly in V1 area, in the primate brain. Furthermore, the expression of occ1 turned out to be activity dependent, because, in the monocularly deprived monkeys injected with TTX into one of the eyes, the expression of occ1 is markedly decreased in the ocular dominance columns of the primary visual cortex (V1). We also demonstrated that occ1 expression was markedly increased postnataly in V1 (Fig 1, Tochitani et al., 2003).
ii) The other gene that showed marked difference within the neocortex, is gdf7, a member of BMP/TGF-b family, which is specifically expressed in the motor cortex of the African green monkey. We are currently examining the detailed expression pattern of the gene.
iii) The third gene that we designated tentatively as 134G which was preferentially expressed in association and higher areas in the neocortex (Komatsu et al., Society for Neuroscience 32^nd Annual Meeting, 2002, being submitted).

3) We have also further isolated several area specific genes with RLCS (Restriction Landmark cDNA Scanning). We are now characterizing these genes to reveal the mechanisms that form neocortical areas.

In summary, our studies thus far revealed the following points.
(1) Genes that are specifically expressed within neocortical areas in the primate neocotex are not different overall.
(2) We have identified several genes that are distinctively different among neocortical areas.
(3) These genes are specific in visual, motor and association areas.
(4) A gene specific in the visual cortex (occ1) is activity dependent and also postnatally regulated.
(5) An association area specific gene is expressed in a complimentary manner to the expression of occ1.
(6) These results suggest that these genes may be useful markers to study the mechanisms underlying neocortical formation.

Fig. 1 Developmental Expression pattern of occ1 in the visual cortex. In situ hybridization pattern of occ1 in the primate visual cortex. occ1 is markedly expressed in the layer IVcb and moderately in the layers of II, III and IVa in area V1 (e). This adult expression pattern is regulated by postnatal development. (a, b) newborn; (c, d) three month; (e, f) adult monkey. (a, c, d) with occ1 antisense probe and (b, d, f) with thionin for cell bodies. (See. Tochitani et al., Eur. J. Neurosci:, 13, 297-307, 2001)

II. Gene expression under a declarative and a non-declarative memory

In order to study informational processing underlying the declarative and non-declarative memory at molecular and cellular levels in the brain, we employed c-Fos mapping techniques, for which we used gene expression of c-Fos. There have been an increasing number of studies using c-Fos as markers to examine neuronal activities ever since c-Fos induction by electrical stimulation was found by Morgan and Curran. However, many sensory stimuli per se are now known to cause c-Fos induction. So, we should be very careful to distinguish the c-Fos expression that is caused by learning process from that caused by sensory stimuli. For this purpose, it is necessary to use behavioral systems that are able to distinguish the two. Although a few behavioral systems in rodents have been successfully used for physiology, animal behavior and recently for analyses of knockout mice, little behavioral systems in fact distinguish it. Therefore, we prepared ourselves for using two behavioral systems, which represent declarative and non-declarative memory, respectively.

In collaboration with professor Yoshio Sakurai (Kyoto University) who developed audio-visual discrimination task (AVD-task). In this task, a rat was asked to choose either an audio cue (a high tone or low tone) or a visual cue (a light from the right or the left) to obtain a food pellet . We found that the visual and audio tasks enhanced the specific expression of c-Fos in the visual and audio cortices, respectively. Among the early visual and auditory pathways examined, c-Fos was specifically induced in the cortices but not in the earlier pathways, suggesting the neural modulation of the neocortex depending on the types of the tasks. Interestingly, the task-dependent Fos expression was only observed in excitatory neurons in the relevant sensory cortices.

Fig.2. AVD tasks. A) A rat face to a panel with a visual cue (a light from right or left) or an auditory cue (a high or low tone) was shown to push the panel with the time sequence shown in (B). The combination of possible cues by auditory and visual tasks to obtain a food pellet is shown in (C). (See Sakata et al., Eur. J. Neurosci., 15, 735-743, 2002, for the detail of the tasks)

Although this AVD task system is quite powerful to analyze a kind of task above described and presumably very useful for studying underlying molecular and cellular mechanisms because of advantages using rodents, one problem is that the auditory stimuli and visual stimuli are in different positions. Thus we cannot exclude the possibility that the difference between the auditory task and the visual task may not completely depend on the modality (i.e., visual Vs auditory) difference.

We wanted to solve this problem by placing auditory and visual stimuli in the same position. We also use nose-poking to measure the reaction time in which a rat responds to stimuli. By using this behavioral system, we were able to confirm amodal recognition of space which means that a rat can respond to a different modality (visual or auditory) if the stimuli is in the same position and previously reported in other systems. We also confirm multisensory enhancement is indeed observed in rats. These resuluts suggests that this new modified AVD system can be used to explore the molecular and cellular mechanisms underlying multisensory processing in rats.

The other task we developed is a wheel running system in which a water-deprived mouse is asked to run to obtain water in front because the wheel with the pegs is turning to the other direction (Kitsukawa et al., Society for Neuroscience 32nd Annual Meeting, 2002). The pegs can be changed with various patterns as desired. The task required for the mouse thus can be regarded to represent a non-procedural learning. We examined a various areas of brains following to the change of the peg pattern. Among the areas examined, we found marked c-Fos expression in the striatum. The striatum, which is composed of projection neurons and several distinguished types of interneurons, is known to play an important role in a reward-based learning. The characterization of these subtypes of interneurons has been progressed. However, their roles in behavioral tasks have been little known. We hope our system combined with c-Fos mapping technique reveals it. We are currently doing collaborative work with Professor Ann Graybiel (MIT) to record with tetrode from the striatum under the task (supported by US-Japan exchange program). We hope our wheel running system combined with c-Fos mapping technique and electrophysiology reveals molecular and cellular mechanisms under the Wheel Running task.

III. CNTF is specifically expressed in the developing pineal gland

CNTF, a member of the IL-6 family, attracts quite attentions of developmental neuroscientists because it shows various effects on neurons and glial cells. CNTF knockout mice, however, only indicate moderate motor neuron deficiency in the adult, but no apparent phenotype in the development. In order to explore the function of the IL-6 family, we extensively examined the expression of members of the family and their receptors and found the specific expression of CNTF in the embryonic pineal glands and eyes. This was to our knowledge for the first time to show a clear expression of CNTF in the development. Next question to be asked is the functional role of CNTF in the pineal development. Because of no reported phenotype of CNTF knockout mice in development, it has been difficult to know the developmental roles of CNTF if any. In fact, we also found there seemed no apparent difference in the pineal development of CNTF knockout mice compared to that in wild type mice. This is presumably because the CNTF gene knockout is largely compensated by other CNTF-like factors.

We therefore studied cultured pineal organs in rodents to ask if there were any effects of CNTF on it following to a previous study that shows that cultured neonatal pineal organs develop photoreceptor-like cells (Araki, 1992). We found that CNTF inhibits photoreceptor-like cells (Fig. 3, Hata et al., Dev Brain Res., 2003). Our observation raises an interesting possibility that CNTF plays a critical role in the pineal development to suppress a certain phenotype such as photoreceptors. It also raises a question what roles CNTF play in developing pineal glands if it play any roles in the development. CNTF has been studied for nearly thirty years and still the role in vivo remains to be studied. We hope that our study gives a clue to solve this long standing question.

Fig. 3. CNTF suppress Rhodopsin expression in pineal glands in vitro. CNTF was added to the medium of rat pineal organ culture. (Right) Control culture without CNTF on day 10. Rhodopsin-positive photoreceptor-like cells with many processes are seen at a higher magnification. (Left) CNTF (100 ng/ml) was added for 10 days. Photoreceptor-like cells of rhodopsin –positive cells are very rare. If any, they do not show long processes (lower, a higher magnification). These results were published in Hata et al., Dev. Brain Res., 143, 179-187, 2003.