Among the foremost investigators of the chemistry and biochemistry of
the central nervous system is Holger Hyden at the University of
Göteborg, Sweden. He and others (refs. [ref.118]-[ref.120]) have for
many years performed elegant microanalytical studies of single nerve
cells. The evidence which Hyden has obtained is consistent with the
hypothesis that the initial electrical reverberations in the brain
induce a change in the molecular structure of the ribonucleic acid (RNA)
of the neurones which, in turn, leads to a subsequent deposition of
specific proteins. It is well known from other investigations that a
major role of RNA in any type of cell is to specify and mediate
synthesis of the protein enzymes of the cells. Thus, in this hypothesis,
it is only necessary to postulate the modification of brain RNA by the
activities associated with reverberational memory. Particularly
pertinent to this hypothesis are observations that—
(1) Large nerve cells have a very high rate of metabolism of RNA and
proteins, and, of the somatic cells, are the largest producers of
RNA.
(2) Vestibular stimulation by passive means leads to an increase in
the RNA content of the Deiters nerve cells of rabbits ([ref.121]).
The protein content of these cells is also increased.
(3) Changes in the RNA composition of neurones and glia of the
brainstem occur during a learning situation. Animals were trained
over a period of 4 to 5 days to climb a steeply inclined wire to
obtain food. The big nerve cells and the glia of their lateral
vestibular apparatus were analyzed, since the Deiters neurones
present in this structure are directly connected to the middle
ear. The amount of RNA was found to be increased in the nerve
cells; and, more significantly, the adenine-to-uracil ratio of
both the nuclear RNA of nerve cells and glia cells became
significantly increased ([ref.119]). A variety of control
experiments were conducted. Although there was an increase in RNA
content of these cells in animals exposed to passive stimulation,
there was no change in the ratio of adenine to uracil. Nerve cells
from the reticular formation, another portion of the brain, had
only an increased content of RNA with no base-ratio change.
Animals subjected to a stress experiment involving the vestibular
nucleus showed only an increase in content of RNA. Littermates
living in cages on the same diet as learning animals showed no
change in content of RNA. Thus, it would appear that the change in
the base ratio of the RNA synthesized is not due to increased
neurone function per se, but is more directly related to the
learning process. The fact that this was nuclear RNA implies that
it was immediately related to chromosomal DNA.
(4) Neuronal RNA with changed cytosine-guanine ratios synthesized
during a short period of induced protein synthesis could be
blocked by actinomycin D. It was concluded, therefore, that the
RNA was immediately DNA dependent and directly related to the
genetic apparatus.
Rats which were normally right handed were forced to modify their
handedness in order to obtain food. The RNA of nerve cells in that part
of the cortex, whose destruction destroys the ability to transfer
handedness, was analyzed. A significant increase in RNA of nerve cells
of the fifth to sixth cortical layers on the right side of the brain was
observed. The corresponding nerve cells on the opposite side of the same
brain served as controls. There was an increase in RNA and a significant
increase in the purine bases relative to the pyrimidine bases in the
learning side of the cortex. When the animals were not forced to learn a
new procedure, only an increase of RNA was observed, with no change in
base ratio.
Frank Morrell, head of the Neurology Department at Stanford Medical
School, has also been active in this field during the past 6 years. He
has found that if a primary epileptic lesion is induced on one side of
the cortex, a secondary mirror lesion eventually develops in the
contralateral homologous cortex. This secondary lesion, which showed
self-sustaining epileptiform discharge, could be isolated, whereupon the
epileptiform discharge disappeared. This was interpreted as learned
behavior of the secondary lesion. From changes in the staining
properties of the secondary lesion, Morrell concluded that changes in
RNA had occurred in the cell. Changes in the composition of the RNA
could not be shown by these techniques.
At the University of California at Berkeley, Drs. Rosenzweig, Bennett,
and Krech have conducted extensive studies related to this topic. These
investigators have directed their efforts toward demonstrating
alterations in the cerebral cortex of animals exposed to continuing
learning situations or continuously deprived of sensory stimulation. In
a recent publication ([ref.116]), which also summarizes a considerable
amount of previous work, they report studies which demonstrate the
following:
(1) Rats given enriched experience develop, in comparison with their
restricted littermates, greater weight and thickness of cortical
tissue and an associated proportional increase in total
acetylcholinesterase activity of the cortex.
(2) The gain in weight of cortical tissue is relatively larger than
the increase in enzymatic activity. Acetylcholinesterase activity
increases in other portions of the brain even though tissue weight
decreases.
(3) The changes appear in a variety of lines of rats, although
differing in amount between strains.
(4) The changes are observed in both the young and adult animals.
The previous studies were comparisons between experience-enriched
animals and animals maintained in isolation. Animals which were housed
in colonies, but given no special treatment, showed intermediate effects
in those situations studied.
The Berkeley group emphasized that the finding of changes in the brain
subsequent to experience does not prove that the changes have anything
to do with memory storage, but do establish the fact that the brain can
respond to environmental pressure. However, the results are compatible
with the hypothesis that long-term memory storage involves the formation
of new somatic connections among neurones. Calculations of the amount of
additional material required to permit this to exist are compatible with
the increases observed.
A number of investigators have studied the effects of antimetabolites
and drugs on the learning process. Since their specific metabolic
effects are known in other tissues, the rationale is that if these
materials do interfere with memory, then specific types of metabolic
activities may be implicated in the deposition of the engram.
One of the initial studies of this type was conducted by Dingman and
Sporn ([ref.122]), presently at the National Institute of Mental Health.
They showed that 8-azaguanine, a purine antagonist, injected
intra-cisternally was incorporated into the RNA of the brains of rats.
Associated with this incorporation was an impairment of the
maze-learning ability of the animals. These findings have been
confirmed.
Flexner and his associates injected puromycin, an inhibitor of protein
synthesis, into the brains of mice, which were then trained to perform
in a maze. Losses of short-term or long-term memory were obtained,
depending upon the site of the injection. The results indicate that the
hippocampal region is the site of recent memory.
The hippocampal region is of interest in connection with memory
processes for a number of other reasons. Adey et al. ([ref.123]) and his
group observed a transient fall in electrical impedance in this region
when cats learned to perform in a T-maze in response to a visual cue. It
was supposed that the electrodes were situated within glial cells of the
dendritic zone of the hippocampal pyramidal cell layer. Extinction of
the learned habit abolished the briefly evoked impedance changes, which
subsequently reappeared with retraining.
A number of other studies more or less indirectly implicate RNA in the
learning processes. For instance, in retinal cells of rabbits raised in
darkness, there was virtually no ribonucleoprotein as compared with
normal amounts in the cells of animals raised in light ([ref.124]).
Further, maintenance of normal electrical activity of isolated perfused
cat brains is highly dependent upon the presence of the ribonucleic acid
precursors, uridine and cytidine, in the perfusate ([ref.125]), and
severe derangements occur if any of a variety of pyrimidine antagonists
are added ([ref.126]). Brief electrical stimulation of cat cortical
tissue causes an increase in nucleic acid cytidine and adenine, thus
indicating a synthesis of altered polynucleotides. Finally, injections
of RNA in animals have shown interesting effects. When given at a dose
of 116 mg/kg daily for 1 month, rats showed an enhanced response and
greater resistance to extinction in a shock-motivated behavioral
response. It has been shown by another group that injections of RNA
enhance the ability of young animals to learn various tasks.
Planaria have been used in a variety of studies which seem to bear on
the problem of memory. Quite recent evidence by Bennett, Calvin, and
their associates has cast somewhat of a pall over the studies;
nevertheless, the work may have some validity. Interest in the use of
flatworms, particularly planaria, for study of memory began with a
demonstration by McConnell that these simple animals could undergo
conditioning ([ref.127]). Subsequently, it was found that some
conditioning was retained when the animal was transected and allowed to
regenerate. The retention of training was found in both new animals,
although the very simple brain, really only two ganglia, was in the head
section ([ref.128]).
Apparently, some diffusely distributed component of the animal was
responsible for retention of learning. Evidence has accumulated to
indicate that this material is RNA. Among this evidence is the
following:
(1) The two halves of a trained planaria were allowed to regenerate in
a solution containing RNA-destroying enzymes. Whereas the head
ends retained some training, no retention was observed in the
animals derived from the tail end ([ref.129]).
(2) When pieces of trained planaria were fed to untrained animals, the
untrained cannibal required a shorter time to become trained to a
criterion. It would appear that the digestive system of planaria
is so simple that the material responsible for the transfer of the
information was not broken down.
(3) When RNA, obtained from trained planaria, is injected into the
digestive tract of untrained animals, there is a transfer of
information.