Ann. Rev. Pharmacol. Toxicol. 1988. 28:1-23
Copyright © 1988 by Annual Reviews Inc. All rights reserved

AN UNEXPECTED LIFE IN
RESEARCH

Julius Axelrod

Laboratory of Cell Biology, National Institute of Mental Health, Bethesda, Maryland
20892

BEGINNINGS

Successful scientists are generally recognized at a young age. They go to the
best schools on scholarships, receive their postdoctoral training fellowships at
prestigiots laboratories, and publish early. None of this happened to me.

My parents emigrated at the beginning of this century from Polish Galicia.
They met and married in America, where they settled in the Lower East Side
of New York, then a Jewish ghetto. My father, Isadore, was a basketmaker
who sold flower baskets to merchants and grocers. I was born in 1912 in a
tenement on East Houston Street in Manhattan.

I attended PS22, a school built before the Civil War. Another student at that
school before my time was I. I. Rabi, who later became a world-renowned
physicist. After PS22 I attended Seward Park High School. I really wanted to
go to Stuyvesant, a high school for bright students, but my grades were not
good enough. Seward Park High School had many famous graduates, mostly
entertainers: Zero Mostel, Walter Mathau, and Tony Curtis. My real educa-
tion was obtained at the Hamilton Fish Park Library, a block from my home. I
was a voracious reader and read through several books a week—from Upton
Sinclair, H. L. Menken, and Tolstoy to pulp novels such as the Frank
Merriwell and Nick Carter series. .

After graduating from Seward Park High School, I attended New York
University in the hope that it would give me a better chance to get into
medical school. After a year my money ran out, and I transferred to the
tuition-free City College of New York in 1930. City College was a proletarian
Harvard, which subsequently graduated seven Nobel Laureates. I majored in
biology and chemistry, but my best grades were in history, philosophy, and
literature. Because I had to work after school, I did most of my studying

1
0362-1642/88/04 15-0001$02.00
2 AXELROD

during the subway trip to and from uptown City College. Studying in a
crowded, noisy New York subway gave me considerable powers of concen-
tration. When I graduated from City College, I applied to several medical
schools but was not accepted by any.

In 1933, the year I graduated from college, the country was in the depths of
a depression. More than 20% of the working population was unemployed, and
there were few jobs available for City College graduates. I had heard about a
laboratory position that was available at the Harriman Research Laboratory at
New York University, and although the position paid $25 a month, I was
happy to work in a laboratory. I assisted Dr. K. G. Falk, a biochemist, in his
research on enzymes in malignant tumors. I also purified salts for the prepara-
tion of buffer solutions and determined their pH. The instrument used to
measure pH at that time was a complex apparatus; the glass electrode occu-
pied almost half a room. In 1935 the laboratory ran out of funds, and I was
fortunate to get a position as a chemist in the Laboratory of Industrial
Hygiene. This laboratory was a nonprofit organization and was set up by New
York City’s Department of Health to test vitamin supplements added to foods.
1 worked in the Laboratory of Industrial Hygiene from 1935 to 1946.

My duties there were to modify published methods for measuring vitamins
A, B, Bz, C, and D so that they could be assayed in various food products that
city inspectors randomly collected. Vitamins had just been introduced at that
time, and the New York City Department of Health wanted to establish that
accurate amounts of vitamins were added to milk and other food products.
The methods used for measuring vitamins then were chemical, biological, and
microbiological. It required some ingenuity to modify the methods described
in the literature to assays of food products. This experience in modifying
methods was slightly more than routine, but it proved to be useful in*my later
research. The laboratory subscribed to the Journal of Biological Chemistry,
which I read with great interest. Reading this journal made it possible to keep
up with advances in the enzymology, nutrition, and methodology. During the
time I was in the Laboratory of Industrial Hygiene, I received a MS degree in
chemistry at New York University in 1942 by taking courses at night. My
thesis was on the ester-hydrolyzing enzymes in tumor tissues. Because of the
loss of one eye in a laboratory accident, I was deferred from the draft during
World War II. In 1938, I married Sally Taub, a graduate of Hunter College
who later became an elementary school teacher. We had two sons, Paul and
Alfred, born in 1946 and 1949.

FIRST EXPERIENCE IN RESEARCH: GOLDWATER
MEMORIAL HOSPITAL

I expected that I would remain in the Laboratory of Industrial Hygiene for the
rest of my working life. It was not a bad job, the work was moderately

AN UNEXPECTED LIFE IN RESEARCH 3

interesting, and the salary was adequate. One day early in 1946 the Institute
for the Study of Analgesic and Sedative Drugs approached the president of the
Laboratory of Industrial Hygiene with a problem. The president of the
Laboratory at that time was George B. Wallace, a distinguished pharmacolo-
gist who had just retired as Chairman of the Department of Pharmacology at
New York University. Many analgesic preparations contained nonaspirin
analgesics, such as acetanilide or phenacetin. Some people who became
habituated to these preparations developed methemoglobinemia. The Institute
for the Study of Analgesic and Sedative Drugs offered a small grant to the
Laboratory of Industrial Hygiene to find out why acetanilide and phenacetin
taken in large amounts produced methomoglobinemia. Dr. Wallace asked me
if I would like to work on this problem. I had little experience in this kind of
research, and he suggested that I consult Dr. Bernard “Steve” Brodie. Dr.
Brodie was a former member of the Department of Pharmacology at New
York University and was doing research at Goldwater Memorial Hospital, a
New York University Division.

I met with Brodie in February 1946 to discuss the problem of analgesics. It
was a fateful meeting for me. Brodie and I talked for several hours about what
kind of experiments could be done to find out how acetanilide might produce
methemoglobinemia. Talking to Brodie about research was one of my most
stimulating experiences. He invited me to spend some time in his laboratory
to work on this problem. Qne of a number of possible products of acetanilide
that would cause the toxic effects was aniline. It had previously been shown
that aniline could produce methemoglobinemia. Thus, one approach was to
find out whether acetanilide could be deacetylated to form aniline in the body.
With the help and guidance of Steve Brodie, I developed a method for
measuring aniline in nanogram amounts in urine and plasma. After the
administration of acetanilide to human subjects, aniline was found to be
present-in urine and plasma. A direct relationship between the level of aniline
in blood and the amount of methemoglobin present was soon observed (1).
This was my first taste of real research, and I loved it.

Very little acetanilide was found in the urine, suggesting extensive
metabolism in the body. Since acetanilide was almost completely transformed
in the body, we looked for other metabolic products. Methods to detect
possible metabolites, p-aminophenol and N-acetyl-p-aminophenol, were de-
veloped that were specific and sensitive enough to be used in the plasma and
urine. Within a few weeks, we identified the major metabolite as hydroxy-
lated acetanilide N-acetyl-p-aminophenol and its conjugates. This metabolite
was also found to be as potent as acetanilide in analgesic activity. By taking
serial plasma samples, acetanilide was shown to be rapidly transformed to
N-acetyl-p-aminophenol (1). After the administration of N-acetyl-p-
aminophenol, neglible amounts of methemoglobin were produced. As a result
of these studies, Brodie and I stated in our paper (1), “the results are
4 AXELROD

compatible with the assumption that acetanilide exerts its action mainly
through N-acetyl-p-aminophenol [now known as acetaminophen]. The latter
compound administered orally was not attended by the formation of methe-
moglobin. It is possible therefore, that it might have distinct advantages over
acetanilide as an analgesic.” This was my first paper, and I was determined to
continue doing research.

Soon after Brodie and I examined the physiological disposition and
metabolism of acetanilide, we turned our attention to a related analgesic drug,
phenacetin (acetophenetidin). I spent some time developing sensitive and
specific methods for the identification of phenacetin and its possible metabo-
lite, p-phenetidine. Brodie and I soon found that in humans, the major
metabolic product was also N-acetyl-p-aminophenol arising from the de-
ethylation of the parent compound (2). A minor metabolite was p-
phenetidine, which we found was responsible for the methemoglobinemia
formed after the administration of large amounts of phenacetin to dogs. After
the administration of phenacetin to human subjects, N-acetyl-p-aminophenol
was rapidly formed. The speed and the amount with which N-acetyl-p-
aminophenol was formed in the body suggested that the analgesic activity
resided in its deethylated metabolite. :

The laboratories at Goldwater Memorial Hospital where I began my re-
search career were set up during World War II to test newly synthesized
antimalarial drugs for their clinical effectiveness. Early in the war, the
Japanese had cut off most of the world’s supply of the antimalarial quinine.
James Shannon, then a renal physiologist at New York University, was put in
charge of this program. Shannon had the remarkable capacity to pick the right
young people to carry out research in the antimalarial project. Members of the
team that worked at Goldwater in addition to Steve Brodie were Sid Uden-
friend, Robert Berliner, Bob Bowman, Tom Kennedy, and Gordon Zubrod.
The atmosphere at Goldwater was highly stimulating, and an outpouring of
important new findings resulted. It was in this atmosphere that, in a period of
a few years, I became a researcher.

After completion of the studies on acetanilide and phenacetin, Brodie
invited me to stay on at Goldwater to study the fate of other analgesic drugs.
We received a small grant from the Institute for the Study of Analgesic and
Sedative Drugs, and the Laboratory of Industrial Hygiene paid my salary.
Another drug we investigated was the analgesic antipyrine. A sensitive
method for the detection of this drug was developed, which has since been
used by other investigators as a marker to determine the activity of drug-
metabolizing enzymes in vivo. We identified 4-hydroxyantipyrine and its
sulfate conjugate as metabolites of antipyrine. We also observed that anti-
pyrine distributed in the same manner as body water. Because of this prop-
erty, antipyrine has been used for the measurement of body water. Another

AN UNEXPECTED LIFE IN RESEARCH 5

analgesic we studied was aminopyrine. We found that this drug was de-
methylated to aminopyrine and N-acetylated to N-acetylaminopyrine. Many
of the drugs whose fate Brodie and I studied were later used by many
investigators as substrates for the microsomal drug-metabolizing enzymes:
aminopyrine for N-demethylation, phenacetin for O-dealkylation, and aniline
for hydroxylation. Together with Jack Cooper, we developed a method for
measuring the anticoagulent dicoumerol in plasma. In a study on the disposi-
tion of dicoumerol in humans, an exceedingly wide difference in the plasma
levels of this drug was found, suggesting genetic differences in drug metabo-
lism.

MOVE TO THE NATIONAL HEART INSTITUTE

Because I did not have a doctoral degree, I realized that I would have little
chance for advancement in any hospital attached to an academic institution. |
had neither the inclination nor the money to spend several years getting a
PhD, so I decided to join the National Heart Institute as a research chemist. In
1949, Shannon was chosen as the director of the newly organized National
Heart Institute in Bethesda, and he offered me a position. Also coming to the
National Institutes of Health (NIH) at that time were many members of the
Goldwater staff—Brodie* Udenfriend, Berliner, Kennedy, and Bowman.

At the National Heart Institute from 1950 to 1952, I collaborated with
Brodie and his staff on the metabolism of analgesics and adrenergic blocking
agents and the actions of ascorbic acid on drug metabolism. After a while, |
became dissatisfied with working with a large team and was allowed to work
independently. The first problem I chose was an examination of the physio-
logical disposition of caffeine in man. Very little was known about the
physidlogical disposition and metabolism of this widely used compound. A
method for measuring caffeine in biological material was developed, and the
plasma half-life and distribution were determined (3). Because of my work on
analgesics and caffeine, I was delighted to be elected without a doctorate as a
member of the American Society of Pharmacology and Experimental Ther-
apeutics in 1953. K. K. Chen and Steve Brodie were my sponsors.

At that time, I became intrigued with the sympathomimetic amines. In
1910, Barger and Dale reported that numerous B-phenylethanolamine de-
rivatives simulated the effects of sympathetic nerve stimulation with varying
degrees of intensity and precision, and they coined the term sympathomimetic
amines. Sympathomimetic amines such as amphetamine, mescaline, and
ephedrine also produced unusual behavioral effects. In 1952, very little
information concerning the metabolism and physiological disposition of these
amines was known. Because of my experience in drug metabolism, I decided
6 AXELROD

to undertake a study on the fate of ephedrine and amphetamine. In retrospect,
this was an important decision.

The first amine that I studied was ephedrine. Ephredrine, the active princi-
ple of Ma Huang, an herb used by ancient Chinese physicians, was introduced
to modern medicine by Chen and Schmidt in 1930. I soon found that
ephedrine was transformed in animals by two pathways (demethylation and
hydroxylation) to yield metabolic products that had pressor activity. Various
animal species showed considerable differences in the relative importance of
these two metabolic routes. The next sympathomimetic amines I examined
were amphetamine and methylamphetamine. These compounds were shown
to be metabolized by a variety of metabolic pathways including hydroxyla-
tion, demethylation, deamination, and conjugation. Marked species varia-
tions in the transformation of these drugs were also observed.

THE DISCOVERY OF THE MICROSOMAL DRUG
METABOLIZING ENZYMES

‘When amphetamine was given to rabbits, it disappeared without a trace. This
puzzled me, so I decided to look for enzymes that metabolized this drug. I had
no experience in enzymology, but there were many outstanding enzymolo-
gists in Building 3 on the NIH campus where my laboratory was located.
Gordon Tomkins, who occupied the lab bench next to mine, offered me good
advice. Gordon had the capacity for demystifying enzymology and told me
that all I needed to start in vitro experiments was a method for measuring
amphetamine, an animal liver, and a razor blade. I did my first in vitro
experiment with rabbit liver in January 1953. When rabbit liver slices were
incubated in Krebs Ringer-buffer solutions with amphetamine, the drug was
almost completely metabolized. Upon homogenization of the rabbit liver,
amphetamine was not metabolized unless cofactors such as DPN (NAD),
TPN (NADP), and ATP were added. I then decided to examine which
subcellular fraction was responsible for transforming amphetamine. Hoge-
boon and Schneider had just described a reproducible method for separating
the various subcellular fractions by homogenizing tissue in isotonic sucrose
and subjecting the homogenate to differential centrifugation. After separation
of nuclei, mitochondria, microsomes, and the cytosol, none of these fractions
were able to metabolize amphetamine, even in the presence of added cofac-
tors. However, when the microsomes and cytosol were combined, amphet-
amine rapidly disappeared upon the addition of DPN, TPN, and ATP. At that
time Bert La Du, a colleague at the NIH, observed that the demethylation of
aminopyrine in a dialyzed rat liver whole homogenate required TPN. In a
subsequent experiment I found that amphetamines were metabolized in a
dialyzed preparation of microsomes and cytosol in the presence of TPN, but

AN UNEXPECTED LIFE IN RESEARCH 7

not DPN or ATP. However, when the microsomes and cytosol were sepa-
rately incubated, little or no drug was metabolized, despite the addition of
TPN. I realized then that I was dealing with a unique enzymatic reaction.

Before I went further, I decided to identify the metabolic products of
amphetamine produced when the combined microsomes and cytosolic frac-
tion were incubated with TPN. One of the possible metabolic pathways might
be deamination, leading to the formation of phenylacetone. After incubation
of amphetamine with the above preparations, phenylacetone and ammonia
were identified. These results indicated that amphetamine was deaminated by
an oxidative enzyme requiring TPN either in the microsomes or cytosol to
form phenylacetone and ammonia. Because of its properties and the structure
of the substrate, it was apparent that this enzyme differed from another
deaminating enzyme, monoamine oxidase.

Where was the enzyme located, in the microsomes or the soluble super-
natant fraction? An approach that I used to locate the enzyme was to heat each
fraction for a few minutes at 55°C, a temperature that would destroy heat-
sensitive enzymes. When the cytosol was heated to 55°C and then added to
unheated microsomes and TPN, amphetamine was deaminated. When the
microsomes were heated and added to the cytosol fraction together with TPN,
amphetamine was not metabolized. This was a crucial experiment, which
demonstrated that a heat-labile enzyme that deaminated amphetamines was
localized in the microsome$ and that the cytosol provided factors involving
TPN necessary for this reaction.

Bernard Horecker, then working in Building 3, prepared several substrates
for the TPN-requiring dehydrogenase for his classic work on the pentose
phosphate pathway. He generously supplied me with these substrates, which I
could test on my preparations. I found that the addition of glucose-6-
phosphate, isocitric acid, or phosphogluconate acid, together with TPN, to
unwashed microsomes transformed amphetamines. A reaction common to
these substrates is the generation of TPNH, suggesting that the enzymes in the
cytosol fraction were reducing TPN. Incubating microsomes with a TPNH-
generating system using glucose-6-phosphate and glucose-6-phosphate de-
hydrogenase resulted in the deamination of amphetamines. Upon incubation
of chemically synthesized TPNH, microsomes, and oxygen, amphetamine
was deaminated. At about the same time, I also found that ephedrine was
demethylated to norephedrine and formaldehyde by enzymes present in rabbit
microsomes that required TPNH and oxygen. By the end of June 1953, I felt
confident that I had described a new enzyme that was localized in the
microsomes, required TPNH and oxygen, and could deaminate and de-
methylate drugs. I reported these findings at the 1953 fall meeting of the
American Society of Pharmacology and Experimental Therapeutics (4, 5).

After the description of the TPNH-requiring microsomal enzymes that
8 AXELROD

deaminated amphetamine and demethylated ephedrine, several members of
the Laboratory of Chemical Pharmacology at the NIH described similar
enzyme systems that could metabolize other drugs by a variety of pathways,
N-demethylation of aminopyrine (La Du, Gaudette, Trousof, and Brodie),
oxidation of barbiturates (Cooper and Brodie), and the hydroxylation of
aniline (Mitoma and Udenfriend) (6). In a study of the N-demethylation of
narcotic drugs that I made soon after, it became apparent that there were
multiple microsomal enzymes that required TPNH and O, (7). Research on
the microsomal enzymes (now called cytochrome-P450 monooxygenases) has
expanded enormously and has had a profound influence in biomedical sci-

ences, ranging from studies of metabolism of normally occurring compounds .

to carcinogenesis. In retrospect, the discovery of the microsomal enzymes is
among the best work I did.

Brodie and I were struck by the findings of investigators at Smith Kline &
French that SKF525A, a compound with little pharmacological action of its
own, prolonged the duration of action of a wide variety of drugs. We
conjectured that the compound might exert its effects by inhibiting the
metabolism of drugs. The effects of SKF525A on the metabolism of ephed-
rine in dogs and on the metabolism and duration of action of hexabarbital
were examined. We found that SKF252A slowed the demethylation of ephed-
rine in the intact dog. It also prolonged the presence of hexabarbital in the
plasma and the sleeping time in rats and dogs. Thus, the ability of SKF525A
to prolong the action of drugs could be explained by its ability to slow their
metabolism. As soon as the microsomal enzymes were described, it was
observed that SKF525A inhibited this class of enzymes. Subsequently,
SKF525A was widely used as an inhibitor of the microsomal enzymes.

The effect of the microsomal enzymes on the duration of drug action’ was
examined with the collaboration of Gertrude Quinn, a graduate student at
George Washington University, and Steve Brodie. Since sleeping time of
hexabarbital was easy to measure, we chose that drug to make this study.
Cooper and Brodie had found that hexabarbital was metabolized by micro-
somal enzymes in the liver (6). The sleeping time of a given dose of
hexabarbital was compared with its plasma half-life and with the activity of a
liver enzyme preparation using the barbiturate as a substrate in a number of
mammalian species. There were considerable differences in the plasma half-
life, sleeping time, and enzyme activity among the various species (8). A high
correlation was observed between the plasma half-life and sleeping time of the
barbiturate. There was also an inverse relationship between the duration of
action of hexabarbital and its ability to be metabolized by the microsomal
enzymes.

In 1956, I reported that narcotic drugs such as morphine, meperidine, and
methadone were N-demethylated by the liver microsomes requiring TPNH

AN UNEXPECTED LIFE IN RESEARCH 9

and O, (7). Differences in the rate of N-demethylation of various narcotic
drugs in several species made it apparent more than one enzyme was involved
in their N-demethylation. There was also a marked sex difference in the
N-demethylation of narcotic drugs by rat liver microsome enzymes. Micro-
somes obtained from male rats were found to N-demethylate narcotic drugs
much faster than those from female rats. When testosterone was administered
to oophorectomized female rats, the activity of the demethylating enzyme was
markedly increased. Estradiol given to male rats decreased the enzyme activ-
ity. Subsequent work by many investigators found similar sex differences in
microsomal enzyme activity for many metabolic pathways.

While working on the metabolism of narcotic drugs, I observed that the
repeated administration of narcotic drugs not only produced tolerance to these
drugs, but also markedly reduced the ability to N-demethylate them enzymati-
cally (9). There was also a correlation between the rate of demethylation of
opiate substrates and their cross-tolerance to morphine. Opiate antagonists not
only blocked the development of tolerance, but also prevented the reduction
of enzyme activity. On the basis of these observations, a mechanism for
tolerance to narcotic drugs was proposed. In a paper reporting these ex-
periments, the following statement was made: “The changes in enzyme
activity in morphine-treated rats suggests a mechanism for the development of
tolerance to narcotic drugs, if one assumes that enzymes which N-
demethylate narcotic drugs-and the receptors for these drugs are probably
closely related. The continuous interaction of narcotic drugs with the de-
methlyating enzymes inactivates the enzymes. Likewise the continuous in-
teraction of narcotic drugs with their receptors may inactivate the receptors.
Thus, a decreased response to narcotic drugs may develop as a result of
unavailability of receptor sites.” This hypothesis stimulated considerable
critical reaction, mostly negative.

Although I had just described the physiological disposition of caffeine,
demonstrated the variety of metabolic pathways of amphetamine and ephed-
rine, and independently described the microsomal enzymes and their role in
drug metabolism, it was difficult for me to obtain a promotion to a higher rank
at the National Heart Institute because I had no doctorate. I decided to get a
PhD degree at George Washington University, since few courses were re-
quired if a candidate already had an MS degree. However, it would be
necessary to take demanding comprehensive examinations in several subjects.
Paul K. Smith, then Chairman of Pharmacology, accepted me as a graduate
student in his department. He allowed me to submit my work on the metabo-
lism of sympathomimetic amines and the microsomal enzyme for my disserta-
tion. I took a year off to attend courses at George Washington University, and
I found going back to school pleasant and challenging. A few of the medical
students did better than I did in the pharmacology examinations. On one
10 AXELROD

occasion a multiple-choice question on antipyrine, a compound on which I
published several papers, was asked, and I gave the wrong answer. After a
year’s study, I passed a tough comprehensive examination, and my thesis The
fate of phenylisopropylamines was accepted. In 1955, at the age of 42 years, I
received my PhD.

SETTING UP A LABORATORY AT THE NATIONAL
INSTITUTE OF MENTAL HEALTH

While studying for my PhD, I was invited by Edward Evarts to set up a
Section of Pharmacology in his Laboratory of Clinical Sciences at the Nation-
al Institute of Mental Health (NIMH). To get started on my new position at
the NIMH I took a few afternoons off my classes at George Washington
University to do laboratory work. I thought that a study of the metabolism and
distribution of LSD would be an appropriate problem for my new laboratory
in the NIMH. LSD was then used as an experimental drug by psychiatrists to
study abnormal behavior. Bob Bowman at the NIH was in the process of
building a spectrofluorometer. He was kind enough to let me use his ex-
perimental model, which allowed me to develop a very sensitive fluorometric
assay for LSD. This made it possible to measure the nanogram amounts found
in brain and other tissues. This instrument later became the well-known
Aminco Bowman spectrofluorometer. The availability of this instrument
made it possible for many laboratories to devise sensitive methods for the
measurement of endogenous epinephrine, norepinephrine, dopamine, and
serotonin in brain and other tissues. These newly developed methods for
biogenic amines were crucial in the subsequent rapid expansion in neuro-
transmitter research.

Just before I left the Heart Institute, I read a report in the literature that
uridine diphosphate glucuronic acid (UDPGA) was a necessary cofactor for
the formation of phenolic glucuronides in a cell-free preparation of livers.
Jack Strommiger, a biochemist then at the NIH, and I discussed the possible
mechanism for the enzymatic synthesis of UPDGA. We suspected that it
would arise from the oxidation of uridine diphosphate glucose (UDPG) by
either TPN or DPN. We obtained a sample of UDPG from Herman Kalckar
and did a preliminary experiment in which I measured the disappearance of
morphine in guinea pig liver. When morphine was incubated with guinea pig
liver microsomes and soluble fraction with DPN and UDPG, morphine was
metabolized; TPN had no effect. When either DPN, UDPG, soluble fraction,
or liver was omitted, the disappearance of morphine was negligible. After a
period of incubation during which the mixture was heated in IN HCl, the
morphine that disappeared was recovered. These experiments suggested that
morphine was enzymatically conjugated in the presence of UDPG and DPN,

AN UNEXPECTED LIFE IN RESEARCH i

presumably by the formation of UDPGA followed by morphine glucuronide. I
had little time to continue this problem because I was in the process of getting
my PhD. Strominger and coworkers then went on to purify an enzyme UDPG
dehydrogenase that formed UPDGA from UDPG and DPN.

After completion of my PhD, I returned to the glucuronide problem in my
new laboratory at the NIMH. As expected from my preliminary experiment
with morphine, I found that morphine and other narcotic drugs formed
glucuronide conjugates by an enzyme present in liver microsomes that re-
quired UDPGA. Working together, Joe Inscoe, a graduate student at George
Washington University, and I showed that glucuronide formation could be
induced by benzpyrene and 3-methylcholanthrene.

The work on glucuronide conjugation led to a study on the role of glucuron-
ic acid conjugation on bilirubin metabolism. Rudi Schmid, then at the NIH,
made the interesting observation that bilirubin was transformed to a glucuro-
nide. Schmid and I then went on to describe the enzymatic formation of
bilirubin glucuronide by enzymes in the liver requiring UDPGA. This con-
jugating enzyme served as a mechanism for inactivating bilirubin. This led to
an interesting clinical observation concerning a defect in glucuronide forma-
tion. In congenital jaundice there is a marked elevation of free bilirubin in the
blood. This suggested to us that something might be wrong with glucuronide
formation in this disease. The availability of a mutant strain of rats (Gunn
rats) that exhibited congenital jaundice made it possible to examine whether
the glucuronide-forming enzyme was defective. We then went on to demon-
strate that these rats showed a marked defect in the ability to synthesize
glucuronides from UDPGA (10). Glucuronide formation was also examined
in humans with congenital jaundice by measuring the rate and magnitude of
plasma acetominophen glucuronide after the administration of the acetomi-
nophen. A defect in glucuronide formation in this disease was demonstrated.

CATECHOLAMINE RESEARCH

When I joined the NIMH, I knew very little about neuroscience. My impres-
sion of neuroscience then was that it was mainly concerned with
electrophysiology, brain anatomy, and behavior. These subjects were to me
somewhat strange and esoteric and concerned with complicated electronic
equipment. I believed that an investigator had to be a gifted experimentalist
and theorist to do research in the neurosciences. Ed Evarts, my lab chief,
assured me that I could work on whatever problem I thought would be likely
to yield new information. The philosophy of Seymour Kety, then head of the
Intramural Programs of the NIMH, was to allow investigators working in the
laboratories of the NIMH to do their research on whatever was potentially
productive and important. Kety believed that without sufficient basic knowl-
12 AXELROD

edge about the life processes, doing targeted research on mental illness would
be a waste of time and money.

Instead of working on a neurobiological problem, I thought it would be best
to work on one that I knew something about, and that might be appropriate to
the mission of the NIMH. I began to experiment on the metabolism and
physiological disposition of LSD and the enzymes involved in the metabolism
of narcotic drugs. I also worked on the enzymatic synthesis of glucuronides
described above.

Although the NIMH administrators were supportive of the type of research
I was doing, I still felt guilty that I was not working on some aspect of the
nervous system or mental illness. Dr. Kety, in a seminar to our laboratory,
gave a fascinating account of the findings of two Canadian psychiatrists. They
reported that adrenochrome produced schizophreniclike hallucinations when
it was ingested. Because of these behavioral effects, they proposed that
schizophrenia could be caused by an abnormal metabolism of epinephrine to
adrenochrome. I was intrigued by this proposal. In searching the literature, I
was surprised to find that little was known about the metabolism of
epinephrine at that time, in 1957. In view of the provocative hypothesis about
the abnormal metabolism of epinephrine in schizophrenia, I decided to work
on the metabolism of epinephrine. Epinephrine was then believed to be
metabolized and inactivated by deamination by monoamine oxidase. Howev-
er, with the introduction of monoamine oxidase inhibitors by Albert Zeller
and coworkers, it was observed that, after the inhibition of monoamine
oxidase in vivo, the physiological actions of administered epinephrine were
still rapidly ended. This indicated that enzymes other than monoamine ox-
idase metabolized epinephrine. A possible route of metabolism of epinephrine
might be via oxidation. I spent several months looking for oxidative enzymes
for epinephrine without any success.

An abstract in the March 1957 Federation Proceedings gave me an impor-
tant clue regarding a possible pathway for the metabolism of epinephrine. In
this abstract, Armstrong and coworkers reported that patients with
norepinephrine-forming tumors (pheochromocytomas) excreted large
amounts of an O-methylated product, 3-methoxy-4-hydroxymandelic acid
(VMA) (11). This suggested that this metabolite could be formed by the
O-methylation and deamination of epinephrine or norepinephrine. The O-
methylation of catecholamines was an intriguing possibility that could be
experimentally tested. A potential methyl donor could be S-adenosylmethio-
nine. That afternoon I incubated epinephrine with a homogenate of rat liver,
ATP, and methionine. I did not have S-adenosylmethionine available, but
Cantoni had shown that an enzyme in the liver could convert ATP and
methionine to S-adenosylmethionine (12). I found that epinephrine was rapid-
ly metabolized in the presence of ATP, methionine, and liver homogenate.

AN UNEXPECTED LIFE IN RESEARCH 13

When either ATP or methionine was omitted or the homogenate was heated,
there was a negligible disappearance of epinephrine. This experiment sug-
gested that epinephrine was O-methylated in the presence of a methyl donor,
presumably S-adenosylmethionine. In a following experiment, I obtained
S-adenosylmethionine and observed that incubating liver homogenate with the
methyl donor resulted in the metabolism of epinephrine. The most likely site
of methylation would be on the meta hydroxy! group of epinephrine to form
3-O-methylepinephrine. I prevailed on my colleague Bernhard Witkop, a
bioorganic chemist, to synthesize the O-methy] metabolite of epinephrine. A
few days later Sero Senoh, a visiting scientist in Witkop’s laboratory, syn-
thesized meta-O-methylepinephrine, which we named metanephrine. After
incubating liver and S-adenosylmethionine, the metabolite formed from epineph-
rine was identified as metanephrine, indicating the existence of an O-methylating
enzyme. The O-methylating enzyme was purified and found to O-methylate
catechols, including norepinephrine, dopamine, L-DOPA, and _ synthetic
catechols, but not monophenols (13). In view of the substrate specificity, the
enzyme was named catechol-O-methyltransferase (COMT). The enzyme was
found to be widely distributed in tissues, including the brain.

Injecting catecholamines into animals resulted in the excretion of the
respective Q-methylated metabolites. We soon identified normally occurring
O-methylmetabolites such as normetanephrine, metanephrine, 3-methoxy
tyramine, and 3-methoxy-4-hydroxyphenylglycol (MHPG) in liver and brain.
As a result of the discovery of the O-methylation metabolites, the pathways of
catecholamine metabolism were clarified (13). Catecholamines were metabo-
lized by O-methylation, deamination, glycol formation, oxidation, and con-
jugation. As as result of these findings, I then considered myself a
neurochemist. This work also gave me a long-lasting interest in methylation
reactions that I describe later. The metabolites of catecholamines, particularly
MHPG, have been used as a marker in many studies in biological psychiatry.

A major problem in neurobiology research is the mechanism by which
neurotransmitters are inactivated. At the time I described the metabolic
pathway for catecholamines in 1957, it was believed that the actions of
neurotransmitters were terminated by enzymatic transformation. Acety!cho-
line was already known to be rapidly inactivated by acetylcholinesterase.
However, when the principal enzymes for the metabolism of catecholamines,
catechol-O-methyltransferase and monoamine oxidase, were almost com-
pletely inhibited in vivo, the physiological actions of injected epinephrine
were rapidly ended. These experiments indicated that there were other mech-
anisms for the rapid inactivation of catecholamines.

The answer to the question of the inactivation of catacholamines came in an
unexpected way. When the metabolism of catecholamines was described,
Seymour Kety and coworkers set out to examine whether or not there was an
14 AXELROD

abnormal metabolism of epinephrine in schizophrenic patients. To carry out
this study, Kety asked the New England Nuclear Corporation to prepare
tritium-labeled epinephrine and norepinephrine of high specific activity. The
first batch of *H-epinephrine that arrived in late 1957 was labeled on the 7
position, which we found to be stable. Kety was kind enough to give me some
of the *H-epinephrine for my studies. I thought it would be a good idea to
examine the tissue distribution and half-life of 7H epinephrine in animals.

At about that time, Hans Weil-Malherbe spent three months in my labora-
tory as a visiting scientist, and together we developed methods for measuring
3H-epinephrine and its metabolites in tissues and plasma. To our surprise,
when *H-epinephrine was injected into cats, it persisted unchanged in the
heart, spleen, and the salivary and adrenal glands long after its physiological
effects were ended. This phenomenon puzzled us. We also found that 3H-
epinephrine did not cross the blood-brain barrier. Just about this time Gordon
Whitby, a graduate student from Cambridge University, came to our labora-
tory to do his PhD thesis. I suggested that he use methods for assaying
3H-norepinephrine similar to those we used for 3H-epinephrine to study its
tissue distribution. As in the case of *H-epinephrine, *H-norepinephrine
remained in organs rich in sympathetic nerves (heart, spleen, salivary gland).
These studies gave us a clue regarding the inactivation of catecholamine
neurotransmitters: uptake and retention in sympathetic nerves.

The crucial experiment that established that catecholamines were selective-
ly taken up in sympathetic neurons was suggested by Georg Hertting from the
University of Vienna, who joined my laboratory as a visiting scientist. In the
next experiment, the superior cervical ganglia of cats were taken out on one
side, resulting in a unilateral degeneration of the sympathetic nerves in the
salivary gland and eye muscles. Upon the injection of 7H-norepinephyine,
radioactive catecholamine accumulated on the innervated side, but very little
appeared on the denervated side (13, 14). This simple experiment clearly
showed that sympathetic nerves take up and store norepinephrine. In another
series of experiments, Hertting and I found that injected *H-norepinephrine
taken up by sympathetic nerves was released when these nerves were stimu-
lated (15). As a result of these experiments, we proposed that norepinephrine
is rapidly inactivated by reuptake into sympathetic nerves. Other slower
mechanisms for the inactivation of catecholamines proposed were removal by
the blood stream, metabolism by O-methylation, and/or deamination at effec-
tor tissue or by liver and kidney.

In 1961, the first postdoctoral fellow, Lincoln Potter, joined my laboratory
via the NIH Research Associates Program. .The NIH Research Associate
Program and the Pharmacology Research Associate Program provided an
opportunity for recent PhD or MD graduates to spend two or three years in
Bethesda doing full-time research. Because of the number of applicants for

AN UNEXPECTED LIFE IN RESEARCH 15

this program, the investigators in the Intramural Program at the NIH would
get the best and brightest postdoctoral fellows. During the past 25 years more
than 60 postdoctoral fellows joined my laboratory to do full-time research.
With one or two exceptions, most of the postdocs who worked in my
laboratory went on to productive careers in research.

When a postdoc joins my laboratory I try to start him on a problem that has
a good chance of success but is not trivial or pedestrian. There is an open and
free exchange of ideas between my postdocs and myself, which makes it
possible to try novel approaches to problems. By the time postdocs are ready
to leave the laboratory, they are independent investigators. I found the
interactions with bright and motivated young people stimulating and highly
conducive to productive and original research.

When Linc Potter joined my laboratory, we directed our attention to the
sites of the intraneural storage of norepinephrine. We suspected that 7H-
norepinephrine, already shown to be taken up by sympathetic neurons, would
label intracellular storage sites. 7H-norepinephrines was injected into rats, and
their hearts were homogenized in isotonic sucrose; then the various sub-
cellular fractions were separated in a continuous sucrose gradient. There was
a sharp peak of radioactive norepinephrine in a fraction that coincided with
endogenous-catecholamines and dopamine B-hydroxylase, the enzyme that
converts dopamine to norepinephrine. The norepinephrine-containing parti-
cles exerted a pressor response only when they were lysed. In another
experiment, 7H-norepinephrine was injected, and the pineal gland, an organ
rich in sympathetic nerve terminals, was subjected to radioautography and
electron microscopy (16). Photographic grains of *H-norepinephrine were
highly localized over dense core-granulated vesicles of about 500 angstroms.
All these experiments indicated that norepinephrine in sympathetic nerves
was stored in small, dense core vesicles.

Subsequent studies with another postdoc, Dick Weinshilboum, showed that
upon stimulation of the hypogastric nerve of the vas deferens, both nore-
pinephrine and dopamine-B-hydroxylase were disscharged from the nerve
terminals. This suggested that norepinephrine and dopamine-f-hydroxylase
were colocalized in the catecholamine storage vesicles of sympathetic nerves
and were then discharged together by exocytosis (17). These findings led us to
the postulation that the released dopamine-f-hydroxylase would appear in the
blood, which was soon confirmed. Later, our laboratory and others found
abnormally low levels of plasma dopamine-B-hydroxylase in familial
dysautonomia and Down’s syndrome, and high levels in patients with torsion
dystonia, neuroblastoma, and certain forms of hypertension.

As soon as it was found that catecholamines could be taken up and
inactivated by reuptake into sympathetic nerve terminals, I and my coworkers
turned our attention to the effect of adrenergic drugs on this process. We
16 AXELROD

designed relatively simple experiments for this study, injecting the drug into
rats and then measuring the uptake of injected *H-norepinephrine in tissues.
Cocaine was the first drug we examined. It had been postulated that cocaine
causes supersensitivity to norepinephrine by interfering with its inactivation.
After pretreatment of cats with cocaine, there was a marked reduction of
3H-norepinephrine in tissues that were innervated by sympathetic nerves after
the injection of the radioactive catecholamine (18). This experiment indicated
that cocaine blocked the reuptake of norepinephrine in nerves and thus
allowed large amounts of the catecholamine to remain in the synaptic cleft and
act on the postsynaptic receptors for longer periods of time. Using a similar
approach, we observed that antidepressant drugs and amphetamine and other
sympathomimetic amines also blocked the uptake of norepinephrine. In an-
other type of experiment, using an isolated, perfused beating rat heart whose
nerves had previously been labeled with 7H-norepinephrine, we found that the
physiological action of sympathomimetic amines, such as tyramine, was
mediated by releasing the norepinephrine from sympathetic nerves (19). After
repeated treatment of the isolated heart with tyramine, the heart rate and
amplitude of contraction were gradually reduced, presumably by the depletion
of the releasable stores of the neurotransmitters. After replenishing the iso-
lated heart with exogenous norepinephrine, the heart rate and amplitude of
contraction of the isolated heart were restored. Amphetamine also released
norepinephrine, and it was later shown by others that the physiological effects
of the amine were due to the release of dopamine.

Most of my early work in catecholamines was done in the peripheral
sympathetic nervous system. Hans Weil-Matherbe and I had found that
catecholamines did not cross the blood-brain barrier. This made it impossible
to study the metabolism, storage, and release of norepinephrine in the brain
by peripheral administration of 7H-norepinephrine. It was Jacques Glowinski,
a visiting scientist from France, who circumvented this problem. He devised a
technique to introduce *H-norepinephrine directly into the brain by injection
into the lateral ventrical. Subsequent experiments showed that 3H-
norepinephrine was mixed with the endogenous catecholamines in the brain.
As in the peripheral nervous system, the *H-norepineprine was found to be
metabolized by O-methylation and deamination. In a series of experiments we
established that 7H-norepinephrine could serve as a useful tool in studying the
activity of brain adrenergic nerves (13).

After labeling the brain adrenergic neurons, Glowinski and I examined the
effect of psychoactive drugs on brain biogenic amines. We found that only the
clinically effective antidepressant drugs block the reuptake of 37H-
norepinephrine in adrenergic nerve terminals (20). This, together with the
observation that monoamine oxidase inhibitors have antidepressant actions
and that reserpine, a depleter of biogenic amines, sometimes causes depres-
sion, led to the formulation of the catecholamine hypothesis of depression

AN UNEXPECTED LIFE IN RESEARCH 17

(21). We also found that amphetamines block the reuptake as well as the
release of *H-norepinephrine in the brain. Other investigators later showed
that paranoid psychosis caused by excessive ingestion of amphetamines is due
to the release of the catecholamine dopamine. One of the reasons that Les
Iversen came to my lab as a postdoctoral fellow was to learn about the brain
and its chemistry. Iversen and Glowinski worked extensively together in my
laboratory on the effects of drugs on the adrenergic system in different areas
of the brain. To conduct this study they devised a method of disection of
various parts of the brain that has become a classic procedure.

For several years our laboratory was concerned with adaptive mechanisms
of the sympathoadrenal axis. One such mechanism, the induction of the
catecholamine’s biosynthetic enzyme, tyrosine hydroxylase, was observed in
an unexpected manner, as often happens in research. Hans Thoenen, then
working in Basel, asked to spend a sabbatical year in my laboratory. He and
Tranzer had observed that injected 6-hydroxydopamine selectively destroys
catecholamine-containing nerve terminals (22). I invited Thoenen to join my
laboratory and bring 6-hydroxydopamine. The first experiment that Thoenen
tried was to examine the effects of the destruction of peripheral sympathetic
nerves on tyrosine hydroxylase, After the injection of 6-hydroxydopamine, as
expected, tyrosine hydroxylase almost completely disappeared from sym-
pathetically innervated nerves. A surprising observation was a marked eleva-
tion of tyrosine hydroxylase ii the adrenal medulla. 6-Hydroxydopamine was
known to cause persistent firing of nerves. We suspected that tyrosine hy-
droxylase was elevated in the adrenal medulla by continuous firing of the
splanchic nerve innervating the adrenals. This supposition was confirmed
when other drugs that caused prolonged nerve firing, such as reserpine and
a-adrenergic blocking agents, also increased tyrosine hydroxylase (23). Sub-
sequent experiments showed that increased nerve firing induced the synthesis
of new tyrosine hydroxylase molecules in nerve cell bodies and the adrenal
medulla in a transsynaptic manner. Similar results were obtained with another
catacholamine biosynthetic enzyme, dopamine-8-hydroxylase (13).

Another regulatory mechanism for catecholamine synthesis was found by
asking the right questions rather than by serendipity. The ratio of epinephrine
to norepinephrine in the adrenal medulla was known to be dependent on how
much of the medulla was enveloped by the adrenal cortex. In species in which
the cortex is separated from the medulla, norepinephrine is the predominant
catecholamine, while in species in which the medulla is surrounded by the
adrenal cortex, the methylated catecholamine, epinephrine, is by far the major
amine. Dick Wurtman, a research associated in my laboratory, suggested an
elegant experiment to determine the role of the adrenal cortex in regulating the
synthesis of epinephrine. He removed the rat pituitary, a procedure that
depleted glucocorticoid in the adrenal cortex, and then measured the effect on
the levels of the epinephrine-forming enzyme, phenylethanolamine-N methyl-
18 AXELROD

transferase (PNMT), in the medulla. I had just characterized PNMT and
found that it was highly localized in the adrenal medulla. The ablation of the
pituitary caused a profound decrease in PNMT in the medulla after several
days (24). The administration of ACTH, a peptide that increases the forma-
tion of glucocorticoids in the adrenal cortex, or the injection of the synthetic
glucocorticoid, dexamethasone, increased PNMT in hypophysectomized rats
almost to normal values.

METHYLTRANSFERASE RESEARCH

After the description of catechol-O-methyltransferase, I became very much

involved with methyltransferase enzymes (25). I spent most of my time at the -

lab bench working on methylating enzymes for many years. Soon after
describing COMT, I turned my attention to the enzymatic N-methylation of
histamine. A major pathway for histamine metabolism occurs via N-
methylation. This prompted a search for a potential histamine-methylating
enzyme. As in the case of other methyltransferases, I suspected that the most
likely methyl donor would be S-adenosylmethionine. To make the identity of
the histamine-methylating enzyme possible, Donald Brown, a postdoc in the
lab of a colleague, and I synthesized ('4C-methyl]-S-aderiosylmethionine
enzymatically from rabbit liver with '*C-methylmethionine and ATP. Be-
cause of its ability to label the O or N groups of potential substrates by the
transfer of 7H-methylmethionine, the availability of '4C-S-adenosylmethio-
nine led to the discovery of a number of methyltransferase enzymes. Hista-
mine N-methyltransferase was soon found and purified and its properties
described. The enzyme is highly localized in the brain, and it also has an
absolute specificity for histamine. Other methyltransferases soon discovered
using ['4C-methyl]-S-adenosylmethionine were PNMT, hydroxyindole
O-methyltransferase, the melatonin-forming enzyme, a protein carboxymeth-
yltransferase, and a nonspecific N-methyltransferase. This latter enzyme was
found to convert tryptamine, a compound normally present in the brain, to
N-N-dimethyltryptamine, a psychotomimetic agent.

These methyltransferase enzymes, together with (7H-methy!]-S-adenosyl-
methionine of high specific activity were used in developing very sensitive
methods for the measurement of trace biogenic amines. We were able to
detect, localize, and measure octopamine, tryptamine, phenylethylamine,
phenylethanolamine, and tyramine in the brain and other tissues. The methyl-
transferases and [°H-methyl]-S-adenosy!methionine also made it possible to
measure norepinephrine, dopamine, histamine, and serotonin in 130 separate
brain nuclei. Because of the sensitivity of the enzymatic micromethods, my
colleagues and I were able to show the coexistence of several neurotransmit-
ters in single identified neurones of Aplysia (26). Later, Thomas Hokfelt,
using immunohistofluorescent techniques, demonstrated the coexistence of
neurotransmitters in many nerve tracts (27).

AN UNEXPECTED LIFE IN RESEARCH 19

THE PINEAL GLAND

I was struck by an article from Aaron Lerner’s laboratory, published in 1958,
that described the isolation of 5-methoxy-N-acetyltryptamine (melatonin)
from the bovine pineal gland, a compound that had powerful actions in
blanching the skin of tadpoles (28). This compound attracted my attention for
two reasons: it had a methoxy group and a serotonin nucleus. The methoxy
group of melatonin had a special attraction for me. Also, at that time,
serotonin was believed to be involved in psychoses because of its structural
resemblance to LSD. I thought it would be fun to spend some time working on
the pineal gland, an organ that was a mystery to me. The best way to start was
to concentrate my efforts on aspects of the problem that 1 was familiar with,
such as O-methylation.

Herbert Weissbach expressed an interest in collaborating with me in work-
ing out the biosynthetic pathway for melatonin. Weissbach had already made
important contributions on the metabolism of serotonin. The availability of
S-adenosyl-L-methionine with a radioactive methyl group provided an op-
portunity to examine whether the pineal gland could form labeled melatonin
from potential precursor compounds. When we incubated bovine pineal
extracts with N-acetylserotonin and ['4C-methyl]-S-adenosyl-L-methionine, a
radioactive product that we soon identified as melatonin was found (29).
Weissbach and I then purified the melatonin-forming enzyme, which we
named hydroxyindole-O-methyltransferase (HIOMT), from the bovine
pineal gland. We also found another enzyme that converted serotonin to N-
acetylserotonin in the rat pineal. From these observations, we proposed that
the synthesis of melatonin in the pineal proceeds as follows: tryptophan >
5-hydroxytrypotophan — serotonin — N-acetylserotonin — melatonin (30).
Irwin Kopin, Weissbach, and I also found that melatonin was mainly meta-
bolized by a microsomal enzyme via 6-hydroxylation. In a study of the
tissue distribution of HIOMT we observed that the enzyme was highly local-
ized in the pineal. This convinced me that the pineal was a biochemically
active organ containing an unusal enzyme and product and was worth further
study.

During 1960-1962 I spent little time doing pineal research. Most of my
efforts were directed towards the biochemistry of catacholamines and the
effect of psychoactive drugs. In 1962, when Wurtman joined my laboratory, I
thought that he should devote most of his time to catecholamine research. As
a medical student Wurtman had already made an important finding that
bovine pineal extracts blocked gonadal growth in rats induced by light.
Although pineal research was not a fashionable subject for research then,
Wurtman and I were caught up by the romance of this organ, so we decided to
spend our spare time working on the pineal. We thought that a good place to
start was the isolation of the gonad-inhibitory factor of the pineal. Neither of
20 AXELROD

us wanted to go through a tiresome isolation and bioassay procedure, and we
decided to take a chance and examine the effects of melatonin. We soon found
that melatonin reduced ovarian weight and decreased the incidence of estrus
in the rat (30).

Wurtman and I turned our attention to the effects of light on the biochemis-
try of the pineal. We found that keeping rats in the dark for a period of time
increased HIOMT activity, compared to those kept in continuous light. This
experiment gave Wurtman and me a biochemical marker to study how light
transmits its message to an internal organ. Ariens Kappers had found that the
pineal is innervated by sympathetic nerves arising from the superior cervical
ganglia. This finding suggested an experiment to determine the effects of light
on the pineal by removing the superior cervical ganglia and examining the
effects of light and dark on the HIOMT. When the superior cervical ganglia
were removed, the effects of light on HIOMT were abolished. This experi-
ment told us that the effects of light on melatonin synthesis were mediated via
sympathetic nerves arising from the superior cervical ganglia.

In 1964, Sol Snyder joined my laboratory as a postdoc, and he too was
‘fascinated by pineal research. Quay had just made an important observation
that the levels of serotonin, a precursor of melatonin in the pineal, are high
during the day and low at night. Snyder and I developed a very sensitive assay
for measuring serotonin in a single pineal. This gave us the opportunity to
study how the serotonin rhythm, which can serve as a marker for the melato-
nin rhythm, is regulated by light in a tiny organ such as the pineal. We found
that in normal rats in continuous darkness, or in blinded rats, the daily
serotonin rhythm in the pineal persisted (31). This indicated that the in-
dolemine rhythms in the pineal were controlled by an internal clock. Keeping
rats in constant light abolished the circadian serotonin rhythm, showirg that
light somehow stopped the biological clock. These experiments were the first
demonstration that the rhythms of indoleamines in the pineal were
endogenous and that they were synchronized by environmental lighting. We
found that the circadian serotonin rhythm was abolished after ganglionectomy
and also after decentralization of the superior cervical ganglion, indicating
that the circadian clock for the serotonin and presumably the melatonin
rhythm resided somewhere in the brain. Wurtman and I published an article in
Scientific American in which we suggested that the pineal serves as a neuroen-
docrine transducer, converting light signals to hormone synthesis via the brain
and noradrenergic nerves (32).

Shein, a psychiatrist at McLean Hospital, Wurtman, who was then at MIT,
and I decided to see whether the rat pineal in organ culture metabolized
tryptophan to melatonin, and it did. This finding provided an opportunity to
examine whether the neurotransmitter of the sympathetic nerve, nore-
pinephrine, could affect the synthesis of melatonin in pineal organ culture.
The addition of norepinephrine to rat pineals in organ culture increased the

AN UNEXPECTED LIFE IN RESEARCH 21

synthesis of melatonin from tryptophan. Shein and Wurtman then showed that
noradrenaline specifically stimulated the B-adrenergic receptor.

For two years after 1970 I did little work on the pineal until Takeo Deguchi,
a biochemist from Kyoto, joined my laboratory. Because interest in receptors
was beginning to grow at that time, we decided that the pineal gland would be
a good model to study the regulation of the @-adrenergic receptor. The
activity of the G-receptor could be determined by measuring changes in
serotonin N-acetyltransferase (NAT). David Klein previously showed that
pineal serotonin N-acetyltransferase had a marked circadian rhythm that was
controlled by a B-adrenergic receptor (33). Deguchi and I devised a rapid
assay for N-acetyltransferase and soon confirmed Klein’s findings. We then
found that the nighttime rise in NAT was abolished by B-adrenergic blocking
agents, reserpine, decentralization, ganglionectomy, and agents that inhibit
protein synthesis (30). This told us that noradrenaline released from sym-
pathetic nerves innervating the pineal gland stimulated the f-adrenergic
receptor, which then activated the cellular machinery for the synthesis of
NAT. Blocking the B-adrenergic receptor with propranolol at night or expos-
ing rats to light also caused a rapid fall of NAT. These results indicated that
unless the B-adrenergic receptor is stimulated by norepinephrine at a relative-
ly high frequency, NAT rapidly decays. We thought that the rapid synthesis
and fall of NAT would provide a useful model to study the molecular events
in receptor linked synthesis. of a specific protein (NAT) leading to the
formation of a hormone (melatonin).

The regulation of supersensitivity and subsensitivity of receptors is an
important biological problem. The rapidly changing pineal NAT provided a
productive approach to study the mechanism of super- and subsensitivity of
the B-adrenergic receptor (30). Procedures that depleted the neuronal input of
noradrenaline in the rat pineal (denervation, constant light, or reserpine)
caused a superinduction of NAT when rat pineals were cultured and treated
with the B-adrenergic agonist, 1-isoproterenol. When pineal 6-adrenergic
receptors were repeatedly stimulated by injections of 1-isoproterenol into rats,
the cultured pineals became almost unresponsive to the B-adrenergic agonist.
In collaboration, my postdoctoral fellows Jorge Romero and Martin Zatz and I
showed that the regulation of NAT and subsequent melatonin synthesis
consists of a complex series of steps involving: B-adrenergic receptor, cyclic
AMP, cyclic GMP, protein kinase, specific activation of mRNA for NAT,
and synthesis of NAT (30). Decreased nerve activity induced by light caused
an increase in receptor number and adenylate cyclase and kinase activity. This
cascade of events then explained why a small change in release of noradrena-
line from nerves causes a large change in pineal NAT. With the onset of
darkness, there is an increase in sympathetic nerve activity that acts on the
supersensitive receptor, cyclase, kinase, etc. This, we believe, considerably
amplifies the signal (norepinephrine) to cause the large nighttime rise in NAT
22 AXELROD

formation. Klein later showed that norepinephrine acting on an a-adrenergic
receptor further amplified the NAT levels.

THE LAST TEN YEARS

Because of space limitations I can only give a brief description of my research
during the past ten years. Most of the research in my laboratory was con-
cerned with how neurotransmitters transmit their specific messages, About
ten years ago Fusao Hirata, a visiting scientist in my laboratory, and '
observed that the occupation of certain receptors stimulated the methylation °
phospholipids. On the basis of these findings we proposed a mechanism or
the transduction of biological signals (34). This proposal generated consider-
able controversy, and the role of phospholipid methylation in signal transduc-
tion still remains to be resolved. Later, with the collaboration of several
postdoctoral fellows, we reported on the interaction of stress hormones
(catecholamines, ACTH, and glucocorticoids) and the multireceptor release
35).
°F ae s onculy retired from government service at the age of 72. The
NIMH allows me to keep my small laboratory and generously supports my
research. With the help of postdoctoral fellows, 1 continue to be actively
engaged in studying transduction mechanisms of neurotransmitters and hor-
mones. We have evidence for a receptor-mediated release of arachidonic acid
and its many metabolites via the activation of a GTP-binding protein linked to
phospholipase A» (36). This pathway promises to be an active area of future
oF Seat Fitzgerald once stated that there are no second acts in American
lives. After a mediocre first act, my second act was a smash. Soefar the third

act has not been so bad.

Literature Cited

1. Brodie, B. B., Axelrod, J. 1948. The
fate of acetanilide in man. J. Pharma-
col, Exp. Ther. en a

. ie, B. B., Axelrod, J. . Th E

7 te of acetophenetidin (phenacetin) in and other foreign compounds. Ann. Rev.
man and methods for the estimation of Biochem. te Th nymatic N-
acetophenetidin and its metabolites in 7. Axelrod, J. 1956. The hi yina “a
biological materials. J. Pharmacol. Exp. Shonen, n arco - a ies .
Ther. 97:58-67 a . . . : 3

3. Axelrod, J., Rochen : 1953. re 8. es etic iin i. 7 gr
fate of caffeine in man and a met ‘or : 8. .. :

imati i i i : bolism of hexobarbiton,
its estimation of biological material. J. ences in metabolism | pa pei
I. Exp. Ther. 107:519-23 amidopyrine, antipyrine ;
4. nied I . 1958. An enzyme for the Biochem. Pharmacol. 1:152-59

zymes. Trends Pharmacol. Sci. 3:383-
86

6. Brodie, B. B., Gillette, J. R., LaDu, B.
1958. Enzymatic metabolism of drugs

inati thomimetica- 9. Axelrod, J. 1956. Possible mechanism
deaminatios areucol Ep Ther. 110:2 of tolerance to narcotic drugs. Science
5. Axelrod, J. 1982. The discovery of 124:263-64

the microsomal drug-metabolizing en- 10. Axelrod, J., Schmid, R., Hammaker, L.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21,

22.

AN UNEXPECTED LIFE IN RESEARCH 23

1957. A biochemical lesion in con-
genital, non-obstructive, non-hemolytic
Jaundice. Nature 180:1426-27

. Armstrong, M. D., McMillan, A. 1957.

Identification of a major urinary metabo-
ie of norepinephrine. Fed. Proc. 16:
1

Cantoni, G. L. 1953. Adenosyl
methionine: A new intermediate formed
enzymatically from |-methionine and
adenosine triphosphate. J. Biol. Chem.
187:439-52

Axelrod, J.: Noradrenaline: Fate and
control of its biosynthesis. In: Les Prix
Nobel. Imprimerieal Royal P. A. Nor-
stedt and Soner, Stockholm, 1971, pp.
189-208; Science 173:598-606, 1971
Hertting, G., Axelrod, J., Kopin, I. J.,
Whitby, L. G. 1967. Lack of uptake of
catecholamines after chronic denerva-
tion of sympathetic nerves. Nature
189:66 :
Hertting, G., Axelrod, J. 1961, The fate
of tritiated noradrenaline at the sym-
pathetic nerve-endings. Nature 192:
172-173 , .
Wolfe, D. E., Potter, L. T., Richardson,
K. C., Axelrod, J. 1962. Localizing
tritiated norepinephrine in sympathetic
axons by electron microscopic auto-
radiography. Science 138:440-42
Weinshilboum, R., Thoa, N. Be: John-
son, D. G., Kopin, I. J., Axelrod, J.
1971. Proportional release of nore-
pinephrine and dopamine-8-hydroxylase
from sympathetic nerves. Science
174:1349-5! ‘ t
Whitby, L. G., Hertting, G., Axelrod,
J. 1960. Effect of cocaine on the disposi-
tion of noradrenaline labelled with tri-
tium. Nature 187:604-5

Axelrod, J., Gordon, E., Hertting, G.,
Kopin, I. J., Potter, L. T. 1962. On the
mechanism of tachyphylaxis to tyramine
in the isolated rat heart. Br. J. Pharma-
col. 19:56-63

Glowinski, J., Axelrod, J. 1964. Inhibi-
tion of uptake of tritiated-noradrenaline
in the intact rat brain by imipramine and
structurally related compounds. Nature
204:13 18-19

Schildkraut, J. J. 1965. The catechola-
mine hypothesis of affective disorders:
A review of the supportive evidence.
Am. J. Psychiatry 122:509-22
Thoenen, H., Tranzer, J. P. 1968.
Chemical sympathectomy by selective
destruction of adrenergic nerve endings
with 6-hydroxydopamine — Naunyn-
Schmiedebergs Arch. Pharmacol.
261:271-88

23.

24,

25.

26.

27.

28.

29.

30.

31.

32.
33.

' 34,

35.

36.

Thoenen, H., Mueller, R. A., Axelrod,
J. 1969, Increased tyrosine hydroxylase
activity after drug induced alteration of
sympathetic transmission. Nature
221:1264
Wurtman, R. J., Axelrod, J. 1966. Con-
trol of enzymatic synthesis of adrenaline
in the adrenal medulla by adrenal corti-
cal steroids. J. Biol. Chem. 241:2301-5
Axelrod, J. 1981. Following the methyl
group. In Psychiatry and the Biology of
the Human Brain, A symposium dedi-
cated to S. S. Kety, ed. S. Matthysee,
pp. 5-14. New York: Elsevier/North-
Holland
Brownstein, M. J., Saavedra, J. M.,
Axelrod, J., Zeman, G. H., Carpenter,
D. O. 1974. Coexistence of several
putative neurotransmitters in single iden-
tified neurons of Aplysia. Proc. Natl.
Acad. Sci. USA 71:4662-65
Hokfelt, T., Johansson, A., Ljungdahl,
A., Lundberg, H. M., Schultzberg, M.
{980. Peptidergic neurons. Nature
284:515-21
Lermer, A. B., Case, J.D., Takahashi,
Y., Lee, T. H., Mori. W. 1958. Isola-
tion of melatonin, the pineal gland factor
that lightens melanocytes. J. Am. Chem.
Soc. 80:2587
Axelrod, J., Weissbach, H. 1961.
Purification and properties of hydroxyin-
dole-O-methyl transferase. J. Biol.
Chem. 236:211-13
Axelrod, J. 1974. The pineal gland: A
neurochemical transducer. Science
184:1341-48
Snyder, S. H., Zweig, M., Axelrod, J.,
Fischer, J. E. 1965. Control of the circa-
dian rhythm in serotonin content of the
rat pineal gland. Proc. Natl. Acad. Sci.
USA 53:301-5
Wurtman, R. J., Axelrod, J. 1965. The
ineal gland. Sci. Am. 213:50-60

lein, D. C., Weller, J. L. 1970. Indole
metabolism in the pineal gland: A circa-
dian rhythm in N-acetyltransferase. Sci-
ence 169:348-53
Hirata, F., Axelrod, J. 1980. Phospholi-
pid methylation and biological signal
transmission. Science 209:1082-90
Axelrod, J., Reisine, T. 1984. Stress
hormones: Their interaction and regula-
tion. Science 224:452~59
Burch, R. M., Luini, A., Axelrod, J.
1986. Phospholipase A, and phospholi-
pase C are activated by distinct GTP-
binding proteins in response to ay-
adrenergic stimulation in FRTLS thyroid
cells, Proc. Natl, Acad. Sci. USA
83:7201-5