BRAIN PEPTIDES AND FEEDING BEHAVIOR 
Jeffrey M. Friedman, M.D., Ph.D., Assistant Investigator 
Numerous experiments in humans and in other 
mammals have suggested that feeding behavior and 
body weight are tightly regulated. These studies 
have suggested that specific central and peripheral 
neural circuits sense and react to both the overall 
nutritional state and recent food intake of an or- 
ganism. From this work, a hypothesis has emerged 
that states that each individual has a "set point" 
that determines how much he or she should weigh. 
Deviations in weight from the set point result in 
compensatory changes in food intake and energy 
expenditure that generally return the individual's 
weight to some genetically determined level. 
This set point hypothesis predicts that the level 
of peripherally synthesized molecules (hormones 
perhaps) reflects the nutritional state of an individ- 
ual and that the levels of these "satiety factors" are 
sensed by feeding control centers in the hypothala- 
mus and elsewhere. The identity of these satiety 
factors, however, remains to be elucidated. Dr. 
Friedman's laboratory is taking two approaches to 
the elucidation of factors involved in the control of 
feeding behavior. The first approach is aimed at the 
molecular cloning of two recessive mouse muta- 
tions, obese (ob) and diabetic (db), that result in 
profound obesity. The obesity in these animals is 
the result of abnormalities in feeding behavior and 
energy expenditure and may reflect a genetic defect 
in the set point. The second approach involves a 
detailed molecular analysis of the regulation and 
function of a candidate satiety factor, the brain-gut 
peptide cholecystokinin (CCK). 
I. Molecular Cloning of the Mouse ob 
and db Genes. 
In 1950, during the course of experiments aimed 
at generating inbred mouse lines, Dr. George Snell 
(Jackson Laboratories) identified a recessive muta- 
tion, obese (ob), that resulted in massive obesity. 
Mutant animals often weigh three times as much as 
their lean litter mates at adulthood. Subsequent ge- 
netic studies localized this mutation to mouse chro- 
mosome 6. In 1964 a second mutation, diabetic 
(db), was identified by Dr. Douglas Coleman; when 
bred on the same genetic background this mutation 
has a phenotype identical to ob. This mutation was 
genetically localized to mouse chromosome 4. On 
the basis of studies employing parabiosis (crossed 
circulation studies) between obese and non-obese 
animals, Dr. Coleman concluded that animals carry- 
ing the ob mutations were missing a circulating fac- 
tor that suppresses appetite and that animals with 
the db mutation were missing a receptor for this sa- 
tiety factor. It has proven difficult, however, to iden- 
tify the ob and db gene products directly, because 
the profound obesity present in these animals re- 
sults in numerous secondary endocrine and bio- 
chemical abnormalities. The primary defect in these 
animals does not appear to be an abnormality in 
CCK metabolism, since Dr. Friedman and his col- 
leagues have shown, in collaboration with Dr. Uta 
Francke (HHMI, Stanford University), that this gene 
is present on mouse chromosome 9. Furthermore, 
Dr. Friedman's group has not found a difference in 
the levels of CCK in the brain and intestine of 
obese and non-obese animals. 
The technology of reverse genetics is being used 
to clone the ob and db genes (in collaboration with 
Dr. Rudolph L. Leibel). The first step in this strategy 
involves the localization of the mutant gene to spe- 
cific regions of the genome, using restriction frag- 
ment length polymorphisms (RFLPs). 
A standard backcross of 500 animals (to date) be- 
tween ob/ob C57BL/6J mice and DBA mice has been 
generated. Polymorphic markers (RFLPs) can be 
easily identified among the progeny of this cross 
and tracked relative to the obese phenotype. One 
marker present on chromosome 6, met, has been 
shown to map to within 3 centimorgans (cM) of 
the ob trait. Additional polymorphic markers pres- 
ent on proximal chromosome 6, including carboxy- 
peptidase A (cpa), grl, and procollagen have been 
mapped among the backcrossed animals to gener- 
ate a molecular map of this mutation. These experi- 
ments have demonstrated that cpa and met flank 
the ob gene and are ~5 cM apart. A similar cross 
has also been established for the db mutation. Dr. 
Friedman and his colleagues have mapped db to a 
position 3 cM distal to the interferon-a gene, 2 cM 
distal to c-jun, and 5 cM proximal to the liver glu- 
cose transporter gene. To be most effective, this ap- 
proach requires the genetic mapping of numerous 
clones so as to saturate the regions in the vicinity of 
ob and db. To generate a bank of clones specific for 
mouse chromosomes 4 and 6, Dr. Friedman derived 
a cell line from a mouse that carried a Robertsonian 
4:6 translocation. Robertsonian translocations are 
centromeric fusions of two chromosomes. This cell 
line is useful because the Robertsonian chromo- 
Continued 
201 
