M. García-Barros, J. Devesa, and V.M. Arce
as a prohormone that would need to be cleaved to
low mol wt variants in order to be able to exert its
full range of biological actions (6, 7).
showed identical subcellular localization, although
cleavage pattern obtained was different, and their
specific activity toward GH was 2-fold lower than
that of muscle or adipose tissue homogenates and
was not impeded by the same protease inhibitors.
Interestingly, all these tissues (with the only excep-
tion of thyroid gland) are important target tissues
for GH action. In contrast, GH processing activity
was absent in brain, an organ in which GH actions
are not so clearly documented to the present.
Thirdly, the specificity of GH cleavage also accounts
for the substrate, since only 22K-GH was suscepti-
ble to undergo proteolytic cleavage, while the 20
kDa variant was resistant to it. The 20 kDa GH vari-
ant is originated in the pituitary by alternative splic-
ing of the pre-mRNA, which results in the loss of
part of exon 3. 20K-GH accounts for about 5-10%
of pituitary GH, but its biological role remains un-
known (1). Although the tertiary structure of 20K-
GH has not been completely elucidated, the hor-
mone has a reduced affinity for the GHR, thus sug-
gesting that it has a different conformational fold-
ing than the 22 kDa variant. Although we do not
have any readily explanation for the resistance to
degradation showed by the 20K-GH variant, it is
tempting to speculate that a specific protein folding
is needed to allow the action of enzymes involved
in the breakdown of the 22 kDa GH molecule. In
this context, it is also interesting to point out that
when the 22 kDa variant was denatured prior to in-
cubation with the S9 pellet fraction, a different
cleavage pattern was observed, thus suggesting
the involvement of a different enzymatic activity.
This substrate specificity even exists within 22K-GH,
since only a fraction of this variant was susceptible
to proteolysis by peripheral tissue homogenates,
at least under our experimental conditions. Micro-
heterogeneity of pituitary-derived 22K-GH was rec-
ognized a long time ago, and to the present, two
deamidated, one acetylated and two glycosylated
forms have been reported within the 22 kDa range
(1, 26, 27). Although the physiological significance
of these variants is not yet understood (1), it has
been shown that deamidation of pituitary-derived
GH results in an increased susceptibility of the 22
kDa variant to metabolization by serine-proteases;
a process that can be counteracted by the protease
inhibitor PMSF (28). Hence, it is tempting to spec-
ulate that the presence of microheterogeneity in
22K-GH of recombinant origin may explain the par-
tial breakdown observed in the present work.
Over the years, numerous GH fragments have been
synthesized or generated by proteolytic digestion
of the hormone in the laboratory, and experimen-
tal evidence exists suggesting that the multiple ac-
tions of GH are produced by different parts of the
molecule (6, 7, 17-19). Moreover, GH fragments can
be isolated from the pituitary, although at fairly
abundant concentrations; and from plasma, where
the 22 kDa form may even not be the predominant
GH variant (20, 21). The greater concentration of
GH fragments in plasma than in the pituitary sug-
gests that they originate by proteolytic breakdown
of the hormone in the periphery. The existence of
proteolytic activity toward GH has been reported
for several tissues in vitro, including thyroid gland,
skeletal muscle, liver and adipose tissue (9, 22-24
and the present study). These GH fragments would
recirculate and exert their specific actions through
binding to the GH receptors or to separate recep-
tors, not identified yet (25); or, alternatively, they
could exert their actions by acting directly within
the cell nucleus as proposed for the intact hor-
mone.
Nevertheless, the existence of GH breakdown in
peripheral tissues in vitro does not necessary reflect
a mechanism of GH activation in vivo, since cellular
integrity is not conserved in such experiments, and
the hormone may interact in vitro with enzymes that
would be taken apart in vivo. Therefore, one of the
major concerns arising from this kind of experi-
ments is whether the existence of GH cleavage in
peripheral tissues represents a physiological mech-
anism of generation of active peptides, or is a first
step in a general degradation pathway.
To our knowledge, evidence regarding the physio-
logical importance of GH processing in target tis-
sues is several fold. First, GH fragments with a mol
wt similar to those isolated in vitro can be found in
serum after im administration of the intact hormone
(9). Since GH is very stable in plasma, these mole-
cules are thought to be formed in the peripheral
tissues following receptor-mediated endocytosis
(1). Second, the enzymes involved in GH processing
show a rather specific distribution. Inhibitor studies
provided initial characterization of at least two dif-
ferent types of enzymatic activity. A chymotripsin-
like serine protease activity, localized exclusively to
the 9000 g pellet fraction, was found in skeletal
muscle and adipose tissue homogenates. A similar
enzymatic activity has been previously reported in
muscle homogenates, but also in thyroid gland ho-
mogenates (9). On the other hand, liver enzymes
Finally, the fact that GH cleavage shows important
age-related changes also supports the physiologi-
cal relevance of this process. GH levels undergo
important changes throughout life in both humans
752
Bretschneider
