S. Safarians et al., ECTOPIC G-CSF EXPRESSION IN HUMAN-MELANOMA LINES MARKS A TRANSDOMINANT PATHWAY OF TUMOR PROGRESSION, The American journal of pathology, 150(3), 1997, pp. 949-962
Using a human melanoma/Scid xenograft model with the C8161, M24-met, L
D-1 and other human melanoma lines to investigate spontaneous metastas
is, we made the observation of marked splenomegaly (up to five times n
ormal weight and size) in only those xenografts exhibiting high degree
s of spontaneous metastasis, Evaluation of this revealed the cause to
be massive myelopoiesis due to ectopic granulocyte/colony-stimulating
factor (G-CSF) production by the melanoma cells, Because of these obse
rvations linking G-CSF expression with metastasis of human melanoma, w
e decided to investigate the mechanism of this ectopic production. No
gross amplification or rearrangement of the G-CSF gene could be detect
ed as the basis for the increased transcriptional activity in any of t
hese lines. Human-human somatic cell hybridization studies carried out
between the metastatic C8161 and several different nonmetastatic non-
G-CSF-expressing lines revealed, in addition to metastatic dominance,
3- to 10-fold enhancement of G-CSF transcription and expression in the
fusions compared with C8161 itself. The suggestion of a trans-dominan
t mechanism was further supported by transfection studies with a human
G-CSF promoter-CAT-reporter construct, which revealed 3- to 5-fold in
creased reporter activity in only those melanoma lines and hybrids exp
ressing G-CSF, Furthermore, no obvious autocrine or paracrine effects
of this ectopic G-CSF expression on the melanoma lines' growth or meta
stasis were apparent, as all of the G-CSF-expressing lines lacked the
G-CSF receptor and injections of purified recombinant G-CSF exerted no
stimulatory effects on their tumorigenicity, latency, growth, or meta
stasis in Scid mice. Thus, we advance the hypothesis that G-CST expres
sion is serving as a marker of a more generalized trans-dominant pathw
ay linked to tumor progression and metastasis. This hypothesis has dir
ect relevance to many human cancers where ectopic hormone or growth fa
ctor production occurs with no obvious autocrine or paracrine benefit
to the tumor.