Biomedical Engineering Reference
In-Depth Information
potentially more malleable liver progenitors? Unfortunately, in vitro experiments
are rather uninformative because, like most other epithelial cell types subjected to
regular culture conditions, hepatocyte populations rapidly become fibroblastic
when attached to plastic. Hence, the significance of the experiments designed by
Horb and collaborators [
76
]. These investigators generated transgenic frogs in
which the gene Xlhbox8 (the amphibian equivalent to the mammalian Pdx1 [
77
])
was placed under the control of the early liver-specific promoter for transthyretin
(TTR). Of note, the Xlhbox8 gene was ''reinforced'' by the addition of a VP16
fusion domain. VP16 is a powerful transcriptional transactivator from the herpes
simplex virus (HSV) [
78
,
79
], and its inclusion in the cassette obeyed to the
rationale that non-pancreatic cells may lack developmentally relevant molecular
Pdx1 adjuvants that may be needed for it to be fully functional. A green fluorescent
protein (GFP) marker was placed under the control of the pancreatic elastase
promoter and added to the transgenic construction in order to tag cell derivatives in
which the liver-to-pancreas conversion had been successful. Up to 60 % transgenic
tadpoles showed not only GFP, but also insulin, glucagon, and amylase expression
in their liver. Of note, the approach did not work when using a VP16-less Xlhbox8
cassette, which adds weight to the notion that, while necessary, Pdx1 may not be
sufficient by itself to promote pancreatic transdifferentiation. Another important
point worth mentioning is that, because TTR is expressed very early during liver
specification, it could be argued that the expression of Xlhbox8 at that time merely
helped redirect the normal course of development of multipotent progenitors
toward a pancreatic fate. Strictly speaking, this would not be a reprogramming
phenomenon. To address this concern, the authors proceeded to transfect
immortalized human hepatocytes (HepG2) with the same construct. As previ-
ously seen in the transgenic setting, up to 65 % of the transfected cells activated
elastase expression. Of these, nearly 15 % were insulin-positive. Confirming
earlier observations, the hepatic phenotype was found to be lost in a Pdx1-
mediated manner, suggesting that a dedifferentiation step precedes reprogram-
ming toward beta cells. The resulting insulin-producing cells were also PC 1/3,
C-peptide, and glucagon-like receptor 1 (GLP-1)-positive, responded to glucose
stimulation and increased the secretion of insulin following treatment with beta-
cellulin and GLP-1 [
80
].
The first systematic analysis of Pdx1 alone versus ''enhanced'' versions of Pdx1
with VP16 was conducted by Tang and colleagues [
81
], who, in contrast with the
results previously reported by the team that developed the transgenic Xlhbox8-
VP16 frogs, were successful at reprogramming rat hepatic cells into pancreatic-
like beta cells both with Pdx1 and Pdx1-VP16. Insulin-expressing cells generated
through both types of constructs were able to correct diabetes in mice. In the short
term, however, the VP16 version of the gene proved slightly superior.
Additional studies showed the synergistic effect of combining Pdx1 with other
pancreatic transcription factors such as BETA2/NeuroD, Ngn3, or MafA [
82
-
88
],
once more confirming that Pdx1 may require the help of other molecular partners
to accomplish full transdifferentiation. An interesting study by Wang and collab-
orators [
82
] added an unexpected twist to the story by describing the inability of
