We thank Dr. Hakvoort and colleagues for their comments, and we appreciate the opportunity to clarify some of the points they raised. In our recent publication,1 we described the binding of STAT5 purified from liver nuclear extracts to a consensus site in the 5′-upstream enhancer of the glutamine synthetase (GS) gene in close proximity to a lymphoid enhancer factor/T cell-specific factor (LEF/TCF)-binding site. On the basis of deletion experiments in respective reporter constructs we concluded that binding of STAT5 may mediate enhancement of expression of GS, whereas the LEF/TCF site may function as a silencer of growth hormone–mediated enhancement in normal hepatocytes. This interpretation is in line with known effects of growth hormone2 and provides attractive possibilities for the cross-talk with the β-catenin signaling pathway, because LEF/TCF may act as a suppressor in the absence of β-catenin.3
The data presented for the first time by Hakvoort et al. in their letter nicely illustrate the long-known fact that growth hormone does not influence zonation of GS in liver parenchyma.2, 4 Indeed, we do not suggest that growth hormone and STAT5 might be involved in the spatial restriction of GS gene expression, because we recently published evidence that a silencer element in the first intron of the GS gene is responsible for GS zonation5 rather than the main 5′-enhancer. Furthermore, these findings make us also believe that the LEF/TCF site described by us1 may not mediate the important role of β-catenin signaling and of APC (adenomatous polyposis coli) in controlling zonation described by Benhamouche et al.6 Instead, this site may contribute to the strength of GS gene expression,7 and we maintain that growth hormone, acting via STAT5, may also influence GS mRNA levels in this way. At first glance, this view seems to contrast with the semiquantitative data of Hakvoort et al. on mRNA abundance, suggesting that either STAT5a or STAT5b, or both, are dispensable for maintaining almost normal levels of GS mRNA. However, it must be taken into account that the measurements taken by Hakvoort et al. were done in constitutive knockout mice after several weeks of age. Because of the plethora of alternative mechanisms regulating GS expression,7 it cannot be excluded that other regulatory pathways compensate for the absence of STAT5. Hakvoort et al. already referred to a possible redundancy at the STAT5 site. If so, it would be most interesting to know what mechanisms might be involved. Recently, we pointed out possible mechanistic details of the interaction of STAT5, LEF/TCF, and β-catenin, but also stressed that the structure of the GS 5′ enhancer is more complex, containing binding sites for other transcription factors as well.7 Thus, further work is needed to elucidate the full potential of this enhancer that seems to constitute an interesting precipitation point for the influence and cross-talk of several regulatory pathways.