The H19 locus belongs to a cluster of imprinted genes that is linked to the human Beckwith-Wiedemann syndrome. The expression of H19 and its closely associated IGF2 gene is frequently deregulated in some human tumors, such as Wilms' tumors. In these cases, biallelic IGF2 expression and lack of expression of H19 are associated with hypermethylation of the imprinting center of this locus. These observations and others have suggested a potential tumor suppressor effect of the H19 locus. Some studies have also suggested that H19 is an oncogene, based on tissue culture systems. We show, using in vivo murine models of tumorigenesis, that the H19 locus controls the size of experimental teratocarcinomas, the number of polyps in the Apc murine model of colorectal cancer and the timing of appearance of SV40-induced hepatocarcinomas. The H19 locus thus clearly displays a tumor suppressor effect in mice.

PNAS 2008 105:12417-12422; published ahead of print August 21, 2008, doi:10.1073/pnas.0801540105

Engineered mice play an ever-increasing role in defining connections between genotype and phenotypic expression. The potential of magnetic resonance microscopy (MRM) for morphologic phenotyping in the mouse has previously been demonstrated; however, applications have been limited by long scan times, availability of the technology, and a foundation of normative data. This article describes an integrated environment for high-resolution study of normal, transgenic, and mutant mouse models at embryonic and neonatal stages. Three-dimensional images are shown at an isotropic resolution of 19.5 μm (voxel volumes of 8 pL), acquired in 3 h at embryonic days 10.5–19.5 (10 stages) and postnatal days 0–32 (6 stages). A web-accessible atlas encompassing this data was developed, and for critical stages of embryonic development (prenatal days 14.5–18.5), >200 anatomical structures have been identified and labeled. Also, matching optical histology and analysis tools are provided to compare multiple specimens at multiple developmental stages. The utility of the approach is demonstrated in characterizing cardiac septal defects in conditional mutant embryos lacking the Smoothened receptor gene. Finally, a collaborative paradigm is presented that allows sharing of data across the scientific community. This work makes magnetic resonance microscopy of the mouse embryo and neonate broadly available with carefully annotated normative data and an extensive environment for collaborations.

PNAS 2008 105:12331-12336; published ahead of print August 19, 2008, doi:10.1073/pnas.0805747105

In the simplest view of transcriptional regulation, the expression of a gene is turned on or off by changes in the concentration of a transcription factor (TF). We use recent data on noise levels in gene expression to show that it should be possible to transmit much more than just one regulatory bit. Realizing this optimal information capacity would require that the dynamic range of TF concentrations used by the cell, the input/output relation of the regulatory module, and the noise in gene expression satisfy certain matching relations, which we derive. These results provide parameter-free, quantitative predictions connecting independently measurable quantities. Although we have considered only the simplified problem of a single gene responding to a single TF, we find that these predictions are in surprisingly good agreement with recent experiments on the Bicoid/Hunchback system in the early Drosophila embryo and that this system achieves ∼90% of its theoretical maximum information transmission.


PNAS 2008 105:12265-12270; published ahead of print August 21, 2008, doi:10.1073/pnas.0806077105

Purhonen et al. (1) have refuted the data published in >50 reports (2, 3), neglecting to quote key articles or utilize relevant models, and have drawn unsubstantiated conclusions about the contribution of endothelial progenitor cells (EPCs) to tumor angiogenesis that are not supported by their nonquantitative data and superficially executed experiments. Their study (1) is flawed in experimental design and data interpretation. For example, they do not cite their own publication demonstrating the existence of VEGFR2+ EPCs (4) and neglect mentioning clinical validation (5, 6) and acknowledging mouse genetic models (2, 3), which provide convincing evidence for functional incorporation of EPCs into neovessels. Every figure lacks stereoconfocal-microscopic quantification of vessels that are presented as poorly defined longitudinal–linear streaks. Plasma VEGF-A levels were not measured in vivo in mice treated with VEGF-A, questioning their low level of VEGFR2+ EPC detection (3). Indeed, their FACS analysis is inaccurate because of (i) unconvincing CD31/VE-cadherin/VEGFR2 expression detected on MS-1 endothelium used as positive control and (ii) failure to show long-term marrow engraftment of donor-derived hematopoietic and authentic VEGFR2+LacZ+ colony-forming EPCs. APCmin mice develop only obstructive adenomas, rather than adenocarcinomas; therefore, it is an inappropriate model to study EPC incorporation, as Spring et al. (7) (not quoted) demonstrate that EPCs do not contribute to adenomas but contribute only to carcinomas/metastatic tumors. In the parabiotic model, wild-type EPCs compete with GPF+ EPCs, which underestimates EPC recruitment. Finally, study of 6-month-old VEGF-A-loaded Matrigel plugs in mice is impossible because Matrigel plugs are degraded within 2 months, particularly when VEGF-A by itself does not induce neoangiogenesis. No quantification of patent vessels in Matrigel plugs was provided. This article fails to disprove the established role of EPCs in supporting neoangiogenesis in certain tumors (3, 5) and metastatic transition (2)

PNAS 2008 105:E54; published ahead of print August 20, 2008, doi:10.1073/pnas.0804876105