IGF binding proteins: Make another little piece of my heart now, baby

Insulin-like growth-factor-binding protein 4 (IGFBP-4) promotes cardiac development by binding to Wnt co-receptors and inhibiting canonical Wnt signaling.

Insulin-like growth-factor-binding proteins (IGFBPs) regulate IGF signaling. IGFBPs also possess IGF-independent activities, but their function in this capacity is not yet well understood. In Nature, Komuro and colleagues now report that IGFBP-4 binds and inhibits components of the canonical Wnt signaling pathway to promote cardiomyocyte differentiation.

IGFBP-4 was identified in a screen to uncover soluble compounds that induced cardiomyocyte differentiation. Exogenous IGFBP-4 induced differentiation, whereas siRNA-mediated knockdown of IGFBP-4 blocked differentiation. This effect was independent of IGF-I and IGF-II, as an IGFBP-4 mutant unable to bind IGF was still able to promote differentiation.

Canonical Wnt signaling is necessary for cardiomyocyte differentiation; therefore, a potential role for IGFBP-4 in modulating Wnt signaling was evaluated. Immunoprecipitation analyses revealed that the carboxy-terminal region of IGFBP-4 bound to the extracellular domain of the Wnt co-receptors Lrp6 (low-density lipoprotein receptor-related protein 6) and Frz8 (Frizzled 8). Furthermore, IGFBP-4 competitively inhibited Wnt3a binding to Frz8 and blocked Frz8/Lrp6–Wnt3a–β-catenin-induced transcription. These results suggest that IGFBP-4 promotes differentiation by blocking Wnt-mediated Frz8/Lrp6 signaling.

The effect of IGFBP-4 on canonical Wnt signaling in vivo was evaluated in developing Xenopus embryos. IGFBP-4 expression was detected in the liver, suggesting that IGFBP-4 may regulate cardiomyocyte differentiation via paracrine signaling. Morpholino oligo (MO)-mediated knockdown of Xenopus IGFBP-4 (XIGFBP-4) blocked formation of the heart, which was rescued by expression of MO-resistant XIGFBP4 or dominant negative LRP6. Furthermore, injection of Lrp6 or Xwnt8 mRNA caused formation of a secondary developmental axis, which was blocked by co-injection of XIGFBP4. These results indicate that IGFBP-4 antagonizes canonical Wnt signaling downstream of LRP6 in vivo.

It is interesting to note that IGFBP-4 is expressed in liver tissue adjacent to the heart and is also detected in the cells that surround cardiomyocytes, but not in cardiomyocytes themselves. A previous study indicated that FGFs secreted by cardiac tissue induce liver development, and these data provide evidence for reciprocal paracrine signaling between the heart and liver. In addition, the IGFBP-1, 2 and 6 proteins can also bind LRP6 and Frz8. As their expression is low in cardiac tissue, it will be important to determine if the IGFBP–LRP/Fzd interaction regulates Wnt signal transduction in other developmental settings.

Emily J. Chenette
Signaling Gateway

Original Reference:
Zhu, W. et al.
IGFBP-4 is an inhibitor of canonical Wnt signalling required for cardiogenesis
Nature 454, 345-349 (2008)
Full text | PDF | Subscribe to Nature

Post-translational modifications: The active TGF-β receptor is sumoylated

Transforming growth factor-β receptor type I is sumoylated, which regulates Smad3 activation and affects invasion and metastasis.

Transforming growth factor-β (TGF-β) signaling is propagated by the heterotetrameric TGF-β receptor, which consists of two type I (TβRI) and two type II (TβRII) serine/threonine kinases. Following ligand binding, TβRII transphosphorylates TβRI, which stimulates additional trans- and autophosphorylation and phosphorylates the downstream effector molecules Smad2/Smad3. TβRI/TβRII are known to undergo ubiquitination to promote their degradation. In Nature Cell Biology, Kang et al. now report that active, phosphorylated TβRI is also sumoylated, which facilitates Smad3 binding and TGF-β signaling.

SUMO (small ubiquitin-like modifier) is a small molecule that is post-translationally added to many nuclear proteins to regulate transcription and nuclear transport. The authors found that TβRI, but not other TGF-β receptor family members, was specifically sumoylated in vitro and in vivo. TGF-β increased sumoylation of the receptor in vivo, suggesting that the active form of the receptor is preferentially sumoylated. Indeed, constitutively active TβRI was sumoylated more efficiently than wild-type TβRI, and pharmacologic inhibition of TβRI kinase activity, or dephosphorylation of TβRI, decreased sumoylation. Similarly, a kinase-dead TβRI^K230R or TβRII^K277R mutation prevented efficient sumoylation.

Site-directed mutagenesis showed that Lys 389, which lies close to the Smad3 binding site, was required for sumoylation, as a TβRI^K389R mutant was not sumoylated. Furthermore, the TβRI^K389R mutant did not robustly phosphorylate Smad3 in vitro and was unable to stimulate transcription of Smad-responsive genes. Lack of sumoylation affected TGF-β signaling in vivo, as Tgfbr1^-/- cells that expressed exogenous TβRI showed decreased proliferation in response to TGF-β, whereas cells that expressed TβRI^K389R were not affected. Sumoylation of TβRI also appears to contribute to invasion and metastasis. HRas-transformed Tgfbr1^-/- mouse embryonic fibroblasts (MEFs) that expressed TβRI^K389R were less invasive in vitro and formed fewer, smaller lung metastases in mice than MEFs expressing wild-type TβRI. Thus, sumoylation of TβRI appears to regulate its biological activities in part by potentiating Smad3 binding. However, it is not yet know which SUMO-conjugating enzymes regulate TβRI sumoylation, or how activation of TβRI regulates this process.

These studies define a role for TβRI sumoylation in regulating invasion and metastatic growth. Interestingly, a TGFBR1 mutant linked to breast and head-and-neck cancer metastases was a poor substrate for sumoylation both in vitro and in vivo. This mutant also showed limited Smad-mediated transcriptional response to TGF-β. Whether this mutation provides a competitive advantage by stimulating invasion and metastasis but disrupting other aspects of TβRI signaling, such as Smad-independent signaling pathways, awaits further investigation.

Emily J. Chenette
Signaling Gateway

Original Reference:
Kang, J. S., Saunier, E. F., Akhurst, R. J. and Derynck, R.
The type I TGF-β receptor is covalently modified and regulated by sumoylation
Nature Cell Biology 10, 654-664 (2008)
Full text | PDF | Subscribe to Nature

Miyazono, K., Kamiya, Y. and Miyazawa, K.
SUMO amplifies TGF-β signalling
Nature Cell Biology 10, 635-637 (2008)
Full text | PDF | Subscribe to Nature Cell Biology

Cell cycle: Spinning a transcription factor web

A cyclin-CDK-independent network involving successive waves of transcription factor activity regulates the periodic expression of a significant proportion of cell cycle genes.

Progression through the cell cycle is controlled by the activity of cyclin-dependent kinases (CDKs) and the periodic expression and activation of transcription factors (TFs). However, genome-wide transcription studies and TF binding analyses have suggested that the cell cycle may actually be regulated by the successive expression of TFs, which acts as an intrinsic oscillatory network. In Nature, Steven Haase and colleagues now report that periodic transcription is an inherent property of the cell cycle and is governed by sequential waves of TF activity.

A budding yeast strain devoid of all S-phase and mitotic cyclins (clb1,2,3,4,5,6) was used to determine the relative contribution of cyclin-CDK complexes versus intrinsic TF networks in periodic transcription. As previously reported, these cells arrested at the G1/S phase transition yet cyclically activated some G1-phase events such as bud emergence. Genome-wide transcription in wild-type and cyclin-mutant cells was analyzed over the course of two cell division cycles. A set of 1,271 genes was transcribed periodically in wild-type cells. Surprisingly, despite arresting at G1/S, 882 of those genes nonetheless maintained cell-cycle periodicity in the cyclin-mutant cells; about half of those genes showed slight changes in amplitude of expression. This observation points to a cyclin-CDK-independent mechanism that maintains periodicity in cycling cells.

To determine whether an intrinsic TF network controls the observed periodic transcription program, the authors developed a synchronously updating Boolean network model. In this model, TFs were arrayed around a schematic of the cell cycle based on their observed peaks of expression. Links between TFs were drawn based on known interactions between one TF and the promoter of a TF subsequently expressed – a TF expressed in G1 would regulate transcription of an S-phase TF, and so on. When endowed with Boolean logic functions, the model predicted five distinct states in one complete cell cycle that were independent of the starting state, thus reflecting the in vivo transcriptional program. Furthermore, the models for wild-type cells and cyclin-mutant cells were virtually identical. This network model supports the conclusion that a TF-based network oscillator can regulate the periodic transcription of a significant proportion of cell cycle genes.

Cyclins and CDKs are required for successful cell cycle progression, but these data show that an intrinsic oscillatory network of TFs can regulate some key biological processes in yeast mitosis, such as bud emergence. The detailed nature of the TF clock, and its interplay with the canonical cyclin-CDK control machinery, awaits further investigation.

Emily J. Chenette
Signaling Gateway

Original Reference:
Orlando, D. A. et al.
Global control of cell-cycle transcription by coupled CDK and network oscillators
Nature 453, 944-947 (2008)
Full text | PDF | Subscribe to Nature

Toll-like receptor signaling: IRAK2 stays the course

The interleukin-1 receptor-associated kinase 2 (IRAK2) is dispensable in the early phase of Toll-like receptor signaling but is required for sustained NF-κB activity.

Toll-like receptors (TLRs) are an integral part of the innate immune defense system. TLRs recognize microbial components and associate with adaptor proteins that recruit interleukin-1 receptor-associated kinases (IRAK1, IRAK2, IRAK-M and IRAK4). IRAKs stimulate cytoplasmic signaling cascades to affect NF-κB-mediated transcription of inflammatory cytokines, including interleukin-6 (Il6) and tumor necrosis factor (Tnfa). Overexpression of IRAK2 activates NF-κB, but its precise function in the innate immune response is unknown. In Nature Immunology, Shizuo Akira and colleagues now reveal that the activity of IRAK2 is dependent on an intact kinase domain, and show that IRAK1 and IRAK2 have sequential roles in transducing TLR signaling.

Irak2 knockout mice were generated to assess the role of IRAK2 in TLR signaling. Macrophages from Irak2^-/- mice showed impaired Il6 expression following TLR2 stimulation. Tnfa and cyclooxygenase-2 (Cox2) were induced normally, but their expression unexpectedly declined four hours after TLR2 stimulation. Similarly, NF-κB DNA binding was not sustained beyond four hours after TLR2 stimulation, suggesting that IRAK2 is required for prolonged NF-κB activity.

IRAK1 and IRAK4 both possess intrinsic kinase activity, but whether IRAK2 is catalytically active is unclear. Exogenous expression of kinase-defective IRAK2 in Irak2^-/- macrophages could not rescue IL-6 or TNF-α expression, whereas expression of wild-type IRAK2 permitted normal cytokine induction following TLR2 activation. Furthermore, wild-type, but not kinase-deficient IRAK2 was phosphorylated in vitro. Therefore, an intact kinase domain is required for IRAK2 phosphorylation and function.

Similar to Irak1^-/- mice, IRAK2-deficient mice were able to mount an impaired response to infection, suggesting functional redundancy with another IRAK family member. TLR2 stimulation induced rapid but transient activation of IRAK1. In contrast, phosphorylation and activation of IRAK2 began 30 minutes after TLR2 stimulation and was sustained for up to eight hours. Microarray analysis showed that wild-type and IRAK2-deficient macrophages expressed a similar cluster of TLR2-inducible genes two hours after TLR2 activation, but expression of TLR2-inducible genes was significantly reduced in Irak2^-/- macrophages after eight hours. Macrophages from Irak1/Irak2 double-knockout mice showed severe deficiencies in Tnfa, Il6 and Cox2 expression following TLR2 stimulation. Furthermore, these macrophages did not express the TLR2-inducible gene set.

In this study, Akira and colleagues have shown that IRAK2 induces NF-κB-mediated transcription of inflammatory cytokines following TLR activation, and that IRAK2 activity is dependent on an intact kinase domain. While IRAK1 and IRAK2 have similar functions in the innate immune response, they display discrete temporal activation following TLR stimulation, as IRAK2 is dispensable in the early phase of signaling but required for sustained NF-κB activity.

Emily J. Chenette
Signaling Gateway

Original References:
Kawagoe, T. et al.
Sequential control of Toll-like receptor-dependent responses by IRAK1 and IRAK2
Nature Immunology 9, 684-691 (2008)
Full text | PDF | Subscribe to Nature Immunology

Meylan, E. & Tschopp, J.
IRAK2 takes its place in TLR signaling
Nature Immunology 9, 581-582 (2008)
Full text | PDF | Subscribe to Nature Immunology