Report

Regulation of Chamber-Specific Gene Expression in the Developing Heart by Irx4

See allHide authors and affiliations

Science  19 Feb 1999:
Vol. 283, Issue 5405, pp. 1161-1164
DOI: 10.1126/science.283.5405.1161

Abstract

The vertebrate heart consists of two types of chambers, the atria and the ventricles, which differ in their contractile and electrophysiological properties. Little is known of the molecular mechanisms by which these chambers are specified during embryogenesis. Here a chicken iroquois-related homeobox gene,Irx4, was identified that has a ventricle-restricted expression pattern at all stages of heart development. Irx4 protein was shown to regulate the chamber-specific expression of myosin isoforms by activating the expression of the ventricle myosin heavy chain–1 (VMHC1) and suppressing the expression of the atrial myosin heavy chain–1 (AMHC1) in the ventricles. Thus, Irx4 may play a critical role in establishing chamber-specific gene expression in the developing heart.

During embryonic development, the ventricles and atria of the heart arise from a single tubular structure (1, 2). Mature atria and ventricles differ in their contractile and electrophysiological characteristics and express distinct sets of genes (3, 4). Most of the known chamber-specific genes encode isoforms of contractile proteins, including the myosin heavy chains and light chains (4). The correct expression of these myosin isoforms is essential for embryonic survival and proper function of the mature heart (5). The mechanisms involved in regulation of these chamber-specific patterns in the developing heart tube are unknown. Here we show that in chick hearts, this process requires the proper regulation of theiroquois-related homeobox gene Irx4.

The Irx4 gene was identified by a low-stringency hybridization screening of a chick embryonic day 6 to 8 (E6–E8) retinal cDNA library, with probes derived from mouse and human EST clones that span the Iroquois homeodomains (6). The predicted open reading frame contains a homeodomain highly homologous to those in the DrosophilaIroquois proteins (7, 8), and it is most closely related to the human IRX4 (83% amino acid identity overall and 70% amino acid identity outside of the homeodomain) (Fig. 1, and supplementary data available atwww.sciencemag.org/feature/data/985642.shl). We have also identified a mouse IRX4 with 71% overall amino acid homology to chick Irx4 (9).

Figure 1

(A) The open reading frame of the chick Irx4 gene predicts a 485–amino acid protein (28) that includes a homeodomain (box), a putative transactivation domain rich in acidic amino acids (solid underline), and an iro box (dotted underline) (8). These features are found in all known Iroquois family members. The GenBank accession number for chick Irx4 is AF091504. Expression patterns of Irx4 in the developing heart at stage 19 (B), E4 (stage 24) (C), and E9 (stage 35) (D). In situ hybridizations on whole mount (B and C) and on a paraffin section of heart (D) are shown. Note that Irx4 is only expressed in the ventricles (V) of the heart, not in the atria (A) or in the distal outflow tract (arrowheads). The section plane in (C) does not include the outflow tract. Scale bar in (C) is common to (B), 100 μm; bar in (D), 0.5 mm.

By in situ hybridization (10), we detected Irx4expression in the retina, a subset of nuclei in the hindbrain, the developing feather buds, and the heart (Fig. 1, B to D) (11). Irx4 was highly expressed in the ventricular myocardium, but expression was absent from the atria or the distal outflow tract in the developing heart. Low levels of expression were observed in the proximal outflow tract. The ventricle-specific expression pattern was observed as early as stage 10 in the prospective ventricular region and persisted in all developmental stages examined (Fig. 1, B to D, and Fig. 2). The expression of the mouse Irx4 gene was similarly restricted to the ventricles in all stages of developing heart and adult heart (9).

Figure 2

Expression patterns of Irx4,AMHC1, and VMHC1 mRNAs in the early chicken heart. All panels show whole-mount in situ hybridizations. Early expression of Irx4 (B, E, andH), VMHC1 (A, D, andG), AMHC1 (C, F, andI) at stage 9+ (A and C), stage 10 (B), stage 10+ (D, E, and F), and stage 13 (G, H, and I). The prospective ventricular (PV) region is anterior to the prospective atrial (PA) region. Note that expression of the Irx4 gene is restricted to the prospective ventricular region as early as stage 10, whereas bothAMHC1 and VMHC1 are expressed throughout the heart tube initially. Chamber-specific expression of AMHC1and VMHC1 began after stage 12. Scale bar, 100 μm.

We next compared Irx4 with other genes known to have a chamber-restricted expression pattern in the developing chick heart.Irx4 expression was first observed at Hamburger-Hamilton (HH) stage 10, when the developing heart is a linear tube. At stage 10,Irx4 expression was already restricted to the middle portion of the heart tube, which corresponds to the prospective ventricles (Fig. 2B). The ventricle-restricted pattern persisted to later stages (Fig. 2, E and H). In contrast to Irx4, ventricle myosin heavy chain–1 (VMHC1) gene expression was detected earlier and in regions of the heart tube destined to become both atria and ventricles (Fig. 2, A and D) (12). As with VMHC1, early atrial myosin heavy chain–1 (AMHC1) gene expression also was not restricted (Fig. 2F). Although previous studies suggested that AMHC1 is expressed only in the prospective atrial region (posterior) at stage 9 (13), we observed posterior and anterior expression of AMHC1 during its onset at late stage 10 to stage 12 (an ∼10-hour window). As development proceeded, AMHC1 RNA became more abundant in the posterior and less abundant in the anterior. The restriction of AMHC1 and VMHC1expression to different regions of the developing heart was obvious by stage 13 (Fig. 2, G and I) when the morphological constriction between atria and ventricles became visible. A similar pattern of expression has been described for the quail homolog of AMHC1,slow MyHC 3 (14).

Two features of the spatiotemporal expression patterns ofIrx4 and myosin heavy chain (MHC) isoform genes are noteworthy. First, Irx4 expression was confined to the prospective ventricular region by stage 10, suggesting that the molecular specification of atria and ventricles occurs by this stage. Second, because the initial expression of AMHC1 andVMHC1 overlaps in the entire heart tube, chamber-specific expression observed later in development is probably achieved by down-regulation of AMHC1 in the ventricles and down-regulation of VMHC1 in the atria. Consistent with this hypothesis, slow MyHC 3 contains a negative regulatory element, mutation of which results in ectopic expression of slow MyHC 3 in ventricular cells (15).

To define the role of Irx4 in cardiac development, we used both gain-of-function and loss-of-function approaches. A replication-competent avian retrovirus (RCAS-Irx4) (16), encoding full-length mouse Irx4cDNA, was injected into the precardiac regions of stage 7-8 embryos (17). Injected embryos were incubated for another 5 days (E6) and analyzed by in situ hybridization. At E6 (stage 29), expression of VMHC1 and AMHC1 is normally completely restricted to ventricles and atria, respectively. However, in hearts infected with RCAS-Irx4 virus, substantial amounts ofVMHC1 mRNA were observed in the atria (Fig. 3A), and the normally high level of atrial AMHC1 mRNA was markedly reduced (Fig. 3B). These phenotypes were observed in more than 90% of the injected hearts (n = 28). Hearts injected with a control virus encoding alkaline phosphatase did not show altered expression ofVMHC1 or AMHC1, demonstrating that the effect of RCAS-Irx4 is due to Irx4 expression, and not to viral infection alone. In addition, VMHC1 was never induced in the distal outflow tract in RCAS-Irx4–infected hearts, even though the infection was throughout the heart, suggesting that Irx4 function is limited to a predetermined field. Thus, misexpression ofIrx4 in the atria up-regulates VMHC1 expression and down-regulates AMHC1 expression.

Figure 3

Functional analyses of Irx4 in heart development. Recombinant retroviruses were injected into the cardiogenic regions at stage 7–8 in ovo. The injected hearts were harvested at E6 for further analyses. In situ hybridization on whole mount or cryosections of hearts were performed with VMHC1(A, F to H) or AMHC1(B, C to E) probes. Results of misexpression of full-length Irx4 are shown in (A) and (B). Compared with the uninjected controls (CTL) (left), hearts infected with RCAS-Irx4 virus (right) showed up-regulation of VMHC1 mRNA [arrow in (A)] and down-regulation of AMHC1 mRNA in the atria [arrow in (B)]. Results of misexpression of the dominant negative Irx4 construct are shown in (C) to (H). The control hearts were injected with the virus encoding the Engrailed repressor domain only (RCAS-enr) [CTL in (C) and (F)]. Hearts infected with RCAS-H+enr showed ectopic induction ofAMHC1 mRNA [arrow in (C)] and down-regulation ofVMHC1 mRNA in the ventricles [arrow in (F)]. The results were confirmed by in situ hybridization on cryosections of hearts injected with RCAS-enr (D and G) or RCAS-H+enr(E and H), probed with AMHC1 (D and E), or VMHC1(G and H). Cross sections through ventricles are shown. Induction ofAMHC1 in the ventricles of RCAS-H+enr–injected hearts was within the myocardium. Scale bar in (B) is common to (A), (C), and (F), 0.5 mm; bar in (D) is common to (E), (G), and (H), 25 μm.

To perturb the function of Irx4 in the ventricles, we introduced a putative dominant negative construct, RCAS-H+enr, by retroviral infection. This construct encodes a fusion protein composed of the chick Irx4 homeodomain and the repressor domain of theDrosophila Engrailed protein. Several studies (18) have shown that fusion of a DNA binding domain such as a homeodomain with the repressor domain of Engrailed can create a protein that interferes with transcriptional activation by the wild-type protein. This creates a dominant negative effect, as well as a potential gain-of-function effect due to active repression. At E6 (stage 29), AMHC1 expression was restricted to atria in uninjected embryos as well as in embryos injected with control viruses encoding either an irrelevant homeodomain fused with the Engrailed repressor domain or the Engrailed repressor domain alone (Fig. 3). In contrast, AMHC1 was abundantly expressed in the ventricular myocardium of the hearts injected with RCAS-H+enr (Fig. 3, C to E). Ventricular expression of VMHC1 was suppressed in >80% of the hearts (n > 40) injected with RCAS-H+enr (Fig. 3, F to H). These results demonstrate that Irx4 function is also required to maintain a ventricular profile of myosin heavy chain gene expression.

Both RCAS-Irx4–injected hearts and RCAS-H+enr–injected hearts displayed grossly normal morphology in both atria and ventricles. There were no detectable abnormalities in trabeculation or in the thickness of the myocardium layer in either chamber (11). This suggests that Irx4 does not control morphological aspects of atrial or ventricular identity. Alternatively, the misexpression experiments may have missed a critical time window, produced insufficient amounts of proteins, or the injected embryos may have died too early for a morphological defect to be manifested. We also examined the expression of endogenousIrx4 in the misexpression experiments by using a probe specific for the 5′ untranslated region. No change of expression was detected in either RCAS-Irx4– or RCAS-H+enr–injected hearts, suggesting that Irx4 does not regulate its own expression (11).

Because the initial expression of AMHC1 andVMHC1 is not chamber-specific, the ectopic expression ofAMHC1 and VMHC1 in the misexpression experiments could have resulted from early derepression of these genes, rather than later reinduction. We therefore examined the infected embryos at earlier stages. At 24 hours after infection (stage 13), when restricted expression was first detected, neither RCAS-Irx4 nor RCAS-H+enr affected the expression of AMHC1 orVMHC1. This was expected because ∼24 hours is required for viral propagation and viral gene expression. At 48 hours after infection, ∼20% of the hearts showed some ectopic induction ofAMHC1 or VMHC1. Thus, Irx4 can affect MHC gene expression after the initial segmental restriction is completed. A minimum delay of 48 hours may also explain the observation that only a subset of cells responded to the viral misexpression (Fig. 3, E and H) even though the hearts were completely infected.

Several transcription factors, Nkx2-5, GATAs, MEF2s, dHAND, and eHAND, have been shown to play critical roles in heart development (2), but none is expressed in the heart in a manner that suggests a role in establishment or maintenance of atrial versus ventricular characteristics. Irx4 is expressed only in the ventricles at all stages of heart development. Moreover, chick, mice, and zebrafish all show ventricle-specific expression of Irx4(9, 19), suggesting that Irx4 has an evolutionarily con- served role in heart development. Although fate mapping studies indicate that atrial and ventricular lineages in the precardiac mesoderm are separated during or shortly after gastrulation (stage 5 in chick) (20), cells are not committed to the atrial or ventricular fates until much later (21). Once committed, the cardiac cells maintain the expression of the proper contractile isoforms even when they are grafted into the opposite cardiac chamber, suggesting that an intrinsic factor (or factors) is responsible for maintaining this chamber-specific expression pattern (22). Our results indicate that Irx4 can alter the MHC expression profile after the chamber-specific pattern is established. Thus, Irx4 may be a maintenance as well as determining factor in the development of the cardiac chambers.

The roles that Irx4 plays in regional specification within the heart tube is reminiscent of the functions of other Iroquoisfamily members in regionalization of the Drosophila wing discs (7), eye disc (8), and neural precursor domain in Xenopus (23). As noted for other family members (7, 24), additional factors may act in concert with Irx4 to specify multiple aspects of the chamber identities. It is intriguing that a homeodomain protein expressed only in the ventricle establishes the myosin isoform profiles in both atrial and ventricular chambers. This implies that the contractile characteristics of the cardiac chamber may be determined by imposition of a ventricular phenotype over a default atrial fate, by the action of Irx4 in the ventricles.

  • * To whom correspondence should be addressed. E-mail: cepko{at}rascal.med.harvard.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article