Full text Figures and data Side by side Abstract eLife digest Introduction Results Discussion Materials and methods References Decision letter Author response Article and author information Metrics Abstract Proper organogenesis depends upon defining the precise dimensions of organ progenitor territories. Kidney progenitors originate within the intermediate mesoderm (IM), but the pathways that set the boundaries of the IM are poorly understood. Here, we show that the bHLH transcription factor Hand2 limits the size of the embryonic kidney by restricting IM dimensions. The IM is expanded in zebrafish hand2 mutants and is diminished when hand2 is overexpressed. Within the posterior mesoderm, hand2 is expressed laterally adjacent to the IM. Venous progenitors arise between these two territories, and hand2 promotes venous development while inhibiting IM formation at this interface. Furthermore, hand2 and the co-expressed zinc-finger transcription factor osr1 have functionally antagonistic influences on kidney development. Together, our data suggest that hand2 functions in opposition to osr1 to balance the formation of kidney and vein progenitors by regulating cell fate decisions at the lateral boundary of the IM. https://doi.org/10.7554/eLife.19941.001 eLife digest The human body is made up of many different types of cells, yet they are all descended from one single fertilized egg cell. The process by which cells specialize into different types is complex and has many stages. At each step of the process, the selection of cell types that a cell can eventually become is increasingly restricted. The entire system is controlled by switching different genes on and off in different groups of cells. Balancing the activity of these genes ensures that enough cells of each type are made in order to build a complete and healthy body. Upsetting this balance can result in organs that are too large, too small or even missing altogether. The cells that form the kidneys and bladder originate within a tissue called the intermediate mesoderm. Controlling the size of this tissue is an important part of building working kidneys. Perens et al. studied how genes control the size of the intermediate mesoderm of zebrafish embryos, which is very similar to the intermediate mesoderm of humans. The experiments revealed that a gene called hand2, which is switched on in cells next to the intermediate mesoderm, restricts the size of this tissue in order to determine the proper size of the kidney. Switching off the hand2 gene resulted in zebrafish with abnormally large kidneys. Loss of hand2 also led to the loss of a different type of cell that forms veins. These findings suggest that cells with an active hand2 gene are unable to become intermediate mesoderm cells and instead go on to become part of the veins. These experiments also demonstrated that a gene called osr1 works in opposition to hand2 to determine the right number of cells that are needed to build the kidneys. Further work will reveal how hand2 prevents cells from joining the intermediate mesoderm and how its role is balanced by the activity of osr1. Understanding how the kidneys form could eventually help to diagnose or treat several genetic diseases and may make it possible to grow replacement kidneys from unspecialized cells. https://doi.org/10.7554/eLife.19941.002 Introduction Organs arise from precisely defined territories containing progenitor cells with specific developmental potential. Distinct progenitor territories often abut one another, and communication at the interfaces between neighboring territories acts to refine their boundaries (Dahmann et al., 2011). This process delineates the final dimensions of each territory and, subsequently, influences the sizes of the derived organs. Boundary refinement is generally thought to be mediated by interplay between opposing inductive and suppressive factors (Briscoe and Small, 2015). In many cases, however, the identification of and interactions among these factors remain elusive. Kidney progenitor cells originate from the intermediate mesoderm (IM), a pair of narrow bilateral stripes within the posterior mesoderm, flanked by lateral mesoderm that gives rise to vessels and blood and by paraxial mesoderm that gives rise to bone, cartilage, and skeletal muscle. The mechanisms that determine the dimensions of the stripes of IM are not fully understood. Several conserved transcription factors are expressed in the IM and are required for its development, including Lhx1/Lim1, Pax2, and Osr1/Odd1 (Dressler and Douglass, 1992; James et al., 2006; Krauss et al., 1991; Toyama and Dawid, 1997; Tsang et al., 2000; Wang et al., 2005). Studies in chick have indicated essential roles for the lateral mesoderm, paraxial mesoderm, and surface ectoderm in regulating the expression of these transcription factors (James and Schultheiss, 2003; Mauch et al., 2000; Obara-Ishihara et al., 1999). Furthermore, TGF-beta signaling acts in a dose-dependent manner to pattern the medial-lateral axis of the posterior mesoderm: for example, low levels of BMP signaling promote IM formation while high levels of BMP signaling promote lateral mesoderm formation (Fleming et al., 2013; James and Schultheiss, 2005). Beyond these insights, the pathways that set the boundaries of the IM and distinguish this territory from its neighbors are largely unknown. Several lines of evidence have suggested that a carefully regulated refinement process is required to sharpen the boundary between the IM and the lateral mesoderm. In chick, mouse, and Xenopus, Osr1 and Lim1 are expressed in both the lateral mesoderm and the IM before becoming restricted to the IM, implying the existence of a mechanism that acts to exclude IM gene expression from the neighboring lateral territory (Carroll and Vize, 1999; James et al., 2006; Mugford et al., 2008; Tsang et al., 2000). Additional data have hinted at an antagonistic relationship between the IM and the blood and vessel lineages (Gering et al., 2003; Gupta et al., 2006): for example, overexpression of vascular and hematopoietic transcription factors (tal1 and lmo2) induces ectopic vessel and blood specification while inhibiting IM formation (Gering et al., 2003). Along the same lines, zebrafish osr1 morphants exhibit disrupted pronephron formation together with expanded venous structures (Mudumana et al., 2008). Despite these indications of interconnections between IM and vessel development, the network of factors that link these processes has not been fully elucidated. Here, we establish previously unappreciated roles for the bHLH transcription factor Hand2 in both IM and vessel formation. Prior studies of Hand2 have focused on its functions in other tissues, including the heart, limb, and branchial arches (e.g. Charité et al., 2000; Fernandez-Teran et al., 2000; Funato et al., 2009; Miller et al., 2003; Srivastava et al., 1997; Yanagisawa et al., 2003; Yelon et al., 2000). Although Hand2 is also expressed in the posterior mesoderm (Angelo et al., 2000; Fernandez-Teran et al., 2000; Srivastava et al., 1997; Thomas et al., 1998; Yelon et al., 2000; Yin et al., 2010), its influence on patterning this tissue has not been extensively explored. Through both loss-of-function and gain-of-function studies, we find that hand2 limits the size of the kidney by repressing IM formation while promoting venous progenitor formation. hand2 is expressed laterally adjacent to the IM, and a set of venous progenitors arise at the interface between the hand2-expressing cells and the IM. Ectopic expression of IM markers within the hand2-expressing territory in hand2 mutants suggests that hand2 establishes the lateral boundary of the IM through direct inhibition of IM fate acquisition. Finally, genetic analysis indicates that hand2 functions in opposition to osr1 to control kidney dimensions. Together, our data demonstrate a novel mechanism for defining territory boundaries within the posterior mesoderm: hand2 represses IM formation to establish its lateral boundary while promoting venous progenitor formation in this region. These important functions of Hand2 help to define the precise dimensions and components of the kidneys and vasculature. Moreover, these findings have implications for understanding the genetic basis of congenital anomalies of the kidney and urinary tract (CAKUT) and for developing new approaches in regenerative medicine. Results hand2 limits pronephron dimensions by repressing pronephron formation Our interest in the role of hand2 during kidney development began with an observation arising from our previously reported microarray analysis of hand2 mutants (Garavito-Aguilar et al., 2010). We compared gene expression profiles at 20 hr post fertilization (hpf) in wild-type embryos and hans6 mutant embryos, which contain a deletion that removes the entire coding region of hand2 (Yelon et al., 2000). Surprisingly, 11 of the 26 transcripts that were increased in hans6 relative to wild-type were expressed in the pronephron, the embryonic kidney. (For a full list of differentially expressed genes, see Garavito-Aguilar et al., 2010.) We therefore sought to understand the effect of hand2 function on the pronephron. The six genes with the most elevated expression in hans6 mutants (Garavito-Aguilar et al., 2010) are all expressed in the pronephron tubules. We examined the expression of two of these genes that are expressed throughout the tubules -- atp1a1a.4, which encodes a subunit of the Na+/K+ ATPase (Thisse et al., 2004), and cadherin17 (cdh17) (Horsfield et al., 2002) -- and observed an increase in tubule width in hans6 (Figure 1A–C, E–G). A similar increase in width of expression was seen for another gene upregulated in hans6 but only expressed in a single tubule segment (Wingert et al., 2007), slc12a3 (Figure 1I–K), which encodes the thiazide-sensitive sodium-chloride cotransporter solute carrier 12a3. Notably, in contrast to its widened expression, the anterior-posterior extent of slc12a3 expression appeared unaltered. Last, by examining expression of lhx1a and pax2a, which mark glomerular precursors at 24 hpf (O'Brien et al., 2011), we found that the populations of glomerular precursors, like the tubules, were expanded in hans6 (Figure 1M–O and Figure 8B,C). Thus, the elevated gene expression detected by microarray analysis in hans6 mutants corresponded with broadened expression of genes throughout the pronephron. Figure 1 Download asset Open asset hand2 inhibits pronephron formation. (A–P) Dorsal views, anterior to the left, of pronephron schematics (A, E, I, M), wild-type embryos (B, F, J, N), hans6 mutant embryos (C, G, K, O), and hand2-overexpressing embryos (D, injected with hand2 mRNA; H, L, P, carrying Tg(hsp70-hand2-2A-mcherry), abbreviated hs:hand2) at 24 hpf. In schematics (A, E, I, M), colored regions correspond to area of pronephron gene expression. In situ hybridization demonstrates that atp1a1a.4 (A–D) and cdh17 (E–H) are expressed throughout the pronephron tubules, slc12a3 (I–L) is expressed in the distal late segments of the pronephron tubules, and lhx1a (M–P) is expressed in the glomerular precursors (arrows, N), as well as overlying spinal neurons (asterisks, N). Compared to wild-type (B, F, J, N), gene expression is expanded in hans6 mutants (C, G, K, O) and reduced in hand2-overexpressing embryos (D, H, L, P). Of note, injection of a hand2 translation-blocking morpholino caused effects on pronephron formation similar to those seen in hans6 mutants (data not shown). Scale bars represent 100 μm. (Q, R) Transverse sections through wild-type (Q) and hans6 mutant (R) pronephron tubules at 24 hpf. Dashed lines outline the tubule and asterisks indicate individual tubule cells. (S, T) Bar graphs indicate the average number of tubule cells per cross-section (S) and the average tubule area per cross-section (T) in wild-type and hans6 mutant embryos; error bars indicate standard deviation. Asterisks indicate statistically significant differences compared to wild-type (p<0.0001, Student’s t test; n = 18). https://doi.org/10.7554/eLife.19941.003 Figure 1—source data 1 Number of tubule cells per cross-section. Number of tubule cells observed in transverse sections from embryos at 24 hpf were quantified at three different anterior-posterior levels along the pronephron, with two sections analyzed per level in each embryo. Three animals for each genotype (wild-type and hans6) were examined. Average number of cells and standard deviation are represented in Figure 1S. https://doi.org/10.7554/eLife.19941.004 Download elife-19941-fig1-data1-v3.xlsx Figure 1—source data 2 Area of pronephron tubule in cross-section. Pronephron tubules in transverse sections from embryos at 24 hpf were measured at three different anterior-posterior levels along the pronephron, with two sections analyzed per level in each embryo. Three animals for each genotype (wild-type and hans6) were examined. Average tubule area and standard deviation are represented in Figure 1T. https://doi.org/10.7554/eLife.19941.005 Download elife-19941-fig1-data2-v3.xlsx To determine if this broadened expression was due to an increase in cell number, we next analyzed the structure of the pronephron in more detail. We found that the number of cells seen in cross-section of the tubule was increased in hans6 (Figure 1Q–S). This increase in cell number was observed at multiple regions along the anterior-posterior axis, suggesting that the entire pronephron is expanded. Concomitant with this increase in the number of cells, the total tubule area in cross-section was increased in hans6 mutants (Figure 1T). Together with the expanded expression of pronephron genes, these findings implicate hand2 in inhibiting the accumulation of pronephron cells. Does hand2 simply prevent excessive expansion of the pronephron or is it potent enough to repress pronephron formation? To differentiate between these possibilities, we examined the effects of hand2 overexpression on the pronephron. Injection of hand2 mRNA inhibited pronephron formation, as assessed by atp1a1a.4 expression (Figure 1D). To bypass any potential effects of overexpression on gastrulation, we also utilized our previously characterized Tg(hsp70:FLAG-hand2-2A-mCherry) transgenic line to drive hand2 expression at the tailbud stage (Schindler et al., 2014). Heat shock-induced overexpression resulted in pronephron defects comparable to those caused by mRNA injection, as assessed by atp1a1a.4, cdh17 and slc12a3 expression (Figure 1H,L and Figure 3A,B). Furthermore, like tubule gene expression, lhx1a and pax2a expression in glomerular precursors was dramatically inhibited by hand2 overexpression (Figure 1P and data not shown), again emphasizing the broad effect of hand2 function on pronephron formation. In contrast to these effects of hand2 overexpression on the pronephron, most other regions of expression of these markers, such as expression of atp1a1a.4 in mucus-secreting cells and expression of lhx1a in spinal neurons, were unaffected (Figure 1D,P). Thus, the effect of hand2 overexpression seems to reflect its particular impact on pronephron development, as opposed to a general influence of hand2 on the expression of each of these genes. Of note, however, we did observe a dramatic reduction in otic vesicle expression of both atp1a1a.4 and pax2a (data not shown), suggesting a shared susceptibility to hand2 overexpression in the otic vesicles and the pronephron. Taken together, our loss-of-function and gain-of-function studies show that hand2 is necessary to constrain the size of the pronephron, likely through an ability to repress pronephron formation. hand2 limits intermediate mesoderm dimensions by repressing intermediate mesoderm formation We next sought to determine the origin of the effect of hand2 on the pronephron, and we hypothesized that the pronephron defects observed in hans6 mutants might reflect a requirement for hand2 to limit IM dimensions. Indeed, using two established IM markers, pax2a and lhx1a, we observed that the width of the IM was expanded in hans6 mutants and hand2 morphants (Figure 2A,B and Figure 2—figure supplement 1A,B), at stages shortly after gastrulation (i.e. at the 6, 10 and 12 somite stages). In contrast, there was no substantial difference in the length of the IM seen in wild-type and hans6 mutant embryos (Figure 2A,B). Furthermore, compared to wild-type embryos, hans6 mutants typically had ~50% more Pax2a+ IM cells (Figure 2D,E,G). This accumulation of IM cells did not appear to be associated with heightened proliferation within the IM (Figure 2—figure supplement 2). Thus, our data suggest that hand2 restricts the initial formation of the IM. Figure 2 with 2 supplements see all Download asset Open asset hand2 inhibits IM production. (A–C) Dorsal views, anterior to the left, of the posterior mesoderm at the 11 somite stage. In situ hybridization depicts normal expression of pax2a in the IM (arrows) of wild-type embryos (A), widened expression in hans6 mutants (B), and lack of expression in hand2-overexpressing embryos (hs:hand2) (C). Expression in the spinal neurons (asterisk) is unaffected by altered hand2 function. Scale bar represents 100 μm. (D–F) Pax2a immunofluorescence in the posterior mesoderm of wild-type (D), hans6 mutant (E), and hs:hand2 (F) embryos at the 12 somite stage. Dorsal views, anterior to the left, are three-dimensional reconstructions of flat-mounted embryos from which the yolk and anterior tissues have been dissected away. Scale bar represents 100 μm. (D'–F') Magnification of 250 µm long regions from (D–F) used for quantification of the number of Pax2a+ cells in wild-type (D'), hans6 mutant (E'), and hs:hand2 (F') embryos. White dots indicate Pax2a+ nuclei. Intensity of staining varied from strong (for example, green arrows) to weak (for example, yellow arrows). Scale bar represents 25 μm. (G) Bar graph indicates the average number of Pax2a+ cells per 100 µm of IM in wild-type, hans6, and hs:hand2 mutant embryos; error bars indicate standard deviation. Asterisks indicate a statistically significant difference compared to wild-type (p<0.0001, Student’s t test; n=13 for wild-type, n=10 for hans6, and n=19 for hs:hand2). https://doi.org/10.7554/eLife.19941.006 Figure 2—source data 1 Pax2a+ cells in wild-type, hans6, and hs:hand2 intermediate mesoderm. The number of Pax2a+ cells was quantified on the indicated dates of analysis. For wild-type and hans6 embryos, representative 250 μm long regions of IM were analyzed, while for hs:hand2 embryos, 500 μm long regions of IM were analyzed. For hs:hand2 embryos, the IM on both the left and right sides of the embryo were analyzed independently when the dissection and preparation of the sample allowed. All values were normalized to represent the number of cells per 100 μm. Average number of Pax2a+ cells per 100 μm and standard deviation are represented in Figure 2G. https://doi.org/10.7554/eLife.19941.007 Download elife-19941-fig2-data1-v3.docx To determine whether hand2 is sufficient to repress IM formation, we examined the effects of hand2 overexpression. Both hand2 mRNA injection and heat shock-induced overexpression of hand2 resulted in loss of IM, as determined by pax2a and lhx1a expression (Figure 2C and Figure 2—figure supplement 1C). More specifically, hand2-overexpressing embryos exhibited phenotypes ranging from a significantly reduced number of Pax2a+ IM cells to a complete absence of Pax2a+ IM (Figure 2F,G). Taken together with our loss-of-function analyses, these results suggest that hand2 limits the dimensions of the IM by repressing its initial specification. To define the time window during which hand2 overexpression can inhibit IM formation, we utilized Tg(hsp70:FLAG-hand2-2A-mCherry) embryos to induce hand2 expression at different times after gastrulation (Figure 3). Unlike overexpression at the tailbud, 2 somite, or 6 somite stages (Figure 3A–D), overexpression at the 10 somite stage failed to inhibit pronephron development (Figure 3E). Furthermore, we found progressively less severe defects as heat shock induction was performed at successively later stages between tailbud and 10 somites. For example, while inhibition at the tailbud or 2 somite stages resulted in the loss of the majority of pronephric atp1a1a.4 expression (Figure 3A,B), inhibition at the 6 somite stage resulted in only a mild reduction (Figure 3D). Overall, our analysis suggests that there is a hand2-sensitive phase of IM specification prior to the 10 somite stage. Figure 3 Download asset Open asset Pronephron development is susceptible to hand2 overexpression prior to the 10 somite stage. (A–E) Dorsal views, anterior to the left, of in situ hybridization for atp1a1a.4 at 24 hpf depict a range of severity of pronephron defects, ranging from absence of the pronephron (A) to unaffected (E). Tg(hsp70:FLAG-hand2-2A-mCherry) embryos were subjected to heat shock at the tailbud, 2 somite, 6 somite, or 10 somite stages, and the consequences on pronephron development were scored at 24 hpf. Percentages indicate the distribution of phenotypes produced by each treatment; the number of embryos examined is in the right-hand column. Heat shock at later stages resulted in more mild loss of atp1a1a.4 expression in the tubule, and heat shock at the 10 somite stage did not disrupt atp1a1a.4 expression in the tubule. Scale bar represents 100 μm. https://doi.org/10.7554/eLife.19941.011 hand2 is expressed beside the lateral boundary of the intermediate mesoderm Considering the strong effect of hand2 function on repressing IM formation, we sought to define the location of hand2 expression relative to the IM. Prior studies had demonstrated bilateral hand2 expression in the posterior mesoderm of zebrafish, mouse, chick, and Xenopus embryos (Angelo et al., 2000; Fernandez-Teran et al., 2000; Srivastava et al., 1997; Thomas et al., 1998; Yelon et al., 2000; Yin et al., 2010), but the precise localization of this expression relative to the IM had not been determined. During the hand2-sensitive phase of IM formation, we found hand2 to be expressed in bilateral regions immediately lateral to the IM (Figure 4A,B). At these stages, hand2 was also expressed lateral to multiple markers of blood and vessel progenitors, including etv2, tal1, and gata1 (Figure 4C and data not shown). However, a gap lies between these markers and the hand2-expressing cells, consistent with the IM residing between the lateral hand2-expressing cells and the medial blood and vessel progenitors (Figure 4C,D). Figure 4 Download asset Open asset hand2 expression in the posterior lateral mesoderm. (A–D, F–I) Two-color fluorescent in situ hybridization (A, C, F, H) and immunofluorescence (B, D, G, I) label components of the posterior mesoderm in dorsal views, anterior to the left, of three-dimensional reconstructions, as in Figure 2D–E. In embryos containing transgenes in which GFP expression is driven by the regulatory elements of hand2 (B, G) or etv2 (D, I), anti-GFP immunofluorescence was used to enhance visualization. Tg(etv2:egfp) expression was also observed in the midline neural tube, as previously reported (Proulx et al., 2010). Scale bar represents 100 μm. (E, J) Schematics depict posterior mesoderm territories, dorsal views, anterior to the left; hand2-expressing cells, IM, and medial vessel/blood progenitors are shown at 2–10 somites (E) and at 11–13 somites (J), together with lateral vessel progenitors. (A–D) hand2 is expressed lateral to the IM at the 2–10 somite stages. Embryos shown are at the 10 somite stage; similar expression patterns were seen at earlier stages. (A, B) hand2 is expressed lateral to the IM markers lhx1a (A) and Pax2a (B). (C) hand2 is expressed lateral to tal1, a marker of blood and vessel progenitors; note the unlabeled gap between expression territories. (D) A marker of vessel progenitors, Tg(etv2:egfp), lies medially adjacent to Pax2a. (F–I) Vessel progenitors arise at the interface between hand2-expressing cells and the IM at the 11–13 somite stages. (F, G) hand2 is expressed lateral to the IM marker pax2a. (H) hand2 is expressed lateral to a second territory of tal1 expression; note the presence of lateral tal1-expressing cells (arrows) immediately adjacent to hand2 expression. (I) The IM lies between two territories of etv2 expression; the more lateral etv2-expressing cells (arrows) are lateral to the IM. https://doi.org/10.7554/eLife.19941.012 We also found that the relationship between hand2 expression, the IM, and blood and vessel progenitors changed after the completion of the hand2-sensitive phase of IM formation. At approximately the 11 somite stage, a second, lateral population of vessel progenitors arises at the interface between the IM and the hand2-expressing cells (Figure 4F–I). These lateral vessel progenitors are likely to be venous progenitors, based on the results of a prior study that suggested that the medial, earlier-forming vessel progenitors contribute to the dorsal aorta and that the lateral, later-forming vessel progenitors contribute to the primary cardinal vein (Kohli et al., 2013). This set of results defines the general location of hand2 expression relative to other territories within the posterior mesoderm (Figure 4E,J). However, these observations do not exclude the possibility of transient overlapping expression at the boundaries of each territory (e.g. overlapping etv2 and pax2a expression). Furthermore, we note that gene expression patterns are not necessarily uniform within each territory (e.g. tal1 and etv2 expression patterns are neither uniform nor equivalent within their territory [Kohli et al., 2013]). Nevertheless, our data suggest that, during the hand2-sensitive phase of IM formation, hand2 may exert its repressive effect on IM formation by constraining the lateral boundary of the IM. Furthermore, our findings suggest the possibility of close interactions between hand2, the IM, and the lateral population of venous progenitors. hand2 promotes lateral venous progenitor development in the posterior mesoderm The appearance of lateral venous progenitor cells at the interface between the IM and the hand2-expressing cells raised the question of whether hand2 regulates the development of these venous progenitors. To address this possibility, we first assessed the early expression of etv2, tal1, and gata1 in the medial population of vessel and blood progenitors, and we observed no differences between wild-type and hans6 mutant embryos at either the 6 or 10 somite stages (Figure 5—figure supplement 1 and data not shown). In contrast, after the 11 somite stage, expression of both etv2 and tal1 in the lateral venous progenitor population was absent in hans6 mutants (Figure 5A,B,F,G). Two-color fluorescent in situ hybridization confirmed the medial location of the remaining territory of etv2 and tal1 expression in hand2 loss-of-function embryos (Figure 5D,E,I,J). Thus, the formation of the lateral venous progenitors, the appearance of which coincides with the end of the hand2-responsive phase of IM formation, requires hand2. Figure 5 with 1 supplement see all Download asset Open asset hand2 promotes vessel progenitor development. (A–C, F–H, K–M) In situ hybridization depicts etv2 (A–C), tal1 (F–H) and gata1 (K–M) expression in wild-type (A, F, K), hans6 mutant (B, G, L) and hand2-overexpressing (C, H, M) embryos; dorsal views, anterior to the left, at the 12 somite stage. (A, F) etv2 and tal1 are expressed in relatively medial and lateral (arrows) territories on each side of the wild-type embryo. In hans6 embryos (B, G), only the medial territory is present. In hand2-overexpressing embryos (C, H), expression of both etv2 and tal1 is increased, but it is not possible to distinguish whether this represents an increase in the medial or the lateral vessel progenitor populations, since no markers exist that distinguish these two groups of progenitors. (K–M) gata1 expression is equivalent in wild-type (K) and hans6 (L) embryos, but it is decreased in hand2-overexpressing embryos (M). Scale bar represents 100 μm. (D, E, I, J) Fluorescent in situ hybridization depicts the relationship of etv2 and pax2a expression in wild-type (D, D') and hand2 morphant (hand2 MO) embryos (E, E'), and the relationship of tal1 and hand2 expression in wild-type (I, I') and hand2 MO (J, J') embryos; dorsal views, anterior to the left, at the 12 somite stage. Medial and lateral (arrows) territories of etv2 expression flank pax2a in wild-type embryos (D, D'). The lateral territory of expression is absent in hand2 morphants (E, E'). The lateral territory (arrows) of tal1 expression is located at the medial border of hand2 expression in wild-type embryos (I, I'), but is absent in hand2 morphants (J, J'). Note that we observed variable thickness of tal1 expression in its medial territory of expression. https://doi.org/10.7554/eLife.19941.013 To gain insight into whether hand2 directly promotes the formation of vessel progenitors, we assessed the consequences of hand2 overexpression. Overexpression of hand2 resulted in an expansion of both etv2 and tal1 expression in the posterior mesoderm (Figure 5C,H). In contrast, hand2 overexpression resulted in a reduction of gata1 expression (Figure 5M). Thus,