Few recent studies are available reporting on the biochemical changes of the differentiating cardiomyocyte, specifically preceding the initiation of heartbeats and circulatory function. Evidence, relating to timing of human placenta development, as well as hypoxia requirements for in vitro culture of mouse embryos at early gestational stages, suggested that the early developing heart may primarily rely on glycolytic metabolism for its energy supply, a necessary adaptation for the environmental hypoxia that exists during early embryogenesis. We analyzed immunohistochemically relative hypoxia levels of the heart fields using a reliable marker for hypoxia EF5 (Koch and Evans, 2003; Mahy et al., 2003). Additionally, we localized immunohistochemically during early cardiomyocyte differentiation a member of the monocarboxylate cotransporter (MCT-) family, MCT-4, beginning with cell sorting of the nonmyogenic mesoderm from the cardiomyogenic during epithelialization of the anterior bilateral cardiogenic mesoderm (Linask, 1992, 2003; Linask et al., 1997) and ending at stage 12, when a beating, looping heart is present.
Monocarboxylates such as lactate and pyruvate have an important role in cellular metabolism. Lactic acid is produced as the end product of glycolysis. All tissues become dependent on glycolysis during conditions as hypoxia and ischemia. Lactic acid transport across the plasma membrane is fundamental for the metabolism and pH regulation of all cells, by removing lactic acid produced by glycolysis and allowing uptake by those cells that utilize it as a respiratory fuel, as does the heart. The monocarboxylate transporters (MCTs) catalyze the proton-linked transport of metabolically important monocarboxylates (Halestrap and Meredith, 2003). The adult rat heart expresses the MCT-1 isoform, but not the MCT-4 (Bonen, 2001). For cells that rely on glycolysis for their normal energy metabolism, MCT-4 appears to be the major isoform (Bonen, 2001). Both MCT-1 and MCT-4 require an ancillary protein known as 5A11/Basigin (also identified as CD147, OX-47, EMMPRIN, neurothelin), which facilitates targeting of MCT-1 and MCT-4 to the plasma membrane, where they remain tightly bound to each other (Kirk et al., 2000). This association appears to be important in determining their activity and location.
Techniques to detect hypoxia in tissues have been developed and are of clinical significance in oncology for specific tumor types. One such technique makes use of the so-called “hypoxic cell chemical markers,” such as 2-2-nitro-1H-imidazol-1yl-N-(2,2,3, 3,3-pentafluoropropyl)-acetamide (known as EF5). EF5 allows relative quantitation of tissue oxygen levels both in vitro and in vivo by immunofluorescence, or by flow cytometry. Because of the importance of spatiotemporal localization within the embryo, we chose to use immunohistochemistry to define hypoxic regions. EF5 is very stable in vivo and is substantially lipophilic (Mahy et al., 2003). EF5 molecules are nitroimidazoles compounds that undergo biochemical reduction, forming covalent adducts to intracellular protein thiols. EF5 reduction and binding are inhibited, as a function of increasing oxygen concentration. EF5 binding to cells is detected by use of monoclonal or polyclonal antibodies against EF5 (Lord et al., 1993).
Detection of Hypoxia
We first determined whether the stage 4 or 5 chick embryo incubated at atmospheric conditions (21% O2) at 38°C shows any EF5 labeling. Control embryos were immunostained with a hapten-competed antibody for the primary step, followed by the Cy3 secondary antibody. This control allows for determination of background levels of fluorescence at these stages. EF5 labeling was readily detectable at stage 4 and 5 (primitive streak stages) in the chick embryo using the EF5 antibody in comparison to control embryos (Fig. 1A,C and controls in Fig. 1B,D), indicating that the cellular microenvironment at these stages just before the cardiac compartment forms is normally relatively hypoxic. Interestingly, when these stage 4 and 5 embryos were sectioned, it was evident that not all of the cells expressed this marker, but the mesoderm within the heart field was a mixture of EF5 expressing and non-expressing cells (Fig. 1C). The reason for this is not presently known. Hapten-competed heart fields showed no fluorescent signal (Fig. 1D). At stage 7 during cardiac cell sorting, all cells within the heart field localize EF5, including dorsal mesoderm, ectoderm, and endoderm (Fig. 1E). EF 5 localizes also in embryonic neural and cardiac tissues at 42 hr. EF5 localization is generally apparent, specifically in the neural tissue and to a much lesser degree in the looping heart at 42 hr (Fig. 1F). By 72 hr, the developing eye in the diencephalon region of the brain (Fig. 1G) shows an especially high level of EF5 indicative of hypoxia. At 72 hr, cardiac tissue showed reduced EF5 signal (not shown).
Spatiotemporal Localization of MCT-4 and 5A11/Basigin
The mRNA for the MCT-4 isoform is expressed by the early differentiating cardiomyocytes beginning with stage 5 (see Fig. 4). It is first seen immunohistochemically during the formation of the bilateral cardiac compartments in the chick embryo at early stage 7 (2 to 3 somite pairs) (sections in Fig. 2A,B, and whole mount in Fig. 2C). The left heart forming region is known to differentiate slightly before the right. This is also apparent in the more prevalent MCT-4 expression apparent in Figure 2B. As the ventral myogenic population of cells sorts out from the dorsal nonmyogenic population, MCT-4 is expressed in the epithelial cardiac cell population in the left heart field (L; arrows Fig. 2B), and at higher levels than in the right (R; Fig. 2A). At this time of cardiac compartmentalization, there is a higher level of localization in the apical part of the cardiomyogenic cell population. After cell sorting, the myogenic cardiac compartment as well as the nonmyogenic dorsal mesoderm, continue to express MCT-4. At the 6-somite stage (Fig. 2D and E), the epithelialized cardiac compartments (shown here at the level of the anterior intestinal portal), as well as the cells of the neural tube, express MCT-4. At stage 11, the myocardial wall of the anterior part of the tubular heart that is beating displays a high level of MCT-4 expression (Fig. 2F), as well as in posterior regions where the bilateral heart compartments are still epithelial sheets (Fig. 2G). Note in Figure 2G that MCT-4 is expressed by both the dorsal nonmyogenic mesoderm that will form the lining of the pericardial cavity, as well as the ventral cardiomyogenic mesoderm. MCT-4 remains highly expressed in the stage 12 beating heart (sections in Fig. 2I,J shown at higher magnification, and the whole mount in Fig. 2K).
The ancillary protein 5A11/Basigin is expressed only at low levels in some cells within the mesoderm of the chick heart fields at stage 5 (Fig. 3A). Neural plate cells are already expressing higher levels. By stage 7, its presence increases in association with the cell surfaces of the cardiomyogenic population at the dorsal boundary of the heart compartment where cell sorting shortly will be discernible (Fig. 3B, section, and Fig. 3C, whole mount of same embryo). This coincides with the apical cell surface localization where MCT 4 expression was apparent in Figure 2A and B. It is also expressed within the underlying endoderm. After the tubular heart has formed at stage 10 (Fig. 3D, section, and Fig. 3E, whole mount), 5A11 is primarily expressed by the cells of the ventral myocardial wall (Fig. 3D) and in the outer convex surface of the beating heart at stage 12 (Fig. 3F and G, whole mount), and in the extraembryonic splanchic mesoderm that is closely associated with the heart. High levels of the ancillary protein are detectable in neural tissue at all stages that were analyzed.
The immunohistochemical results were confirmed using RT-PCR for MCT-4, MCT-1, and 5A11/Basigin in the heart (Fig. 4). Analysis beginning with stage 5 (16 hr post fertilization) to stage 42 (day 16) of development showed consistent expression of MCT-4 (Fig. 4, top row). MCT-1 message (Fig. 4, middle row) is first detectable on day 3 and remains expressed throughout the late fetal stages. Although the signal was weak, 5A11 message (Fig. 4, bottom row) is detectable at stage 5 and continues to be expressed throughout embryonic and fetal periods, as would be expected, when MCT-4 is localized at the plasma membrane. The confirmed presence of MCT-4 protein, a marker of glycolysis, and its chaperone 5A11/Basigin during precardiac mesoderm cell differentiation into cardiomyocytes, substantiate the EF5 evidence and that the differentiating cardiomyocytes possess the metabolic components necessary for functioning in a hypoxic environment.
These studies were initiated to define specific cellular characteristics of cardiac differentiation in the apparent hypoxic environment of early embryogenesis. For example, culture conditions described for rat and mouse embryos indicate that the early primitive streak- and early somite- mammalian embryo requires hypoxic conditions for development (5% oxygen recommended at primitive streak stages); increasing to 20% at 10–15 somites (Cockroft, 1990). The O2 requirement increases to 40% at 20–30 somites, and to 95% when ∼30 or more somites are present (Cockroft, 1990). All gas mixtures contain 5% CO2 to maintain pH and the balance is nitrogen.
Studies relating to placenta formation indicate that the human embryo and human placenta develop in a hypoxic environment during the first trimester (Chen and Aplin, 2003). During the first two months of human gestation, the placenta surrounds the whole gestational sac, the villi contain only a few capillaries, the trophoblastic layer is twice the thickness it will be in the second trimester, and the exocoelomic cavity occupies most of the space inside the gestational sac. The exocoelomic cavity contains no oxygen transport system, but anti-oxidant molecules are present that may provide protection to the embryo from oxidative damage (Jauniaux et al., 2003). Overall, these features provide indirect evidence that the architecture of the human first trimester gestational sac limits embryonic oxygen exposure.
The results reported here using EF 5 localization in the in ovo developing chick embryo indicate the bilateral heart fields, as well as neural plate, are relatively hypoxic. Interestingly, not all the mesoderm cells localized the hypoxia marker EF5 at stages 4 and 5, but they did by stage 7. These results could explain, however, why early embryos are relatively tolerant of low oxygen levels: studies have described effects of transient hypoxia on early chick embryos, e.g., Day 2, 3, and 4 chick embryos in which embryos were exposed to 10% hypoxia for 2, 4, and 6 hr and then incubated further at normoxia (21% O2) until Day 9 (Altimiras and Phu, 2000). The results indicated that an acute hypoxic episode did not have an adverse effect on development when exposed on day 2 or 3. By day 4, however, survivability is largely decreased. In other studies, ambient oxygen was decreased to 10% O2 for a 2- or 4-hr period between day 3 to day 9 in chick heart development in ovo (Akiyama et al., 1999). Young embryos showed a variety of moderate responses to hypoxia, but all survived. Notably there is quite a difference in recovery of hearts after anoxia between stages of chick development (Sedmera et al., 2002): tubular hearts (embryonic day 2) in vitro showed very little response to anoxia. In comparison, trabeculated hearts at embryonic day 5 were no longer able to recover completely. These cited studies would suggest that during early stages of development, i.e., stages 7–10 (2 to 10 somites), development can continue normally at moderate hypoxia and even in low oxygen. By 47 hr of development, levels of EF5 localization in the heart were decreasing; MCT-4 continues to be expressed. In later stage chick embryos, EF5 has been successfully used to monitor changes in hypoxia levels in the outflow tract of the heart during morphogenesis (Ivnitski-Steele et al., 2004; Sugishita et al., 2004).
As we report here, the differentiating cardiomyogenic mesoderm cells express both MCT-4 and 5A11/Basigin early during cardiac cell sorting, as the cardiogenic cell population sorts out from the dorsal nonmyogenic cells to form the ventral, epithelial, cardiac compartment in the bilateral plate regions. This expression continues into late fetal stages and neonatal stages. Our data indicate that MCT-4 is first expressed concomitant with initial steps of cardiac cell differentiation. The initial expression shows left-right differences seemingly associated with the left heart field differentiating slightly ahead of the right. This asymmetric pattern of expression is similar to what we reported for the sodium calcium exchanger NCX-1(Linask et al., 2001). After heart tube rotation, the left side is at the outer curvature of the heart that is also the region that first begins to display contractility. At day 3, both MCT-4 and MCT-1 are present for the remainder of gestation. In the postnatal rat heart, 10 days after birth, both MCT-1 and MCT-4 protein and mRNA continue to be expressed (Hatta et al., 2001). After day 10, MCT-4 is down regulated and no longer detectable, while MCT-1 continues to be expressed in the adult.
These cited studies indicate that cardiomyogenic cells during early stages of development, between stages 6 to 12 somites can function normally at moderate hypoxia and even at low oxygen levels. This means they must have the capability to generate energy through glycolysis. The localization of MCT-4 and 5A11/Basigin indicates that the proteins necessary for glycolysis are already expressed at primitive streak stages. The underlying cellular and molecular mechanisms that regulate MCT expression are poorly understood even in the adult in which most of the studies have been done. It is known that changes in physiological states, resulting in changes in substrate concentrations and/or cell signaling associated with changes in cell metabolism, alter MCT expression. For example, in retinal angiogenesis using retinal explants, MCT-1 was induced by hypoxia and by the addition of exogenous vascular endothelial growth factor (VEGF) (Enerson and Drewes, 2003). Regulation by both pre- and post-translational mechanisms is implicated for MCT-1 expression in skeletal muscle. MCT-4 expression in skeletal muscle, however, appears to be regulated by different signaling mechanisms from MCT-1 (Enerson and Drewes, 2003; Halestrap and Meredith, 2004). There is little data on the regulation of these genes in the embryo. Our data suggests, however, there may be regulation of MCT-4 in early embryos by low oxygen, thus possibly being induced by HIF1α.
5A11/Basigin null mice resulted in viable mice that can breed (Philp et al., 2003). These mice have eye-related defects, however, associated with abnormal photoreceptor cell function and degeneration. It is possible that another isoform of the MCT- family can compensate for MCT-4 activity in the heart of these null mice that is not as reliant on the 5A11 chaperone function as is MCT-4 (or also MCT-1). Alternatively, other basigin family members may compensate for 5A11. For example, GP70 shares the ability of 5A11/Basigin (CD147) to interact with MCT-1 and is strongly and broadly expressed during early stages of embryogenesis (Fan et al., 1998). These possibilities need to be addressed in this transgenic mouse model, as effects on the heart have not been analyzed in the 5A11/Basigin null mice.
Low oxygen concentrations induce an entire spectrum of cellular and systemic responses (Semenza, 2001a, b; Marx, 2004). Cell and organ oxygen concentrations are tightly regulated by pathways that affect the expression and activity of numerous cellular proteins. Sensing and responding to changes in oxygen tension is an important variable in physiology and tissues have developed a number of essential mechanisms by which to respond to low physiological oxygen levels. Hypoxia-inducible factor-1 (HIF-1) is critical in the transcriptional response of cells to hypoxia. HIF1α expression in the chick embryo has been reported to be expressed at the 10-somite stage within the pharyngeal endoderm (Etchevers, 2003), but it may be present earlier. Expression of HIF1α and HIF1β mRNA have been detected in all adult and embryonic mouse and human tissues that have been analyzed.(Semenza, 2001b). It has been estimated that 5% of the human genome comes under HIF1 control, although the exact set of genes regulated varies depending on the cell type (Marx, 2004). Known target genes coming under HIF1 regulation include those involved in glycolysis, including some of the MCTs under certain conditions (Enerson and Drewes, 2003), angiogenesis, cell growth, division, survival, and migration. Similarly, other hypoxia-inducible genes as Cited2 associated with left-right patterning of the early embryo and heart fields (Weninger et al., 2005) and VEGF (Dor et al., 2001) have been documented to be important in heart development. HIF1α knockout mice demonstrate that this gene is necessary for embryonic development and survival. Developmental arrest occurs in HIF1α-deficient mouse embryos by embryonic day (ED) 9 of gestation and they die by ED 10.5 with severe neural tube and cardiovascular defects as well as massive cell death, specifically in the cephalic and branchial arch regions (Iyer et al., 1998; Kotch et al., 1999). We suggest that the initial low hypoxia levels normally present within the heart fields and neural tissue activate HIF1α-inducible genes for normal development.
The fundamental effects of oxygen levels on development, physiology, and disease pathophysiology are now increasingly recognized. Oxygen levels appear to play a critical role in early cardiac cell differentiation and the cardiovascular system in general. Additionally, it has been reported that the adult cardiomyocytes' genetic response to stress is to revert back to expressing developmental genetic patterns associated with heart development (Schwartz et al., 1992). Thus, the early hypoxic developmental history of the differentiating embryonic cardiac cell may underlie the cardiac cellular genetic response to ischemia in the adult myocardium, by their re-expression of genes that were relevant and adaptive in the early embryonic low oxygen environment.
The MCT-4/MCT-1/5A11 expression data using chick embryos is in agreement with the general concept that the early vertebrate embryo and placenta have limited access to oxygen that can only occur through diffusion. Thus, development takes place in a physiologically low O2 environment and cardiac cellular energy demands rely on glycolysis. It is reasonable to conclude that results from hypoxia experiments using vertebrate models, including the chick model, can be extrapolated to the human embryo.
Chick embryos (Gallus gallus) were used in these studies. Fertile eggs (SPAFAS) were obtained from Charles River Laboratories (Wilmington, MA) and incubated in our laboratory at 38°C to the specified stages (staging according to Hamburger and Hamilton, 1951). For in ovo experiments, 100 μl of 10 μM EF5 was injected into the air chamber of chicken eggs at different stages and the eggs were incubated at 38°C for 3 hr after injection.
After removal from the yolk, the embryos were fixed in 4% paraformaldehyde-PBS and prepared for immunostaining. A whole embryo, pre-embedding, immunohistochemical procedure that we routinely use with embedding in plastic has been published in detail (Linask and Tsuda, 2000). The primary antibodies used were: an isoform-specific mouse MCT-4 monoclonal antibody; A511/Bsg antibody was produced in rabbits against an extracellular domain of basigin; the EF5 specific antibody and the hapten-competed antibody were purchased from the Koch laboratory, University of Pennsylvania (Philadelphia, PA). The secondary antibody was anti-mouse IgG or anti-rabbit, conjugated to Cy3 (Jackson ImmunoResearch Laboratory, Co. West Grove, PA). Embryos were sectioned at 2 μm through the heart region and analyzed with a Nikon Optiphot II fluorescence microscope. Digital images were acquired with a Princeton MicroMax cooled CCD camera interfaced with MetaMorph software (Universal Imaging, West Chester, PA) and made into figures using Adobe Photoshop 7.
Total RNA was isolated from HH stages 5, 7, 10, 12, and embryonic days 3, 5, and 16 chicken heart-forming regions or embryonic hearts using STAT60 (Tel-Test, Inc. Friendswood, TX). First-strand cDNA synthesis was carried out for 1 hr at 42°C in 20 μl of reverse transcriptase buffer supplemented with 0.1 mM each dNTP, 4 mM dithiothreitol, 5 U of RNase inhibitor, 1 μl PowerScript reverse transcriptase (ClonTech, Palo Alto, CA), and 500 ng Oligo (dT) 12-18. Polymerase chain reaction was carried out in a volume of 50 μl with PCR buffer and 5 U Taq DNA polymerase (Promega, Madison, WI) with 2.5 mM MgCl2, 0.4 mM each dNTP, 0.2 pmol of each primer, and 4 μl of first-strand cDNA. The following primers were used in the PCR reaction: chicken MCT-4 (accession no. NM204663) forward primer 5′-gccttcgctcatcatgttaa -3′, reverse primer 5′-aacccacttgagaccagctac -3′. Mouse MCT-1 (accession no. BC014777) forward primer 5′-tcagccttccttctttccat -3′, reverse primer 5′-cccttttctgcttctcctcc -3′. Chicken 5A11 (accession no. AY 248696) forward primer 5′-caccatccaaacctctgtcc-3′, reverse primer 5-acgtaggtgccagggtcaac -3′. The amplification sequence consisted of an initial DNA denaturation at 94°C for 2 min followed by 30 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 1 min, and extension at 72°C for 1 min. In addition, a final extension at 72°C for 5 min was performed. In each RT-PCR experiment, 10 μl of each reaction product was separated on the agarose gel.
We gratefully acknowledge Dr. Nancy J. Philp (Thomas Jefferson University, Philadelphia, PA) and Dr. Judith D. Ochrietor (University of Florida, St. Augustine, FL) for generously providing us with the antibodies for chick MCT-4 and for mouse 5A11/Basigin, respectively. Carl Coleman, a senior undergraduate student from Holy Family College in Philadelphia, PA, did some of the initial immunohistochemical localization studies for MCT-4.