XB-ART-38824Proc Natl Acad Sci U S A October 28, 2008; 105 (43): 16608-13.
Modulation of potassium channel function confers a hyperproliferative invasive phenotype on embryonic stem cells.
Ion transporters, and the resulting voltage gradients and electric fields, have been implicated in embryonic development and regeneration. These biophysical signals are key physiological aspects of the microenvironment that epigenetically regulate stem and tumor cell behavior. Here, we identify a previously unrecognized function for KCNQ1, a potassium channel known to be involved in human Romano-Ward and Jervell-Lange-Nielsen syndromes when mutated. Misexpression of its modulatory wild-type beta-subunit XKCNE1 in the Xenopus embryo resulted in a striking alteration of the behavior of one type of embryonic stem cell: the pigment cell lineage of the neural crest. Depolarization of embryonic cells by misexpression of KCNE1 non-cell-autonomously induced melanocytes to overproliferate, spread out, and become highly invasive of blood vessels, liver, gut, and neural tube, leading to a deeply hyperpigmented phenotype. This effect is mediated by the up-regulation of Sox10 and Slug genes, thus linking alterations in ion channel function to the control of migration, shape, and mitosis rates during embryonic morphogenesis. Taken together, these data identify a role for the KCNQ1 channel in regulating key cell behaviors and reveal the molecular identity of a biophysical switch, by means of which neoplastic-like properties can be conferred upon a specific embryonic stem cell subpopulation.
PubMed ID: 18931301
PMC ID: PMC2575467
Article link: Proc Natl Acad Sci U S A
Genes referenced: sox10 snai2 kcnq1 kcne1 dct atp6v0c gal.2 trpc2
Antibodies: Kcne1 Ab1 Kcnq1 Ab1 Kcnq1 Ab2
Article Images: [+] show captions
|Fig. 1. KCNE1 overexpression induces hyperpigmentation. (A) Microinjection of KCNE1 mRNA at the one-cell stage induces 32% of embryos grown to stage 45 to exhibit hyperpigmentation compared with controls (<2%). (B) Controls. (C) Hyperpigmented embryos have far more melanocytes in the head but very normal overall development. Yellow rectangle indicates region in which melanocytes were counted (Table 1). (Magnifications: B and C, ×9.)|
|Fig. 2. KCNE1 inhibits KCNQ1 currents and depolarizes embryonic cells. (A) Tail currents were analyzed at −120 mV and normalized to the value followed after the 60-mV depolarizing pulse to estimate the voltage dependence of channel activation (n = 10 or 11). At voltages of 0 mV or less, addition of KCNE1 reduced the currents compared with KCNQ1 alone. (B) KCNQ1 (5 ng of mRNA) was expressed alone or together with 0.2 ng of KCNE1 mRNA. Each lane was loaded with equal amounts of protein (as estimated by Bradford test and Ponceau staining). Four consecutive biotinylation Western blots detected no clear alterations of KCNQ1 protein in the plasma membrane (densitometric analysis of all four gels normalized to the respective KCNQ1 band gave 1.0061 ± 0.1025 for KCNQ1 + KCNE1 injections). Analysis of membrane potential in vivo using DiSBAC voltage reporter dye reveals that KCNE1 mRNA expression depolarizes cells in the neurulating embryo. (C) Histograms showing normalized frequencies corresponding to normalized fluorescence intensities were created for each embryo (n = 37). One example each of the histograms from controls (injected with mRNA encoding Xenopus Ductin, a subunit of the V-ATPase ion pump) and KCNE1-injected embryos are shown. The peak value is the value of relative intensity corresponding to a frequency of 1.0. Also indicated is the half-maximum (frequency = 0.5) position at which width was measured for statistical comparisons. Area is computed as the peak times the width. These two histograms were chosen because their peak and width values are closest to the mean values; however, there was large variation in shapes of the histograms. (D) Bar chart showing the difference between three descriptors of histograms from xDuctin- and KCNE1-injected embryos. Control xDuctin-injected histograms had a mean peak at 0.255 ± 0.093 (mean ± 99% confidence intervals), a mean width of 0.150 ± 0.042, and a mean area of 0.044 ± 0.030, whereas the mean peak for KCNE1-injected embryos was at 0.786 ± 0.134, the mean width was 0.627 ± 0.192, and mean area was 0.552 ± 0.308. Two-tailed t tests revealed highly significant differences for each comparison: peaks, P = 8 × 10−10; widths, P = 2 × 10−6; areas, P = 4 × 10−4. Control n = 15; KCNE1 n = 22.|
|Fig. 3. KCNE1 expression induces a neoplastic-like phenotype in melanocytes. Control larvae sectioned through the brain (A) and tail (A′) possess a small number of melanocytes at the dorsal surface of the neural tube (NT); these cells have the normal rounded morphology (red arrows). In contrast, embryos injected with KCNE1 exhibit higher numbers of melanocytes that spread out into a more dendritic morphology (green arrows), observed in both the brain (B) and the tail (B′, where the melanocytes penetrate the somite and surround the spinal cord). (C) The level of sections for A, A′, B, B′, and D. (D) Melanocytes of KCNE1-injected larvae often invade the dense nervous tissues of the neural tube (green arrow). Blue stain in B, B′, and D indicates lineage label coinjected with KCNE1 mRNA. At 3 months, the ectopic melanocytes colonized the blood vessels (control embryo in E, white arrows, vs. KCNE1-injected in F, red arrows). (Magnifications: A, B, and D, ×12; A′ and B′, ×20; E and F, ×7.5.)|
|Fig. 4. KCNE1-induced hyperpigmentation phenotype is a non-cell-autonomous effect involving up-regulation of Sox10 and Slug expression. (A) Embryos were injected with a mixture of KCNE1 and β-gal mRNAs and were lightly bleached at stage 43, allowing evaluation of melanocytes and clear detection of the β-gal lineage label in the area dorsal to the eyes, which normally has few or no melanocytes (all of the melanocytes in this region are ectopic). The vast majority of the excess cells did not themselves contain the lineage label (B, white arrows indicate lack of β-gal signal; blue arrow indicates cells positive for β-gal), illustrating the non-cell-autonomous mechanism of hyperpigmentation induction by KCNE1. Embryos were injected with KCNE1 + β-gal mRNAs at the one-cell stage (resulting in mosaic expression throughout the embryo), processed for in situ hybridization, and sectioned. Note the ectopic expression of Sox10 (C) and Xslug (D, compare with contralateral side showing very little Sox10 expression on the side where KCNE1 was not injected). Ectopic domains lie adjacent to KCNE1-misexpressing cells. Red signal (and red arrowheads) indicate β-gal lineage label of injected cells. Blue arrows indicate the positive in situ hybridization signal (purple).|
|Fig. 5. A model of KCNQ1/KCNE1 function in embryogenesis. (A) A parsimonious model of the data proposes that KCNE1 modifies the function of KCNQ1, which up-regulates Sox10 and its downstream targets, such as XSlug. These factors are known to be necessary and sufficient for the observed up-regulation of proliferation, change in cell shape, and induction of invasive migration behavior in melanocytes. (B) Embryonic regions expressing KCNE1 induce up-regulation of key transcription factors, such as Sox10 and Slug, in other cells in the neural crest population, which confers upon them a hyperproliferative, invasive phenotype. This illustrates a non-cell-autonomous mechanism by which ion flows are transduced into canonical transcription cascades that control cell behavior. Black dashed lines indicate migration of hyperproliferating offspring of the target (green) cell. The original event takes place in the cell expressing KCNE1 (blue), which does not itself contribute all of the proliferative offspring. The target cell may induce the XSox10 pathway in neighbors by conventional biochemical signals (e.g., BMPs, FGFs, and WNTs). The direct biophysical interaction occurs in the target cell, where KCNE1-mediated loss of membrane polarization is relayed to the downstream signaling machinery by cell-autonomous mechanisms.|
|Fig. S1. Endogenous expression of KCNQ1 mRNA. In situ hybridization with an antisense probe to native Xenopus KCNQ1 was performed on embryos at various stages. (D–F) Embryos were embedded in JB4 medium and sectioned at 30 m after hybridization in whole mount. During neurulation, anterior neural plate cells exhibited a strong signal, and expression also was detected in the neural folds (A) and in the mandibular, hyoid, and branchial neural crest streams (B). At tailbud stage, embryos exhibited strong expression in the tail, branchial arches, face, and dorsal neural tube (C); stain in tailbud (D), neural tube (E), and underlying cells around the cement gland (F) was confirmed in section. The extending tail maintained a focus of expression in under the spinal cord (G). Red arrows indicate positive signal. cg indicates cement gland.|
|Fig. S2. Endogenous expression of KCNE1 mRNA. Expression was not detected during neurulation (A). During organogenesis (B), zygotic transcription was detected in the eye (blue arrowhead), cement gland (yellow arrowhead), otic vesicle (green arrowhead), and hindgut (red arrow). Sections confirmed expression in cells underlying the cement gland (C) and within the eye (D), possibly in the internuclear layer. Red arrows indicate positive signal.|
|Fig. S3. Hyperpigmentation can be induced even after neural crest specification. (A) A wild-type embryo (control) at stage 45. (B) A typical hyperpigmented embryo (induced by KCNE1 overexpression), photographed here without tricaine, which causes the pigment granules in each cell to contract into the center, facilitating counting (used in Fig. 1). Larvae exposed to Chromanol 293B (a specific blocker ofKCNQ1channels) at stage 41 acquired a hyperpigmented phenotype within 6 days of culture, first in the region overlying the hindbrain (C, blue arrow) and then spreading throughout the head (D, red arrows indicate ectopic melanocytes in the periocular region).|
|Fig. S6. KCNE1 expression induces neoplastic-like phenotypes in melanocytes. Embryos exposed to RL-3 (an activator of KCNQ1) express slightly fewer melanocytes in the neural tube (A) and many fewer melanocytes in the tail (A ), compared with Fig. 3 A and A . Closer examination of sections taken through the neural tube (B) revealed that melanocytes of KCNE1-injected larvae often dropped down and invaded the dense nervous tissues of the neural tube (green arrows). Blue stain in B indicates lineage label that was coinjected with KCNE1 mRNA. Sections through the gut indicate the presence of numerous melanocytes (D, arrowheads) in KCNE1-injected larvae, where control embryos have none (C). During the second week of development, the excess melanocytes continue to migrate and spread out. They leave the trunk core and invade the dorsal fin (E, red arrowheads). Occasionally, KCNE1-misexpressing animals exhibited small growths on their outer surface, which often contained a core of melanocytes within (F, green arrowheads). When larvae were grown for 3 months, the hyperpigmentation and metastasis phenotype continued, and the ectopic melanocytes were seen to colonize the blood vessels: controls are shown in G (white arrows indicate lack of melanocytes in blood vessels), whereas KCNE1-injected larvae are shown in whole mount in H (red arrows indicate melanocytes throughout the ventral vessels). Higher magnification (H ) reveals melanocytes completely encircling the vessel. Melanocytes also spread throughout the epidermal layers of the animal (compare control in I with KCNE1-injected shown in J). Several particular internal organ primordia are frequently invaded, such as the pronephros and lateral ganglia (higher-magnification indicated by green arrow, shown in J ). Yellow plus signs indicate tissues from KCNE1-injected embryos; minus signs indicate control tissues. NT, neural tube; LP, lateral plate; vp, velar plate.|
|Fig. S7. Histological characterization of ectopic melanocytes. Embryos injected with KCNE1 mRNA were embedded in paraffin, sectioned, and stained with hematoxylin/eosin by using standard methods. Two specific organs were chosen for in-depth analysis: liver and skin. In the liver, KCNE1-injected larvae (A) possessed many more melanocytes compared with controls (B). In the skin, we observed far more melanocytes in the connective tissue under the dermis of KCNE1-overexpressing animals (C, C ) than of controls (D). These melanocytes occurred in groups, although it is not yet known whether they travel in groups or whether they proliferate in situ after reaching their destination compared with controls. In C (higher magnification of region from C), green arrowhead indicates epidermal melanocyte (not seen in controls), yellow arrows indicate dermal melanocytes (rarely seen in controls), and blue arrows indicate subdermal/peritoneal melanocytes.|
|Fig. S8. KCNE1 does not induce ectopic cells to enter melanocyte lineage. Embryos were injected with low levels of a mixture of KCNE1 and -gal mRNA at the one-cell stage (resulting in random mosaic expression throughout the embryo), fixed at early tailbud stages (before melanocyte migration), and processed for in situ hybridization to reveal the expression of Trp-2, a marker of early melanocyte cell lineage. Compared with controls (A), injected embryos revealed no ectopic expression (B, red -gal revealing lineage label in a majority of injected cells) beyond the normal endogenous expression at the dorsal domain (blue arrows).|
|kcne1 (potassium channel, voltage gated subfamily E regulatory beta subunit 1) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 28, lateral view, anterior left, dorsal up.|
|kcnq1 (potassium channel, voltage gated KQT-like subfamily Q, member 1) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 18, dorsal view, anterior up.|
|kcnq1 (potassium channel, voltage gated KQT-like subfamily Q, member 1) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 22, anterior view, dorsal up.|
|kcnq1 (potassium channel, voltage-gated KQT-like subfamily Q, member 1) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 26, lateral view, anterior right, dorsal up. Key: red arrows point to tailbud and anterior and porterior branchial neural crest streams.|