HCN2 Rescues brain defects by enforcing endogenous voltage pre-patterns

Following the training sessions a significant red light aversion was generated in control tadpoles. Majority of nicotine-exposed tadpoles failed to demonstrate red light aversion learning during testing. Nicotine-exposed tadpoles microinjected with Hcn2-WT mRNA showed a restored ability to learn associative red light aversion

Only ~12% (n = 26) of the nicotine-exposed tadpoles showed learning ability; the majority failed to learn in this automated behavioral test, in comparison to control tadpoles, which showed good learning abilities (100%, n = 25) (Fig. 7b, c, ANOVA p < 0.001), even though only the able-bodied (healthy, morphologically normal) nicotine-exposed animals were tested in this assay. As expected, the brain dysmorphias caused by nicotine were accompanied by severe performance defects. However, nicotine-exposed embryos receiving Hcn2-WT mRNA showed significant improvement in learning ability (85%, n = 20) in comparison to nicotine-exposed tadpoles (Fig. 7b, c, ANOVA, p < 0.001). Thus, HCN2 channels not only repair morphology, but also restore cognitive learning abilities in nicotine-exposed tadpoles.

To determine the effects of HCN2 channel intervention in the hierarchy of known transcriptional regulators of brain patterning, we analyzed the expression of the key transcription factors otx2 (forebrain and midbrain), emx (telencephalon and forebrain), xbf1 (forebrain), and Pax6 (forebrain and eye)59–62.

Embryos with or without Hcn2-WT mRNA microinjection (both blastomeres at 2-cell stage) were exposed to nicotine from stage 10. Uninjected and untreated embryos were used as controls. Embryos were analyzed at stage 25 for expression of otx2, xbf1, emx, and pax6 by in situ hybridization (Fig. 8). The nicotine-exposed embryos showed significantly reduced expression of otx2 (in both area and intensity) and reduced or mispatterned expression of xbf1 in significant number of embryos [46% (χ2, p < 0.001) and 35% (χ2, p < 0.01), respectively] in comparison to control embryos (10% and 4.5%, respectively). In marked contrast, Hcn2-WT mRNA-injected embryos showed normal pattern of otx2 and xbf1 expression at similar levels to that of controls (4 and 4%, respectively). Nicotine exposure had little effect on emx and pax6 expression pattern, with minor reductions in a few embryos (17 and 12%, respectively) in comparison to controls (0 and 3%, respectively). Hcn2-WT-injected embryos exhibited normal emx and pax6 expression pattern at similar levels to controls (9 and 6%, respectively). Thus, HCN2 channel rescue of nicotine-induced defects includes correction of otx2 and xbf1, connecting HCN2-mediated repair of brain to the known molecular-genetic cascade of brain patterning factors.Fig. 8HCN2 channels restore nicotine exposure induced mispatterning of brain markers during neural development. Stage 25 embryos as illustrated with the angle of view marked by the black arrow. Control (untreated/uninjected) embryos (a, e, i, l), embryos exposed to nicotine (0.1 mg/mL – stage 10 onwards) (b, f, j, n), and nicotine-exposed embryos microinjected with Hcn2-WT mRNA (0.75 ng/injection) in both blastomeres at 2-cell stage (c, g, k, o).

In situ hybridization for Otx2 (a–d), Xbf1 (e–h), Emx (i–l), and Pax6 (m–p) show that nicotine exposure leads to significantly mispatterned expression (magenta arrows) of Otx2 [46% n = 24] and Xbf1 [35% n = 23], but has little effect on Emx [17%, n = 29] and Pax6 [12%, n = 33] in comparison to controls [10%, n = 28, 4.5%, n = 22, 0%, n = 28, and 3%, n = 29, respectively]. Nicotine-exposed embryos that were also microinjected with Hcn2-WT mRNA showed largely normal expression of Otx2 [4% n = 23], Xbf1 [4% n = 26], Emx [9% n = 23], and Pax6 [6% n = 34] in comparison to controls

Nicotine is a well−studied developmental neuroteratogen in humans29, non-human primates31, and invertebrates32. A fundamental gap in this area is the lack of mechanistic knowledge of how the effects on individual cell physiology result in impaired cognitive function. We addressed the developmental events that underlie the middle layer between cell behavior and organismal behavior—teratogenic effects of nicotine exposure on morphogenesis of the developing brain.

Bioelectric circuit9–11,63 dynamics are distinct from those of chemical gradients64–66 and are often complex. To bridge the gap between cell-level molecular electrophysiology and organism-level patterning and behavior, we used BETSE,41 a powerful and flexible biorealistic platform that enables predictions of bioelectrical patterns from channel expression and function data. We built a model of bioelectric patterns of neural stage Xenopus embryo, taking into account spatiotemporal dynamics of all different parameters including gap junctions, tight junctions, and relevant biochemical signals.

The endogenous early embryonic Vmem patterns across the neural tube are required for normal brain development28. Nicotine, an ion channel ligand, disrupts brain patterning (Fig. 1). We analyzed the BETSE model to understand how nicotine and IT actions would integrate at the level of embryonic Vmem patterning, and generate predictions and testable hypotheses.

Nicotine exposure diminished the endogenous neural tube hyperpolarization in the model (Fig. 3). Such disruptions in the brain specific Vmem pattern have already been shown to cause serious brain morphology defects28. In the model, HCN2 channels were seen to behave in a unique context-specific manner (Fig. 3f) that specifically amplifies the differential neural tube Vmem pattern across the neural plate despite the presence of nicotine (fitting the functional criteria of the model-derived IT). This predicts that context-specific action of HCN2 would strengthen the Vmem pattern otherwise diminished by nicotine and thus restores downstream transcription factors and brain patterning. Precisely, this was observed in our physiological profiling experiments (Fig. 4). It is also important to note that the model predicts that due to its context specific action, HCN2 channels also enhance the brain specific Vmem pattern of control (non-nicotine exposed) embryos (Fig. 3), consistent with our observation of patterning improvement of even the low incidence of brain defects in control batches of embryos (Fig. 5b). The BETSE model proved essential to quantitatively test this hypothesis and evaluate the properties of HCN2-containing circuits to reveal that they can indeed overcome the effects of nicotine on Vmem and do so appropriately depending on location in the neural tube center and its lateral areas. The balance of Vmem contrast between the neural tube and surrounding tissue has been shown to control the balance of proliferation and apoptosis in neural tube development48. This kind of sharpening of boundaries between distinct bioelectric compartments explains why HCN2 channels are able to correct even background brain defects (Figs. 5a, b).

An alternative case is where nicotine may directly affect endogenous HCN2 channels46, resulting in brain patterning defects. This suggests that normal endogenous HCN2 activity is required for proper brain patterning. However, our experiments with dominant-negative HCN2 channel shows that native HCN2 channel activity is not required for proper brain patterning (Supplementary Figure 4), thus negating this possibility. The effect of nicotine on the embryonic Vmem pre-pattern through nAChR (Figs. 2 and 3) is sufficient to explain the observed change in bioelectric state and the subsequent phenotypes, but our data do not rule out additional endogenous targets of nicotine. This is consistent with bioelectric mechanisms of action suggested for a number of other teratogenic chemicals67 and genetic syndromes68–70.

A critical next step for developmental physiology is to extend the understanding of single-cell electrophysiological properties (and ion channel expression) to spatialized dynamics that can explain the maintenance or disruption of large-scale patterns with relevance to developmental or regenerative patterning. To test predictions of the BETSE model, we generated Vmem maps of neurula stage Xenopus embryos under the different treatment conditions (Fig. 4). As predicted by the BETSE model, nicotine treatment indeed diminished the Vmem pattern (depolarized) by ~15 mV (peak Vmem in the neural tube) and HCN2 channel even in presence of nicotine increased the Vmem pattern (restored hyperpolarization) by ~35 mV (peak Vmem in the neural tube) (Figs. 3 and 4). In the case of HCN2 channel alone, the observed enhancement of Vmem pattern in relation to controls (~15 mV) was much less than the predicted value of increase by ~30 mV (Figs. 3 and 4). Since HCN2 channels are also regulated by cAMP and the context of cell metabolism, this difference may be the result of as yet unknown aspects of cell metabolism. It is also possible that there are other ion channels and pumps at play within the cells, affecting the action of HCN2 or directly influencing Vmem. We found an excellent quantitative match between the BETSE model’s predictions and the observed physiological and functional outcomes. To our knowledge, this is the first use of computational modeling to reveal spatialized bioelectric signaling dynamics that not only correctly predict the effects of a teratogen, but also show how to implement a rescue strategy.

Bioelectric interventions targeting ion flux have been used to induce regenerative response16,17. Here, we established a proof-of-principle extending this strategy to the use of ion channels as a reparative intervention for birth defects. Remarkably, HCN2 overexpression in embryos exposed to nicotine led to a complete rescue of nicotine-induced brain morphology defects (Fig. 5c–f). HCN2 channel overexpression restored not only gross brain morphology, but also the normal relation between brain length and the shape and size of the head (Fig. 6). This suggests that HCN2-mediated rescue involves sensing information from the surrounding regions to develop a brain with the correct size in relation to surrounding tissues and overall embryo proportions. Our prior work had pointed to the existence of a mechanism that coordinates brain sizing with remote tissue regions48. These consequences of HCN2 function are most likely due to its ability to act in a context-specific manner (Figs. 3 and 4)—it does not merely bring every cell to the same Vmem, but effects specific changes of resting potential that are different in distinct regions based on the local pre-existing Vmem. This ability to reinforce spatial differences enables embryos to establish the crucial embryonic brain-specific Vmem pre-pattern28,48 that is otherwise erased by nicotine exposure (Fig. 4).

A notable finding was the ability of HCN2 channels to improve the brain morphology of control embryos (Fig. 5a, b). This effect is likely also attributable to the context-specific actions of HCN2, which significantly enhance the crucial embryonic brain’s Vmem pre-pattern of these embryos (Figs. 3 and 4). These observations suggest that the restorative action of HCN2 is not limited to nicotine-induced brain morphology defects, but could reveal a more general mechanism of action, reinforcing the endogenous bioelectrical signals required for correct developmental brain morphology that might be partially destabilized by a wide range of exogenous teratogens or endogenous physiological/genetic lesions. The full scope of the kinds of defects that can be rescued by HCN2 remains to be delineated and is an active area of investigation in our laboratory.

We tested the effects of HCN2 on learning behavior in nicotine-exposed Xenopus embryos using an automated training apparatus54–56. Our data showed that HCN2 channels largely restore learning abilities in nicotine-exposed tadpoles (Fig. 7). This suggests that HCN2 along with restoring the physiological states necessary for normal brain morphology (Figs. 5 and 6) restores the internal wiring and networks necessary for normal learning abilities (Fig. 7). Given that HCN2 is able to fix learning abilities in addition to gross brain morphology, we suggest that HCN2 or other Vmem-based therapies could represent an important new direction for addressing learning and memory disorders that are related to such defects.

Endogenous Vmem patterns in neurula stage Xenopus embryos are critical for proper patterning and sculpting of brain tissue. Here we established an amphibian model for understanding nicotine neuroteratogenesis at the level of gene expression, electrophysiology, morphology, and behavior. Because of the central role of bioelectric circuits in normal brain development and teratogenesis, we formulated a predictive, biorealistic model of developmental physiology that explained the patterning of resting potentials across the neural tube under wild-type and disease conditions. Our model is the first quantitative explanation of teratogenesis targeting bioelectric controls of growth and form. Most importantly, it predicts that reagents that sharpen borders between bioelectric compartments will reinforce normal patterning, thus identifying a context-sensitive HCN2 ion channel as an intervention tool. We used HCN2 to confirm this prediction in vivo. Forced expression of HCN2 mRNA rescued nicotine-induced malformations, restoring not only the anatomy, but the gene expression in the nascent brain, and ultimately improving behavioral metrics of learning and memory. These data identify a patterning function that can be implemented using HCN2 in synthetic biology or in vivo. Taken together, our data shed light on a new mechanism of brain defects and establish proof-of-principle for using developmental physiology models to understand disorders of pattern and test interventions in silico. We provide a tool for predictive bioelectric modulation and demonstrate its use as a part of a tractable strategy for manipulating neurogenesis and neural patterning. Future work targeting native channels with small molecule compounds to effect rescue without gene therapy are currently on-going in our lab.

Thus, activation or expression of specific channels is a general strategy for biomedical approaches to complex organ patterning in the context of birth defects, regenerative medicine, and bioengineering.

Xenopus laevis embryos were fertilized in vitro according to standard protocols71: in 0.1 × Marc’s Modified Ringer’s (MMR; 10 mM Na+, 0.2 mM K+, 10.5 mM Cl-, 0.2 mM Ca2+, pH 7.8). Xenopus embryos were housed at 14–18 °C (14 °C overnight after injection and subsequently at 18 °C), and staged according to Nieuwkoop and Faber72. We saw the normal very-low background levels of abnormal morphologies (<9%), indicating the good health of our animals and good rearing conditions. For animals used in behavior trials, individuals were raised under 12 h:12 h light:dark cycle at a temperature of 16 °C, at no more than 30 individuals per 100 × 25 mm Petri dish. After stage 46, tadpoles were fed twice per day on standard sera micron powdered food until behavioral testing. All experiments were approved by the Tufts University Animal Research Committee (M2017-53) in accordance with the guide for care and use of laboratory animals.

Capped synthetic mRNAs generated using mMessage mMachine kit (Ambion) were dissolved in nuclease free water and injected into embryos immersed in 3% Ficoll using standard methods71. Each injection delivered between 0.5–1 ng of mRNA (per blastomere) into the embryos, at the indicated stages into the middle of the cell in the animal pole. Hcn2-WT and Hcn2-DN were mammalian (mouse) (HCN) channel 2, modified as detailed in the supplemental documents.

Xenopus embryos were incubated in chemicals or pharmacological blockers dissolved in 0.1 × MMR during the stages of interest as indicated in respective experiments followed by several washes with 0.1 × MMR. The embryos were exposed (from stage 10 to 35 unless otherwise specified, because neural/brain tissue development takes place in this time period, allowing specific testing of effects on these processes while allowing cleavage and gastrulation to proceed normally) to the following: 0.1 mg/mL nicotine (Sigma). The dose of nicotine was titrated to a level at which no general toxicity was observed and survival rate of embryos was similar to untreated controls.

Tadpoles used for morphometric analysis were imaged with a Nikon SMZ1500 microscope with a Retiga 2000R camera and Q-capture imaging software. Landmark data were then recorded using ImageJ software73. Landmarks for morphometric analysis were chosen based on biological relevance and reproducibility across tadpoles with varying brain and head morphologies. Landmarks were: (1) anterior most tip of the head; (2) between the two olfactory bulbs at the beginning of the forebrain; (3) laterally center of the brain at the point of transition between forebrain and midbrain; (4) Laterally center of the brain at the point of transition between midbrain and hindbrain; (5) Transition point between hindbrain and spinal cord; and (6) and (7) Lateral outermost points of the head along the eyes. MorphoJ software74 was used for Canonical Variate Analysis, to quantify and graphically represent change in brain regions in relation to head shape. MorphoJ software was also used to calculate Procrustes distances and perform statistical analysis. Our analysis is conservative as nicotine-treated embryos that had major portions of their brain missing were not used since the same landmarks as controls could not be used (e.g., Fig. 1c). Only those tadpoles that had brain morphology defects, but still could be used to place the landmarks were used (e.g., Fig. 5d).

All behavioral trials were performed with a custom built automated training robot54,56. The device consists of an array of 12 individual chambers, each capable of holding a 60 × 15 mm petri dish filled 15 mls of 0.1 × MMR. Below each dish is a machine vision camera (Insight-Micro 1400, Cognex Corporation, Natick, MA, USA), which uses a background subtraction algorithm to track any animals in the chamber in a Cartesian manner. Illumination is provided to each chamber independently from above, by a control module capable of specifying color and intensity by quadrants within each chamber. Red or blue light is delivered by light-emitting diodes (Osram Semiconductors, blue LED; 470 nm part no. LBW5SM, red LED, 635 nm part no. LRG6SP) and Xenopus see both given the spectral profiles of their three known cone classes75. Each chamber also has a set of six, equidistant, iridium oxide-coated titanium electrodes allowing the delivery of mild to strong electric shocks. All shocks delivered during wavelength-mediated training experiments were 1.2 mA AC currents, pulsed for 100 ms followed by 300 ms of no shock. This value was previously determined to be the lowest that elicits a behavioral response and no animals displayed physiological or behavioral abnormalities upon completion of testing.

Tadpole color-based associative learning assay was performed as documented before56 (Fig. 7a). Individual tadpoles were introduced to the chamber, which was illuminated with half red and half blue light, in absence of any punishment to probe their innate color preferences. Innate testing lasted 30 mins with colors across quadrants inverted after 15 mins to avoid considering stationary tadpole as having a 100% preference for a particular color (if the tadpole is stationary inverting the color would result in 50/50 preference). Tadpole then enters learning acquisition phase where tadpole receives a 1.2 mA shock if it occupies the red quadrant. This duration is 20 mins, with colors in the chamber being inverted every 5 mins. Then tadpoles are given a 90 min rest period where the entire chamber is illuminated with blue light and no shock punishment is delivered. Finally, the tadpole is probed for learning of light preference by giving them a choice between red and blue light for 5 mins with no punishment. The entire block of acquisition-rest-probe is repeated 6 times.

All tadpoles used in behavioral trials were stage ~48. A tadpole was determined to have learned if their preference for red light was below 40%, averaged across the final 3 probe sessions of the experiment. Note, that tadpoles were fed directly before trials and small amount of food was added to each chamber during training, as hungry tadpoles fail to learn.

Xenopus embryos were collected and fixed in MEMFA (1 hour at room temperature)71 and in situ hybridization was performed as previously described71. Briefly, the embryos were washed with PBS 0.1% Tween 20 (PBST) and transferred through series of methanol washes 25–50% to 75–100%. In situ antisense probes were generated in vitro from linearized templates using a DIG-labeling mix (Roche). Chromogenic reaction times were optimized for signal-to-background ratio. Probes used were otx2 (ref. 76), xbf1, and emx (refs. 77,78). Antisense probe for Xenopus HCN2 was generated from X. laevis hcn2.L IMAGE clone 5514485 (Dharmacon): a HindIII fragment was deleted, leaving exons 2–4 and part of exon 5 as probe79.

CC2-DMPE and DiBAC4(3) voltage reporter dyes were obtained from Invitrogen and used as per the standard protocol, including dark-field and flat-field correction63. Briefly, the use of two dyes with opposite emission profiles simultaneously provides an internal control and allows ratiometric normalization. CC2-DMPE stock (5 mM) was dissolved 1:1000 in 0.1 × MMR and the embryos were incubated in the dark in this solution for at least 1 h followed by 5 washes with 0.1 × MMR. DiBAC4(3) stock (1.9 mM) was dissolved 1:1000 in 0.1 × MMR and the CC2-DMPE-stained embryos were then incubated in the dark in this solution for at least 30 min washed thoroughly in 0.1 × MMR, followed by visualization under the microscope. An Olympus BX-61 microscope equipped with a Hamamatsu ORCA AG CCD camera, and controlled by MetaMorph software (Molecular Devices), was used to collect signal. NIH ImageJ software was used to quantify the fluorescence intensities of the CC2-DMPE:DiBAC4(3) signal.

Membrane potentials were measured using an oocyte clamp OC-725C amplifier (Warner Instruments) with a single voltage electrode. Microelectrodes were made from thin-walled borosilicate glass pulled with a flaming/brown micropipette puller (p-97; Sutter Instruments) and back-filled with electrode solution (2 M potassium acetate, 10 mM KCl, and 5 mM HEPES, pH 7.5). Tip resistances were 80–100 MΩ. Electrode penetration of ectodermal cells was by visual guidance on a fixed-stage microscope (Zeiss) using a three axis micromanipulator.

HCN2 channel detection was performed by immunofluorescence for HCN2 channel. Briefly, embryos were fixed overnight in MEMFA at 4 °C (ref. 71). The embryos were permeabilized in PBS 0.1% Triton X-100; blocked with 10% goat serum in PBST for 1 hour at room temperature; and incubated at 4 °C overnight with primary antibody (Anti-HCN2: ThermoFisher Scientific PA1-918) at 1:500 dilution in PBST + 10% goat serum (blocking buffer). Embryos were washed six times in PBST and incubated with Alexa Fluor-conjugated fluorescent secondary antibody (Invitrogen) at 1:500 dilution in PBST + 10% goat serum overnight at 4 °C. Sections were washed six times and photographed using an Olympus BX-61 microscope equipped with a Hamamatse ORCA AG CCD camera, and controlled by MetaMorph software. NIH ImageJ software was used to quantify the fluorescence intensities of the immunostained embryos.

A mouse HCN2 cDNA clone was kindly provided by Dr. Jeanne Nerbonne (Washington University). The HCN2 cDNA was cloned into the eukaryotic expression vector pcDNA3 (Thermo Fisher Scientific). To create the dominant-negative (DN) HCN2 mutation, the entire GYG signature motif in the channel pore was mutated to alanine-alanine-alanine (AAA) by site-directed mutagenesis. The forward primer was: 5’-ATGAGCCACATGC TGTGTATCGCCGCGGCACGACAAGCACCCGAGAGCATGACA-3’ and the reverse primer was: 5’-TGTCATGCTCTCGGGTGCTTGTCGTGCCGCGGCGATACA CAGCATGTGGCTCAT-3’. The WT or DN mutant of HCN2 was then sub-cloned into a bi-cistronic mammalian expression vector, pIRESGFP1 (kindly provided by Dr. David Johns, Johns Hopkins University), to create a CMVp-HCN2-IRES-eGFP fusion for subsequent studies. The HCN2 channel construct function was tested using voltage clamping in HEK293 cells (which do not have endogenous HCN channels) (Supplementary Fig. 2A).

HEK293 cells were cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum, 100 μg/ml penicillin, and 100 μg/ml streptomycin in a 37 °C, 5% CO2, 95% air incubator as previously described (Makielski et al. 2003). Approximately 1 × 105 cells were seeded in a 35-mm-diameter culture dish containing 1.5-ml culture media. A volume of 0.7 µg DNA of HCN2-WT or HCN2-AAA or the empty vector control was transiently transfected into HEK293 cells at 80% confluency using a Fugene6 transfection kit (Roche Applied Science, Indianapolis, IN) following manufacturer’s instructions. The transfected cells were incubated at 37 °C for 24 h to allow protein expression.

Whole-cell HCN2-encoded inward currents (IHCN) were recorded at room temperature from transfected HEK293 cells using voltage clamping as previously described (Ye et al. 2008; Ye and Nerbonne 2009). The extracellular solution contained in (mM): NaCl 110, MgCl2 0.5, KCl 30, CaCl2 1.8, and HEPES 5 while the pipette solution contained (in mM): NaCl 10, MgCl2 0.5, KCl 130, HEPES 5, and EGTA. The cells showing green fluorescence were identified under epifluorescence and selected for IHCN recordings. Recording electrodes were fabricated from borosilicate glass (WP Instruments) on a two stage (P-87; Sutter Instrument) vertical puller. IHCN was evoked in response to hyperpolarizing voltage steps to potentials between −60 and −130 mV from a holding potential of −40 mV. Data were collected using an Axopatch 200B amplifier interfaced to a Digidata1322A acquisition system (Molecular Devices, Sunnyvale, CA). Data were subsequently processed and analyzed using Axon pCLAMP10 Software Suite (Molecular Devices, Sunnyvale, CA). A one-way ANOVA with Bonferroni correction was carried out to determine the statistical significance between experimental groups. P < 0.05 was considered a significant difference between two groups.