Artificial Gap Junctional Channel Inhibiting the Proliferation of Hepatocellular Carcinoma Cells

ABSTRACT

Gap junctions play a crucial role in intercellular communication, participating in both physiological and pathological processes. Although artificial gap junctions have been developed, their physiological functions beyond basic intercellular communication remain largely unexplored. Herein, we report a functional artificial gap junctional channel system constructed via electrostatic co‐assembly of an amphiphilic transmembrane channel and a hydrophilic extracellular channel. The efficient ion transport capability of the amphiphilic channel, coupled with the extensive aggregation of the hydrophilic module within the intercellular gap, enables the formation of stable intercellular junctions. Similar to natural gap junctions, these artificial channels not only mediate intercellular ion transport but also induce G1 phase cell cycle arrest and inhibit the proliferation of hepatocellular carcinoma cells. This work provides a new strategy for intervening in cell–cell communication and offers a potential therapeutic avenue for cancer and gap junction‐related diseases.

We have constructed artificial gap junctional channels within living intercellular gaps and found that they can inhibit the proliferation of liver cancer cells by blocking the cell cycle, functioning as natural gap junction channels.

Natural gap junctional channels are a class of membrane proteins formed by the docking of two hemichannels located in the plasma membranes of adjacent cells [1, 2, 3, 4, 5]. These channels enable the direct exchange of ions, metabolites, and second messengers, facilitating metabolic and electrical coupling between cells [6, 7, 8, 9, 10]. They play key roles in organism development, differentiation, and various physiological and pathological processes [11, 12, 13, 14, 15]. Conversely, dysfunction of gap junctions is linked to a variety of diseases, including cancer development and progression [16, 17]. Over the past decades, numerous artificial transmembrane channels have been constructed to mimic the structure and function of natural channel proteins [18, 19, 20, 21, 22, 23, 24, 25]. However, building artificial analogues of gap junctions remains challenging, primarily due to the obstacle in achieving precise structural alignment and efficient communication between adjacent cells [26, 27, 28]. Recently, our group has achieved in the construction of these complex channels by mimicking the hydrophobic–hydrophilic–hydrophobic triblock structure of natural gap junctional channels [29, 30]. Despite these advances, the physiological functions of such artificial gap junctions, beyond mediating basic intercellular communication, have not been fully explored. Herein, we report the construction of a functional artificial gap junctional channel system via electrostatic co‐assembly. We demonstrate that these channels not only mediate intercellular ion transport but also induce cell cycle arrest at the G1 phase in human hepatocellular carcinoma cells (HepG2), leading to significant inhibition of cell proliferation.

Pillar[n]arenes have emerged as a promising backbone for constructing artificial transmembrane channels due to their rigid columnar structures, electron‐rich cavities, and easily modifiable side chains [31, 32]. In this work, we designed two modular components: a hydrophilic channel 1 and an amphiphilic channel 2 (Figures 1a and S1–S6). Channel 1 was synthesized by appending hydrophilic side chains onto a pillar[5]arene backbone. Channel 2 was designed with a pillar[5]arene core functionalized with hydrophobic tryptophan (Trp)‐containing peptides at both ends and hydrophilic groups in the middle. The Trp residues are known to enhance the membrane incorporation efficiency of channel molecules via interactions with lipid bilayers [33, 34, 35, 36]. We envisioned that channel 2 would incorporate into cell membranes to form transmembrane channels, while the hydrophilic channel 1 would dock electrostatically to the exposed charged regions of 2 within the intercellular gap, leading to the assembly of a complete gap junctional channel (Figure 1b). To facilitate co‐assembly, opposite charges were introduced on the pillar[5]arene backbones of 1 and 2. This design renders 1 highly water‐soluble and promotes its electrostatic interaction with 2. Two‐dimensional nuclear magnetic resonance (2D NMR) analysis of 1 confirms its propensity for ordered self‐assembly in solution. The NOESY spectrum exhibited clear nuclear Overhauser effect (NOE) correlations between the α‐protons of the amino and carboxyl groups, providing direct evidence for ordered stacking of the pillar[5]arene units (Figure S7).

FIGURE 1.

Design and assembly schematic of the artificial gap junctional channel system. (a) Chemical structures of the hydrophilic channel 1 and the amphiphilic channel 2. (b) Schematic illustration of the proposed assembly process: channel 2 inserts into the membranes of adjacent cells to form transmembrane channels, while channel 1 docks electrostatically to 2 within the intercellular gap, leading to the formation of a complete supramolecular gap junctional channel.

The ability of channel 2 to form functional transmembrane channels was first assessed by single‐channel conductance measurements using a planar lipid bilayer. Square‐like current traces are observed upon applying a voltage, confirming the spontaneous insertion of a single 2 molecule into the lipid bilayer and ion transport performance (K⁺, Na⁺, Cl⁻) of 2 (Figures S8–S11). The ion transport activity of 2 was further quantified using vesicle salt transport experiments. Large unilamellar vesicles (LUVs) composed of 1,2‐dioleoyl‐sn‐glycero‐3‐phosphocholine (DOPC) and encapsulating the Cl⁻‐sensitive fluorescent probe 6‐methoxy‐N‐(3‐sulfopropyl)quinolinium (SPQ)[37, 38, 39, 40] were prepared. Upon addition of channel 2 to an external KCl solution, the fluorescence intensity of SPQ inside the vesicles decreases in a concentration‐dependent manner, indicating the symport transport of K⁺ and Cl⁻ into the vesicles (Figure 2a) [41]. Plotting the initial transport rate (V ₀) against the concentration of 2 yielded a slope close to 1 (Figure S13), confirming a unimolecular transport mechanism.

FIGURE 2.

Characterization of channel function and intervesicular junction formation in vitro. (a) Fluorescence quenching traces of LUV^SPQ (λ _ex = 344 nm, λ _em = 443 nm) upon addition of 2, indicating concentration‐dependent Cl⁻ transport from the external KCl solution into the vesicles. (b) Normalized donor fluorescence intensity in a FRET‐based vesicle fusion assay. (c) DLS measurements showing the hydrodynamic diameter of LUVs in the presence of 1, 2, or their mixtures. (d) Normalized SPQ fluorescence traces in a vesicle–vesicle KCl transport assay. The shown images are representative of three independent experiments.

Amphiphilic molecules like 2 can sometimes induce non‐specific vesicle fusion [42, 43, 44], which would hinder specific gap junction formation. We employed a Förster resonance energy transfer (FRET) assay to verify that adding 1 could inhibit vesicle fusion. LUVs were separately labeled with donor (N‐(7‐nitro‐2,1,3‐benzoxadiazol‐4‐yl, NBD) and acceptor (rhodamine, Rho) fluorophores (Figure S14). Fusion between labeled and unlabeled vesicles dilutes the donor–acceptor pairs, leading to a decrease in FRET efficiency and an increase in donor fluorescence. Compared to the addition of 2 alone, the co‐addition of 1 and 2 results in a significantly smaller increase in donor fluorescence (Figure 2b), demonstrating that 1 effectively suppresses 2‐induced vesicle fusion. Dynamic light scattering (DLS) measurements provided further support. While 2 alone causes a moderate increase in vesicle size, the combination of 1 and 2 leads to a more pronounced and stable increase in hydrodynamic diameter (Figure 2c). This shift indicates that 1 redirects the system from fusion toward vesicle adhesion—a necessary precondition for gap junction formation.

To demonstrate the formation of functional gap junctions between vesicles, we performed a vesicle–vesicle salt transport assay (Figure S15). Two populations of LUVs were prepared: one loaded with KCl (LUV^KCl) and the other loaded with the SPQ probe (LUV^SPQ). Their mixture was monitored for changes in SPQ fluorescence, which would indicate KCl flux from LUV^KCl to LUV^SPQ via connecting channels. As shown in Figure 2d, the addition of 2 alone causes only a slight fluorescence decrease, likely due to residual fusion or non‐specific leakage. Strikingly, the co‐addition of 1 and 2 induces a rapid and significant fluorescence quenching. The transport efficiency increases with higher concentrations of 1, validating that the 1+2 co‐assembly is essential for forming efficient intervesicular junctions that mediate directed ion flux. Additionally, the co‐addition of 1 and 2 produces current levels with distinct amplitudes that are different from those observed for 2 alone (Figure S12). These results suggest that supramolecular assembly between 1 and 2 generates a new conductive species, consistent with the formation of the supramolecular channel architecture [30].

The cellular distribution of the channels was visualized using confocal microscopy. HepG2 cells incubated with Cy5‐labeled 2 (2‐Cy5) show bright fluorescence outlining the cell membranes (Figure 3a), confirming its efficient insertion. In contrast, Cy3‐labeled 1 (1‐Cy3) alone shows minimal membrane association and diffuse cytoplasmic staining (Figure 3b). However, when cells were co‐incubated with unlabeled 2 and 1‐Cy3, distinct Cy3 signals became enriched specifically at the interfaces between adjacent cells (Figure 3c). This precise localization demonstrates that membrane‐embedded 2 acts as an anchor, recruiting 1 to cell–cell contact sites—the exact location required for gap junction assembly.

FIGURE 3.

Cellular localization of channel modules visualized by confocal microscopy. HepG2 cells were incubated with (a) 2‐Cy5 (5 µM; purple), (b) 1‐Cy3 (10 µM; orange), or (c) a mixture of 2 (5 µM) and 1‐Cy3 (10 µM). Scale bars: 20 µm. The shown images are representative of three independent experiments.

The formation of gap junctions between living cells was assessed using whole‐cell patch clamp electrophysiology (Figure S16) [45, 46]. Human embryonic kidney cells HEK293, which lack endogenous gap junctions [47], were used to avoid interference (Figure 4a). A coupling coefficient (CC%), defined as the ratio of the transjunctional voltage (V _transjunc) in the recipient cell to the induced membrane potential (V _induc) in the stimulated cell, was calculated to quantify coupling efficiency [48]. HEK293 cells incubated with 1 alone show no detectable electrical coupling (Figure 4b). Cells incubated with 2 alone exhibit weak coupling (21.6 ± 0.8%, Figure 4c), possibly due to limited self‐assembly across the gap. Remarkably, cells co‐incubated with both 1 and 2 display evident electrical coupling, with a coupling coefficient of 61.4 ± 0.7% (Figure 4d). This result unequivocally confirms that our co‐assembly strategy enables the efficient formation of functional artificial gap junctions between adjacent cells.

FIGURE 4.

Electrophysiological demonstration of intercellular coupling mediated by artificial gap junctions. (a) Schematic of the whole‐cell patch clamp recording setup on a pair of adjacent HEK293 cells. Scale bars: 20 µm. (b–d) Representative membrane potential traces (current–clamp mode) recorded from paired cells incubated with (b) 1 (1 µM), (c) 2 (1 µM), or (d) 1+2 (each 1 µM). The alternating application of current steps to one of the paired cells elicited the generation of membrane potential (V _induc) within that cell, concurrently inducing a transjunctional voltage (V _transjunc) in its coupled neighboring cell. The shown images are representative of three independent experiments.

Natural gap junctions are known to regulate cell proliferation, partly by influencing cell cycle progression [49, 50, 51]. We therefore investigated whether our artificial channels could exert a similar effect on HepG2 cells. Cell cycle distribution was analyzed by propidium iodide staining and flow cytometry after 12, 24, and 36 h of incubation. No significant changes were observed at 12 h (Figure 5a). However, after 24 and 36 h, the proportion of cells in the G1 phase increased by 12% and 13%, respectively, in the channel 1+2 group compared to the control (Figure 5b,c). No such arrest was observed in groups treated with 1 or 2 alone, indicating that halting the cell cycle at the G1 phase is specific to the assembled gap junction structure. Moreover, a magnetic bead‐based cell counting assay confirmed the anti‐proliferative effect: after 24 h, the 1+2 group showed a 40% reduction in cell number compared to the control, while other groups showed no significant difference (Figure 5d). The G1 phase represents a critical checkpoint where cells commit to DNA replication and division [52]. Therefore, the ability of artificial gap junction channels to induce G1 phase arrest to slow down cell proliferation suggests their potential as valuable tools for regulating cell growth and inhibiting cancer progression.

FIGURE 5.

Artificial gap junction channels induce G1 phase arrest and inhibit the proliferation of HepG2 cells. (a–c) Cell cycle distribution of HepG2 cells after incubation with 1 (20 µM), 2 (10 µM), or 1+2 for (a) 12, (b) 24, and (c) 36 h. Percentages of cells in G1, S, and G2/M phases are indicated. (d) Relative cell numbers after 24 h of treatment. Data are mean ± S.D. (n = 3). **p < 0.01, *p < 0.05 versus Blank control.

To elucidate the mechanism behind G1 phase arrest, we examined the expression of Cyclin E, a key regulator of the G1/S transition [53, 54, 55]. Downregulation of Cyclin E is known to halt the cell cycle at G1, thereby inhibiting proliferation [56, 57, 58]. Western blot analysis revealed that the 1+2 treatment significantly reduced Cyclin E protein levels in HepG2 cells, whereas 1 or 2 alone had no appreciable effect (Figure 6a,b). This result suggests that the artificial gap junction channels inhibit cell proliferation by interfering with the expression of cell cycle regulators.

FIGURE 6.

Downregulation of Cyclin E protein expression by artificial gap junction channels. (a) Representative Western blot images showing Cyclin E protein levels in HepG2 cells treated for 24 h with 1 (20 µM), 2 (10 µM), or 1+2. GAPDH serves as a loading control. (b) Quantification of Cyclin E band intensity normalized to GAPDH. Data are mean ± S.D. (n = 3). *p < 0.05 versus Blank control. The shown images are representative of three independent experiments.

In summary, we have successfully constructed a functional artificial gap junctional channel system through the electrostatic co‐assembly of a designed amphiphilic transmembrane channel 2 and a hydrophilic extracellular channel 1. This system efficiently incorporates into lipid membranes, mediates specific intercellular ion transport between cells, and localizes preferentially at cell–cell contact sites. Beyond restoring intercellular communication, we have uncovered a novel biological function for these synthetic structures: they induce G1 phase cell cycle arrest in HepG2 hepatocellular carcinoma cells by downregulating the key cell cycle protein Cyclin E, leading to significant inhibition of cell proliferation. This effect is strictly dependent on the co‐assembled channel complex, highlighting the importance of the supramolecular structure for its biological activity. Our findings significantly advance the field of artificial gap junctions by moving beyond structural mimicry to demonstrate functional integration with cellular regulatory pathways.

Author Contributions

Meng‐Ke Zhang: methodology, formal analysis, data curation, investigation. Lu‐Lu Ye: methodology, formal analysis, data curation, investigation. Jun‐Li Hou: conceptualization, funding acquisition, formal analysis, supervision, writing – original draft, writing – review and editing. Mo Sun: funding acquisition, writing – review and editing, formal analysis.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: Materials and instrumentation, experimental details, data analysis methods, and supplementary figures are provided in the Supporting Information. The authors have cited additional references within the Supporting Information [59, 60].

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.