Unraveling the Oncogenic Characteristics of the Cytolinker, Plectin, in Esophageal Squamous Cell Carcinoma

Abstract

Background & Aims

Tissue mechanics involved in carcinogenesis by regulating cell morphology and structure, cell-cell, and cell-extracellular matrix (ECM) interactions are not fully understood. Plectin, a cytolinker and a critical component of the cell-ECM adhesion complex hemidesmosome (HD), plays an important role in the regulation of epithelial tissue mechanics, but its functions in carcinogenesis remain elusive.

Methods

We used cellular and molecular methods and multiple systems, including a 2-dimensional (2-D) esophageal keratinocyte Ca²⁺-dependent differentiation system, a 3-dimensional (3-D) esophageal keratinocyte organoid system, and tissue samples of normal rat and human esophageal stratified squamous epithelium (SSE), N-nitroso-N-methylbenzylamine (NMBzA)-induced rat esophageal squamous cell carcinoma (ESCC), and human ESCC, to determine the role(s) of plectin in regulating SSE homeostasis and ESCC carcinogenesis.

Results

We show that plectin is ubiquitously expressed in all proliferative and differentiative cell types in esophageal SSE. However, the localization of plectin in different cell types is controlled by plectin crosslinking to different macromolecular structures, such as HD, desmosome (DSM), and cytoskeletal filaments, and its expression is regulated by the ESCC oncogenic drivers and transcription factors, p63 and/or Notch1. Plectin functions by coupling plectin-associated HD, DSM, and cytoskeletal components together with plectin regulators p63 and Notch1, to maintain cell anchorage, proliferation/differentiation, and stratification of esophageal SSE tissue homeostasis. Perturbation of plectin expression and localization leads to the disruption of SSE homeostasis and the involvement in ESCC carcinogenesis.

Conclusions

Plectin is involved in maintaining SSE homeostasis, and misexpression of plectin through its genetic alterations or transcriptional dysregulations perturbs the compositions, stoichiometries, and localizations of plectin, HD, DSM, and the cytoskeleton underlying the oncogenic characteristics of plectin.

Graphical abstract

Summary.

Plectin plays a critical role in the regulation of esophageal stratified squamous epithelium homeostasis. Aberrant alterations in the expression and localization of plectin contribute to esophageal squamous cell carcinoma carcinogenesis.

The mammalian esophageal epithelium is an endoderm-derived stratified squamous epithelium (SSE) composed of the basal, suprabasal, and/or keratinized layers that undergo cell proliferation-differentiation to maintain tissue homeostasis.¹ Although the precise regulations of esophageal SSE homeostasis maintenance remain largely elusive, accumulating evidence demonstrates that esophageal SSE, like other ectodermal and endodermal SSE, retains heterogeneous cell populations in the basal and suprabasal layers. The heterogeneous cell populations are involved in stem cell renewal, proliferation, differentiation, fitness, competition, polarity, and damage-repair response via the processes of stemness maintenance or cell lineage and hierarchy commitment at the inter- and intra-layer levels, ultimately achieving the tissue homeostasis.2, 3, 4, 5, 6 In the basal layer, most cells are identified as proliferation-prone cells that express high levels of the cell-cycle related markers for cell proliferation. However, other cell populations have been identified, such as quiescent basal cells (QBCs), which possess stem cell characteristics. These QBCs express low levels of the cell-cycle related markers and high levels of the hemidesmosome (HD) components (eg, collagen17A1 [Col17A1], dystonin [DST], and plectin).6, 7, 8 In addition, another population is shown as the cells expressing high levels of the transcription factor KLF4 with the differentiation-prone fate.5, 6, 7^,9 Similarly, heterogeneous cell populations are also observed in the suprabasal layers of the esophagus. Although most cells expressing differentiation markers in the suprabasal layers are committed to differentiation, we and others have shown that there exists a cell population expressing high levels of the SSE lineage-determining transcription factor Sox2, but not p63 (termed as Sox2⁺p63^- cells). These Sox2⁺p63^- cells serve as the origin cells, playing a dual role both in cell differentiation and rapid tissue damage-repair responses.^10,11 Thus, these studies demonstrate that cells in the esophageal SSE are not only heterogeneous and functionally diverse but also exhibit plasticity in maintaining tissue homeostasis. Disruption of these processes would lead to severe esophageal diseases, including eosinophilic esophagitis (EoE) and esophageal squamous cell carcinoma (ESCC), one of the most aggressive and lethal human cancers.

ESCC, the dominant form of human esophageal cancer, is the 7th most common cancer and the 6th leading cause of cancer death worldwide.¹² Epidemiologic and genetic studies have shown that the incidence, development, and progression of ESCC are influenced by both environmental factors and genetic alterations. Although the environmental factors associated with ESCC have been identified as poor diet, alcohol consumption, and smoking, recent large-scale genetic landscape analyses have identified cancer-driver gene mutations in ESCC.13, 14, 15, 16, 17, 18, 19 As in other squamous cell carcinomas (SCCs), an extremely high frequency of p53 gene mutations is observed in ESCC, indicating a significant disruption in genomic stability. Other frequently mutated genes involved in regulating cell stemness and lineage commitment (Sox2 and p63), cell proliferation (RTK-Ras-PI3K-AKT-MAPK-Myc), cell cycle progression (cyclin D1-Cdk4/6-p16-RB), cell polarity and differentiation (FAT1-4 and Notch/WNT/Hippo), epigenetics (KMT2/4 and KAD6), and redox metabolism homeostasis (NRF2).15, 16, 17, 18^,20 Despite the advances in multi-omics analysis, the molecular mechanisms underlying ESCC carcinogenesis remain largely unknown. As ESCC is characterized by high heterogeneous, we note that “a long tail” of mutated genes with low frequency are recurrently detected in ESCC.¹⁷ However, the roles of these genes involved in ESCC carcinogenesis remain unclear. Among them, plectin, a cytolinker and a crucial component of the HD, has genomic alterations in ∼5% of all ESCC samples analyzed (Figure 1A). Notably, we and others have recently found that mutations in the plectin gene are recurrently detected in 4-nitroquinoline-N-oxide (4NQO)- or N-nitroso-N-methylbenzylamine (NMBzA)-induced SCC/ESCC rodent models.^21,22

Figure 1.

PLEC mutations in human ESCC. (A) Rate of PLEC mutations in 5 human ESCC genomic sequencing data (n = 973). (B) PLEC mutations were detected in 8 human ESCC cell lines. (C) Enrichment of PLEC mutations in its protein functional domains in 465 ESCCs (4 cohorts).

Plectin is a giant protein with a molecular mass of ∼600kDa that belongs to the Plakin family.²³ It contains multiple protein-protein interaction domains that crosslink the cytoskeleton, including intermediate filaments (IFs), microfilaments (F-actins), and microtubules (MTs)²⁴ with various cell membrane-associated macromolecules and structures. Primarily, plectin is a central component of type I HD (plectin, integrin α6/β4, Col17A1, CD151, and DST complex) in the SSE and type II HD (plectin and integrin α6/β4 complex) in the simple columnar epithelium. Type I HDs are positioned at the basement membrane of the basal cells of SSE that anchor epithelial cells to the extracellular matrix (ECM) via the interaction of the integrin α6/β4 (ITGα6/β4) in the HDs with the laminin 332 (LAM332) in the ECM. Meanwhile, plectin/DST of the HDs connects the basement membrane to the intracellular cytoskeleton in basal cells via interacting with the cytoskeletal Ifs.^25,26 Furthermore, HD-independent plectin can also crosslink the cytoskeleton with other cell membrane-associated structures and sub-organelle protein complexes, such as focal adhesions (FAs), desmosomes (DSMs), tight junctions (TJs), adhesion junctions (AJs), and mechanotransduction complexes.25, 26, 27, 28 For instance, plectin not only binds to the desmoplakin (DSP) of DSMs but also, together with DSMs, mediates the F-actin/IF crosstalk required for tension and cohesion in epithelial cells.²⁸ Thus, plectin facilitates the arrangement of intracellular cytoskeletal fiber networks that promote the interactions, the subcellular localizations, and the dynamic stabilities of cellular structures to regulate cell stemness, division cycle, differentiation, shape, morphology, fitness and competition, polarity, or damage-response required for tissue homeostasis.^5,25,26,29, 30, 31

Consistently, plectin is expressed in both basal and suprabasal cells in skin SSE, and its genetic alterations/abnormal protein expressions cause several human SSE-related diseases, including bullous pemphigoid (BP).³² Moreover, in addition to the low frequency of mutated plectin gene observed in ESSCs, high and frequent loss-of-function (LOF) mutations in the plectin gene are detected in liposarcoma and testicular germ cell tumors.^28,33, 34, 35 Paradoxically, overexpression of plectin mRNA and protein is also found in many human tumors, including prostate cancer, bladder cancer, and pancreatic cancer, suggesting that the role(s) of plectin in carcinogenesis is complex.36, 37, 38, 39 In particular, plectin expression is controlled by the oncogenic transcription factor ΔNp63 (referred to as p63), and the function(s) of the plectin protein is regulated by post-translational modifications, such as CDK, AKT, and/or PKC phosphorylation.^40,41 In addition, the Drosophila plectin/spectraplakin ortholog, Shot, is a downstream effector modulated by Notch signaling during foregut development.⁴² However, it is unknown whether Shot is a direct transcriptional target of Notch proteins. Given that LOF mutations in the Notch gene family are commonly detected in ∼20% of ESCCs, and that gene amplification and protein overexpression of p63 occur in over 50% of ESCCs, it is plausible that plectin dysfunction, abnormalities in plectin-associated or -dependent macromolecules and interactions, and/or deficiencies in plectin regulators may significantly contribute to the carcinogenesis of ESCC.15, 16, 17^,20 In this study, we unravel the characteristics of the aberrant oncogenic cytolinker, plectin, in human ESCC.

Results

Frequent and Recurrent Genomic Alterations of Plectin and Plectin-associated HD, DSM, and Cytoskeletal Genes, and Abnormal Expression of Plectin in Human ESCC/SCC

We sought to determine plectin and plectin-associated genes (eg, HD, DSM, and cytoskeletal component genes) involved in the carcinogenesis of ESCC. To this end, we first manually analyzed the genomic sequencing data of 973 ESCC samples, including 639 whole-genome sequencing (WGS) datasets, 215 whole-exome sequencing (WES) datasets, and 119 targeted deep sequencing datasets from 5 published studies.15, 16, 17^,19,20 The results revealed genomic alterations in the plectin gene, including missense mutations, nonsense mutations, and deletion mutations, occurring in approximately 5% of these ESCCs (Figure 1A). In addition, plectin gene mutations were also detected in 2 ESCC cell lines, EC109 and KYSE450, from our WGS data (Figure 1B; Supplementary Table 1) and 2 published studies (Figure 1B).^19,43 Although no hotspot mutations were observed, the mutations were mainly located within the functional domains of the plectin protein, such as the C-terminal plectin repeat domains (PRDs), the rod coiled-coil domain, and the N-terminal plakin domain (Figure 1C), suggesting that they were likely to be LOF mutations, as previously reported.^28,33, 34, 35 Taken together, these results indicated that, although the genomic alterations of the plectin gene were rare, they occurred recurrently in ESCCs.

As plectin was a critical component of HD and served as a cytoskeletal/structural cytolinker, we next expanded our analysis to include HD/HD-associated and DSM/DSM-associated component genes (referred to as plectin-associated genes), using the online available genomic landscape databases of 1640 SCCs, including 629 head and neck SCCs (HNSCCs), 487 lung SCCs (LSCCs), 297 cervical SCCs (CSCCs), and 227 ESCCs.^44,45 Detailed analyses via cBioPortal showed that mutations of HD/HD-associated components (PLEC, DST, COL17A1, ITGA6, ITGB4, CD151, KRT14, KRT5, LAMA3, LAMB3, and LAMC2) and DSM/DSM-associated components (PLEC, DSP, DSG, DSC, JUP, and PKP1-3) were recurrently found in 1% to 8% of the SCC samples (Figure 2A and B). Notably, mutations in individual plectin or plectin-associated genes were rare events, but they were mutually exclusive. When combined with all mutations in plectin and plectin-associated genes, the total number of mutations was approximately 50% of the SCCs examined, suggesting that plectin and plectin-associated gene products could play important roles in ESCC/SCC carcinogenesis.

Figure 2.

Genomic alterations of plectin/plectin-associated genes in human ESCC/SCC. (A) Schematic illustration of HD (left) and DSMs (right) structures in the stratified epithelium. (B) Somatic mutations of plectin/plectin-associated genes, including HD/HD-associated components and DSM/DSM-associated components in 7 SCC cohorts (1640 samples) including 629 HNSCCs, 487 LSCCs, 297 CSCCs, and 227 ESCCs using cBioPortal. (C) Amplification of plectin/plectin-associated genes, including HD/HD-associated components and DSM/DSM-associated components in 7 SCC cohorts (1640 samples) using cBioPortal. (D) PLEC mRNA levels in 102 pairs of ESCC and adjacent tissues by qRT-PCR assay. ∗∗∗P < .001. (E) Plectin mRNA expression in 2 SCC cohorts (1010 samples) analyzed by cBioPortal.

Paradoxically, we also observed genomic amplifications in the plectin gene locus on chromosome 8q24 in 6% of these analyzed SCCs, which was consistent with previous reports (Figure 2C).^16,19,46,47 Because some samples had plectin mutations, whereas others showed plectin amplifications in these SCC cohorts, we determined the mRNA expression levels of plectin in 102 paired ESCCs and their adjacent non-tumor tissues by quantitative reverse transcription-polymerase chain reaction (qRT-PCR). Our results showed that compared with controls, 6.9% of ESCCs had low plectin mRNA expression, whereas 45.1% of ESCCs had plectin mRNA overexpression (Figure 2D). The underexpression or overexpression of plectin mRNA was also detected in SCCs from 2 The Cancer Genome Atlas (TCGA) PanCancer Atlas databases using cBioPortal analysis (Figure 2E). These results indicated that the perturbations of plectin expression in human cancers were complex.

These paradoxical results led us to hypothesize that plectin, as a cytolinker, crosslinked many subcellular structures and macromolecules (eg, HDs, DSMs, and/or intermediate filament cytokeratins [CKs]). Altered plectin expression would affect the composition, stoichiometry, and interactions of plectin with plectin-associated HDs, DSMs, and/or CKs that were critical for their functions. Misexpression of plectin, either by underexpression or overexpression, could disrupt the localizations and/or functions of plectin, thereby affecting the functions of plectin-associated HDs, DSMs, and/or CKs, ultimately leading to dysfunctions and abnormalities in cell-ECM and cell-cell interactions that could promote the initiation and progression of tumorigenesis.

Characterization of Plectin Expression and Regulation by p63 and Notch1 in Esophageal Squamous Epithelial Cells In Vitro

To test the hypothesis, we first determined the expression and regulation of plectin in esophageal squamous epithelial cells in vitro. Because the plectin gene was reported to be a critical downstream target of p63,⁴⁰ we speculated that the overexpression of plectin in ESCCs might be due to gene amplifications and/or the high frequency of amplification or overexpression of p63. Conversely, we speculated that the underexpression of plectin in ESCCs might be due to the LOF mutations of the plectin gene and/or the absence of another upstream regulator of plectin, such as Notch1.⁴² To this end, we utilized an hTERT-immortalized normal rat esophageal keratinocyte cell line, REN-D3 (D3 for short),¹⁰ established in our laboratory using a calcium (Ca²⁺)-dependent keratinocyte proliferation-differentiation assay in vitro. A detailed experimental scheme is shown in Figure 3A. Previous studies and our preliminary results showed that D3 cells were highly proliferative in complete medium (10% of fetal bovine serum [FBS] with 1.16 mM/L Ca²⁺), became quiescent in conditioned medium (0.5% of FBS with Ca²⁺ free), and underwent a differentiation process in differentiation medium (0.5% of FBS with 0.6 mM/L Ca²⁺).^10,48,49 We examined the cell cycle profiles, cell morphologies, and expressions of plectin and plectin-related proteins, as well as proliferation/differentiation marker proteins in D3 cells using fluorescence-activated cell sorting (FACS), microscopy, immunoblotting, and immunofluorescence analyses. D3 cells, similar to primary cultured rat esophageal epithelial basal layer keratinocytes, were highly proliferative, displayed flat morphologies with 18.00 ± 1.10 μm in the z-axis, and expressed the basal layer markers p63, Col17A1, ITGβ4, and CK14 when cultured in complete medium (Figure 3B–E, G). Expression of Notch1 and its activated fragment, Notch intracellular domain 1 (NICD1), could be detected in these cells due to the high Ca²⁺ concentration in medium. High levels of plectin expression were also observed in D3 cells cultured in complete medium (Figure 3C–E).

Figure 3.

Plectin expression during the Ca²⁺-dependent keratinocyte proliferation-differentiation process. (A) Schematic diagram depicting Ca²⁺-dependent keratinocyte proliferation-differentiation in vitro. D3 cells were cultured in the complete medium (1.16 mM/L Ca²⁺). Then, D3 cells were treated with the conditioned medium (without Ca²⁺) for 48 hours. Third, D3 cells were induced to differentiation by switching the conditioned medium to the differentiation medium (0.6 mM/L Ca²⁺) for 48 or 72 hours. (B) Representative bright field images of cell morphology during the Ca²⁺-dependent proliferation-differentiation process and quantification of the thickness of cells in the z-axis direction with confocal microscopy (n = 5 random microscope fields; 63×). W/O Ca²⁺ 48 hours, without Ca²⁺ 48 hours; W/ Ca²⁺ 48 hours, with Ca²⁺ 48 hours. ∗∗P < .01; ∗∗∗P < .001. (C) Representative IF images of plectin/plectin-associated proteins in D3 cells during the Ca²⁺-dependent proliferation-differentiation process in vitro. (D) Quantification of mean fluorescence intensities of plectin/plectin-associated proteins (n = 5 random microscope fields; 40×). ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001. (E) Protein expressions of plectin, p63, Col17A1, CK14, FL-Notch1, NICD1, IVL, CK13, and HES1 in D3 cells during the Ca²⁺-dependent proliferation-differentiation process by Western blotting assay. GAPDH was used as an endogenous control. The ratios were obtained by the comparison of the band intensities of target proteins with GAPDH. (F) Western blotting to examine the protein levels of plectin and Notch signaling after GSI treatment (0 μM, 0.2 μM, 1 μM, and 5 μM) in D3 cells. GSI is a γ-secretase inhibitor. β-actin was used as an endogenous control. The ratios were obtained by the comparison of the band intensities of target proteins with β-actin. (G) Cell cycle profiles of D3 cells were detected during the Ca²⁺-dependent proliferation-differentiation process using FACS (n = 3 experiments). ∗∗∗P < .001.

However, when D3 cells were switched to conditioned medium, microscopy and FACS showed significant morphologic and cell cycle profile changes, with cells adopting cobblestone-like flat morphologies with 12.80 ± 2.31 μm in the z-axis and entering the G0/G1 quiescent phase (Figure 3B and G). Immunoblotting and immunofluorescence analysis also revealed a slight reduction in the expression levels of p63, CK14, Col17A1, and plectin, whereas Notch1/NICD1 showed a significantly reduced expression under the conditioned medium culture condition (Figure 3C–E). When D3 cells were induced to differentiate by switching from conditioned medium to differentiation medium, these cells retained the flat morphologies with 12.60 ± 1.36 μm in the z-axis and remained in the G0/G1 phase of the cell cycle (Figure 3B and G). Immunoblotting and immunofluorescence showed that these cells expressed low levels of p63, CK14, Col17A1, and ITGβ4, but high levels of squamous differentiation markers such as CK13, HES1, and IVL, the downstream targets of Notch (NICD1)/RBPJ, indicating that these cells underwent the Ca²⁺-dependent differentiation process in vitro (Figure 3C–E). Consistently, increased expression of Notch1, especially its cleaved form NICD1, was detected in these differentiated cells. These results confirmed that Notch1 expression and activation were regulated in a Ca²⁺-dependent manner (Figure 3C–E).

To further determine whether Notch1 activation was required for the differentiation process, D3 cells in differentiation medium were treated with GSI, a γ-secretase inhibitor that could suppress Notch1 cleavage to NICD1.^50,51 Immunoblotting showed that GSI treatment blocked Notch1 cleavage to NICD1, resulting in decreased expression levels of CK13 and IVL (Figure 3F). These results demonstrated that Notch1 induction and activation were required for calcium-induced esophageal keratinocyte differentiation in vitro. Notably, high levels of plectin expression were also detected in D3 cells cultured in differentiation medium, similar to the expression patterns of Notch1/NICD1 (Figure 3C–E). Furthermore, the expression levels of plectin in D3 cells in differentiation medium were inhibited by GSI, suggesting that plectin, like CK13, HES1, and IVL, may be a downstream target of Notch1 during esophageal keratinocyte differentiation in vitro (Figure 3F). Thus, these results, together with the expression patterns of plectin detected in ESCCs, revealed the complexities of plectin expression and regulation in esophageal keratinocytes. Plectin was expressed not only in proliferating esophageal keratinocytes but also in differentiating esophageal keratinocytes. Plectin expression was regulated, at least in part, by p63 in proliferating esophageal keratinocytes. In contrast, plectin expression might be controlled, at least in part, by Notch signaling in Ca²⁺-dependent differentiating esophageal keratinocytes in vitro.

Determination of Plectin Expression, Localization, and Regulation in Esophageal SSE and D3 Organoids

Next, we analyzed the expression and localization of plectin in esophageal SSE in vivo. Immunohistochemistry (IHC) and immunofluorescence revealed that plectin was abundantly expressed and localized in both basal layer cells (p63⁺CK14⁺) and suprabasal layer cells (p63^-CK13⁺) of human and rat esophageal SSE tissues (Figure 4A–C).⁶ Co-immunofluorescence further demonstrated that plectin was predominantly localized to the basement membranes with HD components, ITGβ4/Col17A1, and cytoplasm with CK14 in the basal layer cells. Additionally, plectin was detected at the basolateral and basoapical membranes with the DSM component, DSP, in the basal layer cells. In the suprabasal layer cells, plectin colocalized with DSP in the cytoplasmic membranes and with CK13 in the cytoplasm (Figure 4B and C). These results demonstrated that, although plectin was expressed in both proliferating and differentiating esophageal keratinocytes, plectin was associated with different membranous and/or macrostructures, including HDs and CK14 in “proliferating” basal layer cells, and DSMs and CK13 in “differentiating” suprabasal layer cells. The complex expression and localization of plectin, together with its aberrant expression in ESCCs, suggested that plectin could have a crucial role in regulating and/or maintaining tissue homeostasis in esophageal SSE.

Figure 4.

Plectin expression and localization in human and rat esophageal SSE. (A) Representative images to show the expression and localization of plectin in human and rat esophageal epithelium by IHC and IF assays (Left), and visualize the change of plectin localization at basal layer and suprabasal layer by 3D Surface Plot of ImageJ (right). Scale bars, 50 μm. Arrowheads represent the localization of plectin at the basal layer (yellow) and suprabasal (orange) layer of the human esophagus. (B–C) IF images of plectin, CK14, p63, CK13, DSP, ITGβ4, and Col17A1 in human (B) and rat (C) esophageal epithelia. Dotted lines separate the basal layer from the suprabasal layer. B, basal layer; SB, suprabasal layer. Scale bars, 50 μm.

Therefore, we adapted an in vitro D3 organoid culture system developed in our laboratory to test this possibility. The D3 organoid culture system allowed a single D3 esophageal keratinocyte cell with “anchorage” properties to grow in Matrigel for 10 to 12 days, forming three-dimensional (3-D) organoids containing the basal, suprabasal, and keratinized layers, similar to rat esophageal SSE in vivo.¹⁰ Time course experiments showed that the D3-organoid system mimicked the landscape processes of esophageal SSE formation, development, and maturation trajectories (Figure 5A).¹⁰ Immunofluorescence analysis revealed that during D3 cell anchorage to D3 organoid formation (day 1 to day 4), development (day 5 to day 8), and maturation (day 9 to day 12), the expression and localization of plectin protein were observed in both basal layer cells and suprabasal layer cells (Figure 5B), consistent with the results obtained from rat and human esophageal SSE tissues (Figure 4A). The STORM super-resolution microscopy confirmed the subcellular localization of plectin in the D3-organoid (Figure 5C). Notably, plectin was localized in the cytoplasmic membranes and cytoplasm, especially in the basement membranes (ITGβ4⁺), in the formation stage of D3 organoids at day 3, when individual D3 cells grew only to asymmetric undifferentiated small cell clumps (p63⁺CK14⁺) (Figure 5B and D).

Figure 5.

Plectin expression and localization in D3-organoids. (A) Schematic illustration of D3 organoid formation (day 1, 3), development (day 5, 7), and maturation (day 10) trajectory and representative bright field and HE images. Scale bars, 100 μm. (B) Expression and localization of plectin in D3-organoid cultured trajectory by IF assay. Scale bars, 50 μm. (C) Super-resolution IF image showing the localization of plectin at the basal and suprabasal layers of D3 organoid at the nanoscale level using N-STORM system in super-resolution microscopy. B, basal layer; SB, suprabasal layer. Scale bars, 5 μm. (D) IF images of CK14, Col17A1, ITGβ4, plectin, p63, CK13, DSP, and Notch1 in D3 organoid cultured trajectory. Scale bars, 50 μm.

As D3-organoids underwent rapid cell expansion from day 5 to day 8, the SSE began to stratify, with D3-organoid cells proliferating and differentiating into stratified multilayered epithelia. Representative results obtained from D3-organoids on day 5 and day 7 are shown in Figure 5A and D. Immunofluorescence analysis revealed that the outermost layer of D3-organoids (+1 layer) in contact with Matrigel could be characterized as the basal layer (p63⁺CK14⁺), whereas the inner layers of D3-organoids (+2 to +4 layers) represented the suprabasal layers (p63^-CK13⁺). We also occasionally detected a few p63⁺CK14⁺CK13^low or p63^lowCK14⁺CK13⁺ cells in the +1 basal layer and/or the +2 suprabasal layer, suggesting that these cells might represent transitional cells from basal to suprabasal cells. At this stage, the SSE of the D3-organoids began to display stratified layers, with cells in the basal layer (p63⁺) having basement membranes anchored to Matrigel (ITGβ4⁺Col17A1⁺), basolateral membranes connected to neighboring cells (DSP⁺), and apical membranes associated with cells in the suprabasal layer (membrane-bound Notch1⁺DSP⁺). In contrast, the suprabasal layer cells (p63^-) had cytoplasmic membranes that were attached either to basal layer cells (membrane-bound Notch1⁺DSP⁺) or to each other (membrane-bound DSP⁺).¹⁰ During this developmental stage, plectin was predominantly localized to basement membranes (ITGβ4⁺) and cytoplasm (CK14⁺) in basal layer cells, indicating its critical function of HDs. Furthermore, plectin was detected on the basolateral and apical membranes of basal layer cells, the cytoplasmic membranes of suprabasal layer cells (DSP⁺). In addition, a fraction of plectin showed cytoplasmic localization in the suprabasal layer cells (CK13⁺) (Figure 5B and D).

As D3-organoids matured into multilayered SSEs with keratinized layers in the center of the organoids from day 9 to day 12, plectin was again detected in both basal and suprabasal layer cells. Representative immunofluorescence results of mature D3-organoids at day 10 are shown in Figure 5A and D. The +1 basal layer cells were identified as p63⁺CK14⁺, and the +2 to +4 suprabasal layer cells were identified as p63^-CK13⁺. The SSE of mature D3-organoids clearly showed the geometric stratified layers that the location of ITGβ4, Col17A1, Notch1, and DSP were similar to those of multilayered SSEs in D3-organoids during developmental stages. In the mature D3-organoids, the basal cells became more organized and the suprabasal cells became flatter and denser. The SSE structures of the mature D3-organoids closely resembled those of the rat in vivo. However, plectin was predominantly localized to the basement membranes (ITGβ4⁺) and cytoplasmic IFs (CK14⁺) of the basal layer cells. Plectin was also found in the basolateral (DSP⁺) and apical (Notch1⁺DSP⁺) membranes of basal layer cells, as well as in the cytoplasmic membranes (DSP⁺) and cytoplasm (CK13⁺) of suprabasal layer cells, indicating its critical function of DSMs (Figure 5B and D).

Plectin was identified as a downstream target of p63, as evidenced by its co-expression with p63 in the basal layer cells (Figure 5D).⁴⁰ Meanwhile, plectin also showed co-expression with activated Notch1 in the suprabasal layer cells (Figure 5D). To verify that plectin was a direct downstream target of Notch1 in the suprabasal layer, we performed cleavage under targets and tagmentation (CUT&Tag) sequencing analysis using the anti-Notch1 antibody on day 7 D3-organoids containing sufficient suprabasal cells. Motif analysis of the Notch 1 CUT&Tag data sequencing revealed that the plectin gene promoter was significantly enriched for NICD-RBPJ binding sites, especially between −4233 bp and +767 bp from the transcription start site (TSS), indicating that plectin could be directly regulated by Notch signaling (Figure 6A and B). According to the prediction of the Jaspar 2020 database (http://jaspar.genereg.net), 63 putative NICD-RBPJ binding sites were identified in the rat sequence (Figure 6C). The top 10 NICD-RBPJ binding sites are shown in Figure 6D. Among these binding sites, 4 were conserved across human, rat and mouse species (Figure 6D and E). Because there were 20 putative NICD-RBPJ binding sites ranging from −1960 bp to −760 bp in the rat sequence, we speculated that this region would be the core promoter of the plectin gene (Figure 6C). To confirm that the expression of the plectin gene was directly regulated by Notch1, we constructed 2 dual luciferase reporter plasmids using the putative core promoter from −1990 bp to −741 bp with high relative scores of the NICD-RPBJ binding sites (pGL3-PLEC-wild) or the NICD-RPBJ binding sites containing mutations (pGL3-PLEC-mutant), and then co-expressed them with a mammalian NICD1-expressing plasmid in D3 cells, respectively. As shown in Figure 6F, ectopic expression of NICD1 significantly increased the luciferase activity of pGL3-PLEC-wild by dual luciferase assay. In contrast, mutations in the “−1701 bp to −1692 bp” NICD-RPBJ binding site resulted in a significant decrease in the luciferase activity of pGL3-PLEC-mutant (Figure 6F). Thus, these results, together with the calcium-dependent keratinocyte differentiation assay, provided strong evidence for the direct regulation of plectin by Notch1 in differentiated esophageal SSE cells.

Figure 6.

Plectin regulation by Notch signaling in esophageal squamous epithelial cells. (A) Heatmaps of the Notch1 CUT&Tag signal density at promoter (+767∼−4233 bp) of PLEC. (B) Binding motif of the transcription factor NICD-RBPJ according to the CUT&Tag data. (C) Location of predicted NICD-RBPJ binding sites at PLEC promoter (−1∼−4100 bp) by the Jaspar database. T refers to the top 10 NICD-RBPJ binding sites. C represents the conserved NICD-RBPJ binding sites. (D) The tables list the top 10 NICD-RBPJ binding sites according to the relative scores and the conserved NICD-RBPJ binding sites by the Jaspar database. (E) Sequence homology of 4 conserved NICD-RBPJ binding sites in Rattus norvegicus (Norway rat), Mus musculus (house mouse), and Homo sapiens (human). (F) Notch signaling regulates the expression of PLEC. The image on the left shows the wild-type (WT) sequence of putative RBPJ binding site (T4) in the PLEC promoter (−741∼−1990bp) and the corresponding mutant (Mut) sequence. Western blotting was used to examine the level of NICD1 in D3 cells after transfection of pcDNA3.1-NICD1 plasmid for 48 hours (intermediate image). β-actin serves as an endogenous control. The ratios were obtained by the comparison of the band intensities of target proteins with β-actin. The dual luciferase reporter assay was conducted to detect the luciferase activity of D3 cells transfected with pcDNA3.1-NICD1 and pGL3- PLEC-wild or pGL3- PLEC-mutant plasmids for 48 hours (the right image). Relative luciferase activity is calculated as the ratio of firefly luciferase activity to renilla luciferase activity. ∗∗∗P <.001.

Collectively, the results presented in this section together with data reported previously⁴⁰ indicated that, although plectin was ubiquitously expressed in esophageal SSE cells, it exhibited complex localizations in different cell types across different tissue layers. Plectin expression in the basal layer cells was regulated by p63, whereas plectin expression in the suprabasal layer cells was regulated by Notch1 in the esophageal SSE. These specific plectin expressions in different cell types and layers would enable its associations with various macromolecular structures, such as HDs, DSMs, and/or CKs, thereby facilitating the participation of plectin in diverse cell functions during proliferation, differentiation, and homeostasis maintenance of the esophageal SSE.

Elucidation of Plectin and Plectin-associated Components/Factors in the Regulation of D3 Organoid Formation, Development, and Maturation

To determine whether, when, and how plectin was involved in the regulation of tissue homeostasis in the esophageal SSE, we generated two D3 cell lines, D3shPLEC1 and D3shPLEC2, that individually expressed a different plectin small interfering RNA (siRNA) sequence via a short hairpin RNA (shRNA) expression system using lentiviral transduction. Immunoblotting showed that both D3shPLEC1 and D3shPLEC2 cells had significantly reduced expression of endogenous plectin protein compared with parental D3 cells (Figure 7A). We cultured D3, D3shPLEC1, and D3shPLEC2 cells in Matrigel for 10 days. As shown in Figure 7B, these plectin-silencing D3-organoids (D3shPLEC1/D3shPLEC2-organoids, collectively referred to as D3shPLEC-organoids) displayed irregular morphologies with aberrant cell and tissue structures by hematoxylin and eosin (H&E) staining. Given the fact that plectin could affect cytoskeletal interactions via macromolecular structures such as HDs and DSMs for epithelial cell focal adhesion, tension, and cohesion,26, 27, 28 we detected ITGβ4 and Col17A1 as well as DSP in D3shPLEC-organoids. Immunofluorescence analysis showed that D3shPLEC-organoids exhibited the abnormal expression and localization of ITGβ4, Col17A1, and DSP (Figure 7C–F). The membrane-bound ITGβ4, Col17A1, and DSP were reduced, and the cytoplasmic Col17A1 was increased in D3shPLEC-organoids compared with those in D3-organoid cells. These results indicated that knockdown of plectin expression in D3shPLEC-organoids disrupted the organization of plectin-dependent HDs for cell-matrix interactions and DSMs for cell-cell conjunctions. To further verify these results, we performed transmission electron microscopy (TEM) on D3- and D3shPLEC-organoids and examined the ultrastructure of HDs and DSMs in cells of these organoids. Consistently, the ultrastructural HD-IF or DSM-IF plaques connecting cell-ECM or cell-cell in D3shPLEC-organoids were severely disrupted, with smaller HD or DSM plaque sizes and reduced the attachments of CKs to HDs and DSMs when compared with those in D3-organoids (Figure 7G).

Figure 7.

Deficiency of PLEC impairs HD and DSM structures in esophageal SSE. (A) Western blotting to detect the knockdown efficiency of plectin in D3 cells. β-actin serves as an endogenous control. The ratios were obtained by the comparison of the band intensities of target proteins with β-actin. (B) H&E images of D3shPLEC- and NC-organoids. Scale bars, 100 μm. (C) Expression and localization of Col17A1 and ITGβ4 in D3shPLEC- and NC-organoids at day 10 by IF assays. Scale bars, 100 μm. Arrowheads represent the localization of Col17A1 and ITGβ4 at the basement membrane of the basal layer in organoids. (D) Quantification of Col17A1 and ITGβ4 intensity at the basement membrane of the basal layer (n = 5 organoids). ∗P < .05; ∗∗P < .01. (E) Expression and localization of DSP in D3shPLEC- and NC-organoids at day 10 by IF assays. Scale bars, 100 μm. Arrowheads show the localization of DSP at the cell membrane of the suprabasal layer in organoids. (F) Quantification of DSP intensity at the apical membrane of the basal layer and the cell membrane of the suprabasal layer (n = 5 organoids). ∗P < .05; ∗∗P < .01. (G) Representative TEM micrographs of HD and DSM in rat esophagus and organoids showed that HD and DSM ultra-structures were disrupted in D3shPLEC-organoids. Scale bars, 200 nm.

Meanwhile, both D3shPLEC1 and D3shPLEC2 cells also showed a significant reduction in their organoid formation rate (OFR, ∼10%) (Figure 8A and B). These results were similar to the observations from our previous studies, in which the OFR of primary rat esophageal basal layer keratinocytes began to decline significantly after continuous passaging, which might be attributed to the gradual exhaustion of the stemness during the in vitro culture process.¹⁰ Knockdown of PLEC also affected the levels of HD components Col17A1 and ITGβ4 and disturbed the structure of HD at day 3 and day 5 (Figure 8C–E), indicating that plectin depletion affected HD organization, which was required for organoid formation. Moreover, the “p63⁺CK13⁺” cells, which were rarely observed in control D3-organoids, were detected in the abnormally enriched suprabasal layers of D3shPLEC-organoids (Figure 9A and B). These p63=positive cells (p63⁺CK14⁺ and p63⁺CK13⁺ cells) formed additional inward “ring-like” structures in the suprabasal layers with “p63^-CK13⁺” cells inside (Figure 9A and C). These results indicated that, in addition to the affected organoid formation, the processes of multilayered-stratification (ie, from the “p63⁺CK14⁺ basal” cell layer to the “p63^-CK13⁺ suprabasal” cell layer and then to the keratinized cell layer) were disrupted in D3shPLEC-organoids. Furthermore, increased numbers of CK13⁺CK14⁺ cells and Ki67⁺ cells but decreased number of CK13⁺CK14^- cells were observed in D3shPLEC-organoids compared with D3-organoids (Figure 9D–G), indicating that plectin depletion perturbed cell proliferation-differentiation. Taken together, these results demonstrated that plectin depletion perturbed organoid formation, cell proliferation-differentiation, and multilayered-stratification, which impaired the maintenance of SSE homeostasis in D3shPLEC-organoids.

Figure 8.

Effect of plectin deficiency in D3 organoids on organoid formation of esophageal SSE. (A–B) Representative bright field images of D3shPLEC- and NC-organoids cultured at day 10 (A) and quantification of organoid formation rate (B) of D3shPLEC- and NC-organoids (n = 3 experiments). ∗P < .05. Scale bars, 500 μm. (C–E) Expression, localization, and quantification of Col17A1, ITGβ4 in D3shPLEC-, D3shPLEC2-, and NC-organoids at day 3 and day 5 by IF assays (Day 3: n = 7 organoids for NC, 5 organoids for shPLEC1, 6 organoids for shPLEC2; Day 5: n = 5 organoids for NC, shPLEC1, and shPLEC2, respectively). ∗∗P < .01; ∗∗∗P < .001. Scale bars, 50 μm.

Figure 9.

Silencing of plectin disturbs the multilayered-stratification and proliferation-differentiation of esophageal SSE in D3 organoids. (A) Co-staining of p63 with CK13 or CK14 in D3shPLEC- and NC-organoids at day 10 of 3D culture using IF assay. Scale bars, 100 μm. (B) Percentage of CK13⁺ cells in p63⁺ cells per organoid (n = 5 organoids). ∗P < .05; ∗∗P < .01. (C) Silencing of plectin promoted the formation of p63⁺ ring-like structures in the suprabasal layer of D3shPLEC-organoids at day 10 of 3D culture (n = 6 organoids). Scale bars, 50 μm. (D–E) Ki67 staining (D) and percentage of Ki67⁺ cells (E; n = 5 organoids) in D3shPLEC- and NC-organoids. ∗P < .05; ∗∗P < .01. Scale bars, 100 μm. (F–G) Co-staining of CK13 with CK14 (F) and percentage of CK13⁺CK14^-, CK13^-CK14⁺, CK14⁺CK13⁺ cells (G; n = 5 organoids) in D3shPLEC- and NC-organoids at day 10 of 3D culture. ns, not significant; ∗P < .05; ∗∗∗P < .001. Scale bars, 100 μm.

To further dissect the complex functions of plectin in esophageal SSE, we decided to silence another HD component, Col17A1 or a DSM component, DSP, by RNA interference (RNAi) in D3 cells (Figure 10A). Meanwhile, we also inhibited Notch signaling, which was involved in regulating plectin expression of the suprabasal layer, through GSI treatment (Figure 10A). We examined D3 (control), shCol17A1- (D3shCol17A1), shDSP- (D3shDSP), and GSI-treated D3 cells for organoid formation, development, and maturation by snapshot microscopy analyses. D3shCol17A1-organoids exhibited unique non-smooth spheres with bulge or bump morphologies by H&E staining (Figure 10B). Immunofluorescence analysis of ITGβ4, Col17A1, plectin, and DSP showed that the Col17A1 depletion severely disrupted the organization of HDs in the basement membrane of basal layer cells, but did not affect the organization of DSMs in basal and suprabasal layer cells (Figure 10C–G). The OFR of D3shCol17A1 cells was also significantly reduced due to the disruption of the organization of HDs at day 3 and day 5, similar to that of D3shPLEC cells (Figure 11A–E). Furthermore, the “p63⁺CK14⁺ basal” cells were increased in D3shCol17A1-organoids with a reduction in ITGβ4 and plectin basement membrane staining and irregular cell-ECM junctions, resulting in additional layers of “p63⁺CK14⁺ basal” cells (Figures 10C–E, G and 12A). Consistently, increased Ki67⁺ cells were observed in the basal layers of D3shCol17A1-organoids compared with control D3-organoids (Figure 12D and E). In contrast, “p63^-CK13⁺ suprabasal” cells within D3shCol17A1-organoids appeared relatively normal with normal staining of DSP at cell membranes and cell-cell junctions, as Col17A1 was not expressed in the cells of “p63^-CK13⁺ suprabasal” layers (Figures 10D and F, 12A). Similar to D3shPLEC-organoids, an increased number of CK13⁺CK14⁺ cells but a decreased number of CK13⁺CK14^- cells were detected in D3shCol17A1-organoids (Figure 12F and G). These results indicated that depletion of Col17A1, a plectin-associated HD component,^5,6 mainly disrupted the organoid formation and basal layer homeostasis of D3 cells, resulting in proliferation-differentiation ‌perturbation.

Figure 10.

Aberrant plectin-associated genes and Notch signaling perturb HD and DSM structures in esophageal SSE. (A) Western blotting to detect the knockdown or inhibition efficiency of shCol17A1, shDSP, and GSI treatment in D3 cells. β-actin serves as an endogenous control. The ratios were obtained by the comparison of the band intensities of target proteins with β-actin. (B) H&E images of day 10 organoids treated with shCol17A1, shDSP, or GSI. Scale bars, 100 μm. (C) Expression and localization of Col17A1 and ITGβ4 in day 10 organoids treated with shCol17A1, shDSP, or GSI by IF assays. Scale bars, 100 μm. Arrowheads represent the localization of Col17A1 and ITGβ4 at the basement membrane of the basal layer in organoids. (D) Expression and localization of DSP in day 10 organoids treated with shCol17A1, shDSP, or GSI by IF assays. Scale bars, 100 μm. Arrowheads show the localization of plectin at the basement membrane of the basal layer and DSP at the cell membrane of the suprabasal layer in organoids. (E) Quantification of Col17A1 and ITGβ4 intensity at the basement membrane of the basal layer (n = 5 organoids). ns, not significant; ∗∗P < .01; ∗∗∗P < .001. (F) Quantification of DSP intensity at the apical membrane of the basal layer and the cell membrane of the suprabasal layer (n = 5 organoids). ∗P < .05; ∗∗P < .01. (G) Quantification of plectin intensity at the basement membrane and the apical membrane of the basal layer, as well as at the cell membrane of the suprabasal layer (n = 5 organoids). ns, not significant; ∗P < .05; ∗∗P < .01.

Figure 11.

Effect of aberrant plectin-dependent macromolecules/structures in D3 organoids on organoid formation of esophageal SSE. (A–B) Representative bright field images (A) and quantification of organoid formation rate (B) of NC-, D3shCol17A1-, D3shDSP-, and D3GSI-organoids (n = 5 experiments). ns, not significant; ∗∗P < .01. Scale bars, 200 μm. (C–E) Expression, localization, and quantification of Col17A1 (C, D) and ITGβ4 (C, E) in day 3 and day 5 organoids treated with shCol17A1, shDSP, or GSI by IF assays (Day 3: n = 7 organoids for NC, n = 7 organoids for shCol17A1, n = 11 organoids for shDSP, n = 7 organoids for +GSI; Day 5: n = 5 organoids for NC, n = 8 organoids for shCol17A1, n = 7 organoids for shDSP, n = 6 organoids for +GSI). ns, not significant; ∗∗P < .01; ∗∗∗P < .001. Scale bars, 50 μm.

Figure 12.

Regulation of esophageal SSE multilayered-stratification and proliferation-differentiation by plectin-dependent structures in D3 organoids. (A) Co-staining of p63 with CK13 or CK14 in D3shCol17A1-, D3shDSP-, D3GSI-, and NC-organoids at day 10 of 3D culture using IF assay. Scale bars, 100 μm. (B) Percentage of CK13⁺ cells in p63⁺ cells per organoid (n = 5 organoids). ns, not significant; ∗P < .05; ∗∗P < .01. (C) Effect of knockdown of Col17A1, DSP, and GSI treatment on the formation of p63⁺ ring-like structures in the suprabasal layer. The table below lists the number of p63⁺ ring-like structures formed in each organoid, shown as the mean ± SD (n = 6 organoids). Scale bars, 50 μm. (D–E) Ki67 staining (D) and percentage of Ki67⁺ cells (E; n = 5 organoids) in D3shCol17A1-, D3shDSP-, D3GSI-, and NC-organoids. ns, not significant; ∗P < .05. Scale bars, 100 μm. (F–G) Co-staining of CK13 with CK14 (F) and percentage of CK13⁺CK14^-, CK13^-CK14⁺, CK14⁺CK13⁺ cells (G; n = 5 organoids) in D3shCol17A1-, D3shDSP-, D3GSI-, and NC-organoids. ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001. Scale bars, 100 μm. (H) RESC-1-organoids showed the increased p63⁺CK14⁺ cells but the decreased p63^-CK13⁺ cells and perturbed the membrane-associated localization of plectin. Scale bars, 50 μm.

Depletion of the plectin-associated DSM component, DSP, in D3 cells did not significantly reduce the OFR of D3shDSP-organoids because DSP was not a component of HD, and HD formation was unaffected (Figure 11C–E). The OFR of D3shDSP-organoids was similar to that of control D3-organoids (Figure 11A and B). However, D3shDSP-organoids showed clear morphologic and tissue structural alterations with highly disorganized shapes (Figure 10B). Both “p63⁺CK14⁺ basal” cells and “p63^-CK13⁺ suprabasal” cells were disorganized, and abnormal “p63⁺” inward ring-like structures were also detected in the suprabasal layers of D3shDSP-organoids (Figure 12A–C). Although normal localizations of ITGβ4 and plectin to basement membranes were detected in D3shDSP-organoids, aberrant basolateral and apical membrane-associated plectin in “p63⁺CK14⁺ basal” cells and cytoplasmic membrane-associated plectin in “p63^-CK13⁺ suprabasal” cells were greatly reduced in D3shDSP-organoids (Figures 10C–G, 12A and B). Consistently, compared with control D3-organoids, D3shDSP-organoids also displayed increased numbers of CK13^-CK14⁺ and Ki67⁺ cells but a decreased number of CK13⁺CK14^- cells in the basal and suprabasal layers (Figure 12D–G). Thus, these results indicated that silencing of the plectin-associated DSM component, DSP, did not perturb the organoid formation of D3shDSP-organoids. However, depletion of DSP severely impaired the basolateral and apical membrane functions of “p63⁺CK14⁺ basal” cells and the cytoplasmic membrane functions of “p63^-CK13⁺ suprabasal” cells. This disruption significantly affected the proliferation-differentiation of “basal” and “suprabasal” cells and multilayered-stratification, ultimately perturbing the tissue homeostasis in the SSE of D3shDSP-organoids.

GSI treatment in D3-organoids (D3GSI-organoids) also did not affect the OFR of D3GSI-organoids (Figure 11A and B), and the localizations of plectin, Col17A1 and ITGβ4 to the basement membrane remained unchanged in D3GSI-organoids (Figure 10C–E, G). The “p63⁺CK14⁺ basal” cells were detected only in the +1∼+2 layer of D3GSI-organoids (Figure 12A), indicating that the organoid formation and cell lineage commitment of the basal layer (p63⁺CK14⁺ basal cells) with normal HDs and DSMs were not disrupted in D3GSI-organoids. In contrast, the locations and expressions of plectin, DSP, and CK13 were greatly affected in the “p63^-CK13⁺ suprabasal” cells of D3GSI-organoids (Figures 10D, F and G, 12A and B), resulting in the D3GSI-organoids exhibiting irregular, non-smooth bulge/papillary morphologies with disorganized stratification and differentiation of “p63-CK13⁺ suprabasal” cells (Figures 10B, 12A and B). However, GSI treatment did not increase the number of Ki67⁺ cells in D3GSI-organoids (Figure 12D and E). These results were consistent with the observations that Notch1 was mainly localized on the apical membrane of “p63⁺CK14⁺ basal” cells and on the cytoplasmic membrane of “p63^-CK13⁺ suprabasal” cells (Figure 5D) and the reports that the Notch signaling pathway was mainly involved in SSE stratification and differentiation.^52,53

We attempted to overexpress plectin in D3 cells but were unsuccessful due to the molecular weight of the plectin protein (∼600 kDa). Nonetheless, we found that RESC-1, an NMBzA-transformed rat ESCC cell line established in our lab²¹ that expressed extremely high levels of p63 and plectin proteins, formed RESC-1-organoids with non-smooth, crumpled, and disorganized cell and tissue structures. The RESC-1-organoids showed multiple layers of “p63⁺CK14⁺ basal” cells with very few “p63^-CK13⁺ suprabasal” cells (Figure 12H). Decreased membrane-associated plectin and increased cytosolic-associated plectin were observed in the RESC-1-organoids (Figure 12H), combining all the phenotypes observed in the D3shPLEC-, D3shCol17A1-, D3shDSP-, and D3GSI-organoids.

In summary, the results obtained in this section demonstrated that plectin was involved in the regulation of tissue homeostasis in the esophageal SSE of D3-organoids via proper plectin expressions, localizations, and plectin-associated various subcellular macromolecular structures (eg, HDs, and DSMs) in different cell types across different SSE layers. Perturbation of HDs by plectin or Col17A1 depletion in the basal layer diminished organoid formation and disrupted the proliferation-differentiation of the esophageal SSE basal layer. Perturbation of DSMs by plectin or DSP depletion in the basal layer and the suprabasal layers disrupted the proliferation-differentiation and multilayered-stratification of esophageal SSE. Disruption of the Notch signaling in the suprabasal layers perturbed the stratification and the proliferation-differentiation of the suprabasal layers of esophageal SSE (Figure 13).

Figure 13.

Effects of plectin and plectin-dependent macromolecules/structures on homeostasis of D3 organoids. (A) Schematic diagram to illustrate the perturbation of HD or DSM structures by knockdown or inhibition of plectin, Col17A1, DSP, and Notch signaling. (B) Knockdown or inhibition of plectin and plectin-associated proteins impaired the esophageal tissue homeostasis in the SSE of D3-organoids. ↓, decrease of the protein expression; -, unchanged protein expression; ⅹ, affected; √, unaffected.

The Involvement of Plectin in ESCC Carcinogenesis

As plectin played critical roles in regulating esophageal SSE homeostasis and plectin displayed genetic alterations and abnormal expressions in ESCC, we next determined the involvement of plectin in ESCC carcinogenesis. We first examined whether the expressions and localizations of plectin and plectin-associated macromolecular structures, HDs and DSMs, were altered in ESCC carcinogenesis using the NMBzA-induced ESCC rat model we established.²¹ Figure 14A shows representative H&E staining images of normal rat esophagus SSE, benign papilloma, and malignant carcinoma samples from the model. Immunofluorescence analysis demonstrated that similar to D3-organoids, plectin, Col17A1, and ITGβ4 were colocalized at the basement membrane (Figure 14C). In contrast, plectin and DSP were also colocalized at the lateral and apical membranes of p63⁺ basal cells, as well as in the cytoplasmic membrane of p63^- suprabasal cells of normal rat esophagus SSE (Figure 14B and C). The NMBzA-induced benign papilloma specimens, which were thickened and protruded outward to grow into papillary pattern morphologies, were similar to the SSE of D3GSI-organoids. Compared with the normal esophageal SSE, the expression of plectin was gradually increased and the expression of DSP was steadily decreased, leading to the increased cytosolic mislocalization of these proteins in the papilloma SSE samples. Although small increases in “p63⁺ basal” cells were found, very significant decreases in “p63^- suprabasal” cells with enhanced stratified layers were detected in the papilloma SSE samples. In addition, further increased expressions and enhanced cytosolic mislocalization of plectin, Col17A1, and ITGβ4, and decreased expression of DSP were observed in the ESCC samples. These NMBzA-induced ESCC samples clearly showed to have disrupted epithelial cell/tissue structures and a marked increase in “p63⁺” cells (Figure 14B–D). These results indicated that the expressions and localizations of HD or DSM components, especially plectin, were affected during ESCC carcinogenesis, ranging from mild benign papillomas to severe malignant carcinomas. Consistently, TEM revealed that both HD and DSM ultrastructures were compromised in RESC-1-organoids, resulting in weakened anchoring of the basal layer cells and impaired cell-cell conjunctions (Figure 14E).

Figure 14.

Involvement of plectin in rat ESCC carcinogenesis. (A) NMBzA-induced ESCC rat model and pathological progression from normal esophageal epithelium to ESCC by H&E staining. Study design: Animals were treated with NMBzA for 5 weeks, followed by promoter-agent (sorafenib) or vehicle for 20 weeks. Rats were sacrificed at 15, 20, and 25 weeks during the process to obtain esophageal tissue samples at different stages of ESCC carcinogenesis. W, weeks. Scale bars, 50 μm. (B–D) Representative IF images and quantification of p63 (B, D), Col17A1, ITGβ4, plectin, and DSP (C, D) in normal rat esophagus SSE, benign papilloma, and malignant carcinoma (n = 3 fields from 3 slides. Two rats for normal; 3 rats for papilloma; 3 rats for carcinoma. 20×). ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001. Scale bars, 50 μm. (E) Representative TEM micrographs of HD and DSM in rat RESC-1-organoids. Scale bars, 200 nm. Arrowheads represent the HD or DSM structures.

Then, we performed IHC analysis to determine whether HD (plectin/Col17A1), DSM (plectin/DSP), or plectin transcriptional regulators (plectin/p63/Notch1) were altered in human ESCC tissues using a tissue microarray containing 171 paired samples of human ESCC and their adjacent “normal” specimens. Representative images are shown in Figure 15A. Collectively, we found that, similar to normal rat esophageal SSE, the adjacent “normal’ tissues from human ESCCs displayed plectin/Col17A1 (HD) staining on the basement membranes of “p63⁺ basal” layer cells, and plectin/DSP (DSM) staining on both the basolateral and apical membranes of “p63⁺ basal” layer cells and the cytoplasmic membranes of “Notch1⁺ suprabasal” layer cells. In contrast, ESCC samples displayed misexpression (either underexpression or overexpression) and mislocalization of plectin, accompanied by disrupted cell and tissue structures (Figure 15B). The abnormal plectin expression results matched our RT-PCR data shown in Figure 2D. Notably, nine ESCC samples (9/171; ∼5%) exhibited undetectable plectin protein levels compared with their adjacent “normal” samples, while displaying high levels of Notch1 or p63 protein (Figure 15C). These results suggested that these ESCC samples harbored mutations in the plectin gene, consistent with the 5% mutation rate detected through WGS and WES analysis (Figure 1A). We summarized the expression levels of plectin, Notch1, and p63 (Figure 15D), and analyzed the expression correlations of these proteins in this microarray cohort. The results revealed that, compared with the adjacent “normal” controls, abnormal plectin expression was positively correlated with abnormal expression levels of p63 (P = .006; R = 0.207) or Notch1 (P < .001; R = 0.359) in human ESCC samples (Figure 15E). In agreement, the expression data derived from the Gene Profiling Interactive Analysis (GEPIA) database online also showed significant positive correlations of these proteins in esophageal cancer (Figure 15F).⁵⁴ Thus, the abnormal expression of the ESCC oncogene p63 or the tumor suppressor gene Notch1 could lead to the misexpression of plectin in human ESCC. Because the proper expression and localization of plectin, along with plectin-associated HDs and DSMs, and plectin transcriptional regulators, p63 and Notch1, were critical for maintaining esophageal SSE homeostasis. These results indicated that the misexpression and mislocalization of plectin, plectin-associated subcellular macromolecules, and plectin transcriptional regulators could initiate or promote carcinogenesis in human ESCC.

Figure 15.

Aberrant expression and localization of plectin in human ESCC tissues. (A) Expression and localization of plectin, p63, Col17A1, DSP, and Notch1 in human ESCC tissues by IHC assay. (B) Representative IHC images of misexpression and mislocalization of plectin in ESCC. Scale bars, 100 μm. Arrowheads show the localization of plectin in human ESCC tissues. (C) Expression of Notch1/p63 in ESCC samples with plectin mutations. (D) Distribution of plectin, Notch1, p63 expressions in 171 human ESCC samples. (E) Cluster and correlation analysis of the expressions of plectin, Notch1, and p63 in 171 human ESCC samples by Pearson’s test. (F) Correlation analysis of PLEC with TP63 and Notch1 in esophageal carcinoma using GEPIA database. TPM, transcripts per million.

Discussion

As a cytolinker and a critical component of HD, the role of plectin in the regulation of epithelial tissue mechanics has been extensively studied.^26,28 A large body of evidence has implicated that tissue mechanics is involved in cancer development, with changes in cell morphology and structure, cell-cell and cell-ECM interactions, and alterations in the tumor microenvironment all being closely linked to cancer initiation, progression, and metastasis.⁵⁵ Thus, it is conceivable that plectin may play a key role in cancer carcinogenesis. However, to date, only a few studies have been published to elucidate the role of plectin in cancer.⁵⁶ In this study, we unraveled the oncogenic characteristics of plectin involved in ESCC. We show that plectin, a member of the long tails of mutated genes found by large-scale WGS/WES in SCC/ESCC and NMBzA- or 4NQO-induced SCC/ESCC rodent models, is recurrently mutated at a low frequency in human ESCC. We also reveal that other plectin-associated genes are mutated in SSC/ESCC. The mutations in the plectin and plectin-associated genes are mutually exclusive. When all mutations are combined, the total number of mutations can exceed 50% of the SCCs examined. These results indicate that genetic alterations in plectin and plectin-associated HD and DSM components should be a common event in SCC/ESCC.

However, we also paradoxically observed that plectin is frequently overexpressed in ESCC samples, consistent with previous findings that plectin overexpression has been detected in various types of human cancers.^36,39,57,58 Moreover, in addition to previous reports that plectin expression is regulated by the oncogenic transcription factor p63,⁴⁰ we demonstrate that plectin is also a downstream target of Notch1, a frequently mutated tumor suppressor gene in ESCC. Although accumulating evidence supports the critical roles of p63 and Notch signaling involved in SSE tissue homeostasis and ESCC carcinogenesis, the precise downstream targets of p63 and Notch1 in esophageal SSE/ESCC have not been well-defined. Thus, genetic alterations, abnormal expressions, and transcriptional regulations of plectin, as well as aberrant plectin-associated HDs and DSMs in ESCC, led us to hypothesize that misexpression (either underexpression or overexpression) of plectin could disrupt its localization and function. These perturbations disrupt the functions of plectin-associated macromolecular structures, HDs and DSMs, ultimately leading to the dysfunctions of these structures and abnormalities in cell-ECM and cell-cell junctions, as well as signaling and mechanotransduction to drive/promote ESCC carcinogenesis.

To date, sophisticated methods used to investigate SSE tissue homeostasis, including determination of cell stemness, cell lineage tracing, cell competition, polarity, hierarchy, and stratification by in vivo animal models and/or in vitro organoids remain very challenging compared with those used in simple columnar epithelial tissues.^4,6,59 Furthermore, although we and others have been able to determine the tissue heterogeneity of SSE and identify the cell subpopulations of SSE by single-cell RNA sequencing (scRNA-seq), the precise locations of these cells in SSE could not be determined.^6,60, 61, 62, 63 Spatiotemporal in situ transcriptome and proteome analysis at the single cell level in SSE might provide solutions in the future. Nevertheless, we tested our hypothesis in vitro and in vivo via detailed analyses of the expression and localization of plectin, pectin-associated proteins, and regulatory proteins. We demonstrate that plectin is ubiquitously expressed in all proliferative and differentiative esophageal cell types in 2-D cell cultures, 3-D organoids, and esophageal tissue samples. However, the localizations and expression regulations of plectin in different cell types across different tissue layers were differentially regulated by plectin crosslinking to different macromolecular structures (eg, HDs and DSMs) and/or by different transcriptional regulators (eg, by the p63 in proliferative/basal cells or by the Notch1 in differentiative/suprabasal cells).

Previously, we and others have shown that HD components, DST and Col17A1, function as potential stem cell markers and that HD is required to maintain basal cell stemness and tissue homeostasis in the skin and esophageal SSE.^5,6 Plectin is a key protein that supports the crosstalk of HDs and FAs via crosslinking cytoskeletal filaments, CKs, and F-actins, to modulate force generation and mechanotransduction at the basement membranes of basal cells in SSE.²⁶ High-resolution imaging reveals that HDs/FAs form highly ordered interdigitating arrays of chevron patterns in 2-D cultured human keratinocytes.^26,28 Thus, HDs are not only the mainstay hubs but also participate in signaling and mechanotransduction in the basement membranes of SSE. In addition, plectin localizes to cell-cell conjunctions and often colocalizes with the DSM component, DSP, to couple the cortical F-actin belt and the IF rim.²⁸ Consistent with these findings and our hypothesis, depletion of plectin or another HD component, Col17A1, in esophageal keratinocytes D3 significantly reduces the formation of D3-organoids, demonstrating the important role(s) of HD in maintaining cell-ECM anchorage and basal cell stemness of SSE. However, unlike plectin depletion, which disrupts both basal and suprabasal cell proliferation-differentiation and multilayered-stratification, Col17A1 depletion mainly results in enhanced basal cell proliferation during D3-organoid development and maturation, as Col17A1 is not present in the suprabasal cells. In contrast, depletion of DSP, a component of DSM, in D3 cells disrupts both basal and suprabasal cell proliferation-differentiation and multilayered-stratification during D3-organoid development and maturation, as DSM is not only localized to suprabasal cell-cell conjunctions but is also present at basal/suprabasal cell-cell conjunctions.

In contrast to plectin depletion, DSP depletion does not affect the formation of D3-organoids, indicating that DSM does not play the role(s) in D3 cell-ECM anchorage/stemness maintenance. These results are consistent with the fact that DSM is not a subcellular structure at the basement membranes of basal cells in SSE. Taken together, these results demonstrate that plectin acts as a core component to control multilayered stratification of esophageal SSE tissue homeostasis. Plectin could couple HDs and DSMs with CKs to maintain cytostructural dynamics and orchestrate downstream signaling and/or mechanotransduction mediators, such as adaptor proteins (eg, Talin and Mob), receptor or non-receptor tyrosine kinases (eg, EGFR, FAK, and Src), and serine/threonine kinases (eg, MST/LATS and GSK) to regulate these processes. This includes targeting and controlling the nuclear import of transcriptional regulators (eg, YAP and β-catenin) to transcriptionally regulate the downstream targets required for these complex processes. In the future, these precise molecular mechanisms need to be elucidated with advancing technologies, such as high/super-resolution time-lapse microscopy, and in situ single-cell omics, which could be accelerated by artificial intelligence (AI).⁶⁴

A large body of evidence demonstrates that p63 functions as a critical “switch” to determine stem cell maintenance, basal/suprabasal cell lineage specification, and tissue homeostasis in mammalian SSE. As a transcription factor, p63 is thought to control a variety set of downstream targets involved in the regulations of these biological processes, although the precise molecular mechanisms remain elusive.65, 66, 67, 68, 69, 70 As p63 contributes to the maintenance of normal SSE homeostasis, abnormal expression and dysfunction of p63 are found in many SSE-derived cancers.^71,72 Gene amplification and overexpression of p63 are identified in more than 50% of human ESCC samples, resulting in a significant increase in basal-like cells with p63 overexpression and a marked decrease in suprabasal-like cells without p63 expression in ESCC.^17,20,40 Consistently, we show that plectin is regulated by p63 in non-differentiated esophageal keratinocytes, D3. We also show that, in contrast to control D3-organoids, the NMBzA-induced ESCC RESC-1-organoids display an increase in highly overexpressed p63 (p63⁺) basal-like cells and a decrease in p63 negative (p63⁻) suprabasal-like cells. Overexpression of plectin with disturbed localizations is also observed in RESC-1-organoids. Similar results are also observed in ESCC samples from the NMBzA-induced rat model and in tissue array samples from human patients with ESCC. Thus, the regulation of plectin by p63, together with the critical roles of both plectin and p63 in the functions of basal cells in esophageal SSE, indicates that this regulatory contributes to the oncogenic characteristics of plectin in ESCC.

Mutations in the Notch family of genes (Notch1–4) are frequently found in human SCC/ESCC, suggesting that the Notch signaling functions as a tumor suppressor in human SCC/ESCC carcinogenesis.73, 74, 75 However, the recent identification of high mutation rates in the Notch gene family (Notch1–4) in normal human esophageal or other SSE specimens suggests that other genetic and epigenetic alterations, in addition to Notch gene mutations, may be required for SCC/ESCC carcinogenesis. Abby et al have shown that “Notch1 mutations drive clonal expansion in normal esophageal epithelium but impair tumor growth,” providing an alternative explanation.⁷⁶ Given that plectin is a direct downstream target of Notch1, we show that inhibition of Notch signaling by GSI in D3 cells does not affect D3-organoids formation, suggesting that Notch signaling is not required for the cell-ECM anchorage/stemness maintenance in esophageal SSE in vitro. These results are consistent with the absence of activated Notch1 not being detected in the basement membranes of D3-organoid SSE and human esophageal SSE samples.⁷⁶ However, inhibition of Notch1 signaling by GSI in D3-organoids severely perturbs the “p63^-CK13⁺ suprabasal” cell differentiation process, resulting in the formation of benign papilloma-like D3-organoids. Similar morphologies have also been reported in the mouse esophageal organoids in which the Notch1 gene is knocked out of by the CRISPR method and in a genetic mouse model in which the Jag1/2 genes (Notch ligands) are knocked out, impairing activation of Notch signaling could disrupt esophageal SSE homeostasis and promote esophageal tumorigenesis.^77,78 Taken together, these results indicate that the underexpression of plectin, caused by LOF mutations in either the plectin gene or the Notch1 gene, which could abolish plectin expression required for the SSE differentiation, also contributes to the oncogenic characteristics of plectin in ESCC carcinogenesis.

While this manuscript was under review, Outla et al recently reported that plectin plays a crucial role in mechanical homeostasis and mechanosensitive oncogenic signaling that drives hepatocarcinogenesis.²⁸ Although these studies and ours are beginning to reveal the importance of plectin in carcinogenesis, several limitations need to be addressed in the future. These include focusing on more detailed hypothesis testing, such as validating genetic/epigenetic alterations of plectin and plectin-associated components that affect tissue homeostasis through cell lineage tracing both in vitro and in vivo. In addition, the critical downstream mechanical/signaling targets or pathways of plectin involved in carcinogenesis, particularly ESCC carcinogenesis, will be investigated.

In conclusion, our study demonstrates that a cytolinker and a critical component of HD, plectin, which crosslinks to a variety of macromolecular structures in esophageal SSE and whose expression is regulated by p63 or Notch signaling, plays a critical role in SSE tissue homeostasis. Perturbation of plectin expression (underexpression or overexpression) and disruption of plectin localization and function contribute to ESCC carcinogenesis. Therefore, the precise expression and localization of plectin and/or plectin-associated macromolecular structures in esophageal SSE may be utilized as potential biomarkers for ESCC diagnosis and therapeutic strategies.

Materials and Methods

Genomic Alteration and mRNA Expression Analysis of Plectin/Plectin-associated Genes

Genomic alteration rates of the plectin gene were calculated according to the genomic sequencing data of 973 ESCCs, including 639 WGS data, 215 WES data, and 119 targeted deep sequencing data from 5 published studies.15, 16, 17^,19,20 Mutation data of the plectin gene in ESCC cell lines (KYSE30, KYSE150, KYSE180, KYSE410, KYES450, KYES510, EC109, and TE-1) were acquired from the published sequencing studies, Cancer Cell Line Encyclopedia (CCLE) database and our sequencing data (Supplementary Table 1).^19,43,79 The clean data files of whole genome sequencing of EC109 have been deposited on GSA for humans (HRA008578, https://ngdc.cncb.ac.cn/gsa-human/browse/HRA008578).

Somatic mutations and copy-number alterations of HD/HD-associated components (PLEC, DST, Col17A1, ITGB4, ITGA6, CD151, KRT14, KRT5, LAMA3, LAMB3, LAMC2), DSM components (DSP, DSG1, DSG2, DSG3, DSG4, DSC1, DSC2, DSC3, JUP, PKP1, PKP2, PKP3) in 1640 SCC samples including 629 HNSCCs, 487 LSCCs, 297 CSCCs and 227 ESCCs were analyzed using cBioPortal (cBio cancer genomics portal), available at (https://www.cbioportal.org/).^44,45

The mRNA expression levels of plectin in 1010 SCCs including 523 HNSCCs and 487 LSCCs from 2 of TCGA PanCancer Atlas databases using cBioPortal analysis were assessed based on the mRNA expression z-score (threshold ±1.5) relative to normal samples (log RNA Seq V2 RSEM).

Ethics Approval

The present study was approved by the Ethics Committee of the National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences, and Peking Union Medical College. The studies involving animals were also approved by the Animal Ethics Committee of the National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College.

Tissue Samples

Tissue microarrays were prepared by Servicebio Co. Ltd based on primary ESCC tissues and adjacent normal esophageal tissues, including a total of 171 pairs of available samples without any preoperative treatment. The mRNA levels of plectin were detected in 102 paired ESCCs and their adjacent normal tissues. Informed consents were acquired from all patients with ESCC. The clinical characteristics of patients are listed in Supplementary Table 2.

Cell Lines and Cell Culture

hTERT immortalized rat normal esophageal epithelial cell line (REN-D3, D3 for short) was established and preserved by our lab.¹⁰ Rat esophageal squamous carcinoma cell line 1 (RESC-1) was generated from the NMBzA-induced rat ESCCs of our lab using multi-kinase inhibitor sorafenib as a tumor promoter.²¹ D3 and RESC-1 cells were cultured in complete medium, including DMEM/F12 (3:1) medium supplemented with 10% FBS (Thermo Fisher Scientific), 8 ng/mL cholera toxin (CELL Technologies), 5 ng/mL insulin (CELL Technologies), 25 ng/mL hydrocortisone (CELL Technologies), 0.1 ng/mL EGF (PeproTech), and 10 μM Y27632 (Topscience) in a humidified 37 °C incubator supplemented with 5% CO₂.

Human ESCC cell line EC109 was cultured in RPIM-1640 medium containing 10% FBS. It was established in our hospital lab in 1973 and deposited in the Chinese Cell Bank. It was authenticated by short tandem repeat (STR). The STR profile of EC109 was compared with those from the Chinese Cell Bank.

Calcium (Ca²⁺)-dependent Keratinocyte Proliferation-Differentiation Model

The complete medium was prepared as previously mentioned, which contains 1.16 mM/L Ca²⁺ as we tested. The conditioned medium consists of keratinocyte-serum-free medium (Thermo Fisher Scientific, Ca²⁺ free) supplemented with 0.5% dialyzed FBS(VivaCell Biosciences). The differentiation medium consists of conditioned medium and 0.6 mM/L Ca²⁺. The Ca²⁺-dependent keratinocyte proliferation-differentiation model was performed as follows: First, D3 cells were cultured in the complete medium. Then, D3 cells were treated with the conditioned medium for 48 hours. Third, D3 cells were induced to differentiation by switching the conditioned medium to the differentiation medium for 48 or 72 hours.

Z-scan Measurement

D3 cells cultured in the different media were fixed and stained with tubulin antibody (Sigma-aldrich, Cat# T5168) and DAPI. Zeiss confocal microscopy was used to perform z-scan along the z-axis of the cells. The depth of the cells was defined as the thickness within which tubulin fluorescence signals could be detected. Each group contains 5 random microscope fields from at least 3 experimental replicates.

γ-secretase Inhibitor

The usage of Compound E (Calbiochem), a γ-secretase inhibitor (GSI), reconstituted in dimethyl sulfoxide (DMSO) was previously reported that could block the activation of Notch signaling and the differentiation process.^52,80 In 2D cell culture, 0, 0.2, 1, and 5 μM of GSI were added to the culture medium to inhibit Notch signaling activation, and in 3D-organoid culture, 5 μM of GSI was used in the experiments.

Esophageal 3D Organoid Culture

In brief, 2000 cells mixed in 50 μL Matrigel were seeded per well into 24-well plates. The plate was inverted for Matrigel solidification at 37 °C for 20 minutes, and 500 μL of the advanced Dulbecco's Modified Eagle Medium (DMEM)/F12 medium containing R-spondin-1 (R&D Systems), Noggin (Pepro Tech), and ROCK inhibitor Y27632 (Sigma-Aldrich) was added and refreshed every other day. The number of growing organoids was photographed and then quantified using ImageJ software.⁸¹ RESC-1-organoids were cultured as described previously.⁶ All groups, including D3-, shPLEC-, shCol17A1-, shDSP-, and GSI-organoids, underwent at least 3 experimental replicates.

OFR was defined as the number of organoids matured /the number of cells seeded in Matrigel multiplied. The cell population proportions were calculated as the number of studied cells divided by the total number of nucleated cells in the corresponding matrix gel cross-section.

shRNA Construction

The lentiviral vectors stably expressing shRNAs targeting PLEC (shPLEC1, shPLEC2) were obtained from Syngentech Co. Ltd, and Col17A1 (shCol17A1), DSP (shDSP) were from OBiO Technology. The siRNA silencing approach was employed mainly due to the death of plectin-null mice within 2 to 3 days after birth.⁸² The D3-shPLEC, D3-shCOL17A1 and D3-shDSP cell lines were constructed by lentiviral transduction using the following sequences: shPLEC1: 5′-GCACAAGCCCATGCTCATAGA-3′, shPLEC2: 5′-GCGCATTGTGAGCAAGCTACA-3′, shCOL17A1: 5′-GGACCTATCACAACAACATAG-3′, and shDSP: 5′-CCAACAGAAGAATGACTAT-3′.

RNA Extraction and qRT-PCR

Total RNA was extracted using TRIzol reagent (Life Technologies). One μg RNA was reverse transcribed by the PrimeScript RT Master Mix (Takara), and the real-time PCR was performed using the SYBR Premix Ex Taq kit (Takara) in the ABI step-one Detection System (Applied Biosystems). The primers regarding PLEC and GAPDH were as follows: PLEC, Forward: ACACAGAGACTCTGGAGAAGG, Reverse: TTGATACGGTCTACCATGATCTTG; GAPDH, Forward: CATGCCGCCTGGAGAAAC, Reverse: CCCAGGATGCCCTTTAGT, synthesized by Sangon Co. Ltd.

FACS

Cells were digested and fixed according to the steps of the Cell Cycle and Apoptosis Analysis Kit (Beyotime) and were stained with propidium iodide. Flow cytometry (BD LSR II) was used. Cell DNA content analysis and light scattering analysis were performed using BD FACSDiva software.

Western Blotting

Western blotting was performed according to our previous paper.²¹ The primary antibodies were used against plectin (1:5000, Abcam; ab32528), TP63 (1:1000, Abcam, ab124762), Col17A1 (1:1000, Abcam, ab184996), Notch1 (1:1000, Cell Signaling Technology; #3447), HES1 (1:500, Signalway Antibody; 49016), CK13 (1:5000, Proteintech; 66684), IVL (1:1000, Proteintech, 55328), DSP (1:1000, Abcam, ab16434), and β-actin (1:5000, ZSGB-BIO; TA-09), respectively.

H&E and IHC Staining

H&E and IHC staining were performed as described previously.⁸³ The primary antibodies were used against plectin (1:800, Abcam, ab32528), Notch1 (1:200, Proteintech, 20687), Col17A1 (1:100, Abcam, ab184996), DSP (1:100, Abcam, ab16434; 390975) and p63 (1:500, Abcam, ab735, ab124762). The staining of IHC was assessed according to a previously reported study.⁸⁴ The intensity of staining was estimated as follows: negative (0), weak (1+), moderate (2+), and strong (3+). The distribution of positive staining was graded as 0 (<5%), 1+ (5%–30%), 2+ (31%–60%), and 3+ (>60%). Staining scores were calculated by the multiplication of both scores mentioned above, and final scores were defined as follows: 0, defined as a 0 score; 1 to 3 as a 1 score; 4 to 6 as a 2 score; 7 to 9 as a 3 score. Expression of plectin, Notch1, and p63 in human ESCC tissues was normalized with that in adjacent normal tissues.

Immunofluorescence and Super-resolution Microscopy

Cell and organoid or esophageal epithelium immunofluorescence were conducted according to our previously published papers.^10,21 The primary antibodies were used as follows: plectin (1:400, Abcam, ab32528), CK14 (1:400, Abcam, ab7800), CK13 (1:50, Santa Cruz, sc-57003), p63 (1:400, Abcam, ab735; 1:50, Abcam, ab124762), Notch1 (1:100, Cell signaling Technology, D1E11), integrinβ4 (1:50, Abcam, 29042), Col17A1 (1:100, Abcam, ab184996), DSP (1:100, Abcam, ab16434; 1:50, Santa Cruz, 390975), Ki67 (1:100, Abcam, ab625). A fluorescence microscope was used to detect red and green fluorescence at the excitation wavelength of 488 nm and 594 nm, respectively.

Fluorescence staining intensity was quantified using ImageJ software.⁸² In organoids at day 10, the intensities of Col17A1 and ITGβ4 at the basement membrane of the basal layer, as well as DSP at the apical membrane of the basal layer and the cell membrane of the suprabasal layer were quantified by delineating regions of interest (ROIs) in at least 5 organoids and calculating the mean intensity using the formula: Integrated Density/Area. Plectin staining at the basement membrane and the apical membrane of the basal layer, as well as at the cell membrane of the suprabasal layer was similarly analyzed by delineating ROIs from at least 5 organoids. In organoids at day 3 and day 5, the intensities of Col17A1 and ITGβ4 were determined by quantifying the entire organoids and calculating the mean intensity. Fold changes were calculated by normalizing the fluorescence intensities of these markers in the experimental groups (shPLEC-, shCol17A1-, shDSP-, and GSI-organoids) against the control group (NC-organoids). In samples from the rat NMBzA-induced ESCC model, the mean intensities of Col17A1, ITGβ4, DSP, and plectin were quantified, and the percentage of p63⁺ among nucleated cells was calculated and compared across normal rat esophagus SSE, benign papilloma, and malignant carcinoma, with at least 2 samples in each group. Plectin localization within the esophageal epithelium and organoids was visualized using the Interactive 3D Surface Plot of ImageJ software.

Super-resolution imaging was performed and analyzed using the N-STORM system in super-resolution microscopy (Nikon) as described previously.⁶⁴ The primary antibody was used against plectin (1:400, Abcam, ab32528).

TEM

Rat esophageal epithelium and organoids were fixed successively in a 2.5% glutaraldehyde solution and a 1% osmium tetroxide solution. They were then dehydrated using a gradient series of ethanol solutions (30%, 50%, 70%, 80%, 90%, 95%, and 100% concentrations) and pure acetone. Subsequently, the samples were processed with a mixture of embedding agent and acetone, and then with pure embedding agent. After embedding and heating overnight at 70 °C, the samples were sectioned using a LEICA EM UC7 ultramicrotome to obtain sections of 70 to 90 nm, and were stained with lead citrate and uranyl acetate for 5 minutes each and then dried. Observations and photography were conducted using a JEOL JEM-1200EX TEM.

CUT&Tag Sequencing

The organoids were processed into single cells as described previously.¹⁶ CUT&Tag library was performed according to the instructions of NovoNGS CUT&Tag 4.0 High-Sensitivity Kit (for Illumina) (Novoprotein). Briefly, the cells were fixed on the surface of ConA magnetic beads, and then the cell membrane was permeabilized by adding Digitonin. The primary antibody, secondary antibody, and transposon were incubated and cleaved in turn. The reaction products were extracted by DNA extraction and enriched by PCR to complete the construction of the next-generation sequencing library. The library preparations were sequenced on the Illumina Novaseq platform at Tianjin Novogene Bioinformatic Technology Co, Ltd, and 150 bp paired-end reads were generated. Raw data (raw reads) of fastq format were first processed using fastp (v 0.20.0) to obtain clean data. Index of the reference genome was built using BWA (v 0.7.12), and clean reads were aligned to the reference genome using BWA mem -k 32 -T 30 -t 4 -M. All peak calling was performed with MACS2 (v 2.1.0). Peaks were adjusted to the same size (500 bp) centered on peak summits, and motif discoveries of these loci sequences were performed using findMotifsGenome.pl program in HOMER (v 4.9.1) software. Peak-related genes were confirmed by ChIPseeker. The raw data files are available from the GSA database (CRA022955, https://ngdc.cncb.ac.cn/gsub/submit/gsa/subCRA029321/finishedOverview).

Dual-luciferase Reporter Assay

Dual-luciferase reporter assay was conducted by Shanghai Generay Biotech Co, Ltd.. PLEC promoter (-741∼-1990 bp, NC_005106.4:c117256755-117255506, Rattus norvegicus) containing RBPJ binding site (-1701∼-1692 bp: GGTTCCCACC) or mutant site (-1701∼-1692 bp: CCAAGGGTGG) was separately cloned into pGL3-basic luciferase reporter vector to form 2 constructs, pGL3-PLEC-wild and pGL3-PLEC-mutant. 1.5 × 10⁵ cells were seeded per well into 24-well plates. After incubation for 24 hours, 400 ng of pGL3-PLEC-wild or pGL3-PLEC-mutant were transfected into D3 cells, accompanied by the transfection of an equal amount of pcDNA3.1-NICD1 or pcDNA3.1-Empty vector (as a control) using lipofectamine 3000 (Life Technologies) following the manufacturer’s recommendations. 15 ng of pRL-SV40 containing renilla luciferase gene as an internal control, was co-transfected to standardize transfection efficiency. Firefly and renilla luciferase activities were detected using Dual-Luciferase Reporter Assay System (Promega) and Muti-mode Microplate Reader.

The NMBzA-induced Rat ESCC Model

The NMBzA-induced rat ESCC model was established as we previously described.⁶⁴ Briefly, male F344 rats were given subcutaneous injections of NMBzA for 5 weeks. Then, sorafenib was administered at week 6 for 20 weeks. The animals were sacrificed at 15, 20, and 25 weeks to obtain esophageal tissue samples. The esophageal epithelium and tumors were isolated, snap-frozen with liquid nitrogen, and stored at −80 °C. The remaining portion was fixed in 10% buffered formalin and embedded in paraffin.

Statistical Analysis

The Student’s t-test and analysis of variance (ANOVA) were used to analyze significant differences between groups, and P < .05 was considered significant. Data were presented as mean ± standard deviation (SD). GraphPad Prism version 8 was utilized for plotting. Correlation analysis was performed by Pearson’s correlation test using SPSS Version 25. All authors had access to the study data and had reviewed and approved the final manuscript.