2025 Volume 58 Issue 2 Pages 45-57
Squamous cell carcinoma (SCC), a common malignancy affecting the skin, vagina, uterine cervix, anus, larynx, and upper digestive tract, is characterized by significant disruption of cell-cell adhesion in stratified squamous epithelium during tumorigenesis, progression, and metastasis. CALML5, a stratified epithelial-specific protein linked to desmosomal junctions, plays a key role in cell adhesion and is notably downregulated in human papillomavirus (HPV)-associated cervical SCC. Esophageal and pharyngeal cancers, commonly with a squamous cell phenotype, have distinct etiologies: oropharyngeal carcinoma is strongly associated with HPV, whereas esophageal carcinoma is linked to environmental factors such as smoking, alcohol, and diet. To investigate the role of CALML5 in these cancers, we performed immunohistochemical analyses on clinical samples and explored its regulatory mechanisms using in vitro studies with human esophageal SCC cell lines. Our findings revealed that CALML5 expression is suppressed in early-stage esophageal SCC but reactivated at invasive sites in well to moderately differentiated SCC undergoing keratinization. In specialized SCC with sarcomatoid component, CALML5 reactivation occurred alongside aberrant KLF4 expression, highlighting its context-dependent role in tumor progression. Conversely, while HPV-unrelated oropharyngeal SCC exhibited patterns similar to esophageal SCC, HPV-related oropharyngeal SCC consistently showed suppressed CALML5 expression due to impaired KLF4 nuclear translocation. These results suggest that CALML5 functions as a tumor suppressor in HPV-associated cervical SCC but may be reactivated in non-HPV-associated invasive SCC, emphasizing its complex role in SCC pathogenesis and the need for careful interpretation of its expression in clinical contexts.
The epidermis, particularly the squamous epithelium, serves as a critical barrier between external and internal environments [3, 16]. The squamous epithelium undergoes calcium-dependent terminal differentiation, progressing from the basal layer to the surface of the epidermis to form a robust, stratified squamous epithelium [6, 7]. The three main organs of the upper alimentary canal—the oral, pharynx, and esophagus—are lined by a mucosa composed of non-keratinizing squamous epithelium [40]. Malignant tumors in this region are primarily characterized by squamous cell carcinoma arising from the indigenous squamous epithelium [21, 43]. Despite being the same type of carcinoma, however, the tumorigenesis of oropharyngeal carcinoma differs significantly from that of squamous cell carcinoma in other parts of the upper gastrointestinal tract. In the oropharynx, there is a high incidence of tumors associated with human papillomavirus (HPV) infection [14, 30], similar to the development of cervical cancer in women [33]. In contrast, HPV-related tumors are rare in the rest of the upper alimentary tract, where carcinogenesis is mainly attributed to environmental factors such as smoking [14, 15], alcohol [2, 43], and diet [11].
A monoclonal antibody, 11B3 that specifically recognizes differentiating squamous epithelium, and identifies the antigenic target as Calmodulin-like protein 5 (CALML5) has been developed [19]. Calmodulin-like protein 5 (CALML5) belongs to the calmodulin family of calcium-binding proteins; it undergoes a conformational change upon binding calcium and contributes to the formation of a strong squamous structure [39]. Restricted expression of CALML5 gene in the stratified squamous epithelium is controlled by p63, ZNF750 and KLF4 cascade systems [19, 38, 39]. Transcription factor p63 is master transcription factors regulating terminal differentiation of the squamous epithelium [18]. Zinc-finger protein 750 (ZNF750) was reported to function as a transcriptional regulator of epidermal cell differentiation [4, 38]. Krüppel-like factor 4 (KLF4) is a zinc finger transcription factor that plays a critical role in regulating cell differentiation during development and epithelial homeostasis [37, 41]. Regarding CALML5 transcriptional regulation, its gene promoter contains binding sites for ZNF750 and KLF4 that regulate its cell-type and differentiation-specific expression [19, 39]. In previous study, we reported that CALML5 expression is downregulated in squamous cell carcinoma of the cervix from the early stages of dysplasia caused by HPV infection [19], and alteration of the p63-ZNF750-KLF4 axis results in the critical functional loss of CALML5 genes during cervix cancer progression. However, the expression manner of these genes in squamous cell carcinoma at esophagus and oropharynx is not well understood.
The aim of this study is to better understand the histopathological aspects of CALML5 during the tumorigenesis and progression of squamous cell carcinoma at esophagus and oropharynx using the antibody11B3 and elucidate the mechanism of how ZNF750 and KLF4 cascade system contribute to the regulation of CALML5 expression. In the present study, oropharyngeal carcinoma, the most common HPV-related carcinoma outside of the cervix was compared with squamous cell carcinoma of the esophagus that is less frequently associated with HPV. Specifically analyzed was the expression and localization of p63, ZNF750, and KLF4, transcriptional regulators involved in CALML5 expression.
Clinical cases from Ehime University Hospital between 2015 and 2023 were selected, enrolled, and evaluated (Table 1). The study period was in line with strict fixation methods and management protocols for fixation periods, essential for genomic medicine, implemented in 2015. A total of 37 esophageal lesions from 35 patients and 37 oropharyngeal lesions from 37 patients were analyzed. The pathological classification of the T factor (depth of invasion) for esophageal cancer cases included intramucosal (Tis-T1a, 18 lesions, 48.6%), submucosal invasive (T1b, 9 lesions, 24.3%), and invasion to the muscularis propria or beyond (T2-T3, 10 lesions, 27.0%). The pathological classification of the T factor (maximum tumor size) of oropharyngeal cancer cases was p16-positive (T1-T2, 23 lesions, 62.2%) and p16-negative (T1-T2, 14 lesions, 37.8%). T factors are based on T classifications (8th UICC classification system) for the esophagus [32] and the oropharynx [29]. Esophageal cancer cases were selected to match as closely as possible the age and sex of oropharyngeal cancer cases. None of the cases had undergone prior chemotherapy or radiation therapy. The specimens used were collected by mucosal resection or endoscopic surgery. This study was approved by the Ethics Committee of Ehime University Hospital (approval number: 2411003).
Patient Characteristics
| Esophagus (n = 37) | ||
|---|---|---|
| Sex, n | Male 31 | Female 6 |
| Median age, years (Range) | 72 (50–87) | |
| Depth of invasion (T), n | keratinization | CALML5 positive |
| pT1a(-EP, LPM) | 0/10 | 0/10 |
| pT1a(-MM) | 1/8 | 1/8 |
| pT1b(SM) | 9/9 | 9/9 |
| pT2 | 6/7 | 6/7 |
| pT3 | 1/3 | 2/3 |
| Oropharynx (n = 37) | ||||
|---|---|---|---|---|
| p16 negative (n = 14) | p16 positive (n = 23) | |||
| Sex, n | Male 12, Female 2 | Male 19, Female 4 | ||
| Median age, years (Range) | 77 (60–83) | 72 (58–86) | ||
| Depth of invasion (T), n | keratinization | CALML5 positive | keratinization | CALML5 positive |
| pT1 | 8/10 | 8/10 | 3/12 | 3/12 |
| pT2 | 3/4 | 3/4 | 4/11 | 4/11 |
Keratinization: more than 10% of tumor cells show a keratinizing tendency.
CALML5 positive: more than 20% of tumor cells show CALML5 immuno-activity at cell membrane.
The specimens were fixed in 10% neutral buffered formalin for 24 hr then processed for paraffin embedding and sectioning; after hematoxylin and eosin (H&E) staining, the specimens were examined under light microscopy by two pathologists. Tissue sections (5-μm thick) were deparaffinized with xylene and rehydrated through a series of graded alcohol. Antigen retrieval was carried out by microwave treatment at 350 W for 10 min in 10 mM citrate buffer (pH 6.0). Endogenous peroxidase activity was blocked with 0.3% H2O2 in methanol. The primary antibodies were applied, and the slides were incubated for 60 min at room temperature. The avidin-biotin-peroxidase complex method was used for immunohistochemical staining with the DAKO Liquid DAB Substrate Chromogen System (DAKO, cat. no. K3468, Tokyo, Japan). Monoclonal antibody 11B3, developed in our laboratory, was diluted tenfold with PBS (pH 7.4) and used as the primary antibody against CALML5 [19]. Mouse polyclonal antibodies raised against KLF4 (1:200, cat. no. ab106629, Abcam, Cambridge, UK), ZNF750 (1:100, cat. no. 21752-1-AP, Proteintech Group, Chicago, IL, USA), p63 (1:100, cat. no. M7317, DAKO, Glostrup, Denmark), and Roche CINtec® Histology p16 (cat. no. 705-4713, Roche Diagnostics, Heidelberg, Germany) were used for serial sections. Non-immunized, type-matched serum was used as a negative control. The final development of the sections was carried out with 3,3-diaminobenzidine containing 0.03% H2O2. Immunohistochemical staining was graded as negative (no staining), weakly positive (less than 30% of cells positive), positive (30–60% of cells positive), or strongly positive (more than 60% of cells positive with strong immunoreactivity), localized to staining regions (nucleus or cytoplasm).
Cell culture and preparationTE-4 [27] (RRID: CVCL_3337, highly differentiated esophageal squamous cell carcinoma), TE-8 [27] (RRID: CVCL_1766, moderately differentiated esophageal squamous cell carcinoma), TE-9 [27] (RRID: CVCL_1767, poorly differentiated esophageal squamous cell carcinoma), TE-10 [27] (RRID: CVCL_1760, highly differentiated esophageal squamous cell carcinoma), TE-11 [27] (RRID: CVCL_1761, moderately differentiated esophageal squamous cell carcinoma), ME-180 [8] (RRID: CVCL_1401, uterine cervix squamous cell carcinoma), and A431 [9] (RRID:CVCL_0037, skin squamous cells), all from RIKEN, Tsukuba, Japan, were cultured and maintained in α-MEM (Sigma-Aldrich, St. Louis, MO, USA), supplemented with 10% FBS (Sigma-Aldrich, St. Louis, MO, USA), 50 I.U./ml-50 μg/ml penicillin/streptomycin (ICN Biomedicals Inc., Aurora, OH, USA) and 2 mM L-glutamine (ICN Biomedicals Inc., Aurora, OH, USA). The cells at reaching confluence were collected by centrifugation (1500 rpm, 5 min). All the experiments were carried out with mycoplasma-free cells.
Western blottingCells were collected, washed and re-suspended in a lysis buffer containing 1 mM dithiothreitol (DTT), and phosphatase and protease inhibitors (cat. no.78440, Thermo Scientific, Waltham, MA, USA). Samples were loaded onto a precast Criterion TGX gel (Bio-Rad, Hemel Hempstead, UK) and transferred onto nitrocellulose membranes. Blots were blocked with 5% non-fat skim milk in Tris-buffered saline and Tween 20 (TBST) for 1 hr then incubated with the following primary antibodies: 11B3 (undiluted), KLF4 (1:200, cat. no. ab106629, abcam, Cambridge, UK), and ZNF750 (1:100, cat. no. 21752-1-AP, Proteintech Group, Chicago, IL, USA). Anti-mouse IgM-HRP, A-10668 (1:500) purchased from molecular probes (Eugene, OR, USA) and horseradish peroxidase-linked anti-mouse IgG secondary antibody (1:2000, cat. no. 7076S, Cell Signaling Technology, Danvers, MA, USA) were used as secondary antibodies. After three 10-min washes with TBST, the membranes were developed with an ECL kit (cat. no. RPN2232, Thermo Scientific, Waltham, MA, USA). Protein bands were quantified with ImageJ/Fiji software (version 2.1.0/1.53f; National Institutes of Health).
Quantitative real-time polymerase chain reaction (PCR)Total RNA was extracted from TE9 and TE10 cells by standard methods with a RNeasy Protect Mini kit (cat. No. 74106, Qiagen, Valencia, CA, USA). To assess the relative expression level of the mRNA of TRAcP, RANK, and NFATc1, 1 μg of total RNA was reverse transcribed to produce cDNA that was then amplified and quantified by the ABI PRISM 7300 Real Time PCR system (Applied Biosystems, Foster City, CA, USA) with the use of sets of primers and probes (Assay ID; TRAcP, Hs00356261_m1, RANK, Hs00187192_m1, and NFATc1, Hs00542678_m1, Thermo Fisher Scientific Inc, Tokyo, Japan). Rodent GAPDH primers and a probe (Assay ID; Hs00266705_g1, Thermo Fisher Scientific, Waltham, MA, USA) were used for standardization of relative mRNA expression.
Effect of blocking nuclear translocation of KLF4 proteinTo determine whether blocking nuclear translocation of KLF4 affects CALML5 gene expression, TE9 and TE10 cells were pretreated with graded concentrations of 10 uM Cycloheximide (CHX) [5] (cat. No. 037-20991, Fujifilm, Japan) and 10 uM MG132 [5] (cat. No. 139-18451, Fujifilm, Japan). For the morphological detection of KLF4, TE9 and TE10, cells were treated with the vehicle, 10 uM Cycloheximide (cat. No. 037-20991, Fujifilm, Japan) and 10 uM MG132 (cat. No. 139-18451, Fujifilm, Japan) for 5 min, fixed with 100% methanol, permeabilized by exposure to ambient air for 5 min, and treated with an anti-KLF4 antibody for 30 min. After washing 3 times with PBS (−), the cells were treated with Alexa Fluor 594-labeled secondary antibody (1:200, cat. no. A-11012, Thermo Fisher Scientific, Waltham, MA, USA) for 30 min; images were then visualized under a BioRevo microscope (KEYENCE, Tokyo, Japan). The relative amount of KLF4 exposed at the nucleus was estimated from measurements of KLF4 fluorescence signals. Intensities of KLF4 fluorescence signals in regions of interest were quantified using Image-J, then the ratios of KLF4 signals were calculated. Simultaneously, proteins from either the cytoplasmic or the nuclear fraction (at 2 and 5 min after treatment), and total RNA (24 hr after treatment) were extracted from TE9 and TE10 cells and subjected to Western blot analysis for KLF4 protein and real-time PCR, respectively.
Statistical analysisIn in vitro experiments, numerical values and error bars are represented as the means ± standard deviation (SD). Statistical analyses were carried out by one-way analysis of variance (ANOVA) with Scheffé post hoc tests.
Immunohistochemical evaluation of CALML5, p63, ZNF750, and KLF4 in normal nonneoplastic esophageal squamous epithelium (Fig. 1).
H&E stain (A and B), and immunohistochemical evaluation of CALML5 (C), p63 (D), ZNF750 (E), and KLF4 (F) in normal or nonneoplastic esophageal squamous epithelium. (A) Normal esophageal mucosal tissue and (B) magnified view of the basal portion (black box in A) (H&E stain). The following immunohistochemical images (C-D-E and F) correspond to magnified area (B). (C) CALML5 immunohistochemical staining. The magnified image within white box in the upper left corner shows CALML5-positive intercellular bridges (white arrow). (D) p63 immunohistochemical staining. (E) ZNF750 immunohistochemical staining. (F) KLF4 immunohistochemical staining. Bars = 150 μm.
In normal esophageal mucosal tissue (Fig. 1A and B), CALML5 was expressed in squamous cells from parabasal to intermediate layers and localized intercellularly, with staining patterns consistent with intercellular bridges, as described (Fig. 1C) [19]. Notably, cells in the basal layer lack intercellular bridges did not express CALML5 (white box in Fig. 1C). p63 was generally expressed in the nuclei of squamous cells including cells in the basal layer (Fig. 1D). ZNF750 was nuclear-positive in cells that largely overlapped those expressing CALML5 (Fig. 1E), while the nuclei in the basal layer were mostly negative (Fig. 1E). KLF4 was expressed in both the cytoplasm and nuclei of squamous cells, with decreased intensity towards the surface layer. Positive KLF4 staining was also observed in the nuclei of stromal cells (Fig. 1F).
Immunohistochemical evaluation of CALML5, p63, ZNF750, KLF4 in mucosal esophageal carcinoma (Fig. 2, Table 1)In all cases of non-invasive (Tis) or mucosal (T1a) and submucosal (T1b) esophageal carcinoma (n = 27), tumor cells proliferated in a downward-pressuring bulky fashion, while retaining the basement membrane in Tis (Fig. 2A) and the non-invasive area of T1a and T1b. While CALML5 expression was positive in the esophageal mucosal squamous epithelium in nontumorous areas (indicated by asterisk) compressed at the periphery, it was negative in proliferating tumor cells and in invasive or minimally invasive nests (Fig. 2B). p63 expression was positive in the nuclei of both nontumorous and tumorous squamous cells (Fig. 2C). ZNF750 expression pattern was similar to that of CALML5: positive in the nuclei of squamous cells at tumor-nontumor boundaries, but negative in most of the growing tumor (Fig. 2D). Like that of ZNF750, KLF4 expression was positive in the nuclei of squamous cell carcinoma at tumor-nontumor boundaries, but reduced or negative in most of the growing tumor cells (Fig. 2E).
H&E stain (A) and immunohistochemical evaluation of CALML5 (B), p63 (C), ZNF750 (D), KLF4 (E) in noninvasive esophageal carcinoma. In typical cases of noninvasive esophageal carcinoma, tumor cells proliferate in a downward-pressuring bulky fashion, while retaining the basement membrane (A, H&E stain). (B) CALML5 immunohistochemical staining. (C) p63 immunohistochemical staining. (D) ZNF750 immunohistochemical staining. (E) KLF4 immunohistochemical staining. Area within black box magnified in the upper right corner (B–E). Asterisks indicate the nontumorous areas. Bars = 150 μm.
Although CALML5 expression remained negative in most T2 and T3 tumor cells of esophageal cancer at mucosal layer, consistent with its absence in intraepithelial carcinoma, it was, nevertheless, observed in tumor foci displaying a tendency toward internal differentiation, keratinization and the formation of intercellular bridges (Supplementary Fig. S1A). CALML5 expression was observed in all two cases of the well-differentiated type, (Supplementary Fig. S1B), weakly positive in all four cases of the moderately differentiated type, and negative in two out of four cases of the poorly differentiated type (the remaining two poorly differentiated cases, later identified with EMT, were positive). p63 expression was positive in the nuclei of tumor cells (Supplementary Fig. S1C). ZNF750 expression was positive in the nuclei of tumor cells showing a tendency towards keratinization and is negative at the edge of tumor lesion (Supplementary Fig. S1D). The expression pattern of KLF4 was similar to that of ZNF750 (Supplementary Fig. S1E). These results indicate a positive correlation between the level of differentiation in esophageal squamous cell carcinoma and CALML5 expression, consistent with similar observations in p16-negative oropharyngeal cancer, as discussed later.
Immunohistochemical evaluation of CALML5, p63, ZNF750, KLF4 in invasive (T2-T3 stage esophageal carcinoma) with sarcomatoid component (Fig. 3, Table 1)At the invasive front of esophageal cancer, especially in poorly differentiated or undifferentiated phenotypes, cancer cells may adopt a spindle-shaped morphology, indicating sarcomatoid component. In the present study, of 10 cases of T2-T3 invasive esophageal cancer, two cases of poorly differentiated type exhibited this characteristic sarcomatoid component, even in focal areas (indicated by black arrows in Fig. 3A). Interestingly, sarcomatoid component of squamous cell carcinoma was uniquely positive for CALML5 (indicated by black arrows in Fig. 3B). CALML5 expression is spread in large tumor cells, but is more intracytoplasmic than intercellular as seen in normal structures (indicated by red arrows in Fig. 3B). CALML5 expression in these cells, however, was primarily more intracytoplasmic than intercellular, as seen in normal or differentiating structures within cancer foci. p63 expression was observed in the nuclei of nearly all tumor cells, including the spindle-shaped squamous cell carcinoma cells (indicated by black arrows in Fig. 3C). While the number of ZNF750 expressing cells was reduced (Fig. 3D), strong KLF4 expression was evident in squamous cell carcinoma cells exhibiting sarcomatoid component (Fig. 3E).
H&E stain (A) and immunohistochemical evaluation of CALML5 (B), p63 (C), ZNF750 (D), KLF4 (E) in a case of invasive esophageal carcinoma (poorly differentiated or undifferentiated) showing sarcomatous feature. (A) H&E stain of a specimen from a case of invasive esophageal squamous cell carcinoma with poorly differentiated or undifferentiated morphology. (B) CALML5 immunohistochemical staining. (C) p63 immunohistochemical staining. (D) ZNF750 immunohistochemical staining. (E) KLF4 immunohistochemical staining. Bars = 100 μm.
Oropharyngeal carcinoma cases were divided into two groups based on their p16 status and used as surrogate markers of human papillomavirus (HPV) infection (Table 1). The immunohistochemical expression patterns of CALML5, p63, ZNF750, and KLF4 in p16-negative oropharyngeal carcinoma (Fig. 4B) were generally similar to those observed in invasive esophageal carcinoma. CALML5 expression was observed in all four cases of the well-differentiated type, weakly positive in all four cases of the moderately differentiated type, and negative in all six cases of the poorly differentiated type. CALML5 expression was detected intercellularly in areas with a tendency toward keratinization (Fig. 4C). In these areas, ZNF750 expression was observed in both the nucleus and cytoplasm (Fig. 4D, black arrows), while KLF4 expression was weak but localized in the nucleus (Fig. 4E, black arrows). In contrast, immunostaining of p16-positive oropharyngeal carcinoma (Fig. 4G) revealed total absence of CALML5 expression in almost all tumor cells in all 23 cases (Fig. 4H). In this region, ZNF750 expression was mainly nuclear (Fig. 4I, black arrows), while KLF4 expression was strong but restricted to the cytoplasm (Fig. 4J). None of the negative controls exhibited significant staining, as shown in Supplementary Figs. S2 and S3.
Comparison of CALML5, ZNF750, and KLF4 expression in p16-negative (A–E) and p16-positive oropharyngeal carcinoma (F–J). (A, F) H&E staining. Immunostaining of the area in the black box (A) of the oropharyngeal carcinoma with negative p16 (B) is shown in C–E. (C) CALML5 immunohistochemical staining. (D) ZNF750 immunohistochemical staining. (E) KLF4 immunohistochemical staining. Immunostaining of the area in the black box (F) of the oropharyngeal carcinoma with positive p16 (G) is shown in H–J. (H) CALML5 immunohistochemical staining. (I) ZNF750 immunohistochemical staining. (J) KLF4 immunohistochemical staining. Magnified images are shown in the lower right corner (D, I, J). Bars = 100 μm.
CALML5 expression examined during the logarithmic growth phase in various cells disclosed a 16 kDa band in human skin keratinocyte cell line A431 and in poorly differentiated esophageal squamous cell carcinoma cell line TE-9 (Fig. 5A, yellow arrowheads). ZNF750 (Fig. 5B) and KLF4 (Fig. 5C) expression was, in nuclear and cytoplasmic fractions, detected predominantly in the nuclear fraction of TE-9 cells as 77 kDa (Fig. 5B) and 55 kDa (Fig. 5C) bands, respectively. Even in TE-10 and TE-11 cells, where CALML5 expression is absent in the logarithmic growth phase, ZNF750 and KLF4 expression is observed predominantly in the nuclear fraction, indicating that ZNF750 and KLF4, two important transcriptional regulators of CALML5 expression, are requisite but not sufficient for producing CALML5 protein.
Analysis of CALML5 (A), ZNF750 (B) and KLF4 (C) expression in nuclear and cytoplasmic fractions by Western blotting of various squamous cell carcinoma cell lines (A) Western blot analysis of CALML5. Immunoblot signals are detected as a 16 kDa band (indicated by yellow arrowheads). (B) Western blot analysis of ZNF750. Immunoblot signals are detected as a 77 kDa band. (C) Western blot analysis of KLF4. Immunoblot signals are detected as a 55 kDa band.
To examine the relationship between HPV-induced suppression of KLF4 nuclear translocation and the regulation of CALML5 expression in oropharyngeal cancer, the cytopathological condition of inhibited KLF4 nuclear translocation observed in oropharyngeal carcinoma was simulated using the squamous cell carcinoma cell line TE-9. To investigate the effect of inhibiting KLF4 translocation to the nucleus, CHX was administered to TE-9 cells. At 0 hr of the administration, KLF4 (red fluorescence in Fig. 6A) was localized exclusively in the nuclei of TE-9 cells; at 1-, 3-, and 5-hours post-administration KLF4 shifted from predominantly nuclear to both nuclear and cytoplasmic, and eventually to cytoplasmic only. Consistent with these immunocytochemical observations, inhibition of KLF4 nuclear transport by CHX and MG132 resulted in a decrease of CALML5 mRNA expression in TE-9 cells (Fig. 6B).
Effects of KLF4 nuclear transport (A) inhibition on CALML5 mRNA Expression (B). (A) KLF4 immunostaining is shown in the upper panels. DAPI staining is shown in the middle panels. Merged image of KLF4 and DAPI is shown in the bottom panels. (B) CHX and MG132, both inhibitors of nuclear translocation of KLF4, administered to TE-9 cells, and RNA collected after 24 hr. CALML5 expression was compared by Realtime PCR. Data are presented as the means ± standard deviation (SD), n = 4. *p < 0.05, one-way ANOVA.
In this study, the histopathological characteristics of esophageal and oropharyngeal squamous cell carcinoma were compared with a focus on CALML5 and the transcriptional regulators involved in its expression. Monoclonal antibody 11B3 was developed as the antigen with the use of normal human squamous epithelium from archival formalin-fixed, paraffin-embedded specimens [19]. Morphological screening with immunohistochemistry was then implemented to identify antigens with characteristic localization. The 11B3 antibody showed staining between the cells of differentiating squamous epithelium while sparing the basal layer. The antigen recognized by 11B3 was identified through LC/MS analysis as CALML5 [19], a calcium-binding protein related to the calmodulin family. CALML5 undergoes a conformational change upon binding calcium and associates with transglutaminase 3 to promote terminal squamous cell differentiation [1, 24, 25]. CALML5 expression is observed in the epidermis and is restricted to differentiating keratinocytes [19]. Regarding its transcriptional regulation, the CALML5 gene promoter contains binding sites for ZNF750 and KLF4 that regulate its cell-type and differentiation-specific expression [19, 39]. Indeed in the present study, CALML5 expression in the esophageal squamous epithelium was found at the overlapping region of ZNF750 and KLF4 expression. With this background, we investigated how CALML5 expression is altered during the carcinogenesis and progression of squamous cell carcinoma and how this relates to ZNF750 and KLF4.
The regulation of CALML5-ZNF750-KLF4 transcriptional machinery in esophageal and oropharyngeal carcinomaIn esophageal cancer, whole genome analysis has identified a number of cancer driver genes, including TP53, CDKN2A, PIK3CA, NFE2L2, and NOTCH1 [26, 28]. That these genes are involved in processes such as cell cycle, apoptosis, and cell differentiation aligns with histopathological observations that, in the early stages of esophageal squamous cell carcinoma development, differentiation from basal cells in the squamous epithelium is suppressed, and tumor cells exhibit basal cell-like characteristics [35]. As shown in Fig. 2, basal cell-like tumor cells in squamous intraepithelial carcinoma as well as in the basal layer of non-tumorous squamous epithelium formed localized foci that grew compressing their surroundings. In these regions, CALML5 expression was nearly absent and both ZNF750 and KLF4 were also downregulated. This phenomenon shows a basal cell-like phenotype where CALML5 expression is lost, and both ZNF750 and KLF4 are downregulated.
On the other hand, in esophageal invasive carcinoma, there may be significant heterogeneity within the cancerous tissue, leading to aberrant and disorganized keratinization and the formation of so-called cancer pearls [36]. While the mechanisms behind this heterogeneity are not fully understood, it is likely that the tumor microenvironment and interactions with stromal cells contribute to the abnormal redifferentiation of squamous cell carcinoma [12, 23]. The partial reactivation of CALML5 observed during the invasion and progression of esophageal squamous cell carcinoma resembles the pattern seen in p16-negative, non-HPV-associated oropharyngeal carcinoma, where CALML5 expression is localized at the plasma membrane and is consistent with the aberrant expression of ZNF750 and KLF4 (Fig. 7B). Thus, esophageal squamous cell carcinoma and HPV-unrelated oropharyngeal carcinoma show similarities of CALML5 expression patterns.
Schema for regulation of CALML5 expression. (A) Squamous carcinoma showing a basal cell-like phenotype. In the noninvasive esophageal carcinoma, both ZNF750 and KLF4 are decreased, and their nuclear localization is nullified, as observed in the basal cell layer of normal esophageal squamous epithelium (shown as basaloid phenotype). (B) Invasive squamous carcinoma with sarcomatoid component. In poorly differentiated squamous cell carcinoma with sarcomatous feature, CALML5 expression, previously lost, is restored in a small number of cases. CALML5 expression restoration may be due to aberrantly expressed KLF4 in tumor cells sarcomatous feature. (C) Human papillomavirus (HPV)-associated squamous carcinoma. The inhibition of KLF4 nuclear translocation by HPV proteins prevents the activation of CALML5 expression.
In squamous cell carcinoma with sarcomatoid component, a rare variant of poorly differentiated type) where cancer cells lose their focal organization at the site of invasion and invade to dissolve into surrounding stromal tissue, resulting in pseudosarcomatous change [20, 31]. In such cases, CALML5 expression, as shown in Fig. 3, shifts from the plasma membrane to the cytoplasm. In pseudosarcomatous areas, KLF4 expression is increased in the nuclei of large tumor cells, likely reflecting a transformation of the tumor cells toward stromal cell characteristics. Since ZNF750 expression was somewhat maintained in sarcomatoid component (Fig. 3D), the elevated KLF4 expression might trigger the aberrant expression of CALML5. Given that CALML5 expression in this pseudosarcomatous state is confined to the cytoplasm and does not reach the cell surface where it typically functions [39], its apparent expression assessed through immunohistochemistry may have limited functional significance as a tumor suppressor. The present and previous studies [10, 22, 44] have demonstrated that the nuclear translocation of KLF4 is blocked due to the persistent expression of HPV proteins resulting from the integration of HPV DNA into host nuclear DNA (Fig. 7C). In p16-positive or HPV-associated oropharyngeal carcinoma even when ZNF750 and KLF4 are aberrantly expressed, the inhibition of KLF4 nuclear translocation prevents the reactivation of CALML5 [19], leading tumor cells to retain basal cell-like traits or a poorly differentiated phenotype with minimal keratinization. Also, the suppression of KLF4 nuclear translocation caused by persistent HPV infection may also inhibit the high KLF4 expression typically associated with epithelial-mesenchymal transition (EMT) [13], thereby preventing EMT in oropharyngeal carcinoma. These studies could explain why HPV-associated oropharyngeal carcinoma often displays a basal cell-type cancer nest, characterized by poor differentiation in terms of keratinization and intercellular bridges. Therefore, caution needs to be exercised when determining cancer grade based solely on the expression of squamous differentiation markers such as keratinization [34] and CALML5 [17, 42].
In conclusion, oropharyngeal cancers that are not associated with HPV exhibit characteristics similar to those of esophageal squamous cell carcinoma, particularly regarding CALML5 expression. In contrast, in HPV-associated oropharyngeal carcinoma, CALML5 expression remains weak due to the inhibition of KLF4 nuclear migration that preserves basaloid traits.
The authors have no conflicts of interest.
Taniwaki M performed the histopathological diagnosis, immunohistochemical studies and wrote the paper; Kitazawa, R and Ono T carried out the histopathological diagnosis and interpretation of immunohistochemical data; Haraguchi R, and Takaoka Y carried out cell biology experiments and analyzed the data; Kitazawa S designed and coordinated the research.
This study is supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan (22K06981, 24K02251, K2410000). We express sincere thanks to Ms. Mariko Hashimoto for her excellent technical assistance, and to Dr. Tetsuro Nishihira, Tohoku University, for providing esophageal squamous cell carcinoma cell lines.
