This guide is an access and locator index, not an independent source of biological evidence. It records original-paper access and the scope of the experiments used in the EPCAM review (2026-10-24). A publication cache's full_text_available flag describes the automated fetch, which can differ from separately recovered original material. Public links below identify the actual papers, data or author copies; no publication text has been added as quotations.
| Reference | Original material and operative location | Evidence used and limits |
|---|---|---|
| PMID:9382878 | Official original, Methods; Results Figures 4, 8–9 and Mu1 tail-deletion comparison; Discussion | Human EPCAM constructs in murine E-cadherin cells and human HCA cells; quantitative aggregation and cytoskeletal association establish a contextual decrease in cadherin-mediated adhesion. These are not direct cadherin-binding experiments. |
| PMID:19785009 | Published article PDF, pp. 29–35; Methods and Figures 2–6 | Human HES2/HES3 surface expression, transient/stable knockdown, viable-cell counts and lineage-marker output. Main text and captions read; Figure 2 image inspected. Differentiation is from established stem cells; it is not acquisition of stem-cell identity. Supplements were not independently read. |
| PMID:23486470 | Official original, Experimental Procedures; Results Figures 1, 5–10 | Human T84/Caco-2 and intestinal tissue establish lateral EPCAM below tight junctions; endogenous complexes contain claudin-7, with claudin-1 association dependent on claudin-7. Knockdown, lysosomal inhibition and binding-impaired mutant rescue support claudin stabilization; the main Results paragraph following Figure 10B/C explicitly discusses residual stabilization by EpCAMmutHA, leaving residual binding below assay detection or additional mechanisms open. Figures 3 and 9–10 show increased junctional accumulation/barrier resistance after EpCAM depletion. The main Results before Figure 2 reports no tight EpCAM–E-cadherin co-IP, referring to Supplemental Figure 1B; that supplementary image was not independently inspected. Figures 1, 5 and 10 images inspected. These data do not make EPCAM a protease, prove binary CLDN1 binding, or exclude a different localization in every tissue. |
| PMID:19276185 | Publisher; published article PDF, Methods pp. 296–298; Results Figures 1–3 pp. 287–292; Discussion pp. 294–296 | Rat EpCAM/CLDN7 cDNAs expressed in human HEK and rat AS cells. TM-dependent association, oligomerization and cell aggregation/monolayer attachment were measured. Figure 3 shows increased attachment/aggregation with EpCAM alone and suppression when claudin-7 is associated; it does not support an adhesion-promoting complex. The operative binding partner is claudin-7, not a cadherin. Human host cells do not change construct species. |
| PMID:30185875 | Official original, Results Figures 2–5; Methods protein constructs, bead aggregation and FLIM-FRET | Human EpCAM-derived proteins, insect-cell recombinant preparations and tagged full-length constructs in human cells distinguish cis association from proposed trans interactions. E-cadherin positive-control beads aggregate; EpCAM preparations do not significantly do so. Mixed donor/acceptor cells show no corresponding trans FRET. Main Figures 4–5 images inspected. These bounded assays challenge a homophilic trans-adhesion mechanism; they do not refute indirect or claudin-associated contributions to epithelial contacts. Supplementary figures were not independently inspected. |
| PMID:32814053 | Author-institute original PDF; original supplementary archive, Results workflow, STAR Methods Y2H/DULIP/LuTHy; Table S2, mmc3.xlsx, sheet NN1.0 |
Exact EPCAM pairs: CDK5R1 row 2036, CALR row 3071, CDH1 row 3072, DLST row 3073, NEK7 row 6916 (Excel rows including header). Human cDNA-fragment interactions are mapped to genes. Pair-specific fragment boundaries, mammalian validation, endogenous location and native trans orientation were not established. Neither fragment is assumed to be an extracellular-domain construct. The CDH1 hit remains a supported non-core screen interaction, separate from the unresolved inherited cadherin-adhesion role. |
| PMID:18025791 | Publisher; author-uploaded original text, Methods and Results/Figure 7 caption | Human fetal/adult tissues are distinguished from a human-EpCAM transgenic mouse. Figure 7D is fetal human apical staining; Figure 7E is adult human basolateral staining; panels A–C concern the mouse transgene. The operative original text was read, not just the repository abstract. Expression is not alone proof of a developmental process role. |
| PMID:19136966 | Publisher article; actual Figure 1, Figure 4, Figure 5; original supplement, S4–S7 | Main text/complete Methods remain inaccessible; publisher PDF route returned preview HTML. Actual main Figure 1/4 images and captions, Figure 5 caption, supplement S4–S7 captions and S5 image were read. Human EpEX release, EpICD-containing native nuclear assembly, FHL2 interaction and reporter/cell-number effects support bounded processing/signaling claims. Complex-associated EMSA is not proof that EPCAM directly recognizes DNA. Exact endogenous cleavage coordinates are not assigned. |
| PMID:20064925 | Official original, Methods, surface-expression Results and EpICD ChIP/knockdown Results | Human hESC promoter-associated EpICD is distinct from functional knockdown in HCT116 carcinoma cells. ChIP can reflect complex-mediated chromatin association and does not establish direct EPCAM DNA binding. Histone modifiers acting on the EPCAM promoter are not EPCAM enzymatic functions. |
| PMID:19056867 | Official original, Results overview and Concise Methods; authors' NIH/NHLBI data, TACSTD1 row | Pooled healthy-human urine, differential centrifugation and LC-MS/MS. The target row is TACSTD1/NP_002345, one peptide, reference 2 (Gonzales et al., this paper). The historical GO-function column is not treated as an assay. This supports non-core exosomal occurrence, not EPCAM participation in exosome production. |
| PMID:18572020 | Official original, patient and laboratory Methods; Results Figures 2–3 | Human family linkage/segregation, pathogenic splice variants, duodenal RNA and diminished protein establish the congenital tufting enteropathy context. Disease-associated exon skipping is not a normal alternative protein product. |
The official UniProt record and independent FASTA agree exactly on the displayed 314-residue precursor. Signal peptide 1–23 is followed by chain 24–314 (PRO_0000022467); the membrane helix is 266–288. The extracellular and cytoplasmic topology segments are 24–265 and 289–314. The record lists no alternative products. Shed EpEX and released EpICD are processing products; the study's recombinant tail boundaries do not automatically define the endogenous cleavage site.
The actual PTHR14168 PAINT assertions and current PANTHER family tree place human P16422 below PTN008351906, PTN002680918 and PTN000366716. The last lies on the EPCAM side of the EPCAM/TACSTD2 duplication. RGD:621365 is rat Epcam/O55159; MGI:1861606 is mouse Tacstd2/Q8BGV3, a legitimate descendant source for deeper ancestral assertions. Human target evidence in its own IBA source set is expected. The cadherin-adhesion concern is the scope of the donor's experimental assertion, not a missing node, an invalid clade, donor count, or target self-inclusion.
The current HPA record reports supported plasma-membrane localization, with HPA067463 staining in MCF-7 and no staining in the displayed CACO-2/U2OS samples. SL-0153 maps lateral membrane to GO:0016328, SL-0265 maps tight junction to GO:0005923, and IPR043406 maps the EPCAM/Trop-2 family broadly to membrane. Correct source mappings do not resolve every biological specificity question.
Current GO:0048863 concerns acquisition of stem-cell features, distinguishing it from the lineage-output experiment reviewed here. GO:0098641 combines cadherin binding with an adhesion role and inherits the trans-cell contact requirement of GO:0098632; the human Y2H pair does not establish that combined role. GO:0045296 is a possible cadherin partner-class refinement, but native-protein binding remains unresolved given the construct and cellular-association limits above; the screened interaction is retained as non-core. GO:0050839 describes binding to a cell adhesion molecule and does not itself assert that EPCAM forms a trans adhesive bond.