Surface keratin 1, a tumor-selective peptide target in human triple-negative breast cancer

Abstract

Targeting drugs to cancer cells via overexpressed cell-surface receptors has emerged as an effective therapeutic strategy for several cancers. However, identifying cell-surface receptors that allow selective uptake of targeting ligands by cancer cells-while sparing normal cells-remains a challenge, especially for triple-negative breast cancer (TNBC), which lacks a well-defined receptor for targeted delivery. In this study, immunohistochemical (IHC) analysis revealed that human TNBC patient tissues have significantly higher levels of keratin 1 (K1) compared to normal breast tissues. Among TNBC tissues, grade 3 tumors showed significantly higher (threefold) K1 expression compared to grade 2 tumors. We analyzed human TNBC and normal mammary epithelial cells to detect K1 from cell lysates using three methods: mass spectrometry, peptide mass fingerprinting, and Western blot. TNBC cell lysates confirmed the presence and high expression levels of 67 kDa K1. Importantly, intact cells showed that K1 is uniformly present on the surface of TNBC cells, while no or minimal cell-surface K1 was found in normal mammary epithelial cells using immunofluorescence confocal microscopy. Further, we show that cell-surface K1 was utilized by TNBC-selective peptide 18-4 for its uptake via cell-surface receptor (K1)-mediated endocytosis in TNBC cells, and the presence of peptide 18-4 did not affect the assembly of endogenous cytoplasmic K1. Taken together, our results demonstrate that K1 is overexpressed in human TNBC, and cell-surface K1 represents a promising new target for directed delivery in TNBC using targeting ligands such as peptide 18-4.

Keywords: Cell-surface keratin 1; Keratin 1 expression; TNBC; Targeted drug delivery; Tumor-selective peptide.

Fig. 1

Schematic showing the conjugate approach for targeted drug delivery via the cell-surface receptor (target protein) in cancer cells.

Fig. 2

Tumor-selective peptides for targeting keratin 1 (K1). (a) Schematic of K1 showing conserved rod region flanked by N-terminal head and C-terminal tail domains. The rod domain is divided into four α-helical units, 1a, 1b, 2a, and 2b linked via L1, L12, and L2 loops. a.a. stands for amino acid. (b) Sequence of peptide p160 and the engineered peptide analogues that display high binding to breast cancer cells and low/minimal binding to non-cancerous cells. X is Norleucine, lower case stands for D-amino acids, and cy stands for cyclic peptide with N- to C-terminal cyclization. (c) Chemical structure of peptide 18–4. L- and D-amino acids are shown in black and blue, respectively.

Fig. 3

K1 levels are significantly increased in human TNBC compared to normal breast and other breast cancer subtypes. (a) Representative breast tissue images showing IHC staining for K1. IHC analysis of human breast tissue array (TissueArray.com) was performed using rabbit monoclonal [EPR17744] to cytokeratin 1 (Abcam) and goat anti-rabbit IgG H&L (HRP polymer) secondary antibody (Abcam). Images of tissues from two patients each with TNBC, ER + /PR +, HER2 + or normal breast are shown (scale bar 50 μm). (b) IHC images of a full TNBC tissue core (scale bar 200 μm) and a 10X magnified region of the tissue. (c) Data for human tissues (from TissueArray.com) used for the IHC images shown above. IC invasive carcinoma (d) Quantification of K1 staining (dark brown) determined as immunoreactive score (IRS) from IHC images of tissues from normal breast (n = 8), HER2 + (n = 14), ER + /PR + (n = 20), and TNBC (n = 24) breast cancer subtypes (refer to Fig. S1 for patient details). Mean ± SEM; ns not significant; ***p < 0.001, Welch’s one-way ANOVA followed by Dunnett’s T3 post-hoc test. (e) Quantification of K1 staining (IRS) of grade 2 (n = 12) and grade 3 (n = 10) TNBC tissues compared to normal breast (n = 8) tissues. Mean ± SEM; *p < 0.05, **p < 0.01, Welch’s one-way ANOVA, followed by Dunnett’s T3 post-hoc test.

Fig. 4

Human TNBC cells overexpress K1. (a) Mass analysis of cell lysate (MDA-MB-468) using MALDI-TOF mass spectrometry. The mass spectrum was obtained in positive linear mode and shows a mass peak for K1 (found 66819, expected ~ 67 kDa). Peak for housekeeping β-actin (~ 42 kDa) was also observed. (b) Coomassie stained gel showing resolved protein bands from lysates of normal (MCF-10A and MCF-12A) and TNBC (MDA-MB-231, MDA-MB-436, and MDA-MB-468) cells. (c) Identification of keratin 1 (K1) after in-gel trypsin digestion of the 67 kDa band for MDA-MB-468 cell lysate and LC MS/MS analysis of peptide fragments (in red). Thirteen peptide sequences matched K1 which corresponds to 22% protein sequence coverage. (d) Western blot analysis using anti-K1 antibody (LHK1 clone) of cell lysates (15 μg) shows overexpression of K1 (at 67 kDa) in TNBC cells versus no/minimal K1 expression in normal breast cells. β-actin was analyzed for all cell lysates (TNBC and normal breast) as a loading control.

Fig. 5

Immunofluorescence analysis of CSK1 expression (green and orange) in human TNBC cells (MDA-MB-231, MDA-MB-436, and MDA-MB-468). (a) Z-section confocal microscopy images of a few representative cells show higher CSK1 expression in TNBC cells compared to the normal mammary epithelial cells (MCF-10A and MCF-12A). Cells were stained with anti-K1 mAb (green), cell-membrane stain (red), and the overlay is shown in orange. DAPI (blue) was used to stain the nucleus. Scale bar 15 μm. (b) Fluorescence intensity of CSK1 (% of cell membrane stain intensity) for TNBC and normal cells. Mean ± SEM; ***p < 0.001 (c) Representative Z-section confocal microscopy images show overlay (orange) of K1 and cell-membrane staining in a large number of TNBC cells. Scale bar 25 μm.

Fig. 6

CSK1 receptor-mediated endocytosis of peptide 18–4 by TNBC cells. Cells (MDA-MB-231 and MDA-MB-468) were incubated with FITC labeled peptide 18–4 (green) in the absence (-) and presence (+) of K1 mAb, followed by incubation with LysoView 640 (orange) for labeling lysosomes. Nuclei were stained with DAPI (blue). (a) Confocal Z-section images of representative cells along with overlay (yellow) are shown. Scale bar 20 μm. (b) Fluorescence intensity of FITC-peptide (% of lysosomes intensity) in TNBC cells in the absence and presence of K1 mAb. Mean ± SEM; ****p < 0.0001.

Fig. 7

The three negative-stain electron microscopy images show mature keratin 1/10 intermediate filament formation in a cell-free system in the absence or presence of K1-specific peptides (2 or 18–4).

Fig. 8

Schematic of a TNBC cell showing CSK1-mediated endocytosis of a PDC targeting K1. First, the peptide domain of PDC binds CSK1 (1), leading to endocytosis of PDC (2). Once internalized, the K1-PDC complex undergoes proteolytic cleavage and degradation in lysosomes (3) followed by release of the cytotoxic drug in the cytoplasm (4). Finally, the drug enters the nucleus to intercalate with the DNA (5) eventually causing TNBC cell death.