Id3 induces an Elk-1-caspase-8-dependent apoptotic pathway in squamous carcinoma cells

Abstract

Inhibitor of differentiation/DNA-binding (Id) proteins are helix-loop-helix (HLH) transcription factors. The Id protein family (Id1-Id4) mediates tissue homeostasis by regulating cellular processes including differentiation, proliferation, and apoptosis. Ids typically function as dominant negative HLH proteins, which bind other HLH proteins and sequester them away from DNA promoter regions. Previously, we have found that Id3 induced apoptosis in immortalized human keratinocytes upon UVB exposure, consistent with its role as a tumor suppressor. To investigate the role of Id3 in malignant squamous cell carcinoma (SCC) cells (A431), a tetracycline-regulated inducible system was used to induce Id3 in cell culture and mouse xenograft models. We found that upon Id3 induction, there was a decrease in cell number under low serum conditions, as well as in soft agar. Microarray, RT-PCR, immunoblot, siRNA, and inhibitor studies revealed that Id3 induced expression of Elk-1, an E-twenty-six (ETS)-domain transcription factor, inducing procaspase-8 expression and activation. Id3 deletion mutants revealed that 80 C-terminal amino acids, including the HLH, are important for Id3-induced apoptosis. In a mouse xenograft model, Id3 induction decreased tumor size by 30%. Using immunofluorescent analysis, we determined that the tumor size decrease was also mediated through apoptosis. Furthermore, we show that Id3 synergizes with 5-FU and cisplatin therapies for nonmelanoma skin cancer cells. Our studies have shown a molecular mechanism by which Id3 induces apoptosis in SCC, and this information can potentially be used to develop new treatments for SCC patients.

Keywords: Apoptosis; Elk-1; HLH; Id3; SCC; caspase-8.

Figure 1

(A) Immunoblot result showing inducibility of Id3 in A431 stable cell lines (pooled clones and four representative clones) by tetracycline (1 μg/mL) for 24 h. GAPDH was used as a loading control. Clone 5 (Cl5) was chosen to proceed with experiments described. (B) A431/Id3 cells were cultured in low serum (0.1% FBS) and viable cell numbers were recorded for 13 days for both uninduced (no Tet) and induced (Tet) groups. (C) A growth curve for A431/Vc cells was determined under the same conditions as A431/Id3 cells. (D–G) A431/Id3 cells were cultured in 0.1% serum and cell cycle analysis performed to determine percentages of cells in phases of the cell cycle. (H) GFP-expressing A431/Id3 and Ds-Red-expressing A431/Vc were cocultured in 0.1% serum. Growth curves were determined and statistical analyses performed to determine any significant differences in growth. (I) Immunoblot showing Id3 and caspase-3 cleavage products for A431/Id3 cells. FBS, fetal bovine serum. P values ≤ 0.05 are considered statistically significant and represented by asterisks.

Figure 2

(A) Heat map showing 16 genes significantly up- or downregulated by Id3 by at least 1.5-fold. (B) RT-PCR was performed to verify genes that were shown to be upregulated by Id3 in microarray analysis. (C) Immunoblot analysis of proteins corresponding to genes upregulated by Id3.

Figure 3

(A) A431/Id3 cells in 0.1% FBS were treated with either scrambled or Elk-1 siRNA (siElk-1) for 48 or 72 h and assayed for protein levels. Densitometric quantification of immunoblot showing relative cleaved caspase-8 levels compared to those of GAPDH. (B–D) GFP-expressing A431 Id3 cells were treated with vehicle (B, DMSO), caspase-8 inhibitor (C) or pan-caspase inhibitor (D) in 0.1% FBS. Cell growth was assayed. (E) GFP-expressing A431/Id3 cells treated with either vehicle (DMSO), caspase-8, or pan-caspase inhibitor were grown in agarose and assayed for colony formation after 4 days. (F) SCC4, SCC9, and SCC25 cells were treated with scrambled or Id3 siRNA for 24 h, and assayed for Id3 and caspase-8 levels by immunoblot analysis. SCC, squamous cell carcinoma; FBS, fetal bovine serum. P values ≤ 0.05 are considered statistically significant and represented by asterisks.

Figure 4

(A) Schematic representation of Id3 deletions of the C-terminus (ΔC38), N-terminus (ΔN39), C+HLH (ΔC80) domain, or N+HLH (ΔN81) domain. (B) Induction of Id3 mutants by Tet (1 μg/mL for 24 h) were verified. (C–F) Growth curves for Id3-mutant cell lines. (G–J) Cell cycle analyses performed for Id3-mutant cell lines. HLH, helix–loop–helix. P values ≤ 0.05 are considered statistically significant and represented by asterisks.

Figure 5

(A) Mouse xenografts formed from GFP-expressing A431/Id3 cells imaged in vivo. (B) Id3 and Elk-1 protein levels were examined in A431/Id3 tumor lysates. (C,D) Xenograft sizes from GFP-expressing A431/Id3 cells and Ds-Red-expressing A431/Vc cells for both uninduced (−Dox) and induced (+Dox) groups were imaged in vivo. Results are expressed as mean ± SE (A431/Id3, −Dox: N = 10; +Dox: N = 14. A431/Vc, −Dox: N = 5; +Dox: N = 5). (E) Tumor sections from A431/Id3 xenografts were assayed for levels of active caspase-3 (top row) and Ki-67 (bottom row) by immunofluorescent staining in ±Dox groups. DAPI stains cell nuclei. Percentages of cells stained for both markers are shown in bar graphs on right (low Id3: N = 6, high Id3: N = 3). P values ≤ 0.05 are considered statistically significant and represented by asterisks.

Figure 6

(A–B) Cell viabilities were determined for Tet-induced A431/Id3 and A431/Vc cells treated with CDDP or 5-FU for 3 days. (C–D) A431/Id3 mouse xenografts were fed with control diet or with Dox to induce Id3. Dox-induced or -uninduced mice were then left untreated, or treated with CDDP (C), or 5-FU (D). Tumor sizes were determined as described in Materials and Methods. Results are expressed as mean ± SE (N = 6 for each group). Asterisks represent statistically significant differences in tumor sizes after chemotherapy in the presence or absence of Dox-induced Id3. P values ≤ 0.05 are considered statistically significant and represented by asterisks.