Targeting homeostatic mechanisms of endoplasmic reticulum stress to increase susceptibility of cancer cells to fenretinide-induced apoptosis: the role of stress proteins ERdj5 and ERp57

Abstract

Endoplasmic reticulum (ER) malfunction, leading to ER stress, can be a consequence of genome instability and hypoxic tissue environments. Cancer cells survive by acquiring or enhancing survival mechanisms to counter the effects of ER stress and these homeostatic responses may be new therapeutic targets. Understanding the links between ER stress and apoptosis may be approached using drugs specifically to target ER stress responses in cancer cells. The retinoid analogue fenretinide [N-(4-hydroxyphenyl) retinamide] is a new cancer preventive and chemotherapeutic drug, that induces apoptosis of some cancer cell types via oxidative stress, accompanied by induction of an ER stress-related transcription factor, GADD153. The aim of this study was to test the hypothesis that fenretinide induces ER stress in neuroectodermal tumour cells, and to elucidate the role of ER stress responses in fenretinide-induced apoptosis. The ER stress genes ERdj5, ERp57, GRP78, calreticulin and calnexin were induced in neuroectodermal tumour cells by fenretinide. In contrast to the apoptosis-inducing chemotherapeutic drugs vincristine and temozolomide, fenretinide induced the phosphorylation of eIF2alpha, expression of ATF4 and splicing of XBP-1 mRNA, events that define ER stress. In these respects, fenretinide displayed properties similar to the ER stress inducer thapsigargin. ER stress responses were inhibited by antioxidant treatment. Knockdown of ERp57 or ERdj5 by RNA interference in these cells increased the apoptotic response to fenretinide. These data suggest that downregulating homeostatic ER stress responses may enhance apoptosis induced by oxidative stress-inducing drugs acting through the ER stress pathway. Therefore, ER-resident proteins such as ERdj5 and ERp57 may represent novel chemotherapeutic targets.

Figure 1

Fenretinide or thapsigargin induced cell death in SH-SY5Y and A375 cells. Cells were treated with fenretinide at 3 μM (SH-SY5Y cells) or 15 μM (A375 cells), or thapsigargin at 1.5 μM (SH-SY5Y cells) or 7.5 μM (A375 cells) for 24 h and cell death (apoptosis) measured by flow cytometry. Bar heights are mean percentage apoptosis of three replicates+95% confidence interval (error bars).

Figure 2

The induction of ER-stress genes in response to fenretinide or thapsigargin in SH-SY5Y and A375 cells. (A) Western blots of total protein extracted from SH-SY5Y cells (left-hand column) or A375 cells (right-hand column) treated with thapsigargin for 24 h (Thap; 1.5 μM for SH-SY5Y cells or 7.5 μM for A375 cells) or fenretinide (FenR; 3 μM for SH-SY5Y cells or 15 μM for A375 cells) for 0, 6, 18 or 24 h. Blots were probed with antibodies to GADD153, ERp57, calreticulin, calnexin, GRP78 and, as a loading control, β-tubulin. Apparent molecular weights of the relevant bands are given on the left. (B) Reverse transcription–polymerase chain reaction was used to quantify ERdj5 and GAPDH (as a loading control) in RNA extracted from cells treated with thapsigargin or fenretinide. In the PCR reactions the number of cycles used for each primer pair was adjusted so that amplification remained within an approximately linear range. (C) Western blots (as in A) probed using antibodies for eIF2α and phosphorylated eIF2α (P-eIF2α). Cells were treated with fenretinide for 0.25, 0.5, 1, 2 or 4 h, or thapsigargin for 4 h, (D) induction of XBP-1 splicing in response to fenretinide or thapsigargin as in (B) with additional experiments to show lack of XBP-1 splicing in cells treated with 10 nM vincristine (SH-SY5Y cells) or 1 mM temozolomide (A375 cells) for up to 24 h.

Figure 3

Induction of ATF4 mRNA in SH-SY5Y cells (upper graph) and A375 cells (lower graph) in response to treatment with fenretinide (3 or 15 μM respectively) for 24 h, or with thapsigargin (1.5 μM or 7.5 μM, respectively) or 10 nM vincristine (SH-SY5Y cells) or 1 mM temozolomide (A375 cells) for 6, 18 and 24 h. Bar heights and error bars are means and upper range of duplicate samples or means and upper standard deviation of triplicate samples (thapsigargin data) relative to the control (vehicle) treatment.

Figure 4

Blocking ROS with an antioxidant inhibits fenretinide-induced stress responses in SH-SY5Y and A375 cells. In all experiments, fenretinide was used at 3 μM to treat SH-SY5Y cells and at 10 μM to treat A375 cells; vitamin C (100 μM) was added to cells 2 h before adding fenretinide and incubation continued in the presence of both reagents for a further 22 h. (A) A comparison of ROS induction in response to fenretinide (FenR), thapsigargin (Thap), vincristine (Vin; SH-SY5Y cells), temozolomide (Temo; A375 cells) or velcade (Vel)in SH-SY5Y cells (left panel) and A375 cells (right panel). ROS levels are expressed relative to control vehicle treatment and were measured with two different probes for reactive oxygen intermediates: CM-H₂DCFDA or DHE in separate experiments. Bar heights are means plus upper 95% confidence interval of CM- H₂DCFDA fluorescence (black bars) or DHE fluorescence (grey bars) (n=3, but n=6 for control and fenretinide-treated cells). ROS was measured after 6 h for control, fenretinide, thapsigargin, vincristine or temozolomide treatments, but for velcade after 3, 5, 7, 12 and 16 h using the DHE probe and 6 h with CM-H₂DCFDA or DHE probes. (B) left and centre panels: flow cytometry CM- H₂DCFDA-fluorescence profiles for SH-SY5Y (included as a positive control) or A375 cells showing the induction of ROS by fenretinide (FenR) and the reduction in fenretinide-induced ROS in cells after pretreatment with 100 μM vitamin C 2 h before adding fenretinide (VitC+FenR). Vitamin C alone (VitC) did not increase ROS above the background control. Ordinate, event counts; abscissa, fluorescence signal intensity (FL1-H). Right panel: abrogation of fenretinide-induced apoptosis in A375 cells by vitamin C (mean and upper 95% confidence interval of three replicates). (C) Western blots showing that the fenretinide-induced increase in expression of ERp57 and GRP78 protein in response to fenretinide was inhibited by pretreatment of SH-SY5Y or A375 cells with vitamin C. (D) The fenretinide-induced increase in ERp57 and ERdj5 mRNA in SH-SY5Y and A375 cells, as measured by real-time quantitative PCR, was inhibited by pretreatment with vitamin C. Bar heights are means plus 95% confidence limit for triplicates samples. Contrasts were used within one-way ANOVA to compare relative ERp57 or ERdj5 mRNA levels in cells treated with fenretinide or fenretinide plus vitamin C; ^**P<0.01, ^***P<0.001. For ERp57 or ERdj5 in SH-SY5Y cells, one-way ANOVA F_3,8>39, P<0.0001, contrasts: fenretinide versus fenretinide plus vitamin C F_1,8>112, P<0.0001. For ERp57 or ERdj5 in A375 cells, one-way ANOVA F_3,12>6, P⩽0.007, contrasts: fenretinide versus fenretinide plus vitamin C F_1,12>13, P=0.003. In A-D, control, vehicle control; vitC, cells treated with 100 μM alone; FenR, SH-SY5Y or A375 cells treated with 3 μM or 15 μM fenretinide, respectively; vitC+FenR, cells pretreated with 100 μM vitamin C 2 h before adding fenretinide.

Figure 5

SiRNA-mediated knockdown of ERdj5 increased apoptosis in response to fenretinide in SH-SY5Y and A375 cells. In all experiments, SH-SY5Y cells were treated with fenretinide at 3 μM and A375 cells with fenretinide at 10 μM. For the A375 cells, a slightly lower concentration of fenretinide was used to ensure that any increased cell death remained within a 20–80% range. Cell death (apoptosis) was measured by flow cytometry. Two different ERdj5 siRNAs were evaluated against a scrambled siRNA control, and at doses of 20 and 40 nM. The upper graphs show the percentage apoptosis of SH-SY5Y cells (left-hand column) and A375 cells (right-hand column) and the lower graphs show real-time quantitative PCR data to verify ERdj5 knockdown for SH-SY5Y cells and A375 cells, respectively. Cells were transfected with 20 or 40 nM scrambled control siRNA (siRNA control), 20 or 40 nM ERdj5-2 siRNA, or 20 nM or 40 nM ERdj5-3 siRNA, and subsequently treated with (gray bars) or without (black bars) fenretinide. Real-time quantitative PCR data are expressed relative to the 20 nM scrambled siRNA control in the absence of fenretinide. On all graphs, error bars are the upper 95% confidence limits.

Figure 6

SiRNA-mediated knockdown of ERp57 increased apoptosis in response to fenretinide in SH-SY5Y and A375 cells. Experimental details as in the legend to Figure 6. Cells were transfected with 20 or 40 nM scrambled control siRNA (siRNA control), 20 or 40 nM ERp57-1 siRNA, or 20 or 40 nM ERp57-2 siRNA, and subsequently treated with (gray bars) or without (black bars) fenretinide. Real-time quantitative PCR results to verify ERp57 knockdown are shown in the lower graphs. The reduction in expression of ERp57 was also evident on Western blots (data not shown) and shown in independent experiments (Figure 7).

Figure 7

Apoptosis in SH-SY5Y cells (A) and A375 cells (B) after gene knockdown with control, ERdj5-2 or ERp57-3 siRNA. Apoptosis in response to control vehicle, fenretinide (3 or 10 μM), velcade (5 or 30 nM), vincristine (10 nM) or temozolomide (1 mM) (SH-SY5Y or A375 cells, respectively), or thapsigargin (1.5 μM for SH-SY5Y or 7.5 μM for A375 cells) was measured quantitatively by flow cytometry of PI-stained cells; the detection of cleaved caspase-3 on Western blots (a technique that is less sensitive than flow cytometry), relative to β-actin as a loading control, was used as additional verification of apoptosis (upper panels in A and B). Confirmation of ERp57 and ERdj5 protein knockdown was assessed by Western blotting. Bar heights indicate the mean with error bars showing the upper 95% confidence limits. Statistical comparisons were hypothesis tests (Bonferroni-corrected) from within a two-way ANOVA in which the effect of siRNA knockdown of ERdj5 or ERp57 was compared with the scrambled control for that drug treatment; ^*P<0.05; ^***P<0.0001.