Pro‐apoptotic effect of haem oxygenase‐1 in human colorectal carcinoma cells via endoplasmic reticular stress

Abstract Several biological effects of haem oxygenase (HO)‐1, including anti‐inflammatory, antiapoptotic and antioxidative properties were reported; however, the role of HO‐1 in apoptosis is still unclear. In the presence of stimulation by cobalt protoporphyrin (CoPP), an HO‐1 inducer, apoptotic characteristics were observed, including DNA laddering, hypodiploid cells, and cleavages of caspase (Casp)‐3 and poly(ADP) ribose polymerase (PARP) proteins in human colon carcinoma COLO205, HCT‐15, LOVO and HT‐29 cells in serum‐free (SF) conditions with increased HO‐1, but not heat shock protein 70 (HSP70) or HSP90. The addition of 10% foetal bovine serum (FBS) or 1% bovine serum albumin accordingly inhibited CoPP‐induced apoptosis and HO‐1 protein expression in human colon cancer cells. CoPP‐induced apoptosis of colon cancer cells was prevented by the addition of the pan‐caspase inhibitor, Z‐VAD‐FMK (VAD), and the Casp‐3 inhibitor, Z‐DEVD‐FMK (DEVD). N‐Acetyl cysteine inhibited reactive oxygen species‐generated H2O2‐induced cell death with reduced intracellular peroxide production, but did not affect CoPP‐induced apoptosis in human colorectal carcinoma (CRC) cells. Two CoPP analogs, ferric protoporphyrin and tin protoporphyrin, did not affect the viability of human CRC cells or HO‐1 expression by those cells, and knockdown of HO‐1 protein expression by HO‐1 small interfering (si)RNA reversed the cytotoxic effect elicited by CoPP. Furthermore, the carbon monoxide (CO) donor, CORM, but not FeSO4 or biliverdin, induced DNA ladders, and cleavage of Casp‐3 and PARP proteins in human CRC cells. Increased phosphorylated levels of the endoplasmic reticular (ER) stress proteins, protein kinase R‐like ER kinase (PERK), and eukaryotic initiation factor 2α (eIF2α) by CORM and CoPP were identified, and the addition of the PERK inhibitor, GSK2606414, inhibited CORM‐ and CoPP‐induced apoptosis. Increased GRP78 level and formation of the HO‐1/GRP78 complex were detected in CORM‐ and CoPP‐treated human CRC cells. A pro‐apoptotic role of HO‐1 against the viability of human CRC cells via induction of CO and ER stress was firstly demonstrated herein.


| INTRODUC TI ON
Reactive oxygen species (ROS) are major cellular oxidants generated as byproducts of oxygen metabolism. Under some circumstances, ROS generation is greatly provoked by extracellular insults such as ionizing radiation, UV light, xenobiotics and pathogens, leading to an imbalance in the intracellular reduction-oxidation status. Excessive levels of ROS can induce oxidative damage to DNA leading to gene mutations and carcinogenesis. Moreover, ROS may damage cellular structures and induce lipid peroxidation, eventually inducing apoptosis of various cells. 1,2 Clinically, ROS augmentation is a useful approach for cancer treatment, and various chemotherapeutic agents, such as cisplatin, nocodazole, and taxol, were shown to exert their antitumour activities through activating ROS-dependent apoptosis in different tumour cells. 3 Haem oxygenase (HO)-1 is a phase II enzyme that responds to oxidative stress, cellular injury and diseases by metabolizing haem into biliverdin (BV)/bilirubin (BR), carbon monoxide (CO) and ferrous iron. 5 HO-1 is regarded as a survival molecule, as it exerts cytoprotection against various cells in response to stressful conditions. [6][7][8] HO-1 is widely recognized to overcome assaults by augmented oxidative stress from chemotherapeutic agents to prevent cancer cells from undergoing apoptosis and even stimulating cell proliferation.
Both protective and detrimental effects of HO-1 were also reported in different diseases, including kidney injury and neurodegeneration. 9,10 Increasing evidence has shown a dark side of HO-1, as it acts as a critical mediator in ferroptosis and as causative factor in the progression of several human diseases. 5 Elevated HO-1 levels were detected in various human malignancies, indicating its contribution to cancer cell growth, metastasis, and resistance to chemotherapy. 11,12 In contrast, augmented HO-1 expression enhanced the death of many cancer cells. 13,14 Emerging evidence suggests another dark side of HO-1 via inducing ferroptosis through iron accumulation. Although the bright and dark sides of HO-1 have been discussed in different studies, the mechanism by which HO-1 augmentation causes protective and cytotoxic activities in cancer cells is still unknown.
Colorectal cancer (CRC) is one of the leading diagnosed cancers with high mortality, and remains a significant global health problem.
Many chemotherapeutic agents, such as taxol and carboplatin, are used to treat CRC; however, there are side effects with chemotherapy that are associated with high mortality and local recurrence at least in part through ROS production. In humans, haem-iron is more bioavailable than non-haem-iron, and unabsorbed haem reaches colon epithelial cells. 15 Previous studies showed that haem is able to irritate the epithelium of the colon as indicated by mild diarrhoea. 16,17 Feeding haem resulted in significantly increased proliferation of colonic mucosa of rats. 18 This indicates the positive correlation between haem and colon carcinogenesis. HO-1 induction was shown to metabolize haem, accompanied by producing four byproducts: CO, ferric ion, BV and BR The effects of HO-1 overexpression on CRC treatment and the roles that ROS and their byproducts play in the process are still unclear.
Cobalt protoporphyrin (CoPP) is a substrate for HO and was identified as a potent HO-1 inducer. 19 Previous studies indicated that CoPP is able to increase endogenous CO generation against myocardial infarction in vivo, and decrease production of inflammatory molecules in the central nervous system. 20,21 In microglia, CoPP protected against lipopolysaccharide (LPS) interleukin-13-induced apoptosis and reduced the expression of monocyte chemoattractant protein-1 and microglia recruitment in a retinal injury mouse model.
Our previous study demonstrated that CoPP inhibited LPS-or lipoteichoic acid (LTA)-induced inducible nitric oxide (NO) synthase (iNOS) and NO production by microglia. 22 Although several biological activities of HO-1 induced by CoPP have been reported, the effect of HO-1 on the viability of human colon carcinoma cells is still undefined. In the present study, we found that CoPP preferentially reduced the viability of three poorly differentiated CRC cell lines, COLO205, HCT-15 and LOVO, but had less of an effect on the wellto-modestly differentiated colon carcinoma HT-29 cells. The roles of HO-1 protein, ROS production, CO and ER stress in CoPP-induced apoptosis of human CRC cells are investigated in the present study.

| MTT (3-(4,5,-dimethylthiazol)-2-yl-2,5diphenyltetrazolium bromide) assay
Cells were plated at a density of 5 × 10 4 cells/well in 24-well plates. At the end of treatment, the supernatant was removed, and 30 μL of the tetrazolium compound, MTT, and 270 mL of fresh DMEM were added. After incubation for 2 hours at 37°C, 200 μL of 0.1 N HCl in 2-propanol was placed in each well to dissolve the tetrazolium crystals. Finally, the absorbance at a wavelength of 600 nm was recorded using an enzyme-linked immunosorbent assay (ELISA) plate reader.

| In vitro morphology
Human CRC cells were grown at a density of 105 cells/well in 24well plates for 24 hours. Cells were treated with or without CoPP, and cells were fixed with 3.7% formaldehyde. Cell morphological changes were examined under a light microscope.
The agarose gels were run at 100 V for 45 minutes in TBE buffer, then observed and photographed under UV light.

| Detection of hypodiploid cells by CoPP
Cells were plated in 24-well plates in duplicate, then incubated for 24 hours. Media were removed, and different treatments were added to each well. Cells were treated for 12 hours, and the supernatant and cells were harvested by exposing cells to a 0.25% trypsin-EDTA solution for 10 minutes, then centrifuged, washed in PBS, and fixed in 3 mL ice-cold 100% ethanol. All samples were incubated for 30 minutes at room temperature in the dark. The cell cycle distribution and hypodiploid cells were determined using a FACSan flow cytometer (FACScan, Becton Dickinson). 23

| Statistical analysis
Values are expressed as the mean ± SD of triplicate experiments.
The significance of the difference from the respective controls for each experimental was assayed using a one-way analysis of variance (ANOVA) with a post hoc Bonferroni analysis when applicable, and P < 0.05 or P < 0.01 were considered statistically significant.

| The HO-1 inducer, CoPP, reduced the viability of human CRC cells
In order to examine the role of HO-1 in apoptosis of human CRC cells, HO-1 inducer CoPP and four human colorectal carcinoma cell lines, including COLO205, HT-29, HCT-15 and LOVO were used in the study. As illustrated in Figure 1A Figure 1C). We further examined that cytotoxicity of CoPP in normal cells, and two preferential normal cells including WI-38 and normal human lung fibroblast (NHLF) were used. Data of MTT assay showed that human colorectal carcinoma cells COLO205 and HCT-15 expressed high sensitivity to CoPP-induced cytotoxicity, compared with WI-38 and NHFL cells. ( Figure S2).

| CoPP's reduction in cell viability via apoptosis was associated with induction of HO-1 protein expression in human CRC cells
Data of the DNA integrity assay showed that CoPP induction of DNA ladders in human colorectal carcinoma cells and the DNA ladder intensity were higher in COLO205, HCT-15 and LOVO cells than (C), The percentage of hypodiploid cells (sub-G 1 peak) in CoPP-treated human colorectal carcinoma cells was examined by a flow cytometric analysis using propidium iodide (PI) staining. Cells were treated as described in (A), and the percentage of hypodiploid cells in the indicated cells was detected by a flow cytometric analysis. Each data point was calculated from three triplicate groups, and data are shown as the mean ± SD. **P < 0.01 denotes a significant difference between indicated groups in HT-29 cells (Figure 2A). Expressions of heat shock proteins (HSPs), including HSP32 (HO-1), HSP70 and HSP90, in human CRC cells under CoPP treatment were examined by Western blotting using specific antibodies. As shown in Figure 2B, increased levels of HO-1, but not HSP70 or HSP90, protein were observed in human CRC cells. We further examined the effects of serum (FBS) and albumin (BSA) on CoPP induction of HO-1 and cell death in human CRC cells.
As illustrated in Figure 2C, the addition of FBS or BSA significantly reduced the cytotoxicity of CoPP on the viability of human CRC cells in an SF condition. Data of Western blotting indicated that CoPP-induced HO-1 protein expression and cleavage of Casp-3 and PARP proteins were suppressed by the addition of FBS or BSA in human CRC cells ( Figure 2D).

| The VAD and DEVD caspase inhibitors protected human CRC cells from CoPP-induced cell death
We further examined the role of caspase activation in apoptosis by CoPP in human CRC cells. Two caspase inhibitors, including the pancaspase inhibitor, VAD, and the specific Casp-3 inhibitor, DEVD, were used in the study. Data of Western blotting indicated that increased protein was quantitated with normalization to α-TUB by a densitometric analysis (ImageJ), and expressed as folds of control (C). Data from three independent experiments were obtained, and results are shown as the mean ± SD. **P < 0.01 denotes a significant difference between indicated groups cleavage of Casp-3 and PARP proteins were observed in COLO205, HCT-15 and LOVO cells, but not in HT-29 cells ( Figure 3A). Results of the MTT assay in Figure 1B and Western blotting in Figure 3A suggested that HT-29 cells were less sensitive to CoPP induction of apoptosis than were COLO205, HCT-15 and LOVO cells. The addition of VAD or DEVD significantly reduced CoPP-induced cell death in COLO205, HCT-15 and LOVO cells ( Figure 3B). Data of Western blotting indicated that VAD and DEVD inhibited CoPP-induced cleavage of Casp-3 and PARP proteins in these CRC cells ( Figure 3C).

| ROS did not participate in CoPP-induced apoptosis of human CRC cells
We further examined if ROS production is involved in CoPP-induced apoptosis of human CRC cells. The addition of the antioxidant NAC did not affect the decreased viability of COLO205, LOVO and HCT-15 human CRC cells by CoPP according to the MTT assay ( Figure 4A). Data of Western blotting showed that CoPP-induced cleavage of Casp-3 and PARP proteins was not altered by NAC in these three cell lines ( Figure 4B). A positive study using H 2 O 2 as an inducer of ROS-dependent apoptosis indicated that NAC significantly reduced cell death by H 2 O 2 in human carcinoma cells ( Figure 4C). Intracellular peroxide levels were measured by a flow cytometric analysis using DCFHDA as a fluorescent probe, and data ( Figure 4D) showed that intracellular peroxide levels were not altered by CoPP, but were increased by H 2 O 2 in these three cell lines, and NAC showed an ROS-scavenging effect thereby reducing peroxide levels in these cells ( Figure 4D). These results suggested that CoPP-induced apoptosis with increased HO-1 protein expression was likely not related to ROS production in human CRC cells.

| Increased HO-1 protein expression contributes to CoPP-induced cell death of COLO205, LOVO and HCT15 human CRC cells
There are three metal protoporphyrins, including FePP, zinc protoporphyrin (ZnPP) and CoPP, and they were reported to increase

| Differential effects of HO-1-catalysed products including CO, Fe +2 and BV on the viability of human CRC cells
Three products, including CO, Fe +2 and BV, were reported to be in-    The HO-1/BiP complex was examined by immunoprecipitation using an anti-HO-1 antibody followed by Western blotting using an anti-BiP antibody. The intensity of indicated proteins was quantitated with normalization to α-TUB by a densitometric analysis (ImageJ), and expressed as folds of control (C). Data from three independent experiments were obtained, and results are shown as the mean ± SD. **P < 0.01 denotes a significant difference from the control (CON) group Reactive oxygen species production was identified as a mediator leading to apoptosis, and both ROS-dependent and -independent apoptosis elicited by various agents was reported. Shen et al 26  The ER, an intracellular stress response induced by the accumulation of unfolded or misfolded proteins, and three sensors for ER stress, including protein kinase R-like ER kinase (PERK), inositol-requiring protein (IRE)-1 and ATF6 have been identified. 29 One of the major ER chaperones, BiP (GRP78), was able to bind to these sensors in non-stressful conditions. In the response to ER stress, an increased level of the BiP/GRP78 protein was dissociated from the sensors and facilitated their activation via interactions with downstream proteins. 30  showed that excessive production of ROS caused the misfolding of proteins in the ER, which leads to ER stress. 36,37 It indicated that oxidative stress-enhanced accumulation of misfolded proteins in the ER resulted in ER stress. However, Change et al (2018) reported that ROS-independent ER stress promoted the stemness properties of cancer-initiating cells. 38 In the present study, CoPP treatment did not elevate the intracellular peroxide production via flow cytometric analysis using DCHFDA as a fluorescent dye, and addition of NAC showed no inhibition on CoPP-induced HO-1 and apoptosis in human colorectal carcinoma cells. Data of co-immunoprecipitation assay showed that binding between HO-1 and BiP/GRP78 proteins were observed in CoPP-or CORM-treated human CRC cells. It suggested that CoPP-induced ER stress might be through ROS-independent manner, and formation of HO-1/GRP78 complex to block the binding of GRP78 to ER stress sensors might participate in ER stress of human CRC cells under CoPP and CORM stimulation.

| D ISCUSS I ON
Colorectal cancer is a leading cause of cancer deaths and is among the most abundant cancer type that occurs in both sexes.
It was shown that early stages of CRC are highly detectable and can be cured with surgical resection followed by standard therapy.
However, survival rates in stage III and IV CRC patients decrease, and recurrence rates range 40%-60% in the first 3 years. In the present study, four human CRC cell lines, including HT-29 (Duke's type B), HCT-15 (Duke's type C), COLO205 (Duke's type D) and LOVO, were used. It appeared that three human CRC cell lines (ie COLO205, LOVO and HCT-15) in the late stages (Duke's C and D) possessed higher sensitivities to CoPP-induced apoptosis than did early-stage HT-29 cells. Data of the flow cytometric analysis showed that both HCT-15 and COLO205 cells expressed higher endogenous peroxide levels than did HT-29 cells ( Figure S1). This suggested that HO-1 induction may potentially and preferentially treat late-stage CRC at least in part due to higher endogenous peroxide levels related to malignancy of human CRC cells. Due to CoPP-induced apoptosis and HO-1 protein expression were suppressed by BSA or FBS addition, it indicating that CoPP might not be effective for in vivo study. Therefore, development of effective HO-1 inducers deserved for verifying the anti-CRC actions of HO-1 in vivo.

| CON CLUS IONS
To conclude, our results demonstrate that induction of HO-1 contributes to apoptosis through activation of ER stress in human CRC cells. We found that poorly differentiated colorectal cells, such as COLO205, HCT-15 and LOVO, showed greater sensitivity to apoptosis elicited by the HO-1 inducer, CoPP, than did well-differentiated HT-29 cells. Data of the flow cytometric analysis revealed that endogenous peroxide levels in poorly differentiated COLO205 and HCT-15 cells were higher than those in well-differentiated HT-29 cells. This indicated that increased HO-1 protein preferentially induced apoptosis related to higher amount of endogenous ROS levels in poor/malignant human CRC. Our results may provide experimental basis for anti-CRC effect of HO-1 protein, and activation of ER stress is involved herein. Future studies using animal models are needed to assess the in vivo anti-CRC actions of HO-1.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.