High expression of heme oxygenase-1 is associated with tumor invasiveness and poor clinical outcome in non-small cell lung cancer patients

Abstract

Background

Heme oxygenase-1 (HO-1), a rate-limiting enzyme in heme catabolism, is known to play a role in the protection of cells against oxidative stress, inflammation, anomalous proliferation and apoptosis. As yet, the role of HO-1 expression in non-small cell lung cancer (NSCLC) development and metastasis remains unclear and insufficient data are available regarding its impact on the prognosis of NSCLC patients.

Methods

Seventy NSCLC patients who underwent surgical resection were included in this HO-1 expression study and, concomitantly, clinical parameters were collected. Two lung adenocarcinoma cell lines (A549 and H441) were used to assess both invasive and migratory parameters in vitro.

Results

NSCLC patients with a high HO-1 expression ratio (tumor tissue/normal tissue) (> 1) exhibited a significantly poorer prognosis and a higher metastatic rate compared to those with a low HO-1 expression ratio (p < 0.05). The invasive and migratory abilities of A549 and H441 cells significantly increased after exogenous HO-1 over-expression and significantly decreased after siRNA-mediated HO-1 expression silencing. HO-1 up- and down-regulation also positively correlated with the expression of metastasis-associated proteins EGFR, CD147 and MMP-9. In addition, we found that HO-1 expression can be inhibited by PI3K and AKT inhibitors, but not by MAPK inhibitors.

Conclusions

HO-1 is a poor prognostic NSCLC predictor and its over-expression may increase the metastatic potential of NSCLC. Based on our findings and those of others, HO-1 may be considered as a novel NSCLC therapeutic target.

Introduction

Heme oxygenase-1 (HO-1) is a member of the heat shock protein family. Its expression is triggered by diverse stress-inducing stimuli including hypoxia, heavy metals, UV radiation and reactive oxygen species [1–3]. HO-1 can catalyze heme to carbon monoxide and bilirubin with a concurrent release of iron. The physiologic functions of HO-1 entail anti-oxidation, anti-inflammation, anti-proliferation and anti-apoptosis effects [4, 5]. In addition, several studies have shown that HO-1 expression is increased in cancer tissues, including oral squamous cell carcinoma [6], colorectal cancer [2, 7, 8], prostate cancer [9, 10] and lung cancer [11–14].

The processes of tumor invasion and metastasis are complex and involve multiple steps. Tumor cells initially adhere to the extracellular matrix (ECM), degrade the surrounding ECM, migrate and proliferate before, finally, stimulating angiogenesis [15]. The matrix metalloproteinase (MMP) family, the cluster of differentiation 147 (CD147) and the epidermal growth factor receptor (EGFR) play important roles in tumor invasiveness, metastasis and proliferation [16–20]. A growing body of evidence indicates that HO-1 may also affect the growth and metastasis of tumors through regulating these metastasis-associated proteins [15, 21]. However, the exact role of HO-1 in tumor metastasis is, as yet, unclear. It has been found that HO-1 over-expression may increase melanoma cell viability, proliferation and angiogenic potential [22]. In contrast, silencing of HO-1 has been found to decrease cellular invasion properties of prostate cancer [10]. The role of HO-1 in tumor prognosis is generally inconsistent, and mostly depends on the cell of origin, i.e., patients with colon cancer [2] or oral carcinoma [23] and a high HO-1 expression appear to exhibit a relatively good long-term survival rate. On the contrary, patients with prostate cancer [9], melanoma [22] or pancreatic cancer [24] and a high HO-1 expression appear to exhibit a poor prognoses. To date, the role of HO-1 expression in NSCLC prognosis has not been clarified yet [11, 25]. The induction of HO-1 expression can be brought about by intracellular signaling cascades that encompass upstream transcription factors such as NF-κB [26], AP-1/2 [25] and Nrf2 [11]. In addition, phosphatidylinositol 3-kinase (PI3K)/AKT and mitogen-activated protein kinase (MAPK) signaling pathways have been found to play major roles in controlling HO-1 expression in different tumor cell systems [27, 28]. The exact mechanism regulating HO-1 expression in NSCLC has not yet been established.

Lung cancer is the leading cause of cancer mortality in Taiwan. Fast progression from benign and limited growth to invasive and metastatic growth is the major cause of poor clinical outcome in lung cancer patients. The objective of this study was to examine the effect of HO-1 expression on NSCLC growth, invasion, migration, metastasis and clinical prognosis, and to elucidate the mechanism regulating HO-1 expression.

Methods

Tumor sample collection

NSCLC and corresponding normal tissues were collected from 70 patients who underwent surgical resection at the Division of Thoracic Surgery, Department of Surgery, Kaohsiung Medical University Hospital, from 2004 to 2008. Complete staging procedures including chest radiography, bronchoscopy, brain and thoracic computed tomography, sonography, and bone scintigraphy were carried out to precisely determine the characteristics of the primary tumor (T), nodal involvement (N), and metastasis (M) according to the TNM International Staging System for Lung Cancer [29]. All patients were subjected to follow-up until March 2011, including their demographic details and survival data. Informed consent was obtained from all participants, and the study was approved by the Institutional Review Board for research (KMUH-IRB-960446 and KMUH-IRB-980542).

RNA extraction and real-time PCR

Total RNA was isolated from lung cancer tissues and adjacent normal lung tissues of the NSCLC patients, and, subsequently, analyzed by real-time PCR. The following primers were designed using Primer Express software (RealQuant, Roche) based on published sequences: human HO-1 sense primer: 5′-TTC TTC ACC TTC CCC AAC TA-3′; HO-1 antisense primer 5′-GCA TAA AGC CCT ACA GCA AC-3′. Human GAPDH sense primer: 5′-AGC CAC ATC GCT CAG ACA-3′; GAPDH antisence primer 5′-GCC CAA TAC GAC CAA ATC C-3′. PCR conditions included an initial denaturation at 94 °C for 180 s, followed by 40 cycles at 95 °C for 30 s, 60 °C for 25 s, 72 °C for 30 s, and 1 cycle at 72 °C for 7 min. Fluorescence data were acquired after the final extension step. A melt analysis was conducted for all products to determine the specificity of the amplification. In addition, PCR products were run on 1 % agarose gels to confirm their correct sizes. Relative expression levels were calculated as ratio between expression in the tumor and expression in the normal adjacent tissue (high expression: tumor lesion/normal tissue >1; low expression: tumor lesion/normal tissue <1) [30, 31].

Immunohistochemical staining

Tumor specimens were dissected from human lung tissue, fixed in paraformaldehyde, embedded in paraffin and sectioned at 5 μm thicknesses. The paraffin sections were de-paraffinized with xylene and stained with an anti-human HO-1 antibody (1:2000x) (Santa Cruz) for 1 h. After washing in PBS, the sections were incubated for 1 h at room temperature with a horseradish peroxidase-conjugated secondary antibody. For color reactions, diaminobenzidine (DAB) was used in conjunction with hematoxylin counter staining.

Gelatin zymography assay

MMP-2 and MMP-9 activities were determined by gelatin zymography as described previously [32, 33].

Western blot analysis

Samples were assayed according to our previously described method [33] using mouse anti-human HO-1, EGFR, CD147 and MMP-9, and α-tubulin (Sigma-Aldrich) as internal control protein.

Culture of NSCLC cells

Lung adenocarcinoma A549 and H441 cells (ATCC; CCL-185/HTB-174) were cultured according to published methods [34].

HO-1 over-expression and silencing in A549 and H441 cells

For exogenous HO-1 expression, we used the constructs and methods reported by Shiraishia et al. [35]. Briefly, A549 and H441 cells were grown till 80 % confluence and transfected with a HO-1 expression construct using Lipofectin (Invitrogen). To generate various exogenously HO-1 expressing cell lines, cells were diluted and seeded 24 h after transfection and, subsequently, maintained in F12K medium (serum free) until 48 h. Cells transfected with an empty pcDNA3 vector served as a negative control. Real-time PCR or Western blotting were used to assess HO-1 expression in the transfected cells.

A small interfering RNA (siRNA), a specific double-stranded 21-nucleotide RNA sequence homologous to the target mRNA, was used to silence HO-1 expression. This siRNA and a negative control siRNA were designed and synthesized using the computer software of Ambion (Austin, TX) and the Silencer^TM siRNA construction kit according to the manufacturer’s instructions. HO-1 mRNA inhibition and protein expression were assessed by real-time PCR and immunoblot analysis, respectively, following transfection of A549 and H441 cells with HO-1 siRNA. Specifically, cells were transfected with 20 nM siRNA using 8 μl of siPORT Amine (Ambion Inc.) in a total volume of 0.5 ml growth medium. After incubation at 37 °C, 5 % CO₂ for 5 h, 1.5 ml of normal growth medium was added and cells were incubated for 48 h.

In vitro invasion and migration analyses

Cellular invasion was quantified using a modified Matrigel Boyden chamber assay. The BD BioCoat Matrigel invasion chamber (BD Biosciences, Bedford, MA), with 12 mm diameter and 8 μm pore size corning caster, was used according to the manufacturer’s instructions. A549 and H441 cells (4 × 10⁴) suspended in serum-free medium were seeded onto Matrigel (10 μg/cm²)-coated filters [36]. After a 24 h incubation period, the membrane was briefly washed with PBS and fixed with 4 % paraformaldehyde. The degree of A549 and H441 cell invasion was evaluated by microscopy under high-power field (×100) using 6 randomly selected clones. The migratory ability of the cells was assayed in a scratch wound assay as previously described [37], and photographed to analyze the denuded areas.

Nuclear extract preparation and electrophoretic mobility shift assay (EMSA)

After being washed in PBS, cells were scraped off the plates in 0.6 ml ice-cold buffer A [10 mM N-2-hydroxyethyl) piperazine-N’-(2-ethenesulfonic acid) (HEPES), pH 7.9, 10 mM KCl, 1 mM dithiothreitol (DTT), 1 mM PMSF, 1.5 mM MgCl₂, and 2 μg/ml each of aprotinin, pepstatin and leupeptin]. After centrifugation at 300 g for 10 min at 4 °C, cells were resuspended in buffer B (80 μl of 0.1 % Triton X-100 in buffer A), left on ice for 10 min, then centrifuged at 12,000 g for 10 min at 4 °C. The nuclear pellets were re-suspended in 70 μl ice-cold buffer C (20 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 0.42 M NaCl, 1 mM DTT, 0.2 mM EDTA, 1 mM PMSF, 25 % glycerol, and 2 μg/ml each of aprotinin, pepstatin, and leupeptin), then incubated for 30 min at 4 °C, followed by centrifugation at 15,000 g for 30 min at 4 °C. Protein concentrations were determined using the Bio-Rad method. The NF-κB probe used in the gel shift assay was a 31-mer synthetic double-stranded oligonucleotide (5′-ACA AGG GAC TTT CCG CTG GGG ACT TTC CAG G-3′; 3′-TGT TCC CTG AAA GGC GAC CCC TGA AAG GTC C-5′) containing a direct repeat of the κB site. In the electrophoretic mobility shift assay, the digoxigenin (Dig) gel shift kit for 3′-end labeling of oligonucleotides (Roche, Indianapolis, IN) was used for nuclear protein-DNA binding. To confirm the presence of bands specific for NF-κB, unlabelled oligonucleotide controls and a p65-specific antibody were added to the binding mixture for supershift assays.

Statistical analyses

Statistical differences in HO-1 expression ratios and clinical parameters were assessed using a Student’s t-test or a one-way ANOVA test. Survival curves were established using the Kaplan-Meier method. The multivariable-adjusted risk ratios were computed from Cox regression with the additional variables of gender (man versus woman), age (years), metastasis, tumor, lymph node and smoking. Data were analyzed using JMP software (SAS, JMP, Version 8.0, Cary, NC) and presented as mean ± standard deviation. The chi-square test was used for statistical analysis. P < 0.05 was considered to be statistically significant.

Results

Demographic characteristics and HO-1 expression in NSCLC patients

Of 70 NSCLC patients included in this study, 54 (77 %) patients were histologically identified with adenocarcinoma and 16 (23 %) patients with squamous cell carcinoma. The average age at diagnosis was 60.4 ± 9.6 years (range 27–77 years). The TNM staging and HO-1 expression data of the NSCLC patients are summarized in Table 1. NSCLC patients with metastasis and advanced stage disease (stage III–IV) exhibited a higher HO-1 expression ratio as compared to non-metastasis and early stage patients. These differences were found to be statistically significant (p = 0.03 and 0.04, respectively). Although there were no statistically significant differences between T (tumor), N (lymph node) stages, the HO-1 expression ratios increased as stages advanced (data not shown). Using the Kaplan-Meier method, we also found that the low HO-1 expression ratio group of patients showed a notably better survival as compared to the high HO-1 expression ratio group (Fig. 1, p < 0.05). The multivariate-adjusted risk ratios were computed from Cox regression with the additional variables of gender (man versus woman), age (years), metastasis, tumor, lymph node involvement and smoking (Table 2). By doing so, we found that the high HO-1 expression group exhibitd a 2.18 times higher mortality risk (p = 0.03). Other factors were not significantly associated with prognosis.

Table 1.

Relationship between HO-1 expression ratios and clinical parameters, in non-small cell lung cancer patients

Variable	Number	Mean ± SD	P
Stage			P = 0.04
stage (I–II)	34	0.81 ± 0.43
Stage (III–IV)	36	1.07 ± 0.59
Tumor(T)			P = 0.95
1	12	0.88 ± 0.40
2	43	0.94 ± 0.60
3	5	1.02 ± 0.47
4	10	1.00 ± 0.49
Node (N)			P = 0.95
0	37	0.94 ± 0.63
1	15	0.98 ± 0.39
2	13	0.87 ± 0.48
3	5	1.00 ± 0.27
Metastasis (M)			P = 0.03
0	46	0.84 ± 0.43
1	24	1.13 ± 0.66
Smoking			P = 0.28
No	34	1.01 ± 0.52
Yes	36	0.88 ± 0.54
Pathology			P = 0.11
SQ	16	0.76 ± 0.38
AD	54	1.01 ± 0.56
Sex			P = 0.88
Male	45	0.94 ± 0.60
Female	25	0.98 ± 0.37

SQ squamous cell carcinoma; AD adenocarcinoma

Fig. 1.

Survival curves of non-small cell lung cancer patients in high and low HO-1 expression groups using the Kaplan-Meier method

Table 2.

Multivariate Cox regression analysis of mortality

Term		Risk Ratio	Lower 95 %	Upper 95 %	p-value
HO-1 expression ratio >1		2.18	1.07	4.50	0.03*
Age		1.40	0.26	8.08	0.69
Metastasis		1.10	0.48	2.44	0.82
Node	1 vs. 0	1.54	0.64	3.58	0.32
	2 vs. 0	1.09	0.34	2.90	0.87
	3 vs. 0	3.54	0.81	13.31	0.09
Tumor	2 vs. 1	1.93	0.78	5.82	0.16
	3 vs. 1	1.32	0.27	6.36	0.73
	4 vs. 1	0.97	0.25	3.93	0.96
Smoking		1.65	0.58	5.15	0.36
Gender	M vs. F	1.67	0.49	5.71	0.41

* p < 0.05; M Male; F Female

MMP-2 and MMP -9 expression in NSCLC patients

The MMP-9 and HO-1 mRNA expression levels were positively correlated (p < 0.0001), whereas the MMP-2 expression levels showed no such correlation (Fig. 2a). Immunohistochemical staining of lung cancer tissues also revealed an increase in MMP-9 and HO-1 expression as compared to normal tissues, whereas such a correlation was, again, not seen for MMP-2 (Fig. 2b). Based on Western blot analysis, HO-1 expression was again found to be increased in lung cancer tissues (Fig. 2c), consistent with the immunohistochemical staining. Also after gelatin zymography, the activity of MMP-9 was increased in lung cancer tissues compared to normal tissues (Fig. 2d), whereas the activity of MMP-2 did not show such an increase.

Fig. 2.

Relationship between HO-1 expression and MMP-2/MMP-9 expression in NSCLC patients. (a) Correlations between HO-1 expression and MMP-2/MMP-9 expression were analyzed. The Pearson correlation test was used to calculate P values. (b) Lung samples (lung cancer and corresponding normal adjacent lung tissues) were analyzed using antibodies against HO-1, MMP-9 and MMP-2, in conjunction with immunohistochemical staining (DAB staining and hematoxylin counterstaining). For negative controls, antibodies were replaced by control IgG. The results show that the expression of HO-1 and MMP-9 are increased in lung cancer tissues compared to normal lung tissues, and that the expression of MMP-2 is not (scale bar = 100 μm). (c) HO-1 expression was analyzed in lung cancer and normal tissues by Western blotting. HO-1 expression is elevated in lung cancer samples as compared to matched non-cancer tissues. Representative data from three different patients are shown (T = tumor; N = normal). (d) Gelatin zymography analysis of MMP-2 and MMP-9 activities from three different NSCLC patients. The activity of MMP-9 is increased in lung cancer tissues whereas MMP-2 is not. * P<0.05 versus normal lung tissue

HO-1 affects metastasis associated protein expression

A549 and H441 cells were used to evaluate the effects of HO-1 expression on NSCLC cell migratory and metastatic abilities in vitro. First, A549 and H441 cells were transfected with a HO-1 expression vector to induce exogenous HO-1 expression. The resulting HO-1 over-expression was significant (20x) and confirmed by real-time PCR (Fig. 3a–b). Furthermore, A549 and H441 cells were transfected with a HO-1 targeting siRNA vector to silence HO-1 expression. More than 50 % gene silencing was achieved. As a negative control a vector encoding a non-targeting siRNA (NC siRNA) was used.

Fig. 3.

Exogenous induction and silencing of HO-1 expression in A549 and H441 cells (a–d). HO-1 mRNA expression in A549 and H441 cells (a–b) is increased significantly after exogenous HO-1 expression, as revealed by real time PCR. HO-1 expression is decreased significantly after siRNA-mediated HO-1 silencing. NC-siRNA serves as a negative control. Representative examples of three independent experiments are shown. EGFR, CD147 and MMP-9 expression is enhanced after exogenous HO-1 expression in A549 and H441 cells, as revealed by Western blotting, and decreased after HO-1 silencing (c–d). *P < 0.05 versus control pcDNA or NC siRNA. Columns, mean values; bars, SEM. Assays were carried out in triplicate

EGFR and CD147 are genes that promote tumor metastatic and invasive abilities [11, 16, 19]. In A549 and H441 cells, the expression of MMP-9, CD147 and EGFR increased significantly together with exogenous HO-1 over-expression. Similarly, after siRNA mediated down-regulation of HO-1 expression, the expression of MMP-9, CD147 and EGFR decreased, but only that of MMP-9 was significant (Fig. 3c–d). The expression of MMP-2 was not affected by either HO-1 over-expression or HO-1 silencing (data not shown).

HO-1 affects A549 and H441 cell migration and invasion

Cellular invasion and migration characteristics were analyzed using matrigel-coated Boyden chambers and wound scratch assays, respectively. As compared to the control group, cell migration and invasiveness were augmented by approximately 50 % and 30 %, respectively, (p < 0.05) after exogenous HO-1 expression during 48 h in A549 and H441 cells (Fig. 4a–b). The cellular invasive/migratory abilities exhibited a decrease of approximately 30 % and 24 %, respectively, (p < 0.05) after siRNA mediated HO-1 silencing during 48 h.

Fig. 4.

Effect of HO-1 expression on invasiveness/migration of A549 and H441 cells. (a) Cellular migration was determined using a scratch wound assay. The migratory ability of A549 and H441 cells is augmented with increasing exogenous HO-1 expression as compared to the control group. After HO-1 silencing, the migration is decreased. NC-siRNA serves as a negative control. (b) Invasiveness was determined by Matrigel-coated Boyden chamber assays. Exogenous HO-1 expression in A549 and H441 cells increases cellular invasive capacities significantly (p < 0.05), whereas silencing of HO-1 expression results in a significant reduction in cellular invasiveness (p < 0.05). Results of three independent experiments are shown. *P < 0.05 versus control pcDNA or NC siRNA. Assays were carried out in triplicate

The regulatory pathway of HO-1 expression in A549 cells

Finally, we assayed the effect of HO-1 expression on NF-κB. By doing so, we found that the HO-1 inducer CoPPIX increased the nuclear translocation of p65/p50 in conjunction with an up-regulation of NF-κB expression (Fig. 5a), whereas siRNA mediated HO-1 silencing led to a decrease in NF-κB expression. Intracellular signaling pathways, such as the MAPK and PI3K/AKT pathways, have previously been shown to play important roles in tumor invasiveness [38–40]. After exogenous HO-1 expression for 48 h, we treated A549 cells with several MAPK and PI3K/AKT inhibitors for 1 h. We observed that the expression of HO-1 was significantly decreased by selective inhibitors of PI3K (LY294002) and AKT (API-59) (Fig. 5b), but not by various MAPK inhibitors (PD98059, SP600125, and SB203580 for ERK, JNK and p38 MAPK, respectively). These results suggest that HO-1 expression regulation is mediated by the PI3K/AKT signaling pathway. However, the exact relationship between HO-1 expression and these intracellular signaling pathways still remains to be established.

Fig. 5.

Regulatory pathway of HO-1 expression in A549 cells (a) The expression of nuclear p65/p50 is up-regulated by CoPPIX treatment and down-regulated after HO-1 silencing. (b) A549 cells were treated with inhibitors of PI3K (LY294002) (10 mM), AKT (API-59) (3 mM), MAPK inhibitors (PD98059 (10 mM), SB203580 (10 mM), and SP600125 (10 mM) for ERK, p38 MAPK, and JNK, respectively) after exogenous HO-1 expression for 48 h. HO-1 expression is decreased using PI3K and AKT inhibitors. Cobalt protoporphyrin IX (CoPPIX)

Discussion

In the present study, we found that NSCLC patients with a high HO-1 expression ratio exhibit a significantly shorter survival rate as compared to those with a low HO-1 expression ratio. NSCLC patients presenting with metastasis and an advanced stage also had a significantly higher HO-1 expression ratio. We found that HO-1 expression serves as a negative prognostic factor after multivariate Cox regression analysis. Through in vitro studies, we found that the migratory and invasive abilities of lung adenocarcinoma cells (A549 and H441) increased after exogenous HO-1 over-expression, while siRNA mediated HO-1 silencing showed the opposite. The metastasis-associated proteins EGFR, CD147 and MMP-9 were all augmented concomitantly with HO-1 over-expression. These results were consistent with our clinical findings and indicate that HO-1 may serve as a potential therapeutic target in NSCLC.

Heme oxygenases (HOs) are microsomal enzymes that catalyze the oxidative cleavage of heme to biliverdin, free heme iron, and carbon monoxide [41]. HO-1, one of the HO isoforms, is up-regulated in stressful conditions in order to protect cells against oxidative injury [21]. The primary roles of HO-1 are to regulate cellular homeostasis, to modulate inflammatory responses and to promote cell survival [4, 41]. Interestingly, the expression of HO-1 has been found to be high in many tumor types as compared to its surrounding normal tissues. This was shown in hepatomas [19, 42], prostate cancers [9, 10], brain tumors [43], pancreatic tumors [24], oral squamous carcinomas [6, 23] and lung cancers [11, 44–46]. Here we show, by immunohistochemical staining and Western blotting, that HO-1 expression is also increased in NSCLC.

The effects of HO-1 expression on tumor metastasis and clinical prognosis are different in various malignancies. HO-1 expression can, for example, promote lymph node metastasis in oral squamous cell carcinoma patients [6]. On the contrary, in colorectal and tongue cancer, the expression of HO-1 can decrease metastatic potential resulting in a better long-term survival [2, 47]. Others found that HO-1 expression does not seem to be associated with lung tumor progression and/or prognosis [25, 48]. In our study, advanced stage and metastasis status were found to be positively correlated with HO-1 expression. Based on these observations, we hypothesize that HO-1 can promote lung cancer progression and metastasis. We also found that cellular invasive and migratory abilities of A549 and H441 cells can be increased in vitro through exogenous over-expression of HO-1. These results are in accordance with those obtained in pancreatic and prostate cancer cells. Pancreatic tumor cells engineered to over-express HO-1 led to an increase in the occurrence of metastasis, whereas inhibition of HO-1 expression completely deterred this potential [24]. Concordantly, silencing of HO-1 expression in prostate cancer cells reduced invasion in vitro and metastasis formation in vivo [9, 10].

The processes underlying invasion and metastasis of tumors are complex, involving multiple steps affecting cell adherence, degradation of the surrounding matrix, migration, proliferation and angiogenesis [49]. The exact role of HO-1 in enhancing metastatic abilities has not been clarified yet. CD147, a transmembrane protein, is highly expressed in various cancers, including malignant melanoma and hepatocellular carcinoma, and plays an important role in tumor invasion, metastasis and growth [18–20]. The epidermal growth factor receptor (EGFR) is a transmembrane glycoprotein that can activate signal transduction pathways involved in regulating cellular proliferation, differentiation and survival. EGFR is highly expressed in a variety of tumor cell lines and has been associated with tumor metastasis and a poor prognosis [16, 17]. Proteins of the MMP family are involved in breakdown of the ECM, which is an important part of the metastatic process. In our study, we have confirmed that expression levels of EGFR, CD147 and MMP-9 are correlated to HO-1 expression levels in vitro. HO-1 appears to regulate the expression of MMP-9, which further enhances the metastatic potential. The exact role of HO-1 in the regulation of EGFR and CD147 expression, however, requires further investigation.

In addition, we found that HO-1 may regulate nuclear NF-kB expression. HO-1 expression regulation depends on the cell type, its microenvironment, the intensity and duration of exposure to stimuli, and species [50]. In MCF-7 breast cancer cells, HO-1 can be induced by cadmium chloride through the p38 MAPK pathway [51], whereas cyclopentenone prostaglandins induce HO-1 expression through the PI3K and ERK pathways [52]. In the present study, we found that activation of PI3K/AKT, rather than activation of MAPKs, affects HO-1 expression in lung adenocarcinoma A549 and H441 cells. It has been shown by others that through inhibition of the Nrf2-driven HO-1 pathway, the chemosensitivity of cisplatin may be augmented in A549 and H441 cells [11].

In conclusion, HO-1 over-expression can increase NSCLC’s invasive and metastatic ability in vitro and in vivo. The regulation of HO-1 expression is most likely mediated by PI3K/AKT, and not MAPK, pathways. These findings provide new insights into the molecular mechanisms underlying the progression and clinical prognosis of NSCLC. Pharmacologic inhibition of HO-1 expression may become a powerful therapeutic strategy to prevent the progression of lung cancer or as a sensitizing agent, in combination with chemotherapy and radiotherapy, to potentiate its anticancer effects.