Decreased heme oxygenase-1 expression distinguishes human melanomas from melanocytic nevi

Authors


Sook Jung Yun, e-mail: sjyun@chonnam.ac.kr

Dear Sir,

Heme oxygenase (HO) is a microsomal enzyme that catalyzes the first, rate-limiting step in the degradation of heme, and plays an important role in its recycling (Maines, 1997). It cleaves the α-meso carbon bridge of heme, yielding equimolar quantities of carbon monoxide, Fe2+, and biliverdin (Tenhunen et al., 1968). The enzymatic activity of HO results in decreased oxidative stress, an attenuated inflammatory response, and a lower rate of apoptosis. Three distinct mammalian HO isoforms (HO-1, HO-2, and HO-3) have been identified. They are the products of three different genes (Maines, 1988). Heme oxygenase-1, the inducible 32-kDa isoform, also known as heat shock protein 32, is highly expressed in liver and spleen, but can also be detected in many other tissues. It can be induced strongly in response to cellular stress and diverse oxidative stimuli, including its heme substrate, heat shock, ultraviolet (UV) irradiation, reactive oxygen species (ROS), nitric oxide (NO), inflammatory cytokines, prostaglandins, ethanol, heavy metals, and hypoxia (Applegate et al., 1991; Choi and Alam, 1996). Most of the known HO-1 inducers stimulate the production of ROS or lead to the depletion of glutathione levels, suggesting the involvement of HO-1 activity in cellular protection against oxidative stress (Ryter and Choi, 2002).

Heme oxygenase-1 is known to play an important role in resistance to apoptosis and in the rapid growth of several solid tumors. It is often up-regulated in tumor tissues, and its expression can be further increased in response to therapy (Jozkowicz et al., 2007). Induction of HO-1 has been recognized as one of the most promising targets for chemoprevention and chemoprotection research (Prawan et al., 2005). Few studies of HO-1 expression in melanoma have been reported. Okamoto et al. (2006) suggested that HO-1 is an important candidate gene in the pathogenesis and growth of melanomas. Was et al. (2006) found that overexpression of HO-1 increased the viability, proliferation, and angiogenic potential of melanoma cells, augmented metastasis, and decreased survival of melanoma-bearing mice.

In this study, we examined the expression of HO-1 in melanocytic nevi and melanomas by immunohistochemical staining, and in cultured normal human melanocytes and melanoma cell lines. We enrolled 38 patients, 20 with melanocytic nevi and 18 with melanomas. Patient characteristics are shown in Table S1. The mean age of the patients was 42 ± 23 yr (range 6–80). The diagnoses of the 20 melanocytic nevi were as follows: compound nevi (12; 60%), junctional nevi (4; 20%), intradermal nevus (1; 5%), and Spitz nevi (3; 15%). The histopathological types of the 18 melanomas were superficial spreading type (7; 39%), acral lentiginous type (7; 39%), lentigo maligna type (3; 17%), and metastatic melanoma (1; 6%). Immunohistochemical staining was performed using formalin-fixed, paraffin-embedded tissue specimens from 20 patients with melanocytic nevi and 18 with melanomas, using the LSAB2® System-HRP (DakoCytomation, Carpenteria, CA, USA). The intensity of HO-1 immunoreactivity was graded using the following scale: 0 = negative (no staining of melanocytic cells, but staining of the epidermis); 1 = weakly positive (staining intensity was markedly weaker than that of the epidermis); 2 = moderately positive (staining intensity was mildly weaker than that of the epidermis); and 3 = strongly positive (staining intensity was at least equal to that of the epidermis). The Mann–Whitney test (performed using SPSS version 12.0; SPSS Inc., Chicago, IL, USA) was used to analyze differences in HO-1 staining intensity between melanocytic nevi and melanomas. Additionally, HO-1 expression in melanocytic nevi and melanomas from sun-exposed and sun-protected areas was compared. Expression of HO-1 was detected in 19 of 20 melanocytic nevi (95%) and in 11 of 18 melanomas (61%). Of melanomas, one superficial spreading type, four acral lentiginous type, one lentigo maligna type, and one metastatic type did not express HO-1. Note that the intensity of HO-1 expression was markedly different between melanocytic nevi and melanomas (Figure 1). The mean intensity in melanocytic nevi was 1.82 ± 0.86. There was no difference in staining intensity among the types of melanocytic nevi. The intensity of HO-1 expression in the nearby epidermis in melanocytic nevi was 2.30 ± 0.64 (Figure S1). Only 13 of 20 melanomas expressed HO-1. Moreover, the HO-1 staining intensity was much lower in melanomas than in the melanocytic nevi, and the features were variable. Mean intensity of melanoma staining was 0.43 ± 0.45, whereas the intensity of HO-1 staining in the nearby epidermis was high, with an average of 2.2 ± 0.86 (Figure S1). Nine melanocytic nevi were located at sun-exposed sites (patient nos. 1–6 and 17–19). The intensity of HO-1 staining in melanocytic nevi in sun-exposed and sun-protected areas was 2.18 ± 0.25 and 1.52 ± 1.07, respectively. This difference was not significant (P = 0.175). In the epidermis of melanocytic nevi, the HO-1 staining intensity was 2.37 ± 0.48 at sun-exposed sites, and 2.24 ± 0.76 at sun-protected sites. There were eight melanomas at sun-exposed sites (patient nos. 21, 22, 28, 29, and 35–38). The intensity of HO-1 staining in melanomas at sun-exposed and sun-protected sites was 0.17 ± 0.24 and 0.64 ± 0.48, respectively. The heme oxygenase-1 staining intensity was lower in melanomas at sun-exposed sites (P = 0.027). In the epidermis of melanomas, the intensity of HO-1 staining was 1.91 ± 1.08 at sun-exposed sites and 2.43 ± 0.59 at sun-protected sites.

Figure 1.

 (A) Heme oxygenase-1 expression in melanocytic nevi. Compound nevi (CN, patients 1 and 2), intradermal nevus (IN, patient 17), Spitz nevus (SN, patient 18) show strong HO-1 expression. (hematoxylin and eosin, ×100; HO-1 immunohistochemical stain, original magnification, ×100). (B) Heme oxygenase-1 expression in melanomas. Superficial spreading type melanoma (SSM, patients 24, 25 and 21), lentigo maligna type melanoma (LMM, patient 35), acral lentiginous type melanoma (ALM, patient 29), and metastatic melanoma (MM, patient 38) show weak (patients 24, 25, and 21) or no (patients 35, 29, and 38) staining intensity (hematoxylin and eosin (H&E), patients 24 and 38, ×40; patients 25, 21, 35 and 29, ×100; HO-1 immunohistochemical stain, original magnification, patients 24 and 25, ×100; patients 21, 35, 29 and 38, ×200).

Normal human melanocytes were cultured primarily from neonatal foreskin as previously described (Eisinger et al., 1979). The A375 (ATCC no. CRL-1619) and G361 (ATCC no. CRL-1424) melanoma cell lines were purchased from ATCC (Manassas, VA, USA), and A375SM (KCLB no. 80004) and A375P (KCLB no. 80003) from the Korean Cell Line Bank (Seoul, Korea). After immunocytochemical staining of HO-1, images were visualized under a confocal microscopy laser scanning microscope (LSM 510; Carl Zeiss, Jena, Germany) equipped with a 20× objective, and analyzed using LSM 5 imaging browser software (Carl Zeiss, Jena, Germany). We found that normal human melanocytes displayed strong HO-1 expression in their cytoplasm. The A375P, G361, A375, and A375SM melanoma cell lines expressed HO-1, but at lower levels compared with normal human melanocytes (Figure 2A). For Western blot analysis, cultured normal human melanocytes and melanoma cell lines were lysed in RIPA buffer (1% Triton X-100, 150 mM NaCl, 0.1% SDS, 50 mM Tris-HCl (pH 7.5), 2 mM EDTA), and cell extract (30 μg) was analyzed. The result revealed that A375P, G361, A375, and A375SM melanoma cell lines exhibited a markedly decreased expression of HO-1, compared with normal human melanocytes, which yielded a positive band corresponding to HO-1 at 32 kDa (Figure 2B). Heme oxygenase-1 mRNA expression, as measured by RT-PCR, was also decreased in the same four melanoma cell lines, compared with the expression in normal human melanocytes (Figure 2C).

Figure 2.

 (A) Heme oxygenase-1 expression in normal human melanocytes and melanoma cell lines, assessed by laser scanning confocal microscopy performed using an Alexa Fluor 488-conjugated fluorescence. Normal human melanocytes show strong cytoplasmic HO-1 staining (green) and the melanoma cell lines A375P, G361, A375, and A375SM show much weaker HO-1 expression (cell nuclei are stained blue with DAPI, original magnification, ×400). (B) Western blot analysis of HO-1 expression in normal human melanocytes and melanoma cell lines. The melanoma cell lines A375P, G361, A375, and A375SM show markedly decreased expression of HO-1 protein, compared with normal human melanocytes, which yields a positive band corresponding to HO-1 at 32 kDa. (C) RT-PCR analysis of HO-1 mRNA expression in normal human melanocytes and melanoma cell lines. The melanoma cell lines A375P, G361, A375, and A375SM display decreased expression of HO-1 mRNA, compared with normal human melanocytes, which produces a strong positive band for HO-1 (554 bp).

In these experiments we demonstrated that HO-1 expression was markedly decreased in melanomas, irrespective of type and site, whereas it was strong in melanocytic nevi. In addition, we found strong HO-1 protein and mRNA expression in cultured normal human melanocytes, but only weak expression in melanoma cell lines. To our knowledge, this is the first report to compare HO-1 expression in melanocytic nevi and melanomas.

Normal human skin is under continuous oxidative stress from UV. Keratinocytes are known to play a major role in protecting against this, by up-regulating HO-1 expression (Gruber et al., 2007). Melanocytes, in contrast, have not been studied extensively. In our study, HO-1 expression was increased diffusely in the epidermis of melanocytic nevi and melanomas. In human melanocytes, HO-1 expression is induced by exposure to UV and may play a role in protection against photooxidative stress (Marrot et al., 2005). However, in melanomas, redox regulation is lost during melanomagenesis, and inappropriate redox-sensitive transcription factor activation results in enhancement of the anti-apoptotic phenotype in transformed cells (Meyskens et al., 2001). When the antioxidant N-acetylcysteine is administered prophylactically before acute UV exposure in patients with nevi, it prevents pro-oncogenic oxidative stress in nevi, although the ability of antioxidants to reduce long-term melanoma risk remains to be demonstrated (Goodson et al., 2009).

To our knowledge, this is the first reported study to compare HO-1 expression in melanocytic nevi and melanomas. Heme oxygenase-1 expression was increased in the epidermis of melanocytic nevi and melanomas, with higher expression in the upper spinous and granular layers. In melanocytic nevi, there was increased expression of HO-1 in cells located in the epidermis and dermis. There was no difference in melanocytic nevi between sun-exposed and sun-protected areas. Also, there was no difference between Spitz nevi and other common melanocytic nevi. Unfortunately, there were no cases of dysplastic nevi in our study. In contrast, in melanomas there was decreased or weak expression of HO-1 in melanoma cells. Moreover, melanomas in sun-exposed areas had significantly lower HO-1 expression, which may be explained by a more marked reduction in the protective effects of HO-1 in the face of severe photooxidative stress. However, there were no statistically significant differences between the different types of melanoma. In our study, we also detected differences in the HO-1 expression in cultured normal human melanocytes and melanoma cell lines by Western blotting, RT-PCR analysis, and confocal microscopy. While normal human melanocytes displayed strong HO-1 expression, all four melanoma cell lines analyzed exhibited lower HO-1 expression, mirroring the finding that HO-1 was weakly or not expressed in human melanoma tissues. These results suggest that melanocytic nevi preserve the ability to produce HO-1 as an antioxidant, whereas melanomas lose HO-1 expression as a result of loss of protection from oxidative stress. Further studies are warranted to determine the exact role of HO-1 in melanomas.

Acknowledgements

This work was supported by research grants from the Chonnam National University Research Institute of Medical Sciences, and the Chonnam National University Hospital Research Institute of Clinical Medicine (CRI09040-1).

Ancillary