Modulation of MHC class I surface expression in B16F10 melanoma cells by methylseleninic acid

ABSTRACT

The essential trace element selenium (Se) might play a role in cancer prevention as well as for cancer therapy. Its metabolite methylselenol is able to kill cells through distinct mechanisms including induction of reactive oxygen species, DNA damage and apoptosis. Since methylselenol affects innate immune responses by modulating the expression of NKG2D ligands, the aim of this study was to determine whether the methylselenol generating compound methylseleninic acid (MSA) influences the expression of the MHC class I surface antigens and growth properties thereby reverting immune escape.

Treatment of B16F10 melanoma cells expressing low basal MHC class I surface antigens with dimethyldiselenide (DMDSe) and MSA, but not with selenomethionine and selenite resulted in a dose-dependent upregulation of MHC class I cell surface antigens. This was due to a transcriptional upregulation of some major components of the antigen processing machinery (APM) and the interferon (IFN) signaling pathway and accompanied by a reduced migration of B16F10 melanoma cells in the presence of MSA. Comparative “ome”-based profilings of untreated and MSA-treated melanoma cells linked the anti-oxidative response system with MHC class I antigen processing. Since MSA treatment enhanced MHC class I surface expression also on different human tumors cell lines, MSA might affect the malignant phenotype of various tumor cells by restoring MHC class I APM component expression due to an altered redox status and by partially mimicking IFN-gamma signaling thereby providing a novel mechanism for the chemotherapeutic potential of methylselenol generating Se compounds.

Abbreviations

2D-DIGE

two-dimensional differential gel electrophoresis

APM

antigen processing machinery

CALR

calreticulin

CMV

cytomegalovirus

CTL

cytotoxic T lymphocytes

DMDSe

dimethyldiselenide

DNTB

5,5′-dithiobis 2-nitrobenzoic acid

ER

endoplasmic reticulum

EstD/estD

esterase D

G6PD/G6pd

glucose 6-phosphate dehydrogenase

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

GO

gene ontology

GSH

glutathione

Gsta4

glutathione S-transferase a4

HDAC

histone deacetylase

hnRNPA2/B1/Hnrnpa2b1

heterogeneous ribonucleoprotein family A2/B1

IFN

interferon

IRF

interferon-regulated factor

JAK-1/Jak1

janus kinase 1

LMP

low molecular weight protein

mAb

monoclonal antibody

MHC

major histocompatibility complex

MIC

MHC class I polypeptide-related sequences

MSC

methylselenocysteine

NKG2D

natural killer group 2 member D

Nqo1

NAD(P)H dehydrogenase (quinone) 1

Nrf2

nuclear factor (erythroid-derived 2) like 2

PDI

protein disulfide isomerase

PD-L1

programmed death ligand 1

PRXD1/Prdx1

peroxiredoxin 1

ROS

reactive oxygen species

Se

Selenium

SeMet

selenomethionine

STAT1/Stat1

signal transducer and activator of transcription 1

TNBC

triple negative breast cancer cells

TAP

transporter-associated with antigen processing

TPN/Tapbp

tapasin

ULBP

UL-16 binding protein

Introduction

Selenium (Se), an essential trace element, is metabolized to a variety of low molecular weight compounds and selenoproteins. While inorganic Se is mainly represented by selenate and selenite, members of the organic Se forms are selenomethionine (SeMet) and methylselenocysteine (MSC).¹ In contrast, methylseleninic acid (MSA) is a synthetic Se compound, which has been implemented in a number of in vitro and in vivo experimental studies.² The metabolism of the different Se compounds is complex and closely regulated. The two key metabolites selenide and methylselenol have been shown to be crucial for the biological function of these substances. In addition, the cellular thioredoxin (TRX) and glutaredoxin systems known to play an important role in the maintenance of the intracellular redox balance are involved in the reduction of various Se compounds.

Furthermore, Se has been suggested to be chemopreventive and to exert anti-carcinogenic effects on many solid and haematopoietic tumors such as bladder, prostate and breast cancer as well as lymphoma.^3,4 The functional activity of selenoproteins might be due to their anti-oxidative effects or to an altered expression pattern of genes involved in cancer initiation and progression.^5,6 This was further strengthened by epidemiologic studies revealing an inverse correlation of the Se status and cancer incidence as well as cancer mortality.⁷ Furthermore, Se is one of the few trace elements, for which anti-carcinogenic effects have been proven in a randomized, double blinded, placebo controlled study.⁸ Clinical interventional studies demonstrated that a supra-nutritive intake of Se has positive effects in the prevention of several tumor types, which were also confirmed by in vivo and in vitro experiments.⁹ In contrast, the SELECT study studying the chemopreventive potential of Se within a heterogeneous group of healthy man failed to reduce the risk for prostate cancer.¹⁰ These controversial data might be due to the different Se compounds used in these trials. Several lines of evidence have implicated that the Se metabolite methylselenol is the active Se compound for anti-carcinogenic effects.^11,12 Se compounds directly entering the monomethylated pool of Se compounds are more effective in reducing tumorigenesis than those preferentially metabolized to hydrogen selenide.^13,14 Methylselenol induces in combination with intracellular thiols reactive oxygen species (ROS).¹⁵ MSA, a proximal precursor of the redox active Se metabolite methylselenol, is intracellularly reduced to methylselenol by several non-enzymatic steps ¹⁴ and exhibits cytotoxic potential in different tumor entities, e.g., colon,¹⁶ melanoma¹⁷ and mammary carcinoma.⁴ Furthermore, MSA has both anti-carcinogenic and anti-metastatic potential in vivo without affecting the general behavior of animals. The combined treatment of triple negative breast cancer cells (TNBC) with MSA and paclitaxel upon xenoengraftment in mice reduced tumor growth more than paclitaxel alone,¹⁸ which might be due to induction of apoptosis and/or cell cycle arrest. The chemopreventive/chemotherapeutic effects of MSA might be associated with an inhibition of histone deacetylases (HDACs). HDACs are involved in the regulation of the gene expression pattern and often overexpressed in tumor cells. In B cell lymphoma and esophageal cancer, MSA reduces HDAC activity^19,20 by modulation of conserved cysteine residues in the catalytic center of HDACs.²¹

Different models of tumor-initiation showed that natural killer (NK) cells and the interferon (IFN) system are involved in the immunity of tumors. The NK-mediated anti-tumoral responses require the expression of the activating receptor NKG2D (natural killer group 2 member D).²² NKG2D recognizes infected or abnormal cells by binding to ligands, which are induced on stressed or transformed cells thereby leading to their elimination.²² Recently, Hagemann-Jensen and coworkers showed that the selenocompound MSA affects the immunogenicity of tumors by modulating different ligands of the NKG2D receptor, like the MHC class I polypeptide-related molecules (MICA, B) leading to enhanced NK cell responses.²³ Since NKG2D expression is not only limited to NK cells, but also found on human CD8⁺ T cells^24,25 it plays a key role in both NK and T cell-mediated tumor elimination.²⁶

The human major histocompatibility complex (MHC) class I molecules are subdivided into classical HLA-A, -B and -C (MHC class Ia) and non-classical HLA-E, -F and -G (MHC class Ib) antigens and are expressed on the cell surface in association with ß₂-microglobulin. While the expression of HLA-A, -B and -C antigens are frequently lost or downregulated on tumor cells, HLA-G and/or HLA-E antigens are often overexpressed in tumors of distinct origin, e.g., ovarian carcinoma, colon carcinoma and melanoma.^27-29 Since HLA-E can bind to the inhibitory CD94/NKG2A receptors expressed on NK cells and a subset of T cells, HLA-E expression could lead to an escape of tumor cells from immune surveillance.³⁰ Interestingly, the selenocompound selenite is able to reduce HLA-E antigen expression in different tumor cells, which was accompanied by an enhanced NK cell-mediated killing³¹ suggesting that selenite might be able to potentiate the antitumor cytotoxicity in settings of NK cell-based immunotherapies.

In contrast, for proper T-cell responses MHC class Ia surface expression is required, which is controlled by different components of the MHC class I antigen processing machinery (APM). MHC class I molecules present antigenic peptides on the surface of cells, which are predominantly generated by proteasomal degradation of intracellular proteins. The generated peptides are translocated by the transporter-associated with antigen processing (TAP) from the cytosol into the lumen of the endoplasmic reticulum (ER) and there loaded onto MHC class I molecules, which is facilitated by different chaperones, e.g., tapasin (TPN), ERp58, protein disulfide isomerase (PDI). Subsequently, the loaded MHC class Ia molecules are transported via the trans-Golgi apparatus to the cell surface and there presented to CD8⁺ cytotoxic T lymphocytes (CTL). However, total, partial loss or downregulation of MHC class I surface expression or APM components were found in many tumors of distinct origin thereby limiting the respective antitumor immune responses,^32,33 which were often associated with disease progression and reduced patients’ survival.^34-36 The underlying molecular mechanisms of MHC class I abnormalities are diverse and include either irreversible structural alterations or reversible deregulatory processes of APM components. While mutations, deletions and/or loss of heterozygosity rather occur rarely, the expression of APM components in tumors is frequently regulated by transcriptional, epigenetic or post-transcriptional control mechanisms.^32,37,38 Moreover, the expression pattern of MHC class I APM components seems to be linked to changes in the cellular or environmental metabolism. Recently, oxidative stress has been shown to alter cellular glutathione levels, which is associated with changes in the MHC class Ia dimer formation on the cell surface.³⁹ The impact of changes in the cellular redox status on the modulation of the MHC class Ia expression level is further supported by the role of the oxidizing enzyme ER oxidoreductase α (ERO-1a), which is controlled by oxidative folding.⁴⁰

Since one major goal in the field of tumor immunology is to re-establish and to maintain/bolster MHC class Ia surface expression in tumors thereby increasing their immunogenicity the identification of key molecules or substances overriding tumor intrinsic escape routes, such as the downregulation of MHC class Ia surface expression used by tumor cells will help to improve durable tumor rejection. In order to get better insights into the role of redox processes on the immunogenicity of tumor cells, B16F10 melanoma cells were treated with various concentrations of selenocompounds including MSA and their influence on the regulation of MHC class I surface expression level and/or the expression pattern of APM components were determined. The MSA-mediated altered redox metabolism caused a transcriptional upregulation of APM components leading to increased levels of MHC class I cell surface antigens. In addition, MSA mimics interferon-gamma (IFNγ) signaling by also upregulating members of IFNγ responsive genes.

Results

Effects of MSA treatment on cell viability

A prerequisite for the analyses of the immunogenic phenotype of tumors and of their proteome expression profiles in the presence of MSA is to determine the effects of MSA treatment on cell viability. Therefore, B16F10 melanoma cells were treated with different concentrations of MSA for 24 and 48 h, respectively. While lactate dehydrogenase (LDH) activity in the supernatants of B16F10 cells did not differ between the untreated and MSA-treated groups indicating that the viability of these cells was not compromised, XTT assays revealed more metabolic active cells under MSA treatment in comparison to the control group indicating that other enzymes catalyzing the formation of formazan might be upregulated. As formazan is also metabolized by NAD(P)H, the transcription of the NAD(P)H producing enzymes was determined. As shown in Fig. S1, the expression level of both glucose-6-phosphate dehydrogenase and 6-phospho-gluconate dehydrogenase were upregulated in response to MSA treatment. Furthermore, wound healing (Fig. 1A) as well as migration assays (Fig. 1B) support a decreased migration rate of B16F10 cells in the presence of MSA.

Figure 1.

Altered wound healing and migration of B16F10 melanoma cells in response to MSA treatment. (A) To investigate wound-healing processes scratch assays were performed. Therefore, confluent monolayers of B16F10 cells cultured in complete media supplemented with 10% FCS were mechanically disrupted with a sterile pipette tip, left untreated (control) or treated with 2.5 µM MSA and then photographed immediately (0 h) and 8 h after scratching. The assays were performed in triplicates. (B) For migration assays B16F10 cells either left untreated or treated with MSA for 24 h were seeded into trans-well chambers. All assays were independently performed at least three times as described in the Material and Method section. **p < 0.01 (Student's t-test).

Association of the MSA-induced decreased migration of melanoma cells with reduced AKT phosphorylation

Since protein kinase B (AKT) is one of the key regulators of proliferation, cell survival and migration,⁴¹ the phosphorylation status of AKT as well of its negative regulator PTEN were investigated in untreated (control) and MSA-treated B16F10 cells by Western blot analysis. When compare with the control cells MSA treatment resulted in a decreased phosphorylation rate of AKT at serine residue 473, whereas the phosphorylation status at position threonine 308 as well as of the total AKT protein expression level remained unaffected by this treatment (Figs. 2A and B). In contrast, the phosphorylation rate as well as the protein expression level of PTEN increased in response to MSA treatment. Since AKT phosphorylation might also regulate the glucose metabolism, the glucose uptake rates were determined in untreated and MSA-treated B16F10 cells by flow cytometry. As shown in Figs. 2C and D, the glucose uptake rate of MSA-treated cells was slightly reduced, which is in accordance with the reduced AKT phosphorylation rate.

Figure 2.

Changes in the phosphorylation pattern of AKT and glucose uptake after MSA exposure of B16F10 melanoma cells. (A) Representative phosphorylation states and expression levels of AKT and PTEN in B16F10 cells following treatment with 2.5 µM MSA for 24 h were determined by Western blot analyses as described in the method section. (B) Relative quantification of p-AKTSer473, p-AKTThr308, total AKT, p-PTENSer380/Thr382/383 and total PTEN expression levels normalized to GAPDH and thereafter set in relation to the control group. (C) Glucose uptake of B16F10 cells after treatment with 2.5 µM MSA was measured using the fluorescent glucose analog 2-NBDG, which was determined by flow cytometry as described in the Material and methods section. (D) 2-NBDG fluorescent intensities were normalized to the respective unstained control. All assays were independently performed at least three times. *p < 0.05 (Student's t-test). 2-NBDG, 2-(N-(7-Nitrobenz-2-oxa-1,3-diazol-4-yl)Amino)-2-Deoxyglucose.

MSA-mediated changes in the protein expression pattern of B16F10 melanoma cells

To gain further insights into the anti-carcinogenic potential of MSA, 2D-DIGE analysis of untreated (control) and MSA-treated B16F10 cells was performed in an untargeted approach (Fig. 3, Table 1). The consensus map across all experimental samples (three biological replicates/group, n = 6) was comprised of 872 distinct spots. In total, 33 differentially expressed protein spots could be identified, by subjecting them to MALDI-TOF-MS analyses resulting in the identification of 14 unique protein species. The identified proteins were grouped according to their biological function(s) (Fig. 3C). The expression pattern of three of the upregulated proteins are controlled by the transcription factor Nrf2.⁴² These include glucose 6-phosphate dehydrogenase (G6PD), peroxiredoxin 1 (PRXD1) and S-formylglutathione hydrolase (ESTD), all involved in mounting/maintaining the anti-oxidative defense. In addition, two ER-resident enzymes named protein disulfide isomerase (PDI) and calreticulin (CALR) were upregulated in response to MSA treatment. PDI exhibits oxidoreductase functions and together with CALR is involved in the MHC class I assembly pathway.^43,44 In contrast, the heterogeneous ribonucleoprotein A2B1 (hnRNPA2B1), which is involved in splicing processes, was downregulated in the presence of MSA.

Figure 3.

Differential protein expression pattern of untreated and MSA-treated B16F10 melanoma cells. (A) A representative preparative protein expression pattern of B16F10 melanoma cells following two-dimensional gel electrophoresis (2-DE) is shown as described in Materials and methods sections. Differentially expressed proteins as defined by 2D-DIGE were identified by MALDI-TOF-MS analyses as described in the material section and are indicated by their protein IDs (see also Table 1). (B) Representative heat map and cluster analysis for the panel of identified differentially expressed proteins was done using the software TM4 MeV4.9 (http://www.tm4.org/). (C) Classification analysis of the differentially expressed proteins according to their biological processes was performed using the Panther gene ontology (GO) software (http://pantherdb.org/).

Table 1.

Differentially expressed proteins in response to MSA treatment

					Theoretical
Protein	Uniprot ID	Score MS	Sequence Coverage (%)	No. of matched peptides	Mass (kDa)	pI	Ratio MSA/C
Peroxiredoxin-1 (PRDX1)	P35700	268	78	25	22.39	8.26	2.57 ↑
Calreticulin (CALR)	P14211	158	44	21	48.136	4.33	1.96 ↑
Creatine kinase B-type (KCRB)	Q04447	235	75	30	42.971	5.40	1.88 ↑
S-formylglutathione hydrolase (ESTD)	Q9R0P3	116	47	14	31.87	6.7	1.69 ↑
Fructose-bisphosphate aldolase A (ALDOA)	P05064	266	85	27	39.787	8.31	1.64 ↑
Protein disulfide-isomerase (PDIA1)	P09103	262	48	33	57.422	4.77	1.58 ↑
Glucose-6-phosphate 1-dehydrogenase X (G6PD1)	Q00612	407	66	49	59.681	6.06	1.52 ↑
Galectin-1 (LEG1)	P16045	118	61	11	15.198	5.32	1.52 ↑
Eukaryotic translation initiation factor 3 subunit A (EIF3A)	P23116	100	21	28	162.237	6.38	0.38 ↓
Protein CDV3 (CDV3)	Q4VAA2	70	54	9	29.711	5.84	0.38 ↓
Ribose-phosphate pyrophosphokinase 2 (PRPS2)	Q9CS42	57	37	9	35.163	6.15	0.37 ↓
Heterogeneous nuclear ribonucleoproteins A2/B1 (hnRNPA2/B1)	O88569	179	52	25	37.437	8.97	0.33 ↓
Fermitin family homolog 3 (URP2)	Q8K1B8	59	24	11	76.157	6.60	0.32 ↓
Programmed cell death 1 ligand 2 (PD1L2)	Q9WUL5	56	19	5	28.201	8.46	0.28 ↓

Values are given as means (n = 3) in relation to the untreated control group. ↑ indicate significantly upregulated proteins, ↓ indicate significantly downregulated proteins (p < 0.05, Student´s t-test).

Validation of representative differentially expressed proteins in response to MSA treatment

The upregulation of the PRXD1, G6PDH, ESTD, PDI and the downregulation of hnRNPA2/B1 were validated by RT-PCR in untreated (control) and MSA-treated B16F10 cells (Fig. 4A). The mRNA and protein expression levels of PDI were enhanced in the presence of MSA (Fig. 4B). Since PRDX1 and G6PDH are both regulated by Nrf2,⁴² the expression level of the latter was determined by Western blot analysis. As shown in Fig. 4C, no nuclear Nrf2 accumulation was detected in untreated B16F10 melanoma cells, whereas MSA treatment strongly enhanced Nrf2 accumulation in the nucleus. Furthermore, the transcriptomic profiling of additional Nrf2 targets revealed an enhanced Nrf2 activity in response to MSA treatment. Since Nrf2 activation in combination with the transcriptional induction of Nrf2 target genes could be mediated by enhanced ROS levels, intracellular ROS levels were analyzed using a ROS-sensitive reporter dye. As shown in Figs. 5A and B, increased ROS levels were detected within 4 h after MSA treatment. Furthermore, total glutathione (GSH) levels were approximately 2-fold increased in B16F10 cells after 24 h MSA treatment (Fig. 5C) suggesting that MSA might interrupt/interfere with the cellular redox homeostasis.

Figure 4.

Validation of selected MSA targets obtained by proteomic analysis. (A) mRNA expression levels of differentially expressed proteins were determined by qPCR analyses as described in the Materials and methods section. The relative mRNA expression levels were normalized to the reference genes RPL13a, GAPDH and ß-actin and subsequently defined in relation to the control group. (B) Densitometric analysis of the relative PDI1A expression levels based on the Western blot analysis as representatively shown in the lower panel. (C) mRNA analysis of different Nrf2 target genes determined by qPCR (upper panel) and nuclear Nrf2 content of B16F10 cells left untreated or treated with 2.5 µM MSA was determined by Western blot analysis (lower panel). All experiments were independently performed at least three times. *p < 0.05, **p < 0.01, ***p < 0.001 (Student's t-test).

Figure 5.

Changes of the intracellular ROS and GSH levels upon MSA treatment of B16F10 cells. B16F10 cells were left untreated or treated with MSA for 4 h. Intracellular ROS concentrations in these cells were determined using the fluorescent reporter dye DCF by flow cytometry as described in the material and method section. (A) Representative dot blots and (B) MFI of DCF normalized to unstained cells. (C) Intracellular levels of total GSH were measured spectrophotometrically as described in the method section. All experiments were performed at least three times. *p < 0.05 (Student's t-test). DCF, 2′,7′-Dichlorfluorescein.

Identification of MSA-induced alterations in the redox status of proteins and their correlation with MHC class I APM components

In order to determine the MSA-mediated changes in the redox status, redox proteome analysis using iodoTMT tags, which target free cysteine residues, were employed. This led to the identification of proteins with an altered redox status including several ER-resident proteins, which are directly involved in the MHC class I assembly, such as e.g., ERO1a, PDIA1, PDIA3, PDIA4 and CALR. In addition, TRIM28 showed an altered redox-status in response to MSA treatment (personal communications).

By employing the R2: Genomics Analysis and Visualization Platform (http://hgserver1.amc.nl/cgi-bin/r2/main.cgi) correlations between the expression rates of hnRNPa2/b1, TRIM28 and selected components of the MHC class I APM in melanoma were determined (Fig. 6). With this approach, inverse associations of hnRNPa2/b1 to the MHC class I APM component TPN (PC: −0.222, p = 1.1 × 10⁻⁶) and the interferon regulatory factors IRF5 (PC: −0.282, p = 4.7 × 10⁻¹⁰) and IRF9 (PC: −0.221, p = 1.3 × 10⁻⁶) were found. In addition, inverse correlations of TRIM28 with STAT1 (PC: −0.405, p = 5.8 × 10⁻²⁰), STAT2 (PC: −0.232; p = 3.5 × 10⁻⁷), STAT3 (PC: −0.243; p = 9.5 × 10⁻⁸) and with NLRC5 (PC: −2.89; p = 1.7 × 10⁻⁹) were detected, whereas the Nrf2 target NQO1 positively correlated with HLA-A (PC: 0.230; p = 2.4 × 10⁻⁶).

Figure 6.

Correlation analysis. Correlation analysis of selected differentially expressed (A) or redox-modified (B) proteins were performed by employing the R2: Genomics Analysis and Visualization Platform (http://hgserver1.amc.nl/cgi-bin/r2/main.cgi).

Enhanced MHC class I surface antigens of B16F10 cells after MSA treatment

Since Nrf2 has been shown to induce the expression of proteasomal subunits necessary for the generation of peptides⁴⁵ and MSA mediated an altered expression or redox status of CALR, PDI, ERO1a and TRIM28, respectively, it was postulated that MSA upregulates the expression of MHC class I antigen molecules in B16F10 cells, which are per se characterized by low MHC class I cell surface expression levels due to the impaired expression of several APM components. This immune escape phenotype of B16F10 could be restored by IFNγ treatment.⁴⁶ As shown in Fig. 7, a dose-dependent approximately 2- to 4-fold increase of the MHC class I surface expression level in B16F10 melanoma cells treated for 24 h with MSA was detected, which was stable for at least 48 h (Figs. 7A and B). In addition, treatment with DMDSe also resulted in an upregulation of the MHC class I surface expression level comparable to that of MSA. In contrast, treatment with selenite showed no effect, whereas SeMet treatment at least for the higher doses even resulted in a downregulation of the MHC class I surface expression level (Figs. 7C–E).

Figure 7.

Effects of MSA on the MHC class I surface expression of B16F10 melanoma cells. (A) Dot blots of SSC versus anti-MHC class I staining of B16F10 cells treated with different concentrations of MSA (0–2.5 µM) for 24 and 48 h, respectively, are shown. Analyses were performed by flow cytometry as described in Materials and methods. The fluorescence intensity of (B) MSA, (C) DMDSe, (D) selenite and (E) selenomethionine (SeMet) treated cells were normalized to the respective isotype controls. Data represent means ± SD and are expressed in relation to the untreated (0 µM MSA) group. All experiments were performed at least three times. ** indicate significant differences (p < 0.01, Student's t-test) of the treated vs. control group. ## indicate significant differences (p < 0.01, Student's t-test) within the MSA-treated groups. SSC, sideward scatter.

General effects of MSA treatment on the MHC class I expression level of tumor cells

To study whether the effects observed in B16F10 cells in response to treatment with the compound MSA could be also extended to other tumor cells, three human melanoma cell lines (WM1552c, UKRV, Colo875), two human mammary carcinoma cell lines (SK-BR-3 and BT-474) and the human colon carcinoma cell line SW480 were left untreated and treated with MSA as described for B16F10 cells, before their MHC class I expression levels were determined by flow cytometry. As shown in Figs. 8A–F, treatment with MSA increased the MHC class I surface expression levels in all the tested tumor cell lines.

Figure 8.

MHC class I surface expression of tumor cell lines following treatment with 2.5 µM MSA for 24 h as analyzed by flow cytometry. One representative experiment of three independent experiments is shown. Bars represent mean fluorescence intensity (MFI) of MHC-APC normalized to the isotype control within the given independent experiments. ** indicate significant differences (p < 0.01, Student's t-test).

Transcriptional upregulation of MHC class I APM components by MSA

In order to define the underlying molecular mechanism which leads to the observed upregulation of the MHC class I surface expression levels in response to treatment with MSA, the mRNA and protein expression pattern of selected APM components were determined. As shown in Fig. 9A MSA increased the transcript levels of a panel of APM components including the IFNγ-inducible proteasomal subunits low molecular weight protein (LMP)2 and LMP7, TAP1, TAP2, TPN, β₂-m, H-2D and H-2K, yet for LMP10 rather a downregulation of the corresponding transcript level was observed. In addition, the mRNA levels of NLRC5, a critical co-activator of genes of the MHC class I presentation pathway,⁴⁷ were upregulated. To understand the molecular mechanisms of the MSA-mediated regulation mechanism, also the promoter activities of TAP1, TAP2, TPN and LMP2 were determined in response to treatment with MSA and revealed enhanced activation states (Fig. 9B). Furthermore, the protein expression levels of LMP2 and LMP7 were analyzed by Western blot supporting the coordinated expression both at the mRNA and protein levels for these APM components. Moreover, MSA treatment resulted in an enhanced transcript level of signal transducer (STAT)1 and in a shift of STAT1ß toward STAT1α expression at the protein level. In addition, the phosphorylation states of STAT1 and JAK1 were increased following MSA treatment suggesting that MSA enhances the expression of MHC class I surface antigens by inducing APM components and by mimicking IFNγ signaling (Figs. 9C and D).

Figure 9.

Analysis of APM component expression in B16F10 cells after treatment with MSA. (A) mRNA expression levels of APM components were determined by qPCR analyses normalized to the set of reference genes RPL13a, GAPDH and ß-actin as well as in relation to the control group. (B) Different wild-type (WT) APM-luc promoters and the ß-gal promoter were transiently co-transfected into B16F10 cells 24 h prior treatment with MSA for 24 h. Luc activity was determined as described in Materials and methods and normalized to ß-gal activity. Data are expressed in relation to untreated controls. (C) Representative Western blots of APM components and (D) their quantification by measuring the OD normalized to GAPDH in relation to the control group are shown. Data represent means ± SD of at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 (Student's t-test).

Effect of MSA on the IFNγ signal transduction in B16F10 cells

In addition to components of the MHC class I APM the transcript levels of components of the IFNγ signaling pathway were analyzed in response to treatment with either MSA or IFNγ alone as well as to treatment with a combination of both substances (Fig. 10). MSA and IFNγ treatment alone increased the transcript levels of IRF1, IRF5, IRF7 and IRF9, whereas IRF4 was downregulated by MSA and upregulated by IFNγ. Combination of both substances resulted in a further increase of IRF5, STAT1 and STAT2 transcript levels.

Figure 10.

Changes in the mRNA expression levels of the IFNγ signaling pathway in response to MSA and IFNγ treatment. B16F10 cells were treated with 2.5 µM MSA and 200 U/mL IFNγ alone or in combination for 24 h. The mRNA expression levels were determined by qPCR analyses, normalized to the reference genes RPL13a, GAPDH and ß-actin and subsequently defined in relation to untreated control group. Data represent the mean of at least three independent experiments. * indicate significant differences between the treated and control groups (*p < 0.05, **p < 0.01, Student's t-test). # indicate significant differences within the different treatment groups (^#p < 0.05, ^##p < 0.01, Student's t-test).

Discussion

The trace element Se has been discussed to exhibit chemo-preventive potential. Furthermore, newer studies reveal also chemo-therapeutic effects of Se.^3,4 However, the mechanisms, by which Se exerts those beneficial effects are largely unknown, but include their anti-oxidative properties, the induction of apoptosis and inhibition of cell proliferation. There exist two key metabolites of Se, which are critical for chemoprevention/chemotherapy.⁴⁸ Hydrogen selenide necessary for the formation of selenoproteins results mainly from inorganic selenocompounds such as selenate or selenite, respectively, and methylselenol, which is generated from organic selenocompounds. In this study, the pro-oxidant MSA, a synthetic selenocompound and precursor of methylselenol, induced MHC class I surface expression not only in B16F10 melanoma cells, but also on several other tumor cell lines, including human cancer cells. These data were in line with the effects of the pro-oxidative selenocompound DMDSe, which also induces MHC class I surface expression. In contrast, the selenocompounds selenite and SeMet known to produce less amounts of ROS and methylselenol during their metabolic processing/turnover failed to increase MHC class I surface expression. These results indicate that the effects on the MHC class I expression pattern are rather mediated by the formation of ROS or methylated selenocompounds, such as methylselenol, as by Se itself. This hypothesis is in accordance to results by Hagemann-Jensen and coworkers demonstrating a MSA- and DMDSe-mediated upregulation of NKG2D receptor ligands, whereas selenite, selenate, SeMet and SeCys did not influence MICA and MICB expression.²³ In accordance to our study, Enqvist and co-authors³¹ described no effect of selenite on the expression of MHC class Ia molecules, but detected a strong post-transcriptional downregulation of HLA-E antigens, which was associated with an enhanced NK cell-based tumor killing. HLA-E antigens act as both inhibitory and activating ligands for CD94/NKG2 receptors depending on the antigen load: Upon loading with a peptide derived from an HLA-Ia leader sequence, HLA-E can interact with the inhibitory receptor CD94/NKG2A, while upon loading with a HLA-G leader sequence HLA-E can interact with the activating receptor NKG2C.⁴⁹ In contrast to this important role of HLA-E in the innate immunity, only few experiments revealed that MHC class Ib molecules could interact with T cells.^30,50 It could be shown that HLA-E⁺ targets loaded with peptides derived from splice variants of the stress inducible anti-oxidant enzyme peroxiredoxin 5 (Prdx5) are recognized by HLA-E-restricted CD8⁺ T cells.⁵¹ In contrast, complexes composed of HLA-E and Prdx5 did not interact with NKG2A and therefore did not downregulate NK cell function.⁵¹ HLA-E might be upregulated in response to cytokines like IFNγ⁵² suggesting that an upregulation of HLA-E in response to MSA cannot be excluded. Thus, further analyses are required to determine the impact of MSA treatment on the expression of MHC class Ib molecules.

In our study, the induction of MHC class Ia was caused by a transcriptional upregulation of components of the APM, including proteasomal subunits and components of the peptide loading complex. Furthermore, MSA partially mimics IFNγ signaling, such as the upregulation of STAT1, JAK1, IRF1, IRF5, IRF7 and IRF9 on the mRNA and/or protein expression levels. In addition, MSA treatment leads to activation of the transcription factor Nrf2, thus resulting in the initiation of the upregulation of genes involved in the anti-oxidant defense. The observed enhanced Nrf2 accumulation in the nucleus could be mediated by the formation of ROS in response to treatment with MSA. Under normal conditions, Nrf2 is bound to its inhibitor protein Keap1 and marked for proteasomal degradation. ROS or electrophilic agents modify redox-sensitive cysteine residues of Keap1, leading to conformational changes and the release of Nrf2, which in turn could translocate into the nucleus and initiate the transcriptional induction of its target proteins.⁴² In our study, a strong upregulation of Nrf2 targets was observed, such as NQO1. Next to its role in the anti-oxidative defense, NQO1 is described to stabilize p53 and thus prevents its proteasomal degradation.⁵³ Concerning the peptide transporter subunit TAP1, it has been described that p53 regulates its expression and enhances the peptide transport as well as the expression of MHC class I peptides on the cell surface.⁵⁴ This might be a link, given that MSA enhances p53 protein levels as well as its activity.^55,56

It is noteworthy that IFNγ is an important inducer of ROS, via NOX enzymes, that causes oxidative stress and an enhanced formation of oxidative damaged proteins.^57,58 Furthermore, it could be shown that the upregulation of subunits of the so-called immuno-proteasome (LMP2, LMP7, LMP10) might rather be an mechanism to degrade oxidative-damaged proteins.⁵⁹ Also, in our study an upregulation of proteasomal subunits after treatment with MSA could be shown, which might be attributed to the enhanced activity of the transcription factor Nrf2 mediated in response to MSA treatment, which has been found to induce the transcriptional upregulation of LMP2 and LMP7.⁴⁵

Despite the clear association between the observed induction of ROS along with the increased expression of Nrf2 (oxidative stress response) and the elevated MHC class I surface expression levels, which are associated with an upregulation of APM components, the exact molecular mechanisms of the observed effects in response to treatment with MSA have not yet been identified. One explanation might be that MSA has been suggested to inhibit HDACs.²⁰ In line with this hypothesis, HDAC inhibitors have been described to upregulate both MHC class I and class II molecules at the cell surface.⁶⁰ In addition, MSA has been shown to induce the expression of MICA and MICB in Jurkat cells to the same extent as HDAC inhibitors.²³ Therefore, further analysis are required to test whether the modulation of the expression pattern of MHC class I molecules by MSA is indeed mediated at least in part by decreased HDAC activities.

In conclusion, our data provide a new molecular mechanism which might explain the potential anti-carcinogenic properties of Se. As MHC class I molecules are often downregulated in tumor cells and thus regarded as one of the key mechanism by which tumor cells might escape from immune surveillance, treatment with MSA might be in particular beneficial for the treatment of tumors exhibiting low MHC class I surface expression levels, thereby improving and likely reshaping the mounting of an immune response.

Material and methods

Cell culture

The murine melanoma cell line B16F10 and human cell lines WM1552c, UKRV, Colo875, BT474 and SK-BR-3 were routinely cultured in RPMI media supplemented with 10% FCS, 100 U/mL penicillin and 100 µg/mL streptomycin, whereas the human colon carcinoma cell line SW480 was maintained in complete DMEM media. All cells were cultured at 37°C in 5% (v/v) CO₂ humidified air.

Reagents and treatment of cells

MSA, dimethyldiselenide, sodium selenite and SeMet were all purchased from Sigma-Aldrich. IFNγ was purchased from Immunotools (#12343534). Stock solutions were prepared in either water (MSA, selenite, SeMet) or DMSO (DMDSe). Cells were seeded and incubated to reach adherence. After 24 h, cells were supplemented with the indicated concentrations of Se compounds and/or 200 U/mL IFNγ and incubated as indicated prior to the respective analyses. Non-supplemented and DMSO treated cells served as controls.

Cytotoxicity assay

Cytotoxicity was determined by measurement of LDH release using the CytoTox 96® Non-Radioactive Cytotoxicity Assay (#G1781, Promega) according to the manufactures instructions. In brief, 5 × 10³ cells were seeded per 96-well. After 24 h, total media was removed and replaced by media containing different concentrations of MSA for additional 24 h. LDH release was measured at 490 nm using the infinite M200 pro plate reader (Tecan, Germany).

Cell viability assay

To asses cell viability in response to MSA treatment the CellTiter-Glo® Luminescent Cell Viability Assay (#G7571, Promega) was used according to the manufactures instructions. In brief, 5 × 10³ cells were seeded per 96-well. After 24 h, total media was removed and replaced by media containing different concentrations of MSA (0–10 µM) for additional 24 h. Luminescence measurement was done using the luminometer MicroLumatPlus LB96V (EG&G Berthold Technology, Germany).

Wound healing and migration

For monolayer wound-healing assay, cells were seeded in 12-well plates. At a 100% confluence wounds were made using a sterile pipette tip. Wound size was measured after 8 h as described previously.⁶¹ For determination of the cell migration, 1 × 10⁵ of 24 h treated or untreated cells, respectively, cells were seeded into the upper well of the Transwell chamber system (Corning), with medium containing 0.5% FBS. At the lower chamber, 10% FKS was added as an attractant. After incubation at 37°C for 4.5, 18 and 20 h, respectively, the non-migrated cells of the top insert were completely removed, and the cells on the bottom insert surface were lysed with CellTiter-Glo (Promega) before the ATP content was measured in the luminometer (EG&G Berthold Technology, Germany).

2D DIGE analysis

Three biological replicates per sample representing either the control group or the MSA-treated group were subjected to undergo proteomic profiling. DIGE analysis was performed using the minimal labeling approach according to the manufacturer´s instructions (NH DyeAGNOSTICS, Halle, Germany) with the exception that 25 µg of the respective protein lysates as well as the internal protein standard were labeled with 100 pmol of the given G-dye. 2D gel electrophoresis and subsequent MALDI-TOF MS analysis were performed as described previously.⁶² Differentially expressed proteins were clustered according to their functions using the Panther gene ontology (GO) software (http://pantherdb.org/).⁶³

Flow cytometry

Flow cytometric analysis was performed as previously described.⁶⁴ Briefly, cells were stained with following monoclonal antibodies (mAb): APC-conjugated anti-mouse MHC class I (H-2D^b) (#17–5999, ebioscience), APC-conjugated mouse IgG2a (#400222, Biolegend), APC-conjugated anti-human HLA-ABC (#555555, BD biosciences), APC-conjugated mouse IgG2B (# IC00417, RD systems), before fluorescence was determined on a BD LSRFortessa cytometer (BD Biosciences). Data were analyzed on a free tool “Flowing software” (http://www.uskonaskel.fi/flowingsoftware/).

Determination of intracellular ROS and GSH levels

For determination of ROS levels, 2×10⁵ cells were transferred into a FACS analysis tube, washed with PBS and re-suspended in 1 mL PBS containing 1 µM DCFH-DA (AAT Bioquest®, Inc., CA). After 30 min of incubation at 37°C, cells were centrifuged, supernatant discarded and the cell pellet re-suspended in RPMI media supplemented with or without MSA followed by an incubation of 4 h. After washing the cells with PBS analysis was performed in FL channel of the BD LSRFortessa™ cytometer.

Total GSH (reduced and oxidized) concentrations were determined in plasma and liver homogenates according to a standard procedure coupled to GSH reductase and DTNB.⁶⁵ The concentrations were calculated using a GSSG standard curve ranging from 1 to 10 nM GSH equivalent/mL.

Determination of glucose uptake

5 × 10⁵ untreated or MSA-treated cells were re-suspended in 2-NBDG solution (0.1 mM in PBS) and incubated for 20 min at RT prior to analysis by flow cytometry (BD LSRFortessa cytometer).

RT-PCR

Total RNA was extracted using the NucleoSpin RNA extraction kit (MACHEREY-NAGEL) according to the manufactures instructions and converted to cDNA using the cDNA synthesis kit and Oligo dT primer from Thermo Scientific. Amplification of the cDNA was performed using the GoTaq® Real-Time qPCR system (Promega) and SybrGreen as a fluorescent reporter. Target specific primers were conducted using the program Primer3⁶⁶ and listed in Table S1. Analysis of amplification data were performed with the Rotor-Gene 6000TM series software (Qiagen, Hilden, Germany) using the method according to Pfaffl.⁶⁷ Amplifications of GAPDH, RPL13a and ß-Actin were used for normalization. The data are represented as relative mRNA expression levels as x-fold of the control group of at least three separate experiments.

Determination of promoter activity of TAP1, TAP2, TPN and LMP2

Constructs and procedure are described elsewhere.⁶⁸ In brief, 5×10³ cells were seeded into 96-well plates. After 24 h, cells were co-transfected with 0.3 µg pAPM-luc and pSV-galactosidase vector (Promega), which served as transfection control, using Lipofectamine in OptiMEM (Invitrogen) according to the manufacturers’ instructions. After 6 h, cells were washed and cultured in complete media for further 24 h, before treatment with 2.5 µM MSA. Untreated cells served as the control. After 24 h, cells were lysed and Luc activity was determined using the Luc-assay system (Promega) according to the manufactures instructions. The transfection efficiency was determined by measuring ß-gal activity. Relative luc activity was calculated as (luc activity/ß-gal activity)−(mean pGL3 – mean b-gal activity).

Western blot analysis

For Western Blot analysis, cells were washed in 0.1 M sodium phosphate buffer containing 5 mM EDTA and protease as well as phosphatase inhibitors (Halt™, Thermo Scientific) and lysed using a TissueLyser (Qiagen, Germany). After centrifugation (13.000 x g, 30 min, 4°C) the supernatants were collected and stored at −80°C until further analysis. Determination of protein content was performed according to the method described by Bradford⁶⁹ modified for a 96-well plate. Western blot analysis was performed as described previously.⁷⁰ The following primary antibodies were purchases from Cell Signaling Technology and used for immune detection: anti-JAK1 (CST #3332), anti-phospho-JAK1 (CST #3331), anti-phospho-STAT1 (CST #9167), anti-STAT1 (CST #9172), anti-LMP2 (biotrend #PW8205), anti-LMP7 (biotrend #PW8200), anti-PDI1A (CST #3501), anti-phospho-PTEN (CST #9549), anti-PTEN (CST #9556), anti-AKT (CST #9916), anti-phopsho-AKT^Ser473 (CST #9271), anti-phospho-AKT^Thr308 (CST #2965), anti-GAPDH (CST #2218). The anti-Nrf2 antibody (Santa Cruz Biotechnology, sc-722) was kindly provided by the group of Prof. Lillig (University of Greifswald).

The secondary HRP-coupled anti-rabbit/anti-mouse Abs were purchased from cell signaling technology and DakoAgilent, respectively.

Statistical analysis

Mean values were calculated from results of at least three independent experiments and given as means ± their standard deprivation (SD). SPSS 20 software was used to analyze significant differences within the groups, therefore Student's t-test was implemented after analyzing variance homogeneity. Differences between the groups were considered to be significantly different at p < 0.05.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Financial disclosure

The authors would like to state that they have no financial disclosures to report.

Acknowledgements

We thank Manuela Gellert and Christopher Lillig for providing the Nrf2 antibody.

Funding

This work was supported by an interdisciplinary DFG grant (grant numbers: LI1527/3–1, WE1467/13–1 and MU3275/3–1).