Immunosuppressive effects of melanoma ... - Wiley Online Library

14 downloads 0 Views 179KB Size Report
Apr 2, 2001 - tasis from a patient attending the Sydney Melanoma Unit (Sydney,. Australia) and established in the laboratory. The derivation of. MM200 ...
Int. J. Cancer: 92, 843– 850 (2001) © 2001 Wiley-Liss, Inc.

Publication of the International Union Against Cancer

IMMUNOSUPPRESSIVE EFFECTS OF MELANOMA-DERIVED HEAVY-CHAIN FERRITIN ARE DEPENDENT ON STIMULATION OF IL-10 PRODUCTION Christian P. GRAY1, Agustin V. FRANCO1, P. AROSIO2 and Peter HERSEY1* 1 Department of Oncology and Immunology, Newcastle Mater Hospital, Newcastle, New South Wales, Australia 2 Protein Engineering Unit, Department of Biological and Technological Research, San Raffaele, Milan, Italy Cultured melanoma cells release soluble factors that influence immune responses. Screening of a cDNA library with anti-sera from a melanoma patient identified an immunoreactive plaque, which encoded heavy-chain ferritin (H-ferritin). Previous studies have drawn attention to the immunosuppressive effects of this molecule and prompted further studies on its biochemical and functional properties in human melanoma. These studies demonstrated, firstly, that H-ferritin appeared to be secreted by melanoma cells, as shown by immunoprecipitation of a 21.5 kDa band from supernatants. It was also detected in extracts of melanoma cells by Western blotting as 43 and 64 kDa dimers and trimers of the 21.5 kDa fraction. Secondly, flow-cytometric analysis of H- and light-chain ferritin (L-ferritin) expression on melanoma showed a wide variation in L-ferritin expression and consequently of the ratio of H- to L-ferritin expression. Suppression of mitogenic responses of lymphocytes to anti-CD3 showed a correlation with the ratio of H- to L-ferritin in the supernatants and was specific for H-ferritin, as shown by inhibition studies with a monoclonal antibody (MAb) against H-ferritin. Similar results were obtained with H- and L-ferritin from other sources. Suppression of mitogenic responses of lymphocytes to anti-CD3 by H-ferritin was inhibited using a MAb against IL-10, which suggested that the immunosuppressive effect of H-ferritin was mediated by IL-10. Assays of cytokine production from anti-CD3–stimulated lymphocytes showed that H-ferritin markedly increased production of IL-10 and IFN-␥ and had only slight effects on IL-2 and IL-4 production. Our results suggest that melanoma cells may be a major source of H-ferritin and that production of the latter may account for some of the immunosuppressive effects of melanoma. © 2001 Wiley-Liss, Inc. Key words: heavy-chain ferritin; human melanoma; immunosuppression; IL-10; light-chain ferritin; SEREX

The mechanisms by which melanoma avoids destruction by the immune system are of much interest. One possibility is that melanoma cells release soluble factors, which inhibit immune responses in the local tumour environment. Such factors have been found in the supernatants from sarcoma, lung carcinoma, melanoma, head-and-neck carcinoma and colon cancer; they inhibit mitogen and other strong lymphocyte responses.1–3 The nature of the factors responsible for the immunosuppressive effects in many instances are unknown, but in others, a wide variety of factors have been implicated, such as PGE2,4 TGF-␤2,5,6 IL-10,7,8 ␣-MSH,7 IL-6,9 hydrogen peroxide10 and Fas ligand.11 Immunoscreening of a cDNA library from the MM200 melanoma line with sera from a melanoma patient identified several plaques that were shown by cDNA sequence analysis to be heavychain ferritin (H-ferritin). Previous studies have shown that Hferritin may suppress proliferation of T cells,12–14 E-rosette formation by T cells15 and colony formation by normal human macrophages.16 Ferritin is a major tissue iron-binding protein17 and, in its native form, is approximately 500 kDa. It is composed of 24 subunits consisting of acid/heavy (H) and basic/light (L) chains.18,19 The genes encoding H- and L-ferritin are found in different chromosomes and are transcriptionally independent.20 The 24-subunit polymer may form isoferritins, which are either more acidic (Hrich) or more basic (L-rich), depending on the relative proportions of H and L chains. Liver and spleen ferritins are basic because they

are made up mainly of light chains, with very few heavy chains. In contrast, heart, kidney and placental ferritins are highly acidic because they are composed of mostly heavy chains.21 Interestingly, ferritin in cancer cells consists mainly of heavy chains.22,23 Others have suggested that there is another species of ferritin in cancer cells called super-heavy chain or P43.24 In the present studies, we sought to characterize the nature of ferritin in melanoma cells and to investigate its possible effects on immune responses. MATERIAL AND METHODS

Cell lines Mel-JG was isolated from a fresh surgical biopsy of s.c. metastasis from a patient attending the Sydney Melanoma Unit (Sydney, Australia) and established in the laboratory. The derivation of MM200, Me1007, Me4405 and Me10538 is described elsewhere.25 All melanoma cell lines were positive for tyrosinase and MART-1 mRNA by RT-PCR, described elsewhere.26 MRC-5 lung fibroblasts were obtained from BioWhittaker (Walkersville, MD). All melanoma cell lines and MRC-5 were cultured in DMEM containing 5% FCS (Commonwealth Serum Laboratories, Melbourne, Australia). Human umbilical vein endothelial cells (HUVECs) were kindly supplied by Dr. A.D. Hibberd and Mr. D. Clark (Transplantation Unit, John Hunter Hospital, Newcastle, Australia); they were isolated from umbilical veins of placenta by digestion in collagenase, as described elsewhere,27 and grown in M199 medium (Life Technologies, Gaithersburg, MD) supplemented with 100 mg/l L-glutamine (Life Technologies), 20% FCS, 135 mg/l heparin (Sigma, St. Louis, MO) and 16.7 ␮g/l endothelial cell growth supplement (Sigma). Over 95% of the cells expressed CD31 and von Willebrand factor using flow cytometry. Cell lines were tested routinely for Mycoplasma by the Mycoplasma Detection Set PCR method (Takara Shuzo, Tokyo, Japan). Antibodies and recombinant proteins The monoclonal antibody (MAb) rH02, specific for human H-ferritin, was kindly provided by Dr P. Arosio and is described elsewhere.28 Dr T.L. Nagabhushan (Schering-Plough, Bloomfield, NJ) kindly provided the MAb JES3-9D7, specific for human IL-10 and described elsewhere.29 The 10-F10 MAb against H-ferritin was purchased from Fitzgerald (Concord, MA). A MAb against L-ferritin, 1A-4E2, was purchased from Bioclone Sydney, Australia). The IgG1 isotype MAb against TNP was purchased from Pharmingen, (San Diego, CA). The rat IgG isotype control was Grant sponsor: Hunter Melanoma Foundation and Melanoma and Skin Cancer Institute, New South Wales, Australia. *Correspondence to: Oncology and Immunology Unit, Room 443, David Maddison Clinical Sciences Building, Cnr. King and Watt Streets, Newcastle, NSW 2300, Australia. Fax: ⫹61-2-492-36184. E-mail: [email protected] Received 25 August 2000; Revised 13 November 2000; Accepted 19 January 2001 Published online 2 April 2001

844

GRAY ET AL.

purchased from Sigma. Anti-CD3 was the OKT3 MAb. Heart and liver ferritin were purchased from Fitzgerald. Recombinant Hferritin (rH-ferritin) and rL-ferritin were purchased from Calbiochem (San Diego, CA). rIL-10 was purchased from Pharmingen.

(1:2,500 dilution) and the mixture was further incubated for 90 min at room temperature. The membrane was subsequently washed 5 times and developed using a Renaissance chemiluminescent kit (NEN, Boston, MA).

Isolation of plasmid DNA and DNA sequencing Immunoscreening of an MM200 cDNA library (Novagen, Maddison, WI) with melanoma patient Mel-FH was carried out as described elsewhere.30 The host strain BM25.5 was used for automatic subcloning of the pEXLOX vector; 0.1 ml was added to 0.1 ml of the positive plaque diluted in suspension medium. The host-phage mixture was incubated for 30 min at 37°C. LB broth (3 ml) containing 100 ␮g/ml ampicillin was incubated overnight in plaque diluted in suspension medium. The host-phage mixture was incubated for 30 min at 37°C, inoculated with colonies picked from the plate and incubated with agitation overnight at 37°C. The overnight culture was pelleted and subjected to the Wizard plasmid miniprep purification system (Promega, Madison, WI). Purified plasmid DNA was digested with EcoRI and HindIII. The restriction enzyme product of 708 bp was isolated from a 1% agarose gel using the Geneclean Kit (BIO 101, La Jolla, CA). The isolated insert was ligated to pBluescript SK phagemid (Stratagene, La Jolla, CA) using T4 DNA ligase (Promega) for DNA-sequencing purposes. Automated DNA sequencing was performed using the dye-labeled terminator technique on a Perkin-Elmer (Norwalk, CT) ABI prism 377 DNA sequencer by the Biomolecular Research Facility (Newcastle, Australia), using T7 and T3 primers for the pBluescript phagemid (Stratagene) and T7 and SP6 primers for the pGEM-T vector (Promega) ligation products. For RT-PCR, total RNA was isolated from 4 ⫻ 106 melanoma cells using the SV total RNA isolation kit (Promega). First-strand cDNA was synthesized using the Expand reverse-transcriptase system (Boehringer-Mannheim, Mannheim, Germany). cDNA (5 ␮l) was used in a 50 ␮l amplification reaction with primers from the H-ferritin cDNA sequence: sense, 5⬘-GCTCCAGCGCCGCGCAGCCACC-GC-3⬘; anti-sense, 5⬘-GGGGGATCCGCATGCACTGCCTTGGTG-3⬘.31 Three units of Expand high-fidelity Taq polymerase (BoehringerMannheim) were added per reaction. Samples were amplified for 35 cycles in a Perkin-Elmer 9600 thermal cycler. The annealing temperature of H-ferritin cDNA was 59°C. Amplified DNA was subsequently cloned in pGEM-T as described above for sequencing purposes.

Immunoprecipitation of H-ferritin from melanoma supernatants Three milliliters of melanoma supernatants were pre-cleared by rotation with 50 ␮l protein G Sepharose–packed beads (Amersham) at 4°C for 2 hr and then with 50 ␮l of freshly packed beads overnight at 4°C. Ten micrograms of MAb rH02 were then added to the lysate and rotated at 4°C for 2 hr. Fifty microlitres of protein G Sepharose–packed beads were added and rotated for a further 2 hr. Protein G Sepharose–packed beads were washed 4 times with ice-cold lysate buffer before elution of proteins in lysate buffer at room temperature for 1 hr. Subsequently, immunoprecipitates were mixed with 6⫻ denaturing buffer and run on 12.5% SDSPAGE. The gel was electrotransferred to an ECL nitrocellulose membrane before immunoblotting, as described above.

Western blotting Melanoma cells were harvested and washed twice with PBS. The pellet was dissolved in 1 ml of lysis buffer with the protease inhibitors 1,4-dithio-L-threitol and PMSF, added at 1 ␮M and 1 mM, respectively.32 Samples were incubated for 1 hr on ice and subsequently macerated with a mortar and pestle. rH-ferritin was used as a positive control. Twenty micrograms of each sample, measured by protein assay (Bio-Rad, Hercules, CA), were treated with 6⫻ denaturing buffer33 and run on 12.5% SDS-PAGE. To dissociate protein multimers into monomers, lysates were precipitated in tricarboxylic acid (TCA)/guanidinium hydrochloride (GuHCl). Briefly, 20 ␮l of 6 M Gu-HCl were mixed with 20 ␮g of cell lysate. This was then diluted to 100 ␮l with water. To the mixture, 100 ␮l of 10% TCA were added. Subsequently, the mixture was left on ice for 30 min and spun down for 15 min at 12,000 g at 4°C. The pellet was washed with 100 ␮l of ice-cold ethanol, dried, resuspended in 6⫻ denaturing buffer and run on 12.5% SDSPAGE containing 8 M urea. The gel was subsequently electrotransferred to an ECL nitrocellulose membrane (Amersham, Aylesbury, UK)33, washed with PBS and blocked with 1% sodium casein in TBST (20 mM Tris, 500 mM NaCl, 0.05% Tween-20, pH 7.5) for 1 hr at 37°C. The membrane was then washed 5 times in TBST. MAb rH02, specific for H-ferritin, was added to the membrane at a concentration of 1 ␮g/ml in TBST and incubated for 90 min at room temperature. The membrane was subsequently washed 5 times with TBST, horseradish peroxidase– conjugated goat anti-mouse MAb (Bio-Rad) was added to the membrane

Flow cytometry Analysis was carried out using a Becton Dickinson (Mountain View, CA) FACScan flow cytometer. Cells were fixed in 4% paraformaldehyde for 20 min at room temperature. Optimal concentrations of MAbs 10-F10 and A1-4E2 directed against H- and L-ferritin, respectively, were added to the cells in 100 ␮l PBS containing 10% human serum and incubated for 30 min at 4°C. Cells were washed twice with PBS before incubation with F(ab⬘)2 fragment affinity-isolated, FITC-conjugated sheep anti-mouse immunoglobulin (Silenus-Amrad Biotech, Boronia, Australia) plus 10% human serum for 30 min at 4°C. Cells were then washed once in PBS before analysis. A minimum of 5,000 cells were analyzed by flow cytometry. The percentage of cells expressing H- or L-ferritin was calculated as the difference in positive area between the positive and negative control histograms. The positive area was that to the right of the intersection of the 2 curves.34 Assay for inhibition of anti-CD3 stimulation of normal lymphocytes by melanoma supernatants Melanoma supernatants were prepared from 2 ⫻ 106 melanoma cells, cultured in a 5 cm2 flat-bottomed well with 1 ml of AIM-V serum-free medium (Life Technologies) for 24 hr. Melanoma supernatants were filtered using a 0.2 ␮m filter (Sartorius, Go¨ttingen, Germany) before use. Peripheral blood lymphocytes (PBLs) were separated from whole blood on a Ficoll-Paque gradient (Amersham) by centrifugation at 500 g for 30 min at room temperature. PBLs were washed twice in DMEM, with centrifugation at 500 g for 10 min at room temperature. PBLs (2 ⫻ 105), in 200 ␮l of DMEM containing 10% FCS, were cultured with 20 ␮l of melanoma supernatant in a 96-well U-bottomed plate (Falcon, Becton Dickinson, Franklin Lakes, NJ) in triplicate. Anti-CD3 (50 ng) was added, and cultures were incubated for 72 hr in a 37°C incubator supplemented with 5% CO2. Where required, the blocking antibody MAb rH02 was incubated with the melanoma supernatants for 1 hr at 37°C prior to addition to the proliferation assay. For IL-10 blocking experiments, MAb JES3-9D7 was added directly to the proliferation assay. After 72 hr, 2 ␮Ci of [5-125I]iodo21-deoxyuridine ([125I]UDR) (Amersham) was added for a further 4 hr before cell harvesting. [125I]UDR incorporation was then measured on a CompuGamma CS Gamma counter (Wallac, Oy, Finland). The degree of inhibition of anti-CD3 stimulation was calculated as previously described.35 Assay for IL-2, IL-4, IL-10 and IFN-␥ An anti-CD3 proliferation assay was performed as described in Material and Methods. After 24 hr, supernatants were collected from PBLs. IL-2, IL-4, IL-10 and IFN-␥ were then assayed using a biotinylated ELISA system (Pharmingen), according to the manufacturer’s instructions, for each cytokine.

H-FERRITIN IN HUMAN MELANOMA

845

FIGURE 1 – (a) Sequence comparison of the 3⬘-untranslated region of H-ferritin cDNA from MM200 melanoma and from human brain and liver. Bases within boxes are different in the MM200 sequence when compared to the other sequences. The base with the grey box is an insertion in the MM200 sequence. (b) Text graph comparing the composition of the pyrimidine-rich sequences within the 3⬘ region of H-ferritin cDNA38 indicated in (a). The sequence derived from melanoma has 9 purines (Pu) and 9 pyrimidines (Py), while the human brain and liver has 1 purine and 16 pyrimidines. The program GAP in the Genetics Computer Group Package55 was used for comparison.

FIGURE 2 – Detection of H-ferritin in supernatants from melanoma cells. H-ferritin of m.w. 21.5 kDa was detected in supernatants collected from melanoma cell lines by immunoprecipitation with the MAb rH02. Recombinant H-ferritin was used as a positive control. An IgG1 isotype control MAb did not immunoprecipitate a similar band. RESULTS

Detection of H-ferritin by SEREX Immunoscreening of cDNA from the MM200 expression library with sera from patient FH resulted in the detection of several clones. Subsequent restriction enzyme analysis of one of the clones with EcoRI/HindIII revealed an insert of 708 bp. DNA sequencing of this clone showed that the insert encoded for human H-ferritin. The cDNA sequence from bases 1 to 656 was identical to that of human brain (Genbank accession number L20941) and human liver (Genbank accession number M11146) H-ferritin. This included the translated portion of H-ferritin (bases 43 to 594), which appears to be identical and highly conserved when compared with other sequences. However, as shown in Figure 1, the 3⬘-untranslated region of Hferritin cDNA from MM200 showed a number of differences in comparison to previously published sequences.36,37 The MM200 cDNA sequence has been submitted to Genbank (accession number AF088851). We found no similarity between

FIGURE 3 – Western blot analysis of H-ferritin expression in melanoma cells. Both 43 and 64 kDa forms of H-ferritin were identified in melanoma cell lines with MAb rH02 (a). Upon Gu-HCl/TCA precipitation of extracts from MM200, an additional band was identified at 21.5 kDa (b). An IgG1 isotype control MAb did not detect similar bands.

the 3⬘-untranslated region of the H-ferritin sequence with any other sequences when the Genbank databases were searched. There were 10 substitutions and 1 insertion within the pyrimidine-rich region in MM200 H-ferritin cDNA when compared with the human brain H-ferritin sequence.38 Detection of intracellular and secreted H-ferritin We immunoprecipitated supernatants from 5 melanoma lines using MAb rH02 against H-ferritin. As shown in Figure 2, a 21.5 kDa band was detected in all 5 melanoma supernatants. An isotype control MAb against TNP did not identify a similar band. Western blot analysis of lysates from a panel of melanomas using rH02 is shown in Figure 3a. Protein bands of 43 and 64 kDa

846

GRAY ET AL. TABLE I – EXPRESSION OF H- AND L-FERRITIN IN HUMAN MELANOMA CELLS Cell lines

% Inhibition1

H-ferritin % cells

L-ferritin % cells

Ratio of H/L-ferritin

MM200 Me10538 Me1007 Me4405 Mel-JG Lymphocytes MRC-5 HUVECs

22.3 13.3 13.3 9.4 ⫺6.2 ⫺2.5 ⫺9.6 ⫺14.2

58.42 (1.73)3 41.7 (1.37) 53.8 (1.67) 52.8 (1.70) 48.0 (1.72) 44.0 (1.53) 37.9 (1.50) 67.6 (1.94)

1.2 (1.01) 1.0 (1.01) 6.0 (1.06) 9.2 (1.08) 33.9 (1.43) 25.0 (1.21) 34.0 (1.42) 38.8 (1.37)

48.7 43 9 5.7 2.1 1.8 1.1 1.7

1 Percentage inhibition of anti-CD3–stimulated lymphocytes with supernatants collected after 24 hr from the different cell types.– 2 Measured by flow cytometry as % cells stained.–3Numbers in parentheses indicate median fluorescence intensity.

FIGURE 4 – Expression of H- and L-ferritin in Me1007 (a) and Mel-JG (b) cells by flow cytometry (solid line). Expression was detected in intact cells using MAb 10-F10 for H-ferritin and MAb A1-4E2 for L-ferritin. MAb TNP was used as an IgG1 isotype control (shaded area).

were detected in each melanoma line. An isotype control did not detect similar bands. Given that H-ferritin is of m.w. 21.5 kDa, we investigated whether the bands might represent multimers of the monomeric form. A lysate prepared from MM200 melanoma cells was therefore subjected to denaturation with 6 M Gu-HCl, 10% TCA precipitation and analysed by Western blotting. As shown in Figure 3b, this procedure resulted in the appearance of a 21.5 kDa band, suggesting that the higher m.w. bands were dimers and trimers of the monomeric form of H-ferritin. Eamination of H- and L-ferritin expression by flow cytometry Morikawa et al.21 suggested that the biological properties of ferritin might vary depending on the ratio of heavy and light chains. In view of this, H- and L-ferritin expression was examined in melanoma cells and other non-malignant cells by flow cytometry. The histograms shown in Figure 4 for Me1007 and Mel-JG indicate that the concentration of H-ferritin (shown by the shift to the right and a higher median fluorescent intensity) was much higher than that of L-ferritin. A representative study on the 5 melanoma lines is shown in Table I. H-ferritin was expressed in all 5 lines and 3 non-malignant cell types. There was, however, a marked difference in surface expression of L-ferritin between melanoma cell lines. For example, L-ferritin was expressed at low levels in MM200, Me10538, Me4405 and Me1007 (1.2%, 1%, 9.2% and 6% of cells, respectively) and at higher levels in Mel-JG (33.9% of cells). The expression of L-ferritin in Mel-JG was comparable to that in the non-malignant MRC-5 cells, HUVECs and lymphocytes (34%, 38.8% and 25%, respectively). The ratio between H- and L-ferritin expression was calculated by dividing the percentage of cells positive for H-ferritin by the percentage of cells positive for L-ferritin. Regression analyses were carried out to examine the correlation of the ratio between H- and L-ferritin with the percentage inhibition of the anti-CD3 response. There was a correlation between the H/L ratio (r ⫽ 0.61, p ⫽ 0.022) and the percentage inhibition of anti-CD3 response. The results in Table I indicate marked variation in the ratio between the lines.

FIGURE 5 – Titration of anti-CD3 inhibitory activity in melanoma supernatants. Cultures consisted of 2 ⫻ 105 PBLs, 50 ␩g anti-CD3 and serial 2-fold dilutions of supernatants from melanoma cell lines. The proliferative response was measured at 72 hr and percentage inhibition calculated as described in Material and Methods.

Suppressive effects of melanoma supernatants on anti-CD3–stimulated T-cell responses Supernatants were collected from 2 ⫻ 106 melanoma cells after 24 hr and tested for their effects on the mitogenic response to anti-CD3. As shown in Figure 5, there was a dose-dependent suppression of the anti-CD3 response. The titre of the inhibitory activity for MM200, Me10538, Me4405 and Me1007 ranged from a log2 dilution of 1 to 5. However, supernatant from Mel-JG did not suppress proliferation of anti-CD3–activated lymphocytes. As previously reported, MAb rH02, developed against human heart ferritin,39 inhibited the suppressive effect of rH-ferritin on colony formation of normal human granulocyte progenitor cells.40 Similar experiments were therefore carried out using rH02.28 Dose-titration experiments showed that rH02 at 500 ng/ml completely inhibited the suppressive effect of heart and rH-ferritin (data not shown). Figure 6 demonstrates that rH02 reversed the suppressive effects of melanoma supernatants from MM200 (by paired t-test p ⬍ 0.001), Me1007 (p ⫽ 0.013), Me10538 (p ⬍ 0.001) and Me4405 (p ⫽ 0.011). MAb rH02 also reversed the suppressive effects induced by rH-ferritin (p ⬍ 0.001) and heart ferritin (p ⫽ 0.0036). The enhancement of lymphocyte proliferation by Mel-JG (p ⫽ 0.225) and liver ferritin (p ⫽ 0.634) was not affected by rH02. The isotype control MAb TNP had no significant effect. Effects of H- and L-ferritin on anti-CD3 responses To investigate the possible link between the ratio of H- and L-ferritin to the level of suppression, ferritin from different sources was analyzed for the ability to suppress proliferation of anti-CD3– stimulated lymphocytes. As shown in Figure 7, the level of suppression was directly proportional to the concentration of rH-

H-FERRITIN IN HUMAN MELANOMA

847

FIGURE 6 – Reversal of the inhibitory activity of melanoma supernatants and other sources of ferritin by MAb rH02 against H-ferritin at a concentration of 500 ng/ml. The MAb was incubated with melanoma supernatants from melanoma cell lines for 1 hr prior to addition to the lymphocyte proliferation assay. MAb rH02 pre-incubated with rH-ferritin and ferritin from the human heart or liver was used as control. MAb TNP was used as an IgG1 isotype control. The proliferative response was measured at 72 hr and percentage inhibition calculated as described in Material and Methods. Error bars represent the SD for each data set.

FIGURE 7 – Doses responses of recombinant H- and L-ferritin in cultures of anti-CD3–stimulated lymphocytes. Cultures consisted of 2 ⫻ 105 lymphocytes, 50 ng anti-CD3 and serial 2-fold dilutions of rH-ferritin, rL-ferritin and ferritin from heart or liver. The proliferative response was measured at 72 hr and percentage inhibition calculated as describede in Material and Methods. Error bars represent the SD for each data set.

ferritin. In contrast, L-ferritin had a small stimulatory effect on anti-CD3–stimulated lymphocytes. Heart ferritin, which is higher in H-ferritin,21 increased the levels of suppression in a dosedependent manner, whereas liver ferritin, reported to be rich in L-ferritin,21 did not suppress lymphocyte proliferation. As shown in Figure 8, pre-incubation of increasing concentrations of rLferritin with a constant concentration of H-ferritin (45 ␮M) resulted in inhibition of the suppressive effects induced by H-ferritin.

At a molar ratio of approximately 1:1, the inhibitory effect of H-ferritin was reduced by 65%. Mechanism of anti-proliferative effects of H-ferritin on anti-CD3–stimulated lymphocytes The division of lymphocytes in response to anti-CD3 is dependent on signals from the CD3 complex and activation of the transcription factors AP-1, NF-AT and NF-␬B.41 One of the con-

848

GRAY ET AL.

sequences of this is activation of a number of cytokine genes involved in providing proliferation and differentiation signals, e.g., IL-2, IL-4 and IL-10, and production of co-factors, e.g., IFN-␥. We therefore examined the effects of melanoma supernatants as well as heart and liver ferritin on the production of these cytokines. A representative experiment, shown in Table II, indicates that the supernatants increased the production of IL-10 and IFN-␥ but inhibited that of IL-2 and IL-4. Heart ferritin had similar but more marked effects. The induction of IL-10 in anti-CD3–activated lymphocytes by H-ferritin was detectable by ELISA within 6 to 8 hr and peaked at 48 hr (data not shown). Regression analyses were carried out to examine the correlation of the percentage cytokine production with the percentage inhibition of anti-CD3 response. There was a correlation between cytokine production of IL-2 (r ⫽ 0.703, p ⫽ 0.009), IL-4 (r ⫽ 0.72, p ⫽ 0.007), IL-10 (r ⫽ 0.70, p ⫽ 0.009) and IFN-␥ (r ⫽ 0.68, p ⫽ 0.012) and the percentage inhibition of anti-CD3 response. The cytokine response by normal anti-CD3–activated lymphocytes was similar to levels described elsewhere.42 The specificity of the changes for H-ferritin was examined by addition of rH02 to the supernatants. As shown in Table II, the MAb reversed the effects of the supernatants on the increased production of IL-10 and IFN-␥ and the suppression of IL-2 and IL-4 production. The decline in suppression of IL-2 and IL-4 via

FIGURE 8 – Reversal of the inhibitory activity of H-ferritin by addition of increasing concentrations of L-ferritin. A serial 2-fold dilution of rL-ferritin was incubated with a constant concentration of H-ferritin (45 ␮M) for 1 hr prior to addition to the lymphocyte assay. The molar ratio was calculated by dividing the molar concentration of H-ferritin by the molar concentration of L-ferritin. Error bars represent the SD for each data set.

pre-treatment of anti-CD3–activated lymphocytes with MAb rH02 was significant, using a paired t-test. An isotype control MAb against TNP did not inhibit the effects on cytokine production. IL-10 has been previously shown to suppress lymphocyte proliferation in studies using a blocking MAb (JES3-9D7) directed against it.29 To demonstrate whether the induction of IL-10 in anti-CD3–stimulated lymphocytes by H-ferritin suppresses the anti-CD3 response, JES3-9D7 was incubated with rH-ferritin and heart ferritin in an anti-CD3–stimulated lymphocyte proliferation assay. Incubations with 10 ng rIL-10 and liver ferritin were used as positive and negative controls, respectively. Figure 9 demonstrates that MAb JES3-9D7 reversed the suppressive effects of rH-ferritin (by paired t-test p ⬍ 0.001), heart ferritin (p ⬍ 0.001) and rIL-10 (p ⬍ 0.001). The enhancement of lymphocyte proliferation by liver ferritin (p ⫽ 0.336) was not affected by JES3-9D7. The isotype control rat IgG had no significant effect. DISCUSSION

The above results indicate that H-ferritin is expressed by a wide range of cultured melanoma cells. Flow-cytometric studies indicated that it was present in the cell membrane and detectable in supernatants from melanoma cultures. Immunoprecipitation of the latter showed that it was present as a 21.5 kDa fraction, whereas Western blots on cell extracts suggested that it was present in 43 and 64 kDa fractions within the cell, which were most likely dimers and trimers of the 21.5 kDa form. Multimers form between recombinant H- and L-ferritin when unfolded chains are mixed and renatured.43 In melanoma cells, this multimer is presumably composed mainly of H-ferritin since melanoma cells appeared to express more H- than L-ferritin. Others have shown a species of H-ferritin called super-heavy-chain, or P43, with immunosuppressive properties in breast cancer.24,44 However, no known P43 sequence has ever been described, and it is possibly a dimer of H-ferritin. The H-ferritin cDNA sequence from the MM200 melanoma cell line appeared identical in the translated region to that from fetal human and human liver.36,37 However, there appeared to be unique variations in the 3⬘ end of the untranslated region. Ai and Chau38 demonstrated that there is a pyrimidine-rich region in the 3⬘ end between bases 753 and 769. This region presumably binds to a cytosolic protein, which shortens the half-life of the H-ferritin transcript. Mutations in this region increase the half-life of Hferritin mRNA.38 Therefore, we hypothesize that in MM200 mutations within this region may alter the effects of this cytosolic protein and stabilize the ferritin transcript, leading to over-expression of H-ferritin. Over-expression of H-ferritin in ovarian cancer also appears to result from transcriptional regulation.45 Elevated ferritin levels in the serum of patients with different malignancies, including melanoma, has been described; but the source of the

TABLE II – EFFECT OF MELANOMA SUPERNATANTS AND HEART OR LIVER FERRITIN ON CYTOKINE PRODUCTION FROM ANTI-CD3–STIMULATED LYMPHOCYTES

Culture conditions

PBL PBL PBL PBL PBL PBL PBL PBL

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

MM200 sup.3,4 Me10538 sup. Me1007 sup. Me4405 sup. Mel-JG sup. heart ferritin liver ferritin

% Inhibition1

IL-2 (pg/ml)

IL-2 pretreated with rH022 (pg/ml)

IL-4 (pg/ml)

IL-4 pretreated with rH02 (pg/ml)

IL-10 (pg/ml)

IL-10 pretreated with rH02 (pg/ml)

IFN-␥ (pg/ml)

IFN-␥ pretreated with rH02 (pg/ml)

⫺2.5 22.3 13.3 13.3 9.4 ⫺6.2 18.3 ⫺6.8

157 ⫾ 3 127 ⫾ 3 138 ⫾ 8 132 ⫾ 4 130 ⫾ 3 150 ⫾ 1 120 ⫾ 1 142 ⫾ 1

149 ⫾ 8 139 ⫾ 55 139 ⫾ 1 142 ⫾ 3 144 ⫾ 8 145 ⫾ 2 129 ⫾ 1 134 ⫾ 2

115 ⫾ 7 104 ⫾ 4 104 ⫾ 4 100 ⫾ 4 103 ⫾ 5 116 ⫾ 2 101 ⫾ 2 110 ⫾ 6

111 ⫾ 1 110 ⫾ 3 117 ⫾ 3 107 ⫾ 1 110 ⫾ 3 109 ⫾ 3 111 ⫾ 1 109 ⫾ 2

430 ⫾ 100 777 ⫾ 61 605 ⫾ 37 605 ⫾ 42 488 ⫾ 45 478 ⫾ 77 918 ⫾ 32 427 ⫾ 40

311 ⫾ 57 298 ⫾ 2 317 ⫾ 11 321 ⫾ 6 306 ⫾ 2 326 ⫾ 6 328 ⫾ 22 309 ⫾ 4

1,844 ⫾ 7 5,396 ⫾ 127 5,048 ⫾ 137 6,271 ⫾ 105 2,804 ⫾ 38 3,281 ⫾ 160 7,276 ⫾ 178 1,977 ⫾ 15

1,769 ⫾ 48 1,815 ⫾ 15 1,625 ⫾ 22 1,768 ⫾ 65 1,945 ⫾ 15 1,743 ⫾ 8 1,895 ⫾ 28 1,738 ⫾ 16

1 Percentage inhibition of anti-CD3–stimulated lymphocytes with ferritin (1.1 ␮g/ml) or supernatants from melanoma or lymphocytes.–2rH02 (500 ng/ml) was incubated for 1 hr with supernatant or ferritin from other sources prior to addition to lymphocytes.–3Supernatants were collected after 24 hr from melanoma cells.–4Anti-CD3–activated lymphocytes were incubated with melanoma supernatants or ferritin (1.1 ␮g/ml) from other sources for 24 hr prior to cytokine detection.–5Italicized results were significantly different (p ⬍ 0.05) from expression of IL-2 or IL-4 from anti-CD3–activated lymphocytes without pre-treatment with MAb rH02 (paired t-test).

849

H-FERRITIN IN HUMAN MELANOMA

FIGURE 9 – Reversal of the inhibitory activity of ferritin by MAb JES3-9D7 against IL-10 at a concentration of 10 ␮g/ml. MAb JES39D7 pre-incubated with rIL-10 was used as control. Rat IgG purified from serum was used as isotype control. The proliferative response was measured at 72 hr and percentage inhibition calculated as described in Material and Methods. Error bars represent the SD for each data set.

ferritin is unknown.22,46 – 48 It appears from the present studies that melanoma cells may be a source of elevated serum levels. Our studies support the view that H-ferritin has immunosuppressive properties in that it was shown to inhibit the response of lymphocytes stimulated with anti-CD3. Supernatants from melanoma cultures had similar effects, and this correlated with the ratios of H- and L-ferritin expression in the melanoma cells. The specificity of these effects for H-ferritin was shown by reversal of the inhibitory effects by MAb against H-ferritin. In contrast, Lferritin did not inhibit anti-CD3 responses but had weak potentiating effects. It was also shown that addition of rL-ferritin to rH-ferritin reduced the inhibitory effects of H-ferritin. These latter findings may account for the variation in the inhibitory effects of supernatants from different melanoma cultures in that it was shown that those without inhibitory activity (e.g., Mel-JG) had high levels of L-ferritin whereas those with low L-ferritin levels (e.g., MM200 and Me10538) had marked immunosuppressive effects. The ability of ferritin to inhibit lymphocyte responses to anti-CD3 therefore appears to be linked to the ratio of H- to

L-ferritin. Heart ferritin, rich in heavy chains, inhibits anti-CD3– activated lymphocytes, whereas liver ferritin, rich in light chains, enhanced anti-CD3–activated lymphocytes. The mechanism by which H-ferritin inhibits lymphocyte responses is largely unknown. Suggestions include signaling via specific receptors on lymphocytes for H-ferritin49,50 or downregulation of CD2,15 which acts as a co-factor for lymphocyte stimulation. In the present study, we examined the effects of H-ferritin on cytokine production and found small decreases in IL-2 and IL-4 but marked increases in IL-10 and IFN-␥. This cytokine profile is distinct from that of a normal TH0, TH1 or TH2 response. This suggests the involvement of a subset of CD4⫹ T cells with regulatory properties that suppress antigen-specific immune responses in vitro and in vivo.51 These T-regulatory 1 cells (Tr1) are defined by their unique profile of cytokine production and characteristically produce high levels of IL-10 and significant amounts of IFN-␥ but no IL-4 or IL-2.52 Kinetic studies have demonstrated that Tr1 clones produce IL-10 more rapidly than TH0, TH1 or TH2 cells.53 It is therefore possible that the effects on lymphocyte responses to anti-CD3 are due to IL-10 release by Tr1 cells. The role of IFN-␥ production by these cells in immunoregulation is not clear. IL-10 has been shown to inhibit IL-2 production as well as the proliferation of PHA-activated lymphocytes. Furthermore, IL-10 strongly reduces antigen-specific T-cell proliferation through inhibition of the antigen-presenting capacity of monocytes by downregulation of the expression of the co-stimulator molecule B7 and class II MHC expression.54 In the present studies, a role for IL-10 in the suppression of the anti-CD3 response by H-ferritin was supported by reduction of the suppressive effect by MAb JES39D7 against IL-10. IL-10 appeared to be induced rapidly in antiCD3–stimulated PBLs by H-ferritin and was detectable in supernatants of the cultures by 6 to 8 hr (and most probably at an earlier period in a cell-bound form). The kinetics of production is therefore consistent with IL-10 production as a mediator of the inhibitory effects of H-ferritin. Further studies are needed to confirm the nature of the cells producing IL-10 and the mechanism by which H-ferritin induces it production.

ACKNOWLEDGEMENTS

We thank Ms G. Frith for Mycoplasma PCR testing and Ms C. Cook for expert secretarial assistance.

REFERENCES

1.

Roth JA, Grimm EA, Gupta RK, Ames RS. Immunoregulatory factors derived from human tumors. J Immunol 1982;128:1955– 62. 2. Brotherick I, Shenton BK, Kirby JA, Rigg JM, Palmer JM, Yeaman SJ, et al. Production of immunosuppressive factors by a cultured tumour cell line and their effect on lymphocyte proliferation. Surg Oncol 1993;2:241– 8. 3. Bailet JW, Lichtenstein A, Chen G, Michel RA. Inhibition of lymphocyte function by head and neck carcinoma cell line soluble factors. Arch Otolaryngol Head Neck Surg 1997;123:855– 62. 4. Linnemeyer PA, Pollack SB. Prostaglandin E2 induced changes in the phenotype, morphology, and lytic activity of IL-2 activated natural killer cells. J Immunol 1993;150:3747–54. 5. Bodmer S, Strommer K, Frei K, Siepl C, De Tribolet N, Heid I, et al. Immunosuppression and transforming growth factor-␤ in glioblastoma. J Immunol 1989;143:3222–9. 6. Roszman T, Elliott L, Brooks W. Modulation of T-cell function by gliomas. Immunol Today 1991;12:370 – 4. 7. Matsuda M, Salazar F, Petersson M, Masucci G, Hansson J, Pisa P, et al. Interleukin 10 pretreatment protects target cells from tumor and allo-specific cytotoxic T cells and down regulates HLA class I expression. J Exp Med 1994;180:2371– 6. 8. Groux H, Bigler M, Devries JE, Roncarolo MG. Interleukin-10 induces a long-term antigen-specific anergic state in human CD4⫹ T cells. J Exp Med 1996;184:19 –29. 9. Tilg TS, Dinarello CA, Mier JW. IL-6 and APPs: anti-inflammatory and immunosuppressive mediators. Immunol Today 1997;18:428 –32. 10. Kono K, Salazar-Onfray F, Petersson M, Hansson J, Masucci G,

11. 12. 13. 14.

15.

16.

17.

Wasserman K, et al. Hydrogen peroxide secreted by tumor-derived macrophages down-modulates signal-transducing zeta molecules and inhibits tumor-specific T cell and natural killer cell-mediated cytotoxicity. Eur J Immunol 1996;26:1313. O’Connell J, O’Sullivan GC, Collins JK, Shanahan F. The Fas counterattack: Fas-mediated T cell killing by colon cancer cells expressing Fas ligand. J Exp Med 1996;184:1082. Matzner Y, Hershko C, Polliack A, Konjin AM, Izak G. Suppressive effect of ferritin on in vitro lymphocyte function. Br J Haematol 1979;42:345–53. Matzner Y, Konjin AM, Shlomai Z, Ben-Bassat H. Differential effect of isolated placental isoferritins on in vitro T-lymphocyte function. Br J Haematol 1985;59:443– 8. Rosen HR, Ausch C, Reinerova M, Zaspin E, Renner K, Rosen AC, et al. Activated lymphocytes from breast cancer patients express the characteristic of type 2 helper cells—a possible role for breast cancerassociated p43. Cancer Lett 1998;127:129 –34. Wigginton JM. Reversal of ferritin-mediated immunosuppression by levamisole: a rationale for its application to management of the acquired immune deficiency syndrome (AIDS). Med Hypotheses 1995;44:85– 8. Broxmeyer HE, Lu L, Bicknell DC, Williams DE, Cooper S, Levi S, et al. The influence of purified recombinant human heavy-subunit and light-subunit ferritins on colony formation in vitro by granulocytemacrophage and erythroid progenitor cells. Blood 1986;68:1257– 63. Hann H, Levy HM, Evans AE. Serum ferritin as a guide to therapy in neuroblastoma. Cancer Res 1980;40:1413.

850

GRAY ET AL.

18. Arosio P, Adelman TG, Drysdale JW. On ferritin heterogeneity: further evidence for heteropolymers. J Biol Chem 1979;253:4451– 8. 19. Theil EC. Ferritin: structure, gene regulation, and cellular function in animals, plants, and microorganisms. Annu Rev Biochem 1987;56: 285–315. 20. McGill JR, Naylor SL, Sakaguchi AY, Moore CM, Boyd D, Barrett K, et al. Human ferritin H and L sequences lie on ten different chromosomes. Hum Genet 1987;76:66 –72. 21. Morikawa K, Oseko F, Morikawa S. A role for ferritin in hematopoiesis and the immune system. Leukemia Lymphoma 1995;18:429 –33. 22. Hazard JT, Drysdale JW. Ferritinaemia in cancer. Nature 1977;265: 755– 6. 23. Constanzo F, Colombo M, Staempfli S, Santoro C, Marone M, Frank R, et al. Structure of gene and pseudogenes of human apoferritin H. Nucleic Acids Res 1986;14:721–36. 24. Shterman N, Kupfer B, Moroz C. Expression of messenger RNA species coding for a Mr 43,000 peptide associated with ferritin in human leukemia K562 cells and its down regulation during differentiation. Cancer Res 1989;49:5033– 6. 25. Zhang X, Franco A, Myers K, Gray C, Nguyen T, Hersey P. Relation of TRAIL receptor and FLIP expression to TRAIL induced apoptosis of melanoma. Cancer Res 1999;59:2747–53. 26. Curry BJ, Myers K, Hersey P. Polymerase chain reaction detection of melanoma cells in the circulation: relation to clinical stage, surgical treatment, and recurrence from melanoma. J Clin Oncol 1998;16: 1760 –9. 27. Jaffe EA, Nachman RL, Becker CG, Minick CR. Culture of human endothelial cells derived from umbilical veins: identification by morphologic and immunologic criteria. J Clin Invest 1973;52:2745–56. 28. Corsi B, Perrone F, Bourgeois M, Beaumont C, Panzeri MC, Cozzi A, et al. Transient over expression of human H- and L-ferritin in COS cells. J Biol Chem 1998;330:315–20. 29. Chen Q, Daniel V, Maher DW, Hersey P. Production of IL-10 by melanoma cells: examination of its role in immunosuppression mediated by melanoma. Int J Cancer 1994;56:755– 60. 30. Smith M, Bleijs R, Radford K, Hersey P. Immunogenicity of CD63 in a patient with melanoma. Melanoma Res 1997;7:S163–70. 31. Constanzo F, Santoro C, Colantuoni V, Bensi G, Raugei G, Romano V, et al. Cloning and sequencing of a full length cDNA coding for a human apoferritin H chain: evidence for a multigene family. EMBO J 1984;3:23–7. 32. Shuai K, Schindler C, Prezioso VR, Darnell JE Jr. Activation of transcription by IFN-␥: tyrosine phosphorylation of a 91-kD DNA binding protein. Science 1992;258:1808 –12. 33. Ausubel F, Brent R, Kingston RE, Moore DD, Seidman JG, Struhl K. Short protocols in molecular biology: a compendium of methods from current protocols in molecular biology. New York: John Wiley & Sons, 1995. 34. Sharrow CO. Analysis of flow cytometry data. In: Coligan JE, Kruisbeek AM, Margulies DH, Shevach EM, Strober W. Current protocols in immunology. New York: John Wiley & Sons, 1996. 35. Hersey P, Bindon C, Czerniecki M, Spurling A, Wass J, McCarthy WH. Inhibition of interleukin 2 production by factors released from tumor cells. J Immunol 1983;131:2837– 42. 36. Boyd D, Vecoli C, Belcher DM, Jain SK, Drysdale JW. Structural and functional relationships of human ferritin H and L chains deduced from cDNA clones. J Biol Chem 1985;260:11755– 61. 37. Dhar M, Chauthaiwale V, Joshi JG. Sequence of a cDNA encoding the ferritin H-chain from an 11-week-old human fetal brain. Gene 1993;126:275– 8. 38. Ai LS, Chau LY. Post-transcriptional regulation of H-ferritin mRNA. Identification of a pyrimidine-rich sequence in the 3⬘-untranslated

39.

40.

41. 42.

43. 44.

45. 46. 47. 48. 49.

50.

51. 52. 53.

54.

55.

region associated with message stability in human monocytic thp-1 cells. J Biol Chem 1999;274:30209 –14. Cavanna F, Ruggeri G, Iacobello C, Chieregatti G, Murador E, Albertini A, et al. Development of a monoclonal antibody against human heart ferritin and its application in an immunoradiometric assay. Clin Chim Acta 1983;134:347–56. Dezza L, Cazzola M, Bergamaschi G, Stella CC, Pedrazzoli P, Recalde HR. Effects of recombinant human H-subunit and L-subunit ferritins on in vitro growth of human granulocytes-monocyte progenitors. Br J Haematol 1988;68:367–72. Schraven B, Marie-Cardine A, H¨ubener C, Bruyns E, Ding I. Integration of receptor-mediated signals in T cells by transmembrane adaptor proteins. Immunol Today 1999;20:431– 4. Goto S, Sato M, Kaneko R, Itoh M, Sato S, Takeuchi S. Analysis of TH1 and TH2 cytokine production by peripheral blood mononuclear cells as a parameter of immunological dysfunction in advanced cancer patients. Cancer Immunol Immunother 1999;48:435– 42. Santambrogio P, Levi S, Cozzi A, Rovidia E, Albertini A, Arosio P. Production and characterization of recombinant heteropolymers of human ferritin H and L chains. J Biol Chem 1993;268:12744 – 8. Moroz C, Shterman N, Kupfer B, Ginzburg I. T-cell mitogenesis stimulates the synthesis of a mRNA species coding for a 43-kDa peptide reactive with CH-H-9, a monoclonal antibody specific for placental isoferritin. Proc Natl Acad Sci USA 1989;86:3282–5. Tripathi PK, Chatterjee SK. Elevated expression of ferritin H-chain mRNA in metastatic ovarian tumor. Cancer Invest 1996;14:518 –26. Luger TA, Linkesch W, Knobler R, Kokoschka EM. Serial determination of serum ferritin levels in patients with malignant melanoma. Oncology 1983;40:263–7. Iancu TC, Shiloh H, Kedar A. Neuroblastomas contain iron-rich ferritin. Cancer 1988;61:2497–502. Partin AW, Criley SR, Steiner MS, Hsieh K, Simons JW, Lumadue J, et al. Serum ferritin as a clinical marker for renal cell carcinoma: influence of tumor volume. Urology 1995;45:211–7. Fargion S, Fracanzani AL, Brando B, Arosio P, Fiorelli G. Specific binding sites for H-ferritin on human lymphocytes: modulation during cellular proliferation and potential implication in cell growth control. Blood 1991;78:1056 – 61. Moss D, Fargion S, Fracanzani AL, Levi S, Cappellini MD, Arosio P, et al. Functional roles of the ferritin receptors of human liver, hepatoma, lymphoid and erythroid cells. J Inorg Biochem 1992;47:219 – 27. Levings MK, Roncarolo MG. T-regulatory 1 cells: a novel subset of CD4⫹ T cells with immunoregulatory properties. Allergy Clin Immunol 2000;106:S109 –12. Groux H, O’Garra A, Bigler M, Rouleau M, Antonenko S, De Vries J, et al. A CD4⫹ T-cell subset inhibits antigen-specific T-cell responses and prevents colitis. Nature 1997;389:737– 42. Cavani A, Nasorri F, Prezzi C, Sebastiani S, Albanesi C, Girolomoni G. Human CD4⫹ T lymphocytes with remarkable regulatory functions on dendritic cells and nickel-specific TH1 immune responses. J Invest Dermatol 2000;14:295–302. De Waal Malefyt R, Haanen J, Spits H, Roncarolo MG, Velde A, Figdor C, et al. Interleukin 10 (IL-10) and viral IL-10 strongly reduce antigen-specific human T cell proliferation by diminishing the antigen-presenting capacity of monocytes via down regulation of class II major histocompatibility complex expression. J Exp Med 1991;174: 915–24. Needleman SB, Wunsch CD. A general method applicable to the search for similarities in the amino acid sequence of two proteins. J Mol Biol 1970;44:443–53.