Interferon α2b Decreases Hepatic Micrometastasis in a Murine Model of Ocular Melanoma by Activation of Intrinsic Hepatic Natural Killer Cells

Abstract

purpose. To investigate in a murine model the mechanism by which micrometastatic melanoma, which spreads from the eye to the liver, is controlled by interferon (IFN)-α2b.

methods. Major histocompatibility (MHC) class I antigen (H-2, all haplotypes) expression in three murine melanoma cell lines (Queens, B16LS9, B16F10) was determined by flow cytometric immunophenotyping. The cell lines were heterotopically inoculated into the posterior compartments (PCs) of C57Bl/6 mice, and the mice were given intraperitoneal (IP) injections of IFN-α2b or PBS for 1 or 4 days before enucleation at 7 days after inoculation. Groups of mice were made NK deficient or depleted with subcutaneous (SC) injection of anti-asialo GM1. The mice were killed at 28 days or 56 days (survival experiment) after inoculation, and the number of hepatic micrometastases was histologically determined. NK cells were isolated from the spleen and liver at necropsy, and propidium iodide labeled target-specific cytolysis was determined by flow cytometry. The micrometastases were evaluated for apoptosis and proliferation with TUNEL and MIB1 immunostaining, respectively, and TUNEL-to-MIB1 ratios were determined. Hepatic NK cells were immunostained with CD49b.

results. MHC class I antigen was expressed in the three cell lines in the order of Queens < B16LS9 < B16F10. All cell lines grew, were confined to the PC, and formed hepatic micrometastases. A decrease in micrometastases, an increase in target-specific cytolysis, and an increase in survival correlated with decreased HLA class I expression by the melanoma cells. The IFN-α2b treatment resulted in a boost of intrinsic hepatic NK cells, demonstrated in NK-deficient but not NK-depleted mice. The treatment effect corresponded to increased apoptosis (TUNEL)-proliferation (MIB1) ratios in the micrometastases. Immunostaining demonstrated an increased number of intrahepatic NK cells associated with the micrometastases in treated groups.

conclusions. Neoadjuvant IFN-α2b results in decreased hepatic micrometastasis and increased survival time through increased intrinsic hepatic NK cell–mediated tumor apoptosis in a murine model of metastatic ocular melanoma.

There is currently no effective treatment for patients with metastatic uveal melanoma, ¹ which almost exclusively metastasizes to the liver. ^{1 2} At the time of initial diagnosis and treatment of uveal melanoma, most patients have no demonstrable evidence of metastatic disease. ² However, within 5 years, metastases appear in approximately 20% to 35% of patients, with a mean relapse-free interval of 2 to 4 years. ^{2 3} Calculations using estimations of tumor-doubling times have shown that hepatic micrometastases of uveal melanoma occur 2.9 to 5.2 years before the diagnosis of the primary intraocular malignancy. ^{4 5} There is clinical evidence that this is the case, as enucleation versus brachytherapy for medium-sized uveal melanoma results in the same mortality outcomes. ⁶ An evolving concept is that hepatic micrometastases, which are often present at the time of diagnosis of the primary intraocular melanoma, remain dormant, with the potential to grow. One factor that may allow for growth of dormant hepatic micrometastases is the immune status of the host. ⁷  

We have developed a murine model of uveal melanoma that forms hepatic micrometastases. ^{8 9 10} We found in our model that depletion of natural killer (NK) cells results in growth of the micrometastases. ¹¹ In addition, a preliminary study showed that neoadjuvant interferon (IFN)-α2b resulted in decreased micrometastases in our model. ¹² We designed a set of experiments to test the mechanism by which neoadjuvant IFN decreases hepatic micrometastasis. 

Materials and Methods

Mice

Twelve-week-old female C57BL/6 mice were obtained from Jackson Laboratories (Bar Harbor, ME). All experiments were conducted according to the Declaration of Helsinki and Guiding Principles in the Care and Use of Animals and conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Tumors

B16F10 cells (courtesy of Jerry Niederkorn, UT Southwestern, Dallas, TX) are a line of B16 melanoma cells that were obtained after isolation of metastases and serial passages in tissue culture. ¹³ Queens melanoma cells (courtesy of Jerry Niederkorn) were derived from serial passages of B16F10 melanoma cells. ¹⁴ B16LS9 cells (courtesy of Dario Rusciano, Friedrich Miescher Institut, Basel, Switzerland) express high levels of c-Met and metastasize to the liver. ¹⁵ All three cell lines are of cutaneous melanoma origin. All frozen cell lines were thawed and resuspended in 15 mL of minimum essential medium (MEM) supplemented with fetal calf serum, l-glutamine, and sodium bicarbonate. The cell suspension was centrifuged and the pellet was washed and resuspended in 15 mL of supplemental MEM. The suspension was placed in a 75-cm² tissue culture flask (T-75; BD Biosciences, Franklin Lakes, NJ) in a carbon-dioxide incubator (Kendro, Asheville, NC) at 37°C and grown to confluence in 3 to 5 days. The cells were trypsinized, aliquoted, and washed three times in 5 mL Hanks’ balanced salt solution. An aliquot of 10 μL of suspension was placed in a hemocytometer (AO, Buffalo, NY) to calculate the concentration of melanoma cells. 

Inoculation of Melanoma Cells into the Posterior Compartment

Aliquots of 5 × 10⁵cells/2.5 μL were inoculated into the posterior compartment (PC) of the right eyes using a transcorneal technique that allows the inoculated cells to remain in the eye. ⁹ Briefly, the mice were anesthetized and a tunnel was prepared from the limbus within the cornea, sclera, and ciliary body to the choroid with a 30-gauge needle under the guidance of a dissection microscope. The tip of a 10-μL glass syringe with a blunt metal needle (Hamilton, Reno, NV) was introduced into the PC through the needle track, and no cells were inoculated until the needle tip was inside the eye. A 5.0-μL suspension of cells was inoculated. No tumor cell reflux occurred, and the subconjunctival space remained free of tumor cells. The right eyes were enucleated 7 days after inoculation. ¹⁶ For each experiment, there were 15 mice in each group and each experiment was repeated. 

Interferon

Recombinant human IFN-α2b (Intron A; Schering-Plough, Kenilworth, NJ) was used. The mice were given 20,000 IU (20KIU) IFN-α2b per intraperitoneal (IP) injection. 

Treated and Control Groups

There were three groups (4-day IFN-α2b and 1-day IFN-α2b [control]) for each cell line (B16F10, B16LS9, and Queens). The 4-day treatment group received an IP injection of 20KIU IFN-α2b each day for 4 days before enucleation, and the 1-day treatment group received an IP injection 1 day before enucleation. The control groups received an equal volume of IP PBS every day for 4 days before enucleation. Mice were killed at 28 days after inoculation or observed for survival up to 56 days after inoculation. 

Anti-asialo GM1 NK Deficiency and Depletion

Two groups (4-day IFN-α2b/anti-asialo GM1 and 4-day PBS/anti-asialo GM1) were used for three cell lines (B16F10, B16LS9, and Queens). Methods for IP injection and PBS were as described earlier. Anti-asialo GM1 (Rabbit; Wako Pure Chemical Industries, Osaka, Japan) was diluted to 1:4 in distilled water. For NK deficiency, the groups were injected subcutaneously (SC) with 100 μL anti-asialo GM1 every 3 days, starting 2 weeks before enucleation and continuing until the mice were killed. For NK depletion, Queens cells only were used in a separate experiment and 100 μL anti-asialo GM1 was injected SC every day starting 2 weeks before enucleation and continuing until the mice were killed. 

Assay for Micrometastasis

The eyes were routinely processed for light microscopic examination. Serial 5-μm-thick sections were stained with hematoxylin-eosin and evaluated for the presence and location of the melanoma (Olympus BX41; Olympus, Tokyo, Japan). Only mice with melanoma confined to the eye were used in the study. Less than 5% of the eyes had melanoma growing outside the eye. The mice were euthanatized 28 days after inoculation (21 days after enucleation), and necropsies were performed. The livers were grossly examined, submerged in 4% neutral buffered formaldehyde, and processed for light microscopic examination. Three sections through the center of each liver were microscopically evaluated (Olympus) for the presence or absence of micrometastases (<100 μm in diameter), and the average number of micrometastases per section was determined, as previously described. ^{8 11 12}  

MHC-I Flow Cytometry

Major histocompatibility complex (MHC) class I antigen expression for the three melanoma cell lines (B16F10, B16LS9, and Queens) and host natural killer (NK) cell activity were assessed with a cytofluorometer (FACSCAN; BD Biosciences, San Jose, CA). Negative control cells were spleen lymphocytes from normal C57BL/6 mice. For MHC class I expression, rat myeloma monoclonal antibody directed against mouse H-2 (all haplotypes; IgG2a, clone:M1/42.3.9.8; courtesy of Judy Kapp, Emory University, Atlanta, GA) served as primary antibodies. Purified rat myeloma IgG2a (Zymed, San Francisco, CA) served as an isotype-matched negative control antibody. Single-cell suspensions were prepared and washed in buffer (FACS; BD Biosciences) consisting of phosphate-buffered saline (PBS, pH 7.4) with 1% BSA and 0.02% sodium azide. Cells (5 × 10⁵) were incubated with primary antibodies for 30 minutes on ice, washed three times, and incubated with fluorescein-conjugated goat affinity-purified F(ab″)2 fragment to rat IgG (ICN, Aurora, OH) for 30 minutes on ice. Cell suspensions were fixed in 1% paraformaldehyde for analysis. 

NK Cell Isolation and Flow Cytometry

Isolating lymphocytes from the liver and spleen was performed by exposing and catheterizing the hepatic portal vein with a 24-gauge, 0.75-in. needle. Prewarmed RPMI 1640+10% PBS (8–10 mL) was slowly injected to flush blood from the hepatic vasculature. The liver and spleen were excised and homogenized with glass slides. Cells were gently forced through a sterile 70-μm nylon cell strainer (catalog no. 352350; BD Biosciences) with a sterile syringe and placed on ice. The extract was centrifuged at 600g for 10 minutes, and the supernatant was removed. The cells were resuspended with 1 mL PBS. Viable lymphocytes were isolated from the liver cell suspension by density separation. Polysucrose 400 and sodium diatrizoate (5 mL; Lympholyte-M; Cedarlane, Hornby, Ontario, Canada) was added to a 15-mL centrifuge tube. Using a pipette, the cell suspension was carefully layered over the Lympholyte-M with as little mixing as possible at the interface. The gradient was centrifuged at 1500g for 20 minutes at room temperature. The second layer of cells was carefully removed to a new centrifuge tube. The isolated cells were diluted with medium and centrifuged at 800g for 10 minutes, and the supernatant was discarded. The cells were washed two to three times in the medium, counted, and resuspended to 2 × 10⁶/100 μL in staining buffer. 

For flow cytometry, the following antibodies and isotypes were used: FITC anti-mouse NK1.1(NKR-P1C, LY-55; PK136), PE anti-mouse pan-NK cells; isotype-FITC rat IgG2α, λ isotype control (anti-KLH), rat IgM, κ isotype control, and Fc block-purified anti-mouse CD16/CD34 (FCγ III/II receptor; 2.4G2; all from BD Pharmingen, San Diego, CA). The antibodies and isotypes were diluted according to the manufacturer’s directions. Then, 2 × 10⁶ lymphocytes/100 μL cells were isolated from the spleens and livers of PC mouse models using B16F10, Queens, and B16LS9 melanoma cell lines, and the cells were placed in a 96-well plate. Dilute Fc (100 μL) was placed in each well and incubated at 4°C for 15 minutes in the dark and washed once. The cells were stained in 50 μL staining buffer with anti-mouse NK1.1 and anti-mouse pan-NK incubated with an isotype control reagent for 30 minutes at 4°C in the dark. The cells were then washed three times with staining buffer, fixed with 1% paraformaldehyde, and analyzed by flow cytometry (BD PharMingen). 

Cytotoxicity Assay

The target cells used were B16LS9, B16F10, Queens, and YAC-1. There were 15 mice tested in each group. Various numbers of effector cells from the spleen or liver were added to the target cells at effector-target (E:T) ratios ranging from 25:1, 50:1, to 100:1 to a final volume of 500 μL of RPMI1650+10%FCS+1 μg/mL propidium iodide (BD PharMingen). Spontaneous lysis was measured with only target cells and no effector cells. The cells/RPMI/PI were incubated for 2 hours at 37°C, 5%CO₂, washed three times with staining buffer, and analyzed by flow cytometry (BD PharMingen). 

Immunohistochemical Staining

TUNEL Analysis.

Apoptosis in liver sections was detected by TUNEL analysis. Cryopreserved or paraffin-embedded liver sections were routinely processed, according to the protocol of the in situ cell death detection kit (Roche, Mannheim, Germany). 

MIB1: An Anti-Ki67 Protein Expressed during the Cell Cycle.

Paraffin-embedded liver sections were cut (4 μm) and mounted on slides (ProbeOn; Fisher Scientific, Pittsburgh, PA). The sections were dried, deparaffinized, and hydrated with xylene and graded ethanol. They were then washed with Tris-buffered saline with 0.05% Tween 20 (TBST). Heated slides were placed in target retrieval solution at 95°C for 20 minutes. Endogenous peroxidase activity was blocked with 3% H₂O₂ in water. The slides were washed with TBST, incubated with biotin and avidin (Dako, Carpinteria, CA), washed with TBST, and incubated for 1 hour at room temperature with anti-mouse Ki67 Clone TEC-3 (Dako A/S, Glostrup, Denmark) 1:50 with Tris-HCl. The slides were washed three times, incubated with 1:200 secondary biotinylated rabbit anti-rat immunoglobulin (DakoA/S) for 30 minutes at room temperature, and washed three times. The slides were submerged in dilute streptavidin (DAKO) 1:300 and incubated for 30 minutes. They were washed three times, incubated with 3-amino-9-ethylcarbazole (AEC; Sigma-Aldrich, St. Louis, MO) for 5 to 10 minutes, and counterstained with hematoxylin, mounted in aqueous medium, and coverslipped. 

Apoptosis-Proliferation Rates.

The TUNEL (apoptosis) and MIB1 (proliferation) rates were determined for each group by counting the number and determining the percentage of TUNEL- or MIB1-positive cells in micrometastases in adjacent serial sections of the livers. 

NK Cells.

Frozen sections were cut at 5 μm and mounted on microscope slides (ProbeOn Plus; Fisher Scientific) and heated at 37°C overnight. The sections were fixed in acetone, washed in TBST, and incubated with 3% H₂O₂ for 10 minutes. Biotin and avidin (Dako) were applied to the slides. The slides were washed in TBST and blocked with normal rabbit serum for 30 minutes, and 1:25 primary antibody (anti-mouse CD49b; eBioscience, San Diego, CA) was applied. The slides were incubated for 1 hour at room temperature and washed three times, and 1:200 biotin-conjugated anti-Armenian hamster IgG (eBioscience) was applied. The slides were incubated for 30 minutes and washed in TBST. Dilute streptavidin (Dako) 1:300 was applied, and the slides were incubated for 30 minutes then washed in TBST. 3,3′-Diaminobenzidine (DAB; Dako) was added, and the slides were incubated for 2 minutes and counterstained with hematoxylin for 1 minute. The slides were dehydrated with graded alcohol, cleared with xylene, and coverslipped. 

Statistical Analysis

The Wilcoxon Mann-Whitney test was used to determine whether the distribution of number and size of metastatic melanoma differed among combinations of the groups of mice. The number of hepatic micrometastases was compared between the control versus 1-day IFN-α2b treatment versus 4-day IFN-α2b treatment groups. The number of apoptosis (TUNEL⁺) and proliferating (MIB1⁺) cells of the treated and control groups were compared by χ² test. 

Results

Intraocular melanoma grew and was confined to the eye in 95% of each group for each cell line (Fig. 1) . There was no difference in the histologic appearance of the melanoma in the eye after 1 or 4 days of treatment with IFN-α2b. Flow cytometry showed the lowest level of MHC class I expression in the Queens cell line, followed by B16LS9 and B16F10 cells (Fig. 2) . The number of hepatic micrometastases in the treated versus control groups is shown in Table 1 . There was no statistically significant difference among the number of micrometastases in the control groups. There was a decrease in the number of hepatic micrometastases with IFN-α2b treatment for all three cell lines, with the greatest decrease in Queens followed by B16LS9 then B16F10 cells (Fig. 3) . The decrease was statistically significant for all melanoma cell lines in the IFN-α2b 4-day–treated groups (P < 0.01) and the Queens and B16LS9 cell lines in the IFN-α2b 1-day treatment group. Flow cytometry showed the highest level of NK-specific lysis in the Queens melanoma cell line treated with 4- or 1-day IFN-α2b, followed by B16LS9 and B16F10 (Fig. 4) . This reached statistical significance for the Queens melanoma cell line (P < 0.01) and corresponded with the in vitro finding of the largest decrease in micrometastasis in the Queen cell line treatment groups, followed by B16LS9 and B16F10 (Fig. 3) . The probabilities for decrease in number of micrometastasis and percentage of NK lysis in the 4-day IFN-α2b–treated versus control groups were smallest for Queens (0.0051, 0.0083) followed by B16LS9 (0.0093, 0.034) and B16F10 (0.0094, 0.034) cells. Absolute numbers of micrometastases cannot be compared among cell lines, as single lines were used for separate experiments. The experiments were designed to determine differences in absolute numbers in treated versus control groups for a given cell line, not to compare the absolute number of micrometastases among cell lines. Flow cytometry showed increased hepatic NK activity in the 4-day IFN-treated group for all melanoma cell lines (Table 2) . It should be noted that the cell lines tested were of murine cutaneous melanoma origin. Uveal melanoma may or may not yield similar results. In addition, IFN-α2b may upregulate MHC class I antigen. However, there is in vitro evidence that MHC class I antigen is downregulated after IFN-α administration. ¹⁷  

A smaller IFN-α2b treatment effect was also observed in the anti-asialo GM1 systemically pretreated, NK-deficient group (Table 3) . Flow cytometry showed decreased splenic NK cells after anti-asialo GM1 in the NK-deficient groups, although hepatic NK cells remained intact and increased with IFN-α2b treatment (Table 4) . There was a loss of hepatic NK cells and no difference in micrometastases in the NK-depleted versus control mice (Table 5) . Immunohistochemical stains demonstrated increased intrahepatic NK cells associated with micrometastasis in 4-day IFN-α2b–treated versus control mice in most of the micrometastases (Fig. 5) The micrometastases had an increased apoptosis (TUNEL)-to-proliferation (MIB1) ratio (Figs. 6 7) in hepatic micrometastatic melanoma in the 4- and 1-day IFN-treated groups. The apoptosis rates were significantly higher and the proliferation rates were significantly lower (P < 0.01) in all three cell lines in both the 4- and 1-day IFN-treated groups. There was significantly longer survival in days in the 4-day IFN-α2b–treated groups (Queens, B16LS9, and B16F10 groups survived 47.23, 45.71, and 44.5 days, respectively) versus the control groups (Queens, B16LS9, and B16F10 groups survived 21.93, 23.93, and 25.13 days, respectively, P < 0.01; Table 6 ) in all cell lines and in the 1-day IFN-α2b–treated Queens cell line. Necropsies showed hepatic metastases (≥100 μm) in all mice that did not survive. 

Discussion

There is mounting evidence that NK cells are involved in the elimination of human uveal melanoma metastases to the liver. NK cells eliminate cells that do not express HLA class I antigen. ¹⁸ Immunohistochemical and population-based studies have shown that patients with uveal melanomas that do not express HLA class I antigen have a better prognosis than those with uveal melanomas that do. ¹⁹ A large percentage of uveal melanomas exhibit a loss of classic HLA class I antigens, allowing for escape from the immunosurveillance of cytotoxic T lymphocytes. ²⁰ HLA class I antigen expression correlates with aggressive uveal melanoma cell types. ^{21 22} Metastatic (≥100 μm) uveal melanoma that grows from micrometastatic (<100 μm) melanoma to the liver expresses HLA class I antigen, thus allowing for escape from NK cells. ²² Shedding of uveal melanoma cells with low expression of HLA class I antigen into the systemic circulation may facilitate their removal by NK cells and prevent development of micrometastases. ¹⁹ It appears that NK cells eliminate a subset of micrometastases (those with low HLA class I expression) within the liver. 

There is experimental evidence, including in vitro and in vivo studies of murine melanoma, that NK cells’ killing of metastatic uveal melanoma to the liver is IL-2 and TGFβ dependent. ^{23 24} Our laboratory ¹¹ and others ²⁵ have shown that anti-asialo GM1 elimination of NK cells results in an increased number and growth of hepatic micrometastases in murine models using B16-derived melanoma cell lines. In addition, NK cell stimulation with maleic anhydride divinyl ether results in decreased hepatic metastases of experimental murine melanoma. ²⁶  

IFN-α2b is a dimer formed of 166 aa monomers. Each monomer contains five α helices, three 3/10 helices and eight loops. IFN-α2b binds with the cell surface receptor IFN-α/βR and forms a signal complex resulting in antiviral, antiproliferative, and immunomodulatory effects. IFN-α2b enhances NK cell activity in mice. ^{26 27} Injection of mice with IFN-α2b induces results in increased splenic NK cell cytotoxicity against tumor cells. ^{26 27} IP injection of human recombinant IFN decreases the number of pulmonary metastases of IV-injected B16 melanoma cells in mice. ²⁸ IFN-α2b also has antiangiogenic properties ²⁹ by downregulation of bFGF in tumor cells, resulting in decreased blood vessel density. ³⁰ This effect is less important than other effects in our model, as the micrometastases have not developed their own vascular supply ¹¹ and maintain a state of equilibrium (dormancy) where apoptosis equals proliferation. ³¹ The proposed mechanism of action of IFN-α2b in our model is enhanced NK melanoma killing through a series of autocrine and paracrine loops (Fig. 8) . In our experiments, all three cell lines responded to IFN-α2b therapy in vivo. It is possible that intrinsic hepatic NK cell lysis rates of micrometastases, rather than MHC class I expression by the melanoma cell lines, was the most important factor related to efficacy of IFN-α2b therapy. The importance of intrinsic NK cells for killing micrometastatic melanoma in our model was shown. NK-deficient mice had a limited, but present, intrinsic hepatic NK anti-melanoma effect (Table 4) , whereas NK depleted mice had no anti-melanoma effect (Table 5) . 

Our study confirms preliminary evidence that neoadjuvant IFN-α2b decreases hepatic micrometastases from experimental murine uveal melanoma. ¹² Flow cytometric analysis showed that the increased NK cells in IFN-α2b–treated versus control groups is a cell-line dependent phenomenon, with decreased MHC class I expression associated with increased efficacy of IFN-α2b treatment. The percent decrease in micrometastasis, increased NK activity, and increased survival time was greatest in mice inoculated with Queens (low MHC class I) followed by B16LS9 (moderate MHC class I) and B16F10 (high MHC class I) cell lines. Although there may have been a cytotoxic T-cell (CTL) response to MHC class I, which is upregulated by IFN-α2b, this is probably mechanistically less important, in that there is in vitro evidence that uveal melanoma HLA class I expression is downregulated after administration of IFNα. ³²  

We found that there continues to be a treatment effect after systemic anti-asialo GM1. Flow cytometry of circulating and intrahepatic NK cells showed increased intrahepatic NK cells in the IFN-α2b treated mice, even after anti-asialo GM1, indicating that intrahepatic NK cells are involved with in vivo killing of micrometastases. Immunohistochemical staining demonstrated and increased number of NK cells associated with hepatic micrometastasis and increased micrometastasis apoptosis-proliferation after IFN-α2b treatment. Our studies show in our murine model that systemic IFN-α2b decreases hepatic micrometastasis from intraocular melanoma by increasing hepatic NK activity and increasing apoptosis in micrometastatic melanoma. This has therapeutic implications when considering combinations of neoadjuvant and adjuvant therapies for patients with primary uveal melanoma. 

 Supported by National Eye Institute Grant R01 EY13165.

 Submitted for publication December 10, 2003; revised March 3 and 15, 2004; accepted March 19, 2004.

 Disclosure: H. Yang, None; S. Dithmar, None; H.E. Grossniklaus, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked ”advertisement“ in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

PC melanoma model. Heterotopically transplanted melanoma grows in the choroid and subretinal space (A, B) of the PC model and is capable of forming hepatic micrometastases (C). The intraocular tumor is vascularized and metastasizes hematogenously. Hematoxylin and eosin; magnification: (A) ×10, (B) ×25, (C) ×100.

Figure 2.

 

Flow cytometry to determine MHC-I antigen expression in melanoma cell lines, showing mean fluorescence intensity for MHC-1 expression in Queens < B16LS9 < B16F10 (blue) compared with negative control (red). The mean fluorescence intensity for MHC class I for Queens, B16LS9, and B16F10 cells is 0.38, 1.59, and 6.37, respectively (rat anti-mouse H-2, all haplotypes). The x-axis shows side scatter, and the y- axis shows number of cells.

Table 1.

 

Number of Hepatic Micrometastases in Uveal Melanoma Mice, with or without IFNα2b Treatment

Group	Cell Lines
		Queens	B16LS9	B16F10
IFN 4 days	1.84 ± 0.26*	2.50 ± 1.45*	2.40 ± 0.52*
IFN 1 day	2.12 ± 0.81*	3.22 ± 0.72, †	4.11 ± 1.01
PBS	8.08 ± 1.19	7.15 ± 0.98	6.68 ± 1.03

 IFN 4 days, IFNα2b IP, injection for 4 days prior to enucleation; IFN 1 day, IFNα2b IP injection for 1 day prior to enucleation.

*  Significant difference from PBS control group in the number of hepatic micrometastases (P < 0.01; Wilcoxon Mann-Whitney test).

†  Difference from PBS control group in the number of hepatic micrometastases (P < 0.05; Wilcoxon Mann-Whitney Test).

Figure 3.

 

Number of micrometastasis in IFNα2b-treated and control (PBS) groups. IFNα2b treatment for 4 days resulted in significantly fewer micrometastases for all melanoma cell lines compared with the control (P < 0.001). There were significantly fewer micrometastases after IFN 1-day treatment, only with the Queens cell line (P < 0.001). IFN 4d, IFNα2b IP for 4 days before enucleation; IFN 1d, IFNα2b IP for 1 day before enucleation.

Figure 4.

 

Percentage of splenic NK melanoma–specific cell lysis in IFNα2b and control groups. There is significantly higher tumor lysis for Queens melanoma (IFN 1 and 4 day treatments) than the control (P < 0.001). This did not reach statistical significance for the other two cell lines. Definition of key is provided in Figure 3 .

Table 2.

 

Number of Spleen and Liver CD49b+NK1.1+ Cells in 10,000 Spleen or Hepatic Lymphocytes in Uveal Melanoma Mice, with or without IFNα2b

Groups	The Ratio of Spleen and Liver CD49b+NK1.1+
		Queens			B16LS9			B16F10
		Spleen	Liver	Spleen	Liver	Spleen	Liver
IFN 4 days	5.89*	15.30, †	5.26	10.30*	3.43	8.70*
PBS	3.90	6.02	3.71	4.23	1.78	3.85

 Data are the ratio (%) of spleen and CD49b+NK1.1+ cells. IFN 4 days is defined in Table 1 .

*  Difference between PBS and IFN groups in the ratio of hepatic NK cells to hepatic lymphocytes (P < 0.05; chi-square test).

†  Significant difference from the PBS group in the ratio of hepatic NK cells to hepatic lymphocytes (P < 0.01; chi-square test).

Table 3.

 

Number of Hepatic Micrometastases in the IFNα2b–Treated versus Control Uveal Melanoma Mice with Anti-asialo GM1 NK Deficiency

Group	Cell Lines
		Queens	P	B16LS9	P	B16F10Q	P
IFN 4 days	5.78 ± 1.40*	0.0004	5.50 ± 2.43*	0.0073	6.25 ± 2.42, †	0.025
IFN 1 day	6.3 ± 1.35*	0.0026	8.42 ± 2.69, †	0.03	8.57 ± 5.81	0.15
PBS	14.91 ± 2.06		14.07 ± 3.08		17.83 ± 7.45

 IFN 4 days and 1 day as defined in Table 1 .

*  Significant difference from PBS control group in the number of hepatic micrometastases P < 0.01; Wilcoxon Mann-Whitney test).

†  Difference from PBS control group in the number of hepatic micrometastases (P < 0.05; Wilcoxon Mann-Whitney test).

Table 4.

 

Percentages of Spleen and Liver CD49b+NK1.1+ Cells in 10,000 Spleen or Hepatic Lymphocytes in PC Melanoma Mice, with or without IFNα2b Combined with Low-Dose Anti-asialo GM1 (NK Deficient)

Groups	Queens			B16LS9			B16F10
		Spleen	Liver	Spleen	Liver	Spleen	Liver
IFN 4 days	1.72%	6.30%*	1.38%	3.2%, †	1.28%	4.22%*
PBS	1.56%	3.50%	1.37%	2.62%	1.06%	1.80%

 Data are expressed as ratios. Normal CD49b+NK1.1+ splenic lymphocytes = 19.45% of total splenic lymphocytes; normal CD49b+NK1.1+ hepatic lymphocytes = 18.20% of total hepatic lymphocytes in C57B1/6mice.

*  Significant difference from the PBS group in the ratio of hepatic NK cells to hepatic lymphocytes (P < 0.01; chi-square test).

†  Difference between PBS and IFN groups in the ratio of hepatic NK cells to hepatic lymphocytes (P < 0.05; chi-square test).

Table 5.

 

Number of Hepatic Micrometastases in the IFNα2b–Treated versus Control Uveal Melanoma Mice with NK Depletion

Group	Queens	P
IFN 4 days	85.40 ± 12.32
PBS	85.56 ± 15.44	0.9811

 No significant difference from PBS control group in the number of hepatic micrometastases (P = 0.9811; Wilcoxon Mann-Whitney test). IFN 4 days, see Table 1 . All mice were NK depleted with 100μL of 1:4 anti-asialo GM1 per day, injected SC starting 2 weeks prior to enucleation until the mice were killed.

Figure 5.

 

NK cells increased in micrometastases after IFN treatment. Immunohistochemical stain (brown reaction product) demonstrates CD49b positive NK cells associated with hepatic micrometastatic Queens melanoma cells in an IFN 4-day–treated mouse (top) versus a PBS-treated control (bottom). Hepatocytes surrounded the micrometastasis, which measured less than 100 μm in diameter. Anti-CD49b NK1.1; DAB; magnification, ×100.

Figure 6.

 

Increased micrometastatic apoptosis after IFN treatment. Immunohistochemical stain for TUNEL product in hepatic micrometastasis in control (A) and IFN 4-day–treated (B) Queens PC model. The brown reaction product shows apoptotic melanoma cells in treated (B) but not control (A) micrometastases. The graph shows the ratio of apoptotic to nonapoptotic melanoma cells in the micrometastasis in treated versus control groups for all three melanoma cell lines. There was a significantly greater ratio of apoptotic to nonapoptotic cells after IFN treatment in all three cell lines (P < 0.01). Definition of key is provided in Figure 3 .

Figure 7.

 

Decreased micrometastatic proliferation after IFN treatment. Immunohistochemical stain for MIB1 (anti-Ki-67, protein expressed during cell cycle) in control (A) and IFN 4-day–treated (B) Queens PC model. ABC; magnification, ×100. The brown reaction product shows numerous proliferating melanoma cells in control (A) but not treated (B) micrometastases. There was significantly more proliferation in the control versus the 4- and 1-day IFN–treated groups (graph). Definition of key is provided in Figure 3 .

Table 6.

 

Survival Times after IFNα2b Treatment

Groups	Cells
		Queens	B16LS9	B16F10
PBS	21.93 ± 6.19	23.93 ± 9.86	25.13 ± 5.89
IFN 1 day	34.8 ± 13.95*	34.79 ± 12.66	35.57 ± 9.62
IFN 4 day	47.23 ± 9.39*	45.71 ± 12.18*	44.5 ± 12.53*

*  Significant difference from PBS control group in the number of hepatic micrometastases (P < 0.01; Wilcoxon Mann-Whitney Test).

 IFN 4 days and 1 day are defined in Table 1 . Data are expressed in days ± SD.

Figure 8.

 

Proposed effect of IFN-α2b on micrometastatic melanoma in hepatic sinusoid. IFN-α2b enhances NK activity. Melanoma micrometastasis is associated with sinusoidal endothelial damage. The damaged endothelial cells secrete IL-12 and vascular cell adhesion molecule (VCAM)-1, -4 and -6, thus recruiting NK cells. IFN-α2b boosts NK cells, which secrete IFN-γ, thus potentiating macrophages and NK cells in an autocrine loop. IFN-γ also stimulates T lymphocytes, which potentiate NK cells through IL-10. Damaged sinusoidal endothelium secretes IL-12 and -18, potentiating B lymphocytes. B lymphocytes potentiate NK cells. NK cells and IFN-γ have a direct melanoma-killing effect. NK cells also block bFGF and melanoma angiogenesis.

The authors thank Judy Kapp and Linda Kapp for their assistance with the flow cytometry and Kyle McKenna for review of the manuscript.