Reprogramming Nurse-like Cells with Interferon γ to Interrupt Chronic Lymphocytic Leukemia Cell Survival*

Nurse-like cells (NLCs) play a central role in chronic lymphocytic leukemia (CLL) because they promote the survival and proliferation of CLL cells. NLCs are derived from the monocyte lineage and are driven toward their phenotype via contact-dependent and -independent signals from CLL cells. Because of the central role of NLCs in promoting disease, new strategies to eliminate or reprogram them are needed. Successful reprogramming may be of extra benefit because NLCs express Fcγ receptors (FcγRs) and thus could act as effector cells within the context of antibody therapy. IFNγ is known to promote the polarization of macrophages toward an M1-like state that is no longer tumor-supportive. In an effort to reprogram the phenotype of NLCs, we found that IFNγ up-regulated the M1-related markers CD86 and HLA-DR as well as FcγRIa. This corresponded to enhanced FcγR-mediated cytokine production as well as rituximab-mediated phagocytosis of CLL cells. In addition, IFNγ down-regulated the expression of CD31, resulting in withdrawal of the survival advantage on CLL cells. These results suggest that IFNγ can re-educate NLCs and shift them toward an effector-like state and that therapies promoting local IFNγ production may be effective adjuvants for antibody therapy in CLL.

Given the crucial role NLCs play in CLL cell survival, a number of immune modulators have been screened for their suitability as therapy against them. Burger et al. (3) showed that SDF-1-blocking antibodies reduced the protective effects of NLCs on CLL cells. Morande et al. (5) recently showed that NLCs were susceptible to Aplidin-induced death, suggesting that its anti-tumoral effects were from targeting CLL cells and NLCs simultaneously . Schulz et al. (6) showed that treatment with lenalidomide changed the functional and phenotypic nature of NLCs by interfering with their nurturing properties.
Interferons have been widely accepted as modulators of macrophage plasticity and activation, and it is known that IFN␥ is capable of promoting the differentiation of monocytic cells (7). With regard to therapeutic use, Miller et al. (8) have shown that IFN␥ is beneficial for treating immune disorders such as systemic sclerosis and that it displays antitumor and antiangiogenic effects both in vitro and in vivo . IFN␥ treatment has also been shown to induce antineoplastic immune responses by sensitizing tumor cells to apoptosis via up-regulation of both MHC class I and II molecules and by enhancing antitumor immune activity while decreasing M2 characteristics in immune cells (9,10). IFN␥ has been successfully used in cases of ovarian cancer, multiple myeloma (11), and bladder carcinoma and, recently, in malignant gliomas (12).
Here we examined the effects of IFN␥ on the phenotype and function of NLCs. We found that IFN␥ significantly increased the expression of the M1-related markers CD86 and HLA-DR as well as the phagocytic receptor Fc␥RI. Concurrently, the prosurvival ligand CD31 was down-regulated. Consistent with this, IFN␥-treated NLCs showed superior phagocytic ability toward both opsonized sheep RBCs (SRBCs) and rituximab-coated CLL cells as well as withdrawal of support for CLL cell survival. These results show that IFN␥ can reprogram NLCs to function as immune effectors and suggest that therapies that enhance IFN␥ production locally may be valuable treatments for CLL, particularly when combined with monoclonal antibodies such as rituximab.

Experimental Procedures
Patient Samples-Peripheral blood was collected from CLL patients with informed consent in accordance with the Declaration of Helsinki and under approval from the Institutional Review Board of Ohio State University.
CLL Cell Enrichment-CLL-enriched fractions were prepared using the Rosette-Sep B cell kit (StemCell Technologies, Vancouver, BC, Canada) according to the instructions of the manufacturer.
Flow Cytometry Staining of Surface and Intracellular Markers-NLCs were harvested using 0.25% trypsin (Invitrogen) followed by gentle scraping. After Fc receptor blockade, cells were incubated with 1 g of surface marker antibody at 1 ϫ 10 6 cells/ml for 30 min at 4°C. Cells were washed twice with FACS buffer (PBS, 0.1% sodium azide, and 2% FBS). Intracellular staining was done for CD68 using the BD Cytofix/Cytoperm kit following the protocol of the manufacturer (BD Biosciences). An LSRII flow cytometer (BD Biosciences) was used for flow cytometry, and FlowJo (FlowJo, Ashland, OR) software was used for analysis.
Real-time Polymerase Chain Reaction-RNA was isolated using TRIzol reagent (Invitrogen) and chloroform extraction followed by DNase (Invitrogen) treatment and reverse-transcribed, and cDNA was used for qPCR. GAPDH and 18S RNA were used for normalization. The relative copy number was calculated as 2( Ϫ⌬Ct ) (13).
Phagocytosis-Phagocytosis assays were performed as described previously (14). Briefly, NLCs were pretreated with IFN␥ or PBS for 72 h. SRBCs (Colorado Serum Co., Denver, CO) were fluorescently labeled using PKH26 dye (Sigma) according to the instructions of the manufacturer and then opsonized with anti-SRBC antibody (Sigma) according to the instructions of the manufacturer. SRBCs were mixed with NLCs, and then cells were gently pelleted by slow centrifugation and incubated for 30 min at 37°C. Non-ingested SRBCs were lysed with RBC lysis buffer (eBioscience, San Diego, CA) at room temperature for 10 min. Cells were washed with PBS and fixed with 4% paraformaldehyde. Ingested RBCs were counted using fluorescence immersion oil microscopy. The phagocytic index was calculated as the total number of SRBCs ingested by 50 NLCs.
CLL Cell Survival Assay-Nurse-like cells were pretreated with 10 ng/ml IFN␥ or PBS control for 72 h. Cells were counted and left overnight for adherence in quadruplicate in 96-well tissue culture plates (5 ϫ 10 4 /100 l of medium). CLL cells were added at (5 ϫ 10 5 /100 l) and incubated for 24 h. Non-adherent CLL B cells (100 ϳ 5 ϫ 10 5 ) were harvested and stained with Annexin V FITC/propidium iodide (BD Biosciences) using the protocol of the manufacturer. Data are represented as the percentage of total live CLL cells (Annexin V FITC Ϫ propidium iodide Ϫ ).
NLC Phagocytosis of CLL Cells-CLL cells were enriched as described previously, incubated with 10 g/ml rituximab on ice for 2 h, washed with PBS, and then labeled with wheat germ agglutinin conjugated to Alexa Fluor 647 (5.0 g/ml, Life Technologies) for 10 min at room temperature. Simultaneously, NLCs were harvested, washed, and incubated with wheat germ agglutinin conjugated to Alexa Fluor 488 (5.0 g/ml, Life Technologies) for 10 min at room temperature. Following washes, co-incubations of NLCs with CLL cells for 60 min at a ratio of 1:5 (NLC:CLL) were done. Cells were then washed and fixed with 1% formaldehyde for 20 min at 37°C. Cells were placed onto microscope slides with ProLong Gold mounting solution (Thermo Scientific) and then examined using confocal microscopy.
Statistics-For NLC gene expression studies, paired twotailed Student's t tests were used to compare untreated relative copy numbers to IFN␥-treated relative copy numbers. For the phagocytosis results, paired two-tailed Student's t tests were used to compare the mean phagocytic index control versus IFN␥-treated cells. For the inhibitor experiments, mixed-effect modeling was performed using SAS 9.4 (SAS Inc., Cary, NC). Significance was counted as p ]ltequ] 0.05.

Results
Characterization of NLCs-CLL-patient NLCs were derived as described under "Experimental Procedures," and their characteristics were confirmed via flow cytometry (Figs. 1, A-D). Gating based on forward and side scatter was done in accordance with previous work (1-3) (Fig. 1A). The results showed that the cells were CD68 ϩ and CD14 dim (Fig. 1B), with some CD68 ϩ cells also showing expression of CD200R (Fig. 1C). All cells were CD80 Ϫ (Fig. 1D).

IFN␥ Enhances NLC Expression of CD86, HLA-DR, and
Fc␥RI-NLCs have been described as counterparts to tumorassociated macrophages seen within solid tumors (1,15,16). This led to the question of whether they could be reprogrammed away from their tumor-supportive phenotype. Because of their similarities with solid-tumor macrophages, the possibility existed that treatment with cytokines such as IFN␥ (9) might be effective. To test this, we treated NLCs for 72 h with IFN␥ and measured levels of the M1-related markers (17) NOS2 (nitric-oxide synthase 2), HLA-DR, and CD86. Results showed that IFN␥ led to variable effects with NOS2, increasing for some donors and decreasing for others (data not shown). Transcripts for the T cell coactivator CD86 were significantly elevated ( Fig. 2A), with a corresponding increase in surface expression (Fig. 2, B and C). Likewise, HLA-DR was significantly elevated at the transcriptional (Fig. 2D) as well as cell surface levels (Fig. 2, E and F) by IFN␥.
We also examined Fc␥RI, which is the high-affinity IgG receptor and thus can play a role in antibody-mediated responses (18). This receptor has been shown to respond to IFN␥ in healthy donor monocytes and macrophages (19,20) as well as in primary acute myeloid leukemia cells (21). Results showed that treatment of NLCs with IFN␥ significantly increased Fc␥RI transcript (Fig. 2G) and protein (Fig. 2, H and  I). However, Fc␥RIIa, Fc␥RIIb, Fc␥RIIIa, and ␥ chain levels were unaffected (data not shown).
We also tested the effect of IFN␥ on SDF-1 levels because this is a major protumoral factor produced by NLCs (3). However, no effect of IFN␥ on SDF-1 transcript was seen (data not shown). Collectively, these results suggest that IFN␥ does not effect a complete shift toward an M1 phenotype in NLCs but that it does up-regulate molecules involved with effector functions.
IFN␥ Enhances Phagocytosis by NLCs-Because IFN␥ increased expression of the high-affinity IgG receptor Fc␥RI, we tested the effects of IFN␥ on phagocytosis. We treated NLCs with IFN␥ for 72 h and then measured their ability to ingest fluorescently labeled, opsonized SRBCs. The results showed that IFN␥-treated NLCs ingested significantly more SRBCs than untreated NLCs (p ϭ 0.034, Fig. 3A, plotted  in B).
Next, we tested whether IFN␥-treated NLCs would be capable of phagocytosing antibody-coated CLL cells. We treated NLCs with IFN␥ for 72 h as above, membrane-labeled them with fluorescent dye, and then incubated them for 1 h with membrane-labeled CLL cells that were opsonized with the anti-CD20 antibody rituximab, which is commonly used for the treatment of CLL. We examined phagocytosis via confocal microscopy between untreated (Fig. 3C) and IFN␥-treated (Fig. 3D) NLCs and found that IFN␥ significantly increased the number of ingested CLL cells (Fig. 3E). IFN␥ Down-regulates CD31 and Reduces NLC-mediated Survival of CLL Cells-CD38/CD31 interactions in cooperation with CD100 promote survival of CLL cells, and it has been shown that blocking antibodies against CD31 can disrupt CLL cell survival (22). Hence, we tested the effects of IFN␥ treatment on the expression of CD31 in NLCs. The results showed that 72-h treatment with IFN␥ led to a significant reduction in CD31 (Fig. 4A).
The above led us to test whether IFN␥-induced NLC polarization would be sufficient to interfere with NLC-dependent CLL cell survival. We treated NLCs for 72 h and then co-cultured them with CLL cells for 24 h. CLL cell survival was measured by Annexin/propidium iodide staining. The results showed that IFN␥ treatment significantly reduced the survival of CLL cells within the co-cultures (Fig. 4B) despite not accelerating the death of control CLL cells in single culture (data not shown).

Discussion
In this study, we found that treatment of NLCs with IFN␥ could reprogram them toward a more effector-like phenotype and also in such a way that they no longer supported the survival of CLL cells. IFN␥ also significantly enhanced the phagocytic ability of NLCs against opsonized SRBCs as well as rituximab-coated CLL cells. These findings suggest that IFN␥ could serve to improve the outcome of antibody therapy for CLL. They also support the earlier observation that NLCs resemble M2-like tumor-associated macrophages (15) because the latter have been found to respond to IFN␥ (9).
In addition, these data suggest that SDF-1, although important for CLL cell survival, is by itself not sufficient. IFN␥ did not significantly decrease NLC SDF-1 but did decrease CLL cell survival in NLC/CLL cultures. This is in agreement with Burger et al. (3), who found that supplementing CLL cells with SDF-1 offered some but not full protection against apoptosis. Additional survival stimuli such as CD31/CD38 interactions, along with others yet to be tested, are likely to contribute to CLL cell survival. Quantifying the full effects of IFN␥ on NLCs with regard to their interactions with CLL cells will require further study.
Direct administration of IFN␥ continues to be tested for conditions including macular edema, HIV, and various tumor types (http://www.clinicaltrials.gov). A synthetic version of IFN␥ (Actimmune) was approved for the treatment of chronic granulomatous disease as well as to delay the progression of malignant osteopetrosis (http://www.fda.gov/Drugs/DrugSafety/Postmarket DrugSafetyInformationforPatientsandProviders/ucm109130. htm). Our results suggest that such IFN␥ administration may be beneficial against CLL as well. Within the context of antibody therapy, IFN␥ would be induced in natural killer cells, which could act locally upon the NLCs. This could be further strengthened by the co-administration of agents such as IL-12 (23), CpG (24), and TLR8 agonists (25). Such co-treatment would be predicted to significantly enhance antibody-mediated clearance of CLL cells and may also inhibit the development of new NLCs. Given the importance of CD20 antibody-based therapy in prolonging survival of CLL patients, this could represent a major advance for this currently incurable disease. Author Contributions-J. C. B., S. T., and J. P. B. conceived and designed the study and wrote the manuscript. S. G., K. F., S. E., B. F. R., and L. R. performed the experiments, collected the data, and summarized the results. X. M. analyzed and interpreted the data and contributed to manuscript preparation. All authors gave final approval to the manuscript.