Reagents

To understand the role of IL-2 family of receptors on NK behavior we used the following NK cell lines: YT (RRID:CVCL_1797), an IL-2 independent cell line courtesy of Eric Long; YTS, an IL-2 independent cell line originally derived from YT cells (CVCL_D324, RRID:CVCL_D324); NK-92 (CRL-2407, ATCC, Manassas, Virginia, RRID:CVCL_2142) and NKL (LGL cell line, RRID:CVCL_0466) are both IL-2 dependent cell lines. All these cells were cultured in RPMI 1640 (cat# 338080, Invitrogen, Carlsbad, California, United States) supplemented with 12.5% fetal bovine serum (cat# 16000−044, Invitrogen, Carlsbad, California, United States), 1% pennicillin-streptomycin (cat# MT 30–002-CI, Thermo Fisher Scientific, Waltham, Massachusetts, United States), 1mM HEPES (cat# 15630−080, Invitrogen, Carlsbad, California, United States), 1% non-essential amino acids (cat# 25–025-CI, Thermo Fisher Scientific, Waltham, Massachusetts, United States), 1% sodium pyruvate (cat# 25–000-CI, Thermo Fisher Scientific, Waltham, Massachusetts, United States) and L-glutamine (cat# MT 25–005-CI, Thermo Fisher Scientific, Waltham, Massachusetts, United States) If IL-2 dependent, they received 300U/mL recombinant human (rh) IL-2 (cat# 200−02, Peprotech, Cranbury, New Jersey) and 50uM β2 mercaptoethanol (cat# 21-985-023, Thermo Fisher Scientific, Waltham, Massachusetts, United States) every other day. Pathway inhibitors used were: Wortmannin (cat# W1628-1MG Sigma-Aldrich, St. Louis, Missouri, United States), U0126 (cat# 662005 Sigma-Aldrich, St. Louis, Missouri, United States), Ly294002 (cat# 440202 Sigma-Aldrich, St. Louis, Missouri, United States), API-2 (Triciribine hydrate; cat# T3830 Sigma-Aldrich, St. Louis, Missouri, United States), AG 490 (cat# 658401 Sigma-Aldrich, St. Louis, Missouri, United States).

Primary human peripheral blood mononuclear cells (PBMC) were sourced from Florida Blood Services’s (OneBlood, Orlando, Florida, United States) buffy coats and isolated by Ficoll-Paque Plus density gradient (cat# 45001750, GE Healthcare, Chicago, Illinois, United States). Ethical review and approval were waived for this study, due to the use of buffy coats which is considered non-Human research under exemption status 4 since the samples are de-identified and obtained from a commercial source. Therefore, due to lack of characterizing information (i.e.,: age and gender), we did not select the specimens used in our experiments. After PBMC isolation, NK cells were enriched using the negative NK cell enrichment kit from Stem Cell Technologies (cat# 17955, StemCell Technologies, Vancouver, Canada) and purity assessed by flow cytometry of CD56⁺CD3^- cells (see flow cytometry section). PBMC were cultured in complete RPMI as cell lines and NK cells were cultured with ImmunoCult NK cell expansion kit (cat# 100−0711, StemCell Technologies, Vancouver, Canada) following the manufacturer’s protocol and supplemented every other day with 300U/mL rhIL-2 and 50uM β2 mercaptoethanol. Immediately after isolation PBMC and isolated NK cells were used. Recombinant human IL-2 was added at 300U/mL for testing and rh IL-15 (cat# 300-138P, GeminiBio, West Sacramento, California, United States) was added at 10ng/mL as indicated in the results section. Lenalidomide (cat# S1029, Selleckchem, Houston, Texas, United States) was dissolved first in dimethyl sulfoxide (DMSO, cat# D2650, Sigma-Aldrich, St. Louis, Missouri, United States) at a 10mM concentration and aliquots frozen at -80^oC until ready to use.

Flow cytometry

For cell cycle analysis 7.5x10⁵ cells were fixed in 1.5mL of ice cold 100% ethanol (cat# E7023-6X500ML, Sigma-Aldrich, St. Louis, Missouri, United States) while gently vortexing to avoid cell clumping, and allowing for a single cell suspension, and stored at -20^oC until ready to stain. Cells were then washed with 1X PBS (cat# 14190–250, Thermo Fisher Scientific, Waltham, Massachusetts, United States) and incubated in PBS supplemented with propidium iodine (PI, 1 mg/mL, cat# 556463, BD Biosciences, Franklin Lakes, New Jersey, USA) and ribonuclease A (RNAse, 10 mg/mL, cat# BP25391, Thermo Fisher Scientific, Waltham, Massachusetts, United States) at room temperature for 2–3 hours or overnight at 4^oC. After acquisition, data was analyzed using ModFit LT Software (Verity Software House, Topsham, Maine, United States, RRID:SCR_016106). Flow cytometry-based proliferation assessment was done by staining cells with either CFSE (Carboxyfluorescein succinimidyl ester, cat# C34554, Invitrogen, Carlsbad, California, United States) or CellTracker Violet BMQC dye (cat# C10094, Invitrogen, Carlsbad, California, United States). Apoptosis analysis was done by staining experimental cells with Annexin V FITC (cat# 556547, BD Biosciences, Franklin Lakes, New Jersey, USA) and PI according to the manufacturer’s instructions.

Surface phenotyping of NK cells was measured with the use of the following antibodies: anti-human CD3 FITC (HIT3a, cat# 11-0039-42, Thermo Fisher Scientific, Waltham, Massachusetts, United States, RRID:AB_1724043), anti-human CD3 PE (HIT3a, cat# 300308, Biolegend, San Diego, California, United States, RRID:AB_314044), anti-human CD16 FITC (3G8, cat# 555406, BD Biosciences, Franklin Lakes, New Jersey, United States, RRID:AB_395806), anti-human NKG2D PE (1D11, cat# 12-5878-41, Thermo Fisher Scientific, Waltham, Massachusetts, United States, RRID:AB_1659727), NKp44 PE (p44-8, cat# 558563, BD Biosciences, Franklin Lakes, New Jersey, United States, RRID:AB_647239), NKp30/CD337 PE (P30-15, cat# 558407, BD Biosciences, Franklin Lakes, New Jersey, United States, RRID:AB_647240), anti-human CD132/IL-2Rγ chain PE (AG184, cat# 561699, BD Biosciences, Franklin Lakes, New Jersey, United States, RRID:AB_10896131), anti-human CD122/IL-2Rβ PerCP eFluor 750 (9A2-CD122, cat# 46-1229-41, Thermo Fisher Scientific, Waltham, Massachusetts, United States, RRID:AB_10698171), anti-Human CD107a (LAMP-1) PE (eBioH4A3, Cat# 12-1079-42, Thermo Fisher Scientific Waltham, Massachusetts, United States, RRID:AB_10853326), anti-human CD107a PECy7 (H4A3, cat# 561348, BD Biosciences, Franklin Lakes, New Jersey, United States, RRID:AB_10644018), IFN-γ APC (B27, Cat# 554702, BD Biosciences, Franklin Lakes, New Jersey, United States, RRID:AB_398580), IFN-γ eFluor 450 (4S.B3, cat# 48-7319-42, Thermo Fisher Scientific, Waltham, Massachusetts, United States, RRID:AB_2043866), CD56 APC (CMSSB, Cat# 17-0567-41, Thermo Fisher Scientific, Waltham, Massachusetts, United States, RRID:AB_10596498), CD56 FITC (MEM188, cat# 11-0569-42, Thermo Fisher Scientific, Waltham, Massachusetts, United States, RRID:AB_1834372), 2B4/CD244 PE (2−69, cat# 550816, BD Biosciences, Franklin Lakes, New Jersey, United States, RRID:AB_393901), DNAM-1/CD226 Alexa Fluor 647 (DX11, cat# 564797, BD Biosciences, Franklin Lakes, New Jersey, United States, RRID:AB_2738957), NKp46/NCR1/CD335 PE (195314, cat# FAB1850P, R and D Systems, Minneapolis, Minnesota, United States, RRID:AB_2149279), KIR2DL1/KIR2DS5 FITC (143211, cat# FAB1844F, R and D Systems, Minneapolis, Minnesota, United States, RRID:AB_2130402), KIR2DL3/CD158b2 FITC (180701, cat# FAB2014F, R and D Systems, Minneapolis, Minnesota, United States, RRID:AB_2130677).

Data was acquired in BD LSRII equipment from Moffitt’s Flow cytometry core facility, that maintains daily quality control of each cytometer and analyzed using FlowJo (TreeStar, Ashland, Oregon, United States, RRID:SCR_008520). All analyses included gating on live cells which were staining using 7AAD (cat# 51–2359KC, BD Bioscience, Franklin Lakes, New Jersey, United States,) or LIVE/DEAD dyes (cat# L34960, Invitrogen, Carlsbad, California, United States).

Radioactive isotope-based proliferation and cytotoxicity assays

Apart from flow cytometric assessment, proliferation of NK cells after treatment (as described in the results section) was also measured by [3H] thymidine (1.0 μCi per well; cat# MP01240662, MP Biomedicals, Irvine, California, United States) in 96-well U bottom plates (cat# 3879, Corning Life Sciences, Acton, Massachusetts, United States). To assess the cytotoxic potential of NK cells (primary or cell lines) we used K562 (cat# CCL-243, ATCC, Manassas, Virginia, United States, RRID:CVCL_K562), 721.221 (cat# CRL-1855, ATCC, Manassas, Virginia, United States, RRID:CVCL_6263), or MDS-L (RRID:CVCL_A8QV) target cells pre-loaded with 100 μCi of ⁵¹Cr (chromium, cat# NEZ 0305 002MC, PerkinElmer, Waltham, Massachusetts, United States) per 1x10⁶ cells for 60 min at 37°C. After co-culture with NK cells for 5 hours, 200 uL of media was collected for radioisotope release. Quantification was done by measuring radioactivity on a Wizard 1470 gamma counter (PerkinElmer, Waltham, Massachusetts, United States) and plotted as mean counts per minute (cpm) of triplicate wells.

Cytokine production

To measure cytokine release, culture supernatants were collected and frozen at -80C until ready to run the assay. At that point, supernatants were incubated with IL-2, IFN-γ, TNF-α, and IL-6 from the BD Cytometric Bead Array (CBA) Human Th1/Th2/Th17 kit (cat# 560484, BD Biosciences, Franklin Lakes, New Jersey, United States) measured by flow cytometry, at Moffitt’s flow cytometry core facility, following the manufacturer’s protocol. Using the standard provided in each array, the concentration of each cytokine was assessed and plotted accordingly. Each sample was run in triplicates for each experimental determination.

Western Blot

To measure changes in protein expression, after experimentation cells were harvested and lysed for 30 minutes in a buffer containing: 50 mM Tris-HCl, 1% NP-40, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1 mM sodium orthovanadate, 1 mM sodium fluoride supplemented with protease inhibitor cocktail (cat# P 2714 Sigma-Aldrich, St. Louis, Missouri, United States) and phosphatase inhibitor coacktails I and II (cat# P2850 and cat# P-5726, Sigma-Aldrich, St. Louis, Missouri, United States) to prevent protein degradation. Total protein was measured in a Bio-Rad protein assay (cat# 500−0006, Bio-Rad, Hercules, California, United States) prior to running on a 10−12% Tris Acrylamide gel (prepared in house using cat# 161−0158 30% Accylamide solution and cat# 161−0800 TEMED both from Bio-Rad, Hercules, California, United States) to resolve protein by size by running constant current at 40 Volts (V) until the protein reached the border between the stacking and running gels (approximately 15 minutes) followed by 100V for approximately 1.5 hours, until the dye reached the bottom of the gel. A total of 50ug of protein per lane was used to allow normalization to a house keeping gene. The proteins were then transferred to a PVDF membrane (cat# 1620177, Bio-Rad, Hercules, California, United States) by electrophoresis overnight at 20 milliAmps (mA) surrounded by ice. After assessing that the highest band transferred completely membranes were retrieved and blocked with 5% nonfat milk-based tris buffer with 0.05% Tween 20 (cat# BP337−500, Thermo Fisher Scientific, Waltham, Massachusetts, United States). The primary antibodies used for protein expression assessment were: total STAT5 (D3N2B, 1:1000, cat# 25656, Cell Signaling Technology, Danvers, Massachusetts, United States, RRID:AB_2798908), phospho- STAT5 (C11C5, 1:1000, cat# 9359P, Cell Signaling Technology, Danvers, Massachusetts, United States), total AKT (11E7, 1:1000, cat# 4685S, Cell Signaling Technology, Danvers, Massachusetts, United States), phospho-AKT (D25E6, 1:1000, cat# 130385, Cell Signaling Technology, Danvers, Massachusetts, United States), PI3K p85 (I9H8, 1:1000, cat# 4257, Cell Signaling Technology, Danvers, Massachusetts, United States, RRID:AB_659889) and beta-actin as a loading control (AC-74, 1:5000, cat#A5316, Sigma-Aldrich, St. Louis, Missouri, United States, RRID:AB_476743). All primary antibodies were incubated overnight and beta-actin for 1 hour. After blotting with the appropriate secondary bands were detected with the SuperSignal™ West Pico PLUS Chemiluminescent Substrate detection system (cat# 34580, Thermo Fisher Scientific, Waltham, Massachusetts, United States).

Real Time PCR

RNA was isolated from treated cells by Trizol Reagent (cat# 15596018, Thermo Fisher Scientific, Waltham, Massachusetts, United States) lysis and isolation or using the RNeasy kit (cat# 74106, Qiagen, Germantown, Maryland, United States) after lysing in RLT buffer included in the kit. To measure the concentration after isolation we used NanoDrop Spectrophotometer (Thermo Fisher Scientific, Waltham, Massachusetts, United States, RRID:SCR_025369) part of Moffitt’s common equipment. One ug of total RNA was used to prepare cDNA using the qScript cDNA Supermix (cat# 95048–100, Quantabio, Beverly, Massachusetts, United States) followed by real-time semi-quantitative PCR with 1uL of cDNA using the iQ SYBR Green Supermix (cat# 170–8880, Bio-Rad, Hercules, California, United States) for the enzymatic reaction with a total reaction volume of 20uL. After an initial activation step of 3 minutes at 95°C, each reaction continued with 40 cycles of 95°C for 15 seconds, 60°C–62°C for 30 seconds and 95°C for 1 minute with fluorescence measurement at the annealing step of each cycle. To ensure the quality of the primers and reactions, a melting curve was run after cycling finished at a 0.5 °C ramp starting from 55 °C till 95 °C. The reactions were done in a 96-well spectrofluorometric thermal cycler (Bio-Rad CFX96 Real-Time PCR Detection System (RRID:SCR_018064). For each experiment duplicate determinations of each treatment were done.

Primers specific for each respective mRNA used in this study were: Perforin (PRF1): forward 5’-CAGCACTGACACGGTGGAGT-3’; reverse-5’-GTCAGGGTGCAGCGGG-3’. Granzyme B (GZMB): forward- 5’-TCCTAAGAACTTCTCCAACGACATC 3’; reverse- 5’-GCACAGCTCTGGTCCGCT-3’. Actin (PCR control, ACTB)): forward- 5’- TGGCACCCAGCACAATGAA-3’; reverse- 5’-CTAAGTCATAGTCCGCCTAGAAGCA-3’. After selecting the cycle threshold (Ct) quantification was done using the method 2 ^−ΔΔCT where first delta is the difference of each point compared to the experimental control and the second delta is the difference with the internal gene ACTB.

Confocal immunofluorescence microscopy

After appropriate culturing conditions for each experiment 100000 cells were collected on slides in a Cytospin by centrifugation for 3 min cetrifugation at 500 rpm and fixed with BD Cytofix/Cytoperm buffer (cat# 554722, BD Biosciences, Franklin Lakes, New Jersey, United States,) for 15 min at room temperature followed by 3 washes with 1X PBS (pH 7.4) and freezing at −20 °C until use. When ready, slides were allowed to defrost, rehydrated with three 5 min washes with 1X PBS followed by blocking with 2% BSA (cat# J30867-A1, Thermo Fisher Scientific, Waltham, Massachusetts, United States) in PBS for 30 min at room temperature. Appropriate primary antibodies: Granzyme B (C-19, 1:200 dilution, cat# sc-1968, Santa Cruz Biotechnology, Dallas, Texas, United States, RRID:AB_2279405) or Perforin (H-315, cat# sc-9105, Santa Cruz Biotechnology, Dallas, Texas, United States, RRID:AB_2168954)), were diluted in blocking buffer used to incubate the slides overnight at 4 °C. The next day the slides were washed three times with 1X PBS after which the appropriate secondary antibodies (donkey anti-goat A647 cat# A-21447 RRID:AB_141844, and donkey anti-rabbit A488 cat# A-21206 RRID:AB_2535792 respectively, both from Thermo Fisher Scientific, Waltham, Massachusetts, United States) were diluted in blocking buffer and incubated on the slides for 1 hr in the dark at room temperature. Slides were then washed and mounted with ProLong Gold Antifade Reagent with DAPI (cat# P36930, LifeTechnologies, Carlsbad, California, United States) and covered with a coverslip. Fluorescence imaging was measured with a Leica TCS SP5 AOBS Laser Scanning Confocal microscope (Leica Microsystems GmbH, Wetzlar, Germany) at the Microscopy Core Facility of Moffitt Cancer Center where samples were excited with 405, 488, and 638 diode lasers and emissions were tuned to appropriate settings for each dye. Images were captured through a 63X/1.4NA objective lens using two PMT detectors (DAPI and Alexa 488) and one HyD detector (Alexa 647). All system settings remained consistent for all samples within each experiment. Images were viewed and exported in TIF format with Leica Application Suite X software version 3.1.5 (Leica Microsystems GmbH, Wetzlar, Germany RRID:SCR_013673). In each slide (one per treatment per experiment) 10 fields selected at random were captured for analysis and saved as TIF files.

Fluorescence Intensity analysis was performed by the Microscopy Core on the TIF images with Definiens Tissue Studio version 4.7 (Definiens AG, Munich, Germany RRID:SCR_014283). In the software the nucleus detection and cell growth algorithms were used to segment individual cells within each image. Using this cell segmentation the mean fluorescence intensity (MFI) was calculated for each image and exported to Microsoft Excel for analysis.

Virus production and transduction

Plasmids used for our PI3K studies were originally derived from p110α of class IA PI3K: constitutively active CAp110 is M-p110*-myc a chimera containing the iSH2 domain of p85 fused to the N terminus of p110α by a flexible glycine kinker; while dominant negative DNp110 is M-p110Δkin-myc a kinase-deficient p110α in which the arginine at position 802 is mutated to a lysine residue. These vectors were originally described in [21,22] and were a gift from Dr. Julie Djeu who herself obtained them from A. Klippel (Atugen, Berlin, Germany).

The generation of and infection with vaccinia virus have been described previously [23,24]. Briefly, cells were incubated with recombinant vaccinia viruses encoding CAp110 or DNp110 for 1.5 h at 37^oC in serum-free medium at a multiplicity of infection of five. Vaccinia virus expressing CD56 was used as a control for nonspecific effects of viral infection. Transduction was assessed by flow cytometry and confirmed by western blot analysis of PI3K. Cells were not selected post infection and used immediately after testing for expression.

Statistical analysis

Statistical analyses were performed using GraphPad Prism Software (GraphPad Prism, Boston, Massachusetts, United Stated, RRID:SCR_002798). Sample sizes were estimated based on our previous studies and literature. Data was assessed by normality and appropriate tests were selected for analysis. Statistical significance to assess two treatment groups only, such as the cytokine levels (ELISA) of untreated or treated with lenalidomide, were determined by unpaired Student t tests. Analysis of two variables, such as time and cytotoxicity, were analyzed for significance using two-way ANOVA. A one-way ANOVA was applied to other data with multiple comparison analysis when various treatment groups were analyzed. Data of timelines were assessed by calculating the area under the curve followed by comparison of area measurements. All analysis and graphics show standard error of the mean bars (SEM) of at least three biological replicates (denoted as n on figure legends) and p values < 0.05 were statistically significant.

Lenalidomide combination with IL-2 family of cytokines changes NK cell subsets

To assess the effect of combining lenalidomide treatment of NK cells with IL-2 or its relative IL-15, a more potent activator of NK cells [19,25], we treated either peripheral blood mononuclear cells (PBMC, to assess the potential indirect effect of lenalidomide on NK cells), or isolated NK cells (to understand the direct effect of treatment) with lenalidomide and/or IL-2/IL-15 to study viability and apoptosis. We used the effective dose of lenalidomide we used in our previous studies in these experiments [26]. Lenalidomide decreased the viability of PBMC (live cells defined as Annexin V-/PI-) which was overcome by the addition of IL-2 or IL-15 (S Fig 1a, representative figure of n = 5 separate healthy donor PBMC). However, this did not extend to availability of NK cells in the mononuclear population where we saw an increase in the percentage of these cells in the presence of IL-2, IL-15 or lenalidomide (CD56⁺CD3^- cells gated directly from live cell; Fig 1a, representative figure of n = 6 separate healthy donor PBMC). IL-2 and IL-15 maintained the viability of NK cells regardless of lenalidomide treatment suggesting that the effect of lenalidomide and IL-2 family of cytokines are independent of each other on NK cells (Fig 1b-c). Moreover, it suggests that lenalidomide treatment does not induce enough IL-2 production from other cells in the microenvironment to increase the percentages of NK cells in the mononuclear population.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Combination with IL-2 or IL-15 restores impaired proliferation induced by lenalidomide.

a) Flow cytometric phenotyping of primary healthy human PBMC, gated on NK cell subpopulations as shown from total live cells (total NK cells are defined as CD3-CD56+and percentage is shown on the side of gates, NK cell subsets: CD56dim and CD56bright from total NK and percentages are shown inside each gate) after treatment with 10μM lenalidomide (LEN), or DMSO vehicle for 7 days. Data representative of n = 6 healthy donor individual experiments (biological replicates). Pseudo-coloring denotes the density of cells on each marker. b) Isolated primary human healthy NK cells and subsets: (CD56⁺CD3^- total NK cells, CD56^dim and CD56^bright) after treatment with lenalidomide, IL-2, IL-15 or their combination for 7 days. Data representative of n = 6 healthy donor individual experiments (biological replicates). Pseudo-coloring denotes the density of cells on each marker. c) Viability of isolated primary NK cells treated with lenalidomide, IL-2, IL-15 or their combination for 72 hours by flow cytometric analysis using a live/dead dye. Proliferation of isolated primary NK subsets (d) CD56^bright and (e) CD56^dim by flow cytometric analysis of CellTrace Violet from cells gated as in “a” and “b” (n = 6 healthy donor biological replicates). f) Representative figure of proliferation with CellTrace Violet of NK subsets as in d and e from primary isolated NK cells from healthy donor PBMC (representative of n = 6 biological replicates). Grey line shows the recorded data, and the blue lines represent the divisions as calculated by FlowJo. In all figures, error bars represent the SEM of biological replicates and were analyzed by ordinary one-way ANOVA with multiple comparison analysis. P values are shown as asterisk: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001, **** P ≤ 0.0001 as summarized in Graphpad Prism.

https://doi.org/10.1371/journal.pone.0344471.g001

CD56 expression defines the functionality of human NK cell subsets: CD56^dim NK cells are more cytotoxic while CD56^bright NK cells focus on the production of cytokines after monocyte-induced activation [27]. Lenalidomide can modulate the expression of CD56 potentially shifting the activity of NK cells in the tumor microenvironment [5,6,28]. Therefore, we tested the effect of lenalidomide alone or in combination with IL-2 family of cytokines on NK cell subsets. Lenalidomide increased the percentages of total NK cells (percentage shown on the side of the gates representing total CD56⁺CD3^- NK cells), their populations remained with higher CD56^dim proportions (bottom gate labeled “dim” with percentage shown inside the gate). However, treatment with IL-2 increased the proportion of CD56^bright (percentages shown on gate labeled “Bright”) as well changing the ratio of these two different NK cell subsets regardless of the presence of other immune cells in the culture (PBMC versus isolated cells, Fig 1a-b, representative figures of n = 6 healthy donors). IL-15 co-treatment with lenalidomide had a better preservation of the CD56^dim population (Fig 1a-b). This matches the proliferation potential of NK cell subtypes under these conditions: The number of dividing CD56^bright cells decreased with the presence of IL-2 or IL-15 while the combination of lenalidomide with IL-15 did not have a significant reduction, showing the same number of generations as control or lenalidomide only treatments (Fig 1d and f; f is a representative figure of n = 6 healthy donors from tracker dye stained cells gated as in Fig 1b). In comparison, the cytotoxic CD56^dim cells did not significantly change the number of dividing cells with any of the teatment combinations (Fig 1e-f; f is a representative figure of n = 6 healthy donors from tracker dye-stained cells gated as in Fig 1b).

Lenalidomide needs extra IL-2 family of cytokines to move from cell cycle arrest

To confirm that IL-2/IL-15 can increase proliferation restricted by lenalidomide separately from just increased survival, we used the NK cell lines YT and YTS (IL-2 independent cell lines), NK-92 and NKL (IL-2 dependent cell lines) to confirm results observed in primary cells as well as to understand the role of the shared IL-2 receptor in lenalidomide’s effects. The IL-2 independent cell lines validated the observations in primary NK cells with reduced proliferation when the cytokines or cytokines and lenalidomide are present, while IL-2 sensitive cell lines did not (S Fig 1b). This suggests that IL-2 needed for survival leads to a bigger threshold to induce proliferation, supporting the need to enhance lenalidomide treatment with IL-2 family of cytokines. We further interrogated this increased proliferation by selecting YT cells for long term culture, where we observed that lenalidomide treated YT cells have significantly impaired cell proliferation (Fig 2a, average of n = 3 biological replicates) although the addition of exogenous IL-2 improved the lenalidomide-induced proliferation impairment, to levels above the control group in YT cells (Fig 2b, average of n = 3 biological replicates). We also compared the proliferative effects of IL-2, IL-15, or their combination with lenalidomide and found that IL-15, alone or in combination, significantly increased the proliferation of cells compared to IL-2 alone (only day 7 shown Fig 2c, n = 3 biological replicates).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Combination with IL-2 with Lenalidomide’s effects on NK cell cycle.

Proliferation of YT cells measured by overnight tritium (3H) incorporation (1.0 μCi per well) after culture for 1,3,5 or 7 days in the presence of 10 or 20μM lenalidomide (LEN) in the (a) absence or (b) presence of human recombinant (hr) IL-2 (100U/mL) or (c) combined results of YT proliferation at 7 days for coculture of lenalidomide with IL-2 and/or IL-15. Results expressed as mean counts per minute (cpm) of triplicate wells for each experiment, and a total of n = 3 biological replicates. Cell cycle analysis of isolated primary NK cells (n = 6 healthy donor biological replicates), treated as indicated, measured via flow cytometric analysis of propidium iodide (PI) staining and expressed as percent of cells in (d) S, (e) G1, or (f) G2 phase. g) Cell cycle analysis of YT cells cultured with 20μM LEN (for up to 14 days) measured via flow cytometric analysis of propidium iodide (PI) staining and expressed as percent of cells in S, G1 and G2 phase. Error bars represent the SEM and were analyzed by 2way ANOVA. P values are shown as asterisk: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001, **** P ≤ 0.0001 as summarized in Graphpad Prism.

https://doi.org/10.1371/journal.pone.0344471.g002

Next, we assessed if the reduced cell proliferation, in the absence of increased apoptosis, is instead the result of cell cycle arrest. In cultures of primary NK cells, the only demonstrable cell cycle change seen was a significant decrease in the percentage of cells in S-phase in the presence of IL-2, which was only achieved by IL-15 when combined with lenalidomide (Fig 2d-f). Longer incubations using YT cells treated with lenalidomide did have a time-dependent increase in cells in G1, decrease in S-phase and corresponding reduction in G2 (Fig 2g) which were not apparent until day 5 (G1 & S) or day 10 (G2). This may explain the lack of effect seen in the short incubation of isolated NK cells. Similarly, PBMC treated for 7 days also had a G1 arrest, overcome by IL-2 or IL-15 (S1C Fig, average of n = 6 biological replicates).

NK receptor expression is selectively modulated by Lenalidomide

NK activating and inhibitory receptor expression has a direct impact on the cytotoxic efficacy of a NK cell [29], which can be modulated by treatment with lenalidomide [5,30]. Due to functional implications of subset switching, we decided to investigate the phenotypic characterization of NK cell activating and inhibitory receptors after lenalidomide treatment. An analysis of lenalidomide’s impact demonstrates that several key NK receptors are affected by treatment including CD16, NKp46, KIR2DL3, KIR2DL1 (Fig 3a) while others remained unchanged by lenalidomide treatment (NKG2D, 2B4 and DNAM1, S Fig 2a). We found that CD16 expression is increased by lenalidomide in CD56^bright NK cells, similar to other’s work [6] while in the more cytotoxic CD56^dim cells it did not. NKp46, a key NK activating receptor, is downregulated in the CD56^bright population, but unchanged on CD56^dim cells (Fig 3a). The two inhibitory KIRs, KIR2DL1 and KIR2DL3, were both reduced by lenalidomide. KIR2DL3 was reduced in both NK subsets while KIR2DL1 was only in the CD56^dim population (Fig 3a). This suggests that despite a ratio disfavoring CD56^dim NK cell, these are likely more active than untreated NK cells. The other receptors studied; NKG2D, 2B4, or DNAM-1 were not affected in either subset (S2a Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Lenalidomide acts directly on NK cells to modulate the expression of CD16 and NKp46.

(a) Flow cytometric phenotyping of NK cell receptor expression in primary human healthy PBMC, gated on NK cell subpopulations (total NK cells CD3^-CD56⁺, NK cell subsets CD56^dim and CD56^bright from total NK) after treatment with 10μM lenalidomide (LEN), or DMSO. Shown are: CD16, NKp46, KIR2DL3, KIR2DL1. Data representative of n = 5 independent experiments. (b) Western blot analysis of phosphorylated AKT (pAKT) in either primary isolated NK cells or YT cells treated as shown. Full raw gels are shown in supplemental figure 2. (c) Quantification of western blots showing the ratio of phosphor-AKT (phosphorylated threonine 308) to total AKT (n = 3) and representative figure for p85 in YT cells treated with 20μM of LEN, or a DMSO control, for 1, 3, 5, or 7 days. (d) Corroboration of phosphor AKT reduction by LEN treatment measured with phospho-flow. To adjust data across samples to minimize technical variation (batch effects) for biological comparisons the Y-axis overlay normalization in FlowJo was used in this analysis as shown. Error bars represent the SEM and were analyzed by ANOVA with multiple comparison analysis. P values are shown as asterisk: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001, **** P ≤ 0.0001 as summarized in Graphpad Prism.

https://doi.org/10.1371/journal.pone.0344471.g003

The PI3K/AKT/mTOR (Phosphoinositide 3-kinases/ Protein kinase B/ mechanistic target of rapamycin) pathway regulates the expression of crucial NK receptors, including NKp46 [31,32], and lenalidomide can affect PI3K activity [33]. Hence, to define the mechanism by which lenalidomide affects these receptors, we first dissected the role of lenalidomide on this pathway. Lenalidomide inhibits AKT activation in both primary isolated human NK and YT cells, as measured by AKT phosphorylation in western blot, but did not affect the overall protein level of p85 (the regulatory subunit of PI3K, [34]) suggesting a specific inhibition of AKT activation (Fig 3b-c, representative figure of n = 5 healthy donors and n = 3 biological replicates, S2b-S2d Figs shows raw gels). We validated this reduction of phosphorylated AKT in YT cells by intracellular flow cytometry (Fig 3d, representative figure of n = 3 biological replicates) and observed a time-dependent inhibition with the strongest reduction at day 1 and restoration of normal phosphor-AKT levels by day 7 post lenalidomide treatment.

NKp46 expression is selectively modulated by Lenalidomide through PI3K inhibition

To confirm that NKp46 regulation by lenalidomide is linked to AKT we did chemical inhibition of key pathways linked or peripheral to PI3K/AKT pathway to assess their role (schematized in Fig 4a). We have previously used a similar strategy to demonstrate that PI3K is the key upstream regulator of MAPK-ERK in the NK lytic process and that Wortmannin- or LY294002-inhibition of PI3K in NK cells blocked ERK activation and perforin–granzyme B movement towards target cells, and subsequently tumor cytotoxicity [35,36]. While both Wortmanin (IC50 5nM) and LY294002 (IC50 1.4μM,) are highly specific PI3K inhibitors, it has been shown that they have different effects on phosphorylation of the p85 and p110 subunits [37,38]. Wortmannin makes a covalent bond with the catalytic subunit p110 making a conformation change in the protein [37]. LY294002 on the other side binds reversibly and competes directly for the ATP binding site of p110 catalytic subunit [37]. As such, Wortmanin treatment decreases the phosphorylation of both p85 and p110 while LY294002, under the same conditions, has residual phosphorylation of both subunits even at large quantities. Importantly both have been shown to have different effects on apoptosis and proliferation in cells [39]. They affect iNOS and immune responses below it differently too [40]. Lenalidomide decreased the levels of NKp46, even at a lower dose of 5uM (Fig 4b, representative figure of n = 3 biological replicates) which was mimicked by Wortmannin at the two doses tested (0.2nM and 20nM, Fig 4c, representative figure of n = 3 biological replicates). However, LY294002 did not reduce NKp46 at 20nM and only slightly reduced NKp46 at the 2uM dose (Fig 4c, representative figure of n = 3 biological replicates). This suggests that complete abrogation of the PI3K signaling is needed for the effect observed with lenalidomide.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Lenalidomide acts directly on NK cells to modulate the expression of CD56 and NKp46.

(a) Schematic representation of pathways targeted with inhibitors to understand how NKp46 is regulated. Created in BioRender. Eksioglu, E. (2025) https://BioRender.com/rg1nk09. (b) NKp46 expression by flow cytometry in YT cells after treatment with either 20μM or 5μM lenalidomide (figure representative of n = 3 biological replicates). Legend inside gate shows the percentage of events inside the gate shown from total live cells measured. (c) NKp46 expression by flow cytometry in YT cells after treatment with the PI3K inhibitors Wortmannin (doses tested 0.2nM and 20nM) and LY294002 (doses tested 20nM and 2μM). Figure representative of n = 3 biological replicates. Legend inside gate shows the percentage of events inside the gate shown from total live cells measured. (d) NKp46 expression by flow cytometry in YT cells after inhibition of other AKT targeting pathways: U0126 (MEK inhibitor, 0.1μM), AG490 (JAK2 inhibitor, 25μM) and API-2 (AKT inhibitor, 10μM). Figure representative of n = 3 biological replicates. (e) Mechanistic validation of PI3K signaling in YT cells infected with viral vectors expressing a CD56 (control), DNp110, or CAp110; followed by treatment with lenalidomide for 4 days. Graph shows the percent of NKp46 positive cells after transfection with either CD56 (control), CAp110 or DNp110 in the presence or absence of lenalidomide measured by flow cytometry. Error bars represent the SEM of n = 4 biological replicates and were analyzed by ordinary ANOVA with multiple comparison analysis. P values are shown as asterisk: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001, **** P ≤ 0.0001 as summarized in Graphpad Prism.

https://doi.org/10.1371/journal.pone.0344471.g004

Our group previously demonstrated that inhibition of PI3K in NK cells blocks MAPK kinase (MEK) and ERK activation suppresses NK cytotoxicity [35]. To assess if NKp46 reduction induced by PI3K inhibition is linked to MAPK inhibition we treated cells with U0126, a MEK inhibitor [41]. Inhibition of MEK did not affect the expression of NKp46 suggesting that while PI3K inhibition modulates NK function through MEK it does not affect NK receptor expression (Fig 4d, representative figure of n = 3 biological replicates). The other pathway linked to NK function leads to AKT activation via gene expression upregulation in JAK2 pathway which can be blocked by AG-490 (Tyrphostin) [42]. However, blocking JAK2 did not lead to a significant change to NKp46 protein expression at the cell surface in YT cells. Similarly, blocking directly AKT signaling with API-2 [43] did not change this expression. These data suggest that lenalidomide’s induction of NKp46 expression in the surface of NK cells is directly regulated by PI3K inhibition separate from the other NK function linked pathways (Fig 4d, representative figure of n = 3 biological replicates).

To directly validate the role of lenalidomide and PI3K regulation of NKp46 expression, we infected YT cells with vaccinia viral vectors expressing the constitutively active PI3K (CAp110) or a dominant negative PI3K (DNp110) prior to lenalidomide (Fig 4e). Lenalidomide significantly inhibited NKp46 expression in CAp110 expressing cells but not significantly in DNp110 expressing cells, consistent with a role of lenalidomide inhibiting NKp46 through the PI3K pathway. Interestingly, DNp110 also induced an increase in NKp46 positive cells, however, this was not significantly altered by lenalidomide. Lenalidomide had a significant inhibition of NKp46 suggesting that it uses a similar but separate process as PI3K p110 to reduce NKp46, as DNp110 was not capable of completely preventing the effect of lenalidomide on NKp46 expression (Fig 4e). This data provides evidence that lenalidomide reduces NKp46 through inhibition of PI3K on CD56^bright cells, but not CD56^dim.

Lenalidomide does not induce activating cytokines but increases production of cytolytic mediators

NK cells, especially CD56^bright cells, are potent cytokine producers contributing to the inflammatory microenvironment, regulate immune cell activation, and direct response to tumor cells [27,44]. Lenalidomide and thalidomide (its parent compound) are well known for their immunomodulatory effects [15,45]. Further, lenalidomide has diverse effects on cytokine secretion, depending on the environment and cell type [46,47]. We found that lenalidomide treatment of whole PBMCs resulted in decreased IL-10 and IL-4 but significantly increased IL-2 while IFNγ, IL-6 and IL-17 were not significantly changed (S3a Fig, n = 3 separate healthy donor PBMC each run in n = 3 independent determinations). Similar experiments in isolated primary NK cells revealed that secretions of IL-2 and IL-6 were significantly decreased by lenalidomide treatment while IFN-γ was significantly reduced only after 6 days of treatment (Fig 5a, n = 3 separate healthy-donor isolated NK cells each run in n = 3 independent determinations). Furthermore TNF-α was downregulated by lenalidomide in isolated primary NK (Fig 5a, n = 3 healthy donor isolated NK cells each run in n = 3 independent determinations). NK cells produce IFN-γ in response to tumor cells or other stimulatory conditions to stimulate an antitumorigenic immune response [48]. Therefore, we reasoned that without stimulation IFN-γ production might be unaffected by lenalidomide, as observed in PBMC and in isolated NK cells. Hence, we treated isolated healthy NK cells with lenalidomide for 7 days followed by a 5-hour pulse with K562 cells as tumor target cells and assess the intracellular level of IFN-γ in NK cells (Fig 5b, n = 3 separate healthy donor isolated NK cells). However, while exposure to tumor target cells enhanced the activation of NK cells, neither CD107⁺ granule mobilization nor IFN-γ production was significantly altered by lenalidomide.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Lenalidomide modulates NK cytokine production and enhances natural cytotoxicity.

(a). Concentration of cytokines secreted in the supernatants of isolated NK cells treated with 20μM lenalidomide (LEN) or vehicle measured at 1, 3 or 6 days of culture using a cytometric bead array. (b) Primary NK cells cultured in 10μM lenalidomide or DMSO for 7 days followed by pulsing with K562 tumor targets at a 1:1 ratio for 5 hours prior to staining for CD107 granule mobilization or IFN-γ activation by flow. Lysis of K562 targets cells as in “b” by either (c) PBMC (effector to target ratio 10:1), (d) isolated NK cells, or (Supplemental Fig. 4b) YT cells as effectors cultured in the presence or absence of 20μM lenalidomide, or DMSO, with or without hrIL-2 (100U/mL) 7 days after pulsing (effector to target ratio 10:1). (e) Lysis of 721.221 targets cells as in “d” in the presence of 20μM lenalidomide, or DMSO for 7 days after pulsing (effector to target ratio 5:1). (f) NK cells cultured as in “a” to measure the mRNA expression of Perforin (PRF1) and Granzyme B (GZMB) by q-PCR. Values were calculated by the ΔΔCt method with control (DMSO) as the experimental control and the housekeeping gene ACTB as the internal control. (g) Immunohistochemistry of perforin and granzyme B of isolated primary NK cells treated with lenalidomide, IL-2, IL-15 or their combination as shown. Quantification of mean fluorescence intensity (MFI) of (h) granzyme and (i) perforin immunostaining (n = 4) from representative figure shown in Fig 3g. In all experiments in this figure, error bars represent the SEM of the mean and p-value was calculated by paired student t test (a and e) or ordinary one-way ANOVA with multiple comparisons (b, c and d).

https://doi.org/10.1371/journal.pone.0344471.g005

Immune modulating features of lenalidomide include its ability to co-stimulate T cells [49] and enhance NK cell mediated cytotoxicity [6,7]. Lenalidomide by itself did not significantly enhance the killing of K562 target cells by PBMC (Fig 5c, n = 3 separate healthy donor PBMC) or YT cells (S3B Fig, n = 6 separate biological replicates) and only a slight yet significant increase in isolated NK cells (Fig 5d, n = 3 separate healthy donor isolated NK cells). Addition of IL-2 significantly increased the killing ability of PBMC and isolated NK cells (Fig 5c-d), although not in YT cells which are less sensitive to the need for this cytokine (S3b Fig), while combination with lenalidomide increased the killing of K562 target cells in all primary NK cells (Fig 5c-d). Similar results were observed in YT cells with the combination of lenalidomide with IL-2 significantly killing K562 target cells confirming our observations (S3b Fig). Considering the consistently increased cytotoxic activation by IL-2 with lenalidomide it suggests that exogenous addition of IL-2 can enhance NK activity during lenalidomide treatment. To validate our observation, we repeated the isolated NK cell experiment using 721.221 target cells and observed similar results to isolated NK cells targeting K562 cells (Fig 5e, n = 3 separate healthy donor isolated NK cells). We also tested IL-15 which showed the same level of cytotoxicity enhancement as IL-2 (Fig 5e), demonstrating the IL-2 family members have a similar behavior. Since lenalidomide enhanced the killing by NK cells, it may also stimulate the expression of perforin and/or granzyme B, as suggested by others [50]. We found that upon treatment gene expression of perforin (PRF1) and granzyme B (GZMB), evidenced by mRNA levels, increased significantly (Fig 5f, n = 3 separate healthy donor isolated NK cells each run in n = 2 independent determinations). Lenalidomide by itself only significantly increase granzyme protein expression in the primary isolated NK cells after immunostaining of treated cells as measured by mean fluorescence intensity (MFI, Fig 5g-h, n = 3 separate healthy donor isolated NK cells each measured in n = 10 separate images). Granzyme B expression was significantly enhanced by both IL-2 and IL-15 and both enhanced Perforin significantly compared to lenalidomide. Both members of the IL-2 family of cytokines enhanced the expression of both proteins when combined with lenalidomide. This again strengthens the idea that combination of lenalidomide with IL-2 family of cytokines IL-2 or IL-15 can enhance the function of NK cells after treatment.

Lenalidomide combination with IL-15 increases IL-2 receptor subunit expression

IL-2R signaling regulates perforin expression through the activation of its downstream mediator STAT5 [51,52]. Based on our observations, we wanted to determine if lenalidomide can also activate STAT5 in NK cells, despite not being needed for NKp46 expression. Lenalidomide treatment of PBMCs for 7 days induced a robust activation of STAT5 (Fig 6a, representative pictures of n = 3 separate healthy donor isolated NK cells, S4a Fig shows raw gels). Lenalidomide can induce IL-2 production, hence we investigated if IL-2 is contributing to an increase in STAT5 phosphorylation in PBMCs. Indeed, we found that STAT5 activation was further augmented after the addition of exogenous IL-2 to the PBMC culture (Fig 6a). To clarify this cytokine signaling interaction between T cells and NK cells, we isolated NK cells from healthy donors and treated them with IL-2 in addition to lenalidomide. As expected, while NK cells alone had a small amount of phosphorylated STAT5, addition of IL-2 induced a robust activation of STAT5 (Fig 6b, representative pictures of n = 3 separate healthy donor isolated NK cells S Fig 4b shows raw gels), suggesting that pairing lenalidomide with IL-2 leads to restoring full functionality to NK cells and enhancing the effects of lenalidomide alone. Since IL-15 belongs to the same family of cytokines as IL-2, sharing a common receptor, albeit with higher potency at simulating their cytokine expression [25,53], we also determined the effect of IL-15 activation in the activity of lenalidomide. We find that IL-15 enhances the STAT5 activation induced by lenalidomide (Fig 6b-c). This data was validated by flow cytometric analysis of pSTAT5 activation in isolated primary NK cells (Fig 6c, representative figure of n = 5 healthy donors), which was evenly distributed between proliferating and non-proliferating cells. However, while the protein levels of pSTAT5 after treatment with lenalidomide+IL15 were higher than IL-15 alone, the proportion of pSTAT5⁺ cells was not higher (Fig 6b-c), suggesting that while the number of cells with active pSTAT5 are not changed past ~20% the amount of pSTAT5 activity can get higher in each cell with the combined treatment.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Lenalidomide’s activation of STAT5 in NK cells is linked to differential effects on IL-2R subunits.

(a) Western blot analysis of activation of STAT5 and total STAT5 analysis of healthy PBMCs treated with 10μM lenalidomide (LEN) or DMSO, for 7 days stand alone or in combination with 100U/mL of IL-2. (b) Western blot analysis of activation of STAT5 and total STAT5 analysis of flow sorted healthy NK cells treated with 10μM lenalidomide, or DMSO, for 7 days in the presence of 100U/mL IL-2 or 10ng/mL IL-15. B-actin served as the loading control. (c) Representative flow cytometric analysis of pSTAT5⁺ isolated NK cells treated with lenalidomide alone or in combination with IL-2 or IL-15 (representative figure). (d) Analysis of IL-2Rα (CD25), IL-2Rβ or IL-2γ receptor chains by flow cytometric analysis of NK cells (CD3^-CD56⁺) or its subsets (CD56^bright, CD56^dim) in healthy PBMCs treated with 10μM lenalidomide, or DMSO, for 7 days in combination with 100U/mL of IL-2 or 10ng/mL of IL-15 as shown (representative figure of n = 3). To adjust data across samples to minimize technical variation (batch effects) for biological comparisons the Y-axis overlay normalization in FlowJo was used in this analysis as shown.

https://doi.org/10.1371/journal.pone.0344471.g006

Production and activation of STAT proteins is a highly regulated process [54]. Hence, we investigated how lenalidomide and or its combination with IL-2/IL-15 changes the expression of IL-2Rα, β and γ chains as a potential contributing factor to the enhanced STAT5 phosphorylation. We found that IL-2Rα expression was strongly induced by IL-15 alone or in combination with lenalidomide but minimally increased by lenalidomide alone and was unaffected by IL-2 (Fig 6d, representative figure of n = 5 healthy donors). In contrast, IL-2Rβ expression was enhanced by lenalidomide alone, lenalidomide plus IL-15 and IL-15 alone. IL-2 alone also increases expression, but this was reversed in the presence of lenalidomide. Lastly, IL-2Rγ responded similarly to IL-2Rβ in response to IL-15 and lenalidomide, however, IL-2 alone suppressed expression. This is in agreement with other work showing similar divergent effects of IL-2 and IL-15 cytokines on IL-2Rβ and γc subunit expression [55]. Similar trends in response were observed when the analysis was done on either CD56 subset. These results suggest that lenalidomide in combination with IL-15 is most effective in enhancing pSTAT5 and all three subunits of the IL-2R, potentially creating a positive feedback loop.