Mechanistic Insights into CpG DNA and IL-15 Synergy in Promoting B Cell Chronic Lymphocytic Leukemia Clonal Expansion

Abstract

Malignant cell growth within patients with B cell chronic lymphocytic leukemia (B-CLL) is largely restricted to lymphoid tissues, particularly lymph nodes. The recent in vitro finding that TLR-9 ligand (oligodeoxynucleotide [ODN]) and IL-15 exhibit strong synergy in promoting B-CLL growth may be particularly relevant to growth in these sites. This study shows IL-15-producing cells are prevalent within B-CLL-infiltrated lymph nodes and, using purified B-CLL cells from blood, investigates the mechanism for ODN and IL-15 synergy in driving B-CLL growth. ODN boosts baseline levels of phospho-RelA(S529) in B-CLL and promotes NF-κB-driven increases in IL15RA and IL2RB mRNA, followed by elevated IL-15Rα and IL-2/IL-15Rβ (CD122) protein. IL-15→CD122 signaling during a critical interval, 20 to 36-48 h following initial ODN exposure, is required for optimal induction of the cycling process. Furthermore, experiments with neutralizing anti-IL-15 and anti-CD122 mAbs indicate that clonal expansion requires continued IL-15/CD122 signaling during cycling. The latter is consistent with evidence of heightened IL2RB mRNA in the fraction of recently proliferated B-CLL cells within patient peripheral blood. Compromised ODN+IL-15 growth with limited cell density is consistent with a role for upregulated IL-15Rα in facilitating homotypic trans IL-15 signaling, although there may be other explanations. Together, the findings show that ODN and IL-15 elicit temporally distinct signals that function in a coordinated manner to drive B-CLL clonal expansion.

FIGURE 1. IL-15-producing cells are present within B-CLL-infiltrated LN

Five μm sections were taken from a fixed, paraffin-embedded LN determined to be B-CLL infiltrated by clinical pathology laboratory analysis and used for immunohistochemical assessment of IL-15-positive cells. Shown are photos (600× magnification) of a section stained with (left) goat anti-human IL-15 Ab or (right) normal goat IgG. The presence of IL-15-positive cells in B-CLL-infiltrated LN is consistent with findings of IL-15-positive cells in spleen (8).

FIGURE 2. IL-15 signaling does not become critical until ≥ 20 hours after B-CLL activation by ODN

CFSE-labeled cultures were stimulated with ODN at t=0 and provided with rIL-15 (15 ng/ml) either concomitantly or after varying intervals. A set of parallel ODN-stimulated cultures did not receive IL-15. B-CLL populations used for these experiments were selected on the basis of their relatively strong growth response to ODN+IL-15 in prior studies (8). A. CFSE-fluorescence histograms of cultured B-CLL (M-1031) gated for viable (V450-low) or dead (V450-high) cells (dot-pots at left) after 6 days of culture. Markers representative of division-gating are shown for each histogram. (CFSE fluorescence declines by 2-fold with each division). B. Calculated values for the total yield of viable B-CLL in each division subset within d6 cultures of (top) M-1031 cells and (bottom) M-693 cells are shown (mean ± sem of triplicate cultures). Total yield was determined with standardizing beads, as described in Materials & Methods. Left: Top and bottom plots show cell yield in cultures exposed to IL-15 alone, ODN alone, or both ODN+IL15 at t=0. Right: Top and bottom plots compare cell yield in ODN-stimulated cultures that received IL-15 at varying intervals (t=0 to 84h after ODN stimulation). C. Box-plots comparing the relative recovery of gated (left) undivided and (right) divided B-CLL cells, as a percent of the total cells collected (viable + dead). Experiments involved three distinct B-CLL populations (M-1031, M-693, and U-321) that were cultured for 6-7 days with ODN alone or ODN with IL-15 added at varying intervals. Statistical evaluation involved one-way ANOVA using the Holm-Sidak method for multiple comparisons versus a control group. * indicates significant P values (p < 0.05) upon comparing cell recovery in ODN-only cultures with cell recovery in IL-15-supplemented cultures. # indicates significant P values upon comparing recovery in cultures with ODN+IL15 at t=0 with recovery in cultures receiving IL-15 at later intervals. Cultures supplemented with IL-15 at t=20h were not significantly different from those receiving IL-15 at t=0h in terms of recovery of undivided or divided cells. D. Box plots comparing the relative viability within gated (left) undivided and (right) divided B-CLL cells in ODN-stimulated M-1031, M-693, and U-321 cultures without IL-15 or with IL-15 received at varying intervals. Paired, one-sided T-test indicated statistical differences in viability between groups receiving IL-15 at t=0h and groups with IL-15 delayed to 48-70h (P ≤0.03). E. Box plots comparing recovery of total viable progeny within the above cultures of M-693 (●), U-321 (○), and M-1031 (▼), expressed as % of total viable+dead cells collected. Statistical evaluations were performed as in C, above. F. Graph showing absolute yield of viable cells within each division subset in cultures of CFSE-labeled M-1031 cells stimulated with ODN+IL-15 (5 ng/ml) at t=0 with/without anti-IL-15 mAb added at t=0 or t=20h (mean ± sem of triplicate cultures). Similar results were obtained in a replicate experiment. G. Anti-IL-15 mAb (clone ct2nu; 3 μg/ml final) blocks U-CLL770 blastogenesis upon t=0 addition to cultures stimulated with ODN+IL-15. Photographs were taken at 400× magnification using phase microscopy.

FIGURE 3. B-CLL exposure to ODN upregulates receptors for IL-15: IL-15Rα (CD215) and IL-2/IL- 15Rβ (CD122)

A. Fluorescence histograms of U-996 cells stained with anti-IL15Rα Ab (grey fill) or IgG control (dashed line) after varying culture intervals (3- 20h) with medium alone or with ODN. Histograms represent cells gated for viability (dot-plots to left). B. Fluorescence histograms of viability-gated U-996 cells stained with anti-CD122 Ab (grey fill) or IgG control (dashed line) following 20h culture with medium alone (top) or ODN (bottom). C. Level of IL-15Rα staining in 10 distinct B-CLL clones following 20h culture in medium (light grey bars) or ODN (black bars). Values represent the mean ± sem of staining replicates. IL-15Rα expression is quantified on the basis of %-positive cells, with a threshold set at 5% positive within the IgG-control-stained population (dotted line). IL-15Rα staining above background was noted in 9/10 B-CLL clones cultured in medium alone; in 8/10 of these B-CLL, levels of IL-15Rα rose upon ODN activation. (Note: the 2 B-CLL which did not show an ODN-induced increase in IL-15Rα staining (U-1991 and M-2018) at 20h did show a transient increase at earlier intervals (see E, below). D. Level of CD122 staining in 5 distinct B-CLL populations following 20h or 44h culture with/without ODN. CD122 expression was quantified similarly to IL-15Rα in C, above. E. Kinetics of IL-15Rα expression (as % positive cells) in 9 distinct B-CLL evaluated at varying times after culture with medium alone (white circles) or with ODN (black circles); mean ± sem of staining replicates is shown. F. Summarized data from IL-15Rα staining analyses with 10 B-CLL (four left box plots) or with U-CLL and M-CLL subsets segregated (four right box plots). Four left box plots with total B-CLL reveal the % of cells above the threshold for positivity within medium- or ODN-cultures stained with IgG control or anti-IL-15Rα Ab. Statistical significance was determined by 2-sided, paired T-tests. G. Box plots comparing % positive cells in CLL clones stained with anti-CD122 Ab or IgG control after 20-44h culture with medium only or ODN (n=5). ODN-stimulated B-CLL showed a significant increase in % CD122-positive cells (P= 0.04).

FIGURE 4. ODN-stimulated B-CLL show rapid increases in IL15RA and IL2RB (CD122) mRNA that are blocked by NF-kB inhibitor

A. B-CLL clones (U-791, U-1239, U-1953) were cultured for varying intervals (3h, 9h, and 20h) with medium alone or with ODN, prior to RNA isolation. Specific mRNA levels were determined by qRT-PCR performed as described in Materials & Methods, using GAPDH housekeeping control mRNA to calculate ∆Ct values. ∆Ct values within medium versus ODN treated B-CLL cultures for IL15RA (left plot) and IL2RB (CD122) (right plot) are shown. Statistical evaluations for differences were made with paired T-tests. * indicates that ∆Ct values for specific mRNA in medium-treated cells is significantly different (P≤ 0.03) by 1-sided T-tests; ** indicates that ∆Ct values in medium versus ODN treated cells is significantly different (P≤ 0.04) by 2-sided T-tests. B. ∆Ct values were used to calculate ∆∆Ct values, using medium-treated cells as control baseline and fold-change above medium only obtained by 2^(−∆∆Ct) (31). Plotted are the fold-change values for IL15RA mRNA (left) and IL2RB (right) at increasing intervals after ODN stimulation. Represented are data from the 3 individual B-CLL clones. C. Intracellular immunofluorescent staining with APC-labeled anti-p65 NFkB(Ser529) mAb (solid line) and APC-labeled IgG control (filled grey) was used to confirm that p65/RelA is activated in ODN-2006-exposed B-CLL. A gate was placed in the fluorescence histogram to indicate the threshold for positivity (set at 1% positive cells with IgG staining control). P value indicates statistical significance by 1-sided, paired T-test. E and F. NF-kB inhibitor reduces IL15RA and IL2RB (CD122) mRNA. The effect of 30 min pre-exposure to NF-kB inhibitor, BAY-11-9802 (1.25 and/or 2.5 μM) on ODN-triggered increases in specific mRNA was evaluated in two B-CLL clones: (E) M-1031 and (F) U-1239. M-1031 CLL cells were cultured for 3h with medium alone or ODN prior to mRNA isolation, with sets of parallel cultures pre-exposed for 30 min to NF-kB inhibitor or DMSO vehicle control. Specific RNA levels were determined as above and are represented as fold-change above that in medium only cultures with vehicle (shown by dotted line). U-1239 CLL cells were treated as above, but two culture periods were employed (3h and 9h) to permit optimal detection of IL2RB mRNA. * indicates that specific mRNA levels in cultures with NF-kB inhibitor are significantly different from those in parallel cultures with vehicle: P< 0.05 upon data analysis by 2-sided, unpaired T-tests.

FIGURE 5. Neutralizing mAbs specific for IL-15 or CD122 block B-CLL division both when added early and late after ODN+IL-15 stimulation

B-CLL populations used for these experiments were selected on the basis of their relatively strong growth response to ODN+IL-15. A. CFSE fluorescence histograms from 132h-harvested M-CLL2018 cultures, with dot-blots revealing gating for viable/dead B-CLL cells to the left. Cultures exposed to the stimuli shown, with anti-IL-15 or anti-CD122 neutralizing mAb pulsed into sets of ODN+IL-15-stimulated cultures, at varying times after activation (anti-IL-15: clone ct2nu; 3 μg/ml final and anti-CD122: clone TU27; 25 μg/ml final) Histograms of viable-gated cells are overlaid with histograms of dead-gated cells from the same culture, revealing the proportion of viable and dead cells within each division subset. B. Box plots representing results from similar experiments with multiple B-CLL clones, shown in the legend. The yield of viable progeny (cells with CFSE evidence of division) is expressed as a percent of that obtained in ODN+IL-15-stimulated cultures without Ab. Isolated symbols represent yields in the individual clones. TOP: Yields of viable progeny from groups receiving anti-IL-15 at 0, 72, and 96h after ODN+IL-15 stimulation (n=6 B-CLL for all, excepting n=4 for the t=96 group) was found in each case to be significantly different from yield in ODN+IL-15-stimulated cultures without mAb (as assessed by one-way ANOVA, followed by use of the Holm-Sidak method for making multiple comparisons versus a control group; P < 0.001 for each comparison; indicated by *). BOTTOM: Yields from B-CLL groups receiving anti-CD122 at 0, 72, and 96h after ODN+IL15 stimulation and the ODN+IL-15 group with no mAb were initially evaluated by the non-parametric Kruskal-Wallis one way analysis of variance on ranks, showing a statistically significant difference exists between all the groups (P=0.001). When the Dunn’s Method was used to subsequently make multiple comparisons against the control group with no Ab, a statistically significant difference was noted for the group receiving anti-CD122 mAb at t=0h (P<0.001 = *) and a P value of borderline significance (P=0.06) was noted for the group with anti-CD122 mAb at t=72h (n=3 B-CLL for the t=96h pulse, while n=5 for other groups). C. Box plots representing the viability of divided cells within the ODN+IL-15-stimulated B-CLL populations evaluated in (B) above. Values for divided cell viability within the control group without mAb are shown (left-most bar). TOP: While divided cells within the B-CLL groups receiving IL-15 neutralizing mAb generally showed lesser viability that the control group, this was not determined to be statistically significant by one-way ANOVA. However, if the viability data from the various B-CLL experiments were normalized and expressed as a % of the viability within the control group without mAb, then significance (P=0.015, indicated as (*), or borderline significance (P=0.06) was noted for the groups treated with anti-IL-15 at t=0 and t=72h, respectively. BOTTOM: When viability data on divided cells within the B-CLL groups receiving anti-CD122 mAb were compared to that in ODN+IL-15-stimulated cultures without Ab, a statistically significant decline in viability in the t=0h mAb-treated group (P=0.03) was noted by one-way ANOVA, followed by the Holm-Sidak test for multiple comparisons. Evidence for a statistically-significant decline in viability in the t=72h mAb-treated group was only observed upon normalizing the viability values (as above) and application of the Krushal-Wallis one way analysis of variance on ranks, followed by the Dunn’s method for making multiple comparisons. In the latter case P=0.02, which is indicated in the graph as (*). D & E. The yield of viable progeny within each division subset is shown for experiments with 3 B-CLL clones (U-1993, M-2018, and M-1031). D) Shown is the total yield of viable divided cells at culture termination, as determined with standardization beads, in B-CLL populations stimulated with ODN alone or ODN+IL-15 ± anti-IL-15 mAb at t=0 or t=96h after stimulation. Data are expressed as the mean ± sem of triplicate cultures. The asterisk linked to M-1031 serves to indicate that in this experiment, the B-CLL clone was stimulated with a suboptimal dose of IL-15 (5 ng/ml), rather than the routine 15 ng/ml used in other experiments. E) To better appreciate the impact of delayed blockade of the IL-15-signaling pathway on B-CLL clonal expansion, values for absolute yield of viable progeny in D) are expressed as a percent of that in ODN+IL-15-stimulated cultures without mAb.

FIGURE 6. Recently divided B-CLL cells in patient PB show elevated levels of IL2RB mRNA

Levels of IL2RB mRNA were determined by gene expression array analysis (Materials & Methods) of RNA isolated from various B-CLL populations. (A and B) Purified B cells from the PB of patients with M-CLL or U-CLL were sorted by FACS on the basis of CXCR4^DimCD5^Bright (proliferative fraction, PF) or CXCR4^DimCD5^Bright (resting fraction, RF) expression, as described (32). RNA from each of the above, in replicates, was assessed for average IL2RB signal (B), with a ratio of the IL2RB signal in PF versus RF for total CLL (left cluster), U-CLL (middle cluster) and M-CLL (right cluster) shown in (A).

FIGURE 7. Extended B-CLL cycling following ODN+IL-15 stimulation is best observed under conditions of high B-CLL density

A. CFSE-labeled B-CLL clones (n=6) were stimulated at varying cell densities, ranging from 6250 to 100,000 cells per 200 μl flat-bottomed well. Recovery of viable blasts per each division is plotted as a percent of total viable cells recovered at culture harvest (day 5-7); shown is the mean of triplicate cultures per each condition. B. CFSE-labeled cells from U-770 cells (right) and a second experiment with M-618 cells (left) were seeded at varying densities into wells stimulated with either ODN + IL-15 or ODN + IL-2 (cytokines present at 15 ng/ml). Statistical significance was determined with a 2-sided, unpaired T-test. The diminished yield of viable U-770 progeny at high plating density (100,000 cells/well) was associated with increased acidity (overgrowth) and diminished viability of cultures (seen with both IL-15 and IL-2).