Toll-like receptors on B-CLL cells: expression and functional consequences of their stimulation

Abstract

Toll-like receptor (TLR) stimulation plays a crucial role in the homeostasis of human B cells. We investigated the expression of TLRs 1–9 on the cells of B-cell chronic lymphocytic leukemia (B-CLL) and analyzed the functional consequences of TLR stimulation on leukemic cells. We showed that B-CLL cells express similar set of TLRs as memory B cells of healthy donors, i.e. TLR-1, TLR-2, TLR-6, TLR-7 and TLR-9. However, in contrast to memory B cells, B-CLL cells lack TLR-4 expression. Expression of TLRs correlates with their capacity to respond to specific TLR agonists. At the level of phenotype, ODN2006 (TLR-9 agonist) is the most potent stimulus. B-CLL cells also respond to the stimulation with loxoribine, Pam3CSK4 and MALP-2 (TLR-7, TLR1/TLR2 and TLR2/TLR6 agonists, respectively). TLR-7 and TLR-9 stimulation induces production of IL-6 and TNFα. In 47% of tested patients, treatment with ODN2006, MALP-2 and Pam3CSK4 reduced leukemic cells survival. Stimulation of B-CLL cells with TLR-9 agonists, loxoribine, MALP-2 and Pam3CSK4 induces significant proliferation. We report that TLR stimulation induces expression of CD38, a negative prognostic marker, on B-CLL cells. Expression of CD38 is induced by direct stimulation of B-CLL cells through TLR-7 and TLR-9 or CD38 can be induced on B-CLL cells indirectly by a soluble factor induced in non-B-CLL cells after stimulation with TLR-2, TLR-3 or TLR-5 agonists; the nature of this factor remains unknown. Our results argue for cautious evaluation of immunointervention strategies based on the administration of TLR agonists in the treatment of B-CLL as their effects on B-CLL cells may be tumor promoting as well as tumor suppressing.

B-cell chronic lymphocytic leukemia (B-CLL) is one of the most common hematological malignancies with the highest prevalence in patients older than 50 years. The disease is characterized by extremely variable clinical course with survival from months to decades.1, 2 B-CLL is characterized by the progressive accumulation of CD19+, CD5+ B cells with a low proliferative index and prolonged cell survival. Patients with unmutated IgVH genes have significantly shorter median survival than those with mutated IgVH genes. Surface expression of CD38 and expression of ZAP-70 are additional prognostic markers associated with a poor clinical outcome.3

Toll-like receptors (TLRs) are pattern recognition receptors that trigger innate immunity. Signaling via TLRs plays a crucial role in a defense against pathogens. Cells of the innate immune system expressing these receptors recognize conserved pathogen structures and initiate activation of adaptive immune response.4, 5

Besides antibody production, B cells have important antigen presenting functions.6 Mature B cells express costimulatory molecules, produce inflammatory cytokines and secrete factors directly inducing microbe destruction.7, 8

Toll-like receptor signaling plays an important role in the biology of B cells. They might participate in the regulation of B-cell differentiation process.9 Toll-like receptor stimulation is required as a third signal for the activation of human naive B cells.10 In human naive B cells, TLRs are expressed at low levels, but the expression of some TLRs is induced upon B-cell receptor (BCR) triggering. Consequently, memory B cells express several TLRs at constitutively high levels. It has been suggested that constitutive expression of TLR-9 in memory B-cell pool allows polyclonal activation of the entire memory pool and is crucial for the long-term maintenance of long-lived memory B cells.10–12 Besides TLR-9 other TLR agonists were described to induce B-cell activation: several TLR-7 agonists, MALP-2 (macrophages activation lipopeptide 2) or synthetic lipopetid Pam3CSK4, agonists of TLR2/6 and TLR1/2 heterodimer, respectively, activate murine and human B cells.13, 14 In addition to the induction of BCR-independent proliferation of memory B cells, these agonists also increase the immunogenicity of B cells by upregulating costimulatory molecules.

Given the fact that B-CLL cells share some characteristics with their normal counterparts, investigation of TLR's expression, functional consequences of their stimulation and its importance for biology of B-CLL cells is of great interest. Until recently, the majority of the studies focused on TLR7 and TLR-9. It has been suggested that the stimulation of TLRs expressed on B-CLL cells could increase immunogenicity of tumor cells and thus potentially contribute to the induction of leukemia-specific immune response.15–18 Several studies also suggested that TLR-9 agonists decrease viability of B-CLL cells and increase susceptibility of B-CLL cells to apoptosis in the cell culture.19, 20 However, a concise study including the analysis of TLRs expression pattern, comparison with normal B cells and functional consequences of TLRs stimulation on B-CLL cells has not yet been performed.

In our study, we thus present a complex analysis of TLRs expression on B-CLL cells. We also show phenotypic and functional consequences of TLRs stimulation on leukemic cells. Our results argue for cautious evaluation of immune intervention strategies based on the administration of TLRs agonists in the treatment of B-CLL. TLR stimulation of B-CLL cells results in complex changes in B-CLL cells biology and has both antitumor and tumor-promoting effects. While stimulation with TLR agonists induces the apoptosis of B-CLL cells in the subset of patients and increases the immunogenicity of tumor cells, it also promotes the proliferation of leukemic blasts and the expression of CD38, a marker associated with a negative outcome.

Material and Methods

Media and cell cultures

Complete culture medium was used for the culture of B cells and B-CLL cells and consisted of RPMI 1640 (Cambrex, Verviers, Belgium) supplemented with 10% heat-inactivated fetal bovine serum (Cambrex), 2 mM L-glutamine (Cambrex) and 1% penicillin/streptomycin (Cambrex). Cells were cultured at 37°C in a 5% CO₂ atmosphere.

Patients and controls

In our study, 17 B-CLL patients (11 men and 6 women) and 8 healthy donors (4 men and 4 women) were tested. Other 6 patients (2 men and 4 women, no. 18–23) were used in CD38 experiments. The patients' characteristic and clinical data are detailed in Table 1. All patients were untreated or off the treatment for at least 3 months before blood sampling. All patients and healthy controls participating in the study were informed in details about the study and informed consents were signed prior to their enrolment. The study design was approved by the Institutional Review Board and Ethical Committee of the Charles University, 2nd Medical School.

Table 1. Clinical characteristics of CLL patients

C, chlorambucil; P, prednisone; F, fludarabine; R, rituxan; Cy, cyclophosphamide; H, adriamycin; O, vincristine; NA, not analyzed.

B cells and B-CLL cells isolation and stimulation in a culture

Peripheral mononuclear cells were isolated from peripheral blood of B-CLL patients or healthy donors by density gradient centrifugation (Ficoll paque; Amersham Biosciences, Uppsala, Sweden). For mRNA expression experiments, CD19+CD27+ or CD19+CD27− populations from healthy donors and CD19+CD5+ B-CLL cells were sorted using FACS Aria with more than 98% purity. For stimulation experiments, PBMCs from B-CLL patients were used unseparated (50–90% of B-CLL cells) or were isolated from PBMCs by CD19 Microbeads with more than 98% purity. In some experiments, B-CLL were directly isolated from peripheral blood by B cells Rosettesep (Stem Cell Technologies, Vancouver, Canada) with more than 95% purity. For TLR stimulation experiments, B cells from healthy donors were first enriched by using B cells Rosettesep (Stem Cell) and then positively separated by CD19 Microbeads (Miltenyi Biotech, Bergisch Gladbach, Germany) with more then 95% purity. B-CLL cells or purified B cells at the concentration 1 × 10⁶ cells per ml media per well were stimulated with purified TLR agonists as previously described.21 LPS (Sigma): 1 μg/ml; Poly(I:C) (Sigma): 50 μg/ml; CpG ODN2006 and CpG ODN 2006K (synthesis at Tib-Mol biol): 5 μg/ml; flagellin (Invivogen, San Diego, CA): 1 μg/ml; ssPolyU complexed with Lyovec (Invivogen): 10 μg/ml; zymosan (Invivogen): 10 μg/ml; loxoribine (Invivogen): 500 μM; peptidoglycan Staphylococcus aureus (Invivogen): 5 μg/ml; lipoteichoic acid from S. aureus (Invivogen): 2 μg/ml; Pam3CSK4 (Invivogen): 1 μg/ml; MALP-2 (Alexis Biochemicals, Farmingdale, NY): 1 μg/ml.

Cytokine detection

Culture supernatants from stimulated B-CLL cells or B cells from healthy donors were collected on day 5 of stimulation. TNFα and IL-6 production was measured with ELISA kits (Cytosets; Biosource, Camarillo, CA).

Flow cytometry

Fluorochrome-conjugated monoclonal antibodies against the following molecules were used: CD19-PE (Miltenyi) or CD19 PE-Cy7 (Immunotech, Marseille, France), CD5-PE-Cy7, CD40-PE, CD80-FITC, HLA-DR-PE (Immunotech), CD86-PE-Cy5 (BD Biosciences, San Jose, CA), CD38 Alexa 700 (Exbio, Prague, Czech Republic). Cells were stained for surface markers in PBS + 1% FBS for 30 min at 4°C and washed in PBS. For apoptosis analysis, the cells were stimulated 2 days with TLR agonists and stained with Annexin-Dy647 according to the manufacturer's protocol (Exbio). The samples were measured on FACS Aria (BD Biosciences) using FACS Diva Software. The cells were gated according to their FSC and SSC and CD19+ B cells or CD19+CD5+ B-CLL cells were selected for further analysis (FlowJo, TreeStar, CA). Dead cells were discriminated by DAPI staining (Invitrogen, Carlsbad, CA).

Analysis of cell proliferation

Proliferation was tested by 3H-thymidine incorporation. Stimulation with TLR agonists was done in triplicates. On day 4, 3H-thymidine (0.04 MBq/100 μl) was added. After 18 hr, the proliferative capacity of the cells was determined using a beta-counter (Wallace, Finland) as the counts per minute (cpm). Further proliferation was measured by CFSE dilution. Cells were washed in PBS and stained with 1 μM CFSE using Vybrant CFDA SE Celltracer kit (Molecular Probes, Eugene, OR) for 8 min. The reaction was blocked with equal volume of FBS and the cells were twice washed in PBS + 10% FBS. CFSE-stained cells were stimulated with Toll-like receptor agonists for 5 days and the intensity of proliferation was measured by flow cytometry.

Reverse transcription–polymerase chain reaction

Total cellular RNA was prepared from B-CLL or B cells using Rneasy mini or micro kit (Qiagen, Hilden, Germany). One microgram of RNA was treated by 1 U of DNase (Fermentas, St. Leon-Rot, Germany) for 30 min at 37°C. The enzyme was inactivated by heating at 65°C for 10 min in the presence of 2 mM EDTA (Fermentas). RNA was then transcribed into cDNA using M-MLV Reverse Transcriptase (Invitrogen) according to the manufacturer's instructions. Presence of DNA contamination and verification of successful isolation of RNA was tested by PCR amplification of ABL gene with intron spanning primers: forward primer 5′-TTC AGC GGC CAG TAG CAT CTG ACT T-3′, reverse primer 5′-TGT GAT TAT AGC CTA AGA CCC GGA GCT TTT-3′. This reaction results in a 764-bp product when DNA is the template and a 201-bp product in RNA. cDNA was synthesized using Superscript First Strand Synthesis System for reverse transcription–polymerase chain reaction (RT-PCR; Invitrogen).

The following primers were used: TLR2: reverse 5′-CCAGG TAGGTCTTGGTGTTCA-3′, forward 5′-GGCCAGCAAATTA CCTGTGTG-3′; TLR4: reverse 5′-TCCCACTCCAGG TA AGTGTT-3′, forward 5′-CTGCAATGGATCAAGGACCA-3′; TLR8: reverse 5′-AATGTCACAGGTGCATTCAAAGGG-3′, forward 5′-CAGAATAGCAGGCGTAACACATCA-3′. To detect TLR1, TLR3, TLR5, TLR6, TLR7 and TLR9 gene expression, PCR was performed with primers described previously.22 The amplified products were separated on 2% agarose gel and visualized by ethidium bromide staining.

The expression of TLR1, TLR2, TLR6, TLR7 and TLR9 was assessed by real-time quantitative RT-PCR using iCycler (Bio-Rad, Hercules, CA). The following primers were used: TLR1: 5′-ACG AGG AAG AGG GCC TGG TA-3′, 5′-ACT CCC GGA GGC AAT GCT-3′, Taqman probe 5′-TTC ATG AAG ACC CTG GCC ACA AAA ACA G-3′; TLR2: 5′-AGG CGG ACA TCC TGA ACC T-3′, 5′-GGC CAG CAA ATT ACC TGT GTG-3′, Taqman probe 5′-CTC CAT CCC ATG TGC GTG GCC T-3′; TLR6: 5′-GAA GAA GAA CAA CCC TTT AGG ATA GC-3′, 5′-AAG GCA AAC AAA ATG GAA GCT T-3′, Taqman probe 5′-TGC AAC ATC ATG ACC AAA GAC AAA GAA CCT ATT-3′; TLR7: 5′-AAG CTG ATC TTG GCA CCT CTC-3′, 5′-AGA ATT TGT CTC TTC AGT GTC CAC-3′, Taqman probe 5′-TGC TCT CTT CAA CCA GAC CTC TAC ATT CCA-3′; TLR9: 5′-TGA AGA CTT CAG GCC CAA CTG-3′, 5′-TGC ACG GTC ACC AGG TTG T-3′, Taqman probe 5′-AGC ACC CTC AAC TTC ACC TTG GAT CTG TC T-3′; β-actin: 5′-GCT GAT CCA CAT CTG CTG GAA-3′, 5′-ATT GCC GAC AGG ATG CAG AA-3′, Taqman probe 5′-CAA GAT CAT TGC TCC TCC TGA GCG CA-3′. The reactions were amplified and quantified using iCycler (Bio-Rad). The relative quantity of TLR genes was normalized using β-actin reaction as exogenous control.

Statistics

In most experiments, nonparametric Wilcoxon paired test was used to compare the characteristics of B-CLL cells treated or untreated with TLRs agonists. Nonparametric Mann–Whitney test was used for the comparison of cytokines production by B-CLL cells and normal B cells. A p value of <0.05 was considered to be significant.

Results

TLRs expression in normal versus malignant B cells

First, we decided to analyze TLR expression profile in B-CLL cells. We analyzed the expression of TLR-1 through TLR-9 in 10 B-CLL patients by RT-PCR and found the expression of TLR-1, TLR-2, TLR-6, TLR-7 and TLR-9 (Figs. 1a and 1b). To compare TLR expression profile of B-CLL cells with normal B cells, we sorted peripheral blood B cells of healthy donors into CD27+ memory B cells and CD27− naive B cells populations and analyzed TLR expression. TLR expression profile on B-CLL cells resembled TLR expression in memory CD19+CD27+ B cells, i.e. high levels of TLR-7 and TLR-9. However, in contrast to memory B cells, B-CLL cells lacked TLR-4 and expressed only low levels of TLR-6, similarly to naive B cells (Fig. 1). Quantitative real-time PCR revealed that TLR-2 and TLR-7 were expressed at higher levels in B-CLL cells than in memory B cells (Fig. 1b). TLR9 expression levels in B-CLL cells were higher than that in naive B cells but still lower than that in memory B-cells.

Figure 1. B-CLL cells express similar pattern of TLRs as memory B cells and respond to the stimulation with TLR agonists. mRNA expression of TLR receptors in malignant B-CLL cells and in naive (CD27−) and memory B cells (CD27+) from healthy donors. B and B-CLL cells were sorted by FACS with more than 98% purity. (a) TLR1-9 expression detected by RT-PCR and visualized by gel electrophoresis. Representative data from 2 B-CLL patients and 1 healthy donor are shown. (b) Quantification of TLR1, TLR2, TLR6, TLR7 and TLR9 expression in B cells (2 donors) and B-CLL cells (7 donors). TLR expression levels were normalized to the expression of β-actin and mean levels with SD are shown.

Phenotype changes in leukemic and nonmalignant B cells induced by Toll-like receptor agonists

To analyze the consequences of TLR stimulation on B-CLL cells, we stimulated B-CLL cells with TLR agonists and assessed the expression of CD40, CD86, CD80 and HLA-DR molecules. In accordance with the real-time PCR expression data, TLR-9 agonists ODN2006 and ODN2006C were the most efficient in upregulating the expression of evaluated markers, followed by loxoribine (a TLR-7 agonist), macrophage activating lipopeptide 2 (MALP2, TLR2/6 agonist) and Pam3CSK4 (TLR1/2 agonist). Other TLR2 agonists (zymosan, peptidoglycan, lipoteichoic acid) had only little effect on the activation markers' expression. Also, stimulation with LPS (TLR4 agonist), Poly(I:C) (TLR3 agonist), ssRNA (TLR8) and flagellin (TLR5) had no or little effect. LPS and Poly(I:C) induced a slight increase of CD86 expression. However, the upregulation of activation markers induced by CpG, loxoribine, MALP-2 and Pam3CSK4 was 2- to 4-fold higher and was detectable for all activation markers tested. Thus, the expression of TLRs on B-CLL cells correlates with their capacity to respond to the stimulation with respective TLRs ligands (Fig. 2a). Stimulation of normal B cells with the same set of TLR agonists resulted in similar phenotype changes as in leukemic cells. Significant upregulation of CD40, CD86 (Fig. 2b), CD80 and HLA-DR (data not shown) was induced by ODN2006, ODN2006C, loxoribine, MALP-2 and Pam3CSK4.

Figure 2. TLR agonists induce phenotypic changes in B-CLL cells. B-CLL cells (n = 17) and sorted healthy B cells (n = 8) were stimulated with TLR agonists. Expression of CD40, HLA-DR, CD80 and CD86 was analyzed in CD19+ population. Relative expression of analyzed molecules compared to the unstimulated sample is shown. Data were analyzed by Wilcoxon paired test. Statistical significance is indicated with 1 (p < 0.05) or 2 asterisks (p < 0.01). Bars represent median values of all tested samples with 95% confidence interval. (a) Malignant B-CLL cells isolated from the peripheral blood of 18 B-CLL patients. (b) B cells isolated from 6 healthy donors.

Analysis of cytokine production in B-CLL and healthy B cells

Besides phenotype changes, TLR stimulation also leads to the production of various cytokines. We thus measured TNFα and IL-6 production by sorted B-CLL cells and normal B cells after TLR stimulation. In general, B-CLL cells were more heterogeneous in the amount of produced cytokines. TLR-9 agonists induced TNFα production in both B-CLL and normal B cells, whereas loxoribine, a TLR-7 agonist, induced TNFα in B-CLL cells only (Fig. 3a). Interestingly, MALP-2 or Pam3CSK4 did not induce any significant TNFα production despite the fact that their receptors were expressed on B-CLL cells and they induced phenotypic changes. Significant amounts of IL-6 were induced in normal B cells mainly after CpG stimulation in contrast to B-CLL cells that only produced very low quantities of IL-6. Figure 3b shows cytokine production data in individual patients and healthy controls for the most potent agonists, loxoribine, ODN2006 and ODN2006C. B-CLL cells produced significantly more TNFα and IL-6 than normal B cells after loxoribine stimulation. However, normal B cells produced higher amounts of IL-6 after stimulation with TLR-9 ligands. Interestingly, most B-CLL samples also produced significantly less IL-6 after nonspecific PMA + ionomycin stimulation while levels of TNFα did not differ between malignant and nonmalignant cells (Fig. 3b). These results indicate that failure to produce IL-6 could result from an intrinsic defect of B-CLL cells.

Figure 3. Stimulation of B-CLL (n = 17) cells and healthy B cells (n = 8) with TLR agonists induces cytokine production. Cells were isolated by magnetic separation (98% purity) and stimulated by TLR agonists. Concentrations of TNFα and IL-6 were measured in the cell culture supernatants. (a) Median values for TNFα and IL-6 produced by B-CLL patients and healthy B cells. (b) Detailed data for individual B-CLL patients and healthy controls. For statistical analysis nonparametric Mann–Whitney test was used and p values are shown.

B-CLL cells survival after TLR agonists treatment

The effect of TLR agonists on the survival of B-CLL cells in culture was heterogeneous. We detected 2 distinct patterns after stimulation of B-CLL cells with TLRs agonists. In 53% of the patients neither ODN2006 nor other TLR ligands had any effect on viability (Fig. 4a, left column). However, in 47% of patients, the treatment with ODN2006, ODN2006C, MALP-2 and Pam3CSK4 significantly reduced cell survival (viability reduced below 70% in comparison to the unstimulated control) (Fig. 4a, right column). Interestingly, stimulation by TLR-7 agonist, loxoribine, had only negligible effect on leukemic cells survival. Based on annexin V staining, cells underwent apoptosis (Fig. 4c). However, blocking of Fas-Fas-L interaction with anti-FAS antibody did not improve B-CLL cells survival during the 2 days of culture. We also did not observe any effect of anti-TNFα blocking antibody (data not shown). Effect of TLR stimulation on the survival of nonmalignant B cells was completely different as TLR stimulation significantly improved normal B-cell survival (Fig. 4b). Interestingly, unstimulated B-CLL cells had significantly better cell survival after 5 days of culture when compared to normal B cells (Figs. 4a and 4b).

Figure 4. Stimulation of B-CLL cells with TLR agonists affects cell survival. (a) B-CLL cells viability after 5 days of culture in the presence of TLR agonists analyzed by DAPI staining. Two patterns of B-CLL cells survival are illustrated by representative experiments. In 10 of 17 patients (53%) TLR stimulation had a negligible effect on the survival (left column) while in remaining 7 patients (47%) stimulation with some TLR agonists reduced survival below 70% (right column). (b) Viability of normal B cells isolated from healthy controls (n = 8) after 5 days stimulation with TLR ligands. (c) Annexin V/DAPI staining of B-CLL cells undergoing apoptosis after 2 days stimulation with TLR ligands to determine the stage of apoptosis. Percentages of alive (annexin V−, DAPI−), early apoptotic (annexin V+, DAPI−) and late apoptotic (annexin V+, DAPI+) cells are shown.

Cell proliferation

We also tested the effect of TLRs stimulation on the proliferation of B-CLL cells. TLR-9 agonists, MALP-2 and Pam3CSK4 induced significant proliferation of both B-CLL cells and normal B cells (Fig. 5a). Similar to cytokine production, loxoribine stimulation only induced proliferation of leukemic B-CLL cells and had no effect on normal B cells. Despite detectable proliferation of B-CLL cells induced by TLRs stimulation it was evident that most leukemic cells only divided once in contrast to more sustained multiple division induced in non-malignant B-cell population (Fig. 5b).

Figure 5. TLR agonists induce proliferation of B-CLL cells. (a) H3 thymidine uptake was measured in triplicates 72 hr after the stimulation of B-CLL and healthy B cells treated with TLR agonists. Mean values ± SD are depicted. Results from 1 of the 3 representative experiments are shown. (b) Cells were stained with 1 μM CFSE and fluorescent dye dilution in dividing cells was measured 5 days later by flow cytometry. Frequency of dividing cells is shown.

CD38 expression after TLR treatment

CD38, a surface protein whose expression increases upon normal B-cell activation, is a marker of disease aggression in B-CLL.23 We analyzed CD38 expression on B-CLL cells stimulated with TLRs agonists. As shown in Figure 6a, TLR-9 and TLR-7 but not TLR1/2 or TLR2/6 stimulation of CD38− B-CLL cells leads to the upregulation of CD38 expression on sorted B-CLL cells. Surprisingly, stimulation of unsorted B-CLL samples with agonists of TLRs not expressed on B-CLL cells [PGN, LTA, ssRNA, LPS, flagellin, Poly(I:C)] also induced significant CD38 upregulation on B-CLL cells. This suggests the existence of a factor induced by TLRs agonists in nonleukemic cell populations. To test whether this putative factor is soluble, purified B-CLL cells were treated with the supernatants collected from unsorted B-CLL cell cultures stimulated with PGN, Poly(I:C) and ssRNA, i.e. with TLRs agonists that do not activate B-CLL cells themselves. CD38 expression on B-CLL cells significantly increased after stimulation with the supernatants, indicating that CD38 upregulation is caused by a soluble factor (Fig. 6b). We tested whether the addition of the most abundant cytokine in the supernatants of TLRs stimulated B-CLL cells, i.e. TNFα could complement the effect of supernatants on CD38 upregulation. However, TNFα did not induce CD38 expression in B-CLL cells. Thus the identity of the soluble factor remains to be determined.

Figure 6. Induction of CD38 expression on B-CLL cells by stimulation through TLRs. (a) Sorted B-CLL cells (top row) or bulk PBMCs from B-CLL patients (bottom row) were stimulated with TLR agonists and B-CLL cells were analyzed for the expression of CD38. Representative results of 1 of the 3 experiments are shown. Filled histograms represent the negative controls and the open histograms show the staining for CD38. (b) Purified B-CLL cells were stimulated by supernatants from PBMCs of B-CLL patients that were stimulated with TLR agonists PGN, ssRNA and Poly(I:C) or with different concentrations of TNFα. CD38 on B-CLL cells was analyzed 5 days later by flow cytometry.

Discussion

In antigen-presenting cells (APCs), TLRs function as sensors of pathogen presence and triggers for their maturation and cytokine production. Activated APCs stimulate naive T cells and TLRs thus bridge innate to adaptive immunity. Recent studies have revealed a novel role for TLRs in the biology of B lymphocytes. While in human naive B cells, most TLRs are expressed at low levels, their expression increases after BCR triggering and memory B cells acquire the capacity to respond to specific TLR agonists. BCR-independent stimulation of TLR-expressing B cells leads to the polyclonal activation of the memory B-cell pool and allows the maintenance of serological memory.11, 12 Subsequent studies identified that agonists of TLR2/6, TLR1/2, TLR7 and TLR9 can induce proliferation of memory B cells of healthy donors.10

Discovery of the crucial role of TLRs in the homeostasis of memory B cells in humans prompted us to investigate the expression and function of TLRs on B-CLL cells. We first analyzed TLRs expression in B-CLL cells from B-CLL patients in comparison with sorted naive and memory B cells from the peripheral blood of healthy controls. We show, by both qualitative and quantitative PCR, that B-CLL cells express similar set of TLRs as sorted CD19+CD27+ memory B cells of healthy donors, i.e. TLR-1, TLR-2, TLR-6, TLR-7 and TLR-9. However, B-CLL cells lack TLR-4 and express only low levels of TLR-6, similarly to CD19+CD27− naive B cells. There has been a long debate on the origin of a nonmalignant counterpart of B-CLL cells.24 Recent advances in the characterization of both B-CLL and normal B-cell subpopulations by phenotypic analysis and global gene expression profiling25 analyses have shed new light on the phenotype and the cell derivation of B-CLL and provided novel hypotheses concerning its pathogenesis.26 These findings support the concept that B-CLL may be derived from marginal zone B cells with an activated/memory phenotype.27, 28 TLR expression pattern detected in our study is in accordance with the hypothesis that memory B cells represent the most closely related nonmalignant counterpart of B-CLL cells. However, the expression pattern of TLRs on human marginal zone B cells has not yet been studied.

We show that the expression of TLRs on B-CLL cells correlated with their capacity to respond to the stimulation with the respective TLR agonists. At the level of phenotypic changes, ODN2006 was the most potent stimulus. However, a significant upregulation of activation markers was also induced by a control oligonucleotide ODN2006C, lacking unmethylated CpG sequence. This is in accordance with reports showing the activation of B lymphocytes by the thioester bone present in these oligonucleotides.29 B-CLL cells also responded to the stimulation by loxoribine. Higher response of B-CLL cells to loxoribine in comparison with normal B cells might be explained by higher expression of TLR-7 in B-CLL cells as detected by real-time quantitative polymerase chain reaction. In accordance with a recent study by Muzio et al.,30 we show that Pam3CSK4 (a TLR1/TLR2 agonist) and MALP-2 (a TLR2/TLR6 agonist), also induce significant upregulation of activation markers in B-CLL cells. Interestingly, both B-CLL cells and normal B cells express TLR-2 but apparently they do not respond to PGN or LTA, well described TLR-2 agonists. However from recent reports it seems that the immunostimulatory activity of PGN is predominantly mediated by other pattern recognition proteins, such as NOD1 and NOD2,31 and LTA requires LBP and CD14 as coreceptors for its action.32 Taken together, phenotypic changes induced in B-CLL cells by TLR agonists resemble TLR-mediated activation of nonmalignant memory B cells.

Production of TNFα and IL-6 after stimulation with TLRs agonists is another parameter of B-cell activation that we have analyzed in our study. Interestingly, in contrast to CpG and loxoribine, MALP-2 or Pam3CSK4 stimulation never induced substantial levels of TNFα or IL-6 despite their potency to induce phenotypic changes. B-CLL cells stimulated by loxoribine produced significantly higher levels of TNFα when compared to normal B cells in accordance with the higher expression of TLR-7 in B-CLL cells. After TLR9 stimulation B-CLL cells produced similar levels of TNFα but significantly lower levels of IL-6 in comparison to normal B cells. Even PMA/calcium ionomycin stimulation of B-CLL cells induced significantly lower production of IL-6. Thus the low production of IL-6 probably results from an intrinsic defect of B-CLL cells rather than from an impaired response to TLRs stimulation. Indeed, a recent report has also described lower IL-6 production in B-CLL when compared to normal B cells.33 Spaner et al.16 also reported decreased levels of IL-6 after stimulation with TLR-7 agonists in B-CLL in contrast to normal B cells. In B-CLL in vitro experiments, IL-6 was proposed to inhibit the proliferation of leukemic cells while prolonging their survival.34, 35 Thus one can speculate that IL-6 deficiency found in most B-CLL patients might favor B-CLL growth. On the other hand, increased serum levels of IL-6 were reported to indicate the disease progression and correlate with decreased survival of B-CLL patients.36, 37 This is not in contradiction with our results, as we mainly analyzed samples from patients with low Rai score.

The use of TLRs agonists in the treatment of B-CLL has been discussed in recent years.38 Intriguing aspect of TLR stimulation in B-CLL is their ability to induce apoptosis and cell death in leukemic cells which could be of great relevance for their use in the clinics. TLR-7 and TLR-9 agonists were reported to induce apoptosis in B-CLL cells.20 Given the role of TLRs in promoting the proliferation and survival of memory B cells, we analyzed both survival and proliferation of B-CLL cells after TLRs stimulation. As expected, the treatment of normal B cells with TLRs ligands improved their survival in vitro. In contrast stimulation by TLR ligands induced apoptosis in 47% of analyzed B-CLL samples while the rest of the samples were not affected. The most effective agonists in the induction of B-CLL cell death were ODN2006 and Pam3CSK4. Great heterogeneity in TLRs-induced apoptosis was found in patients and the increased apoptotic effect might correlate to specific B-CLL subtype (e.g. defined by the presence of specific chromosomal aberration).19 Fas-Fas-L interaction was suggested to trigger apoptosis induced by ODN2006.20 However, in our study, we did not observe any effect of Fas blocking antibody on the survival of B-CLL cells. Indeed, CD95-expressing B-CLL cells were demonstrated to be resistant to CD95-mediated apoptosis.39 Granzyme B produced by B-CLL cells after activation was also proposed to be involved in B-CLL cell death.19

We also show that the treatment of B-CLL cells with TLR agonists induces proliferation of a subpopulation of B-CLL cells. Although the degree and extent of proliferation induced in B-CLL cells is lower than in B cells, it should be taken into account when considering the use of TLR agonists in the treatment of B-CLL.

We report an interesting observation regarding the effect of TLRs signaling on the expression of CD38 on B-CLL cells. CD38 is a cell-membrane expressed molecule, routinely measured at the time of B-CLL diagnosis because of its correlation with unfavorable disease outcome. In healthy B cells, CD38 is expressed in B cell precursors and plasma cells and it can be induced in mature B cells upon activation.40 Recent report has described CD38 upregulation and differentiation of B cells in plasma cells after stimulation with CpG ODN.41 We show that the expression of CD38 on B-CLL cells could be induced by 2 mechanisms. Either by the direct stimulation of B-CLL cells by agonists of TLR-7 and TLR-9, receptors expressed on B-CLL cells. Alternatively, CD38 can be induced on B-CLL cells indirectly by a soluble factor induced in non-B-CLL cells after stimulation with TLR-2, TLR-3 or TLR-5 agonists. Nature of this factor remains unknown but we excluded the role of TNFα in our experiments. So far the expression of CD38 was considered to be relatively stable in B-CLL cells. Here we show, for the first time, that CD38 expression increases after a short-time TLR stimulation in vitro. Upregulation of CD38 by IFNα in CD38 positive B-CLL samples was published by Pittner et al.42 However, whereas IFNα enhanced CD38 expression only in samples that were CD38+ at the diagnosis, we show that TLR agonists induce CD38 expression in samples that were CD38− prior to the treatment with TLRs agonists. This raises the question whether CD38 expression on B-CLL cells in vivo can vary depending on the degree of tumor-related inflammation and factors present in the tumor microenvironment or during infections. Several groups indeed reported changes in CD38 expression during the course of the disease.23, 43, 44 It is unclear if CD38-expressing B-CLL cells have distinct biological properties than the rest of tumor cell population. It has been suggested that CD38+ B-CLL cells grow more aggressively and they are resistant to apoptosis. Growth and survival signals are delivered through CD38 after interaction with CD31 expressed on the adjacent cells in tumor microenvironment.45 If this was the case, the administration of TLRs ligands for the therapy of B-CLL could have deleterious consequences. Alternatively, CD38 expression on B-CLL cells could only result from a more pronounced inflammatory process associated with the aggressive disease.

Recent studies argued for the potential use of TLR agonists, especially TLR-9 ligands, in the treatment of B-CLL for several reasons.18 Stimulation with TLR agonists could increase the immunogenicity of B-CLL cells through the upregulation of costimulatory molecules. Induction of a transient proliferation could sensitize dormant tumor cells to the action of chemotherapeutics. Finally, in some patients, TLRs agonists induce apoptosis of B-CLL cells. In our study, we describe the expression of TLRs on B-CLL cells and we characterize the phenotype, cytokines production, survival, proliferation and CD38 expression after stimulation with specific agonists of TLRs. We plead for the cautious evaluation of therapeutic potential of different TLR agonists in B-CLL as their effects on B-CLL cells are complex. While TLR agonists indeed increase the immunogenicity of B-CLL cells, induction of inflammatory cytokines production by B-CLL cells and induction of CD38 expression could worsen the course of the disease. Cytotoxic action of TLR agonists on B-CLL cells also has to be better characterized and markers predicting sensitivity of leukemic cells to TLR stimulation need to be identified. It is also important to note the functional differences between various ligands. Loxoribine and CpG induce the expression of costimulatory molecules, proinflammatory cytokines production, B-CLL cell proliferation and CD38 expression. On the other hand, MALP-2 (a TLR-2/6 agonist) and Pam3CSK4 (a TLR-1/2 agonist) do not induce cytokine production in B-CLL cells and do not upregulate CD38. They, however, induce comparable expression of costimulatory molecules to TLR-7 and TLR-9 ligands.

Toll-like receptor stimulation plays a crucial role in the homeostasis of human B cells, especially in the maintenance of memory B cells pool. Apparently, TLRs are also expressed on B-CLL cells and their stimulation leads to the complex changes in the characteristics and function of leukemic cells. Thorough analysis of the effects of TLRs stimulation on B-CLL cells is a prerequisite not only for their potential use in the therapy of B-CLL but also for better understanding of B-CLL biology.