Lymphoma B-cell responsiveness to CpG-DNA depends on the tumor microenvironment

Nuclease-stable phosphorothioate-modified CpG 1826 (CpG) with 5_-TCCATGACG TTCCTGACG TT (the nucleotides in bold represent the immunostimulatory CpG sequences), fluorescein isothiocyanate (FITC)-conjugated CpG 1826 ODNs, and control 1826 ODN with 5_-TCCATGAGCTTCCTGAGCTT were purchased from InvivoGen (Cayla, France).

A20.IIA is an FcγR-negative clone originating from the A20-2 J B-cell lymphoma line [13]. For in vivo experiments, A20.IIA cells were transfected by an electroporation system with the green fluorescent protein (GFP) gene. These cells, hereafter referred to as A20.IIA or A20.IIA-GFP cells as appropriate, were maintained at 37°C, 5% CO₂ in complete Roswell Park Memorial Institute (RPMI) 1640 Medium Glutamax plus (RPMI; Gibco-Invitrogen, France) supplemented with 10% fetal calf serum (FCS; PAA Laboratories, Germany), 100 μg/mL penicillin and 100 μg/mL streptomycin (both from Eurobio, France), 10 mM sodium pyruvate (Gibco-Invitrogen), and 50 μM 2-mercaptoethanol (Gibco-Invitrogen). The A20.IIA-GFP cell culture was also supplemented with 0.5 mg/mL neomycin (G418; Gibco-Invitrogen). To obtain the A20.IIA-luc 2 cell line, A20.IIA cells were transfected with pGL4.50[luc2/CMV/hygro] (Promega), in the AMAXA Nucleofector II device (Lonza, Switzerland) and were cultured in 0.75 mg/mL hygromycin B (Gibco-Invitrogen) medium.

A20.IIA cells at a concentration of 10⁵cells/mL were incubated with serial dilutions of CpG 1826 or control 1826 ODNs at concentrations ranging from 0.0003 to 60 μg/mL or with complete RPMI medium alone. After 3 days, [³H] thymidine (GE Healthcare) was added for the last 4 h. Cells were harvested onto fiber filters and [³H] thymidine incorporation was measured in a scintillation counter (Microbeta, Perkin Elmer).

A20.IIA cells (10⁴) were cultured in complete RPMI medium in 96-well plates in the presence or absence of 3 μg/mL or 30 μg/mL of CpG or control ODNs. Staining with Annexin V/allophycocyanin (APC) and propidium iodide (PI) (BD Biosciences, France) was performed 72 h later and then analyzed by flow cytometry. Apoptotic cells were defined as those positive for Annexin V and PI.

Female BALB/c mice (H-2^d) were obtained from Charles River Laboratories (L’Arbresle, France) and used between 6 and 8 weeks of age. They were provided with sterile food and water ad libitum and kept on a 12-hour light–dark cycle. All procedures involving mice conformed with European Union guidelines, French regulations for animal experimentation (Ministry of Agriculture Act No. 2001–464, May 2001), and the guidelines of the Institut National de la Santé et de la Recherche Médicale Committee on Animal Research, and were approved by the relevant local committees (Charles Darwin Ethics Committee for Animal Experiments, Paris, France; Permit Number: p3/2009/004).

Mice were first anesthetized by intraperitoneal injection of a mixture containing 120 mg/kg of ketamine (Virbac, France) and 6 mg/kg of xylazine (Rompun 2%; Bayer Healthcare). To obtain a subcutaneous lymphoma (SCL) murine model, BALB/c mice were inoculated subcutaneously with 5 × 10⁶ A20.IIA-GFP tumor cells in a final volume of 50 μL of RPMI, at 2 different sites: the right and left abdomen. For the intracerebral tumor implantation, anesthetized mice were immobilized on a stereotaxic frame (David Kopf Instruments, Tujunga, CA, USA). Tumor cells (5 × 10⁴ in a final volume of 2 μL RPMI) were injected into the specific cerebral location (right striatum), located 2 mm to the right of the medial suture and 0.4 mm in front of the bregma, through a Hamilton syringe attached to a penetrating depth controller. The penetrating depth of the syringe was 2.5 mm from the surface of the brain. Each injection delivered the solution slowly, and the syringe was held in place for an additional minute to reduce backfilling of tumor cells. For the intravitreal tumor implantation, we used a 32-gauge needle attached to a syringe to inject 10⁴ cells in a final volume of 2 μL of RPMI into the vitreous under a dissecting microscope. Lacrinorm 2% (Bauch&Lomb) drops were instilled after intravitreal injection. For each tumor model, control mice received either 1× phosphate-buffered saline (pH7.4; PBS) or control 1826 ODNs instead of CpG 1826 ODNs.

Tumor growth in the SCL model was monitored by caliper measurements 3 times a week. Treatment began when the longest tumor diameter reached 0.5 to 0.7 cm. The mice then received daily intratumor injections of CpG-ODNs for 5 days (100 μg per injection in a final volume of 50 μL RPMI) in the right tumor only; the left tumor served as an untreated control tumor. Mice were killed one week after the last treatment injection. Lymphomas established in the brain and eye were treated 7 days after tumor inoculation, by a single local injection of 60 μg (brain) or 20 μg (eye) CpG-ODNs in 2 μL of RPMI (treatment groups) or 2 μL of PBS (control groups). Tumor burden was analyzed in the sacrificed mice one week after treatment administration.

The tumor-injected brains and eyes and the subcutaneous tumors were harvested one week after treatment injection, minced with surgical scissors, incubated for 30 minutes in RPMI containing 0.1 mg/mL DNAse I (Roche Diagnostics, Meylan, France) and 1.67 Wünch U/mL Liberase (Roche), and filtered through a 70-μm membrane (BD Falcon). Mononuclear cells were separated from myelin with a Percoll cell density gradient.

The A20.IIA (1 × 10⁴) cells expressing luciferase (luc2 gene) were injected via subcutaneous, intracerebral or intravitreal routes into immunocompetent 7-week-old BALB/c mice. CpG or control ODNs were administered in situ for each lymphoma model according to the same experimental design and at the time points and doses described above. The tumor burden was thereafter monitored by bioluminescence imaging. Mice were injected intraperitoneally with 150 mg/kg of D-luciferin potassium salt (Interchim) and underwent imaging within the next 10 minutes with the IVIS LUMINA II (Caliper LS) imaging system. The exposure time was set to optimize the signal and obtain the best signal-to-noise ratio. The bioluminescence signal is expressed in photons per second.

Mice were implanted with tumor cells in the brain (PCL), eye (PIOL) or flank (SCL) or injected with PBS in the eye (PIE). Either 14 days later (brain and eye) or when tumor diameter reached 0.5 to 0.7 cm (SCL), the relevant cells were isolated and cultured for 36 h in complete RPMI medium in 96-well (10⁴ cells per well) plates. Supernatant was then harvested from each well.

To analyze TLR9 expression on A20.IIA cells, these cells underwent intracellular staining with the Fixation/Permeabilization solution kit (BD Biosciences) and an anti-TLR9/PE mAb (BD Biosciences).

Tumor burden was analyzed according to the following protocol: Fc receptors were saturated for 20 min with 10 μg/mL of anti-CD16/CD32 mAb (clone 2.4.G2), and then the cells were incubated for 20 min with either rat IgG2a anti-CD19/APC mAb, or the corresponding isotypic mAb control (all from BD Biosciences). The living cells were defined with side scatter (SSC) and forward scatter (FSC) after autofluorescent cells were excluded. Cell phenotypes were analyzed with the LSRII cytometer and Diva software (BD Biosciences).

TLR9 is an intracellular receptor that recognizes CpG-DNA. Cell stimulation by CpG motifs requires that they bind to TLR9. We therefore began by confirming with flow cytometry that A20.IIA B lymphoma cells express TLR9 (Figure 1A).

We next evaluated the direct effect of CpG-ODNs on the proliferation of A20.IIA lymphoma in vitro. Based on our study of its proliferation kinetics (data not shown), tumor cells were incubated for 72 h with CpG 1826 ODNs at concentrations ranging from 0.0003 to 30 μg/mL. Cell proliferation was measured with the [³H] thymidine incorporation assay. The CpG-ODNs inhibited A20.IIA [³H] thymidine incorporation in a dose-dependent manner, whereas control ODNs had no effect on cell proliferation (Figure 1B). The maximum inhibitory effect was obtained from 0.3 to 30 μg/mL of CpG-ODNs.

Based on these results, we analyzed the induction of apoptosis of A20.IIA cells by CpG- ODNs and found that they induced apoptosis of about 30 to 35% of the cells (that is, 30-35% stained positive for Annexin V and PI), compared with 5% for RMPI Medium and 5-10% with the control ODNs (Figure 1C).

After showing that the CpG motif has a direct antiproliferative and proapoptotic effect on A20.IIA lymphoma cells, we sought to explore its effects in vivo when injected intratumorally, by comparing the 3 types of murine models of lymphoma: SCL, PCL and PIOL. A20.IIA-GFP cells were implanted on the left and right flank of the mice for the SCL model. Tumor size was measured by a caliper 3 times a week. When the tumors reached 5–7 mm in diameter, the left site was treated by local injections of CpG- ODNs, while the right one was used as an untreated control tumor. As described by Houot & Levy in 2009 [14] mice did or did not receive daily intratumoral injections of 100 μg/50μL CpG-ODNs for 5 days. Tumor size was then measured daily until sacrifice, one week after the last treatment injection. The tumor burden of mice treated with CpG and control ODNs was compared with a bioluminescence imaging system that assessed total photon influx. The CpG-ODNs inhibited tumor growth very soon after treatment in this SCL model. On day 7 after treatment, the untreated tumor was more than 100 times brighter than the CpG-treated one, and on day 20, 120 times brighter (Figure 2A). Flow cytometric analysis of CD19⁺GFP⁺ cells confirmed that tumor cells decreased significantly more in the treated than the untreated tumors (Figure 2B).

We next addressed the question of whether CpG motifs have the same antitumor effect in cerebral lymphomas. Imaging analysis showed two different profiles. Some mice did not respond to in situ CpG-ODN treatment, and the lymphoma developed in the brain and even developed in lymph nodes at day 21; this timing was nonetheless later than in the control group (Figure 2C – Example 1). Some mice did respond to the treatment; the tumor grew from day 0 to day 7 after treatment, and then decreased until it was undetectable (Figure 2C – Example 2). We also examined the percentage of CD19⁺GFP⁺ cells in the group treated by CpG-ODNs, compared it with the control group and observed a significant decrease in the proportion of tumor cells (Figure 2D).

Next we investigated the antitumor effect of CpG-ODNs on PIOL mice that had a tumor implanted in the right eye only and were then treated with CpG-ODNs (20 μg/2μL) or control ODNs (20 μg/μL). As shown in Figure 2E, CpG-ODNs seem to have had no detectable effects on the primary eye tumor. Nevertheless, they appeared to prevent lymph node invasion at day 27 (Figure 2E). Flow cytometric analysis showed no significant difference in tumor growth between CpG ODN-treated and control (PBS 1X) treated eyes (Figure 2F).

These results suggest that the behavior of tumors in the eye is different from that of systemic lymphomas, but also from that of cerebral lymphoma, and thus, that tumor cells responsiveness to CpG-DNA depend on the tissue microenvironment.

As described above, in vivo experiments showed that the responsiveness of lymphoma B cells to CpG-ODN administration was tissue-dependent. To confirm that the blockade of CpG-ODN antitumor effects was due to the PIOL molecular microenvironment, we tested whether supernatant from PIOL could counteract the inhibitory effect of CpG-ODNs on the proliferation of A20.IIA cells in vitro. A [³H] thymidine incorporation assay was performed as described above, with the addition of supernatant obtained from PBS-injected eyes (PIE) (as control), or from the mouse model SCL, PCL, and PIOL. As shown in Figure 3, the addition of PIE (Figure 3A) and SCL (Figure 3B) supernatants did not modify the ability of CpG-ODN treatment to inhibit tumor growth. PCL supernatant (Figure 3C) increased proliferation, but CpG-ODNs were still active at doses of 30 and 60 μg/mL. In contrast, CpG-ODNs were unable to inhibit tumor cell proliferation after incubation with PIOL supernatant (Figure 3D) and to induce apoptosis (data not shown). Moreover, their antiproliferative effect was recovered upon dilution of the PIOL supernatant (Figure 4). These data confirm our in vivo results and show that a soluble factor, released in eye tumors but not in normal eyes, was able to counteract the antiproliferative effect of CpG motifs.

To investigate the possibility that the loss of the CpG-ODNs antitumor action was associated with modulation of TLR9 expression, we used flow cytometry to compare TLR9 expression on A20.IIA cells after incubation with supernatant from medium alone, PIOL or PIE. No differences were found between these conditions (Figure 5A).

Next we examined whether the PIOL molecular microenvironment inhibited internalization of CpG ODNs by tumor cells. FITC-labelled CpG 1826 ODNs were added for 24 hours at a concentration of 3 μg/mL to A20.IIA lymphoma cells in the presence of PIOL or PIE supernatant. Flow cytometric analysis indicated that FITC expression by tumor cells with PIOL supernatant was similar to that incubated with PIE supernatant (Figure 5B).These findings show that the addition of PIOL supernatant does not modify CpG internalization by lymphoma B-cells, even in vivo in our three model (data not shown).