No interactions between heparin and atacicept, an antagonist of B cell survival cytokines

Abstract

Background and purpose: The TNF family ligands, B cell activating factor of the TNF family (BAFF, also known as B lymphocyte stimulator, BLyS) and a proliferation-inducing ligand (APRIL), share the transmembrane activator and calcium-modulator and cyclophilin ligand (CAML)-interactor (TACI) as one of their common receptors. Atacicept, a chimeric recombinant TACI/IgG1-Fc fusion protein, inhibits both ligands. TACI and APRIL also bind to proteoglycans and to heparin that is structurally related to proteoglycans. It is unknown whether the portion of TACI contained in atacicept can bind directly to proteoglycans, or indirectly via APRIL, and whether this could interfere with the anti-coagulant properties of heparin.

Experimental approach: Binding of atacicept and APRIL to proteoglycan-positive cells was measured by FACS. Activities of heparin and atacicept were measured with activated factor Xa inhibition and cell-based assays. Effects of heparin on circulating atacicept was monitored in mice.

Key results: Atacicept did not bind to proteoglycan-positive cells, but when complexed to APRIL could do so indirectly via APRIL. Multimers of atacicept obtained after exposure to cysteine or BAFF 60-mer bound directly to proteoglycans. Atacicept alone, or in complex with APRIL, or in a multimeric form did not interfere with heparin activity in vitro. Conversely, heparin did not influence inhibition of BAFF and APRIL by atacicept and did not change circulating levels of atacicept.

Conclusions and implications: Lack of detectable interference of APRIL-bound or free atacicept on heparin activity makes it unlikely that atacicept at therapeutic doses will interfere with the function of heparin in vivo.

Figure 1

Atacicept does not bind proteoglycans directly but can do so via APRIL. (a) elisa assay to monitor the binding of Flag‐APRIL to coated atacicept. The experiment was performed three times. (b–d) HEK293 cells and the same cells stably transfected with full‐length human BAFF (HEK293‐BAFF) were stained with atacicept and/or Flag‐APRIL at the indicated concentrations, in the presence or absence of heparin (8 IU·ml⁻¹). Binding of Flag‐APRIL and atacicept were respectively revealed with anti‐Flag and anti‐Fc secondary reagents. Results shown are from one of three similar experiments with similar results. (b) Selected histograms of the experiment described above in which the X‐axis displays fluorescence intensity on a log scale (3 × 10² to 2 × 10⁵). (c) Mean fluorescence intensity (MFI) of atacicept binding to HEK293‐BAFF cells, in the presence or absence of heparin. (d) MFI of Flag‐APRIL or atacicept binding to HEK293 cells in the presence or absence. APRIL, a proliferation‐inducing ligand; BAFF, B cell activating factor of the TNF family

Figure 2

Dimeric recombinant TACI‐Fc does not bind to proteoglycans. (a) Schematic representation of human TACI. Portions of the extracellular domain of TACI contained in the different TACI‐Fc constructs are shown with thick vertical red lines. Numbers at the top refer to plasmid number encoding these proteins (see Table 1). aa, amino acid residues; CRD, cysteine‐rich domain; ID, intracellular domain; TMD, transmembrane domain. (b) UV elution profile of a size exclusion chromatography column loaded with a mixture of standards of indicated MW (in kDa) or with protein A or protein G‐purified TACI‐Fc or BCMA‐Fc constructs, as indicated. Fractions chosen for further analysis are indicated and underlined in red. (c) SDS‐PAGE (10% or 12% acrylamide) and Coomassie blue analysis of 10‐μg protein per lane of size‐fractionated proteins from panel (b). Input: Protein after Protein A or G affinity purification. MW standards indicated on the left are valid for the first gel. For the next ones, refer to standards (Std) shown on the left of each gel, knowing that red proteins correspond to 72 and 28 kDa, respectively. Target proteins are indicated with red asterisks, and presumptive Fc fragments with black asterisks. The insert shows a western blot analysis of 50 ng of the indicated fractions, revealed with anti‐human or anti‐TACI antibodies. (d) HEK293 cells were stained with the indicated proteins at 10 μg·ml⁻¹, in the presence or absence of heparin. Results are expressed as the ratio of mean fluorescence intensity (MFI) in the absence of heparin to that in the presence of heparin. In the presence of heparin, all samples gave background MFI. The experiment was performed three times with similar results. BCMA, B cell maturation antigen; TACI, transmembrane activator and calcium‐modulator and cyclophilin ligand (CAML)‐interactor

Figure 3

Atacicept does not interfere with heparin activity. (a) Heparin activity was measured as its ability to inhibit activated factor X (factor Xa) in the presence of human plasma (containing endogenous anti‐thrombin). The dose‐dependent inhibition of factor Xa by heparin was measured in the presence of the indicated proteins (each at 100 μg·ml⁻¹), or of poly‐lysine (poly‐K; at 90 μg·ml⁻¹). Mean ± SEM of duplicate measures. One representative experiment out of three similar ones with identical results is shown. (b) Reporter cells expressing the BAFFR:Fas fusion receptor were incubated overnight in the presence of a lethal (5 ng·ml⁻¹) dose of Fc‐BAFF, in the presence of increasing concentrations of atacicept, and in the presence of the indicated amounts of heparin. Cell viability was measured with a colorimetric test. Mean ± SEM of duplicates. One experiment out of three is shown. (c) Heparin was incubated overnight at 37°C in medium, or with reporter cells in medium, after which time its ability to inhibit binding of Fc‐APRIL to HEK293 cells was monitored. The X‐axis displays fluorescence on a log scale (3 × 10² to 5 × 10⁴). The experiment was performed three times. (d) Same as panel (b), but using BCMA:Fas reporter cells and a lethal (20 ng·ml⁻¹) dose of Fc‐BAFF. Mean ± SEM of duplicates. The experiment was performed three times. (e) Same as panel (d), but with a lethal (20 ng·ml⁻¹) dose of Fc‐APRIL. The experiment was performed three times with identical results. (f) BCMA:Fas reporter cells were cultured overnight with titrated amounts of Fc‐APRIL, plus or minus heparin. The experiment was performed three times with identical results. APRIL, a proliferation‐inducing ligand; BAFF, B cell activating factor of the TNF family; BAFFR, BAFF receptor; BCMA, B cell maturation antigen

Figure 4

Multimers of atacicept bind to proteoglycans. (a) Size exclusion chromatography elution profile of atacicept incubated for 48 hr at 37°C in phosphate buffer without cysteine. Online monitoring of absorbance at 280 nm is shown as a line. Fractions analysed by SDS‐PAGE and Coomassie blue staining (20 μl) under non‐reducing (top gel) or reducing conditions (bottom gel) are shown in the background. Sizes of MW standards for SDS‐PAGE are shown on the right, and those for the size exclusion chromatography are shown on the top. The vertical dotted line marks the elution position of dimeric atacicept. (b) Same as panel (a), except that 20‐mM cysteine was included in the 48 hr incubation. The experiment was performed five times (but the Coomassie under non‐reducing conditions was performed twice). (c) HEK293 cells, with or without expression of membrane‐bound BAFF, were stained by FACS with atacicept (dimers) or atacicept multimers (from fraction 10, see panel (b)) at the indicated concentrations, in the presence or absence of heparin. MFI, mean fluorescence intensity. The experiment was performed three times. Points are monoplicates. (d) HEK 293 cells were incubated with the indicated concentrations of atacicept and BAFF 60‐mer, in the presence or absence of heparin, and stained by FACS anti‐hFc. The experiment was performed three times. (e) HEK293 cells were incubated with various Fc‐containing proteins (or no Fc‐containing protein: Ø) at 5 μg·ml⁻¹ and with BAFF 60‐mer at 2 μg·ml⁻¹ and then analysed by FACS anti‐hIgG. The ratio of the mean fluorescence intensity (MFI) of staining performed in triplicate without or with heparin are shown. Mean ± SEM. The experiment was performed three times with similar results. Ten micrograms of the Fc‐containing proteins (6 μg for mBAFFR‐Fc) used in these experiments were analysed by SDS‐PAGE and Coomassie blue staining under reducing conditions. BAFF, B cell activating factor of the TNF family; BAFFR, BAFF receptor

Figure 5

Atacicept in serum and atacicept multimers do not interfere with heparin activity. (a) HEK293 cells were stained by FACS, in the presence or absence of heparin, with atacicept contained in sera samples collected at the indicated time points from a healthy subject who received 150 mg atacicept. D, day; PK, pharmacokinetics. The experiment was performed twice with sera of subject 219, and once with sera of subject 221, treated identically, with similar results. (b) Same as panel (a), except that HEK293‐hBAFF cells were stained. (c) Heparin activity assay (inhibition of activated factor X) performed in the presence of PBS, 90 μg·ml⁻¹ of poly‐lysine (Poly‐K), or 100 μg·ml⁻¹ of dimeric atacicept. (d) Same as panel (c), but with two independent preparations of atacicept multimers at 100 μg·ml⁻¹. (e, f) Same as panel (c), but with sera from healthy subjects, before or 24 hr after administration of 150 mg of atacicept. (g, h) Same as panels (e) and (f), but with sera of patients before or after 24 weeks of treatment with weekly administration of 150‐mg atacicept. Experiments in panels (c)–(h) were performed twice with similar results. Mean ± SEM of duplicates (SEM is smaller than symbol size). BAFF, B cell activating factor of the TNF family

Figure 6

No evidence that atacicept binds to proteoglycans in vivo. Wild‐type mice were treated subcutaneously without (n = 5) or with atacicept at 5 mg·kg⁻¹; 24 hr (D1, n = 6) or 72 hr (D3, n = 6) later, plasma was collected, and then 1 IU·g⁻¹ of heparin was administered intravenously, and 20 min later, plasma was collected again. (a) Heparin activity in plasma was measured with the activated factor X (factor Xa) inhibition assay. *P<.05, significantly different as indicated; one‐way ANOVA with Bonferroni's multiple comparison test. (b) 293T cells stained with a complex of atacicept bound to Flag‐APRIL were exposed for the indicated times to the indicated concentrations of heparin, after which time cells were washed and bound atacicept was detected. The experiment was performed three times. (c) The activity of heparin in plasma was measured as its ability to prevent binding of Flag‐APRIL to 293T cells. In this assay, cells were incubated with Flag‐APRIL first, and plasma was added afterwards. X‐axis shows fluorescence on a log scale (10²–2 × 10⁵). The experiment was performed three times. (d) Triglyceride lipase activity was measured in plasma of mice (n = 13) before or after heparin administration, using a quenched fluorescent triglyceride (TG) derivative as substrate. The assay was performed twice in the presence of an excess of heparin. *P<.05, significantly different as indicated; paired t‐test. (e) A standard curve of active atacicept was measured by its ability to protect BAFFR:Fas reporter cells from Fc‐BAFF‐induced death. The experiment was performed three times. (f) Active atacicept was measured in mouse plasma before (n = 5), or 1 day (n = 5), 3 days (n = 5), or 7 days (n = 1) after atacicept administration. Active atacicept concentrations were derived from the EC₅₀ of Fc‐BAFF inhibition. Values obtained in plasma of heparin‐treated mice were corrected by the blood dilution factor caused by heparin administration. The dotted line shows the lower limit of quantification. Measures were performed in duplicate. The experiment was done three times with identical results. (g) A standard curve of atacicept diluted in mouse plasma was measured by AlphaLISA, in the presence or absence of 200 ng·L⁻¹ of recombinant mouse BAFF, using a matched pair of anti‐TACI antibodies. (h) Same as panel (g), except that samples were submitted to an acid dissociation procedure to separate BAFF from atacicept. Experiments of panels (g) and (h) were performed five times. (i) Mice received atacicept (or not, n = 6) subcutaneously at Day 0. Plasma was taken at Days 1 (n = 6), 3 (n = 5, because not enough plasma was left from the sixth mouse), 7 (n = 6), or 14 (n = 10), before (circles) and after (triangles) intravenous administration of heparin. Atacicept was measured by AlphaLisa in plasma samples without (white symbols) or with (black symbols) an acid dissociation step. Values of heparin‐treated mice were corrected by the dilution factor induced by heparin administration. Measures were performed four times on the same set of samples, with similar results. *P<.05, significantly different as indicated; one‐way ANOVA was performed three times for conditions at Days 3, 7, and 14, respectively, with Bonferroni's multiple comparison tests at Days 7 and 14. Statistical analysis was performed on log‐transformed data; because under these conditions, there were no significant variance inhomogeneity. (j) A standard curve of mouse BAFF diluted in serum of a Baff ^−/− mouse was measured by AlphaLISA, in the presence of 0, 100, or 2000 ng·ml⁻¹ of atacicept, using a matched pair of anti‐mBAFF antibodies. The experiment was performed three times. (k) Circulating BAFF levels in plasma of WT mice treated subcutaneously at Day 0 with a single dose of atacicept. Plasma was collected at Day 0 (before atacicept administration), Day 1, Day 3, Day 7, or Day 14 after atacicept administration, and before or after administration of heparin on the day of plasma collection. BAFF levels were calculated using the standard curve of BAFF in the absence of atacicept and are expressed as “unbound mBAFF equivalent.” Samples were measured three times independently, with similar results. APRIL, a proliferation‐inducing ligand; BAFF, B cell activating factor of the TNF family; BAFFR, BAFF receptor; TACI, transmembrane activator and calcium‐modulator and cyclophilin ligand (CAML)‐interactor