CD39 and CD73 Activity Are Protective in a Mouse Model of Antiphospholipid Antibody-Induced Miscarriage

Antiphospholipid antibodies (aPL-ab) including lupus anticoagulant, anticardiolipin (aCL) antibodies and anti-β2-glycoprotein I (β2GP1) antibodies initiate inflammation, tissue factor-activated coagulation, oxidative stress and complement activation. These pathways contribute to the pathophysiology of thrombotic and obstetric manifestations in antiphospholipid syndrome (APS). Not all patients with aPL-ab develop disease phenotype and hence a 'second hit' is postulated to be necessary. Extracellular purines (ATP, ADP, and adenosine) are important signalling molecules with the general acceptance that ATP promotes inflammation, ADP is pro platelet-aggregatory, while adenosine has vasodilatory, antithrombotic and anti-inflammatory effects. Extracellular ATP/ADP levels are tightly regulated by CD39 and CD73. CD39 hydrolyses ATP and ADP to AMP, which is then hydrolysed to adenosine by the ecto-5'-nucleotidase CD73.

We hypothesised that CD39 and CD73 mitigate both vascular thrombosis as well as inflammatory responses in APS. We developed a murine model of APS-miscarriages by passive infusion of aPL-ab harvested from patients into pregnant mice. Comparison made with infusion of non-immune IgG. We applied this model to wild type (WT), CD39-transgenic mice as well as CD39-null and CD73-null mice. Non-immune IgG administration to WT or CD39^-/- C57BL/6 mice did not cause miscarriages, whereas aPL-ab-treated WT mice displayed a mean resorption frequency of 3.67%. In comparison, aPL-ab-treated CD39^-/- mice showed a significantly higher mean resorption frequency of 20.1% (p<0.001 vs aPL-ab-treated WT; Fig. 1a). aPL-ab-treated CD73^-/- mice also had a significantly increased mean resorption frequency of 10.5% (p<0.05 vs aPL-ab-treated WT).

We have previously demonstrated that CD39 over expression in BALB/c transgenic mice is associated with enhanced adenosine generation under prothrombotic conditions. WT BALB/c mice injected with non-immune IgG had a mean resorption frequency of 21.2%, which doubled with aPL-ab injection (39.7%; p<0.05 vs non-immune IgG-treated WT). In transgenic BALB/c mice over-expressing CD39, mean resorption frequencies were significantly lower in the aPL-ab-treated group (11.4% vs 39.7% in aPL-ab-treated WT, p<0.01).

QRT-PCR showed 1.6-fold and 2.4-fold increase in TF expression in aPL-ab administered CD39^-/- and CD73^-/- mice relative to both WT C57BL/6 mice and littermates injected with non-immune sera and. CD39 over expression suppressed TF expression compared with WT BALB/c mice after both non-immune IgG and aPL-ab challenge particularly with 6.3 fold lower TF mRNA levels after aPL-ab infusion. Decidua from aPL-ab injected CD39^-/- and CD73^-/- C57BL/6 mice showed increased TF immunoreactivity compared to aPL-ab-treated or nonimmune IgG-treated WT mice. C3d deposition was increased by aPL-ab treatment relative to non-immune IgG treatment in all groups except the CD39 transgenic mice (Fig. 3). C3d deposition was most marked in aPL-ab-treated CD39^-/- and CD73^-/- mice.

Placental expression of inflammatory cytokines C57BL/6was measured at the mRNA level by qRT-PCR. TNF-α expression was doubled in CD39^-/- and CD73^-/- mice relative to non-immune IgG controls. CD39 Tg mice showed suppression of TNF- α after aPL-ab challenge. Consistent with the previously documented role for IL-10 in exacerbating lupus/auto-immune diseases, we noted a significant reduction in IL-10 mRNA in aPL-ab-treated CD39 Tg mice compared to aPL-ab-treated WT mice.

Therefore, we demonstrate a protective role for CD39 in APS and provide rationale for the development of soluble CD39 as a novel therapeutic for APS and analysis of perturbations in the purinergic pathway to explain human disease