Biologic TNF-α inhibitors reduce microgliosis, neuronal loss, and tau phosphorylation in a transgenic mouse model of tauopathy

Background

Tumor necrosis factor-α (TNF-α) plays a central role in Alzheimer’s disease (AD) pathology, making biologic TNF-α inhibitors (TNFIs), including etanercept, viable therapeutics for AD. The protective effects of biologic TNFIs on AD hallmark pathology (Aβ deposition and tau pathology) have been demonstrated. However, the effects of biologic TNFIs on Aβ-independent tau pathology have not been reported. Existing biologic TNFIs do not cross the blood–brain barrier (BBB), therefore we engineered a BBB-penetrating biologic TNFI by fusing the extracellular domain of the type-II human TNF-α receptor (TNFR) to a transferrin receptor antibody (TfRMAb) that ferries the TNFR into the brain via receptor-mediated transcytosis. The present study aimed to investigate the effects of TfRMAb-TNFR (BBB-penetrating TNFI) and etanercept (non-BBB-penetrating TNFI) in the PS19 transgenic mouse model of tauopathy.

Methods

Six-month-old male and female PS19 mice were injected intraperitoneally with saline (n = 12), TfRMAb-TNFR (1.75 mg/kg, n = 10) or etanercept (0.875 mg/kg, equimolar dose of TNFR, n = 10) 3 days/week for 8 weeks. Age-matched littermate wild-type mice served as additional controls. Blood was collected at baseline and 8 weeks for a complete blood count. Locomotion hyperactivity was assessed by the open-field paradigm. Brains were examined for phosphorylated tau lesions (Ser202, Thr205), microgliosis, and neuronal health. The plasma pharmacokinetics were evaluated following a single intraperitoneal injection of 0.875 mg/kg etanercept or 1.75 mg/kg TfRMAb-TNFR or 1.75 mg/kg chronic TfRMAb-TNFR dosing for 4 weeks.

Results

Etanercept significantly reduced phosphorylated tau and microgliosis in the PS19 mouse brains of both sexes, while TfRMAb-TNFR significantly reduced these parameters in the female PS19 mice. Both TfRMAb-TNFR and etanercept treatment improved neuronal health by significantly increasing PSD95 expression and attenuating hippocampal neuron loss in the PS19 mice. The locomotion hyperactivity in the male PS19 mice was suppressed by chronic etanercept treatment. Equimolar dosing resulted in eightfold lower plasma exposure of the TfRMAb-TNFR compared with etanercept. The hematological profiles remained largely stable following chronic biologic TNFI dosing except for a significant increase in platelets with etanercept.

Conclusion

Both TfRMAb-TNFR (BBB-penetrating) and non-BBB-penetrating (etanercept) biologic TNFIs showed therapeutic effects in the PS19 mouse model of tauopathy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12974-021-02332-7.

Acute and Chronic Dosing of a High-Affinity Rat/Mouse Chimeric Transferrin Receptor Antibody in Mice

Non-invasive brain delivery of neurotherapeutics is challenging due to the blood-brain barrier. The revived interest in transferrin receptor antibodies (TfRMAbs) as brain drug-delivery vectors has revealed the effect of dosing regimen, valency, and affinity on brain uptake, TfR expression, and Fc-effector function side effects. These studies have primarily used monovalent TfRMAbs with a human constant region following acute intravenous dosing in mice. The effects of a high-affinity bivalent TfRMAb with a murine constant region, without a fusion partner, following extravascular dosing in mice are, however, not well characterized. Here we elucidate the plasma pharmacokinetics and safety of a high-affinity bivalent TfRMAb with a murine constant region following acute and chronic subcutaneous dosing in adult C57BL/6J male mice. Mice received a single (acute dosing) 3 mg/kg dose, or were treated for four weeks (chronic dosing). TfRMAb and control IgG1 significantly altered reticulocyte counts following acute and chronic dosing, while other hematologic parameters showed minimal change. Chronic TfRMAb dosing did not alter plasma- and brain-iron measurements, nor brain TfR levels, however, it significantly increased splenic-TfR and -iron. Plasma concentrations of TfRMAb were significantly lower in mice chronically treated with IgG1 or TfRMAb. Overall, no injection related reactions were observed in mice.

Eliminating Fc N-Linked Glycosylation and Its Impact on Dosing Consideration for a Transferrin Receptor Antibody-Erythropoietin Fusion Protein in Mice

Erythropoietin (EPO), a hematopoietic growth factor and a promising therapy for Alzheimer's disease, has low permeability across the blood-brain barrier. The transferrin receptor antibody fused to EPO (TfRMAb-EPO) is a chimeric monoclonal antibody that ferries EPO into the brain via the transvascular route. However, TfRMAbs have Fc-effector function-related adverse effects including reticulocyte suppression. To overcome this, we recently developed an effectorless TfRMAb-EPO fusion protein, designated TfRMAb-N292G-EPO, by eliminating the Fc N-linked glycosylation site at position 292 of the antibody heavy chain. The mutant fusion protein showed enhanced plasma clearance and dramatically reduced plasma concentrations compared with the wild-type (WT) nonmutant fusion protein. This increased clearance of the aglycosylated TfRMAb is expected to increase the injection dose of the mutant fusion protein. To provide a basis for future therapeutic uses of this IgG-neurotrophin fusion protein, the current study aimed to characterize the pharmacokinetic profile of this effectorless TfRMAb-N292G-EPO at different doses following different routes of administration in the mouse. Adult C57BL/6J male mice were injected with a single dose (3, 6, 9, or 20 mg/kg; n = 3-6 per dose) of TfRMAb-N292G-EPO through either the subcutaneous (SQ) or intraperitoneal (IP) route. TfRMAb-N292G-EPO plasma concentrations were determined using an enzyme-linked immunosorbent assay. Mice were sacrificed 24 h after injection, and terminal blood was used for a complete blood count. Brain concentrations in the WT- and mutant fusion protein-treated mice were compared. We observed stark differences in the plasma pharmacokinetics of TfRMAb-N292G-EPO between the IP and SQ routes of administration. Dose escalation from 3 to 20 mg/kg increased the plasma C_max only 3.5-fold for the SQ route, compared with a 35-fold increase for the IP route. The plasma C_max was 15.0 ± 2.0, 21.3 ± 4.1, 21.3 ± 6.4, and 52.8 ± 27.9 ng/mL following SQ injection and 288 ± 47, 389 ± 154, 633 ± 194, and 10,066 ± 7059 ng/mL following IP injection for 3, 6, 9, and 20 mg/kg doses, respectively. The plasma C_max following the SQ route was therefore 19- to 190-fold lower than that following the IP route. This finding is consistent with a 31-fold higher apparent clearance following the SQ route compared with the IP route at the highest dose administered. The brain concentrations in the mice treated with a 3 mg/kg dose of the mutant fusion protein were lower than those in the nonmutant WT-treated mice. No reticulocyte suppression was observed at the 3 mg/kg SQ dose of TfRMAb-N292G-EPO. However, reticulocyte suppression increased with an increase in dose and area under the plasma concentration-time curve (AUC) for both the IP and SQ routes. Overall, elimination of Fc N-linked glycosylation, to mitigate TfRMAb effector function side effects, has a profound effect on the plasma exposure of TfRMAb-N292G-EPO at therapeutic as well as high doses (3-20 mg/kg). This effect is more pronounced following SQ injection. The low plasma concentrations of the mutant fusion protein following a 3 mg/kg dose resulted in negligible brain uptake. The beneficial rescue of reticulocyte reduction by the N292G mutation is a function of AUC and is negated at high doses of the N292G mutant.