Heat-Labile Enterotoxin IIa, a Platform To Deliver Heterologous Proteins into Neurons

Abstract

Cholera toxin (CT) and the related heat-labile enterotoxins (LT) of Escherichia coli have been implicated as adjuvants in human therapies, but reactivity upon intranasal delivery dampened efforts to develop other clinical applications. However, each CT family member variant has unique biological properties that may warrant development as therapeutic platforms. In the current study, a nontoxic variant of the heat-labile enterotoxin IIa (LTIIa) was engineered to deliver heterologous, functional proteins into the cytosol of neurons. As proof of principle, the LTIIa variant delivered two cargos into neurons. LTIIa delivered β-lactamase efficiently into cells containing complex gangliosides, such as GD1b, as host receptors. LTIIa delivery of β-lactamase was sensitive to brefeldin A, an inhibitor that collapses the Golgi compartment into the endoplasmic reticulum, but not sensitive to treatment with botulinum neurotoxin D (BoNT/D), an inhibitor of synaptic vesicle cycling. LTIIa delivered a single-chain, anti-BoNT/A camelid antibody that inhibited SNAP25 cleavage during post-BoNT/A exposure of neurons. Delivery of functional, heterologous protein cargos into neurons demonstrates the potential of LTII variants as platforms to deliver therapies to inactivate toxins and microbial infections and to reverse the pathology of human neurodegenerative diseases.

Importance: This study engineered a protein platform to deliver functional, heterologous proteins into neurons. The protein platform developed was a variant of the heat-labile enterotoxin IIa (LTIIa) which lacked the catalytic domain, yielding a nontoxic protein. As proof of principle, LTIIa variants delivered two functional proteins into neurons, β-lactamase and a camelid antibody. These studies show the utility of LTIIa variants to deliver therapies into neurons, which could be extended to inactivate toxins and microbial infections and potentially to reverse the progression of neurological diseases, such as Alzheimer's disease and Parkinson's disease.

FIG 1

Primary structure of LTIIa. (Top) Schematic of LTIIa, βlac-LTIIa, and βlac_null-LTIIa. (Bottom) Purified LTIIa, βlac-LTIIa, and βlac_null-LTIIa were separated by SDS-PAGE and stained with Coomassie Blue. The ratio of B subunit to A subunit was determined by densitometry (B/A ratio is listed below each lane). Numbers at left are molecular masses in kilodaltons.

FIG 2

Dose-dependent delivery of βlac by βlac-LTIIa. (A) Rat primary cortical neurons were incubated with 40 nM βlac-LTIIa at 37°C for 60 min. Cells were incubated with CCF2-AM at RT for 30 min followed by IF to detect LTIIa bound (anti-HA, red) and βlac (anti-FLAG, magenta). Uncleaved CCF2 is shown in green, and cleaved CCF2 (CCF2_C) is shown in cyan. (B) Rat primary cortical neurons were incubated with 40 nM LTIIa or 0.1 to 40 nM βlac-LTIIa at 37°C for 60 min. Cleavage of CCF2 was quantified using the ratio of fluorescence of cleaved CCF2 to that of CCF2 as a function of added βlac-LTIIa.

FIG 3

Delivery and separation of βlac from the B subunit during βlac-LTIIa entry into neurons. (A) Rat primary cortical neurons were incubated with 40 nM βlac-LTIIa at 4°C or 37°C for 60 min. Cells were washed, followed by IF staining, using anti-HA antibody (green) and anti-FLAG antibody (red). (B) Representative colocalization between HA and FLAG staining was shown by a cytofluorogram with Pearson’s coefficient (PC) determined. Bar, 20 µm.

FIG 4

LTIIa delivers cargo (βlac) more efficiently into GD1b-enriched Neuro-2a cells than GM1a-enriched Neuro-2a cells. (A) Neuro-2a cells were loaded with 10 µg/ml of ganglioside GD1b or GM1a in DMEM with 0.5% FBS at 37°C for 3 h. Cells were washed and incubated with 40 nM βlac-LTIIa or LTIIa at 37°C for 60 min. Cells were loaded with CCF2-AM at RT for 30 min, followed by IF staining using anti-HA antibody (red). Uncleaved CCF2 is shown in green, and cleaved CCF2 (CCF2_C) is shown in cyan. (B) Cleavage of substrate CCF2 was quantified using the CCF2_C/CCF2/HA ratio of fluorescent intensities. (C) Neuro-2a cells were loaded with GD1b and then incubated with 40 nM βlac-LTIIa, βlac_null-LTIIa, or LTIIa at 37°C for 60 min alone or with 0.1 µg/ml of brefeldin A (BFA). Cells were washed and were loaded with CCF2-AM at RT for 30 min. Cleavage of CCF2 was quantified using the CCF2_C/CCF2/HA ratio of fluorescent intensities. Data were analyzed by two-tailed Student’s t test. *, P < 0.05; **, P < 0.005; ***, P < 0.001. Bar, 20 µm.

FIG 5

LTIIa delivers βlac more efficiently to neurons than to Neuro-2a cells and HeLa cells. A 40 nM concentration of βlac-LTIIa was incubated with rat cortical neurons; Neuro-2a cells loaded with GD1b or GM1a; and HeLa cells loaded with GD1b, GM1a, GM2, or GD2 at 37°C for 60 min. Cells were loaded with CCF2-AM at RT for 30 min followed by IF staining using anti-HA antibody (red). Cleavage of substrate CCF2 was quantified using the CCF2_C/CCF2/HA ratio of fluorescent intensities. The dashed line was drawn based on detectable translocation by IF. Data were analyzed by two-tailed Student’s t test. *, P < 0.05; **, P < 0.005; ***, P < 0.001.

FIG 6

βlac-LTIIa cleaves CCF2 in BoNT-intoxicated primary cortical neurons. (A) Rat cortical primary neurons were incubated with 2 nM BoNT/D at 37°C for 16 h. BoNT-treated neurons were incubated with 40 nM βlac-LTIIa at 37°C for 60 min alone or with 0.1 µg/ml brefeldin A (BFA), washed, and loaded with CCF2-AM at RT for 30 min followed by IF staining, using anti-HA antibody (red) and anti-VAMP2 (magenta), which recognizes only full-length VAMP2. Uncleaved CCF2 is shown in green, and cleaved CCF2 (CCF2_C) is shown in cyan. (B) Cleavage of VAMP2 was quantified using the VAMP2/HA ratio of fluorescent intensities. (C) Cleavage of substrate CCF2 was quantified using the CCF2_C/CCF2/HA ratio of fluorescent intensities. Data were analyzed by two-tailed Student’s t test. ***, P < 0.001. Bar, 20 µm.

FIG 7

VHH-B8-LTIIa protects neurons from BoNT/A intoxication. (A) Schematic of VHH-B8-LTIIa. (B and C) A 1 nM concentration of BoNT/A (B) or BoNT/D (C) was incubated with rat cortical neurons at 37°C for 2 h, when toxin was removed and the indicated amount of VHH-B8-LTIIa (B8) was added to neurons for an additional 3 h. Cells were fixed and incubated with Alexa 647-wheat germ agglutinin (WGA; magenta) for 30 min as a cell marker. The IF assay stained for HA (red) and cleaved SNAP25 (SNAP25c, green) in BoNT/A-treated cells or intact VAMP2 (green) in BoNT/D-treated cells. (D) Cleavage was quantified by measuring the SNAP25c/WGA ratio of fluorescence intensities for BoNT/A-treated cells and the VAMP2/WGA ratio of fluorescence intensities for BoNT/D-treated cells. (E) BoNT/A or BoNT/D (1 nM) was incubated with rat cortical neurons at 37°C for 2 h, when toxin was removed and the indicated amount of VHH-B8-LTIIa (B8) was added with neurons overnight. Cleavage was quantified as described for panel D. Data were analyzed as SEM by two-tailed Student’s t test. *, P < 0.05; **, P < 0.005. Bar, 20 µm. NS, not significant; o/n, overnight.

FIG 8

VHH-B8-LTIIa protects neurons from BoNT/A but not BoNT/B intoxication. (A) BoNT/A or BoNT/D (0.5 nM) with or without 20 nM VHH-B8-LTIIa (B8) was incubated with rat cortical neurons at 37°C for 24 h, when neurons were washed with PBS, lysed with radioimmunoprecipitation assay buffer, and separated by SDS-PAGE. WB was performed using anti-SNAP25 antibody (recognizing full-length and cleaved SNAP25) and VAMP2 (recognizing only full-length VAMP2) and β-actin (loading control [ctl]). (B) SNAP25 cleavage was quantified by measuring the SNAP25c/SNAP25 density ratios from three independent experiments. Data were analyzed by two-tailed Student’s t test. *, P < 0.05.