Novel antiinflammatory biologics shaped by parasite-host coevolution

Abstract

Parasitic helminth infections, while a major cause of neglected tropical disease burden, negatively correlate with the incidence of immune-mediated inflammatory diseases such as inflammatory bowel diseases (IBD). To evade expulsion, helminths have developed sophisticated mechanisms to regulate their host's immune responses. Controlled experimental human helminth infections have been assessed clinically for treating inflammatory conditions; however, such a radical therapeutic modality has challenges. An alternative approach is to harness the immunomodulatory properties within the worm's excretory-secretory (ES) complement, its secretome. Here, we report a biologics discovery and validation pipeline to generate and screen in vivo a recombinant cell-free secretome library of helminth-derived immunomodulatory proteins. We successfully expressed 78 recombinant ES proteins from gastrointestinal hookworms and screened the crude in vitro translation reactions for anti-IBD properties in a mouse model of acute colitis. After statistical filtering and ranking, 20 proteins conferred significant protection against various parameters of colitis. Lead candidates from distinct protein families, including annexins, transthyretins, nematode-specific retinol-binding proteins, and SCP/TAPS were identified. Representative proteins were produced in mammalian cells and further validated, including ex vivo suppression of inflammatory cytokine secretion by T cells from IBD patient colon biopsies. Proteins identified herein offer promise as novel, safe, and mechanistically differentiated biologics for treating the globally increasing burden of inflammatory diseases.

Keywords: IBD; helminth; hookworm; protein; therapeutic.

Fig. 1.

Cell-free expression of recombinant hookworm ES proteome. (A) Predicted ES protein sequences were assembled into protein families (Pfam) using Blast2Go. Where Blast2Go did not assign a Pfam status, a manual BlastP search was performed. If manual searches did not assign a Pfam, the protein was deemed to be “hypothetical.” (B) Schematic depiction of the production of the hookworm cell-free secretome library using a L. tarentolae in vitro translation system. (C) Representative IVTRs of nine hookworm recombinant proteins fused to eGFP that were monitored over time by measuring the RFU. (D) Protein expression in IVTRs was visualized by Coomassie-stained SDS-PAGE gel loaded with IVTRs shown in B. (E) the same gel as shown in D but scanned for eGFP fluorescence. The fluorescent bands represent eGFP fluorescence of hookworm recombinant proteins.

Fig. 2.

Twenty of the tested IVTRs alleviated experimental colitis. BALB/c mice were injected intraperitoneally with 200 µL of IVTRs (protein concentration: 100 μg/mL) 1 d prior to intrarectal administration of TNBS and were monitored daily for 3 d before mice were killed. n = naïve. (A) Mean percent change of initial weight from a representative experiment where 12 IVTRs containing different hookworm proteins fused to eGFP were screened for activity compared to eGFP-only IVTR control and naïve mice (n = 5 per group). (B) Clinical scores (mean and individual data points) combining fecal consistency, motility, piloerection and weight change at day 3. (C) Macroscopic colon pathology scores (mean and individual data points) at day-3 necropsy. *P ≤ 0.05, **P ≤ 0.01, Mann–Whitney U test compared to eGFP control. (D) Combined statistical analyses of the efficacy of all 78 expressed hookworm protein-containing IVTRs compared to respective eGFP-IVTR control. The x axis depicts the difference in mean of the combined z-scores of clinical outcomes (weight loss, colon length, clinical scores, and macroscopic scores) between treatment and control (eGFP IVTR), and the y axis shows the P value where P < 0.01 or –log10 > 2 were regarded as significant. Proteins that achieved significant protection in both categories are highlighted in red (or turquoise) and were selected for further validation, whereas proteins that failed to achieve significant protection in both categories are shown in dark blue. Proteins shown in turquoise have previously been reported (in purified recombinant form) to alleviate experimental colitis. The proteins that were subsequently expressed in mammalian cells are annotated with their protein name and Pfam status.

Fig. 3.

Validation of efficacious proteins in TNBS colitis after expression in mammalian cells. BALB/c mice were injected intraperitoneally with 20 µg of recombinant hookworm proteins generated in Expi293F human embryonic kidney cells 1 d prior to intrarectal administration of TNBS and daily thereafter until termination. Control mice received either recombinant BSA expressed using the same production methods or PBS vehicle. (A) Mean percent change of initial weight. (B) Clinical scores (mean and individual data points). (C) Macroscopic scores of colon pathology (mean and individual data points). Data are combined from three individual experiments, where n = 10 to 20 for each treatment group. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, Mann–Whitney U test compared to PBS control.

Fig. 4.

Lead proteins reduce inflammatory cytokine release by intestinal T cells from human colitis patients. Intraepithelial lymphocytes and lamina propria cells were isolated and combined from fresh colon biopsies collected from UC patients. Cells were treated with either PBS control, purified recombinant hookworm proteins, or bovine serum albumin, generated in Expi293F cells, and then stimulated with αCD3/CD28 Dynabeads (TCR stimulation). Patient-to-patient variability was accounted for by normalizing all data to the PBS (vehicle)-treated samples (100%, indicated by a dotted line in Fig. 4 A–C). Ac-FAR-2 was used at 10 μg/mL based on earlier studies in our laboratory, and all other proteins at 50 μg/mL. Positive control cells were treated with CSA or were left unstimulated (PBS). (A–C) Cell-free supernatants were collected and analyzed after 3 d of culture for TNF, IFNγ, and IL-17A secretion. Values were normalized to PBS-treated and stimulated conditions for each respective patient (dotted line). Data are combined from two independent sample collection days, with each colored symbol representing an individual patient sample (n = 5 to 12 per treatment group) and the mean value for each treatment shown as a horizontal line. (D) Heat map depicting the combined mean cytokine values from A–C. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, one-sample t test.

Fig. 5.

Ac-FAR-2 binds hydrophobic ligands. Binding was assessed by spectrofluorometric analysis. (A) Binding of all-trans retinol to Ac-FAR-2, with no or competitive displacement with successive additions of 10-fold increasing concentrations of oleic acid. (B) Binding of DAUDA to Ac-FAR-2, with no or successive additions of 10-fold increasing concentrations of arachidonic acid. (C and D) Failure of individual lead proteins Ac08034, Ac22177, and Ac07727 to bind all-trans retinol (C), DAUDA (D), or ANS (E) compared with Ac-FAR-2. Fluorescence intensity was converted to relative arbitrary units (AU). The small sharp peaks at shorter wavelengths are from water Raman scatter.