The secretome of adult murine hookworms is shaped by host expression of STAT6

Abstract

Co-evolutionary adaptation of hookworms with their mammalian hosts has selected for immunoregulatory excretory/secretory (E/S) products. However, it is not known whether, or if so, how host immunological status impacts the secreted profile of hematophagous adult worms. This study interrogated the impact of host Signal transducer and activator of transcription 6 (STAT6) expression during experimental evolution of hookworms through sequential passage of the life-cycle in either STAT6 deficient or WT C57BL/6 mice. Proteomic analysis of E/S products by LC-MS showed increased abundance of 15 proteins, including myosin-3, related to muscle function, and aconitate hydratase, related to iron homeostasis. However, most E/S proteins (174 of 337 unique identities) were decreased, including those in the Ancylostoma-secreted protein (ASP) category, and metallopeptidases. Several identified proteins are established immune-modulators such as fatty acid-binding protein homolog, cystatin, and acetylcholinesterase. Enrichment analysis of InterPro functional categories showed down-regulation of Cysteine-rich secretory proteins, Antigen 5, and Pathogenesis-related 1 proteins (CAP), Astacin-like metallopeptidase, Glycoside hydrolase, and Transthyretin-like protein groups in STAT6 KO adapted worms. Taken together, these data indicate that in an environment lacking Type 2 immunity, hookworms alter their secretome by reducing immune evasion proteins- and increasing locomotor- and feeding-associated proteins.

Introduction:

Hookworm excreted and secreted (E/S) products have long been hypothesized to direct the host immune response towards a tolerogenic state that favors both parasite survival and reduced host tissue immunopathology. Examples of host immune modulation by helminth E/S isolates are well-documented (1–3) as well as the central role of host Type 2 immune responses in driving host-protective immunity (4). STAT6 is a critical transcription factor that drives Type 2 immune effector mechanisms in myeloid, lymphoid and epithelial cell lineages required for expulsion of enteric-stage parasitic nematodes (5–8) and to protect the host from harmful inflammatory responses (9). We recently demonstrated that hookworms (Nippostrongylus brasiliensis, Nb) adapted to the STAT6 deficient host environment have increased size and fecundity, but decreased mRNA transcripts encoding E/S proteins (10). Curiously, the progeny of these worms elicited stronger inflammatory responses than WT-adapted control worms when reintroduced to wildtype hosts, which was associated with increased host mortality (10). This supports the hypothesis that STAT6 immunity selects for hookworms that express immune-dampening E/S products. Here, we employed LC-MS analysis to profile the secretome of adult-stage Nb, adapted to WT or STAT6 KO mice for up to 26 generations. Data show an overall reduced number of detected proteins in the STAT6 KO condition, with preferential loss of known immune-modulating proteins. These results identify the specific hookworm E/S proteins that are mutable in the presence or absence of STAT6 immune pressure; information that could provide useful therapeutic targets aimed at reducing the prevalence of parasitic helminth infection.

Materials and Methods:

Isolation of hookworm E/S.

Nb were adapted to either WT or STAT6 KO mice as previously described (10), via serial passage of cohorts of worms for 26 (STAT6 KO) or 27 (WT) generations, in compliance with the US National Research Council’s Guide for the Care and Use of Laboratory Animals and the US Public Health Service’s Policy on Humane Care and Use of Laboratory Animals, as approved by the University of Pennsylvania IACUC (protocol number 805911, date of approval December 10, 2021). A total of 4 STAT6 KO mice and 3 WT mice were infected with adapted parasites on 3 different dates, and parasitic adult male and female Nb were isolated from the small intestine at day 7 post infection, as previously described (10). For each experimental date, all isolated worms per condition were pooled together, and further cleaned and maintained ex vivo as follows. Worms were washed 3 times with PBS + penicillin/streptomycin, incubated for 30 min at 37°C in RPMI media supplemented with 10% penicillin/streptomycin, 500 ug/mL gentamicin, and 10% fungizone, washed 3X with sterile PBS, and then resuspended in RPMI media supplemented with 1% glucose, 1% penicillin/streptomycin, 50ug/mL gentamicin, and 1% fungizone. Worms were plated in multiple wells of a 6-well plate at concentration of 2000 worms per mL, with 1 mL per well, and kept in a 5%CO2 37°C incubator, for up to 7 days. Supernatant was collected after 3, 5, and 7 days of culture. Because the total protein yields of individual collections were low, these timepoints were pooled and frozen at −80°C for further analysis. The protein concentration per worm was estimated from these E/S isolates, as follows. For each host infection experimental date (3 WT, and 4 STAT6 KO), all wells were pooled, and concentrated with a 10Kd Amicon filter (milipore), and the total protein was measured from pooled supernatant wells by a Bradford assay. After collection of the supernatant, the number of worms in each well was counted and totaled for the experimental date. Protein per worm was estimated by dividing the total protein by total worms for each experimental date, and means were compared between the two conditions by t-test.

Protein identification by LC-MS.

Hookworm E/S containing supernatant was prepared for LC-MS as follows. Samples were pooled over all experimental dates, concentrated over 10Kd Amicon filters (milipore), to a volume of 60uL, and 5 uL was run on a 10% polyacrylamide gel visualized with silver staining. The remaining volume was dehydrated by speed vac, and samples were digested with trypsin and run on a 90-minute LC standard gradient followed by Mass spectrometry on the Thermo Q Exactive instrument. Protein abundance and identity were computed from mass to charge (m/z) and abundance spectra, using MaxQuant 1.6.17.0., against predicted protein sequences of Nippostrongylus brasiliensis genome (PRJNA994163, accessibility link: https://osf.io/gpyuc/), the UniProt Nippostrongylus brasiliensis proteome reference database (UP000271162), and a contaminant database. The resulting peptide alignments were filtered to remove contaminant sequences and compiled to a list of protein groups of 1–27 predicted Nb proteins with uniquely matching peptide sequences, and abundance of protein groups was computed, based on the MS peak area under the curve of relative peptide intensity using MaxQuant.

Data analysis.

Comparison of relative E/S protein abundances between WT and STAT6 KO host derived worms, was done by computing the scaled Log2(Fold-change) value of STAT6 KO over WT hosts as follows. Protein abundance values for each identified protein group were scaled by adding a constant value corresponding to half the lowest non-zero abundance value, and then Log2 transformed. Fold-change values were plotted using the ggplot package in R. InterPro names were annotated to protein ID’s as previously described (10), and matched to each protein group based on the protein IDs they contained. Over-represented InterPro entries within all measured protein groups across both conditions were identified using a hypergeometric test in the ClusterProfiler package in R, with all Nb predicted proteins as background (11,12). The network interaction plot of enriched protein sets was plotted using the cnetplot function within ClusterProfiler in R.

Results:

To determine if STAT6 KO adaptation leads to an altered repertoire of excreted and secreted (E/S) products in the murine hookworm Nb, we measured relative protein abundance of E/S from day 7 adult stage parasites whose ancestral cohorts were passaged in STAT6 KO or WT mice for 26 or 27 generations, respectively (Fig. 1A), as previously described (10). E/S was collected from the supernatant of day 7 adult parasitic stage Nb (mixed males and females), from 4–5 hosts per condition, repeated for 3 or 4 collections, WT or STAT6 KO, respectively. In the WT condition, 2289 total worms were collected, (763 ±537 average total each replicate), and 5543 total STAT6 KO derived worms (1108 ±171 average total each replicate). The total E/S protein released was normalized to the worm count and trended higher in the WT condition, without reaching statistical significance. (WT 0.081±0.0144 vs STAT6KO 0.047±0.017 average ug/worm, p-value = 0.16, 3 WT or 4 STAT6 KO replicates). E/S isolate proteins ranged from 15 to 220 kDa, with only slight in differences relative band intensities (Fig. 1B, arrows), suggesting a difference in relative abundance of distinct proteins.

Fig. 1:

A. collection of E/S protein. B. Silver stained 10% gel of concentrated supernatant of WT or STAT6 KO mouse adapted Nb ex vivo cultures of day 7 adult males and females, pooled from 3 (WT-adapted) or 5 (STAT6 KO-adapted) independent isolates. The magenta arrow indicates the presence of a band in WT but not STAT6 KO conditions, and the orange arrow indicates the converse.

To analyze the differences in protein composition, pooled isolates were subject to trypsin digest and LC–MS to identify peptide sequences. These peptide sequences aligned with 337 distinct protein groups from predicted Nb protein sequences. Each group contained between 1 and 27 protein identifiers, with a median of 3 protein identifiers per group. The number of identified Nb proteins in the E/S mapped to 1100 predicted protein IDs. Interestingly, 89 out of 337 protein groups were below the limit of detection in the STAT6 KO condition, while 14 were below this limit in the WT condition, suggesting that WT E/S had a more diverse repertoire of secreted proteins. Strikingly, there were 174 protein groups with increased abundance in the WT group (Log2FC < −2), and only 15 increased in the STAT6 KO group (Log2FC > 2) (Fig. 2A). The most increased proteins in STAT6 KO derived E/S included Myosin-3 (5.44 Log2FC), which suggests increased release of muscle components, and Aconitate hydratase (4.91 Log2FC), which is implicated in iron sensing (13). The two most increased proteins in the WT condition include a Metalloendopeptidase (−9.6 Log2FC) and a Sperm coat protein (SCP) domain-containing protein (−9.1 Log2FC), which are both known components of adult stage E/S (14).

Fig. 2:

A. Scatterplot of protein intensity fold-change by protein group ID; color of circles indicates increased intensity in WT (red), or increased intensity in STAT6 KO (blue).

We next sought to understand whether proteins of similar functional groups were enriched in association with presence or absence of STAT6 host immunity in our E/S. We categorized each measured protein group ID by functional or structural InterPro entries (15–18), and then performed enrichment analysis. We found a total of 17 enriched categories (FDR p-value < 0.05). The top 10 enriched InterPro categories included numerous families found in hookworm E/S from previous studies (Fig. 3A) (14). Notably, protein domains falling into the Ansylostoma-secreted protein (ASP) family such as CAP domain and Allergen 5 like proteins were strongly over-represented, overlapping with 30 and 8 protein groups, respectively. The fold-change values of the protein groups within these categories were overwhelmingly negative (Fig. 3A), the median Log2(Fold-Change) for each category was significantly less than 0 for all but two displayed categories. This data suggests that WT adapted worms secrete a greater abundance and diversity of E/S associated proteins, such as ASP, metallopeptidases and transthyretin-like groups.

Fig. 3:

A. Top 10 enriched InterPro protein entries and number of overlapping protein group IDs in parentheses, by enrichment analysis of all identified 338 protein groups, with false discovery rate (FDR) q-value < 0.05 and redundant InterPro entries (i.e. more than 95% identical protein group ID overlapping) excluded. InterPro category (y-axis) plotted by Log2(Fold Change) x-axis, for detected proteins (circles) within each group, overlayed by box-plots. Color of circles indicates increase (Log2(Fold Change)>2) or decrease (Log2(Fold Change)< −2 in STAT6 KO protein abundance. * Bonferroni adjusted p-value < 0.05 for Wilcoxon test of median Log2(Fold Change) less than 0.

A substantial number of protein groups fell into multiple InterPro categories, and we therefore assessed the interactions of the enriched InterPro categories by plotting InterPro entries, protein groups and their Log2FC values as an interaction network (Fig. 4A). This depiction revealed three distinct clusters, containing common protein group IDs, among the top enriched functional groups. First, the bottom right cluster, was characterized by the following five InterPro entries: Peptidase, Metallopeptidase [IPR006026], Peptidase M12A [IPR001506], Metallopeptidase, catalytic domain [IPR024079], Metallopeptidase, nematode [IPR017050], and Astacin-like metallopeptidase domain [IPR034035], implying that this cluster of proteins shares the functions of these categories. The negative Log2(Fold-change) values indicates that this cluster is less abundant when hosts lack STAT6 immunity. Second, the top left cluster reveals InterPro categories CAP superfamily [IPR035940], CAP domain [IPR014044], and Cystine-rich secretory protein-related [IPR001283] sharing protein group IDs and a subset of these falling into the venom allergen 5-like [IPR002413] and Allergen V5/Tpx-1-related [IPR018244] InterPro entries. Finally, protein groups within the Transthyretin-like [IPR001534, IPR038479] entry (Fig. 4A, top right cluster) are not represented in other nodes on the network plot, possibly indicating a distinct role of these proteins. As in Fig. 3A, the protein fold-change abundance for individual protein group IDs within all enriched nodes were largely negative, indicating decreased levels of these functional groups in STAT6 KO adapted hookworm E/S.

Fig. 4:

A. Network enrichment plot of all E/S protein groups (small nodes), colored by Log2(Fold Change); green circular nodes are the top 11 enriched InterPro groups, whose relative size indicates the number of protein ID’s overlapping with the InterPro group.

Given that the highly enriched ASP category of proteins has been hypothesized to contain immune modulating proteins (19), we inquired whether our dataset contained any proteins with prior evidence of immune-modulatory functions. We identified several proteins, all of which were decreased in abundance in the STAT6 KO condition, which displayed immune-modulatory properties (Table 1). These included Fatty acid-binding protein homolog, Cystatin, acetylcholinesterase, and Cathepsin. These have been shown to suppress inflammatory immune responses (Fatty acid-binding protein, cystatin and Cathepsin), or induce immune-regulatory responses (Cystatin), and may suggest a reduced inflammatory property of the WT-derived worm E/S. In addition, 120 identified proteins (60 WT-increased and 4 STAT6 KO increased) had entirely unknown functions, and were not described by existing InterPro entries, which may provide novel candidate molecules for study of host immune modulation.

Table 1:

Table of down-regulated immunomodulatory proteins identified in STAT6 KO-adapted hookworm E/S.

Protein name	UniProt ID	Log2FC (STAT6 KO/WT)	Immunomodulatory role	Evidence
Fatty acid-binding protein homolog	A0A158R0W4	−8.3	Suppression of LPS-induced inflammatory cytokines; Increased induction of M2 macrophages	(1,25,30)
Cystatin	Q966W0	−6.25	Inhibition of antigen processing	(1,24)
acetylcholinesterase	O61587	−3.4	Possible interference with host peristalsis, and increased proinflammatory and reduced type 2 cytokine production in infection with transgenic AChE-expressing trypanosome	(1,26,31)
Cathepsin	A0A0N4YGL1	−6	Degradation of fibrinogen and fibrin	(1,27)

Discussion:

In this study, we tested the hypothesis that sequential passage of Nb within murine hosts lacking STAT6 immune selection would adversely impact the normal adult worm E/S proteome. In a recent study, we generated STAT6 KO adapted hookworms by serial passage for 26 generations, which resulted in worms with increased size and fecundity compared to WT host-adapted controls. Notably, STAT6 KO adapted worms had decreased expression of E/S encoding-genes and this altered Nb line elicited significantly increased immune cell recruitment, cytokine release and immune cell activation than parental Nb when inoculated into WT hosts. In the present study, we found that adaptation of Nb in STAT6 deficient hosts led to a dramatic loss of E/S protein abundance and diversity, with a preferential reduction in the abundance of E/S proteins with known or putative immunomodulatory functions. Interestingly, among the few proteins that were increased in STAT6 KO adapted worms, we found myosin-3, which together with a larger body size (10), suggests increased growth of body-wall muscle. Moreover, we also noted a marked increase in the iron-sensing protein Aconitate hydratase, which is consistent with the increased hemoglobin consumption of STAT6 KO adapted worms as compared to WT worms (10). Taken together, our data in this report support a hypothesis that selection of parasitic nematodes to host immune pressure driven by the canonical Type 2 transcription factor STAT6 increases the abundance and diversity of E/S proteins that support immunomodulation at the expense of protein involved in somatic tissue and feeding.

A key function of hookworm E/S is to modify the immune response of the host, which enables the host to tolerate chronic parasite infection. Our data found the majority of E/S proteins are less abundant when STAT6-related immune pressure is removed, which supports this hypothesis. Hookworms secrete the largest array of E/S molecules once they are established as reproductive adults in the host gut, as compared to earlier stages of development (14), which drives a bias towards Type 2 immunity (1,4,20). Indeed, treatment of mice with E/S isolate alone can increase Th2 cell development from naïve CD4 T cells and suppress Th1 cells through the immunosuppressive cytokine IL-10 (21). In our data, the top protein families that are represented in E/S from Nb include the CAP-domain Cysteine-rich SCP/TAPS family, also called Venom Allergen/Ancylostoma secreted protein-like (VAL) products (14). These were the most overrepresented categories (37 protein group IDs), and less abundant in the STAT6 KO condition (21 Log2FC < −2), suggesting a direct relationship between these molecules and STAT6-dependent immune function. Their immune-modulatory function is hypothesized based on the location and timing of expression, and capacity to bind host fatty acid ligands (22). VAL proteins are inhibitory to neutrophils and platelets, which may reduce inflammatory signaling (23). The next prominent protein family within our dataset that was also down-regulated was the Astacin-like, metallopeptidase, and Peptidase M12A category, which are present in both larval and adult stage E/S, and have purported tissue degradation functions (14), but can also cleave eotaxin, to limit eosinophil recruitment (1). In addition, we found several individual down-regulated proteins with immune-modulatory functions: cystatin impairs dendritic cells (24), fatty acid binding protein homolog promotes alternatively activated macrophages (1,25), acetylcholinesterase in E/S may impact host peristalsis and reduce type 2 cytokine secretion (26), and cathepsin degrades fibrinogen and fibrin (27). Among our 175 STAT6 KO down-regulated proteins, 10 of them had either purported or experimentally tested immune-modulatory functions (1). This collectively suggests that E/S selected by STAT6 immunity functions to drive immune modulation.

Our previous transcriptional profile also demonstrated an overall bias towards down regulated gene expression in STAT6 KO adapted worms, and an increased immunogenicity and mortality when this line was used to infect wildtype hosts (10). In our E/S proteomics dataset, we similarly found that few proteins had increased abundance in the STAT6 KO condition. Among these 15 protein groups, only two were from InterPro categories associated with hookworm E/S, including a metalloendopeptidase and SCP-domain containing protein. Given that such peptides experimentally caused increased release of IFNγ (28), it is possible that our E/S profile reflects a dysregulation of immunogenic versus immune-dampening peptides upon loss of STAT6 pressure. Indeed, we observed increases in IFNγ on re-challenge of wild-type hosts (10). Other protein groups that were more abundant in the STAT6 KO E/S likely reflect an increase in peptides corresponding to functions related to growth and fecundity, which is consistent with our previous findings (10). This includes two myosin protein group IDs, which could correlate to the expected increased body wall muscle mass of larger worms. Additionally, aconitate hydratase (4.9 Log2FC) has iron sensing capabilities (13), which may correspond to the increased amount of host blood consumption. And lastly, regulator of chromosome condensation (A0A158QZ77) (Log2FC = 2.4) is likely associated with cell turnover and embryogenesis (29). The profile of STAT6 KO up-regulated proteins reflects prioritization of growth and feeding which is congruent with a lack of need for immune evasion by these modified parasites.

This study found that selection of hookworms in STAT6 deficient mice dramatically alters the profile of secreted peptides and decreases the abundance of peptides associated with immune modulation. We found that expression of hookworm E/S is highly mutable in the absence of STAT6 immunity, with more than half of the identified proteins being reduced or absent in worms adapted to STAT6 KO hosts. Shortcomings to the current data include a lack of experimental repeats, which was required due to the limited quantity of E/S collected. In addition, as previously noted, the adaptation of Nb to STAT6 KO hosts was only done for a single cohort. Despite these limitations, our work has generated and characterized secreted profiles of hypo- and hyper- inflammatory murine hookworms. It provides further support to the hypothesis that STAT6 immunity selects hypo-inflammatory worms that and may offer insights to treating pathological inflammatory disorders also characterized by STAT6-driven type 2 immune responses.

Data Accessibility:

The data that supports the findings of this study are available in the supplementary material of this article