Assessment of an Immuno-Diagnostic Method for Hookworm-Related Cutaneous Larva Migrans Using Crude Extracts of Ancylostoma caninum

People can become infected with cutaneous larva migrans (CLM) through skin penetration by the infective zoonotic larvae of hookworms. Few studies have investigated CLM’s immunodiagnosis, and the existing studies were limited to crude somatic or excretory/secretory antigens (Ags) from adult worms. Here, we aimed to develop an indirect enzyme-linked immunosorbent assay (ELISA) to differentiate and diagnose hwCLM by detecting immunoglobulin (Ig)E, IgG, and IgG subclasses 1–4 (IgG1–4) against the somatic Ag of adult Ancylostoma caninum checkerboard titrations of adult A. caninum worm extract. Pooled serum controls were immunocharacterized using an indirect ELISA. The IgG1–4 and IgE results were unsatisfactory; however, the use of total IgG achieved results comparable to those of immunoblotting. Thus, we continued to analyze the IgG-ELISA using serum samples from patients with hwCLM and heterologous infections as well as from healthy controls. The sensitivity and excellent specificity of the total IgG-ELISA were 93.75% and 98.37%, respectively, and its positive and negative predictive values were 75% and 99.67%, respectively. Antibodies from five cases of angiostrongyliasis, gnathostomiasis, and dirofilariasis cross-reacted with the somatic Ag of adult A. caninum. This new assay can adequately serodiagnose hwCLM when combined with clinical features and/or histological examination.


Introduction
Cutaneous larva migrans (CLM), also known as creeping eruption, sandworm eruption, plumber's itch, and serpiginous dermatitis, is caused by the intradermal penetration and migration of several larvae of helminths, mainly canine hookworms (zoonotic hookworms) and also including Ancylostoma caninum, A. braziliense, and A. ceylanicum, which are well-known causative agents of hookworm-related CLM (hwCLM), and other helminths such as Gnathostoma spp., Loa loa, Uncinaria stenocephala, Pelodera strongyloides, Strongyloides stenocephala, Dirofilaria repens, Fasciola (ectopic feature), and Schistosoma may mimic similar migratory skin lesions [1][2][3][4][5][6]. HwCLM infections are mainly distributed across tropical and subtropical countries [2,4]. The species of zoonotic hookworms that infect humans vary throughout different regions [3,7]. Globally, 1.3 billion people are infected by hookworms, Clinical manifestations related to cutaneous larva migrans or creeping eruption, such as pruritus, swelling pain, and the appearance of an erythematous track on the skin of the trunk, buttock, hand, leg, etc. History of traveling or/and exposure to soil or beach land where dogs were around Tested negative for gnathostomiasis with an IgG-immunoblotting test and for strongyloidiasis with an IgG-ELISA, including a demonstration of two hwCLM cases and a section of larva structure from the patient via tissue biopsy (Figures 1-3)

ELISA
A checkerboard titration method with a twofold condition was used to optim indirect ELISA protocol. Immunocharacterizations of the ELISA checkerboard curves of the optical density (OD) values showed that total IgG had the best sep between positive and negative pooled controls, but the IgG1-4 and IgE curves wer isfactory ( Figure 4). Therefore, only total IgG was used for the ELISA, carried o cordance with the procedure of Yoonuan et al. [46]. The optimal ELISA conditio 0.25 µg/mL A. caninum Ag in 0.05 M carbonate-bicarbonate buffer at pH 9.6 with dilution of 1:1600 and an anti-human IgG conjugate dilution of 1:2000. Briefly, t of a 96-well plate were coated with the Ag, washed, and blocked with 1% skim m 0.02% NaN3 in phosphate-buffered saline (PBS). Each diluted serum sample reac

ELISA
A checkerboard titration method with a twofold condition was used to optim indirect ELISA protocol. Immunocharacterizations of the ELISA checkerboard curves of the optical density (OD) values showed that total IgG had the best se between positive and negative pooled controls, but the IgG1-4 and IgE curves wer isfactory ( Figure 4). Therefore, only total IgG was used for the ELISA, carried o cordance with the procedure of Yoonuan et al. [46]. The optimal ELISA conditio 0.25 µg/mL A. caninum Ag in 0.05 M carbonate-bicarbonate buffer at pH 9.6 with dilution of 1:1600 and an anti-human IgG conjugate dilution of 1:2000. Briefly, t of a 96-well plate were coated with the Ag, washed, and blocked with 1% skim m 0.02% NaN3 in phosphate-buffered saline (PBS). Each diluted serum sample reac

ELISA
A checkerboard titration method with a twofold condition was used to optimize the indirect ELISA protocol. Immunocharacterizations of the ELISA checkerboard titration curves of the optical density (OD) values showed that total IgG had the best separation between positive and negative pooled controls, but the IgG 1-4 and IgE curves were unsatisfactory ( Figure 4). Therefore, only total IgG was used for the ELISA, carried out in accordance with the procedure of Yoonuan et al. [46]. The optimal ELISA conditions were 0.25 µg/mL A. caninum Ag in 0.05 M carbonate-bicarbonate buffer at pH 9.6 with a serum dilution of 1:1600 and an anti-human IgG conjugate dilution of 1:2000. Briefly, the wells of a 96-well plate were coated with the Ag, washed, and blocked with 1% skim milk-and 0.02% NaN 3 in phosphate-buffered saline (PBS). Each diluted serum sample reacted with Ag at 37 • C for 1.5 h. After washing, the immunoreaction was combined with diluted anti-human IgG Ab conjugated with horseradish peroxidase (SouthernBiotech, Birmingham, AL, USA). After incubating the plate at 37 • C for 1.5 h, the reaction was visualized via incubation for 30 min with ABTS substrate solution (2,2-azino-di-[3-ethyl-benzthiazoline sulfonate]; Sigma-Aldrich Canada Co., Oakville, ON, Canada). The reaction was stopped by adding 1% sodium dodecyl sulfate (SDS), and absorbance was measured using a microplate reader at a wavelength of 405 nm. by adding 1% sodium dodecyl sulfate (SDS), and absorbance was measured using a microplate reader at a wavelength of 405 nm.

Observation of Crude Extract Profiles via Staining and Antibodies
Because the worms were stored at −80 °C for several years, ice crystals may have damaged the internal structures of the frozen worms and proteins during processing, which involved various temperatures from −12 °C to −75 °C [47,48]. Somatic extract patterns were demonstrated by transferring 15 and 30 µg of the extract into SDS-

Observation of Crude Extract Profiles via Staining and Antibodies
Because the worms were stored at −80 • C for several years, ice crystals may have damaged the internal structures of the frozen worms and proteins during processing, which involved various temperatures from −12 • C to −75 • C [47,48]. Somatic extract patterns were demonstrated by transferring 15 and 30 µg of the extract into SDS-polyacrylamide gel via electrophoresis (ATTO Corporation, Tokyo, Japan) with 5% stacking and 13% separating gels, which was followed by simple staining. The two extract concentrations were treated with sample buffer (1.5×; 0.25 M Tris-HCl, pH 6.8, 1.75% SDS, 3.75% mercaptoethanol, 7.5% glycerol, and 0.008% bromophenol blue) at 100 • C for 3 min in a heat block (Accublock™; Labnet International Inc., Edison, NJ, USA); then, the proteins were separated via SDS-PAGE at a constant current of 20 mA, performed in accordance with the protocol of Dekumyoy et al. [49] with slight modifications. Finally, the gel was stained with Coomassie brilliant blue and destained.
As the IgG 1-4 and IgE checkerboard titration results were unsatisfactory (Figure 4), we performed immunoblotting [49] to observe the reactions between the Ag and IgG 1-4 and IgE. A blot of the somatic extracts was analyzed based on the Ab reactions. The electrophoresed extracts (30 µg) were transferred to a nitrocellulose membrane using a semi-dry transfer cell (ATTO Corporation). The nonbinding sites on the membrane were blocked with 3% skim milk-0.02% NaN 3 in PBS for 1 h on a rocking platform. Each strip was incubated in a 1:50 dilution of pooled positive serum from the hwCLM samples in 0.02% NaN 3 -PBS-Tween20 (PBS-T) overnight on a rocking platform. The strips were then incubated with individual anti-human Abs against total IgG, IgG 1-4 , and IgE (SouthernBiotech) and conjugated with horseradish peroxidase at a dilution of 1:1000 in PBS-T for 2 h. Afterward, the membrane strips were treated with the substrate solution (2,6-dichlorophenol indophenols)-H 2 O 2 . The reaction was stopped by washing them with distilled water, and the Ag-Ab complex formations were observed.

Data Analysis
The cut-off values of the indirect ELISA were determined via analysis of a receiver operating characteristic curve (95% confidence interval) using PASW Statistics for Windows, version 18.0. (SPSS Inc., Chicago, IL, USA). Sensitivity, specificity, and predictive values were calculated using the 2 × 2 table method [50]. Scatter plots of the OD values obtained from the indirect ELISA were created using GraphPad Prism, version 6.00 for Windows (GraphPad Software, Inc., La Jolla, CA, USA).

Analysis of Crude Extract Profiles
The concentration of the extract containing both female and male worms was 1 mg/mL.

Correlations between Immunoglobulins and Antigens
After the checkerboard titrations (Figure 4) The reactions produced many strong distinct reactive bands for total IgG but not for IgG [1][2][3][4] or IgE ( Figure 5B). This thereby demonstrates the poor IgE and IgG 1-4 reactions of the positive pooled serum controls at 1:200 to 1:1600 dilutions in the checkerboard titrations via immunoblot.

Full Scale IgG-ELISA
All serum samples were subjected to IgG-ELISA. Of the 16 hwCLM samples, 1 was a false negative, and the OD values of the other 15 samples ranged from 1.111 to 0.452 (cut-off value: 0.451). Most of the hwCLM OD values ranged from 0.452 to 0.556. The false negative case had an OD value of 0.286. There were two high OD values in the hwCLM cases, 1.111 and 1.078, which separated them from the group. Therefore, the sensitivity and specificity of the IgG-ELISA were 93.75% and 98.39%, respectively, and its positive and negative predictive values were 75% and 99.67%, respectively. Only three nematodiases yielded false positives toward the crude adult A. caninum Ag, i.e., angiostrongyliasis (2/15), gnathostomiasis (2/15), and dirofilariasis (1/4). Nonetheless, we achieved satisfactory results because we found no cross-reactions in the serum samples for 24 diseases: nematodiases (ascariasis, capillariasis, Bancroftian filariasis, Brugian filariasis, hookworm infection, strongyloidiasis, trichostrongyliasis, toxocariasis, trichinellosis, and trichuriasis), cestodiases (taeniasis saginata, neurocysticercosis, cystic echinococcosis, sparganosis, and hymenolepiasis nana), trematodiases (paragonimiasis, opisthorchiasis, fascioliasis, minute intestinal fluke infection, and schistosomiasis), and protozoan infections (amoebiasis, giardiasis, Blastocystis infection, and malaria) ( Figure 6). It was found that strongyloidiasis (11 cases), fascioliasis (3 cases), and schistsomiasis (4 cases) can cause CLM symptoms, but no false positive occurred. Negative control sera demonstrated good discrimination from the hwCLM cases; all negative control sera yielded true negative results, and only two samples were less distinct from others in this group.

Discussion
When focusing on IgG, several previous studies assessed total IgG but yielded poor results. In contrast, the selected crude A. caninum somatic Ags after our checkerboard titration reacted less with Abs against cestodiasis, trematodiasis, and protozoan infections, including ten nematode-associated diseases, and all OD values were below the cut-off value. This may be because of the small amount of the Ags used (0.025 µg/100 µL of coating buffer) and their low antigenicity to the Abs in those tested serum samples. Finally, our proposed ELISA conditions for evaluating IgG yielded 93.75% sensitivity and 98.39% specificity. All healthy control sera had low OD-ELISA values in response to the Ags and yielded similar OD values.
One hwCLM sample in our study had few IgG Abs to A. caninum somatic Ag; thus, it gave a false negative result. This patient had a clinical history of subcutaneous swelling and other erythematous skin changes and larval migrans, and his sero-tests were negative for gnathostomiasis and strongyloidiasis. The gnathostomiasis test was done via detection of a 24-kDa diagnostic band of the G. spinigerum Ag on an IgG immunoblot, and the strongyloidiasis test was performed via IgG-ELISA using a molecular weight cut-off Ag of Strongyloides stercoralis infective larvae; the test was performed at the Department of Helminthology, Faculty of Tropical Medicine. In 2003, a previous study of a patient showed clinical features of erythematous linear and serpiginous plaques. The patient tested positive for crude A. caninum Ags and negative for G. doloresi Ags using the IgG-ELISA. Researchers attempted to screen this patient's antibodies for other infections using a multiple-dot ELISA with 12 Ags: Dirofilaria immitis, Toxocara canis, Ascaris suum, Anisakis simplex, G. doloresi, Strongyloides ratti, Paragonimus westermani, P. miyazakii, Fasciola hepatica, Clonorchis sinensis, Spirometra erinacei, and Cysticercus cellulosae. The tests were negative, and the patient was diagnosed with A. caninum infection [7]. Additionally, the sensitivity and specificity of the proposed ELISA for diagnosing a single case of hwCLM in that study were suboptimal.
In our study, A. caninum somatic Ag was screened and did not cross-react with Abs from 24 parasitic diseases or with the negative controls. Only five sera samples from three nematodiasis patients cross-reacted with A. caninum somatic Ag. These false positives may not be encountered for serodifferentiation if the targeted habitats of those helminths are considered, i.e., Angiostrongylus worms mostly infect the brain. Dirofilaria immitis exists as nodules in the lungs or tissues because humans are the dead-end host for this worm, and Gnathostoma worms travel throughout the human body, organs, and especially the cutaneous tissues. All hwCLM cases in the present study were negative for gnathostomiasis ( Table 1).
The larval and immature stages of many helminths, including Gnathostoma, Strongyloides, sparganum, Paragonimus, and Fasciola, can cause CLM cases similar to those of hookworms [2,5,[51][52][53][54][55], but Abs against these helminths in our study did not cross-react with the A. caninum somatic Ag. Additionally, a few Abs in two of fifteen cases of gnathostomiasis showed false positives. However, animal hookworms mostly migrate in the dermis of human skin, but Gnathostoma worms can move in the skin and into deeper tissues and organs.
After several years of frozen storage, the worms were not degraded and demonstrated sharp and defined protein bands on Coomassie brilliant blue-stained gel. Ice crystal damage did not appear to affect the frozen worms. However, the adult worm Ag in our study did not react well with the pooled positive and negative control sera on the ELISA checkerboard titration for individual IgG subclasses and IgE, which is possibly because the serum samples from the controls contained low levels of individual immunoglobulins. The duration of storage of serum samples at −80 • C did not affect antibody reactivity because two high OD values, 1.111 and 1.078, were from two young boys whose sera were kept for 34 and 3 years, respectively. The false negative case was an adult patient, and his serum sample was kept for 11 years. To support low IgE in hwCLM, IgE is produced after all other isotypes in the blood [56], with the lowest concentration among serum Igs of a sequential Ig response [57]. Some reports of hookworm-related CLM presented laboratory findings with normal serum IgE levels [14,39], and the Ags from canine hookworm larvae were mostly confined to the dermis of humans (an abnormal host), thereby revealing a limited IgE response to infection [12]. Hence, the low or absent IgE levels in the positive controls for hwCLM in our study may have led to their being indistinguishable from the negative controls in the checkerboard titration. The increasing IgE levels may have depended on the magnitude of infection such as in schistosomiasis, filarial tropical eosinophilia [58], and N. americanus self-infection, which requires three to four infections to measure increases in IgE [59]. Helminth species other than A. caninum can cause increased IgE levels because their locations are not limited to the dermis [12].
Finding a different result when comparing antigens from A. caninum worms, Loukas and colleagues [13] used a different Ag from A. caninum worms, excretory-secretory (ES) Ag, against four helminthiases, which did not react with Abs from human hookworm infections but reacted with Abs to strongyloidiasis, toxocariasis, and schistosomiasis. In our study, crude A. caninum somatic Ag did not cross-react with Abs against human-associated hookworm diseases, including strongyloidiasis, toxocariasis, and schistosomiasis, which differs from the study using ES-Ag. In Thailand, N. americanus is mainly reported via molecular confirmation, and minor species include A. duodenale and A. ceylanicum in humans at the studied sites [60,61]. Human-associated hookworms (15 sera used) in our study were confirmed only via detection of eggs and larvae; however, Abs from those hookworm infections may be induced by the aforementioned three species. Antibodies of those fifteen cases did not cross-react with the A. caninum Ag and thereby yielded true negative results ( Figure 6). When comparing the crude somatic and crude ES Ags of A. caninum, Loukas and colleagues used an IgG-ELISA to study Abs from human eosinophilic enteritis with A. caninum infections and found a higher detection rate using ES-Ag than using a crude somatic Ag, although cross-reactions still occurred with parasitological soils [40]. Improvement of the diagnostic method using immunoblot with Ac68 of the ES product from A. caninum Ag detected eight ancylostomiasis caninum cases, five of which were positive for IgG and IgG 1-4 , one was positive for IgG and negative for IgG 1-4 , one was negative for IgG and positive for IgG 1-4 , and one was negative for all IgGs. All IgGs showed 75% sensitivity in immunoblot results, and IgG 4 showed the best specificity (100%), which was determined from only three helminthiases (strongyloidiasis, trichinellosis, and schistosomiasis), but all human-associated hookworm infections (A. duodenale, N. americanus and unknown species) reacted with the Ac68 Ag. These authors found that Ac68 Ag was hookworm-specific [41].
Occasionally, when a serodiagnosis from a total IgG-ELISA cannot be interpreted because of high cross-reactivity, IgG 1-4 can be analyzed to obtain a better diagnosis. However, previous studies recommended using an individual IgG subclass or co-interpretation of IgG subclasses for serodiagnosis, such as for gnathostomiasis (i.e., co-interpretation of IgG 1screening tests and IgG 2 -confirmed diagnosis via ELISA using the CSAg of Gnathostoma larvae) [62]. Laummaunwai and colleagues [63] reported that the highest sensitivity for immunoblotting for diagnosing gnathostomiasis using crude Gnathostoma larvae Ag was obtained with total IgG rather than IgG 1-4 , although specificity with IgG 4 was slightly better than with total IgG (93.9% vs. 91.6%, respectively). HwCLM may be difficult to diagnose because of misdiagnoses and prescribed treatments from physicians and laboratory practitioners [64,65]. A previous study reported a misdiagnosis and improper treatment rate of 58% [38], including the observation of reported information over a long duration of increasing Igs because patients often visit a physician at the early onset of clinical symptoms. Thus, our results suggest that IgG 1-4 and IgE from hwCLM sera cannot be used as mediators between diagnoses using ELISA and immunoblotting. However, the crude somatic Ag used in the present study had good sensitivity and excellent specificity once the optimal conditions for the IgG-ELISA were determined. Use of a crude Ag can demonstrate cross-reactivity with Abs of heterologous diseases; thus, proper ELISA conditions can improve the results.

Conclusions
Serodifferentiation with crude A. caninum somatic Ag for the proposed IgG-ELISA can potentially be used to diagnose hwCLM from A. caninum hookworms in humans. However, IgE and IgG 1-4 could not be differentiated via the ELISA for the positive and negative controls following checkerboard titration; also in addition, the antigen-antibody banding pattern via immunoblot used pooled positive controls. Further studies are warranted to evaluate the proposed ELISA using more hwCLM cases and incorporating a purification technique during Ag preparation.