Angie-LAMP for diagnosis of human eosinophilic meningitis using dog as proxy: A LAMP assay for Angiostrongylus cantonensis DNA in cerebrospinal fluid

Background Angiostrongylus cantonensis (rat lungworm) is recognised as the leading cause of human eosinophilic meningitis, a serious condition observed when nematode larvae migrate through the CNS. Canine Neural Angiostrongyliasis (CNA) is the analogous disease in dogs. Both humans and dogs are accidental hosts, and a rapid diagnosis is warranted. A highly sensitive PCR based assay is available but often not readily accessible in many jurisdictions. An alternative DNA amplification assay that would further improve accessibility is needed. This study aimed to assess the diagnostic utility of a newly designed LAMP assay to detect DNA of globally distributed and invasive A. cantonensis and Angiostrongylus mackerrasae, the other neurotropic Angiostrongylus species, which is native to Australia. Methodology/Principal findings Cerebrospinal fluid (CSF) from dogs with a presumptive diagnosis of A. cantonensis infection (2020–2022) were received for confirmatory laboratory testing and processed for DNA isolation and ultrasensitive Angiostrongylus qPCR targeting AcanR3390. A newly designed LAMP assay targeting the same gene target was directly compared to the reference ultrasensitive qPCR in a diagnostic laboratory setting to determine the presence of A. cantonensis DNA to diagnose CNA. The LAMP assay (Angie-LAMP) allowed the sensitive detection of A. cantonensis DNA from archived DNA specimens (Kappa = 0.81, 95%CI 0.69–0.92; n = 93) and rapid single-step lysis of archived CSF samples (Kappa = 0.77, 95%CI 0.59–0.94; n = 52). Only A. cantonensis DNA was detected in canine CSF samples, and co-infection with A. mackerrasae using amplicon deep sequencing (ITS-2 rDNA) was not demonstrated. Both SYD.1 and AC13 haplotypes were detected using sequencing of partial cox1. Conclusions/Significance The Angie-LAMP assay is a useful molecular tool for detecting Angiostrongylus DNA in canine CSF and performs comparably to a laboratory Angiostrongylus qPCR. Adaptation of single-step sample lysis improved potential applicability for diagnosis of angiostrongyliasis in a clinical setting for dogs and by extension, to humans.


Introduction
The 'rat lungworm' (Angiostrongylus cantonensis) is a nematode with a unique affinity for the central nervous system (CNS). It is found throughout the Asia-Pacific region, including Hawaii, and more recently in mainland United States and Europe [1][2][3][4][5][6][7][8]. Currently, A. cantonensis is recognised globally as the leading cause of eosinophilic meningitis in humans, which is a serious condition observed when larvae migrate through the CNS [5,9,10]. It has been recorded as an aetiological agent of meningitis in humans in at least 30 countries, including Australia, making it a parasite of increasing importance [11]. Humans are an accidental host when they deliberately or inadvertently ingest infected intermediate or paratenic hosts [9]. The rat lungworm traditionally cycles between the intermediate host-molluscs, and the definitive host, any rat species, where it completes the life cycle and reaches the pulmonary arteries via an obligatory migration through the rat's spinal cord and brain [12]. The infective larvae are generalists and will rapidly invade the CNS of many animals including primates, marsupials, bats, horses, dogs and birds [13][14][15][16]. The overt host response in the CNS leads to the associated neurological signs and symptoms, as well as severe inflammation of the meninges and brain [17]. Canine Neural Angiostrongyliasis (CNA) is a condition in dogs analogous to the human disease eosinophilic meningitis [18]. CNA manifests as severe neurological dysfunction, with signs such as hindlimb and tail paresis to paralysis, muscle wasting often associated with urinary incontinence. Critically, it is characteristically associated with marked hyperaesthesia [19]. Naturally occurring cases of CNA are reported throughout most of coastal Queensland and eastern New South Wales in Australia [18][19][20]. There appears to be a "peak season" of CNA, with most cases occurring after unusually wet weather between April and June in QLD [18]. A similar pattern occurs in NSW where a spike in cases occurs around April/May [19]. Infection disproportionately affects younger dogs, speculated to be due to their higher propensity for eating slugs and snails out of curiosity, and possibly their lack of acquired immunity to nematodes [19,20]. Diagnostics remain challenging in both dogs and humans due to the rapid onset of the clinical signs and symptoms [20][21][22][23][24]. Laboratory diagnostics rely on the application of DNA amplification tests, with the recent introduction of ultrasensitive AcanR3390 real-time qPCR assay having been demonstrated to be superior to previously used PCR and qPCR-based assays targeting rDNA in both humans and dogs [23,25]. In an effort to further improve the accessibility of this assay, it was adopted for Recombinase Polymerase Assay (RPA) that has the potential to become a point-of-care molecular assay [25,26]. As a compromise to qPCR and RPA, loop-mediated isothermal DNA amplification (LAMP) using dsDNA binding fluorescent dye that provides fast amplification, approximate quantification and product identity control, thus ensuring both speed and reliability [27,28]. Such assays can be performed either in qPCR thermocycler or a portable device with appropriate fluorescence detection system [29].

PLOS NEGLECTED TROPICAL DISEASES
This study aimed to assess the diagnostic utility of a newly designed LAMP assay targeting AcanR3390 in a diagnostic laboratory setting in a direct comparison with the reference ultrasensitive qPCR for determining the presence of A. cantonensis infection. We utilised archived (2020-2022) cerebrospinal fluid (CSF) and DNA isolated from canine CSF from patients living along the east coast of Australia. The LAMP developed in this work (Angie-LAMP) was shown to be both sensitive and specific at detecting DNA from A. cantonensis; and also the only other neurotropic Angiostrongylus species that is native to Australia, A. mackerrasae. The potential diagnostic applicability of the Angie-LAMP extends to other species and material, including human CSF from suspect neural angiostrongyliasis cases.

Ethics statement
The use of residual clinical samples was in accordance with the University of Sydney' Animal Ethics Committee Protocol (2023/2278).

Diagnostic samples
Material was provided for the purpose of diagnostics used in decision-making by the veterinary practitioner. Results and material used for LAMP were de-identified. Material (CSF) was collected by registered veterinary practitioners in accordance with the Veterinary Practice and Animal Welfare Acts. Samples were submitted for parasitological diagnostics (Sydney School of Veterinary Science, The University of Sydney). Cerebrospinal fluid (CSF) from 111 domestic dogs were provided by Vetnostics (Laverty Pathology-North Ryde Laboratory, Sydney) and Veterinary Pathology Diagnostics Services (University of Sydney). In total, 108 samples were submitted for molecular confirmatory of A. cantonensis infections, with signalment suggesting an infectious cause and with the clinical suspicion for CNA (S1 Table). The remaining three CSF samples did not have any clinical suspicion of CNA (Ag70, Ag71, Ag107) (S1 Table).

DNA isolation from cerebrospinal fluid (CSF)
Genomic DNA was purified from 100μL of CSF samples using the Monarch Genomic DNA Purification Kit (New England Biolabs, Australia) according to the manufacturer's instructions, using the Mammalian Whole Blood protocol. Up to 100μL of CSF was lysed, with insufficient samples being topped up to the required volume with phosphate-buffered saline (PBS). Each batch of samples processed included a 'blank' extraction control. Samples were eluted with 75μL of pre-heated gDNA Elution Buffer, and the final centrifugation step was repeated for maximum gDNA yield. DNA and remaining unused CSF were stored at -20˚C prior to further processing.

Ultrasensitive AcanR3390 Angiostrongylus qPCR
A probe-based qPCR assay targeting a repeat sequence on contig 3990 (AcanR3390) of A. cantonensis described by Sears et al. [25] was locally optimised and used as a veterinary diagnostic assay at the Laboratory of Veterinary Parasitology, University of Sydney. Duplicate reactions were run at a final volume of 20μL, including 2μL of gDNA and 10μL Luna Universal Probe qPCR Master Mix (New England Biolabs, Australia). Final concentrations of the forward (S0947) and reverse (S0948) primers were 0.4μM, and FAM probe (S0949) at a final concentration of 0.1μM. Primers and probes were sourced from Integrated DNA Technologies, Inc. (Australia). All qPCR runs included a 'blank' extraction control, ddH 2 O no-template control, and positive A. cantonensis control for the detection of contamination and assurance of qPCR run success. qPCRs were performed in a CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.) with the following cycling conditions: 95˚C for 3 minutes; and 40 cycles of 95˚C for 5 seconds, and 60˚C for 15 seconds. Amplification curves and cycle threshold (C t )-values were recorded using CFX Maestro Software 2.3 (Bio-Rad Laboratories, Inc.).
A region of the mammalian glyceraldehyde 3-phosphate dehydrogenase (G3PDH) gene was targeted by a separate probe-based qPCR assay, which was run in parallel with the AcanR3390 assay to ensure that DNA was successfully isolated. An additional positive control containing DNA isolated from whole dog blood was included in these runs. Reagent volumes and concentrations were the same as outlined above, but instead using G3PDH forward and reverse primers (S1072 and S1073, respectively), and Cy5 probe (S1074), as described by Peters et al. [30].
Currently used interpretation workflow for ultrasensitive AcanR3390 Angiostrongylus qPCR utilised by the Sydney School of Veterinary Science is as follows. C t values resulting from duplicate qPCR of the ultrasensitive Angiostrongylus DNA, alongside C t values from qPCR for the presence of mammalian DNA are both recorded. Sample(s) are run alongside the isolated 'blank' extraction control sample. (1) Mammalian DNA qPCR is positive (C t < 40) for the sample DNA, while the 'blank' is negative for both mammalian DNA and Angiostrongylus DNA (C t � 40). All samples in this set passed this step. (2) If duplicate Angiostrongylus DNA qPCR returned C t � 35 in both qPCRs, the sample is considered a 'strong positive' for Angiostrongylus DNA. (3) If one of the qPCRs returns C t � 35 and the other C t > 35 (including C t � 40), the sample is still considered 'positive' for Angiostrongylus DNA. (4) If duplicate Angiostrongylus DNA qPCR returns C t � 38 but C t > 35 for both qPCRs, the sample is considered 'weak positive' for Angiostrongylus DNA. (5) The sample is considered 'equivocal' for Angiostrongylus DNA if one C t > 38 and the other does not amplify (return C t � 40), or one is C t > 35 but the other does not amplify, or both C t > 38. (6) If both qPCRs do not amplify (i.e. return C t � 40), the sample is considered 'negative' for Angiostrongylus DNA.

Molecular determination of Angiostrongylus species in CSF
We adopted an ITS-2 rDNA qPCR diagnostic assay for amplicon metabarcoding and Next-Generation Sequencing (NGS) to allow us to determine Angiostrongylus species. The qPCR primers adapted from Fang et al. [31] target a~140bp region of ITS-2 rDNA; [S1060] AngioITS2_FOR (TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG CCA GTT TTG GTG AAG AAT AA) and [S1061] AngioITS2_REV (GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GAC ACG ACG GTA ACA ATG ACA), nucleotides in italic represent the Nextera (Illumina) adapter. This region includes a single nucleotide polymorphism (SNP) (G-to-A) at the 83 rd position which discriminates A. cantonensis from A. mackerrasae [32]. A selection of Angiostrongylus qPCR-positive individual dog DNA CSF samples (n = 35), and DNA from voucher A. cantonensis (n = 6) and A. mackerrasae (n = 2) specimens from Valentyne et al. [33] were subjected to barcoding amplification and processing for NGS at the Ramaciotti Centre for Genomics, University of New South Wales, Australia using Illumina MiSeq v2 250 PE. Obtained FastQ files were processed using a local DADA2 pipeline and curated database of reference sequences via CLC Main Workbench v22 (Qiagen, CLC bio). The output (in the form of ASV counts per sample) was manually analysed in Excel to identify samples with >200 reads matching a reference Angiostrongylus sequence. FastQ files were deposited in Sequence Read Archive (SRA) under the BioProject: PRJNA912228.
To determine the cox1 haplotype of A. cantonensis, a partial gene was amplified and sequenced according to Mallaiyaraj Mahalingam et al. [32]. Briefly, primers AngiCOI_forward (S0963) and AC1R (S0966) were used in a qPCR reaction with Luna Universal qPCR Mastermix (New England Biolabs, Australia). The qPCR reactions were run on the CFX96 Touch Real-Time PCR Detection System (BioRad, Australia) and analysed using the corresponding CFX Maestro 2.3 software (BioRad, Australia). Cycling conditions were as follows: 95˚C for 60 s, followed by 40 cycles at 95˚C for 15 s and 55˚C for 30 s. Results were considered positive if visual inspection of the melt curve profile corresponded to that of the A. cantonensis control and reactions were submitted for bidirectional sequencing at Macrogen Inc. (Seoul, Korea). The DNA chromatographs were visually inspected in CLC Main Workbench v22 (Qiagen, CLC bio) and matched against the reference alignment of all known cox1 haplotype sequences from GenBank. Newly obtained cox1 were deposited in GenBank (OQ029497-OQ029501). Additional associated files are available from LabArchives (https://dx.doi.org/10.25833/z7zq-9e13).

Loop-mediated isothermal DNA amplification for detection of Angiostrongylus DNA-Angie-LAMP assay
Primers for Angie-LAMP were designed using the GLAPD online tool [34] using the published Acan3990 tandem repeats sequence as a template [25]. The selected primers and optimal concentrations used are presented in Table 1. All Angie-LAMP reactions (25 μL) were run using Isothermal Master Mix (ISO-004, Optigene, UK) according to manufacturer's instructions in a portable machine Genie II and/or Genie III (Optigene, UK) equipped for reading, recording and visualisation of the FAM channel fluorescence.
To determine the analytical specificity, the Angie-LAMP assay was tested against DNA extracted from various lineages of A. cantonensis (Fatu Hiva lineage FH.1, lineage in Sydney SYD.1, and Thai lineage in Sydney AC13), and other Angiostrongylus species (A. mackerrasae, A. chabaudi, A. costaricensis, A. daskalovi and A. vasorum). In addition, relevant DNA from hosts (human, dog) and DNA from migratory parasites that can be present in CSF of dogs (Toxocara, Ophidiascaris, Ancylostoma, Uncinaria) were also tested. The limit of detection was calculated by testing 10-fold serial dilutions of DNA from A. cantonensis L1 larvae.
Two Angiostrongylus DNA positive CSF samples (Ag66, Ag77) were initially used to evaluate a 'point-of care' applicability of the Angie-LAMP. We tested (i) neat CSF sample, (ii) CSF sample heated at 100˚C for 10 minutes, and (iii) CSF sample treated with PrepMan Ultra (Thermo-Fisher Scientific) and heated at 100˚C for 10 minutes. The selected samples were analysed at different dilutions (undiluted to 1:1,000) and tested immediately for Angie-LAMP.
Archived DNA isolated from dog CSF samples (n = 111) and available neat CSF samples from a subset of the DNA samples for which retained CSF was available (n = 57) were used to compare the standard laboratory diagnostic procedure (DNA isolated using extraction kit, followed by Angiostrongylus qPCR). Neat CSF samples (10 μL) were mixed with equal volume PrepMan Ultra (Thermo-Fisher Scientific) reagent in 0.2 mL PCR 8-well strips and heated to 100˚C for 10 min (T100; BioRad, Australia) and used immediately. Angie-LAMP reactions included either 5 μL of DNA isolated from CSF or 5 μL of the CSF treated with PrepMan Ultra reagent and run in Genie III (OptiGene, UK) with 30 minutes amplification at 65˚C and subsequent anneal temperature analysis (98˚C to 80˚C). Negative (ddH 2 O) and positive (A. cantonensis SYD.1) DNA controls were included. Two values obtained for each tested sample were the time of amplification detection in minutes (in principle comparable to C t in qPCR) and annealing temperature (when primers attach to the DNA product and so cause measurable change in fluorescence of the dsDNA intercalating dye). The processing of raw data in Genie Explorer software (OptiGene, UK) included control of fluorescence change during amplification and annealing analysis (shape, angle of curves). Time of detected amplification calculation had to be adjusted in some runs to eliminate erroneous readings caused by background changes in fluorescence (smoothing by 2-3 points, detection threshold 0.010-0.008 ratio). Anneal derivative analysis (smoothing 2-5 points, peak detection full, threshold varied from 4,000-5,000 derivative). If amplification read more values due to atypical shape of the amplification derivative, the correct value was obtained manually in Genie Explorer software, reading the peak value. A sample was considered 'positive' if amplification was detected before 25 th minute and quantified anneal temperature was within 0.5˚C of the positive controls. If only a single reading of either amplification time or anneal temperature occurred, or late amplification occurred (>25 minutes), we interpreted the result as 'equivocal'. No amplification within 30 minutes or amplification of a product with anneal temperature outside the set range was considered as a 'negative' LAMP result.

Data analysis
Data were imported and analysed in GraphPad Prism 9.4.1 (GraphPad Software, US). Data were visualised as scatterplots and histograms, analysed using Bland-Altman plot and linear regression to find the best line that predicts Y from X and interpreted using contingency tables.
Means were compared using t-tests. Data were tested for normality (D'Agostino & Pearson test), 95% confidence intervals calculated, and significance was set at 0.05. Kappa was interpreted using the scale of Landis and Koch [35] and Bland Altman plot according to Giavarina [36].
To further consider the diagnostic utility of Angie-LAMP, we attempted to bypass DNA extraction as a separate step and utilised a single step protocol for cell lysis with PrepMan Ultra reagent. We assessed 57 canine CSF samples previously used for DNA isolation and tested for presence of Angiostrongylus DNA (32 strong positive, 3 positive, 21 negative, 1 equivocal) (Fig 1E and Table 2). The agreement of the Angio-LAMP on CSF samples treated with PrepMan Ultra with Angiostrongylus qPCR showed 'substantial agreement' (Kappa = 0.77; 95%CI 0.59 to 0.94) using the scale of Landis and Koch [35]. Bland-Altman bias was 15 (SD = 4.4) with 95% limit of agreement 6.7 to 24 (Fig 1E inset). Using the qPCR as the reference test, the Angie-LAMP assay using CSF treated with PrepMan Ultra reagent had a sensitivity of 85.7% (95%CI 69.7% to 95.2%) and specificity of 94.7% (95%CI 74.0% to 99.9%) (Fig 1C). Linear regression demonstrated that the slope between qPCR and Angie-LAMP on CSF treated with PrepMan Ultra was not significantly 'non-zero' (F = 2.59, DFn = 1, DFd = 18, P-value = 0.13; R 2 = 13%) (Fig 1F).
The use of PrepMan Ultra reagent was initially evaluated using serial dilutions of two CSF samples with Angiostrongylus DNA (Angiostrongylus qPCR: Ag66, avg. C t value = 27.0; Ag77, avg. C t value = 27.0). We tested (i) neat CSF, (ii) boiled CSF and (iii) boiled CSF mixed with PrepMan Ultra reagent in the Angie-LAMP. Neat, undiluted samples returned positive, but

Confirmed presence of Angiostrongylus cantonensis in CSF of dogs
Angiostrongylus species identification was successful in 40% (14/35) of samples that were processed for ITS-2 rDNA deep sequence analysis. Of the 64 Angiostrongylus AcanR3990-positive individual dog CSF samples, 35 were processed using the ITS-2 rDNA amplification. Overall, 23/35 (66%) produced ITS-2 rDNA amplicon with primers adopted for deep sequencing. All 14 deep-sequenced dog CSF samples produced raw sequence reads that matched Angiostrongylus ITS-2 rDNA (AB700700) by BLAST analysis. All (14/14) were further determined to contain only A. cantonensis DNA (i.e. no G-to-A SNP at the 83 rd position) when aligned to reference A. mackerrasae (NC_046586) and A. cantonensis (AB700700) partial cox1 sequences. The mean C t value of the successfully identified DNA was~8 cycles lower than those that were unsuccessful (paired t-test, P value < 0.001). Amplification and deep sequencing of voucher DNA of A. cantonensis and A. mackerrasae confirmed the G-to-A SNP within ITS-2 rDNA. For 5/33 (15%) Angiostrongylus qPCR positive samples we were able to PCR amplify and DNA sequence partial cox1. Four perfectly matched the SYD.1 haplotype (Ag8, Ag11, Ag13, Ag72) and one perfectly matched AC13 haplotype (Ag6) mtDNA haplotype of A. cantonensis.

Discussion
The recently developed ultrasensitive qPCR targeting the AcanR3390 repetitive element can detect A. cantonensis in samples with as little as 0.00001 of a single third stage larva [25]. Such performance improved the ability to diagnose rat lungworm disease and outperforms previously used assays in clinically confirmed case of CNA [23]. While the assay is highly sensitive, it requires collection of CSF and DNA isolation in a specialised laboratory setting before the assay can be performed. Our approach adopted LAMP technology to target the same genetic target, because LAMP can be used in field-or point of care conditions and works using various samples, i.e. CSF. The performance of LAMP in terms of sensitivity is comparable with qPCR

PLOS NEGLECTED TROPICAL DISEASES
(N-1 χ 2 , P-value = 0.15). Analogous to previous studies targeting AcanR3390, Angie-LAMP is analytically highly specific for the globally emerging A. cantonensis and so does not cross react with any other viruses, bacteria, protozoa and helminths, except A. mackerrasae [25,26]. A closely related sibling species native to Australia, A. mackerrasae, shares a life cycle and obligatory neurotropic sojourn to that of A. cantonensis; however, A. cantonensis is universally assumed to be the causative agent of human eosinophilic meningitis and CNA [21,37]. Although, it is speculated that A. mackerrasae could cause CNA, especially after a patent infection was discovered in a black flying fox (Pteropus alecto) in 2013 [38,39]. In this study, using a deep sequencing approach and recent canine CSF samples (2020-2022), we demonstrate an absence of A. mackerrasae DNA which supports the dogma that A. cantonensis is the sole culprit in CNA cases [32]. Unlike the previous study by Mallaiyaraj Mahalingam et al. [32], we adopted the ITS-2 assay for deep amplicon sequencing to detect potential co-infections between the two species. The absence of A. mackerrasae DNA in dogs with CNA in the current study may result from A. mackerrasae's innate preference for native rat species, such as Rattus fuscipes (bush rat) and R. lutreolus (Australian swamp rat), due to their co-evolution [37]. Bush and swamp rats are known to avoid dispersal through developed areas where invasive Rattus spp. thrive, which may reduce cross-over to domestic dogs and urban rats [40][41][42]. Alternatively, A. mackerrasae is simply a specialist not capable of accidently infecting dogs.
Globally, many cox1 haplotypes of A. cantonensis have been characterised [33,43]. The analysis of the partial cox1 region from dog CSF in our study confirms that Australian dogs currently host at least two known haplotypes; the established SYD.1, and the more recently recognised AC13 from Thailand [32,44]. The presence of additional haplotypes locally is not yet known, but is speculated due to the ongoing invasion of A. cantonensis strains into previously non-endemic countries [45,46]. Experimental infection of rats with various Taiwanese (H and P strains) and Brazilian (ac8 and ac9) A. cantonensis haplotypes demonstrated differences in infectivity and fecundity [47,48]. Further work should aim to determine if the two now endemic Australian haplotypes have different virulence properties, geographical ranges, host preferences or propensity to infect non-permissive hosts, such as dogs, tawny frog mouths, bats and humans.
The newly designed LAMP assay targeting AcanR3390 demonstrated substantial agreement without the need to extract purified DNA, and thus possesses good diagnostic utility under laboratory conditions. Both the correlation and the Bland-Altman plot suggest poor agreement between the quantitative readings from qPCR (C t values) and Angie LAMP (minutes) tests, nevertheless the qualitative observations provide robust Kappa agreement outcome that is clinically relevant. One, however, has to consider if LAMP is a useful alternative diagnostic modality for clinical scenarios in both veterinary and medical settings. Infections with A. cantonensis are medical as well as veterinary emergencies that requires rapid intervention to alleviate the overt host response in the meninges as well as administration of antiparasiticides [24,49]. The signalment and clinical signs associated with the syndrome often lead to a presumptive diagnosis and initiation of treatment regardless of early laboratory data, as CSF collection can be financially prohibitive or not feasible due to the locations where these patients are situated. In situations where CSF is collected, the typical finding is eosinophilic pleocytosis, or an increased percentage of eosinophils [22,50]. Diagnostic PCR, including the ultrasensitive AcanR3390, are not 'rule out' tests, because in clinically confirmed cases of CNA summarised by Lee et al. [23] only 81% (48/59) of affected dogs tested positive for presence of Angiostrongylus DNA, the remainder being diagnosed on the basis of detecting specific antibodies directed against A. cantonensis in CSF using ELISA. In the veterinary setting, DNA tests often take 1-2 days to be conducted and reported and no private veterinary hospitals have molecular diagnostic capacity on site. Availability of the portable instrument utilised in this study and the ease of preparation of the samples without the need for DNA extraction as a separate procedure can enable on site diagnostics for moderate to large referral hospitals. The turnaround time from collecting the CSF to receiving the result can be achieved in under 1 hour (CSF preparation 10 minutes, LAMP 30 minutes, 10 minutes sample and reagents manipulation). In fact, our results show that utilising LAMP will only slightly increase the false negative rate compared to reference laboratory results.

Conclusion
We have developed the new Angie-LAMP that performs favourably on CSF of dogs when compared to a referral diagnostics laboratory qPCR method following DNA isolation. Replacing the DNA isolation using single-step preparation of CSF is feasible in current clinical settings and thus Angie-LAMP can enable 'point-of-care' diagnostics using CSF. One of the main reasons for this study was to consider a 'One Health' approach to diagnostics, because the serious emergency consequences of this parasite for both humans and dogs. Our archival material of CSF samples from 2020-2022, where CNA was presumed, potentially represents one of the largest sample sizes of CSF specimens to evaluate new test such as Angie-LAMP. While CNA is frequently recorded in eastern Australia [23], the equivalent human condition eosinophilic meningitis is rare and sporadic, with far fewer detected cases and more difficult access to archived CSF collections. The analogous pathogenesis of the disease in dogs and humans enables us to consider the performance of Angie-LAMP on canine CSF for human CSF diagnostics. In both species, the CSF has abnormally high number of cells that are mostly eosinophils. The assay does not amplify human or canine DNA, nor does it amplify other potential neuropathogens including other parasites that could be present in CSF. Our results show that Angie-LAMP is a robust tool suitable for use in either canine or, by extension, to human diagnostics. Furthermore, Angie-LAMP is ideally suited for testing in the field for screening water supplies, food (crustaceans and molluscs as food source), snail infested waters (eDNA) or rat faecal pellets for presence of Angiostrongylus. As far as veterinary and human diagnostics, the retrieval of CSF via cisternal or spinal tap is an invasive and potentially traumatic procedure. Evaluation of the validated Angie-LAMP assay on other more accessible diagnostic material (such as well-timed paired whole blood, sera and urine specimens) is urgently needed.
Supporting information S1