A dirofilariasis mouse model for heartworm preclinical research

Use of experimental cats and dogs in veterinary heartworm preclinical drug research is increasing. As a potential alternative primary in vivo heartworm preventative drug screen, we assessed lymphopenic mice with ablation of the interleukin-2/7 common gamma chain (γc) as susceptible to the larval development phase of D. immitis. Non-obese diabetic (NOD) Severe Combined ImmunoDeficient (SCID)γc-/-(NSG / NXG) mice consistently yielded viable D. immitis larvae at 2-4 weeks post-infection across multiple experiments, different batches of infectious larvae inoculates, different isolates of D. immitis and at independent laboratories. Mice did not display any overt clinical signs associated with infection up to 4 weeks. Developing larvae were found in subcutaneous and muscle fascia tissues, the natural site of this stage of heartworm in dogs. Larvae retrieved from NSG / NXG mice were mid-L4 stage of development. Compared with 14-day in vitro propagated larvae, in vivo derived L4 were significantly larger and contained expanded Wolbachia endobacteria titres, determined by QPCR and Fluorescent in situ Hybridisation (FISH). We established an ex vivo 6-day L4 paralytic screening system against nematodicidal agents (moxidectin, levamisole) which highlighted discrepancies in relative drug sensitivities in comparison with in vitro reared L4 D. immitis. We demonstrated effective depletion of Wolbachia by 70-90% in D. immitis L4 following 2-7 day oral in vivo exposures of NSG / NXG infected mice with doxycycline or the rapid-acting investigational anti-Wolbachia drug, AWZ1066S. We validated the NSG / NXG mouse model as a filaricide drug screen by in vivo treatments with single injections of moxidectin, which mediated 60-88% reduction in L4 larvae at 14-28 days. Future adoption of the mouse model as a first-line efficacy screen will benefit end-user laboratories conducting research and development of novel heartworm preventatives via increased access, rapid turnaround and reduced costs whilst simultaneously decreasing need for experimental cat or dog use.


INTRODUCTION
Dirofilaria immitis is a major veterinary filarial parasitic nematode causing chronic heartworm disease (HWD) in dogs.Dirofilariasis is spread primarily by mosquito species of the Culicidae family, including the invasive tiger mosquito, Aedes albopictus [1].HWD develops following the establishment of adult nematodes in the right chambers of the heart associated vessels following larval migrations in subcutaneous and muscle tissues.Adult infections can persist in the heart for greater than five years [2].Pathology is chronic-progressive, associated with enlargement and hyper-proliferation of endocardium and physical blockage of adult worms in the pulmonary artery contributing to vessel narrowing, hypertension and ultimately heart failure [3].Dirofilaria immitis causes a more acute immunopathology in cats where arrival of immature worms often triggers an overt inflammatory reaction in the lungs leading to heartworm-associated respiratory disease [2].Both cats and dogs are at risk of acute, fatal thromboembolisms when dead adult worms lodge in pulmonary vasculature [3].Dirofilaria spp.can also cause abbreviated zoonotic infections in humans, whereby arrested development of immature adults can cause subcutaneous nodules and lung parenchyma disease [4].Dirofilaria repens is the most widely reported dirofilarial zoonosis, noted to be increasing across Europe, Asia and Sri Lanka; although, D. immitis, D. striata, D. tenuis, D ursi and D. spectans also infect humans [5].In 2012, 48,000 dogs tested positive for heartworm in the USA and in 2016 over one million pets were estimated to carry disease [6].Incidence of HWD in the US is increasing both within endemic areas and into erstwhile HW-free, westerly and northerly regions, including Canada [3].A similar epidemiological pattern of increased dirofilariae incidence in The Mediterranean and spread into northern latitudes of Central and Western Europe has also been documented [7,8].Heartworm disease is controlled mainly by preventative chemotherapy and curative treatment of diagnosed cases.Chemo-prophylaxis with macrocyclic lactones (ML), namely: ivermectin, milbemycin oxime, moxidectin and selamectin, are effective at targeting L3-L4 larvae during subcutaneous tissue development and before immature adults reach the pulmonary artery to .CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is The copyright holder for this preprint this version posted April 3, 2023.; https://doi.org/10.1101/2023.04.03.535321 doi: bioRxiv preprint establish pathological adult infection [9,10].After more than 40 years of use in veterinary medicine, ML drug resistance is prevalent in veterinary nematode parasites, with several D. immitis isolates formally determined as resistant to ML ,whereby timed experimental infections and accurate prophylactic dosing have failed to prevent the development of fecund adult HW infections [10].The only regulatory approved cure available for HWD is the injectable, melarsomine dihydrochloride.However, issues of this therapy include: lengthy treatment regimens requiring in-clinic administrations, potential steroid pre-treatment, exercise restriction, and the risk of severe adverse events.Melarsomine is unsafe for use in cats, with no alternative curative therapies currently approved.Alternative curative therapies include the use of moxidectin and doxycycline ('moxi-doxy') [11] with the latter antibiotic validated as a curative drug targeting the filarial endosymbiont, Wolbachia, demonstratable in human filariasis clinical trials [12].However, due to concerns of veterinary applications of doxycycline, in companion animals, such as long treatment timeframes, dysbiosis side-effects, and antibiotic stewardship of a human essential medicine, development of short-course, narrow-spectrum anti-Wolbachia heartworm therapeutics without general antibiotic properties, may offer a potential future alternative [13].ML preventatives, costing typically between $266-329 a year for a pet's treatment in the USA, represent a potential multi-billion dollar global market [14].Due to emergent spread of Dirofilaria immitis infections, the growing concerns of ML prophylactic failure in USA and the current inadequacies of curative treatments, new therapeutic strategies are being intensively investigated.The only fully validated in vivo preclinical models currently available for heartworm anti-infectives are laboratory reared cats and dogs.Lymphopenic and type-2 immunodeficient mice have been recently developed and validated as in vivo and ex vivo drug screens for the medically important filarial parasite genera; Brugia, Onchocerca and Loa [12,[15][16][17].Here, we describe that lymphopenic immunodeficient mice with ablation of the interleukin-2/7 common gamma chain (c) are susceptible to the initial

RESULTS
Selection of a susceptible mouse model of tissue-phase heartworm infection NSG and RAG2c mouse strains were initially selected to investigate permissiveness to D. immitis tissue-phase larval infections, based on our previous success in establishing long-term infections of the related filarial species, Brugia malayi, Loa loa and Onchocerca ochengi, utilising lymphopenic immunodeficient mice [15,17].In these models, the additional knockout of the interleukin-2/7 common gamma chain within lymphopenic mice is essential for susceptibility to Loa adult development in subcutaneous tissues [17] and bolsters both Brugia and Onchocerca adult infections within the peritoneal cavity [18].We also trialled methylprednisolone acetate (MPA) administrations to evaluate whether steroid suppression of residual innate immune responses could increase survival and yields of D. immitis larvae in vivo, as has been reported for experimental Strongyloides stercoralis infections [19].Initially, we used a Missouri (MO) isolate of D. immitis (NR-48907, provided by the NIH/NIAID Filariasis Research Reagent Resource Center, FR3, for distribution through BEI Resources).Infectious L3 were isolated 15 days after membrane feeding of D. immitis mf in dog blood to A. aegypti (Figure 1A).Following inoculations of 200 L3 under the skin, at 14 days post-infection, we successfully recovered D. immitis parasites from subcutaneous and muscle fascia tissues in all (5/5) NSG and RAG2c+MPA mice (Figure 1B).Multiple tissues were dissected to locate parasites (heart, lungs, peritoneal cavity, gastrointestinal tract, liver, spleen) but no evidence of infection was found in these ectopic locations.Infection success was lower in NSG+MPA (3/4 mice) and RAG2c (4/5) mouse groups.Yields significantly varied between groups with RAG2c+MPA mice yielding higher numbers of D. immitis developing larvae (L4) compared with either RAG2c or NSG+MPA groups (Kruskal-Wallis 1 Way ANOVA P=0.0033, Dunn's post-hoc tests P<0.05).Median recovery rates were similar between NSG and RAG2c+MPA groups (median recovery 9% vs 12%, ns).Due to the simpler infection regimen in NSG mice, the international commercial availability of the model, the potential for further humanisation, and to avoid potential welfare or drug-drug interactions arising with long-term MPA administrations, we selected this immunodeficient mouse model for extensive characterisation.We evaluated the infection success and yields of D. immitis L4 across multiple independent experiments utilising different batches of MO isolate D. immitis shipped from USA to UK as mf in dog blood and passaged to the infectious L3 stage in A. aegypti mosquitoes (Figure 1C).In six independent experiments, using a total of 21 mice, we were reproducibly able to recover D. immitis larvae at 14dpi (21/21 mice) with a 5% median yield of the initial 200 L3 infectious inoculate (range 1.5% -23.5%).In experiments where >2 mice were infected, we compared yields between batches and determined that batch-to-batch variability of the initial L3 infectious inoculate significantly influenced the yields of larvae recovered at 14dpi in NSG mice (Kruskal-Wallis 1Way ANOVA P=0.0017, Dunn's post-hoc tests P<0.01).We then measured recoveries utilising NXG mice; a similar Severe Combined Immunodeficient mouse line on the Non-Obese Diabetic strain background with additional c ablation, developed independently and recently commercialised by Janvier Laboratories [20] (Figure 1D).As with the original NSG mouse line, in seven independent experiments, with a total of 18 experimental mouse infections, we consistently recovered developing larvae at 14dpi (18/18 NXG mice) with a 6% median proportion of the initial 200 L3 infectious inoculate (range 1% -23%).Similar to NSG infections, in individual experiments where >2 NXG mice were available for analysis, batch-to-batch variability of the initial L3 infectious inoculate significantly influenced the yields of larvae recovered at 14dpi (Mann-Whitney P<0.05).We then repeated experiments with NSG mice in TRS Labs (Georgia, USA), accessing an inhouse parasite life-cycle and using a unique 'Georgia III' (GA III) isolate of D. immitis.We infected batches of five mice and evaluated yields of L4 at 14, 21 or 28 days post inoculation with 200 L3 (Figure 1E).All mice, irrespective of time point post-inoculation, yielded GA III D. immitis developing larvae.Yields were typically 6-fold higher than those derived at the LSTM laboratory at 14 days post-infection (median = 29.5%,range 13.3-80.5%,Figure 1F).Yields did not significantly deviate between 14, 21 and 28 days post-infection (Figure 1F).However, the distribution of larvae in mouse tissues had changed between 14 and 28 days-post infection, with relatively more larvae recovered in muscle tissues by 28 dpi (P<0.01,Fisher's Exact Test, Figure 1G).These experiments demonstrate the reproducible success of NSG/NXG mouse models as susceptible to D. immitis tissue-stage infection using different isolates of heartworm, in independent laboratories and when shipping larvae internationally between sites.Further, time-course data indicates that tissue-phase heartworm larvae persist without significant decline in yields within NSG mice whilst initiating their natural migratory route through subcutaneous and muscle tissues over the first 28 days of infection.NSG mouse-derived developing larvae demonstrate superior morphogenesis, Wolbachia content and reduced drug assay sensitivities compared with in vitro cultured D. immitis Mosquito-derived infectious L3 larvae are traditionally utilised in serum-supplemented 37C mammalian cultures to induce moulting and morphogenesis into fourth-stage developing larvae [21][22][23].This technique has been utilised to study D. immitis larval biology and for applied applications such as biomarker and preventative drug discovery [24][25][26].The survival of various filarial parasite life-cycle stages can be extended when utilising co-cultures with mammalian 'feeder cell' monolayers or trans-well compartments [16, [27][28][29][30][31].We therefore compared survival and motility of MO isolate D. immitis larvae between cell-free and LL-MCK2 (monkey) or MDCK (dog) kidney cell co-cultures in 10% calf-serum cultures (Figure 2A).The 50% survival time of cell-free cultures was d18 and subsequently, all larvae had died by d28 in culture (Figure 2B).Conversely, co-cultures with both LL-MCK2 and MDCK cells significantly increased survival, whereby >80% of D. immitis larvae were viable up to d28 (P<0.0001,Mantel Cox survival analysis).We noted a reduction in motility in all larval cultures after the first week in culture, which persisted to end-point, apart from MDCK co-cultures which returned to full-motility by d32 in culture (Supplementary figure 1).Selecting MDCK co-cultures as supportive of long-term larval motility and survival, we directly compared morphogenesis, growth and Wolbachia endobacterial expansions between in vitro propagated MO D. immitis larvae and MO larvae derived from NSG mouse infections at the 14-day time point (Figure

2A
).Both in vitro and in vivo derived d14 D. immitis larvae displayed the blunted and widened anterior extremities characteristic of the L4 developmental stage [32,33], compared with the tapered, narrow anterior of filariform infectious L3 (Figure 2C).However, anterior morphogenesis was partially arrested in in vitro compared with NSG mouse-derived larvae (Figure 2C).In the dog, larvae complete the L3-L4 moult rapidly, the vast majority by 3 days post-infection [34].In our cultures, approximately 50% of the day 14 L4 had completed moulting, with cuticle casts evident in the culture media.The other ~50% of in vitro cultured larvae displayed partial moulting of the third stage cuticle (Figure 2C).There were obvious microscopic degenerative features of the in vitro larvae by d14 compared with in vivo larvae, including malformed cuticle, hypodermis, buccal cavity, oesophagus and intestine (Figure 2C).Despite their high survival rate and continued motility, 14-day old in vitro propagated larvae were also significantly stunted compared with larvae derived from NSG mice (mean = 1020 vs 1880 M, 1 way ANOVA F=57.7,P<0.0001, Tukey's multiple comparisons test), and had not grown significantly in comparison to the L3 infectious stage (mean = 870 m) (Figure 2D).Wolbachia titre analysis by QPCR further highlighted disparities between in vitro and in vivo reared larvae (Figure 2E).The MO D. immitis in vivo larvae had undergone a significant, 66fold average Wolbachia expansion during the 14-day NSG mouse infection in comparison to iL3 (median = 4.2x10 4 vs 6.2x10 2 Wolbachia/larva, Kruskal Wallis 1-way ANOVA 26.4,P<0.0001 Dunn's multiple comparisons test), whereas MO larvae cultured for 14 days in vitro had failed to expand Wolbachia content (median = 8.7x10 2 Wolbachia/larva).Figure 2

.2. Bars represent mean +/-SEM (D, G) or median (E, H) values with individual larva data plotted. Significant differences were determined by Mantel Cox log-rank tests (B) One-way ANOVA with Tukey's multiple comparisons tests (D, G), or Kruskal Wallis with Dunn's multiple comparisons tests (E, H). Significant differences (P values <0.05) are indicated in bold. Data is one individual experiment except (N) which is two independent experiments.
Utilising GA III D. immitis, we further examined length and Wolbachia expansions between d14 and d28 post-infection in NSG mice (Figure 2F).GA3 L4 continued to grow in length between d14, d21 and d28 post-infection in NSG mice (Figure 2G; means = 1335, 1713 & 2211m, respectively, 1 Way ANOVA F=87.4,P<0.05 -P<0.0001,Tukey's multiple comparisons tests).Similarly, Wolbachia titres continued to expand within the NSG-derived GA III D. immitis larvae (Figure 2H) with significant differences evident between d14 and d28 (median = 1.7x10 5 vs 3.1x10 6 Wolbachia/larva, Kruskal Wallis statistic = 8.5, P<0.05 Dunn's multiple comparisons test).We corroborated quantitative PCR Wolbachia data, visualising that time-dependent Wolbachia multiplication was occurring within the hypodermal chord cell syncytia from a posterior to anterior direction in NSG mouse-derived, but not in vitro cultured, D. immitis L4 specimens, utilising fluorescent in situ hybridisation (FISH) of Wolbachia 16S rRNA and confocal microscopy (Figure 2I).We then examined the in vitro vs ex vivo paralytic susceptibilities of MO isolate D. immitis L4 following 6-day exposures to the standard preventative drug, moxidectin, using cultured L3-L4 larvae at 0-6d, 15-21d or 28-35d compared with NSG mouse L4 larvae isolated at 14dpi and exposed to drug ex vivo between 15-21d in matching culture conditions (Figure 2J).The IC 50 concentrations inhibiting motility of D. immitis were 1.7M for 0-6d L3-L4 larvae (Figure 2K).Sensitivity to moxidectin had increased in d14-d35 larvae with IC 50 ranging between 300-330nM (Figure 2L&M).In comparison, ex vivo larvae derived from NSG mice were relatively insensitive to the in vitro paralytic activity of moxidectin with IC 50 ranging between 48-66M (Figure 2N).This equated to a >28-fold decrease in moxidectin susceptibility compared with D. immitis L3-L4 cultures and >140-fold decreased sensitivity compared with longer-term L4 cultures.We further examined relative paralytic susceptibilities of D. immitis MO in vitro vs ex vivo L4 to 6-day exposures of the anthelmintic, levamisole, commencing at 15 days after iL3 culture / infection (Supplementary Figure 2).Whilst in vitro larvae were susceptible to high doses of levamisole (IC 50 13.2M), ex vivo L4 maintained full motility for 6 days in the presence of the top dose of drug (100 M).These data demonstrate the developmental superiority of D. immitis larvae derived from the subcutaneous and muscle tissues of NSG mice compared with standard in vitro cultures, reflected in a profound lowered sensitivity to direct-acting nematodicidal agents when used in ex vivo drug titration assays.D. immitis NSG mouse infections can be used to evaluate anti-Wolbachia drug efficacy Because we established rapid Wolbachia expansions occur during the L4 tissue development phase of D. immitis following infections of NSG mice, we next investigated the validity of the D. immitis NSG mouse model as an anti-Wolbachia drug screening model.We infected batches of 4-6 mice with MO isolate D. immitis iL3 and randomised mice into 7-day 50 mg/kg oral treatment with doxycycline or matching vehicle controls, commencing at infection with a further 7-day washout period to 14dpi (Figure 3A).We selected this regimen and timing of dose based on proven significant depletion of B. malayi L3-L4 Wolbachia in vivo in a SCID mouse model [35].We then randomised a further four mice into a 2-day bi-daily 200mg/kg treatment of our fast-acting anti-Wolbachia azaquinazoline clinical candidate, AWZ1066S [36] or vehicle control, to compare relative anti-Wolbachia activity.Doxycycline treatment mediated a 70% median reduction in Wolbachia titres in d14 MO D. immitis larvae when compared against vehicle control levels (0.29x10 4 vs 9.5x10 4 Wolbachia/larva, Mann-Whitney test P=0.014, Figure 3B).The short-course AWZ1066S 2-day oral treatment mediated a more profound 90% median efficacy in depletion of Wolbachia from d14 MO D. immitis larvae (0.49x10 4 vs 4.5x10 4-Wolbachia/larva, Mann-Whitney test P<0.0001, Figure 3C).The effect of Wolbachia depletion via doxycycline and AWZ1066S on larval growth was also evaluated (Figure 3D&E).We found that depletions of Wolbachia by 7-day doxycycline or 2-day AWZ1066S were associated with a 15.9% and 15.3% mean stunting effect on 14d MO D. immitis larvae, which was significant for doxycycline treatment (Student's T-Test, P=0.0182).We repeated the validation of the D. immitis NSG mouse model as an anti-Wolbachia drug screening system in an independent laboratory, utilising GA III isolate D. immitis (Figure 3A).In this dosing study, the 7-day oral regimen of doxycycline mediated a significant, 89% median depletion of Wolbachia in d14 GA3 larvae (0.25x10 5  Whitney test, P=0.0098, Figure 3F).We corroborated the clearance of Wolbachia from posterior hypodermal chord cells by FISH staining (Figure 3G).Stunting of GA III D. immitis larvae was also apparent following 7-day doxycycline exposures in NSG mice (mean reduction in length 30%, Figure 3G).Together, these data demonstrate the utility of the D. immitis tissuephase NSG mouse model to screen for efficacy of oral anti-Wolbachia regimens in vivo.

DAPI Wolbachia
D. immitis NSG mouse infections can be used to evaluate preventative drug efficacy Moxidectin is a front-line ML preventative used in various oral, topical or injectable formulations as monthly, biannual or annual heartworm prophylaxis in dogs [37].We selected a single high dose subcutaneous injection of moxidectin (2.5mg/kg) for evaluation of larvicidal efficacy in NSG/NXG mice (emulating route of delivery and dose of long-acting injectable formulations of moxidectin in dogs).Matched pairs of mice were infected with batches of 200 D. immitis MO or GA III isolate larvae and the next day were randomised into vehicle control or moxidectin treatment (Figure 4A).After 14 days post infection (13 days post treatment), we recorded 65-80% reductions in MO D. immitis L4 in NSG mice (n=3 pairs, P=0.03, paired ttest, Figure 4B).A similar range of larvicidal efficacy were evident in NXG mice 14 days after infection with MO isolate D. immitis and treatment with a single injection of moxidectin (range 46-88% n=6 pairs, P=0.008, Figure 4C).When using the GA III isolate of D. immits for NSG infections, level of moxidectin efficacy ranged between 29-73%, evaluated at 14 days postinfection (P=0.022,n=5 pairs, Figure 4D).We examined extended washout periods after moxidectin single dosing in NSG mice infected with GA III D. immitis.At 21 days post infection, the range of moxidectin efficacy was 45-94% (P=0.032,n=5 pairs, Figure 4E) whilst at 28 days post-infection, efficacy ranged between 35-91% (P=0.0078,n=5 pairs, Figure 4F).The median efficacy for all studies is shown in Figure 4G.In summary, single injection of moxidectin delivered a median efficacy between 65-80% at two weeks in NSG/NXG mice infected with MO isolate, and 60, 73 and 75% efficacy at two, three or four weeks in NSG infected with GA III isolate.All mice on drug studies displayed typical behaviour and gained weight over the 2-4 week period of infection and dosing (Supplementary figure 3).

DISCUSSION
We have determined that ablation of both the B & T lymphocyte compartment and additional cytokine signalling via the IL-2/7 common gamma chain receptor in mice allows permissiveness to D. immitis tissue phase larval development over the first 28 days of infection.A polarised type-2 adaptive immune response with associated type-2 tissue macrophage activation leads to eosinophil entrapment and degranulation as the basis of immune-mediated filarial larvicidal activity in mice [38][39][40][41].However, experimental infections with B. malayi, L. sigmodontis and O. ochengi in lymphopenic mouse strains (SCID / RAG2 -/- ) with additive c gene ablations have illustrated bolstered chronic susceptibility [18,42], whilst in L. loa subcutaneous infections, only combination of lymphopenia and c deficiency is sufficient to allow permissiveness to adult infections [43].Thus, an additional layer of innate immune resistance operates which can eliminate establishing larval filarial infections.In B. malayi infections, p46+ NK cells with an activated / memory phenotype and residual eosinophilia are implicated in the innate immune resistance to chronic infection in RAG2 -/-mice [18] whereas in Litomosoides infections, CD45-/TCRβ-/CD90.2+/Sca-1+/IL-33R+/GATA-3+type-2 innate lymphoid cells (ILC2) are required for innate immune resistance to microfilarial blood infections [44].It remains to be determined which of these multiple innate and adaptive immune processes are operating to control larval establishment of D. immitis in mice.Because we detected increases in D. immitis larval burdens in NOD.SCID-vs BALB/c RAG2 -/--c -/-mice which could be improved by steroid treatments in the latter model, this may indicate residual innate immune differences between background strains.For instance, NOD mice are deficient in complement humoral immunity due to a 2-bp deletion in the hemolytic complement (Hc) gene, which encodes the C5 complement protein [45].Our model now affords an opportunity for reconstitution of innate or adaptive immune cell types and humoral immune components to dissect the mechanisms of immunity to D. immitis migrating larvae, as has recently been attempted for L. sigmodontis with CD4+ T cell transfers into RAG2 -/-c -/-mice [46].This application of the model may be useful in determining minimally sufficient immune pathways necessary to mediate sterilising immunity.Similarly, the model may be valuable in evaluating efficacy of neutralising sub-unit vaccine target antibody responses (e.g., via passive transfer of purified specific antibodies or isolated B cell clones adoptively transferred from immunocompetent NOD mice).We evaluated that D. immitis larval parasitism and development in our mouse model accurately tracts the natural course of infection in definitive hosts over the first month.All larvae were recovered from the subcutaneous tissues and muscle fascia, in line with previous observations of natural parasite locations in both ferret and dog infections of D. immitis at this time-interval [33,47].We demonstrate that in vivo larvae complete cuticle moulting and undergo 4 th stage larval morphogenesis.L4 growth lengths in NSG/NXG mice are also within the range of those prior documented in dog and ferret infections at matching point of infection, at 14-15 days (NSG = 1.2-2.8mm, dog = 1.7-2.2mm, ferret = 1.6-2.7 mm) [33,47].
We also demonstrate that D. immitis expand Wolbachia titres significantly during parasitism of NSG mice.From PCR analysis, we ascertain that Wolbachia are doubling approximately every 42h for MO isolate to every 55h for GA III isolate over the first 14-day infection timecourse of D. immitis L3-L4 larvae in vivo.This is the first record of early Wolbachia expansions in D. immitis developing larvae and is comparatively slower compared with the average doubling time (32h) over the first 14 days of L3-L4 development in vivo for the human filariae, B. malayi [48].The establishment of the D. immitis mouse model now allows for tractable comparative endosymbiotic biology of this clade C nematode Wolbachia (also found in the causative agent of river blindness, O. volvulus) versus the clade D Wolbachia of human lymphatic filariae, most commonly used in basic and applied nematode Wolbachia research.Until now, a ready source of in vivo D. immitis L4 propagations for onward 'ex vivo' basic and translational research has been unavailable.Mosquito stage L3 can be induced to moult rapidly into the early L4 stage, with as much as 95% moulting success, and survive for 3 weeks in calf-serum supplemented cultures [21].We recapitulated this early L4 morphogenesis and improved L4 longevity to greater than one month in culture if larvae were co-cultured with dog or monkey kidney cells.However, comparisons with in vivo reared larvae highlighted several defects in growth, incomplete morphogenesis and, most strikingly, an almost complete failure to expand Wolbachia endosymbiont titres.Therefore, the failure of larvae to thrive in vitro may be linked with a deficit in Wolbachia-produced haem, riboflavin, nucleotides, or other biosynthetic pathways identified as relevant in the Wolbachia -nematode symbiosis [49].
Whether environmental cues are lacking in vitro for Wolbachia expansion is currently not known.Sub-optimal neo-glucogenesis in cultured D. immitis larvae could lower available carbon energy sources necessary for Wolbachia expansion [50].Alternatively, because autophagic induction in filariae regulates Wolbachia populations residing within host vacuoles [51], failure of Wolbachia growth may be the result of starvation / stress in culture inducing autophagy.Certainly, filarial stress responses are demonstrably upregulated in ex vivo adult worm culture systems [52].
We exemplify that use of sub-optimal L3/L4 grown in vitro for pharmacological screening leads to significantly >28-fold increased sensitivities to the paralytic activities of two nematodicical agents, moxidectin and levamisole, compared with larvae of same age derived from NSG/NXG mouse infections.Thus, we argue that reliance on in vitro larvae may lead to spuriously artificial sensitivities to new preventatives in development and, onwards, to incorrect selection of candidates or dose levels for in vivo preclinical evaluations with consequences for incorrect cat and dog usage.The new mouse model now affords a facile method of generating physiologically relevant L4 larvae for more accurate pharmacological assessments prior to a decision to advance into in vivo preclinical screening justifying protected animal use.
Our data determining the failure of Wolbachia to expand in vitro within D. immitis developing larvae precludes use of cultured L3/L4 in evaluating activities of novel anti-Wolbachia compounds.We thus demonstrate utility of the NSG/NXG mouse model as an in vivo anti-Wolbachia drug screen.L4 larvae could be reproducibly depleted of Wolbachia using a sevenday regimen of the established anti-Wolbachia antibiotic, doxycycline, with confirmatory experiments run in an independent laboratory with a different D. immitis isolate.Excitingly, we demonstrate that a two-day in vivo treatment with the novel investigational azaquinazoline drug, AWZ1066S, is a rapid and profound D. immitis Wolbachia depleting agent with 90% efficacy achieved.This benchmark of 90% efficacy has been determined as clinically relevant in terms of sustained Wolbachia reductions and subsequent long-term anti-parasitic activities in human filariasis clinical trials [12].The unique rapid activity of AWZ1066S has been determined through time-kill assays with B. malayi Wolbachia, whereby a near maximum kill rate can be achieved with 1-day exposure compared with six days for standard classes of antibiotics including tetracyclines [36].Thus, azaquinazolines or other novel anti-Wolbachia chemistry with similar rapid killing activity, as identified in high throughput industrial screening [53], might hold promise as new heartworm preventative or curative candidates and now can be triaged for activity utilising our novel D. immitis NSG/NXG mouse model.We used a high single parenteral dose of moxidectin, mimicking extended-release formulations used in dogs [37], to evaluate the D. immitis NSG/NXG mouse model as a preventative drug screen.We demonstrated, using multiple batches of different MLsusceptible D. immitis isolates, in different evaluating laboratories, that injected moxidectin mediated significant 65-89% reductions in larvae assessed between 13-27 days postexposure.Whilst drafting this manuscript, an independent study was published reporting a similar success of establishing D. immitis subcutaneous larval infections in NSG mice [54].Hess and colleagues also measured ivermectin and moxidectin responses in infected NSG mice following oral doses ranging between 0.001-3mg/kg given on day 0, 15 and 30 postinfection.Their studies determine that the MO isolate and an ivermectin resistant JYD-34 isolate of D. immitis were equally sensitive to moxidectin with high but incomplete efficacy demonstrable after 0.01mg/kg dosing.They also show high levels of ivermectin efficacy against the MO, but not JYD-34 isolate, in dose titrations ranging between 0.01-3mg/kg.Thus, we conclude the D. immitis NSG/NXG model is robustly validated by multiple independent laboratories as a screening tool for assessing direct-acting nematodicidal agents over at least a 28-day infection window.Our ex vivo and in vivo drug response evaluations of D. immitis L3/L4 in NSG/NXG mice demonstrate the flexibility to establish this model in independent laboratories with different commercially available NSG/NXG lines.Further, we determine feasibility of international shipping of live mf in dog blood to produce L3s for onward experimental infections in NSG/NXG mice.Mouse infections utilising shipped L3s may allow for increased accessibility to this model in order to expand experimental D. immitis research out of the few specialist reference centres which maintain the full life cycle of the parasite and/or the mosquito vector.Main limitations of our study are: 1) lack of data on full permissiveness to adult infections within murine cardiopulmonary vasculature, 2) lack of validation within female NSG/NXG mice and 3) less than 100% achievable moxidectin preventative efficacy response.In Hess et al, evaluation periods were extended and, whilst larvae continued to grow and mature within NSG mice, a divergence in growth compared with comparative dog studies was apparent after the first month of infection.Further, there was no evidence of immature adults arriving in the heart and lungs by 15 weeks [54].The authors conclude either physiological or anatomical deficiencies may prevent the full development of D. immitis in NSG mice.However, full development of the highly-related subcutaneous filaria, Loa loa, is possible in NSG mice [43] and thus it remains to be tested whether full D. immitis development may be achieved over an extended time-frame.We selected use of male mice due to observations that even in immunodeficient systems [55], as well as in outbred gerbils [56], male biased sex-specific susceptibility is a feature of rodent filarial infections.For future pharmacological investigations, to fully enable characterisation of interactions between sex and the drug pharmacokineticpharmacodynamic (PK-PD) relationship, it would be useful to assess whether D. immitis are able to develop within female NSG mice.In natural hosts, injectable formulations of moxidectin are proven to mediate 100% preventative efficacies [37].The substantial yet incomplete moxidectin responses in our NSG/NXG model may reflect full efficacy evolves over an extended time-period, particularly considering that this ML depots in fatty subcutaneous tissues and delivers a long-tail of systemic exposure, detectable over one month [57,58].Alternatively, an immunopharmacological mode-of-action involving disruption of immunosuppressive parasite secretions and an activated host-immune response has been proffered as one rationale why filarial larvae are differentially sensitive to ML-drugs at physiological levels in vitro vs in vivo [59].Thus, we currently cannot rule out a potential synergy with adaptive immune-mediated responses (such as the development of opsonising or neutralising antibodies) contributing to the complete efficacy of moxidectin.As previously discussed, passive transfer of antibodies into NSG/NXG mice may determine whether such a mechanism contributes to ML preventative efficacy at physiologically relevant dose levels.It is widely accepted that the use of specially protected, highly sentient species in preclinical research, including cats and dogs, should be strictly minimised wherever possible.This has not been plausible for veterinary heartworm preventative R&D due to a lack of a tractable small animal laboratory model.Typical drug screening has relied on in vitro potency testing against D. immitis larvae, potentially combined with initial preclinical evaluation in a surrogate rodent filarial infection model, before deciding to proceed into experimental dog infection challenge studies.Vulnerabilities of this approach include: differential drug sensitivities between larvae being tested in vitro versus in vivo, vagaries in filarial species larval migration routes/ parasitic niches, and variability in drug target expression / essentiality across different filarial parasite species and life-cycle stages, all of which may drive artefactual efficacy information.Our model, with international commercial supply of NSG/NXG strains and shipping of mf or L3 from donating laboratories, provides universal access to accurate and facile PK-PD assessments of preventative D. immitis drug candidate responses against the prophylactic L3-L4 larval target.Evaluations of drug larvicidal activities over the first month of infection, whilst larvae are developing in subcutaneous and muscle tissue, allows for rapid assessments whilst avoiding risk of welfare issues associated with arrival of adult parasites in the cardiovascular system.We observed no overt welfare issues of mice after parasitism, with mice gaining weight and displaying typical behaviour.If adopted, our model would accelerate drug research timelines and enable precise dose-fractionation studies for clinical selection.We conclude that a D. immitis NSG/NXG mouse model is established for more efficient heartworm drug discovery which will reduce the requirements for long-term cat and dog experimentation with the risk to cause severe harm, in line with an ethos of 'replacement, refinement and reduction' of animals in scientific research.

Animals
Male NOD.SCIDγc −/− (NSG; NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ) and BALB/c RAG2 −/− γc −/− (RAG2ɣc; C;129S4-Rag2 tm1.1FlvIl2rg tm1.1Flv /J) mice were purchased from Charles River UK. Male NXG mice (NOD-Prkdc scid -IL2rg Tm1 /Rj) were purchased from Janvier Labs, France.Mice were group housed under specific pathogen-free (SPF) conditions at the biomedical services unit (BSU), University of Liverpool, Liverpool, UK.Male NSG mice used at TRS laboratories were purchased from The Jackson Laboratory, USA, and group housed within filter-top cages.Mice were age 5-7 weeks and weighed 21-32 g at the start of experiments.Animals had continuous access to fresh sterile food and water throughout experiments.Weight was monitored twice weekly and welfare behaviour monitored daily.Study protocols were approved in the UK by LSTM & University of Liverpool Animal Welfare and Ethics Review Boards and licensed by The UK Home Office Animals in Science Regulation Unit.In the USA, studies were approved by TRS Institutional Animal Care and Use Committee.Dirofilaria immitis parasite production Missouri isolate (MO) Dirofilaria immitis microfilariae in dog blood were fed to female Aedes aegypti mosquitoes (Liverpool strain) at a density of 5000 mf/ml through an artificial membrane feeder (Hemotek, UK).Blood-fed mosquitoes were reared for 15 days with daily sugar-water feeding to allow development to the L3 stage.At day 15, DiL3 were collected from infected mosquitoes by crushing and concentration using a Baermann's apparatus and Rosewell Park Memorial Institute (RPMI) 1640 with 1% penicillin-streptomycin (both Sigma-Aldrich, UK).For validation studies at TRS Labs, USA, and in-house Georgia III (GAIII) isolate of D. immitis was utilised.Dirofilaria immitis mf were used to infect female Aedes aegypti mosquitoes (Liverpool strain) in dog blood using a glass feeder at a density of 1000 -2500 mf/ml.At day 14, DiL3 were collected from infected mosquitoes using crushing and straining with RPMI 1640 and 1% penicillin-streptomycin. Dirofilaria immitis experimental infections Highly motile infectious stage larvae (DiL3) retrieved from mosquitoes were washed in RPMI 1640 with 1% penicillin-streptomycin and 1% amphotericin B (Sigma-Aldrich, UK), and injected subcutaneously into the flank of male NSG/NXG or RAG2ɣc mice at a density of 200 DiL3 per mouse.Cohorts of mice also received a single intraperitoneal injection of 2 mg methylprednisolone acetate (MPA; Sigma-Aldrich, UK) immediately prior to infection and after one week post-infection.Mice were humanely culled between 14-28 days post-infection.To retrieve parasites, skins were removed and subcutaneous tissue scarified with a sharp scalpel blade.Muscle tissues were similarly scarified.Visceral organs were dissected and viscera, skin (pellet side-up), muscle tissues and carcass soaked in warm Eagle's minimum essential media (EMEM; Sigma-Aldrich, UK) with 1% penicillin-streptomycin and 1% amphotericin B for 2 hours to allow active larvae to migrate from tissues.Skin, muscle and carcasses were incubated for a further 24-hour period allowing residual larvae to migrate out of tissues.

In vitro larval cultures
Madin-Darby Canine Kidney (MDCK) cells and Rhesus Monkey Kidney Epithelial Cells (LLCMK2) cells were passaged in T-75 flasks in EMEM with 10% foetal bovine serum (FBS), 1% penicillin-streptomycin, 1% amphotericin B and 1% non-essential amino acid solution (NEAAS) (Sigma-Aldrich, UK).Cells were seeded onto 12-well plates to reach confluent monolayers 24-48 hours prior to parasite addition.For parasite cultures, washed MO DiL3 from mosquitoes were plated onto cell monolayers, or the cell-free media (EMEM) control at a density of 10-20 iL3 per well with 4 ml media.Larvae were monitored over a 35-day timepoint for survival and motility and at 14 days post-culture to evaluate development, length and Wolbachia titres.In vitro and ex vivo drug screening assays MO DiL3 larvae were transferred onto MDCK monolayers and allowed to develop to 14 (earlymid L4) or 28 (mid-late L4) day old larvae.For comparative ex vivo assays, L4 stage larvae were recovered from male NSG mice 14-days post infection and washed in sterile EMEM prior to the addition of drugs.All larval stages were plated into 12-well plates at densities of 3-5 larvae per well per drug concentration in 4ml of EMEM with 10% FBS, 1% penicillinstreptomycin, 1% NEAAS and 1% amphotericin B for drug screening assays.Moxidectin (Sigma-Aldrich, UK) was solubilised in phosphate buffered saline (PBS, Fisher Scientific) and 10-fold serial dilutions ranging from 0.0001-100µM were prepared in EMEM with 1% penicillinstreptomycin, 1% NEAAS and 1% amphotericin B. Vehicle controls were included using the equivalent percentage PBS added to the cultures.Assays were incubated for 6 days in which larvae were continuously exposed to drug at 37°C, 5% CO 2 , and scored daily for motility and survival.
In vivo drug screening validation Paired groups of 1-5 male NSG mice were subcutaneously inoculated with 200 DiL3 into the right flank on day 0. They were then randomised into treatment groups with a single subcutaneous dose of moxidectin prepared at 2.5 mg/kg in saline, or a saline only control, in the nape of the neck on day 1.Mice were monitored daily for weight change and culled at day 14 to evaluate efficacy based on parasite recoveries.Alternatively, immediately following infection, groups of 4-6 mice were randomised into a 7-day oral regimen of doxycycline at 50 mg/kg prepared in ddH 2 O followed by a 7-day washout period, a 2-day oral bi-daily regime of AWZ1066S prepared in standard suspension vehicle (SSV; PEG300/propylene glycol/H 2 O (55/25/20), or matching vehicle controls.Mice were monitored daily (weight and welfare) and culled between day 14-28 post-infection to evaluate parasitology, L4 length, and/or Wolbachia depletion using qPCR.

Wolbachia titre analysis
To assess Wolbachia titres across different developmental timepoints (2-, 3-and 4-weeks post infection), to compare in vitro reared larvae in parallel to in vivo reared controls and to investigate drug activity against Wolbachia, individual larvae were taken, and their DNA .CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is The copyright holder for this preprint this version posted April 3, 2023.; https://doi.org/10.1101/2023.04.03.535321 doi: bioRxiv preprint extracted using previously published methods [15].Wolbachia single copy Wolbachia surface protein (wsp) gene quantification was undertaken by qPCR using the following primer pair: F-TTGGTATTGGTGTTGGCGCA and R-AGCCAAAATAGCGAGCTCCA, under conditions used to determine Brugia malayi wsp copy numbers [15].Fluorescence in-situ hybridisation Fluorescence in-situ hybridisation (FISH) was used for detecting Wolbachia in DiL3 and L4 larvae using two different DNA probes specific for Wolbachia 16S rRNA: W1 -/5ATTO590N/AATCCGGC-GARCCGACCC and W2 -/5ATTO590N/CTTCTGTGAGTACCGTCATTATC, as previously described Walker, Quek (60).L3 and L4 larvae were stored in 50% ethanol at room temperature until further processing.For FISH staining, frozen larvae were fixed using 4% paraformaldehyde (PFA) and incubated with 10ug/ml pepsin for 10 min at 37°C.After thorough wash using PBS, samples were hybridised overnight in hybridisation buffer with probes (or without probes for negative controls).Hybridisation buffer consisted of 50% formamide, 5xSCC, 0.1M dithiothreitol (DTT), 200g/L dextran sulphate, 250mg/L poly(A), 250mg/L salmon sperm DNA, 250mg/L tRNA and 0.5x Denhardt's solution.Larvae were then washed twice in 1xSSC and 0.1xSSC 10Mm DDT before mounting with VECTASHIELD antifade mounting medium containing DAPI (4',6-diamindio-2-phenylinole) (Vector laboratories).L4 were visualised using brightfield microscopy for length measurements, calculated using Fiji (ImageJ), USA.FISHstained larvae were imaged using Zeiss laser scanning confocal microscope and changes in larval morphology was visualized using brightfield and DAPI nuclear staining.

Statistical analysis
Continuous data were tested for normality using the using D'Agostino & Pearson omnibus Shapiro-Wilk normality tests.Where data were skewed, non-parametric analyses were used to compare statistical differences between groups using Dunn's post-hoc tests.Where data passed normality tests, Tukey's post-hoc tests were applied.Categorical data was analysed by Fisher's Exact Tests.Survival of larvae in culture (frequency motile vs immotile) was evaluated by Log-rank (Mantel-Cox) test.Moxidectin / levamisole IC 50 values were derived 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is The copyright holder for this preprint this version posted April 3, 2023.; https://doi.org/10.1101/2023.04.03.535321 doi: bioRxiv preprint

Figure 4 :
Figure 4: Validation of D. immitis NSG mouse model as an in vivo larvicidal drug screen.
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is larval development phase of D. immitis and can be successfully utilised as a primary drug screening model for evaluation of direct-acting preventatives and anti-Wolbachia therapeutics.(whichwas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is tissue (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is Figure 1.Susceptibility of compound immunodeficient mouse strains to D. immitis.Experimental schematic using Missouri (MO) isolate D. immitis (A).Percentage recovery of initial infectious load of D. immitis L4 at 14 days post-infection in indicated immunodeficient mice (B-D).Experimental schematic using Georgia (GA)-3 isolate D. immitis (E).Percentage recovery of initial infectious load (F) and tissue distributions (G) of D. immitis L4 at indicated time-points in NSG mice.Bars indicate median values with individual mouse data plotted.Significant differences were assessed by Kruskal Wallis One-way ANOVA with Dunn's multiple comparison's tests except (G) where difference in proportions were tested byFisher's Exact Test.Significant differences (P values <0.05) are indicated in bold.Data is combined of two or more independent experiments (B-D) or individual experiments (C,D,F) with between 1-5 mice per group.
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is vs 2.4x105Wolbachia/larva, Mann-.CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is Figure 3: validation of the D. immitis NSG mouse model as an in vivo anti-Wolbachia drug screening system.Experimental schematic using MO or GA III isolates of D.immitis L3 (A).Wolbachia loads determined by PQCR in d14 MO D. immitis larvae exposed to doxycycline (B) or AWZ1066S (C).Length changes in d14 MO D. immitis larvae exposed to doxycycline (D) or AWZ1066S (E).Wolbachia loads determined by QPCR in d14 GA III D. immitis larvae exposed to doxycycline (F).Visualisation of Wolbachia depletion in hypodermis of d14 GA3 D. immitis larvae exposed to doxycycline by FISH (G).Length changes in d14 GA III D. immitis larvae exposed to doxycycline (H).Bars represent median (B, C, F) or mean +/-SEM (D, E, H) values with individual larva data plotted.Significant differences were determined by Mann-Whitney tests (B, C, F) or unpaired Student's T Tests (D, E, H).Significant (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is The copyright holder for this preprint this version posted April 3, 2023.; https://doi.org/10.1101/2023.04.03.535321 doi: bioRxiv preprint from percentage immotile larvae per drug concentration on day 6 of the assay.Non-linear curves were generated by 3-parameter least squares fit with [IC 50 ] calculated.All tests were performed in GraphPad Prism 9.1.2software.Significance is indicated at or below alpha = 0.05.10.1016/j.cub.2021.03.056.PubMed PMID: 33857432; PubMed Central PMCID: PMCPMC8210651. .CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is The copyright holder for this preprint this version posted April 3, 2023.; https://doi.org/10.1101/2023.04.03.535321 doi: bioRxiv preprint