Evaluation of in vitro culture systems for the maintenance of microfilariae and infective larvae of Loa loa

Background Suitable and scalable in vitro culture conditions for parasite maintenance are needed to foster drug research for loiasis, one of the neglected tropical diseases which has attracted only limited attention over recent years, despite having important public health impacts. The present work aims to develop adequate in vitro culture systems for drug screening against both microfilariae (mf) and infective third-stage larvae (L3) of Loa loa. Methods In vitro culture conditions were evaluated by varying three basic culture media: Roswell Park Memorial Institute (RPMI-1640), Dulbecco’s modified Eagle’s medium (DMEM) and Iscove’s modified Dulbecco’s medium (IMDM); four sera/proteins: newborn calf serum (NCS), foetal bovine serum (FBS), bovine serum albumin (BSA) and the lipid-enriched BSA (AlbuMax® II, ALB); and co-culture with the Monkey Kidney Epithelial Cell line (LLC-MK2) as a feeder layer. The various culture systems were tested on both mf and L3, using survival (% motile), motility (T90 = mean duration (days) at which at least 90% of parasites were fully active) and moulting rates of L3 as the major criteria. The general linear model regression analysis was performed to assess the contribution of each variable on the viability of Loa loa L3 and microfilarie. All statistical tests were performed at 95% confidence interval. Results Of the three different media tested, DMEM and IMDM were the most suitable sustaining the maintenance of both L. loa L3 and mf. IMDM alone could sustain L3 for more than 5 days (T90 = 6.5 ± 1.1 day). Serum supplements and LLC-MK2 co-cultures significantly improved the survival of parasites in DMEM and IMDM. In co-cultures with LLC-MK2 cells, L. loa mf were maintained in each of the three basic media (T90 of 16.4–19.5 days) without any serum supplement. The most effective culture systems promoting significant moulting rate of L3 into L4 (at least 25%) with substantial maintenance time were: DMEM + BSA, DMEM + NCS, DMEM-AlbuMax®II, DMEM + FBS all in co-culture with LLC-MK2, and IMDM + BSA (1.5%), DMEM + FBS (10%) and DMEM + NCS (5%) without feeder cells. DMEM + 1% BSA in co-culture scored the highest moulting rate of 57 of 81 (70.37%). The factors that promoted L. loa mf viability included feeder cells (β = 0.490), both IMDM (β = 0.256) and DMEM (β = 0.198) media and the protein supplements NCS (β = 0.052) and FBS (β = 0.022); while for L. loa L3, in addition to feeder cells (β = 0.259) and both IMDM (β = 0.401) and DMEM (β = 0.385) media, the protein supplements BSA (β = 0.029) were found important in maintaining the worm motility. Conclusions The findings from this work display a range of culture requirements for the maintenance of Loa loa stages, which are suitable for developing an effective platform for drug screening. Electronic supplementary material The online version of this article (10.1186/s13071-018-2852-2) contains supplementary material, which is available to authorized users.


Background
Loiasis is a parasitic disease caused by the filarial nematode Loa loa that is transmitted through the bite of an infected Chrysops fly. Loiasis is endemic in the rainforest areas of West and Central Africa [1]. The common clinical signs of loiasis are the subconjunctival migration of the adult worm, reported for the first time by Mongin in 1770 [2], Calabar Swelling, pruritis, oedemas and arthralgia. Interest in this filarial species, which has long been considered to be less pathogenic than related species [3], came from several reports in Cameroon indicating that high microfilaraemia of L. loa is associated with severe and sometimes fatal encephalopathic reactions in patients who had taken ivermectin for onchocerciasis treatment [4][5][6][7]. Loiasis is a neglected tropical disease (NTD) which has attracted only limited attention in drug research and development. Apart from surgical removal of adult worms moving under the skin or across the eye that can be done to relieve anxiety, only two medications have so far been employed for clinical cases since the last century, namely diethylcarbamazine (DEC) and albendazole. The latter is sometimes used in patients who are not cured with multiple DEC treatments. Several cases of brain inflammation, coma and death have been reported in people with heavy infections when they are treated with DEC [8,9]. The risk of side effects has limited the deployment of mass drug administration of ivermectin in areas where the L. loa prevalence exceeds 20% [10][11][12], impeding the goals stated by the African Programme for Onchocerciasis Control (APOC) in areas of co-endemicity. Progress in drug research and development for loiasis requires suitable screening systems both at in vitro and in vivo levels. Though innovations in filarial animal models have recently been achieved [13,14], in vitro maintenance systems of the different stages of L. loa have not been established. The present study aimed to design suitable in vitro culture systems for drug screening against both infective larvae (L3) and microfilariae (mf ) of L. loa.

Methods
Isolation and purification of L. loa L3 Loa loa L3 were obtained from dissected Chrysops flies that had previously fed on a consented microfilaremic individual at Ediki Forest (South West region, Cameroon). Engorged Chrysops were kept in captivity for 12 days, to allow development to the infective stage (L3). The flies were fed daily with 15% sucrose solution soaked in cotton wool. After 12 days of rearing, the flies were dissected in Petri dishes containing RPMI 1640 medium (Sigma-Aldrich, St Louis, USA). The head, thorax and abdomen were separated and teased apart in three different Petri dishes. Fly tissues were incubated for 20 min to allow L3 larvae to migrate out. A sterile pipette was used to pick the larvae and pooled in a shallow convex glass dish [15]. The worms were transferred into 15 ml centrifuge tubes (Corning, Kennebunk-ME, USA) for purification. Only L3 harvested from the head (where more mature larvae are expected to be found) were used in this study. The remaining larvae were frozen to be used in other studies (immunology and molecular biology). The L3 were washed using a Percoll® (GE Healthcare, Pharmacia, Uppsala, Sweden) technique. The L3 suspension concentrated in less than 1 ml RPMI was slowly layered on the surface of a 15 ml tube containing stock iso-osmotic Percoll® and centrifuged (Humax 14k human, Germany) at 800× rpm for 10 min. The process was repeated to remove microbial contaminants. At the end, the L3 were washed twice with RPMI-1640 by centrifugation at 1500× rpm for 10 min to remove Percoll® remnant.
Isolation and purification of L. loa mf Loa loa mf were obtained from baboons (Papio anubis) experimentally infected with human strain of L. loa reared in Kumba Medical Research Station (South West region, Cameroon). Peripheral blood samples of hypermicrofilaraemic baboons were collected as described in the previous reports [13]. Microfilaraemic loads were determined microscopically on thick films. Calibrated thick blood smears were prepared by spreading a 50 μl venous blood sample from a 75 μl non-heparinised capillary tube, onto a clean slide over an area of 1.5 × 2.5 cm [16]. After drying, films were dehemoglobinized and stained with Giemsa. The Percoll® density centrifugation method previously described [17] was used to purify mf from infected blood samples.

Monkey kidney cell co-culture
Monkey kidney epithelial cells (LLC-MK 2 ) (ATCC, USA) were cultured in flasks at 37°C in a CO 2 incubator (Sheldon Mfg. Inch, Cornelius, OR, USA) at 5% CO 2 until the cell layer became fully confluent. For new inoculations and other cell manipulations, trypsin was used to detach cells from the walls of the flasks. Cells were then dislodged with trypsin solution (25%) containing EDTA, the mixture was kept at 37°C for less than 1 h. The cell suspension was centrifuged at 1,500 rpm for 10 min, the supernatant was discarded, and the pellet resuspended and diluted to 10 5 cells/ml in complete culture medium. Aliquots (100 μl) of cell suspensions were plated into a 48-well culture plate and kept in the incubator for cells to become fully confluent.

Assessment of parasite viability
The viability of the parasites was assessed daily, by visual inspection (by two individuals) under an inverted microscope until they die. Their motility was scored on a 4point scale [18,19]: 0, no movement or immotile; 1, intermittent shaking of head and tail; 2, sluggish (shaking of the whole worm on a spot); 3, vigorous movement (shaking of the whole worm and migration from one spot to was considered).

Data processing and analysis
Three different batches of L3 larvae and microfilariae were used for each culture system. For each batch of parasites, 4 replicate wells were used per system. Raw data collected daily on record sheets were entered into a template designed on Microsoft Excel 2007. Three variables were defined and computed to assess the viability of the parasites (mean motility and mean mortality, T 90 ).
Motility variable was computed based on the scoring system described above, and using the following formula.
where Si is the score of point scale i and Ni is the total number of worms at a point scale i.
The variable T 90 was defined as the duration at which 90% of the worms were still fully active (score 3 above) in the well. This variable was set as one of the major indicators of the suitability of the culture system, with relevance to drug screening for loiasis. From values obtained after testing each system on three batches of parasites, T 90 values were expressed as mean ± standard deviation.
The Kruskal-Wallis test was used to assess the global significant differences between the distribution of the median T 90 across media and supplements, and the pairwise multiple comparisons of the ranked data was performed using the Pairwise Multiple Comparisons of Mean Rank (PCMR) package in R version 3.1.4. Mann-Whitney U-test was used to compare the the medians of the T 90 between the cell free and cell containing culture. Statistical tests were interpreted using a 5% significance level. It was considered that a valid appreciation of the effect of any drug could be possible only in a system where at least 90% of parasites motility were sustained till the end of the experiment.
Factors that promoted parasite survival were identified using the multiple linear regression. The general linear model (GLM) was built using the hierarchical stepwise method. A total of 5 blocks were achieved with the 5 factors (incubation time, presence of feeder cells, basic medium, serum/protein, protein concentration) and those that contributed significantly to the improvement of the model were identified based on the F-statistics and the adjusted R-square (Additional file 1: Table S1). The incubation time was treated as a metric factor. Dichotomous variables such as the presence of monkey kidney cells were coded using binary figures. For each nominal or ordinal factor (Basic culture media, protein or protein concentration), sets of dummy variables were created and compared to one of the categories defined as reference. While RPMI-1640 was used as a reference against DMEM and IMDM, the four sera (Albumax II, BSA, FBS and NCS) were compared to the serum free culture (No serum). The three concentrations of each serum (0.5%, 1% and 1.5% for Albumax II and BSA; 5%, 10% and 15% for FBS and NCS) were labelled using the three ordinal levels: low, medium and high concentration; they were also compared to the serum-free culture. Interaction factors were created between explanatory variables and added to the models. The prediction of the motility by the protein concentration was poor, and there was a non-statistical difference between different concentrations of protein as will be discussed latter. Based on experimental observations, interactions were expected between the other three variables. Therefore, two ways and three ways interaction terms were created between those three experimental parameters: the presence of feeder cells, the culture medium used in reference to RPMI1640 and the type of protein supplement.
The passage of the L. loa larvae from the third (L3) to the fourth (L4) stages was further considered the second target product profile in assessing the suitability of the culture systems tested. For each of the 78 culture systems designed and evaluated, the moulting rate (percentage of moulted worms) as well as the timeframe was computed.

Results
Evaluation of the effect of each basic culture medium on the viability of L. loa L3 and mf The T 90 values of the parasite motility in different basic media were evaluated and results are presented in Fig. 1 and the statistical report in Table 1. The values of T 90 ranged from 3.0 day to 6.5 days for L3. In the absence of protein supplement and feeder cells, there were significant difference between the T 90 value of the 3 media vis a vis the L3 (χ 2 = 38.793, df = 2, P < 0.0001) and mfs (χ 2 = 38.793, df = 2, P < 0.0001). Pairwise comparison indicated that L3 survived longer particularly in IMDM compared to DMEM (P < 0.0001) and RPMI (P < 0.0001), whereas mf hardly exceeded 3 days survival, irrespective of the culture medium tested in absence of supplements. However, mf survival time was significantly higher in DMEM than in RPMI (P < 0.0001) or IMDM (P = 0.0010).
Evaluation of the effect of serum/protein supplementation on the viability of L. loa larvae and microfilariae in culture Effect of serum/protein on the viability of L. loa larvae AlbuMax® II Generally, the supplementation of DMEM and IMDM with AlbuMax® II promoted the viability of larvae, increasing T 90 by 1.5-2.5-fold. RPMI supplemented with AlbuMax® II inhibited the larvae viability (T 90 < 1 day, Fig. 2a).

Fetal bovine serum (FBS)
This serum could support the viability of larvae for up to 8.7 days when IMDM was supplemented with 10% FBS. T 90 > 5 days was observed in all concentrations of FBS except in the RPMI supplement medium (Fig. 2b).

Bovine serum albumin (BSA)
All media supplemented with BSA also improved larvae viability with T 90 ≥ 5 days, except for DMEM supplemented with 1% BSA and all concentrations of BSA with RPMI (Fig. 2c).    All concentrations of AlbuMax® II supplement improve the viability of mf regardless the basic medium except for RPMI. Although the improvement was noticeable, all T 90 were less than 5 days as shown in Fig. 3a.

Fetal bovine serum (FBS)
All concentrations of fetal bovine serum improved parasite viability but T 90 > 5 days was reported only with FBS supplemented DMEM and IMDM (Fig. 3b).

Bovine serum albumin (BSA)
Generally, BSA boosted the microfilariae viability. The T 90 values above 5 days were reported with all concentrations of BSA supplemented DMEM and IMDM + 0.5% BSA (Fig. 3c).

Assessment of the importance of monkey kidney cells as feeder layer
The findings on the effect of the monkey kidney epithelial cells as feeder on the survival of the L. loa L3 and mf are summarized in Fig. 4 and Fig. 5, respectively and the summary of the statistical report is presented in   Table  S3, with an illustration in Fig. 9. The proportion of moulting recorded with RPMI was below 25%. This contrasted with DMEM and IMDM where significant moulting rates were noted, both in cell-free and coculture of L3 with LLC-MK2. Apart from the isolated case of IMDM with up to 27.27% (24 moulted worms out of a total of 88), serum/protein supplementation was found indispensable for the transition from L3 to L4 in vitro. The higher moulting rates were generally observed with DMEM, compared to IMDM. The highest moulting rate was observed in protein and feeder layer supplemented DMEM as illustrated in Fig. 6. Loa loa L3 moulting began on day 9 and observation was continued until day 29, supplementation with BSA scored highest moulting rates, with 57.32, 70.37 and 58.62% larvae moulted in 0.5, 1 and 1.5% BSA supplement, respectively.
Linear regression analysis of different factors that influence the in vitro maintenance of L. loa microfilariae and L3 Bivariate analysis indicated strong association between motility and incubation time (Spearman's rho = -0.674, P < 0.001). This was the first to be introduced in the single linear regression analysis (R 2 = 0.452), and the remaining variables were successively added to construct the final GLM. Before interaction terms were introduced in the models, the important factors that contributed to the improvement of worm motility were identified separately based on their standardized coefficient (Fig. 10). From the built models, the important variables required to meet the threshold of T 90 are as followed and classified with respect to their importance. Feeder cells were found as the most important. In combination to DMEM (DMEM-LLCMK2), the unstandardized coefficient decreased although the interaction was the best among others. In addition, the combination of three variables (feeder layer-basic culture medium-protein/serum supplement) was not mandatory as the interaction LLCMK2-Albumax instead weakened the model with a negative coefficient. The model was diagnosed by assessing the assumptions of normal distribution and homoscedasticity. The histogram of the residuals (errors) in the model was used to check if they are normally distributed (Additional file 4: Figure S1). Although not perfect, the frequency distribution of the residuals has a shape close to that of the normal Gauss curve, indicating evidence of normal distribution. Additionally, P-P plot was used for further check (Additional file 5: Figure S2). Here, the expected and observed cumulative probabilities were closed suggesting that the assumption of normal distribution of the residual was far to be not violated. The scatterplot of standardized residuals against standardized predicted values were used to assess the assumption of homoscedasticity (Additional file 6: Figure S3). The variance of residuals were random distributed indicating that the assumption of homoscedasticity was likely to be safe.

Discussion
Drug discovery research for L. loa has so far attracted only very limited attention compared to other filarial diseases. Repurposing attempts have been conducted with existing drugs with only limited success. The effect of several antimalarial drugs (quinine, chloroquine, amodiaquine and artesunate) was investigated on loiasis in a randomized, placebo-controlled approach in central Cameroon [20]. This study recorded no significant change in parasite loads in any of the treatment groups. Another study tested different intermittent doses of albendazole on Loa loa microfilaraemia, the reduction in mf load obtained was insufficient to prevent the risk of severe adverse reactions during ivermectin mass drug administration in loiasis co-endemic areas [21]. These observations demonstrated that repurposing of existing antiparasitic therapies may not be a suitable approach to develop drugs with satisfactory therapeutic window. The traditional approach starting from standard in vitro discovery through preclinical and clinical testing necessitates in vitro maintenance of L. loa stages for a minimum duration required for drug screening. Herein we designed and tested the effect of varying 78 culture conditions on both L3 and mf viability starting with three basic culture media, four serum/protein supplements and one feeder cell. In general, L. loa L3 survived for longer periods than mf in the different basic culture media with neither protein nor feeder cell supplementation. For the maintenance of L3, IMDM exhibited the best performance, whereas on microfilariae, DMEM had the highest T 90 though only three-day survival. Considering a minimum cut-off point of five-day maintenance of 90% highly active larvae or microfilariae for in vitro drug screening on L. loa microfilariae and infective  larvae, IMDM was the only basic medium that could be employed as such without any need of protein supplement/feeder layer. However, none of the media would be suitable for in vitro investigations requiring longer periods of incubation.
Four different serum/protein supplements were therefore applied at increasing concentrations to the three basic culture media in order to improve their nutritional potencies for both L3 and mf. The results obtained were highly diverse. In RPMI, serum/protein supplementation rather caused drop in T 90 which was more pronounced for AlbuMax® II followed by FBS, NCS and BSA, in both L3 and mf. Consequently, none of the formulations based on RPMI supplemented with serum/protein (without feeder layer) was successful in keeping mf alive and active for more than five days. Mengome et al. [22] recently reported on screening of 12 methanolic extracts of nine traditional plant remedies employed in Gabon, on L. loa mf maintained in vitro using modified Eagle's medium supplemented with 10% foetal calf serum with five-day incubation time. Our findings showing an average T 90 value of 7.5 ± 2.9 days and thus corroborate the data reported by these authors, confirming the suitability of the culture system employed.
A drastic increase in the survival time and viability of the L. loa mf was obtained with addition of LLC-MK2 cells as feeder layer, except for AlbuMax® II in RPMI. The T 90 values were extended for all culture formulations (except AlbuMax® II in RPMI), even exceeding 20 days.
The hypothesis that LLC-MK2 cells alone can be enough to sustain the viability of the L. loa parasites in vitro was tested. In all the three basic culture media, RPMI, DMEM and IMDM, without serum/protein supplement, parasites could still survive 15 days and beyond. Toback et al. [23] reported that monkey kidney cells during their growth in vitro produce growth factors such as the epidermal growth factor (EGF), interleukin growth factor (IGF) and transforming growth factor (TGF-β). Thus, these factors are possibly supportive of L. loa larval growth and/or survival, which facilitate their in vitro maintenance.
Culture systems capable of sustaining moulting of infective larvae present an additional advantage for in vitro investigations on the parasite, including both physiological studies and the exploration of drug targets. In addition to the survival of parasites, the effect of the different culture systems on the moulting of L3 was further examined. Maintenance and moulting of filariae larvae using different culture systems have been reported for Onchocerca spp. [24,25], Wuchereria bancrofti [26][27][28][29] and Brugia malayi [30,31]. The present study is the first attempt to optimise such systems for L. loa larvae.