High Plasmodium Infection Rate and Reduced Bed Net Efficacy in Multiple Insecticide-Resistant Malaria Vectors in Kinshasa, Democratic Republic of Congo

High and multiple resistance to insecticides are recorded in the 2 main malaria vectors in the Democratic Republic of the Congo, leading to a significant loss of efficacy of conventional bed nets in the presence of alarmingly high Plasmodium infection rate, suggesting high malaria transmission.

With 11% of all worldwide malaria cases occurring in the Democratic Republic of the Congo (DRC) [1], this country is key for malaria elimination. Control efforts in DRC extensively rely on the use of long-lasting insecticidal nets (LLINs). However, the development of insecticide resistance is threatening the effectiveness of this control tool, calling for urgent action to implement suitable resistance management strategies [2]. Unfortunately, the scarcity of information on the extent and impact of resistance prevents the design of such strategies in DRC.
Resistance to pyrethroids and organochlorines has previously been reported in DRC, although exclusively in the major malaria vector Anopheles gambiae [1,3,4]. Nevertheless, the extent and intensity of this resistance remain unclear and completely unknown for other major vectors such as Anopheles funestus. Additionally, the resistance profile to other insecticide classes has not been clearly established, limiting the ability of national malaria control programs to make informed decisions for resistance management. Indeed, so far, only limited evidence of resistance to carbamates and organophosphates has been reported in An. gambiae from DRC [1], with the exception of possible resistance to malathion observed in a single locality [3]. Furthermore, it also remains unclear how pyrethroid resistance in malaria vectors currently impacts the efficacy of both conventional and piperonyl butoxide (PBO)-based LLINs, although a low efficacy has been reported previously for the OlysetNet [5].
Limited investigation of resistance mechanisms in DRC has detected the presence of the kdr mutation in An. gambiae in Kinshasa [5]. The role of metabolic resistance has not been explored further. To support and facilitate the success of the ongoing and future vector control programs in DRC, the present study extensively evaluated the current insecticide resistance profile of An. gambiae and An. funestus, the 2 main malaria vectors in Kinshasa, the capital city. Furthermore, the efficacy of several LLINs was assessed, in addition to the molecular mechanisms driving resistance. Additionally, the Plasmodium infection rate in malaria vectors was also assessed. and An. funestus, through the Ndjili River and its flooded shores. Blood-fed and half-gravid female Anopheles, morphologically identified as belonging to the An. funestus or An. gambiae complex [6], were kept in holding cages for 4-6 days and forced to lay eggs individually during 2-3 extra days [7]. One hundred thirty-three An. gambiae sensu lato (s.l.) and 79 An. funestus s.l. laid eggs. F 0 dead females and F 1 eggs were transported to the Liverpool School of Tropical Medicine for the subsequent analysis under a license from the Department for Environment, Food and Rural Affairs (DEFRA) (PATH/125/2012, UK).

Species Identification
Genomic DNA from 111 An. gambiae s.l. and 81 F 0 An. funestus s.l. whole female mosquitoes was extracted, using the Livak protocol [8], and identified to species level using a cocktail polymerase chain reaction (PCR) assay [9,10]. Larvae were transferred to plastic trays according to species group for rearing, as previously described [7,11].

Plasmodium Infection Rate
The Plasmodium infection rate was estimated using the mosquito's whole body extracts by detecting the presence of Plasmodium falciparum (F+) and/or Plasmodium ovale, Plasmodium vivax, and Plasmodium malariae (OVM+) in 109 An. gambiae s.l. and 81 An. funestus sensu stricto (s.s.) field-collected F 0 females individually using the TaqMan assay, as previously described [12,13]. Results of TaqMan assay were confirmed by performing a nested PCR assay as previously described [14].
WHO tube bioassays were used to assess the mortality of An. gambiae s.l. after extended periods (3 and 6 hours) of exposure to DDT and deltamethrin. Furthermore, in addition to the 60-minute exposure described above, mortality after 30-and 90-minute exposures to bendiocarb was also assessed. The mortality rates were determined 24 hours later.

Synergist Assays
Synergist assays were performed with PBO using An. gambiae s.l. Four replicates of 20-25 adult mosquitoes (2-5 days old) were preexposed using WHO tube bioassays with PBOimpregnated papers (4 %) for 1 hour. Thereafter, the mosquitoes were immediately exposed to permethrin (0.75%), deltamethrin (0.05%), or DDT (4%) for 60 minutes. Control assays using only 4% PBO papers for 60 minutes were also performed. Mortality was scored after 24 hours, and the results obtained were compared with the mortality without PBO exposure using unpaired Student t test.

Insecticide-Treated Bed Net Efficacy
The efficacy of conventional bed nets against both mosquito populations was estimated by 3-minute exposure cone bioassays following the WHO guidelines with minor modifications, with respect to the number of pieces of net and the number of mosquito per replicate tested [16]. In brief, 5 replicates of 10 F 1 females (2-5 days old) were placed in plastic cones attached to different commercially produced nets newly purchased: OlysetNet, OlysetPlus, Permanet 2.0, Permanet 3.0-side and -roof, and an untreated net (as a control). Due to the low number of F 1 mosquitoes available, only 1 piece per type of net was tested. After exposure, the mosquitoes were placed in holding paper cups with cotton soaked in 10% sugar solution. Mortality was recorded 24 hours after exposure.

An. gambiae
To assess the role of target-site knockdown resistance in An. gambiae s.l., 111 F 0 female field-collected mosquitoes were genotyped for the L1014F and L1014S kdr mutations by TaqMan assay as previously described [17]. Furthermore, the genetic diversity of the voltage-gated sodium channel (VGSC) gene was investigated for An. gambiae. A fragment of intron 19 of the VGSC gene spanning a portion of exon 20 (including the 1014 codon associated with kdr) was amplified in 11 field-collected An. gambiae s.s. females, cleaned, and sequenced as previously described [7,18]. Sequences were aligned using ClustalW [19], whereas haplotype reconstruction and polymorphism analysis were done using DnaSPv5.10 [20]. DRC haplotypes were compared to the 4 kdr haplotypes previously detected across Africa as containing either the 1014F (H1/H4) or the 1014S (H2/H3) mutations [21], as well as to a susceptible haplotype from Cameroon [22]. All DNA sequences have been submitted to GenBank (accession numbers KY700707-KY700728).

Genotyping of Resistance Markers in An. funestus
The role of the An. funestus s.s. L119F-GSTe2 and A296S-RDL resistant markers, involved in DDT/permethrin and dieldrin resistance, respectively, was also evaluated by TaqMan assay, genotyping 66 F 0 female mosquitoes collected from the field. L119F-GSTe2 and A296S-RDL TaqMan reactions were performed as previously described [23,24].

Transcription Profile of Resistance Genes in An. funestus
Total RNA was extracted from 3 batches of 10 adults from F 1 female An. funestus s.s. mosquitoes nonexposed to insecticides and the FANG susceptible strain, as previously described [24,25]. The expression patterns of key resistance genes including CYP6P9a, CYP6P9b, CYP6M7, and GSTe2 were assessed by quantitative reverse-transcription PCR (qRT-PCR) [24]. After normalization with housekeeping genes Actin (AFUN006819) and RSP7 (AFUN007153-RA), the relative expression for each gene was calculated according to the 2 -ΔΔCT method [26]. The statistical significance between gene expression estimates was performed using unpaired Student t test.

Synergist Assays
The synergist assay results showed a slight recovery of susceptibility after PBO preexposure for the 2 pyrethroids tested  Figure 3A). Tests with DDT also revealed a lack of impact of PBO preexposure with only 2.7% ± 2.7% mortality (P = .29) observed with PBO exposure vs no mortality without exposure, suggesting that DDT resistance is conferred by other mechanisms or gene families. No mortality was observed in control mosquitoes exposed to the synergist PBO only.

Insecticide-Treated Bed Net Efficacy
Cone assays were conducted to evaluate the efficacy of conventional bed nets ( Figure 3B)  In addition, sequencing of a 510-bp fragment of VGSC gene spanning the 1014 codon revealed a reduced genetic diversity with only 4 polymorphic sites and 3 haplotypes, including a predominant 1014F haplotype (19/22), in line with the near fixation of the 1014F allele in this population. Comparison of the DRC-VGSC haplotypes with 4 kdr bearing haplotypes previously detected across Africa revealed that all 1014F haplotypes from Kinshasa belong to the H1-1014F haplotype, predominant in West/Central Africa [21] ( Figure 4B). The 1014S haplotypes belong to the H3-1014S haplotype previously described in East Africa [21], whereas the other haplotype (CG63) exhibited a single mutational step difference (position 91) from the previously described H2-1014S. This is further supported by the Templeton-Crandall-Singh haplotype network tree ( Figure 4C), showing that the 1014S haplotypes are separated by 1 or 2 confirmed mutational steps ( Figure 4D), suggesting an independent occurrence of the CG63-1014S haplotype in DRC, potentially from local selection.

Transcription Profile of Resistance Genes in An. funestus
The transcription analyses of 3 cytochrome P450 genes, CYP6P9a, CYP6P9b, and CYP6M7, known to be involved in pyrethroid resistance in An. funestus [30][31][32], and 1 glutathione-s-transferase, GSTe2, previously shown to confer DDT and permethrin resistance [23], were assessed by qRT-PCR in nonexposed mosquitoes ( Figure 5C). The results reveal that these genes are significantly up-regulated in the field-collected An. funestus s.s., in comparison with the susceptible laboratory strain FANG.

Contribution of Vectors to Malaria Transmission
The high number of An. gambiae and An. funestus infected with Plasmodium (37.6% and 30.9%, respectively) suggests a high burden of malaria in this location. These infection rates are far higher than those obtained by Coene in 1993 in urban and rural areas of Kinshasa using the enzyme-linked immunosorbent assay (ELISA) detection method [33]. However, it should be noted that we were detecting all stages of the development of Plasmodium, whereas ELISA detects only the infective sporozoite stage. Furthermore, the infection rates observed in Kinshasa are considerably higher than those recently reported in other locations using the same method (Benin [ [13,24,[34][35][36][37][38][39]. Our findings suggest that P. falciparum is the predominant malaria parasite in Kinshasa, although control and elimination efforts should not ignore P. malariae and P. ovale, present at low frequency, as previously reported [1,40].

Multiple Insecticide Resistance in Malaria Vector in Kinshasa
This study revealed a high frequency of resistance to multiple insecticide classes in An. gambiae and An. funestus which, together with their high level of Plasmodium infection rate, calls for urgent actions to be taken to control malaria in this region. Both malaria vectors exhibit resistance toward pyrethroids, the only insecticide class recommended for use on LLINs [41]. The insecticide resistance pattern observed in An. gambiae differs from that found in the locality of Kingasani (province of Kinshasa) in 2009, which showed higher mortalities [1,3]. This difference in resistance may be due to the effect of environmental and physiological factors such as the temporal selection caused by insecticide pressures, as well as to the genetic heterogeneity of the mosquito populations in DRC. The Kinshasa An. funestus population also exhibits resistance to pyrethroids and DDT, but at a lower level than An. gambiae. In contrast, An. funestus is susceptible to the carbamate bendiocarb. The resistance profile of this An. funestus population is similar to the situation observed in East Africa, where full susceptibility to bendiocarb is reported [7,18]. As observed in other African populations of malaria vectors, organophosphates are the only insecticide class to which there is full susceptibility, and these could be recommended for indoor residual spraying around Kinshasa.
The commonly used OlysetNet and Permanet 2.0 LLINs presented a very low bioefficacy under laboratory conditions. The low efficacy of OlysetNet, treated with permethrin only, is particularly evident against An. gambiae. This observation correlates well with the very high permethrin resistance observed for An. gambiae in Kinshasa. A previous assessment of efficacy of the OlysetNet in other parts of Kinshasa had reported a higher efficacy of the OlysetNet, suggesting either a heterogeneity of the response to this net or that resistance has worsened with time [5]. Even when the synergist PBO is combined with permethrin (OlysetPlus), the mortality of An. gambiae was low, indicating that P450 genes might not be the major drivers of the observed permethrin resistance, but rather the kdr mutation, which is nearly fixed in this population.

Molecular Mechanisms Involved in Insecticide Resistance
The extremely high resistance to permethrin and DDT in An. gambiae correlates with the high frequency of the L1014F kdr mutation (87.8%). This mutation has been previously detected in DRC [3,5]. Furthermore, the fact that 17% of mosquitoes possess the 1014S resistance allele further contributes to maintain the high resistance level. The likely independent selection of the 1014S mutation suggests that both migration and local selection forces are driving the development of pyrethroid/ DDT resistance in DRC. A de novo occurrence of kdr mutations has previously been described across Africa with two 1014F and two 1014S haplotypes reported [21]. Similar to the low efficacy of Olyset Plus, the lack of notable recovery after PBO exposure in the synergist assays further supports the assumption that cytochrome P450s only play a minor role in this resistance, with kdr mutations likely driving resistance to pyrethroids/DDT. Although both type I and II pyrethroids have the same molecular target, the VGSC, resistance to permethrin (type I) was higher than to deltamethrin (type II). Several studies based on modeling, electrophysiology, and in vivo bioassay analyses have shown that the different L1014 polymorphisms may present different responses to the pyrethroids type I and type II [42][43][44]. For An. gambiae, similar results have been reported in a population with a high 1014F frequency in Burkina Faso (West Africa) [45], but also in Tanzania and Uganda (East Africa), where the presence of L1014S is predominant [46,47]. Nevertheless, it is necessary to bear in mind that the differences in response to the type of pyrethroids may be also influenced by other minor mutations in the VGSC that may enhance the level of resistance associated with L1014 polymorphism [43,48].
In the absence of kdr mutations in An. funestus s.s. [49], this study explored the contribution of metabolic resistance in the multiple resistance observed. Previous reports have confirmed the important role of the duplicated CYP6P9a/CYP6P9b and CYP6M7 genes in pyrethroid resistance [30,31]. The significant up-regulation of these genes in this population supports the predominant role of metabolic enzymes in the pyrethroid resistance. Nevertheless, the FC values obtained here are considerably lower than in southern African populations [24], suggesting that mechanisms are different as recently shown between African regions for this species [50]. On the other hand, up-regulation for GSTe2 gene and the action of L119F-GSTe2 mutation have been associated mainly with DDT resistance [23]. The 119F-GSTe2 mutation, which has been shown to play a key role in DDT resistance in West and Central Africa (23), was found in high frequency, which is consistent with the DDT resistance observed in Kinshasa. Due to this, and along with the high frequency of kdr in An. gambiae, the use of DDT as an alternative to pyrethroids would not be recommended for vector control. Additionally, the presence of the 296S-RDL resistant to dieldrin allele is reported for the first time in DRC and could be a consequence of past use of this insecticide or ongoing use in the agricultural sector.

CONCLUSIONS
Besides the high Plasmodium infection rate in both An. funestus and An. gambiae, the results from this study reveal high and multiple insecticide resistance patterns, together with an alarmingly low efficacy of conventional LLINs without the synergist PBO under laboratory conditions. This study highlights the urgent need for actions to better manage the issue of insecticide resistance in order to prolong the effectiveness of the ongoing and future malaria control programs in DRC. Acknowledgments. John Gimnig is thanked for his critical reading of the manuscript.
Disclaimer. The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the Centers for Disease Control and Prevention or the institutions with which the authors are affiliated.
Financial support. This work was supported by a Wellcome Trust Senior Research Fellowship in Biomedical Sciences to C. S. W. (award number 101893/Z/13/Z). S. I. is funded by the US President's Malaria Initiative.
Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.