**4. Evaluation of fitness cost of insecticide resistance**

The main approach to investigate the fitness cost of resistance in field populations is to monitor the levels of resistance along the time in environments distinctly exposed to insecticides. Moreover, if the principal mechanism selected for resistance is known, the genotyping of resistance genes in place and time scales render important assumptions about their fitness cost. It is very difficult to access this kind of data from the field, however, since there are many variables occurring simultaneously.

For example, one population of *A. gambiae* from M'Bé, Côte D'Ivoire, used to be considered susceptible to most insecticides up to 2002, when a civil crisis broke and the monitoring was discontinued. Ten years later, a new study revealed important changes of the resistance mechanisms among *A. gambiae* populations from that locality. The main mechanisms that led them to become highly resistant to OC, PY, and CA were the L1014F *kdr* mutation and elevated activity of MFO and EST. The only well-known contexts that might explain this severe shift from susceptible to highly resistant were the pressure with deltamethrin-based products from rice paddles and the distribution of long-lasting PY impregnated nets (LLINs) since 2006 [56]. The alteration in the resistance profile over the time would suggest a low cost of the resistance alleles. However, little was known regarding the actual levels of insecticide pressure, migra‐ tion from vicinity areas, and about the extent of the influence of surrounding environment. In this case, controlled laboratory assays could help to estimate the fitness costs of the selected resistance mechanisms.

For fitness studies in the laboratory, population cage experiments can evaluate the fluctuation of resistance itself and the selected mechanisms over successive generations, under an environment clearly free of insecticide and without interference of migration. In this matter, the cost of resistance can be measured according to the velocity that the resistance alleles decrease in confined lineages along the time. A laboratory lineage of *A. aegypti* resistant to PY due to the NaVR2 *kdr* mutation4 presented deleterious effects in a series of life-trait parameters. Population cage assays corroborated these negative costs, showing that the *kdr* allele severely decreased from 75% to almost zero along 15 generations [57]. Most of the studies have been making use of an opposite direction: populations from the field are confined and submitted to a selection pressure in the laboratory. In another example, also with *A. aegypti*, populations

<sup>4</sup> NaVR2 is the kdr allele mutant in both 1016 (Val to Ile) and 1534 (Phe to Cys) of the voltage gated sodium channel (NaV), found in American populations of *A. aegypti*.

from distinct Mexican localities were pressured with the PY permethrin for at least five generations in the laboratory. All the lineages had an increase in the resistance levels, corre‐ lated with an augment in the frequency of the Val1016Ile *kdr* mutation and with a number of detoxifying genes differentially transcribed, generally distinct at each lineage. Interestingly, the lineages that reached the highest frequencies of the *kdr* mutation presented a lower number of altered detoxifying genes [58]. These results strongly suggested that this *kdr* mutation had a lower fitness cost compared to the metabolic resistance genes occurring at each genetic background.

One has to consider that the evaluation of the overall fitness effects of a given mutation is very challenging, once it is difficult to separate their own effects from those caused by other mechanisms possibly coselected for resistance. In these aforementioned *rdl* examples, the reduced fitness might be related to the A302S mutation itself, and/or to metabolic resistance mechanisms. Similarly, the persistence of the resistance allele in an environment free of dyeldrin might be explained by the rdl cross-resistance with other insecticide that had been continually applied, as well as by the selection of modifiers genes, as previously discussed.

The main approach to investigate the fitness cost of resistance in field populations is to monitor the levels of resistance along the time in environments distinctly exposed to insecticides. Moreover, if the principal mechanism selected for resistance is known, the genotyping of resistance genes in place and time scales render important assumptions about their fitness cost. It is very difficult to access this kind of data from the field, however, since there are many

For example, one population of *A. gambiae* from M'Bé, Côte D'Ivoire, used to be considered susceptible to most insecticides up to 2002, when a civil crisis broke and the monitoring was discontinued. Ten years later, a new study revealed important changes of the resistance mechanisms among *A. gambiae* populations from that locality. The main mechanisms that led them to become highly resistant to OC, PY, and CA were the L1014F *kdr* mutation and elevated activity of MFO and EST. The only well-known contexts that might explain this severe shift from susceptible to highly resistant were the pressure with deltamethrin-based products from rice paddles and the distribution of long-lasting PY impregnated nets (LLINs) since 2006 [56]. The alteration in the resistance profile over the time would suggest a low cost of the resistance alleles. However, little was known regarding the actual levels of insecticide pressure, migra‐ tion from vicinity areas, and about the extent of the influence of surrounding environment. In this case, controlled laboratory assays could help to estimate the fitness costs of the selected

For fitness studies in the laboratory, population cage experiments can evaluate the fluctuation of resistance itself and the selected mechanisms over successive generations, under an environment clearly free of insecticide and without interference of migration. In this matter, the cost of resistance can be measured according to the velocity that the resistance alleles decrease in confined lineages along the time. A laboratory lineage of *A. aegypti* resistant to PY

Population cage assays corroborated these negative costs, showing that the *kdr* allele severely decreased from 75% to almost zero along 15 generations [57]. Most of the studies have been making use of an opposite direction: populations from the field are confined and submitted to a selection pressure in the laboratory. In another example, also with *A. aegypti*, populations

4 NaVR2 is the kdr allele mutant in both 1016 (Val to Ile) and 1534 (Phe to Cys) of the voltage gated sodium channel

presented deleterious effects in a series of life-trait parameters.

**4. Evaluation of fitness cost of insecticide resistance**

variables occurring simultaneously.

250 Insecticides Resistance

resistance mechanisms.

due to the NaVR2 *kdr* mutation4

(NaV), found in American populations of *A. aegypti*.

The knowledge of alterations in physiological and reproductive aspects is generally achieved by comparing life-trait parameters between susceptible and resistant individuals. As the result of pleiotropic effects of an altered gene will depend on the whole genomic structure, it is important that susceptible and resistant groups have the most similar genetic background as possible. The parameters usually evaluated are larval developmental time, adult longevity, ability to avoid predators, fecundity, fertility, mating competitiveness, and reproductive potential. When treating of blood-sucking insects, probing, acceptance of blood meal, and amount of ingested blood can also be tested. Such studies demand well-controlled conditions and are generally highly laborious, so that most of them follow few parameters at a time. In addition, the knowledge of the biology of the species under investigation is a prerequisite for the definition of which aspects would be more informative.

Fitness studies in the laboratory necessarily have to consider a well-representative collection from the field, in order to contemplate most part of the whole amplitude of variable traits from the original population. An F1 offspring of this sampling may then be raised in the laboratory to sufficiently amplify the number of individuals to be tested, as well as to normalize the physiological condition among the different populations. A laboratory lineage control of susceptibility and vigor should also be raised in parallel, as an endogenous control of experi‐ mental conditions, whenever possible.

Selection pressure for insecticide resistance in the laboratory has the advantage of controlled strength of selection and environmental conditions, population size, and absence of migration. On the other hand, if a monogenic key-mechanism for resistance was under lower frequency in the field, it is likely that this gene is not present in the sampling that established the first generation in the laboratory. For this reason laboratory pressures tend to result in polygenic resistance, where several resistance traits of minor effects are selected, but with a larger response when emerged together in the same genetic background [30]. This could also explain the different patterns of selected mechanisms to the same class of insecticide, especially metabolic resistance, in a same species.

Another important issue to be aware of when evaluating fitness costs in the laboratory environment is that most of the studies have investigated the possible life-trait alterations under optimal conditions. The amount and quality of food, the composition of substrate (or water in case of aquatic insects), density of individuals along life cycle phases, and mainly temperature and humidity are usually controlled. By contrast, insects are continually exposed to a wide range of abiotic or biotic stresses in the field. Therefore, the physiological costs of resistance alleles are probably underestimated in laboratory optimal conditions [41, 59]. The

evaluation of the fitness costs in resistant insects under stress conditions (in terms of nutrition, temperature, and larval density, for instance) can bring forth relevant data related to the evolution of resistance in the field. However, such investigations are still scarce [60-62].
