Ethological and Physiological Harmful Effects of Metformin, An Antidiabetic Used to Treat Type 2-Diabetes: A Study on Ants as Biological Models

Metformin is a drug mainly used for caring of persons suffering from type 2-diabetes. Over time, it was found to be efficient for treating others illness. Its use increasing, it is nowadays the second active pharmaceutical ingredient more present pollutant in natural water. It could therefore affect the biology of the aquatic fauna through harmful physiological effects. Using ants as biological models, we studied the potential physiological and ethological adverse effects of Metformin. We found that it largely reduced or impacted these insects’ food intake, activity, audacity, social relationships, state of stress, cognition, and learning abilities. No adaptation and no habituation to the effects of Metformin were observed, and ants developed some dependence on its consumption. After weaning, the effect of Metformin became significantly lower than its initial one as soon as after four hours and fully vanished in a total of 13 hours. Metformin could thus harmfully impact the freshwater fauna rich in insect species, especially if chronic exposures occur. As regards patients treated with Metformin, our study suggests that they may suffer from side effects not mentioned in the drug notice. For instance, they may develop dependence and increase their daily dose, accentuating so the drug side effects, e.g., they may suffer from anorexia. Practitioners should know the side effects of Metformin and monitor patients as for their occurrence.


Pollution of Freshwater Ecosystems by Pharmaceutical Products
Pollutions of freshwater ecosystems by pharmaceutical products and active pharmaceutical ingredients (APIs) occur in all countries in the world regardless of the level of GDP or the specificities of the economy (Wilkinson et al., 2022). Numerous products used or consumed by humans are present in the environment and all ecosystems (Arnold et al., 2014), mainly in aquatic ecosystems (freshwater but also in coastal ecosystems (Prichard and Granek, 2016)). These products affect all species and ecological processes (Richmond et al., 2017), including in the most remote areas (Krief et al., 2022). As it has been shown that these APIs can have (potentially in some case study) harmful impacts on ecosystems health (Richmond et al., 2017;Grenni et al., 2017;Prichard andGranek, 2016, Arnold et al., 2013), the effects of these substances on aquatic organisms and ecological processes are a worldwide subject of concern (Wilkinson et al., 2022;EEB, 2018 and references therein;Jones et al., 2004).
Investigations on potential harmful effects of APIs on freshwater organisms are not easy to conduct, due to experimental constraints and logistical challenges of this type of study, although if small-scale (microcosm) experiments using, for example, daphnia (Damasceno de Oliveira et al, 2015;Kim et al., 2007), or other freshwater invertebrates, such as mussels (Aguirre-Martí nez et al., 2013) are common. A growing number of mid-scale experiments (using mesocosm experimental designs) are being carried out (Bartmentlo et al., 2021), but they remain very expensive, time-consuming and difficult to replicate.
If physiological impacts of APIs can be quite well evaluated through the use of microcosm experimental designs, it is often more complicated to investigate ethological effects of drugs on freshwater vertebrates and invertebrates (Brodin et al., 2014). The majority of the latter ones are insects, and some of them display complex, even almost social, behaviors. This is the case for some caddis flies or stoneflies for instance (Johnstone, 2009;Hanada et al., 1994;Stewart & Zeigler, 1984). It is therefore relevant to rely on a biological model organism easy to experiment in order to bio monitor harmful effects of drugs and APIs on ethological and physiological traits of freshwater invertebrates. To carry out this type of study, we used a species of ants, Myrmica sabuleti Meinert, 1861, we have previously used in the course of many experimental works during which we examined the impact of APIs and drugs on several ethological and physiological traits. For instance, we so studied the effects of some drugs that are largely distributed in freshwater ecosystems, namely fluoxetine (Cammaerts & Cammaerts D, 2015a), carbamazepine (Cammaerts & Cammaerts D., 2015b), paroxetine (Cammaerts & Cammaerts, 2016), ibuprofen (Cammaerts &Cammaerts, 2018), and fluvoxamine (Cammaerts & Cammaerts, 2021). We present here the result of a study dedicated to the second most common drug found worldwide in freshwater ecosystems: Metformin, a drug largely used for caring of patients with type 2-diabetes, and the use of which nowadays increases.

Metformin, the Second More Frequent Polluting API in Freshwater Worldwide
Metformin, the complete name being Metforminhydrochlorid, the chemical structure of which is given in Figure 1, is a hypo-or normo-glycerinate, i.e., a drug which reduces the amount of sugar in blood (Silvio et al, 1998;Ferrannini, 2014). Its mode of action is not yet fully elucidated. It does not increase the production of insulin, but increases the insulin sensitivity (in other words, it reduces the insulin resistance) of muscles and fat tissues (Silvio et al, 1998). Metformin also reduces the neoglucogenesis, inhibiting the glycerophosphate dehydrogenase, and decreases the absorption of glucose (Mdiraju et al, 2014). In addition, it inhibits the propagation of glucagon into the blood (Silvio et al., 1998). Over time, practitioners and researchers have found other properties of Metformin. It is sometimes used for treating polycystic ovary syndrome (Lord et al., 2003) but the effect seems to be very limited (Legro et al., 2007). It may be efficient for preventing and/or reducing some cancers (Ben Sahra et al., 2010). Metformin has also been shown to be efficient for losing weight (Frieling et al., 2021). However, though being a useful drug, Metformin appears to present some side effects, the most divulgated ones (in the notice joined to the drug package, and on several internet links, e.g. https://www.revmed.ch › revue-medicale-suisse-394 › t...) being digestive problems, decrease of the amount of vitamin B12, metallic taste in the mouth, anorexia, and kidney dysfunction (Wei-Hao et al. 2018).
Since over time Metformin was found to be efficient for treating other illness and symptoms than type 2-diabetes, it progressively became more and more used. This drug is nowadays the secondly most present API pollutant in the natural water, the first most present one being Carbamazepin (Wilkinson et al., 2022). Metformin is for instance widely distributed in the German water cycle (Trautwein et al., 2014) and in south-eastern United States wadable streams (Bradley et al., 2016). In drinking water system treatment plants, Metformin is transformed thanks to chlorination process. The chlorination by-products of Metformin have led to serious problems as for human's and animal's health due to direct toxicity and cell tissues damages (Zhang et al., 2021). These products also cause endocrine disruption in reproductive tissues of fishes, leading to the development of intersex gonads in males, and reduction of fecundity in treated pairs of fishes (Niemuth and Klaper, 2015). Some negative effects of Metformin on photosynthetic activity of the freshwater chlorophyte, Chlorella vulgaris have also been reported (Cummings et al., 2018). Nevertheless, another study revealed no physiological consequence on the health of a freshwater gastropod -the big ramshorn snail (Planorbarius corneus) -due to an exposition to Metformin and to its transformed product guanylurea at environmentally relevant concentrations (Jacob et al., 2019).

Aim of the Study
We aimed to know which kind of harmful effects Metformin can cause to animal species and to ecological process when released in the environment. Alongside to potential physiological disorders, Metformin appears to also cause some ethological impairments to freshwater invertebrates (Godoy et al 2018). As the latter include numerous insect species, we aimed to study potential ethological effects of Metformin on terrestrial insects, what is easier than on aquatic ones. More precisely, we intended to make our research on our usual biological model, the ant M. sabuleti, which is very suitable to do so for several reasons reported in the subsections 1.4 and 2.1. To summarize, in the present paper, we report our study of some potential harmful ethological and physiological effects of Metformin on an ant species used as a biological model organism. In the subsection 2.2., we precise the ethological and physiological traits on which we focused our investigation. In total, 13 biological traits were considered, from food intake to dependence on the drug consumption and to the decrease of its effect after its consumption was stopped.

Ants as Biological Organism Models
Fundamental biological processes, such as embryology, proteins synthesis, nervous functioning, muscular functioning, conditioning acquisition, are similar in most animal species including humans. Therefore, several animal species are used as biological models for physiological and ethological studies (Wehner & Gehring, 1999;Russel & Burch, 2014). Invertebrates are generally preferred because they have a small size, can be easily maintained out of their natural environment, and have a short generation time (Wolf & Heberlein, 2003). Insects are often used, e.g., locusts, mealworms, fruit flies, bees (Andre et al., 1989). Ants can thus be used, the more so since they can easily be maintained in any room, at low cost, and since they detain a lot of evolved skills on which the effect of events or substances can be studied (Passera & Aron, 2005). They can memorize visual and olfactory cues and use them for navigating; they use pheromones to communicate with congeners; they are able to rapidly recruit congeners for killing enemies, collecting food, hunting for prey; they differently mark zones of their territory. They take care of their brood and queens; they build complex nests; they clean their nests and manage cemeteries at the limits of their territory.

Which Species We Used and What We Know on It
We worked on the species Myrmica sabuleti Meinert, 1861. We know rather well its biology as regards to physiological, ethological and cognitive traits. Indeed, we have studied its visual perception, conditioning acquisition, recruitment system (Cammaerts & Cammaerts D., 2014), ontogenesis of some of their skills , their recognition in a mirror , as well as several of their numerosity abilities (Cammaerts & Cammaerts, 2020a, 2020b, 2022. For example, we showed that they have a number line, can acquire the notion of zero, can count and add numbers of elements, and can expect the next element of an increasing or decreasing arithmetic or geometric sequence. The distance effect, the size effect and Weber's law can be applied to their perception (Cammaerts & Cammaerts, 2020c, 2020d. The physiological and ethological effects of drugs and APIs can thus valuably be examined on such a species detaining so many sophisticated skills, even if the latter always stay at a concrete level. Effectively, until now, we have successively examined on the workers of the ant M. sabuleti ants the side effects of 54 products or situations (e.g., Cammaerts, 2016Cammaerts, , 2017Cammaerts, , 2018aCammaerts, , 2019Cammaerts, , 2021aCammaerts, , 2022.

Which Traits We Intended to Examine
Here we aimed to study the adverse effects of Metformin on M. sabuleti workers' meat and sugar water consumption, general activity, linear and angular speeds, orientation ability, audacity, tactile perception, brood caring behavior, congeners' relationships, stress, cognition, learning and memory, adaptation to these side effects, and dependence on its consumption. We also aimed to precise the manner according to which Metformin loses its effect after its consumption was stopped. The materials and methods were similar to those used for conducting our previous works (e.g., Cammaerts, 2021b, c;Cammaerts & Cammaerts D, 2015b, c;Cammaerts & Cammaerts R, 2016, 2020e, 2021. For the readers' convenience, we however briefly related them, but without avoiding some inevitable self-plagiarism.

Collection and Maintenance of Ants
The present experimental work was conducted on two colonies of M. sabuleti collected since a year in Ardenne (Belgium) from an old quarry situated near the Aise. These colonies contained 500-600 workers, 1-2 queens, and brood. Each one was maintained in one to three glass tubes half-filled with water, a cotton plug separating the ants from the water. The nest tubes of each colony were set in a tray (34 cm x 23 cm x 4 cm), the borders of which having been covered with talc to prevent ants from escaping. These trays served as foraging areas. In them, pieces of Tenebrio molitor larvae (Linnaeus, 1758) were provided three times per week, and a cotton-plugged tube filled with sugar (~ 15% of sugar) water was permanently set. The lighting of the laboratory varied between 330 and 110 lux, the ambient temperature constantly equaled 20-21°C, the humidity 80%, and the electromagnetic field 2 μWm2. These conditions are favorable to M. sabuleti. The word 'ant' is here often replaced by 'worker' or 'nestmate' as do researchers on social insects.

Solution of Metformin Given to the Ants
One tablet of Metformin 500 mg (Metformin chlohydrate, Sandoz®; autorisation: viatris sante, 1 rue de Turin, 69007 Lyon; manufacturer: MC Dermott Laboratories T/A Gerard Laboratories, Dublin, Ireland) was furnished by the pharmacist Wera (1170 Bruxelles). People treated with this drug are advised to consume minimally one tablet of 500 mg per day. As most mammals, they consume a rather large amount of water. Generally, one human consumes about at minimally one litter of water (not including the water contained in his food) per day. Insects, and thus ants, due to their anatomy and physiology (cuticle, excretory apparatus) generally drink ca 10 times less water. Therefore, for living under a diet with Metformin similar to that of people treated with this dug, the ants must be provided with a solution of 1 tablet 500 mg Metformin in 100 ml water. We thus duly crushed such a tablet of Metformin and dissolved the obtained powder into 100 ml of the sugar water provided to the ants ( Figure 2). The obtained solution was delivered to the ants in their usual cotton-plugged tubes. The plugs of these tubes were refreshed every 2-3 days, and the entire content of the tubes was renewed every seven days. Several times per day, we checked if ants drunk the delivered solution of Metformin, and they did. All the control experiments were first conducted on the two colonies normally maintained. Then, the tubes filled with sugar water were replaced by those filled with the sugared solution of Metformin, and the test experiments started after the ants had the latter tubes at their disposal for one day. On ants living under normal diet, then while they consumed Metformin, we separately counted those sighted on the meat food, sighted at the entrance of the sugar water tube, and being active in the foraging area, near the food, at the nest entrance and inside the nest, all this twice during the day and twice during the night (n° of counts = 4 x 2 colonies = 8 counts per day for each kind of count). Such counting was made during 6 successive days. For each kind of count and each kind of diet, the daily mean was established (Table 1, lines 1 -6). For each kind of considered trait, these six mean values obtained for ants intaking Metformin were compared to the six mean values recorded for ants normally maintained by using the non-parametric test of Wilcoxon (Siegel & Castellan, 1988). In addition, for each kind of diet and of count, the mean of the six daily means was calculated (Table 1, last line I-VI). Note that, each time the Wilcoxon test was used in this work, we gave, in the results section, the values of N, T, P as defined by Siegel and Castellan (1988).

Linear and Angular Speeds, Orientation
These three traits were quantified as usually (Cammmaerts, 2021bCammaerts & Cammaerts, 2016, 2020e, 2021 on ants walking in their foraging area, the speeds without stimulating the ants, the orientation when stimulating them with a nestmate tied to a piece of paper (Figure 3 a). Such a tied nestmate emits its attractive alarm mandible glands pheromone, attracting so the ants located at 1 to 8 cm from it. For quantifying the ants' speeds, then their orientation, 40 ant trajectories were collected and treated with adequate software (Cammaerts et al., 2012). The latter assessed the three required variables on the basis of the three following criteria. An individual's linear speed (assessed for instance in millimeter per second = mm/s) equals the length of its trajectory divided by the time it spend to travel it; its sinuosity (quantified for instance in angular degrees per centimeter = ang.deg./cm) is the sum of the angles made by the successive parts of its trajectory divided by the length of this trajectory; its orientation (quantified for instance in angular degrees = ang. deg.) towards a given point is the sum of the successive angles it makes with its own direction and that to the given point, divided by the number of measured angles. An orientation value lower than 90° means that the animal tends to orient itself toward the location; an orientation value higher than 90° means that the animal tends to avoid the location. For the three kinds of quantified variable, the median and the quartiles of the 40 recorded values were established (Table 2, lines 1, 2, 3), and the distribution of values obtained for ants consuming Metformin was compared to that obtained for ants living under normal diet using the non-parametric χ² test (Siegel & Castellan, 1988).

Audacity
This trait was studied, as usually (Cammmaerts, 2021b, c: Cammaerts & Cammaerts, 2015a, b;Cammaerts & Cammaerts, 2016, 2020e, 2021 by quantifying the ants' tendency to come onto an unknown apparatus. To do so, a cylinder (height = 4 cm; diameter = 1.5 cm) orthogonally attached to a square (9 cm 2 ), both of Steinbach® white paper, was deposited in the foraging area of each colony, and the ants coming onto this apparatus were counted 10 times over 10 minutes (Figure 3 b). The mean and the extremes of these counts were established (Table 2, line 4). The numbers obtained for the two colonies were correspondingly added. Then, the obtained sums relative to every two successive minutes were added, what provided five successive sums. The five sums obtained for ants consuming Metformin were compared to those obtained for ants under normal diet using the non-parametric Wilcoxon test (Siegel & Castellan, 1988).

Tactile Perception
An ant perceiving the uncomfortable character of a substrate moves on it slowly, sinuously, and often touches the substrate with its antennae (Figure 3 c1). An ant unable to correctly perceive the rough character of a substrate walks on it rather easily, rapidly, not vey sinuously, and seldom touches the substrate with its antennae. Therefore, for quantifying the ants' tactile perception, we assessed their linear and angular speeds while they walked on an uncomfortable substrate exactly as we assessed these two traits on ants moving in their foraging area (see the subsection relative to the ants' speeds). For making such an assessment, as in previous works (Cammmaerts, 2021bCammaerts & Cammaerts, 2016, 2020e, 2021, a piece (3 cm x 2 + 7 + 2 = 11 cm) of n° 280 emery paper was inserted in a tray (15 cm x 7 cm x 4.5 cm) in order to divide this tray in a first small 3 cm long zone, a second 3 cm long one covered with the emery paper, and a third 9 cm long zone. For each colony, 12 ants were transferred in the first small zone of the tray, and their trajectories were recorded while they walked on the emery paper. The ants' linear and angular speeds were then quantified. The 24 recorded values of linear and of angular speeds were characterized by their median and quartiles (Table 2, lines 5, 6). The distributions of the values of these two variables obtained for ants consuming Metformin were compared to those obtained for ants normally maintained using the non-parametric χ² test (Siegel & Castellan, 1988).

Brood Caring Behavior
This social task, which occurs in the nest, was evaluated through the ants' behavior in front of larvae experimentally removed from the nest. Normally, the ants soon re-entered such larvae. To make the required evaluation, as in e.g., Cammmaerts, 2021b, c: Cammaerts and Cammaerts, 2015a, b;Cammaerts and Cammaerts, 2016, 2020e, 2021, for each colony, a few larvae were taken out of the nest and set in front of its entrance. For each colony, the workers' behavior towards five of these larvae was observed during five minutes, and the not re-entered ones were counted after 30 seconds, 1, 2, 3, 4, 5 minutes (n° of counts = 5 x 2 =10) ( Table 3, line 1; Figure 3 d). We looked only at five larvae for each colony because we had to look at all of them at the same time. The experiment was made only once because removing brood from the nest perturbs the ants and imperils the larvae's survival. The six numbers obtained for the two colonies were correspondingly added, and the six sums obtained for ants consuming Metformin were compared to those obtained for ants normally maintained using the non-parametric test of Wilcoxon (Siegel & Castellan, 1988).

Social Relationships
Ants belonging to the same colony are not aggressive towards each other. Several factors may alter this peaceful social behavior. For examining if Metformin has such an impact, as in e.g., Cammmaerts, 2021b, c: Cammaerts and Cammaerts, 2015a, b;Cammaerts and Cammaerts, 2016, 2020e, 2021, five dyadic encounters were performed for each colony in a small cup (diameter = 2 cm, height = 1.6 cm) the borders of which having been slightly covered with talc to prevent ants from escaping. During each encounter, one ant of the pair was observed during 5 minutes, and we recorded its amounts of 'level 0 = doing nothing', 'level 1 = making antennal contacts with its congeners', 'level 2 = opening its mandibles', 'level 3 = gripping its congener', and 'level 4 = trying to sting or stinging its congeners' (Table 3, line 2, Figure 3 e). The numbers recorded for each ant and each colony were correspondingly added, and the distribution of values obtained for ants consuming Metformin was compared to that obtained for ants maintained under normal conditions using the non-parametric χ² test (Siegel & Castellan, 1988). For each kind of diet, the ants' social interactions were also assessed by a variable 'a' equaling the number of aggressiveness levels 2 + 3 + 4 divided by the number of aggressive levels 0 + 1 (Table 3, line 2).

Stress and Cognition
Escaping from an enclosure requires staying calm, not stressing, meticulously looking for an exit, and detaining some cognitive ability. Therefore, to assess the state of stress and the cognitive ability of the ants, for each colony, six ones were enclosed under a reversed polyacetate cup (height = 8cm, bottom diameter = 7 cm, ceiling diameter = 5 cm, the inside surface of the cup having been slightly covered with talc) set in their foraging area. This protocol has been often used (e.g., Cammmaerts, 2021b, c: Cammaerts & Cammaerts, 2015a, b;Cammaerts & Cammaerts, 2016, 2020e, 2021. A notch (3 mm height, 2 mm broad) was created in the bottom rim of the enclosure for allowing the ants to escape (Figure 3 f). For each colony, the numbers of ants which could escape after 2, 4, 6, 8, 10 and 12 minutes were recorded, then the numbers obtained for the two colonies were correspondingly added (Table 3, line 3). The six sums obtained for ants consuming Metformin were compared to those obtained for ants normally maintained using the non-parametric Wilcoxon test (Siegel & Castellan, 1988 & Cammaerts, 2016& Cammaerts, , 2018& Cammaerts, , 2020e, 2021, through the ants' skill in crossing a twists and turns path. For each colony, two pieces of Steinbach® paper (12 cm x 4.5 cm) duly folded were inserted into a tray (15 cm x 7 cm x 4.5 cm) in order to create a twists and turns path between a 2 cm long zone in front of this path and an 8 cm long zone beyond it. For making an experiment on a colony, 15 ants were deposited into the zone lying in front of the twists and turns path, and the ants still there as well as those having reached the zone lying beyond the 'difficult' path were counted after 2, 4, 6, 8, 10 and 12 minutes (Figure 4 a, Table 3, line 4). The numbers obtained for the two colonies were correspondingly added, and for each zone of the design, the numbers obtained for ants consuming Metformin were compared to those obtained for ants normally maintained using the non-parametric Wilcoxon test (Siegel & Castellan, 1988).

Conditioning Acquisition, Memory
We proceeded as previously (e.g., Cammmaerts, 2021b, c: Cammaerts & Cammaerts, 2015a, b;Cammaerts & Cammaerts, 2016, 2020e, 2021. For each colony, at a recorded time, a green hollow cube (constructed in strong paper (Canson®)) was deposited above the entrance of the tube filled of sugared solution of Metformin and the meat food was relocated near that cube (Figure 4 b a). Since this deposit, the ants underwent operant visual conditioning. The control experiment on ants normally maintained, was previously performed on another similar colony of M. sabuleti collected in the same site. Such a doing was required because since a worker is conditioned to a stimulus, it keeps its conditioning during several days, and even after having lost it, it acquires it again more rapidly than initially, and its conditioning acquisition can no longer be validly assessed. Over the ants' conditioning acquisition, then after the green cube removal, over the ants' loss of conditioning, 10 workers of each colony were tested in a Y-apparatus. A Y-maze was built for each colony in strong white paper, with its sides slightly covered with talc, and was set in a separated tray (15 cm ×7 cm × 5cm). A green hollow cube was deposited randomly in the left or the right branch of these Y-apparatus. For making a test on a colony, 10 workers were transferred one by one in the area lying in front of the Y-maze division into its two branches. The first choice made by each experimented ant between one or the other branch of the Y-apparatus was recorded (Figure 4 b b). Of course, choosing the branch containing the green cube was considered as giving the correct response. After having been tested, each ant was kept in a cup, until 10 ants of its colony were tested to avoid testing twice the same ant. After having tested 10 ants of a colony, all of them were transferred back into their foraging area. For each considered time period (Table 4), the responses obtained for the two experimented colonies were correspondingly added. The 10 control responses and the 10 summed responses obtained for ants consuming Metformin were compared to each other using the non-parametric test of Wilcoxon (Siegel & Castellan, 1988). Also, for each considered time period and each diet, the ants' conditioning score (i.e., the proportion of correct responses) was established (Table 4).

Adaptation to a Side Effect of Metformin
The reasoning is identical to that used in previous works (e.g., Cammaerts, 2021bCammaerts & Cammaerts, 2016, 2020e, 2021). An individual adapts itself to the side effects of a drug when, over this drug consumption, it less and less suffers from these side effects. To study such an adaptation, a trait affected by the drug has to be assessed soon after the start of the consumption, then again after some time of such a consumption, and the results of the two assessments must be compared. In the present work, Metformin appeared to affect the ants' locomotion. Therefore, the ants' linear and angular speeds were assessed after the ants had the drug at their disposal during seven days, exactly as they had been after one day, and the distribution of the values each time recorded were compared to one another using the non-parametric χ² test (Siegel & Castellan, 1988). Also, for each kind of speed, the median and quartiles of the values recorded after seven days of ants' maintenance under a diet with Metformin were established as they had been after one day of such a maintenance (Table 5, upper part).

Habituation to a Wanted Effect of Metformin
The reasoning is identical to that used in previous works (e.g., Cammmaerts, 2021b, c: Cammaerts & Cammaerts, 2015a, b;Cammaerts & Cammaerts, 2016, 2020e, 2021). An individual becomes habituated to an expected effect of a drug when it less and less perceives this effect over its drug consumption. To evaluate such a habituation, an expected effect of the drug has to be assessed soon after the individual begins consuming the drug, and then after it had consumed the drug during some time, and the two assessments must be compared. In the present work, the ants' less intake of sugar water may be an expected effect of Metformin. Consequently, to evaluate the ants' habituation to Metformin, their sugar water intake was assessed after they had that drug at their disposal during eight days, exactly as it had been assessed during the first six days of their maintenance under a diet with Metformin, except that the six assessments were made over 24 hours (6 x 8 counts) instead of over six days. The values of the six assessments made after the ants had the drug at their disposal for eight days were compared to the six values of the assessments performed over the six first days of the ants' drug ijb.ccsenet.org International Journal of Biology Vol. 14, No. 2;2022 consumption, using the non-parametric test of Wilcoxon (Siegel & Castellan, 1988). Also, the mean of the last six assessments was established as had been the mean of the first six assessments (Table 5, middle part).

Dependence on Metformin Consumption
An individual develops dependence on a drug when it "enjoys" consuming this drug, tries to have it at its disposal at any time, consumes the drug even if it suffers from some adverse effects, and becomes unable to live without consuming the drug. In the present work, the ants' dependence on Metformin was studied after they had this drug at their disposal during 9 days, this using a protocol already often used (e.g., Cammmaerts, 2021b, c: Cammaerts & Cammaerts, 2015a, b;Cammaerts & Cammaerts, 2016, 2020e, 2021. The cotton plug of the tubes containing the sugared solution of Metformin was not refreshed on the tenth day, and at the end of that day, 15 ants of each two colonies were deposited in an own tray (15 cm ×7 cm × 5 cm) containing two cotton-plugged small tubes (h = 2.5 cm, diam. = 0.5 cm), one full of sugar water, the other full of the sugared solution of Metformin used all over the work. The tube containing the drug was located on the right in one tray and on the left in the other tray (Figure 4 e). This having been done, for each sample of 15 ants of each colony, those present near each two presented tubes were counted 15 times over 15 minutes, and the 15 counts obtained for each colony were correspondingly added. Also, the 15 added counts were summed for each kind of tube (i.e., of each solution) what allowed establishing the proportions of ants having visited one and the other tubes (Table 5, lower part). In addition, the sums of the 15 added counts corresponding to each presented tubes were compared to the numbers expected if ants randomly approached each tube, using an adequate non-parametric test, i.e., the χ² goodness-of-fit one (Siegel & Castellan, 1988).

Decrease of the effect of Metformin After Its Consumption Was Stopped
This decrease was studied after the ants had consumed Metformin during 10 days, and the trait used for conducting this study was the ants' angular speed which was largely impacted by the drug consumption. The protocol was similar to that already often used (e.g., Cammmaerts, 2021b, c: Cammaerts & Cammaerts, 2015a, b;Cammaerts & Cammaerts, 2016, 2020e, 2021. One day before starting the study, a fresh sugared solution of Metformin was delivered to the ants, and the day after, the ants' angular speed was assessed as it had been assessed after 1 and 7 days of this drug consumption, except that 20 instead of 40 ants' trajectories were recorded and analyzed. This sample reduction was done for having time enough to analyze the recorded data all along the decrease of the effect of the drug, and so to evaluate the level of the decay of the impact of Metformin on the ants' angular speed. Just after this first assessment made at t = 0, the weaning started: the ants' tubes full of the sugared solution of Metformin were replaced by tubes full of pure sugared water. From then on, the ants' sinuosity was quantified each two hours until this trait value was identical to that obtained for ants normally maintained (= to the control values). For each quantification, the median and quartiles of the 20 recorded values of sinuosity were established (Table 6). Also, each successively obtained distribution of values was compared to that obtained at t = 0 and to the control one using the non-parametric χ² test (Siegel & Castellan, 1988). In addition, the mathematical function best describing the decrease of the effect of Metformin on the ants' angular speed was empirically sought and statistically analyzed, and is given in the text. The result of the present study is illustrated in Figure 5.

Food Consumption, General Activity
These physiological traits were impacted by Metformin consumption (Table 1). Under that drug diet, the ants eat less meat (N = 6, T = 21, P = 0.016), drunk less sugar water (N = 6, T = 21, P = 0.016), and were less active (N = 6, T = 19, P = 0.047) than when living under normal diet. The most impacted trait was the sugar water intake; the less impacted trait was the individuals' activity. These effects of Metformin should be taken into account when caring of patients with this drug, although eating less, and essentially intaking less sugar, may be one of the expected and wanted effects of the drug.

Linear and Angular Speeds
The ants' locomotion was affected by Metformin consumption (Table 2, lines 1, 2). While consuming this drug, the ants walked more slowly (χ² = 28.99, df = 2, P < 0.001) and essentially more sinuously (χ² = 59.02, df = 2, P < 0.001) than when living under normal diet. This was very obvious to observers: the walking ants continuously changed of direction, and made some abnormal movements with their legs and their body. Such a potential effect of Metformin on the individuals' displacements should be considered when caring of humans with this drug.

Orientation
While consuming Metformin, the ants less well oriented themselves towards a tied nestmate than when living under normal diet (Table 2, line 3; Figure 3 a). This was obvious to observers and was statistically significant (χ² = 24.04, df = 2, P < 0.001). The experiment was repeated, and a median of 53.2 angular degrees was obtained, confirming thus the results of the initial experiment. This decrease of orientation ability may be due to the ants' large sinuosity of movement (see the above subsection), but may also result from a weaker ants' olfactory or general sensory perception, a hypothesis examined thanks to a following experiment (see below the subsection devoted to the ants' tactile perception).

Audacity
Metformin impacted the ants' tendency to come onto the presented unknown apparatus (Table 2, line 4; Figure 3 b). While ants under normal diet rather frankly came onto such an apparatus, those consuming the drug were reluctant to do so. They hesitated to come on the apparatus; if they did so, they soon went away from it; if they tempted to climb on the tower, they soon went back or even felt down. Such an observation was in agreement with that made for the ants' locomotion (see the above subsection relative to the ants' linear and angular speeds). The numbers of ants sighted on the apparatus over the experimental time statistically differed between the ants consuming Metformin and those normally maintained (N = 5, T = 15, P = 0.031). Such an impact of the drug should be considered when caring of patients thanks to Metformin.

Tactile Perception
This important physiological trait was not affected by Metformin consumption (Table 2, lines 5, 6; Figure 3 c). When ants under normal diet walked on a rough substrate, their linear speed was smaller and their angular speed was higher than when they walked on their foraging area, and these differences were statistically significant (linear speed as well as ijb.ccsenet.org International Journal of Biology Vol. 14, No. 2;2022 angular speeds: χ² = 64.00, df = 1, P < 0.001). Such differences also occurred for ants consuming Metformin, the differences being simply somewhat smaller because these ants' linear speed was already smaller and their angular speed already larger when they walked in their foraging area. The differences between their walking on one and the other kinds of substrate were however highly significant: linear speed: χ² = 23.39, df = 1, P < 0.001; angular speed: χ² = 29.57, df = 1, P < 0.001). This result is in favor of the drug use.

Brood Caring
This ethological trait was affected by Metformin consumption (Table 3, line 1; Figure 3 d). When living under normal diet, the ants soon found the larvae removed from the nest, held them with their mandibles, and re-entered them in the nest. While consuming Metformin, the ants had some difficulty in holding the larvae; they tried several times to do so and did not succeed each time. If they finally succeeded in holding a larva, they presented obvious difficulties for transporting them. All this leaded to only few larvae re-entered in the nest over the five experimental minutes. The difference as for the numbers of not re-entered larvae over time between the ants maintained under one and the other kinds of diet was significant: N = 6, T = 21, P = 0.016. This result was in agreement with that on the ants' locomotion and that on their capability in climbing on a tower (see the two above subsections relative to the ants' speeds and to the ants' audacity).

Social Relationships
This ethological trait appeared to be somewhat affected by Metformin consumption (Table 3, line 2; Figure 3 e). Ants normally maintained generally stayed near its congeners, touching them with their antennae. Ants consuming the drug did not stay side by side during long time periods, though not really avoiding their nestmates. When they were very near from each other, they rather often opened their mandibles, but never gripped or tried to sting their opponent. Nevertheless, the difference as for the numbers of presented levels of aggressiveness between the ants under one and the other kinds of diet was significant: χ² = 10.60, df = 2, 0.001 < P < 0.01. The variable assessing the ants' aggressiveness equaled 0.05 and 0.21 for ants maintained under respectively a normal diet and a diet with Metformin. This slight but significant impact of the drug on the individuals' social interactions should be taken into account when caring of patients with Metformin.

Escaping Ability
This ethological trait was impacted by Metformin consumption (Table 3, line3; Figure 3 f). When maintained under normal diet, the ants walked along the rim of the enclosure, found the exit, and escaped. When consuming Metformin, the ants also walked along the rim of the enclosure and also found the exit, but then, they hesitated to escape, they seemed undecided and they often walked back then away from the exit. Consequently, only 4 among 12 ants were escaped after the 12 experimental minutes while, when living under normal diet, the 12 experimented ants could escape. The difference between the ants living under one and the other kinds of diet as for the numbers of escaped ants over time was statistically significant: N = 6, T = 21, P = 0.016. This result was in agreement with that relative to the ants' audacity (see the above subsection concerning this trait), and should be considered when treating patients with Metformin.

Cognition
This trait was impacted by Metformin consumption (Table 3, line 4; Figure 4 a). Ants living under normal diet soon entered the twists and turns path, and 10 among 30 ants could cross this difficult path over the twelve experimental minutes. When consuming Metformin, the ants hesitated to enter the twists and turns path, and delayed in progressing inside this difficult-path. They often came back on their way. Consequently, after the twelve experimental minutes, only 3 ants could cross the twists and turns path, and were beyond it. The difference between the ants maintained under one and the other kinds of diet as for their presence in front and beyond the difficult path was statistically significant: in front: N = 6, T = 21, P = 0.016; beyond: N = 4, T = 10, P = 0.063). This result was in agreement with those on the ants' audacity and escaping ability (see the subsections relative to these two traits): the ants hesitated, were undecided, and seemed reluctant to make novel tasks. Such an impact of Metformin should be considered when caring of patients with this drug.

Conditioning Acquisition, Memory
These abilities were affected by Metformin consumption (Table 4; Figure 4 b2). Ants maintained under normal diet soon acquired conditioning: they reached a score of 70% and 85% after respectively 31 and 72 training hours; they still detained a score of 80% 72 hours after the cue removal. When consuming Metformin, the ants never acquired conditioning; they still presented a score of 45% after 72 training hours. The difference of reached conditioning scores ijb.ccsenet.org International Journal of Biology Vol. 14, No. 2;2022 over time between the ants living under one and the other kinds of diet was significant: N = 6, T = 21, P = 0.016. The ants consuming the drug having learned nothing, their middle-term memory could not be assessed. When such ants were in the Y-maze, they were reluctant in entering a branch of the maze, above all that containing the green cube. Their hesitation recalled that presented near the risky apparatus (see the subsection relative to the ants' audacity), near the exit of the enclosure (see the subsection relative to the ants' escaping ability), as well as in front and inside the twists and turns path (see the subsection relative to the ants' cognition). Obviously, Metformin induced hesitation, reluctance in performing novel tasks, and this should be considered when treating patients with this drug.   Vol. 14, No. 2;2022 hesitated to approach the green cube; c: ants did not adapt themselves to the impact of the drug on their locomotion; d: ants consuming the drug still drunk little sugar water; e: ants developed dependence on the drug consumption

Adaptation to the Impact of Metformin on the Ants' Locomotion
After seven days of Metformin consumption, the ants had not adapted themselves to the impact of Metformin on their linear and angular speeds (Table 5, upper part; Figure 4 c). Indeed, at that time, the ants' linear speed was even smaller, though statistically not smaller than the speed they presented after one day of this drug consumption (χ² = 0.27, df = 1, 0.50 < P < 0.70). Also, after seven days of the drug consumption, the ants' angular speed was even larger than that presented after one day of consumption and this increase was statistically significant (χ² = 12.34, df = 2, 0.001 < P < 0.01). Such a non-adaptation to the side effects of the drug is not in favor of its use. This effect is not mentioned in the notice for use joined to the drug package (on the contrary, it is reported that side effects may occur only at the start of treatment). This should be known by practitioners and taken into account while treating patients with Metformin.

Habituation to the Impact of Metformin on the Ants' Sugar Water Consumption
After having consumed Metformin during eight days, the ants went on drinking very little sugar water (Table 5,   Numerical results are reported in Table 5, lower part, and two photos are shown in Figure 4 e. Ants developed dependence on Metformin consumption. In the course of the experiment devoted to the exam of such a potential dependence, 69 ants of colony A were counted on the drug solution and 28 ones on the drug-free solution, while 12 ants of colony B were counted on the drug-free solution and 33 ones on the drug solution. In total, 102 ants were sighted on the drug solution and 40 ones on the drug-free solution, what leaded to 71.83% of ants having chosen the drug solution and 28.17% having chosen the drug-free solution. The recorded numbers (102 vs 40) were statistically different from the numbers (71 vs 71) which should be obtained if ants had randomly gone onto the two provided solutions (χ² = 13.31, df = 1, P < 0.001). This revealed dependence, and consequently the 'non-stop' consumption of the drug, may, at least partly, explain the occurrence of anorexia and other symptoms in patients treated with this drug. Such patients should thus be imperatively monitored as for their potential development of dependence on the drug. Table 6 gives the numerical and the statistical results; Figure 5 graphically illustrates the results. In short, the effect of Metformin totally vanished in a total of about 13 hours after its consumption was stopped. In details, 2 hours after weaning, the effect of the drug was still statistically similar to its initial one, but 4 hours after weaning, it was already lower than its initial effect, and this could be perceived by consumers. The effect of Metformin went on decreasing over time after weaning. It still differed from the control situation until 12 hours after weaning (after 10 hours: P < 0.001; after 12 hours: P < 0.01), and became statistically similar to the control situation 14 hours after weaning (P < 0.50). It could thus be stated that the effect of Metformin vanished in a total of about 13 hours after its consumption was stopped. The assessment made 20 hours after weaning confirmed this vanishing. From 0 to 6 hours after weaning, Metformin lost '51 ang. deg./cm' of its effect, so 8.5 ang.deg./cm per hour. From 6 to 12 hours after weaning, Metformin lost '70 ang. deg./cm' of its effect, so 11.67 ang.deg./cm per hour. From 12 to 14 hours after weaning, Metformin lost '25 ang. deg./cm' of its effect, so 12.5 ang.deg./cm per hour. Metformin effects presented thus a decrease which slightly increased over time, but this slight increase was not statistically significant, and in fine, a mathematical and statistical analysis of the recorded data revealed that the decrease of the angular speed values recorded over time nearly obeyed to a linear function, and could best obey to the following function: E t = E Ia t or E t = 254 -10.21 t with E t = effect at time 't'; E I = initial effect; t = time (in hours) Note that, using the experimental values of 'a' here above reported (8.5, 11.67, 12.5), we obtained, for a linear function, a value of 'a' equaling 32.67/3 = 10.89, what approached the value provided by the mathematical analysis.

Decrease of the Effect of Metformin After Its Consumption Was Stopped
Such a rapid decrease of the effect of Metformin after weaning could account for the development of dependence on that drug consumption (Cammaerts, 2018b). Table 6. Decrease of the effect of Metformin after its consumption was stopped. The table gives the median (and quartiles) of ants' angular speed values (in angular degrees per centimeter) obtained over the decrease (column 2), as well as the results of the statistical analysis (column 3). Briefly, the effect of the drug soon became different from its initial one, and total vanished in about 13 hours. These results are illustrated in Figure 5 Time ( Figure 5. Decrease of the effect of Metformin after its consumption was stopped. Numerical and statistical results are given in Table 6. Briefly, the effect of the drug rapidly decreased, becoming significantly lower than its initial one as soon as 4 hours after weaning, and totally vanishing in about 13 hours. This decrease could be best described thanks to a linear function of the passing time

Discussion, Conclusion
Nowadays, Metformin is used for caring of persons suffering from type-2-diabetes (as initially) and several other illnesses, and is the second more frequent API pollutant of the natural watercourses. Few adverse effects are reported for Metformin, and most of them are not easily available to general public. On ants, this drug largely impacted the individuals' food consumption (mainly reducing the sugar water consumption), activity, orientation, audacity, social interactions, state of stress, cognition, learning and memorization. Ants did not adapt and did not habituate to the effect of Metformin, and they developed dependence on its consumption. After weaning, the effect of Metformin soon differed from its initial one and fully vanished in 13 hours. The results of our different experiments agreed with one another, and were also in agreement with those of other researchers (see the introduction section), considering the fact that we used a different model organism that other researchers. Before concluding, we below give some more bibliographical information on Metformin.
Metformin is effectively very efficient for caring of persons suffering from type 2-diabetes, and is so thanks to several modes of action more and more elucidated over practitioners' researches (Wiemsperger & Bailey, 2012). It is efficient for adults (same reference as above), and also for children (Kenneth et al., 2002). Metformin has also been proved to be efficient for treating obesity: the obtained results are really scientifically valid, and its modes of action are now known (Paolisso et al., 1998). Metformin also appeared to be useful for helping caring of persons suffering from some kinds of cancer (Jun et al., 1998;Legros et al., 2007). Research on this promising subject is going on, and, generally, best efficiency is obtained by combining Metformin with other drugs (Jun et al., 1998;Kailin et al., 2020).
Concerning the adverse effects of Metformin, the best-known ones are those on the intestine and on the liver; they are considered as being very important and are still now largely studied (Tongzhi et al., 2017). As for the effect of Metformin on the kidneys, let us recall (see the introduction section) that Hsu and co-authors have shown that this drug has an adverse effect on the renal function in patients suffering from type 2-diabetes and moderate chronic kidney disease (Wei-Hao, 2018). However, other researchers have shown that, in some cases of kidney illness, thanks to its numerous clinical effects, Metformin can act as a therapeutic drug (Heaf, 2014).
As for its environmental impact, being a largely used drug, Metformin is present in wastewater, and, together with its degraded compound guanylurea, contributes to the natural water pollution by API (Kosma et al., 2015: a study made in Greece and lasted four years). In their review, Elizalde-Velazquez and co-authors explain how Metformin is present in natural water and is toxic, report perspectives for reducing the adverse impacts of this drug on the environment, but do not relate any ethological or physiological effect of the drug on the aquatic fauna (Elizalde-Velazquez & Gomez-Olivan, 2020). In the same year, several co-authors found that Metformin present in water can be degraded by photocatalysis, and that the degradation products are not toxic (Carbuloni et al., 2020). Although we looked for published works on the biological impacts of Metformin on the aquatic fauna or on any other animals used as models, using several actual systems of research, we could not find any. The present work brings thus novel unpublished and useful information.
In the present paper, we reveal that, in experimental conditions and under a rather high concentration, Metformin caused adverse physiological and ethological harmful effects to ants. Therefore, the drug could very likely affect any insects. Since most species constituting the macrobenthos and inhabiting watercourses are insects, there is a risk that a prolonged exposition to Metformin induces changes in the behavior and/or in some physiological traits of these insect species, with consequences on the whole community living in the river bed and on the ecological processes which there occur. For instance, a decrease of general activity, orientation, audacity, and food consumption (as we observed in ants), occurring in preys as well as in predators, could affect the whole trophy chain in watercourses, reducing the velocity of organic matter processing and thus of self-depuration efficiency. Other products used by humans also affect the environment. We have studied some of them (see the introduction section) and, to cite one work among the numerous ones made on the subject, a study has been recently made on the ant Lasius niger, the authors concluding that the use of neonicotinoids induce severe environmental and ecological risks (Schlä ppi at al., 2020).
To conclude, Metformin, with its several medicinal properties, must be used by practitioners. However, the patients must imperatively be monitored as for their general activity, audacity, social interactions, cognition, memorization, and above all their potential dependence on this drug consumption and the occurrence of anorexia. Also, wastewater should be submitted to photocatalysis in order to degrade the drug for not affecting the health of organisms, those living in the water and those consuming natural water, this including thus humans.