*2.5. Data Analysis*

Raw respirometric data was normalized by *FCR* to allow direct comparison with the results presented in Lemieux et al. [13]. The pathway control ratios of oxygen flux with substrate types provided separately for the N- and S-pathways were normalized for flux through the combined NS-pathway (i.e., N/NS and S/NS). All respiratory states were corrected for residual oxygen consumption. For treated cells, MCPA was added after ADP and raw data were collected after MCPA addition. The different nominators and denominators were calculated by subtracting residual oxygen consumption from every calculated raw data. Regarding our ratios, CI*<sup>P</sup>* and CI + CII*<sup>P</sup>* are relative to OXPHOS capacity whereas CI + CII*<sup>E</sup>* and CII*<sup>E</sup>* are relative to ET capacity. Also, it is assumed that S-pathway linked respiration is not influenced by uncoupling [13,21] and therefore CII*<sup>E</sup>* = CII*P*. The ratios obtained for SUIT1 and SUIT2 in MCPA-treated cells were compared to ratios calculated in MCPA affected horses. Two vials from the same horse were used for respirometric analysis and runs were performed in duplicate (N = 2, n = 2), when not specified otherwise. Additionally, for each raw parameter in MCPA treated cells, a percentage from the control cells was calculated. Similarly, the effect of MCPA was evaluated with SUIT3, SUIT4, and SUIT5.

For a same respirometric parameter, a two-tailed *t*-test was performed between control and treated cells. Statistical significance was set at *p* < 0.05. In the tables, means are represented ± standard error of the mean (SEM). Additionally, ratios of MCPA treated cells were compared to previously published ratios of AM affected horses in Lemieux et al. via unpaired two-tailed *t*-test. For this comparison, ratios of different passages were pooled.

Regarding toxicity assays, data were analyzed by GraphPad Prism and a non-linear regression (sigmoidal, 4PL, least squares fit) was performed. For cell reduction potential, a non-linear regression (sigmoidal, 4PL, least squares fit) was also performed for each tested MCPA concentration at different times.

#### **3. Results**

#### *3.1. Toxicity Assays*

When analyzed independently, the kinetics of recorded fluorescence show a steep augmentation reaching a plateau at approximatively 10 h after toxin exposure. Plateau values were grouped for each concentration. After a 24-h exposure, wells submitted to 15 mM MCPA displayed between 78 and 105% of maximal response compared to the lysis solution (Promega Benelux, Leiden, The Netherlands). An important increase of cytotoxicity was observed between 10 and 15 mM (Figure 2). Best-fit values for total data, regardless of the passage indicated an IC50 of 15.7 mM. This value should however be interpreted with caution since a complete confidence interval could not be calculated. Also, it is notable that the differences between individual passages do result in very different

responses to MCPA. Indeed, cytotoxicity ranged widely from 25 to 67% for 8 mM and from 29 to 76% for 10 mM. Lowest toxicity was measured at passage 8 with 25 and 28% for 8 mM and 29 and 31% for 10 mM.

**Cytotoxicity vs. MCPA concentrations**

Regarding cell reducing potential, a non-linear regression (sigmoidal, 4PL, least squares fit) was also performed for each tested MCPA concentration at different times. Bestfit values indicated an IC50 of 8.09 ± 0.55 mM depending on the time recorded (Figure 3). In neither toxicity assay, HGA addition to cell cultures at concentrations ranging from 0.25 to 1 mM resulted in changes compared to control cells (data not shown). According to best-fit values, MCPA addition for subsequent experiments (i.e., HRR) was set at 9 mM.

#### **Cell redox capacity vs. MCPA concentrations**

**Figure 3.** Cell reduction capacity depending on methylenecyclopropylacetyl (MCPA) concentrations at different time points.

≤

#### *3.2. High-Resolution Respirometry*

**0**

**20**

**40**

**60**

**% of max response measured** 

**via luminescence**

**80**

**100**

At any passage, the addition of cytochrome *c* resulted in a slight (i.e., ≤10%) increase in O<sup>2</sup> flux thus confirming the preservation of the outer mitochondrial membrane integrity. When comparing obtained *FCR* of control vs. MCPA treated cells in the SUIT1 protocol (with glutamate and malate), a significant difference was observed for the ratio CI*P*/CII*E*, which was increased with 9 mM MCPA (Figure 4). For the three ratios with CI + CII at the denominator, there were no significant changes with MCPA treatment.

≤

1h 8h 16h 24h

**1 4 6 810 15**

**MCPA concentrations (mM)**

**Cell redox capacity vs. MCPA concentrations**

**Figure 4.** Ratios in control cells and cells treated with 9 mM methylenecyclopropylacetyl (MCPA) for the substrate-uncoupler-inhibitor titration protocol 1 (SUIT1). Each dot represents one passage, and the same passage are represented with or without MCPA. The bar represents the mean. \* Significantly different from controls (*p* < 0.05) with a t-test were indicated with a \*. Abbreviations: CI*<sup>P</sup>* = Complex I linked OXPHOS capacity; CII*<sup>E</sup>* = Complex II linked ET capacity; CI + CII*<sup>P</sup>* = Complex I&II linked OXPHOS capacity; CI + CII*<sup>E</sup>* = Complex I&II linked ET capacity.

In SUIT2, significant differences between groups were also detected. Compared to the control cells, the MCPA treated cells showed an increase in the ratios of CI*<sup>P</sup> <sup>a</sup>*/CI + II*<sup>P</sup>* and CI*<sup>P</sup> <sup>a</sup>*/CII*<sup>E</sup>* and a decrease in the ratio CII*E*/CI + II*<sup>P</sup>* (Figure 5). Interestingly, the ratio CI*P*/CI + II*<sup>P</sup>* was significantly different between groups when pyruvate and malate were used as substrates, but not when glutamate was added (Figure 5). The ratio of CII/CI + II was significant only when the denominator (CI + II) was taken under the *P* state, but not when it was taken under the *E* state (Figure 5).

Without considering cell passages, a decrease in oxygen flux of 46% in average (SEM ± 0.08) was recorded after MCPA addition to the chamber in SUIT1. This decrease remained relatively constant for the rest of the protocol: CI + II*<sup>P</sup>* = 43% of control (SEM ± 0.15), CI + II*<sup>E</sup>* = 40% of control (SEM ± 0.2), CII*<sup>E</sup>* = 32% (SEM ± 0.13) (Table 2). Similarly, a decrease in oxygen flux to 53% of control cells (SEM ± 0.11) was recorded after MCPA addition to the chamber (CI*P*). A constant decrease was noted for the rest of the protocol: CI + II*<sup>P</sup>* = 43% of control (SEM ± 0.10), CI + II*<sup>E</sup>* = 40% of control (SEM ± 0.17), CII*<sup>E</sup>* = 29% (SEM ± 0.09).

**Table 2.** Respirometric value percentage of MCPA treated cells compared to control cells.


**Figure 5.** Ratios in control cells and cells treated with 9 mM methylenecyclopropylacetyl (MCPA) for the substrateuncoupler-inhibitor titration protocol 2 (SUIT2). Each dot represents one passage, and the same passage are represented with or without MCPA. The bar represents the mean. \* Significantly different from controls (*p* < 0.05) with a t-test were indicated with a \*. <sup>a</sup> using glutamate and malate as N pathway substrates. <sup>b</sup> using pyruvate, malate and glutamate as N pathway substrates.

Similarly, for fatty acid protocols using acetylcarnitine as substrate, oxygen flux immediately after MCPA addition was on average 67% of the control value (SEM ± 0.2). When Succinate was added, 61% of control (SEM ± 0.24) and after FCCP addition 72% of control (SEM ± 0.08). Finally, after rotenone addition, 51% (SEM ± 0.12) of control were attained. When the fatty acid substrate was octanoylcarnitine, oxygen flux immediately after MCPA addition was on average 74% of the control value (SEM ± 0.24). For the rest of the protocol: CI + II*<sup>P</sup>* = 62% of control (SEM ± 0.14), CI + II*<sup>E</sup>* = 56% of control (SEM ± 0.24), CII*<sup>E</sup>* = 49% (SEM ± 0.15). Fatty acid protocols using palmitoylcarnitine as substrate, oxygen flux immediately after MCPA addition was on average 67% of the control value (SEM ± 0.07). For the rest of the protocol: CI + II*<sup>P</sup>* = 60% of control (SEM ± 0.11), CI + II*<sup>E</sup>* = 58% of control (SEM ± 0.31), CII*<sup>E</sup>* = 37% (SEM ± 0.25). Hence, a decrease in respiration was recorded in all SUIT protocols used. Also, whatever the protocol, initial LEAK respirometric parameters were similar between control and treated cells.

Regardless of the substrate sustaining electron flow used, the addition of MCPA depressed respiration in treated cells. This effect was noted at each cell passage.

#### *3.3. Total Protein Content*

In order to compare macroscopic cell count to an internal measure, total protein content was analyzed (Table 3). Despite a similar cell count, the total protein content seemed to vary between runs (*no statistical analysis performed*). Since total protein content was not measured for each vial used, internal normalization by ratios was preferred. These ratios are independent of cell count or tissue mass [28].

**Table 3.** Total protein content (±SEM) at each passage performed in duplicate (n = 2). Protein concentration is expressed in µg/µL).


#### **4. Discussion**

Regardless of the protocol, an immediate effect of MCPA on mitochondrial electron transfer system (ETS) complexes was observed. This is interesting as it corroborates the in vivo observations obtained in horses with AM [11]. Similar to findings in Lemieux et al. [13], mitochondrial respiration seemed to be more depressed with glutamatemalate sustained respiration (SUIT1) compared to pyruvate-glutamate-malate (SUIT2). Indeed, a severe depression of both OXPHOS and ET capacity could be reproduced in vitro. In all five SUIT protocols, the addition of MCPA resulted in an immediate effect on the N- and S-pathway but also on the F-pathway, sustained by fatty acid ß-oxidation. When analyzing SUIT2, the contribution of the S-pathway was similar to the N-pathway in affected and control horses [13]. However, in our study, the S-pathway seemed to be more affected and therefore resulted in a strong diminution of OXPHOS and ET capacity, no matter which protocol was used. This finding is probably linked to a longer exposure time to the toxin since CII-linked activity is measured after CI and CETF sustained O<sup>2</sup> flux.

Regarding fatty acid substrates, it is also important to note that ß-oxidation supplies electron transfer through the N-junction as well as the rate-limiting F-junction pathway branches. Even though a progressive decrease in respiration was also recorded with SUIT3, 4 and 5, F-pathway combined with S-pathway sustained respiration resulted in better supported respiration than the NS-pathway.

Overall, MCPA addition to the oxygraphy chamber resulted in a generalized inhibitory effect, acting either on all ETS complexes, or having a specific target in downstream ETS components as Q, CIII or CIV. In order to address this question, enzymatic assays testing either the individual complexes' response to the toxin or related enzymes upstream (e.g., PDH) and other SUIT protocols targeting specific downstream complexes will need to be applied in future studies. Also, it cannot be excluded that the IC50 calculated on the basis of toxicity and viability assays was too high to identify the first target and resulted in a general decrease of O<sup>2</sup> flux.

It is also noteworthy that despite the recording of a depression in mitochondrial respiration, without considering the passage and the cell batch, the differences between the different passages should be further explored in order to define the cause of variability. The use of undifferentiated equine primary myoblasts implies that metabolism and mitochondrial function may significantly differ compared with differentiated myotubes [30]. Therefore, it seems plausible to suspect that the oxidative phenotype, which depends on oxidative capacity and fiber type composition [30], would be impacted by myogenic differentiation and therefore is not completely reflected in situ metabolism. While the toxic effect of MCPA and MCPrG on ß-oxidation is well documented [9,10] it is worth noting that in this study as well as in the study performed by Lemieux et al. [13], the pathologic pathway is not restricted to an inhibition of ß-oxidation since SUIT1 and SUIT2 both result in a severe depression of the OXPHOS and ETS homeostasis.

The concentrations used in the cytotoxicity/viability study ranged from 1 to 15 mM MCPA and from 0.25 to 1 mM HGA. Regarding the latter, no effect was observed at the aforementioned concentrations. This might indicate a lack of metabolization by the cell culture used within the time of the experiment. When analyzing both cytotoxicity and cell reduction capacity, it appears that a wider range of concentration must be tested in future experiments. Indeed, at 8 mM MCPA, cell reduction capacity, which is cumulative, increases. This may be imputed to the immediate effect of MCPA on the cells. It is, however, unclear if only some cells are less impacted by low concentrations of MCPA. In any event, at the lowest concentration after several hours the reductive capacity of the cells is restored and similar to control cells. However, it is clear that even at 1 mM, there is an immediate reaction to the toxin indicating cellular impairment. Also, best-fit value calculated for IC50 in our cytotoxicity experiment was 15.7 mM. This concentration would indicate that almost two times more MCPA is necessary to induce a cytotoxic effect compared to an effect on cell reductive capacity. When analyzed in detail, the calculated cytotoxicity IC50 value for passages 6 and 7 is 9.02 mM (±0.58 mM), similar to the IC50 obtained for the reducing potential. This raises the question of the fluctuation of cell types and maturity between passages as well as the sensitivity of these assays for these types of cells. Concentrations of MCPA (i.e., 9 mM) induced a severe decrease in mitochondrial respiration by HRR, compatible with a 50% decrease depending on the evaluated mitochondrial complex. However, the range of concentrations needs to be larger in future experiments in order to determine the minimal concentration at which a toxic effect of MCPA can be recorded. Also, despite a similar cell count, total protein content varied between runs. The differences between passages may also be related to a variable response on the cellular level to the toxin.

In any case, our study manages to replicate mitochondrial alterations in response to MCPA intoxication. However, current mitochondrial activity assessment still relies on endpoint assays, which yield limited kinetic and therefore prognostic information [31]. Indeed, our assays are based on oxygen flux recordings depending on substrates, pathways, and oxidative phosphorylation. Since in this case the clinical picture is a myopathy secondary to poisoning, it is essential to determine if the mitochondrial damage is consequential to the ß-oxidation defect and therefore toxic lipid accumulation. To determine if altered oxidative phosphorylation/mitochondrial respiration is causal or consequential to the clinical symptoms observed in AM affected horses, tissue-tissue interactions might need to be monitored to detect early if onsets of mitochondrial stress precede acute rhabdomyolysis. Even though cellular adaptations might be far-fetched and unrealistic in such an acute disease, defining the onset of stress in the first affected tissue will enhance chances of therapy. Since the mainly affected muscle fibers are oxidative, a mitochondrial dysfunction leading to a shift from oxidative phosphorylation to glycolysis, these cells may be less equipped to assume their role because of their limited ability to generate ATP by alternative means or because of the ultrastructural mitochondrial changes [19].

So far, many factors have been cited as potential contributors in the pathophysiology of mitochondrial dysfunction involved in a wide variety of disorders as decreased mitochondrial content, altered substrate delivery, muscle inflammation, morphological distortion of mitochondria due to glycogen cytoplasmic accumulation, oxidative damage and mitochondrial damage induced by gluco- and lipotoxicity secondary to intracellular substrate accumulation [32–34]. Since the direct effect of MCPA can be replicated on a cellular model, a down-regulation of nuclear and mitochondrial genes in AM does not seem plausible. However, the larger scale consequences on organs and organelles of HGA and MCPrG metabolization are to date unknown and may also constitute a therapeutic target strategy. Indeed, if the toxins have an impact on mitochondrial proteostasis, the damage may occur at different scales; the horse's whole metabolism can be impacted, the mitochondrion's interaction with the cell and the mitochondrion itself may be damaged, which will activate pathways to counteract the damage [35].

In the same line, MCPA-carnitine concentrations quantified in serum of AM-affected horses went up to several thousand nmol/L [7,15]. It is therefore imperative to compare our results to a direct dosage of MCPA-carnitine in muscle of AM affected horses as well as to more sensitive techniques, able to detect event slight augmentations of low concentrations. Additionally, purchased cells originated from one donor horse, and the reaction of these cells are therefore not to be extrapolated to all animals susceptible to HGA and MCPA intoxication. Through an immortal cell line, an easy-to-use and alternative can be found [36,37], providing a pure population of cells to reproduce results obtained in this preliminary study. A standardized cell culture with an immortal cell line will also minimize horse-associated reactions to the toxin as well as passage-dependent responses.

In conclusion, our cellular in vitro model reproduced MCPA linked toxicity to a certain extend. For result reproduction, cytotoxicity assessment and in fine high throughput screening of therapeutic molecules, the use of an immortalized cell line is the next step.

**Author Contributions:** C.-J.K. and D.-M.V.: Conceptualization and methodology; C.-J.K. and A.N.: investigation; C.-J.K. and D.S.: data curation; C.-J.K. and D.-M.V.: writing—original draft preparation; H.L., D.-M.V. and C.-J.K.: review and editing; A.M.-M., A.N. and D.-M.V.: proofreading; C.-J.K. and H.L.: visualization; D.-M.V. and T.A.: supervision and project administration; D.-M.V. and T.A.: funding acquisition. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by "Fonds spéciaux à la recherche" of the University of Liège. The first author is the recipient of a "Fonds De La Recherche Scientifique—FNRS" grant.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Equine primary myoblasts were obtained from RevaTis and are commercially available.

**Acknowledgments:** The authors gratefully acknowledge the valuable help received from T. Arnould for protein extraction at the Laboratory of Biochemistry and Cell Biology (URBC) from the Namur Research Institute for Life Sciences (NARILIS) of the University of Namur (Belgium). Our acknowledgements extend to L. Gillet and the Laboratory of Immunology and Vaccinology of the University of Liège (Belgium) for the use of the EnSpire® Multimode plate reader and Clovis Wouters financed by a fellowship from Pommier Nutrition and LABÉO (UniCaen and ULiège) for proofreading.

**Conflicts of Interest:** The authors declare no conflict of interest.
