Memory impairment in the D2.mdx mouse model of Duchenne muscular dystrophy is prevented by the adiponectin receptor agonist ALY688

Abstract Memory impairments have been well documented in people with Duchenne muscular dystrophy (DMD). However, the underlying mechanisms are poorly understood, and there is an unmet need to develop new therapies to treat this condition. Using a novel object recognition test, we show that recognition memory impairments in D2.mdx mice are completely prevented by daily treatment with the new adiponectin receptor agonist ALY688 from day 7 to 28 of age. In comparison to age‐matched wild‐type mice, untreated D2.mdx mice demonstrated lower hippocampal mitochondrial respiration (carbohydrate substrate), greater serum interleukin‐6 cytokine content and greater hippocampal total tau and Raptor protein contents. Each of these measures was partly or fully preserved after treatment with ALY688. Collectively, these results indicate that adiponectin receptor agonism improves recognition memory in young D2.mdx mice.

in murine models of DMD, including C57BL/10ScSn-mdx and D2.mdx mice (Hayward et al., 2022;Vaillend et al., 2004).Such cognitive impairments might be related to disruptions in brain energy homeostasis, given that separate investigations using non-invasive magnetic resonance spectroscopy reported elevated inorganic phosphate (P i ) to ATP and P i to phosphocreatine (PCr) in boys with DMD (Tracey et al., 1995).Similar observations were reported in C57BL/10ScSnmdx mice (Tracey et al., 1996), who found that maximal activities of mitochondrial enzymes in the brain were not different from wildtype (WT) mice, suggesting that the capacity of specific mitochondrial pathways might be preserved.However, functional assessments of oxidative phosphorylation using intact mitochondria have not been performed.Such approaches have revealed attenuations in mitochondrial substrate oxidation in muscle from mdx mice and muscle from people with DMD (reviewed by Bellissimo et al., 2022).
Glucocorticoids (GCs) are the current standard of care for slowing the progression of muscle weakness in males with DMD (Balaban et al., 2005;DeSilva et al., 1987;Marden et al., 2020) by reducing inflammation (Kim et al., 2015).The effects of GCs on memory in DMD are not fully resolved in the literature, but memory impairments attributed to GC therapy have been identified more broadly in other populations (reviewed by Brown et al., 2007).As such, there is an unmet need to develop new therapies that improve memory in addition to muscle function in DMD.In this regard, adiponectin receptor (AdipoR) agonism was recently shown to improve markers of muscle quality in mdx mice (Abou-Samra et al., 2020).The agonist ALY688 (also referred to as ADP355; Otvos et al., 2011), in particular, attenuates inflammation in various inflammatory disorders, such as dry eye and liver diseases (Kumar et al., 2014;Wang et al., 2016), and reprograms substrate metabolism (Da Eira et al., 2020).Given that adiponectindeficient mice have impaired recognition memory (Bloemer et al., 2019), we hypothesized that a recently developed slow-release formulation (ALY688-SR) with improved pharmacokinetic properties enabling once-daily injection (Allysta Pharmaceutical, unpublished observations) would improve memory during early disease stages in young D2.mdx mice, in association with lower inflammation and enhanced mitochondrial metabolism.(Bellissimo et al., 2023;Coley et al., 2016;Hughes et al., 2020Hughes et al., , 2019;;Ramos et al., 2020).Mice were killed 20-24 h after the last dose.

Highlights
• What is the central question of this study?
Can adiponectin receptor agonism improve recognition memory in a mouse model of Duchenne muscular dystrophy?
• What is the main finding and its importance?
Short-term treatment with the new adiponectin receptor agonist ALY688 improves recognition memory in D2.mdx mice.This finding suggests that further investigation into adiponectin receptor agonism is warranted, given that there remains an unmet need for clinical approaches to treat this cognitive dysfunction in people with Duchenne muscular dystrophy.
Breeding of WT mice was unsuccessful, similar to a previous report (Holmes, 2003).Instead, 3-week-old WT DBA/2J WT mice (D2A) were obtained from Jackson Laboratories (stock number 000671; Bar Harbor) and allowed to acclimate in the same room as the D2.mdx mice for 1 week before they were killed.As shown in the Results, the discrimination index of these WT mice was similar to previous work from our group (Hayward et al., 2022).Mice were housed in standard 12 h-12 h light-dark cycles and were allowed access to standard rodent chow and water ad libitum.Four days before the end of the protocol, all groups underwent novel object recognition (NOR) acclimation and testing as described below.
Mice were anaesthetized under 5% isoflurane vaporized in medical air (21% oxygen) at a flow rate of 2 L/min, then maintained at 2%-3% before exsanguination.Muscles were removed and used for another investigation (under review at the time of this publication).

Novel object recognition test
The NOR test was performed as previously described (Bagdatlioglu et al., 2020;Denninger et al., 2018;Hayward et al., 2022;Leger et al., 2013).Briefly, animals were placed in an open field arena (40 cm × 40 cm × 40 cm).Sessions were recorded with a cell-phone video camera secured above the apparatus.Testing was performed in three stages: acclimation, habituation and testing.During acclimation, mice were placed into the arena and allowed to move freely for 5 min each day for 4 days preceding testing.The next day, habituation was performed, whereby two identical objects (object 1) were placed in opposite corners of the arena, and mice were left to explore for 10 min.
Objects were similar in size to the animals and were chosen to ensure novelty in all trials.After a habituation period, mice were returned to a neutral cage for 30 min.Thereafter, mice were placed back in the arena with one familiar object and one novel object (object 2) and allowed to explore for 10 min.Object exploration time was defined as the time the mouse interacted with the object, defined by sniffing or touching the object when the mouse is <2 cm from the object.Sitting or standing on the object was not included unless the mouse sniffed the object while climbing on it.For the trial to be considered valid, the animal must have interacted with the object for >20 s.The discrimination index was calculated as follows: [Time (object 2) − time (object 1)]/total time.

High-resolution respirometry
Mitochondrial oxygen consumption (respiration) was measured using the in situ brain permeabilization previously described (Herbst & Holloway, 2015), with some modification.After weighing, a portion of the hippocampal samples was quickly minced with scissors in a chilled tube containing BIOPS buffer followed by immediate placement into an Oxygraph-2K respirometer (Oroboros Instruments, Austria) containing MiR05 respiration medium (0.5 mM EGTA, 3 mM MgCl 2 , 10 mM KH 2 PO 4 , 20 mM HEPES, 60 mM potassium lactobionate, 110 mM sucrose and 1 g/L bovine serum albumin; pH 7.2) at 37 • C, with constant stirring at 750 r.p.m.Samples were equilibrated in respiration buffer for 10 min before the addition of 50 μg/mL saponin to facilitate permeabilization of the tissue.After permeabilization, pyruvate-stimulated respiration was examined in the brain using 5 mM pyruvate and 2 mM malate to generate NADH and saturate electron entry into complex I.To examine state III respiration as an index of oxidative phosphorylation, ADP was then added at a concentration of 15 μM to approximate the concentrations reported in human brains using non-invasive magnetic resonance spectroscopy assessments (Roth & Weiner, 1991).Cytochrome c was added as a test of mitochondrial outer membrane integrity.All experiments demonstrated <10% increase in respiration.Each protocol was initiated with a starting [O 2 ] of ∼350 μM and was completed before the oxygraph chamber [O 2 ] reached 150 μM, as done previously (Perry et al., 2011;Ydfors et al., 2016).Polarographic oxygen measurements were acquired in 2 s intervals, with the rate of respiration derived from 40 data points and expressed as picomoles per second per milligram wet weight.Chemicals and reagents were purchased from Sigma (St Louis, MO, USA) or BioShop (Burlington, ON, Canada).

Cytokine profiling
Protein contents of TGF-β, TNF-α, IL-1β, IL-6 and IL-10 in serum were measured by flow cytometry using the LEGENDplex Mouse Custom Panel (BioLegend, San Diego, CA, USA).Serum, collected from all groups, was diluted fourfold in the assay buffer and loaded onto the provided assay plate in a blinded manner.The assay was completed according to the manufacturer's instructions on an Attune NxT flow cytometer (Thermo Fisher).The FCS files generated on the flow cytometer were analysed using the LEGENDplex cloud-based analysis software.

Western blotting
Hippocampal tissue (collected from the same mice used in highresolution respirometry experiments) was homogenized in cold lysis buffer containing (mM: 20 Tris-HCl, 150 NaCl, 1 EDTA, 1 EGTA, 1% Triton X-100, 2.5 Na 4 O 7 P 2 and 1 Na 3 VO 4 ; pH 7.0) supplemented with protease and phosphatase inhibitors (Sigma) according to the protocol established by Hughes et al. (2019).Detection of electron transport chain complex subunits was performed according to the procedure described by Hughes et al. (2019)

Statistics
Results

RESULTS
The D2.mdx-VEH mice demonstrated impaired recognition memory, as shown by a lower discrimination index in the NOR test (Figure 1a,b).
Hippocampal mitochondrial pyruvate-supported respiration stimulated by ADP at physiological concentrations (see Materials and Methods) was lower in D2.mdx-VEH but completely preserved by ALY688-SR (Figure 1c).These differences between groups are likely reflect adaptive reprogramming intrinsic to mitochondria, given that electron transport chain subunit contents were similar in all groups (Figure 1d,e).
We then assessed cytokine markers, given that the parent compound of ALY688-SR (ALY688) has been shown to have antiinflammatory effects.No significant changes were observed in hippocampal mRNA of IL-6, IL-1β, TNF-α or IL-10 (Figure 2a-d).
Cytokine protein contents were assessed in serum owing to tissue limitations in the hippocampus.Serum IL-6 and TNF-α were significantly elevated in D2.mdx-VEH mice versus D2A control animals (Figure 2e,g).ALY688-SR attenuated IL-6 (Figure 2e) and increased IL-1β relative to D2.mdx-VEH, while also increasing IL-10 versus WT (Figure 2f,h). We

DISCUSSION
Here, we show that the new AdipoR agonist, ALY688-SR, improved recognition memory in young D2.mdx mice after a short-term, 21-day treatment protocol.This effect was associated with a complete restoration of hippocampal pyruvate-supported mitochondrial respiration (index of glucose oxidation) when stimulating ATP synthesis with an ADP concentration matching the level reported in vivo in the human brain (Tracey et al., 1995).This finding is interesting, given the high dependence of neurons on glucose as a fuel source.
Given that this protocol, in part, on complex I oxidation of NADH, the finding might also imply that complex I dysfunction occurs in the hippocampus, similar to reports in muscle from D2.mdx mice (Bellissimo et al., 2023;Hughes et al., 2019Hughes et al., , 2020;;Ramos et al., 2020) and in other regions of the brain in C57BL/10ScSn-mdx mice (Tuon et al., 2010).Hippocampal mitochondrial relationships to NOR F I G U R E 3 AMPK and downstream protein markers related to autophagy and neurofibrillary tangles.(a) Theoretical cascade linking AMPK (an adiponectin receptor target) to factors related to autophagy and neurofibrillary tangles.The AMPK target, Raptor, is rendered inactive upon its phosphorylation, allowing for induction of autophagy through phosphorylation of ULK-1, leading to clearance of dysfunctional mitochondria and proteins (Lee et al., 2010;Norwitz & Querfurth, 2020).Alternatively, activation of the mTORC1 complex leads to increased protein synthesis via p70s6k1, which has been implicated in tau phosphorylation, destabilization of microtubules and generation of neurofibrillary tangles (Iqbal et al., 2010;Pei et al., 2006).Made with BioRender.(b,c) Western blot markers for targets outlined, assessed in hippocampal tissue (a).Data are expressed as the mean ± SD, with n = 6-12 per group.*P ≤ 0.05 WT versus D2.mdx-VEH; § P ≤ 0.05 D2.mdx-VEH versus D2.mdx-ALY688-SR.Abbreviations: AMPK, AMP kinase; mTORC1, mammalian target of rapamycin complex 1; p70s6k1, ribosomal protein S6 kinase β1; Raptor, regulatory-associated protein of mTOR; ULK1, Unc-51 like autophagy activating kinase; WT, wild-type.
have also been reported previously.For example, reduced NOR and hippocampal mitochondrial structural dynamics were altered in ageing-related neurodegeneration (Mishra & Thakur, 2022), whereas improvements in pyruvate-supported respiration and NOR were observed in a model of erythropoietin overexpression (Jacobs et al., 2021).
The precise relationship between dystrophin mutations and mitochondrial stress in the brain is not clear but might be informed by previous discoveries in muscle from dystrophin-deficient animal models (Bellissimo et al., 2022;Nghiem et al., 2017), which generally implicate calcium overload, redox stress and inflammation as possible links (Guiraud & Davies, 2017).Increases in serum IL-6 and reductions in mitochondrial pyruvate oxidation in the present study align with this possibility, particularly given that ALY688-SR partly attenuated IL-6 while completely preserving respiration and recognition memory.
Although hippocampal inflammatory markers were unchanged, the reductions in serum IL-6 with ALY688-SR present the possibility that systemic inflammation caused by the well-characterized muscle damage in D2.mdx mice (Bellissimo et al., 2023;Coley et al., 2016;Fukada et al., 2010;Hughes et al., 2019) might contribute a trigger separate from the direct effects of the dystrophin mutation in the hippocampus.Indeed, pharmacological inhibition of IL-6 in the hippocampus improved long-term memory in rats (Balasubramanian et al., 2022), but the effects on mitochondrial bioenergetics were not examined.
Improved recognition memory was also related to increased contents of tau protein and the inhibitory complex of mTOR, Raptor, both of which are downstream of AdipoR signalling ( Domise et al., 2016;Gwinn et al., 2008).The lack of change in their phosphorylation state does not necessarily rule out a potential role for these pathways in mediating the disease or treatment effects, given that the tissues were harvested 20-24 h after the last dose.Assessing the activity of these and other signalling cascades shortly after dosing might provide more insight into their potential contributions.This timing consideration might also explain why AMPK itself was not phosphorylated in response to the drug despite being a known downstream target of AdipoR (Yamauchi et al., 2002;Zhang et al., 2022) and considering that both AMPK and p38MAPK are phosphorylated after 30 min of ALY688 treatment in L6 muscle cells (Sung et al., 2022).Future experiments could extend these findings to incubations in neurons.

Limitations and future directions
In addition to the hippocampus, the NOR test is dependent on multiple regions of the brain not investigated in the present study, such as the prefrontal cortex.Also, the lack of change in protein markers of autophagy and neurotrophic factors does not reflect the activity of these pathways, which require assessments of their kinetic aspects with additional methodologies.As such, these processes could still be considered in future research with respect to how dystrophin mutations impair recognition memory and how ALY688-SR restores this cognitive function, given that AdipoRs and/or AMPK (a target of AdipoR) have been linked to autophagy, neurofibrillary tangles or amyloidogenesis (Chan et al., 2012;He et al., 2021;Pei et al., 2006;Zhang et al., 2011).
Lastly, the relative contributions of direct effects of ALY688-SR on brain AdipoRs (Clain et al., 2022;Thundyil et al., 2012) or indirect effects remain to be determined.

CONCLUSION
ALY688-SR prevented impairments in recognition memory after 21 days of treatment in young D2.mdx mice.Given that there is an unmet need for therapies that improve memory and other cognitive functions in DMD, these findings provide a foundation for continued exploration into the mechanisms and potential of adiponectin receptor agonism in treating cognitive dysfunction separately from the myopathy in this debilitating disease.

AUTHOR CONTRIBUTIONS
Study design and conception were prepared by Catherine A.
are expressed as means ± SD.The level of significance was set to P < 0.05 for all statistics.D'Agostino-Pearson normality tests (GraphPad Prism software; GraphPad, La Jolla, CA, USA) confirmed that all data were normally distributed, and one-way ANOVAs were used with the two-stage step-up method of Benjamini, Krieger and Yekutieli post hoc analyses for false discovery rate corrections in multiple-group comparisons.All reported P-values are false discovery rate-adjusted P-values (traditionally termed 'q').A required sample size of 12 was calculated based on the exploration time results of our previous work in the D2.mdx mouse model(Hayward et al., 2022) using a desired power of 0.80 and α of 0.05.
next assessed markers of potential pathways linking AMPK or inflammation to neural function, given that AdipoR agonism activates AMPK signalling in various models(Abou-Samra et al., 2020;Da Eira et al., 2020) and lowers inflammation, as noted above.As shown in Figure3b,c, total AMPK protein was lower after treatment with ALY688-SR.Although difficult to explain, the degree of activation (p-AMPK or p-AMPK/AMPK) was not altered.Given that tissues were harvested ∼20-24 h after the last injection, it is possible that AMPK activation might have occurred earlier, given that previous work showed rapid increases in AMPK phosphorylation, albeit in L6 muscle cells after 30 min of treatment(Sung et al., 2022).Alternatively, it is possible that a different drug dose would be required to see an effect on APMK phosphorylation.ALY688-SR also completely prevented the increases in protein contents of total tau (marker of neurofibrillary tangles) and total Raptor (upstream regulator of tangles and autophagy) seen in D2.mdx versus WT.However, phosphorylation (both absolute and relative to total protein) of tau and Raptor (serine 792; an AMPK-specific phosphorylation site;Gwinn et al., 2008) were similar in all groups (Figure3a-c).Markers of amyloidogenesis (Figure4a-c) were similar in all groups, as were brain-derived neurotrophic factor (BDNF) and neuronal nuclear protein (NeuN) (data not shown).ALY688-SR did not alter body weight, tibial length, spleen mass, kidney mass or liver mass (data not shown).
Bellissimo, Laura N. Castellani, Shivam Gandhi, Gary Sweeney, Christopher G. R. Perry and Allysta Pharmaceuticals, and all authors contributed to the rationale for specific measurements.Data acquisition, analysis and interpretation of the work were performed by all authors.The manuscript was written by Catherine A. Bellissimo and Christopher G. R. Perry.All authors contributed to the manuscript, approved the final version and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.
translational modifications of pyruvate dehydrogenase (noting the use of pyruvate as a substrate in our respirometric protocol) in addition to alterations to mitochondrial morphology-linked changes in the regulation of oxidative phosphorylation and global mitochondrial content measures beyond the OXPHOS markers that were assessed.
(Hayward et al., 2022)e to obtain sufficient sample of this region, given the small size of the brain at 4 weeks of age, prior research has shown that greater markers of amyloidogenesis occur in this region compared to only one marker in the hippocampus at 8 weeks of age in D2.mdx mice(Hayward et al., 2022).The pre-