Justifying and reporting the use of stress protocols in pre‐clinical experiments that use animals

We thank Dr. Stanford and colleagues for their Letter regarding our recent publication by Michael et al., in the British Journal of Pharmacology (Michael et al., 2025). We share and applaud their commitment to protecting animal welfare in preclinical research. In line with that, we confirm that our study investigating (2R,6R)-hydroxynorketamine's potential as a novel pharmacotherapy in opioid use disorder (OUD) using animal models rigorously adhered to international ethical standards, the 3Rs (Replacement, Reduction, Refinement), and ARRIVE 2.0 guidelines. We note that our manuscript included procedural citations, consistent with BJP regulations. We welcome this opportunity to further clarify how our studies align with ARRIVE 2.0 items 9, 12, and 16, explaining procedures, protocol justification, and animal welfare monitoring. The 10-day chronic social defeat stress (CSDS) protocol was selected to model the interaction between psychosocial stress susceptibility and OUD vulnerability, as in previous studies (Cooper et al., 2017). As a gold-standard protocol for stress susceptibility research, the CSDS, standardized in 2011 by the Nestler group (>1200 papers used this protocol; cited >1600 times) (Golden et al., 2011), reliably induces stable susceptible or resilient phenotypes in mice, replicating key aspects of human inter-individual behavioural variability in stress responses, thus establishing it as an ethologically-relevant model. This is critical, as this protocol is based on both psychosocial and physical stress components, better mimicking the human conditions. For Replacement, we evaluated non-animal alternatives, including cell lines and computer-based methodologies, but determined that intact neural circuits are required for assessing depression- and opioid-related behavioral endophenotypes, including reward-related behaviours, emotional deficits during abstinence and drug reinstatement. Reduction was achieved via a priori power analyses based on our extensive experience using these experimental procedures, including the CSDS model (Georgiou et al., 2022; Zanos et al., 2016). We aimed to minimize numbers of animals being used, while preserving rigour. Refinement included detailed 3-day pre-screening of the aggressive mice (ensuring attack latency being <60 s to induce adequate stress, but avoiding to inflict significant wounds in the test mice). We also kept physical encounter to 10 min per day with the presence of a trained experimenter, blind to the treatment groups. This physical stressor was followed by a sensory exposure via perforated divider for the rest of the day for 10 days in total. These steps ensure the least physical harm intensity for depression-like endophenotype stability. As part of our protocol, we allowed ~3 weeks for full sucrose preference recovery. CSDS procedures are described in our manuscript, including encounter parameters and the daily use of novel aggressive mice among others, by also referring to the standardized protocol published by Golden et al. (2011), as well as our prior publications in Nature, Nature Neuroscience, PNAS and others (e.g., Georgiou et al., 2022; Zanos et al., 2016). To enhance transparency regarding animal welfare (ARRIVE 2.0, Item 16), we note that daily monitoring of mice was carried out by trained staff and veterinarians assessing cage locomotion, posture, piloerection, weight (>20% loss trigger), and presence or absence of wounds via direct visual inspection (no formal scoring system was required, as standard for approved CSDS protocols). Physical encounters involved brief attacks (latency <60 s) with immediate submissive posture by the test mouse, defined as submissive upright posture, stretching forepaws, exposing belly, and/or freezing. Attacks, by the aggressive mouse, were ceased upon the test mouse displaying submissive behaviour. Although it did not occur in our studies, if severe fighting (i.e., continuous attacking/biting despite submissive behaviour) had been observed by the trained experimenter present and blind to the treatment groups, the test mouse would have been immediately transferred to a different aggressor cage. The most severe cases included occasional minor scratches (leading to max. 1–2 of <0.5 cm non-severe cuts) that healed naturally with a brief application of a triple-antibiotic cream (as approved for non-severe wounds). No wounds exceeded 1 cm, no cumulative severe wounding (e.g., multiple 0.5 cm cuts), barbering or dominance behaviours beyond protocol design occurred, and no animals exhibited excessive distress or met predefined humane endpoints (euthanasia for wounds >1 cm or >20% weight loss, persistent piloerection, hunching posture, reduced or abnormal locomotion, or interference with normal food/water intake; per the Golden et al. (2011) study, and our previous publications (see Georgiou et al., 2022; Zanos et al., 2016). No lethality was observed (Michael et al., 2025). As already stated in our manuscript, the 10-day CSDS protocol used in our studies was approved by the University of Maryland, Baltimore IACUC under AAALAC-compliant oversight (covering studies that included the 10-day CSDS exposure). We agree that precise animal welfare reporting strengthens scientific research and this should be included with appropriate details in all scientific publications, or in cases of established protocols, via proper reference to previous published work. It is important to note that alternative stress models were considered by our research team, such as the chronic unpredictable mild stress; however, these protocols were not considered suitable for the objectives of our studies since they do not reliably lead to sociability deficits and do not induce the full spectrum of depression-related endophenotypes in C57BL/6J mice (e.g., Strekalova et al., 2011). Indeed, sociability impairment represents a critical phenotype occurring in individuals abstinent from opioid use. Therefore, such an endophenotype in our preclinical research was deemed necessary to serve as a translational behavioural marker of social withdrawal, plausibly leading to heightened relapse during protracted opioid abstinence. In our manuscript we explain the selection of our paradigms by highlighting the need for understanding the interplay between stress susceptibility, sociability deficits and other depressive-like behaviours, such as anhedonia, and their influence on the propensity for opioid conditioning and/or relapse. Thereby, our intention and rationale for using CSDS was to model key vulnerabilities observed in human OUD. The behavioural assays used in our studies directly assessed disorder-related endophenotypes, not subjective states. Susceptibility was determined not only via social avoidance, but also through reduced sucrose preference to ensure our findings better replicated depression-related manifestations. Sucrose preference deficits (consummatory anhedonia) and female urine sniffing reductions (sexual anhedonia) have been previously used to complement each other for a more robust depression-like endophenotype in mice (Georgiou et al., 2022; Zanos et al., 2016). Importantly, ketamine, as the prototype rapidly-acting antidepressant with established efficacy in treatment-resistant depression in humans, reverses CSDS-induced sucrose preference deficits (e.g., Georgiou et al., 2022; Zanos et al., 2016) and female urine sniffing reductions (e.g., Georgiou et al., 2022) in susceptible mice. Therefore, we were able to exclude the possibility that olfactory confounds were the underlying cause of such behavioural changes. The opioid crisis leading to at least 300,000 deaths yearly, underscores our study's timeliness. Relapse rates of up to 91% persist, despite the availability of pharmacotherapies. Our findings position a novel medication (now in Phase II clinical trials in humans) as a repurposed option for these gaps, further justifying the necessity for our model's use and the translational validity of our findings. We would like to emphasize that we are in favour of a more thorough journal scrutiny for preclinical studies reporting stress-related animal protocols and we appreciate the letter from Stanford et al., for starting such an important discussion. Such an exchange will certainly facilitate the goals we share for the implementation of ethically- and transparently-conducted scientific research. Andria Michael: Writing—review and editing; writing—original draft. Anna Onisiforou: Writing—review and editing. Polymnia Georgiou: Writing—review and editing; writing—review and editing. Morfeas Koumas: Writing—review and editing; writing—original draft. Chris Powels: Writing—review and editing. Elmar Mammadov: Writing—review and editing. Andrea N. Georgiou: Writing—review and editing; writing—review and editing. Panos Zanos: Writing—review and editing; writing—original draft. CONFLICT OF INTEREST STATEMENT P.Z. is listed as co-inventor on patent application for the synthesis, polymorph forms, and use of (2R,6R)-hydroxynorketamine and (2S,6S)-hydroxynorketamine in the treatment of depression, anxiety, anhedonia, suicidal ideation, and post-traumatic stress disorders. P.Z. has assigned his patent rights to the University of Maryland, Baltimore, but will share a percentage of any royalties that may be received by the University of Maryland, Baltimore.