Sleep Deprivation Enhances Cocaine Conditioned Place Preference in an Orexin Receptor-Modulated Manner

Synthesis

The reviewers were positive about the study, but they made several pertinent comments that can be seen in their review below. Additional experiments are not required; all of the reviewers’ critiques can be addressed with text changes.

Author Response

The reviewers were positive about the study, but they made several pertinent comments that can be seen in their review below. Additional experiments are not required; all of the reviewers’ critiques can be addressed with text changes.

We thank the reviewers for their thoughtful critiques and helpful suggestions.

Reviewer 1

The authors set out to test whether sleep-deprivation had an effect on cocaine-induced CPP, and if so, the degree to which hypocretin/orexin modulation participates in sleep-deprived enhancement of cocaine-CPP. The question of the effects of sleep deprivation on cocaine consumption is an interesting one, and as the authors point out, has ramifications on treatment plans for stimulant substance use disorders. The hypotheses posed are likely of high interest to the addiction research field. Although a considerable amount a work went into the execution of these studies, they appear not entirely well-planned or designed, tempering this reviewer’s enthusiasm for the results. In particular, some of the study design choices (e.g., length of sleep-deprivation, length of conditioning for CPP, dose of hypocretin-antagonist tested), are not adequately justified or thought out. Similarly, the analyses of the data and interpretation are simple, yet confused. The authors need to re-work this manuscript with a richer analyses and interpretation of the data. Please consider the following comments:

1. The title does not correctly state their results given they did not manipulate the “dose” of sleep-deprivation, but rather they tested effects of sleep-deprivation on dose-dependent cocaine-induced CPP.

We modified the title to reflect the dose-dependence of cocaine not sleep deprivation duration. The new title is ‘Sleep deprivation enhances conditioned place preference in a cocaine dose-dependent manner’.

2. All text referring to “orexin antagonism” should be changed to “orexin-receptor antagonism.”

We apologize for the sloppy wording. We now use orexin-receptor (OX1R) antagonism.

3. The authors should be careful about relating cocaine-withdrawal-associated sleep literature to that of acute-cocaine-associated sleep literature (particularly in the introduction) and how it relates to their sleep-deprivation studies using acute cocaine exposure.

We agree and have added the sentence ‘Thus, chronic sleep disturbance emerges after more extensive exposure to cocaine, while limited cocaine induces a sleep deprivation plus recovery response.’

4. Limited cocaine exposure (2 of 4 days) in the establishment of CPP used in these studies is a weakness of the study. The authors may want to justify such a short phase of place conditioning.

There is no effect of the number of conditioning trials (1-4) on effect size for cocaine CPP (Bardo et al., 1995) and a 2 trial CPP is commonly used (for example, McClung et al., 2005; Graham et al., 2009; Schindler et al., 2012; Lui et al., 2016). We have added the statement ‘While conditioning protocols vary, two trials are expected to be sufficient to support the development of cocaine preference (McClung et al., 2005; Graham et al., 2009)’ as justification for the use of this design within the Cocaine Conditioned Place Preference section of the methods. Furthermore, since we expected sleep deprivation to enhance cocaine CPP such that sleep deprived animals would be more sensitive to the rewarding properties of cocaine (as in the case of the subthreshold dose of 3 mg/kg) lower cocaine exposure (in the case of the number of conditioning trials) would be a better test of this hypothesis. An expert in rodent behavior was consulted during the design of this project and is listed in the acknowledgements statement.

4. (continued). It appears in Figure 2 that they achieved cocaine-induced CPP only at the high dose of 15mg/kg. This is concerning for the interpretation of the sleep-deprivation data.

Preference was achieved by both groups at the 8 mg/kg dose as indicated by the asterisks.

5. CPP testing/pairing was conducted during light phase. Participation of hypocretin-related modulation of CPP may be affected by light/dark cycle. Could your hypocretin manipulations be compromised because of light-phase testing times (ZT4-7)?

As the reviewer points out, orexin (hypocretin) is modulated by behavioral state which itself is modulated by circadian phase with orexin levels higher during the dark (active) phase compared to the light (inactive) phase. However, orexin levels and orexin neuronal activity increase with sleep deprivation (Estabrooke et al., 2001; Yoshida et al., 2001) such that orexin would be expected to be higher in sleep deprived animals compared to controls, which is why we hypothesized that sleep deprivation may enhance cocaine CPP via the orexin system given the known role of orexin in motivated behavior. CPP is typically done during the light phase and previous experiments have shown that manipulations of the orexin system during the light phase can influence CPP (Sartor and Aston-Jones, 2012). We don’t believe that light phase testing compromised the orexin manipulations because OX1R antagonism was sufficient to prevent cocaine CPP acquisition and reduce the SD-induced increase in CPP acquisition and expression.

6. It is not clear why a one-tailed t-test is used instead of two-tailed for either the CPP results or the weight validation test? Couldn’t animals show aversion (decrease in preference compared to 0 or to control group) and therefore need a two-tailed test?

We used one-tailed T tests since we hypothesized that sleep deprivation would enhance cocaine CPP and OX1R antagonism would reduce feeding thereby reducing weight based on the literature. Specifically, sleep deprivation increases preference for the stimulant methylphenidate in humans (Roehrs et al., 1999) and induces preference to a low dose of amphetamine in mice (Berro et al., 2018). This information has been added to the introduction.

Regarding OX1R antagonism and weight, orexin influences food intake (Sakurai et al., 1998), while SB334867 administration reduces weight under regular chow (Ishii et al., 2005). In response to the comment below (#7) about validation of SB 334867 by weight loss as irrelevant, this experiment has been removed.

We did not test for significance of aversion since the animals were given cocaine immediately prior to entry into the CPP apparatus. Conditioned aversion to cocaine has been shown when cocaine exposure precedes chamber entry by 15 min, while preference is seen when cocaine was given immediately prior to entry or following a 5 min delay (Ettenberg et al., 1999).

7. The “validation of SB334867 dose by weight loss” is somewhat irrelevant given the complicated role of hypocretin peptides and feeding behavior (as opposed to energy homeostasis). This should be removed to the paper, not only as it is confusing and irrelevant to the study, but also in the fact that they did not even elicit the ‘confirmation’ of hypocretin-receptor antagonism resulting in a significant reduction of overall weight. (The authors would have more accurately tested this with feeding behavior following administration of SB334867.)

We agree that feeding is a more direct indicator of SB334867 efficacy and since the OX1R antagonist used in our experiments was not from the same manufacturer from which hydrolysis was noted by Lewin and colleagues (McElhinny et al., 2012) we removed this experiment.

8. Regarding Figure 1: the text states an increase in preference (compared to a hypothetical value of 0) at the 8mg/kg dose in the first CPP test (figure 1), but this does not appear to be indicated on the graph (only a sign difference BETWEEN the groups is shown). Was this increase in preference at 8mg/kg in both groups statistically significant?

The significant difference from 0 for the 8 mg/kg dose was provided in the table. We had tried to avoid duplicating statistical information by providing it only in text or in tables with group differences shown in the figures when present, but this was deemed confusing by both reviewers. We have added symbols for the within group difference from 0 comparison for all doses in the figures and have revised the table to be more complete at the cost of duplicating some information that is also provided within the text.

There was a significant increase in preference between groups at the 8 mg/kg dose, while animals sleep deprived prior to cocaine conditioning trials showed preference (as indicated by a significant difference from 0) at 3, 8, and 15 mg/kg doses. Animals sleep deprived prior to saline conditioning trials showed preference at 15 mg/kg with a failure to achieve preference at 8 mg/kg even though preference values were similar to those of the undisturbed animals in the CPP expression design, likely due to slightly higher within group variability. This variability appears to be driven by a single animal, though the animal is not an outlier. We added a sentence to the results section of Study 1, as follows: ‘Controls (noSD Coc, SD Sal) showed preference at 15 mg/kg but not at 8 mg/kg, possibly due to high variability driven by one animal (not identified as an outlier when using the ROUT method [GraphPad Prism]).’

We also added the following paragraph to the discussion section regarding the unexpected lack of preference to the 8 mg/kg dose in noSD Coc, SD Sal, as follows: ‘Most groups showed preference for the 8 mg/kg dose of cocaine, a dose in which preference is expected (Campbell et al., 2000); however, animals sleep deprived prior to saline conditioning trials (noSD Coc, SD Sal, Study 1) did not reach statistical significance for preference despite the majority of animals showing an increase in time spent on the cocaine-paired side from the pre to post-conditioning tests (8/12). This lack of preference is likely attributable to high variability in preference scores relative to the group average (126.4 +/- 83.6) and is driven by a single animal as can be seen in Figure 1B, though this animal does not qualify as an outlier. As can be seen from the raw data plots with the 8 mg/kg dose across studies, most animals show an increase in relative time in the cocaine-paired side from pre to post-conditioning, though not all animals do so. We cannot explain the source of the individual differences, but these are consistent with individual differences seen with locomotor sensitization to cocaine (Hooks et al., 1991; Allen et al., 2007) and cocaine self-administration (Glick et al., 1994; Griffin et al., 2007).’

Additionally, we modified the results section for Study 1 to better distinguish the general increase in relative time spent on the cocaine-paired side from pre to post-conditioning tests from preference itself and did the same for other studies for consistency. We now use ‘relative time values’ instead of ‘preference’ for pre- and post-conditioning test values for clarity.

8. (continued) Panel B (showing just 8mg/kg group) is superfluous information that can be relayed purely in text. It is also not clear why the 8mg/kg groups are singled from other doses.

We include the relative time of pre and post conditioning test plots (Panel B in figures 1-4) as a way to show raw data and chose the 8 mg/kg dose since it is common across all studies. While unnecessary since the statistical comparison information is provided, we do not think these plots are superfluous since they illustrate variability seen when using this behavioral task. We added the following statement to the methods section ‘For all studies, pre and post-conditioning relative time values are shown (1B, 2B, 3B, 4B) for the 8 mg/kg dose to illustrate variability between animals within each group alongside the general pattern of increased time in the cocaine-paired side.’

9. Regarding Cocaine CPP, Study 2: saline/saline (0mg/kg cocaine) controls should have been included here as the sleep deprivation test was done prior to the final test (which is different than the 0mg/kg testing group from study 1, so not a suitable control).

We agree that the 0 mg/kg control used in Study 1 is not a suitable control for Study 2. The purpose of the saline/saline only control of Study 1 was to determine if sleep deprivation was sufficient to induce preference on its own. An inducement of preference would be seen as a significant increase in time spent on the side in which saline was received following sleep deprivation (i.e. a significant increase from 0). It is unclear how a change in preference would be observed with sleep deprivation immediately prior to the post-conditioning test since both sides are associated with saline, i.e. why would sleep deprivation prior to the post-conditioning test increase time in one saline-paired side over the other? Furthermore, unless most sleep deprived animals spent more time in the same side of the box (for example the side paired with saline on days 1 and 3) an emergence of group preference would likely not be observable. We used an unbiased apparatus such that there was no difference in the average time spent on either side across pre-conditioning tests so there should be no intrinsically preferred side (as could be seen with biased apparatuses that feature one dark and one bright side). Animals that showed an innate preference for one side of the box during the pre-conditioning test were excluded from experiments. When using an unbiased design with an unbiased apparatus, sleep deprivation prior to the post-conditioning test following saline/saline conditioning would be challenging to interpret.

10. Regarding Figure 2: the text again states an increase in preference for the 8mg/kg groups but this is not reflected (if statistically significant) in the graph (Panel C).

There was an asterisk with a line between groups indicating a significant group difference. Based on feedback from reviewer #2, we now use different symbols to distinguish between comparisons and within group difference from 0 comparisons. Group differences are now labeled with a carrot symbol.

10 (continued). And again, Panel B (showing just 8mg/kg group) is superfluous information that can be relayed purely in text. Should be removed from the figure.

As stated above, we believe the raw data plots are useful for showing variability across animals alongside the general pattern of increased time in the cocaine-paired side post-conditioning.

11. It is not clear why two of three doses were chose for the Study 3 and only one dose was chose for study 4 for tests of hypocretin-receptor antagonism. Please explain.

We used OX1R antagonism for 3 and 8 mg/kg doses for Study 3 on the basis that sleep deprivation induced preference to the 3 mg/kg dose and increased preference for the 8 mg/kg dose as found in Study 1 (CPP acquisition design used in Studies 1 and 3). For Study 4, we used OX1R antagonism for the 8 mg/kg dose since sleep deprivation increased preference to this dose but did not include the 3 mg/kg dose since only a non-significant trend towards preference with the 3 mg/kg dose was found in Study 2 (CPP expression design used in Studies 2 and 4). We have added the following statements to the methods sections for Study 3 and 4: ‘The cocaine doses chosen were based on the doses in which there were significant group differences in Study 1 (3, 8 mg/kg).’ and ‘The cocaine dose chosen was based on the dose in which there was a significant group difference in Study 2 (8 mg/kg).’

12. Tables: The organization of the tables is utterly confusing and not acceptable. It would be helpful to match the order of the columns with the order of the bars in related graphs. The headings of the columns do not match the group labels used in legends of figures, and some are errantly labeled all together. All p-values within the table for t-test versus hypothetical preference value of 0 should be report (some are, and some are not). Each unpaired t-test p-value should be reported as table subscript (some are, some are not).

We thank the reviewer for pointing out the incompleteness and confusing nature of the tables. We remade the tables with more complete information resulting in some information repeated in the text and tables, but we believe this change greatly improves readability. We have also added additional statistical information including 95% confidence intervals for all comparisons, difference between means for between group comparisons, and the specific statistical comparisons used. The order of the groups in the tables now matches that of the corresponding figures.

13. The interpretation of data in Figures 3 and 4 of “Orexin antagonism prevented the SD-induced enhancement of both acquisition and expression.” makes no sense without appropriate vehicle controls.

All groups in Study 3 received both the OX1R antagonist and its vehicle control with two of the groups serving as controls (noSD Veh Coc, SD SB Sal and noSD SB Coc, noSD Veh Sal) for the experimental group (SD SB Coc, noSD Veh Sal), though an additional control of noSD Veh Coc, noSD SB Sal could have been added for completeness. We did not include any vehicle control groups in Study 4, which was also raised as a potential weakness by reviewer #2, instead using noSD SB as the control for SD SB. We agree that this is a potential limitation since the Veh could have an effect different than that of no injection (as used in Study 2) and at certain concentrations DMSO (used to dissolve the OX1R antagonist) can influence behavior; however, we diluted DMSO to reduce the concentration below that which behavioral effects are observed (Cavas et al., 2005). We have added a paragraph in the discussion section for potential limitations and include the lack of vehicle only controls as one limitation.

14. Figure 5- Panel B: The most direct test of effects of orexin-receptor antagonism on the enhanced effects of sleep deprivation on cocaine-induced cpp are a comparison between “noSD Coc, SD Sal” (gray) and “noSD Veh Coc, SD SB Sal” (green), and between “SD Coc, noSD Sal” (red) and “SD SB Coc, noSD Veh Sal” (Purple). These statistical analyses (two-way anova) should be carried out and added to Figure 5B (or made into a bar graph with only these four groups). The preference scores of controls, in this case gray and red, may not be significantly different/attenuated by treatment of SB of related treatment groups (green and purple, respectively). This is important information in the interpretation of the data and the role of hypocretin/orexin in sleep deprivation-enhanced cocaine-CPP acquisition. Similar analyses should be completed and interpreted for sleep-deprivation-derived enhancement of established cocaine-CPP.

We did not perform statistical comparisons across studies since there was a considerable time gap between data collected for Study 1 and Study 3 and Study 2 and Study 4 along with some differences in the actual boxes used for CPP conditioning as noted in the methods section. Within each study, control and experimental animals were littermates and data were collected concurrently across groups with multiple sets of control and experimental animals collected for each study; however, there was a considerable time lag between data collection of related studies. We acknowledge that a superior design would include concurrent data collection for related studies (1&3, 2&4) in order to allow direct comparisons instead of the piecemeal design used here. We have included this point in the discussion of potential limitations. We also softened the language within this paragraph by modifying ‘orexin-dependent’ to ‘orexin-influenced’ and ‘prevented’ to ‘reduced’ since these terms are more consistent with observational comparisons.

14 (continued). These results may suggest the need for further dose testing of the SB compound.

The dose of OX1R antagonism used was sufficient to reduce SD-induced enhancement of cocaine CPP. While a higher dose may be able to prevent the acquisition or expression of CPP following SD, as was seen in the absence of SD in Study 3, higher doses would also be expected to reduce locomotor activity (Rodgers et al., 2001) which would be a potential confound.

15. Figure 5- Panel D: What dose are these data derived from? It is also no clear which groups these data are derived from. Statistical analyses should be performed on these data.

Date were derived from Study 2 and 4 in which preference score for the experimental group was normalized to preference score for the control group for each study. We modified the figure legend (Study 2, Study 4) and Y axis label (8 mg/kg) to make the data source clearer. Also, we modified the description of this comparison to provide more details and added a statement to explicitly note that this and other graphs in figure 5 are observational comparisons not statistical comparisons. The end of this paragraph now reads ‘Unexpectedly, OX1R antagonism in undisturbed animals prior to the post-conditioning test [Study 4; noSD SB] led to preference values similar to that of sleep deprived animals in the absence of OX1R antagonism [Study 2; SD], thereby reversing the polarity of the effect of SD as determined by dividing the group average preference score of the sleep deprived group by the group average preference score of the undisturbed group (i.e. [Study 2, SD/noSD]; [Study 4, SD SB/noSD SB]). A score above 1 indicates that SD results in a higher relative preference score compared to the undisturbed condition, while a score below 1 indicates that SD results in a lower relative preference score compared to the undisturbed condition. Statistical comparisons across studies were not performed because of data collection constraints (see discussion section) so these comparisons are observational in nature and should be interpreted with caution.’ This addition also modifies a couple of terms used in the original manuscript that have been revised for accuracy (reversing instead of inverting and polarity instead of direction).

16. Can the authors provide a justification for the use of the 4 hours of sleep deprivation in their model?

We apologize for not including the rationale for the SD duration. Four hours of sleep deprivation was used since this duration reliably induces a homeostatic response as measured by an increase in slow wave activity (SWA, 0.5-4.5 Hz) during slow wave sleep (Bjorness et al., 2018b). Furthermore, this duration does not increase expression of glucocorticoid-related genes as determined by transcriptome analysis of cortical tissue (Bjorness et al., 2020). The Bjorness et al., 2020 citation is a BioRvix preprint; this citation will be updated if this paper is accepted for publication (currently conditionally accepted and under re-review) while the current manuscript is being considered.

17. The Discussion is a very cursory and superficial summary and interpretation of these data. This Discussion needs significant improvement.

We acknowledge that the discussion was quite short, but we tried to avoid speculating much beyond interpretation of the results of the current studies. We have added two new paragraphs based on the concerns raised in this reviewer’s earlier points and an additional paragraph (given below) that puts the findings of the current manuscript in context with related experiments and suggests future experiments that could shed additional light on the ability of SD to enhance the rewarding properties of drugs. ‘The SD-induced enhancement of cocaine CPP is consistent with previous studies in which SD increases preference of methylphenidate in humans (Roehrs et al., 1999) and induces preference to a low dose of amphetamine in rodents (Berro et al., 2018). However, there are several additional studies that would be of interest in further delineating the ability of sleep loss to influence reward. First, thus far all studies have used stimulants so the generalizability of the SD-induced enhancement of preference across drugs is unknown. Additionally, the time course of this enhancement preference is unclear. A long term enhancement of preference would likely be more relevant to the development of addiction than if the SD-induced enhancement is quickly lost. Finally, it is unknown if SD-induced enhancement of stimulant reward is sustained in drug experienced animals since existing studies have included drug naïve rodents or non-dependent humans.’

Reviewer 2

Sleep-mediated regulation of drug reward is increasingly recognized, while the underlying mechanisms are poorly understood. The orexin system represents an important hub that integrates arousal and reward/motivation regulation. This study reports that acute sleep deprivation enhances both the acquisition and expression of cocaine-induced place-preference, and examines the role of orexin receptor activity in regulating this effect. The experiments are thoughtfully designed and carefully conducted. The manuscript is carefully written and mostly clear and logical. There are a few concerns as follows, mostly minor ones.

Results:

1. There is no mentioning of whether changes in CPP may be associated with locomotor activity changes. This may be relevant given orexin’s role in arousal, especially after SD.

We do not believe that the change in CPP is related to locomotor activity changes since it has been previously shown that acute (4h) SD does not influence the magnitude of locomotor sensitization to cocaine (Bjorness and Greene, 2018b). There is no difference in chamber entries during the post-conditioning test with any of the doses for Studies 1, 3, 4 and for 15 mg/kg cocaine for Study 2; we do not have chamber entry information for 3 and 8 mg/kg for Study 2. We also do not have general activity measures outside of chamber entries, itself a relatively gross indicator of activity, for 3 and 8 mg/kg cocaine in Studies 1 and 2 so we cannot directly assess activity in the CPP chambers for the specific conditions in which there is a group difference. For Study 1 with 15 mg/kg cocaine, there was no difference in activity as determined by beam breaks, between animals sleep deprived immediately prior to the first cocaine conditioning trial compared to those that were undisturbed. While we do not believe that locomotor activity increases are associated with the SD-induced enhancement of CPP, we cannot exclude this possibility. We include the following statement in a newly added limitations paragraph: ‘Additionally, since activity measures are not available for conditionings performed with all of the CPP boxes, we cannot exclude a possibility that enhanced preference is associated with an increase in locomotor activity; however, it has been shown that acute SD does not influence the magnitude of locomotor sensitization to cocaine (Bjorness and Greene, 2018b) so an SD-dependent increase in locomotor activity is not expected.’

2. To examine cocaine x sleep interactions, two-way ANOVA is more appropriate than T tests. Cocaine x sleep x SB may need multi-factor ANOVA.

We acknowledge the possibility of falsely finding a significant effect due to the use of multiple t tests instead of using a two-way ANOVA with corrections for multiple comparisons. Many CPP experiments in which a two-way ANOVA is used include phase as a factor such that both pre-conditioning and post-conditioning values are present. This makes it easier to see a significant effect since preference scores should be very low during the pre-conditioning phase. We chose to subtract pre-conditioning values from post-conditioning values to ensure that any increase in preference cannot be attributed to slight preferences in the cocaine-paired chamber prior to drug exposure. We also used an unbiased design such that some animals received cocaine in their initially more preferred compartment which would also serve to limit the possible increase in preference that occurs following cocaine as is apparent in the plots of relative time from pre to post-conditioning (Figures 1B, 2B, 3B, 4B). We have added 95% confidence intervals and difference of means information to the revised tables for each of the statistical comparisons to provide more robust information than p values alone.

There is also no explanation why some used one-tail and others used two-tail T tests (Tables 1-3). ?

We used one-tailed T tests for 3 mg/kg, 8 mg/kg, and 15 mg/kg cocaine with Study 1 and 2 and with weight (now removed based on the suggestion of reviewer #1) since we hypothesized that sleep deprivation would enhance cocaine CPP and orexin-receptor antagonism (OX1R) would reduce feeding thereby reducing weight based on the literature. Conversely, we used two-tailed T tests for comparisons in which there is not direct literature support of the hypothesis (as in the case of the ability of OX1R to influence SD-mediated increases in cocaine CPP). We added information pertaining to the literature support of our hypothesis that SD would enhance cocaine CPP to the introduction, as follows: ‘SD also influences reward in that SD increases preference for the stimulant methylphenidate in humans (Roehrs et al., 1999) and induces preference to a low dose of amphetamine in mice (Berro et al., 2018).’ We also revised the statistical comparisons information in the methods section to clarify statistics used in each Study and provide the rationale for the use of one and two tailed T tests.

Minor concerns:

3. Figure 1: significant differences for within-group and between-group comparisons are both labeled with “*”. This appears confusing and can be changed to use different symbols and be consistent with the other figures (e.g. Figure 2).

We appreciate the reviewer pointing out this confusing oversight and have changed the between-group comparison significant difference symbol to a carrot.

4. Figure 4: There is no control group “Veh noSD” to compare for SB effect alone (on cocaine CPP), in the absence of SD. The authors have cited published studies for this, but it is not clear if the experimental conditions match. If the authors intend to examine an orexin x sleep “interaction", then the design needs to be more complete. Otherwise, the phrasing needs to be revised.

This point was also raised by reviewer #1 and we have added a limitations paragraph to the discussion section in which we discuss the lack of a vehicle-only control group, as follows: ‘First, Study 4 lacked a vehicle control group; a control group of noSD SB was used for the experimental group SD SB in which the ability of OX1R-antagonism to counter SD-induced enhancement of cocaine CPP to an 8 mg/kg dose of cocaine (from Study 2) was tested. DMSO was diluted in order to reduce the concentration below that which behavioral effects are observed (Cavas et al., 2005). However, the effect of vehicle alone on CPP expression was not determined so the possibility that SD-induced enhancement of cocaine CPP was reduced by the vehicle cannot be excluded.’ We also rephrased the introductory statement of Study 4 within the results section from ‘however an interaction with SD’ to ‘however OX1R antagonism following SD’ to more accurately describe the comparison that was conducted.

5. Figure 5: There is no show of statistical significance in this figure.

We do not include statistical significance within the dose-response figure since we did not do statistical comparisons across studies. We added a statement within the results section for this section to explicitly note that the graphs in figure 5 are not statistical comparisons but are instead observational in nature and that they should be interpreted with caution. We also discuss the lack of between study comparisons as a potential limitation within the newly added ‘limitations’ paragraph.

Conclusions (minor):

6. The authors pointed out that “Notably, the relative preference of <1 under orexin antagonism indicates that orexin antagonism in the presence of SD reduces relative preference, while orexin antagonism in the absence of SD increases relative preference.” Therefore, the authors concluded that “Unexpectedly, orexin antagonism..., inverting the direction of the effect of SD.” However, results in Figure 4C shows that in the presence of SB (orexin antagonism), there is no significant difference between SD and non-SD groups, therefore does not support this conclusion. Although Figure 5D shows that SD/noSD ratios are >1 and <1 respectively, there is no statistics to validate this conclusion.

We agree with the lack of statistics and have added a statement that this and other graphs in figure 5 are observational in nature, not based on statistical comparisons such that comparisons should be interpreted with caution. While we did not directly compare outcomes across studies due to data collection constraints, we wanted to visualize that not only did OX1R antagonism reduce the SD-induced enhancement of CPP (as shown in Figure 4C), but also that the group average preference values were non-significantly lower compared to the control group and show this by dividing the preference score of the experimental group by that of the control group. When normalizing the experimental group preference value to that of the control group, polarity is reversed between Study 2 (> 1) and Study 4 (<1). We modified the result description to more explicitly state the transformation being used after which the statement regarding the lack of statistical comparisons is given.

We address the lack of direct comparisons between related studies (Study 1 & 3 and Study 2 & 4 in figure 5) as a potential limitation within the newly added limitations paragraph.

7. The authors concluded that “Orexin antagonism prevented the SD-induced enhancement of both acquisition and expression.” However, in Figure 3C (CPP acquisition), “SD SB Coc” (purple) shows preference to Coc-paired side, whereas “noSD SB Coc” (blue) shows no preference to Coc-paired side, thus it appears that the SD-induced enhancement may be reduced rather than entirely prevented by orexin antagonism.

We agree and revised the results statement to reflect this interpretation as follows: ‘Finally, SB 334867 prevented preference in the absence of SD and blocked SD-induced enhancement of preference to a 8 mg/kg dose of cocaine; however SD animals treated with SB 334867 did show preference for the cocaine-paired side suggesting that SD-dependent enhancement is reduced but not entirely prevented (Figure 3C; Table 3).’ We also changed ‘prevented’ to ‘reduced’ within the section for figure 5 (summary graphs) and discussion section.

Possible typo related to results (minor):

8. “15 mg/kg dose of cocaine (Figure 2C; Table 2), but the sleep deprived group showed a non-significant trend towards preference for the cocaine-paired side.” Likely should be “towards reduced preference for ...”

For the 15mg/kg cocaine dose, the SD group did not achieve preference based on the statistical analysis, but rather approached preference based on the p value (p = 0.1). We heavily revised the tables to provide fuller statistical information, including details that are given elsewhere. We believe that these new tables will make the within and between group comparisons easier to follow.

Citations added to the manuscript:

Allen RM, Everett CV, Nelson AM, Gulley JM, Zahniser NR (2007) Low and high locomotor responsiveness to cocaine predicts intravenous cocaine conditioned place preference in male Sprague-Dawley rats. Pharmacol Biochem Behav 86:37-44.

Bjorness TE, Greene RW (2018b) Sleep deprivation alters the time course but not magnitude of locomotor sensitization to cocaine. Sci Rep 8:17672.

Bjorness TE, Kulkarni A, Rybalchenko V, Suzuki S, Bridges C, Harrington AJ, Cowan CW, Takahashi JS, Konopka G, Greene RW (2020) An essential role for MEF2C in the cortical response to loss of sleep. bioRvix.

Cavas M, Beltran D, Navarro JF (2005) Behavioural effects of dimethyl sulfoxide (DMSO): changes in sleep architecture in rats. Toxicol Lett 157:221-232.

Glick SD, Raucci J, Wang S, Keller RW, Jr., Carlson JN (1994) Neurochemical predisposition to self-administer cocaine in rats: individual differences in dopamine and its metabolites. Brain Res 653:148-154.

Graham DL, Krishnan V, Larson EB, Graham A, Edwards S, Bachtell RK, Simmons D, Gent LM, Berton O, Bolanos CA, DiLeone RJ, Parada LF, Nestler EJ, Self DW (2009) Tropomyosin-related kinase B in the mesolimbic dopamine system: region-specific effects on cocaine reward. Biol Psychiatry 65:696-701.

Hooks MS, Jones GH, Smith AD, Neill DB, Justice JB, Jr. (1991) Individual differences in locomotor activity and sensitization. Pharmacol Biochem Behav 38:467-470.

McClung CA, Sidiropoulou K, Vitaterna M, Takahashi JS, White FJ, Cooper DC, Nestler EJ (2005) Regulation of dopaminergic transmission and cocaine reward by the Clock gene. Proc Natl Acad Sci U S A 102:9377-9381.

Roehrs T, Papineau K, Rosenthal L, Roth T (1999) Sleepiness and the reinforcing and subjective effects of methylphenidate. Exp Clin Psychopharmacol 7:145-150.

Citations used in the response but not added to the manuscript:

Bardo MT, Rowlett JK, Harris MJ (1995) Conditioned place preference using opiate and stimulant drugs: a meta-analysis. Neurosci Biobehav Rev 19:39-51.

Ettenberg A, Raven MA, Danluck DA, Necessary BD (1999) Evidence for opponent-process actions of intravenous cocaine. Pharmacol Biochem Behav 64:507-512.

Griffin WC, 3rd, Randall PK, Middaugh LD (2007) Intravenous cocaine self-administration: individual differences in male and female C57BL/6J mice. Pharmacol Biochem Behav 87:267-279.

Liu X, Chen Y, Tong J, Reynolds AM, Proudfoot SC, Qi J, Penzes P, Lu Y, Liu QS (2016) Epac Signaling Is Required for Cocaine-Induced Change in AMPA Receptor Subunit Composition in the Ventral Tegmental Area. J Neurosci 36:4802-4815.

Rodgers RJ, Halford JC, Nunes de Souza RL, Canto de Souza AL, Piper DC, Arch JR, Upton N, Porter RA, Johns A, Blundell JE (2001) SB-334867, a selective orexin-1 receptor antagonist, enhances behavioural satiety and blocks the hyperphagic effect of orexin-A in rats. Eur J Neurosci 13:1444-1452.

Schindler AG, Messinger DI, Smith JS, Shankar H, Gustin RM, Schattauer SS, Lemos JC, Chavkin NW, Hagan CE, Neumaier JF, Chavkin C (2012) Stress produces aversion and potentiates cocaine reward by releasing endogenous dynorphins in the ventral striatum to locally stimulate serotonin reuptake. J Neurosci 32:17582-17596.