Enhanced Circadian Entrainment in Mice and Its Utility under Human Shiftwork Schedules

Round 1

Reviewer 1 Report

In “Enhanced circadian entrainment in mice and its utility under human shiftwork schedules”, Walbeek et al. test the effects of bifurcation and T cycle manipulation on entrainment following phase shift. The goal of the study is to determine whether adaptation to shift work in humans can be modulated by such behavioral modifications. The article follows up on previous work describing a beneficial reentrainment effect of bifurcation and dim light at night on a model of permanent shiftwork, and attempts to determine whether there might also be beneficial effects in models of more complex shift work, such as irregular schedules. The results presented include those of several studies, the first employing 24, 30, or 36 LDLD cycles with bifurcated mice prior to phase delay or advance. The second study addresses whether specific entrainment conditions with bifurcation assists in adapting to a rodent model of DuPont style work demands. Finally, the third experimental paradigm tests the effects of T-cycle entrainment on shift work adaptation.

The article is well written and the figures adequately present the data and experimental designs. I have only a couple of minor comments.

The article cites papers discussing the potential sex specificity of shift work adjustment in humans. Because of the apparent discrepancy (but also the degree to which females differ from males here), the study should discuss whether mechanisms for this have been studied in the rodent model, and if so, what has been determined regarding mechanisms of sex specificity. The effects of food intake on entrainment (or internal synchronization at least), while not the primary focus of the study, is completely ignored. Is there any evidence that ad libitum feeding poses a competing zeitgeber for any of these conditions? Could restricted feeding be expected to enhance entrainment for any of the models used? Some discussion of this should be included.

Author Response

In “Enhanced circadian entrainment in mice and its utility under human shiftwork schedules”, Walbeek et al. test the effects of bifurcation and T cycle manipulation on entrainment following phase shift. The goal of the study is to determine whether adaptation to shift work in humans can be modulated by such behavioral modifications. The article follows up on previous work describing a beneficial reentrainment effect of bifurcation and dim light at night on a model of permanent shiftwork, and attempts to determine whether there might also be beneficial effects in models of more complex shift work, such as irregular schedules. The results presented include those of several studies, the first employing 24, 30, or 36 LDLD cycles with bifurcated mice prior to phase delay or advance. The second study addresses whether specific entrainment conditions with bifurcation assists in adapting to a rodent model of DuPont style work demands. Finally, the third experimental paradigm tests the effects of T-cycle entrainment on shift work adaptation.

The article is well written and the figures adequately present the data and experimental designs. I have only a couple of minor comments.

Comment: The article cites papers discussing the potential sex specificity of shift work adjustment in humans. Because of the apparent discrepancy (but also the degree to which females differ from males here), the study should discuss whether mechanisms for this have been studied in the rodent model, and if so, what has been determined regarding mechanisms of sex specificity.

Response: We have not yet studied mechanisms for sex differences in our rodent models, but agree this is very interesting. We have updated to text to reflect this.

Comment: The effects of food intake on entrainment (or internal synchronization at least), while not the primary focus of the study, is completely ignored. Is there any evidence that ad libitum feeding poses a competing zeitgeber for any of these conditions? Could restricted feeding be expected to enhance entrainment for any of the models used? Some discussion of this should be included. 

Response: We have updated the text to acknowledge the potential influence of multiple zeitgebers on behavioral rhythms

Reviewer 2 Report

The authors address the important question of how light can be used to facilitate entrainment to different "work" schedules (DuPont and Continental) in mice.  They measure entrainment following a variety of treatments (e.g. 7h light:5h dark schedules (termed "Bifurcation"), dim light at night, and non-24h light cycles) and find that all treatments improve entrainment compared to a standard 12h light:12 h dark cycle.  The authors are applauded for their careful, quantitative assessment of entrainment using metrics such as BSI, EQ, and wavelet-based period estimation.  The data are compelling and support the authors' conclusion that the mouse circadian system "can be rendered sufficiently flexible to adapt to...changing shiftwork schedules." The results will be of interest to shift work researchers and industry representatives.   The manuscript will be improved when the authors address the following minor concerns:

The authors conclude that bifurcation and dim night lighting "minimized" circadian misalignment. Ideally, the paper would compare the entrainment rates to the different work schedules with each treatment to address the question: how much does each treatment speed entrainment?  Minimally, the authors should be clear that their treatments "reduced" (but did not eliminate or necessarily minimize) circadian misalignment.  

The wavelet plots are particularly creative and helpful in visualizing how mice responded to changing light cycles, but leave us just short of a clear conclusion.  Can the authors envision ways to summarize these plots to estimate "time needed to entrain" and/or "time spent entrained" for each mouse, condition and group?

The authors use appropriate, established and novel metrics for entrainment, but run the risk of losing the readers with too much jargon.  This reader could not easily discern details of the treatments, data analysis and presentation.  The authors have clearly worked hard in designing and describing their complex light cycles and findings, but more information is needed in the figures and results. Here are a few examples. 1) Please include the exact light:dark parameters (duration, relative intensity) in figure legends rather than challenging the readers to find the information in the methods, 2) label N1 and N2 and Phase 1 and Phase 2 in the appropriate actograms 3) label "entrained" or "free-running" lines on the wavelet plots.   

Please add error bars to the average results in Figure 2.

Figure 7 A-C appears to show something really cool, but I am not sure what!  Please flesh out the figure legend and consider complete labels (e.g. are 1-14 the individual mice? What is an "averaged activity profile for a weekly repeat" and are you showing 4 weeks of data for each mouse? Are 500 and 1000 in wheel revolutions/hour blocks?

The authors choose to expect us to remember Study 1, 2 and 3 as a simplification of the complex treatments and measurements done.  Unfortunately, I can not. 

An important strength of this work is the exploitation of light to entrain (e.g. over drugs, mutations, deep brain stimulation, etc.). The introduction nicely summarizes the work from the Gorman lab in this regard, but fails to cite work from others and in other model systems on treatments that enhance or impair circadian entrainment.  For example, the paper will be strengthened when they compare their findings to the long-day vs. short-day effects on entrainment by Dr. Johanna Meijer, flickering light effects by Dr. Rob Lucas, desynchronizing light pulse effects by Dr. Erik Herzog and Todd Holmes and light intensity and color effects by Dr. Ken Wright and others. 

Author Response

The authors address the important question of how light can be used to facilitate entrainment to different "work" schedules (DuPont and Continental) in mice.  They measure entrainment following a variety of treatments (e.g. 7h light:5h dark schedules (termed "Bifurcation"), dim light at night, and non-24h light cycles) and find that all treatments improve entrainment compared to a standard 12h light:12 h dark cycle.  The authors are applauded for their careful, quantitative assessment of entrainment using metrics such as BSI, EQ, and wavelet-based period estimation.  The data are compelling and support the authors' conclusion that the mouse circadian system "can be rendered sufficiently flexible to adapt to...changing shiftwork schedules." The results will be of interest to shift work researchers and industry representatives.   The manuscript will be improved when the authors address the following minor concerns:

The authors conclude that bifurcation and dim night lighting "minimized" circadian misalignment. Ideally, the paper would compare the entrainment rates to the different work schedules with each treatment to address the question: how much does each treatment speed entrainment?  Minimally, the authors should be clear that their treatments "reduced" (but did not eliminate or necessarily minimize) circadian misalignment.  

Response: We agree and have edited the text to be clearer that we are referring to reduced misalignment rather than attaining some absolute minimal level of misalignment. As discussed further below, we do not feel that we can fairly speak to "entrainment rates" in these constantly changing conditions. Rather, we can fairly characterize the overall match/mismatch aggregated over the entire schedule as reflected, for instance, in the amount of activity in the light.  On its own, that measure is a quite non-specific measure of entrainment, but we find great utility here because it captures an easy-to-understand parameter that can be transparently calculated over an extended interval of changing photoperiod conditions.  

The wavelet plots are particularly creative and helpful in visualizing how mice responded to changing light cycles, but leave us just short of a clear conclusion.  Can the authors envision ways to summarize these plots to estimate "time needed to entrain" and/or "time spent entrained" for each mouse, condition and group?

Response: We spent many hours thinking about whether we could assess these regimens in terms of "time to entrain." This has been a key measure in the literature and has great appeal. In the end, we determined that we could not because the cycles were changing too quickly and the animals were probably never in a steady state.  The wavelets, however, let us ask what the periods are particular moments.  Figure 6 gets at the issue of “time spent entrained” by reflecting period estimates during each week/weekend block separately. To facilitate our conclusion, we have added GLME test results comparing the number of animals from dim versus dark that fall within entrainment range during each phase of the cycle to the figure.

The authors use appropriate, established and novel metrics for entrainment, but run the risk of losing the readers with too much jargon.  This reader could not easily discern details of the treatments, data analysis and presentation.  The authors have clearly worked hard in designing and describing their complex light cycles and findings, but more information is needed in the figures and results. Here are a few examples.

a) Please include the exact light:dark parameters (duration, relative intensity) in figure legends rather than challenging the readers to find the information in the methods,

Response: We appreciate this comment urging us to clarify the presentation. We agree that we can improve clarity by updating the figure legends..

b) label N1 and N2 and Phase 1 and Phase 2 in the appropriate actograms

Response: We have updated figures and legends where appropriate.

c) label "entrained" or "free-running" lines on the wavelet plots. 

Response: We have updated the figure indicating entrainment and free-running as well as statistical results

Please add error bars to the average results in Figure 2.

Response: The figure does contain error bars! They are, however, so small that they are obscured by the data points in the formatted figure (The dots span about 40 minutes, while the SE is ~ 10 min). We have updated the figure legend to acknowledge this issue.

Figure 7 A-C appears to show something really cool, but I am not sure what!  Please flesh out the figure legend and consider complete labels (e.g. are 1-14 the individual mice? What is an "averaged activity profile for a weekly repeat" and are you showing 4 weeks of data for each mouse? Are 500 and 1000 in wheel revolutions/hour blocks?

Response: We have included an extra panel in the figure and updated the figure legend to guide the reader.

The authors choose to expect us to remember Study 1, 2 and 3 as a simplification of the complex treatments and measurements done.  Unfortunately, I can not. 

Response: We have updated the references to the studies to include a name for the study to remind the reader.

An important strength of this work is the exploitation of light to entrain (e.g. over drugs, mutations, deep brain stimulation, etc.). The introduction nicely summarizes the work from the Gorman lab in this regard, but fails to cite work from others and in other model systems on treatments that enhance or impair circadian entrainment.  For example, the paper will be strengthened when they compare their findings to the long-day vs. short-day effects on entrainment by Dr. Johanna Meijer, flickering light effects by Dr. Rob Lucas, desynchronizing light pulse effects by Dr. Erik Herzog and Todd Holmes and light intensity and color effects by Dr. Ken Wright and others. 

Response: We have updated both introduction and discussion to include reference to work from others on enhanced circadian resetting.

Round 2

Reviewer 2 Report

The authors addressed most of my prior concerns with additional methods, revised figures, and attempts to reduce the jargon.  The results are interesting and offer a creative set of lighting paradigms that could reduce jetlag in some real-world work settings.  The authors should ensure they keep results in past tense (e.g. legends for Figures 3 and 5 now report some findings in a confusing mix of present and past tenses). The paper would be stronger with more quantitation of how much jetlag is reduced (e.g. between the four tested lighting strategies); the authors have elected to not follow this path.